Proceedings of the 27th Annual International Pittsburgh Coal Conference

TWENTY - SEVENTH ANNUAL INTERNATIONAL

PITTSBURGH COAL CONFERENCE FINAL TECHNICAL PROGRAM COAL - ENERGY, ENVIRONMENT AND SUSTAINABLE DEVELOPMENT October 11 - 14, 2010 Hilton Istanbul Istanbul, TURKEY Co-Sponsors and Organizers

University of Pittsburgh Swanson School of Engineering

Istanbul Technical University

Sponsors

gold b

Turkish Coal Enterprises

2

GENERAL INFORMATION

WELCOME! On behalf of the Conference Advisory Board, Conference Committees, and the University of Pittsburgh we welcome you to the Twenty-Seventh Annual International Pittsburgh Coal Conference held October 11 - 14, 2010 at the Hilton Istanbul in Istanbul, Turkey. The Conference is hosted by the University of Pittsburgh. The theme of this year’s conference is “Coal - Energy, Environment and Sustainable Development” which covers a wide spectrum of important topics on coal technology, synfuel and environmental issues. The topics cover energy and environmental issues and technologies related to coal and its byproducts. Over 300 technical papers and posters will be presented throughout the conference. The Poster Sessions will be held on Tuesday, October 12 from 18:20 - 20:20. For detailed information on technical sessions, papers and speakers, please turn to page 6 in the Technical Program. The invited Plenary Speakers include: Charles Taylor, Director, Chemistry and Surface Science Division, U.S. Department of Energy, National Energy Technology Laboratory, USA; Carlos A. Cabrera, President and CEO, National Institute of Clean and Low Carbon Energy (NICE), CHINA; Selahaddin Anaç, General Director, Turkish Coal Enterprise (TKI), TURKEY; Richard Winschel, Director, Research Services, CONSOL Energy Inc., USA; John Topper, Managing Director, IEA Clean Coal Centre, UNITED KINGDOM; Ekrem Ekinci, Rector, Isık Universty, TURKEY; John Wheeldon, Manager of Science and Technology, National Carbon Capture Center, EPRI, USA; Volkan Ediger, İzmir Technology Institute, TURKEY; and Robert Beck, Executive Vice President and Chief Operating Officer, The National Coal Council, Inc., USA. We express our sincere gratitude to the contributors for their support and involvement, to all the authors and co-authors of the technical papers and to all the members of the Program Committee, Awards Committee, International Committee and Membership Committee. Special thanks go to our Turkish host and coordinator Dr. Güven ÖNAL, Technical Program Chairs, Evan Granite of NETL-DOE, USA, and Jim Hower of the University of Kentucky, CAER, USA as well as to all session chairs, speakers and international delegates for their contributions to the 2010 technical program. As the chair and vice chair of the Advisory Board of the Conference, we deeply appreciate your participation and interest in this year’s Conference and we invite you to join us next year for the Twenty-Eighth Annual International Pittsburgh Coal Conference, which will be held in Pittsburgh, PA, USA at the David L. Lawrence Convention Center. Sincerely,

Robert Beck, Chair The National Coal Council, Inc., Washington, DC, USA

Richard Winschel, Vice -Chair CONSOL Energy Inc., Pittsburgh, PA, USA

CONFERENCE REGISTRATION On-Site Registration will begin Monday, October 11, from 18:00 - 20:00 and continues Monday, Tuesday, and Wednesday from 08:00 until 17:00. Please check in even if you have Pre-Registered!

Tours and reservations for the 2010 Pittsburgh Coal Conference are provided by: Aktuel Tourism Company Contact Person: Süleyman Göncü Ortaklar Cad. Unal Apt. No.7/Daire. 8 Mecidiyeköy/Istanbul/Turkey Tel. : +90(212)273. 06 90 Fax : +90(212)213. 63 95 E-mail: [email protected] http://www.aktueltours.com/

The International Pittsburgh Coal Conference EXECUTIVE DIRECTOR: Dr. Badie I. Morsi CONFERENCE ORGANIZER: Ms. Heidi M. Aufdenkamp University of Pittsburgh Swanson School of Engineering 1249 Benedum Hall Pittsburgh, PA 15261 USA Tel: +1-412-624-7440 FAX: +1-412-624-1480 Email: [email protected] www.engr.pitt.edu/pcc

PITT AWARD The Award for Innovation in Coal Conversion was founded by the Chemical and Petroleum Engineering Department, University of Pittsburgh in 1983 with industrial support. Since 1992, it has been fully funded by CONSOL Energy Inc.

3

GENERAL INFORMATION PLENARY SPEAKERS CONFERENCE OVERVIEW MONDAY, OCTOBER 11, 2010 Technical Tour Registration Reception

07:00 - 22:30 18:00 - 20:00 19:30 - 21:30

TUESDAY, OCTOBER 12, 2010 Registration Opening Ceremony Plenary Session – 1 Concurrent Tech. Sessions Conference Luncheon Concurrent Tech. Sessions Poster Session Dinner

08:00 - 17:00 09:00 - 09:30 09:30 - 11:15 11:30 - 13:10 13:10 - 14:25 14:25 - 18:20 18:20 - 20:20 20:20 - 22:20

WEDNESDAY, OCTOBER 13, 2010 Registration Plenary Session – 2 Concurrent Tech. Sessions Conference Luncheon Concurrent Tech. Sessions Gala Dinner

08:00 - 17:00 09:00 - 10:45 11:00 - 12:40 12:40 - 13:55 13:55 - 17:50 20:30 - 22:30

THURSDAY, OCTOBER 14, 2010 Registration Plenary Session – 3 Concurrent Tech. Sessions Awards Luncheon Concurrent Tech. Sessions Advisory Board Meeting

08:00 - 17:00 09:00 - 10:45 11:00 - 12:40 12:40 - 13:55 13:55 - 17:50 18:00 - 21:00

TUESDAY, OCTOBER 12, 2010 Energy Production/Policy Speakers Charles Taylor Director, Chemistry and Surface Science Division U.S. DOE/NETL, USA “U.S. Department of Energy’s Carbon Capture and Storage Efforts and Results”

Carlos A. Cabrera President and CEO National Institute of Clean and Low Carbon Energy (NICE), CHINA “A Refined Approach to Coal”

Selahaddin Anaç General Director Turkish Coal Enterprise (TKI), TURKEY “Coal in Turkish Energy Policy and Clean Coal Technologies”

WEDNESDAY, OCTOBER 13, 2010 International Issues Richard Winschel Director, Research Services CONSOL Energy Inc., USA “FutureGen – the World’s First Near-Zero Emission Coal-Based Power Plant”

John Topper Managing Director IEA Clean Coal Centre, UNITED KINGDOM “Sustainable Low Emissions Coal for our Grandchildren?”

Ekrem Ekinci Rector Isık Universty, TURKEY “Realities and Constraints of Coal in Energy and Environmental Perspectives”

THURSDAY, OCTOBER 14, 2010 Environmental Issues John Wheeldon Manager of Science and Technology National Carbon Capture Center, EPRI, USA “CO2 Capture and Storage for Coal-Based Power Generation”

Volkan Ediger İzmir University of Economics, TURKEY “Geopolitics of Coal and Global Climate”

Robert Beck Executive Vice President and Chief Operating Office The National Coal Council, Inc., USA “Low Carbon Coal: 21st Century Technologies and Policies”

4

TECHNICAL PROGRAM SCHEDULE

5

LOCATION GUIDE Hilton Istanbul Convention Center

ROOM DIRECTORY Registration Hyde Park Monday Evening Reception Convention Center Opening Ceremony Convention Center Plenary Sessions Convention Center Conference Luncheons Convention Center Poster Presentations Convention Center Exhibits Convention Center A/V & Speaker Preparation Room Orman Park

SESSION MEETING ROOMS Convention Center Sessions 1, 7, 13, 19, 25, 31, 37, 43, 49 Convention Center Sessions 2, 8, 14, 20, 26, 32, 38, 44, 50 Convention Center Sessions 3, 9, 15, 21, 27, 33, 39, 45, 51 Saturn Sessions 4, 10, 16, 22, 28, 34, 40, 46, 52 Lalezar Sessions 5, 11, 17, 23, 29, 35, 41, 47, 53 Jupiter Sessions 6, 12, 18, 24, 30, 36, 42, 48, 54

6

TECHNICAL PROGRAM ORAL SESSIONS

Tuesday, October 12, 2010 11:30 - 18:20 SESSION 1 COMBUSTION: OXY-COAL DEVELOPMENT – 1 John Wheeldon and İskender Gokalp Development of OxycoalTM Technology Resulting from Testing Conducted at Doosan Power Systems’ Clean Combustion Test Facility (CCTF), Peter Holland-Lloyd, David Fitzgerald, Doosan Power Systems, UNITED KINGDOM Co-Firing of Coal and Wood Biomass in Oxy-Fuel Combustion, Seong Yool Ahn, Jae woo An, Yon Mo Sung, Cheor Eon Moon, Gyung Min Choi, Duck Jool Kim, Pusan National University, SOUTH KOREA NO x Reburning in Oxy-Fuel Combustion - An Experimental Investigation, Daniel Kühnemuth, Fredrik Normann, Klas Andersson, Filip Johnsson, Bo Leckner, Chalmers University of Technology, SWEDEN Oxy-Combustion of Pulverized Coal: Modeling of Char-Combustion Kinetics, M. Geier, C. R. Shaddix, Sandia National Labs, USA; B. S. Haynes, University of Sydney, AUSTRALIA

SESSION 3 CARBON MANAGEMENT: GHG MANAGEMENT STRATEGIES AND ECONOMICS – 1 Leslie Ruppert and Ender Okandan

Design and Operational Strategies for IGCC with CO2 Capture, Chris Higman, Higman Consulting GmbH, GERMANY; George Booras, Electric Power Research Institute; Dan Kubek, Gas Processing Solutions LLC; Jim Sorensen, Sorensenergy LLC; Doug Todd, Process Power Plants, LLC, USA

The Creation of Georeactor Global Scientific Network, Jan Rogut, Jozef Dubinski, Aleksandra Tokarz, GIG, Central Mining Institute, POLAND; Marc Steen, Institute for Energy, Joint Research Centre; Hans Bruining, Delft University of Technology, THE NETHERLANDS; Hema J. Siriwardane, West Virginia University; Tomasz Wiltowski, Southern Illinois University; Elizabeth Burton, Lawrence Livermore National Laboratory; Subhas K. Sikdar, US EPA, USA; Thomas Kempka, German Research Centre for Geosciences (GFZ), GERMANY; Sohei Shimada, University of Tokyo, JAPAN

Envisioning CO2 Distribution Networks for Carbon Capture and Storage (CCS) in the United States: Strategies for CO2 Pipeline Deployment at a Regional Scale, Nils Johnson, Joan Ogden, University of California, Davis, USA

SESSION 6 COAL-DERIVED PRODUCTS: CHEMICALS/MATERIALS Belma Demirel and Bekir Zühtü Uysal

Processes to Produce Value Added Products from CO2, Belma Demirel, Deena Ferdous, Rentech, Inc., USA

Linking Economic and Technological Modeling of CCS and Legislative Policy for Coal Mining Companies, Pratt Rogers, Sean Dessureault, University of Arizona, USA SESSION 4 COAL SCIENCE: COAL CHEMISTRY – 1 Ashok K. Singh and Karol Koster

Study on Thermodynamic Calculation for O2/CO2 Flue Gases Recycled Combustion Boiler, Li-Qi Zhang, Can-Zhi Li, Fang Huang, Ji-Hua Qiu, Chu-Guang Zheng, Huazhong University of Science and Technology, CHINA

Carbonaceous Emissions Reflected in Deposits on Building Stones: Case Study in Prague Castle, Ivana Sýkorová, Martina Havelcová, Hana Trejtnarová, Petra Matysová, Alexandr Šulc, Institute of Rock Structure and Mechanics AS CR, v.v.i.; Antonín Zeman, Institute of Theoretical and Applied Mechanics, AS CR, v.v.i., CZECH REPUBLIC

SESSION 2 GASIFICATION: GENERAL SESSION – 1 Ke Liu and Hasancan Okutan

Visualization of Coal Conversion Using X-Ray Computed Tomography, QP Campbell, HWJP Neomagus, North-West University, SOUTH AFRICA

Coal: Biomass Gasification - A Pathway for New Technology Development of Oxygen Blown Co-fired Gasification with Integrated Electrolysis, Tana Levi, R. Witney, Y Iwasaki, Tony Clemens, CRL Energy Ltd.; S Pang, Q Xu, University of Canterbury; AI Gardiner, Industrial Research Limited, NEW ZEALAND Hydrogen Generation from Water by Using Plasma, Beycan İbrahimoğlu, İbrahim İbrahimoğlu, Anadolu Plazma Teknoloji Merkezi; Fırat Şen, Vestel Defence Industry R&D Department; Şahika Yürek, Türkiye Kömür İşletmeleri; Orhan Demirel, Turkish Coal Enterprises (TKI), TURKEY Technology and Operational Experience – The Shell Perspective, Jay Wang, Shell Global Solutions International BV, THE NETHERLANDS Controlling the Synthesis Gas Composition from Catalytic Gasification of Hypercoal and Coal by Changing the Gasification Parameters, Atul Sharma, Toshimasa Takanohashi, National Institute of Advanced Industrial Science and Technology, JAPAN A Technico-Economical Feasibility Study of Plasma Assisted Coal Gasification Compared to a Reference Auto-Thermal Gasifiction Process, Nazim Merlo, Iskender Gökalp, ICARE-CNRS, FRANCE

Reducing Greenhouse Gas Emissions from Coal Combustion by Adding Micro-Algal Biomass, Jaco Brink, Sanette Marx, North-West University, SOUTH AFRICA

Organic Sulphur Functionality Changes in Biotreated Coals, Lenia Gonsalvesh, Stefan Marinov, Maya Stefanova, Institute of Organic Chemistry, Bulgarian Academy of Sciences, BULGARIA; Robert Carleer, Jan Yperman, Hasselt University, BELGIUM Surface Coal Mine Planning Against Large Landslides, Celal Karpuz, Levent Tutluoglu, Arman Kocal, Middle East Technical University; Mustafa Ozdingis, Kıvanc Het, Turkish Coal Enterprises (TKI), TURKEY SESSION 5 SUSTAINABILITY AND ENVIRONMENT – 1 Brenda Pierce and Fevzi İsbilir A Solution to Water Crisis in Energy Production: Feasibility of Using Impaired Waters for Coal-Fired Power Plant Cooling, Radisav D. Vidic, Heng Li, ShihHsiang Chien, Jason D. Monnell, University of Pittsburgh; David Dzombak, Ming-Kai Hsieh, Carnegie Mellon University, USA Leaching Characteristics of Waste from PF Utilities and Transitional Technologies using Australian Coal, D.H. French, K.W. Riley, CSIRO Energy Technology; C.R. Ward, L.G. Stephenson, L.W. Gurba, University of New South Wales, AUSTRALIA Borovac Coal Cleaning Process, Branislav Grbovic, Borovac International Pty Ltd, AUSTRALIA; Miloljub Grbovic, Borovac International Pty Ltd; Jelenko Micic, Mining Basin Kolubara d.o.o.; Miroslav Spasojevic, “Nikola Tesla” Power Plants, SERBIA

Converting Brown Coal into Chemicals and Hydrogen by Steam Cracking and Gasification, Nozomu Sonoyama, Idemitsu Kosan Co., Ltd; Kazunari Nobuta, Tokuji Kimura, Sou Hosokai, Teruoki Tago, Takao Masuda, Hokkaido University; Jun-ichiro Hayashi, Kyusyu University, JAPAN Pilot Scale Production of Humic Substances from Turkish Leonardites, Bekir Zühtü Uysal, Ufuk Gündüz Zafer, Ö. Murat Dogan, Duygu Öztan, Gazi University; Zeki Olgun, Mustafa Ozdingis, Selahaddin Anac, Turkish Coal Enterprises (TKI), TURKEY Heavy Metal Removal by Using Chemically Crosslinked Turkish Coal Based Humic Acid, Tulay Inan, Hacer Dogan, Murat Koral, TUBITAK Marmara Research Center; Selahattin Anaç, Zeki Olgun, TKI(Turkish Coal Enterprises), TURKEY Analysis of the Effect of Internal Defect on Coke Fracture Behavior by Rigid Bodies-Spring Model, Kenichi Hiraki, Hayashizaki Hideyuki, Yoshiaki Yamazaki, Tetsuya Kanai, Xiaoqing Zhang, Masakazu Shoji, Hideyuki Aoki, Takatoshi Miura, TOHOKU University, JAPAN SESSION 7 COMBUSTION: CHEMICAL LOOPING DEVELOPMENT – 1 John Wheeldon and Mehmet Tombul Ilmenite as an Oxygen Carrier in a Chemical Looping Combustion System: Reaction Kinetics and Fluidized Bed Performance, Christopher K. Clayton, Gabor Konya, Edward M. Eyring, Kevin J. Whitty, The University of Utah, USA Application of Inorganic Remains Originating from Water Purification and Sewage Sludge Ashes in Chemical Looping Combustion Process, Ewelina Ksepko, Grzegorz Łabojko, Marek Sciazko, Institute for Chemical Processing of Coal, POLAND Chemical Looping with Oxygen Uncoupling: Design Calculations and Process Engineering Simulations Using Kinetic Data, JoAnn S. Lighty, Adel F. Sarofim, Asad H. Sahir, Edward Eyring, Gabor Konya, University of Utah, USA Chemical Looping Combustion and Gasification – A Novel Technique to Produce Concentrated Stream of Hydrogen and Carbon Dioxide from Victorian Lignites, Chiranjib Saha, Ali Akhavan, Sankar Bhattacharya, Monash University, AUSTRALIA

7

TECHNICAL PROGRAM SESSION 8 GASIFICATION: UNDERGROUND COAL GASIFICATION – 1 Rohan Courtney and Ömer Sezgin Bloodwood Creek UCG Pilot 2008 – 2010, Cliff Mallet, Carbon Energy Pty. Ltd., AUSTRALIA; Burl E. Davis, Carbon Energy Pty Ltd, USA Studies on Gasification of Turkish Lignite via Underground Coal Gasification, Şahika Yürek, Kıvanç Het, Directorate of Turkish Coal Enterprises (TKİ), TURKEY Underground Coal Gasification and Applicability to Thrace Basin Lignite in Turkey, Ayşe Yildirim, Serdar Dogan, Turkish Petroleum Company, TURKEY SESSION 9 CARBON MANAGEMENT: GHG MANAGEMENT STRATEGIES AND ECONOMICS – 2 Leslie Ruppert and Bahtiyar Unver CO2-Reduction through Biomass Co-Firing in Coal Fired Power Plants, Klaus-Dieter Tigges, Roland Jeschke, Alfred Gwosdz, Alfons Leisse, Hitachi Power Europe GmbH, GERMANY Ventilation Air Methane Abatement at CONSOL Energy’s Enlow Fork Mine, Richard A. Winschel, Deborah A. Kosmack, William P. Fertall, CONSOL Energy Inc.; Jerry Gureghian, Green Holdings Corp., USA Novel Methods of Coal Seam Gas Content Determination for Estimation of Greenhouse Gas Emissions from Mining, Abouna Saghafi, CSIRO Energy Technology, AUSTRALIA Swelling of Moist Coal in Carbon Dioxide and Methane, Richard Sakurovs, Robyn Fry, Stuart Day, CSIRO Energy Technology, AUSTRALIA Evaluation of Total Porosity and the Amount of Inaccessible Pores in Coal Using Small-Angle Neutron Scattering, Yuri B. Melnichenko, L. He, Oak Ridge National Laboratory; M. Mastalerz, Indiana University, USA; R. Sakurovs, CSIRO Energy Technology; T. Blach, Griffith University, AUSTRALIA SESSION 10 COAL SCIENCE: COAL FIRES Susan J. Tewalt and Levent Ergun Early Stage Detection of Coal Spontaneous Combustion in View of Pretreatment of the Coal, Boleslav Taraba, Zdenek Pavelek, Jiri Janek, Ostrava University, CZECH REPUBLIC SEM Study of Some Indian Natural Cokes (Jhama), Ashok K. Singh, Nandita Choudhury, Central Institute of Mining & Fuel Research, CSIR; Mamta Sharma, National Metallurgical Laboratory, CSIR; Mahendra P. Singh, Banaras Hindu University, INDIA Examination of Low Temperature Air Oxidation Mechanism of Brown Coal for Supressing Self Ignition Tendency, Kouichi Miura, Ryota Okajima, Mitsunori Makino, Ryuichi Ashida, Kyoto University, JAPAN Beneficiation Prospects of Baked Coking Coals from Seam XV, Jharia Coalfield, Damodar Valley, India, Ashok K Singh, N. K. Shukla, N. Choudhury, Central Institute of Mining & Fuel Research, CSIR; Mamta Sharma, National Metallurgical Laboratory, CSIR, INDIA

SESSION 11 SUSTAINABILITY AND ENVIRONMENT – 2 Brenda Pierce and Aysel Atimtay Emissions from Cofiring Coal with Renewable Materials Such as Biomass and Sewage, Lesley Sloss, IEA Clean Coal Centre, UNITED KINGDOM Element Leaching from Coal Stockpiles – Case Studies from the Sydney and Collie Basins, Australia, Colin R. Ward, Leanne Stephenson, Zhongsheng Li, University of New South Wales; David French, Ken Riley, Owen Farrell, CSIRO Energy Technology, AUSTRALIA Coal Resource Estimation in Isiklar-Kisrakdere (Soma, Manisa, Turkey), A. Erhan Tercan, Bahtiyar Ünver, Mehmet Ali Hindistan, Hacettepe University; Perihan Çorbacı, Kıvanç Het, Turkish Coal Enterprises, TURKEY Coal Explorations in Turkey: New Projects and New Reserves, İlker Şenguler, MTA, TURKEY SESSION 12 COAL-DERIVED PRODUCTS: ACTIVATED CARBON PRODUCTION – 1 Belma Demirel and Hasan Heperkan Activated Carbon from Brown Coal by Chemical Activation, Luguang Chen, Sankar Bhattacharya, Monash University, AUSTRALIA

Effect of Coal Volatile Matter on Emissions of Boiler Combustion, Hasancan Okutan, Nalan Erdöl Aydın, Erhan Böke, İstanbul Technical University, TURKEY SESSION 14 GASIFICATION: UNDERGROUND COAL GASIFICATION – 2 Rohan Courtney and Sibel Özdogan The Improvement of UCG Processes, Karol Kostur, Technical University of Košice, SLOVAK REPUBLIC An Integrated 3-D UCG Model for Predicting Cavity Growth, Product Gas, and Interactions with the Host Environment, John J. Nitao, David W. Camp, Souheil M. Ezzedine, Thomas A. Buscheck, S. Julio Friedmann, Lawrence Livermore National Laboratory, USA Quantification of the Effects of Various Thermal Boundary Conditions in the Underground Coal Gasification Cavities Using a Compartment Model, Sateesh Daggupati, Ramesh Naidu Mandapati, Sanjay M Mahajani, Anuradda Ganesh, Preeti Aghalayam, IIT Bombay; Sapru R.K., Sharma R.K., ONGC, INDIA Estimation of Chemical Reaction Occurred in Underground Coal Gasification, Osamu Yamada, Mamoru Kaiho, National Institute of Advanced Industrial Science and Technology (AIST); Sohei Shimada, The University of Tokyo; Shouji Fujioka, Japan Coal Energy Center; Jie Liang, China University of Mining and Technology, JAPAN

Low-Temperature Catalytic Graphitization of Carbon Material, Ch. N. Barnakov, A.P. Kozlov, V.Yu.Malysheva, Institute of Coal and Coal Chemistry SB RAS; S.K. Seit-Ablaeva, Kemerovo Technological Institute of Food Industry; Z. R. Ismagilov, M.A.Kerzhentsev, Boreskov Institute of Catalysis SB RAS, RUSSIA

Computational Flow Modeling of Underground Coal Gasification (UCG) Process, Sateesh Daggupati, Ramesh Naidu Mandapati, Sanjay M Mahajani, Preeti Aghalayam, Anuradda Ganesh, IIT Bombay; Sapru R.K, Sharma R.K., UCG Group, IRS, ONGC, INDIA

Research on the Preparation of Highthermalconductivity Carbon Block by the Ordered Growth of Self-assembled Mesophase, Ming-Lin Jin, Rong-Hua Liu, Qing-Zhong Cheng, Jingxia Hu, Shanghai Institute of Technology; Zong-Hong Bao, Nanjing University of Technology, CHINA

SESSION 15 CARBON MANAGEMENT: CO2 SEQUESTRATION Leslie Ruppert and Hanzade Acma

SESSION 13 COMBUSTION: MERCURY AND TRACE ELEMENTS John Wheeldon and Mustafa Ziypak UNEP Coal Combustion Partnership Area Activities Prior to 2013 Global Mercury Treaty, Wojciech Jozewicz, ARCADIS, USA; Lesley Sloss, IEA Clean Coal Centre, UNITED KINGDOM; Gunnar Futsaeter, United Nations Environment Programme, SWITZERLAND Direct Measurement of Mercury in Simulated Flue Gas, Bihter Padak, Jennifer Wilcox, Stanford University, USA Modeling Trace Element Release from Included and Excluded Pyrite during Pulverized Coal Combustion, Wayne S. Seames, Esam I. Jassim, Steven A. Benson, University of North Dakota, USA Online Monitoring of Boron in Flue Gas Desulfurization Effluents by Fully Automated Measuring Equipment, Seiichi Ohyama, Keiko Abe, Hitoshi Ohsumi, Hirokazu Kobayashi, Central Research Institute of Electric Power Industry; Naotsugu Miyazaki, Koji Miyadera, Kin-ichi Akasaka, DKK-TOA Corporation, JAPAN Mercury Sorption on Brominated Activated Carbon, Erdem Sasmaz, Jennifer Wilcox, Stanford University, USA

CO2 Sequestration in Unminable Coal with Enhanced Coal Bed Methane Recovery: The Marshall County Project, Richard A. Winschel, James E. Locke, Ravi S. Srivastava, CONSOL Energy Inc.; Richard A. Bajura, Tom Wilson, Hema J. Siriwardane, Henry Rauch, Douglas Patchen, Brad D. Hega, Raj K. Gondle, West Virginia University; Arthur W. Wells, NETL/DOE, USA A GIS-DSS for a CO2-ECBM Project Feasibility Study: Case of Sulcis Coal Basin (Sardinia, Italy), Raimondo Ciccu, Alessandro Mazzella, Caterina Tilocca, University of Cagliari; Paolo Deiana, Sezione Impianti e Processi ENEA - Agenzia Nazionale per le Nuove Tecnologie, ITALY Effect of Rock Composition on Mineralization in Sequestration, Prashanth Mandalaparty, Milind Deo, Joseph Moore, University of Utah, USA Geological CO2 Storage in Coal-Bearing Formation, Sohei Shimada, Zhenjie Chai, Naoto Sakimoto, The University of Tokyo, JAPAN Experimental Study on Carbon Dioxide Sequestration by Mineral Carbonation, Jun-Ying Zhang, Heng Yan, Yong-Chun Zhao, Chu-Guang Zheng, Huazhong University of Science and Technology, CHINA CO2 Sequestration for the Shenhua DCL Plant in China, Qingyun Sun, Jerald J. Fletcher, US-China Energy Center, West Virginia University, USA

8

TECHNICAL PROGRAM SESSION 16 COAL SCIENCE: COKING Susan J. Tewalt and Erdoğan Yuzer

Clean Fuel Production Works from Canakkale-Can Coals, Oguz Altun, Akan Gulmez, Ayşe Erdem, Selami Toprak, Mineral Research and Exploration Directorate in Turkey; Zeki Olgun, Turkish Coal Enterprises, TURKEY Thermoplasticity Improvement of Coal Blends by Adding Solvent-Extracted Coal, Noriyuki Okuyama, Hiroki Shishido, Koji Sakai, Maki Hamaguchi, Nobuyuki Komatsu, KOBE STEEL, Ltd.; Haruo Kumagai, Hokkaido University, JAPAN SESSION 17 SUSTAINABILITY AND ENVIRONMENT – 3 Brenda Pierce and Mustafa Cetin Carbon Capture and Integration: An Alternative Perspective to CO2 Emissions and Carbon Capture and Sequestration, Catherine A. McGanity, University of Richmond, USA Future Coal Production Outlooks in the IPCC Emission Scenarios: Are They Plausible?, Mikael Höök, Uppsala University, SWEDEN The European Coal Market, a Prosperous Future?, Manfred Rumberger, ER-Consult GmbH, GERMANY Methane Enrichment from Anaerobic Digestion Gas (ADG) Using Polymeric Hollow Fiber Membrane, Hyung-Keun Lee, Dae-Hoon Kim, Ki-Hong Kim, YoungMo An, Hang-Dae Jo, Korea Institute of Energy Research; Gang-Woo Lee, Yoo Sung Co. R&D Center, KOREA; KiJun Baik, Yanbian University of Science and Technology, CHINA

Institute of Chemistry and Chemical Technology of Siberian Branch of Russian Academy of Sciences, RUSSIA Research on the Preparation of Thermal Conductivity C/C Composites by One-step Hot Press Molding, Jin Ming-Lin, Qingzhong Cheng, Shanghai Institute of Technology; Yan-Wen Zhang, Xiao-Long Zhou, East China University of Science and Technology, CHINA

ORAL SESSIONS Wednesday, October 13, 2010 11:00 - 17:50 SESSION 19 COMBUSTION: OXY-COAL DEVELOPMENT – 2 Steven Carpenter and Yucel Ozden Comparative Study of Coal Ash and Deposits from Air and Oxy-Fuel Combustion, Jost O.L. Wendt, Dunxi Yu, William J. Morris, University of Utah; Andrew Fry, Constance L. Senior, Reaction Engineering International, USA Oxy-Fuel Combustion: A Technological Option for Retrofitting Existing Pulverized Lignite Fired Power Plants in Turkey, İskender Gökalp, CNRS-ICARE, FRANCE; Mücella Ersoy, Turkish Coal Enterprises, TURKEY Oxy-Fuel Combustion Chemistry – Implications on Corrosion Related Issues, Klas Andersson, Daniel Fleig, Fredrik Normann, Filip Johnsson, Chalmers University of Technology, SWEDEN

Effect of Introduction of Clean Coal Technology on Future Asian Energy Supply, Sohei Shimada, Yuta Koyama, The University of Tokyo, JAPAN

SESSION 20 GASIFICATION: FUNDAMENTALS – 1 Ke Liu and Hüsnü Atakul

SESSION 18 COAL-DERIVED PRODUCTS: ACTIVATED CARBON PRODUCTION – 2 Belma Demirel and Özcan Gulsoy

Modeling of Coal Char Gasification in Coexistence of CO2 and H2O, Satoshi Umemoto, Shiro Kajitani, Saburo Hara, Central Research Institute of Electric Power Industry (CRIEPI), JAPAN

Investigation of Carbonization Kinetic of Tunçbilek Lignite Used for the Preparation of Activated Carbon, Burcu Özdemir, Nilgün Karatepe, Reha Yavuz, Istanbul Technical University, TURKEY

Investigation of Component Release During Pressurized, High Heating Rate Devolatilization of Coal and Petroleum Coke, David Wagner, Kevin J. Whitty, University of Utah; Glenn L Shoaf, Paul Fanning, Eastman Chemical Company, USA

Synthesis of Nitrogen-Doped Carbon Materials from Coal-Tar and Petroleum Pitches and Nitrogen Containing Organic Precursors, Z. R. Ismagilov, M.A.Kerzhentsev, I.Z.Ismagilov, Boreskov Institute of Catalysis SB RAS; Ch. N. Barnakov, A.P. Kozlov, Institute of Coal and Coal Chemistry SB RAS; E.I.Andreikov, Institute of Organik Synthesis UB RAS, RUSSIA Effect of Mineral Matter of Brown Coals on the Reactivity of Char Steam Gasification and on the Properties of Activated Carbons, P.N.Kuznetsov, Kolesnikova S.M., L.I.Kuznetsova, Institute of Chemistry and Chemical Technology of Siberian Branch of Russian Academy of Sciences; Yu.F.Patrakov, Institute of Coal and Coal Chemistry of Siberian Branch of Russian Academy of Sciences, RUSSIA Study of the Properties of Coal from Mongolian SaikhanOvoo Deposit and the Char and Carbons Produced, B.Purevsuren, Ya.Davaajav, Kh.Serikjan, S.Batbileg, Institute of Chemistry and Chemical Technology, Mongolian Academy of Sciences, MONGOLIA; P.N Kutsnezov,

High Pressure TGA Studies on German Brown Coal under Carbon Dioxide Atmosphere, Abhishek Bhargava, Patrick J. Masset, Freiberg University of Mining and Technology, GERMANY Studies on Gasification of Char in Fixed Bed Reactor, Ramesh Naidu Mandapati, Preeti Aghalayam, Sateesh D, Narseh D, Sanjay M Mahajani, Anuradda Ganesh, IIT Bombay; Sharma R.K., IRS, ONGC, INDIA SESSION 21 CARBON MANAGEMENT: PRE-COMBUSTION CO2 CAPTURE Richard Sakurovs and Mustafa Ozdingis Development of Dry Regenerable CO2 Sorbent and WGS Catalyst for SEWGS Process, Joong Beom Lee, Tae Hyoung Eom, Jungho Ryu, Jeom-In Baek, Dong-Hyeok Choi, Keun Woo Park, Seong Jegarl, Seug-Ran Yang, Korea Electric Power Research Institute, KOREA

Development of Dry Regenerable CO2 sorbent for Fluidized-Bed CO2 Capture Process from Coal Power Plant, Chong Kul Ryu, Joong Beom Lee, Tae Hyoung Eom, Bok Suk Oh, Jeom-In Baek, Kyeongsook Kim, Young Ho Wi, Jegarl Seong, Won Sik Jeon, Korea Electric Power Research Institute, KOREA SESSION 22 COAL SCIENCE: CBM/CARBON DIOXIDE Richard Winschel and Işık Özpeker A Comparison of Experimental and Theoretical High Pressure CO2 Isotherms on Coals from the Upper Silesian Basin, Czech Republic, Zuzana Weishauptová, Oldřich Přibyl, Martina Švábová, Institute of Rock Structure and Mechanics, Academy of Sciences of the Czech Republic, CZECH REPUBLIC Conversion of Ukrainian Low Grade Solid Fuels with CO2 Capture, О.М. Dudnyk, I.S. Sokolovska, Coal Energy Technology Institute of National Academy of Sciences of Ukraine, UKRAINE Coalbed Gas Potential in the Miocene Soma Basin (Western Turkey), Sedat İnan, Aynur Dikbaş, Semih Ergintav, Fırat Duygun, TÜBİTAK Marmara Reasearch Center; M.Namık Yalçın, Kübra Tırpan, Korhan Yaşar, İstanbul University; Ender Okandan, Mustafa Baysal, Middle East Technical University; Yuda Yürüm, Sabancı University; Ruhi Saatçılar, Sakarya University; Murat Yılmaz, Ali Rıza Toygar, Turkish Petroleum Corporation; Ayhan Kösebalaban, Selahaddin Anaç, Hakkı Duran, Mehmet Onbaşı, Mücella Ersoy, Mehmet Atasayar, İsmail Ergüder, Turkish Coal Enterprises, TURKEY Underground Coal Determination by Integrated (Reflection & WVSP) Seismic in the Miocene Soma Basin (Western Turkey), Ruhi Saatçılar, Sedat İnan, Fırat Duygun, Semih Ergintav, Aynur Dikbaş, TÜBİTAK Marmara Reasearch Center; Murat Yılmaz, Ali Rıza Toygar, Turkish Petroleum Corporation; Ayhan Kösebalaban, Selahaddin Anaç, Hakkı Duran, Mehmet Onbaşı, Mücella Ersoy, Mehmet Atasayar, İsmail Ergüder, Turkish Coal Enterprises; M.Namık Yalçın, İstanbul University; Ender Okandan, Middle East Technical University; Yuda Yürüm, Sabancı University; Emin Demirbağ, İstanbul Technical University, TURKEY SESSION 23 SUSTAINABILITY AND ENVIRONMENT – 4 Brenda Pierce and Orhan Kural The Thermal Treatment Study of Pyrite from South Brazil Coal Mining Industry, Michael Peterson, Jussara P. Pizzolo, Morgana Bom, Deise Tramontin, Gabriela B. Vieira, Keli V.S. Damin, Universidade do Extremo Sul Catarinense; Adilson Oliveira, Formula Chemical Company, BRAZIL Pyrolysis of the Various Types of Fuels - Credit Cards, Dagmar Juchelkova, Helena Raclavska, Adela Cízkova, Vaclav Roubicek, VSB-Technical University of Ostrava, CZECH REPUBLIC SESSION 24 COAL-DERIVED PRODUCTS: DIRECT COAL-TO-LIQUIDS Rachid Oukaci and Fatma Arslan Alliance DCL Technology for Producing Ultra Clean Transportation Fuels, Theo L.K. Lee, Headwaters CTL, LLC.; John Duddy, Axens North America, Inc., USA

9

TECHNICAL PROGRAM Direct Coal to Liquids (DCL): High Quality Jet Fuels, W. Weiss, H. Dulot, A. Quignard, N. Charon, M. Courtiade, IFP New Energy, FRANCE Extraction of Brown and Sapropelitic Coals with Toluene and Water Containing Fluids under Supercritical Conditions, P.N.Kuznetsov, S.M.Kolesnikova, L.I.Kuznetsova, E.S.Kamensky, Institute of Chemistry and Chemical Technology of Siberian Branch of Russian Academy of Sciences; V.A.Kashirtsev, Trofimuk Institute of Petroleum Geology and Geophysics SB RAS, RUSSIA Research and Development of New Type Reactor for Direct Coal Liquefaction, Hu Fating, Li Peilin, Shi Shidong, Liu Min, Beijing Research Institute of Coal Chemistry, China Coal Research Institute, CHINA Characterization and Distribution of Phenolics in Direct Coal Liquefaction Oils, Zhennan Gao, Xuefeng Mao, Beijing Research Institute of Coal Chemistry, China Coal Research Institute; Wenhua Li, GE China Technology Center, CHINA SESSION 25 COMBUSTION: CHEMICAL LOOPING DEVELOPMENT – 2 Steven Carpenter and Murat Özbayoglu Development of Real-Time Dynamic Simulation of Chemical Looping Process for Advanced Controls, Xinsheng Lou, Hao Lei, Abhinaya Joshi, Carl Neuschaefer, Alstom Power Inc., USA Effect of H2S on Chemical Looping Combustion of CoalDerived Synthesis Gas over Fe2O3 - MnO2 Supported on ZrO2/Sepiolite/Al2O3, Ewelina Ksepko, Marek Sciazko, Institute for Chemical Processing of Coal, POLAND; Ranjani V. Siriwardane, Hanjing Tian, Thomas Simonyi, James A. Poston Jr, U.S. DOE, USA Effect of Coal Blending Method on Combustion Characteristics and NOx Emission in a Drop Tube Furnace, Song-gon Kim, Chun-sung Lee, Byoung-Hwa Lee, Ju-Hun Song, Young-June Chang, Chung-Hwan Jeon, Pusan National University, SOUTH KOREA TGA and DTF Studies on Coal Blends to Assess Combustion Performance, P. Sarkar, A. Mukherjee, S. G. Sahu, A. Choudhury, A. K. Adak, M. Kumar, N. Choudhury, S. Biswas, Central Institute of Mining and Fuel Research; S. Ghosal, Jadavpur University, INDIA SESSION 26 GASIFICATION: FUNDAMENTALS – 2 Ke Liu and Ayhan Sirkeci Effect of Operation Parameters on Gasification for the Production of Synthesis Gas, Atilla Biyikoğlu, Bekir Zühtü Uysal, Gazi University; Afşin Güngör, Niğde University; Coşku Kasnakoğlu, Murat Özbayoğlu, TOBB University of Economics and Technology, TURKEY High-Pressure and High-Temperature Gasification of Upgraded Brown Coal Char Using a Mini Direct Heating Reactor, Kouichi Miura, Ryuichi Ashida, Mitsunori Makino, Xian Li, Kyoto University, JAPAN C F D Si mulation of Proces s - dr i v e n Pa r t i c l e Fragmentation in a Coal Bed Gasifier, Franz Holzleithner, Roland Eisl, Markus Haider, Institute for Energy Systems and Thermodynamics, Vienna University of Technology; Georg Aichinger, Siemens VAI Metals Technologies Gmbh&Co, AUSTRIA Experimental Investigation into Primary Fragmentation of Large Coal Particles, Adam Luckos, Roelof L.J. Coetzer, Ed L. Koper, Sasol Technology R&D, SOUTH

AFRICA; Monika Kosowska-Golachowska, Częstochowa University of Technology, POLAND

Joubert, Trudie Brittz, Sasol Technology R&D, SOUTH AFRICA

Investigation of the Pyrolysis and Gasification of a Turkish Coal Using Thermal Analysis Coupled with Mass Spectrometry, Sibel Özdoğan, Ugur Özveren, Mehmet Beypınar, Marmara University; Aylin Boztepe, Yildiz Technical University, TURKEY

Dry Coal Cleaning Using the FGX Separator, Mehmet Saracoglu, Rick Q. Honaker, University of Kentucky, USA

SESSION 27 CARBON MANAGEMENT: POST-COMBUSTION CO2 CAPTURE – 1 Richard Sakurovs and Mehmet Cagil Development of Post Combustion Carbon Capture Technology, Matthew Hunt, F. D. Fitzgerald, S. Black, R. A. Gardiner, Doosan Power Systems, UNITED KINGDOM Lignite Derived Carbons for CO2 Capture, Alan L Chaffee, Seamus Delaney, Gregory Knowles, Monash University, AUSTRALIA Physical Properties and Reactivities of Mg-Based Dry Regenerable CO2 Sorbents Prepared by Spray-Drying Method, Jeom-In Baek, Tae Hyoung Eom, Joong Beom Lee, Won Sik Jeon, Chong Kul Ryu, Korea Electric Power Research Institute, KOREA SESSION 28 COAL SCIENCE: COAL CHEMISTRY – 2 Frans Waanders and Dündar Ergunalp Carbonaceous Particles from the Incomplete Combustion, Martina Havelcová, Ivana Sýkorová, Hana Trejtnarová, Alexandr Šulc, Institute of Rock Structure and Mechanics AS CR,v.v.i., CZECH REPUBLIC Trace Element Partitioning and Leaching in Solids Derived from Gasification of Australian Coals, Alexander Ilyushechkin, Daniel Roberts, David Harris, Ken Riley, CSIRO Energy Technology, AUSTRALIA Effect of Coal Rank on Carbon Oxides Formation via the Low Temperature Atmospheric Oxidation Process, Uri Green, Zeev Aizenstat, Hebrew University of Jerusalem; Haim Cohen, Ariel University Center at Samaria and Ben-Gurion University of the Negev, ISRAEL; Christoph Wiedner, Sven Stark, Ariel University Center at Samaria, ISRAEL and TU Bergakademie Freiberg, GERMANY Quantitative Determination of Minerals in Coal by CQPAC Method, Zdeněk Klika, VŠB-Technical University Ostrava, CZECH REPUBLIC The Catalytic Effect of Added Sodium- and Potassium Carbonate to an Acid Treated Inertinite Rich South African Bituminous Coal Char, Christien A Strydom, Lucinda Klopper, John R Bunt, North-West University, SOUTH AFRICA; Harold H Schobert, Penn State University, USA SESSION 29 COAL SCIENCE: BENEFICIATION – 1 B.K. Parekh and Selçuk Buyurgan Pre-Combustion Cleaning of Pulverized Fine Coal at Power Plant Using Novel RTS Dry Separation Technology, Daniel Tao, Ahmed Sobhy, Qin Li, Rick Honaker, University of Kentucky, USA The Prediction of Caking Propensity of Gondwanaland Coals Using Petrography, Daniel Van Niekerk, Johan

Improving the Efficiency of Lignite Drying, Wayne S. Seames, Carlos J. Bucaram, Steven A. Benson, University of North Dakota; Srivats Srinivasachar, Envergex, LLC, USA SESSION 30 COAL-DERIVED PRODUCTS: COAL-TO-LIQUIDS/FISCHER-TROPSCH – 1 Rachid Oukaci and Alp Gurkan Deactivation of Iron Based Fischer-Tropsch Catalyst: A Critical Problem, Belma Demirel, Deena Ferdous, Rentech, Inc., USA Product Distribution and Reaction Pathways during Fischer-Tropsch Synthesis on an Iron Catalyst, Dragomir B. Bukur, Texas A&M University at Qatar, QATAR; Lech Nowicki, Technical University of Lodz, POLAND SESSION 31 COMBUSTION: FLUIDIZED-BED COMBUSTION AND CO-FIRING – 1 Steven Carpenter and Gündüz Atesok Smartsheet Tool Applied to Boiler Performance Analysis and Economic Optimization of a Circulating Fluidized Bed Boiler, Abhinaya Joshi, Xinsheng Lou, Carl Neuschaefer, Paul Panos, Alstom Power Inc.; Weikko Wirta, AES Thames, USA A Model of Primary Fragmentation of Coal Particles in Fluidized-Bed Combustion, Adam Luckos, Sasol Technology R&D, SOUTH AFRICA; Monika KosowskaGolachowska, Częstochowa University of Technology, POLAND Main Problems Concerning Co-Firing Biomass Mixture with Hard Coal in Pulverized-Fuel Boilers, Krzysztof Jesionek, Henryk Karcz, Marcin Kantorek, Wrocław University of Technology, POLAND Te c h n i c a l a n d E c o n o m i c E v a l u a t i o n o f t h e Desulphurization Processes at Power Stations Using Lignite, Hasancan Okutan, Bülent D. Çift, İstanbul Technical University, TURKEY Desulfurization Characteristics of Powdered Hydrated Lime for Flue Gas Sorbent Injection Process, Hyok Bo Kwon, Kyungnam University; Sang Whan Park, KIST; Hyung Taek Kim, Ajou University, KOREA Status of Large Circulating Fluidized Bed Boiler Operation in China, Jie Yu, Beijing Research Institute of Coal Chemistry, China Coal Research Institute, CHINA SESSION 32 GASIFICATION: FUNDAMENTALS – 3 Ke Liu and Gülhan Özbayoglu Performance of a 500 KWTH Pressurized EntrainedFlow Coal Gasifier, Kevin J. Whitty, Randy Pummill, David R. Wagner, Travis Waind, David A. Wagner, The University of Utah, USA Characterization of a Small Scale Slurry-Fed, OxygenBlown Entrained Flow Gasifier: How Injector Geometry Affects Flame Stability and Performance, Travis Waind, Kevin Whitty, University of Utah, USA

10

TECHNICAL PROGRAM

Analysis of Fines Produced from Non-Slagging Coal Gasifier and Evaluation of Economic Usage, Yongseung Yun, Seok Woo Chung, Na Rang Kim, Institute for Advanced Engineering, KOREA Development of Gas, Power and Tar Co-Generation System with Circulating Fluidized Bed Technology, Qinhui Wang, Mengxiang Fang, Zhongyang Luo, Mingjiang Ni, Kefa Cen, Zhejiang University, CHINA

Research and Exploration Directorate in Turkey; Zeki Olgun, Turkish Coal Enterprises, TURKEY

Tests on Drop Tube, Radim Paluska, Marian Bojko, VSB – Technical University of Ostrava, CZECH REPUBLIC

Drying Kinetics of Çanakkale – Çan Lignite, Ufuk Gündüz Zafer, Ö. Murat Dogan, Duygu Öztan, Bekir Zühtü Uysal, Gazi University; Zeki Olgun, Mustafa Ozdingis, Ömer Sezgin, Selahaddin Anac, Turkish Coal Enterprises (TKI), TURKEY

SESSION 38 GASIFICATION: FUNDAMENTALS – 4 Johan van Dyk and Mevlüt Kemal

Thermal Chemical Process Study on Chemical Reaction Network of Jet-Fluidized Bed Gasifer Reaction System, Jie Feng, Xuecheng Hou, Wenying Li, Xiao-Hui Chen, Taiyuan University of Technology, CHINA

Innovative High Energy Efficiency Brown Coal Drying based on Self-Heat Recuperation Technology, Muhammad Aziz, Chihiro Fushimi, Yasuki Kansha, Kazuhiro Mochidzuki, Shozo Kaneko, Atsushi Tsutsumi, The University of Tokyo, JAPAN

Optimization of Canadian Petroleum Coke, Coal and Fluxing Agent Blends via Slag Viscosity Measurements and Models, Marc A. Duchesne, Arturo Macchi, University of Ottawa; Ben Anthony, CanmetENERGY, CANADA; Alexander Y. Ilyushechkin, CSIRO Energy Technology, AUSTRALIA

SESSION 33 CARBON MANAGEMENT: POST-COMBUSTION CO2 CAPTURE – 2 Richard Sakurovs and Mustafa Tiris

SESSION 36 COAL-DERIVED PRODUCTS: COAL-TO-LIQUIDS/FISCHER-TROPSCH – 2 Rachid Oukaci and Ali İhsan Arol

An Efficient Membrane Process to Capture Carbon Dioxide From Power Plant Flue Gas, Bilgen Firat Sercinoglu, Tim Merkel, Xiaotong Wei, Haiqing Lin, Jenny He, Richard Baker, Karl Amo, Hans Wijmans, Membrane Technology and Research, Inc., USA

Using Pyrolysis Tar to Meet Fuel Specifications in Coal-to-Liquids Plants, Jaco Schieke, Foster Wheeler, UNITED KINGDOM

Effect of Dense Medium Separation of a South African Coal Source on Slag-Liquid Formation: An Experimental and Factsage Approach, JC van Dyk, SASOL Technology; FB Waanders, North West-University, SOUTH AFRICA

CO2 Capture by Condensed Rotational Separation, R.J. van Benthum, H.P. van Kemenade, J.J.H. Brouwers, M. Golombok, Eindhoven University of Technology, THE NETHERLANDS Influence of Pressure on Dry Reforming of Methane over Carbonaceous Catalyst, Bingmo Zhang, Yongfa Zhang, Guojie Zhang, Fengbo Guo, Taiyuan University of Technology, CHINA SESSION 34 COAL SCIENCE: COAL CHEMISTRY – 3 Frans Waanders and Neset Acarkan Predetermination of the Fault Crossing the Underground Coal Mine Galeries by Seismic Reflection Method: An Application at a Longwall Coal Mine in Turkey, G.G.U. Aldas, B. Kaypak, B. Ecevitoglu, Ankara University; A. Can, General Directorate of Mineral Research and Exploration, TURKEY Alberta’s 2 Trillion Tonnes of ‘Unrecognized’ Coal, R.J. Richardson, CanZealand Geoscience Ltd., NEW ZEALAND Temperature as a Factor Affecting Adsorption Behavior of Coal to Lead (II) Ions, Boleslav Taraba, Petra Vesela, Roman Marsalek, Zuzana Navratilova, Ostrava University, CZECH REPUBLIC Bioflotation of Coal, Peter Fecko, Tana Kantorkova, Radomir Michniak, Lukas Koval, Alena Kasparkova, VSB - Technical University of Ostrava, CZECH REPUBLIC SESSION 35 COAL SCIENCE: BENEFICIATION – 2 B.K. Parekh and Ali Güney Suitability of the Sulcis Coal for CWS Preparation, Raimondo Ciccu, Giovanni Mei, Caterina Tilocca, University of Cagliari; Paolo Deiana, Sezione Impianti e Processi ENEA - Agenzia Nazionale per le Nuove Tecnologie, ITALY

Overview of the Rentech Process, Belma Demirel, Harold Wright, Rentech, Inc., USA

Compositional Variations in Pilot Gasifier and Laboratory-Produced Slags and their Impacts on Slag Viscosity and Coal Assessment, Alexander Ilyushechkin, D. Roberts, D. Harris, CSIRO, Energy Technology, AUSTRALIA

Comparative Evaluation of Different Coal to Liquid Process Conditions via Fischer-Tropsch Synthesis, Serhat Gul, Atilla Ersoz, Murat Baranak, Omer Faruk Gul, Fehmi Akgun, TUBİTAK Marmara Research Centre, Energy Institute, TURKEY

Shaping Slag Flow in an Entrained Flow Gasifier: Numerical Simulation and Physical Experiments, Randy Pummill, Gabriel Hansen, Kevin Whitty, University of Utah, USA

Technoeconomic and Environmental Life Cycle Analysis of Coal and Coal/Biomass to Liquids Facilities, Anastasia M Gandrik, Idaho National Laboratory; Vivek P. Utgikar, University of Idaho, USA

Influence of Gasification Conditions on the Properties of Fly Ash in a Bench-Scale Opposed Multi-Burner Gasifier, Qinghua Guo, Guangsuo Yu, Fuchen Wang, Zhenghua Dai, East China University of Science and Technology, CHINA

Conversion of Waste Biomass to Transportation Fuels: Energy for the Future, S.K.Srivastava, S.R.K.Rao, Amlendu Sinha, Central Institute of Mining and Fuel Research, INDIA

ORAL SESSIONS Thursday, October 14, 2010 11:00 - 17:50 SESSION 37 COMBUSTION: OXY-COAL DEVELOPMENT – 3 John Wheeldon and Ersan Putun The Effect of Coal Composition on Ignition and Flame Stability in Co-Axial Oxy-Fuel Turbulent Diffusion Flames, Dadmehr M. Rezaei, Eric G. Eddings, Kerry E. Kelly, Jingwei Zhang, Jost O.L. Wendt, University Of Utah, USA; Yuegui Zhou, Shanghai Jiao Tong University, CHINA Study on the In-Furnace Desulfurization in Oxy-Fuel Combustion using Drop Tube Furnace with Limestone, Hyung-Keun Lee, Wook Choi, Hang-Dae Jo, Won-Kil Choi, Korea Institute of Energy Research; Sang-In Keel, Korea Institute of Machinery & Materials, KOREA Ignition Loss and Ultrafine Particle and Soot Emissions From Air and Oxy-Coal Flames, William J. Morris, Dunxi Yu, Jost O. L. Wendt, University of Utah, USA

Comparative Study of Oil Agglomeration and Flotation of Low Grade Coals, Feridun Boylu, Fırat Karakas, Istanbul Technical University, TURKEY

High Speed Video Analysis of Oxycoal Combustion in 40kw Coaxial Turbulent Diffusion Flames, Terry A. Ring, Jingwei Zhang, Husam el Gendy, Jost O.L. Wendt, Kerry Kelly, Eric G. Eddings, University of Utah, USA

Briquetting Studies of Canakkale-Can Coals, Oguz Altun, Akan Gulmez, Ayşe Erdem, Zafer Gencer, Mineral

Comparison of the Mathematical Model of Pulverized Coal Burnout with Results Gained from Experimental

SESSION 39 GASIFICATION: GENERAL SESSION – 2 Ting Wang and Muammer Bulut GTI’s Sampling and Analysis Systems for Gas Streams of Gasification and Downstream Processes, Osman M. Akpolat, Tanya S. Tickel, Rachid B. Slimane, Chun W. Choi, Gas Technology Institute (GTI), USA Design of Comminution Unit for the Gasification Pilot Plant, N. Acarkan, G. Onal, A. A. Sirkeci, G. Atesok, Istanbul Technical University, TURKEY Preparation of Coal Water Mixture with High Concentration from Low Rank Coals and Lignite by Dry Fine Coal with Optimum Particle Distribution, Baoqing Li, Institute of Coal Chemistry, Chinese Academy of Sciences; Feng He, Yulin Western Coal Technology Research Center, CHINA Fluidised Bed Co-Gasification of Coal and Biomass Under Oxy-Fuel Conditions, Marcos Millan, Nicolas Spiegl, Nigel Paterson, Cesar Berrueco, Imperial College London, UNITED KINGDOM Co-Gasification of Footwear Leather Waste and High Ash Coal: A Thermodynamic Analysis, Rodolfo Rodrigues, Nilson R. Marcílio, Jorge O. Trierweiler, Federal University of Rio Grande do Sul (UFRGS); Marcelo Godinho, University of Caxias do Sul (UCS); Adriene M. S. Pereira, Pontifical Catholic University of Rio Grande do Sul (PUC-RS), BRAZIL

11

TECHNICAL PROGRAM SESSION 40 COAL SCIENCE: COAL CHEMISTRY – 4 Ashok K. Singh and Omer Unver Mineralogical, Petrographic and Geochemical Features of the Achlada and Mavropigi Lignite Deposits, NW Macedonia, Greece, Colin R. Ward, Zhongsheng Li, University of New South Wales; Stavros P. Kalaitzidis, BHP Billiton Mitsubishi Alliance, AUSTRALIA; Nikolaos Koukouzas, Centre for Research and Technology Hellas, Institute for Solid Fuels Technology and Applications, GREECE X-Ray Computer Tomography on Coal Particles, Patrick J. Masset, Freiberg University of Mining and Technology, GERMANY; Heikki Suhonen, European Synchrotron Radiation Facility, FRANCE Mineralogy, Geochemistry, and Petrography of Upper Permian Bituminous and Carboniferous Anthracite Coals from Xuanwei County, Eastern Yunnan Province, China, Harvey E. Belkin, U.S. Geological Survey; James C. Hower, Jordan W. Drew, University of Kentucky, CAER, USA; Linwei Tian, Chinese University of Hong Kong, CHINA SESSION 41 COAL SCIENCE: BENEFICIATION – 3 B.K. Parekh and Steven Carpenter Pyrolysis Residue from Waste Materials in Black Coal Flotation, Peter Fecko, Alena Kasparkova, Vlastimil Kriz, Josip Isek, Tien Pham Duc, VSB – Technical University of Ostrava, CZECH REPUBLIC Studies of a Multi Gravity Separator (MGS) to Produce Clean Coal from Turkish Lignite and Hard Coal Fine Tailings, Eyüp Sabah, Selçuk Özgen, M. Fatih Can, Afyon Kocatepe University, TURKEY Petrographic Characterisation of Beneficiated Material of Tailings from Soma Işıklar Derekoy Coal Washing Plant (TURKEY),by Multi Gravity Seperator (MGS), Selami Toprak, Ayşe Erdem, Akan Gulmez, Oguz Altun, Mineral Research and Exploration Directorate in Turkey; Sarper Alyildiz, Turkish Coal Enterprises, TURKEY Soma Region’s Coals Washing at Dereköy Washery with Working 800 TPH and it’s Washing Performance Evaluation, S.I. Alyildiz, S. Gurkan, S. Tuncer, Directorate of the Aegean Lignite Establishment, TURKEY SESSION 42 COAL-DERIVED PRODUCTS: H2 PRODUCTION/SNG Dragomir Bukur and Vedat Arslan Development of Hydrogen Transport Membranes for Separating Hydrogen from Coal Gasification Stream, U. (Balu) Balachandran, T. H. Lee, C. Y. Park, Y. Lu, S. E. Dorris, Argonne National Laboratory, USA HTGR-Integrated Coal to Liquids Production Analysis, Anastasia M Gandrik, Rick A. Wood, Idaho National Laboratory, USA Methane Production from Coal, Coal-Biomass Mixtures, Arun SK Raju, Christophe Capelli, Viresco Energy LLC, USA Methanation of Syngas over Coral Reef-Ni/Alumina Catalysts, Yizhuo Han, Yisheng Tan, Institute of Coal Chemistry, Chinese Academy of Science; Shengli Ma, Graduate University of the Chinese Academy of Sciences, CHINA

SESSION 43 COMBUSTION: FLUIDIZED-BED COMBUSTION AND CO-FIRING – 2 John Wheeldon and Daman Wallia

Entrained Flow Slagging Slurry Gasification and the Development of Computational Fluid Dynamics Models at CanmetENERGY, Robin Hughes, Dennis Lu, Adrian Majeski, Ben Anthony, CanmetENERGY; Andrew Corber, National Research Council, CANADA

Sulphur Capture Under Fluidised Bed Combustion Conditions Using Coal Ashes as Sorbents, Rufaro Kaitano, Dursman Mchabe, Raymond C Everson, Hein W J P Neomagus, North-West University, SOUTH AFRICA

Numerical Simulation Analyses of an Entrained-Bed Gasification Reactor, Ming-Hong Chen, Tsung Leo Jiang, National Cheng Kung University; Yau-Pin Chyou, ChangBin Huang, Institute of Nuclear Energy Research Atomic Energy Council, TAIWAN, ROC

CO 2 Reduction Potential and Co-Combustion Possibilities of the FBC-Boilers on the Czech Conditions, Dagmar Juchelkova, Helena Raclavska, Jiri Bilik, Pavlina Pustějovská, VSB-Technical University of Ostrava, CZECH REPUBLIC

SESSION 46 COAL SCIENCE: COAL CHEMISTRY – 5 Quentin Campbell and Frantisek Verbich

Co-Combustion of Various Biowastes with a HighSulfur Turkish Lignite in a Circulating Fluidized Bed Combustor, Aysel T. Atimtay, Murat Varol, Middle East Technical University; Hayati Olgun, Alper Unlu, Berrin Bay, Ufuk Kayahan, TUBITAK-MRC, Energy Institute; Hüsnü Atakül, Mustafa C. Çelebi, Istanbul Technical University, TURKEY Co-Combustion Performance of Oil Shale and Biomass Fuels, Emre Özgür, Mustafa Verşan Kök, Middle East Technical University, TURKEY; Sharon Falcone Miller, Bruce G. Miller, Penn State University,USA SESSION 44 GASIFICATION: FUNDAMENTALS – 5 Johan van Dyk and Deniz Üner Influence of Steam on the Release of Alkali Metal, Chlorine, and Sulphur Species During High Temperature Gasification of Lignite, Marc Bläsing, Michael Müller, Institute of Energy Research (IEF-2), GERMANY Kinetics of Char and Catalyzed Char Gasification under High H2 and Steam Partial Pressure, Katsuhiro Nakayama, Yoshizo Suzuki, National Institute of Advanced Industrial Science and Technology; Shiying Lin, Japan Coal Energy Center, JAPAN Modeling of Steam Gasifier in Dual Fluidized Bed Gasification, Toshiyuki Suda, Zhihong Liu, Makoto Takafuji, Masahiro Narukawa, IHI Corporation, JAPAN Steam Gasification of Low Rank Coals with IonExchanged Sodium Catalysts Prepared Using Natural Soda Ash, Yasuo Ohtsuka, Yuu Hanaoka, Enkhsaruul Byambajav, Takemitsu Kikuchi, Naoto Tsubouchi, Tohoku University, JAPAN Separation of Pyrolysis from Fluidized Bed Steam Gasification: Its Conception and Application, Masahiro Narukawa, Makoto Takafuji, Toshiyuki Suda, IHI Corporation, JAPAN

Comparison of Measured and Calculated Viscosities of German Lignite Based Slags, Arne Bronsch, Patrick J. Masset, Freiberg University of Mining and Technology, GERMANY Drying Mechanism of Low Rank Coal with Different Reacting Conditions: Fixed Bed vs. Fluidized Bed, Hyungtaek Kim, Taejin Kang, Doman Jeon, Ajou University; Sihyun Lee, Sangdo Kim, Korea Institute of Energy Research, SOUTH KOREA The Natural Technology for Pretreatment and Utilization of the Energetic Fly Ash, Maria Kusnierova, Maria Prascakova, Institute of Geotechnics of Slovak Academy of Sciences, SLOVAK REPUBLIC; Peter Fecko, Rudolf Matysek, VSB-Technical University of Ostrava, CZECH REPUBLIC Physical Structure and Chemical Properties of Organic Matter of Brown Coals from Different Fields in Relation to the Composition of Mineral Components, P.N.Kuznetsov, L.I.Kuznetsova, Institute of Chemistry and Chemical Technology of Siberian Branch of Russian Academy of Sciences, RUSSIA SESSION 47 COAL SCIENCE: BENEFICIATION – 4 B.K. Parekh and Ekrem Yuce A Study to Recover Coal from Turkish Lignite Fine Coal Tailings: Comparison of Falcon Concentrator and Multi Gravity Separator (MGS), Eyüp Sabah, M. Fatih Can, Selçuk Özgen, Afyon Kocatepe University, TURKEY Evaluation of Dense Medium Separation Performance of Imbat Coal Preparation Plant, G.Özbayoğlu, Atilim University; Ü. Atalay, Ali İ.Arol, O.Sivrikaya, Middle East Technical University, TURKEY Effect of Shape Factor on Coal Flotation, G. Bulut, O. Güven, K.T. Perek, Istanbul Technical University, TURKEY

SESSION 45 GASIFICATION: MODELING – 1 Ting Wang and Ziya Cosar

SESSION 48 COAL-DERIVED PRODUCTS: GENERAL SESSION – 1 Dragomir Bukur and Hüseyin Özdag

Process Simulation - The Way from Pilot Plant to a Training-Centre for a 500 MW Gasifier System, Friedemann Mehlhose, Julia Kittel, S. Stoye, H. Kotthaus, Siemens Fuel Gasificaiton Technology GmbH & Co.KG, GERMANY

Reforming of Low Rank Coal by Solvent Treatment at Around 350OC, Xian Li, Ryuichi Ashida, Hiroyasu Fujitsuka, Kouichi Miura, Kyoto University, JAPAN

A Dynamic Simulator of a Commercial-Scale IGCC Plant, Mi-Yeong Kim, Yong-Jin Joo, In-Kyu Choi, JoongWon Lee, Si-Moon Kim, Min-Churl Lee, Korean Electrical Power Corporation, KOREA

An Experimental Investigation of Factors Related to Coke Strength Degradation in Coke Milli-Structure, Tetsuya Kanai, Yoshiaki Yamazaki, Kenichi Hiraki, Xiaoqing Zhang, Masakazu Shoji, Hideyuki Aoki, Takatoshi Miura, TOHOKU University, JAPAN

12

TECHNICAL PROGRAM

An Experimental Study on the Effect of Metallic Iron Particles on Strength Factors of Coke after CO2 Gasification Reaction, Yoshiaki Yamazaki, Kenichi Hiraki, Tetsuya Kanai, Xiaoqing Zhang, Masakazu Shoji, Hideyuki Aoki, Takatoshi Miura, Tohoku University, JAPAN SESSION 49 COMBUSTION: ASH DEPOSITION AND HEAT TRANSFER John Wheeldon and Oktay Erbatur Spectral Emissivities of Ni and Fe based Boiler Tube Materials with Varying Chromium Content at High Temperature Atmospheres, Miki Shimogori, BabcockHitachi K.K. Kure Research Laboratory, JAPAN; Fabian Greffrath, Viktor Scherer, Ruhr University of Bochum; Alfred Gwosdz, Christian Bergins, Hitachi Power Europe GmbH, GERMANY Effect of MGO Additive on the Reduction of Ash Deposition of Upgraded Brown Coal, Katsuya Akiyama, Haeyang Pak, Kobe Steel, Ltd.; Yasuaki Ueki, Ryo Yoshiie, Ichiro Naruse, Nagoya University, JAPAN Modeling and Optimization of NOx Emission and Pulverized Coal Flame in Utility Scale Furnaces, Srdjan Belosevic, Miroslav Sijercic, Branislav Stankovic, Nenad Crnomarkovic, Institute of Nuclear Sciences Vinca, Laboratory for Thermal Engineering and Energy; Slobodan Djekic, Electric Power Industry of Serbia, SERBIA Observation of Heat Release Region as Functions of Coal Properties in Turbulent Jet Pulverized Coal Flames, Yon Mo Sung, Cheor Eon Moon, Seong Yool Ahn, Jae Woo An, Gyung Min Choi, Duck Jool Kim, Pusan National University, SOUTH KOREA

SESSION 51 GASIFICATION: MODELING – 2 Ting Wang and Volkan S. Ediger

SESSION 54 COAL-DERIVED PRODUCTS: GENERAL SESSION – 2 Dragomir Bukur and Mehmet Canbazoglu

Investigation of Coal Gasification Process under Various Operating Conditions Inside a Two-Stage Entrained Flow Gasifier, Ting Wang, Armin Silaen, University of New Orleans, USA

Natural-Gas-Level Emissions when Burning Naphtha (without Water Injection) in a Commercial Gas Turbine using the LPP Technology, Creating a “Clean Power” Alternative for an Integrated Gasification Combined Cycle (IGCC) Polygen Plant, Leo D. Eskin, LPP Combustion, LLC., USA

Start-Up Behavior of a Fixed Bed Gasifier: One Dimensional Modeling, Giampaolo Mura, Mariarosa Brundu, University of Cagliari, ITALY Entrained Flow Coal Gasification: Modeling, Simulation & Experimental Uncertainty Quantification for a Laboratory Reactor, Philip J. Smith, Charles Reid, Julen Pedel, Jeremy Thornock, Institute for Clean and Secure Energy, The University of Utah, USA Numerical Simulation of the Hydrodynamics of a Fluidized Bed Combined with an Entrained Bed Gasifier, Jiantao Zhao, Jiejie Huang, Yitian Fang, Yang Wang, Institute of Coal Chemistry, Chinese Academy of Sciences, CHINA Numerical Simulation of the Gasification Process inside a Cross-Type Two-Stage Gasifier, Yau-Pin Chyou, ChangBin Huang, Yan-Tsan Luan, Institute of Nuclear Research, Atomic Energy Council, TAIWAN, ROC; Ting Wang, University of New Orleans, USA SESSION 52 COAL SCIENCE: COAL CHEMISTRY – 6 Quentin Campbell and Mustafa Aktas

Mathematical Model of the Low-Temperature Oxidation of Coal in Coal Stockpiles and Dumps, Marian Bojko, Milada Kozubkova, VŠB-Technical University; Zdeněk Michalec, Institute of Geonics AS CR, v. v. i., CZECH REPUBLIC

Uranium and Some Other Trace Metal Element Concentration of Some Turkish Coal Ashes, Isik Ozpeker, Fikret Suner, Mehmet Maral, Tahsin Aykan Kepekli, Istanbul Technical University, TURKEY

SESSION 50 GASIFICATION: GAS CLEANUP Johan van Dyk and İsmail Boz

Transformations of Karaman -Ermenek Lignites of Turkey under Accelerated Electrons Impact, Islam Mustafayev, Fethullah Chichek, Azerbaijan National Academy of Sciences, AZERBAIJAN; Guven Onal, Istanbul Technical University, TURKEY

Slipstream Tests of Palladium Sorbents for High Temperature Capture of Mercury, Arsenic and Selenium from Fuel Gas, Hugh G.C. Hamilton, Liz Rowsell, Stephen Poulston, Andrew Smith, Johnson Matthey Technology Centre, UNITED KINGDOM; Tony Wu, Subhash Datta, Robert C. Lambrecht, John Wheeldon, National Carbon Capture Center; Evan J. Granite, Henry W. Pennline, U.S. DOE/NETL, USA Mercury Measurement and Removal from an Entrained Flow Slagging Coal Gasifier, Dennis Lu, Robin Hughes, Ben Anthony, CanmetENERGY/Natural Resources Canada; Karl Abraham, Environment Canada, CANADA Performance Improvement of a Desulfurization Sorbent for Warm Synthesis Gas Cleanup, Jeom-In Baek, Jungho Ryu, Tae Hyoung Eom, Joong Beom Lee, Yong-Ho Wi, Chong Kul Ryu, Korea Electric Power Research Institute, KOREA

Desulfurization and Kinetics of Removal of Sulfur from High Sulfur Coal under Hydrogen Atmosphere, Guojie Zhang, Yongfa Zhang, Fengbo Guo, Bingmo Zhang, Taiyuan University of Technology, CHINA

Laminar Flame Speed Study of Syngas Mixtures (H2CO) with Straight and Nozzle Burners, İskender Gökalp, Nicolas Bouvet, Christian Chauveau, CNRS-Institut de Combustion, FRANCE; Seong-Young Lee, Robert J. Santoro, The Pennsylvania State University, USA Structural Changes in Bituminous Coal Fly Ash Due to Treatments with Aqueous Solutions, Roy Nir Lieberman, Ariel Goldman, Ariel University Center of Samaria; Haim Cohen, Ariel University Center at Samaria and Ben-Gurion University of the Negev, ISRAEL; Roy Nitzsche, TU Bergakademie Freiberg, GERMANY Influence Factors on Density and Specific Surface Area (Blaine Value) of Fly Ash from Pulverized Coal Combustion, Hiromi Shirai, Michitaka Ikeda, Kenji Tanno, Central Research Institute of Electric Power Industry, JAPAN

POSTER SESSIONS Tuesday, October 12, 2010 18:20 - 20:20 POSTER SESSION 1 COMBUSTION Ionic Liquids with Amine Functional Group: A Shortcut to Improve the Performance of Ionic Liquids for CO2 Scrubbing, Jelliarko Palgunadi, Jin Kyu Im, Antonius Indarto, Hoon Sik Kim, Minserk Cheong, Kyung Hee University, KOREA Absorption of Sulfur Dioxide in Task Specific Ionic Liquids Containing SO2-Philic Groups on the Cation, Sung Yun Hong, Jelliarko Palgunadi, Hoon Sik Kim, Minserk Cheong, Kyung Hee University, KOREA

SESSION 53 COAL SCIENCE: BENEFICIATION – 5 B.K. Parekh and Mehmet S. Celik

Reaction Characteristics of New Oxygen Carriers for Chemical Looping Combustion, Ho-Jung Ryu, Jaehyeon Park, Gyoung-Tae Jin, Korea Institute of Energy Research; Moon-Hee Park, Hoseo University, KOREA

Preparation of Alternative Fuel from Compost and Coal Slurries, Dagmar Juchelkova, O. Zajonc, H. Skrobankova, H. Raclavska, K. Raclavsky, VSB – Technical University Ostrava, CZECH REPUBLIC

Combustion Reactivity of Char Derived from Solvent Extracted Coal, Sihyun Lee, Hokyung Choi, Sangdo Kim, Jeongwhan Lim, Youngjoon Rhim, Korea Institute of Energy Research (KIER); Woosik Park, Hanyang University, KOREA

Investigation of Effect of Reagents on the Coal Recovery from Coal Washing Plant Tailings by Floatation, Oktay Bayat, Huseyin Vapur, Cukurova University; Metin Ucurum, Nigde University, TURKEY Column Flotation of Fine and Coarse Coal using a Novel Approach, B.K. Parekh, D.P. Patil, University of Kentucky CAER; Edgar B. Klunder, NETL,USA Study of the Lignite Qualitative Parameters Modification, During its Storage, Sanda Krausz, Nicolae Cristea, University of Petrosani; Ion Bacalu, Mihail Dafinoiu, Daniel Burlan, National Society of Lignite, Oltenia, ROMANIA

Forced Flame Response Measurement in a Gas Turbine Combustor with High Hydrogen Fuel, Kyu Tae Kim, University of Cambridge, UNITED KINGDOM; Jong Guen Lee, Bryan D. Quay, Dom A. Santavicca, Pennsylvania State University, USA Development of Commercial CFBC Boiler for Refuse Derived Fuel, Dowon Shun, Dal-Hee Bae, Jaehyeon Park, Seung Yong Lee, Korea Institute of Energy Research, KOREA Simplified Quantification of Tetrafluoroborate Ion in Flue Gas Desulfurization Effluent for Management of Fluorine Emission, Seiichi Ohyama, Hiroyuki Masaki,

TECHNICAL PROGRAM Shinji Yasuike, Kazuo Sato, Central Research Institute of Electric Power Industry, JAPAN Enhancing Thermal Efficiencies in Steam Power Plants by Utilizing the “W2” Prime Mover as Auxillary Equipment, Jerry F. Willis, Admiral Air, Inc., USA Comparing Efficiencies of the Steam Turbine Versus the “W2” Prime Mover, Jerry F. Willis, Admiral Air, Inc., USA Optimization of Fuel Properties with Utilization of Biodegradable Municipal Wastes for Combustion Units, Dagmar Juchelkova, Martina Hajkova, Helena Raclavska, VSB – Technical University Ostrava; L. Tararik, Frydecka skladka, a.s., CZECH REPUBLIC Sulfur Retention in the Ash During Combustion of Tuncbilek Briquettes, Ayfer Parlak, Mustafa Ozdingis, H. Köksal Mucuk, Selahaddin Anac, Turkish Coal Enterprises (TKI); Bekir Zühtü Uysal, Gazi University, TURKEY Development of an Analytical Solution for Jet Diffusion Flame Equations, F. N. Pereira, A. L. de Bortoli, N. R. Marcilio, UFRGS – Universidade Federal do Rio Grande do Sul, BRAZIL

POSTER SESSION 2 GASIFICATION Suitability of a South-African High Ash Content and High Ash Flow Temperature Coal Source for Entrained Flow Gasification, JC van Dyk, SASOL Technology, SOUTH AFRICA; R Stemmer, Corus Technology, THE NETHERLANDS Continuous Experiments of Hot Gas Desulfurization Process Using Zn-Based Solid Sorbents in a Pressurized Condition, Sung-Ho Jo, Young Cheol Park, Ho-Jung Ryu, Chang-Keun Yi, Korea Institute of Energy Research; Jeom-In Baek, Korea Electric Power Research Institute, KOREA Biomass Gasification in Dual Fluidized Reactors: Process Modeling Approach, Thanh D. B. Nguyen, Young-Il Lim, Hankyong National University; Byung-Ho Song, Kunsan National University; Won Yang, Uendo Lee, Young-Tai Choi, Korean Institute of Industrial Technology; Jae-Hun Song, Gi-Chul Myoung, Yong-Soo Cho, SeenTec Co., Ltd., KOREA The Role of O2/COG Ratio on Non-Catalytic Partial Oxidation Process of Coke Oven Gas, Haizhu Cheng, Sufang Song, Yongfa Zhang, Taiyuan University of Technology, CHINA Numerical Simulation of Carbon Catalytic Reforming Reactor, Haizhu Cheng, Yongfa Zhang, Sufang Song, Taiyuan University of Technology, CHINA A Study on the Temperature Profile and Heat Transfer Coefficients in Underground Coal Gasification Cavities, Sateesh Daggupati, Ramesh Naidu Mandapati, Sanjay M Mahajani, Anuradda Ganesh, Preeti Aghalayam, IIT Bombay; Pal AK., Sharma R.K., UCG Group, IRS, ONGC, INDIA Some Results of UCG Ex-Situ Trials from HBP Company Point of View, Peter Cicmanec, Frantisek Verbich, Jaroslav Belacek, Hornonitrianske bane Prievidza, a.s.; Karol Kostur, Technical University of Kosice, SLOVAK REPUBLIC

13

POSTER SESSION 3 SUSTAINABILITY AND ENVIRONMENT

POSTER SESSION 5 COAL-DERIVED PRODUCTS

The Coagulation in Electric Field of the Argillaceous Suspensions from the Wastewater Resulted in Coal Processing, Romulus Sarbu, Diana Marchis, University of Petrosani; Adrian Corui, SC AQUATIM SA Timisoara, ROMANIA

Reactions of Coal Structures with Polymers Leading to Hydrogen Production, Pavel Straka, Institute of Rock Structure and Mechanics, v.v.i, Academy of Sciences of the Czech Republic, CZECH REPUBLIC

Thar Coal Mining Challenges, Farid A. Malik, EMRConsult, PAKISTAN Submerged Sequencing-Batch Membrane Bioreactor to Treat the Coke Wastewater, Wen-Ying Li, Jingwen Wu, Baojie Zhao, Jie Feng, Taiyuan University of Technology, CHINA

POSTER SESSION 4 CARBON MANAGEMENT Options of CO2 Capture in Oxyfuel Coal Combustion Technologies, C. Clemente-Jul, J. Rodrigo -Naharro, Universidad Politécnica de Madrid, SPAIN Installation and Operation of 0.5 MW-Scale Dry Sorbent CO2 Capture Pilot Plant Integrated with Real Coal-Fired Power Plant, Chang-Keun Yi, Sung-Ho Jo, Young Cheol Park, Korea Institute of Energy Research; Chong Kul Ryu, Korea Electric Power Research Institute, KOREA Reaction Characteristics of Water Gas Shift Catalysts for SEWGS Process in a Bubbling Fluidized Bed, SeungYong Lee, Ho-Jung Ryu, Dowon Shun, Dal-Hee Bae, Korea Institute of Energy Research, KOREA Carbon Dioxide Capture of Flue Gases from Coal-Fired Power Plant Using Enzymes Originated Marine Life, Sihyun Lee, Soonkwan Jeong, Kyungsoo Lim, Jeonghwan Lim, Mari Vinoba, Korea Institute of Energy Research (KIER); Daehoon Kim, Korea University, KOREA Preparation and Characteristics of Formed Active Carbons for Natural Gas Storage, Grzegorz Łabojko, Aleksander Sobolewski, Institute for Chemical Processing of Coal; Leszek Czepirski, AGH – University of Science and Technology, POLAND The Kinetics of the CO 2 Reforming of CH 4 over Carbonaceous Catalyst, Fengbo Guo, Yongfa Zhang, Guojie Zhang, Bingmo Zhang, Taiyuan University of Technology, CHINA

Catalytic Performance in Fixed-Bed and Bubbling Fluidized-Bed Reactor during Fischer-Tropsch Synthesis on the Iron-Based Catalysts, Jong Wook Bae, Ki-Won Jun, Yun-Jo Lee, Kyoung Su Ha, Korea Research Institute of Chemical Technology (KRICT), KOREA Operation of Slurry Reactor for Fischer-Tropsch Synthesis, Ho-Tae Lee, Jung-Il Yang, Jung Hoon Yang, Dong-Hyun Chun, Hak-Joo Kim, Heon Jung, Korea Institute of Energy Research, KOREA Process Simulation of Steam Hydrogasification to Produce F-T Products and Electricity, Xiaoming Lu, Chan S Park, Joseph M Norbeck, University of California, Riverside, USA Further development of the PSRK Model for the Prediction of the Vapor-Liquid-Equilibria of Direct Coal Liquefaction System at High Temperatures and High Pressures, Xuefeng Mao, Shidong Shi, Wenbo Li, Zhennan Gao, China Coal Research Institute, CHINA Arsenic and Mercury Removal by Using Iron Humate Prepared from Turkish Coal Based Humic Acid, Hacer Dogan, Murat Koral, Tulay Inan, TUBITAK Marmara Research Center; Selahattin Anaç, Zeki Olgun, TKI(Turkish Coal Enterprises), TURKEY Heavy Metal Adsorption of Turkish Coal Based Humic Acid/Epoxy Composites, Emel Yildiz, Hacer Dogan, Murat Koral, Tulay Inan, TUBITAK Marmara Research Center; Selahattin Anaç, Zeki Olgun, TKI(Turkish Coal Enterprises), TURKEY Numerical Simulation of Syngas Production by Partial Oxidation of Coke Oven Gas under Non-Premixed Condition, Honggang Chen, Kai Zhang, North China Electric Power University; Hui Zhao, China University of Petroleum; Yongfa Zhang, Taiyuan University of Technology, CHINA Coal Supply Agreements and Competition, Değer Boden Akalın, Boden Law Office, TURKEY Chemicals from Turkish Lignites, Vedat Mihladiz, Turkish Coal Enterprises, TURKEY

Pd-Free Composite Membrane for Pre-Combustion Capture, Jung Hoon Park, Sung Il Jeon, Korea Institute of Energy Research; Young Jong Choi, Innowill Corporation, KOREA

Methane Cracking over De-ashed Coal Chars and the Effect of the De-ashing Conditions, Yizhuo Han, Yisheng Tan, Jiantao Zhao, Hongjuan Xie, Institute of Coal Chemistry, Chinese Academy of Science; Ling Wei, Graduate School of the Chinese Academy of Sciences; Jinhu Wu, Qingdao Institute of Bioenergy and Bioprocessing Technology, Chinese Academy of Sciences, CHINA; Dongke Zhang, The University of Western Australia, AUSTRALIA

Composite Ceramic Membrane for Oxygen Separation, Jung Hoon Park, Soo Hwan Son, Korea Institute of Energy Research; Jong Pyo Kim, Chungnam National University, KOREA

CeO2-K2O Promoted Co-Mo Sulfur-Tolerant Shift Catalyst for the Shift Reaction of CO in Coke Oven Gas, Yuqiong Zhao, Yongfa Zhang, Guojie Zhang, Taiyuan University of Technology, CHNA

Fixed-Bed Adsorption of Carbon Dioxide-Nitrogen Mixtures onto Activated Carbon: Characteristics of CO2 Adsorption and Modeling, Regina F.P.M. Moreira, Tirzhá L.P. Dantas, Federal University of Santa Catarina; Francisco Murilo T. Luna, Ivanildo J. Silva Jr., Diana C. S. de Azevedo, Federal University of Ceará, BRAZIL; Carlos A. Grande, Alírio E. Rodrigues, University of Porto, PORTUGAL

Effects of Preparation Conditions on Ru/Al2O3 Catalyst for Coal-Based Syngas Methanation Reaction, Liping Wang, Yongfa Zhang, Yaling Sun, Xianglan Li, Taiyuan University of Technology, CHINA

A Study on the Absorption Characteristics of CO2 with a Vortex Tube Type Absorber, Keun-Hee Han, Woo-Jung Ryu, Jong-Ho Park, Won-Kil Choi, Jong-Sub Lee, ByoungMoo Min, Korea Institute of Energy Research, KOREA

The Production of Organic Fertilizers from Göynük, Ilgin and Elbistan Lignites with H2SO4 Oxidation, M.Çöteli, A.Güntürk, A.Yavuz, A.Köker, G.Yıldırım,

14

TECHNICAL PROGRAM

S.Atlıhan, General Directorate of Mineral Research and Exploration, TURKEY Modeling, Scaleup and Optimization of Slurry Bubble Column Reactors for Fischer-Tropsch Synthesis, Laurent Sehabiague, Badie I. Morsi, University of Pittsburgh, USA Biogasification of Soma Lignite (A Preliminary Study), Mustafa Baysal, Yuda Yürüm, Sabanci University; Sedat İnan, TÜBİTAK Marmara Research Centre, TURKEY

POSTER SESSION 6 COAL SCIENCE Modeling of Coal Drying in a Pneumatic Dryer, Sihyun Lee, Sangdo Kim, Kyoungsoo Lim, Soonkwan Jeong, Youngjoon Rhim, Korea Institute of Energy Research (KIER), KOREA Characterization of Chars Made of Solvent Extracted Coals, Sihyun Lee, Jiho Yoo, Hokyung Choi, Sangdo Kim, Jeongwhan Lim, Thiruppathi Raja, Korea Institute of Energy Research (KIER); Wantaek Jo, Yonsei University, KOREA Upgrading of Low Rank Coal by Hybrid Flash Dryer, Sihyun Lee, Sangdo Kim, Hokyung Choi, Kyoungsoo Lim, Sangyoung Lee, Korea Institute of Energy Research (KIER), KOREA Drying Kinetics of Low Rank Coal in Multi-Chamber Fluidized Bed, Jaehyeon Park, Dowon Shun, Dal-Hee Bae, Sihyun Lee, Jeong Hak Seo, Korea Institute of Energy Research; Jaehyeok Park, Hanyang University, KOREA The Influence of the Temperature on Adsorption of SDS on Coals, Boleslav Taraba, Roman Marsalek, Ostrava University, CZECH REPUBLIC Assessment of Elemental Sulphur in Biodesulphurized Coals, Lenia Gonsalvesh, Stefan Marinov, Maya Stefanova, Institute of Organic Chemistry, Bulgarian Academy of Sciences, BULGARIA; Robert Carleer, Jan Yperman, Hasselt University, BELGIUM Study of Biodesulphurized High Sulphur Coals from Bulgaria, Stefan Marinov, Maya Stefanova, Lenia Gonsalvesh, Nadezda Kazakova, Institute of Organic Chemistry, Bulgarian Academy of Sciences; Veneta Groudeva, M.Iliev, University of Sofia; Petyo Gadjanov, Technical University of Sofia, BULGARIA; Robert Carleer, Jan Yperman, Hasselt University, BELGIUM Samples in the World Coal Quality Inventory - A USGS Compilation on Global Coal, Susan J. Tewalt, Harvey E. Belkin, U.S. Geological Survey, USA Model Structure of High Sulphur N.E. Region Indian Coals, Sunil Kumar Srivastava, Atma Ram Singh, Central Institute of Mining and Fuel Research, INDIA Influence of the Surface Treatment with O3 And NH3 on the Physical and Chemical Characteristics of Dried Low Rank Coal, Gi Bo Han, Yongseung Yun, Changsik Choi, Institute of Advanced Engineering, KOREA Energy, Natural Gas, Türkiye & Ankara, İbrahim Halil Kirsan, Başkent Doğalgaz Dağıtım A.Ş., TURKEY An Alternative Application to the Centrifugal Dryer at a Coal Preparation Plant, Ahmet Gitmez, Mustafa Yılmaz, Western Lignite Establishment (GLI), TURKEY The Evaluation of the Contributions to the Productivity of the Process Changes at Tuncbilek Coal Preparation Plant, Ahmet Gitmez, F. Zehra Taksuk, Fatih Albayrak, Western Lignite Establishment (GLI), TURKEY

The Dump Truck Requirement Planning Studies of Turkish Coal Corporation, Mustafa Ziypak, Turkish Coal Corporation, TURKEY Investigation of Radioactive Contents Soma Coals, İsmail Demir, İlgin Kurşun, İstanbul University, TURKEY Effect of Triboelectrostatic Separation on Coal Desulfurization and Deashing, Byoung-Gon Kim, HoSeok Jeon, Sang-Ho Back, Chong-Lyuck Park, Korea Institute of Geoscience and Mineral Resources (KIGAM), KOREA Remove of Ash and Sulfur Minerals from Coal by Triboelectrostatic Separation, Ho-Seok Jeon, ByoungGon Kim, Korea Institute of Geoscience and Mineral Resources (KIGAM); Woo-Zin Choi, The University of Suwon, KOREA Recovery of Valuable Metallic and Non-Metallic Minerals from Coal Mine Wastes, Sang-Bae Kim, SooBok Jeong, Ho-Seok Jeon, Chong-Lyuck Park, Korea Institute of Geoscience and Mineral Resources (KIGAM), KOREA Manufacture of Fired Clay Brick from Coal-Preparation Refuse and its Characteristics, Soo-Bok Jeong, Ho-Seok Jeon, Chong-Lyuck Park, Byoung-Gon Kim, Korea Institute of Geoscience and Mineral Resources (KIGAM), KOREA

15

GENERAL INFORMATION / PROCEEDINGS

2010 Session Chairs ADVISORY BOARD

PROCEEDINGS Orders can be placed online. Credit cards, checks and money orders are accepted. Shipping: CD-ROM (Vol. 14 - 25): $5.00 Domestic and International Paper proceedings (Vol. 1 - 13): Domestic - $5.00 International - $25.00

Year Edition and Cost 1st Annual - 1984 2nd Annual - 1985 3rd Annual - 1986 4th Annual - 1987 5th Annual - 1988 6th Annual - 1989 7th Annual - 1990 8th Annual - 1991 9th Annual - 1992 10th Annual - 1993 11th Annual - 1994 12th Annual - 1995 13th Annual - 1996 14th Annual - 1997* 15th Annual - 1998* 16th Annual - 1999* 17th Annual - 2000* 18th Annual - 2001* 19th Annual - 2002* 20th Annual - 2003* 21st Annual - 2004* 22nd Annual - 2005* 23rd Annual - 2006* 24th Annual - 2007* 25th Annual - 2008* 26th Annual - 2009* * CD-ROM

$30.00 $30.00 $30.00 $30.00 $40.00 $40.00 $50.00 $50.00 $50.00 $50.00 $50.00 $50.00 $50.00 $30.00 $30.00 $30.00 $30.00 $30.00 $30.00 $30.00 $30.00 $30.00 $30.00 $30.00 $30.00 $50.00

Robert Beck, Chair, The National Coal Council, Inc., USA Richard Winschel, Vice Chair, CONSOL Energy Inc., USA Richard Bajura, West Virginia University, USA Francois Botha, Illinois Clean Coal Institute, USA Francis Burke, CONSOL Energy Inc., USA Vann Bush, Gas Technology Institute, USA Tarunjit Butalia, The Ohio State University, USA Shiao-Hung Chiang, University of Pittsburgh, USA Dan Duellman, American Electric Power, USA Shannon Fraser, U.S. Dept. of Commerce, USA Evan Granite, U.S. DOE/NETL, USA Yizhou Han, Chinese Academy of Sciences, CHINA Gerald Holder, University of Pittsburgh, USA Jim Hower, University of Kentucky - CAER, USA Mike Jones, Univeristy of North Dakota EERC, USA Francis Lau, Synthesis Energy Systems, USA Ke Liu, National Institute of Clean & LowCarbon Energy, CHINA Chuck McConnell, Battelle, USA Robert Miller, Air Products & Chemicals, USA Kouichi Miura, Kyoto University, JAPAN Badie I. Morsi, University of Pittsburgh, USA Masakatsu Nomura, Osaka University, JAPAN Guven Onal, Istanbul Technical University, TURKEY Brenda Pierce, U.S. Geological Survey, USA Massood Ramezan, Leonardo Technologies, Inc. (LTI), USA Leslie Ruppert, U.S. Geological Survey, USA Richard Sakurovs, CSIRO, AUSTRALIA Alan Scaroni, Pennsylvania State University, USA Alan Singleton, Energy Technology Partners LLC, USA Chunshan Song, Pennsylvania State University, USA Gary Stiegel, U.S. DOE/ NETL, USA Johan van Dyk, SASOL, SOUTH AFRICA Frans Waanders, North-West University, SOUTH AFRICA Ting Wang, University of New Orleans, USA John Wheeldon, EPRI, USA

Combustion

John Wheeldon, National Carbon Capture Center, USA Evan Granite, U.S. DOE/NETL, USA Steven Carpenter, Marshall Miller & Associates, USA

Gasification

Gary Stiegel, U.S. DOE/NETL, USA Jenny Tennant, U.S. DOE/NETL, USA Johan van Dyk, SASOL, SOUTH AFRICA Ke Liu, NICE, CHINA Rohan Courtney, UCG Partnership, UNITED KINGDOM Ting Wang, University of New Orleans, USA

Sustainability and Environment

Jim Hower, University of Kentucky, USA Brenda Pierce, U.S. Geological Survey, USA

Carbon Management

Bob Miller, Air Products, USA Leslie Ruppert, U.S. Geological Survey, USA Richard Sakurovs, CSIRO, AUSTRALIA

Coal-Derived Products

Rachid Oukaci, Energy Technology Partners, LLC, USA Belma Demirel, Rentech, Inc., USA Dragomir Bukur, Texas A&M University at Qatar, QATAR

Coal Science

Jim Hower, University of Kentucky, USA B.K. Parekh, University of Kentucky, USA Frans Waanders, North-West University, SOUTH AFRICA Susan J. Tewalt, U.S. Geological Survey, USA Quentin Campbell, North-West University, SOUTH AFRICA Richard Winschel, CONSOL Energy Inc., USA Ashok K. Singh, Central Institute of Mining & Fuel Research, CSIR, INDIA

INTERNATIONAL VICE CHAIRS Duke Du Plessis, Alberta Energy Research Institute, CANADA Juan Jose Garcia, Vice Minster Energy, VENEZUELA Yizhou Han, Chinese Academy of Sciences, CHINA Hung-Taek Kim, Ajou University, SOUTH KOREA Evgeny Kuzmin, Moscow State Mining University, RUSSIA Bernd Meyer, Freiberg University Mining & Technology, GERMANY Kouichi Miura, Kyoto University, JAPAN Geoff Morrison, IEA Clean Coal Centre, UNITED KINGDOM Selahaddin Anac, Turkish Coal Enterprise (TKI), TURKEY Richard Sakurovs, CSIRO, AUSTRALIA Harry Schreurs, NOVEM, THE NETHERLANDS Marek Sciazko, Institute of Chemical Processing of Coal, POLAND Sunil Srivastava, Central Fuel Research Institute, INDIA Johan van Dyk, SASOL Technology, SOUTH AFRICA

ANNOUNCING:

TWENTY - EIGHTH ANNUAL INTERNATIONAL PITTSBURGH COAL CONFERENCE DAVID L. LAWRENCE CONVENTION CENTER PITTSBURGH, PA, USA SEPTEMBER 12 - 15, 2011

Abstracts must be submitted by March 1, 2011. Please forward paper title, intended topic area, authors, affiliations, contact information with valid email address and a one-page abstract to: Conference Secretary [email protected] Please visit the PCC WEBSITE: www.engr.pitt.edu/pcc

DEVELOPMENT OF OXYCOALTM TECHNOLOGY RESULTING FROM TESTING CONDUCTED AT DOOSAN POWER SYSTEMS’ CLEAN COMBUSTION TEST FACILITY (CCTF)

F D Fitzgerald and P Holland-Lloyd Doosan Babcock United Kingdom

Abstract Oxyfuel combustion technology is one of several Carbon Abatement Technologies (CATs) currently being developed. The technology offers a means of generating carbon dioxide rich flue gas requiring minimal treatment prior to sequestration or beneficial application. Doosan Power Systems are aiming to develop a competitive oxyfuel firing technology suitable for full plant application post 2015, and is taking a phased approach to the development and demonstration of oxyfuel technology. Doosan Power Systems is leading a number of collaborative projects that are investigating ‘Oxyfuel Combustion Fundamentals and Underpinning Technologies’, the ‘Demonstration of an Oxyfuel Combustion system’, the ‘Modelling and Testing of the 40 MWt OxyCoal™ Burner’, and the ‘Optimisation of Oxyfuel PF Power Plant for Transient Behaviour’. This paper outlines progress on the ‘Demonstration of an Oxyfuel Combustion system’ project, which is demonstrating an oxyfuel combustion system of a type and size (40MWt) applicable to new build and retrofit advanced supercritical boiler plant. Installation and commissioning are complete and testing is in progress. Preliminary results have shown safe and smooth transitions between air firing and oxyfuel operation, with economiser outlet CO2 concentrations greater than 85% v/v dry being achieved. Background To achieve the global target reduction in CO2 emissions of some 60%, compared to 1990 levels, by 2050 Carbon Capture and Storage (CCS) will be necessary. Carbon Abatement Technologies (CATs) will be required for retrofit to existing power plant and installation on new build power plant. Doosan Power Systems have taken a proactive approach to developing CATs, leading and supporting techno-economic studies on both Track 1 and Track 2 approaches of the UK DTI’s Carbon Abatement Technology Strategy

(1)

. The Track 1 approaches are available now and

reduce CO2 emissions per unit of electricity generated by means of improved cycle efficiency and biomass co-firing (i.e. carbon neutral generation). The Track 2 approach, CCS, achieves much larger reductions, of up to 95%, by means of oxyfuel combustion, Post Combustion Capture (PCC) or Pre-Combustion capture via Integrated Gasification Combined Cycle (IGCC) technologies. Doosan Power Systems and Doosan Heavy Industries are active in all three CCS approaches.

Oxyfuel Combustion The oxyfuel process is based on excluding the inert components (mainly nitrogen) of air from the combustion process. In oxyfuel combustion, nitrogen is largely absent from the flue gas, since the fuel is combusted with a mixture of nearly pure oxygen (~95% O2 - separated from air in an air separation unit {ASU}) and CO2 rich recycled flue gas. It is recognised as a leading Carbon Capture technology for new and retrofit power plant. Figure 1 presents the comparison between conventional air and oxyfuel combustion on a utility boiler.

© Doosan Power Systems 2010 Figure 1: Schematic Diagram of Oxyfuel versus Air Firing The combustion temperature of a fuel with high purity oxygen is too high (3500°C) for any conventional combustion process used in utilities, therefore the temperature is controlled by recycling a portion of the flue gas to the furnace. The impact of varying flue gas recycle (FGR) rate on oxyfuel combustion efficiency, boiler thermal performance and pollutant formation is of major interest to manufacturers and utility operators alike. Demonstration of an Oxyfuel Combustion System Doosan Power Systems is leading a number of UK Government supported collaborative projects that are developing oxyfuel combustion technology, including the £7.4M UK Department of Energy and Climate Change (DECC) Hydrogen Fuel Cells and Carbon Abatement Technologies (HFCCAT) Demonstration Programme collaborative project: Demonstration of an Oxyfuel Combustion System (OxyCoal 2), which is demonstrating the Doosan Power Systems’ 40MWt OxyCoalTM burner. The project partners comprise: Doosan Power Systems (Lead), Imperial College London and the University of Nottingham.

Scottish and Southern Energy plc, Air Products plc,

ScottishPower Limited, E.ON UK plc, EDF Energy plc, Drax Power Limited, DONG Energy

Power, UK Coal plc and Vattenfall AB are sponsor participants, with Scottish and Southern Energy plc acting as prime sponsor. The phases of the project comprise development of a purpose designed oxyfuel demonstration facility (Doosan Power Systems’ 90MWt Clean Combustion Test Facility (CCTF)) including detailed design and HAZOP study, installation and commissioning of the oxyfuel equipment and design, manufacture and parametric testing of the 40MWt OxyCoal™ burner.

The

specific test objectives are demonstration of safe operation and successful performance of the full-scale 40MWt OxyCoal™ burner firing at conditions pertinent to the application of an oxyfuel combustion process in a utility power plant, in terms of flame stability, NOX, flame shape and heat transfer characteristics over a reasonable operational envelope with respect to start-up, turndown, shutdown and the transition between air and oxyfuel firing. Clean Combustion Test Facility The 90MWt Clean Combustion Test Facility (CCTF) is designed primarily for the development of burners for fossil fuel firing applications and is one of the largest and most modern single burner test rigs in the world. The CCTF was commissioned in 2000 and has been designed to enable burners to be developed, optimised and performance tested at fullscale prior to application in industrial plant. The front of the CCTF furnace is shown in Figure 2.

© Doosan Power Systems 2010 Figure 2: Clean Combustion Test Facility

The main component is a horizontal, water-jacketed test furnace that is partly lined with high temperature refractory. The furnace is 17 m long and 5.5 m square in section. Observation ports are arranged along one side wall of the furnace on the burner centerline and flame probing access ports, are located on the opposite side wall.

Photographic and video

equipment are fitted for flame observation, monitoring and image recording. An adaptable windbox is fitted at the front of the furnace to accommodate single test burners with throat diameters up to a maximum of 2 m. The facility is able to fire a wide variety of fuels; coals (ranging from low volatile semianthracites – 8% volatile matter to high volatile bituminous coals with 50% volatile matter, up to 35% ash, and up to 20% inherent moisture), heavy fuel oil and natural gas. A storage silo, loss-in-weight feeder and pneumatic transport system is fitted for the supply of pulverised coal to the burner. The system has a maximum feed rate capability of 12 to 14 tonnes per hour depending on the bulk density of the material. The draught plant consists of forced draught (FD), transport air, primary air (PA), core air and induced draught (ID) fans and blowers. Gas-fired airheaters are installed to raise fuel transport stream and combustion air temperatures to plant representative values. A multi-cyclone grit collection system is fitted for cleaning of flue gases prior to the ID fan. A tri-drum boiler with superheater and economiser is installed to cool the flue gases from approximately 1200°C at the furnace exit to approximately 230°C at the grit collector inlet. Two-stage combustion has been implemented on the CCTF by means of installing overfire air ports to the furnace sections to allow the testing of burners operating under staged combustion conditions. OxyCoalTM Clean Combustion Test Facility To allow testing of an OxyCoalTM burner in oxyfuel firing mode the CCTF has been upgraded with the addition of equipment and instrumentation required for oxyfuel firing.

The

conversion to oxyfuel included the addition of an oxygen storage facility comprising three liquid oxygen (LOX) storage tanks each with a capacity of approximately 50 tonnes, and eight ambient vaporisers to supply gaseous oxygen for injection into the primary and secondary flue gas recycle (FGR) streams. The primary and secondary FGR streams replace the primary air and main combustion air respectively, each having a dedicated fan.

A

transport FGR stream replaces the transport air stream. The transport and primary FGR streams have additional flue gas cooling systems fitted to condense moisture, followed by in-

duct heating as a means to mitigate against PF feeding problems in a high moisture flue gas. A proportion of the secondary FGR stream can be redirected to an overfire FGR system for two-stage combustion.

A schematic diagram of the CCTF, including the oxyfuel

configuration is presented in Figure 3.

© Doosan Power Systems 2010 Figure 3: Schematic Diagram of CCTF OxyCoal™ Configuration The construction and installation of the oxyfuel equipment have been completed, with OxyCoalTM burner testing being carried out during late 2009 and early 2010. OxyCoal™ Burner The first generation of oxyfuel burners will most likely be based on current low NOX air-fired burner technology in order to ensure compatibility with existing plant for retrofit purposes. Additionally, a uniform ‘simulated air’ flue gas composition and a design flue gas recycle rate, based on the consideration of radiant and convective heat transfer being theoretically similar to air firing, is a logical first operating condition. With these points in mind, a 40 MWt OxyCoalTM burner was designed to best exploit a range of potential operating conditions for both oxyfuel and air firing. For oxyfuel operation the volumetric flow rate and molar oxygen content of the primary gas is maintained as per air firing. The design overall stoichiometric ratio is 1.2 and the flue gas recycle rate has been chosen to give an adiabatic flame temperature equivalent to air operation.

Burner Modelling Prior to testing, computational fluid dynamics (CFD) modelling was undertaken to support the burner design process. The CFD model of the 40 MWt OxyCoal™ burner extends from the windbox and primary gas inlets to the outlet of the CCTF furnace, and consists of 3.6 million cells. An image of the modelled geometry is shown in Figure 4. Two design operating conditions were considered for this burner arrangement: 1. Air Firing, Existing Plant. 2. Oxyfuel Firing.

© Doosan Power Systems 2010 Figure 4: Modelled Geometry of 40 MWt OxyCoal™ Burner in CCTF Figure 5 shows the predicted gas temperature profiles in the furnace. Although Furnace Exit Gas Temperature (FEGT) is predicted to be broadly similar under oxyfuel and air firing combustion, the peak temperature is predicted to be higher under the particular oxyfuel firing conditions modelled. Despite the flame being well rooted to the flame holder, it is seen that oxyfuel firing is predicted to lead to a narrower flame. This may be due to the reduced momentum resulting from the lower volumetric flow rate of FGR when operating in oxyfuel mode, arising from the higher density of CO2 compared with N2 (1.8 kg/m3 versus 1.14 kg/m3 at 25°C).

Air Firing

Oxyfuel Firing

© Doosan Power Systems 2010 Figure 5: Predicted Gas Temperature (°C) under Air and Oxyfuel Operation Figure 6 presents the predicted CO concentrations under air and oxyfuel operation. It is seen that local CO concentrations within the flame are also predicted to be much higher under oxyfuel firing, which agrees with the findings of a number of experimental studies (2, 3).

Air Firing

Oxyfuel Firing

© Doosan Power Systems 2010 Figure 6: Predicted CO Concentration (ppm) under Air and Oxyfuel Operation

OxyCoal™ Burner Commissioning and Testing The OxyCoalTM burner is mounted in a windbox which supplies either combustion air or secondary FGR to the burner outer annuli.

Combustion air enters the windbox

perpendicularly from the right side wall, looking at the front of the windbox/burner back plate, and the FGR enters perpendicularly from the opposing wall (Figure 3). Before any testing of the burner could commence, the existing and new equipment related to the oxyfuel modifications required cold and hot commissioning.

As part of the cold

commissioning activities, the OxyCoalTM burner underwent isothermal testing to determine swirl numbers and pressure loss factors of the burner outer annuli. The hot commissioning was split into three parts. The first part successfully proved that the test facility was still capable of operating under air firing conditions and provided baseline air firing OxyCoalTM burner, furnace and boiler performance results. Figure 7 shows a typical OxyCoalTM burner air firing flame. The flame is anchored to the burner, within the quarl, and has a bright root.

© Doosan Power Systems 2010 Figure 7: Typical Coal Air firing Flame on OxyCoalTM Burner The second part successfully proved the test facility in oxyfuel oil firing operation. The safe transition from oil firing on air to oxyfuel and establishment of a stable oxyfuel oil flame were demonstrated. The third and final part successfully proved the test facility under oxyfuel operation. The safe transition from coal firing on air to oxyfuel and establishment of a stable oxyfuel flame were demonstrated. Figure 8 shows a typical OxyCoalTM burner oxyfuel flame. The oxyfuel flame

is visually practically indistinguishable from the flame obtained during air firing, again exhibiting a bright root that is strongly anchored to the burner.

© Doosan Power Systems 2010 Figure 8: Typical Coal Oxyfuel Firing Flame on OxyCoalTM Burner The testing of the OxyCoalTM burner was split into two phases. Phase 1 concentrated on ‘burner proving’ in terms of burner characterisation, flame stability and control/operability, while Phase 2 aimed to characterise the achievable performance of the burner. Phase 1 of testing was successfully completed in January 2010 with the following main outcomes: Three air to oxyfuel transition methods were investigated: 1. Transition from Oxyfuel Oil to Oxyfuel Coal Combustion This is the preferred transition method on the CCTF, since it maximises the test time on coal. Oil air firing is switched to oxyfuel oil firing: secondary air is switched to SFGR followed by the introduction of coal with primary air. The coal flow rate is increased, with corresponding reductions in oil flow until coal is at full load and oil support is off. Primary air is then switched to PFGR. This method of transition on the CCTF is reasonably quick to complete and is tolerant to fairly large fluctuations in fuel flow, without affecting the flame stability. 2. Transition from air, starting with SFGR stream, to Oxyfuel This approach starts with coal air firing (primary and secondary air), with a low level of oil support at reduced load. Secondary air is switched to SFGR, followed by increases in coal flow, with corresponding reductions in oil flow until coal is at full load and oil support is off. Primary air is then switched to PFGR. A longer period is required to complete the transition using this method, since small changes in coal flow

during the coal load increments have a significant impact on economiser outlet oxygen and therefore SFGR oxygen concentration. As a result, smaller and slower increments in coal load are required, which consume more test coal and, therefore, reduce the coal test time available. 3. Transition from air, starting with PFGR stream, to Oxyfuel This approach also starts with reduced load coal air firing (primary and secondary air) with a low level of oil support. Primary air is switched to PFGR. Coal flow is increased, with corresponding reductions in oil flow until coal is at full load and oil support is off. The secondary air is then switched to SFGR. This method proved to be the most difficult to perform, since even the smallest possible coal load increments had an appreciable impact on PFGR oxygen concentration.

Flame stability was

adversely affected, since PFGR is supplied at the flame root and provides oxygen at the initial combustion stage. Although a preferred method of CCTF air to oxyfuel transition has been selected, to maximise coal oxyfuel test time, all three of the above methods have been safely demonstrated and are perceived to be applicable to oxyfuel boiler plant. The sensitivity of Methods 2 and 3 to small changes in coal flow rate on the CCTF’s single burner is likely to be less of a concern in multi-burner plant. The importance of air ingress has been investigated. Initial testing, with substantial air ingress, resulted in a CO2 concentration of around 50% v/v dry at full load OxyCoalTM operation. Air ingress minimisation resulted in CO2 concentrations of around 85% v/v dry at full load OxyCoalTM operation. At the time of writing, Phase 2 of testing is in progress and is providing positive results. The remaining test programme aims to establish an operational envelope for the OxyCoal™ burner design, investigating the effects of parameters including FGR rate and stoichiometric ratio on flame stability, heat release and flue gas composition. Conclusions •

Modelling of a full scale OxyCoal™ burner in air and oxyfuel firing modes shows acceptable flame characteristics and emissions performance for both conditions.

•

The 90 MWt CCTF oxyfuel conversion has been successfully completed, all equipment has been commissioned and testing is in progress.

•

Safe and smooth transitions between air firing and oxyfuel operation have been demonstrated.

•

Economiser outlet CO2 concentrations greater than 85% v/v dry are achievable during oxyfuel operation.

•

Successful demonstration of the OxyCoalTM burner will form the foundation for the development of an OxyCoalTM boiler reference design.

References [1]

A Strategy for Developing Carbon Abatement Technolgies for Fossil Fuel Use, Department for Trade & Industry, DTI/Pub URN 05/844, June 2005

[2]

Goh, B, ‘E.ON-UK’s Pilot-Scale Oxy-Fuel Combustion Experience: Development, Testing and Modelling’, IEA GHG International Oxy-Combustion Network, Third Workshop, Yokohama, Japan, March 2008

[3]

Hjärtstam, S et al., ‘Combustion Characteristics of Lignite-fired Oxy-fuel Flames’, presented at the 32nd International Technical Conference on Coal Utilization and Fuel Systems, Clearwater, Florida, USA, 2007

Acknowledgements The authors gratefully acknowledge the grant funding provided by the UK Department of Energy and Climate Change (DECC) and the technical and financial contributions made by the project collaborators: University of Nottingham, Imperial College London, Scottish and Southern Energy plc (Prime Sponsor), Air Products plc, DONG Energy Power, Drax Power Limited, EDF Energy plc, EON UK plc, ScottishPower Limited, UK Coal plc and Vattenfall AB. The authors also acknowledge the support of the members of Doosan Power Systems’ oxyfuel research and development team.

2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010

PROGRAM TOPIC: COMBUSTION NOX REBURNING IN OXY-FUEL COMBUSTION ­ AN EXPERIMENTAL INVESTIGATION

Daniel Kühnemuth, Fredrik Normann, Klas Andersson, Filip Johnsson, Bo Leckner Division of Energy Technology, Department of Energy and Environment, Chalmers University of Technology, SE – 412 96 Göteborg, Sweden This work investigates the reburning reduction of nitric oxide (NO) in a 100 kW propane-fired oxyfuel flame. The conducted experiments include a comprehensive parameter study: NO was injected into the recycled flue-gas, the inlet oxygen concentration was varied between 25 and 37 vol. % and the stoichiometric ratios at the burner inlet ranged from 0.7 and 1.15. The respective influence of inlet oxygen concentration and burner stoichiometry on once-through and total reduction of NO was measured. Furthermore, concentration and temperature in the furnace were mapped to identify important differences between oxy and air-fired conditions. The furnace measurements show that the peak concentration of carbon monoxide may be more than twice as high as in air-fired conditions. The formation paths of CO and its influence on the NOx chemistry are therefore discussed. The results of the parameter study show that reburning is favored by decreased burner stoichiometry. The effect of inlet oxygen concentration on once-through NO reduction is of minor importance. Changes in stoichiometry and oxygen inlet concentration are associated with changes in recycle ratio. The influence of the recycle ratio on the NO reduction is of great importance and is investigated as separate parameter. Keywords: Reburning, Oxy-fuel, NOx, CO2/O2.

Introduction Oxy-fuel combustion is a promising near-term technology for CO2 capture in power plants. Presently, this technology is in a stage where the first pilot-scale plants (tens of MW) are in operation and demo-scale plants (hundreds of MW) are being planned for commissioning (Toftegaard et al. 2010). Also in oxy-fuel power plants combustion of coal will be associated with

1

NOx formation, and the presence of NOx may have to be considered both in the gas entering the flue gas treatment and in the gas emitted to the atmosphere, as well as in the storage gas (Normann et al. 2009). Thus, it is important to examine the possibilities for NOx removal. The changed conditions within the oxy-fuel process compared to air-combustion present new opportunities for NOx control, both by methods applied in the furnace (e.g. reburning of recycled NOx (Andersson et al. 2008, Mackrory 2008, Okazaki and Ando 1997) and high-temperature NOx reduction (Normann et al. 2008)), as well as during the flue-gas treatment (e.g. absorption of NO2 in water (White et al. 2010, Kühnemuth et al. 2008)). Further work is needed to identify strategies to be applied in commercialscale oxy-fuel plants.

In oxy-fuel combustion, pure oxygen is added to a stream of flue gas recirculated to the furnace to control the combustion temperature (among other parameters). NO, which is also contained in the flue gases is recirculated to the combustion zone and a reburning situation is inherent without any special arrangements. The present work investigates the reduction of recycled NO by reburning reactions in oxy-fuel combustion. The reburning mechanism is a route for reduction of NOx by hydrocarbon radicals, whose global reaction can be expressed as, 𝑁𝑂 + 𝐶𝐻𝑖 ↔ HCN + ⋯

(1)

Reduction of NO by reburning has been extensively investigated for air-fired conditions (see e.g. Dagaut and Lecomte 2003, Frassoldati, Faravelli and Ranzi 2003, Bilbao, Alzueta and Millera 1995) with the general conclusion that fuel-mixing, temperature, stoichiometry, and gas residence-time need to be optimized to successfully apply reburning. Two main factors change the conditions for reburning in oxy-fuel combustion: the recycle of NO-containing flue gas, which increases the total residence time of NO in the flame zone, and the changed combustion environment, which affects combustion and nitrogen chemistry.

Previous experimental work has shown that oxy-coal combustion reduces the emission of NOx with around 70%(in mg/MJ of fuel supplied), compared to a corresponding air-fired unit (Andersson et al. 2008). The main part of the reduction achieved is commonly considered to be caused by reburning of NO, recycled with the flue gas to the flame (Andersson et al. 2008, Mackrory 2008, Okazaki and Ando 1997). The most obvious indication of the change in combustion chemistry in oxy-fuel combustion is the elevated CO concentration in the flame, which is reported in several experimental studies (Hjärtstam et al. 2009, Dhungel 2009, Tan et al. 2006, Liu, Zailani and Gibbs 2005). However, literature disagrees on the reasons to the increased CO level and its importance to nitrogen 2

chemistry. Three reaction routes for CO formation have been discussed. Firstly, for coal-fired experiments, char gasification by CO2, Reaction (2), is commonly mentioned (Zhang et al. 2010, Li et al. 2009, Toporov et al. 2008), 𝐶 + 𝐶𝑂2 ↔ 2CO + 𝑂

(2)

However, there is also work (Mackrory 2008) indicating this reaction to be of minor importance for CO formation. Instead, Mackrory suggests that CO is formed by thermal dissociation of CO2 at temperatures above 1500 K, Reaction 3, 𝐶𝑂2 ↔ CO + 𝑂

(3)

However, thermal dissociation of CO2 is commonly claimed to be of importance at much higher temperatures (the conversion of CO2 is around 1% at 2000K) (Jin et al. 2006, Fan et al. 2003, Itoh et al. 1993, Nigara 1986), but it is concluded that the conversion can be significantly improved by removal of the final reaction products CO and O2.

Another formation route of CO is the water-gas-shift reaction, Reaction (4), as e.g. suggested by Dhungel et al. (Dhungel 2009) 𝐶𝑂 + H2 O ↔ CO2 + 𝐻2

(4)

Reaction (5), which is an elementary reaction included in the mechanism of the water-gas shift reaction, is shown to be reversed to produce CO, when CO2 is added to an air-fired flame (Liu et al. 2001). For the high CO2 concentrations in oxy-fuel combustion, similar conclusions have been made, e.g. (Glarborg and Bentzen 2008) CO + 𝑂𝐻 ↔ 𝐶𝑂2 + H

(5)

Two possible routes have been proposed for the impact of elevated CO concentration in oxy-fuel combustion on the nitrogen chemistry: an increased direct catalytic effect on the heterogeneous reaction of NO and char (Dhungel 2009), as earlier observed for air combustion (Levy et al. 1981), and an indirect effect on the radical pool involved in the nitrogen chemistry. The latter effect is explained by the impact of formation as well as oxidation of CO by Reaction (5) on the critical chain branching Reaction (6) (Mendiara and Glarborg 2009b, Normann et al. 2010), 𝐻 + O2 → OH + O

(6) 3

the resulting change in the radical pool under oxy-fuel conditions leads to an altered oxidation of hydrocarbon radicals (Normann et al. 2010), which are the most important reburning agents according to Reaction (1). Furthermore, changes in the radical pool may have an important influence on the oxidation of intermediate nitrogen species in the flame. For instance, Mendiara and Glarborg (Mendiara and Glarborg 2009a) mention that the oxidation behavior of NH3, is significantly influenced by the OH/H ratio, and Giménez-López et al. (Giménez-López et al. 2010) make similar conclusions regarding the oxidation of HCN. Thus, the elevated in-flame concentration of CO may have several in-direct effects on the nitrogen chemistry in oxy-fuel combustion; in this work the homogenous reactions are captured while heterogeneous effects are omitted. The practical implication of the changed combustion chemistry for the reduction of recycled NO during oxy-fuel combustion has been investigated recently. Commonly, experiments are conducted with coal, which contains fuel-bound nitrogen that contributes to the formation of NO and, hence, do not permit separating reburning from NO formation. In such experiments the emission of NO usually decreases with increasing oxygen concentration in the oxidizer (Liu and Okazaki 2003, Croiset and Thambimuthu 2001, Hu et al. 2000). Reduction of the stoichiometric ratio at the burner inlet is reported to be an effective measure for controlling NOx emission also in oxy-fuel combustion (Mackrory 2008), (Liu et al. 2005). It is important to note that both inlet oxygen concentration and burner stoichiometry affect combustion temperature, and that the NOx emission usually increases with temperature (Normann et al. 2010, Hu et al. 2000). Mendiara and Glarborg performed benchscale experiments focused on reburning of NO during oxy-methane combustion, thus excluding the effects of fuel-bound nitrogen. Their once-through experiments showed that oxy-fuel combustion has higher NO-reduction efficiency under fuel-lean conditions than air firing, but that the difference is negligible under fuel-rich conditions. The aim of the present investigation is to complement the present knowledge on reburning in oxyfuel combustion. A parameter study is performed to analyze the sensitivity of reburning reduction to changes in stoichiometric ratio of the burner, inlet oxygen concentration and recycle ratio. All results are compared to air-fired conditions.

Experiments The experiments were performed in a 100 kWth oxy-fuel combustion test unit, shown in Figure 1. The unit can operate in air or oxy-fuel combustion mode, the latter employing dry flue-gas recycling. The furnace is a cylindrical, top-fired, combustion chamber measuring 800 mm in inner diameter and 2400 mm in height. The swirl burner is equipped with two oxidant registers: a primary register

4

with guide vanes of an angle of 45° and a secondary oxidizer-register with 15° angle. A bluff body stabilizes the flame.

Figure 1. Schematic of the 100 kWth oxy-fuel test facility, including the ports for sampling (R2-R5; L1-L3) and NO injection. Measurement Ports R2 – R5 are located at 215, 384, 553 and 800 mm from the burner. Gases were extracted from the flame zone by a water-cooled gas-sampling probe presented in Figure 2. The gas-sampling probe consists of a heated inner tube with controlled temperature to prevent condensation inside the probe and an outer protecting water-cooled jacket with a diameter of 45 mm. The gas is sucked through a 4 mm thick ceramic filter at the tip of the probe. The extracted gases are transported through hot sampling lines (200°C) to gas analysis. Cold (ambient temperature) measurements of NO, O2, CO2, CO were performed with on-line gas analyzers, specified in Table 1. Gas temperature in the furnace was measured with a water-cooled suction pyrometer with a Type-B thermocouple, shielded by a ceramic tube to prevent radiative heat loss from the thermocouple junction, Figure 3.

5

Figure 2. Gas sampling probe.

1900 mm Cooling water outlet

A Thermocouple Type B: D=3mm

Suction inlet flue gas: D=6mm

200mm Cooling water inlet

A

Section A-A

Ø45,0/41,0mm Ø30,0/26,0mm Ø17,2/13,0mm 10,0x1,0 mm Ø13,0mm

Figure 3. Suction Pyrometer

Table 1. Gas analyzers Manufacturer/model NO O2 CO2 CO

Measurement principle Chemoluminescence

Range

Detection limit

Span gas

0 - 0.1 %

≤ 1 %*

417 ppm

Fisher Rosemount (NGA 2000)

Paramagnetic

0 - 25 %

≤ 1 %*

9.00 %

Binos (100 2M)

Paramagnetic

0 - 50 %

≤ 1 %*

9.00 %

Sick Maihak (Sidor)

NDIR

0 - 20 %

≤ 1 %*

18.00 %

Binos (100 2M) Sick Maihak (S 710)

TC NDIR

0 - 100 %

≤ 2 %*

90.00 %

0 - 1.0 %

≤ 1 %*

0.90 %

Sick Maihak (S 710)

NDIR

0 - 20 %

≤ 1 %*

9.00 %

ECO (CLD 700 EL)

* Percent of the full range

Propane was used as fuel in all experiments (Table 2), to exclude heterogeneous effects. The fuel input was 81 kW. Two stoichiometric ratios are defined: the stoichiometric ratio of the burner (λburner), and the global stoichiometric ratio (λglobal). The global stoichiometric ratio was kept at 1.15 6

in all tests by injecting pure oxygen for burn-out (comparable to over-fire air) in Port R5 (Figure 1), 800 mm downstream of the burner inlet. Thereby, a region with adjustable stoichiometry is created between the burner and Port R5, while the burnout of the fuel was secured, and accumulation of CO and hydrocarbons in the recycle loop was avoided. Table 2. Fuel composition Fuel analysis Propane

98.0

mol %

Butane

1.1

mol %

Ethane

0.7

mol %

Pentane

CO + C_s (R3) C_s + O2 => O2_s+ C(b) (R4) O2_s + 2C(b) => C_s + CO2 CO2 gasification reaction: (R5) C_s + CO2 => CO + O_s + C(b) Steam gasification reaction: (R6) C_s + H2O => H2 + O_s + C(b)

A (g/cm2 s)

E (kJ/mol)

3.3E+15 1.0E+08 9.5E+13 1.0E+08

167.4 0. 142.3 0.

3.60E+15

251.0

4.35E+14

222.8

In order to compare predictions and observations, char particle temperatures were computed for the size range 10 - 200 μm. The SKIPPY simulations used GRIMECH 3.0 [19] without nitrogen reactions and used the heterogeneous reactions listed in Table 2, with the rate parameters of R5 and R6 varied to assess the importance of these reactions. Internal specific surface area was held

constant for all particle sizes and O2 concentrations; a value of 10 m2/g was used to crudely reproduce temperatures for 100 μm particles burning in 12% O2. At this condition, particle temperatures are around 1900 K, which is relatively low to expect appreciable effect of the high activation energy gasification reactions [13-14]. Further simulation input parameters include bulk density = 560 kg/m3, tortuosity = 5, void fraction = 0.4, particle thermal conductivity = 1.33 W/m·K, emissivity = 0.8, wall temperature (for radiative heat transfer) = 500 K, pressure = 101 kPa and gas temperature = 1690 K.

3.2. Simplified Char Consumption Model Kinetic parameters for calculating the rate of char consumption can be obtained from measured data and an appropriate char consumption model. In the case of optical measurements on individually burning char particles carried by a hot gas flow, the char consumption process is governed by mass and heat transfer to and from the particle, as well as chemical reactions, both on the particle surface and in the gas phase. Mathematical models of considerable complexity can certainly be formulated to fully describe the overall process. However, using these models to estimate the rate parameters of heterogeneous reactions (a) requires accurate specification of imperfectly known model details (such as gas phase reaction rates, transport properties, pore structure, etc.), and (b) yields kinetics rate parameters that can produce poor predictions if utilized directly with simplified models of CFD codes. For those reasons, we compare measured temperature-size data with predictions from a relatively simple burnout model, as outlined below. A simplified model, similar to those typically employed in CFD codes, was derived from the instantaneous energy balance on a homogeneous, chemically reacting, spherical particle. Assuming negligible effect of reactions in the boundary layer, average gas-mixture properties, and the overall char reaction C + (1+ψ)/2 O2 → ψ CO2 + (1−ψ) CO, the instantaneous energy balance is given by (see [25] for more detail):

27th Annual International Pittsburgh Coal Conference, Istanbul, Turkey, Oct. 11-14, 2010

5 d p  p c vp dTp 6

dz

   (Tp4  Tw4 ) 

2   2  (Tp  Tg )  q h d p  e / 2  1

The left-hand term in this equation represents the thermal inertia of the particles, with particle diameter dp (m), particle bulk density ρp (kg/m3), specific heat c (J/kg·K), particle speed vp (m/s), surface temperature Tp (K), and spatial coordinate z (m). For the current set of measurements, the particle temperatures are nearly independent of the distance from the burner face, allowing this term to be neglected. The first term on the right represents radiative heat loss, with char emissivity ε and wall temperature Tw (K). The second term on the right represents convective heat loss of the particle, with mixture thermal conductivity λ (W/m·K), free-stream gas temperature Tg (K) and κ = (−q·dp/λ) ∑iνi cg,i, which characterizes the heat transfer correction due to Stefan flow [25] (νi are stoichiometric coefficients, with νO2 = (1+ψ)/2, νCO = −(1−ψ), and νCO2 = −ψ; cg,i are the corresponding specific heat capacities (J/kg·K)). The chemical heat release is represented by the far right-hand term, where q denotes the overall burning rate per unit external surface area (kg/m2·s), and Δh = (1−ψ) ΔhCO + ψ ΔhCO2 is the overall heat of reaction (J/kg). The CO2/CO production ratio at the char particle surface is modeled as CO2/CO ≡ ψ/(1−ψ) = 0.02 p(O2,s)0.21 exp(3070/Tp) [24], where p(O2,s) is the oxygen partial pressure at the particle surface (atm). p(O2,s) follows from solution of the gas-phase diffusion equation [25] as p(O2,s)/p = γ + [p(O2,∞)/p − γ] exp[−q/(γ·kd·p)], where γ = −(1+ψ)/(1−ψ), p(O2,∞) is the free-stream oxygen partial pressure (atm), kd = 2·WC·DO2,mix/ (dp·R·Tf·νO2) is the oxygen mass transfer coefficient (kg/m2·s·atm) with WC = 12 g/mol, DO2,mix is the mixture-averaged oxygen diffusion coefficient (m2/s), R = 8.3144 J/mol·K is the universal gas constant, and Tf = (Tp + Tg)/2 is the film temperature (K). Gas properties (λ, DO2,mix, and cg) were calculated for T = Tf, and the free-stream gas composition as outlined in [26] (κ was approximated as κ ≈ −q·dp·cg·WC/(γ·λ·νO2)). The surface-specific burning rate q, is typically expressed as q  k s (Tp ) pon2 ,s

where ks(Tp) = WC·A·exp(−E/RTp) in units of kg/m2·s·atmn, n is the reaction order, and the preexponential factor A is expressed in units of kmol/m2·s. This is the “nth-order Arrhenius” expression of global char kinetics. For Zone-II combustion conditions, as are typically found for pulverized coal combustion, the diffusion of oxygen through the boundary layer has some influence on the burning rate of the particle, but the influence is not controlling the rate. However, the oxygen diffusion profile needs to be accounted for when calculating the surface concentration of oxygen. For the study presented here, temperature-size characteristics were calculated with the described model. The required rate parameters A and E to specify q were determined for three reaction orders n = 1, 0.5 and 0.1 such that observed temperatures of 100 μm particles burning in N2 with 12% and 36% O2 in the free stream were reproduced. Particle temperatures were then computed for sizes from 10 to 200 μm for the experimental conditions studied (either CO2 or N2 diluent, 16% H2O and three O2 bulk concentrations (12%, 24%, 36%)). A gas temperature Tg = 1690 K, radiative boundary Tw = 500 K, and char particle emissivity ε = 0.8 were assumed. The obtained rate parameters are summarized in Table 3. Table 3 Estimated rate parameters for North Antelope and Utah Skyline for the simplified steady-state burning model. Fixed n = 1 0.5 0.1

North Anthelope: A E (kmol/m2·s) (kJ/mol) 0.44 0.00 0.67 36.50 1.70 76.26 Utah Skyline:

1 0.5 0.1

0.39 1.05 2.31

11.95 51.75 83.54

4. Results and Discussion 4.1. Single-film char consumption models Figure 3 shows the results from the simplified model calculations for different reaction orders together with measured data for char particles

27th Annual International Pittsburgh Coal Conference, Istanbul, Turkey, Oct. 11-14, 2010

6 burning in N2 diluent. Temperatures are significantly higher for the subbituminous North Antelope char (top), for which considerably smaller sizes were also measured (most likely due to a stronger tendency to fragment than the bituminous Utah Skyline char (bottom)). The overall spread of temperatures of Utah Skyline char particles is larger for a given reactor condition, which may reflect a higher variability in mineral contents of the coal particles. For both North Antelope and Utah Skyline char, the calculated temperatures agree well near the “design” diameter 100 μm, but the diameter range for which predictions are reasonably good appears to be wider for reaction orders of n = 0.5 and n = 1. According to classical Thiele analysis, apparent reaction orders are constrained to lie between 0.5 and 1 for Zone II combustion [27,28], even though smaller reaction orders have been obtained in regression procedures for estimation of rate parameters (e.g. [25]) and ash-inhibition or similar processes that decrease reactivity with burnout can result in low effective reaction order [29]. The results shown in Figure 3 suggest that the predictions with n = 0.1 hold for sizes larger than 100 μm, and thus could be used in CFD models as long as the kinetics rates are not applied to particle sizes outside that range. Choosing a smaller particle size when estimating the rate parameters would have improved the predictions at smaller sizes, but at the cost of over-predicting temperatures for larger sizes. Given the observed nearly size-independent temperatures, it is recommended to keep reaction orders in the range between 0.5 and 1, to obtain more realistic profiles over the entire size range. However, the activation energies that are required when fitting the simulations to the data for high reaction orders can be very small (e.g. for n =1, E = 11.95 kJ/mol and 0 kJ/mol for North Antelope and Utah Skyline, respectively) and should be viewed as pure fitting parameters without practical meaning. But fitted parameters that strongly deviate from a physically realistic range indicate that the fitted model does not fully capture all relevant physical processes. Figure 4 shows predicted and measured particle temperatures as a function of particle diameter for char particles burning in CO2-dominated gas mix-

tures. Compared with the data for N2 diluent (Figure 3), the measured temperatures for both chars are lower in the CO2 environment, consistent with previous findings [15]. Interestingly, for each free-stream oxygen concentration the char temperatures show a much wider spread about an average temperature-size profile than for the N2 environment. This may be caused by differences in properties of in situ char produced in N2 and CO2, but further investigation is necessary to clarify this point.

Figure 3 Predicted and measured (symbols) particle surface temperatures for three free-stream O2 concentrations, 16% moisture and N2 bath gas. Solid lines n = 0.5, dotted lines n = 1, dashed lines n = 0.1.

As shown in Figure 4, predicted temperatures for reaction orders of n = 0.5 and n = 1 are too high for North Antelope char particles but agree reasonably well for Utah Skyline char. Although

27th Annual International Pittsburgh Coal Conference, Istanbul, Turkey, Oct. 11-14, 2010

7 the predictions are slightly better for n = 1, both models yield similarly good predictions. In interpreting the results for Utah Skyline char, two issues must be kept in mind. First, the measured temperatures for both N2 and CO2 experiments are ~ 150 K lower than for North Antelope char, thus potential contributions by gasification reactions are less significant for Utah skyline char. Second, due to the higher variability in measured temperatures (presumably a consequence of higher variability of mineral contents), average temperatures of 100 μm particles seem biased towards lower temperatures (see Figure 3). The increased spread of temperatures of char burning in a CO2 bath further obscures clear interpretation of discrepancies between model and experimental data. The over-prediction of temperatures for North Antelope char shown in Figure 4 clearly points to inadequacy of the simplified model. From a CFD modeling perspective, several modifications of the model without the need to include time-consuming calculations of temperature and multi-species profiles in the boundary layer could be conceived. The simplest option may be to use the simplified model as it is, but with a different set of rate parameters for CO2 environments. While this is straightforward, the model may remain valid only in a narrow range of moisture and CO2 concentrations if gasification reactions considerably affect particle temperatures and/or char consumption rates. The next level of complexity might thus be to include char gasification reactions in CFD codes.

between 30 and 70 μm, with peak temperatures shifting towards smaller sizes with increasing freestream oxygen concentration.

Figure 4 Predicted and measured (symbols) particle surface temperatures for three free-stream O2 concentrations, 16% moisture and CO2 bath gas. Solid lines n = 0.5, dotted lines n = 1, dashed lines n = 0.1.

4.2. SKIPPY Simulations Simulated particle temperatures for combustion in N2-dominated environments and three different heterogeneous reaction mechanisms are shown in Figure 5. The first mechanism consists of oxidation reactions R1-R4 only, which, as mentioned earlier, results in vast over-prediction of particle temperatures at elevated oxygen concentrations. The temperatures of smaller particles are controlled by convective heat loss, whereas radiative losses cause the temperature to fall at the largesize end of the studied size range. Highest temperatures are predicted in the intermediate size range

The second mechanism includes R1-R4 and the steam gasification reaction, R6; however, in this case a different set of rate parameters was used for R6 than listed in Table 2: A = 1.16E16 g/cm2·s, and E = 251.0 kJ/mol. This parameter set was found to produce the best match of predicted and observed temperatures of 100 μm particles and is still well within the range of activation energy and overall reaction rate that appears in the literature for steam gasification [14]. The relatively good fit for all oxygen bulk concentrations suggests that the steam gasification reaction plays an important

27th Annual International Pittsburgh Coal Conference, Istanbul, Turkey, Oct. 11-14, 2010

8 role in determining the char particle temperature and thereby the char oxidation rate. The third model adds the CO2 gasification reaction, R5, and thus uses the full mechanism specified in Table 2. Although the fit for 100 μm particles is slightly better than for the second model, the necessity of including this reaction is not obvious, based on the data shown in Figure 5 for N2dominated environments.

when adding the heterogeneous CO2 reaction. This implies that CO2 gasification plays an important role also when insignificant amounts of CO2 are present in the bulk gas. Under those conditions, carbon dioxide formed at the char surface and in the boundary layer participates in gasification reactions (because of relative slow diffusion of CO2 in N2, appreciable concentrations of CO2 can be present at the particle surface despite low CO2 production rates). For application to CFD codes, the suggested relevance of steam and CO2 gasification reactions implies that lumped kinetics to model char consumption as a single-step oxidation process may only be valid for narrow ranges of combustion environments (in terms of steam and carbon dioxide concentrations). Therefore, it may be necessary to account for both steam and CO2 concentrations to produce appropriate estimates of char temperatures and burning rates.

Figure 5 SKIPPY predictions of particle surface temperatures for three free-stream O2 concentrations, 16% moisture and N2 diluent compared with measured data for North Antelope char. Model I: oxidation only (R1-R4), Model II: oxidation with steam gasification (R6 with E = 251 kJ/mol, A = 1.16E16 g/cm2 s, Model III: R1-R6 as specified in Table 2.

Comparison of experimental data and model prediction for combustion in CO2 diluent (Figure 6) clearly shows that both steam and CO2 gasification reactions have to be included to produce an acceptable fit for sizes near 100 μm. Both reduced models (oxidation alone and oxidation together with steam gasification) over-predict temperatures for that size range. The suggested importance of CO2 gasification implies that CO oxidation in the boundary layer, which produces additional CO2 in the vicinity of the particle, may be more important then previously assumed. Similarly, the water-gas shift reaction (CO + H2O → CO2 + H2) might play a subtle role due to its impact on the gas composition near the particle surface. For combustion in conventional environments, the rate parameters for the steam gasification reaction had to be reduced

Figure 6 SKIPPY predictions (solid lines) of particle surface temperatures for three free-stream O2 concentrations, 16% moisture and CO2 diluent compared with measured data for North Antelope char. Model I: oxidation only (R1-R4), Model II: oxidation with steam gasification (R6 with E = 251 kJ/mol, A = 1.16E16 g/cm2 s, Model III: R1-R6 as specified in Table 2.

5. Summary Combustion of char particles of bituminous (Utah Skyline) and subbituminous (North Antelope) coal in N2 and CO2 dominated gas enviroments with 12 to 36 mole-% O2 and 16% H2O were investigated both experimentally and through numerical simulations. Particle temperatures from

27th Annual International Pittsburgh Coal Conference, Istanbul, Turkey, Oct. 11-14, 2010

9 two-color pyrometer measurements were compared with predictions from (1) a simplified single-film apparent kinetics model, and (2) a more complex model that includes both homogeneous reactions of gas phase species and heterogeneous reactions in pores and external particle surface. Single-film model predictions suggest that optimized rate parameters for any reaction order between 0.5 and 1 produce reasonably good fits to the data and may thus be employed in CFD codes for wide size and temperature ranges. Particle temperatures for CO2-dominated environments are generally over-predicted when using rate parameters derived with N2 data, i.e. the simplified model investigated does not fully capture the characteristics of char consumption in oxy-fuel combustion conditions. As an alternative to the option of estimating separate sets of rate parameters for CFD applications in N2-dominated and CO2-dominated environments, heterogeneous steam and CO2 gasification reactions could be included to improve the quality of the predictions. Using the reacting particle simulation code SKIPPY (Surface Kinetics in Porous Particles), temperatures for particles around 100 μm were reasonably well predicted for both N2 and CO2 environments if both steam and CO2 gasification reactions were included and a fixed reactive surface area of 10m2/g assumed. The results suggest that oxidation and both steam and CO2 gasification reactions contribute to the char consumption process. An important conclusion therefore is that simplified (single-film) models must account for gasification reactions to maintain predictive capability in both combustion environments. To further substantiate these conclusions, experiments with externally prepared char (to remove effects of the combustion environment on devolatilization characteristics (and hence char properties)) will be conducted. In addition, a more sophisticated oxidation mechanism will be implemented to improve the quality of SKIPPY predictions. Acknowledgement This research was sponsored by the U.S. Department of Energy’s Carbon Sequestration Program under award number DE-NT0005288. This work is part of a project led by Reaction Engineering International and is managed by Mr.

Timothy Fout of the National Energy Technology Laboratory. Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy’s National Nuclear Security Administration under Contract DE-AC04-94AL85000.

Disclaimer This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.

References [1] B.J.P. Buhre, L.K. Elliott, C.D. Sheng, R.P. Gupta, T.F. Wall, Prog. Energy Combust. Sci. 31 (4) (2005) 283-307. [2] R. Tan, G. Corragio, S. Santos, Oxy-Coal combustion with flue gas recycle for the power generation industry. A literature review, Report No. IFRF Doc. No. G 23/y/1, International Flame Research Foundation, 2005. [3] M.B. Toftegaard, J. Brix, P.A. Jensen, P. Glarborg, A.D. Jensen, Prog. Energy Combust. Sci. 36 (2010) 581-625. [4] R. Payne, S.L. Chen, A.M. Wolsky, W.F. Richter, Combust. Flame 67 (1989) 1-16. [5] H. Liu, R. Zailani, B.M. Gibbs, Fuel 84 (16) (2005) 2109-2115. [6] T. Kiga, S. Takano, N. Kimura, K. Omata, M. Okawa, T. Mori, M. Kato, Energy Conv. Managmt. 38 (0) (1997) S134. [7] T. Nozaki, S. Takano, T. Kiga, K. Omata, N. Kimura, Energy 22 (2-3) (1997) 199-205. [8] H. Farzan, S.J. Vecci, F. Chatel-Pelage, P. Pranda, A.C. Bose, Pilot-scale evaluation of coal combustion in an oxygen-enriched recycled flue gas. The 30th international technical conference on coal utilization & fuel systems. Clearwater, FL, USA, 2005. [9] H. Liu, R. Zailani, B.M. Gibbs, Fuel 84 (7-8) (2005) 833-840. [10] S.P. Khare, T.F. Wall, A.Z. Farida, Y. Liu, B. Moghtaderi, R.P. Gupta, Fuel 87 (2008) 1042-1049. [11] X. Huang, X. Jiang, X. Han, H. Wang, Energy & Fuels 22 (2008) 3756-3762. [12] P. Heil, D. Toporov, H. Stadler, S. Tschunko, M. Förster, R. Kneer, Fuel 88 (2009) 1269-1274. [13] E. S. Hecht, C. R. Shaddix, A. Molina, B. S. Haynes Proc. Combust. Inst., in press.

27th Annual International Pittsburgh Coal Conference, Istanbul, Turkey, Oct. 11-14, 2010

10 [14] C.R. Shaddix, E.S. Hecht, M. Geier, A. Molina, B.S. Haynes, Effect of gasification reactions on oxy-fuel combustion of pulverized coal char, Proceedings of the 35th International Technical Conference on Clean Coal & Fuel Systems, Clearwater FL, June 6–10, 2010. [15] C.R. Shaddix, A. Molina, Effect of O2 and High CO2 Concentrations on PC Char Burning Rates during Oxy-Fuel Combustion, Proceedings of the 33rd International Technical Conference on Coal Utilization and Fuel Systems, Clearwater, FL, June 1-5, 2008. [16] D.A. Tichenor, S. Niksa, K.R. Hencken, R.E. Mitchell, Proc. Combust. Inst. 20:1213-1221 (1984). [17] C.R. Shaddix, “Correcting thermocouple measurements for radiation loss: A critical review,” 33rd National Heat Transfer Conference, Albuquerque, NM, USA, 1999, p. 1150. [18] P.J. Ashman, B.S. Haynes, Improved techniques for the prediction of NOx formation from char nitrogen, Project No. C4065, Australian Coal Association. CSIRO Energy Technology, 1999. [19] G.P. Smith, S.D. Golden, M. Frenklach, N.W. Moriaty, B. Eitener, M. Goldenberg, C.T. Bowman, R.K. Hanson, S. Song, W.C. Gardiner, V.V. Lissianski, Z. Qin, (2001) GRIMECH 3.0. http://www.me.berkeley.edu/gri_mech (accessed December 2001). [20] M. Geier, E.S. Hecht, C.R. Shaddix, 26th Annual International Pittsburgh Coal Conference, Pittsburgh PA Sept. 20-23, 2009 [21] A. Molina, A.F. Sarofim, W. Ren, J. Lu, G. Yue, J.M. Beer, B.S. Haynes, Combust. Sci. Technol.174 (2002) 43-71. [22] N.M. Laurendeau, Prog. Energy Combust. Sci. 4 (1978) 221-270. [23] R.H. Essenhigh, Energy & Fuels 5 (1991) 41-46. [24] L. Tognotti, J.P. Longwell, A.F.Sarofim, Proc. Combust. Inst. 23 (1990) 1207-1213. [25] J.J. Murphy, C.R. Shaddix, Combust. Flame 144:710729 (2006). [26] P.H. Paul, J. Warnatz, Proc. Combust. Inst. 27 (1998), 495–504. [27] E.W. Thiele, Ind. Eng. Chem. 31 (1939) 916–920. [28] M.F.R. Mulcahy, I.W. Smith, Rev. Pure. Appl. Chem. 19 (1969) 81–108. [29] J.J. Murphy, C.R. Shaddix, Combust. Flame 157: (2010) 535-539

27th Annual International Pittsburgh Coal Conference, Istanbul, Turkey, Oct. 11-14, 2010

2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 STUDY ON THERMODYNAMIC CALCULATION FOR O2/CO2 FLUE GASES RECYCLED COMBUSTION BOILER Li-Qi Zhang*, Can-Zhi Li, Fang Huang, Ji-Hua Qiu, Chu-Guang Zheng State Key Lab. of Coal Combustion, Huazhong University of Science and Technology 1037 Luoyu Road,Wuhan, Hubei, 430074, P.R.CHINA *Corresponding author. Tel: +86 27 87542417-8316 Fax: +86 27 87545526 Email: [email protected]

Abstract: Oxy-fuel combustion technology is considered as an effective approach to capture CO 2 from the combustion of fossil fuels. The traditional boiler must be retrofitted in order to adapt to this new technology and the conventional thermodynamic calculation method also needs to be modified for obtaining higher accuracy. A 50MW traditional pulverized coal boiler was calculated in this paper, which adopts modified thermal calculation method to calculate the thermodynamic in the different O2/CO2 recycle mode and in different recycle rate, compares the results with that of traditional air and optimizes the recycle rate in O2/CO2 recycle mode. The results indicated that the best operation condition will be achieved when the oxygen volume of oxidant is 30% (recycle rate 0.633) in hot recycle mode. KEY WORDS：O2/CO2; thermal calculation; recycle rate

1. Introduction O2/CO2 recycle combustion is considered as the most potential new combustion technology on CO 2 capture currently. The main characteristic of this technology is that the air oxidants is replaced by the pure oxygen and recycled part of the flue gas, which can greatly increase the concentration of CO2 in flue gas and the CO2 can be used and disposed without separation, consequently, sharply decrease the emission of CO2 into the atmosphere. Meanwhile, the recycled flue gas significantly reduce the exhaust (only 1/5 of traditional air), which greatly reduce exhaust heat loss and apparently increase the boiler efficiency; besides, this new technology has a good ability of removing SOx and NOx. A lot of studies on this new technology have been done until now, including combustion characteristic, pollutant emission, mineral behavior and the theory of radiant heat transfer et al at laboratory scale; besides, including the technological and economic feasibility of using O2/CO2 recycle combustion on retrofitted traditional boilers, pilot scale study and numerical simulation are also included [1-12]. In the near 5 years (before 2015) O2/CO2 recycle combustion will be soon accomplished from 30MW to full industrial scale, 565MW [13]. Due to the change of combustion atmosphere, the thermal calculation method of traditional boiler is no longer suitable for industry application as for the application of this new combustion technology, thus in urgent need of developing new calculation theory and method. This paper has modified the thermal calculation method of traditional boiler and made it useful in O2/CO2 recycle mode compared the results between recycle mode and air in order to find out the best O2/CO2 recycle mode and the best recycle rate.

2. The Calculated Boiler and Coal The calculated boiler is a pulverized coal boiler whose model is F220/100-W and its parameters are showed in tab.1. Along with the flow direction of flue gas, the components of the boiler are furnace, screen superheater, dregs prevent tubes, high temperature superheater hot section and cold section, low temperature superheater, high temperature economizer, high temperature pre-heater, low temperature economizer and low temperature pre-heater. The coal used in this paper is HeGang bituminous coal and its properties are showed in tab.2. Tab.1 The parameters of the boiler F220/100-W Pulverized coal boiler Foursquare once-through burner, Dry-bottom furnace, Natural circulation Rated load 220 t/h Superheat steam pressure 9.8MPa Superheat steam temperature 540 ℃ Feedwater temperature 215 ℃ Feedwater pressure 11.57 MPa Steam drum pressure 11.08 MPa Discharge rate 2% Ambient temperature 30 ℃ Drum-type ball mill, intermediate storage bunker hot air milling system Tab.2 Proximate and ultimate analysis of bituminous coal Proximate analysis/% Ultimate analysis/% Qnet,ar/(kJ/kg) Aar Mar Var Car Har Oar Nar 20.07 9.00 34.73 22676 58.14 3.86 8.09 0.62

Sar 0.22

3. Thermal Calculation Method in O2/CO2 Recycle Mode Fig.1 is the schematic of boiler in air, fig.2,3 are schematic of boiler in cold and hot recycle mode respectively. Air separation unit, condenser and compressor are added to cold recycle mode compared to air but the original boiler structure and parameters remain unchanged. Reversely, the boiler structure in hot recycle mode change a lot. Behind low temperature superheater, part of the flue gas recycled after got through the high temperature precipitator and delivered into the furnace as part of the secondary oxidant. A high temperature precipitator is added because the hot recycle flue gas need to remove the dust, meanwhile, the recycled of hot flue gas result in the reduction of the flue gas that get through the economizer, so it can replace the high and low temperature economizers by using only one economizer; the amount of primary air and oxygen in O 2/CO2 recycle mode which get through the pre-heaters is less than that in air's, so the high and low temperature pre-heaters can be replaced by one pre-heater as well.

Fig.1 Schematic of boiler in air

Fig.2 Schematic of boiler in cold recycle mode

Fig.3 Schematic of boiler in hot recycle mode

The volume difference of each component in the flue gas between O2/CO2 recycle mode and air is quite large, especially CO2. The volume differences of the components definitely change the properties of flue gas in O2/CO2 recycle mode when compared to that in air. Also, the oxidant in O2/CO2 recycle mode is composed of recycle flue gas and pure oxygen, so that the volume of the components has great distinction with air, hence the thermal calculation method in air needs to be modified in order to be used in O2/CO2 recycle mode. The modifications are mainly caused by the different components volume of the flue gas and oxidant, such as the calculation formula of density and enthalpy of flue gas and oxidant, the variation of heat transfer caused by the change of physical parameters et al. Among all of the changes the heat transfer is the most influential, including radiant and convection heat transfer [14,15].

4 Results and Analysis The oxidant in O2/CO2 recycle mode is composed of recycle flue gas and pure oxygen, when operating in recycle mode the newly rejoin gas is only oxygen and the amount of flue gas is controlled by recycle rate, the recycle rate also decides the combustion atmosphere in O2/CO2 recycle mode. In order to stabilize the operation of the boiler in certain recycle rate, it needs to find out each component volume of gas in a state of equilibrium in this certain recycle rate. This paper has set up corresponding oxidant-flue gas cycle calculation program according to different O2/CO2 recycle mode. In the cycle calculation program the basic calculated unit is the component of gas, so each component volume can be attained when the state of equilibrium will be achieved.

4.1 Definition of Oxygen Purity, Excessive Oxygen Coefficient and Air Leakage Coefficient The relation between oxygen purity and cost of the ASU is inverse, the higher the purity the higher the cost, but it will be harmful for the separation of CO2 if the purity is too low. The oxygen purity chosen here is 95% after having considered the factors above. The excessive oxygen coefficient chosen here is 1.05[11]. The N2 volume in flue gas is greatly influenced by air leakage coefficient in O 2/CO2 recycle mode. Fig.4 is the relation between air leakage coefficient and N2 volume in flue gas in a state of equilibrium in O2/CO2 recycle mode when the recycle rates are 0.7 and 0.8, respectively. It can be found out that the N2 volume is linear increased with the increasing of air leakage coefficient and the linear

slope increase with the increasing of recycle rate. The linear slope is up to 209 and 322 respectively by fitting the straight lines, and the N2 volume is up to 44.45% and 67.06% when air leakage coefficient is 0.2 (approximates the air leakage coefficient in air). The excessive high of N2 volume obviously is not the original intention of O2/CO2 recycle combustion technology, because this will enhance the difficulty of separating CO2. The air leakage coefficient chosen here is 0.02, under which the N 2 volume can be controlled within 10%, approximately. 32

70

recycle rate 0.7 recycle rate 0.8

50

Oxygen volume in oxidant/%

N2 volume in flue gas/%

60

40 30 20 10 0 0.00

30

cold recycle mode hot recycle mode

(0.633,30)

28

26

24

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

0.18

0.20

0.62

Air leakage coefficient in cold recycle mode

Fig.4 Relation between air leakage and N2 volume in flue gases

0.64

0.66

0.68

0.70

Recycle rate

Fig.5 Relation between oxygen volume in oxidants and recycle rate

4.2 The Adiabatic Temperature and Boiler Exhaust Temperature Fig.5 is the relation between oxygen volume in oxidant and recycle rate, the trend of this two recycle mode is linear and the lines are basically coincided with each other, when the oxygen volume in oxidant is 30%, the recycle rate is 0.633. Fig.6 is the relation between adiabatic temperature and boiler exhaust temperature of different recycle rates in cold and hot recycle mode. As can be seen from the figure, in both recycle modes the adiabatic temperature decreases with the increasing of recycle rate, and when the recycle rates are 0.679 and 0.683 respectively the theoretical combustion temperature equal to that of in air. This is mainly because small recycle rate generates less flue gas, but fuel consumption changes little along with the change of recycle rate. Although the specific heat capacity of CO2 is higher than N2, the specific heat capacity impact can not offset the impact of the decreased flue gas, resulting in the lower adiabatic temperature in recycle mode compared to that in air; the higher CO2 specific heat capacity start to work when the recycle rate increase gradually and decrease the adiabatic temperature. The boiler exhaust temperature increase first and decrease later along with the increasing of recycle rate in both recycle mode, but the maximum temperature is basically the same as that in air, the reason of this is very similar to that of adiabatic temperature, when adiabatic temperature is high, radiant heat transfer will be strengthened but the amount of flue gas will be few. On the contrary, when the adiabatic temperature is low, radiant heat transfer will be decreased but the amount of flue gas will be increased, which is a counter-balance, interactive process.

2400

2400

adiabatic temperature in cold recycle mode boiler exit temperature in cold recycle mode

adiabatic temperature in hot recycle mode boiler exit temperature in hot recycle mode

2200

2200

2000

(0.679,2034)

1800

adiabatic temperature in air

1100

boiler exit temperature in air

(0.683,2034)

1800

Temperature/℃

Temperature/℃

2000

(0.706,1061) 1050

adiabatic temperature in air

1600

boiler exit temperature in air

1080

(0.759,1061)

(0.700,1061)

1060 1040 1020

1000 0.60

0.62

0.64

0.66

0.68

0.70

0.72

1000 0.60

0.74

0.62

0.64

0.66

0.68

0.70

0.72

0.74

0.76

0.78

Recycle rate in hot recycle mode

Recycle rate in cold recycle mode

a) cold recycle mode b) hot recycle mode Fig.6 Comparison of adiabatic temperature and boiler exhaust temperature in different recycle mode

4.3 H2O volume in flue gas and CO2 volume in flue gas after condensation Fig.7 is the comparison of H2O volume in different recycle mode and in air, it can be found that in hot recycle mode is above 10% higher than in air and in cold recycle mode the increment is about 5%. This is because of the accumulation of H2O in recycle mode. Part of the flue gas is recycled after high temperature precipitator without getting through the condenser in hot recycle mode, which results in higher H2O volume in oxidant when compared to that in cold recycle mode, ultimately increases the H2O volume in flue gas. Fig.8 shows the CO2 volume in flue gas after condensed in different recycle mode and recycle rate, in hot recycle mode the CO2 volume is up to about 89%, while in cold recycle mode is only about 86%. The higher CO2 volume in hot recycle mode is because the higher H2O volume in hot recycle mode which can reduce the impact of N 2, so after the condensation of flue gas the effect is reflected in the increasing of CO2. 89

CO2 volume in flue gas after condensation/%

20

Steam volume in flue gas/%

18

cold recycle mode hot recycle mode

16 14 12

10

air

8 0.61

0.62

0.63

0.64

0.65

Recycle rate

Fig.7 Comparison of H2O volume in flue gas after condensed

88

87

air recycle rate 0.614 recycle rate 0.633 recycle rate 0.648

86

85 15 10 5 0

cold recycle mode

recycle rate

hot recycle mode

Fig.8 Comparison of CO2 volume in flue gas

4.4 Amount of Flue Gas, Exhaust Heat Loss and Boiler Efficiency The air leakage coefficient is usually reached 20% or more in air, but in the O2/CO2 recycle mode the coefficient must be controlled in a rather small range, so the amount of flue gas in air is 40% higher than that in recycle mode; The exhaust is just a little in recycle mode, for most of the flue gas is recycled, thus reducing the exhaust heat loss to about 2%, almost 4% lower than that in air, as shown in Fig.9. The reduction of exhaust heat loss raise the boiler efficiency from about 92% in air to above 95% in recycle mode, and at the same time reduce the boiler fuel consumption.

6

Exhaust heat loss/%

5

air recycle rate 0.614 recycle rate 0.633 recycle rate 0.648

4

3

2

1

0 cold recycle mode

hot recycle mode

Recycle rate

Fig.9 Comparison of exhaust heat loss in difference recycle mode

4.5 Flue Gas Enthalpy and Heat Absorption of Every Part of the Boiler Fig.10 is the relation between flue gas enthalpy and flue gas temperature in different recycle mode. When the recycle rate is 0.633 (oxygen volume is 30%), the changing curve of flue gas enthalpy is coincident to that in air. In this recycle rate the heat transfer of flue gas rate is more closed to that in air. The results show that the amount of hearth heat absorption accounted for the proportion of the total boiler heat absorption is 62% in air, while the proportion in recycle mode is a few percentage points higher than that in air. This is because after N2 is replaced by CO2 the flue gas total emissivity and the adiabatic temperature will increase, thus increasing the radiant heat transfer of the furnace. As the radiant heat transfer has a dominant place in heat transfer in furnace, it will increase the heat absorption in furnace. The increment of heat absorption in furnace will inevitably lead to lower heat absorption in economizer. For a certain boiler, the absorbed heat that makes the feedwater change into steam can be divided into pre-heat, heat of vaporization and overheat. Usually the ratio of this three part is fixed, but when the ratio changes, it will often lead to operate unstable and even cannot operate normally, so in recycle mode the ratio should remain in a relatively unchanged level. The increasing heat absorption in furnace reduces the heat absorption in other parts of the boiler accordingly, the reduction of the economizer pre-heat absorption will enhance the gap between the economizer exit temperature and the saturation temperature in steam drum, which is harmful to the operating stability of a boiler. 25000

25000

air recycle rate 0.614 recycle rate 0.633 (oxygen volume30%) recycle rate 0.648

15000

air recycle rate 0.616 recycle rate 0.633 (oxygen volume30%) recycle rate 0.648

20000

Enthalpy of flue gas/kJ/kg

Enthalpy of flue gas/kJ/kg

20000

10000

5000

15000

10000

5000

0

0 0

500

1000

1500

2000

Temperature of flue gas in cold recycle mode/℃

2500

0

500

1000

1500

a) cold recycle mode b) hot recycle mode Fig.10 Relation between flue gas enthalpy and temperature in different recycle mode

4.6 The Total Water of the Spray-type Attemperator

2000

Temperature of flue gas in hot recycle mode/℃

2500

The role of the spray-type attemperator is to regulate the temperature of superheated steam. Due to the limited range and small amount that it will affect the safety of the boiler operation; besides, as the amount is too large, it will not only affect its safety but also will be wasteful. Hence, usually the total spraying water is about 3% of the boiler capacity (in this paper the capacity is about 6000 kg/h). As can be seen from Fig.11, the total spraying water increase in both recycle rate along with the increasing of recycle rate, the total spraying water in hot recycle mode is lower than that in cold recycle mode in the same recycle rate but the difference is not too much. When the recycle rate is 0.633(oxygen volume is 30%), the total spraying water is about 3% of boiler capacity, in this recycle rate a more reasonable temperature reduction effectiveness will be achieved by using the spray-type attemperator.

锅炉减温喷水总量/kg/h

20000

cold recycle mode hot recycle mode

15000

10000

recycle rate 0.633 oxygen volume 30%

5000

0 0.61

0.62

0.63

0.64

0.65

0.66

0.67

0.68

Recycle rate

Fig.11 Relation between total spraying water and recycle rate

5. Conclusion The combustion atmosphere in O2/CO2 recycle mode is mainly controlled by regulating recycle rate. In recycle mode, the increment of tri-atom gas volume (mainly CO2 volume) and boiler efficiency are very obvious, the reduction of the exhaust heat loss is obvious as well; besides, the fuel consumption decreases a little, the furnace blackness increases because of the increment of the emissivity of flue gas; the air leakage coefficient has great effect on CO2 volume; adiabatic temperature will decrease as the recycle rate increases and equal to that in air when recycle rate is 0.679 (cold recycle mode) or 0.683 (hot recycle mode). The boiler structure remains basically unchanged except a few apparatus are added in cold recycle mode. Though the total results meet the requirements, some of them are manifestly unreasonable in practical operation. For example, the heat absorption of the economizer is decreased which result in much lower exit temperature and make the thermal stress of junction with the steam drum too excessive, affecting the boiler’s safety. Meanwhile, the exhaust gas temperature will be reduced to about 100℃ due to the enhancement of convection heat transfer, which is manifestly unreasonable. The original boiler economizer and air pre-heater have been changed a lot in hot recycle mode. The changed version will adjust better according to different recycle rate. Therefore, hot recycle mode is the better choice to O2/CO2 combustion technology. Through comparing two recycle modes in different recycle rate, it is found that it can get the best results when in hot recycle mode and the oxygen volume in oxidant is about 30%(recycle rate is 0.633).

Acknowledgement:

The authors express their great thanks to the National Natural Science Foundation of China (Grant No. 50704017, 50936001, 50721005), and National Basic Research Program of China (“973” Program) (Grant No. 2006CB705806).

Refrences: [1] Zheng CG, Zheng Y, Li F, et al. Greenhouse effect and its control measures [M]. Beijing, China Electric Power Press, 2001. [2] Zhang LQ, Huang ZJ, Zou C, Zheng CG. Study on changes in pore structure of pulverized coal combustion in O 2/CO2 Atmosphere. International Pittsburgh Coal Conference，2007，Johannesburg，South Africa, September 10-14, 2007. [3] Zhou C, Huang ZJ, Chu K, Gui XL, Qiu JH, Zhang LQ, Zheng CG．A pilot scale study on SO2 and NOx emission control in O2/CO2 recycled coal combustion. Proceedings of the Chinese Society for Electrical Engineering, 2009, 20(2): 20-24. [4] Payne R, Chen SL, Wolsky AM, Richter WF. CO2 recovery via combustion in mixtures of oxygen and recycled flue gas [J]. Combustion Science & Technology, 1989, 67: 833-40. [5] Kimura N, Omata K, Kiga T, Tanako S, Shikisima S. The characteristics of pulverized coal combustion in O2/CO2 mixtures for CO2 recovery [J]. Energy Conversion & Management. 1995, (36): 805-8. [6] Kiga T, Tanako N, et al. Characteristic of pulverized-coal combustion in the system of oxygen/recycled flue gas [J]. Energy Conversion & Management. 1997, (38): S129-134. [7] Liu H, Ramlan Z, Bernard M G. Comparisons of pulverized coal combustion in air and in mixtures of O 2/CO2[J]. Fuel, 2005, 84(7-8): 833-840. [8] Klas A，Robert J, Stefan H, Filip J, Bo L. Radiation intensity of lignite-fired oxy-fuel flames[J]．Experimental Thermal and Fluid Science. 2008, 33(1): 67-76. [9] Stefan H, Klas A, Filip J, Bo L. Combustion characteristics of lignite-fired oxy-fuel flames[J]．Fuel．2009, 88(11): 2216-2224. [10] Buhre B J P, Elliott L K, et al. Oxy-fuel combustion technology for coal-fired power generation [J]. Progress in Energy and Combustion Science，2005，31(4)：283-307. [11] Wal1 T. Combustion processes for carbon capture [J]．Proceedings of the Combustion Institute, 2007: 31-47. [12] Wall T, Yinghui Liu, et al. An overview on oxyfuel coal combustion-State of the art research and technology development [J]． Chemical Engineering Research and Design，2009,87(8): 1003-1016. [13] WALL Terry, TU Jianglong. Coal-fired oxyfuel technology status and progress to deployment [C]．34th International Technical Conference on Coal Utilization & Fuel Systems, Florida, USA: Coal Technology Association,2009. [14] Leckner B. Spectral and total emissivity of water vapor and carbon dioxide [J]. Combustion and Flame. 1972, (19): 133-148. [15] Feng JK, Shen YT, editor in chief. Principle and calculation of boiler[M]．Beijing, Science Press. 1992.7.

Coal:Biomass Gasification - A Pathway for New Technology Development of Oxygen Blown Co-fired Gasification with Integrated Electrolysis Presenting Author: Dr Tana Levi, Technology Operations Manager, [email protected] CRL Energy Ltd., 68 Gracefield Road, Lower Hutt 5040, New Zealand A I Gardiner, [email protected], Manager Hydrogen and Distributed Energy, Industrial Research Limited, 5 Sheffield Crescent, Bishopdale, Christchurch 8053, New Zealand R. Whitney, CEO, [email protected] CRL Energy Ltd., 68 Gracefield Road, Lower Hutt 5040, New Zealand Y Iwasaki, Scientist, [email protected] CRL Energy Ltd., 68 Gracefield Road, Lower Hutt 5040, New Zealand S Pang, Professor, [email protected] Department of Chemical and Process Engineering, University of Canterbury, Christchurch, New Zealand . Q Xu, PhD Student, [email protected] Department of Chemical and Process Engineering, University of Canterbury, Christchurch, New Zealand . and The Late Dr Tony Clemens CRL Energy Ltd., 68 Gracefield Road, Lower Hutt 5040, New Zealand

This paper is dedicated to the memory of Tony Clemens who sadly recently passed away. At the time of his death Tony was a highly respected and greatly esteemed energy scientist. He had a strong belief that it was important the world have alternative, sustainable energy technologies available and led the Government supported project to build a gasifier to produce hydrogen from coal and biomass. Over the years Tony has won several awards at the Pittsburgh International Coal Conference, the most recent in 2009 as a member of a team for their scientific paper on the atomic level processes that occur when carbon dioxide is injected into various coal types. Tony died unexpectedly on 19 February 2010, he was a valued colleague and friend and his dedication to research and his helpfulness to fellow scientists will continue to be an inspiration to us all.

Abstract: This paper describes progress in new research that builds on the air blown gasification of lignite for hydrogen production technology that has been developed over the past several years. The new technology package is designed to assist New Zealand meet the challenges of peak oil and global climate change. The programme uses the results from gasifying a range of pellet blends of lignite or sub-bituminous coal with P.radiata or E.nitens, in bench scale and then small pilot scale gasifiers, to develop a fundamental understanding of the chemical processes underpinning gasification behaviour of New Zealand lignite and coal / woody biomass blends. Significant attention is given to the development of a new technology comprised of oxygen blown co-fired gasification with the integration of high efficiency electrolysis for production of low carbon footprint syngas and/or and hydrogen and ultimately synfuels from New Zealand’s coal, biomass and renewable electricity resources.

Syngas produced can be used for Fischer Tropsch production of liquid transport fuels. If shown to be successful the new technology would offer: i) Low carbon footprint transport fuels for the present and hydrogen for future transport fleets, ii) Increased flexibility in

tailoring syngas for Fischer Tropsch production, iii) A new option for matching the uptake of New Zealand intermittent renewable generation (including wind, marine and hydro) with a new deferrable load in the form of stored hydrogen, and iv) in the extreme case of critical global oil constraints the possibility of protecting New Zealand’s energy security by providing an alternative hydrogen feedstock for hydro-cracking of low grade oil supplies.

1.

Introduction

As a result of environmental and other policy considerations, there is increasing world-wide interest in the use of biomass resources as feedstocks for producing power, fuels, and chemicals. Biomass resources are a major component of strategies to mitigate global climate change. Plant growth recycles CO2 from the atmosphere, and the use of biomass resources for energy and chemicals results in low net emissions of carbon dioxide. The emissions of NOx and SOx from biomass facilities are also typically low. The use of these locally produced energy resources also results in new markets for agricultural and forestry products and provides a mechanism for rural economic development[1].

Gasification technologies provide the opportunity to convert renewable biomass feedstocks into clean fuel gases or synthesis gases. These gaseous products can be burned to generate heat or electricity, or they can potentially be used in the synthesis of liquid transportation fuels, hydrogen, or chemicals. Gasification offers a combination of flexibility, efficiency, and environmental acceptability that is essential in meeting future energy requirements.

Currently, the majority of biomass-based power generation plants are small and relatively inefficient, (< 2 MW) with efficiencies generally in the ~20-25% range. These plants use several feed stocks such as agricultural wastes, landfill methane, wood wastes in various energy conversion devices such as steam turbines, internal combustion engines, and, more recently, fuel cells. Larger biomass combustion systems used in industrial plants, generally in the wood products industry, and in municipal solid waste disposal can range from 5 - 75 MW, but again are not very efficient. With these factors in mind, research programmes to identify methods to more efficiently use the biomass resources to generate power have developed[2] , requiring that gasification of biomass be integrated with the power plant into a system that offers significant improvements in thermal efficiency and environmental performance. The cost of power from these facilities must be competitive with local circumstances.

When looking at biomass conversion it is instructive to look at coal conversion, as there are many similarities. For coal gasification the minimum temperature required is about 900oC. About the same maximum temperature of 800-900oC is required to gasify the most refractory part of almost any biomass, i.e the temperature required for the complete thermal gasification of biomass is of the same order of magnitude as for coal. This high temperature, in combination with the impurities present, whether sulphur or ash components, is why indirectly heated coal and biomass gasification processes in which external heat has to be transferred via a metal surface have not yet achieved any commercial success[3].

There are however a number of significant differences between coal gasification and biomass gasification, which are directly attributable to the nature of the feedstock. These include the quality of biomass ash, which has a comparatively low melting point but

in the molten state is very aggressive; its reactivity; and particularly with vegetable biomass, it’s fibrous characteristics. Finally, particularly in the lower temperature range, biomass gasification has a very high tar make[3].

Although an entrained-flow process might seem an attractive option for generating a clean, tar-free gas as required for chemical applications, where the low melting point of the ash would keep the oxidant demand low, there are problems with this approach. The aggressive quality of the molten slag negates such a solution, whether using a refractory or a cooling membrane for containment protection. Furthermore the short residence times of entrained-flow reactors require a small particle size, to ensure full gasification of the char. No method of size reduction has yet been found, which will perform satisfactorily on fibrous biomass.

A number of fixed-bed processes have been applied to lump wood, but they are limited to this material. They would not work on straw, miscanthus or other materials generally considered for large-scale biomass production unless these were previously briquetted. Furthermore in a counter-flow gasifier, the gas would be heavily laden with tar. The alternative of co-current flow could reduce the tar problem substantially, but the necessity to maintain good control over the blast distribution in the bed restricts this solution to units of very small size.

With this background it is probably not surprising that most processes for biomass gasification use fluidized beds and aim at finding a solution to the tar problem outside the gasifier. In co-firing applications where the syngas is fired in an associated large-scale fossil fuel boiler, the problem can be circumvented by maintaining the gas at a temperature above the dewpoint of the tar. This has the added advantage of bringing the heating value of the tars and the sensible heat of the hot gas into the boiler. There are many biomass processes at various stages of development[3]. Recent research shows increasing interest in co-gasification of biomass with coal[4] and for large scale applications, blending coal with biomass is becoming a preferred option. It not only reduces the cost of the raw materials for energy supply but also increases the overall conversion of coal and increases the calorific value of the feedstock. The blended feed is also considered as a more sustainable energy resource due to its reduced global environmental impact.

Some of the benefits of using oxygen instead of air in the gasification process are: the heating value of the syngas is higher, the volume of gas is approximately half that required for an air blown system for a similar amount of gasification energy, gas handling and cleanup units can be smaller, as can the heat exchangers. The advantages of using pure or enhanced oxygen content in the combustion process are well recognized. However the energy costs associated with producing oxygen in an air separation unit are in the order of 15-25% of the electrical output or 5-10% of the total energy input[5] . This has kindled interest in the use of electrolysis to produce oxygen [6,7] for the gasification process. In this study[6] electrolysis is used to produce oxygen at 64% Higher Heating Value (HHV) efficiency and the study acknowledges that efficiency can be improved and costs lowered further. Another study[7] proposes methanol production from a combination of energy resources including electrolysis to produce oxygen for gasification and hydrogen to increase the H:CO ratio.

The benefits of and problems with existing electrolyser technology are well known. Electrolysis provides an elegant and relatively simple means of producing high purity oxygen and hydrogen in a ratio of 1:2, but the technology is expensive, the cost of feedstock (electricity) is high, and the production efficiency is at present of the order of only 60% HHV. Recent advances in materials technology can potentially reduce each of these barriers. Combined with the changing environment of electricity supply, improvements could alter the economics of electrolysis in applications where both oxygen and hydrogen have a high value. In[8], a number of hydrogen production technologies including electrolysis are assessed for potential cost reduction based on technology learning rates. For electrolysis this is estimated as 18%, which is substantial, and similar to growth energy technologies such as photovoltaics and wind.

A lot of effort is applied within conventional gasification processes to enhance or upgrade the hydrogen content of the syngas. Coproduced hydrogen from electrolysis provides a potential feedstock for more flexible blending or upgrading, reducing the criticality of the water gas shift process. Pilot scale integration of electrolysis, and analysis of how it fits into the overall co-production process to improve syngas quality is an aspect of the research. This approach is particularly relevant to New Zealand since there is already a high content of renewable resources in the electricity mix (60-70% depending on rainfall patterns) and intermittent wind energy will progressively become a significant component of the generation capacity. An overview of the process is shown in Figure 1.

Hydrogen-syngas co-production diagram Lignite and biomass energy sources

Advanced Cogasification

Flexible syngas production for low GHG content F-T liquid fuels

Peaking electricity gen. Chemical energy buffer storage

Intermittently available renewable electricity eg wind

O2

H2

Electrolyser Plant

WGS and CO2 Capture

Carbon sequestration

Hydrogen fuel for FCVs and stationary use

Figure 1. The overall co-fired co-production gasification-electrolysis process.

Hydrocarbon feedstock is applied to the gasifier. Hydrogen and oxygen are produced in the electrolyser at an average rate sufficient to supply the gasifier with oxygen. If the electrolyser has a relatively fast response and overdrive capacity, buffer storage for the product gases could be provided and use of high cost peak electricity could be reduced. The buffer store of hydrogen is effectively a fuel gas, as is the hydrocarbon feedstock. In a large plant, and if the payback could justify the investment, fuel (in the form of hydrogen enriched synthetic gas) could be burnt in a gas turbine for peaking electricity production. The combined system represents a

form of flow battery, with hydrogen and oxygen as the working fluids. The main emission from the plant is carbon dioxide, which along with the carbon monoxide in the syngas contains both the biomass and lignite sourced carbon. If the plant carbon dioxide emissions are sequestered, then the ratio of the biomass/lignite carbon input, the emissions factor for the electricity production, and the plant efficiency all affect how “green” the resulting syngas is.

2.

Initial Objectives

Before the full pilot scale fluidized bed gasifier and electrolysis co-production plant can be brought together and evaluated as an integrated system several individual objectives must be implemented. These are: •

Develop a process in which pellets of a known coal:biomass ratio could be made

•

Conduct fundamental studies on the char activity and catalytic effect on the co-gasification of coal:biomass pellets

•

Develop a model of the co-gasification processes

•

Understand how these pellets perform in an already commissioned 50kw air or air with 30% O2 blown fluidized bed gasifier

•

Analyse electrolysis needs and develop an electrolysis process matched to the requirements of the integrated system

•

Construct and commission a pilot scale oxygen-hydrogen production plant

•

Design, build and commission an integrated O2 blown fluidized bed gasifier where the O2 is provided by an electrolyser.

3.

Experimental Details

3.1

Pellitizing

Prior to pelletizing the coal and biomass in the required percentages the raw materials had to be prepared. Figure 2 shows some of the raw materials used. Preparation of E. nitens was time consuming and required air drying prior to several passes through a knife mill, followed by several hours in a ball mill to break it to a size where it was possible to pass it through the pellitizer. P. radiata, lignite and sub-bituminous coal did not present such a problem and only required ball milling in the appropriate percentages. Once all components were < 1mm in diameter the pelletization procedure could proceed. This consisted of the milled feed being passed through the pellitizer twice. The first pass heated the feed, this was found necessary for the pellets to bind, and the second pass formed compact pellets of the required size for fluidization. Figure 3 details some of the equipment used to prepare the pellets. The pellets need to be robust enough to be fed into a 50kw gasifier via a drop lock hopper and auger. To meet this criteria the pellets were subjected to a quality control “drop-shatter” test.

a) E. nitens Figure 2. Starting materials for pellitizing.

b) P. radiata

c) lignite

a) Ball mill

b) Pellitizer

c) 20% lignite:80% P. radiata pellet

Figure 3. Some of the equipment used during the pellitization processes.

3.2

Bench scale gasifier

The bench scale gasifier comprised a vertical tube into which char samples derived from the pelletised blends were placed and exposed to temperatures ranging from 700 to 950°C whilst being subjected to a gas flow containing entrained steam. The assembly is integrated with an online gas chromatograph for real time analysis of the syngas produced and accurate determination of carbon consumption. This design recognises the fact that while gasification is a complex process involving a number of stages including devolatilisation, the rate determining step, and therefore the factor of critical importance, is most commonly the endothermic reaction between the char arising from devolatilisation and steam: C + H2O => CO + H2. Samples of char (1.00g) were placed into a small, bench scale gasifier, under a flow of nitrogen and then heated to a pre-selected temperature. Chars derived from pellets comprised of E. nitens or P. radiata with (0 , 20, 50, 80 or 100%) added lignite or subbituminous coal were used. After stabilizing at the selected temperature, degassed, distilled water was injected at 1.77 – 1.81 ml/min through a peristaltic pump into the heated nitrogen stream and passed through the char. After exiting the bench scale gasifier, the gas mixture passed through a condenser assembly, drying tower and particulate filters before being drawn into an MTI M200 gas chromatographic gas analyzer. The concentrations of product gases (hydrogen, carbon monoxide, methane and carbon dioxide) were measured. Samples were withdrawn and measured every 90 seconds. Gasification was continued until at least 60% carbon was consumed. 3.3

50kw gasifier

The gasification suite is made up of several integrated modules as shown in Figure 4 and Figure 5. The main modules of interest for this work are are the gasifier, cyclone, venture scrubber and gas cleaning tower.

In the first stage of development air is used as the oxidation medium, with higher content of oxygen progressively introduced later. The combustion mode can be started using a variety of feed from 100% lignite or sub-bituminous coal to a blend of 80% coal : 20% woody biomass pellets. A feed rate is maintained at an average of 4 - 5 kg/h and an air feed of 85 – 90 m3/h. – these parameters will

vary depending of feed stock. The bed depth remains relatively constant at just below 300 mm and the temperature generally settles at 900oC. Combustion was held in a stable state for 1 h prior to switching to gasification. A Delta V system is the main control system. Transition from combustion mode to gasification is normally easily achieved in a matter of minutes. This is done by increasing the coal feed to 17 – 19 kg/h and decreasing the air to 60 m3/h and adding a feed of 3kg/h steam. Temperature is generally maintained by adjusting the feed rather than the air or oxygen. Gas from the gasifier passes through a heat exchanger and cyclone prior to the venturi scrubber. Next it is further scrubbed in a counter current alkaline wash tower where H2S is removed completely (cannot be detected on 0.2ppm draegger tubes). Depending on the test requirements the gas is then either flared, or sent through the WGS and membrane systems. A more detailed description of the operation can be found in a previous publication[9].

Figure 4., Gasification suite (not to scale).

4.

Figure 5. CRL Energy 50kw Gasifier.

Results

The results detailed here are from the first phase of the work and only provide a brief overview of the results to date. Numerous experiments have been carried out over the course of programme, most of which were repeated several times for statistical validation and for use in the modelling. Only a small representation of these results are presented here.

4.1

Bench scale gasifier

The char consumption rates of the pure matrials and blends were ascertained. It was found that the time to 20% and 50% char consumption increased with increasing temperature. This was due to the increasing temperatures enhanceing the active kinetic energy of the reactions. The fastest rate for the 100% pure materials was observed for 100% lignite followed by E. nitens > P. radiata > subbituminous coal > acid washed lignite. For the blended chars it was found that adding lignite to E. nitens or P. radiata increased the reactivities at 850 and 900oC with no change observed at 950oC. In comparison, adding sub-bituminous coal to E. nitens or P.radiata decreased the observed char reactivities at all temperatures.

The effect of alkali and alkaline earth metallic (AAEM) in lignite gasification expressed as the gas production rate (L/s) is shown in Figure 6a for H2 and in Figure 6b for CO. From these Figures it is clearly seen that with the same gasification temperature, the peak values of both H2 and CO generation from lignite gasification were much higher than that from gasification of washed lignite (AAEM-free lignite). When the curves decayed exponentially from the peak value, AAEM-free lignite gasification took approximately twice as long to tend to zero than the gasification of original lignite.

CO production in gasification

H2 production in gasification

0.002000

0.007 0.0065

H2 (L 850)

0.001800

H2 (AWL 850)

CO (L 850)

CO (AWL 850)

CO (L 900)

CO (AWL 900)

CO (L 950)

CO (AWL 950)

0.006 0.001600

0.0055 H2 (L 900)

H2 (AWL 900)

H2 (L 950)

H2 (AWL 950)

0.001400

0.0045

CO amount (L/s)

H2 amount (L/s)

0.005 0.004 0.0035 0.003 0.0025

0.001200 0.001000 0.000800 0.000600

0.002 0.0015

0.000400

0.001 0.000200

0.0005 0

0.000000

0

1000

2000

3000

4000 Time (Sec)

5000

6000

a) Hydrogen

7000

8000

0

1000

2000

3000 4000 Tim e (Sec)

5000

6000

7000

8000

b) Carbon monoxide

Figure 6. The effect of AAEM in lignite gasification.

4.2

Modelling co-gasification

Some of the initial results obtained for the modeling work are described. The model developed for char gasification was firstly solved to predict the gasification rate and producer gas composition at various operational conditions and varying coal/biomass ratios. Gas phase equilibrium was used in calculating the simulated gas components. The resultant gas profiles and carbon consumption rate are then compared with the experimental data to validate the model. Some of the initial results are shown in Figure 7a for the gasification of pure wood and in Figure 7b for pure coal.

a) 100% E. nitens

b) 100% lignite o

Figure 7. Gas production profiles at 900 C.

4.3

50kw air blown gasifier

Good regular quality syngas can be produced from the system. The use of sized coal (+3mm, -10mm) or pellets improved steadiness of operation whilst the injection of steam into the bed slightly increased the hydrogen concentration in the syngas. Typical results for the gasifier running on lignite or sub-bituminous coal and biomass pellets are shown in Table1.

Table 1. Typical syngas compositions.

Fuel 100% lignite 80% lignite – 20% P.radiata 80% lignite – 20% E. nitens 100% subbituminous coal 80% subbituminous coal – 20% P. radiata 80% subbituminous coal – 20% E. nitens

5.

% Gas H2 12 9

CO 12.5 11

CO2

CH4

15

1.5

8

10

14

1.5

11

15

12

1

14

16

13

2

11

13

14

1.5

O2 blown fluidized bed gasifier and electrolyser

The anticipated H2:CO ratio from the gasifier (without WGS promotion) is 1.2:1. In order to upgrade it to 2:1, 0.8 additional mole of H2 is required per mole of CO. Oxygen production from the electrolyser provides 2 moles of hydrogen for each mole of oxygen used in the gasifier. This oxygen is used to produce heat by combusting some of the carbon in the fuel to CO2. The CO in the syngas mostly comes from injected steam. We estimate (using earlier results from the air blown gasifier) that 4Nm3/hr of oxygen is required per 100kW HHV of fuel. This provides 8 Nm3/hr of hydrogen, which based on the above raw syngas composition fairly closely matches the amount required to increase the molar ratio to 2:1. Hydrogen surplus to this requirement can be sold. The inclusion of a modest level of WGS would allow for larger amounts of surplus hydrogen.

5.1

Design of O2 gasifier

The new fluidized bed gasifier has been designed as a 50kw unit capable of running on 100% O2, air or a combination. It has also been designed to have the capability of using different feedstock from a high percentage of woody biomass to 100% lignite or subbituminous coal. A schematic of the system is shown in Figure 9. From the experience with the initial pilot gasifier this unit has been designed in a modular form so that maintenance, repair or installation of new units is easily achieved. The gasifier will run at ambient pressure and have a maximum temperature of 1150oC. As part of the government funded research programme the O2 will be supplied from an electrolyser designed and commissioned by Industrial Research Limited and described later in this paper.

tar removal system

he a t exchanger

cyclone b ypass line

lo c k hoppe r fe e d

venturi scrubber

existing systems

g a s i fi e r gas preh e a te r

LPG preh e a te r H2 electrolyser

N2

a ir

O2

flare

- counter flow caustic wash - WGS reactor - H2 separation membrane systems

s te a m

Figure 9. Schematic of the O2 fluidized bed 50kw gasifier with integrated electrolysis.

5.2

Electrolysis Requirements

For the purposes of this pilot system, the electrolyser must be able to respond to half hour electrcity pricing signals such as might result from load-following of the power output profile of a wind farm. It is required to produce both oxygen and hydrogen at low pressure. The estimated oxygen requirement for the pilot gasifier is 4 Nm3/hr O2. High overall electrolysis plant efficiency is vital to minimize feedstock cost. In a scale up implementation, low maintenance costs and durability are also very important. Basic electrolyser theory shows that the HHV electrical efficiency of an electrolyser can potentially be above 100%. At 20oC the required energy to split the water molecule to form the hydrogen and oxygen gases is made up from 83% electrical energy (the reversible potential) and 17% endothermic heat of reaction. In existing practice all of this heat (and more) is produced from electrical losses within the cell. Most of the loss is due to the sum of the electrode surface activation energy for both gas evolution reactions. For the electrolysis reaction in alkaline electrolytes the oxygen evolution reaction (OER) provides the majority. There are a number of other loss mechanisms including electrode and electrolyte ohmic conduction and bubble blocking that particularly impact on high current performance. To achieve a very low loss design these also cannot be ignored. Recent advances in materials fabrication and processing have raised the possibility of increasing practical electrolyser efficiency, without compromising durability and cost requirements. The project intention is to demonstrate and evaluate oxygen and hydrogen production based on an electrolysis process developed from first principles with these needs in mind.

5.3

Electrolyser Technology

It is not intended that the electrolyser technology developed for this pilot scale system will be scaled up to full production plant level. If the system concept proves to be attractive and can subsequently be shown to be commercially viable, international suppliers of large electrolyser technologies would most likely be engaged. The purpose of building this electrolyser technology is two fold – similar characteristics are required for an alternative application in a different field, and hands-on experimental development is a means by which to builds a deeper understanding of the potential for of electrolyser technology in other hydrogen energy applications. The planned first step is to produce a module operating at nominal 50Vdc, which will produce 0.4 Nm3/hr O2 (0.8Nm3 H2). This module is currently going through prototype testing and the main parts are illustrated in Figure 10. The module is fully self contained and produces oxygen and hydrogen at the required quality without further clean-up. A combination of design features has been introduced to achieve a wide operating range and fast turn-up and turndown, with very low peripheral power demand. The initial prototype achieves modest efficiency of 70% HHV without any special electrode surface preparation. This module is to now be used as a platform for an efficiency enhancement programme with a target module

Figure 10. Exploded assembly diagram showing the main parts in the integrated electrolyser module

level efficiency of over 80% HHV. The 4Nm3/hr O2 electrolyser design for the full pilot plant will consist of a scale up system based on replication of the prototype module.

6.

Summary

There is a significant role for coal to play in the future eceonomy of New Zealand. The research carried out to date within this programme has enabled a significant step in the coal to hydrogen proof of concept technology package to be developed. The facility has grown and developed and become a focus for other hydrogen research initiatives such as the integrated gasifier-electrolyser system. Future scenarios encompass the laboratory as being a test-bed for other hydrogen conversion technologies and gas clean-up options.

The cost of electricity has generally been seen as a barrier to the use of electrolysis for hydrogen production. If modestly priced electricity is available and electrolysis efficiencies are improved, electrolytically produced oxygen may be a realistic alternative to other methods of oxygen production such as air separation units, which also consume substantial amounts of electricity. This may be further supported by lower carbon emission costs from the process.

In the New Zealand context, the production costs appear competitive, and the H2:CO ratio appears suitable for F-T production of liquid fuels. The effects of lignite to biomass ratio on syngas composition and process carbon emissions is as yet unknown.

7.

Future Work

The overall programme raises several interesting energy system integration tradeoffs that will be explored later in the research. If for example there is a strong need to address the greenhouse gas impact of synthetic liquid fuels, by using CCS at the gasification plant it might be possible to sequester most of the fossil component of carbon, resulting in much of the carbon in the liquid fuel being biomass sourced. If the added hydrogen is renewable electricity sourced, the liquid fuel produced becomes relatively green.

8.

Acknowledgements

We are grateful to the New Zealand government through the Foundation for Research, Science and Technology for their investment in this research. We would also like to acknowledge the support and input from our colleagues at Industrial Research Limited and The University of Canterbury.

9. 1.

References Stevens, D.J., “Hot gas conditioning: Recent progress with larger-scale biomass gasification systems,” NREL/SR-510-29952 August 2001

2.

Liscinsky, D., “Biomass gasification and power generation using advanced gas turbine systems,” Final Report United Technologies Research Centre DE-FC26-01NT41354 September 2002

3.

Gasification, C. Higman and M. van der Burgt, Elsevier 2003

4.

Robert C. Brown, Qin Liu, Glenn Norton, “Catalytic effects observed during the co-gasification of coal and switchgrass”, 1999

5.

F. Starr, E. Tzimas, S. Peteves, “Critical factors in the design, operation and economics of coal gasification plants: The case of the flexible co-production of hydrogen and electricity”, Hydrogen Energy 32 (2007)

6.

P C Hulteberg, H T Karlsson, “A study of combined biomass gasification and electrolysis for hydrogen production”, IJHE 34 (2009)

7.

Katayama Y and Tamaura Y, “Development of new green-fuel production technology by combination of fossil fuel and renewable energy” , Energy, Vol30, Aug.-Sep. 2005

8.

K Schoots, F Ferioli, G J Kramer, BCC van der Zwaan, “Learning curves for hydrogen production technology: an assessment of observed cost reductions”, IJHE 33 (2008)

9.

Anthony Clemens, Tana Levi, Alister Gardiner, Ruben Smit and Jonathan Leaver, “Development and Testing of a Coal to Fuel Cell Grade Hydrogen Technology Package and a Pathway for Hydrogen Uptake in New Zealand”, Pittsburg Coal Conference, Pittsburg, USA (2008)

Manuscript Not AVAILABLE

HYDROGEN GENERATION FROM WATER BY USING PLASMA Beycan İbrahimoğlu1, İbrahim İbrahimoğlu1, Fırat Şen2, Şahika Yürek3 , Orhan Demirel4 1

1Anadolu Plazma Teknoloji Merkezi, Ankara Vestel Defence Industry R&D Department, Ankara 3 Türkiye Kömür İşletmeleri, Ankara

2

Abstract Plasma is the key to the development of new advanced technologies of producing hydrogen from different sources – water, hydrogen sulfide, a variety of hydrocarbons (including natural gas) and even coal. Plasma processes are characterized by extremely high specific productivity (more than 100 times in comparison with catalytic processes), low metal capacity and absence of inertia, they are ecology friendly. 21’st era’s the new energy carrier hydrogen can be produced from different sources. Today though different methods have been used for hydrogen production from water; it has not been produced in great amount and cheaply. Recently, to produce hydrogen from water, plasma method is used.

Keywords: Hydrogen, Hydrogen Generation, Plasma



E-mail: [email protected]

2nd International Conference on Nuclear and Renewable Energy Resources 4-7 July 2010 Ankara TURKEY 1. Introduction Plasma is the fourth state of matter and in quasi-neutral field is formed of neutral and charged particles. It is possible to generate plasmas of solid, liquid and gasses by thermal, electromagnetic, laser and other methods. It can exist over an extremely wide range of temperature and pressure. It can be produced at low-pressure or atmospheric pressure by coupling energy to a gaseous medium by several means such as mechanical, thermal, chemical, radiant, nuclear, or by applying a voltage, or by injecting electromagnetic waves and also by a combination of these to dissociate the gaseous component molecules into a collection of ions, electrons, charge-neutral gas molecules, and other species [1]. It is more or less an electrified gas with a chemically reactive media that consists of a large number of different species such as electrons, positive and negative ions, free radicals, gas atoms and molecules in the ground or any higher state of any form of excited species [1]. Plasma is the key to the development of new advanced technologies of producing hydrogen from different sources – water, hydrogen sulfide, a variety of hydrocarbons (including natural gas) and even coal. Plasma processes are characterized by extremely high specific productivity (more than 100 times in comparison with catalytic processes), low metal capacity and absence of inertia, they are ecologically friendly. 2. Principle Of Hydrogen Generation From Water By Plasma Today, in order to produce hydrogen several methods are used [2]. However, these methods are not satisfactory to produce hydrogen in a cheap and easy way. Producing hydrogen from water is the most popular method among others [3]. In this manner, ecologically clean method, plasma is used as the hydrogen production method. Related to this topic, new test apparatus is developed from us as seen in figure 1. With the help of this test apparatus, every kind of water can be used in the hydrogen production with plasma method.

1. Propellant 2. Main Body 3. Spring 4. Valve 5. Quartz Pipe 6. Sponge

7. Anode 8. Cathode 9. Vapor Room 10. Nozzle 11. High Temperature Vapor Room 12. Magnet Figure 1. Test Apparatus

In this test apparatus, electric arc is produced between wire cathode and porous anode when electrical load is applied at this poles and very high pressure is obtained while water is evaporating. Pressurized water vapor changes state to the plasma and is ejected to the outside. This ejected matter contains hydrogen and oxygen ions and water vapor. With the help of the magnetic field, this ejected matter is splitted into the hydrogen and oxygen. 4. Conclusions

2

2nd International Conference on Nuclear and Renewable Energy Resources 4-7 July 2010 Ankara TURKEY As a result, great amount of hydrogen can be produced by using plasma method. Via this method hydrogen splitting process is fast and clean. References [1] Vijay N., Kumar, A. and Dwivedi, H. K., “Atmospheric Non-Thermal Plasma Sources,” International Journal of Engineering. 2 (1),(2008). [2] B. İbrahimoğlu, Ş.Yürek ve O. Demirel, (2009), “Plazma Yöntemi ile Kömürün Gazlaştırılması” IV. Ulusal Hidrojen Enerjisi Kongresi ve Sergisi (Uluslararası Katılımlı) 15–16 Ekim Kocaeli Üniversitesi, Kocaeli, Türkiye. [3] M. D. Kozlu, S.Aksongur, B. İbrahimoğlu, (2009), “Plazma Teknolojilerinin Hidrojen ve Yakıt Pilleri Üzerinde Uygulamaları”, IV. Ulusal Hidrojen Enerjisi Kongresi ve Sergisi (Uluslararası Katılımlı) 15–16 ekim 2009, Kocaeli Kocaeli Üniversitesi, Kocaeli, Türkiye.

3

Manuscript Not AVAILABLE

Technology and operational experience – the Shell perspective In our presentation we will discuss Shell’s experience of deploying SCGP in China, and how learning from our experience has enabled new levels of operational excellence. Up to date, Shell has sold nineteen licences in China, some of which are repeat customers. We will illustrate how we have learnt from experience on the ground and were able to make a wave of several start-ups a success: in one plant less than 12 hours elapsed between initial coal feeding and actual production of methanol products. We will show that we are not just focusing on deployment enablers for today, but continue to invest in both R&D and deployment capabilities that will enable the growth of coal gasification over the coming decades. We’re harnessing excellence in technology and design – especially in India and China – while constantly seeking out new opportunities to deploy gasification. Jay Wang Senior Engineer Shell Global Solutions International BV P.O. Box 38000, 1030 BN Amsterdam, The Netherlands Tel: +31 20 630 2024 Mobile: +31 65 209 7693 Fax: +31 20 630 3964 Email: [email protected] Internet: www.shell.com/globalsolutions

CONTROLLING THE SYNTHESIS GAS COMPOSITION FROM CATALYTIC GASIFICATION OF HYPERCOAL AND COAL BY CHANGING THE GASIFICATION PARAMETERS

Atul Sharma* and Toshimasa Takanohashi Advanced Fuel Group, Energy Technology Research Institute, National Institute of Advanced Industrial Science and Technology, 16-1, Onogawa, Tsukuba, Ibaraki, JAPAN. Email: [email protected]

ABSTRACT. Catalytic gasification of coal is an efficient way to achieve high gasification rates at as low as 700 oC temperatures. The problem of deactivation of catalyst due to the interaction of catalyst with mineral matter in the coal was overcome by using HyprCoal, an ash less product of solvent extraction process as feed coal for catalytic process. Synthesis gas is the main desirable product and its composition H2/CO ratio is important for its use in the downstream FT process. However, in a catalytic gasification process it is difficult to control the gas composition because of the effect of catalyst on water-gas shift reaction. Effect of temperature and gasifying agent composition on gasification rate and synthesis gas composition were investigated. Experiments were carried out with pure steam, pure CO2 and mixture of steam and CO2 as gasifying agents in the 600~700 oC temperature range to investigate the effects. Results showed that by adjusting the steam to CO2 ratio of the gasifying agent it is possible to control the synthesis gas composition. Effect of CO2 addition on reaction kinetics was discussed along with the calculated gas compositions. A new single step process to produce a desired synthesis gas from catalytic gasification has been proposed. KEYWORDS. Coal gasification, Catalyst, Steam, CO2, Synthesis gas. 1

Introduction Gasification is a primary conversion route to produce synthesis gas (H2 and CO) from coal1-12. Gasification is an endothermic reaction and requires temperature above 1000 °C to achieve acceptable rates for commercial application1-6. The product gas obtained at such high temperatures usually has low H2/CO (0.5~0.7) and it is also difficult to control the product gas composition due to equilibrium constraints at such high temperature. Introduction of steam at some stage of the gasifier can increase H2/CO of the product gas but only marginally (~0.9). A more common process approach is to upgrade the synthesis gas from the gasifier to the desired H2/CO (~1, 2) by water gas-shift reaction at a lower temperature (300~400 °C) in a second stage often called as sweet shift process. Catalytic gasification of coal has been widely considered as an effective mean to decrease the gasification temperature8-12. Exxon mobile1 developed a K2CO3 catalyzed steam coal gasification process at high pressure to produce methane. Nearly all catalytic gasification processes developed were to produce either H2 or CH4. Not many attempts were made to produce synthesis gas by catalytic gasification12. This was partly because alkali catalysts catalyzed both gasification and water-gas shift reaction under atmospheric pressure condition leading to formation of H2 and CO2 as main products while higher pressure favors CH4 formation. The most favored catalysts are alkali metal salts especially K2CO3

1-16

. The major

drawback in catalytic gasification of coals is the interaction of catalyst with the mineral matter (ash) present in the coals leading to the formation of compounds from which recovery of catalyst is difficult1-8,12. To overcome the problem of loss of catalyst, our research group has developed a process to remove mineral matter from coal by solvent extraction13-17. The solvent extracted coal from hereon called HyperCoal (HPC) has less than 500 ppm of ash. Because of its ashless characteristics, a catalytic gasification process for coal may be 2

developed by using HyperCoal as a feed material leading to low gasification temperature, easy recovery and recycling of catalyst. In our first study13 we reported high gasification rates at temperatures as low as 775 °C, no catalyst deactivation, feasibility of catalyst recovery and recycling, and H2 selectivity from catalytic gasification of Oaky Creek HyperCoal. In a second study14, effect of catalyst addition on gasification reactivity of HyperCoal and parent coal was compared at 700 and 775 °C and it was found that HyperCoal and coal have nearly same rates at 700 and 775 °C but at different catalyst loadings. In the following study15, gasification rates of HyperCoals prepared from three different ranks of parent coals in the 600~775 °C range were compared. In a subsequent study16, effect of catalyst mixing procedure and catalyst loading ratio on gasification rate was examined. Production of synthesis gas (H2/CO) at such low temperatures would also be an attractive application for HyperCoal. Results of our previous studies13-16 showed that irrespective of gasification temperature and coal type, the product gas from the K2CO3-catalysed steam gasification of HyperCoal contained H2 and CO2 as the major products gases with little CO and was suitable for H2 production. In the most recent study17 we reported production of synthesis gas from catalytic gasification of HyperCoal at 700 °C by changing the steam partial pressure. In this study, we report production and control of H2/CO ratio of synthesis gas by gasifying coal and HyperCoal in steam and carbon dioxide mixed environment. Coal and HyperCoal were gasified in a steam and CO2 mixed environment as a gasifying agent at 700, 650 and 600 °C with K2CO3 as a catalyst. Effect of temperature and ratio of steam to CO2 on reaction rate and on composition of the gas produced was investigated. A single step process to produce and control the composition of synthesis gas from coal at 700~600 °C has been proposed. 3

Experimental Procedures A subbituminous coal, Pasir (PAS) from Indonesia was selected for the investigation. HyperCoal production method18 has been described in detail elsewhere. Briefly, HyperCoal (HPC) was produced by the solvent extraction of the coal with 1-methylnaphthalene at 360 °C and subsequently separating the extract (HyperCoal) from the solvent. The extraction yield was 51 % for Pasir coal. The properties of the Pasir coal and HyperCoal produced from Pasir coal are shown in Table 1. HyperCoal has nearly no mineral matters. Because of almost no mineral matter, all the inorganically associated sulfur will be removed. The only sulfur in HyperCoal will be the organically associated sulfur. A detailed characterization of catalyst, HyperCoal, original coal, and chars using XRD, NMR, and SEM-EDX mapping techniques had been carried out and reported elsewhere13,15-16. The experimental setup is shown in Figure 1. Samples for catalytic gasification experiments were prepared with 50 % catalyst loading. Catalyst loading was on dry and ash free wt % basis of coal and HPC. Both dry mixing and wet mixing methods were investigated16. Catalyst mixing method has been described in detail elsewhere16. Briefly, a desired amount of K2CO3 was added on the top of a measured sample already loaded into a test crucible as solid particles and stirred with a small spatula until white K2CO3 disappears by capturing moisture from the air. Common procedure is to mix K2CO3 as an aqueous solution for homogeneous dispersion7-11 but could not be applied as HyperCoal does not mix with water. The particle size of coal and HyperCoal sample was under 75 µm. The gasification experiments were carried out with and without K2CO3 as a catalyst at 700, 650 and 600 °C with different steam to carbon dioxide (H2O/CO2) ratios as gasifying agent. Experiments were carried out in a thermogravimetric (TG-DTA 2020S, MAC) apparatus with about 20 ml/min argon (Ar) as TG carrier gas flowing from the bottom 4

(Figure 1). At the start of the experiment, 100 ml/min Ar was mixed with 20 ml/min oxygen (O2) and flowed from the top into the TG-DTA. Pre-oxidation was required for HyperCoal samples because of their extremely high swelling propensity. To keep the same experimental conditions coal samples were are subjected to pre-oxidation. In actual process, pre-oxidation may not be necessary depending on the type of feeding system. A desired amount of water was pumped by a HPLC pump to a steam generator held at 250 °C. CO2 was flowed to the steam generator as a steam carrier gas. By changing the amount of water pumped by the HPLC pump to the steam generator and the flow rate of CO2 as the carrier gas, different steam to CO2 ratio were achieved. For pure steam gasification, argon gas instead of CO2 was used as carrier gas. In case of steam and CO2 only conditions, partial pressure of steam and CO2 were kept at 0.5 atm by mixing argon gas. A 4-way valve at the inlet of the TG-DTA was used to change (Ar+O2) flow to (CO2+steam) flow. The flow lines were kept at 250 °C by using ribbon heaters. First, a desired amount of sample was heated in (Ar+O2) flow up to 200 °C and held for 5 min to remove moisture and reduce the swelling propensity of the HyperCoal by mild pre-oxidation. After 5 min hold at 200 °C, the gas was switched to pure argon flow for 60 min to remove the O2 from the reaction zone. After 60 min hold, the sample was heated to the desired temperature at 20 °C /min in pure argon. When the desired temperature was reached without any hold time the pure argon gas was switched to the preset steam/CO2 gas mixture. The steam+CO2 mixture flowing from top comes into contact with the sample in the crucible. The evolved gases flow out together with the purge gas from the side into an ice cooled tar trap to remove tar before injecting to the micro gas chromatograph (Agilent 3000A). The total gas flow rate at the outlet was measured every 3 min by a film flow meter.

Results and Discussion 5

Figure 2 shows a typical weight loss curve for HPC+50 % K2CO3 sample pyrolyzed in argon up to 700 °C followed by gasification with [50% steam+50% carbon dioxide (vol/vol)] as gasifying agent (from hereon in this manuscript steam to carbon dioxide ratio will be addressed as H2O/CO2). In a previous study16 gasification rate and gas composition was investigated at 10, 20, 40, and 50 % catalyst mixing ratio. It was reported that gasification rate was affected by the catalyst amount up to 50 % loading and above this catalyst loading rates were almost independent of catalyst amount. In addition, unlike gasification rate, gas composition was not found to be affected by the catalyst amount. Therefore, in the present study only one catalyst mixing ratio, 50 % catalyst loading has been selected. In actual process 40-50 % catalyst loadings would be very high as it may cause problem to the access of reactant gas to the reaction surface at medium to high conversion level and lower loadings such as 20-30 % would be more appropriate. All experiments were carried out at atmospheric pressure. The weight loss curve can roughly be divided into three stages; moisture removal or drying stage, devolatilization stage and fixed-carbon gasification stage. The initial weight loss (up to 200 °C) during heating from room temperature to 200 °C in O2+Ar mixture is mainly due to moisture captured from air. Pre-oxidation was done to reduce the extremely high swelling propensity of HyperCoal. After pre-oxidation stage, sample was switched to 100 % argon for 60 min. The coal/HPC conversion on dry, ash, catalyst and volatile free basis (dacvf) (from hereon called char conversion) was calculated during the fixed-carbon gasification stage by the following equation: X (char conversion, % dacvf) =

W0 − W × 100 W0 × (1 − Wash − Wcat )

[1]

where W0 is the weight when the gasification begins (db, mg) (weight at t= 45, 41 and 39 min for T= 700, 650 and 600 °C, respectively), W is weight at any gasification time (db, mg, >39

6

min), Wash is weight fraction of ash content in coal or HPC, Wcat is weight fraction of catalyst content. Results of gasification rate and gas composition only in the char gasification stage will be discussed further. Figure 3 shows the gasification profiles and gas composition of produced gas from K2CO3 catalyzed HPC at 700, 650 and 600 °C as a function of H2O/CO2 ratio. In general, at any given temperature within the temperature range investigated, rate decreased with increasing CO2 fraction in the gas mixture. Similarly at any given temperature, H2 decreased and CO increased with increasing CO2 fraction in the gas mixture. Under H2O/CO2 mixed gas environment, three reactions as shown below: C-H2O reaction (1), C-CO2 reaction (2) and water-gas shift (WGS) reaction (3) are expected to take place. While the rate of carbon loss or gasification rate can be predicted primarily by reaction (1) and (2), for proper prediction of gas composition, inclusion of WGS reaction (3) is essential. This is because WGS reaction is one of oxygen exchange reactions which are known to be catalyzed by alkali-carbon system and therefore plays a major role in determining the gas composition12. It is well known that C-H2O reaction (1) is faster than C-CO2 reaction (2)2-5. If overall carbon loss due to gasification is assumed to be the sum of individual contributions of carbon loss by reaction (1) and reaction (2), a decrease in overall gasification rate with increasing CO2 fraction in the mixed gas can be expected due to the increased contribution of reaction (2) to the overall rate. The observed gasification trend in Figure 3 (a,c,e) is in accordance with this explanation. As mentioned before, reaction (3) is essential for predicting the gas composition. It is well known1-5,7-17 and also observed in this study that under pure steam (H2O/CO2=100/0) H2 and CO2 are the main gases and very little CO is produced. This is because the WGS reaction is catalyzed by alkali-carbon system resulting in conversion of nearly all CO produced by reaction (1) to H2 and CO2. In case of pure CO2 (H2O/CO2=0/100), because of absence of 7

steam WGS reaction does not take place and only reaction (2) occurs resulting in CO as main gas produced. The little H2 produced is mainly from the inherent hydrogen in the coal. However, in a H2O+CO2 mixed gas, all three reactions take place and gas composition is determined by these competing reactions12. Ideally, if reaction (3) is the only reaction controlling the gas composition, CO produced by reaction (1) and (2) is expected to be consumed by reaction (3) as WGS reaction can not differentiate between CO from reaction (1) or reaction (2). However, the observed gas composition showed that CO is produced and H2 decreased when CO2 is introduced and fraction of CO increased with increasing CO2 fraction in the H2O+CO2 mixed gas. This may be explained by considering [1] contribution of reaction (2), [2] change in equilibrium state of the reaction (3) on addition of CO2 and [3] steam becoming rate limiting reactant with increasing CO2 ratio. As discussed before, with increasing CO2 fraction the contribution of reaction (2) to overall reaction increases thus increase in CO and decrease in H2 due to decreased contribution of reaction (1) can be expected. However, the above explanation will be valid only if the three reactions are independent. Since reaction (3) is also taking place, only reaction (1) and (2) could be considered as independent but not reaction (3). Therefore, effect of CO2 addition on WGS reaction equilibrium should also be considered. Addition of CO2 may shift the equilibrium of WGS reaction (3), which has been reported to achieve the equilibrium state above 650 °C under catalyzed conditions12, towards left thus reducing the consumption of CO to CO2 and H2. In this case one can expect increase in CO and decrease in H2 which is observed in the present study. The reason [3] is also possible as partial pressure of steam is decreased with increasing CO2 fraction. In our previous report17, we reported the effect of steam partial pressure on gas composition. However, the partial pressure of steam below which any significant change in gas composition was observed (500 years at current consumption rate) and therefore there is a strong incentive for development of efficient technologies, such as chemical looping, for power generation from lignites. Metal oxides of Nickel, Copper, Cobalt, Manganese and Cadmium have been discussed in literature for Chemical looping combustion of gaseous fuels such as natural gas and CH4 (Rubel et al., 2009). Pure Hematite (Fe2O3) and externally impregnated support metal oxides with this has been investigated for application in chemical looping mostly with gaseous fuel (Lyngfelt et al., 2008). Coal gasification experiments in a fluidized bed proved the feasibility of coal CLC with Fe2O3 as the oxygen carrier. It was observed the amount of CO and CO2 produced is consistent with the amount of coal added (Siriwardane et al., 2009; Lyon et al., 2000). Natural iron titanium oxide or Ilmenite has been investigated as oxygen carrier with South African Coal and Petroleum coke by Berguerand & Lyngfelt (Berguerand et al., 2008 & 2009). Synthetic particle of 60% active Fe2O3 and 40% MgAl2O4 has also been investigated with solid fuel by Leion et al. (Leion et al., 2008). Dennis & Scott reported gasification of lignite char under steam and CO2 atmosphere with Fe2O3 as oxygen carrier (Dennis & Scott 2009). It is evident that application of Fe2O3 as oxygen carrier for CLC of lignites has not been investigated widely. Moreover iron oxide is inexpensive in Australia and when used in CLC it is expected to generate concentrated stream of CO2 and H2. There are issues which need to be solved for CLC application of Victorian lignite with Fe2O3. Much is unknown about the rate and yield of products (such as H2, CO2, CO, Char etc.) from this process as a function of time, temperature, particle size and type of lignites. The fate of externally added Fe2O3 particles through the reaction process is unknown. The durability of oxides to sustain repeated cycles through oxidation and reduction in presence of solid residues is another area of concern. Issues related to agglomeration of solid particles over the period of time with variation in temperature during reduction in real reactor situation are also unknown. In the present work some of the issues identified above are explored. The reduction and reoxidation properties of Fe2O3 have been investigated in a Thermo gravimetric Analyser (TGA). Scanning electron microscopy (SEM) along with EDS of fresh solid reactants is compared with used reactants to understand the changes in surface morphology and mineral composition. Surface elemental information of agglomerates as a function of temperature and time are also being investigated. This paper presents preliminary results from this ongoing study.

2. Experimental section Coal samples Chemical looping experiments have been performed primarily with a Victorian brown coal - Loy Yang. The raw coal samples have been first sieved to particle size of 100-150 µm. Similar coal particle size have been used by Siriwardane et al. and Cao et al. in their Chemical looping combustion experiments with coal (Siriwardane et al., 2009; Cao et al., 2006). The sieved coal is then dried in an oven for 6 hours at 105°C. Experiments have also been conducted using a Brazilian lignite to compare the weight loss and reactivity characteristics between a high (Brazilian lignite) and low (Loy Yang) ash content coal. The Brazilian lignite used is from Santa Catarina state and prepared similar to the Loy Yang lignite for experimental purposes. Oxygen carriers The oxygen carrier used in this study is particles of Fe2O3. Fe2O3 (Ferric oxide calcined) is directly from the manufacturer - BDH Chemicals Ltd. with a purity of more than 99%. It has a maroon colour. Initially the particles were in larger lumps. It was broken and particles were sieved in size fraction of 100-150 µm (same as coal particle size) to provide a good physical mixing with coal samples. The procedure was repeated until a sufficient quantity of particles in the size range was obtained. 120 mg of Fe2O3 particles were mixed with coal particles to maintain the desired mass ratio. Samples preparation Coal was physically mixed with metal oxides for Chemical looping combustion experiments. The mass ratio used (metal oxide mass/coal mass) was 6:1which correspond to stoichiometric oxygen supply. 20 mg coal samples have been used for all the experiments. In case of Redox experiments, same amount of coal has been mixed with reacted metal oxide/ash in the TGA crucible. Ash could not be separated after each cycle and hence was accumulated after each cycle in TGA crucible. Instruments Tests were conducted using Thermo gravimetric analyser (TGA). Scanning electron microscope (SEM) was used for solid characterization. The solid residue samples were collected carefully after each test for SEM analysis.

Thermogravimetric Analyser (TGA)

The effect of consecutive reduction-oxidation cycles on the reactivity of metal oxides using various experimental conditions and gas compositions was assessed in a TG-DTA/DSC apparatus with steam injection capability. Experiments were carried out in an alumina crucible of 18 mm ID and 1 mm wall thickness with a tolerance value of ±0.5. The coal and metal oxide mixture was heated from ambient temperature to 950°C at a heating rate of 10°C/min under

different experimental environment and gas compositions. Different test runs were performed to optimize the heating rate and based on that 10°C/min heating rate has been finalized. All the experiments included a 2 hour isothermal hold at the end of the heating segment. Several test runs proved that at 950°C, after a 2 hour isothermal, the loss in weight of samples are almost constant. In case of Redox experiments, the sample was kept isothermal at 950°C for 1 h and afterward air was introduced to oxidize the reduced particles for 1 more hour. The reducing environments used were pure air and 20%CO2 + 80% N2. The metal oxide particles have been re-oxidized using air. Scanning Electron Microscopy (SEM)

The reaction residues were carefully placed on carbon tape for secondary electron imaging with energy dispersive spectra (SEM-EDX). The analysis was performed using a JOEL 840A SEM instrument.

3. Results and Discussions The composition of the dried Loy Yang brown coal and the Brazilian lignite used in this study are given in Table 1. Table 1: Component analysis of Loy Yang Brown Coal Loy Yang Coal Properties (% db) Ash

Weight (%)

1.5-1.7

Volatile Matter

50.5-51.3

Carbon

68.3-69.2

Hydrogen

4.8-4.9

Sulphur

0.4

Nitrogen

0.5

Oxygen

25

TGA Baseline test with Coal Baseline tests were conducted with Loy Yang (without oxygen carrier) by heating up to 950°C in air. Fig. 2 show the weight loss and reaction rate profile at different reaction temperature of Loy Yang coal.

0.0035 Weight Loss (%) Reactivity (S-1)

100

0.0030

0.0025

0.0020

60

0.0015 40

Reactivity (S-1)

Weight Loss (%)

80

0.0010 20 0.0005

0

0.0000 0

200

400

600

800

Reaction Temperature (0C)

Fig. 2. TGA test of Loy Yang in air. A peak was observed at 830°C for Loy Yang coal in case of combustion in air. Ash content was about 2% which means 98% of the weight loss was due to combustion in air. It is evident from Fig. 2 that coal volatilization and combustion is initiated at 250°C and proceeds until 850°C. After 860°C the weight loss is constant for Loy Yang coal. A relatively smaller peak is observed for at around 100°C which indicates the moisture release from coal. Multi cycle redox experiments using metal oxides and coal A five cycle TGA test was conducted with Fe2O3-Loy Yang coal mixtures in N2-CO2 environment to evaluate the coal combustion and metal re-oxidation process. After each reoxidation process at 950°C, reacted metal oxide/ash was mixed with same amount of coal used in previous cycles. Ash could not be separated after each cycle and hence accumulated TGA crucible. The result of redox cycles in CO2 gas are shown in Fig. 3 with Fe2O3. The extent of coal combustion and re-oxidation of metal oxides, as shown by the weight loss and weight gain respectively, during TGA tests show a small decrease for the Fe2O3-coal mixture. However, NiO shows progressively less mass loss and gain over multiple cycle operation. More over, during experiments, no mass loss of Fe2O3 particles was observed from initial value with each cycle. However, even with ash accumulation in TGA crucible, the percentage of combustion at fifth cycle was 89% for Fe2O3 oxygen carrier.

105 Reduction and Oxidation of Fe2O3 and Loy Yang

Weight Loss (%)

100

95

90

85

80

75 0

200

400

600

800

1000

1200

Time (Min)

Fig. 3. Five (5) cycles TGA redox test of Loy Yang coal with Fe2O3 in N2-CO2 environment. The maximum reactivity during each reduction and re-oxidation of oxygen carrier is plotted in Fig. 4 (A)-(B) for Fe2O3. The reduction reactivity is maximum at second cycle for the oxygen carrier tested and shows a decrease in trend in consecutive cycles. This can be attributed to the fact that the oxygen carrier particles give off oxygen more easily for combustion at second cycle due to the structural changes during operation in first cycle. But during progressive cycles the continuous ash accumulation in the TGA crucible may restrict the contact between fresh coal and the re-oxidized metal oxide. One of the major focuses of this study will be how to avoid the ash interference to fresh coal-oxygen carrier interaction in a practical system. Detailed investigation pertaining to this will be addressed in future work. The maximum reactivity in each re-oxidation cycle for Fe2O3 shows maximum value at second cycle similar to maximum reduction reactivity curve. But after that reactivity is almost constant as ash accumulation does not play any role during reaction of oxygen carrier and fresh air. 0.0060

0.0112

0.0056

-1

Reactivity (S )

Reactivity (S-1)

0.0108

0.0104

0.0052

0.0048

0.0044

0.0100

0.0040

0.0036

0.0096 0

1

2

3

4

5

6

0

1

2

(A)

3

4

5

6

Cycle

Cycle

(B)

Fig. 4. Maximum Reactivity Comparison of Fe2O3 with Loy Yang coal during Redox in N2-CO2 environment (A) Reduction (B) Oxidation.

SEM Images of fresh and used oxygen carriers Figure 5 (A) - (D) shows the SEM images of the fresh and reacted Fe2O3 oxygen carrier particles in CO2 gas composition. The images A and C are at lower magnification (1500X) for fresh and used Fe2O3 respectively whereas B and D are at higher magnification of 6000X for the same. The edge of the fresh particle is rougher as compared to the used particle that can be seen from figures A and C. The crystalline structure of fresh particle is absent from the used one rather used particle seems to be highly agglomerated and increased in size as compared to the fresh one. The fresh particle is comprised of loose smaller particles and with less porosity as can be seen from figure (B). However, the porosity increased in case of the used particle (figure D) due to oxygen transfer in multicycle operation. The melting point of Fe2O3 is about 1500°C and the temperature range selected for experiment was 950°C. So severe sintering of oxygen carrier is not expected and absent from these images. But only possibility is that at this temperature range ash particles can melt and stick with metal oxides. However Loy Yang coal has only 2% ash content (Figure 2). There is no evidence of agglomeration with ash after 5 cycle operation from images. This is further explained through EDX semi quantitative analysis.

(A)

(B)

(C)

(D)

Fig. 5. SEM Images of fresh & used Fe2O3 (A) & (B) – Fresh Fe2O3, (C) & (D) – Used Fe2O3.

EDX and Semi-quantitative component analysis of the fresh and used oxygen carrier The EDX analysis provided semi-quantitative weight composition of the major elements detected on the surface of the fresh and used oxygen carrier particles. For the fresh particles of Fe2O3, Fe and O are detected and the weight percentages are 18.18 and 81.82 respectively as shown in figure 6 (C). For used Fe2O3 particles (after five redox cycles) these weight percentages are 17.76 and 82.24 as can be seen from Figure 6 (D). This indicates that there is little ash deposition on the used particles of Fe2O3 under the conditions of the experiments. Thus there is no appreciable mass loss of Fe2O3 during the five redox cycles, a fact backed up by experimental observations (Figure 3).

(A)

(B)

(C)

(D) Fig. 6. SEM – EDX analysis of the oxygen carrier fresh and used particles

4. Conclusions The in-situ CLC/G with a low-ash Victorian brown coal was investigated in a TGA using Fe2O3 as oxygen carriers. CO2 was used as gasification agent. Fe2O3 showed a high coal combustion percentage of more than 90% under CO2 gas in a single cycle operation. No agglomeration between ash and Fe2O3 particles was observed. No mass loss of Fe2O3 was observed for the duration of the experiments. However, at 5th cycle the percentage of combustion achieved by Fe2O3 was 89% in CO2 environment. Cycle 2 showed maximum reactivity (during reduction) with a decreasing trend during the subsequent cycles. Though these initial experiments did not reveal much agglomeration between ash and oxygen carrier, longer duration experiments are required to explore this issue further. REFERENCES Adanez, J., Cuadrat, A., Abad, A., Gayan, P., Diego, L. F. de., & Garcia-Labiano, F., “Ilmenite Activation during Consecutive Redox Cycles in Chemical-Looping Combustion”, Energy & Fuels, 24, 1402-1413, 2010. Berguerand, N., & Lyngfelt, A., “Chemical Looping Combustion of Petroleum Coke Using Ilmenite in a 10 kWth Unit-High Temperature Operation”, Energy & Fuels, 23, 5257-5268, 2009. Berguerand, N., & Lyngfelt, A., “Design and Operation of a 10 kWth chemical-looping combustor for solid fuels – Testing with South African Coal”, Fuel, 87, 2713-2726, 2008. Cao, Y., Casenas, B., & Pan, W-P., “Investigation of Chemical Looping Combustion by Solid Fuels. 2. Redox Reaction Kinetics and Product Characterization with Coal, Biomass, and Solid Waste as Solid Fuels and CuO as an Oxygen Carrier”, Energy & Fuels, 20, 1845-1854, 2006. Dennis, J. S., & Scott, S. A., “In Situ Gasification of a Lignite Coal and CO2 Separation using Chemical Looping with a Cu-based Oxygen Carrier”, Fuel, Article in Press, 2009.

Diego, L. F. de., Garcia-Labiano, F., Adanez, J., Gayan, P., Abad, A., Corbella, B. M., & Palacios, J. M., “Development of Cu-based Oxygen Carriers for ChemicalLooping Combustion”, Fuel, 83, 1749-1757, 2004. Leion, H., Mattisson, T., & Lyngfelt, A., “Solid Fuels in Chemical Looping Combustion”, International Journal of Green House Gas Control, 2, 180-193, 2008.

Lyngfelt, A., Johansson, M., & Mattisson, T., “Chemical-Looping Combustion – Status Development”, 9th International Conference on Circulating Fluidized Beds (CFB-9), May 13-16, 2008, Hamburg, Germany. Lyon, R. K., & Cole, J. A., “Unmixed Combustion – An Alternative to Fire”, Combustion and Flame, 121, 249-261, 2000.

Rubel, A., Liu, K., Neathery, J., & Taulbee, D., “Oxygen Carriers for Chemical Looping Combustion of Solid Fuels”, Fuel, 88, 876-884, 2009. Siriwardane, R., Tian, H., Richards, G., Simonyi, T., & Poston, J., “Chemical-Looping Combustion of Coal with Metal Oxide Oxygen Carriers”, Energy & Fuels, 23, 3885-3892, 2009.

2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11-14, 2010 Program Topic: Underground Coal Gasification (UCG) BLOODWOOD CREEK UCG PILOT 2008 – 2010 Presenting author Dr Cliff Mallett Technical Director Carbon Energy Limited Level 12, 301 Coronation Drive, Milton QLD 4064/ PO Box 2118 Toowong DC, QLD 4066 AUSTRALIA +61419753719 [email protected]

Co-author Burl E. Davis UCG Process Consultant Carbon Energy Limited 245 Lynn Ann Drive New Kensington, PA 15068 412-793-6058 [email protected] EXTENDED ABSTRACT BLOODWOOD CREEK UCG PILOT 2008 - 2010 Carbon Energy Limited is a listed Australian public company focused on underground coal gasification. It uses technology developed by CSIRO, the Australian Government scientific and industrial research organisation. Carbon Energy has been producing syngas from its pilot UCG plant at Bloodwood Creek in southeast Queensland since October 2008. It is completing the construction of a 5MW power station to be fueled with UCG syngas. It is developing a second UCG project near Valdivia in southern Chile. Since its formation in 2006, Carbon Energy has worked to bring a small demonstration project into production. In its first year it explored its coal permits and selected a site for a UCG trial. After another year of site investigations and design development, a UCG panel and associated surface plant was constructed, and ignited in October 2008. During a 100 day experimental period the UCG reactor was operated under a variety of conditions and injection gas types including air, oxygen and steam. Following this the reactor was operated in low flow mode on air injection, through 2009 and 2010. Decommissioning of panel 1 and clean up were commenced in August 2010.

The Bloodwood Creek site is located approximately 200km west of Brisbane in the Surat Basin. The region has limited rural industry developments, mainly pastoral, native forest or feedlots with soils unsuitable for intensive agriculture. Groundwater is unsuitable for human or animal consumption or for irrigation. The UCG pilot panel and surface facilities are constructed in a cleared paddock beside Bloodwood Creek. Underground operations are surrounded by three series of monitoring wells. The first set of operational monitoring wells are located within 10m of the UCG reactor and panel. The second set of sentinel wells are located approximately 50m from the UCG reactor. The third set of monitoring wells are environmental compliance wells located on the boundary of the permitted UCG pilot area which is approximately 1km2in extent. Carbon Energy has developed a UCG panel design that can produce 1PJ of syngas per year, and will operate for 4-5 years. The panel is defined by two horizontal boreholes 30m apart, in the base of the coal seam, that travel in the seam for 5-600m. These horizontal boreholes turn in and intersect a vertical ignition well. The coal is ignited at the base of the ignition well, and after the burn is established, this well is sealed and air/oxygen is injected down one of the horizontal wells, and the product gas is extracted from the other horizontal well. Hot gases move from the injection well to the product well across a face of coal, that is progressively gasified, causing the face to retreat as in a conventional longwall mine. At Bloodwood Creek, the coal is 200m deep and the seam is around 10m thick. The pilot was commenced on the 8th October 2008, initially with air injection. In January 2009 the injection gases were changed to steam and oxygen, and an experimental program carried out with these gases throughout January. Following these experiments the panel continued to operate with low flow air injection. On Day 140 of the trial the UCG cavity grew to engulf monitoring well No. 6, which had been placed in the centre of the panel. This confirmed the position of the reactor and provided data on ground conditions. Temperature sensors had been installed in Well 6 at points 60m above the coal, two at 12m above the coal, and within the coal. These showed that sensors 12m above the coal started heating 10 days before burn through, and continued to heat until they both failed at around 100oC, 10 - 12 days after the burn through. The sensor 60m above the coal has shown no heating. The sensor in the coal was very slightly affected the day before burn through, but showed no significant rise in temperature until a few hours before it was consumed by the face advance. The rate of face advance appears to be faster than the propagation of heat in front of the face, so there is no significant heating in the coal in advance of the cavity. Temperature from a cross-section through the reactor cavity shows that heat migration is mainly in the roof above the reactor, but heating does not reach to 60m above the coal. Temperature sensors in the coal at 10m from the reactor cavity showed no heating for almost a year. The reactor grew to very close to the 1 series of monitoring wells placed to the west of the reactor, and sensor 1M gradually heated from ambient of 26 oC to 70 oC between July 2009 and August 2010. Well 1L which was cased through the coal seam showed heat effects that indicated that the margin of the reactor cavity had affected the

casing and monitoring equipment. Well 2M, 10m to the east of the reactor cavity showed a temperature increase of less than 2 oC over the 22 months of operation of Panel 1. A 3D seismic survey was conducted to locate faults in the coal and allow the design of UCG panel layouts. The coal seam is relatively flat with faults of 20 to 40m throw spaced at approximately 1km intervals. Experimental processing of the seismic data was carried out by Peters et al 2010 to see if the UCG cavity could be identified in the seismic records. This found that the RMS amplitude showed an anomaly centered on the UCG cavity. While the anomaly was centered over the cavity, neither its outline or extent precisely coincided with the position of the cavity based on borehole location and the Well 6 burn through data. The results are very encouraging however, and represent the first example of geophysics pinpointing a UCG reactor, and it is hoped that further analyses will refine the predictive capacity of this method. The key operating requirement to safely operate a UCG reactor is the control of the reactor pressure to less than surrounding hydrostatic pressure. This allows the surrounding water to constrain reaction byproducts and prevents the escape of potential pollutants. Water pressures are continually monitored in the wells around the reactor to ensure that these conditions are maintained. If reactor pressure is allowed to fall too low, excess water runs into the reactor, and it is possible to create a draw down cone of depression around the reactor, that cannot be recharged quickly enough by regional recharge. A balance is required that allows sufficient water to flow into the reactor at all levels to prevent escape from the reactor, but to let only enough water flow in that will not depress the hydrostatic pressure surrounding the reactor. UCG is new as a commercial operation and is yet to achieve acceptance from the public or regulators. A recent event in Queensland has illustrated the potential for problems when other stakeholders are not fully informed on UCG activities. A UCG company, Cougar Energy, had commenced a pilot burn near Kingaroy in SE Queensland, but discontinued the trial for operational reasons after a short period. A sample from their monitoring wells provided a high contaminant level result which was subsequently established as a laboratory error. The original result became known and used by local activists to create public concern and the State Government shut down the UCG operation and constrained other UCG operators far from Cougars operations. After a month of extensive testing around all three UCG operations, the Government announced that it could find no indications of anything associated with the UCG operations that were a threat to people or stock. However the damage to the industry has been done as it is now associated with poisoning of farm water supplies, despite all the Government evidence that it is not. Carbon Energy is aiming to demonstrate consistent syngas production that can fuel a commercial 5MW power station. This can be done with one Carbon Energy UCG module, and a power station is nearing completion at the Bloodwood Creek site. This power station is using reciprocating engines fuelled with syngas, and will feed the electricity into the local electricity network.

Future plans are to increase power generation in stages, next with a 25MW unit and subsequently up to 300MW. The highest value can be extracted from UCG syngas through using it as a feedstock for chemicals, and it is planned to ultimately construct an ammonia and synthetic natural gas plant at Bloodwood Creek. UCG is an attractive development option for Queensland’s abundant coal deposits as it extracts most energy from the in situ coal resource, and creates rural industries in the on site conversion of syngas into high products such as fertilizers and explosives. Coal seam gas competes for access to the coal resource, but can extract only one twentieth of the energy compared with UCG operations. Although UCG is still in its infancy as a commercial venture in western economies, it holds the promise of much higher recovery of in situ resources, it avoids the safety and environmental problems of conventional coal mines, and provides the energy in a conveniently utilized form for a range of applications such as power, chemicals and liquid fuels, and allows for economic capture of carbon dioxide, an essential for a future clean coal future.

Manuscript Not AVAILABLE

Studies on Gasification of Turkish Lignite via Underground Coal Gasification Project Director Şahika Yürek , M.Sc.Mining Engineer Kıvanç Het, Directorate of Turkish Coal Enterprises (TKİ), Yenimahalle, Ankara, Turkey [email protected], [email protected]

ABSTRACT It is known that with nearly 12.3 billion tons of reserve coal, Turkey produces nearly 22% of its electricity needs via the burning of these reserve fuels. TKI, Turkey’s largest mining organization and the sixteenth largest national company, owns a nearly 2.5 billion ton share of this reserve. TKI, engaged in extracting coal from mines for heating and thermal purposes, is also engaged in studies directed towards utilizing coal for alternative purposes, following the latest technological advances in the world. One of the spearheading studies is geared towards obtaining synthetic gas via the underground gasification of coal. The process of underground coal gasification is an operation in which steam, composed of air and water, is directed with the assistance of injection wells into underground coal deposits and the resulting gas from this reaction is directed to the surface by conduction channels. This technique aims to render feasible what has been a difficult and costly means of production for accessing reserves in the sea and at great depths. Furthermore, this technique would render possible the above-ground processing of obtained gas into H2, CH4, NH3, CO, and CO2 for a variety of applications. However, these are not the only advantages of underground gasification of coal; this process also aims to contain CO2, one of the key factors in the level of greenhouse gases, within the areas from which coal is extracted from beneath the surface of the earth. It has been widely accepted that the process of underground coal gasification is a means of clean coal production which minimizes environmental impacts with low costs and high productivity. Within the scope of these studies TKI has carried on contacts with both local and foreign firms, and the most rigorous study in this area has been conducted under the guidance of the leading expert in the field, the American firm Lawrence Livermore Laboratories. If a suitable site replete with the necessary conditions for an experimental initial run involving the underground gasification of coal in Turkey is identified by the advising firm, then plans will be laid for the undertaking of larger scale underground gas production domestically. In conclusion, without a doubt this technique could once again bring to the fore those lignite coal resources of ours which are currently unexploited, which entails the financial

inactivity of those sites. Within the scope of Clean Coal Technology, this technology would be an important step in ensuring a prudential means of meeting of our future energy needs.

Underground Coal Gasification and Applicability to Thrace Basin Lignite in Turkey

Ayşe Yıldırım, Advisor, Turkish Petroleum Company [email protected] Serdar Doğan, Chemical Engineer, Turkish Petroleum Company sedoğ[email protected]

Key Words: UCG, IGCC, Gasification

Abstract As it is known that, hydrocarbon reserves have been declined rapidly and with other problems for energy demand, there have been expanded and renewed interests in new alternative technologies in worldwide. Underground Coal Gasification (UCG) is the one of those technologies for new energy sources. UCG is a gasification process carried on low calorific value, non mined or no minable coal seams due to the geological conditions (high fracture frequencies, volcanic, complex storage/tectonic structures). UCG process converts coal in situ into product gas (syngas) by using oxygen/ steam mixture (air, enriched air, oxygen/water and carbon dioxide/oxygen). Here the coal beds react as a chemical reactor, thus gasification process is maintained underground rather than conventional gasification methods. In this process; coal, steam and oxygen are brought together to the combustion temperature for coal by adjusting the amount of oxygen carefully, the coal is not completely burned but decomposed chemically. The process is a partial oxidation rather than combustion. The resulting mixture (carbon monoxide, hydrogen, carbon dioxide, methane) is UCG gas (syngas) can be used Integrated Gasification Combined Cycle (IGCC) configuration as a supplement and substitute fuel for electricity generation and chemical synthesis resulting in manufacturing of synthetic liquid fuel or chemicals by FisherTropsch process. Turkish Petroleum Company (TPAO), has a vision being an energy company and taking in the commission to assess the low calorific value lignite reserves in Turkey by clean coal technologies under the roof of Ministry of Energy and Natural Resources (ETKB) conjunction with General Directorate of Mineral Research and Exploration (MTA), Turkish Coal Enterprises (TKI), Electric Power Resources Survey and Development Administration (EIEI), Electricity Generation CO.INC (EUAS), ETI MINE Works General Management (ETI Maden), Turkish Hard Coal Enterprises (TTK) and General Directorate of Mining Affairs (MİGM) . In this review, geological, geophysical and chemical studies have been done on the lignite beds, which were determined during natural gas drilling in Thrace Basin in Turkey and their applicability is being discussed to UCG processes.

Introduction The defining feature of global energy markets remains high and volatile prices, reflecting a tight balance of supply and demand. This has put issues such as energy security, energy trade and alternative energies at the forefront of the political agenda worldwide. Energy consumption growth slowed in 2007 compared with 2006, it was still above 10-year average for the fifth consecutive year. World primary energy consumption increased by 2.4% in 2007-down from 2.7% in 2006, but still the fifth consecutive year of above-average growth. Globally, oil consumption rose, but at the weakest rate of all the fossil fuels, reflecting the pressure from high prices. The oil price has been on an upward path for more than six years now. Global oil consumption grew by 1.1% in 2007. Natural gas consumption rose by3.1 in 2007, slightly above the 10-year average. Coal, seen as affordable and locally produced in many parts of the world, was the fastest growing fossil fuel for the fifth year in a row. Renewable energy remains a small share of total global energy use, but most renewable sources experienced rapid growth in 2007. Ethanol output rose by 27.8%. Global capacity for wind and solar electricity generation

grew broadly in line with average of 28.5% and 37%, respectively. The ongoing growth in fossil fuel consumption suggests that global carbon dioxide emissions are still rising. Turkey primary energy consumption increased by 2.9% in 2009. Turkey has limited reserves of oil and natural gas, but proven reserves of lignite in the order of 8.4 billion tones. Combustible renewable, especially wood and the country’s water sources are the important indigenous energy sources. Turkey should use its domestic lignite sources in industry and power generation. In this paper, it is recommended that UCG will be a new technology for new energy and power generation source for Turkey.

Underground Coal Gasification Process Description Underground coal gasification converts coal in-situ into a gaseous product, commonly known as synthesis gas or syngas through the same chemical reactions that occur in surface gasifies. Gasification converts hydrocarbons in to a synthesis gas at elevated pressures and temperatures and can be used to create many products (electric power, chemical feedstock, liquid fuels, hydrogen). Gasification provides numerous opportunities for pollution control, especially with respect to emission of sulfur, nitrous oxides and mercury. UCG could increase the coal resource available for utilization enormously by gasifying otherwise non minable deep or thin coal reserves. In the process gas is produced and extracted through wells drilled down into the coal seam, to inject air or oxygen to combust the coal in-situ, and to produce the coal gas to the surface for further processing, transport or utilization. The process relies on the natural permeability of the coal seam to transmit gases to and from the combustion zone, or on enhanced permeability created through reversed combustion, an in-seam channel or hydro-fracturing.

Underground Coal Gasification

The overall chemistry underlying coal gasification processes is well understood. There are two main chemical reactions. 1.Oxidation Almost all the oxygen from the blast is depleted upon contact with the coal Oxidation zone should be kept as short as possible Oxidation zone should spread out laterally from the injection point C + O2 → CO2

- 393.5 kJ mole-1

C + ½O2 → CO

- 110.5 kJ mole-1

CO + ½O2 → CO2

- 283

kJ mole-1

Both of these reactions are exothermic (generates heat) 2. Reduction The hot gas leaves oxidation zone and passes towards the reduction zone. into contact with hot char. C + CO2 ↔ 2CO

CO2 is reduced to CO as it comes

+172.3 kJ mole-1

With the introduction of steam, H2 is also formed C + H2O ↔ CO +H2

+131.4kJ mole-1

Both of these reactions are endothermic (consumes heat) Too much water (groundwater ingress) can quench the system

Key Considerations for UCG 1.

Thickness of coal seam >1m – any thickness (in theory)

2.

Depth of coal seam >300m

3.

Rank of the coal 200-500m

5.

Proximity to built up areas >500m (In UK)

6.

Proximity to mines both active and historic >500m

Why Consider Underground Coal Gasification UCG has numerous advantages over conventional underground or strip mining and surface gasification: 

Reducing operating costs, surface damage and eliminating mine safety issues such as mine collapse,



Coals that are un minable (too deep, low grade, thin seams) are exploitable by UCG, thereby greatly increasing domestic resource availability,



No surface gasification systems are needed, hence capital costs are substantially reduced,



No coal is transported at the surface, reducing cost, emissions, local foot print associated with coal shipping,



Most of the ash in the coal stays underground, thereby avoiding the need for excessive gas clean-up, and the environmental issues associated with fly ash waste stored at the surface,



There is no production of some criteria pollutants (SOx, NOx) and many other pollutants (mercury, particulates, sulfur species) are greatly reduced in volume and easier to handle.



UCG eliminates much of the energy waste associated with moving waste as well as usable product from the ground to the surface UCG compared with conventional mining combined with surface combustion, produces less greenhouse gas and has advantages for geologic carbon storage.

Potential Limitations and Concerns for UCG Even though UCG has a number of advantages, the technology is not perfect and has several limitations. 

UCG can have significant environmental consequences: aquifer contamination, and ground subsidence,



While UCG may be technically feasible for many coal resources, the number of deposits that are suitable may be much more limited because some may have geologic and hydrologic features that increase environmental risks to unacceptable levels,



UCG operations cannot be controlled to the same extent as surface gasifies. Rate of water influx, the distribution of reactants in the gasification zone, growth rate of cavity are the important process variables.



The economics of UCG has major uncertainties,



UCG is inherently an unsteady-state process, flow rate and the value of the product gas will vary over time.

Comparatıve Costs and Fınance for UCG 

Large-scale UCG with power generation (300MWe) undertaken remote from the gasification site has a generation cost comparable to, and possibly less than, Integrated Gasification Combined Cycle (IGCC) technologies.



However, small scale developments (~50MWe) are not likely to be economically viable as standalone project



UCG with CCS has a power generation and capture cost which lies between two estimates for IGCC with CCS and is comparable with gas turbine combined cycle (GTCC) with CCS (assuming postgasification capture).

Turkey Lignites Almost half of Turkey’s energy production is represented by coal (43% lignite). The lignite deposits are widespread with reserve estimates of more than 8 million metric tons. Turkish Coal/Lignite reserve is shown in the following graph.

Turkish Coal/Lignite Reserves

Potential Issues of Shallow Turkish Lignite for UCG 

Almost of all the Turkish lignite is Tertiary age.



Shallow lignite is likely to cause surface subsidence which may lead to gas escape.



Lignite with a moisture content >40% may be considered unsuitable due to quenching of oxidation zone.



Shallow lignite may be situated close to valuable groundwater bodies and UCG may be considered too marginal due to groundwater contamination issues.



Careful selection of lignite reserve necessary to avoid coal with low calorific value.



However- these issues may be reduced or mitigated by careful site selection.

Applicability of UCG to Thrace Basin Lignite There are seven onshore and four major offshore basins in Turkey. The onshore ones are called SE Turkey (Anatolia) Basin, Thrace Basin, Adana Basin, Tuz Gölü Basin, East Anatolia Basin (including several subbasins) and the onshore Black Sea Basin (Zonguldak and Sinop). The offshore basins can be named as; Black Sea, Marmara Sea, Aegean and Mediterranean Sea. The most active onshore basins are, as far as the exploration and production concern; firstly SE Turkey (Anatolian) and Thrace, secondly Adana and Tuz Gölü Basins, whereas SE Turkey (Anatolian) Basin is known to be the oil prone and the Thrace is the gas prone ones. All the oil and gas fields are located in these two basins where the total production reaches 50M bbl for oil and 60MM SCFD for gas. The European side of Turkey is called the Thrace Basin. It is the largest and thickest Tertiary sedimentary basin in Turkey. The basin is triangular shaped, trends WNW-ESE and was formed by extension in late Middle Eocene to latest Oligocene times. This gas prone basin is located northwest of Istanbul, Turkey. In basin About 9,000 meters of Eocene-Miocene marine clastics and continental sediments were deposited. Several basement faults were reactivated and underwent strike-slip motion in late Miocene times where the famous North Anatolian Fault bounds the southern part of the basin. The Tertiary sedimentary succession, overlying Paleozoic to Mesozoic metamorphic basement, comprises interbeded fine to coarse grained clastics from a variety of depositional environments, turbidities, muddy carbonates with local reef developments, river channels and tuffs. Thrace Basin is one of the most important basins because of coal and hydrocarbon potential. The basin is surrounded on the north by the Istranca, on the west by the Rodop, and on the south by the Menderes Massifs. Coal explorations have been conducted by General Directorate of Mineral Research and Exploration (MTA). important coal formations in the basin arewithin Danişmen Formation of Oligocene age. Danişmen Formation has a lithology including gray-green claystone, sandstone, conglomerate, tuff and lignite, and was named as lignitic sandstones in the region by fırst studies. The lignite occurences have mostly been developed at lagoonal, deltaic, flood plain and lacustrine marshes but economical coal occurrences take place at delta plain, lagoonal and lacustrine marshes. Investigated samples show that the huminite group, the coals are seen to be abundant of huminite maceral group as well as gelinite maceral. From the vitrinite reflection values, the coals seem to be classified as sub bituminous in rank and the coals were determined to have deposited in limnic-paralic environments. There are lignite occurrences on foothills of Istranca Mountains nort of the Thrace Basin, and these are mentioned as Saray (Edirköy, Safaalan, Küçükyoncalı), Vize (Topçuköy), Kırklareli and Demirhanlı fıelds in general, and Keşan, Malkara, Uzunköprü and Meriç fields in the south. Lignites exposed at N and S of the basin gradually deepen toward center of the basin, and reach to 11 000 meters of sedimentary sequence in central parts of the basin and extend to depths below 600m.

Geological Map of Thrace Basin is shown in the following graph.

GEOLOGİCAL MAP of THRACE BASIN

VAKIFLAR

Chemical and Geological Properties of the lignite reserves in Thrace Basin are: 

The lignites are in Tertiary age,



The coal thickness is between 30-50m,



The average coal seam depth is 200-400m,



Lignite are almost with a moisture content >40 %.



Lignite reserves are in low calorific value.

In Thrace Basin with these properties the UCG process might be done under following conditions: 

A geological/hydro geological model should be build for an existing resource,



The size of the deposit should be in excess of 100 million tonnes



If all parameters are successful, pilot site selection and plant should be considered.

References: 1.

Aiman, W.R., R.J. Cena, R.W. Hill, C.B. Thorsness, 1980, Highlights of the LLL Hoe Creek No.3 Underground Coal Gasification Experiment., Lawrence Livermore National Laboratory, Livermore, CA. UCRL-83768.

2.

Ambrozic, T, and Turk G., 2003, Prediction of subsidence due to underground mining by artificial neural networks Computers an&Geosciences 29,627-737

3.

Blinderman, M.S., and Jones, R.M., 2002, The Chincchilla IGCC Project to date: UCG and Environment, 2002 Gasification TECHNOLOGİES Conference, San Francisco, USA, October 27-30, 2002.

4.

Şengüler, İ. (2003) Öz Kaynaklarımız İçinde Linyitin Yeri ve Önemi. Türkiye 9. Enerji Kongresi Bildiriler Kitabı, 59-67, İstanbul.

5.

Uysal, B.Z., (2008) Temiz Kömür Teknolojileri. Türkiye 16. Kömür Kongresi, Bildiriler Kitabı, 335340, Zonguldak.

CO2-Reduction through Biomass co-firing in Coal Fired Power Plants Dr. Roland Jeschke*, Dr. Klaus-Dieter Tigges*, Dr. Alfred Gwosdz*, Alfons Leisse* *

Hitachi Power Europe GmbH, Schifferstraße 80, 47059, Duisburg, Germany

1

Introduction

Facing the threatening background of global warming the impetus for more and more activities to reduce CO2 emissions is given. Biomass co-firing, being regarded as CO2 neutral, has a potential to reduce CO2 emissions generated by mankind. This particularly applies for power generation based on fossil fuels. Besides that operators of utility boilers are always interested in reducing production cost with a special focus on fuels. Biomass is deployed increasingly in power stations as wood pellets, chips and others. Moreover most of the new power stations being under construction right now envisage biomass combustion in the short or mid term. Hitachi Power Europe has been involved in co-firing for a couple of years starting with grinding of wood pellets. In consequence the development of biomass firing technology was extended to cover the entire value chain from storing and grinding up to firing a wide range of biomass products. This technology can be applied for both lignite and hard coal fired steam generators. It is based on HPE’s well proven firing technology for hard coal and lignite and was refined under full scale conditions in a 35 MW test facility. The combustion tests provided a comprehensive set of operating data to evaluate the technology and validate the combustion models especially adapted to biomass firing conditions. This paper outlines the biomass technology for co-firing woody biomass in hard coal or lignite fired steam generators up to an amount of 100% and shows possible applications taking different biomass features into account. Results of combustion tests are described and calculated and measured data are shown.

2

Overview

Considering co-firing solid biomass in coal-fired steam generators an overview containing the relevant aspects of secondary fuel storage, grinding, feeding, injection or admixture as well as the burner technology is depicted in Figure 1. Depending on the chosen co-firing design the feeding or injection type may vary. This is indicated by the multiple arrows starting from the storage icon and pointing to various positions between the grinding and the burner. The options for co-firing design are depending on the main fuel used, because this determines the boiler design of the respective power plant. The boiler design itself determines which biomass fuel is most appropriate for being co-fired, because it affects the residence time of the particles in the burner chamber. Furthermore the change of burner and/or mill configuration from design coal case to biomass configuration may be complicated. The choice depends on the design and the dimensions of the existing boiler on the one hand and the physical and chemical characteristics of the biomass fuel on the other hand.

Page 1 of 12

Figure 1

Overview of relevant aspects for co-firing biomass in coal fired boilers

In the following Figure 2 some of the options for feeding biomass into a coal-fired power plant are shown. Option 1. (BL) stands for direct injection via burner for a complete burner level. Option 2. shows the implementation of separate burners. In option 3. (B) the biomass is admixed to the pulverized fuel piping close to the burner. Option 4. (S) is realised through the admixture on to the conveyer before the mill and option 5. (S) accounts for pre-conditioning e.g. pre-crushing, carbonisation,…

Figure 2

Feeding/Injection types for biomass co-firing

The bracketed (S) after options 4. and 5. stands for ‘small’, indicating that the selection type may also be determined by the amount of biomass to be co-fired. Examples: for co-firing less than 10% of full thermal load the options 4. and eventually 5. can be applied. If one needs to get up to 100% of full thermal load on a burner level option 1. has to be chosen. With this option the possible biomass heat input depends e.g. on burner geometry, because the used burner type may not be capable / suitable of firing biomass fuel with considered amounts. An amount of 10-100% of full thermal load on a single burner needs option 3. (B). As mentioned before the options applicable depend on the fuel characteristics of the chosen biomass fuel as well. A high moisture content (e.g. forest residue), a high ash content (e.g. sewage sludge), a high content of corrosive elements (e.g. straw), the existence of elements that deactivate the DENOX catalyst (e.g. phosphorus) or the tendency to increase slagging and fouling (e.g. via ash fusion temperature) are important restricting factors. Page 2 of 12

This paper deals with wood pellets (WP), wood chips and sawdust. Concerning the restricting factors depending on the fuel properties these three avoid some of the limiting properties discussed above and are therefore suitable for co-firing in a large coal-fired steam generator.

3

DS®-Burner technology

The DS®-Burner technology from Hitachi Power Europe GmbH, used for hard-coal fired steam generators, is able to burn up to 100% biomass. The burner configuration as well as the mill configuration of the MPS mill is easily adaptable to a biomass mode. The DS®-Burner is of swirl staged type and has a concentrically design. A basic set up is shown in Figure 3. DS®-Burners mainly consist of the following items: • Ignition equipment containing light oil gun and high energy sparking rod. • Core air tube. • Primary air tube with integrated fuel nozzle, adjustable swirl vanes and impact elbow at the inlet. • Secondary air tube with adjustable swirl vanes and air deflection throat. • Tertiary air nozzle with adjustable swirl vanes. • SA / TA wind box with connection to the air supply.

Figure 3

DS®-Burner basic burner set-up

DS® burners are focussed on the local definition of the ignition point considering pyrolysis and oxidation processes. Oil gun and sparking rod are guided inside the core air tube and the core air tube is centrically arranged in the pulverized fuel tube. The p.f. tube is connected with the coal line through an impact elbow at the inlet. The impact elbow and the impact table located on the core air tube take care for highly uniform distribution of the pulverized coal in the primary air tube. Page 3 of 12

Function of DS®-Burner Adjustable swirl vanes generate a rotary primary air/coal flow resulting in particle concentration at the outer circumference of the pulverised fuel tube. Before leaving the burner the fuel particles are slowed down by collision against the inner fuel nozzle ring. As a result the flow velocity of the particles is reduced below the fuel’s characteristic flashback velocity by means of the impact on that ring. For inducing and maintaining the ignition process a sufficient amount of oxygen is available in the primary air stream only. The impact of the pulverised fuel particles on the fuel nozzle ring results in reflection and acceleration in flow direction again by diversion into the primary air flow, which discharges into the furnace. Devolatilisation of the pulverised fuel occurs in a controlled way under low-oxygen conditions, so that generation of nitrogen oxide is already severely restricted in the flame and thus the total emission is noticeably reduced. The stable swirl flow formed by the secondary- and tertiary- air which surrounds the core flame causes oxygen enrichment in the peripheral burner zone and thereby ensures an oxygen rich flue gas atmosphere near the furnace walls. Firing 100% Biomass Based on this DS® Burner experience, a special version of this burner was developed that is capable of dealing with fuels being fed in a dense phase i.e. 4-10 kg pulverized, dried fuel per kg carrier gas. This burner type deals with pulverized dried lignite and biomass and is the suitable burner for high fuel concentration in the centre. A further special feature of this burner is its extended control range. A basic set-up is shown in Figure 4. The core air tube is shortened to allow the core air being mixed with the primary air having a high dust load. This happens just before the inner fuel nozzle ring, where the particles entrained in the primary air flow are slowed down by collision with this ring. Helical baffles on the outer diameter of the core air tube support the rotation of the primary air/coal flow. Secondary and tertiary air handling is as for the DS® burner.

Figure 4

DS® T Burner basic set-up

Page 4 of 12

3.1

Testing and refinement in 35 MW test facility

A series of tests with different fuels using the DS® as well as for the DS®-T-burner were carried out at the 35 MWth test facility of the Centro Combustione Ambiente (CCA, an ANSALDO Caldaie company) in Gioa del Colle, Italy. Figure 5 shows a photo of the burning chamber. The natural circulation type boiler has a horizontally shaped combustion chamber. The burner is installed at the front wall. The combustion chamber is partially refractory lined in order to obtain similar heat transfer and flame temperatures as in a real boiler. The fuel supply can be controlled in different ways depending on whether the firing is direct or indirect. The direct system includes a coal silo, a ball mill, a rotating classifier and the PA fan. In case of indirect firing, grinded fuel is transported by a pneumatic conveyer. The flue gas is cleaned from dust by a bag house filter before leaving the chimney. Observation ports are located on the left side wall of the combustion chamber at various distances from the burner. Water-cooled probes can be introduced here into the flame perpendicular to the flow axis. In this way it is possible to measure the temperature distribution at different distances to the burner and at different distances from the centre of the flame. In the test series the concentrations of flue gas species as O2, CO and NOx as well as the release of hydrocarbons were measured in the flame. All air flow rates are measured separately and adjusted by means of control dampers. Figure 5

35 MWth combustion chamber at the Centro Combustione Ambiente (Ansaldo Caldaie company). Length app. 12 m, width app. 4.3 m, height app. 6 m (w/o hopper), ash hopper 2.5 m

The following Figure 6 clearly shows a stable flame and an early ignition very close to the burner even at a sawdust firing rate of 86.6 % for the DS®Burner. The sawdust is characterized by a Residue of 2.6 % on the 1000 micron sieve and 90.7 % on the 90 micron sieve. The stability of the biomass flame over a wide load range assures good conditions for low NOx emission figures as well as low loss on ignition (LOI) figures.

Figure 6

View of the flame of a DS® Burner burning 86.6% sawdust Page 5 of 12

3.2

Calculated data

Figure 7 and Figure 8 show the temperature field (dimension Kelvin) of a DS® – Burner operating with 50% hardcoal and 50% WP or 100% coal as fuel, respectively. The figures show similar temperature field contours so that the well proven fuel range for the DS®-Burner is extended for co-firing biomass that fulfils the fuel characteristics of WP.

Figure 7 50% Coal and 50 % Wood Pellets [K]

Figure 8

100% Coal [K]

Figure 9 and Figure 10 display the 3D temperature field comparison for 50% WP co-firing and for pure hardcoal combustion (dimension degree Celsius). As depicted, the temperature fields look similar for both cases.

Figure 9 50% Coal and 50 % Wood Pellets [°C]

Figure 10

100% Coal [°C]

As shown in the figures, the different behaviour of combustion of coal and biomass and the different fineness of both fuels on the other hand are well compensated to a large extent by the DS®-Burner characteristics. Due to the stable ignition of the fuels by the DS®-burners and the controlled combustion in the furnace, the simulations show that co-firing of biomass is admissible in a wide range. The predictions meet reality for the test facility very well. Co-combustion in utility boilers show the same trends as predicted.

Page 6 of 12

4 4.1

Grinding technology MPS mill

The MPS coal mill, usually deployed for grinding hard coal, can easily be modified into a MPS wood pellets mill, including explosion suppression and safety system for bio-fuel. Reference for this type of modification is AMER 9 in the Netherlands. The MPS wood pellets mill can be re-modified into a MPS coal mill in a couple of days, if desired. For AMER 9 PS, the quantity of WP co-firing is up to 15% of full thermal load with one of six mills, the WP net calorific value was 16.9 MJ/kg, the water content 9.6%ar and the ash content 2.12%ar. The biomass is completely fed into one burner level which means that this burner level runs on 100% biomass combustion. The modification of the MPS mill includes installation of explosion detection and suppression systems, employing relatively cold primary air, carbon monoxide detection and quick-acting isolation systems as shown in Figure 11. As an example explosion detectors mounted within the mill body are capable of triggering a set of extinguishing systems located throughout the mill. Furthermore a double wall is necessary to let the biomass particles be carried out in an annular gap providing sufficient effective lift velocities to prevent unnecessary recirculation of biomass back into the mill.

Figure 11

Sketch of MPS-mill modified for wood pellet operation

Page 7 of 12

Function of MPS mill A motor drives a grinding bowl with exchangeable grinding bowl segments via a gear. The material centrically introduced from the metering hopper onto the rotating grinding bowl through a discharge chute is carried to the grinding track by centrifugal forces and rolled over by the three fixed grinding rollers. The grinding force is a function of the mill load and is set using the hydro pneumatic grinding force system. The process of combined drying and pulverizing follows the airflow principle. The hot primary air enters the grinding chamber around the grinding bowl through the nozzle ring to take over drying and transportation of the pulverized material to the electrically driven lamella classifier (static classifier shown in Figure 11) arranged above the mill housing. Separation of the coarse grain from the fine grain takes place at this point. The dust-air flow rising from the pulverizer is deflected once it has passed the gravity screening/cleaning zone. Using a dynamic classifier a radial acting force is transmitted through the rotating lamella cage against the particles in the bearing air stream which deflects the coarse dust particles from their present flow direction and repels them outward. There, these coarse particles get into a downward directed flow field that transports them back centered into the grinding bowl for post-pulverization via the funnel-shaped oversize material return flow. The separative capacity of the dynamic lamella classifier is determined by the variable rotational frequency of the rotor. For static classifiers it is adjusted by positioning the classifier vanes. The finished dust leaves the classifier via several single dust pipes leading to the burners.

4.2

DGS mill

The DGS mill is usually deployed for grinding of lignite. Depending on the Amount of Biomass to be (co-) fired the feeding/injection can be carried out by direct feeding into the mill by adding the crushed/grinded biomass to the coal on the conveyer for the DGS mill. Up to about 5% of full thermal load biomass share may be added. DGS technology involves the raw coal and flue gases being initially put through a beater section for precrushing. There will be no further pulverisation of the wood fibres in the mill. This ensures excellent air and coal dust allocation into the beater wheel. As a result, there is often no need for any classifier something both raising the pressure balance and cutting back on energy requirements. Figure 12

DGS Mill

Page 8 of 12

4.3

Edge mill The Edge Mill is a vertical spindle mill, see Figure 13, and belongs to the group of 'Applied Force Mills'. It may be used for pre-crushing wood chips to a size, which can be fed to e.g. a DGS mill. The mill is equipped with a lower die, which, depending on the mill size, is connected to one or two drives. Above the lower die grinding rollers are arranged which are stationary and are pressed to the lower die hydraulically. When the die is set to rotation it forces the grinding rollers start crawling on top and overrunning the wood chips, which have been fed in-between. The wood chips are pressed through the stencil and are reduced to small pieces. This procedure allows the wood chips not only to be cut parallel but also transverse to the wood fibres. That is not possible with a beater mill. Figure 13

4.4

Edge Mill by Fa. Kahl

Comparison of fineness measurements

For comparison the fineness achievable for wood chips pre-crushed in an edge mill and of WP after treatment in a MPS mill is shown in the next table. These Residues taken from measurements show the need for utilizing a burn off grate for the co-firing of wood chips. [micron] 6300 4000 2000 1000 500 200 125 90 63

Chips after edge mill [%] 2.5 25.2 51.1 76.4 94.9 97.2 97.7 97.9

Pellets after MPS mill [%] 0.9 14.9 54.8 85 92 95.5 97.4

Table 1 Fineness of pre-crushed wood chips and wood pellets milled by MPS mill

5 5.1

Possible Applications for co-firing woody biomass Co-firing with hard coal

For a hard coal fired boiler, usually using roller mills and swirl burners, the biomass fineness requirement is challenging, because there is no grate in the hopper. For this case, an appropriate biomass is wood pellets. These pellets consist of compressed sawdust particles, which can be disaggregated in roller mills. Page 9 of 12

Figure 14

Wood pellets

As shown in Table 2 the ash content of the WP is rather low, the content of VM is very high and the sulphur content is low. This influences UBC, NOx and SO2 emissions directly. The slagging and fouling tendency of the WP shown beneath are different; the WP2 have low melting temperatures, so that the risk of slagging is much higher than for WP1 pellets. NCV Water Ash Volatile matter S Cl N IDT Softening Temp. Hemisph. Temp. Fluid Temp.

[MJ/kg] [%] [%ar] [% ar] [% ar] [% ar] [% ar] [°C] [°C] [°C] [°C]

WP 1 18,07 4,85 0,30 80,14 0,05 0,04 0,17 1430 1465 1470 1490

WP 2 17,46 9,02 0,86 75,51 0,05 0,03 0,13 1167 1170 1172 1177

Table 2: Exemplary properties of wood pellets

Applying option 1. (BL) shown in Figure 2 ‘direct injection via burners of a complete burner level’ to a hard coal fired boiler, i.e. 100% biomass on one burner level means running a mill totally with WP. The conversion of such a mill to enable pellet combustion involves measures and solutions for dosing to the milling system, storage, optimisation of the firing system, heat and mass balances as well as CFD calculations. This option was realized at AMER 9 PS.

Storage concept for admixture to p.f. piping For firing from 10% up to 100% biomass in a hard coal fired boiler option 3 (B) according to Figure 2 is most suitable. Admixture to pulverized fuel piping can be done e.g. by 3 way valves or by direct injection with non-return valve installed before inlet. Here is the challenging question how to store and transport the pulverized fuel to the burners. In the following example a dense phase transport to each burner is considered. The biomass is injected in a straight-lined pipe section near to the DS®-Burners. Page 10 of 12

The storage concept for this example may look as shown in Figure 15: • Pneumatic conveying of the wood dust to the burners at a high dust load (blue lines). • The carrier-air mass flow for the biomass remains constant for all loads but the dust loading changes with the variation of the quantity of wood. • Wood dust feeding to the burners is made - using rotary piston blowers, - from silos, - via dosing rotary seals, - and , - via conveying lines to the burners. • A silo and a blower supply one burner level. • The feeding into the combustion chamber is made via the DS® Burners.

Silo

Rotary seal

Piston blower

Figure 15

5.2

Mixing device

Storage concept according to Sil & Atex for dense phase transport

Co-firing with lignite

Considering a lignite fired boiler the biomass fineness requirements are not as challenging as for the hard coal case. This is mainly because of the burn off grate. Coarse particles which are not lifted by the upstream gas flow fall onto the grate and incinerate. For this case, especially concerning the less challenging fineness requirements, wood chips instead of WP can be chosen for co-firing. Despite of the afore mentioned grate the chips have to be crushed to smaller peaces before they can be burned. The resulting fineness of the comminuted chips is considerably coarser than for WP. That means the altered combustion behaviour, especially regarding burnout time, has to be taken in consideration. Wood chips, according to Figure 16, have different shape than WP. The water content is also higher and ranges usually between 15% and 50%, so the NCV can vary between 5 and 18 MJ/kg. This influences the appropriate type of mills. A suitable feedstock size for co-firing with lignite is G30 according to Ö-Norm 7133, where the main fraction is 2.8 mm to 16 mm with 60-100 % wt. The coarse fraction has dimensions of more Page 11 of 12

than 16 mm to a maximum of 85mm with maximum 20 % wt. share. The fine fraction consists of particles smaller than 2.8 mm with a maximum of 20% wt. share. Less than 4 % wt. share have dimensions smaller 1 mm. This nearly corresponds to the definition of class P16 in Prenorm CEN/TS 14961. The size of the crushed wood chips will typically be in the range of some millimetres. Therefore a sufficient residence time in the furnace must be ensured or a burn off grate has to be available, which is true for lignite fired units. Applying option 5. (S) ‘Pre-conditioning’, i.e. pre-crushing in this case, and option 4. (S) ‘Admixture before mill on conveyer’ to a lignite fired boiler allows an amount of 35% of full thermal load to be co-fired. Figure 16

Wood chips

For biomass co-firing with lignite the HPE-DGS mill is used with upstream Edge Mill for biomass pre-grinding. A defibration plant, consisting of edge mill and silos, is used for crushing wet wood chips which are stored in silos before feeding on the coal feeder that supports the DGS mill. The finished material is transported mechanically to the boiler house and is supplied to the coal feeders. Laying on top of the coal both fuels will be conveyed together to the mills. The feeder is equipped with a belt weigher to determine the coal mass introduced into the pulverizer. The mill supplies coal and biomass to the boiler. This process is well suited for wet raw material. A drying process is not necessary. On the contrary, the higher the water content, the lower the power required for grinding. As a result the energy consumption of the total process is relatively low. To ensure the wet condition of raw material the minimum water content is limited to 20%. For wood chips with water content between 20 - 30% a humidification of the woodchips is intended. Another advantage of the high water content of the finished property is that there are no problems with ATEX.

6

Outlook

The technology for grinding and co-firing biomass with coal is available. Different types of biomass can be co-fired. The amount of thermal input determines the options for grinding and co-firing that should be chosen. 100% biomass firing is possible for a complete burner level. Several options and applications were shown. It obviously makes sense to use this experience and design and built Power plants capable of firing up to 100% biomass. It could be expected that new plants to be build are able to account for a broad spectrum of biomass. Limitations arise mainly from the properties of the biomass chosen which lead to corrosion, slagging or fouling problems.

Page 12 of 12

2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010

VENTILATION AIR METHANE ABATEMENT AT CONSOL ENERGY’S ENLOW FORK MINE Richard A. Winschel, Director Research Services, CONSOL Energy Inc. 4000 Brownsville Rd., South Park, PA, 15129 USA [email protected], 412-854-6683 Deborah A. Kosmack and William P. Fertall, CONSOL Energy Inc. Jerry Gureghian, Green Holdings Corp.

Abstract: CONSOL Energy Inc. and Green Holdings Enlow, Inc. (a subsidiary of Green Holdings Corp.), are developing one of the largest coal mine ventilation air methane (VAM) emission abatement project in the United States at CONSOL's Enlow Fork Mine in southwestern Pennsylvania, USA. The status of the project is discussed herein.

Background and Introduction: Coal mining, and particularly coal mine ventilation air, is a major source of anthropogenic methane emissions, a greenhouse gas that is 21 times more potent than carbon dioxide (CO2). Globally, VAM emissions from coal mines amount to approximately 300 million tCO2e (metric tonnes of CO2 equivalent) each year. Oxidation of methane to carbon dioxide and water reduces its global warming potential by about 87%. The VAM abatement equipment to be installed at the mine, based on regenerative thermal oxidation (RTO) technology, will capture and destroy methane released during the mining process that would otherwise escape to the atmosphere through the mine's ventilation system. The project is designed to reduce the mine's VAM emissions by the equivalent of 190,000 tCO2e each year. It is expected that the system will be operational in early 2011. Green Holdings will supply the capital, will operate the unit and will be responsible for selling the emissions reduction credits. CONSOL will provide the ventilation air, the site and technical support.

Description of Mine and Bleeder Fan: Enlow Fork Mine is an active underground coal mine located in Washington County, Pennsylvania, USA, approximately thirty-five miles southwest of Pittsburgh. The topography is rolling hills typical of the Allegheny Plateau region of North America. Enlow Fork Mine produces approximately 11 million short tons (10 million metric tonnes) of clean coal per annum from the great Pittsburgh Seam primarily via the longwall mining method (two separate longwall systems are employed), and it is one of the largest underground coal mines in the United States. The ventilation systems in CONSOL’s longwall coal mines use “bleeder” fans to ventilate each longwall section. The bleeder fans draw and exhaust air from the mine through vertical shafts that connect the surface and the underground workings. The bleeder shafts at the mine are

drilled from the surface with a finished diameter of 6 to 8 feet (1.8 – 2.4m), and each one provides ventilation for one of the two longwall mining systems at the mine. At Enlow Fork Mine, typical bleeder fans will exhaust about 300,000 standard cubic feet per minute, scfm (510,000 standard m3 per hour) of air. This ventilation air will contain methane at a typical concentration of 0.6% to 1.0% by volume. Bleeder fans will have a typical life of six to eight years before the mining district they were designed to ventilate is permanently sealed. Figure 1 shows a typical bleeder fan layout at the Enlow Fork Mine with a proposed methane emission abatement system installed. The methane abatement system will be installed on the area where the materials removed during the drilling of the bleeder shaft have been placed and covered with earth; this area is known as the cuttings pit, and it provides a large, flat and level space for the installation in a region where flat, level space is rare.

Description of Regenerative Thermal Oxidation (RTO) System: The RTO plant is designed to safely destroy at least 96.5% of the very low concentrations of methane gas (1.4% or less) contained in high volumes (140,000 to 300,000 scfm or 240,000 to 500,000 m3 per hour) of ventilation air exhausted by underground coal mines, commonly termed ventilation air methane, or VAM. The plant design, based on proven emission control technology, is composed of three individual RTO units, each capable of processing 47,000 scfm or 80,000 m3 per hour of VAM with methane concentrations between 0.5% and 1.2%. Each RTO unit is equipped with: -

two heat-exchange ceramic media beds an oxidation chamber (with propane burners) an inlet/exhaust valve housing direct drive supply fan exhaust stack temperature and methane sensors

The RTO units are arranged in an array connected to the mine evase by a 100-foot-long (30 m) duct, which captures no more than 80% of the mine fan’s output, see Figure 1. At one end, a collection box, designed not to restrict the exhaust of the ventilation air from the mine, connects the duct to the mine evase. At the other end of the duct a damper shuts off all VAM supply to the RTO array in the event methane concentrations exceed the operational limits (0.3% to 1.4%) of the RTO units. The damper is activated by methane sensors located within the collection box. When within operational limits, the VAM is safely carried through the duct to each RTO unit with the assistance of the RTO supply fans to overcome the pressure drop caused by the heatexchange beds. The VAM enters the RTO through the first media bed to the oxidation chamber, which is preheated to 1800 °F (1000 °C) using the propane burners, which are shut off after preheating. The methane contained in the VAM oxidizes in an exothermic reaction, which, above 0.3%, generates sufficient heat to maintain the chamber at 1800 °F (1000 °C). The VAM

exits the oxidation chamber via the second media bed where the air is cooled to 480 °F (250 °C) as heat is transferred to the ceramic bed. The directional flow of VAM is alternated every two minutes to prevent overheating of the media beds and to maintain the temperature of the oxidation chamber within operational limits. Data from methane sensors located at the inlet of the RTO and in the RTO exhaust stack are used to determine the actual quantity of methane which has been destroyed and serves as the basis for claiming the carbon credits which will be issued to the project. Once operational, the RTO plant is expected to function with minimal human intervention although a modem programmed to call or page a remote service center or personnel to broadcast any upset condition.

Permitting: In order to construct, install, and operate the equipment, several permits are required from various government agencies. The authorizing agencies include local, state, and federal groups. There are five organizations that must either approve or inspect the system before it is permitted to be constructed or operated. The status of each of the permits is discussed below. The permit required at the federal level is an addendum to the Mine Ventilation Plan. It is submitted and approved through the U.S. Mine Safety and Health Administration (MSHA). A modification to the existing plan which details the safety features of the abatement system is required. This plan is also reviewed by the Pennsylvania Department of Mine Safety, which is responsible for inspecting the equipment before it is operational. The addendum was formally submittal to MSHA District 2 on August 9, 2010. The existing Mining Activities Permit, issued by the state of Pennsylvania, also required a modification because the project changes the activity on the mine site. The installation of the equipment will remain within the existing permitted area and there will be no impact to road access and water drainage, therefore the modifications to the permit were minor. On July 14, 2010, the Bituminous Coal Mining Activity Permit including the installation of a ventilation air methane regenerative thermal oxidizer on the exiting bleeder shaft was officially approved by the Pennsylvania Department of Environmental Protection (PA DEP) Bureau of Mining and Reclamation. A Plan Approval and Operating Permit from the PA DEP Bureau of Air Quality is required before the system can be operated. The focus of this permit is to evaluate the impact of the operations on air quality. Although the abatement system is an air-cleaning device that will reduce the emissions of methane entering the atmosphere, there will be some minor emissions of nitrogen oxides and carbon monoxide. The application was submitted on June 23, 2010. A Conditional Use Permit is required from the local municipality, Morris Township, where the bleeder fan is located. The focus of this permit is to assure that the project will comply with local ordinances and minimize disturbances such as noise levels and increased traffic. On July 6, 2010, the Supervisors of Morris Township granted the Conditional Use Application for a

ventilation air methane abatement system to be operated as an adjunct facility in connection with the Enlow Fork Mine bleeder shaft.

Status and Plans: Green Holdings issued a request for price to a pre-screened selection of RTO manufacturers in the first quarter of 2010. After all the permits for the project have been issued, Green Holdings will notify the selected equipment supplier and they expect the RTO Plant to be operating at the Enlow Fork site within 22 weeks of such notification. The project is expected to begin generating credits in the second quarter of 2011.

Figure 1. Conceptual layout of bleeder fan site with VAM abatement system installed.

Pittsburgh coal conf. 2010, 11-14 October 2010, Istanbul, Turkey Program Topic 4 – Carbon Management, SESSION 9 - GHG management strategies and economics – 2

NOVEL METHODS OF COAL SEAM GAS CONTENT DETERMINATION FOR ESTIMATION OF GREENHOUSE GAS EMISSIONS FROM MINING Abouna Saghafi CSIRO Energy Technology, P.O. Box 52, North Ryde, NSW 1670, Australia Email: [email protected]

Coal seam naturally contains greenhouse gases, dominantly methane but also carbon dioxide and to a lesser extent higher hydrocarbons. With coal mining most gas volumes trapped in coal seams and strata are liberated which mostly end up in atmosphere. In order to assess fugitive emissions from mining new methods are devised. In situ gas content of coal seams is a primary parameter required in these methods. Traditionally the purpose of gas content determination in coal mines has been the safe operation of mining and workers. Over the years various methods of measurement have been developed for ungrounded mining. However, these methods are not always adequate for requirement of emissions assessment for greenhouse gas inventory and new definition of gas content and more accurate measurement methods are required. The liability of the coal producer or coal user or both in relation to the emitted and remaining gas in coal can be a major factor for the way gas content is defined and measured. In order to accommodate the future emission trade scheme (ETS), one of the first steps is to debate the definition of gas content and developing more adequate methods of measurement. In this paper the authors looks at possible definitions of gas content in relation to the purpose of its use and suggests novel methods of its determination. Keywords: Coal seam gas, gas content, greenhouse gas, fugitive emissions, methane, carbon dioxide

INTRODUCTION Coal seams are high capacity gas reservoirs and to some degree, most coals contain some gas. The mine gas consists generally of methane (CH4) and carbon dioxide (CO2) which are partly or totally released during mining. Mining leads to large disturbance of coal seam reservoir as the fracturing develops both in coal and rocks. Gas escapes to the atmosphere via fractures and exposed coal surfaces. In 2009 the annual greenhouse gas (GhG) emissions from anthropogenic sources in Australia were about 539 Mt CO2-e. About 40 Mt CO2-e of these were fugitive emissions. The coal mining is the main source and represents more than 70% of the total fugitive emissions. The emissions are mostly due to release of the

1

Pittsburgh coal conf. 2010, 11-14 October 2010, Istanbul, Turkey Program Topic 4 – Carbon Management, SESSION 9 - GHG management strategies and economics – 2

trapped coal seam gas which is liberated during and after the completion of mining. The intensity of emissions depends on flow properties of strata and diffusivity and matrix permeability of coal. The method of mining affects the extent and density of induced fractures. Permeability could increase by several orders of magnitude. Moreover, fracturing of strata causes the discharge of water from mining area leading to further increase of permeability and acceleration of gas desorption from coal. Though the rate and intensity of gas release from mining at a given time is primarily a function of temporal gas content and flow properties of coal and strata, the total volume of gas liberated over the life of mine is a function of the virgin, pre-mining magnitude of gas content of coal seams and gas trapping strata. Novel methods are currently being developed to estimate GhG emissions from coal mining. For the case of open cut mining, where the emission is diffuse, direct measurement is difficult and in many cases virtually impossible. In a novel method, developed by author for Australian mines (Saghafi, 2010), the emissions estimate is based on a number of parameters related to the lithology of the strata and coal seam properties. Among these parameters the in situ gas content of coal is of primary importance. Accurate measurement of gas content is therefore primary for reducing the uncertainty of estimation. The appropriate definition and corresponding measurement method of gas content can influence, sometime largely, the magnitude and accuracy of emissions estimation. While measurement of low gas content may not be of any importance to safety issues in underground mining it is quite important for accurate estimation of GhG emissions from both open cut and underground mines. In the next sections after discussing various modes of gas storage in coal various definitions of gas content are provided; this is followed by description of the standard method of the gas content determination and proposed method for low gas content testing. Error associated with the new method will be discussed.

MECHANISMS OF GAS STORAGE IN COAL AND GAS CONTENT Coal is a porous rock where gas can be stored in the large space available on the pores’ internal surface. Gas is stored both in free and adsorbed phases. The adsorbed phase is held on the pore surfaces and constitutes the greatest portion of the stored gas for shallow to medium depths (?5( " & ' # " # 3;5(

' && " $ + ! # ') " ! # 2+ ) # " #

$

)

& #

(

$ " ! ! &

$#

" ' && "

"

! &( 3A5B

$# + '

$&

)

#$" " "

#$" "

# " ( $ ! +) $ ' C " + ) ) )& " &

$# + " " ! # & # $ # ! " " # " ' ( 4( C " ( ! & # 0 !( 0 0 C " !0 #' ) "

! # &

"

$" # $

$ + 4 . &. → $& $ + &. → $&. $& + 4 . &. → $&. , . + 4 . &. → , .& $& + , .& → $&. + , . $ + , .& → $& + , . $ + $&. → .$& $ + . , . → $, 6 $, 6 + , .& → $& + 9, .

$ C "

" "

1!

*$ & ) $! " *( 1@2+ # #

" + 47>4; ! " " &

< + 4;>.; ! $ $+ #$"

#& &&$

; ! #

!

' &

" &

"

( "$&

+'

!

!"

)

&"

$ # #$"

(

!

" !) 0 C

+ &

3@5(

$# '

'

$

&

+D

& " $ & &

"

&& '

#$" # 2

!$&

$

"

"

+

!

) #

"

0

'

#$" # !

"

& (

#$" " ! ! 0 " # (

#

' -

$ #

C

! #!0 #' ! ) & ! !(

*$ & ) $! "

)& & #(

#

" 0 ! " "

B

142 1.2 192 162 1;2 1D2 1@2 1?2 1A2

" "

" ! 1C $" # " "

#$" ' # ) & " & ! # ! !0 #' C " !0 #' ! #$" $& " !

" !)$ " + # " " ' " +@ 2 " + #A " !

"

! ' "

345B ' + *( 1?2

$!)

"

#) ' ; #? < #$ # *$ & ) $! # " ! " (

-

-

1C 2+ *( 1;2+ )& 4 !

"

$

" +?

" "

!

4( " ! "

'

"

$ C + *( 1D2+ < #$ #+ 677 # 4;77 E(

"

$

B; #

)& 4( *$ & ) $! " $ E+ 1E2 *( 1;2 677 6(7;0479 D77 .(@70474 ?77 6(760477 4777 4(9?0477 4;77 9(@047-4

!

" " & !# # " )& . $ # K # $ I) & %

)& .( 182 182 182 182 182 182 $ 182 L M 1E" &/% 2 M 1E" &/% 2

E +'

*( 1D2 @(@047-44 ;(4047-; 6(6047-. .(D.0477 D(7?047.

345 E +! *( 1?2 .(AA047; A(.60474 4(960477 A(D047-. .(;047-9

" & !# .@(46 .(7; 47(4D 7(D4 4(9D ;(6. ;9(.D .7;D .6@;

!

$ # !# .6(?. .(?. 6(4@ 7(D. ;(9; 4D(;? 6;(D6 ..?6 .DA?

#

&

$& J& #

(

&

%

& "+

K " & )& .(

'

# !

# " $% $ ! $%!" &' 9.(6? ;D(?A .(@A 6(.? 4.(A4 @(A@ 7(DD .(4D 7(?? 4(6@ D(66 49(@. 69(?6 49(;4 .DA6 ;6.6 97AD ;@.?

) '

# '

)

"

E +) *( 1@2 ;(.047-46 4(A047-D 4(4047-. 4(A70477 4(D.0479

#$" 0 ! #$ # " # ! $% # & , # & ) $% & ( " & ( & F -4 # & F -. ! # " & % ! G $HI$ ! $ I) & % H &F # " ( & F -. & "

"

(

"

& 0 # "

'

&

> *$ ! # &(

+

' # & # &$ ! # & *$ & #$ $ > # & ! #( L < ' +" # ! !, -! # $ #( & '" ' " &"$& " ! ! & ! # " #

! ! " " &+ " & + " !$& ! # & " &"$& " # ( C " + 4 # .+ # !$! &$ ) + ( $ # 0 " #+ ! # & $& & # & &$ ) " "% #( " &# " "

!

! ! !

" ! ) $ # $ + /$ & + " ! ! # ./ " " ) " && # ) #=$ ! + # 4 $ C " + !

" + ! # & ! $" ! " &"$& ( ' ! + C " " + !(

*$ ! + # ( 4 . && ' & # !" ! $

"

*$ N # L 3600C) varies from 8-33 % of total liquid product. The CO conversion varies from 73-95%. The selectivity of C5+ hydrocarbon at 500 h-1 GHSV of FT-Cat 12, FT-Cat 12A and FT-Cat 14 were 77, 72 and 66% respectively. It has been observed that CO conversion increases and C5+ hydrocarbon liquid selectivity decreases with increase in space velocity of feed. The C5+ Hydrocarbon Liquid selectivity is maximum at low space velocities, which may be due to larger residence time. Also, as per SIMDIST result, IBP of liquid hydrocarbons of product varies from 70-75 0C and FBP varies from 400450 0C. As per the yield and selectivity pattern of the developed catalyst system it can be emphasized that these may be successful formulations leading to commercialization for converting the huge Natural Gas Reserves of our country and adding up in the current oil pool. First time in India, 50 liter/day capacity pilot plant has been set up wherein dried Babool wood was gasified in a 2100 kg/hour feed gasifier, giving rise to real gases containing mainly Hydrogen 19.1%, Carbon Monoxide 17.2%, Nitrogen 50.4% along with other gases viz. methane, carbon dioxide and oxygen. The gaseous mixture thus produced is cleaned so as to free from moisture and oxygen. Additional Hydrogen has been added to the product gas to make the ratio of Carbon Dioxide to Hydrogen was 1:2. This synthesis gas was then converted into transportation fuels in Fisher-Tropsch Process using catalyst in a fixed bed reactor at the desired temperature. Typical Products

19

obtained were: Straight run Gasoline 48-50%, Jet Fuel 21-25% and Soft Wax (Pharmaceutical Grade) 1-10%. Reproducible results were obtained. This process was technically established but its economic feasibility, catalyst life etc. are yet to be ascertained. Further, in the SSRC Subcommittee, a project titled, “Development of Coal to Liquid Technology” has been cleared. For the CTL technology it has been targeted to develop suitable catalysts and system/process for converting the coal derived syngas into liquid hydrocarbons specifically the middle distillate which is a mixture of diesel and kerosene.

Conclusion:

In order to avert climate change due to atmospheric green house accumulations, global emissions ought to be significantly reduced. The use of renewable fuels offers the potential to rectify this problem as carbon cycles can be closed and additional carbon dioxide emissions avoided. Currently Biomass is being used to produce power and heat. As transportation accounts for a significant share of CO2 emissions the use of biomass should be extended beyond heat and power production. Through biomass gasification transport fuels such as gasoline and diesel can be produced to replace existing transport fuels. While the latter represents proven technologies, the production of liquid fuels via Fischer Tropsch synthesis has received growing interest.

Different process concepts for the production of Fischer Tropsch transportations fuels from Biomass were designed & evaluated economically & ecologically. The considered concepts were a small scale plant, a stand alone large scale plant & a large scale plant directly connected to petroleum refinery. The essential advantage of the small scale plant was the nearness to the producing area of the biomass which avoided long transport routes. The gasification systems of these plants are presented, as well as their gas cleaning and conditioning concepts Fischer Tropsch Synthesis, reactors, product separation & upgrading & off - gas utilization. The economic analysis showed that under the given condition the transportation fuels produced by this method from the small scale plant were the most expensive while both types of large scale plant could produce transportation fuels in a more economic manner. The ecological analysis demonstrated that the biomass derived fuel could reduce the output of Greenhouse gases drastically. References: 1. Middle Oil distillates from Fischer-Tropsch process; V.A.Krishna Murthy, A.N.Basu, N.G. Basak, A. Lahiri; Journal of Scientific & Industrial research: 21A (7), 338 (1962). 2. Conversion of coal to oil by Fischer –Tropsch process; Samiran Basu, V.A.Krishna Murthy; Seminar proceedings, Need for coal based chemical Industries, The Inst. of Engineers, 5-6 Dec. 1981, p 48-55.

20

3. Semi Pilot Scale investigations of FT synthesis for conversion of syn gas to Middle Distillates and Wax.; Samiran Basu, G.C.Nandi, S. K.Mitra; Chemical Engineering World, 29, (1), 123-125, 1994. 4. Design, synthesis, and use of cobalt-based Fischer-Tropsch synthesis catalysts; Enrique Iglesia; Applied Catalysis A: General 161 (1997) 59-78. 5. The Fischer–Tropsch process: 1950–2000; Mark E. Dry, Catalysis Today 71 (2002) 227–241. 6. GTL Technology-Challenges and Opportunities in Catalysis; P. Samuel; Bulletin of The Catalysis society of India, (2), 2003, 82-99. 7. Fischer-Tropsch synthesis: relationship between iron catalyst composition and process variables Burtron H. Davis; Catalysis Today 84 (2003) 83–98.

21

THE EFFECT OF COAL COMPOSITION ON IGNITION AND FLAME STABILITY IN CO-AXIAL OXY-FUEL TURBULENT DIFFUSION FLAMES

Dadmehr M. Rezaei, Ph.D. Candidate, Department Of Chemical Engineering, University Of Utah Email: [email protected] Yuegui Zhou, Associate Professor, Shanghai Jiao Tong University, Shanghai, China, Email: [email protected] Jingwei Zhang, Ph.D., Department of Chemical Engineering, University of Utah Email: [email protected] Kerry E. Kelly, Research Associate, Department of Chemical Engineering, University Of Utah Email: [email protected] Eric G. Eddings, Professor, Department of Chemical Engineering, University Of Utah Email: [email protected] Jost O.L. Wendt, Presidential Professor, Department of Chemical Engineering, University Of Utah Email: [email protected]

Abstract Past research on flame stability and stand-off distance under oxy-coal combustion conditions has used a 100kW pulverized coal test rig with a co-axial turbulent diffusion burner, and has been described at previous Pittsburgh Coal Conferences. These studies were for one specific coal, namely a Utah Bituminous Coal. The purpose of the research described in this paper is to extend the previous work and to explore how coal composition changes affect the following dependencies that control flame stand-off distance and flame ignition, namely: 1) the effect of partial pressure of oxygen (PO2) in the primary stream with differing preheat temperatures in the secondary stream; and 2) the effect of PO2 in the secondary stream with zero O2 in the primary stream. The results of this new study were designed to extend previously obtained knowledge on effects of secondary preheat temperature, turbulent mixing, PO2 in various streams, from one single coal to other coals of differing compositions. This paper, therefore, explores the effects of coal composition on ignition in oxy-coal, coaxial, turbulent diffusion flames. In this research, the stability and stand-off distance of the flame were studied for the following three types of coal: Utah Skyline Bituminous, Illinois #6 Bituminous, and a Powder River Basin (Black Thunder) coal. To this end we investigated: 1) the effect of PO2 in the primary stream, 2) the effect of PO2 in the secondary stream, and 3) the effect of preheat temperature in the secondary stream, on flame stand-off distance, using the same photo-imaging methodology described elsewhere. The results of the ignition and flame stability analysis for these three coals under oxy-firing conditions are compared, and the effects of coal composition are elucidated.

1. Introduction The increase of green house gases, and particularly carbon dioxide, for energy production has resulted in new technologies with lower emissions of NOx, and SOx. These techniques are capable of complying with carbon dioxide capture and sequestration. Oxy-Fuel Combustion technology has been suggested as the shorter remedy for the conventional power plants.

In Oxy-Coal combustion, coal burns with pure oxygen instead of air, and the combustion gases are diluted using the recycled flue gas (RFG) which essentially contains CO2 and H2O. The gas mixture has higher emissivity and subsequently greater heat transfer, at the same time a lower volume. Previous study at the University of Utah on Utah Skyline Bituminous in 100 kW vertical oxy-fuel combustor explained the effect of PO2 in the primary stream at different preheat temperature in the secondary stream [1]. Also, this work provided information on the results of increasing of PO2 in the secondary stream without O2 in the primary stream, and flame stability. The result of the study was presented in the 2009 Pittsburgh Coal conference. However the influence of coal composition which is one of the key criteria of combustibility on this set of experiments is not well known. This research explores the effect of coal composition on combustion behavior in the similar oxy-fuel operating conditions. The present study was conducted to compare the ignition behavior of three coals representing a spectrum of rank and coal composition in oxy-coal combustion by 100 kW pilot scale combustor with a co-axial turbulent diffusion burner.

2. Experimental 2.1. Coal selection and sample preparation Three types of coal with different rank have been used in this study. The ultimate, approximate, and ash analysis of the prepared coals is provided in Table 1 and Table 2 where coals are listed by increasing rank as determined by carbon, and volatile matter. In order to see the effect of moisture clearly the PRB (Black Thunder) was chosen as one of the samples.

Table 1 LOD

Ash

C

H

N

S

O (by diff.)

Volatile matter

Fixed carbon

HHV BTU/lb

Utah Skyline

3.18

8.83

70.60

5.41

1.42

0.53

13.21

38.60

49.39

12606

Illinois #6

9.65

7.99

64.67

5.59

1.12

3.98

16.65

36.78

45.58

11598

PRB (Black Thunder)

23.69

4.94

53.72

6.22

0.78

0.23

34.11

33.36

38.01

9078

Table 2 Ti

Al

Ca

Fe

Mg

Mn

P

K

Si

Na

S

Element as

Al2O3

Cao

Fe2O3

MgO

MnO

P2O5

K2O

SiO2

Na2O

SO3

TiO2

Utah Skyline Illinois #6

14.52

6.11

5.09

1.39

0.02

0.59

0.57

60.89

1.41

2.33

0.88

17.66

1.87

14.57

0.98

0.02

0.11

2.26

49.28

1.51

2.22

0.85

PRB (Black Thunder)

14.78

22.19

5.20

5.17

0.01

1.07

0.35

30.46

1.94

8.83

1.30

Three samples of coals have been pulverized and prepared for steady feeding to the combustor. Since that size of the coal particle is one of the indices of the coal combustion, we tried to have the same coal particle size. The coal particle size is smaller than 150 µm. The particle size distribution has been calculated and summarized in Fig 1. According to data shown in the Fig.1, the mass average diameter was calculated 62 µm for Utah Skyline coal particle, and 68.5 µm for the Illinois #6 coal particle. As it is obvious, the size and quantity of both Illinois #6 and Utah Skyline coals are similar; therefore, the study is able to focus only on the effect of coal composition.

50

50

45

45

Utah Skyline coal particle size distribution

40

35

30 25 20

30

25

20

15

15

10

10

5

5

0

0

20

40

60

80

100

120

Coal particle size (µm)

140

160

180

Illinois #6 coal particle size distribution

40

Quantity(%)

Quantity (%)

35

200

0

0

20

40

60

80

100

120

140

160

180

200

Coal particle size (µm)

Fig. 1.Particle size distribution of Illinois #6 bituminous, and Utah Skyline bituminous coals

2.2. Combustion furnace The Oxy fuel Combustor (OFC) consists of three sections, namely: 1) Burner zone or the main chamber 2) Radiant zone which is located in the bottom of the main chamber 3) The convective zone benefits from the heat exchangers to resemble the industrial furnace conditions. The dimension of the Burner zone is 0.61m I.D, 0.91m O.D. and, 1.22m as the height of the burner zone. The chamber has been insulated 76 mm thickness of the Fiberboard that is able to tolerate the temperature of 1700°K. The temperature of the furnace can be monitored by three high temperature resistant type K thermocouples that are located along the height of the chamber. Also, to have an optical access to the flame regarding optical diagnostic, four quartz windows have been provided on the quadrants of the cylindrical chamber The chamber of the OFC has been equipped with 24 ceramic electric heaters that have been arranged in three rows. These 840 watt heaters allow us to control the wall temperature of the chamber in an accurate way. In this study the wall temperature was kept at 1850 °F as one of the parameters which has an important role on ignitibility of coal. 2.3 Gas analyzers, and OPTO22 Gas probe is located at the end of convective zone of the OFC. This probe gives us the ability to monitor the exhaust gas composition during the combustion process using oxygen, carbon dioxide, NOx, SOx analyzers. The furnace is controlled and monitored by the OPTO22 control system. All the data from the furnace and analyzers have been plotted, and saved on charts during the experiment. This system helped us to have a more accurate results and control on the OFC. 2.4 Burner and feeder The schematic of the burner drawn by Solidworks is shown in the Fig. 2. The burner consists of 1) Primary stream 2) Secondary stream. The primary line is located in the center of the burner. Carrier gas which is a mixture of O2 and CO2 carry the gas through this line and spray the coal into the chamber. The secondary stream contains mixture O2 and CO2 and flows around the primary jet. Also the secondary stream is equipped with a heat exchanger that gives us the ability to preheat the secondary flow up to 700°K. The flows and temperatures of both streams are accurately automated and under control using the OPTO22 control system. Having a steady coal feeding system is one of the most effective indices on the flame stability data. In this study it was decided to use a twin-screw coal feeder with an eductor. In the eductor system, the primary stream gases get mixed with coal fed by the screw feeder. It lets us have a steady stream of pulverized coal feeding through the burner.

Fig. 2. The schematic of OFC burner

2.5 Methodology of flame stability measurement In this research, it was determined to indicate the flame stability based on the flame stand-off distance concept. Flame stand-off distance is defined as the distance between the tip of the burner and visible part of the flame. In order to measure stand-off distance, a high speed camera was used to capture the flame images. It is obvious that because of turbulent nature of the co-axial flame, fluctuations are present. The camera operating conditions chosen in this set of experiments are 8.3 ms exposure time and 30 fps. For every combustion condition, 6000 images were taken. For this exposure time, it was found out that there are stand-off distance fluctuations that need to be considered in the method. It was suggested to include all the fluctuations in the flame images by the Probability Density Function (PDF). In this paper, Flame stand-off distance plots are shown based on the PDF. 2.6 Combustion Operating Conditions In order to have a reasonable comparison to investigate the effect of coal composition on ignition and flame stability, the experiment jet aerodynamics parameters showed in Table 3 and Table 5 were kept constant for each type of coal. However to keep the total Stoichiometric Ratio (SR), it was necessary to change the coal feeding rate based on the SR. Table 3 Constant key parameters constant Parameters Total S.R.

Quantity 1.15

Pri. stream temp

305 °K

Pri. stream vel.

6.3 (m/s)

Sec. stream vel. & temp. Wall temp.

1283°K a

Overall O2 VF. a

Refer to Table 5

40%

Volume fraction

Table 4 Coal feeding rate based on SR. Coal type

Feeding rate based on SR. (Kg/hr)

Ratio of Coal weigh/ Carrier Gas volume (Kg/m3)

Utah Skyline Bituminous

4.84

1.266

Illinois # Bituminous

5.26

1.376

PRB ( Black Thunder)

8.07

2.111

Table 5 Combustion aerodynamic operating conditions for both Utah Skyline and Illinois #6 coal Experiment

Kg/hr 15.91 16.27 16.60 16.92 17.35

Average Vel. m/s 13.2 13.2 13.2 13.3 13.3

Residence time ms

Pri.a PO2

Pri. Vel.b

Sec.c Vel.

Sec.Td

Pri. O2

Pri. CO2

Sec. O2

Sec. Co2

m/s 6.3 6.4 6.3 6.4 6.3

m/s 14.9 14.9 14.9 14.9 14.9

K 489 489 489 489 489

Kg/hr 0.00 0.27 0.50 0.73 1.04

Kg/hr 6.84 6.48 6.16 5.87 5.40

Kg/hr 11.05 10.80 10.58 10.33 10.01

A A A A A

1 2 3 4 5

M.F.e 0.000 0.055 0.101 0.146 0.210

B B B B B

1 2 3 4 5

0.000 0.055 0.101 0.146 0.210

6.3 6.4 6.3 6.4 6.3

16.6 16.6 16.6 16.6 16.6

544 544 544 544 544

0.00 0.27 0.50 0.73 1.04

6.84 6.48 6.16 5.87 5.40

11.05 10.80 10.58 10.33 10.01

15.91 16.27 16.60 16.92 17.35

14.5 14.6 14.6 14.6 14.6

83 82 82 82 82

C C C C C C C C

1 2 3 4 5 6 7 8

0 0 0 0 0 0 0 0

6.3 6.3 6.3 6.3 6.3 6.3 6.3 6.3

14.9 15.3 15.6 15.9 16.3 17.8 18.6 19.5

489 489 489 489 489 489 489 489

0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

6.84 6.84 6.84 6.84 6.84 6.84 6.85 6.85

11.05 11.52 12.02 12.53 13.03 15.34 16.60 17.96

15.98 15.98 15.98 15.98 15.98 15.98 15.97 15.97

13.2 13.5 13.8 14.1 14.4 15.7 16.5 17.3

91 89 87 85 84 76 73 69

91 91 91 91 90

a

Primary stream Velocity c Secondary stream d Temperature e Mole fraction b

In this study there are three sets of experiments. Experiments A, and experiments B essentially are determined to research the effect of oxygen in the primary stream. The second benefit of this experiment is that we are able to investigate the role of turbulent diffusion mixing in the burner jet by increasing the oxygen in the primary stream. The overall oxygen concentration is kept at 40%; however, in each case the amount of oxygen gets deducted from the secondary stream, and added to the primary stream. The flow rates in both experiments A and experiments B are identical. In order to evaluate the effect of secondary stream preheat temperature on the combustibility of the coals, two temperatures were determined. For the experiments A the secondary stream temperature was kept at 489°K, and for the experiments B of the experiment it was at 544°K. The experiments C were considered to see the importance of overall oxygen concentration in the furnace. In this case the primary stream oxygen concentration is zero; however, the amount of oxygen in the secondary stream increases. The overall oxygen increases until the attached flame is obtained. The secondary stream temperature in the experiments C is kept at 489°K.

3. Results 3.1 PRB (Black Thunder) In order to measure the flame stand-off distance, it is necessary to have a visible flame. In PRB case, due to high moisture composition of the coal, visible flame was never observed. Thus, we could not measure stand-off distance of this type of coal. To have combustion, each coal particle first needs to be dried. The second step is the coal pyrolysis that causes the ignition of the coal particle, and then propagation of the flame happens. For the PRB case, the residence time was not sufficiently long enough for the coal particles with 23.69% moisture to be dried. Therefore, moisture has a significant effect on the coal ignitibility.

3.2 Combustion of Utah Skyline and Illinois #6 coals at 489°K preheat secondary stream temperature and Overall 40% Oxygen concentration The results of the experiments A are shown in the PDF plots for Illinois #6 and Utah Skyline coals. It is noticeable that at these conditions, we obtained an attached flame from Illinois #6 coal at PO2=0.144 in the primary stream. However, for Utah skyline coal, the attached flame was obtained at PO2=0.207 in the Primary stream.

0.12

Illinois #6, Preheat T=489 K, Primary PO2=0, Overall PO2=40%

Utah Skyline, Preheat T=489 K, Primary PO2=0, Overall PO2=40%

0.11

Probability Density (1/cm)

Probability Density (1/cm)

0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

5

10

15

20

25

30

35

0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

5

10

Standoff Distance (cm)

Utah Skyline, Preheat T=489 K, Primary PO2=0.054, Overall PO2=40% 0.14 0.13 0.12

Probability Density (1/cm)

Probability Density (1/cm)

0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 10

15

20

25

30

0.07 0.06 0.05 0.04 0.03 0.02 0.01 5

10

Utah Skyline, Preheat T=489 K, Primary PO2=0.099, Overall PO2=40%

Probability Density (1/cm)

Probability Density (1/cm)

0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 20

25

30

35

0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

5

10

Standoff Distance (cm)

0.11

Probability Density (1/cm)

Probability Density (1/cm)

0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 10

15

20

25

Standoff Distance (cm)

30

35

15

20

25

30

35

Illinois #6, Preheat T=489 K, Primary PO2=0.144, Overall PO2=40%

Utah Skyline, Preheat T=489 K, Primary PO2=0.144, Overall PO2=40%

5

25

Standoff Distance (cm)

0.12

0 0

20

Illinois #6, Preheat T=489 K, Primary PO2=0.099, Overall PO2=40%

0.1 0.09

15

15

Standoff Distance (cm)

0.11

10

35

0.1 0.09 0.08

0 0

35

0.12

5

30

0.11

Standoff Distance (cm)

0 0

25

Illinois #6, Preheat T=489 K, Primary PO2=0.054, Overall PO2=40% 0.15

0.11

5

20

Standoff Distance (cm)

0.12

0 0

15

30

35

0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

5

10

15

20

25

Standoff Distance (cm)

30

35

Illinois #6, Preheat T=489, Primary PO2=0.207, Overall PO2=40%

Utah Skyline, Preheat T=489 K, Primary PO2=0.207, Overall PO2=40% 0.12 0.11

Probability Density (1/cm)

Probability Density (1/cm)

0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

5

10

15

20

35

30

25

0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

25

20

15

10

5

Standoff Distance (cm)

30

35

Standoff Distance (cm)

Fig. 3. Comparison of PDF of stand of distance for Utah Skyline (Left column) and Illinois #6 (Right column) coals at 489°K preheat temperature.

3.3 Combustion of Utah Skyline and Illinois #6 coals at 544°K preheat secondary stream temperature and Overall 40% Oxygen concentration This set of experiments was selected to evaluate the effect of secondary stream temperature on the combustibility of the coals. The temperature of the secondary was kept at 544°K during the experiment. The results of this set are provided in Fig. 4. According to the results, for Utah Skyline, a semi-attached flame was obtained at PO2=0.054 and it was fully attached at PO2=0.099 in the primary stream. However, Illinois #6 coal had an attached flame at PO2=0.144. Also, it is noticeable that the results of the Illinois #6 coal at 544°K preheat temperature did not change considerably compared to the results at 489°K of the secondary stream preheat temperature. Illinois #6, Preheat T=544 K, Primary PO2=0, Overall PO2=40%

Utah Skyline, Preheat T=544 K, Primary PO2=0, Overall PO2=40% 0.12 0.11

Probability Density (1/cm)

Probability Density (1/cm)

0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

5

10

15

20

25

30

35

0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

5

10

Utah Skyline, Preheat T=544 K, Primary PO2=0.054, Overall PO2=40% 0.11

Probability Density (1/cm)

Probability Density (1/cm)

0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 5

10

15

20

25

30

35

0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

5

10

Standoff Distance (cm)

Probability Density (1/cm)

Probability Density (1/cm)

0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 10

15

20

25

Standoff Distance (cm)

30

35

15

20

25

30

35

Illinois #6, Preheat T=544 K, Primary PO2=0.099, Overall PO2=40%

Utah Skyline, Preheat T=544 K, Primary PO2=0.099, overall PO2=40% 0.12

5

25

Standoff Distance (cm)

0.11

0 0

20

Illinois #6, Preheat T=544 K, Primary PO2=0.054, Overall PO2=40%

0.12

0 0

15

Standoff Distance (cm)

Standoff Distance (cm)

30

35

0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

5

10

15

20

25

Standoff Distance (cm)

30

35

Illinois #6, Preheat T=544 K, Primary PO2=0.144, Overall PO2=40%

Utah Skyline, Preheat T=544 K, Primary PO2=0.144, Overall PO2=40% 0.12 0.11

Probability Density (1/cm)

Probability Density (1/cm)

0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

5

10

15

20

25

30

35

0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

Standoff Distance (cm)

Utah Skyline, Preheat T=544 K, Primary PO2=0.207, Overall PO2=40%

Illinois #6 , Preheat T=544 K, Primary PO2=0.207, Overall PO2=40%

0.12 0.11

Probability Density (1/cm)

Probability Density (1/cm)

0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

5

10

15

20

25

35

30

25

20

15

10

5

Standoff Distance (cm)

30

35

0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

5

10

Standoff Distance (cm)

15

20

25

30

35

Standoff Distance (cm)

Fig. 4. Comparison of PDF of stand of distance for Utah Skyline (Left column) and Illinois #6 (Right column) coals at 544°K preheat temperature

3.4 Combustion of Utah Skyline and Illinois #6 coals at 489°K secondary stream temperature with increasing of overall oxygen concentration in the secondary stream In this set of experiment, the secondary stream preheat temperature was kept at 489°K. However, to see the effect of overall oxygen on the combustion, the amount of oxygen in the secondary stream was increased. It is important to know that in this test, it was tried to keep the primary streams aerodynamics constant, in order to be able to study only the influence of the overall oxygen concentration. Therefore, the amount of oxygen in the primary steam was zero. The results of this test are shown in Fig.5 in PDF form.

Illinois #6, Increasing Secondary O2, Primary PO2=0, Secondary PO2=40%

0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

Probability Density (1/cm)

Probability Density (1/cm)

Utah Skyline, Increasing Secondary O2, Primary PO2=0, Overall PO2=40% 0.15 0.14 0.13 0.12 0.11 0.1

5

10

15

20

25

30

35

0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

5

10

Probability Density (1/cm)

Probability Density (1/cm)

0.13 0.12 0.11

5

10

15

20

25

Standoff Distance (cm)

20

25

30

35

Illinois #6, Increasing Secondary O2, Primary PO2=0, Overall PO2=41%

Utah Skyline, Increasing Secondary O2, Primary PO2=0, Overall PO2=41% 0.15 0.14

0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

15

Standoff Distance (cm)

Standoff Distance (cm)

30

35

0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

5

10

15

20

25

Standoff Distance (cm)

30

35

Illinois #6, Increasing Secondary O2, Primary PO2=0, Secondary PO2=42%

Utah Skyline, Increasing Secondary O2, Primary PO2=0, Overall PO2=42%

0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

Probability Density (1/cm)

Probability Density (1/cm)

0.15 0.14

5

10

15

20

25

30

35

0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

5

10

Probability Density (1/cm)

Probability Density (1/cm)

0.13 0.12 0.11

5

10

15

20

25

30

35

0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

5

10

Standoff Distance (cm)

10

15

20

25

30

35

Probability Density (1/cm)

0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

5

10

Standoff Distance (cm)

Probability Density (1/cm) 15

20

25

30

35

0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

5

10

20

25

30

35

15

20

25

30

35

15

20

25

30

35

Standoff Distance (cm)

Standoff Distance (cm)

Illinois #6, Increasing Secondary O2, Primary PO2=0, Secondary PO2=50%

Probability Density (1/cm)

Probability Density (1/cm)

0.13 0.12 0.11

10

15

Illinois #6, Increasing Secondary O2, Primary PO2=0, Secondary PO2=48%

Utah Skyline, Increasing Secondary O2, Primary PO2=0, Overall PO2=48%

5

35

Standoff Distance (cm)

0.15 0.14

0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

30

Illinois #6, Increasing Secondary O2, Primary PO2=0, Secondary PO2=44%

Probability Density (1/cm) 5

25

Standoff Distance (cm)

Utah Skyline, Increasing Secondary O2, Primary PO2=0, Overall PO2=44% 0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

20

Illinois #6, Increasing Secondary O2, Primary PO2=0, Secondary PO2=43%

Utah Skyline, Increasing Secondary O2, Primary PO2=0, Overall PO2=43% 0.15 0.14

0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

15

Standoff Distance (cm)

Standoff Distance (cm)

0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

5

10

15

20

25

Standoff Distance (cm)

30

35

Probability Density (1/cm)

Illinois #6, Increasing Secondary O2, Primary PO2=0, Secondary PO2=52% 0.15 0.14 0.13 0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0

5

10

15

20

25

30

35

Standoff Distance (cm)

Fig. 5. Comparison of PDF of stand of distance for Utah Skyline (Left column) and Illinois #6 (Right column) coals at 489°K preheat secondary temperature with increasing of overall oxygen concentration

3.5 TGA Analysis In order to have information regarding devolatilization, and structure of Illinois #6 and Utah Skyline coal, we decide to perform TGA analysis. The TGA test carried in both Oxygen and Nitrogen environments. The temperature ramp of the TGA was 20 °K/min. Also the amount of the purging gas was chosen 100 ml/min. In the nitrogen environment, there is ability to see the weight loss of the coal due to drying and devolatilization in an inert gas. However, in Air we have the chance to explorer the influence the oxygen on the pyrolysis as well. The TGA plots are provided in the Fig 6 and Fig 7.

0.4

100 ––––––– Illinois #6 coal ( pure N2 ) – – – – Utah Skyline coal ( pure N2) 447.6°C

90 0.3

Weight (%)

80 0.2 70

Deriv. Weight (%/°C)

464.3°C

0.1

73.8°C

60

50

0

200

400

600

Temperature (°C)

0.0 1000

800

Universal V4.7A TA Instruments

Fig. 6. TGA analysis for Utah Skyline and Illinois #6 coal in Nitrogen environment

0.4

100 – – – – Utah Skyline coal (Air) ––––––– Illinois #6 coal (Air)

448.1°C

458.1°C

80

0.3

60

0.2

Weight (%)

251.3°C 295.7°C 78.3°C

40

0.1

0.0

20

0

Deriv. Weight (%/°C)

364.3°C

0

200

400

600

-0.1 1000

800

Temperature (°C)

Universal V4.7A TA Instruments

Fig. 7. TGA analysis of the Utah Skyline and Illinois #6 coal in Air environment

4. Discussion According to the results in Fig 3, by increasing the oxygen in the primary stream from the secondary stream, the turbulent mixing develops, and the flame stability increases. Better flame stability implies a better ignition. The first criterion of ignition is the amount of volatile matter; however, it is believed that more correct way to determine the combustibility is the index of Fuel Ratio (FR) which is the ratio of (fixed carbon content / volatile matter content) [2,3,4]. General trend is observed that the higher fuel ratio correspond with less carbon burnout. The PDF data agree with the fuel ratio. However, the difference is not that significant. Therefore, in order to justify the results, it was decided to use the TGA tests. The TGA plots and data are presented in Fig 6 and Fig 7. The fuel ratio of Utah Skyline coal and Illinois #6 is provided in the Table.6. Table 6 Fuel ratio of coals Coal type

Fuel ratio

Utah Skyline Bituminous

1.28

Illinois # Bituminous

1.24

TGA plot in Fig. 6 and Fig. 7 show the weight loss of both Utah Skyline and Illinois #6 coals in nitrogen environment. The first peak indicates the moisture lost, and the second peak manifest the devolatilization of both coals. As it is obvious, pyrolysis of the Illinois #6 coal occurs at lower temperature than Utah Skyline. The TGA experiment was carried in air environment as well. Results are exhibited in Fig.7. There are three peaks in the plots. It is believed the first peak is for moisture loss, the second peak corresponds with oxidization of aliphatic structure of the coals. Aliphatic compounds have weaker bonds with smaller activation energy; therefore, they get released at lower temperatures. The pre-devolatilization of the aliphatic compounds of coals boosts heating the particle. The third peak which is largest peak of the plot shows the major devolatilization of the coal. This devolatilization is due to the oxidation of the aromatic compounds of coals. Looking more accurately to large devolatilization peak, we can see a left shoulder peak. This peak is also an indication of existence of more volatile compounds in Illinois #6 coal

than Utah Skyline coal. The result of TGA is a reliable evidence for the two types of coals combustion behaviors in Experiment A and B. Illinois #6 pyrolysis occurs at lower temperature than Utah Skyline; therefore, it is seen that in Experiment A carried out at 489°K secondary stream preheat temperature, flame stability of Illinois #6 develops by increasing the primary stream oxygen concentration. However this effect is not significant for Utah Skyline coal. Experiment B was performed at 544°K secondary stream preheat temperature. It is notable that only 55°K increasing temperature of the secondary stream stimulates the devolatilization of Utah Skyline. Comparing Fig.3 and Fig.4, It is shown Utah coal flame stability increases tremendously. However, Illinois #6 is already able to be pyrolyzed at lower temperatures, and increasing the temperature will not have any more considerable influence on the ignitibility of the Illinois #6 coal. As it is shown in Fig.3 and Fig.4, PDF’s of the experiments B-4 and A-4 look the same. In experiment C, the amount of overall oxygen was increased to able to obtain an attached flame. The amount of oxygen in the secondary stream was increased in each case of experiment C; however, the concentration of oxygen in the primary stream was kept at zero. Fig.5 shows the influence of increasing of overall oxygen is more important for Utah Skyline than Illinois #6. The Utah coal flame has some indication of an attached flame at 42% overall O 2 concentration, and it is fully attached at 44% overall concentration of O2. However; there was not any sign of an attached flame for the Illinois #6 coal until the concentration of O2 was increased to 48% and even at 52% of overall O2, a fully attached flame was not witnessed. According to the approximate analysis of both types of coal, the moisture content of Utah Skyline it 3.18% and for Illinois coal is 9.65% which is three times larger. It is important to note that when the overall oxygen concentration increases, the velocity of the burner jet in the secondary stream rises as well. Therefore, the residence time decreases. The values of both velocity and residence for each case of experiment C has been calculated and tabulated in Table5. The lack of sufficient residence time for the high moisture content coal retards the rate of both heat transfer and mass transfer for drying and devolatilization of the coal particles. So the flame stability of Illinois #6 coal lowers compared to Utah skyline coal, even though O2 concentration was increased.

5. Conclusion Studies carried out in the pilot scale Oxy fuel combustion and TGA on three types of coal with different rank reveals the important role of moisture content and volatile matter on the ignitibility of the coal particles. The existence of the moisture in the coal particle retards the heat transfer and mass transfer for the devolatilization; therefore, the ignitibility of the particles decreases. Higher residence time for high moisture content coals requires to be considered in the burner design. Fuel ratio is a good index to estimate the combustibility of the coal; However, It is not an accurate way to anticipate the combustibility of the coals. It is important to note that coal structure can have a significant effect on the ignition of the coal particle in the oxy-coal combustion. The details of the chemical structure of three coals from the Argonne Premium Coal Sample Bank (Wyodak, Blind Canyon, and Illinois #6) which are quite similar to the three coals used in this study (Black Thunder, Utah Skyline, and Illinois #6) suggest that the differences in the chemical structures of these coals could be the source of the variability in the pyrolysis/combustion properties of these coals [5]. TGA analysis is a suitable method to investigate the devolatilization of the coal for the prediction of the combustion behavior [6]. Secondary stream preheat temperature has an important influence on the flame stability of the coals. However, it is valuable to know that increasing the temperature higher than required temperature for the devolatilization of the coal particle will not have any considerable development on the ignitability of the coal.

6. Acknowledgment This material is based upon work supported by the Department of Energy under Award Number DE-NT0005015. Also; hereby, I would like to express my appreciation for the advices from Prof. Terry Ring, and Prof. Ronald Pugmire. Help in running the experiment, and preparing the data were provided by Ryan Okerlund, Colby Ashcraft, and Charles German.

References: [1] Jingwei Zhang, Kerry E Kelly, Eric G Eddings, Jost O.L Wendt. Ignition in 40kW co-axial turbulent diffusion oxy-coal jet flame, Proc. Combust. Inst. (2010), doi: 10.1016/j.proci.2010.06.106 ( In press) [2] N.Oka. T. Murayama, H. Matsouka, S. yamada, S. Shinozaki, M. Shibaoka, C.G. Thomas. The influence of Rank and maceral composition on the ignition and char burnout of pulverized coal. Fuel Processing technolog, 15 (1987) 213-224 [3] S.Su, J.H. Pohl, D. Holcombe, J.A. Hart. A Proposed maceral index to predict combustion behavior of coal. Fuel, 80 (2001) 699-706 [4] S. Pregermain. Rank and maceral effect on coal combustion characteristic. Fuel Processing technology, 20 (1988) 297-306 [5] M. S. Solum, R. J. Pugmire, D. M. Grant. 13C Solid-State NMR of Argonne Premium Coals, Energy & Fuel, 3 (1989), 187-193. [6] S.Su, J.H. Pohl, D. Holcombe, J.A. Hart. Techniques to determine ignition, flame stability and burnout of blended coals in p.f. power station boilers. Progress in Energy and Combustion Science, 27 (2001) 75-98

Study on the In-furnace Desulfurization in Oxy-fuel Combustion Using Drop Tube Furnace with Limestone Hyung-Keun Lee, Wook Choi, Hang-Dae Jo, Won-Kil Choi, Korea Institute of Energy Research Greenhouse Gas Research Center, 71-2, Jang-dong, Yuseong-gu, Daejeon 305-343, Korea Phone: +82-42-860-3647, Fax: +82-42-860-3134, e-mail: [email protected] Sang-In Keel Korea Institute of Machinery & Materials 171, Jang-dong, Yuseong-gu, Daejeon 305-343, Korea

ABSTRACT Oxy-fuel combustion uses high-purity oxygen as combustion oxidant instead of air used in conventional air combustion to produce pure CO2 stream as combustion products for easy separation and storage of CO2. Oxy-fuel combustion with many advantages like high combustion efficiency, low flue gas flow rate and low NOx emission has emerged as a promising CCS technology for coal combustion facilities. In this study, the effects of limestone types and characteristics, reaction temperature, Ca/S molar ratio, the concentrations of CO2, O2, SO2 on SO2 removal efficiency and decomposition of CaSO4 were investigated in a drop tube furnace under typical oxy-fuel combustion conditions represented by high concentrations of CO2 and SO2 formed by gas recirculation to control furnace combustion temperature. SO2 removal efficiency increased with reaction temperature, but over around 1250 ℃ decreased with reaction temperature due to promoted decomposition of CaSO4 formed by sulfation reaction. And SO2 removal efficiency increased with SO2 concentrations, because the increased SO2 concentrations suppressed the decomposition of CaSO4. The increased SO2 removal efficiency by increased CO2 and O2 concentrations showed that SO2 removal by limestone is mainly done by the direct sulfation reaction under oxy-fuel combustion conditions. Also, it was proved experimentally that the increased concentrations of CO2 and O2 have inhibited the decomposition of CaSO4.

1. INTRODUCTION Oxy-fuel combustion uses high-purity oxygen as combustion oxidant instead of air used in conventional air combustion to produce pure CO2 stream as combustion products for easy separation and storage of CO2. Oxy-fuel combustion with many advantages like high combustion efficiency, low flue gas flow rate and low NOx emission has emerged as a promising CCS technology for coal combustion facilities.(1,2)

Most coal has sulfur compounds that exist in various forms, so the combustion process for coal inevitably produces and emits sulfur oxides (SOx). Moreover, in oxy-fuel combustion, SO2 concentration increases about three times as high as that of conventional air combustion system owing to the flue gas recirculation for controlling the combustion temperature. Such a high concentration of SO2 in oxy-fuel combustion system is very harmful for the combustion system itself, CO2 transportation system and storage facilities due to corrosive potential, therefore it should be treated.(3,4) In-furnace desulfurization can be applied to high temperature region like upper part of oxy-combustion furnace as dry desulfurization process to reduce SOx emission using solid particle alkali reagent like limestone. In general, it is known that the desulfurization efficiencies of in-furnace desulfurization for conventional air combustion system are lowered by the decomposition of CaSO4 which is product of sulfation reaction between alkali reagent and SOx in combustion flue gases. On the other hand, much higher concentrations of CO2, SO2 in oxy-fuel combustion conditions would affect the decomposition behaviors of CaSO4, so higher SOx removal efficiency can be expected. Hao Liu et al. explained that enriched SO2 inside the furnace owing to flue gas recirculation under oxy-fuel combustion conditions suppressed the decomposition of CaSO4. And also CO2-rich atmosphere accelerates CaSO4 decomposition in the absence of oxygen.(5,6) Higher CO2 concentration formed in oxy-fuel combustion conditions enhance the sulfation reaction of SO2 with less calcination, so-called direct sulfation. (7) In this study, in-furnace desulfurization reactions for oxy-fuel combustion were simulated using DTF (Drop Tube Furnace). Using DTF system, the experiments were performed to investigate the effects of limestone type and characteristics, the experimental conditions including reaction temperature, the concentrations of CO2, O2, SO2 on the decomposition of CaSO4 and SO2 removal efficiencies. 2. EXPERIMENTAL 2.1 Drop tube furnace system and experimental conditions The experiments were performed using DTF system as shown in Fig.1 which consists of several parts: a reactor, a limestone feeder, gas supplying system, a reaction products collecting system and gas analyzing system. The reactor was drop tube furnace type and had physical dimensions of 500 mm length, 500 mm width, and 1200 mm height. In the center of DTF, a ceramic tube was located, which was heated electrically and had physical dimensions of 1200 mm height, 50 mm diameter and 5mm thickness. The temperatures of the reactor were controlled by three divided sections in the range of 600 - 1400 ℃ and were measured by thermocouples. Reaction gases used in these experiments were O2, CO2, SO2 to simulate the typical flue gas compositions of oxy-fuel combustion. The exact compositions of reaction gases were controlled by MFCs (Mass Flow Controller) for each gas. The simulated reaction gases flowed into the reactor at the top of the ceramic tube after passing the gas preheater to reduce the temperature difference between gases and the reactor. Reacted gases pass through the sintered metal filter equipped in the collecting system and flow to the gas analyzing system consisting of gas sample conditioner and infrared type gas analyzer. Gas concentrations of reacted gases were measured using gas analyzer (Siemens, Model: Ultramat 23) after removing the moisture and the dust using gas sample conditioner (Model: WE-GSC4P). Solid particles were captured by collecting system at the

exit of the reactor. Table.1 shows the experimental conditions for in-furnace desulfurization tests by limestone and the decomposition characteristics of CaSO4.

Fig.1 Schematic diagram of drop tube furnace type experimental system.

Several type of limestone were collected from various regional mines and characterized. We selected four of them depending on the grade and characteristics for these experiments. The selected two types of limestones were labeled KS and UR1. KS was group label for 2 limestones labeled KS1and KS2 sampled at different times from same mine. Table.2 shows the chemical composition and the average particle size data of the limestone samples used in these experiments. The reagent grade anhydrous CaSO4 with 10.92 ㎛ of the average particle size and 99% of purity manufactured by Alfa Aesar company was used for these experiments.

Table.1 Experimental conditions for oxy-fuel combustion. Experimental Condition

Desulfurization

CaSO 4 decomposition

Total gas flow rate (L/min)

5 ~ 20

8.0

Reaction temperature (℃) Ca/S molar ratio

1000 ~ 1350 1~4

1000 ~ 1350

20 ~ 80 0 ~ 20 1800 ~ 3600 0~40

20 ~ 80 0.5 ~ 20 0 ~ 3000

Gas Concentration CO2 (%) O2 (%) SO2 (ppm) H2O (%)

Table.2 The chemical composition and the average particle size data of the limestones Compositions Limestones

KS

Al2O3 Fe2O3 (%) (%)

CaO (%)

MgO Ig-loss Purity Crystal size (%) (%) (%) (㎛)

Mean diameter (㎛)

KS1

0.09

0.12

0.04

55.70

0.26

42.50

99.5

50~120

19.69

KS3

0.97

0.14

0.23

54.90

0.37

42.60

98.0

50~120

23.17

0.59

0.30

0.076 54.82

0.83

43.35

97.9

20~50

9.86

UR1 *

SiO2 (%)

Purity data were calculated CaCO3 contents based on CaO contents of limestones.

3. RESULTS and DISCUSSION 3.1 Desulfurization SO2 is absorbed through sulfation reaction by the calcined limestone in the in-furnace desulfurization. Since the calcination of limestone is an endothermic reaction and is enhanced at high temperatures, the reaction temperature is a very important factor in the desulfurization reaction. It is important to identify the type of limestone favorable desulfurization reaction and favorable conditions to improve the performance and the utilization of limestone experimentally. Fig.2 shows the effects of the reaction temperature and the type of limestones on SO2 removal efficiency. The experiments were carried out at the conditions of the reaction temperature of 1000-1350 ℃, the inlet gas concentrations of CO2, O2 and SO2 were 80%, 20%, 2400ppm, respectively. Ca/S ratio was fixed at 2.0. SO2 removal efficiencies increase with the reaction temperature for most of the limestone samples until the reaction temperature approach to 1200 ℃. Over 1200 ℃, SO2 removal efficiency is found to be reduced, because of the sintering and the clogging of the active surfaces and pores of the absorbents and the decomposition of CaSO4 due to temperature rise. Fig. 3 shows the effects of SO2 concentration and Ca/S ratio on SO2 removal efficiency for KS3 limestone. The experiments were carried out at the conditions of the reaction temperature of 1200 ℃. The inlet SO2 concentrations were varied in the range of 2400~3600ppm, and Ca/S ratio were varied in the range of 1~4. The results showed that SO2 removal efficiencies increased with SO2 concentration and Ca/S ratio. Fig. 4 shows the effects of CO2 concentration changes on SO2 removal efficiency using KS3 limestone. The experiments were carried out at the conditions of the reaction temperature of 1200 ℃, 2400ppm of the inlet SO2 concentrations and 20% of O2. CO2 concentrations were varied in the range of 40~80%, and Ca/S ratio were varied in the range of 1~4. The results showed that SO2 removal efficiency increased with the increase of CO2 concentrations from 40% to 80% due to the enhancement of direct sulfation over the reduction of active surface area of the calcined limestone by increased CO2 concentration. Fig. 5 shows the effects of O2 concentrations on SO2 removal efficiency for KS1 limestone. The experimental conditions were 1200 ℃ of the reaction temperature,

2400ppm of the inlet SO2 concentrations and 80% of CO2 concentration. O2 concentrations were varied in the range of 0.2~20%. In the case of very low O2 concentrations, SO2 removal efficiency showed very low value. SO2 removal efficiency increased rapidly with O2 concentrations over this point, but the increase rate of SO2 removal efficiency decreased over 5 % of O2 concentration. From these results, it was found that oxygen was needed for the direct reaction of SO2 with CaCO3 forming CaSO4, so-called direct sulfation, as the following equation. Therefore the higher concentration of oxygen entering the reactor promote the formation of CaSO4 and finally increase SO2 removal efficiency. CaCO3 + SO2 +1/2 O2 → CaSO4 + CO2

70

35

SO2 removal efficiency (%)

30

SO2 removal efficiency (%)

KS1 KS3 UR1

25

20

15

10

65

SO2=2400ppm

60

SO2=3000ppm

55

SO2=3600ppm

50 45 40 35 30 25 20 15 10

5

5

0 950

0

1000

1050

1100

1150

1200

1250

1300

1350

0

1400

1

2

Temperature℃

Fig. 2 Effect of the reaction temperature and limestone type on SO2 removal efficiency.

(Qg=8

L/min,

CO2=80%,

O2=20%,

3

4

5

Ca/S

SO2=2400ppm,

Fig. 3 Effect of Ca/S ratio and SO2 concentration on SO2 removal efficiency. (Qg=8 L/min, CO2=80%, O2=20%, 1200℃, KS3)

Ca/S=2)

20

55

CO2=80%

50

CO2=60%

18

CO2=40%

SO2 removal efficiency (%)

SO2 removal efficiency (%)

60

45 40 35 30 25 20 15

16 14 12 10 8 6 4

10 2

5

0

0 0

1

2

3

4

5

Fig. 4 Effect of Ca/S ratio and CO2 concentration on SO2 removal efficiency. 1200℃,

O2=20%,

5

10

15

20

25

O 2 concentration (%)

Ca/S

(Qg=8L/min, KS3).

0

SO2=2400ppm,

Fig. 5 Effect of O2 concentration on SO2 removal efficiency. (Qg=8L/min, 1200℃, Ca/S=2, KS1)

CO2=80%,

SO2=2400ppm,

3.2 CaSO4 decomposition In in-furnace desulfurization reaction for the air combustion, the desulfurization efficiency was lowered by the decomposition of CaSO4, the product of the sulfation reaction, in the high temperature region. On the other hand, the higher CO2, SO2 concentrations kept in oxy-fuel combustion suppress the decomposition of CaSO4. Fig. 6 shows the decomposition rate of CaSO4 according to the reaction temperature. The experiments were carried out at the conditions of the reaction temperature of 10001350 ℃. The inlet gas concentrations of CO2, O2 were 80%, 20%, respectively. The feed rate of CaSO4 was 1.2 g/min. The results show that the decomposition rate of CaSO4 increase with the reaction temperature and the decomposition reaction is activated rapidly above 1250 ℃. Fig. 7 shows the decomposition rate of CaSO4 according to SO2 concentration and the reaction temperature. In this figure, the decomposition rate of CaSO4 decreased with SO2 concentration and increased with the reaction temperature. For the high temperature conditions, the decomposition rates of CaSO4 were significantly reduced with increasing SO2 concentrations. In addition, it was found that there were no remarkable differences of the decomposition rate of CaSO4 with the reaction temperature in high SO2 concentration. [6] From these results, it was found that the high SO2 concentration conditions offered by the flue gas recirculation in oxy-fuel combustion inhibited the decomposition of CaSO4, therefore, higher SO2 removal efficiency could be maintained. 0.06

0.14

Decomposotion rate of CaSO4 (1/s)

Decomposotion rate of CaSO4 (1/s)

0.16

0.12 0.10 0.08 0.06 0.04 0.02

1200℃ 1250℃

0.05

0.04

0.03

0.02

0.01

0.00

0.00 -0.01

950

1000

1050

1100

1150

1200

1250

1300

1350

0

1400

Temperature(℃)

500

1000

1500

2000

2500

3000

3500

SO2 concentration (ppm)

Fig. 6 Effect of temperature and on the decomposition rate of CaSO4. (8L/min, CO2=80%, O2=20%, CaSO4=1.2g/min)

Fig. 7 Effect of reaction temperature and SO2 concentration on the decomposition rate of CaSO4. (8L/min, CO2=80%, O2=20% CaSO4=1.2g/min)

4. Conclusions In this study, the effects of limestone types and characteristics, reaction temperature, Ca/S molar ratio, the concentrations of CO2, O2, SO2 on SO2 removal efficiency and decomposition of CaSO4 were investigated in a drop tube furnace under typical oxy-fuel combustion conditions represented by high concentrations of CO2 and SO2 formed by gas recirculation to control furnace combustion temperature. 1) SO2 removal efficiencies increase with the reaction temperature until the reaction temperature approach to 1200 ℃. Above 1250 ℃, SO2 removal efficiency is found to be

reduced, because of the sintering and the clogging of the active surfaces and pores of the absorbents and the decomposition of CaSO4 due to temperature rise. 2) SO2 removal efficiencies increased with Ca/S ratio and the concentration of SO2. It has been proven experimentally that the increased concentration of SO2 contributed the inhibition of the decomposition of CaSO4. 3) The increase of SO2 removal efficiencies with CO2 and O2 concentrations demonstrated that the direct sulfation reaction was primary reaction for SO2 absorption under oxy-fuel combustion conditions. In addition, it has been proven experimentally that the increased concentrations of CO2 and O2 inhibited the decomposition of CaSO4. 4) The SO2 concentration has larger impact on the decomposition rate of CaSO4, especially in high temperature conditions. 5) From these results, it was found that the decomposition of CaSO4 had larger impact on the desulfurization efficiency in applying the in-furnace desulfurization for the oxy-fuel combustion and the major factors affecting the decomposition of CaSO4 were SO2 concentration and the reaction temperature rather than CO2 concentration and O2 concentration. References

1. IPCC Fourth Assessment Report(AR4), (2007). 2. B.J.P Buhre, L.K. Elliott, “Oxy-fuel combustion technology for coal-fired power generation", Progress in Energy and Combustion Science, 31, 283-307(2005). 3. Yewen Tan, Eric Croiset, “Combustion characteristics of coal in a mixture of oxygen and recycled flue gas", Fuel. 85, 507-512(2006). 4. Molburg JC, Doctor RD, “CO2 capture from PC boilers with O2-firing"(2001). 5. Hao Liu, Ken Okazaki, “Drastic SOx Removal and Influences of Various Factors in O2/CO2 Pulverized Coal Combustion System", Energy & Fuel, 15, 403-412(2001). 6. Hao Liu, et al. "Decomposition behavior and mechanism of calcium sulfate under the condition of O2/CO2 pulverized coal combusion" Chem, Eng, Comm, 199-214(2001). 7. Jun Cheng, Junhu Zhou, “Sulfur removal at high temperature during coal combustion in furnaces: a review", Progress in Energy and Combustion Science, 29, 381-405(2003) .

Manuscript Not AVAILABLE

2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission

PROGRAM TOPIC: 1. COMBUSTION: NOVEL TECHNOLOGIES (OXYFUEL, CHEMICAL LOOPING, ETC) EXACT TITLE OF PAPER: EFFECTS OF COAL COMPOSITION ON IGNITION LOSS, SOOT, AND ULTRAFINE PARTICLE DISTRIBUTION IN AIR AND OXY-FIRED COAL FLAMES William J. Morris, Ph.D Graduate Student, Department of Chemical Engineering, University of Utah 50 S. Central Campus Dr., MEB, Rm. 3290, Salt Lake City, UT 84112, USA e-mail: [email protected] Dunxi Yu, Visiting Research Associate, Department of Chemical Engineering, University of Utah 50 S. Central Campus Dr., MEB, Rm. 3290, Salt Lake City, UT 84112, USA e-mail: [email protected] Jost O.L. Wendt, Presidential Professor, Department of Chemical Engineering, University of Utah 50 S. Central Campus Dr., MEB, Rm. 3290, Salt Lake City, UT 84112, USA e-mail: [email protected]

Abstract: A 100kW maximum design down-fired laboratory combustor was used to determine effects of switching from air to oxy firing on soot, unburned carbon and ultrafine particle emissions from practical pulverized coal flames. Of interest here were potential practical effects of substitution of the N2 in air by CO2 in practical pulverized coal flames. Therefore, the focus is on effects of using oncethrough CO2, simulating cleaned flue gas recycle with all contaminants removed. Three coals, a western bituminous, PRB sub bituminous, and a high sulfur eastern bituminous, were fired at 36.6kW in a) air, b) 27% O2/ 73%CO2, c) 32% O2/68%CO2, respectively. Tests were conducted at (nominally) 3%, 2%, 1% and 0% O2 in the exhaust. For each condition, particulate samples were iso-kinetically withdrawn far from the radiant zone, and analyzed using a scanning mobility particle sizer (SMPS) for ultra-fine particles, a photo-acoustic analyzer (PA) for “black carbon”, and a total sample loss on ignition (LOI) method for unburned carbon in ash. Data suggest that at low stoichiometric ratios ultra-fine particles consist primarily of black carbon, which is produced in lesser amounts under oxy-fired conditions than under air-fired, even when adiabatic flame temperatures are matched. For the three coals, significant differences in the ultrafine particle distributions were noted indicating that particles formed in this region are affected by coal rank, moisture content, and sulfur content. However, significant changes in mineral matter vaporization were not observed unless the flames were hotter. These and other results are interpreted in the light of available mechanisms.

High Speed Video Analysis Of Oxycoal Combustion In 40kw Coaxial Turbulent Diffusion Flames Terry A. Ring, Professor, University of Utah Institute for Clean and Secure Energy Chemical Engineering Department 50 S. Central Campus Drive, MEB 3290 University of Utah Salt Lake City, UT 84112 Phone: 801-581-6915, Fax: 801-585-9291 [email protected]

Jingwei Zhang, Husam el Gendy, Jost O.L. Wendt, Kerry Kelly and Eric G. Eddings Institute for Clean and Secure Energy Chemical Engineering Department University of Utah Salt Lake City, UT 84112 Phone:801-581-6915 , Fax: 801-585-5705 [email protected], [email protected],, [email protected], [email protected], [email protected]

Abstract: A 61 cm diameter, down-flow, axial-flame combustion system utilizing various ratios of oxygen to CO2 for combustion of various types of coal has been studied. In companion work flame stability and the standoff distance between the burner and the point of flame ignition has been determined using low-speed video analysis of the flame. A large number of images were analyzed and probability density functions (PDFs) for flame detachment has been determined. In addition, high speed video analysis of the flame has also been performed. These images, performed at 3,000 f/s and with a shutter speed of 1/500,000 s, show hot coal particles less than 100 microns in diameter and flamelets of hot soot generated by eddies of volatiles reacting with oxygen that range in size from several hundred microns in size to centimeters in size and have temperatures that range from 1600 K to 2300 K. The size and shape of the flamelets are analyzed giving fractal shapes for the larger structures. Finally, frequency analysis of the video images were performed giving a Fourier transform power spectra with resonance characteristics associated with the frequency at which coal particles and flamelets pass by, and power spectrum decay that is characteristic of isotropic turbulence. Introduction: The objective of this experimental program is to gather data that will allow comparison to high-fidelity coal combustion simulations. These computational models provide high temporal and spatial information for temperature, pressure and gas and coal and soot particle velocities, as well as chemical species of various types, in a highly turbulent environment. These models are expensive to run taking multiple days to weeks on 1,000 cpu computer clusters. The challenge to experimentalists is to provide experimental data of a similar highfidelity nature for comparison with these simulations. In this work, we provide a first view of some highfidelity experimental data – high temporal and spatial resolution soot temperature data. A challenge to both experimentalists and simulators is how to condense the massive amount of high-fidelity data into a form that can be used for direct comparisons between experiment and theory. Here we suggest several methods including: population distribution functions (PDF) and Fourier transform power spectra of temporal data for direct comparison. Experimental Setup A 61 cm inside diameter, 1.219 m tall down-flow, axial-flame oxy coal combustion system utilizing various ratios of oxygen to CO2 for combustion has been used, see Figure 1. The experimental apparatus is

1

described in detail in a companion paper in this conference1. The walls in the main burner zone, 76 mm thick, are insulated with Insboard 2600 ceramic board material, and the downstream sections are insulated with a castable refractory material. In the main burner zone, twenty four 840 W electrical heaters are placed within the Insboard material to maintain the wall temperature at 1301±10 K using K-type themocouples (Omega) and individual controllers. The heaters enable the 100 kW system to approximate the near-burner conditions of larger boilers in the field. Illinois bituminous coal with the analysis shown in Table 1 has been used for combustion after it was pulverized and classified to 80% minus 200 mesh (74 μm). The coal was burned at a nominal flow rate of 5.26 kg/hr with a nominal 15% excess oxygen, and the incoming oxidant had an oxygen mole fraction of 40%, the balance being CO2 delivered from a liquid storage tank. The burner consisted of coannular flow of coal and transport gas in the primary stream through a 15.8 mm ID tube and oxidizer gas only in the secondary flow through the annular region between the OD (21.34 mm) of the primary and the ID (35.05 mm) of the secondary tube. The primary flow temperature was 305±2 K and that of the secondary was 544±2 K. For these experiments, the mole fraction of oxygen in the primary flow was varied from 0 to 0.209 keeping the total oxygen mole fraction for both the primary and secondary the same at 0.40. The coal was metered with a screw feeder into the primary gas stream. To keep the primary flow at a constant so that the feed rate of the coal could be kept constant, additional CO2 flow was added to maintain the velocity of the primary constant at 6.338 m/s. The secondary flow compensates for variations in the O2 content of the primary, but since the flow of the secondary is much larger (typically 3:1), it only varies by a small amount keeping the ratio of gas velocity between the secondary and the primary gas flows at a constant of 2.4. The specific experimental conditions used in this work are given in Table 2. These experiments are a subset of those performed in the companion paper1. The results of the companion paper suggest that the conditions under study here show detached (A3), partially attached (A4) and fully attached (A6) flame ignition behavior.

C o a l

Primary

Secondary

Figure 1 Schematic of Down-fired Coal Combustion System with view of flame in top-most windows indicated by vertical blue rectangles.

1

Rezaei, D.M., Eddings, E.G., Kelly, K.E., Zhang, J-W, Shou, Y-G and Wendt, J.O.L., “The Effect of Coal Composition on Ignition and Flame Stability in Co-axial Oxy-fuel Turbulent Diffusion Flames,” Proceedings Pitts burgh Coal Conference, Istanbul, Turkey, October 11-14, 2010.

2

Table 1 Analysis of Illinois Bituminous Coal LOD@ Ash@ C H N S O Volatile Fixed HHV 105ºC 750 ºC Matter Carbon (BTU/lb) 9.65% 7.99% 64.67% 5.59% 1.12% 3.98% 16.65% 36.78% 45.58 11598 Ash Composition (wt %) Al2O3 CaO Fe2O3 MgO MnO P2O5 K2O SiO2 Na2O SO3 TiO2 17.66% 1.87% 14.57% 0.98% 0.02% 0.11% 2.26% 49.28% 1.51% 2.22% 0.85% Table 2 Experimental Conditions for Oxy Combustion Experiments. The primary flow temperature was 305±2 K, secondary temperature is 544±5 K and wall temperature is 1301±10 K and the coal (80% -200 mesh Illinois Bituminous) had a flow rate of 5.26±0.3 kg/h and the total oxygen mole fraction was always 0.401 corresponding to a stoichiometric ratio of 1.15. Primary PO2 Primary Secondary Primary O2 Primary Secondary O2 Secondary Velocity Velocity CO2 CO2 m/s kg/s kg/s kg/s kg/s A Mole fraction. m/s 3

0.100

6.338

14.908

1.39E-04

1.71E-03

2.94E-03

4.61E-03

4

0.146

6.338

14.909

2.02E+04

1.63E-03

2.87E-03

4.70E-03

6

0.209

6.338

14.908

2.90E-04

1.50E-03

2.78E-03

4.82E-03

High Speed Video Analysis Video analysis was performed with a Photron high-speed, 1024x1024 pixels, black and white camera [Itronx Imaging Technology, Westlake Village, CA] at 3,000 f/s and a shutter speed of 1/500,000 of a second. This very fast shutter speed essentially stops the turbulent action of the flame and prevents saturation of any of the pixels in the field of view allowing full resolution of the image. A 70 - 300mm macro zoom lens [Tamron USA, Inc., 10 Austin Blvd., Commack, NY 11725] set at 300 mm (macro), an f-stop of 4 and a focal length of 1.3 m was used. The depth of field for these settings was ~3 cm. At these settings a 5 cm x 5cm field of view at the center line of the combustion chamber was in focus. At these camera settings each pixel observes the light from a 50 μm x 50 μm section of the camera’s view. The camera was positioned to observe at the center line of the burner two different axial locations below the tip of the burner – one 7.6 to 12.6 cm and the other 30.6 to 35.6 cm from the burner tip. The intensity of light captured is converted to an electrical current by the amplifier attached to the photo detector. Each pixel in this camera can detect a range in the number of electrons generated by the photo detector. These electrons pass through an on-pixel amplifier and into the pixel’s resistor producing a voltage that is measured –the pixel intensity scales linearly with the measured voltage and the number of electrons generated in the photo detector. The spectral resolution for the camera is between 400 nm and 1,000 nm with various quantum efficiencies at each wavelength. Using the spectral resolution obtained from the camera manufacturer under confidentiality agreement, the pixel intensity can be scaled albeit nonlinearly with temperature using Planck’s black body radiation equation assuming an emissivity of 1.0. Work done by Lou, et. al.2 using two-color pyrometry of similar flames has shown that the emissivity is typically 0.99 to 1.0. Spectroscopic measurements for visible wavelengths fit to Planck’s grey body radiation model gave a flame emissivity of 0.989. As a result, the assumption of black body radiation (emissivity of 1.0) is not a bad assumption. The resulting pixel calibration curve is given in Figure 2. In this figure the equilibrium flame temperature [2468 K] for Illinois coal in this atmosphere was also plotted. Typically 512 frames of video are taken covering a time frame of 0.17 s and creating 0.5 Gb of data for each experimental run.

2

Lou, C. Zhou, H-C, Yu, P-F, Jiang, Z-W., “Measurements of the flame emissivity and radiative properties of particulate medium in pulverized coal fired boiler furnaces by image processing of visible radiation,” Proceedings of the Combustion Institute, 31(2007), 2771-78.

3

300

Pixel Intensity

240 180 120 60 0 500

3

1×10

3

1.5×10

3

2×10

3

2.5×10

Temperature (K) Figure 2 Temperature Calibration based upon Pixel intensity for 1/500,000 s shutter speed and specific camera setup. Green half circle is the equilibrium flame temperature (2468 K) for Illinois coal at stoichiometric conditions. Measurements were made at an axial distance from the burner of 7.6 cm and 30.1 cm corresponding to the top and bottom of the first window indicated by the yellow circles in Figure 1. The windows were cleaned with compressed air just prior to taking high-speed video images, and the images were recorded after the disturbance due to cleaning cleared the combustion system.

a b c Figure 3 Three Frames of 3,000 f/s High-Speed Video for a 5 cm x 5 cm view at 30.1 to 35.1 cm down stream from the burner tip for experimental conditions A6. Results Viewing with a traditional video camera at 30 f/s, we find that the image of the flame, while not static, is rather constant in its color and intensity, see single color frame in Figure 1. An example of three frames from the 3,000 f/s high speed video for a 5 cm x 5 cm viewing area at 30.1 to 35.1 cm downstream from the burner tip, where the flame is well established for experimental conditions A6 is shown in Figure 3. The high speed image is far from constant in intensity but consists of discrete wispy bright areas of various shapes and roughly circular spots of light. The circular spots of light are from less than 100 μm to 300 μm in size, with the smaller ones emitting less light and are likely to be coal particles or clusters of coal particles (larger ones) heated to radiating conditions. The whispy bright areas cover a broad range of sizes from 100’s of microns to centimeters in length. Given the bright edges of these curved shapes, we can surmise that reaction proceeds at the interface between a turbulent eddy containing volatiles generated from the heated coal, with another turbulent eddy rich in oxygen. The reaction between coal volatiles and oxygen forms soot which is most likely responsible for the light emissions observed. Depending upon the local mixing conditions, the fuel to oxygen ratio will vary

4

significantly in these flamelets giving rise to various flamelet temperatures. Observing the individual images in a sequence, the bright zones tend to get hot over a period of 2 to 10 images or about 2 ms and cool over a period of 15 to 25 images or about 7 ms. Flame velocity can also be assessed using several individual frames of this high speed video. Image Analysis Analyzing all 512 of the images taken at one time using the temperature calibration in Figure 2, converting pixel intensity to temperature, we can determine the temperature distribution for any image. An example of a temperature distribution plot is shown in Figure 4 where the vertical axis is temperature. As we can see the temperature profile ranges from 1900 K to 2200 K which is lower than the equilibrium flame temperature (2468 K) for the coal at stoichiometric conditions. During heat up the coal undergoes drying then devolatilization. The mass released is initially water, followed by coal volatiles, which consist of varying sizes of hydrocarbon molecules that are released as the temperature of the coal particle increases. When a turbulent eddy containing coal volatiles collides with eddies having a range of oxygen concentrations, then the adiabatic flamelet temperature for those various collisions will vary and the amount of soot produced will vary. If the local mole fraction of oxygen due to eddy mixing is 0.2 and not that of the overall mole oxygen fraction of 0.4 then the equilibrium flame temperature decreases to ~1900 K. In addition, after the flamelet has reacted the soot produced will cool back down to the reaction-increased gas temperature as time proceeds. It is for this reason that we have such a broad distribution of flamelet temperatures in Figure 4.

T(K) 2246 2134 2049 1912 0

Figure 4 Temperature distributions for the image shown in Figure 3a. Taking the temperatures as in Figure 4 for all locations in all 512 frames of video taken, a temperature probability distribution function (PDF) can be developed for each experimental condition. Temperature PDFs will allow direct comparison between high-fidelity experiments and high-fidelity simulations. The temperature PDFs for all three experimental conditions for 7.6 to 12.6 cm (Top) and 30.1 to 35.1 cm (Bottom) from the burner tip are shown in Figure 5. These results show that with a detached flame, experimental conditions A3, there are slightly higher temperatures near the top of the flame. With this detached flame there is significant pre-mixing of coal with oxidant that, once ignited, burns very hot with significant heat release relative to the downstream region due to the combustion of coal volatiles. There is also a very high-temperature tail with the bottom PDF, but these higher temperatures are much more dispersed as evidenced by the lower probability magnitude. This high temperature tail is likely a result of the onset of char oxidation, with the dispersed higher temperatures associated with glowing char particles. For the partially attached flames (A4) shown in Figure 5B, the PDFs show essentially the same temperature distribution for the top and bottom with only the magnitude of the PDFs being different in these two locations; e.g., more of the frames are filled with bright areas further from the burner tip. For the partially attached flame there is less premixing before ignition, which slows down the

5

initial heat release and makes it more uniform. For the fully attached flame (A6) shown in Figure 5C, the PDFs show a slight decrease in temperature as it moves away from the burner tip from top to bottom. In the fully attached flames, stable ignition occurs near the burner tip which provides significant heat release in the early portion of the flame as a portion of the fuel is consumed with the oxidant available at that location. With continued entrainment of oxidant along the length of the flame, additional fuel is consumed to provide continual heat release. Some heat loss to the furnace walls and observation window provides for limited cooling of the combustion gases along the length of the flame. The cooling in Figure 5C is from 1840 K to 1815 K over the 23.5 cm axial distance between the top location and the bottom location. Another means to correlate a large amount of either experimental or simulation data is to perform frequency analysis using Fourier transform power spectra of temperature data from individual pixel locations. The 512 frames from the high speed camera operating at 3,000 f/s correspond to information over a 0.17 sec time interval analyzed. The resulting power spectrum probes frequencies from 12 Hz to 3 kHz. These singlepixel Fourier transform power spectrums are rather noisy so we have chosen to average the power spectrums from 128 individual pixels spread horizontally across the center of the image. These averaged Fourier transform power spectra are shown in Figure 6. In Figure 6A we see the results from the top view 7.6-12.6 cm from the burner tip and the bottom view 30.1-35.1 cm from the burner tip for experimental conditions A3. These results show that the power is low for the top view and significantly higher for the bottom which is consistent with the increase temperature away from the burner when the flame is detached. The bottom view shows a power spectrum that has a -5/3 power law slope (dashed black line) which is consistent with isotropic turbulence. Using the primary and secondary gas velocity we find that objects will pass through the view of the camera at 125 Hz and 293 Hz based upon the primary or secondary gas velocity, respectively. In the red (top) line in Figure 6A we see a peak in the power spectrum at 120 Hz. When the gases are heated to 1800 K (see Figure 5A) from 544 K the velocities will increase by a factor of 3.3 increasing the frequency with which object will pass through the view of the camera. In the blue (bottom view) curve in Figure 6A we see peaks in the power spectrum at both 130 Hz and at 410 Hz. In Figure 6B we see the results from the top view 7.6-12.6 cm from the burner tip and the bottom view 30.1-35.1 cm from the burner tip for experimental conditions A4. With this partially attached flame, the power spectrum for the top view decays according to the -5/3 power law above 200 Hz, while that for the bottom view does so only for a limited range of frequency, 200 Hz to 800 Hz. The high frequency (> 1 kHz) results show essentially a constant power for the bottom view which suggests that the newly reacting material with its higher temperatures is responsible for these high frequency results. In Figure 6C we see the results from the top view 7.6-12.6 cm from the burner tip and the bottom view 30.1-35.1 cm from the burner tip for experimental conditions A6. The top view shows higher power than the bottom view for these experimental conditions. Both views show high frequency power law slopes of -5/3 indicating isotropic turbulence. Based upon these very different Fourier transform power spectra, we can clearly see the differences between detached, partially attached and fully attached turbulent flames.

6

A3-PDF Data for all Images 0.0001 0.00009 0.00008 0.00007

PD F

0.00006 0.00005 0.00004 0.00003 0.00002 0.00001 0 1000

1200

1400

1600

1800

2000

2200

2400

2000

2200

2400

2000

2200

2400

Temperature (K) Bottom -2

Top -6

5A) Data from Experimental Conditions A3 – Detached Flame A4 - PDF Data all Images 0.00003

0.000025

P D F

0.00002

0.000015

0.00001

0.000005

0 1000

1200

1400

1600

1800

Temperature (K) Bottome -13

Top -11

5B) Data from Experimental Conditions A4 – Partially-Attached Flame. A6 PDF Data all Images 0.00005 0.000045 0.00004 0.000035

PD F

0.00003 0.000025 0.00002 0.000015 0.00001 0.000005 0 1000

1200

1400

1600

1800

Temperature (K) Bottom -14

Top -18

5C) Data from Experimental Conditions A6 – Fully-Attached Flame. Figure 5 Temperature PDFs for all experimental conditions given in Table 2. Figure 5A corresponds to experimental conditions A3, Figure 5B corresponds to experimental conditions A4 and Figure 5C corresponds to experimental conditions A6. Top corresponds to a view from 7.6 to 12.6 cm from burner tip and Bottom corresponds to a view from 30.1 to 35.1 cm from burner tip.

7

Power

1 ×10

7

1 ×10

6

1 ×10

5

1 ×10

4

1 ×10

3

Fourier Tran s form Power Spectrum

100 10 10

100

1 ×10

3

1 ×10

4

Fre quency (Hz )

Power

6A) Data from Experimental Conditions A3, Red line – Top, Blue line – Bottom 1×10

7

1×10

6

1×10

5

1×10

4

1×10

3

100 10

Fourier Tran sform Power Spectrum

100

1×10

3

1×10

4

Frequency (Hz)

Power

6B) Data from Experimental Conditions A4, Red line – Top, Blue line – Bottom 1×10

7

1×10

6

1×10

5

1×10

4

1×10

3

Fourier Tran sform Power Spectrum

100 10 10

100

1×10

3

1×10

4

Frequency (Hz)

6C) Data from Experimental Conditions A6, Red line – Top, Blue line – Bottom Figure 6 Fourier Transform Power Spectra for Experimental Conditions A3, A4 and A6. Red line – Top View 7.6 to 12.6 cm from burner tip, Blue line – Bottom View 30.1 to 35.1 cm from burner tip. Dashed black lines correspond to a power law slope of -5/3 corresponding to isotropic turbulence.

8

Conclusions High speed B&W video has been used to measure the soot temperature in an oxy-coal combustion system operating at 40 mole % oxygen. The video images have been converted to temperature using Plank’s black body radiation equation assuming an emissivity of 1.0. The temperature profiles show that individual coal particles and clusters of coal particles are present as well as clouds of soot produced by the reaction of oxygen with volatiles generated from the heating of coal. Experimental conditions consist of differing amounts of oxygen in the primary burner gas flow containing the coal. The range of experimental conditions provides for detached, partially attached and fully attached flames. From the high-speed video data temperature PDFs and Fourier transform power spectra were determined. With detached flames the temperature PDF is initially of low, near constant temperature and increases in breadth to higher temperatures moving away from the burner and the power spectrum increased from low power that is essentially constant with frequency to one that follows the -5/3 power law decay law corresponding to isotropic turbulence further away from the burner tip. With the attached flame the temperature PDF is high in temperature, relative to the detached flame, near the burner tip and decreases in temperature and narrows in breadth moving away from the burner tip. The power spectrum for an attached flame only shows agreement with the -5/3 power law decay law corresponding to isotropic turbulence at higher frequencies. Partially attached flames show no decrease in temperature moving away from the burner tip and isotropic turbulence near the burner tip decaying away to an essentially constant power spectrum away from the burner. Thus the various flame characteristics - detached, partially attached and fully attached – are easily identified with a combination of temperature PDFs and Fourier transform power spectra measurements. Acknowledgement This material is based upon work supported by the Department of Energy under Award Number DENT0005015. Help in running the experiments was provided by Dadmehr Rezaei and Dr. Yuegui Zhou.

9

Comparison of the mathematical model of pulverized coal burnout with results gained from experimental tests on drop tube Ing. Radim PALUSKA, listopadu 15/2172, 708 [email protected]

33

Energy Research Center of VSB-Technical University of Ostrava, 17. Ostrava-Poruba, Czech Republic. tel. (+420) 597323845, E-mail:

Ing. Marian BOJKO, Ph.D, Department of Hydromechanics and Hydraulic Equipment, Faculty of mechanical Engineering, VŠB-Technical University of Ostrava, tř. 17. listopadu 15, 708 33 Ostrava-Poruba, tel. (+420) 597 324 385, e-mail [email protected]

1. ABSTRACT

In connection with construction of new supercritical power plants which burn pulverized lignite coal it was started the research of kinetic parameters of coal reserves in Czech Republic. Experimental facility and methodology of pulverized coal termokinetic properties determination with use of mathematical modelling is described in the paper. Thermokinetic properties as a mean for better understanding the nature of combustion process can be determined by experiment using the Drop Tube Test Facility (DTTF) described further in the text. DTTF provides conditions occurring in pulverized coal fired boiler by emulated oxygen concentration, temperature and velocity of reaction gas. The DTTF presented in the paper was build recently at the Energy Research Center. Experimental data acquired from DTTF are planned to be used in mathematical modelling using the code Fluent. The paper describes in detail main differences between used Fluent models of particles combustible fraction and reaction rate. Program Fluent can define different mathematical models of volatile evolution (devolatilization model) and char combustion (surface combustion model) to simulate coal combustion. The single kinetic rate devolatilization model assumes that the rate of devolatilization is first-order and the kinetic/diffusion-limited rate model assumes that the surface reaction rate is determined either by kinetics or by a diffusion rate. Temperature field and distribution of species mass fraction is evaluated for comparing with experimental tests. Results from adjusted mathematical model should provide closer information about combustion process in real operation. 2. INTRODUCTION

Combustion of pulverized coal in large-scale boilers is a complicated process, which is bound with a lot of partial reactions depending on many factors. Main influences on combustion progress are temperature of reaction gas, concentration of oxygen in environment of a particle and then surface area and composition of burned coal. Burning proceeds in several stages. In the first stage, a particle warm-up occurs, then, volatile fraction evolution and combustion proceeds and char (fixed carbon) included in a particle burns as a final stage. Study of the last stage of particle combustion is the main theme of this contribution. For this purpose, a unique experimental stand DTTF has been built (see Fig. 1) which provides relevant data for a mathematical model. 3. EXPERIMENTAL FACILITIES AND INSTRUMENTATION

In order to achieve the most precise setting of boundary conditions of the whole fuel burn-out process, a complete system has been built. This system consists of reaction gas preparation, heating and keeping the temperature on the required level within the reaction chamber. At the end of sample trajectory, an instant cooling and sampling section is situated (see Fig. 1). The system is divided into 5 sections. Major component parts are automatically controlled by PLC unit via visualization on PC station. Parameters of the DTTF are given in Table 1.

Table 1: Parameters of ERC Drop tube test facility Temperature of reaction gas 600 - 1200 °C O2 concentration in reaction gas 0 – 21 %vol. Reaction gas velocity 1 - 4 m.s-1 Vertically hinged, electrically heated 4,800 mm long metal reaction chamber with inner diameter of 66 mm is made of Kanthal APM (Fe75Cr20Al5). This material is able to resist high temperatures in an oxidation environment for a long time. Eight batching holes (with a distance 500 mm one from another) are situated on sides of the drop tube in which a water cooled batching unit charges a fuel sample. The distance between these batching holes and the sampling point combined with velocity of reaction gas gives residence time of a pulverized coal particle in the reaction chamber. For more information see [2] and [8].

Fig. 1. ERC drop tube and schematic view of whole DTTF gas system (I. reaction gas preparation to desired oxygen concentration and flow rate II. Heating of reaction gas mixture to desired temperature III. Reaction chamber IV. Fuel sample batching into the reaction chamber V. Sampling with simultaneous cooling of a sample to temperature below 50°C)

The experimental tests were performed with pulverized black coal with main properties shown in Table 2. Table 2: Proximate analysis of the coal (as fired) Proximate analysis (by weight) Moisture, H0 0.5 % 0 Ash, A 26.6 % 0 Volatiles, V 28.5 % 0,fix Fixed carbon, C 44.4 % This coal was sieved on ASTM sieves with the aim to obtain 80 - 90μm granulometry.

Particle-size distributions were measured using a laser diffractometer Malvern Mastersizer as shown in Fig. 2.

Fig. 2. Size distribution of the coal sample used in the tests

4. TESTS

Approx. 5 g of a sample is continuously batched into the drop tube for the period of 15 minutes long single test. A small sample quantity and extended batching duration are used in order to reduce substantially the effect of boundary conditions of reaction environment by contributing intrinsic heat power of the reaction fuel. After passing through the reaction zone, the sample is quenched using liquid nitrogen to temperature below 50°C and sampled on an ashless filter. Then, the sample is analyzed and the unburnt U fraction (quantity of unburnt fixed carbon) is determined according to equation (1) in dependence of residence time in the reaction chamber. m3  m1  m4 (1) U  100 [%] m4  m2  100  m4  m2 A

These values serve to draw the burn-out curve that characterizes fuel with given granulometry (particle dimensions) for the set conditions, ergo temperature and oxygen concentration in the reaction chamber. These results can be also used to set (specify) a mathematical model of particle burnout, as described in the following chapter. 5. MATHEMATICAL MODEL OF PULVERISED COAL COMBUSTION IN THE DROP TUBE

Problems of flow in the drop tube can be characterized as multidimensional compressible flow of gaseous species (oxygen, carbon dioxide, nitrogen, water vapor and volatile) and flow of solid particles (pulverized coal) in the drop tube including heat transfer between gaseous phase and discretion phase (solid particles), devolatilization of volatile and combustion of pulverized coal char. Move of solid particles (the discrete phase of pulverized coal) is modeled as multiphase flow in a mathematical model based on the Lagrangian discrete phase. In the Lagrangian model, the gaseous phase is considered a continuum (continuous phase) and solid coal particles as a dispersed phase. The resulting defined mathematical model presents a system of partial differential equations that were solved by the ANSYS Fluent 12.0 program which is based on the finite volume method where the computational area (in this case the drop tube) is filled with elements of finite volumes (hexahedral, tetrahedral) in which numerical simulation of the mathematical model defined and described above is performed.

6. COMPUTATIONAL GRID OF THE DROP TUBE, BOUNDARY CONDITIONS, PHYSICAL PROPERTIES

The computational grid of the real drop tube has been constructed by the GAMBIT 2.4.26 code according to dimensions of the experimental equipment. A scheme of geometry is shown in Fig. 3. including the specification of types of the boundary conditions defined on the individual boundaries of the model. T=927°C (temperature of wall) stěny) Outlet from drop tube

Inlet of pulverised coal (diameter 5mm)

Inlet to drop tube O2+CO2+N2 Qm=0.004kg/s T=927°C

 66

600 4900

Fig. 3. Scheme of geometry of the drop tube for construction of the computational grid including the type of boundary conditions

On the base of the scheme, the gaseous mixture conditions at the inlet into the drop tube (composition, temperature, mass flow rate) are evident. The final model of the drop tube is a 3D model of dimensions shown in Fig. 3.Then, inlet of pulverized coal particles through the probe of ds=5mm outlet diameter can be seen. Pulverized coal mass flow rate (Qm=5.5555e-06 kg.s-1) into the reaction area of the drop tube is measured continuously. Composition of pulverized coal is shown in Tab. 1. In Fig.5, the actual computational grid in the longitudinal section view along the axis of the drop tube in a detail of pulverized coal inlet is shown. Compaction of the computational grid in the area of coal particle inlet and in the predicted area of pulverized coal solid particle flow in the longitudinal section view can be seen in Fig. 4.

Fig. 4 Computational grid in the area of pulverized coal inlet into the reaction area

Physical properties of gaseous mixture and pulverized coal are defined in the next step. Computation of density is defined by ideal gas law (2) due to consideration of compressible flow of gaseous mixture (O2, CO2, N2, H2O).



pabs

Yi (2) i M i The mixture’s specific heat capacity is defined as mass fraction average of the pure species heat capacities (3). c p   Yi c p ,i i (3) RT 

Viscosity and thermal conductivity of mixture are defined on the basis of kinetic theory for ideal gas law. Physical properties of pulverized coal are defined in Table 3. Table 3: Physical properties of pulverized coal  1400 kg.m-3 cp 1680 J.kg−1.K−1 λ 0.0454 W.m−1.K−1 7. MODEL OF CHAR BURNOUT (KINETIC/DIFFUSION SURFACE REACTION RATE MODEL)

The kinetic/diffusion-limited rate model (The Kinetic/Diffusion Surface Reaction Rate Model) was defined for numerical simulation of burnout of pulverized coal char where the reaction rate of char burnout is defined by kinetics or by diffusion rate (4). dm p dt

  Ap

RTYox Do r M ox

(4)

Do  r

where Ap=π*d2 is the surface area of a spherical particle. Kinetic rate r is defined by Arrhenius expression  E / RTp  (5) r  Ce Constants C (pre-exponential factor) and E (activation energy for the reaction) are determining parameters of char burnout process according to the above defined model. Activation energy defines minimum of energy needed to start the char burnout process and the pre-exponential factor influences the burnout process velocity. The particle heat balance during surface reaction is defined by the following equation mpc p

dTp

 hAp T  Tp   f h

dm p

H reac (6) dt dt For coal combustion, if the char burnout product is CO2, then ƒh=0.3. The final 3D mathematical model of gaseous species flow with pulverized coal flow, except of a char burnout model, contains a model of volatile devolatilization and a water vaporization model. Detailed characterization and definition of the devolatilization model is introduced in [7]. 8. „INTRINSIC“ MODEL

„kinetic/diffusion-limited rate model“ is in detail described in0. This model assumes, that rate of char consumption (combustion) is defined by kinetics or diffusivity.

D0  C1 where C1  5 10 constant r.

T

 T  / 2

0.75

p

(7)

dp

12

je diffusion constant. Consumption of char by kinetics is defined by r  C2 e



 E1 / RT p



(9)

where C2  0.00035 is pre-exponential factor and E1  7.4  10 is Activation energy. 7

Rate of fixed carbon consumption (and thus of particle mass) is defined by combination of kinetics and diffusion on basics of next equation:

dm p dt

  Ap pOX

D0 R D0  R

(10)

„Intrinsic“ model involves influence of reaction rate kinetics and effect of volume diffusion in regards to equation (4). Diffusion constant is defined by equation (2), but reaction rate constant r is defined with equation (5): d R   p  p Ag ki (11) 6 where factor  considers effect of diffusion on particle pores and thus porosity, ki is intrinsic reaction rate defined by Arrhenius:  E / RT  ki  Ai e i p (12) Input pre-defined parameters of intrinsic model in ANSYS Fluent 12.1 code are in table 2 with adapted porosity value. Table 2: Input pre-defined parameters of „Intrinsic“ model Diffusivity constant C1

5*10-12

Pre-exponential factor A1

0.030198

Activation energy E1

1.794*108

Porosity 

0.2

Mean diameter of pore rp

6*10-8

surface area of a particle Ap

300000

Pores crimp 

1.414214

Mathematical modeling goes from set parameters of reaction gas in reaction chamber of DTTF. Next chapters describe 2 main tested conditions. During the evaluation of numeric simulation of combustion gave discrepancy in defined volatiles content (28,5%) with respect to measurement results (see fig. 4). Accordingly, it was changed volatiles content to 10% and reached conformity with experimental results (fig. 5).

Fig. 5 – Comparison of particle burnout for 2 different volatiles content

The same results give sample in case of grain size d=55m (fig. 6).

Fig. 6 – Comparison of particle burnout for 2 different volatiles content with granulometry of d=55m

Modeling of particle (d=55m) burnout with reaction gas concentration of oxygen 2%vol. and temperature 1000°C. For numeric simulation was used „Intrinsic“ model with modification of pore r  3 109 m mean diameter p . Volatiles content is 10% and porosity   0.2 . Particle burnout of in comparison to experimental measurement is presented in figure 7.

Fig. 7 – Comparison of particle burnout for adapted „Intrinsic“ model

9. CONCLUSION

This paper describes new experimental facility which serves for combustion tests of pulverized coal. Methodology and experiments in conditions (temperature and oxygen concentration of reaction gas) similar to combustion process give relevant data for mathematical model of pulverized coal burnout in drop tube. Mathematical model differentiates devolatilization process from burnout of fixed carbon apparently on course of curves, which can be considerably changed by modification of kinetic constants as preexponential factor and activation energy. This contribution gives first results gained from mathematical modeling of combustion of pulverized black coal. There are described two models of fixed carbon burnout and comparison of model with real experiments. Results are different with volatile fraction content which lead to modification of mathematical model definition.

[1] Carpenter A.M., Skorupska N.M.: Coal Combustion - Analysis and Testing, IEACR/64, IEA Coal Research, London, 1993. [2] Žídek M., Horák J., Paluska R.: Experiences with building of drop tube. Int. Conf. „Power engineering and environment 2008“, Ostrava, 11th to 12th September 2008, ISBN 978-80-2481832-0. [3] FLUENT: Fluent 12.0 User’s Guide, Fluent Inc. 2007. [4] Baum M.M., Street P.J.: Predicting the Combustion Behavior of Coal Particles. Combust. Sci. Tech., 3(5): p. 231-243, 1971. [5] Badzioch S., Hawksley P.G.W.: Kinetics of Thermal Decomposition of Pulverized Coal Particles. Ind. Eng. Chem. Process Design and Development, p. 521-530, 1970. [6] Field M. A.: Rate of Combustion of Size-Graded Fractions of Char from a Low Rank Coal between 1200 K and 2000 K. Combustion and Flame, p.237-252, 1969. [7] Sahajwall V., Eghlimi A., Farell K.: Numerical Simulation of Pulverized Coal Combustion. International Conference on CFD in Mineral  Metal Processing and Power Generation CSIRO. 1997, p. 197-204. [8] Paluska R., Bojko M., Horák J.: Determination of pulverized coal thermokinetic properties with use of mathematical modeling Transactions of the VŠB - Technical University of Ostrava, Mechanical Series, issue 2009 ISSN: 1210-0471, p. 157-166. [9] PALUSKA R., BOJKO M., HORÁK J.: Evaluation of thermokinetic parameters influence on pulverized coal burnout in drop tube. Rynek Energii nr. 2, 2010. ISSN: 1425-5960, p. 113-117

OPTIMIZATION OF CANADIAN PETROLEUM COKE, COAL AND FLUXING AGENT BLENDS VIA SLAG VISCOSITY MEASUREMENTS AND MODELS Marc A. Duchesne, PhD candidate, University of Ottawa 1 Haanel Drive, Ottawa, Ontario, CANADA, [email protected], (613) 9470287 Alexander Y. Ilyushechkin, Research Scientist, CSIRO Energy Technology, Technology Court, Pullenvale, QLD, 4069 AUSTRALIA [email protected] (617) 33274187 Arturo Macchi, Associate Professor, University of Ottawa 161 Louis Pasteur St., Room A409, Ottawa, Ontario, CANADA, [email protected], (613) 562-5800 ext. 6939 E.J. Anthony, FBC and Gasification Group Leader, CanmetENERGY 1 Haanel Drive, Ottawa, Ontario, CANADA, [email protected], (613) 996-2868

Abstract The slagging behavior of petroleum coke must be known to determine suitable feedstock blends for entrained-flow slagging gasification. To increase the amount of slag formed and maintain a low viscosity, petroleum coke may be blended with coal and/or a fluxing agent such as limestone or dolomite. Viscosity measurements were performed for various blends of artificial Genesee coal ash, Suncor petroleum coke ash, limestone and dolomite in a neutral gas atmosphere. Adding petcoke to the coal provided a moderate reduction in viscosity, while limestone and dolomite additions were very effective for viscosity reduction. FactSage phase equilibrium predictions and quenched sample analysis via SEM and EPMA were used to link solids formation to changes in the viscosity-temperature relation. Slag blends without limestone or dolomite showed glassytype behaviour, while those with limestone or dolomite showed crystalline-type behaviour. Predictions from several slag viscosity models were compared to measured values. The viscosity model which provided the most accurate predictions was utilized for optimization of fluxing agent addition to various petcoke and coal blends. Keywords: Slag, Viscosity, Coal, Petcoke

1 Introduction Production of petroleum coke (petcoke), a by-product of the oil refining industry, has been and is expected to continue increasing [1]. Major factors driving this increase include the rising demand for transport fuels, the use of heavier crude oils and new environmental regulations pushing for reduced waste and highly refined fuels. Due to petcoke’s high heating value and low cost, there is much interest in its use as a primary fuel or in a coal-petcoke blend. Although various technologies can utilize petcoke as a fuel for power generation, gasification may provide lower gaseous emissions and less solid waste [2]. Another advantage of gasification for petcoke conversion is the trapping of hazardous metals such as vanadium and nickel in non-leachable slag. Furthermore, gasification may be used to provide steam and hydrogen in addition to power generation. This may be a huge benefit, particularly if the gasifier is located near the petcokeproducing oil refinery which requires steam, hydrogen and power [3]. The majority of gasifiers in operation today are of the entrained-flow type. In this type of gasifier, most of the inorganic component of the fuel (ash) is partially or fully melted, sticks to the reactor wall and flows to the bottom as slag. To determine the suitability of a fuel and appropriate operating conditions for gasification, it is important to know its slagging properties. A slag which is too viscous may accumulate on the reactor wall till the reactor is plugged, ceasing operation. As a rule of thumb, slag viscosity should not exceed 25 Pa⋅s at the slag tapping temperature [4,5]. If a fuel’s slag is too viscous, a fluxing agent such as limestone or dolomite may be added, and/or the fuel may be blended with another fuel. Blending fuels may also be necessary if there are issues with fuel availability, flame stability or ash content. In the present study, the viscosities of artificial Genesee coal ash, reduced Suncor petcoke ash, limestone and dolomite blends were measured. Observed and predicted phases in the blends were investigated to determine a correlation, if any, between solids precipitation and changes in viscosity. Finally, a carefully selected slag viscosity model was applied to different scenarios involving blends of the studied samples.

2 Experimental 2.1 Sample preparation Artificial Genesee coal ash, reduced Suncor petcoke ash, limestone and dolomite compositions are shown in Table 1. These compositions exclude components which represent less than 1 wt% of the ash. Artificial Genesee coal ash was prepared by mixing laboratory or analytical grade Al2O3, CaO, Fe2O3, K2CO3, MgO, Na2CO3, S, SiO2 and TiO2 powders. Composition was based on an ash analysis of real Genesee coal. Petcoke

ash was prepared by repeatedly heating Suncor petcoke up to 800°C in air. The ash mass was recorded after each heating cycle. This process was continued until the mass change between cycles was less than 2 wt%. The petcoke ash was then reduced by heating to 1000ºC in the presence of syngas (10% CO2, 25% H2, 50% CO and balance N2). The purpose of reduction prior to slag viscosity measurements is to avoid unwanted reactions involving vanadium in the melted slag. Limestone and dolomite were used as provided. Table 1. Ash compositions (in wt%, excluding components representing 25mm were crushed to –25mm and screened at 0.5mm before washing. 3.2

FACTSAGE

FactSage modelling provides the opportunity to calculate and manipulate phase diagrams, but has been established mainly in the field of complex chemical equilibrium and process simulations. For example, with FactSage it is possible to access both Fact (slag, matte, salt, ceramic and aqueous) and alloy databases, import and export streams and mixtures and also import ChemSage data files. Another advantage of FactSage is that it can also handle carbon reactions together with the minerals present in such a sample, whilst varying the gas composition and atmosphere. GTT Technologies are the developers and databases administrators of the thermodynamic software and packages, such as FactSage (www.factsage.com). FactSage is the fusion of two well-known software packages in the computational thermochemistry – Fact-Win and ChemSage. The Fact development which already started in the late 70’s, is today the largest thermochemical package and database available. FactSage modelling supply insight into specific mineral interactions, slag formation and liquid temperatures of mineral compositions. The specific value for Sasol in this study will be that these thermochemistry models can be used to analyze equilibrium conditions for reactions occurring between inorganic and/or organic materials, as well as providing insight into the mineral formation and slag formation speciation. The database will assist in understanding, as well as predicting, what might happen with specific coal and mineral sources inside the gasification process. In a study conducted by the University of Queensland [Jak and Hayes, 2002] the focus was on the relationship between ash fusion temperature and FACT liquidus calculations. It has been demonstrated that the application of FACT computer software and database for the prediction of AFT is successful. Thermo-equilibrium simulations also supply detailed insight into the slagging behavior and associated changes in mineral composition during a gasification process even in the presence of the organic components. The focus of this study will be on slag-liquid formation and mineral changes as function of temperature for any gasification process. The critical temperature zone for various gasification processes are given in Figure 2-4.

4

Operating zone

FIGURE 2 FIXED BED GASIFICATION [Holt, 2006]

Operating zone

FIGURE 3 FLUIDIZED BED GASIFICATION [Holt, 2006]

5

Operating zone

Operating zone

FIGURE 4 ENTRAINED FLOW GASIFICATION [Higman and Van der Burgt, 2007]

4.

RESULTS

The experimental results based on dense medium separation, as well as the FACT equilibrium results will be discussed herewith. 4.1

Effect of dense medium separation on mineral composition and AFT

The coal composition of the original coal sample, together with the 5 prepared fractions by means of dense medium separation are given in Table 2. This data will also be used for modelling purposes in the FactSage thermodynamic equilibrium calculations.

6

TABLE 2 CHARACTERISTICS OF PREPARED FLOAT FRACTIONS ORIGINAL Description COAL F2.1 F1.95 F1.8 Yields 100.0 90.3 86.8 84.4 PROXIMATE ANALYSES (mass % Inherent H2O 4.0 4.17 4.22 4.25 Ash 28.4 23.1 21.9 21.2 Volatile matter 21.4 22.3 22.6 22.7 Fixed carbon 46.3 50.4 51.3 51.9 ASH COMPOSITION (mass %) SiO2 46.8 45.5 44.9 44.5 Al2O3 25.5 25.9 25.9 26 Fe2O3 2.9 2.61 2.58 2.5 P2O5 1.2 1.32 1.35 1.38 TiO2 1.5 1.49 1.47 1.47 CaO 10.8 11.5 11.8 12 MgO 3.3 3.61 3.71 3.78 K2O 0.7 0.69 0.68 0.67 Na2O 0.6 0.61 0.63 0.64 SO3 5.6 5.86 6.03 6.12 LOI* 1.0 0.97 0.97 0.96 * Loss of ignition ** F1.4 to F2.1 denotes the different DMS fractions.

F1.6 63.4

F1.4 12.6

4.41 17.4 23.8 54.4

4.6 9.5 28.6 57.3

42.2 25.8 2.22 1.61 1.47 13.3 4.07 0.62 0.71 6.99 1.04

36.8 25.8 2.3 2.4 1.9 14.4 3.5 0.6 1.0 9.6 1.6

From Table 2 it is clear that significant differences in ash content, as well as ash composition, were obtained by dense medium separation. These characteristics (proximate and ash composition) of the fractions will be used individually to quantify the slag-liquid formation and differences based on FactSage modelling. The individual fractions will be treated as individual coal sources as if gasified individually per prepared fraction. The wash curve of the coal sample, as well as the ash content of the coal, is given in Figure 5.

7

90 Cummulative yield (mass %)

23.1

Yield (mass %)

21.9

21.2

Ash content (mass %)

90

87

84

80

25

20

17.4

70 60

15

63

50 9.5

40

10

30 20

5

10

13

0

Cummulative ash content (mass %)

100

0 F1.4

F1.6

F1.8

F1.95

F2.1

Relative density FIGURE 5 CUMMULATIVE YIELD AND ASH CONTENT OF THE COAL SAMPLE

The yields above RD=1.8 were relative high (>80%), where after the yield decreased significantly towards washing at lower relative densities. The ash content decreased from 21.9% at a RD=1.95 to as low as 9.5% at a RD=1.4. Figure 6 depicts the changes in mineral composition with dense medium eparation, i.e. changes in Si, Ca and percentage basic components. 48

21

46

19

44

17

42

15

40

13

38

11

36

9

34

SiO2 CaO

7

% SiO 2 (cummulative mass %)

% CaO and % Basic components (cummulative mass %)

23

32

% Basic components 5

30 F1.4

F1.6

F1.8

F1.95

F2.1

Relative density

FIGURE 6 EFFECT OF DENSE MEDIUM SEPARATION ON Si, Al, Ca AND Fe CONTENTS

8

From Figure 6 the following observations can be highlighted: • The SiO2 concentration decreases with decreasing relative density, implying and confirming the removal of high density rock fragments or extraneous particles with destoning. • The percentage basic components (Ca, Fe, Mg and Na) increased with decreasing relative density. A similar trend is observed for CaO, which contributes more than 50% of the total basic component. Ca-particles are inherent to the coal structure and destoning only removes a limited percentage of Ca, actually increasing the Ca-content in relation to other elements.

1600

o

Ash flow ( C) (oC) Measured ash temperature flow temperature

An ash flow temperature versus % basic component (Ca, Mg, Na, K and Fe) graph is given in Figure 7. Based on these data points for each relative density fraction, a clear decrease in the flow temperature is visible with increasing % basic components up to ±20%, after which the flow temperature is starting to increase. More detail discussions and evaluation will follow with the FactSage modelling results.

1550 1500 1450 1400 1350 0

5

10

15

20

25

% Basic components (Ca, Na, Mg, K and Fe) FIGURE 7 CHANGE IN FLOW TEMPERATURE VERSUS % BASIC COMPONENTS

This correlation is also confirmed by work done by Microbeam Technologies (2003) on Sasol’s coal blend indicating that the ash composition (oxide elemental composition) has an acceptable fit to the ash flow prediction curve (Figure 8).

9

___ Actual ----- Target 2900 T250 (°F) Ash flow temperature

2800 2700 2

2600

y = 1.1914x - 87.066x + 3867 2 R = 0.9489

2500 2400 2300 2200 2100 2000 10

20

30 40 % Basic components percent basic

50

60

FIGURE 8 ASH MELTING TEMPERATURE PREDITION CURVE [MICROBEAM TECHNOLOGIES, 2003]

Taking into account the results discussed above, it can be concluded that significant differences in mineral composition can be obtained by dense medium separation. 4.2

Effect of dense medium separation on slag-liquid formation based on FactSage thermodynamic equilibrium calculations

The amount of slag-liquid formation and associated anorthite (CaAl2Si2O8) formation, at a randomly selected temperature of 1250oC as determined by thermo-equilibrium simulation, is given in Table 3 and graphically illustrated in Figure 9. A temperature of 1250oC was selected to obtain partial slag-liquid formation, and simultaneously have crystalline material, i.e. anorthite, present in the mixture. To determine the correlations between slag-liquid formation (based on FactSage simulation) of coal ash and different mineral ratios, the following ratios from literature [Slegeir, et. al. 1988, Seggiani, 1999 and Collet, 2002] were decided on and calculated for use in this study, where the ash composition is expressed as mass percentages: % Base = Na2O + K2O + CaO + MgO + Fe2O3 % Acid = SiO2 + Al2O3 + TiO2 Silica value = SiO2 / (SiO2 + Fe2O3 + CaO + MgO) Dolomite ratio = (CaO + MgO) / (Fe2O3 + CaO + MgO + K2O + Na2O) Acidity = (SiO2 + Al2O3) / (Fe2O3 + CaO + MgO + Na2O + K2O) Base-acid ratio = % Base / % Acid

(1) (2) (3) (4) (5)

The calculated values, according to equations 1 to 5, together with the slag-liquid and anorthite formation, are given in Table 3.

10

TABLE 3 EXPERIMENTAL COAL AND COAL ASH RATIOS AND FACTSAGE THERMODYNAMIC EQUILIBRIUM OUTPUTS (ANORTHITE AND SLAG-LIQUID FORMATION AT 1250oC) Original Description coal F2.1 F1.95 F1.8 F1.6 F1.4 % Base 18.4 19.0 19.4 19.6 20.9 21.8 % Acid 73.8 72.9 72.3 72.0 69.5 64.5 Silica ratio 0.7 0.7 0.7 0.7 0.7 0.6 Dolomite ratio 0.8 0.8 0.8 0.8 0.8 0.8 Acidity 3.9 3.8 3.7 3.6 3.3 2.9 Acid/Base ratio 4.0 3.8 3.7 3.7 3.3 3.0 % Anorthite* 30.5 33.1 33.5 33.9 35.3 36.9 % Slag-liquid* 58.1 55.4 54.6 53.9 50.5 43.8

* FactSage results 70

% Anorthite* % Slag-liquid*

60

Mass %

50

40

30

20

10

0 F1.4

F1.6

F1.8

F1.95

F2.1

Original coal

Relative density

FIGURE 9 ANORTHITE AND SLAG-LIQUID FORMATION BASED ON FACTSAGE THERM-EQUILIBRIUM CALCULATIONS

From Table 3 and Figure 9 it is clear that the amount of anorthite increased with decreasing relative density and that the amount of slag-liquid present at 1250oC decreased with decreasing relative density. In order to explain these phenomena, a detailed look at the mineral composition (Table 2) as well as mineral ratios (Table 3) will have to be made. From Figure 10 a direct correlation can be made between anorthite formation and the CaO content of the prepared fractions. The detail mechanism of anorthite formation is published by Van, Waanders and Hack (2008).

11

40 y = 1.6x + 14.4 R2 = 0.9

38

Mass % Anorthite

36 34 32 30 28 26 24 22 20 10

11

12

13

14

15

Mass % CaO

FIGURE 10 CaO-CONTENT VERSUS ANORTHITE FORMATION

The higher concentration of CaO seems to result in a higher amount of anorthite formation. The anorthite formed as a product between the SiO2, Al2O3 and Ca-containing species. Take note that anorthite is a crystallized product from the slag formed at lower temperatures. Dense medium separation of this coal sample also supported the findings of another study on oxygen capture tendencies [Van Dyk and Waanders, 2008] of minerals during gasification. From that study, and in support of this study and findings, the following can be concluded: • The coal fractions with the highest concentration of CaO and acidic components (Al2O3 and SiO2) resulted in the highest percentage of Ca-Al-Si minerals (CaAl2Si2O8 - anorthite plus CaAl4Si2O10(OH)2 margarite) formation. In this study the float fraction at RD=1.4 yields these minerals. • The free-SiO2 in the mineral structure of coal sources resulted in forming minerals containing Mg, Na or Ca such as KAl3Si3O10(OH)2 (muscovite), Mg5Al2Si3O10(OH)8 (clinochlore), or other high oxygen molecule-capturing mineral compounds. Thus, if the free-SiO2 is decreased, i.e. by an increase in anorthite formation as with the float fraction at RD=1.4, the concentration of Si-oxygen capture compounds were then relatively low, with a high concentration of anorthite forming, as in the prepared fraction at RD=1.4, from the present investigation. Another output from FactSage that was studied for each prepared fraction was the temperature predicted by the FactSage thermodynamic equilibrium calculation where 100% of the mineral content will be slag-liquid. This information is given in Table 4 and compared to the flow temperature as obtained by the AFT analyses. TABLE 4 FLOW TEMPERATURE AND SLAG-LIQUID TEMPERATURE OF MINERALS FLOAT 2.1 FLOAT 1.95 FLOAT 1.8 Flow temperature obtained by AFT analyses (oC) 1550 1440 1410 FACTSAGE 100% slag-liquid 1565 1565 1550 prediction (oC) Difference between AFT analyses and prediction (oC) 15 125 140

12

FLOAT 1.6

FLOAT 1.4

1400

1410

1580

1600

180

190

From Table 4, a correlation between the difference in actual and predicted temperatures and the CaO content can be obtained, which is shown in Figure 10.

Difference between flow temperature (AFT) and slag-liquid (Factsage) simulation ( o C)

250 y = -10.6x2 + 302.1x - 1959.9 R2 = 1.0

200

150

100

50

0 10.0

10.5

11.0

11.5

12.0

12.5

13.0

13.5

14.0

14.5

15.0

Mass % CaO

FIGURE 10 CORRELATION BETWEEN THE DIFFERENCE IN FACTSAGE FLOW TEMPERATURE PREDICTION AND MEASURED FLOW TEMPERATURE VERSUS CaO CONTENT

From Figure 10 it can be seen that for CaO contents of 11-15% the predicted slag-liquid temperature from FACT is comparable with the standard flow temperature according to an AFT analysis. The difference of less than 10oC is within the experimental error of an AFT analyses of ±30oC. However, the increasing difference with increasing CaO content should be further discussed. According to Jak (2002) a different behaviour between predicted liquid temperatures from FACT and the AFT were also observed in other studies for higher Ca and Fe containing coal sources. It was found that high CaO containing ash cones melt and change shape at higher temperatures relative to the rest of the coal sources.

13

FIGURE 11 CaO CONTENT VERSUS SLAG-LIQUID FORMATION [JAK, 2002]

From Figure 11 it can be seen that the so-called turning point is also reported at about 10% CaO. For the coal used in this study (Figure 12), a decrease in slag-liquid formation for increasing CaO content was also observed.

70.0

Slag-liquid (mass %)

65.0 60.0 55.0 50.0 45.0 40.0 35.0 30.0 10

12

14

16

18

20

CaO content (mass%) FIGURE 12 SLAG-LIQUID FORMATION AT 1250oC versus CaO CONTENT

In an AFT test the material was found to be heterogeneous and consisted of particles with various compositions. As the temperature increases different diffusion processes take place within the ash cones. The rates of reaction, dissolution and homogenisation strongly depend on whether the intermediate phase is solid or liquid [Jak, 2002]. The presence of any liquid phase or low melting mineral significantly increases the dissolution since diffusion in the liquid is much

14

faster. The formation of the liquid and solid phases is directly related to the compositional and temperatures stability ranges of the phases that in turn are described by the phase equilibrium science [Jak, 2002]. These phenomena can be explained, as stated by Jak (2002), that the formation of the intermediate stable phase (anorthite) in the high CaO ash slags makes dissolution and melting rates slower and moves slower towards equilibrium. The higher CaOcontaining coal sources (such as the fractions prepared at lower relative densities), contain higher concentrations of anorthite (CaAl2Si2O8) and the proportion of the liquid lower, as also observed and confirmed in this study. In high CaO content coal sources, the Al-Si-Ca particles causes delay in the formation of slag-liquid. 6.

CONCLUSIONS

The amount of anorthite increased with decreasing relative density and the amount of slagliquid present at 1250oC during gasification decreased with decreasing relative density. The higher concentration of CaO seems to result in a higher amount of anorthite formation. The anorthite formed as a product between the SiO2, Al2O3 and Ca-containing species. Thus, a minimum amount of fluxing agent, which enhances slag-liquid formation and anorthite crystallization and a high concentration of acid component SiO2, which suppresses slag-liquid formation, are then present, to form a slag-liquid material. Dense medium separation of this coal source also supported the findings on the oxygen capture tendencies of minerals during gasification where the coal float fractions at RD=1.4 yielded the highest concentration of CaO and acidic components (Al2O3 and SiO2) resulting in the highest percentage of Ca-Al-Si minerals (CaAl2Si2O8 - anorthite plus CaAl4Si2O10(OH)2 margarite) formation.The free-SiO2 in the mineral structure of coal sources resulted then in forming minerals containing Mg, Na or Ca to form new mineral compounds such as KAl3Si3O10(OH)2 (muscovite), Mg5Al2Si3O10(OH)8 (clinochlore), or other high oxygen molecule-capture mineral compounds. Thus, if the free-SiO2 is decreased or not present after gasification, i.e. by an increase in anorthite formation as with the float fraction at RD=1.4, the concentration of Sioxygen capture compounds were then relatively low, with a high concentration of anorthite forming, as in the prepared fraction at RD=1.4, from the present investigation. For a CaO content of less than 10%, the predicted slag-liquid temperature from FACT calculations is comparable with the standard flow temperature according to an AFT analyses, with a difference of less than 10oC, but for an increasing CaO content the difference increased. The formation of the liquid and solid phases is directly related to the compositional and temperature stability ranges of the phases that in turn are described by the phase equilibrium science. These phenomena can be explained by the fact that the formation of the intermediate stable phase (anorthite) in the high CaO ash slags makes dissolution and melting rates slower and moves slower towards equilibrium. The higher CaO-containing coal sources (such as the fractions prepared at lower relative densities), contain higher concentrations of anorthite (CaAl2Si2O8) and the proportion of the liquid lower, as also observed and confirmed in this study. In high CaO content coal sources, the Al-Si-Ca particles causes delay in the formation of slag-liquid. 7.

REFERENCES

Alpern, B., Nahuys, J., Martinez, L., Mineral matter in ash and non-washable coals – Its influence on chemical properties, In Symposium on Gondwana Coals, Lisbon – proceedings and papers, vol. 70 1984, p299-317 Comun. Serv. Geol. Portugal, 1984, t. 70, fasc. 2, pp. 299-317, 1988.

15

Bale, C.W., Chartrand, P., Degterov, S.A., Eriksson, G., Hack, K., Manfoud, R. B., Melancon, J., Pelton, A.D. and Peterson, S., FactSage Thermochemical Software and Databses, GTT-Technologies, Germany, Calphad, 2002, 26, p. 189-228. Collet, A.G., Matching gasifiers to coal, IEA Clean Coal Centre, 2002, p.1-64. Gray, V.R., Prediction of ash fusion temperature from ash composition for some New Zealand coals, Fuel 66 (1987), p.1230-1239. Higman, C. and Van der Burgt, M., Gasification, Gulf Professional Publishing, Amsterdam, 2007. Holt, N. Gasification technology status – December 2006, Electric Power Research Institute, Palo Alto, 2006 IEA Clean Coal Centre, Coal quality assessment – the validity of empirical tests, September 2002. Jak, E. and Hayes, P.C., Applications of the new F*A*C*T database to the prediction of melting behaviour of coal mineral matter, Coorperative Research Centre for Black Coal Utilisation, Pyrometallurgy Research Group, The University of Queensland, Australia, p.1-9, 2002. Jak, E., Prediction of coal ash fusion temperatures with the FACT thermodynamic computor package, Fuel, 81, 2002, p. 1655-1668. Keyser, M.J., Personal communication, Sasol Technology, [email protected], 2006. Keyser, M.J., van Dyk, J.C., 17th International Pittsburgh Coal Conference, 2000, Pittsburgh, USA, Full Scale Sasol/Lurgi Fixed Bed Test Gasifier Project: Experimental Design and Test Results. Keyser, M.J., van Dyk, J.C., Coetzer, R.L.C., Wagner, N.J., 8th Coal Science and Technology Conference of the Fossil Fuel Foundation of Africa, South Africa, 15-17 October 2002, Full Scale Sasol/Lurgi Fixed Bed Test Gasifier Project: Impact of particle size distribution, destoning and gasifier operating conditions on sulphur production. MICROBEAM TECHNOLOGIES, INC., www.microbeam.com, 2003. Ross, D.P., Kosminski, A. and Agnew, J.B., Reactions between sodium and silicon minerals during gasification of low-rank coal, 12th International Conference on Coal Science, 2003, Australia, p. 1-9. Seggiani, M., Empirical correlations of the ash fusion temperatures and temperature of critical viscosity for coal and biomass ashes, Fuel 78 1999, p. 1121-1125. Slegeir, W.A., Singletary, J.H. and Kohut, J.F., Application of a microcomputor to the determination of coal ash fusibility characteristics, Journal of Coal Quality, 1988, Volume 7, Number 2, p. 48-54. Van Dyk, J.C., Keyser, M.J., van Zyl, J.W., Suitability of feedstocks for the Sasol-Lurgi Fixed Bed Dry Bottom Gasification Process, GTC Conference, San Francisco, USA, October 2001. Van Dyk, J.C., Melzer, S. and Sobiecki, A., Mineral matter transformations during Sasol-Lurgi fixed bed dry bottom gasification – utilization of HT-XRD and FactSage modelling, Minerals Engineering 19 (2006), 1126-1135. Van Dyk, J.C., PhD Thesis – Manipulation of gasification coal feed in order to increase the ash fusion temperature of the coal to operate the gasifiers at higher temperatures, North West University, 2006. Van Dyk, J.C., Waanders, F.B. and van Heerden, J.H.P. Quantification of oxygen capture in mineral matter during gasification, Fuel 87, 2008, p2735-2744.

16

2010 Pittsburgh Coal Conference October 11-14, 2010

SHAPING SLAG FLOW IN AN ENTRAINED FLOW GASIFIER: NUMERICAL SIMULATION AND PHYSICAL EXPERIMENTS Randy Pummill, Graduate Research Assistant, University of Utah 50 Central Campus Dr Rm 3290 Salt Lake City, UT 84112 [email protected] Gabriel Hansen, Research Assistant, University of Utah 50 Central Campus Dr Rm 3290 Salt Lake City, UT 84112 [email protected] Dr. Kevin J. Whitty, Associate Professor, University of Utah 50 Central Campus Dr Rm 3290 Salt Lake City, UT 84112 [email protected]

INTRODUCTION The aggressive environment inside a high temperature slagging gasification reactor can make getting reliable data from the reactor difficult. Any probes used to obtain data are subject to the extreme temperatures and reducing environment of the reactor and will have a very short lifetime. One way to overcome this limitation is to use a non-invasive measuring device such as a tunable diode laser. By firing a tunable laser through the reactor, data such as temperature and major species concentrations can be obtained [1]. In order to function properly, the laser would need a clear line of sight across the reactor. As coal is converted, much of the mineral matter present in the coal remains as ash. In a gasification environment, the high temperatures cause the ash to become molten slag. The slag accumulates on the walls of the reactor and runs downward. This slag flow would interfere with the line of sight necessary for the laser measurements. It is the goal of this paper to demonstrate a possible solution that would divert the slag flow around sight ports in the reactor so that a clear line of sight is maintained during operation of the gasifier. Using a physical diverter attached to the refractory above the sight port would be inefficient, as the aggressive environment would destroy any such device in short time and the device would need to be replaced frequently. Instead, it is thought that by employing a focused stream of purge gas the slag can be diverted around the sight port. Experiments simulating flow of high viscosity fluid around sample ports and a numerical model were developed in order to

simulate and model the slag flow inside the reactor. The results of these experiments are presented here. NUMERICAL MODEL Theory There have been previous successful attempts to model slag inside a gasifier [2] [3]. In the model developed by Bockelie et al., the reactor was divided into two-dimensional vertical strips and solved simultaneously. These strips were then patched together into a pseudo-three-dimensional result. However, this approach assumed a smooth wall and so for the problem of modeling fluid diversion, it was necessary to take a different approach. To solve this problem numerically, a two-dimensional solution of the Navier-Stokes equations was formulated. The equations used in the solution were the standard continuity and momentum balance equations; also, since the thickness of the film is important in this case, the equations needed to include the oft-excluded z term. Figure (1) is minimal schematic of the model domain and also clarifies the coordinate system used. In this domain, the slag flow moves down a plane. It impacts and obstruction to simulate the fluid diverted by the purge stream and then is free to flow out the bottom of the domain.

1

The projection method was employed on a staggered grid to solve this problem. A staggered grid offsets the variables so that no two variables occupy the same grid point, as illustrated below in Figure (2). Here the x and y velocities are represented by the squares and triangles and the pressure is represented by the circles. The fluid thickness was calculated at the pressure nodes and then linearly interpolated to the x and y-direction velocity nodes. Red lines are used to mark the domain boundaries. The staggered grid separates the pressure nodes from the boundaries, which results in easier conversion of the pressure matrix while solving the problem.

Figure 1: Cartoon of the simulation domain which also shows the orientation of the coordinate system used.

Assuming constant density, the continuity equation for this system can be written as Figure 2: Illustration of a staggered grid. The squares mark grid points where the velocity in the x-direction is calculated. The triangles represent the velocity in the y-direction and the circles represent the pressure calculation point. The red grid lines indicate the boundary.

where Rd is the mass flux due to slag deposition. Assuming Newtonian fluid behavior, the momentum equation for the x-direction can be written as

The steps taken to solve this problem will be discussed in detail. The algorithm used to solve the problem is: 1. 2. 3. 4.

Additional simplifications can be made by removing the dimensional dependence of the equations. All velocities are scaled by g/w (gravitational constant divided by the length of one side of the domain), distances are scaled by w, and so on. This is a common approach to this kind of problem [4] and results in

where all the variables have been appropriately scaled and

5. 6. 7.

Set initial velocity and film thickness Impose boundary conditions Set initial matrix values for new time step Solve for intermediate velocity (v*) while neglecting pressure effects Iteratively solve the Poisson equation to determine the pressure gradient Use pressure to calculate the velocity of the next time step Perform a mass balance to calculate film thickness of next time step

The first step is only done once and the other steps are repeated for the requested number of time steps. Two matrices are initialized, one to represent the staggered grid and the other

2

to represent the fluid thickness. It was proposed set the fluid thickness in the empty grid locations, but this would require more interpolation calculations to be performed. A separate matrix is used to track the fluid thickness, as it is thought this will increase overall accuracy. For the second step, the top boundary is a Dirchlet condition set to have zero x and y-direction velocity. This represents the top wall of the reactor. All other boundaries, including the pressure boundaries enforced in Step (5), are Neumann boundary conditions, also with a value of zero. This allows the fluid to freely flow out the bottom or sides of the domain. For the tests reported here, the flow of the system was entirely gravity-driven. For step (4), the motion equation (for the x-direction) is discretized as

and

where

which is quite simple compared to some of the previous equations. The last step, Step (7), is to perform a mass balance on each pressure node to calculate the new fluid thickness. This is based on the continuity equation,

where ω is the Successive Over Relaxation (SOR) constant. Solving this equation gives the pressure field for the next time step. Tests in which the SOR constant changed from 1.00 to 2.00 in order to minimize the number of iterations required to find a solution to the pressure matrix. The optimal value of  was found to be 1.93 for this problem. The next step, Step (6), is to use the pressure field to solve for the new velocity field. This is done (for the x-direction) using

and

and, finally,

A second similar equation is used for the y-direction. For step (5), the Poisson equation is solved iteratively using a Gauss-Seidel with Successive Over-Relaxation (GSSOR) method [5]. The discretized form of this step is solved as

where

These values are then interpolated linearly to the x and yvelocity nodes. The time step used was determined by a Von Nuemmann analysis [6]. The system was found to be stable for

Results and Discussion The code was run for two scenarios: a simple layer of slag flowing under gravitational pull without any obstruction and the same scenario with a rectangular obstruction. The major parameter of the simulation is the Reynolds number. By setting the Reynolds number of the system, most of the other parameters are also set. For the cases presented here the Reynolds number was set to 20. A one meter square plane was set as the domain with 25 grid point in both the x and y directions. The time step was then set at 1/20 of a second as per Equation (14). For these tests, the fluid was considered to be isothermal and of uniform viscosity and density. The first simple case was run for both a short amount of time (500 time steps) and a long time (5000 time steps). For

3

this simple flow case, the velocity field after 500 time steps is shown in Figure (3).

Figure 3: Velocity field for the short-timed case. Notice that the velocity at the top of the domain is zero and the quickly accelerates to a constant velocity distribution in the y-direction.

At the top of the domain the velocity is zero as per the boundary condition. The velocity begins to increase in the ydirection and reaches a constant gradient in the y-direction. Even in this short-timed case, a defined line is apparent where the fluid thickness is changing. This becomes obvious in Fig (4), which shows the fluid thickness for the simple, short-timed case. Looking at the two figures together, it can be seen that the areas with less fluid move more slowly than the thicker areas. This was the expected result. The plots make the difference look large, but examining the scales will show that the change in thickness is quite small (~3 mm).

For the long-timed case, the results are similar. As would be expected, the velocity and thickness distributions are even more pronounced. Figure (5) shows the velocity distribution for this case.

Figure 5: Velocity distribution for the long-timed case. Here the velocity gradient is even more pronounced than in the previous case.

The velocity field shows that as the fluid leaves the top section of the domain the velocity slows. Figure (6) shows the thickness distribution for this case. Again, the velocity and thickness distributions mirror each other. The fluid has continued to move down the wall and is exiting the bottom of the domain. The total thickness change is greater for this case, around 12mm.

Figure 6: Fluid thickness distribution for the long-timed case. The fluid has continued flowing downward. The fluid at the top boundary is lower than that observed in the first case; barely above the initial condition despite the constant mass deposition. Figure 4: Fluid thickness distribution for the short-timed case. The fluid flow is moving in the direction of gravity, as expected.

The fluid at the top boundary in the second case is about 3mm less thick than the fluid at the top of the first case. Even with the constant mass deposition across the domain, the fluid

4

has begun to recede. The fluid in the level part of the domain is much higher than in the first case, due to the flow of the fluid and the mass deposition. In both of these cases, the velocity in the x-direction was negligible. Since this is a gravity-driven flow, it was expected that there would be no x-direction velocity. For the second case with the rectangular obstruction, the code was run for 500 time steps. The velocity field is shown in Figure (7). The boundaries of the obstruction are shown with red lines.

meaning in the real gasifier. A heated system employing real slag running down refractory face was considered, but it was rejected due to the difficulty of keeping the slag above its temperature of critical viscosity through the entire flow circuit. Instead, a system was designed to try and simulate the most important aspects of the gasifier system. It was decided to test the effects of fluid viscosity and surface tension and the effect of the purge gas density on the ability of the gas to divert the flow around the sight port. For the physical experiments, a section of refractory was cast in the same geometry present in the entrained flow gasifier located at the University of Utah’s Industrial Combustion and Gasification Research Facility. The gasifier has an inner diameter of eight inches in the reaction section with 1.5 inch sight ports. The refractory was cast so that it also had these same characteristics. The cast piece was 3 feet in height. A photograph of the setup is shown in Figure (9).

Figure 7: Velocity field of the fluid with a rectangular obstruction. The obstruction is shown as a red box. The fluid moves around the obstruction as expected.

The fluid thickness distribution is shown in Figure (8). It can be seen that some fluid is predicted to build up against the leading edge of the obstruction and the fluid thickness drops behind the trailing edge of the obstruction. This makes intuitive sense and matches observations of the real experimental system.

Figure 8: Fluid thickness distribution for the obstructed case. The fluid builds up against the obstruction and thins in its wake.

PHYSICAL MODELING Experimental Setup The first consideration of designing the physical experiments was to ensure that the data collected would have

Figure 9: Photo of the experimental apparatus. The refractory is curved with an eight inch diameter and has a 1.5 inch hole to act as a sight port. A wooden frame supports the refractory in an upright position.

5

Water and silicone oils of four different viscosities were obtained to simulate the slag flow. Molten slag has a wide viscosity range that is heavily temperature dependent. Below a certain temperature, called the temperature of critical viscosity or Tcv, the slag behaves as a crystalline solid. Above the T cv the slag behaves approximately as a Newtonian fluid [7]. In order to flow freely enough to be tapped, the slag must have a viscosity of less than 25000 cP, with an optimal viscosity of 15,000 cP [8]. In order to include this range in the experiments, the oil viscosities chosen were 1, 100, 5000 and 30,000 cSt. Silicone oil was chosen as it is easy to obtain, has greatly modifiable viscosity, and its other physical properties, such as surface tension, are well known. Water was chosen as it has a high surface tension. By comparing the results for water with the results for the 1cSt oil, inference could be made concerning the effect of surface tension on the results. The fluid was pumped up to the top of the refractory section and allowed to flow downward. A piece of metal tubing with 1/8” holes drilled in it was used to distribute the oil at the top of the refractory. The 30,000 cSt oil also required the three additional white tubes (visible at the top of the refractory in Figure 1) in order to achieve a fairly uniform distribution of oil across the refractory face. The fluid thickness was approximately 1/8 inch for the water, and the 1, 100 and 5,000 cSt oils; the 30,000 cSt oil had an approximate thickness of 3/16 inch. The pump used was a Hydra-Cell membrane pump, which is capable of pumping very high viscosity fluids. A red laser was fixed inside the sight port of the refractory and directed at a piece of white paper. Having the paper made it easy to determine if the silicone oil was interfering with the line of sight as any oil interfering with the line of sight would cause the laser point to blur. To divert the slag around the sight port, a purge stream of inert gas was used. Eleven different devices were constructed to deliver the purge gas to the port. These included a simple setup allowing the gas to travel through the sight port diffusely and using tubes to focus the gas stream at the silicone oil at different angles and with different tube diameters. A table listing the purge devices is shown in Table (1). Figure (10) shows the last purge device, numbered (11) in the table, installed in the refractory section.

Table 1: Table of devices used to deliver the purge gas to the sight port. A short description of each device is also given.

Tube Size

Description

1

None

Diffuse flow through sight port

2

1/4"

Tube at the top of the hole, shooting forward

3

3/8"

Tube at the top of the hole, shooting forward

4

3/8"

5

1/2"

6

1/4"

7

3/8"

8

3/8"

9

1/2"

10

3/8"

Flattened tube, shot horizontally from top of hole Flattened tube, shot horizontally from top of hole Tube along bottom of port, angled to shoot upward Tube along bottom of port, angled to shoot upward Flattened tube along bottom of port, angled to shoot upward Flattened tube along bottom of port, angled to shoot upward Tube in center, laser shot down center of tube

11

1/4"

Tube in center of port, angled to shoot upward

#

Figure 10: Close up photograph showing device #11 installed in the refractory face. The tube can be seen protruding in the sight port, angled up towards the top edge of the port. Purge gas pushed through this tube served to divert the flow around the hole.

6

The purge gas flow rate was controlled using an adjustable rotameter. The flow was started at zero flow and increased manually over time. Both compressed air and carbon dioxide were used as purge gases in separate tests so that any relationship between good fluid diversion and gas density could be observed. Additionally, in order to simulate the convective forces present inside the reactor, a pair of nozzles was used to flow air downward across flowing oil. This permitted the collection of additional qualitative data regarding the stability of the flow diversion with convection occurring. Results and Discussion During the course of testing it was found that four behaviors were displayed by the silicone oil as it came into contact with the purge stream. These behaviors were: 1. 2. 3. 4.

Passing over the stream with little or no diversion or disturbance to the fluid flow. Breaking into droplets upon contact with the purge stream, causing line of sight interference. Diverting around the purge stream in an unstable flow pattern. Diverting around the purge stream while maintaining a stable flow pattern.

Figure (11) shows a photograph of 30,000 cSt fluid being diverted around the sight port by purge device (7) with a purge flow rate of 3.6 pounds per hour.

ridge at the point where the purge stream impacts the fluid is similar to the behavior predicted by the computational model.

For each purge gas, gas delivery device and oil viscosity, the regimes of purge flow rates for each of these behaviors was noted. As the 30,000 cSt fluid is the most relevant to slag behavior studies, the results for each nozzle and the 30,000 cSt fluid are shown below in Table (2). The results for the other fluids are not given in this paper, but follow a very similar pattern. In the table, the red squares indicate that for the corresponding mass flow rate and purge device, the fluid did not diverge, as per the first behavior regime. An orange square indicates that the settings produced an unstable diversion with regular observed line of sight interference. Yellow squares indicate that for the corresponding purge gas flow rate and nozzle combination, the diversion was sufficient, but the flow was unstable and could possibly have intermittent blockage of the line of sight. These two colors cover the second and third behavior regimes listed above. Lastly, the green squares indicate that the fluid diversion was stable and sufficient for a clear line of sight; which behavior falls under the fourth regime. The flows were exposed to a parallel gas flow to simulate the flow of the reacting gases past the slag. The flow rate was about 0.5 feet per second, which is similar to that calculated to exist in the actual reactor. This parallel flow often resulted in stabilizing the flow patterns in the fluid flow, though it did sometimes cause instability with some of the devices. Flows that reacted well to the parallel gas flow are marked with an asterisk in the table. Over the course of the experiments, it was found that aiming the purge stream for the apex of the circular sight port gave the best results for diversion. The table shows that device (11) worked very well at low flow rates and maintained a stable diversion even when exposed to the parallel gas flow. Devices (2), (6) and (8) also performed well, but device (11) provided the widest stable range of diversion.

Figure 11: 30,000 cSt fluid being diverted around the sight port using purge device (7). The purge flow rate is 3.6 pounds per hour. Notice the

7

Table 2: Data for the 30,000 cSt fluid. Again, device design #11 gave the best results.

Purge Flow Rate [lb/hr] 0.0 0.6 1.2 1.8 2.4 3.0 3.6 4.2 4.8 5.4 6.0 6.6 7.2 7.8 8.4 9.0 9.6 10.2 10.8 11.4 12.0 12.6 13.2 13.8 14.4 15.0 15.6 16.2 16.8 17.4 18.0

Purge Device 1

2

3

* * * *

4

* * * *

5

6

7

* * * * * * *

* * *

* * * * * * * * * * *

* * * * * * * * * * * * * * * * * * * * *

* * = Most stable at these flow rates when simulating convection

8

8

9

* * * * * *

10

11

* * * * * *

Water has a viscosity of 1 cSt and a relatively high surface tension of 72.0 dyne/cm. The results of water in the experimental system were compared with those of 1 cSt silicone oil, which has a surface tension of 17.2 dyne/cm. For these tests, the fluid film with a flow rate of 5.8 liters per minute was purged with dry air using device (11). Divergence with water was achieved with a purge rate of 2.23.4 lbs/hr while the oil showed divergence with a purge rate of 1.9-2.3 lbs/hr. The higher surface tension of the water held the fluid together at higher purge rates than the oil. The oil would form a mist at lower purge rates than the water. However, the water also required greater purge rates to diverge, also due to the higher surface tension. Most slag has a surface tension near 350 dyne/cm [9], it is expected that the purge rates required to divert the slag flow will be much greater than that required for oil of a similar viscosity.

CONCLUSIONS AND FUTURE WORK From the work presented here, it appears that it is possible to divert slag flow around a sight port in an entrained flow gasifier. These results will be used to design a purge system for use in a real gasification environment. The model developed for this paper was able to give good predictions for fluid behavior. The flow predications of the model were consistent with those found with the physical experiments. However, this method also has several weaknesses. Complex geometry was difficult to implement and including a purge flow perpendicular to the 2D computational field seems impossible to accurately implement. The model also currently lacks any sort of temperature prediction, which is very important in predicting slag properties. A three dimensional approach with an energy balance will need to be developed. The physical experiments provided a better understanding of how slag in a gasifier will behave when contacted with a purge stream. Diversion of the fluid depends on the mass of the purge gas use, not the volumetric flow rate. Also, the higher viscosity fluids require higher purge flow rates to be diverted, but are also more stable when subjected to parallel gas flows. Finally, fluids with higher surface tension require greater purge flow rates to get a good diversion.

[3] Seggiani, M. (1998). Modelling and Simulation of Time Varying Slag Flow in Prenflo Entrianed-flow Gasifier. Fuel , 1611-1621. [4] Bird, R. B., Stewart, W. E., & Lightfoot, E. N. (2002). Transport Phenomena. Hoboken: John Wiley & Sons. [5] Tannehill, J. C., et al., (1997). Computational Fluid Mechanics and Heat Transfer. Philidelphia: Taylor & Francis. [6] Ferziger, J. H., & Perić, M. (2002). Computational Methods for Fluid Dynamics. New York: Springer. [7] Reid, W. T., & Cohen, P. (1944). Factors Affecting the Thickness of Coal-Ash Slag on Furnace-Wall Tubes. Transactions of the ASME , 685-690. [8] Song, W., Tang, L., Zhu, X., Wu, Y., Zhu, Z., & Koyama, S. (2009). Flow Properties and Rheology of Slag from Coal Gasification. Fuel . [9] Hanao, M., et al., (2007). Evalutation of Surface Tension of Molten Slag in Multi-Component Systems. ISIJ International , 935-939. ACKNOWLEDGEMENTS This work was supported by a U.S. Department of Energy Cooperative Agreement via Stanford University, DEFE0001180. The author also wishes to thank Dave Wagner of the University of Utah Industrial Combustion and Gasification Research Facility, for his help with this project.

REFERENCES [1] Jeffries, J., et al., (2009). Tunable Laser Diode Absorption Temperature Measurements in a Fluidized Bed Gasifier. Pittsburgh Coal Conference Proceedings. Pittsburgh: PCC. [2] Bockelie, M., et al. (2004). A Computational Workbench Environment for Virtual Power Plant Simulation, Appendix F: Entrained Flow Gasifier - Slagging Wall Model. Salt Lake City: Reaction Engineering International.

9

INFLUENCE OF GASIFICATION CONDITIONS ON THE PROPERTIES OF FLY ASH IN A BENCH‐SCALE  OPPOSED MULTI‐BURNER GASIFIER  Qinghua Guo, Guangsuo Yu, Fuchen Wang, Zhenghua Dai  (Key Laboratory of Coal Gasification, Ministry of Education, East China University of Science and  Technology, Shanghai, 200237, China)  Abstract:  Entrained‐flow  coal  gasification  offers  a  high  efficiency  and  low  pollutant  emission  way  in  coal’s  utilization.  Opposed  Multi‐Burner  (OMB)  entrained‐flow  coal  gasification  technology  is  now  widely used. During the coal gasification process, fly ash and slag are the main solid by‐products. In  this  study,  experimental  work  has  been  carried  out  to  characterize  the  fly  ash  particles  which  were  generated  in  the  bench‐scale  OMB  gasifier.  The  composition  of  fly  ash  was  determined  by  SEM  and  EDS.  The  particles  size  distribution  was  measured  by  Malvern  particle  size  analyzer.  The  influencing  factors of the fly ash particles properties with coal water slurry gasification, such as gasifier operation  temperature  and  O/C  ratio,  were  considered  during  the  gasification  experiments.  The  results  show  that  the  shape  of  the  ash  particles  has  irregular  shape,  and  C  in  the  fly  ash  is  the  major  content  element.  O/C  ratio  and  operation  temperature  affect  the  shape  of  particles  and  the  particle  size  distribution  significantly.  At  a  fixed  operation  condition,  the  concentration  of  unburned  carbon  decreases  along  the  gasifier.  At  various  O/C  ratios,  fly  ash  particle  size  distribution  had  a  bimodal  distribution. The investigation on the characteristics of fly ash will be beneficial to the operation and  optimization of the OMB entrained‐flow gasifier.  Keywords: fly ash; gasification; particle size distribution   

1. Introduction Coal gasification technology, especially high pressure and large scale entrained flow coal gasification, is the leading and key technology of coal cleaning utilization. A new type gasifier named as opposed multi-burner (OMB) gasifier has been widely used in China[1]. At the present time, the largest coal handling capacity of OMB commercial gasifier is 2000t/d, the design pressure involve both 4.0MPa and 6.5MPa. Fly ash and slag are the main by-products during the coal gasification process[2]. Ahmaruzzaman[3] reviewed the utilization of fly ash in construction, removal of organic compounds, heavy metals, dyes, and zeolite synthesis which can help a great deal in the reduction of environmental pollution. Fly ash is easy enriched with toxic heavy metal (i.e. Pb and Cd) due to its large specific surface area and subsequently damages to human health and pollutes the environment[4]. Many researchers have been focused on the properties of particulate matter and its formation mechanism during coal gasification and combustion [5-7].Wu Tao et al.[8] investigated particulate matter characteristics and formation mechanism during coal gasification in an entrained-flow gasifier. Hiromi Shirai et al.[9] studied particulate matter properties during pulverized coal combustion. Stanislav et al.[10] summarized the methods that used for characterization of the phase, mineral and chemical composition of fly ash from coal-fired power stations. However, the literature about fly ash particular matter properties at different operation conditions during coal water slurry gasification is rare. In this paper, the effect factors of operation temperature and oxygen carbon atom ratio on properties of fly ash particulate matter have been analyzed in a laboratory scale OMB gasifier by using Scanning Electron Microscopy (SEM), Energy Dispersive x-ray Spectrum (EDS) and Malvern Particle Size Analyzer.

2. Experimental work 2.1 Experimental setup A schematic diagram of the laboratory-scale opposed multi-burner entrained-flow gasifier is shown in Fig. 1. The gasifier that made by stainless-steel includes gasification chamber and quench chamber two parts. The inner diameter and length of the gasification chamber composed of a 15mm thick cast refractory wall are 300 and 2200mm, respectively. The refractory wall, wrapped with 235mm fiber blanket to reduce the heat transfer. The stainless-steel column shell of gasification chamber is 800mm in diameter and 2500mm in height, and the quench chamber is 500mm in diameter and 2000mm in height. Oxygen and coal water slurry (CWS) were fed in the gasifier through four symmetric opposed dual-channel burners. Oxygen is fed into the outer channel of burners from oxygen cylinder and measured by gas mass flowmeters. Diesel oil and CWS are fed into the inner channel of burners by the gear pumps and single-screw pumps respectively. High relative velocities between oxygen and fuel in the central region of gaisifer provide good conditions for active diffusion and convection at the droplet surface, which together with the high temperature results in fast combustion and gasification reactions. Then, the raw syngas flows down to the quench chamber, where it is cooled and partially scrubbed by water, and finally flows to the downstream processes. Nitrogen is fed from nitrogen cylinder to clean the inner channel of burners at the end of the experiment.

Fig.1

Schematic diagram of the experimental setup

The analysis of the coal sample which is used in the experiments is show in Table 1. The coal ash fusion temperature is 1255℃. Coal was milled into various particle sizes, with water and additive added, the concentration of CWS is about 61% during the CWS pulping process. Before feeding the coal water slurry, the gasifier was preheated to the reaction temperature (>1300℃) with diesel oil fuel. The main experimental conditions in this study were to control the oxygen carbon atom ratio at 0.9, 1.0, 1.1. Each burner’s CWS flow rate varified from 9kg/h to 12 kg/h, oxygen flow rate from 4Nm3/h to 5.6Nm3/h.

Table 1 Proximate analysis M A V FC

4.04 8.38 32.56 55.02

Analysis of experimental coal sample (mass%) Ultimate analysis C H O N S

71.66 3.87 10.49 0.94 0.80

Ash analysis SiO2 CaO SO3 Al2O3 Fe2O3 MgO

34.56 20.16 16.77 14.40 8.69 1.87

K2O Na2O TiO2 SrO MnO BaO

1.15 0.70 0.68 0.45 0.23 0.11

2.2 Fly ash collection and analysis Fly ash sampling system included a high-temperature water jacket sampling tube, a filter, a flow meter and a pump. Sampling places were located at the thermocouple ports along the gasifier. In this work, the main sampling position was 3# which was closed to the slag outlet of the gasification chamber. Particle size distribution was analyzed by using Malvern Particle size analyzer. Fly ash samples were pretreated in an ultrasonic cleaning machine in order to make particles dispersed uniformly in the bottle, after formed a uniform suspension solution, then took part of it to analysis using Malvern Particle Size analyzer. Morphology and element distribution were analyzed by SEM and EDS respectively. Suspension solution was firstly filtered and then dried for two hours under condition of 105°C before SEM and EDS analysis.

3. Results and discussion 3.1 Effect of operation temperatures 3.1.1

SEM Images

SEM images of fly ash at 1200℃(below ash fusion temperature) and 1350℃(above ash fusion temperature) are presented in Fig. 2 and highlight important structural and morphological features. Ash generated at high temperature was found to be much finer than ash generated at low temperature due to the differences in porous structure. It can be observed that the fly ash has porous structure both in 1200℃ and 1350℃. Temperatures affect the morphologies of particles significantly, the number of fine particles in 1200℃ less than that in 1350℃. Furthermore, more volatile element vaporized at higher temperature, and then condensed to form more spherical particles. On the contrary, there were less spherical particles at the operation temperature of 1200℃.

(a) 1200℃ (b) 1350℃ Fig. 2 SEM results at different temperatures

Surface element distribution at different temperatures was illustrated in the photos. It can be seen that the content of S, Fe and Na at the temperature of 1350℃ were significantly higher than that at 1200℃. Volatile element such as Na and S were more significantly vaporized, and there was a reducing atmosphere in the gasifier, iron was reduced to metal iron which is more volatility, which will result in a higher content of these elements on the surface of particles.

3.1.2 Particle size distribution The fly ash particle size distribution at different temperatures was shown in Fig. 3. Comparison with the experimental coal particles, the particles size became smaller when they reacted in the gasifier. The peak of the particle size distribution shift to left position significantly and the fraction of fine particles increased with the temperature increasing. Temperatures not only affect the ratio of particle size but also the peak of particle size distribution. The research has shown that temperature affect fine particle formation more significantly under a condition of higher oxygen content. Temperature played an important role in the fine particles formation process. More fine particles were formed at higher temperature than that formed at lower temperature. The results agreed with the previous work well. 5

Vol. %

4

1350℃ 1200℃ Experimental CWS

3 2 1 0

0.1

1

10

100

1000

Paticle size /mm

Fig. 3

Particle size distribution at different temperatures

3.2 Effect of oxygen carbon ratio 3.2.1 SEM Images Morphology of fly ash at different oxygen carbon ratio was shown in Fig. 4. When the oxygen carbon ratio is 0.9, coarse particles were mainly irregular and with less spherical shape. When the oxygen carbon ratio increased to 1.0, the number of spherical particles increased relatively and most fine spherical particle exist as single. For the case of oxygen carbon ratio is 1.1, fine particles mainly exist as agglomerate status. Conversation of coal particles increased as oxygen content increased, at the same time, diameter of particles decreased and content of vaporized mineral increased, subsequently formed more spherical particles. So we can conclude that with the oxygen carbon ratio increased, the diameter of coal particles decreased, resulting higher possibility for included mineral coalescence. And finally leading to more fine particles aggregated.

3000

3000 C 2500

2500

2000

2000

3000

C 2500

1500 1000 500

O

Al

Si

Ca 0 0.0

1.5

1500 1000 500

S 3.0 4.5 Energy (keV)

Fe 6.0

7.5

9.0

Intensity (a.u.)

Intensity (a.u.)

Intensity (a.u.)

C

0 0.0

O

Si Al S 1.5

2000 1500 1000 500

Ca 3.0

Fe

4.5 6.0 Energy (keV)

7.5

9.0

0 0.0

O

Al

Si

S Ca

1.5

3.0

Fe

4.5 6.0 Energy (keV)

7.5

9.0

(a) O/C=0.9 (b) O/C=1.0 (c) O/C=1.1 Fig. 4 SEM and EDS results at different oxygen carbon ratios The EDS results of fly ash surface element distribution at different oxygen carbon ratio were also shown in Fig.4. It can be seen that, carbon is the dominant element in the gasification fly ash. Oxygen carbon ratio was significantly affected particle carbon content and slightly affected content of elements such as S, Fe, Al and Si. Carbon content in the fly ash decreased gradually with oxygen carbon ratio increased. At the oxygen carbon ratio increased from 0.9 to 1.1, the carbon content decreased from 72.2% to 59.6%. The carbon content at various sampling ports had been tested at a fixed oxygen carbon ratio (O/C=1.0). The results show that, fly ash carbon content in position 1# is the highest 78.6%, positions 2# and 3# carbon content are 50.3% and 48.8% respectively. The concentration of unburned carbon decreases along the gasifier because of the residence time of fly ash increased with the high temperature reaction when it flows down to the slag outlet.

3.2.2 Particle size distribution Particle size distribution at different oxygen carbon ratio was shown in Fig. 5. The results indicate that, the fly ash had a bimodal distribution having two peaks. However, the curves of particle size distribution were obviously difference with various oxygen carbon ratios. More fine particles formed under the oxygen carbon ratio of 1.0 (more particles were ranged in size of 10-100µm, less particles were larger than 100µm), and particles formed under the oxygen carbon ratio of 1.1 were mainly coarse particles. For the oxygen carbon ratio of 0.9, particle size distribution was between the two other cases of distribution. It could be interpreted as low carbon conversion and more residual carbon in particles when the oxygen carbon ratio was 0.9. For the case oxygen carbon ratio is 1.1, besides the highest carbon conversion, many minerals were vaporized and condensed to formed fine particle aggregation. Furthermore, included mineral was more likely to coalescence under this condition.

5 4

O/C=0.9 O/C=1.0 O/C=1.1

Vol. /%

3 2 1 0 0.01

0.1

1

10

100

1000

Particle size /μm

Fig. 5

Particle size distribution at different oxygen carbon ratios

4. Conclusion Effect of temperature and oxygen carbon ratio on fly ash properties in a laboratory-scale OMB coal water slurry gasifier was investigated. The main conclusions are as follows: Fly ash particles in the gasifier performed irregular shape and carbon is the main element. Operation temperatures affect the morphology, surface element distribution and size distribution of fly ash particles. Higher temperature is beneficial to form spherical particles and increase content of element such as S, Fe, Al etc. Oxygen carbon ratio mainly affected carbon content of fly ash, carbon content decreased as the oxygen carbon ratio increased. At a fixed oxygen carbon ratio, the concentration of unburned carbon decreases along the gasifier. Fly ash particles had a bimodal distribution. The peaks of fine particles were almost coincident, and the peaks of coarse particles changed with various oxygen carbon ratios.

Acknowledgement This work is financially supported by National Key State Basic Research Development Program of China (973 Program, 2010CB227006), National Nature Science Foundation of China (20876048).

References [1] Fuchen Wang. Progress on the Large-scale and High-efficiency Entrained Flow Coal Gasification Technology[J]. China Basic Science, 2008, 10(3): 4-13. [2] R. H. Matjie, C. Van Alphen, Mineralogical features of size and density fraction in Sasol coal gasification ash, South Africa and potential by-products[J]. Fuel, 2008, 87(10):1439-1445. [3] M. Ahmaruzzaman. A review on the utilization of fly ash[J]. Progress in Energy and combustion Science. 2010, 36(3): 327-363. [4] Chow J C, Watson J G. Review of PM2.5 and PM10 apportionment for fossil fuel combustion and other sources by the chemical mass balance receptor model[J]. Journal of Energy and Fuels, 2002, 16(2): 222-260.

[5] M. Aineto, A. Acosta, J. Ma. Rincón, M. Romero. Thermal expansion of slag and fly ash from coal gasification in IGCC power plant[J]. Fuel, 2006, 85(16): 2352-2358. [6] L. Bartoňová, Z. Klika, D.A. Spears. Characterization of unburned carbon from ash after bituminous coal and lignite combustion in CFBs[J]. Fuel, 2007, 86(3): 455-463. [7] N.J. Wagner, R. H. Matjie, J.H. Slaghuis, et al. Characterization of unburned carbon present in coarse gasification ash[J]. Fuel, 2008, 87(5): 683-691. [8] T. Wu, M. Gong, E. Lester, et al. Characterisation of residual carbon from entrained-flow coal water slurry gasifiers[J]. Fuel, 2007, 86(7): 972-982. [9] Hiromi Shirai, Hirofumi Tsuji, Michitaka Ikeda, et al. Influence of combustion conditions and coal properties on physical properties of fly ash generated from pulverized coal combustion[J]. Energy and Fuel, 2009, 23(7): 3406-3411. [10] Stanislav V. Vassilev, Christina G. Vassileva. Methods for Characterization of Composition of Fly Ashes from Coal-Fired Power Stations: A Critical Overview[J]. Energy & Fuels, 2005, 19(3): 1084-1098.

GTI’S SAMPLING AND ANALYSIS SYSTEMS FOR GAS STREAMS OF  GASIFICATION AND DOWNSTREAM PROCESSES    Osman Mehmet Akpolat,  Gas Technology Institute  Des Plaines, IL, U.S.A.  Tanya S. Tickel, Gas Technology Institute, U.S.A.  Rachid B. Slimane, Gas Technology Institute, U.S.A.  Chun W. Choi, Gas Technology Institute, U.S.A. 

Abstract:   Since early 2004, GTI has been operating its state-of-the-art pilot-scale gasification facility, the Henry R. Linden Flex Fuel Testing Facility (FFTF), to investigate the renewed interest in various gasification technologies. The use of FFTF in the recent years has been a key point in GTI’s focus on the development and commercialization of various programs ranging from research in evaluation of selected coal and biomass feedstock gasification characteristics, to programs where performance evaluation of downstream syngas processing units take the front stage. Syngas analysis of different constituents is important at different process locations. During these research programs, GTI has developed and implemented innovative analytical approaches to meet the challenging task of characterizing the composition of process streams. In addition to measuring major gas species, techniques are applied to analyze select contaminant species both upstream of cleaning, at higher concentrations and downstream of cleaning at trace levels. These systems ensure precise and accurate data that provide for the performance evaluation of the complete process. With every completed project on the FFTF, GTI has optimized these systems further and continues to develop flexible, scalable and reproducible analytical solutions for sampling and analysis for gas streams of gasification and downstream processes. In this paper, these systems are described to illustrate the effectiveness of these innovative systems. The main objective of the discussion is to illustrate the various unique analysis equipment capability of GTI. During the development process many analysis instruments, such as gas chromatographs (GC), mass spectrometers (MS and GC/MS), Fourier transform infrared spectrometers (FTIR) and multiple wet chemistry methods, that are commonly found in much less challenging laboratory environments were adapted to be used in a pilot plant environment. Multiple disciplines of chemistry, chemical engineering and mechanical engineering were merged together to interface with some of the most challenging sampling locations to condition and then deliver representative samples to delicate analysis equipment that were expected to operate almost autonomously. This paper shows how GTI has developed its analytical solutions capability and how such innovative systems can add significant value to any research that is on the pilot scale where gas composition analysis is critical.

Introduction:  Over the last 6 years, GTI has been operating FFTF for various clients to evaluate the performance of different gasification technologies. The common challenge between all of these projects has been collecting and reporting

precise and accurate data. The collected data must be representative of the process conditions in order for the evaluation of the process to be meaningful. Over the years GTI has developed and optimized sampling and analysis solutions to successfully combine some of the most challenging sampling conditions in gasification processes, to some of the most delicate analytical equipment usually reserved for chemical laboratory use. These systems can be separated to three parts: • Sample probe • Sample conditioning • Sample analysis

PROCESS

ANALYZER

Process

Sample Probe

Sample Conditioning

Sample Analysis

Figure 1. Syngas Sampling System Components

This paper will focus on only the sample analysis portion, and will give detailed descriptions of the equipment used and how they are setup. Further information on sample probes and conditioning can be obtained by contacting coauthors Tanya Tickel and Rachid Slimane. The methods developed by GTI will be discussed grouped by compounds of interest. Each group will describe the equipment used and modification & customization performed to operate them almost autonomously. Almost all of the analytical equipment used in the FFTF has been designed for chemical laboratory use. GTI has successfully integrated these delicate analytical equipments to process environment and has developed its analytical solutions capability over many years of testing at pilot plant conditions. This has produced precise and accurate data that has been very valuable to pilot scale research where gas composition analysis is critical.

  Methods:  Major Syngas Species  The first group of compounds of interest is the major syngas species. This group contains H2, CO, CO2, CH4, N2, H2S and COS. During initial stages of operation, O2 is also analyzed to monitor the combustion process; however it is not considered one of the major species analyzed during gasification. The concentrations of these compounds can change from a few percents by volume to 50% (v) depending on the process and the stages of the process.

Over the years GTI has used multiple Varian CP-4900 Micro Gas Chromatographs with thermal conductivity detectors (TCD) to measure the major species. These Micro GC’s were very efficient in analyzing the gas streams and produced very accurate and precise data when compared to spot samples taken and analyzed by GTI’s Chemical Research Services department. The typical detection limit for compounds of interest was about 10ppmv to 20ppmv depending on the compound of interest. The micro GC’s were configured with dual channels. Each channel is configured to analyze a portion of the compounds of interest. This resulted in a typical micro GC analysis of less than two minutes. The micro GC’s had dual carrier gas capability and used He and Ar as carrier gases. Channel one used Ar carrier gas and analyzed for H2, O2, N2, CH4 and CO. This channel used a Molsieve 5A column with a heated and back flushed injector. The back flush option allowed the unwanted compounds or compounds that would elute too late to be flushed out of the pre column and not injected to the analytical column. Restriction

System Pressure

Pre-column

Analytical Column

Detector

Pressure Regulator Pressure Point

Injector

Backflush Vent

Figure 2. Varian Micro GC Back Flush Option Flow Diagram Channel two used He carrier gas and analyzed for CH4, CO2, H2S, COS. It used a PorapakQ column and it also had a heated and back flushed injector. In this channel it was possible to analyze acetylene, ethylene and ethane but calibration for these compounds were not done. During syngas sampling the concentrations of these compounds were observed to be below detection limit and they were not reported in the major gas species group.

Figure 3. Varian CP-4900 Micro GC During regular testing, micro GCs at multiple sampling locations would run continuously to determine syngas gas composition. This online measurement would produce about 700 samples a day that was recorded and charted using an in house developed Excel macro package. The operator would start sampling and then the data was logged on an Excel spreadsheet and charted during the pilot plant testing.

The micro GC method and calibration was tested before and after each run. Four calibration gas mixtures covering the concentration range for each major gas species reported were used. Also on channel two, He was injected to record a “blank baseline”. This was later subtracted from each analysis on that channel to flatten the baseline and have more accurate integration of the compound peaks. During long term pauses in operation an intermediate calibaration check would be done. Later this was omitted since the calibration check before and after each pilot plant test showed very little change in the response factors. Using Molsieve column in the micro GC made that channel very susceptible to moisture. The accumulated water would reduce the effective length of the column by occupying the active sites on the column. This resulted in retention time shifting, and caused peaks to disappear from the charts. Since the peak would shift outside the retention time window, it would not be identified and therefore the Excel macros would not report the concentration. To prevent this, the micro GCs were kept at high temperature “bakeout” modes which restored the column to full efficiency. For long test periods a dual system micro GC with two of each channel was acquired. When a pair of channels was in use the second pair was in bakeout. This minimized the time without any data due to column regeneration. In addition to micro GCs, other equipment such as FTIR and GC-FID were used in early test campaigns to cross check results. The variations in measured syngas concentrations between reported values from different instruments were about 2%. Later due to its ease and unattended operation micro GCs were used as the primary online major species analysis instrument. Spot samples were still used to verify the gas concentrations that were reported. The spot samples were taken either daily or once every steady state period in the early pilot plant runs. Water Content of Syngas  Water content of the syngas was measured by various methods. GTI’s experience during other projects has been that water is one of the most difficult compounds to measure accurately. Most available moisture analyzers do not work at gasifier sample point conditions, and conditioning gas streams to work for these analyzers usually removes most of the water. A typical syngas sample right after the gasifier can be 1600°F, 70 psig and can contain around 20% water. Temperature of the sample stream makes it a very challenging sample point. GTI’s first approach was to dilute the sample gas with an inert gas. When this diluted gas was cooled and depressurized, any condensable compounds that were present in the sample stream would stay in gas phase without condensation. The resulting sample stream was then analyzed using an IMACC process FTIR with a specially developed method to match the syngas sample gas concentrations. This data was then logged and charted with the major gas species data from the micro GCs. A fairly extensive custom macro was developed to automate this process and chart the resulting data using Microsoft Excel. In addition to this dilution approach, a gravimetric moisture collection method was developed. The hot gas stream was sampled through water absorbent material and the dry gas exiting the absorbent was measured with a dry gas meter. Water mass collected over time was divided by the volume of sampled gas to determine the water concentration. This method proved reliable however resulted in very rapid saturation of the absorbent material due to high water content of the syngas. The average sampling period for a gravimetric water sample was around one hour. The syngas flow rate was set at around one liter per minute. During initial method development it was observed that water collected during sampling was too much for efficient use of the absorbent material. A water condensation vessel in an ice bath was installed to collect most of the water before the sample gas contacted the absorbent material. This proved to be very efficient with minimal variation in reported values. The sampling time was adjusted to achieve the desired accuracy. After multiple pilot plant runs, it was decided that only gravimetric water sampling was satisfactory to report the water content of the sample gas. The period of sampling was set to every 3-4 hours with a sample collection time of around one hour.

In addition to these methods, another process water analyzer was installed in process line exiting the gasifier and tested. This was a moisture analyzer commonly used at boiler stack moisture measurement. This instrument was found to be quite hard to implement in the syngas stream right after the gasifier due to the location’s harsh environment. The probe for the analyzer was hard to keep clean. Dust and particulates accumulated on and around the probe preventing proper sampling. Heavy hydrocarbons & light tars  Initial studies in heavy hydrocarbon and light tar analysis started with projects that were performed to characterize the syngas composition during biomass gasification. Initially a standard HP 5890 GC with a flame ionization detector (FID) was used. A non polar DB1 analytical column was used to monitor benzene, toluene, ethyl benzene, o,m,p-xylenes (BTEX) and naphthalene. Sample conditioning systems were developed to transfer sample gas from high temperature sample probes in the process lines to the analyzer in the analytical area. The GC was customized in GTI to inject directly to column. The HP 5890 was integrated to Varian chromatography software and data collected was merged with major gas species results using the GTI custom macros. The results were reported in Microsoft Excel.

Figure 4. HP 5890 GC with Flame Ionization Detector After successfully monitoring BTEX and naphthalene, a detailed characterization of biomass syngas was done with an Agilent 6890 GC and an Agilent 5973 Mass Selective Detector (MSD). This analyzer was assembled by Wasson Ece and used a proprietary Wasson 2318 analytical column (KC1 phase was similar to DB1) to separate 23 compounds. The analyzer was calibrated at the vendor. The syngas sample was delivered to the analyzer using GTI’s sample conditioning systems and injected at 200°C into the GC using a 1 ml sample loop and a 6 port injection valve. The detection limit was 0.01 ppmv.

Figure 5. Agilent 6890 GC with Mass Selective Detector The Agilent 6890 GC used Agilent Chemstation MSD software to collect data. Each run was about 1 hour and the resulting analysis was integrated to the major gas species using the custom GTI macros. The monitored compounds were: Table 1. Compounds Analyzed by Agilent GC-MSD 1. Methanethiol 2. Carbon Disulfide 3. Ethanethiol 4. Benzene 5. Thiophene 6. Toluene 7. Ethylbenzene 8. p-Xylene 9. m-Xylene 10. o-Xylene 11. Naphthalene 12. Benzothiophene 13. 2-Methylnaphthalene 14. 1-Methylnaphthalene 15. o-Cresol 16. Phenol 17. p-Cresol 18. m-Cresol 19. Acenaphthene 20. Acenaphthylene 21. Fluorene 22. Phenanthrene 23. Anthracene The same GC was also used with a flame ionization detector (FID) to measure light hydrocarbons. Using 1 ml sample injection and a set of Wasson Ece 3018+2378 analytical columns (KC153 and KC5 phases) the following light hydrocarbons were analyzed: Table 2. Compounds Analyzed by Agilent GC-FID 1. Methane 2. Ethane 3. Ethylene 4. Propane

5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.

Propylene Acetylene Isobutane Propadiene n-Butane t-2-Butene 1-Butene Isobutylene c-2-Butene Neopentane Isopentane Methyl Acetylene

After initial characterization of heavy hydrocarbons and light tars, it was observed that about 95% of these compounds were single or double aromatic ring compounds. For later pilot plant testing a simpler to operate analyzer was investigated. It was decided that an Agilent 7890 GC with dual FID will be able to sample two streams and analyze them at the same time. This would allow for online comparison of hydrocarbons before and after a process point. The syngas sample was conditioned and delivered to the analyzer at 200°C and injected using 1 mL sample loops and Agilent sampling valves. The sampling valves were connected to Agilent GS-GasPro columns. The analyzer was calibrated to monitor benzene, toluene and naphthalene. It was used in multiple biomass gasification campaigns and provided online benzene, toluene and naphthalene concentrations.

Figure 6. Agilent 7890 GC with Dual Flame Ionization Detectors In addition to online analysis equipment, GTI has used impingers to collect and dissolve hydrocarbons in syngas samples. The solutions that the compounds are captured were analyzed using GC-FID by GTI Chemical Research Services after the test campaign. This would generate the largest number of compounds and has the lowest detection limits. A typical syngas sample capture would require one to two hours of sample flow. This would flow enough sample gas to satisfy the detection limits of the laboratory instruments. This period was modified when capturing samples with suspected high concentrations of hydrocarbons to prevent overloading the laboratory equipment. Low concentration sulfur & other contaminants  One of the earlier projects for FFTF involved H2S removal to trace levels in the syngas. To evaluate the performance of this process, a highly specialized and sensitive sulfur analyzer was ordered. This Multi Purpose Sulfur Analyzer (MPSA) was custom made by Varian using a Varian 3800 GC and a Varian Pulsed Flame Photometric Detector. The standard detection limit of PFPD was further improved by using a proprietary cold trap. The MPSA used

multiple sample loops depending on the sample concentration. The operator would select the proper method to run and the MPSA would select proper valve positions for the transfer of that sample loop to the cold trap. Then the cold trap would be blocked and heated very rapidly evaporating all the sample that had been collected. Finally this sample would be injected into a Varian CP-sil5 non polar column and analyzed by PFPD.

Figure 7. Varian 3800 GC based Multi Purpose Sulfur Analyzer This analyzer was extremely sensitive. When the largest sample loop was used the detection limit was about 1ppt H2S. At the same time using the smallest loop samples containing over 1000ppm H2S were analyzed without reconfiguration. However the operational requirements for the analyzer were high and it took about one hour for one sample to be captured and analyzed. Varian MPSA delivered trace level H2S measurements during pilot plant testing. But later it was replaced by an Agilent 6890 GC with a Sievers 355 Sulfur Chemiluminescence Detector (SCD). During most of the pilot plant testing, this GC-SCD was used as the sulfur containing compound analyzer. Both coal and biomass gasification testing used this instrument to provide online data at 30 minute intervals. The GC used a 1 ml sample injection loop and a Wasson Ece 2332 analytical column (KC68 phase) to identify 15 sulfur compounds. Standard detection limits were at 0.005 ppmv. Agilent Chemstation software was used to integrate the results with the major gas species using custom macros developed in GTI. Final data was represented as Microsoft Excel charts. Table 3. Compounds Analyzed by Agilent GC-SCD 1. Hydrogen Sulfide 2. Carbonyl Sulfide 3. Methanethiol 4. Ethanethiol 5. Dimethyl Sulfide 6. Carbon Sulfide 7. 2-Proanethiol 8. 2-Methyl-2-Propanethiol 9. 1-Propanethiol 10. Thiophene 11. Diethyl Sulfide 12. 1-Butanethiol 13. Diethyl Disulfide 14. Benzenethiol 15. Benzothiophene

During some pilot plant testing the H2S and COS concentrations were high enough to use micro GCs instead of these specialized sulfur analyzers. The detection limit when a micro GC was used increased to about 20 ppmv due to the limitations of the micro GC. Miscellaneous  In addition to process monitoring at gasifier or downstream locations to generate project reports, GTI has used different analyzers for process monitoring only. One of these analyzers was a Rosemount NDIR that monitored H2, CO, CH4 and O2. This analyzer was mainly used to monitor the change from combustion to gasification during the initial startup stages. The sample conditioning systems were the same as the rest of the analyzers. The readout from this analyzer was connected to the plant control system to give operators and indication of some of the major species during the test campaign. Another such analyzer was a Honeywell lead acetate tape H2S analyzer. This was located between two sulfur guard beds in the pilot plant. These guard beds were in series and the analyzer sampled after the first guard bed. The purpose of this analyzer was to determine if the first sulfur guard bed was used up. When the first guard bed was used up, the operators would make a decision to stop or continue the test based on test plan, predicted life of the second bed etc.

Discussion & Conclusion:  Henry R. Linden Flex Fuel Testing Facility has been a state-of-the-art pilot scale gasification facility for GTI to investigate the renewed interest in various gasification technologies. Since 2004, with the addition of FFTF to GTI facilities, GTI has been focusing on the development and commercialization of various programs ranging from research in evaluation of selected coal and biomass feedstock gasification characteristics, to programs where performance evaluation of downstream syngas processing units take the front stage. During these research programs, GTI has developed and implemented innovative analytical approaches to meet the challenging task of characterizing the composition of process streams. In this paper, we have tried to describe these systems in an effort to illustrate the various unique analysis equipment capability of GTI. Representative syngas analysis at multiple process locations is challenging and at the same time necessary to evaluate the process properly. GTI’s analytical systems ensure precise and accurate data that provide for the performance evaluation of the complete process. With every completed project on the FFTF, GTI has optimized these systems further and continues to develop flexible, scalable and reproducible analytical solutions for sampling and analysis for gas streams of gasification and downstream processes. This paper shows how GTI has developed its analytical solutions capability over the several years at FFTF pilot scale testing facility and how real world experience modified and focused these systems to be more efficient with each completed test campaign. Multiple iterations of sample conditioning systems have been developed and modified over the different projects to improve and increase efficiency and design. Together with the developments in sample conditioning many instruments that very originally designed to be operated in much less challenging laboratory environments were adapted to be used in a pilot plant environment. Some were modified and kept such as the micro GCs and some more specialized equipment now wait for new projects where they can add significant value to project deliverables. In addition to measuring major gas species, different techniques were applied to analyze select contaminant species both upstream of cleaning at higher concentrations and downstream of cleaning at trace levels. Operating FFTF through multiple projects, GTI has been able to understand how to sample and analyze gas compositions in pilot plant gasifier operations and was able to use highly specialized equipment to determine the most dominant areas to focus during plant operations.

As pilot scale testing continues at FFTF, GTI has been and will be investigating and evaluating new process monitoring equipment that might replace the modified chemical laboratory analyzers that are currently in use with highly sensitive process analyzers that are being developed now.

Acknowledgement:  The author would like to acknowledge co-authors Tanya S. Tickel, Rachid B. Slimane and Chun W. Choi and the GTI FFTF Analytical Team for their hard work and contributions.

Manuscript Not AVAILABLE

DESIGN OF COMMINUTION UNIT FOR THE GASIFICATION PILOT PLANT

N. ACARKAN, G. ONAL, A. A. SIRKECI, G. ATESOK [email protected]

ABSTRACT

The comminution unit for the gasification pilot plant to be constructed for Turkish Coal Enterprises has been designed at the Department of Minerals Processing of Istanbul Technical University.

It has been designed that the comminution unit will be fed with coal below 100 mm and grind it under 100 m whose moisture content should be 1% maximum prior to be fed to the gasification unit. The capacity has been chosen to be 1800 kg/h. The run of mine coal should have 18% moisture, 25% ash and 38% volatile matter at maximum.

The primary crushing in the unit will be accomplished using a hammer crusher below -20 mm. Followed by crushing the moisture of the product will be decreased less than 10% from about 18% of initial value. The crushed and partially dried product will be stored in a silo before feeding to a vertical mill. Coal will be ground below 100 µm in the vertical mill while dropping the moisture content under 1%.

The project for a commiution process has been designed according to the information given above.

27th PCC, Oct. 11-14, 2010, Istanbul, Turkey

Preparation of coal water mixture with high concentration from low rank coals and lignite by dry fine coal with optimum particle distribution Feng He1 and Baoqing Li2 1

Yulin Western Coal Technology Research Center, Shan’Xi, Shenmu，719300 China

2

State Key Lab of Coal Conversion, Institute of Coal Chemistry, CAS, Taiyuan, 030001 China

Abstract: A new technology for preparation of coal water slurry (CWS) with high concentration from low rank coals and lignite by dry fine coal with optimum particle distribution was developed. According to coal characteristics the micronized coal was achieved by pre-drying and dry grinding of coal using several modified mills to obtain various fineness coals. CWS powder, which can prepare high concentration CWS, was produced by mixing different ratio of various fineness coals in optimum condition. CWS with concentration of up to 68% of eligible slurry using low rank coals has been produced by this technology in the industrial CWS powder plant with capacity of 750,000 t/a. 63% CWS using lignite as raw material has also been produced. Compared with conventional wet process for CWS preparation, the new technology provides less power consumption of about 5 kWh/t CWS and a half amount of additives. The reduction of the relative cost is about 8 CHYuan/t CWS with obvious economic and social benefits. This new technology will greatly promote the clean conversion and utilization of low rank coals and lignite. Keywords：coal water slurry; low rank coal; lignite; coal gasification

Introduction Development and promotion of clean coal technology to China's implementation of energy saving and sustainable development are of great significance. Coal water slurry (CWS), as a new coal-based liquid fuels due to its energy saving, substitution of oil, high-efficiency combustion, clean and easy-to-pump spray, etc., has been listed as one of clean coal technology and of national encouraging and developing key industries . Gasification technology is a core technology of clean coal technologies, which is leading technology for the development of coal-based chemicals (ammonia, methanol, dimethyl ether, alkylene, etc.), coal-based liquid fuel, multi-generation systems, hydrogen and other key process. Most modern gasification technology uses entrain bed slag gasifier including CWS and dry feeding gasification. Thus, the low rank coals with low ash melting point such as Shenmu coal is needed as raw material. CWS feeding GE (Texaco) gasification and multiple nozzle CWS gasification technology developed by the East China University of Science and Technology have been widely used in domestic coal-based ammonia and methanol production. However, Shenmu coal is featured by poor slurryability of CWS (the highest concentration of 60%) due to its high inherent moisture and surface area with high content of polar functional groups. Thus, the energy consumption and costs are increased，affecting the economic benefits. Meanwhile, lignite is an abundant resource in China, but its conversion and use are a major problem because it is also characterized by poor slurryability of CWS (about 50%) and difficult to prepare CWS with high concentration. To enhance the CWS concentration using low-rank coals, many institutes, universities and

27th PCC, Oct. 11-14, 2010, Istanbul, Turkey

companies did a lot of effort over the years, but with little success. Yulin Western Coal Research Center broke the traditional wet technology of CWS preparation and developed a new technology named dry fine coal with optimum particle distribution for CWS preparation with high concentration using low rank coals and lignite. More than 68% and 63% of eligible CWS can be produced using Shenmu coal and Inner Mongolia Baorixile lignite as raw materials, respectively. This technology will improve the economic and social benefits in the coal-based methanol and chemical enterprises by reducing energy consumption and amount of additives, and greatly promote the conversion and utilization of low rank coals and lignite especially their CWS gasification and related chemical synthesis. This article describes the development of the new technology of CWS preparation. The energy consumption and related cost are compared with the conventional technology.

1. Situation of CWS Preparation CWS is normally prepared by wet technology, namely coal, water and additives are added in a ball mill to produce CWS. The process is simple and CWS with 65% slurryability can be produced using the wet technology for most of coals by adjusting the distribution of various balls in the ball mill. However, it showed poor adaptability for low rank coals and lignite. The conventional technology is not suitable for coals with high inherent moisture and less grindability, since it is difficult to increase grain fineness to achieve the optimum particle size distribution (PSD) accompanied by preparation of high concentration CWS. The Raymond Mill machine is available for dry CWS preparation. However, the design of the classifier is unreasonable. All the leaves in the classifier are in strict accordance with the radial layout, which results in the high residence for air flow when the classifier is run at high speed. Large amount of qualified powders cannot be exited, leading to the blockage and a decrease in the processing capacity of the mill. The situation is more sever for raw materials with relatively high moisture such as lignite. The lower speed of air flow makes difficulty to loose the materials and to raise the particle fineness for the optimization of PSD. Meanwhile, the existing Raymond Mill machine depends mainly on the blower for transportation of materials. It brings a problem that air flow inside the mill must keep strict balance, otherwise, it will easily lead to positive pressure, resulting in powder leakage with environmental pollution. 2. Development of CWS preparation technology with high concentration using low rank

coals and lignite by dry fine coal with optimum PSD. After the analysis of existing CWS preparation technology combined with the features of low rank coals and lignite, we modified the Raymond machine and developed a new technology named dry fine coal with optimum PSD for CWS preparation with high concentration using low rank coals and lignite. A semi-industrial plant with annual output of 200,000 tons of dry CWS product has been established in 2009, and run and operated smoothly. After reformation, the capacity of the plant has been reached to 750,000 tons of dry CWS product per year. About 20,000 tons of dry CWS product has been produced and used for preparation of 69% CWS with viscosity of 800 mPa.s (100s-1) using Shenmu coal. The schematic sheet is shown in Fig. 1.

27th PCC, Oct. 11-14, 2010, Istanbul, Turkey

Raw coal bunker Hammer 0-8mm PLC feeding system Ultra-fine mill

Mid fine mill

Mid coarse mi

Coarse mill

Bunker

Bunker

Bunker

Bunker

Elec. scale

Elec. scale

Elec. scale

Elec. scale

Dry fine product bunker 30%additive jar

CWS storage tank

CWS Mix-stiring tank CWS trasport

Fig. 1 Schematic sheet of dry technology for high concentration CWS preparation Raw coal (without pre-drying for lignite with humidity 4 个月

>4 个月

>4 个月

>4 个月

>4 个月

>4 个月

Pseudoplastic

Pseudoplastic

Pseudoplastic

Pseudoplastic

Pseudoplastic

Pseudoplastic

Amount of additive CWS stability

No hard precipitation

CWS rheological ability

Table 1 shows that high concentration CWS with good rheological ability and stability with less amount addition of additive has been obtained from Shemu coal and lignite. By using this new technology, higher than 68% of CWS was produced for Shenmu low rank coal with moisture of 4.16%-6.85% in analytical base and 63% for Inner Mongolia Baorixile lignite with moisture of 15.76%. 2.1 Technical Features 1）With several self-improved dry mills, the average particle size of coal powder is less than 10 microns (typically 40-50 microns by conventional wet process) with the best PSD and the highest accumulation density. Thus, the quality of CWS is improved and the adaptability of different coals is expanded. High concentration CWS can be prepared using low rank coals and lignite. 2）CWS can be prepared for coals with higher inherent moisture even in the frozen season due to its high grinding efficiency. Thus, there is no frozen and agglomeration during transportation and storage because the dry fine powder contains 12% of humidity. 3）The fine powder containing less humidity as the product is packed by bag to avoid the specialized tanker for CWS transportation and its transport weight is also greatly reduced, leading to a significant decrease in transportation costs 4）CWS is prepared by end-user in situ and, thus, the stability of CWS is not an important factor. It is suitable for some coals that cannot produce CWS with high stability. The cost using additives can also be reduced. 5）CWS powder preparation is more suitable for North China due to coal-rich but water shortage situation. 2.2 Technological Innovations 1）Being different from the conventional wet technology, a new technology of high concentration CWS preparation using low rank coals and lignite, named dry fine coal with optimum PSD, is developed. Higher than 68% of qualified CWS is produced for Shenmu low rank coal and 63% for Inner Mongolia lignite with less energy consumption and amount of additives. The production costs are greatly reduced and the adaptability of different coals is expanded. It will also promote the conversion and utilization of low rank coals and lignite. 2）The design of the existing classifier in the mill is improved with operation under negative pressure to improve the adaptability of raw material, the powder quality and the processing capacity of the mill, and to eliminate environmental pollution. 3）An anti-static device with recovery of ultrafine coal under negative pressure is developed to ensure that the temperature rising caused by the friction between the internal grinding plate and cylinder during operation of the dry-mill can be controlled to less than 20 ℃. 95 % of the hot air was sucked away under negative pressure, and thus there is no dust leakage. 4）Multi-dry mills are used to produce pulverized coal with different yields and fineness to achieve the required optimum PSD related to coal used. 2.3 Comparison of energy consumption and relative costs of different CWS preparation

27th PCC, Oct. 11-14, 2010, Istanbul, Turkey

processes The Comparison of energy consumption and relative costs of different CWS preparation processes are shown in Table 2. Table 2 Comparison of energy consumption and relative cost between dry and wet CWS preparation technologies Electricity Technology

Yuan/ kWh

Yuan/t t/tCWS Yuan/t

Yuan/ tCWS

t/tCWS Yuan/t

Yuan/ Total cost tCWS

76%

20

0.46

9.2

0.003

4000

12

0.002

2000

4

25.20

Shenhua

73%

18

0.46 8.28

0.002

4000

8

0.002

2000

4

20.28

68%

18

0.46 8.28

0.010

4000

40

0.002

2000

4

52.28

63%

18

0.46 8.28

0.004

4000

16

0.002

2000

4

28.28

66%

18

0.46 8.28

0.003

4000

12

0.002

2000

4

24.28

63%

18

0.46 8.28

0.005

4000

20

0.002

2000

4

32.28

59%

23

0.46 10.58

18

28.58

59%

22

0.46 10.12

18

28.12

58%

23

0.46 10.58

18

28.58

59%

23

0.46 10.58

18

28.58

(Hailaer)

Dry

kWh/t

Additive-II

Shenhua

Lignite

Lignite (Hailaer) Lignite (yunnan) Lignite (Baodou) Shanghai (Shenmu) Shenmu Ch.Eng.

Wet

Conc.

Additive-I

Yulin Mining （Yulin） Hualu Hensheng (Shenhua)

It is found from Table 2 that compared with CWS wet process, the energy consumption can be decreased by 5 kWh /t CWS with half amount of additive addition using dry technology. Thus, the relative cost can be reduced about 8 CHYuan / tCWS. At the same time, due to the increase in the concentration of CWS, coal and oxygen consumption per unit product in gasification will be reduced and the effective components in synthesis gas will be increased to enhance the efficiency of coal gasification, which will also reduce the energy consumption and costs, and increase the social and economic benefits.

3 Conclusions A new technology of high concentration CWS preparation using low rank coals and lignite, named dry fine coal with optimum PSD, was developed. A demonstration plant with annual output capacity of 750,000 tons of CWS dry powder was established and operated. By this technology, higher than 68% of CWS was produced for Shenmu low rank coal with moisture of 4.16%-6.85% in analytical base and 63% for Inner Mongolia Baorixile lignite with

27th PCC, Oct. 11-14, 2010, Istanbul, Turkey

moisture of 15.76%. Also, energy consumption and amount of additives were reduced with decrease of production costs. The development of the technology resolves the problem of high concentration CWS preparation using western China rich coal resources of high quality low rank coals and lignite. The use of this technology can greatly enhance the economic and social benefits for coal chemical industry using CWS gasification. It also opens up broad prospects for the clean conversion and utilization of lignite.

References 1. Jiashan Wu, Qingjun Ji. Effect of properties of Shenmu coal on characterization of its coal water slurry. Coal Conversion, 1992, 15, 69-75 2. Chenggong Sun, Jiashan Wu, Baoqing Li. Characterization of the rheological behaviour of thermally upgraded low-rank coal water slurry. I Effect of coal upgrading temperature. J Fuel Chem & Technol., 1996, 24, 131-136 3. Xianrong Ye, Dinping Liu, Qizhong Chen, Hongming Cai. Effect of particle size distribution on concentration and viscosity of coal water slurry of blending coals. Coal Conversion, 2008, 31, 28-30

PROGRAM TOPIC: GASIFICATION FLUIDISED BED CO-GASIFICATION OF COAL AND BIOMASS UNDER OXYFUEL CONDITIONS

Nicolas Spiegl, Nigel Paterson, Cesar Berrueco and Marcos Millan* Department of Chemical Engineering, Imperial College London, London SW7 2AZ, UK *Presenting and Corresponding Author Contact Information: Department of Chemical Engineering, Imperial College London, London SW7 2AZ, UK Phone: +44(0)20 7594 1633 [email protected]

ABSTRACT Fluidised bed gasifiers are the preferred option to utilise low value coal, biomass and waste. However, fluidised bed gasifiers are traditionally air rather than oxygen-blown to avoid high temperatures in the gasifier leading to ash melting and loss of fluidisation. Therefore the flue gas of a possible FB-IGCC plant would be diluted by nitrogen, making expensive N2-CO2 separation technology necessary for subsequent capture and storage of the CO2 (CCS). To overcome this disadvantage, an oxy-fuel process is proposed, where the bed is fluidised with recycled flue gas (mainly CO2) and oxygen. A laboratory scale fluidised bed gasifier capable to operate up to 1000°C and 20 bar was set up to study the implications of oxy-fuel firing on flue gas composition and overall operability of the gasifier. High carbon conversions were achieved at 950°C with CO2 as gasification agent. At that temperature, the addition of O2 generated some agglomeration in the bed, leading to lower conversions. At lower temperatures, O2 addition increased carbon conversion, but not the heating value of the gases. The addition of steam enabled the gasifier to operate at lower temperatures without decreasing conversion and to increase the hydrogen content of the fuel gas. Pressure on the other hand produced a decrease in carbon conversion, mainly above 10 bars, due to a larger formation of secondary char with the subsequent loss in reactivity.

These results show that oxy-fuel firing of a fluidised bed gasifier could be a promising route to avoid N2 dilution of the fuel gas and enable integration of fluidised bed gasification with CCS technology.

INTRODUCTION Entrained flow gasifiers currently represent the leading technology for coal gasification due to their high throughput rates and conversions. However, in view of the increasing interest in gasification of low grade coals as well as co-gasification of coal with biomass and waste, fluidised/spouted bed gasifiers present some advantages. Firstly, they can process a broad range of fuels, including high ash and high sulphur coals. Secondly, they have less stringent needs regarding fuel preparation, making them suitable for processing biomass and certain types of waste. In addition, fluidised beds have a high tolerance to load changes, which enables them to process feedstocks with changing properties.

Air-blown fluidised bed gasification processes are not particularly well-suited for CO2 capture and storage, as combustion of the fuel gas produces a stream of CO2 diluted with large volumes of N2, whose separation is costly and complex. In order to make this process compatible with CO2 capture, it is proposed in this work the use of an O2-fired gasifier with the injection of CO2 and/or steam as both gasification agents and diluent needed to replace the role of N2 in avoiding high temperatures of the bed, which would cause ash melting and agglomeration and eventually a loss of fluidization.

Replacing N2 with CO2 and steam will affect the gasification performance and the operation of the gasifier. The main changes are an increase in the thermal capacity of the reactor, the tendency to increasing agglomeration of coal/char in atmospheres with a high partial pressure of CO2 and the reactive nature of CO2 and steam in gasification reactions in comparison with an inert gas as nitrogen, which will change the chemistry of the system. This paper focuses on the impact of the latter on the gasification of coal and coal-biomass mixtures.

EXPERIMENTAL

An existing laboratory scale fluidized bed gasifier was adapted to operate with a continuous feed under oxy-fuel conditions. A diagram of the system is shown in Figure 1. The modifications made to this reactor have been described in detail elsewhere [1]. Briefly, the reactor is made of Incoloy 800HT, 34 mm (i.d.), 504 mm (tall), which acts as both pressure shell and resistance heater. The reactor shell is heated by a high current, low voltage ac power supply, with copper electrodes connecting the reactor body to a transformer. A quartz glass liner contains the bed, protecting the reactor shell from corrosion and avoiding sample contamination by the metal shell. The feedstock used in this study was a German lignite.

T1

T2

E1 K1

V2 F3

F3 F1

F2

V9 BFA

V6

V5

F5

P2

Z1

X1 V8

I1

V7 A1

A2

CO CO2

A3

A4

A5

A6

H2

CH4

O2

CO

V13 Ventilation

I2

V4

V3

V13

MFC1

V10 R1 H1 MFC2 G1

P1 P1

O1 S1

V11

T3

V12

D1

T4 C1

U1

C2

C3

C4

T5 V1

U2

C1 C2 C3 C4 C5 C6 P1/2 T1-5 U1,2 P1 E1 K1 F1,2 F3-5

CO2 N2 20% O2/CO2 Air Calibration Gas Propane Pressure Transducer Thermo Couple Heating Tape HPLC Pump Extraction Hood Incinerator Absorber Filter

H1 D1 O1 MFC1,2 I1,2 S1 G1 R1 Z1 X1 A1-6 V9 V6 V13 BFA

Feed Hoper DC Motor Rotary Valve Mass Flow Controller Ice/Water/Salt Bath Spout Line Glass Liner Reactor Shell Tar Trap Gas Meter Online Gas Analysis H2S Measurement Point Pressure Let Down Valve Pressure Relieve Valve Back flow arrestor

Figure 1. Flowsheet of the continuous feed high pressure fluidised bed gasifier.

C5

C6

The tars were collected in a wire mesh-packed trap that uses ice/water/salt mixtures to decrease the temperature of the gases to less than 20°C. The composition of the tar-free gases was then determined by means of online analysers based on infra-red (CO, CO2, CH4), thermal conductivity (H2) and paramagnetic properties (O2).

RESULTS AND DISCUSSION 1) Effect of temperature and O/C ratio. Experiments at 750°C, 850°C and 950°C using different O2/CO2 mixtures as inlet gas were carried out at atmospheric pressure. The molar ratios between inlet O and C in the feed used in these experiments were: 0, 0.1 and 0.2, which correspond to O2 molar fractions in the gas phase of 0%, 5% and 10% respectively. The total inlet flowrate to the reactor was kept constant at 1.76 NL/min and the resulting superficial velocity was 0.16 - 0.18 m/s.

Figure 2. Carbon Conversion (left) and LHV of the fuel gas (right) as a function of the O/C ratio for gasification experiments at three different temperatures (

950°C, □ 850°C, ∆ 750°C).

Carbon conversion markedly increased with temperature in the range 750-950ºC. At 950ºC, the conversion reached 85% (Figure 2, left). Although there was virtually no residual carbon in the bed, entrainment of fines in the outlet gases prevented the system from reaching higher conversions. The LHV of the gases increased steadily with temperature as CO2 became more reactive and reached a value of 10 MJ/m3 for gasification at 950ºC.

For the lower two temperatures used in this study, the conversion increased with the addition of oxygen, as a result of the larger extent of char combustion. At 750°C it can be estimated that approximately 60-70% of the injected O2 reacted to produce CO2, while the rest formed CO. This increased conversion of fuel carbon to CO resulted in an increase in the overall energy conversion from 31 – 36% and in the heating value of the fuel gas from 5.1 to 5.7 MJ/m3 (Figure 2, right). At 850ºC, the LHV of the gases stayed constant at 8MJ/m3 despite the enhanced carbon conversion, as the addition of O2 did not substantially change the concentrations in the gas phase (Figure 3).

Figure 3. Oxy-fuel gasification of German lignite at different temperatures and O/C ratios – Fuel gas composition (□ CO2,
CO, ∆ H2, × CH4).

An altogether different trend occurred at 950ºC, where a drop in carbon conversion was observed with increasing O/C ratio. This was found to be an effect of bed material agglomeration. Analysis of the bed material recovered at different experimental times revealed that at 950ºC with 0.1 and 0.2 O/C ratios, large agglomerates had formed in the bed. These agglomerates built up in the cone of the bed around the injection spout (Figure 4), growing firstly upwards along the spout and then in width. De-fluidisation of the bottom part of the bed and related decrease in heat transfer from the walls explains the decrease in temperature detected by the bottom thermocouple. Due to the high spout velocity

(approximately 4 times bed velocity) the agglomerating material did not block the injection of solid material.

Figure 4. Bed material agglomerates produced by the use of oxygen in experiments at 950°C.

Agglomeration of silica bed material in fluidised bed reactors is a well known phenomenon. Na and K from the lignite ash together with the Si from the sand can form low melting silicates which act like a glue between the sand particles [2, 3]. The build up of agglomeration is directly related to the availability of free oxygen. No agglomeration was detected for experiments with pure CO2 and agglomerated particles built up around the spout, where levels of oxygen was at its highest. The temperature in the spout region where oxygen reacts with char and volatiles was likely to be significantly higher than the overall bed temperature and the main reason for the agglomeration. The build-up of static material in the cone of the bed prevents the recirculation of the bed material by entrainment in the spout gas. Therefore, the advantage of the spouted bed design in increasing mixing is lost, which in turn affects conversion. However, the increasing amount of agglomerates in the bed did not have an adverse effect on the overall stability of the gasification process.

2) Effect of CO2/C ratio Experiments at atmospheric pressure were carried out at 750°C, 850°C and 950°C using different CO2/C ratios. The resulting carbon conversions and fuel gas LHV are summarised in Figure 5. All experiments were carried out with 100% CO2 as fluidising gas. Different CO2/C ratios were achieved by changing the amount of CO2 injected into the reactor and keeping the fuel feeding rate constant. This changes the superficial velocity but separate experiments established that changes in superficial velocity do not significantly alter conversions under these conditions.

Figure 5. Carbon Conversion (left) and LHV of the fuel gas (right) as a function of the CO2/C ratio for gasification experiments at three different temperatures (

950°C, □ 850°C, ∆ 750°C).

The effect of increasing CO2/C ratio was sharpest at 850ºC with a clear increase in carbon conversion from to 60% to 85%. Conversions at 950ºC were independent from the flowrate of CO2 as they had already reached a maximum level due to loss of fines as discussed above. On the other hand at 750ºC, the increase in carbon conversion was less pronounced as CO2 is less reactive at this temperature. By contrast, at 850°C, the char-CO2 reaction seems to be fast enough to be influenced by the partial pressure of CO2. At a CO2/C ratio of 1.8, 85% conversion was reached, the same as that obtained at 950°C. However, it was found that an increase in inlet CO2 dilutes the fuel gas and decreases its heating value at all temperatures. These results show that it would be possible to decrease operating temperature to 850°C and still achieve high carbon conversion by increasing the CO2 to C ratio. The LHV of the gas however drops, from 8 to around 7MJ/m3 in these experiments.

3) Injection of Steam and co-feeding of biomass. Different CO2/steam ratios (0, 0.3, 1 and 2.3) were used in gasification experiments at atmospheric pressure with GL as fuel. No O2 was added in this set of experiments. The total flow rate entering the reactor was kept constant at 2.64 NL/min and the superficial velocity was kept in the range 0.24 - 0.27 m/s. The resulting fuel gas compositions are presented in Figure 6.

Figure 61. Oxy-fuel gasification of German lignite at different temperatures and Steam/CO2 ratios (

CO, □CO2, ∆ H2, × CH4).

With increasing steam/CO2 ratio, the char-CO2 reaction is partially replaced with charsteam reaction. This produced an increase in the H2 concentration and a decrease in the CO concentration in the fuel gas. As shown in Figure 7, replacing 25% of CO2 by steam increased the carbon conversion at 750°C and 850°C while at 950°C no change occurred. Similarly to what was found in the case of CO2/C ratio, at 950°C an upper limit in carbon conversion was achieved, independently of steam/CO2 ratio. By adding 25% steam at 850°C, carbon conversion was increased to a level comparable to that of experiments carried out at 950°C. Further increases in steam to CO2 ratio did not change the carbon conversion at 850°C, although the hydrogen production continued increasing.

Figure 7. Carbon Conversion as a function of the steam/CO2 ratio for gasification experiments at three different temperatures (

950°C, □ 850°C, ∆ 750°C).

Experiments run with a partial substitution of the German lignite by olive bagasse showed this as a way to increase the carbon conversion without negatively affecting the gasification performance. For example, when the gasifier is operated at 850°C with GL the carbon conversion is 75%. In order to increase the carbon conversion either temperature, the CO2/C ratio or the steam concentration would have to be increased. All this would significantly reduce the efficiency of the gasifier. However replacing 30% of the fuel with OB would increase the carbon conversion to nearly 100%, while maintaining the gasification efficiency. A synergistic effect between coal and biomass was observed. In addition, co-gasification is a potential pathway to utilise waste for power generation and to reduce the CO2 emissions of the plant.

4) Effect of Pressure Experiments were also performed at pressures up to 20bar. Carbon conversion decreased significantly with increasing pressure mainly above 10 bar. These results may be a combination of a slight decrease in primary volatile release, as established in a separate experiment using a Wire Mesh Reactor, and char deactivation due condensation of tars on the char particles and the formation of secondary char. The increasing fuel feeding rate and the subsequently higher particle density in the injection region of the FBR were identified as main reason for the increased coverage of char particles with secondary char. Similar trends to those described for atmospheric pressure operation were obtained regarding variation in carbon conversion with temperature, CO2 to C ratio and steam addition.

Figure 8. Carbon Conversion as a function of pressures for gasification experiments at different conditions (□ 850⁰C- CO2,
850⁰C -O2/ CO2, ○ 850⁰C-steam/O2/CO2, ∆ 950⁰C-O2/CO2).

The fuel gas compositions are shown in Figure 9. With increasing pressure the concentration of CO2 increased (60 – 70%) while that of CO diminished (30 – 20%). At the same time less fuel-hydrogen was converted to H2 (43 – 30%) and more to H2O. The concentration

of

CH4

in

the

fuel

gas

increased

slightly

with

pressure.

Figure 9. Oxy-fuel gasification of GL at 850C with 3.7% O2/CO2 (O/C = 0.1) at different pressures: Fuel gas composition (□CO2,
CO, ∆ H2, × CH4).

CONCLUSIONS A laboratory scale fluidised bed gasifier was used to study the gasification performance of a German lignite under O2/CO2/steam–blown conditions. A high carbon conversion, only limited by loss of fines entrained in the fuel gas, was achieved when the gasifier was operated at 950°C. To achieve conversions closer to 100%, recirculation of fines would be necessary. Similar carbon conversions can be reached at 850°C if either the inlet CO2 to C in the coal ratio is increased or if a fraction of the inlet CO2 is replaced with steam. At the lower temperatures (750 and 850°C), conversion can also be enhanced by using a higher O2/CO2 ratio in the inlet. At these temperatures, addition of O2 slightly increases the heating value of the gases. However, increasing the O2 content of the inlet gases at 950°C produced a drop rather than an increase in carbon conversion due to agglomeration of bed material in the zone near the spout.

Using a higher flowrate of CO2 produces a decrease in LHV. The increasing addition of steam generates a linear increase in the H2 fraction in the fuel gas, even after the carbon

conversion has reached its maximum. This is accompanied by a decrease in CO as predicted from the water-gas shift reaction equilibrium.

Increasing pressure above 10 bars produced a drop in carbon conversion due to a drop in char reactivity. It is thought that this is due to deposition of tars on the char surface leading to the formation of secondary char, which is enhanced by the higher coal throughputs used at higher pressures. Overall, these results show oxy-fuel firing of a fluidised bed gasifier as a promising route to avoid N2 dilution of the fuel gas and enable integration of fluidised bed gasification with CCS technology.

REFERENCES [1] Spiegl, Sivena, Lorente, Paterson, and Millan, “Investigation of the Oxy-fuel Gasification of Coal in a Laboratory-Scale Spouted-Bed Reactor: Reactor Modifications and Initial Results”, Energy Fuels 2010, 24, 5281–5288.

[2] Kikuchi, Suzuki, Mochizuki, Endo, Imai, and Tanji, “Ash Agglomeration Gasification of Coal in a Spouted Bed Reactor”, Fuel 1985, 64, 368-372.

[3] Bartels, Lin, Nijenhuis, Kapteijn, Ommen, “Agglomeration in Fluidized Beds at High Temperatures: Mechanisms, Detection and Prevention”, Progress in Energy and Combustion Science, 2008, 34, 633-666.

THE 27TH ANNUAL INTERNATIONAL PITTSBURGH COAL CONFERENCE. ISTANBUL, 2010.

1

Co-Gasification of Footwear Leather Waste and High Ash Coal: A Thermodynamic Analysis Rodolfo Rodrigues∗ , Nilson R. Marcilio∗ , Jorge O. Trierweiler∗ , Marcelo Godinho† and Adriene M. S. Pereira‡ ∗ Federal University of Rio Grande do Sul (UFRGS), Department of Chemical Engineering. Porto Alegre (RS), Brazil. E-mail: [email protected], [email protected], [email protected] † University of Caxias do Sul (UCS), Department of Chemical Engineering. Caxias do Sul (RS), Brazil. E-mail: [email protected] ‡ Pontifical Catholic University of Rio Grande do Sul (PUC-RS), Department of Chemical Engineering. Porto Alegre (RS), Brazil. E-mail: [email protected]

Abstract—The leather and shoe industry is one of the sectors generates more wastes at same time that has a high polluting potential. In Brazil, the majority of these wastes are disposed of in landfills and less than 5% are recycled. It corresponds to 62.5% of about 190,000 tons per year of hazardous wastes have been generated there. Since 1997 the Laboratory of Residues Processing (LPR) at the UFRGS has been developing practical projects on thermal treatment of leather wastes (biomass) based on the combined gasification-combustion technology for cogeneration. This work evaluates the co-gasification of leather wastes together a fraction of local coal over the fuel gas (syngas). The coal is subbituminous and high ash content (nearly 50%). A thermodynamic equilibrium model is applied for analysis of the co-gasification process. A sensitivity analysis of biomass cogasification related to usual operational parameters (air, steam and blending ratios) is done. That allows to identify efficiencies of ∼64% for middle air demands (∼60% of stoichiometric demand) at any leather-coal blending. Whereas an efficiency of ∼75% should be reached with steam-to-carbon ratios higher than 1.5. From future studies the model can be used to evaluated the formation of N and S compounds in the flue gas. Index Terms—Co-gasification, footwear leather waste, high ash coal, equilibrium model.

I. I NTRODUCTION The leather and shoe industry is one of the sectors generates more wastes at same time that has a high polluting potential. This environmental problem is proven the danger of waste containing chromium, derived from the salt utilized to tan hides. In Brazil, the majority of these wastes are disposed of in landfills and less than 5% are recycled [1]. Southern region locates most of those industries since more than 50% of Brazilian leather is produced in the State of Rio Grande do Sul (RS). It corresponds to 62.5% of about 190 thousand tons per year of hazardous wastes have been generated there [2]. Studies from Godinho et al. [1] shown nearly 50% wastes generate by those industrial sectors are footwear leather residues. The products contain chromium source from tanning substances used in hide processing. The Brazilian regulation [3] classifies chromium compounds as hazardous components in waste materials due to carcinogenic affect under trivalent state — Cr(VI) [4]. These residues contain high quantities of volatile matter and low ash content, while their heating value

is high sufficiently to enables combustion under auto-thermal conditions [1]. Since 1997 the Laboratory of Residues Processing (LPR) at the UFRGS has been developing practical projects on thermal treatment of leather wastes. In 2003 a semi-pilot unit based on the combined gasification-combustion technology has started to operate and results to 350 kW of thermal generating. From new investments the thermal capacity of the unit increases to 600 kW that starts to operation on the second half of 2010. The unit will produce power from the heat generated in the combustion. With the gasification-combustion technology a reducing environment followed by an oxidative environment decreases the possibility of formation of critical components such as Cr(VI) and PCDD/F [1], [5]. Otherwise leather waste as biomass makes the gasification into attractive method of treatment in result a neutral carbon emission. The Brazilian coal mines are placed in the South and the power plants are installed near these mines [6]. Rio Grande do Sul has 89.25% of coal reserves from 7 billion tons total of Brazil [7], [8]. The coal is subbituminous at Candiota and Le˜ao-Buti´a, and high volatile bituminous at Santa Terezinha [9]. Besides that, all of them are high ash content (about 48.5%) typical of Brazilian coals. It is low nitrogen content (0.61%) and relativity high sulfur content (2.15%). And it is poor in carbon (25.1%) compared to worldwide coals (65.7– 85.8%) [6] that limits the thermal yield. As a biomass, the leather wastes are high nitrogen content (12.42%) and they have a sulfur composition (1.83%) around the sulfur content of coal. Its calorific value (18.45 MJ/kg) is 37.7% higher than coal calorific value (13.4 MJ/kg) which should be justified for a higher H/C ratio (17.3% against to 10.4% of coal). The proximate and ultimate analyses and the high heating value of the footwear leather wastes and a mean coal sample are given in Table I. In literature a few number of studies on the thermochemical processing of leather wastes are represent and all of them use leather alone [1], [10]–[14]. This work evaluate the possibility of co-gasifying leather wastes with a fraction of local coal to take advantage of combining the features of each feedstock alone. Low-grade coals are usually difficult to gasify [15], but

2

THE 27TH ANNUAL INTERNATIONAL PITTSBURGH COAL CONFERENCE. ISTANBUL, 2010.

TABLE I C HARACTERIZATION OF THE FEEDSTOCKS CONSIDERED IN THIS STUDY. Leather waste1

Coal2

Proximate analysis (%wt) Moisture 12.4 11.0 Volatile matter 67.7 20.3 Fixed carbon 14.8 25.1 Ash 5.1 42.7 Ultimate analysis (%wt, d.b.) C 49.31 33.39 H 8.52 3.47 O 24.70 16.68 N 12.42 0.61 S 1.83 2.15 Cl 0.45 800◦ C). That range also corresponds to maximum conversion of solid carbon (dotted lines), reaching values of 19.6% for leather waste and 29.5% for coal. A sensitivity analysis of biomass co-gasification related to usual operational parameters is done varying leather-coal blending. The operational parameters are the air supplied (25◦ C and 1 atm) and the steam supplied (200◦ C and 1

4

THE 27TH ANNUAL INTERNATIONAL PITTSBURGH COAL CONFERENCE. ISTANBUL, 2010.

atm) contents. That allows to identify the performance of the co-processing of biomass-coal blends regarding syngas heat values. The feed air flow rate is related to the feedstock by a parameter called equivalence ratio (φ, Eq. 3), which indicates the oxygen used relative to that required for complete combustion (stoichiometric oxygen). In a similar way, the steam flow rate is quantified by a steam-to-carbon ratio (stm, Eq. 4): φ= 

nO2 1+

b 4

−

stm =

c 2

+e−

f 2



(3) nfuel

nH2 O(vap) nC

(4)

The responses to parametric changes were observed in output parameters: higher heating value (HHV ) of the energetic gas product (CO, H2 , and CH4 ) (Eq. 5) and cold gas efficiency (ηcg ) (Eq. 6), according to the equilibrium model. These output parameters are important to quantify the performance of the process.

(a) Coal with no ash


◦ ◦ HHVproducts = nCO ∆Hf,CO − ∆Hf,CO + 2   ◦ ◦ ◦ +nCH4 ∆Hf,CH4 − ∆Hf,CO2 − 2∆Hf,H2 O(l) − ◦ −nH2 ∆Hf,H 2 O(l)

ηcg =

HHVproducts HHVfuel ·Mwfuel

(5) (6)

As first analysis of gasifying process it is presented a comparative between a coal sample with no ash fraction (Fig. 2a) and a full coal sample (Fig. 2b). Since 42.7%wt of coal is ash, the decreasing of efficiency could be seen. Both cases show the peak of maximum efficiency at point where the solid carbon fraction is fully consumed. It happens at φ = 0.67 for efficiency of 68.5% regards to coal with no ash and φ = 0.57 for efficiency of 66.0% to full coal. In absolute terms they correspond to 4.04 Nm3air /kgcoal and 2.52 Nm3air /kgcoal , respectively, that is, a higher ash amount means a higher air demand for approximately the same efficiency because of the higher initial carbon fraction. Figures 3 and 4 compare the main output parameters of gasifying process for both samples. Leather waste (dashed lines) has been a higher efficiency (69.15%) for a lower temperature (1,472◦ C) and air-to-fuel ratio (φ = 0.54). These values are close to those simulated in previous study (φ = 0.34–0.53) using a stoichiometric approach [24]. Coal (dotted lines), in its turn, reaches the maximum efficiency of 66% at 1,537◦ C (φ = 0.57). Notwithstanding those similar temperature profiles (Fig. 3), coal (dotted lines) and leather (dashed lines) reach the peak by means of different absolute air demands, respectively, 5.68 Nm3air /kgleather and 4.40 Nm3air /kgcoal Last point of analysis suggests the study of co-gasifying process involving leather waste and coal. In Figure 5, the efficiencies are calculated to different operational conditions. Figs. 5a to 5f show a set of inlet steam amount (steam-tocarbon ratio) covering a range of air-to-fuel ratio (φ = 0.4–1.0). Here it should be noted that each equivalence ratio is related

(b) Coal with ash Fig. 2. Gasification of coal at 1 atm considering (a) with no ash content and (b) with ash content, where dotted lines represent the cold gas efficiencies.

to the specific blend composition. The evaluation of estimated data suggests that a minimum efficiency of ∼64% should be attained with φ = 0.58–0.63 at any leather-coal blending. However, an increasing of efficiency in +6% (70% total) can be taken with a 57.5%–32.5%wt leather-coal blend (or richer in leather content) and lower air-to-fuel ratio 2.0 and lower air demands 30 wt % (db), contains a higher proportion of inorganic material than the Mavropigi deposit (ash values 97

50– 0.9– 1823 65.90 22.90 1.08 4.58 0.79 0.47 1.37 0.62 0.31 1.37 0.07 0.03

Coal B 27.00 6.46 34.60 58.94 1.70 80.94 4.75 1.03 0.17 13.11 1433 1473 1493 1373 1453 1473 28.85 17.53 24.97 8.79 1.22 0.59 0.22 12.38 0.17 0.56 0.01 0.25

Coal C 29.94 8.53 38.63 52.84 1.37 78.61 5.27 1.42 0.16 14.54 1393 1423 1453 1393 1423 1453 53.04 10.03 14.54 5.25 2.49 2.15 0.78 6.96 0.11 0.56 500 MWth), as well as to express the emission in corresponding unit: [mg/Nm3]. A reasonable degree of prediction ability was demonstrated. However, it should always bear in mind that it is not simple to trace the origins of eventual predictive difficulties, because the defects in the NOx post-processor cannot be separated easily from those of the parent code for the flow and combustion. Table 3. Comparison between predicted and measured NOx emission Numerical predictions of NOx emission from the utility boiler furnace [mg/Nm3] Power plant Kostolac B-1 Test-case 1 1037

Test-case 3 1148

NOx emission [mg/Nm3] Power Plant Kostolac – * Limit values periodical measurements

Power plant Kostolac B-2

B-1

B-2

Serbian regulations

European Union Directive 2001/80/EC

Test-case 14 931

1051

893

450

500 (**200)

*

Emission limit values for NOx (measured as NO2, dry basis, corrected to 6% O2 in flue gases, for solid fuels and the boiler facility power output > 500 MWth) [23] ** From 2016 on

On the other hand, it is clear that the NOx emissions detected at the boiler units are fairly above the emission limits, especialy those which will become valid in the near future (from 2016 on). Thus it is necessary to reduce the emission from the analyzed boiler unit in the forthcoming period, by means of different measures, in the first place primary ones. In doing this, numerical simulations can help a great deal in examination, evaluation and selection of alternative solutions, with respect to their applicability and efficiency. 4.2.2. Numerical results for the reference test-case Numerical predictions given in Figure 5, for the test-case 1, show the dependence of NO content in the furnace of the Kostolac B boiler units on the flue gases temperature and the concentration of reactants, in this case HCN and oxygen, that produce fuel NO by homogeneous reactions, Eq. (4). Reactions of NO reduction with HCN, Eq. (6), are also taking place. Since fuel NO is by far more important than thermal NO (in the temperature domain for the furnace considered), the dependence of fuel NO actually determines the character of the total NO concentration field. Figure 5 suggests significant influence of the concentration of HCN (as intermediate compound released by devolatilization) and thus also of the nitrogen content in the coal which is the source of HCN, on the NO content. A narow 13

zone of higher NO concentrations, placed in the region in which the major part of the fuel is introduced through the lowest-stage burners, corresponds to the maximum of the HCN content. Wider zone of high NO concentrations can be seen in downstream (upward) direction and is related to the region of intensive chemical reactions of HCN depletion and NO formation. In contrast to thermal NO, the content of fuel NO (and consequently also the total NO) is less affected by temperature. In addition to the nitrogen content in the fuel, it is strongly influenced by air to fuel ratio [46] (air excess), i.e. oxygen concentration. This can be clearly seen from the comparison between the concentrations, given in Figure 5. So, the NO concentration field does not follow only the temperature and the HCN concentration field, but even more the O2 concentration field. In spite of high temperatures, there is no high NO concentration in the central region of the furnace, because of the oxygen depletion by intensive reactions of the fuel combustion. In the model, this strong dependence of NO content on oxygen is described by the Eq. (5), that gives dependence of the coefficient α on the local oxygen concentration, where α is exponent of the mole fraction XO2 in Eq. (4) for fuel NO formation reaction rate.

Figure 5. Gas temperature field and O2 , HCN and NO concentration fields in the Kostolac B utility boiler unit furnace, test-case 1

Analysis of the results showed that the contribution of thermal NO was of the order of several percents of the total NO. The region in which thermal NO might have been imortant was restricted to the narrow zone in which it was produced, corresponding to the maximal local gas temperatures in the 14

furnace (T=1650 K-1800 K), where the presence of thermal NO could have been only expected, as already explained. Since in the cases considered, as clarified by more detailed analysis of numerical data, the local gas temperatures rarely exceeded this temperature limit, the contribution of thermal NO to the total NO was not so significant. 4.2.3. The effect of selected operating conditions on NOx emission: a numerical analysis 4.2.3.1. The effect of the air-coal dust mixture distribution over the burner tiers on NOx emission Developed mathematical model of NOx formation/destruction was applied to predict and analyze several test-cases, with respect to NOx emission from furnaces of the Kostolac B utility boiler units. In the test-cases, there were the same flow rates of the air-coal dust mixture and the combustion air, both in total and per individual burner, as for the reference test-case 1. They differed to each other with respect to the working parameters given in Tables 1-2. The main difference was the pulverized coal distribution over the burner tiers. Although both FEGT and NOx emission are given in Table 1 for a number of test-cases, only a few situations are presented in Figures 6-11, in order to provide a concise analysis. Figures 6-11 present numerical results for the field of local concentrations, i.e. mass fractions of NO in characteristic sections of the boiler furnace. In the test-cases 1 and 3 the centrifugal separators are used, while in the test-cases 11 and 14-16 the louver separators are applied. While the air-coal dust mixture distibutions for the test-cases 1 and 3 (Kostolac B-1) and 14 (Kostolac B-2) were obtained during the examinations of the coal mills in existing operating conditions, the test-case 11 was performed as an atempt to optimize this distribution numerically, with respect to the reduction of NOx content in the furnace flue gases. Figures 10-11 show numerical simulations for two cases with opened movable elements of the louvers (test-cases 15 and 16), in comparison with the situation when they are closed, test-case 14, Figure 9. As given in Table 1 and summarized in the titles of Figures 6-8 for the test-cases considered, increase of the fraction of the pulverized coal flow rate through the lower-stage burners provides a decrease of both FEGT and NOx emission. Similar can be concluded for the test-cases given in Figures 9-11. Figures 6-11 also suggest that higher FEGT and NOx emission correspond, in these cases, to the higher position of the flame (i.e. to the wider flame at the furnace exit) and vice versa. However, one should be very carefule in trying to draw some general conclusions. Working situations in furnaces are very complex and require comprehensive analysis. For example, fraction of the pulverized coal flow rate through the lower-stage burners (so called main burners) is the same for the test-cases 1 and 15, i.e. 70%, but FEGT and NOx emission are lower in the test-case 1, with considerably lower position of the flame, Figure 6, compared to Figure 10. This is probably due to the fact that in the test-case 1 considerably more fuel is supplied through the lower tier of both the main burners and the higher-stage burners, as given in Table 1. Moreover, there is lower NOx emission, slightly lower FEGT and lower position of the flame in the test-case 3 than in the test-case 16, Figures 7 and 11, although there is to some extent lower fraction of the pulverized coal flow rate through the lower-stage burners in the testcase 3. The explanation is most likely similar to the one given for the previous two cases. As demonstrated, the analysis has to account for the fuel distribution over each of the burner tiers: two lower-stage and two upper-stage burners. Anyhow, the numerical study is to be performed for each case individually, because there are a lot of mutually dependent parameters influencing the situation in the boiler furnace. This kind of comparative analysis of the gas temperature and the NOx concentration field, in conjunction with the values of FEGT and NOx emission from the furnace, offers the possibility of optimizing both the flame and the emission. Higher position of the flame causes the higher values of FEGT, which suggests also the lower efficiency coefficient of the boiler. For the test-cases presented in Figures 6-11, the lowest emission achieved by numerical optimization (in the test-case 11) was approximately 32% lower that the highest one (in the test-case 16). This is not to be neglected, especially because in this example the emission reduction is attained simply by using an appropriate combustion organization within the furnace considered, with no need for construction changes in the furnace and the boiler.

15

Figure 6. NO content in the furnace and gas temperature field for the test-case 1 (fraction of the pulverized coal flow rate through the lower-stage burners 70%, FEGT=1021 OC, NOx emission=1037 mg/Nm3)

Figure 7. NOx content in the furnace and gas temperatur field for the test-case 3 (fraction of the pulverized coal flow rate through the lower-stage burners 56.2%, FEGT=1096 OC, NOx emission= 1148 mg/Nm3)

Figure 8. NOx content in the furnace and gas temperature field for the test-case 11 (fraction of the pulverized coal flow rate through the lower-stage burners 90%, FEGT=997 OC, NOx emission=841 mg/Nm3)

16

Figure 9. NOx content in the furnace and gas temperature field for the test-case 14 (fraction of the pulverized coal flow rate through the lower-stage burners 76%, FEGT=1047 OC, NOx emission=931 mg/Nm3)

Figure 10. NOx content in the furnace and gas temperature field for the test-case 15 (fraction of the pulverized coal flow rate through the lower-stage burners 70%, FEGT=1075 OC, NOx emission= 1118 mg/Nm3)

Figure 11. NOx content in the furnace and gas temperature field for the test-case 16 (fraction of the pulverized coal flow rate through the lower-stage burners 64%, FEGT=1100 OC, NOx emission= 1244 mg/Nm3)

17

4.2.3.2. The effects of the fuel and air distribution over the burners, the cold air ingress and the combustion air distribution over the burner tiers on NOx emission The test-case 17 examines what happens with the emission when there is an uneven distribution of the fuel and the combustion air-secondary air over the individual burners, used in order to centralize the flame within the furnace cross section, Figure 4. It is important that this, aerodynamically and thermally favorable-central postition of the flame, does not cause the emission increase (but the opposite, see Table 1), in the case considered. In the test-case 18, compared with the test-case 6, with the rest of the operating conditions being the same, an excesive ingress of the cold air slightly reduces the emission (due to the additional cooling of the gases), but also causes an excesive reduction of FEGT. The other interesting situation is given in the test-case 19, which differs from the test-case 5 only by 20% decrease of the fraction of secondary air injected through the lower-stage burners, namely 60% instead of 74.2%, Table 1. Compared to the test-case 5, with already achieved favorable conditions (values of FEGT and the emission, as well as the flame positioned mainly in the central-burners region of the furnace), further decrease of both the FEGT and the emission are provided in the test-case 19. With these modifications of the combustion air distribution, one of the lowest values of the emission is achieved, Table 1, with the flame additionally descended, Figure 12.

Figure 12. NO content in the furnace and gas temperature field for the test-cases 5 and 19 (fractions of the secondary air flow rate through the lower-stage burners: 74.2% and 60%, FEGT=1020 OC and 991 OC, NOx emission= 854 mg/Nm3 and 808 mg/Nm3)

Finally, the test-cases 20 and 21 demonstrate also the influence of different distribution of secondary air over the burner tiers, for the same distribution of coal. Here, however, all air-coal dust mixture is introduced through the lower-stage burners, while the upper-stage burners are used only for injecting the preheated air. Injection of 20% less secondary air through the lower-stage burners decreases FEGT and considerably decreases NOx emission. Although the contribution of thermal NO (strongly dependent on the gas temperature) is not so significant in the cases considered, it may still be supposed 18

that the part of the explanation for the obtained trend may be that less combustion air in the lower-stage burners region will delay, to some extent, the NO formation reactions to the upper regions of lower temperatures. Nevertheless, it seems that the emission reduction obtained by the corresponding modifications of the combustion air distribution over the burner tiers in the test-cases 19-21 could be explained mainly by a proper tuning of the local air excess in the burners region. Performed numerical analyses demonstrate the possibility of applying the developed predictive tool for NOx emission in finding the solutions and selecting between alternative combustion modifications and other primary measures that may be appled for reduction of the emission from pulverized coalfired boiler furnaces. This kind of analysis enables an efficient optimization of the operating conditions and the case-study furnace operation with respect to both the emission reduction and the possibility of keeping, or even increasing the energy efficiency of the facility. CONCLUSIONS A comprehensive numerical study was performed on the effects of different operation parameters on the pulverized coal flame position and NOx emission from tangentially-fired furnaces of 350 MWe utility boiler units. The predictions of flow, combustion and heat transfer were performed by means of previously developed CFD model and computer code. NOx emission predictions were done by recently developed NOx formation and destruction submodel and the post-processor, using simplified chemical models in conjunction with detailed CFD calculations. The coal and the combustion air distribution over the burner tiers, the cold air ingress and different operation regimes of individual burners were investigated and found to affect the flame and the emission considerably. It was possible to achieve favorable vertical position of the flame within the furnace by means of careful numerical optimization of the pulverized coal and the combustion air distribution over the burner tiers. Non-uniform distribution of coal and air over the individual burners offered an option to attain a central position of the flame within the furnace cross section, which might have affected the thermal load of the water-walls and the conditions for fouling and slag deposition. For prediction of thermal NO simplified Zeldovich expression was used, while the reactions of fuel NO formation and destruction were modeled as well. Fuel NO formation was considered through oxidation of HCN as the main intermediate compound from volatiles, taking into account strong influence of local oxygen content. For fuel NO formation reaction rate, correction of pre-exponential factor was adopted, according to relevant references, for predominantly fuel-lean conditions in pulverized coal-fired furnaces. The developed NOx post-processor was applied, along with a comprehensive model of pulverized-coal combustion, to predict NO emissions from the utility boiler furnaces. NOx submodel was preliminary validated by comparison between the predictions and the available results of measurements of the emission from power plant Kostolac B-1 and B-2 boiler units. The results suggested that the present NO model was capable of reproducing the different trends in NO formation/destruction, associated with the dissimilar operating conditions. Numerical results for the gas mixture temperature field and the O2, HCN and NO concentration fields are given as well, for the reference test-case. The numerical study was performed to achieve both NOx emission reduction and favorable position of the flame in the case-study furnace, by investigating the corresponding combustion modifications, without need for construction changes in the facility. The study showed that it was possible to achieve this task by careful optimization of the coal and the combustion air distribution over the burner tiers. In general, both FEGT and NOx emission were decreased and the flame was descended when introduced larger fractions of pulverized coal flow rate through the lower-stage burners and/or the lower stages of both the main burners and the upper-stage burners. The emission reduction could be obtained also by a proper tuning of the local air excess in the burners region. However, the complex situation in the boiler furnace requires accounting for the fuel and air distribution over each of the burners and the burner tiers and performing the numerical study for each case individually. Numerical predictions enable impact-analysis of many operation parameters, optimization of combustion, heat transfer, flame and emission, thus solving critical operation problems, like excessively high or low flame position, or high emission. This kind of numerical analysis enables optimization of the operating conditions in the boiler furnaces both with respect to the emission reduction and the energy efficiency improvements.

19

ACKNOWLEDGMENTS This work has been supported by the Republic of Serbia Ministry of Science and Technological Development (project: TR-18007) and the Electric Power Industry of Serbia. REFERENCES [1] L. M. R. Coelho, J. L. T. Azevedo, M. G. Carvalho, Application of a global NOx formation model to a pulverized coal fired boiler with gas reburning, Proceedings, Fourth International Conference on Technologies and Combustion for a Clean Environment, Lisbon, 7-10 July 1997, Paper 9.4, pp. 1/8-8/8. [2] J. Fan, P. Sun, X. Zha, K. Cen, Modeling of combustion process in 600 MW utility boiler using comprehensive models and its experimental validation, Energy & Fuels 13 (1999), pp. 1051-1057. [3] M. Xu, J. L. T. Azevedo, M. G. Carvalho, Modelling of the combustion process and NOx emission in a utility boiler, Fuel 79 (2000), pp. 1611-1619. [4] Ch. Zheng, Zh. Liu, X. Duan, J. Mi, Numerical and experimental investigation on the performance of a 300 MW pulverized coal furnace, Proceedings of the Combustion Institute 29 (2002), pp. 811-818. [5] R. He, T. Suda, M. Takafuji, T. Hirata, J. Sato, Analysis of low NO emission in high temperature air combustion for pulverized coal, Fuel 83 (2004), pp. 1133–1141. [6] K. Li, S. Thompson, J. Peng, Modelling and prediction of NOx emission in a coal-fired power generation plant, Control Engineering Practice 12 (2004), pp. 707–723. [7] J. Pallares, I. Arauzo, L. I. Diez, Numerical prediction of unburned carbon levels in large pulverized coal utility boilers, Fuel 84 (2005), pp. 2364-2371. [8] R. Vuthaluru, H. B. Vuthaluru, Modelling of a wall fired furnace for different operating conditions using FLUENT, Fuel Processing Technology 87 (2006), pp. 633-639. [9] A. Bosoaga, N. Panoiu, L. Mihaescu, L. I. Backreedy, L. Ma, M. Pourkashanian, A. Williams, The combustion of pulverized low grade lignite, Fuel 85 (2006), pp. 1591-1598. [10] S. Belosevic, M. Sijercic, D. Tucakovic, Three-dimensional modeling of utility boiler pulverized coal tangentially fired furnace, International Journal.of Heat and Mass Transfer 49 (2006), pp. 3371-3378. [11] E. I. Karpenko, V. E. Messerle, A. B. Ustimenko, Plasma-aided solid fuel combustion, Proceedings of the Combustion Institute 31 (2007), pp. 3353–3360. [12] M. Kumar, S. G. Sahu, Study on the effect of the operating condition on a pulverized coal-fired furnace using computational fluid dynamics commercial code, Energy & Fuels 21 (2007), pp. 3189-3193. [13] N. Hashimoto, R. Kurose, H. Tsuji, H. Shirai, A numerical analysis of pulverized coal combustion in a multiburner furnace, Energy & Fuels 21 (2007), pp. 1950-1958. [14] L. I. Dıez, C. Cortes, J. Pallares, Numerical investigation of NOx emissions from a tangentially-fired utility boiler under conventional and overfire air operation, Fuel 87 (2008), pp. 1259–1269. [15] J. Makovička, Mathematical model of pulverized coal combustion, Dissertation thesis, Czech Technical University in Prague, Faculty of Nuclear Sciences and Physical Engineering, Prague, Czech Republic, 2008. [16] S. Belosevic, M. Sijercic, D. Tucakovic, N. Crnomarkovic, A numerical study of a utility boiler tangentially-fired furnace under different operating conditions, Fuel 87 (2008), pp. 3331-3338. [17] S. Belosevic, M. Sijercic, P. Stefanovic, A numerical study of pulverized coal ignition by means of plasma torches in air-coal dust mixture ducts of utility boiler furnaces, International Journal.of Heat and Mass Transfer 51 (2008), pp. 1970-1978. [18] S. Belosevic, M. Sijercic, N. Crnomarkovic, B. Stankovic, D. Tucakovic, Numerical prediction of pulverized coal flame in utility boiler furnaces, Energy & Fuels 23 (2009), pp. 5401-5412. [19] R. Straka, Mathematical model of pulverized coal fired boiler, Dissertation thesis, Czech Technical University in Prague, Faculty of Nuclear Sciences and Physical Engineering, Prague, Czech Republic, 2009 (through private communication). [20] E. H. Chui, H. Gao, Estimation of NOx emissions from coal-fired utility boilers, Fuel 89 (2010), pp. 29772984. [21] H-Ch. Zhou, Ch. Lou, Q. Cheng, Zh. Jiang, J. He, B. Huang, Zh. Pei, Ch. Lu, Experimental investigations

20

on visualization of three-dimensional temperature distributions in a large-scale pulverized-coal-fired boiler furnace, Proceedings of the Combustion Institute 30 (2005), pp. 1699-1706. [22] D. Tucakovic, T. Zivanovic, V. Stevanovic, S. Belosevic, R. Galic, A computer code for the prediction of mill gases and hot air distribution between burners’ sections at the utility boiler, Applied Thermal Engineering 28 (2008), pp. 2178-2186. [23] M. Gavrić, A. Vlajčić, B. Čeperković, Green Book of the Electric Power Industry of Serbia, Belgrade, June 2009, Published by: Electric Power Industry of Serbia, Sector for Public Relations, Belgrade, http://www.eps.rs/publikacije/ZK/EPS – Zelena knjiga.pdf [Accessed 17.02.2010] (in Serbian). [24] C. P. Fenimore, Formation of nitric oxide in premixed hydrocarbon flames, Thirteenth Symposium (International) on Combustion, The Combustion Institute, 1971, pp. 373-380. [25] D. Iverach, K. S. Basden, N. Y. Kirov, Formation of nitric oxide in fuel-lean and fuel-rich flames, Fourteenth Symposium (International) on Combustion, The Combustion Institute, 1973, pp. 767-775. [26] V. Quan, F. E. Marble, J. R. Kliegel, Nitric oxide formation in turbulent diffusion flames, Fourteenth Symposium (International) on Combustion, The Combustion Institute, 1973, pp. 851-860. [27] F. V. Bracco, Nitric oxide in droplet diffusion flames, Fourteenth Symposium (International) on Combustion, The Combustion Institute, 1973, pp. 831-842. [28] C. T. Bowman, Kinetics of nitric oxide formation in combustion processes, Fourteenth Symposium (International) on Combustion, The Combustion Institute, 1973, pp. 729-738. [29] C. T. Bowman, Kinetics of pollutant formation and destruction in combustion, Progress in Energy and Combustion Science 1 (1975), pp. 33-45. [30] G. G. De Soete, Overall reaction rates of NO and N2 formation from fuel nitrogen, Fifteenth Symposium (International) on Combustion, The Combustion Institute, 1975, pp. 1093-1102. [31] J. M. Levy, L. K. Chan, A. F. Sarofim, J. M. Beer, NO/char reactions at pulverized coal flame conditions, Eighteenth Symposium (International) on Combustion, The Combustion Institute, 1981, pp. 111-120. [32] P. R. Solomon, M. B. Colket, Evolution of fuel nitrogen in coal devolatilisation, Fuel 57 (1978), pp. 749755. [33] J. A. Miller, C. T. Bowman, Mechanism and modelling of nitrogen chemistry in combustion, Progress in Energy and Combustion Science 15 (1989), pp. 287-338. [34] G. G. De Soete, Heterogeneous N2O and NO formation from bound nitrogen atoms during coal char combustion, Twenty-Third Symposium (International) on Combustion, The Combustion Institute, 1990, pp. 1257-1264. [35] I. Aarna, E. M. Suuberg, A review of the kinetics of the nitric oxide-carbon reaction, Fuel 76 (1997), pp. 475-491. [36] J. W. Mitchell, J. M. Tarbell, A kinetic model of nitric oxide formation during pulverized coal combustion, AIChE Journal 28 (1982), pp. 302-311. [37] E. E. Khalil, Modelling of Furnaces and Combustors, Abacus Press, UK, 1982. [38] P. J. Smith, S. C. Hill, L.D. Smoot, Theory for NO formation in turbulent coal flames, Nineteenth Symposium (International) on Combustion, The Combustion Institute, 1982, pp. 1263-1270. [39] R. D Boardman, L. D. Smoot, Prediction of nitric oxide in advanced combustion systems, AIChE Journal 34 (1988), pp. 1573-1576. [40] F. C. Lockwood, C. A. Romo-Millares, Mathematical modeling of fuel NO emissions from PF burners, Journal of the Institute of Energy 65 (1992), pp. 144-152. [41] J. M. Jones, P. M. Patterson, M. Pourkashanian, A. Williams, Approaches to modelling heterogeneous char NO formation/reduction during pulverised coal combustion, Carbon 37 (1999), pp. 1545-1552. [42] M. Xu, Y. Fan, J. Yuan, C. Sheng, H. Yao, A simplified fuel-NOx model based on regression analysis, International Journal of Energy Research 23 (1999), pp. 157-168. [43] L. J. Muzio, G. C. Quartucy, Implementing NO control: research to application, Progress in Energy and Combustion Science 23 (1997), pp. 233-266. [44] S. Li, T. Xu, Sh. Hui, X. Wei, NOx emission and thermal efficiency of a 300 MWe utility boiler retrofitted by air staging, Applied Energy 86 (2009), pp. 1797–1803.

21

[45] Sh. Su, J. Xiang, L. Sun, S. Hu, Zh. Zhang, J. Zhu, Application of gaseous fuel reburning for controlling nitric oxide emissions in boilers, Fuel Processing Technology 90 (2009), pp. 396–402. [46] B. G. Miller, Sh. F. Miller, J. L. Morrison, A. W. Scaroni, Cofiring coal-water slurry fuel with pulverized coal as a NOx reduction strategy, http://www.energy.psu.edu/factssheets/55_Cofiring_CWSF.pdf [Accessed: 25.02.2010]. [47] A. M. Eaton, L. D. Smoot, S. C. Hill, C. N. Eatough, Components, formulations, solutions evaluation, and application of comprehensive combustion models, Progress in Energy and Combustion Science 25 (1999), pp. 387-436. [48] S. C. Hill, L. D. Smoot, Modeling of nitrogen oxides formation and destruction in combustion systems, Progress in Energy and Combustion Science 26 (2000), pp. 417-458. [49] T. V. Vilenskij, D. M. Hemaljan, Dynamics of Pulverized Fuel Combustion, Energija, Moscow, 1978 (in Russian). [50] A. V. Saljnikov, B. S. Repic, P. T. Radulovic, L. L. Jovanovic, Combustion kinetics of coal dust, Journal of Engineering Physics and Thermophysics 68 (1995), pp. 225-229. [51] B. Perković, D. Adamović, V. Joksimović, M. Erić, Adjustment and optimization of the operation of power plant Kostolac-B boiler unit 1, after overhaul and reconstructions done in 2002/2003, Technical Report NIVITE 297, Institute Vinca and Power Plant Nikola Tesla, 2005 (in Serbian). [52] P. Radovanović, B. Perković, D. Adamović, Thermotechnical investigations and operation analysis of the boiler units B1 and B2 in power plant Kostolac before rapairs done in 2007 and 2008, Technical Report NIVLTE 373, Institute Vinca and Power Plant Nikola Tesla, 2008 (in Serbian). [53] P. Radovanović, B. Perković, D. Adamović, Thermotechnical investigations of the boiler units B1 and B2 in power plant Kostolac after rapairs done in 2008, Technical Report NIV-LTE 409, Institute Vinca and Power Plant Nikola Tesla, 2009 (in Serbian).

22

OBSERVATION OF HEAT RELEASE REGION AS FUNCTIONS OF COAL PROPERTIES IN TURBULENT JET PULVERIZED COAL FLAMES Yon-Mo Sung, Cheor-Eon Moon, Seong-Yool Ahn, Jae-Woo An School of Mechanical Engineering, Pusan National University, 30 Jangjeon-dong, Gumjeong-ku, Busan, 609-735, Republic of Korea Gyung-Min Choi*, Duck-Jool Kim Pusan Clean Center, Pusan National University, 30 Jangjeon-dong, Gumjeong-ku, Busan, 609-735, Republic of Korea *

Corresponding Author: [email protected] Telephone: +82-51-510-2476 FAX: +82-51-512-5236

Presented at The 27th Annual International Pittsburgh Coal Conference Istanbul, Turkey October 11-14, 2010 ABSTRACT One issue of interest is to develop diagnostic methods for the monitoring and control of the pulverized coal flames in power plants. The purpose of this study is to establish visualization and diagnostic methods in the pulverized coal combustion fields. An advanced instrumentation and research methodology was employed to observe the structure of pulverized coal flame in a laboratory scale burner. The effects of pulverized coal properties, volatile matter, particle size and moisture content, on the heat release region in turbulent jet pulverized coal flames were investigated experimentally. To understand the accuracy of line of sight measurement in the two-dimensional (2-D) visualization, point measurements of chemiluminescence intensity by Cassegrain optics were also conducted. The heat release region for the structure of pulverized coal flame was observed through visualization by CH* chemiluminescence image with an intensified high-speed camera, and by CH* chemiluminescence intensity for local point measurements. The streamwise length of the heat release region based on 2-D visualizations was about 11.4% longer than that of point measurements and increased proportionally to the volatile matter content. The temperature rise for 35~45 µm coal particles was faster than that for 75~90 µm particles, which resulted in a shift of reaction region toward upstream direction. The -1-

coal moisture content less than 15%, however, had little effect on the structure of the pulverized coal flame. The obtained results give us useful information for evaluating practical pulverized coal flames.

Keywords Pulverized coal combustion, Heat release region, Chemiluminescence intensity, Twodimensional visualization, Point measurement, Volatile matter, Particle size, Water contents

Introduction The combustion of pulverized coal is a complex process involving conductive, convective, and radiative heat transfer; turbulent fluid motion; coal particle devolatilization; volatile reaction; char reaction; particle dispersion; ash formation; and other processes, interactively and simultaneously [1]. Thus, the fundamental mechanism of pulverized coal combustion and the structure of pulverized coal flame have not fully understood. To obtain optimal conditions of pulverized coal combustion, the major processes such as ignition, drying, devolatilization, and volatile and char reactions of pulverized coal particles must be understood. All these processes depend on burning condition and coal properties [2]. Coal properties and combusting condition include volatile matter content, particle size, moisture content, coal rank, gas temperature, system pressure and residence time [3-9]. Many studies have been conducted to understand the detail characteristics of pulverized coal combustion [10] such as the structure of pulverized coal flame, pulverized coal particle ignition [11-13], devolatilization [14,15], and volatile and char combustion [16]. The yield of volatile is likely to be fairly high in actual pulverized coal flame, and the volatile reaction is also of great importance for the ignition of coal particles [17]. The interactions of volatile flame and char combustion of coal particles were investigated experimentally and numerically [17]. A surrounding mantle of volatile products was observed during the early stages of combustion when bituminous coal particle are more sensitive to the volatile yield than to the kinetics of devolatilization, and relatively insensitive to the kinetics of char combustion for mediumvolatile coal flame [18]. The phenomena of coal combustion can be described by considering three major regions: a particle heatup region, a devolatilization region, and a soot growth region [19]. There is an extensive review for diagnostic techniques for the monitoring and control of practical flames [20]. Burning coal may give luminous (carbon) flames, and at times the base of such flames may show C2 and CH emission [21]. The OH* radical seems more suitable for lean -2-

flames, which CH* and C2* have a more monotonic behavior and stronger dynamics for richer flames [22]. Thus, luminosity flame structure is observed through CH * chemiluminescence intensity and is considered to be a good representative of the heat release rate [23]. CH * chemiluminescence signals also provide a better indicator for the onset of pulverized coal particle ignition and the volatile reaction region than a direct photograph of the pulverized coal flame [24]. The measurement of CH* chemiluminescence from ensemble-averaging of ICCD image data can reveal the underlying trends in particle ignition and devolatilization [25]. Because the line of sight method by high speed camera might cause measurement error due to superposition, the verification of diagnostic methods is needed to obtain reliable information from the pulverized coal combustion. In this study, to estimate the availability of 2-D visualization method using an intensified high speed camera, pulverized coal jet flames were visualized and the results were compared with those of point measurements with a multi-color integrated cassegrain receiving optics (MICRO) system [26] of chemiluminescence intensity. In addition, the relationship between pulverized coal properties and the structure of pulverized coal flame was investigated. To investigate the effects of pulverized coal properties on the heat release region in turbulent jet pulverized coal flame, various pulverized coal properties including five different types of pulverized coal, two particle sizes of bituminous coal, and three values of moisture contents for sub-bituminous coal were taken into consideration.

Fig. 1. Schematic of pulverized coal burner and supply system.

Experimental Pulverized coal combustion burner Figure 1 shows a schematic diagram of a pulverized coal combustion burner and supplying system. The burner has a coaxial dual piping structure with a main burner port (inner diameter: 6 mm) and an annular slit burner (width: 0.5 mm) installed outside of the main burner port. In -3-

this study, the feed rate of pulverized coal particles were consistently regulated and maintained by a screw-type coal feeder (FEEDCON-µM, Nisshin engineering) and coal particles were carried by the main air flow. Methane was supplied from the slit burner port outside the main burner port in the dual pipe structure to stabilize the two-phase jet flame. The air and methane flow rates were controlled by a mass flow controller (KOFLOC-3660, Kojima instruments). Table 1 shows the proximate analysis, ultimate analysis, and heating values for five different types of coal (NCA, ECM, Vitol, PCC, and MHU). The pulverized coal feed rate, air flow rate and CH4 flow rate was 1.04 × 10-4 kg/s, 5.00 × 10-4 m3/s and 5.00 × 10-5 m3/s, respectively. Table 1 Properties of five coals used in this study

Coals a

M

Proximate analysis

Ultimate analysis

(wt. %, air-dry)

(wt. %, dry)

b

VM

c

e

CV

d

FR

FC

Ash

C

H

O

N

S

Ash

(kcal/kg, air-dry)

NCA

2.88

24.77

57

15.35

72.62

3.99

5.69

1.55

0.35

15.8

2.30

6907

ECM

2.58

32.51

49.86

15.05

68.9

4.54

8.6

1.65

0.87

15.44

1.53

6977

Vitol

3.76

32.87

50.29

13.08

70.68

4.49

9.22

1.46

0.56

13.59

1.53

6954

PCC

15.42

37.83

41.83

4.92

70.41

5.38

16.02

1.39

0.99

5.81

1.11

6918

MHU

14.5

40

40.1

5.4

71.4

4.97

14.78

1.47

1.06

6.32

1.00

6930

a

M: Moisture; bVM: Volatile Matter; cFC: Fixed Carbon; dFR: Fuel Ratio; eCV: Calorific Value

Fig. 2. Schematic of simultaneous optical measurement system. Simultaneous optical measurement system Figure 2 shows a schematic diagram of the simultaneous optical measurement system. To visualize the pulverized coal flame and to measure mean particle temperature, we used a high-4-

speed camera (FASTCAM-1024PCI, Photron) coupled with an image intensifier (UVi Camera Intensifier, Invisible Vision) and two-color radiation pyrometer (METIS-MQ11, Sensortherm), respectively. An image intensifier and a 431 nm filter (CH* chemiluminescence) were used in conjunction with a high-speed camera to obtain high magnification images of the pulverized coal flame.

In order to estimate the availability of 2-D visualization method, we measured

chemiluminescence intensity in pulverized coal flame along streamwise axis using a MICRO system [27,28]. Because this cassegrain optics has quite high spatial resolution, superposition due to line of sight can be eliminated effectively. The collected emission was sent to the photomultiplier spectroscopic unit which consists of an optical band-pass filter (center wavelength, FWHM, transmitting efficiency) and a photomultiplier (R106UH, Hamamatsu) for detecting the CH* (431.4 nm, 1.5 nm, 49.2%) emission intensity. The detailed optical system for local point measurement is described in [26-29]. The analog current signals from the photomultipliers were converted into a voltage signal by an I/V converter, and digitized by an A/D converter (NI-DAQ, National Instruments) at a sampling rate of 5 kHz. The two-color radiation pyrometer [30], which measure temperature from the ratio of radiation signals of two adjacent wavelengths (0.7 to 1.1 µm), was used to measure the temperature of the pulverized coal particles. The particle temperature was measured at the center of the flame in the axial direction in interval of 5 to 10 mm. The analog current signal on the basis of the ratio of the radiation intensity of two adjacent wavelengths were converted into a voltage signal by an I/V converter, and digitized by an analog-to-digital converter at a sampling rate of 5 kHz. The methane flow rate was the minimum amount required to form a stable pulverized coal flame.

Results and Discussion Effect of coal type (volatile matter) The pulverized coal combustion phenomena proceed in the following order. First, pulverized coal is preheated from the burner port and the moisture is dried. Then its volatile matter is emitted and ignited simultaneously. Both volatile matter and fixed carbon burn after reaching the maximum temperature of the particles. Finally, the temperature is rapidly increased because of radiative and conductive heat transfer and then the coal transforms to ash through complete combustion [31]. To investigate the effects of pulverized coal properties on the heat release region in a turbulent jet pulverized coal flame, five different types of coal were selected as shown in Table 1. In this experiment, because turbulent pulverized coal flame was opened to atmosphere, char reaction could not maintain in the downstream. Visualized emission region, therefore, indicates from ignition region to end of volatile matter reaction including partial char reaction. -5-

Figure 3 shows CH* chemiluminescence intensity by the MICRO system and that by intensified high-speed camera measured at the center of the flame in the axial direction. The emission intensity distribution of CH* chemiluminescence can be represented by a ninth-order polynomial equation. From the intersection of the straight line (dotted thick line) with the baseline, the onset of ignition (assumed to correspond to the onset of devolatilization) and the end of volatile reaction can be estimated [24]. The onset of heat release region, defined as the onset of the CH* signal, occurs around 20 mm downstream from the burner rim. The length of heat release region by intensified high-speed camera was about 30 mm (13%) longer than that by MICRO system. This overestimation by the high-speed camera is ascribed to a superposition of emission along the line-of-sight. MICRO system was lifted to keep horizontal balance, but camera was kept at same position.

Fig. 3. Averaged CH* chemiluminescence image and its intensity profile including polynomial fitting based on 2-D measurements by intensified high-speed camera, and point measurements by MICRO system for PCC coal flame. Top (no filtered) and bottom (431 nm filtered) flame images obtained from high-speed (average of 500 images) camera. Figure 4 shows history of heat release region and instantaneous images obtained for PCC coal flame. The onset position and the end of heat release region fluctuated temporally due to complexity of coal ignition process and turbulence.

The averaged onset position of heat

release region, the length of heat release region and the end of heat release region were 38 mm, 234 mm and 272 mm, respectively, though the width of heat release region fluctuated spatially. Because a methane pilot flame was adopted for flame stabilizing, the fluctuation of onset of heat release region was relatively smaller than that of end of heat release region. Because the heat release region of pulverized coal flame fluctuated significantly even for constant mass flow rate, -6-

a multi-dimensional time-series measurement, such as high speed camera, can give us useful information.

Fig. 4. History of heat release region and instantaneous photographs for PCC coal flame. Figure 5 shows estimated lengths of heat release region by an intensified high-speed camera and the MICRO system for five different types of coal in Table 1. The length of heat release region measured by MICRO system (point measurements) increased slightly with increasing volatile matter. The results from 2-D measurements (intensified high-speed camera) showed similar tendency with those from point measurements though averaged discrepancy between two measurement methods was about 11.4%. The discrepancy between point measurement and 2-D measurement increased in proportional to the length of heat release region. Note that some compensation is needed in order to estimate the heat release region using high speed camera.

Fig. 5. Estimated length of heat release region according to volatile matter content (%). -7-

Effect of coal particle size To clarify the effects of coal particle size on the heat release region, experiments with coal particle sizes of 35~45 µm and 75~90 µm were conducted. The coal (Vitol in Table 1) samples from the pulverizer mills in thermal power stations were size classified using standard sieves (325 to 450 mesh and 170 to 200 mesh). Figure 6 shows direct photographs of pulverized coal flames and the effects of coal particle size on the heat release region based on the MICRO system. The length of heat release region of averaged 500 images for 35~45 µm and 75~90 µm were 220 mm and 270 mm, respectively. For the smaller particle size, because the total particle reaction surface area increased and chemical reaction and heat transfer rates also increased, the flame propagation velocity increased definitely. The length of the heat release region for the smaller particle size was reduced 17 mm (8%) compared to the larger one. CH* chemiluminescence intensity and flame luminosity of smaller particle coal flame were stronger at the upstream of the flame compared with those of larger particle one. This strong intensity of flame emission resulted from enhanced chemical reaction rate.

Fig. 6. Examples of CH chemiluminescence intensity and polynomial fit for Vitol coal flame 75~90 µm particle diameter, 35~45 µm particle diameter. Top (75~90 µm) and bottom (35~45 µm) flame images (no filtered) obtained from high-speed (average of 500 images) camera. Figure 7 shows the mean particle temperature with different particle sizes on the central axis of the pulverized coal flame. The temperature rise of the 35~45 µm particle size was faster than the 75~90 µm particle size, and the peak temperature was 35 K (3%) higher. These results indicate that mixing of the volatiles with air and volatile reactions at the particle surfaces are more active in the case of the smaller particle size.

-8-

Fig. 7. Size effects on mean particle temperature profiles for Vitol coal flame. Effect of coal moisture content The coal drying process can improve combustion performance such as an increased flame temperature and low unburned carbon compared to that of raw coals [5,32,33]. To investigate the effect of coal moisture content on the heat release region, experiments for three cases (1.52, 9.49 and 15.42%) of PCC coal were conducted. The PCC coal was characterized as subbituminous with an inherent moisture content of 15.42% in Table 1. The distribution of CH* chemiluminescence intensity measured by the MICRO system at the center of the axial direction of the flame with coal moisture content is presented in Figure 8. The length of heat release region for the three cases was almost identical regardless of moisture content of coal. The discrepancy between the three cases was only 3 mm (1.1%).

Fig. 8. Moisture effects on heat release region for PCC coal flame. Figure 9 shows profiles of mean particle temperature at the center in the axial direction of the flame as a function of coal moisture content. The averaged particle temperatures for each case -9-

were 1401 K, 1422 K, and 1444 K, respectively. The moisture content had little effect on the length of the heat release region and the mean particle temperature compared to volatile matter and coal size when it is less than 15%. That is, coal drying process may be effective up to 15%.

Fig. 9. Moisture effects on mean particle temperature for PCC coal flame.

Conclusions The effects of pulverized coal properties on the heat release region were investigated in turbulent jet pulverized coal flame. Five types of pulverized coal (two particle sizes for a bituminous coal and three moisture contents for a sub-bituminous coal) were considered. The main results of this study are summarized as follows: 1. The length of the heat release region by 2-D measurements was about 11.4% longer than by point measurements, and increased proportionally to the volatile matter content. 2. The temperature rise of the 35~45 µm coal particles was faster than the 75~90 µm particles, and the peak temperature was 3% higher. The length of the heat release region was reduced by about 10%. 3.

Moisture content had little influence on pulverized coal flame when the moisture content was below 15%.

Acknowledgement This work is supported by research [2007-C-CD12-01-3-010] of the Korea Energy Management Corporation and project of the Pusan Clean Coal Center Research Grant.

References [1] L. D. Smoot, P. J. Smith, Coal combustion and gasification, Plenum press, New York, 1985. [2] A. Williams, M. Pourkashanian, J. M. Jones, Prog. Energy Combust. Sci. 27 (2001) 587-610. [3] H. Katalambula, J. Hayashi, K. Kitano, Journal of Chemical Engineering of Japan 33 (2000) - 10 -

49-56. [4] H. Katalambula, J. Hayashi, K. Kitano, T. Chiba, Energy and Fuels 11 (1997) 1033-1039. [5] R. Kurose, H. Tsuji, H. Makino, Fuel 80 (2001) 1457-1465. [6] Y. Zhao, H. Y. Kim, S. S. Yoon, Fuel 86 (2007) 1102-1111. [7] K. Annamalai, W. Ryan, S. Dhanapalan, Prog. Energy Combust. Sci. 20 (1994) 487-618. [8] S. C. Saxena, Prog. Energy Combust. Sci. 16 (1990) 55-94. [9] T. Suda, K. Masuko, J. Sato, A.Yamamoto, K. Okazaki, Fuel 86 (2007) 2008-2015. [10] J. O. L. Wendt, Prog. Energy Combust. Sci. 6 (1980) 201-222. [11] M. Zhu, H. Zhang, G. Tang, Q. Liu, J. Lu, G. Yue, S. Wang, S. Wan, Proc. Combust. Inst. 32 (2009) 2029-2035. [12] L. D. Timothy, A. F. Sarofim, J. M. Beer, Symposium (international) on Combustion 19 (1982) 1123-1130. [13] M. Gieras, R. Klemens, P. Wolanski, S. Wojcicki, Symposium (international) on Combustion 21 (1986) 315-323. [14] W. J. Mclean, D. R. Hardesty, J. H. Pohl, Symposium (international) on Combustion 18 (1981) 1239-1248. [15] H. Kobayashi, J. B. Howard, A. F. Sarofim, Symposium (international) on Combustion 16 (1977) 411-425. [16] J. B. Howard, R. H. Essenhigh, Symposium (international) on Combustion 11 (1967) 399408. [17] J. Yu, M. Zhang and J. Zhang, Proc. Combust. Inst. 32 (2009) 2037-2042. [18] S. M. Choi, KSME International Journal. 4 (1990) 71-77. [19] W. R. Seeker, G. S. Samuelsen, M. P. Heap, J. D. Trolinger, Symposium (international) on Combustion 18 (1981) 1213-1226. [20] J. Ballester, T. García-Armingol, Prog. Energy Combust. Sci. 36 (2010) 375-411. [21] A. G. Gaydon, The Spectroscopy of Flames, 2nd ed., Wiley, New York, 1974. [22] N. Docquier, S. Belhalfaoui, F. Lacas, N. Darabiha, Proc. Combust. Inst., 28 (2000) 17651774. [23] F. Lacas, B. Leroux, N. Darabiha, Proc. Combust. Inst. 30 (2005) 2037-2045. [24] A. Molina, C. R. Shaddix, Proc. Combust. Inst. 31 (2007) 1905-1912. [25] C. R. Shaddix, A. Molina, Proc. Combust. Inst. 32 (2009) 2091-2098. [26] F. Akamatsu, T. Wakabayashi, S. Tsushima, M. Katsuki, Y. Mizutani, Y. Ikeda, N. Kawahara, T. Nakajima, Meas. Sci. Technol. 10 (1999) 1240-1246. [27] S. M. Hwang, R. Kurose, F. Akamatsu, H. Tsuji, H. Makio, M. Katauki, JSME International Journal 49 (2006) 1316-1327. [28] S. M. Hwang, R. Kurose, F. Akamatsu, H. Tsuji, H. Makio, M. Katauki, Energy and Fuels - 11 -

19 (2005) 382-392. [29] J. R. Kim, F. Akamatsu, G. M. Choi, D. J. Kim, Thermochim. Acta 491 (2009) 109-115. [30] S. M. Godoy, F. C. Lockwood, Fuel 77 (1998) 995-999. [31] H. Tsuji, A. K. Gupta, T. Hasegawa, M. Katsuki, K. Kishimoto, M. Morita, High Temperature Air Combustion, CRC Press LLC, 2003. [32] A. Bosoaga, N. Panoiu, L. Mihaescu, R. I. Backreedy, L. Ma, M. Pourkashnian, A. Williams, Fuel 85 (2006) 1591-1598. [33] B. W. Asay, L. D. Smoot, P. O. Hedman, Combust. Sci. Technol. 35 (1983) 15-31.

- 12 -

Mathematical Model of the Low-Temperature Oxidation of Coal in Coal Stockpiles and Dumps Ing. Marian BOJKO, Ph.D, Department of Hydromechanics and Hydraulic Equipment, Faculty of mechanical Engineering, VŠB-Technical University of Ostrava, tř. 17. listopadu 15, 708 33 Ostrava-Poruba, tel. (+420) 597 324 385, e-mail [email protected]

Doc. RNDr. Milada KOZUBKOVÁ, CSc.,

Department of Hydromechanics and Hydraulic Equipment, Faculty of mechanical Engineering, VŠB-Technical University of Ostrava, tř. 17. listopadu 15, 708 33 Ostrava-Poruba, tel. (+420) 597 323 342, e-mail [email protected]

Ing. Zdeněk MICHALEC,

Institute of Geonics AS CR, v. v. i., Studentská 1768, 708 00 OstravaPoruba, tel. (+420) 596 979 228, e-mail [email protected]

Abstract Article defines mathematical model of the low-temperature oxidation of bituminous coal. The mathematical model defines single phase mathematical model with porous zone as coal where consumption of oxygen, production of smoke exhaust and heat are solved as source terms in transport equations. The rate constant defines by Arrhenius expression. Parameters of Arrhenius equation (activation energy and pre-exponential factor) are determined from experimental measuration. For numerical calculation method of finite volume (software ANSYS FLUENT 12) was used. 1

INTRODUCTION

Problem of low-temperature oxidation of coal stockpiles and dumps is presented as dangerous risk with respect to environment, when influence of atmospheric conditions causes self ignition of coal stockpile. The spontaneous ignition of coal matter in coal stockpile and dump is strongly dependent on wind velocity. Flow in coal stockpile or dump is laminar and flow in boundary layer is fully turbulent. Then the simulation includes two diametrically different regimes of flow. This problem is characterized by changes of coal temperature, humidity and oxygen consumption [1], [2], [3], [4]. Low-temperature oxidation of coal is heterogeneous chemical reaction producing carbon dioxide ( CO2 ), carbon monoxide ( CO ), water vapor ( H 2 O ) and oxy-coal. Solution of this problem can be solved by two methods, i. e. solution of two-phase mathematical model with heterogenous reaction or simplified by single phase model. The simplified single phase approach is based on experimental data and kinetics of reaction and will be tested. 2

THERMO-KINETICS DATA OF COAL FROM LABORATORY EXPERIMENT

Low-temperature oxidation of bituminous coal proceeds in two regimes. In the first one the gradual accumulation of heat is observed (so called incubation process) which can keep on several weeks. Subsequently coal oxidation is noticeably accelerated and in case of optimal conditions the temperature accumulation passes into spontaneous combustion. Turning point is characterized by critical temperature which is changed in interval 70°C ÷ 100°C. For these two regimes there are two different chemical reaction mechanisms defined using reaction rate determined by Arrhenius equation. Coal oxidation reaction is a function of temperature and oxygen concentration (O2). The basic parameters of Arrhenius equation are thermo-kinetic data (activation energy and pre-exponential factor) defined often by 1

experimental measurement. From literature sources it is evident the great variance of thermokinetic data see [5], [6], in particular in case of coal types. In this paper the thermo-kinetic data from laboratory experiment using method of adiabatic oxidation [7] for numerical simulation are applied. Activation energy E  and pre-exponential factor  A are input parameters in Arrhenius equation and are used for definition of rate constant k as follows: 

E

k  A  e RT where R  8.3145J / molK  is universal gas constant.

(1)

Critical temperature from experimental measuration is defined Tkr=350.24 K. For first and second oxidation regime the activation energy and pre-exponential factor are evaluated in table 1. Tab. 1 - Activation energy E  and pre-exponential factor  A Temperature T≤Tkr T≥Tkr E [kJ.mol-1] 9.98 55.4 -1 A [s ] 0.00345 20600 The above evaluated values were used for numerical simulation. 3

MATHEMATICAL MODEL OF SINGLE PHASE LOW-TEMPERATURE OXIDATION

The region is supposed as porous region consisting of coal particles in case of laminar flow with superficial velocity (our problem) using porosity and inverse values of permeability, which is accounted for in continuity, momentum and all scalar transport equations:  continuity equation    u j   0 (2) t  xj where  [kg.m-3] is density of mixture and u j [m.s-1] is the component of velocity vector.  momentum equation    ui   ui u j   ui   p          f i  ui   t  xj  xi  x j  xj  

(3)

where p [Pa] is static pressure,  [Pa.s] is molecular viscosity, g [m.s-2] is gravity acceleration vector,  [m2] is permability of porous zone.  Local mass fraction of specie Yi Error! Reference source not found. is defined in the form m V  Yi  i  i i  i  i (4) m V  where mi [kg] is mass of specie i  , m [kg] is total mass of mixture,  i [1] is volume fraction of specie i  in mixture. Another variable which is using in theory of species is molar concentration Ci [kmol.m-3].  equation of species transport in conservative form is in form [8]

2





Yi     u jYi  J j ,i   Ri  Si   t x j xi

where J j ,i   Di,m

(5)

Yi [kgm-2s-1] is diffusion flux, Di,m [m2s-1] is diffusion coefficient, Ri x j

[kg.m-3.s-1] is the net rate of production of species i  by chemical reaction, S i is the rate of creation by addition from the dispersed phase plus any user-defined sources [8]. An equation of this form will be solved for N  1 species where N is the total number of fluid phase chemical species present in the system. The mass fraction of the species must sum to unity.  energy equation T  g c pg  (1   )  s c ps    c pgu jT      jl u j   t x j  xl (6)    T   c pTJ j ,i   S h     x j   x j  x j where T [K] is temperature,  is porosity of the medium,  s ,  g [kg.m-3] is solid resp. fluid medium density, c ps , c pg [J.kg-1.K-1] is solid resp. fluid medium specific heat, s ,  g [W.m.K-1] is solid resp. fluid medium thermal conductivity,  [W.m-1.K-1] is effective thermal conductivity   g  1   s and S h is source term. To respect the influence reaction mechanism between porous zone and fluid it is essential to add the source terms of reactant and product and energy including dependence on kinetic energy of reaction rate. The source terms are related to reaction rate of coal oxidation expressed by Arrhenius formula. Relation reaction rate of coal oxidation on temperature must be derived from experiment [7]. SO   YO A exp  E RT 1

2

2





where A is pre-exponential factor, E is activation energy for the reaction, R is universal gas constant. 4

SIMULATION OF LOW TEMPERATURE OXYDATION IN REAL TERRAIN WITH COAL STOCKPILE

CFD simulation perform in 4,3 x 4,3km and high 400m domain. This domain includes real Earth's surface, mine buildings and coal stock, Fig.1. The mine and coal stack is situated in Orlová Lazy near city Karviná.

Fig. 1 – Coal stock and mine buildings

3

4.1 Grid, physical properties, boundary conditions

Domain included entering surfaces, which were used to unification of air inlet and outlet altitude and flow area respectively. This modification make possible the definition of turbulent quantities. The mesh size ratio and cell size near the wall must be observed, because turbulent quantities equilibrium is strongly dependent on quality of mesh [8]. The mesh size is relatively extensive (ca 1mil. cells) if the mesh criteria are within accepted limits, see Fg. 2. The number of cells must be smaller with respect to implementation of coal oxidation and spontaneous ignition. Mine buildings can be substituted by simply porous domain, which has pressure drop equal to origin domain with buildings see Fig.2. Complicated geometry of the buildings was replaced by simply porosity volume. Resistant coefficient of the porous domain was specified by means of thin domain near the buildings. Pressure resistance coefficient is calculated by means of pressure drop and change of reference velocity of wind for real geometry with buildings. N E

W

Inlet 4

Inlet 3

S

Fig. 2 Grid of terrain with coal stock and detail of mine buildings Physical properties Physical properties of coal (density, specific heat, thermal conductivity) are defined in Tab 2. Values of permeability   and porosity   of coal are gained from literature [9] and moreover there exit dependence between permeability and porosity for spherical particles in form: d 2 3  (7) 2 1501    In this table the constants used in source terms in mathematical model. Tab. 2 – Physical properties of coal, porosity and permeability of coal stock Unity Coal -3 Density kg.m 1300 -1 -1 Specific heat J.kg .K 1090 Thermal conductivity W.m-1.K-1 0.275 Porosity [-] 0.15

4

[m-2]

Permeability Boundary conditions

7.785.10-12

CFD simulation is calculated with mesh see Fig 2. Only two dominant directions of the wind are simulated. Two dominant directions of the wind are South and East. Velocity profile in boundary layer is described by power function, and temperature profile is described by linear function. Atmosphere is stable, so the temperature of air degreasing with altitude. The air compressibility influence is needed for boundary condition definition, because the domain high is inconsiderable and density of air is function of altitude. Velocity profile is defined by power function (8) with reference velocity vref  5 m/s. Reference velocity is defined in zref  10 m altitude. Power coefficient for stable state of atmosphere is p  0.14 . v  vref

 z     10 

p

(8)

where: z – altitude m ,

m v ref – reference velocity in 10 m altitude   , s p – power coefficient - , Temperature profile is defined by linear function, which describes degreasing of temperature as function of altitude. Temperature of Earth's surface is chosen t 0  19 °C and T0  292 K respectively. Turbulent kinetic energy and altitude

k

and dissipation  are defined by frictional velocity v  0.2258



3

v 0.4  z

(9)

2

k

v 0.3

(10)

where:

m v* – frictional velocity   , s Previous formulas is used for formulation of boundary profiles on inlets. Inlet of domain is defined by means of mass flow rate and direction vector. 4.2 Results of numerical calculation

Results of numerical calculation are evaluated by velocity vectors in two cross sections from East direction of wind, see Fig. 2. From results is evidently that maximum of velocity is 1.29 m/s. Then from result you can see detail of velocity field around of coal stock.

5

Fig. 3 – Velocity vectors from East direction of wind with detail of coal stock Next result is the temperature field through coal stock from East direction of wind, see Fig. 4. The range of temperature is . Maximum temperature is near Earth surface and surface of coal stock.

Fig. 4 – Temperature field through coal stock from East direction of wind

Fig. 5 – Temperature field through coal stock from South direction of wind 6

Fig. 5 shows the temperature field through coal stock from South direction of wind. You can see that maximum of temperature is 297K and position of maximum is below the surface of coal stock. Comparison between simulations of various direction of wind is show in Fig. 6. 410

Temperature [K]

390

370

South

North

East

West

350

330

310

290 0

50

100

150

200

250

300

Time [days]

Fig. 6 – Maximum of temperature histories inside of coal stock for all directions of wind 5

CONCLUSIONS

This paper occupies with numerical solution of low-temperature oxidation of coal. In the first part are using experimental results of spontaneous combustion kinetics and rate in two temperature intervals, where the parameters of Arrhenius rate (activation energy and preexponential factor) were derived. Then the singe phase mathematical model of lowtemperature oxidation was defined. This model used the source term of oxygen consumption (O2) and production of carbon dioxide (CO2), carbon monoxide (CO), water (H2O) and heat. These terms were formed by user´s defined functions. Subsequently this model was used in real terrain with coal stockpile. Numerical calculation was realized for four direction of wind (North, South, East, West). From results of maximum of temperature histories inside of coal is evidently that direction of wind is important for spontaneous heating. Work is financed by project Czech Science Foundation - GA 105/08/1414 Mathematical modelling of coal spontaneous in coal stockpiles and dumps REFERENCES [1]

[2] [3] [4] [5]

KRISHNASWAMY, S., GUNN, R. D., AGARWAL, K. P. Low-temperature oxidation of coal 2. An eprimental and modelling investigation using a fixed-bed isothermal flow reactor, Fuel, 1996, Vol. 75, No. 3, pp. 344-352. ISSN 0016-2361. SUJANTI, W., ZHANG, D. A laboratory study of spontaneous combustion of coal: the influence of inorganic matter and reactor size. Fuel, 1999, Vol. 78, pp. 549-556. ISSN 0016-2361. YUAN, L., SMITH, A. Numerical study on effects of coal properties on spontaneous heating in longwall gob areas, Fuel, 2008, Vol. 87, pp. 3409-3419. ISSN 0016-2361. ARISOY, A. AKGUN, F. Modelling of spontaneous combustion of coal with moisture content included, Fuel, 1994, Vol. 73, pp. 281-286. ISSN 0016-2361. KRAJČIOVÁ, M. JELEMENSKÝ, L., KIŠA, M., MARKOŠ, J. Model prediction on self-heating and prevention of stockpiled coals. Journal of Loss Prevention in the Process Industries, 2004, (17), pp. 205-216. ISSN 0950-4230.

7

[6] [7]

[8] [9]

ROSEMA, A., GUAN, H., VELD, H. Simulation of spontaneous combustion, to study the cause of coal fires in the Rujigou Basin. Fuel, 2001, Vol. 80, pp. 7-16. ISSN 0016-2361. CYGANKIEWICZ, J. Badanie skłonności polskich węgli do samozapalenia metodą testu adiabatycznego. In Větrání a bezpečnost dolů, mezinárodní kongres, Ostrava, VŠB-TU Ostrava, 2000, str, 22-57. FLUENT: Fluent 12 - User’s guide, Fluent Inc. 2007. VŠB-TU Ostrava. URL:http:// http://spc.vsb.cz/portal/cz/documentation/manual/index.php. LIMING, Y., SMITH, A. C. CFD modeling of spontaneous heating in a large-scale coal chamber. Journal of Loss Prevention in the Process Industries. 22, 2009, p. 426-433. ISSN: 0950-4230.

8

Manuscript Not AVAILABLE

2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission PROGRAM TOPIC : 2. GASIFICATION – SYNGAS CLEANUP SLIPSTREAM TESTS OF PALLADIUM SORBENTS FOR HIGH TEMPERATURE CAPTURE OF MERCURY, ARSENIC AND SELENIUM FROM FUEL GAS Hugh G.C. Hamilton (Dr.) Johnson Matthey Technology Centre e-mail: [email protected] Tony Wu : [email protected] Subhash Datta : [email protected] Robert C. Lambrecht : [email protected] John Wheeldon : [email protected] National Carbon Capture Center Southern Company, Wilsonville, Al 35186 U.S.A. Tel : 00 1 205 670 5875 Fax : 00 1 205-670-5843 Evan J. Granite : [email protected] Henry W. Pennline : [email protected] National Energy Technology Laboratory, United States Department of Energy P.O. Box 10940, Pittsburgh, PA 15236-040 USA Tel : 00 1 412 386 4607 Fax : 00 1 412 386 6004 Hugh Hamilton : [email protected] Liz Rowsell : [email protected] Stephen Poulston : [email protected] Andrew Smith : [email protected] Johnson Matthey Technology Centre, Blounts Court, Sonning Common, Reading, BerkshireRG4 9NH UK Tel : 00 44 (0)118 924 2000 Fax : 00 44 (0)118 924 2254

Keywords: National Carbon Capture Center (NCCC), mercury, arsenic, selenium, gasification, sorbent Abstract: In gasification for power generation, the removal of mercury and other trace elements such as arsenic, selenium and phosphorus by sorbents at elevated temperatures preserves the high thermal efficiency of the integrated gasification combined cycle system. Most commercial sorbents display poor capacity for elemental mercury at elevated temperatures. Palladium is an attractive sorbent candidate for the removal of mercury and the trace elements from fuel gases at elevated temperatures. The National Carbon Capture Center at the Power Systems Development Facility (PSDF) in Wilsonville, Alabama, is a large-scale flexible test facility established to develop and demonstrate a wide range of advanced power generation technologies that are critical to developing highly efficient power plants that capture carbon dioxide. The palladium-based sorbents have been tested for

extended periods of time in slipstreams of fuel gas at the NCCC. These results will be described, and possible future testing will be discussed.

Manuscript Not AVAILABLE

2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission

PROGRAM TOPIC: GASIFICATION EXACT TITLE OF PAPER: MERCURY MEASUREMENT AND REOMOVAL FROM AN ENTRAINED FLOW SLAGGING COAL GASIFIER

Dennis Lu, Research Scientist, CanmetENERGY/Natural Resources Canada 1 Haanel Drive, Ottawa, Ontario, CANADA K1A 1M1, [email protected], (613) 996-2760 Robin Hughes, Research Engineer, CanmetENERGY/Natural Resources Canada 1 Haanel Drive, Ottawa, Ontario, CANADA K1A 1M1, [email protected], (613) 996-0066 Karl Abraham, Senior Program Engineer, Environment Canada 1 351 St-Joseph Blvd., 9th Floor, Gatineau, QC K1A 0H3 CANADA, [email protected], (819) 953-2079 Ben Anthony, FBC and Gasification Group Leader, CanmetENERGY/Natural Resources Canada 1 Haanel Drive, Ottawa, Ontario, CANADA K1A 1M1, [email protected], (613) 996-2868

Abstract: In typical synthesis gas (syngas) from entrained flow slagging coal gasification thermodynamic calculations predict that only the elemental form of mercury (Hg0) is stable rather than the bivalent oxidized form (Hg ++), such as is present in HgCl2. Therefore, Hg0 is expected to be dominant in such a reducing environment. However, the chemical and physical processes governing the interactions of mercury forms with syngas components are poorly understood, particularly the results of heterogeneous reactions involved in gasification syngas are lacking. Data on Hg emissions from gasification systems have not been sufficiently reliable and the mass balance closures have high associated error ranges because of problems with sampling and analysis, which make understanding mercury characteristics under gasification conditions difficult. This paper presents studies on mercury measurement specifically applicable to an entrained flow slagging gasifier at the CanmetENERGY 0.6MW pilot scale gasification plant. Mercury speciation has been successfully measured directly from a high-P and high-T gasifier vessel and as well quenched downstream syngas containing nitrous and sulfurous species. A bench-scale fixed-bed system was also used to investigate the Hg removal performance of sorbents, including commercial activated carbon, sulphur- and alkali-doped activated carbons, limestone and dolomite. The fixed bed was operated above the dew point temperature of the synthesis gas for the activated carbon sorbents, and for CaO-based sorbents at a higher temperature in the range of 500-700°C, which has been chosen to match the operating conditions of the CaO-sorbent looping process for CO2 capture.

PERFORMANCE IMPROVEMENT OF A DESULFURIZATION SORBENT FOR WARM SYNTHESIS GAS CLEANUP Jeom-In Baek, Jungho Ryu, Tae Hyoung Eom, Joong Beom Lee, Yong-Ho Wi, Chong Kul Ryu* KEPCO Research Institute 305-380 Munji-ro 65, Yuseong-gu, Daejeon, Republic of Korea Tel: 82-42-865-5250, Fax: 82-42-865-5202, e-mail: [email protected]

Abstract KEPCO Research Institute has improved the performance of a solid regenerable desulfurization sorbent prepared by spray-drying method. Here, we present a newly developed desulfurization sorbents which showed improved physical properties and reactivity compared to our previous desulfurization sorbents. The attrition resistance of the new desulfurization sorbents was much higher than the previously developed sorbents. Other physical properties such as average particle size, tap density, and shape were suitable for the fluidized-bed applications. Sulfur sorption capacity of the new sorbent, which was measured by thermogravimetric analyzer using a simulated synthesis gas containing 1% H2S, was around 10 wt% at the reaction temperatures of 500 and 650 oC for absorption and regeneration, respectively. In the future works, an in-depth sorbent analysis and reactivity test according to the reaction temperature change will be carried out to improve the performance of the spray-dried desulfurization sorbent and a fluidized-bed desulfurization process. Introduction Integrated Gasification Combined Cycle (IGCC) is considered as a clean power generation technology with a high thermal efficiency and low emission. The higher cost of electricity (COE) of IGCC power plant than that of pulverized coal (PC) power plant has been barrier to the widespread use of IGCC plant. However, when considered CO2 capture together the COE of coal IGCC with CO2 capture is evaluated less than that of a PC power plant with CO2 capture (DOE/NETL, 2007; IEA, 2008). The international cooperation to reduce CO2 emission for the mitigation of climate change is demanding more cost-efficient power plants with CO2 capture process. Therefore, it is expected that the world IGCC capacity will be rapidly increased in the near future. It was suggested that the efficiency of IGCC without CO2 capture would decrease around 69% and the COE would increase 3240% when the currently available commercial CO2 capture process, Selexol, was adopted (DOE/NETL, 2007). Minimizing the COE increase and efficiency loss is required to accelerate the deployment of IGCC with CO2 capture. New technologies such as

advanced air separation unit, warm gas clean-up and CO2 capture at warm temperatures should be developed to reduce the COE increase of IGCC caused by CO2 capture. A raw synthesis gas from a coal gasifier contains hydrogen sulfide (H2S). It is necessary to remove H2S to avoid the poison of the catalyst used in downstream and to protect the plants from corrosion. Current commercial-technologies to remove H2S are using wetscrubbing method which uses physical solvents or chemical amines. The temperature of gas stream in this capture process should be cooled down to around 40 oC, which results in a efficiency loss. Warm synthesis gas clean-up is attractive because it can be operated at an elevated temperature, reducing the efficiency loss. Therefore, warm synthesis gas cleanup is considered as a key technology to develop an advanced IGCC system with high thermal efficiency. Solid regenerable sorbents are used for H2S removal in warm synthesis gas clean-up. A lot of work has been done to develop solid sorbents for desulfurization at warm temperatures (Bu et al., 2008; Gupta and Gangwal, 1992; Gupta et al., 1997; Lee et al., 2008; Sanchez-Hervas et al, 2005). RTI/Eastman has been developing desulfurization process using ZnO-based sorbents. In the tests at a pilot plant installed in Eastman’s Kingsport coal gasification plant, sulfur removal more than 99.9% was proved during operation over 3000 h (Gupta et al., 2008). KEPCO Research Institute also has been developing solid regenerable desulfurization sorbents for warm synthesis gas clean-up. ZAC-32N prepared by spray drying method showed >99% sulfur removal from coalderived synthesis gas during continuous steady state operation for 720 h in a bench scale circulating fluidized bed process, which was developed by Korea Institute Energy Research (KIER). To improve the performance of ZAC-32N, several new desulfurization sorbents including PS-2 were designed and tested (Baek et al, 2009). Although some properties such as surface area and sulfur sorption capacity of PS-2 were improved, the mechanical strength of PS-2 was less than ZAC-32N. In this study, a new zinc-based sorbent was designed to improve the performances of our previously developed desulfurization sorbents. Physical properties and H2S sorption capacities of newly spray-dried desulfurization sorbent were investigated and compared with previous ones. Experimental Section Sorbent Preparation: New desulfurization sorbent, PS-8, was designed based on the PS-2 composition to improve the mechanical strength. Raw materials for the preparation of PS-8 were composed of 50 wt% ZnO as an active material, 50 wt% matrices containing multi binders, and a Ni-based promoter by 7.5 wt% to improve the attrition resistance and regenerability of the sorbents. Commercially available powder type raw materials were well mixed with pure water. The mixed slurry was comminuted with ball mill to control particle size and make homogeneous colloidal slurry. The homogenized slurry was spray-dried to form free-flowing solid sorbents with suitable shape, particle size, bulk density suitable for fluidized-bed applications in a 10 kg batch scale. The spray-dried green body was calcined at 650 C in a muffle oven for 5 h after pre-drying at 120 C overnight to enhance mechanical strength with burning out the organic additives added during slurry preparation.

Physical Characterization: The shape of calcined desulfurization sorbents was obtained using scanning electron microscope (SEM). Average particle size of the sample was measured using a MEINZER II sieve shaker. The measuring procedure for particle size and size distribution was followed by the American Society for Testing and Materials (ASTM) E-11. Packing density of each sorbent was determined using the Autotap instrument (Quantachrome) proposed in ASTM D 4164-88 to measure the tapped density of formed catalyst. The packing density were obtained by dividing the sorbent mass by its tapped volume. Brunauer-Emmett-Teller (BET) surface area of the sample was determined by N2 physisorption. Pore volume and porosity was obtained with Hg porosimetry. The Hg intrusion volume between particles was subtracted from the total Hg intrusion volume to calculate porosity contributed by the pore volume only within the particle. Attrition resistance of the calcined oxygen carrier was measured with a modified three-hole air-jet attrition tester based on the ASTM D 5757-95. The attrition resistance was determined at 10 standard liters per minute (slpm) over 5 h as described in the ASTM method. The attrition index (AI) is the percent fines generated over 5 h. The fines are particles collected at the thimble after 5 h from the start, which was attached to the gas outlet. AI = [total fine collected for 5 h/amount of initial sample (50 g)] x 100 %

(4)

A lower AI indicates a better attrition resistance of the bulk particles. The AIs of fresh Akzo fluid catalytic cracking (FCC) catalyst, which was used as a reference for comparison, was 22.5%, under the same measurement condition. TGA Chemical Reactivity: Chemical reactivity of the desulfurization sorbents was measured with a thermal gravimetric analyzer (Thermo Cahn, TherMax500) which was modified to supply water vapor into the reactor tube. The overall schematic diagram of TGA system was shown in Figure 1. An alumina crucible was used to hold the desulfurization sorbents. The gas flow rate was controlled by mass flow controller. Reacting gas mixture was introduced into the lower part of the reactor tube and flowed upward. An inert gas, N2, was fed into the balance chamber and flowed downward to the outlet of TGA furnace to protect the balance from the diffusion of reactant gas and reduce the buoyancy effect. The weight and furnace temperature data were continuously recorded by data acquisition unit. H2S absorption and sorbent regeneration were conducted under atmospheric pressure. A simulated synthesis gas composed of 1% H2S, 3% CO2, 30% H2, 64% CO and 2% H2O was used for absorption. Air was used as a regeneration gas. Before the sulfidation step, clean synthesis gas (without H2S) was supplied for 30 min to obtain weight baseline. After absorption, the reactor was purged with N2 and heated to the regeneration temperature. The amount of samples used for reactivity tests were around 20 mg scale. A total flow rate was 0.15 slpm in all steps. To prevent water vapor condensation, electric line heaters were installed along the gas line and gas stream was heated above saturation point. The sulfur sorption capacity was determined from the weight increase during H2S absorption.

Data acquisition

Microbalance E/B

Sytem outlet Vent

Air for pressure control

Back pressure regulator

Balance purge gas

CO2 TGA furnace H2O

N2

N2

CO H2 H2S/CO

Vent

Furnace gas Thermocouple

H2O vapor N2 Air

Figure 1. Schematic diagram of TGA system for desulfurization tests. Results and Discussion Physical Properties: The physical properties of the spray-dried desulfurization sorbents were summarized in Table 1. All three sorbents had spherical shape. The average particle size (115 m) of newly developed sorbent, PS-8, was similar to that of previous sorbents, ZAC-32N and PS-2. Particle size of PS-8 ranged from 58 to 196 m for more than 99 wt% of the sorbents. PS-8 had much narrow particle size range compared to the previously developed sorbents, which will improve the operating stability of a circulating fluidizedbed system. The bulk density of PS-8 was also increased to 0.92 g/cm3, which is more preferable value for fluidized-bed applications. The BET specific surface area of PS-8 (52.9 m2/g) was similar to that of PS-2 (51.3 m2/g) and higher than that of ZAC-32N (39.3 m2/g). The porosity and pore volume of PS-8 were similar to those of ZAC 32N and PS-2. The attrition resistance is an important physical property which decides the applicability of the sorbent for fluidized-bed process. Sorbents with higher AI above 60% are not suitable for bubbling bed reactor and those with AI above 30% are not suitable for transport reactor. The mechanical strength of PS-8 was much higher than that of ZAC 32N and PS-2. And

Figure 2. SEM images of spray-dried desulfurization sorbents.

Table 1. Physical properties of the desulfurization sorbents calcined at 650 °C ZAC 32N PS-2 PS-8 Shape Sphere Sphere Sphere Avg. particle size/µm 111 124 115 Size distribution/µm 38 – 250 57 – 303 58 – 196 Bulk density / (g/cm3) 0.77 0.86 0.92 2 BET / (m /g) 39.3 51.3 52.9 Hg porosity/% 35.6 32.1 32.8 3 Hg pore volume / (cm /g) 0.48 0.44 0.40 Attrition index (AI) / % 31.4 36.0 11.2 it was also higher than that of commercial FCC catalyst (22.5%). The high attrition resistance of PS-8 could be obtained by changing the multi-binder matrices. The increased attrition resistance will contribute to the reduction of sorbent make-up requirement.

Sulfur sorption capacity / wt%

TGA Chemical Reactivity: The sulfur sorption capacity of PS-8 at an absorption temperature of 500 C was presented in Figure 3, along with sorption capacities of ZAC32N and PS-2. The sulfur sorption capacities of sorbents were stabilized from the second cycle. PS-8 showed the sulfur sorption capacity of around 10 wt% after second cycle, which is similar to that of PS-2 and higher than that of ZAC-32N. The higher sorption capacity of PS-8 can be ascribed to the increased surface area resulted from the support matrix change because more active sites can be exposed to the surface. There was significant reduction of sorption capacity at the second cycle, compared with the first cycle. Several explanations can be given for this reduction although they are not clear at this moment. The assumption that zinc migrated to the surface and evaporated during cyclic can be one of the explanations. Another explanation is that sulfur was irreversibly absorbed on the support of sorbents. Detailed analysis on the structure of the fresh and used sorbents should be carried out to find out the accurate cause of the sulfur sorption capacity reduction at the second cycle. 16

ZAC-32N PS-2 PS-8

14

12

10

8 1

2

3

Cycles

Figure 3. Sulfur sorption capacities of the spray-dried desulfurization sorbents

Summary A dry regenerable zinc-based desulfurization sorbent for H2S removal from warm synthesis gas was prepared by spray-drying method. The physical properties of newly developed sorbent, PS-8, were characterized in terms of average particle size (115 µm), size distribution (58196 µm), bulk density (0.92 g/cm3), attrition resistance (AI, 11.2%), surface area (53 m2/g), and porosity (33%). The physical properties of PS-8 were suitable for fluidized bed-applications. Its attrition resistance was greatly improved compared with our previously developed sorbents, ZAC-32N and PS-8, and the sulfur sorption capacity of PS-8 was also improved compared to that of ZAC-32N. In our future works, the sulfur sorption capacity according to the reaction temperature change and the reason for the sorption capacity reduction at the second cycle will be investigated to improve the performance of PS-8 and to provide better operating condition of a warm synthesis cleanup process. Acknowledgement This work was supported by Energy Efficiency and Resources R&D program (2008CCD11P040000) under the Ministry of Knowledge Economy, Republic of Korea. The authors also would like to thank the Korea Electric Power Corporation (KEPCO) and the Korea Western Power Company, Ltd. for supporting the program. References Baek, J.-I.; Eom, T. H.; Lee, J. B.; Wi, W.-H.; Ryu, C. K., “Solid regenerable sorbents for H2S removal from syngas at a warm temperature”, 26th Annual International Pittsburgh Coal Conference, September 2023, 2009. Bu, X.; Ying, Y.; Zhang, C.; Peng, W., “Research improvement in Zn-based sorbent for hot gas desulfurization”, Power Technology 2008, vol. 180, 253258. DOE/NETL, “Cost and performance baseline for fossil fuel plant”, DOE/NETL-2007/1281, USA, 2007. IEA , “CO2 capture and storage”, OECD/IEA, 2008. Gupta, P. R.; Gangwal, S. K., “Enhanced durability of desulfurization sorbents for fluidized-bed applications”, DE93000247, 1992 Gupta, R. K.; Turk, B. S.; Vierheilig, A. A., “Desulfurization sorbents for transport-bed applications”, In proceedings of advanced Coal-based power and environmental systems ’97 conference, Pittsburgh, PA, July 2224, 2997. Gupta, R.; Turk, B.; Lesemann, M., Schlather, J.; Denton D., “Status of RTI/Eastman warm gas clean-up technology and commercialization plans”, Gasification Technologies Conference, October 8, 2008 Lee, J. B.; Baek, J.-I., Ryu, C. K.; Yi, C. K.; Jo, S. H.; Kim, S. H., “ Highly attritionresistant zinc-based sorbents for H2S removal by spray-drying technique”, Ind. Eng. Chem. Res. 2008, vol. 47, 44554464. Sanchez-Hervas, J. M.; Otero, J.; Ruiz, E., “A study on sulphidation and regeneration of Z-

sorb III sorbents for H2S removal from simulated ELCOGAS IGCC syngas”, Chem. Eng. Sci. 2005, vol. 60, 29772989.

Proceedings of the 27th International Pittsburgh Coal Conference, Istanbul, Turkey, October 11-14, 2010

INVESTIGATION OF THE COAL GASIFICATION PROCESS UNDER VARIOUS OPERATING CONDITIONS INSIDE A TWO-STAGE ENTRAINED FLOW GASIFIER Armin Silaen and Ting Wang * [email protected], [email protected] Energy Conversion & Conservation Center University of New Orleans New Orleans, Louisiana, USA through an incomplete combustion. Feedstock is partially combusted with oxygen and steam at high temperature and pressure with only less than 30% of the required oxygen for complete combustion being provided. The syngas produced can be used as a fuel, usually for boilers or gas turbines to generate electricity, or can be used to make a synthetic natural gas, hydrogen gas or other chemical products. The gasification technology is applicable to any type of carbonbased feedstock, such as coal, heavy refinery residues, petroleum coke, biomass, and municipal wastes.

ABSTRACT Numerical simulations of the coal gasification process inside a generic 2-stage entrained-flow gasifier fed with Indonesian coal at approximately 2000 metric tone/day are carried out. The 3-D Navier-Stokes equations and eight species transport equations are solved with three heterogeneous global reactions, three homogeneous reactions, and two-step thermal cracking equation of volatiles. The Chemical Percolation Devolatilization (CPD) model is used for the devolatilization process. Finite rates are used for the heterogeneous solid-to-gas reactions. Both finite rate and eddy-breakup combustion models are calculated for each homogeneous gas-to-gas reaction, and the smaller of the two rates is used. The water-shift reaction rate is adjusted to match available syngas composition from existing operational data without catalyst. This study is conducted to investigate the effects of different operation parameters on the gasification process including coal mixture (dry vs. slurry), oxidant (oxygen-blown vs. air-blown), and different coal distribution between two stages. In the two-stage coal-slurry feed operation, the dominant reactions are intense char combustion in the first stage and enhanced gasification reactions in the second stage. The gas temperature in the first stage for the dry-fed case is about 800 K higher than the slurry-fed case. This calls for attention of additional refractory maintenance in the dry-fed case. One-stage operation yields higher H2, CO and CH4 combined than if a two-stage operation is used, but with a lower syngas heating value. High heating value (HHV) of syngas for the one-stage operation is 7.68 MJ/kg, compared to 8.24 MJ/kg for two-stage operation with 72%-25% fuel distribution and 9.03 MJ/kg for two-stage operation with 50%-50% fuel distribution. Carbon conversion efficiency of the air-blown case is 77.3%, which is much lower than that of the oxygen-blown case (99.4%). The syngas heating value for the air-blown case is 4.40 MJ/kg, which is almost half of the heating value of the oxygen-blown case (8.24 MJ/kg).

1.1 Global Gasification Chemical Reactions This study only deals with global chemical reactions of coal gasification (Smoot and Smith, 1985) that can be generalized in reactions (R1.1) through (R1.9) below: Heterogeneous (solid and gas) phase C(s) + ½ O2 → CO,

ΔH°R = -110.5 MJ/kmol

(R1.1)

C(s) + CO2 → 2CO, ΔH°R = +172.0 MJ/kmol (R1.2) (Gasification, Boudouard reaction) C(s) + H2O(g) → CO + H2, ΔH°R= +131.4 MJ/kmol (R1.3) (Gasification) C + 2H2 → CH4,

ΔHoR = -87.4 MJ/kmol (Direct methanation)

(R1.4)

ΔH°R = -283.1 MJ/kmol

(R1.5)

Homogenous gas phase CO + ½ O2 → CO2,

1.0 INTRODUCTION

CO + H2O(g) → CO2 + H2 , ΔH°R = -41.0 MJ/kmol (Water-shift)

(R1.6)

CO + 3H2 → CH4 + H2O, ΔHoR = -205.7 MJ/kmol (Methanation)

(R1.7)

CH2.121O0.5855 → 0.5855CO + 0.2315H2 + 0.4145CH4 (R1.8) (Volatiles cracking) CH4 + ½ O2 → CO + 2H2 (R1.9) (Volatiles gasification via CH4)

Gasification is the process of converting various carbonbased feedstocks to clean synthetic gas (syngas), which is primarily a mixture of hydrogen (H2) and carbon-monoxide (CO) with minor methane (CH4) and inert nitrogen gas,

1

In this study, the methanation reactions are not considered. Reactions (R1.8) and (R1.9) involve volatiles. The volatiles are modeled to go through a two-step thermal cracking (R1.8) and gasification processes (R1.9) via CH4 as the intermediate product. The coal used in the study is subbituminous from Indonesia, whose compositions are given in Table 1a. It has a moisture content of 8.25%. Its moisturefree (MF) proximate and ultimate analyses compositions are listed in Table 1b. The compositions of volatiles are derived from the values of coal heating value, proximate analysis, and ultimate analysis.

configurations: single-stage down fired system and two-stage with multiple feed inlets. The model was constructed using GLACIER, an REI in-house comprehensive coal combustion and gasification tool. The basic combustion flow field was established by employing full equilibrium chemistry. Gas properties were determined through local mixing calculations and are assumed to fluctuate randomly according to a statistical probability density function (PDF), which is characteristic of the turbulence. Gas-phase reactions were assumed to be limited by mixing rates for major species as opposed to chemical kinetic rates. Gaseous reactions were calculated assuming local instantaneous equilibrium. The particle reaction processes include coal devolatization, char oxidation, particle energy, particle liquid vaporization and gasparticle interchange. The model also includes a flowing slag sub-model.

Table 1a Compositions of Indonesian sub-bituminous coal.

Volatile H2 O ash C H N S O Total, wt % HHV, kcal/kg

Weight % 38.31% 8.25% 3.90% 37.95% 2.68% 0.69% 0.31% 7.91% 100.00% 5690

The U.S. Department of Energy/National Energy Technology Laboratory (NETL) developed a 3D CFD model of two commercial-sized coal gasifiers [Guenther and Zitney (2005)]. The commercial FLUENT CFD software is used to model the first gasifier, which is a two-stage entrained-flow coal slurry-fed gasifier. The Eulerian-Lagrangian approach is applied. The second gasifier is a scaled-up design of transport gasifier. The NETL open source MFIX (Multiphase Flow Interphase eXchanges) Eulerian-Eulerian model is used for this dense multiphase transport gasifier. NETL also developed an Advanced Process Engineering Co-Simulator (APECS) that combines CFD models and plant-wide simulation. APECS enables NETL to couple its CFD models with steady-state process simulator Aspen Plus.

Table 1b Moisture-free (MF) compositions of Indonesian sub-bituminous coal.

Proximate Analysis (MF), wt% Volatile 51.29 Fixed Carbon (FC) 47.54 Ash 1.17 100.00

Ultimate Analysis (MF), wt% C 73.32 H 4.56 O 20.12 N 0.72 S 0.11 Ash 1.17 100.00

Silaen and Wang (2006) carried out a study that focused on the effect of flow injection directions on the gasification performance using the same generic two-stage entrained flow gasifier as studied by Chen et al. and Bockelie et al. Horizontal injection direction was compared to downward and upward direction. The results revealed that the horizontally tangential injection direction gave the best gasifier performance. Changing the direction of the first-stage injectors downward resulted in a carbon fuel conversion reduction, but produced more H2. Changing the direction of the second-stage injectors, however, did little to affect the overall flow patterns due to the smaller-quantity of coal injection (25%); therefore the gasifier performance was essentially insignificantly affected.

1.2 Recent Research Chen et al. (2000) developed a comprehensive threedimensional simulation model for entrained coal gasifiers which applied an extend coal gas mixture fraction model with the Multi Solids Progress Variables (MSPV) method to simulate the gasification reaction and reactant mixing process. The model employed four mixture fractions and separately track the variable coal off-gas from the coal devolatilization, char-O2, char-CO2, and char-H2O reactions. Chen et al. performed a series of numerical simulations for a 200 ton per day (tpd) two-stage air blown entrained flow gasifier developed for an IGCC process under various operation conditions (heterogeneous reaction rate, coal type, particle size, and air/coal partitioning to the two stages).

Silaen and Wang (2010) conducted a study that investigates the effects of different parameters on gasification performance including five turbulence models, four devolatilization models and three coal particle sizes. The Eulerian-Lagrangian approach with finite global reaction rates was applied. A two-step decomposition model was applied to volatiles cracking and gasification via benzene as the intermediate product. The results revealed that the standard kε and the Reynolds Stress Model (RSM) models gave consistent results. Smaller particles have a higher surface/volume ratio, react faster than larger particles, and produce syngas with higher heating value than larger particles. High inertia possessed by larger coal particles propel the

Bockelie et al. (2002) of Reaction Engineering International (REI) developed a CFD modeling capable of entrained flow gasifiers that focuses on two gasifier

2

solid species mass fraction on the surface, and particle surface area. The reaction rates are all global net rates, i.e., the backward reaction, calculated by equilibrium constants, are included in the global rate. Therefore, the finite rate employed in this study implicitly applies to the local equilibrium approach. Reaction rate constants used in this study are summarized in Table 2.

particles cross the gas streamlines and increase particle-gas mixing which result in enhanced reaction rate, but they take longer to complete reaction of the entire particle. The single rate devolatilization model and the chemical percolation model produced moderate and consistent devolatilization rate. This study is the continuous work of Silaen and Wang (2006, 2010) and focuses on investigating the effects of different operation parameters on the gasification process including coal mixture (dry vs. slurry), oxidant (oxygen-blown vs. air-blown), and different coal distribution between two stages.

The reaction rate of the water-shift, adopted from Jones and Lindstedt (1988), is found to be too fast in this study because the rate is obtained with the presence of catalyst. Considering no catalyst is added in a typical gasifier, the water shift reaction rate is purposely slowed down to make the syngas composition consistent with that in the actual production of a commercial entrained-flow gasifier with coalslurry feed from bottom.

2.0 COMPUTATIONAL MODEL The computational model and submodels (devolatilization, reactions, particle dynamics, gasification) used in the study are the same as developed by Silaen and Wang (2010), so all equations and detailed modeling intricacies are not repeated here, but they are briefly summarized below. The time-averaged steady-state NavierStokes equations as well as the mass and energy conservation equations are solved. Species transport equations are solved for all gas species involved. The standard k-ε turbulence model is used to provide closure. Silaen and Wang (2010) applied five turbulence models (standard k-ε, k-ω, RSM, k-ω SST, and k-ε RNG) and reported that the standard k-ε turbulence model yields reasonable results without requiring very much computational time when compared to other turbulence models. Enhanced wall function and variable material property are used. The P1 model is used as the radiation model.

For liquid droplets, water evaporates from the particle’s surface when the temperature is higher than the saturation temperature (based on local water vapor concentration). The evaporation is controlled by the water vapor partial pressure until 100% relative humidity is achieved. When the boiling temperature (determined by the air-water mixture pressure) is reached, water continues to evaporate even though the relative humidity reaches 100%. After the moisture is evaporated, due to either high temperature or low moisture partial pressure, the vapor diffuses into the main flow and is transported away. Please refer to Silaen and Wang (2010) for details. Table 2 Summary of reaction rate constants used in this study Reaction Rate Constant Solid-gas heterogeneous reactions: n k = AT exp(-E/RT) C(s) + ½O2 → CO

C(s) + CO2 → 2CO

n

k = AT exp(-E/RT)

(Gasification, Boudouard reaction) C(s) + H2O(g) → CO + H2

n=0 2

(Combustion)

The flow (continuous phase) is solved in Eulerian form as a continuum while the particles (dispersed phase) are solved in Lagrangian form as a discrete phase. Stochastic tracking scheme is employed to model the effects of turbulence on the particles. The continuous phase and discrete phase are communicated through drag forces, lift forces, heat transfer, mass transfer, and species transfer. The finite-rate combustion model is used for the heterogeneous reactions, but both the finite-rate and eddy-dissipation models are used for the homogeneous reactions, and the smaller of the two is used as the reaction rate. The finite-rate model calculates the reaction rates based on the kinetics, while the eddy-dissipation model calculates based on the turbulent mixing rate of the flow. Gasification or combustion of coal particles undergoes the following global processes: (i) evaporation of moisture, (ii) devolatilization, (iii) gasification to CO and (iv) combustion of volatiles, CO, and char. The Chemical Percolation Devolatilization (CPD) model [Fletcher and Kerstein (1992), Fletcher et. al (1990), and Grant et. al (1989)] is chosen as the devolatilization model based on the finding by Silaen and Wang (2010) that the Kobayashi two-competing rates devolatilization model [Kobayashi et. al. (1976)] is very slow, while the CPD model gives a reasonable result. For solid particles, the rate of depletion of the solid, due to a surface reaction, is expressed as a function of kinetic rate,

Parameters

n

k = AT exp(-E/RT)

(Gasification)

-0.5

A = 0.052 kg/m .Pa 7 E = 6.1x10 J/kmol n=0 2

-0.5

2

-0.5

A = 0.0732 kg/m .Pa 8 E = 1.125x10 J/kmol n=0 A = 0.0782 kg/m .Pa 8 E = 1.15x10 J/kmol

Gas phase homogeneous reactions: n k = AT exp(-E/RT) CO + ½ O2 → CO2

n=0 12

CO + H2O(g) → CO2 + H2 (Watershift)

n

k = AT exp(-E/RT)

A = 2.2x10 E = 1.67x108 J/kmol n=0 2

A = 2.75x10 E = 8.38x107 J/kmol Eddy-dissipation only CH2.121O0.5855 → 0.5855CO + 0.2315H2 + 0.4145CH4 Eddy-dissipation only CH4+ ½O2 → CO + 2H2

The computation is carried out using the finite-volumebased commercial CFD software FLUENT 12.0 from ANSYS, Inc. The simulation is steady-state and uses the pressurebased solver, which employs an implicit pressure-correction scheme and decouples the momentum and energy equations. SIMPLE algorithm is used to couple the pressure and velocity. Second order upwind scheme is selected for spatial discretization of the convective terms. For the finite rate model where the Eulerian-Lagrangian approach is used, the iterations are conducted alternatively between the continuous

3

and the dispersed phases. Initially, two iterations in the continuous phase are conducted followed by one iteration in the discrete phase to avoid the flame from dying out. Once the flame is stably established, twenty iterations are performed in the continuous phase followed by one iteration in the dispersed phase. The drag, particle surface reaction, and mass transfer between the dispersed and the continuous phases are calculated. Based on the dispersed phase calculation results, the continuous phase is updated in the next iteration, and the process is repeated. Converged results are obtained when the residuals satisfy mass residual of 10-3, energy residual of 10-5, and momentum and turbulence kinetic energy residuals of 104 . These residuals are the summation of the imbalance in each cell, scaled by a representative for the flow rate. The computation was carried out in parallel processing on two dual-core Pentium clusters with 12 nodes each. 2.1 Physical Assumptions

Characteristics

of

the

Model

and

This paper studies a two-stage entrained flow coal gasifier as shown in Fig. 1. Meshed computational domain is shown in Fig. 2. The grid consists of 1,106,588 unstructured tetrahedral cells. In the simulations, the buoyancy force is considered, varying fluid properties are calculated for each species and the gas mixture, and the walls are assumed impermeable and adiabatic. The flow is steady and no-slip condition (zero velocity) is imposed on the wall surfaces.

Fig. 2 Meshed computational domain of the two-stage entrained-flow gasifier.

3.0 BOUNDARY AND INLET CONDITIONS Raw Syngas

Indonesian sub-bituminous coal is used as feedstock in this study; its composition is given in Table 1. Boundary conditions for the baseline case are shown in Fig. 3. The summary of the studied cases are listed in Table 3. In the baseline (Case 1) of this study, coal-slurry-fed and two-stage configuration is used with fuel distribution of 75%-25% between the first and the second stages. Total mass flow rates of the coal slurry and the oxidant are 21.9 kg/s and 9.92 kg/s, respectively. The total mass flow rate of the dry coal powder case (Case 2) is 12.90 kg/s. The difference in fuel mass flow rates is caused by water added for slurry. The moisture in the coal is included in both slurry and dry feed cases. The coal/water weight ratio of the coal slurry is 60%-40%. Oxidant/coal slurry feed rate gives O2/coal equivalence ratio of 0.4. The equivalence ratio is defined as the percentage of oxidant provided over the stoichiometric amount for complete combustion.

Top view of second stag injectors 9m

2nd Stage

Coal Slurry

1st Stage

2.25 m

Top view of first stage injectors

The oxidant is considered as a continuous flow and coal slurry is considered as a discrete flow. The discrete phase only includes the fixed carbon and water from the moisture content of coal (8.25% wt) and water added to make the slurry. The slurry coal is treated as particles containing both coal and liquid water. Other components of the coal, such as N, H, S, O, and ash, are injected as gas, together with the oxidant in the continuous flow. N is treated as N2, H as H2, and O as O2. S and ash are not modeled and their masses are lumped into N2. The coal slurry size is uniformly given as 50

Coal Slurry & Oxygen 0.75 m 1.5 m 0.75 m

Fig. 1 Schematic of the two-stage entrained-flow gasifier.

4

μm for the purpose of conveniently tracking the particle size reducing rate. Investigation of effects of coal particle size on gasification performance has been performed by Silaen and Wang (2009 and 2010) and is not repeated here.

4.0 RESULTS AND DISCUSSIONS The following five cases are studied: Case 1: Baseline case, oxygen-blown, coal slurry, 75%-25% distribution in 2 stages Case 2: Oxygen-blown, dry coal, 75%-25% distribution in 2 stages. Case 3: Oxygen-blown, coal slurry, 50%-50% distribution in 2 stages. Case 4: Oxygen-blown, coal slurry, 100% distribution in the 1st stage. Case 5: Air-blown, coal slurry, 75%-25% distribution in 2 stages

The walls are assigned as adiabatic with internal emissivity of 0.8. The boundary condition of the discrete phase at walls is assigned as “reflect”, which means the discrete phase elastically rebound off once reaching the wall. At the outlet, the discrete phase simply escapes/exits the computational domain. The gasifier is operating at 24 atm. P = 24 atm

4.1 Baseline Case (Case 1) • • • •

The baseline case (Case 1) is the two-stage oxygen-blown operation with coal slurry distribution of 75%-25% between the first and the second stages. Gas temperature and species mole fraction distributions on the horizontal and center vertical planes in the gasifier are shown in Fig. 4. The gas temperature is seen higher in the region between the first stage and second stage injection locations than in the region above the second stage injection location. Maximum gas temperature in the first stage reaches 2400K (3860°F). The dominant reaction in the first stage is the intense char combustion (C + ½ O2 → CO and CO + ½ O2 → CO2) in the first stage and gasification reactions (mainly char-CO2 gasification, C + ½ CO2 → CO) in the second stage. Oxygen is completely depleted through the char combustion in the first stage. Char gasification is enhanced in the second stage with the injection of the remaining coal without oxygen. Char is gasified with CO2 produced in the first stage through reaction C + CO2 → CO and with H2O through reaction C + H2O → CO + H2.

Pressure: 24 atm No slip condition at wall Adiabatic walls Inlet turbulence intensity 10%

2nd stage Coal: 5.47 kg/s, 300 K

1st stage Oxidant: 9.92 kg/s, 425 K Coal: 16.43 kg/s, 300K

Fig. 3 Boundary conditions of the baseline case of the two-stage entrained-flow gasifier. Table 3 Parameter and operating conditions of the studied cases. The changed parameters are highlighted.

Parameters Type Fuel Oxidant Stage Distribution Fuel Oxidant Mass flow rate Fuel (kg/s) Oxidant (kg/s)

1

Case 1

Case 2

Case 3

Case 4

Case 5

Slurry Oxygen

Dry Oxygen

Slurry Oxygen

Slurry Oxygen

Slurry Air

2

75% 100%

25% 0%

16.43 9.92

5.47 0.00

Total

1 75% 100%

21.9 9.92

9.47 9.92

2

Total

1 50% 100%

25% 0%

2

Total

50% 0%

3.16 12.90 10.95 10.95 21.90 0.00 9.92 9.92 0.00 9.92

* Oxidant for Case 5 is air (78% N2, 22% O2).

5

1

2

Total

1

100% 100%

0% 0%

21.90 9.92

0.00 21.90 8.12 0.00 9.92 21.00

75% 100%

2

Total

25% 0% 2.70 10.82 0.00 21.00

Fig. 4 Gas temperature and species mole fraction distributions for Case 1 (2-stage, 75%-25%, coal slurry, oxygen-blown).

Mass-weighted averages of gas temperature and species mole fractions along the gasifier height for Case 1 are shown in Fig. 5. The dips in the graphs occur at the injector elevations at heights of 0.75 m for the first stage and 3 m for the second stage. The CO2 mole fraction and the gas temperature decrease from roughly 27% to roughly 19% as the gas flows from the first stage to the second stage. On the other hand, CO mole fraction increases from 12% to 20%, due to the endothermic char-CO2 (R1.2) gasification mentioned above. Meanwhile, the increase in the average mole fraction of H2 in the second stage is negligible. This may indicate that char-CO2 gasification is more dominant than char-H2O gasification in the second stage. 0.4

1st stage fuel injection 2nd stage fuel injection

significant change in the syngas temperature and compositions. The significant temperature drop from roughly 1900K (2960°F) to 1500K (2240°F) across the second stage clearly indicates the advantage of injecting only coal at the second stage to protect the refractory liner and reduce the maintenance cost. Fig. 6a shows helical flow pathlines inside the gasifier; the helical pattern lengthens the flow residence time to allow more time for the reactions to take place. Velocity vectors on vertical midplane and horizontal injection levels are presented in Fig. 7. Due to the vortex generated by the tangential fuel injections in the first stage, strong upward flow occurs near the wall, and weak downward flow occurs in the center. The central core near the second stage exhibits an almost stagnant region due to the opposing fuel injections at the second stage. The flow below the first stage injection level is weak, which could result in some gas being trapped. The momentum of each jet in the second stage is not strong enough to reach the center, and hence the jets are swept upward by the strong main flow from the first stage. Figure 6b shows the coal particle distribution. The particles injected in the first stage are depleted fairly quickly, while those injected in the second stage are depleted at a slow rate.

2000

Gas mole fraction

0.3 H2O

CO

1800

1600

0.2 CO2 1400

Gas temperature (K)

H2

0.1 T O2

Exit syngas temperature and mole fraction compositions are listed in Table 4. Carbon conversion efficiency is 99.4%, which is the comparison of the total mass of carbon injected into the gasifier to the total mass of carbon at the gasifier exit. The high heating value (HHV) of the exit syngas is 8.24 MJ/kg.

1200

CH4

0

1000 0

1

2

3

4

5

6

7

8

9

Gasifier height (m)

Fig. 5 Mass-weighted averages of gas temperature and species mole fraction distributions along gasifier height for Case 1 (2-stage, 75%-25%, coal slurry, oxygen-blown).

At the gasifier height of 8.5 m, the graphs for the average gas temperature and gas mole fractions flatten out. This indicates that the rates of reactions are slowing down. Making the gasifier longer or higher would probably not make

6

The distribution of gas temperature presented in Fig. 8 shows that the local highest temperature in the first stage is approximately 3200 K (5300°F), which is 800 K (1440°F) higher than the coal slurry case (Case 1). Unlike the coal slurry case, the dry coal case does not have a lot of H2O to absorb the heat released by the char combustion, nor does much water react with char through the char-H2O gasification. H2O presented in Case 2 comes from the moisture contained in the coal, while H2O in Case 1 comes from both the moisture contained in the coal and water added to the coal to make the slurry. This higher gas temperature means that the fuel injectors and refractory liner in the first stage will experience more severe thermal loading and maintenance issues than in the coal slurry operation. As seen in Fig. 9, the average CO mole fraction in the first stage is slightly higher than in the coal slurry case (Case 1), approximately 19% versus 12%. The same is observed for the average CO2 and H2 mole fractions, 30% for CO2 and 34% for H2 in the dry coal case compared to 27% for CO2 and 31% for H2 in the coal slurry case.

(a)

Similar to the coal slurry operation in Case 1, char gasification is enhanced in the second stage by injecting the remaining fresh coal. But because the coal injected is dry coal, char gasification that occurs is mainly char-CO2 gasification.

(b)

Fig. 6 (a) Flow pathline colored by the residence time temperature and (b) particle distribution for Case 1. Table 4 Exit syngas temperatures and compositions. Case 1 Fuel distribution Oxidant Fuel type Exit syngas: T (K) T (°F)

Case 2

Case 3

Case 4

Case 5

2-stage 2-stage 2-stage (75%-25%) (75%-25%) (50%-50%)

1-stage

2-stage (75%-25%)

oxygen slurry

oxygen slurry

air slurry

oxygen dry

oxygen slurry

1310 1898

1882 2928

1250 1790

1407 2073

1143 1598

31.7% 20.2% 18.9% 1.2% 26.7% 1.3% 0.0%

33.8% 31.4% 19.0% 1.7% 0.8% 13.3% 0.0%

31.1% 19.7% 19.2% 1.3% 27.4% 1.3% 0.0%

32.2% 21.5% 18.0% 0.7% 26.3% 1.3% 0.0%

19.0% 7.6% 12.5% 0.4% 16.4% 44.1% 0.0%

99.4% 8.24

100.0% 9.45

98.3% 9.03

94.8% 7.68

77.3% 4.40

Mole fraction: H2 CO CO2 CH4 H2O N2 O2 Carbon conversion efficiency HHV at 25°C (MJ/kg)

4.2 Effects of Coal Mixture (Slurry vs. Dry) Case 2 is conducted to investigate the effects of using dry coal as the fuel. Coal and oxidant feed rates are kept the same as for Case 1. Nitrogen is used as the transport gas for the coal powder. The amount of N2 transport gas used is 0.3 times the mass of coal powder. The same fuel and oxidant distributions as in Case 1 are used, which is two-stage operation with 75%25% fuel distribution between the first and second stages and 100% oxidant injected into the first stage with no oxidant injection at the second stage.

Fig. 7 Velocity vectors and temperature field on the center vertical plane and injection planes for Case 1.

7

Fig. 8 Gas temperature and species mole fraction distributions for Case 2 (2-stage, 75%-25%, dry coal, oxygen-blown).

respectively. The syngas HHV of the dry coal case is also higher than the coal slurry case, 9.45 MJ/kg versus 8.24 MJ/kg. Of course, a higher heating value is better. However, in addition to potential increased maintenance issue related to fuel injectors and refractory liner, the higher syngas temperature of the dry coal case means that thermal efficiency will reduce when the syngas temperature is cooled down to the acceptable level for operating the downstream gas clean-up system. Although syngas cooler can transfer the thermal energy of the high raw syngas temperature to high-pressure steam, degradation of the energy quality will inevitably affect the overall plant thermal efficiency.

3000 1st stage fuel injection 2nd stage fuel injection

0.4

0.3

CO

2500

0.2 CO2 2000 0.1

Gas temperature (K)

Gas mole fraction

H2

T O2

4.3 Effects of Fuel Distribution

CH4

H2O 0

1500 0

1

2

3

4

5

6

7

8

9

In the baseline case (Case 1), fuel is distributed by 75%25% between the first and the second stages. Cases 3 & 4 are conducted to study the effects of different fuel distributions. In Case 3, the fuel is evenly distributed between the first and the second stages, i.e. 50%-50%. In Case 4, all (100%) of the fuel is injected in the first stage. In other words, Case 4 simulates the one-stage operation of the gasifier. The same total feed rate of coal slurry and oxidant in Case 1 is used in Cases 3 & 4. As in Case 1, all of the oxidant is injected in the first stage.

Gasifier height (m)

Fig. 9 Mass-weighted averages of gas temperature and species mole fraction distributions along gasifier height for Case 2 (2-stage, 75%-25%, dry coal, oxygen-blown).

Both Figs. 8 and 9 show a significant increase in CO (from approximately 19% to 31%) and decrease in CO2 (from approximately 30% to 19%) in the second stage, due to the charCO2 gasification. Minor char-H2O reaction also occurs in the second stage. The small decrease in H2 in the second stage is due to dilution by the additional coal.

Fig. 10 presents the comparison of average gas temperature and species mole fractions for Cases 1, 3, and 4. Higher massweighted average gas temperature 2500 K (4040°F) occurs in the first stage for Case 3 (50%-50%) compared to 1900 K (2960°F) of Case 1 (75%-25%) and is due to the higher O2/char ratio in the first stage for Case 3. Higher O2/char causes more char to burn, resulting in a higher average gas temperature. However, counter-intuitively, lower O2/char ratio in Case 4 (100%-0%) in the 1st stage also produces higher average gas temperature than Case 1. A plausible explanation would be that the higher temperature in Case 3 is not actually caused by rich combustion as first thought, but it is caused by less water

The average temperature of the exit syngas listed in Table 4 is 1882 K (2928°F), which is 572 K (1030°F) higher than the syngas for the coal slurry case (Case 1), due to lack of steam in the dry coal operation. Compared to the coal slurry case, there is less H2O to absorb the heat from the char combustion and less H2O to react with C through the endothermic char-H2O reaction. H2 and CO2 contents of the syngas are higher than those of the coal slurry case, 33.8% and 31.4% versus 31.7% and 20.2%,

8

turbine combustor without going through the gas cleanup. In reality, the sensible heat will be used to produce steam to produce power through the steam turbine because the syngas temperature needs to be reduced for cleaning and desulfurization.

presence, and hence, less heat capacity to absorb heat generated by combustion. This explanation can be supported by the high oxygen and CO2 concentrations but low CO and H2 concentration in the first stage of Case 3 shown in Fig. 10. This means combustion in the 1st stage in Case 3 is complete (i.e. high CO2) but the gasification process is less productive (i.e. low CO and H2). On the other hand, in Case 4 when 100% coal is injected in the 1st stage, oxygen is quickly consumed (i.e. low O2) to produce CO with high temperature. The relatively lower average gas temperature in the injector area for Case 1 (75%25%) has the benefit of helping prolong the life of fuel injectors and refractory liners.

Based on the syngas temperature and composition, the 50%-50% fuel distribution (Case 3) gives the best result. It has the highest syngas HHV (9.03 MJ/kg) even though its carbon conversion efficiency (98.3%) is slightly lower than that of the 75%-25% case (Case 1 with carbon conversion efficiency of 99.4%). Besides the highest syngas HHV, Case 3 has the lowest syngas temperature (1250 K, 1790°F). This lowest syngas temperature compared to the other cases means that there will be less energy loss during the syngas clean-up process. However, its mass-weighted average of gas temperature (2500 K, 4040°F) in the first stage is highest compared to those of the other cases, 1900 K (2960°F) for Case 1 and 1500 K (2240°F) for Case 4. This high gas temperature will put the gasifier's fuel injectors and wall refractory bricks in a higher thermal loading; consequently, they will be more prone to failure and require more maintenance.

The graph of O2 mole fraction for Case 3 shows that a little amount of O2 still exists when the gas reaches the second stage injection level. This means that char has a good opportunity to react with the abundant O2 at the first stage. Meanwhile, for Case 1 (75%-25%) and Case 4 (100%-0%), O2 is quickly completely depleted in the first stage. The comparison of CO and CO2 mole fractions for all three cases confirms that char combustion is more intense in Case 3. Case 3 has the highest CO2 mole fraction and the lowest CO mole fraction in the first stage. It implies that a large amount of char in the first stage goes through complete combustion. Case 4 (100%-0%), which has the lowest O2/char ratio in the first stage, has the lowest CO2 mole fraction and the highest CO mole fraction.

Velocity vectors on vertical midplane and horizontal injection levels for Case 3 are presented in Fig. 10. With 50% of the fuel injected in the second stage, the fuel jests are stronger than in Case 1 (Fig. 7) and are able to penetrate deeper to the center crossing the upcoming flow from the first stage.

The exit syngas composition listed in Table 4 indicates that among the three cases, Case 4 (100%-0%) yields the highest H2 production – 32.2% compared to 31.7% for Case 1 (75%-25%) and 31.1% for Case 3 (50%-50%). Case 4 also has the highest CO production – 21.5% compared to 20.2% for Case 1 and 19.7% for Case 3. However, Case 4 has the highest exit syngas temperature at 1407 K. Syngas temperature for Cases 1 and 3 are 1310 K and 1250 K, respectively.

4.4 Effects of Oxidant (Oxygen-Blown vs. Air-Blown) Case 5 simulates the air-blown two-stage operation of the gasifier. Air with composition of 22% O2 and 78% N2 by weight is used as the oxidant. The O2/C mole ratio is maintained the same as in Case 1 (oxygen-blown) which is 0.4. Total feed rate of coal and oxidant combined is the same as for Case 1. Similar to Case 1, the fuel is distributed 75% and 25% between the first and the second stages.

Even though Case 4 has the highest H2, CO and CH4 combined, its syngas high heating value is the lowest among three cases. Case 4’s HHV is 7.68 MJ/kg, compared to 8.24 MJ/kg for Case 1 and 9.03 MJ/kg for Case 3. This is due to the lower carbon conversion efficiency of Case 4 (94.8%) compared to the other two cases (99.4% for Case 1 and 98.3% for Case 3). The exit syngas of Case 4 contains the most unreacted char. Thus, combined with its high temperature, it has the lowest HHV. Note that when the syngas exit temperature is high more chemical energy has been converted to the sensible heat of the syngas and less chemical energy is reserved in the syngas. This sensible heat could be effectively used in the gas turbine combustor if the syngas could be fed directly into the gas

As expected, the mass-weighted average of gas temperature in the first stage shown in Fig. 12 is lower than in Case 1 (oxygen-blown) due to the abundance of N2 as a diluent in the air-blown case. The maximum cross-sectional mass weighted average gas temperature is approximately 1450 K (2150°F), while the maximum average gas temperature in the oxygenblown case is 2000 K (3140°F).

9

0.4

1st stage fuel injection

2900

1st stage fuel injection 2nd stage fuel injection

2400

H2 mole fraction

Gas temperature (K)

2nd stage fuel injection

1900

1400 Case 3 (50%-50%)

900

0.3

Case 3 (50%-50%)

0.2

Case 1 (75%-25%)

Case 1 (75%-25%)

Case 4 (100%-0%)

Case 4 (100%-0%)

400

0.1

0

1

2

3

4

5

6

7

8

9

0

1

2

Gasifier height (m) 0.3

3

4

0.4

1st stage fuel injection

7

8

9

CO2 mole fraction

2nd stage fuel injection

0.2

0.1 Case 3 (50%-50%) Case 1 (75%-25%)

0.3

0.2

Case 3 (50%-50%)

0.1

Case 1 (75%-25%)

Case 4 (100%-0%)

Case 4 (100%-0%)

0

0

0

1

2

3

4

5

6

7

8

9

0

1

2

0.15

1st stage fuel injection

O2 mole fraction

0.3

Case 3 (50%-50%)

0.2

4

5

6

7

8

9

1st stage fuel injection 2nd stage fuel injection

2nd stage fuel injection

0.4

3

Gasifier height (m)

Gasifier height (m)

H2O mole fraction

6

1st stage fuel injection

2nd stage fuel injection

CO mole fraction

5

Gasifier height (m)

0.1

0.05

Case 3 (50%-50%) Case 1 (75%-25%)

Case 1 (75%-25%) Case 4 (100%-0%)

Case 4 (100%-0%)

0

0.1 0

1

2

3

4

5

6

7

8

0

9

1

2

3

4

5

6

7

8

9

Gasifier height (m)

Gasifier height (m)

Fig. 10 Mass-weighted averages of gas temperature and species mole fraction distributions along gasifier height for Cases 1, 3 and 4. 0.6

1500

1st stage fuel injection 2nd stage fuel injection

0.5

Gas mole fraction

0.4 1300 0.3 1200

T

H2O

0.2

H2

1100

CO2

0.1

Gas temperature (K)

1400

N2

CO

O2

CH4

0

1000 0

1

2

3

4

5

6

7

8

9

Gasifier height (m)

Fig. 12 Mass-weighted averages of gas temperature and species mole fraction distributions along gasifier height for Case 5 (2-stage, 75%-25%, coal slurry, air-blown). Fig. 11 Velocity vectors and temperature field on the center vertical plane and injection planes for Case 3 (50%-50%).

10

through the char-H2O gasification. This higher gas temperature means that the fuel injectors and refractory walls in the first stage will experience higher thermal loading than in the coal slurry operation. The syngas HHV of the dry coal case is also higher than the coal slurry case -- 9.45 MJ/kg vs. 8.24 MJ/kg. However, the higher syngas temperature of the dry coal case would result in a lower plant thermal efficiency because it needs to be cooled before it goes through the gas clean-up system downstream of the gasifier. Consequently, a lot of energy will be downgraded (i.e. loss of exergy) via waste heat exchanger even though part of the energy can be recovered to produce superheated steam to generate electricity through the steam turbine.

Table 5 Comparison of exit syngas temperature and composition between Cases 1 and 5 after N2 is removed from the syngas.

Fuel distribution Oxidant Fuel type Exit syngas: T (K)

Case 1

Case 5

2-stage (75%-25%) oxygen slurry

2-stage (75%-25%) air slurry

1310

1143

H2 CO CO2 CH4 H2O

32.1% 20.5% 19.1% 1.2% 27.1%

34.0% 13.6% 22.4% 0.7% 29.3%

O2 Carbon conversion efficiency HHV at 25°C (MJ/kg)

0.0%

0.0%

99.4% 8.25

77.3% 7.26

Mole fraction:

Effects of Fuel Distribution between Two Stages Due to less water to absorb heat, reducing the fuel feed in the first stage does result in higher gas temperatures in the first stage. One-stage operation yields higher H2, CO and CH4 combined than if a two-stage operation is used but with a lower syngas heating value. The 50%-50% fuel distribution case yields the highest syngas HHV and lowest syngas exit temperature among the studied cases. The exit syngas of onestage operation contains the most unreacted char, combined with its high exit temperature, results in the lowest heating value.

The syngas composition listed in Table 4 shows that the mole fraction ratio of CO/H2 is 0.4 for the air-blown case (Case 5), which is much lower than those of the oxygen-blown case (Case 1). The syngas HHV for Case 5 is approximately only half of Case 1, 4.40 MJ/kg vs. 8.24 MJ/kg. The syngas of Case 5 is diluted with N2, which causes this low heating value. However, its low carbon conversion efficiency at 77.3% also contributes to this low syngas heating value. Low carbon conversion efficiency is due to the lower overall gas temperature inside the gasifier, where less energy is available to drive the endothermic gasification reactions.

Effects of Oxidant (Oxygen-Blown vs. Air-Blown) Gas temperature inside the gasifier for the air-blown case is lower than in the oxygen-blown gasifier due to the abundant presence of N2. Lower than the oxygen-blown case (99.4%), the carbon conversion efficiency of the air-blown case is 77.3%. The syngas heating value for the air-blown case is 4.40 MJ/kg, which is almost half of the heating value of the oxygen-blown case (8.24 MJ/kg). Even when N2 is removed for comparison, the HHV of the air-blown case is still about 1 MJ/kg less than the oxygen-blown case.

To give a fair comparison between the syngas in Cases 1 and 5, syngas compositions and heating values for both cases are recalculated after the N2 contained in the syngas are removed. The recalculated compositions are compared in Table 5. The mole fraction of H2 (34.0%) for the air-blown case (Case 5) becomes slightly higher than the oxygen-blown (Case 1, 32.1%), but the CO mole fraction for the air-blown (13.6%) is 6.5 percentage points lower than the oxygen-blown case. As expected, the heating value of the syngas increases from 4.40 MJ/kg to 7.26 MJ/kg after N2 is removed. Nonetheless, this recalculated syngas heating value is still lower by roughly 1 MJ/kg than that of the oxygen-blown case (8.25 MJ/kg) even after N2 is removed.

6.0 ACKNOWLEDGEMENTS This study was supported by the Department of Energy contract No. DE-FC26-08NT01922 and the Louisiana Governor's Energy Initiative via the Clean Power and Energy Research Consortium (CPERC), administered by the Louisiana Board of Regents. 7.0 REFERENCES

5.0 CONCLUSIONS

Bockelie, M.J., Denison, K.K., Chen, Z., Linjewile, T., Senior, C.L., and Sarofim, A.F., CFD Modeling For Entrained Flow Gasifiers in Vision 21 Systems, Proceedings of the 19th Annual International Pittsburgh Coal Conference, Pittsburgh, PA. September 24-26, 2002.

Five cases of different operating conditions are simulated and the results show: Effects of Coal Mixture (Slurry vs. Dry)

Chen, C., Horio, M., and Kojima, T., Numerical Simulation of Entrained Flow Coal Gasifiers, Chemical Engineering Science, 55, 3861-3833, 2000.

The temperature in the first stage for the dry-fed case is approximately 2800 K (4580°F), which is 400 K (720°F) higher than the slurry-fed case. Unlike the slurry-fed case, the dry-fed case does not have a lot of H2O to absorb the heat released by the char combustion, nor does much steam react with char

Fletcher, T.H., and Kerstein, A.R., Pugmire, R.J., Grant, D.M., Chemical Percolation Model for Devolatilization: 2.

11

Temperature and Heating Rate Effects on Product Yields, Energy and Fuels, 4, 54, 1990. Fletcher, T.H., and Kerstein, A.R., Chemical Percolation Model for Devolatilization: 3. Direct Use of 13C NMR Date to Predict Effects of Coal Type, Energy and Fuels, 6, 414, 1992. Grant, D.M., Pugmire, R.J., Fletcher, T.H., and Kerstein, A.R., Chemical Percolation of Coal Devolatilization Using Percolation Lattice Statistics, Energy and Fuels, 3, 175, 1989. Guenther, C., and Zitney, S.E., Gasification CFD Modeling for Advanced Power Plant Simulation, Proceedings of the 22th International Pittsburgh Coal Conference, Pittsburgh, Pennsylvania, September 12-15, 2005. Jones, W.P., and Lindstedt, R.P., Global Reaction Schemes for Hydrocarbon Combustion, Combustion and Flame,73,233,1998. Kobayashi, H., Howard, J. B., and Sarofim, A. F., Coal Devolatilization at High Temperatures, 16th Symp. (Int'l.) on Combustion, The Combustion Institute, 1976. Silaen, A. and Wang, T., Effects of Fuel Injection Angles on Performance of A Two-Stage Coal Gasifier, Proceedings of the 23rd Pittsburgh Coal Conference, Pittsburgh, Pennsylvania, Sept. 25-28, 2006. Silaen, A. and Wang, T., Comparison of Instantaneous, Equilibrium and Finite Rate Gasification Models in an Entrained Flow Coal Gasifier, Paper 14-3, Proceedings of the 26th International Pittsburgh Coal Conference, Pittsburgh, Pennsylvania, Sept. 21-24, 2009. Silaen, A. and Wang, T., Effect of Turbulence and Devolatilization Models on Gasification Simulation, International Journal of Heat and Mass Transfer, 53, 2074-2091, 2010. Smoot, D.L., and Smith, P.J., Coal Combustion and Gasification, Plenum Press, 1985.

12

Manuscript Not AVAILABLE

2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission

PROGRAM TOPIC: GASIFICATION: GASIFICATION SCIENCE AND MODELING EXACT TITLE OF PAPER: START-UP BEHAVIOR OF A FIXED BED GASIFIER: ONE DIMENSIONAL MODELING Giampaolo Mura, Prof., University of Cagliari - Department of Chemical Engineering and Material Science [email protected]] Mariarosa Brundu, PhD, University of Cagliari - Department of Chemical Engineering and Material Science [email protected]

Abstract: This work copes with the development of a mathematical model for the investigation of the transient behavior of a countercurrent fixed bed gasifier. The phenomenological model is based on heat and mass continuity equations. Heterogeneousness is somehow considered by the insertion of two separated heat balances, one for the gas and one for the solid phase. All the main phenomena involved in the gasification process are inserted: drying, pyrolysis, gasification and combustion reactions of the solid phase and homogeneous gas phase reactions including secondary pyrolysis reactions. The system is described with a pseudo homogeneous approach. Moisture loss calculation is carried out by the introduction of a first order kinetic on the moisture content of the bed; a competitive reaction model is used for primary pyrolysis; heterogeneousness of the system is considered for gas phase reactions by the introduction of a shrinking core reaction model where the external diffusion and the kinetic resistance are considered. The solid phase is constituted by four pseudo components: coal, ash, char and moisture. The ash behavior is described by the introduction of a shell progressive model with variable particle diameter. Gas species considered by the model are: CO, CO2, H2, H2O, CH4 and tar. Input for the model are flowrates, temperature and composition profiles at the initial state for both: the gas and the solid phase. The model was used to study the start up of an air blown atmospheric gasifier in the case of a Pittsburg n°8 coal seam feedstock. The initial conditions chosen for the dynamic simulation are in accordance with the start up procedure of an existing gasification pilot plant. Output for the model are the variation with time of temperature, composition and spatial velocity profiles of the system. In particular the dynamics is analyzed with reference to the bed behaviour in a long time investigation. Scarce information about this topic was before present in the literature. The influence of steam injection also revealed the presence of multiple steady states for this system.

2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11–14, 2010 PROGRAM TOPIC: GASIFICATION

Entrained Flow Coal Gasification: Modeling, Simulation & Experimental Uncertainty Quantification For a Laboratory Reactor Institute for Clean and Secure Energy, The University of Utah Philip J. Smith, professor & director ([email protected]) Charles Reid, graduate student ([email protected]) Julen Pedel, graduate student ([email protected]) Jeremy Thornock, asst. research professor ([email protected])

Abstract: Modeling and simulation on petascale computing platforms offers unprecedented opportunities to explore new transformative technology options in entrained-flow coal gasification. However, before these advanced simulation tools can be used with confidence, formal validation and uncertainty quantification is required. In this paper we explore the uncertainty in predictions from large eddy simulations (LES) of the Brigham Young University (BYU) pilot-scale entrained flow reactor1 . The specific focus of this paper is to produce predictive capability for gasification with quantified uncertainty bounds through a formal validation and uncertainty quantification (V/UQ) analysis. We have employed the Data Collaboration methods of Michael Frenklach and coworkers at the University of California-Berkeley in our V/UQ analysis 2,3. Data Collaboration requires consistency between simulation results and experimental data. Having developed the ARCHES code, that combines LES with the Direct Quadrature Method of Moments (DQMOM) for the gasification simulation, we now perform V/UQ employing the BYU experimental data. The simulation produces temporally and spatially resolved data of the reacting, multiphase flow field, including the moment description and evolution of the full particle number density function. Specifically, we have studied three key particle behaviors; particle size segregation (Stokes number effects), particle clustering, and particle pyrolysis. We perform the V/UQ analysis with the spatially resolved compositional data measured in the BYU gasifier. In this study we have extracted experimental error for each of the measurements taken. Used prior ARCHES validation information to produce prior uncertainty bonds on the most sensitive simulation parameters. We have included uncertainties in numerical parameters, models, and scenario parameters. The resulting V/UQ analysis produced posterior uncertainty bounds on both the quantities of interest and the uncertain parameter space studied.

Introduction: Utilization of domestic sources of fuel such as coal is growing increasingly important, but under increasing scrutiny. There is a critical need to retrofit existing industrial and applied-scale facilities that use coal to meet regulatory requirements, and to explore new technologies and techniques for carbon emission reduction. With the continually increasing power of computational resources, predictive simulation tools are playing an increasingly important role in this process. As computing moves from peta- to exa-scale, simulation codes and models that can utilize these resources are needed. The large-eddy simulation tool Arches is one such simulation tool. In the current work, a dispersed-phase model is implemented in Arches, and preliminary results are shown. A procedure for implementing the direct quadrature method of moments in a large-eddy simulation code, and applying the method to a complex, multiphysics problem (coal gasification), has been developed. Simulations of the gasifier of Brown et al1 are performed. The gasifier is , and simulations are performed using a grid, with a resolution of . A two-step Kobayashi-Sarofim devolatilization model4 is used, with rate constants given by them. To model the particle-laden gasifier, the direct quadrature method of moments is coupled with large eddy simulation. Recent developments in the direct quadrature method of moments have resolved issues encountered with DQMOM when incorporating complex gasification physics. The DQMOM formulation is explained, the technique used to address the numerical issues in DQMOM is explained, and simulation results are compared to results from Brown et al.

Large Eddy Simulations: Simulation of turbulent flows, particularly large-scale and high Reynolds number flows, is a difficult and expensive problem. The cost of computations in which all length and time scales are resolved scales as Re11/4. In addition, that cost increases for complex geometry. Resolving all scales of the turbulent flow in a direct numerical simulation (DNS) has not yet been attempted for these multiphysics problems. In order to move computations to the scale of applied and industrial systems, it is necessary to model some or all scales of the flow. Most coal combustion simulations have resorted to a Reynolds-Averaged Navier Stokes (RANS) analysis to circumvent the massive computational burden. In contrast to RANS, the large eddy simulation (LES) technique resolves the largest scales of the flow up to a cutoff length ∆, but models the smallest scales of the flow. This is done by applying a low-pass filter to the velocity field to reduce high-frequency processes, allowing for a balance between computational cost and increased accuracy. The smallest (filtered) scales must then be modeled. However, the modeling is simplified over RANS turbulence models because, as stated by the Kolmogorov Hypothesis, at high enough wavenumbers, the turbulence becomes statistically homogeneous, meaning the turbulence at these scales can be modeled without introducing significant modeling error. Almost 99% of the computational cost of DNS is spent on resolving the smallest 1% of flow scales, so LES can substantially reduce computational costs by eliminating the cost of resolving turbulence at small scales and using a simplified model instead. For complex multiphysics problems like coal gasification, there are a wide range of physical processes happening. Each of these have a characteristic length (l) or wavenumber (k). If the wavenumber associated with a particular physical process is greater than the Nyquist limit, kN = 1/∆, then it falls in the resolved spectra. Thus, it is essentially DNS, for that process, on the LES grid. In the case of chemical reactions, if the Damköhler number is small enough, the reactions are sufficiently slow to be completely resolved. Likewise, particle motion can be completely resolved if the particles are ballistic, and the Stokes number of the particles is large enough, because small (unresolved) fluctuations in velocity will have no affect on particle momentum. Our analysis shows that for the entrained flow gasification cases we have studied, the Stokes Number of the essentially all of the pertinent coal particles and the Damköhler number of the coal particle devolatilization and char oxidation processes are all above the Nyquist limit for the simulations that we can perform on reasonable computer clusters. Thus, the coal process are fully resolved (ie. essentially DNS). Of course, there are still gas phase turbulent mixing and reaction processes that are not resolved and must be modeled in traditional LES

fashion. For this reason, large eddy simulation holds great promise for pulverized coal combustion and entrained-flow gasification applications with multiple, coupled physical processes.

Coal Particle Approach: � � jet,� � Nξ Nξ 3 Number Density Function. 3 In the distribution of particles can∂ be�described � � � ∂f ∂ ∂ a particle-laden ∂f ∂ ∂f using a number + density((NDF), f )− denoted D Gj,f >and f )+ +h(ξ, x, t)The (4) f(x,t), a function(< of space time and hasΓunits of [#/volume]. xi which=is− ξj ∂t i=1NDF ∂xiis a function of a i=1 ∂x ∂x ∂ξ ∂ξ ∂ξ i i j j j set of independent variables j=1for particles. This vector j=1of internal coordinates, denoted ξ;

may include temperature, composition, or size of particles. In this way the NDF can describe the distribution of

whichthe represents the spatial and with temporal evolution of ξ.the NDF,density including all particle number density of particles particular values of Thethe number is written as f(ξ; x,phenomena t): drag force, particle size change, breakage, etc.) Nξ � �∂f (ξ;� �� Nξ Nξ 3 x, t) ∂ ∂� � � ∂ ∂f ∂+ ∂ ∂f (�u (ξ; x, t))+ Γξ (�G (ξ;(4) x, t)) = h (ξ; x, t) (1) i |ξ�f j |ξ�f Dxi = −∂t (< f )+ +h(ξ, x, t) j,f > j ∂ξ Quadrature Approximation: Direct Quadrature Method of Moments ∂xG i j ∂x ∂x ∂ξ ∂ξ ∂ξ i i j j j j=1 =1 j=1 j=1

n the Direct Quadrature Method NDF is approximated by � � the � � a sum � of delta functions, as dx N Niξ ξ 3 3 of Moments, � � � is the ξ-space convection term j = dξ j /dt. Each is∂term conditioned on the ,value ξ because the values of j∂f ial and Gtemporal the the including all particle phenomena ∂evolution ∂ GNDF, ∂f ∂ andofG ∂f where ui>isfof)− the x-space convection ui >=fof thex, ξ-space j is +model (< u D = − (< G )+ Γ +h(ξ, existing t) (4) Lagrangian hown onthese Figure 1. The integral of the NDF becomes a summation a set of N weighted abscissas: i,fchanges for different xi j,f values ξj dt terms particles and different of ξ. Because many ∂t etc.) ∂xi ∂xi ∂x ∂ξj ∂ξj ∂ξj ge, breakage, i=1 j=1 j=1 dξji derivatives, modelsi=1 give expressions for these temporal existing Lagrangian models can be “plugged in” to the convection term Gjˆ = . Each � is N conditioned on the value of ξ because NDFwhich transport equation framework, anddttherefore can Eulerian well. represents the spatial and temporal evolution of be theextended the NDF, to including allmodels particleas phenomena w(x)g(x)dx wα < >α (5) the values of these model terms ≈changes forg different particles and different (drag force, particle size change, breakage, etc.) ximation: Direct Quadrature Method of Moments Direct Quadrature Method of Moments. In the Direct Quadrature Method of Moments, the NDF is approximated α=1 values of ξ. Because many existing Lagrangian models give expressions for these by a sum of delta functions, as shown on temporal derivatives, existing models can beas“plugged in” to the thod Moments, NDF iswith approximated byLagrangian a sum of delta functions, For of aFigure multivaritae NDF, i.e. multiple interQuadrature Approximation: Direct Quadrature Method of Moments 1. The the integral of the NDF becomes a NDF transport equation and therefore can be extended to Eulerian ralcoordinates, of the NDFthe becomes aof summation ofas: aframework, set of N weighted abscissas: summation ofNDF a set N be weighted abscissas: al can written In the Direct Quadrature Method of Moments, the NDF is approximated by a sum of delta functions, as models as well. shown on Figure 1. TheNintegral of the NDF becomes a summation of a set of N weighted abscissas: ˆ �   w(x)g(x)dx ≈ w < g > (5) N ˆ ξ N α α N � � 3.2 Eulerian � � � vs. Lagrangian wα (x, t) α=1δ ξj − �ξj (x,w(x)g(x)dx wα < g >α (5) f (ξ; x, t) ≈ t)�  ≈ α

α=1

j=1 e. with multiple The inter-most common and widely-used models for the dispersed phase are LaFor For a multivaritae NDF, i.e.i.e. with multiple intera multivaritae NDF, with multiple inter(6)track a large number of characteristic particles models. These models be written as: grangian the NDF can be written as: nal coordinates, the NDF can be written as: wherenal Ncoordinates, is the number of internal coordinates ξ as they move through the domain and interact with the fluid; an ordinary difnd wα is called a ferential weight and represents the form number in the of    equation N N associated particles per volume with the abscissa � � � � � � x,t)� t) ≈  wα (x, t) δ ξj − �ξj (x, t)�α  .δ ξj − �ξf j(ξ; (x, α α=1 dx DQMOM solves transportj=1equations for the = Sx (2) (6) dt where Nξ is number of internal coordinates eights and abscissas of the quadrature approxima(6) where Nξthe is the number of internal coordinates and w is a weight and represents the number α on. These transport equations are simply scalar dξ = Sξ andcoordinates wα is called a weight and represents the number of internal (3) ofofequations, particles per volume volume associated with the dt particles per associated with the abscissa ansport and are easily implemented represents the number abscissa α. code.α. The transport equations for the to a with CFD is solved for each particle, and information from all particles is then collapsed ated the abscissa DQMOM solves transport equations for the eights w weighted abscissas ςj,α = w FigureThese 1: DQMOM approximation NDF with 3 DQMOM solves transport for the onto the Eulerian grid to yield the methods are describedofinthe greater α and and α �ξj � α NDF. weights abscissas of the equations quadrature approximaweights and abscissas of the quadrature an be expressed as: quadrature nodes rt equations fordetail the in [11, 10]. are simply scalar tion. These transport equations approximation. Theseand transport equations are transport equations, are easily implemented uadrature approximaAlternatively, Eulerian methods are based on solving scalar transport equa∂ ∂ simply scalarcode. transport equations, and areforeasily into a CFD The transport equations the ons are simply scalar for (wwαtions ) +intoweighted (ustatistical ) = ainformation (7)(moments) about the solid phase. Moments of i,f w α implemented the LES code. αςj,α = wα �ξj �α Figure 1: DQMOM approximation of the NDF with 3 weights abscissas α and ∂t ∂x i e easilycan implemented an NDF be expressed as: are defined as quadrature nodes port equations for the ∂ ∂ ´ ∞ ∂ ∂ (u(7) , jthe = t) 1NDF .dξ . . Nwith (8) i,f ςjα ) = ξ scissas ςj,α = wα �ξ �αα ) +Figure ξ kbjα f of (ξ; x, 1:w(ς DQMOM 3 (ui,f )jα =)a+ j(w α α ∂x approximation −∞ ∂t i ∂t ∂xi (4) quadrature nodes mk = ´ ∞ f (ξ; x, t) dξ where ui,f is the exact velocity of∂ the weight w or weighted abscissa ς . The source terms a −∞ α α α and bj,α ∂ (ς ) + (u ς ) = b , j = 1 . . . N (8) jα i,f jα jα ξ re obtained by solving a linear system For more details on the full DQMOM derivation, see ∂t of equations. ∂x (unless (7) k = 0, in which casei there is no denominator) and denote the expected f wα ) = aα ]. where uvalue velocity of the weight abscissa ςαand . The source terms aα and bj,α of exact ξ over the entire NDF.wThe most common least expensive Eulerian i,f is the α or weighted The LES simulations of particle-laden coaxial jet presented in this work were done using four internal are obtained by solving a linear system of equations. For more details on the full DQMOM derivation, see moment method is the multi-fluid model [5, 14], which treats the solid phase ∂ ∂ [3]. oordinates: the particle d, pjand the three velocity components the particle velocity: upx , upy , upz ., as(u ai,fseparate fluid. method solves transportofequations for the number (ςjα )The + LES ςdiameter = 1This ...N (8) jα ) =ofbjα ξ simulations particle-laden coaxial jet presented in this work were done using four internal ∂xi distribution, nd a∂tmondisperse i.e. one quadrature node. th st α=1

ξ

nto a CFD code. The transport∂equations ∂for the (ςjα ) + α �ξ(u (8) i,f ςjα ) = bjα , j = 1 . . . Nξ weights wα and weighted abscissas ∂t ςj,α = w ∂xi j �α Figure 1: DQMOM approximation of the NDF with 3 can be expressed as: quadrature nodes where ui,f is the exact velocity of the weight wα or weighted abscissa ςα . The source terms aα and bj,α re obtained∂by solving∂a linear system of equations. For more details on the full DQMOM derivation, see (wα ) equations + (ufor = aα wα and(7) i,f w α )weights The transport the weighted abscissas ςj,α = wα (ξj)α can be expressed as: 3]. ∂t ∂xi The LES simulations of particle-laden coaxial jet presented in this work were done using four internal oordinates: the particle diameter∂dp(ςand the∂three of the particle velocity: upx , upy , u(8) pz ., (ui,fvelocity ςjα ) = bcomponents jα ) + jα , j = 1 . . . Nξ ∂t ∂x i nd a mondisperse distribution, i.e. one quadrature node. where ui,fui,fisisthe velocity of the weight ςα . The source α or weighted where the exact exact velocity of the weight wα or w weighted abscissaabscissa ςα . The source terms aα andterms bj,α areaα and bj,α are obtained bybysolving a linear linearsystem system of equations. For details more on details the fullderivation, DQMOM obtained solving a of equations. For more the fullon DQMOM seederivation, see Particle Velocity 5 3]. Marchisio and Fox . The LES simulations of particle-laden coaxial jet presented in this work were done using four internal Particle flow, Velocity. a gas-solid flow, theisparticle motion is affected by the drag force, which can be described n a gas-solid theIn particle motion affected by the drag force, which can be described the coordinates: particle diameter dp and the three velocity of the mass particle velocity: uby upy ,Stokes upz ., px , the by the the Stokes drag law. For a meso-scale size particle, whencomponents the other additional forces are omitted, drag law. For a meso-scale size particle, when the other additional mass forces are omitted, the momentum and a mondisperse distribution, i.e. one node. momentum equation for the particle canquadrature be expressed by an ordinary differential equation. quation for the particle can be expressed by an ordinary differential equation. � � Particle Velocity dui,p = � fdrag (ui,g − ui,p ) + gi (ρp − ρg ) + Fi,v (9) dt τ ρ m p p p th here i denotes the i direction, g iis the gravity force acting on the particle, Fv are the other body force n a gas-solid flow, the particle motion is affected by the drag force, which can be described by the Stokes cting on thethe particle, u is the gparticle velocity, and facting is the coefficient ofother dragforces force, which drag edrag i denotes ith direction, isg isthe gravity force the Fthe are other bodyhas for where the pith direction, thewhen gravitythe force acting on theon particle, Fforces bodythe acting v are the law. Fori denotes a meso-scale size particle, other additional massparticle, are the momentum vomitted, ose relationship with particle Reynolds number [6]: on the particle, u is the particle velocity, and f is the coefficient of the drag force, which has a close p drag gequation on the particle, iscan thebe velocity, and fondrag is the coefficient of the drag force, which ha for thetheparticle expressed byforce an acting ordinary differential where irelationship denotes ithup direction, gparticle is the number gravity the particle, Fequation. v are the other body forces 6: with particle Reynolds  relationship with particle Reynolds [6]: is the coefficient of the drag force, �number � which has a acting on the particle, up is the particle velocity, and fdrag �  1 Re < 1 du f g (ρ − ρ ) F p  i,p drag i p g i,v close relationship with particle Reynolds = number + (9)  [6]: (ui,g − u0.687 i,p ) + = dt fdrag ρ m 1 < Re < 1000 i 1τ1p + 0.15Rep p p p Rep < 1   1 0.0183ReRep < 1 Re > 1000 0.687 p p 0.687 f =

1 + 0.15Re < Rep < 1000 fdrag = 1 + 0.15Re drag 1p< Rep 1000 p dpp > p− Rep =

.

ρp dp |up − ug | µg Rep = . p − ug | ρp dp |u Re =µg .

n Equation (9), τp is the particle relaxation ptime

µg

In Equation (9), τp is the particle relaxation time

uation (9), τp is the particle relaxation time ρp d2p τp =

τp . =

ρp d2p . 18µg

(10)

18µg 2 ρp dmomentum where τp is the particle relaxation time The conservation law of the gas phase is, p The momentum momentum conservation law of theof gasthe phase The conservation law gasis, phase is, τp = . ∂ρu

18µg

∂ρu ∂t

+ ∇ · ρu = ∇ · D∇u − ∇p + Sdrag he momentum conservation law phase ∂t of the + ∇gas · ρu = ∇ ·is, D∇u − ∇p + Sdrag

(11)

(10

(

(11

where Sdrag is the momentum exchange between the particles and the gas phase.

∂ρu here Sdrag is the momentum exchange particles and+the gas phase. + ∇between · ρu = the ∇ · D∇u − ∇p Sdrag ∂t is the momentum exchange between the particles and the gas phase.

(

e Modeling Sdrag is the momentum exchange between the particles and the gas phase. where Sdrag

Modeling Arches: Experimental setup

The simulation tool used in this work is ARCHES. ARCHES is a three-dimensional, Large Eddy Simulation In this study, applied to by thepersonnel particle-laden coaxial jet thatfor wasClean used by [7] for his experiments. (LES) LES codeisdeveloped within the Institute andBudilarto Secure Energy within the University of A description of the nozzle is low-Mach given in Figure Particles are conveyed with air in to thesimulate primaryheat nozzle Utah. ARCHES uses a number2a. (M < 0.3), variable density formulation and mass whereas the secondary nozzle is fed with air. Air and particles are injected at room temperature in a 18x18 ninch thischamber. study, LES is applied to theon particle-laden coaxial jeton that usedatbythe Budilarto [7]field for his experiment The experiments focus the effects of velocity ratio thewas flow-field near region region of a coaxial velocity is defined ratio of the to central description of jet. theThe nozzle is ratio, givenwhich in Figure 2a.as the Particles areannular conveyed withinlet airvelocity, in the primary nozz U /U = V R, was varied at three different values: 0, 1.0, 1.5 by varying the volumetric flow rate of the a 0 hereas the secondary nozzle is fed with air. Air and particles are injected at room temperature in a 18x1 annular jet. The Reynolds number based on the average central inlet velocity is 8,400. Table 1 lists the flow s study, LES The is applied to the particle-laden coaxial jet that ratio was used byflow-field Budilartoat[7] fornear his experime nch chamber. experiments focus on the effects of velocity on the the region fiel conditions for the experiments.

Experimental odeling setup

perimental setup

transfer in reacting flows 7. The LES algorithm is composed of the numerical differencing scheme and a solution method for solving the filtered, density-weighted, time-dependent coupled conservation equations for mass, momentum, mixture fraction, coal particle moment equation(s), and enthalpy in a Cartesian coordinate system. This set of filtered equations is discretized in space and time and solved on a staggered, finite volume mesh. The staggering scheme consists of four offset grids, one for storing scalar quantities and three for each component of the velocity vector. An explicit time-stepping scheme is used to advance the simulation forward in time. For the spatial discretization of the LES scalar equations, flux limiting and upwind schemes for the convec- tion operator are used to ensure that scalar values remain bounded. For the momentum equation, a central differencing scheme for the convection operator is used. All diffusion terms are computed with a second-order approximation of the gradient. Overall, ARCHES is second-order accurate in space and time. The ARCHES code is massively parallel and highly scalable due to the integration of ARCHES into the Uintah Computational Framework (UCF)8.

Verification, Validation and Uncertainty Quantification: The process of V&V/UQ is to quantify the uncertainty of a predictive simulation tool by requiring a consistent data set as defined by Feely et al.9 for the intended use of the simulation tool. Here, a data set consists of a set of experimentally observed data and simulation data. Within this data set, uncertainty is defined as ym(x)−ye ≤u where ym is the simulation prediction as a function of input parameters x, ye are the experimentally observed data and u is the total uncertainty. The uncertainty is derived from two sources; uncertainty arising from imprecise physical measurements and uncertainty arising from inputs (x) to the model. Experimental uncertainty is composed of bias errors, reproducibility errors, and instrument error. Each experimental observation, ye,i, has lower (le) and upper (ue) uncertainty bounds. The model inputs, x, are classified as one of three categories; 1) model parameters 2) scenario parameters or 3) numerical parameters. Model parameters are generally physical constants that are determined from a priori knowledge (ie, a reaction rate constant.) Scenario parameters consist of initial and boundary conditions (eg, inlet flow rate.) Numerical parameters relate to the discrete representation of the continuous model (eg, spatial discretization scheme.) Each component of x, like the experimental observations, is also associated with an estimated level of uncertainty consisting of upper (β) and lower bounds (α). Given x with its associated uncertainty, results from the simulation tool maps out a region Rs. Any value within Rs is feasible if it satisfies the criteria le ≤ ym(xi) − ye ≤ ue. Values of xi, such that αi ≤ xi ≤ βi satisfying this criteria are feasible parameters because they lie within a region that has been observed experimentally. In this manner, the experimental observations inform the simulation tool and potentially effect the bounds of x. By the same token, we can also examine when individual observations within ye are consistent with each other through the model. For example, consider two experimental observations, ye,1 and ye,2, contained within ye. The experiments ye,1 and ye,2 are said to be consistent with each other if they can both be described with ym given a set of feasible parameters, x. In this manner, the simulation tool can inform the experiment by explaining two independent experimental observations, thus describing consistency between experiments through the model. The process of discovering the consistent data set with the feasible parameter space has been termed data collaboration10 and reduces to a constrained optimization problem. As with most optimization problems, one must perform function evaluations to determine the optimal parameters. This, however, presents a unique challenge for the size and scope of the coal gasification problem that we wish to address here. That is, each function evaluation of our simulation tool ARCHES requires a substantial amount of computational resources. Consequently, we utilize methods to minimize the total function evaluations by choosing them in optimum locations within x and using presumed function descriptions (surrogates) to describe ym(x) . In this manner we attempt to perform as few function evaluations as necessary.

Results: Single Phase Coaxial Jet. In this study, LES is applied to the non-reacting particle-laden coaxial jet that was used by Budilarto11 for his experiments. Simulations were performed for the three velocity ratios cases without particle. The two main objectives are first to validate our modeling approach by ensuring that mean velocities are predicted correctly and secondly to have a reference case to then understand the effect of particles on the gas dynamics. Figure 2(a) (b) (c) are instantaneous profiles of the x-component of the gas velocity. Figure 2(d) compares the simulation results to the experimental data for the mean gas velocity along the centerline.

(a) Gas x-velocity profile (m/s) - VR = 0

(b) Gas x-velocity profile (m/s) - VR = 1.0

1

Ug/U0

0.8

0.6

0.4

VR = 0, LES VR = 0, Exp. VR = 1, LES VR = 1, Exp. VR = 1.5, LES VR = 1.5, Exp.

0.2

0

0

2

4

6

8

10

12

14

16

x/d

(c) Gas x-velocity profile (m/s) - VR = 1.5

(d) Comparison between LES and exprimental data of gas averaged x-velocity along the centerline for different velocity ratios / No particles

Figure 3: Simulation results of gas x-velocity without particles

Velocity profiles are predicted with a maximum error less than 10%. Results were considered accurate enough to validate the modeling approach. Possible ways to improve predictions are specifying better inlet boundary conditions and increasing mesh resolution. Two-Phase Coaxial Jet. Again, using the Budilarto experiments, simulation results for a coaxial jet for different velocity ratios and particle size are presented in Figures 3 through 7. Figure 3 shows instantaneous profiles for VR = 0. The gas velocity profiles for 25-micron and 70-micron particles look similar [(a), (b) ]. The particle velocity profiles [(c),(d)] are however very different from the gas phase velocity. First, it should be noticed that, for numerical reason, particle velocity and number density are set to zero when the particle concentration is six orders of magnitude smaller than the inlet concentration. Then, it can be observed that the large particles have more inertia; their speed decreases more slowly along the centerline and they are relatively unaffected by eddies in the gas phase. Small particles have a velocity profile which is closer to the gas velocity and are clearly affected by eddies. This is also confirmed by profiles of the number density (i.e. number of particles per volume) [(e),(f)]; the small particles spread more than the large ones and also cluster faster. Figure 4 and 5 presents profiles respectively for VR = 1.0 and VR = 1.5. The correlation between the gas phase and the particle phase for small particles is explicit. Effects of the annular jet can be observed on the small particles velocity at the edges of particle jet. Large particles velocity is less influenced by the gas phase and decreases slowly along the centerline. Figure 6 presents comparisons of the mean gas and particles velocity along the centerline between the simulation predictions and experimental data. For VR = 0 [(a),(b)], small particles velocity is almost the same as the gas velocity. Initially, their velocity is slightly slower than the gas velocity, but around x/d =4, the two lines cross and because of inertia the particle velocity becomes larger than the gas velocity. Large particles have a velocity profile very different from the gas velocity. Initially, their speed is lower and when x/d 98%) and high effective gas

(CO+H2) content (>90%) due to its operation at higher temperature (1200~1500oC). However, the entrained flow gasifier has the higher oxygen consumption, lower thermal efficiency and higher cost. Based on many years’ R&D work on the Ash-agglomerating Fluidized Bed gasification process, an integrated coal gasifier has been developed by the Institute of Coal Chemistry, Chinese Academy of Sciences (ICC, CAS). It is a kind of new gasifier which combines the advantages of high temperature entrained flow gasification and moderate temperature fluidized bed gasification. The relatively lower temperature semi-coke in the dense phase of the fluidized bed and the gas produced by fluidized bed part have the function to cool down the high-temperature product gas and slag coming from the entrained flow part so that the slag particles can be mixed into the dense phase and then discharged together with fluidized bed’s coarse ash in dry mode. This kind of integrated gasifier possesses has many advantages like simpler structure, lower cost, higher carbon conversion, lower oxygen consumption and wider coal adaptability. At present, the conception confirmation of this integrated coal gasifier has been accomplished [3], and the R&D on process optimization and scale-up is now carrying out. It is well know that the traditional scale-up process in chemical industry usually needs a long time and is very costly. In recent years, the rapid development of Computational Fluid Dynamics (CFD) and the computer technology has made it possible to study gas-solid two phase flow more accurately and economically. Moreover, the reliable simulation software has been presented for the description of fluidized bed and entrained flow reactor. Therefore, CFD method is used to study the gas-solid flow of the coupling process of the fluid bed and the entrained flow bed. The objective of this work hopes to provide some instructive information for the fundamental research and the technical development of this new process. 2. Mathematical model 2.1 Physical model The pilot-scale integrated gasifier attached an entrained flow bed on the side of the fluidized bed, as shown in Fig.1. Both the fluidized bed and the entrained flow bed have the inner diameter of 300mm. The lower part of the fluidized bed is height of 1300mm and the length of the entrained bed is 2100mm, respectively. The free board has the inner diameter of 500mm and its height is 1000mm in the simulation. The properties of the gas and particulates are listed in Table 1.

Fig.1. Structure and mesh scheme of the integrated gasifier

Table 1. Properties of particulate phases and gas phase Values Coarse particle Fine particle

Parameters Particle density /(kg/m3)

1200

900

Particle mean diameter /mm

1.1

0.06

Minimum fluidization velocity /(m/s) Voidage at minimum fluidization Gas dynamical viscosity /(Pa·s) Gas density /(kg/m3)

0.4 0.42 1.79×10-5 1.225

---

2.2 CFD model equations Numerical schemes based on the Eulerian-Eulerian and the Eulerian-Lagrangian approachs have been used for CFD simulation of the gas-solid multiphase flow. The Eulerian-Eulerian approach, namely two-fluid hydrodynamic model, considers the particulate phase to be a continuous fluid interpenetrating and interacting with the fluid phase, which is suitable to deal with the dense solid flow such as the fluidized bed process. The Eulerian-lagrangian approach considers the particulates as the discrete phase, which is fit for the simulation of the dilute solid phase. The integrated gasifier is composed of the fluidized bed with dense particulate phase and the entrained bed with dilute particulate phase. However, it is difficult to deal with the different gas-solid flow with the two approaches in a single unit. Therefore, the Eulerian-Eulerian model was use for its simulation [4, 5]. This approach results in mass, momentum, and energy balance equations for both gas

and solid phases, as given below: Continuity equations:   q  q      q  q vq   0 ; t Momentum equations: For gas phase:

p or q  s, g , but p  q

     ( g  g v g )    ( g  g v g v g )   g p     g   g  g g  t        g  g ( Fg  Flift , g  Fvm , g )  K sg (v s v g )  msg v sg

(1)

(2)

For solid phase: In the integrated gasifier, there are two different solid phases. One is the coarse particle phase with larger size in the fluidized bed; the other is the fine particle phase with smaller size in the entrained bed. For the coarse particle phase, the interaction of particles should be considered in the dense gas-solid flow.      ( s1  s1v s1 )    ( s1  s1v s1 v s1 )   s1p  p s1     s1   s1  s1 g  t        s1  s1 ( Fs1  Flift ,s1  Fvm ,s1 )  K s1, g (v s1 v g )  ms1g v s1g

For Fine particle phase, the interaction of particles can be neglected.

(3)

        ( s 2  s 2 v s 2 )    ( s 2  s 2 v s 2 v s 2 )   s 2 p     s 2   s 2  s 2 g   s 2  s 2 ( Fs 2  Flift ,s 2  Fvm,s 2 )  t    K s 2 g (v s 2 v g )  m s 2 g v s 2 g

(4)

The fluid-solid momentum exchange coefficient between gas and solid phase can be expressed as follows: 1 2 (5) K s 2, g   p d s 2 C D v f  v p ; 8 The drag force coefficient can be determined as follows:



C D  max 24(1  0.15 Re 0.687 ),0.44



(6)

Turbulence model: The RNG k-ε model [6] is used for the turbulent flow of the gas phase.     ( U )       t  

      C 1 Pk  C 2     k

(7)

    ( Uk )       t k 

  k   Pk    R  

(8)

R

C   3 1   0   2 1  

 eff     t ;

3

k

;

t  C 

  S k  ; S  2Sij Sij k2



;

Sij 

1  vi  v j  2 x j xi

,

; Pk   t U  (U  U T )

C   0.085 ;  k  0.7179 ;    0.7179 ; C  1  1.4 ; C 2  1.68 ;  0  4.38 ;   0.012

Table 2. Simulating conditions of the integrated gasifier Case Parameters Values T /K 293 P /MPa 0.1 H0 /m 0.65 0-3 Uf /(m/s) 0.85 0 Ue/Uf 0 1 Ue/Uf 0.25 2 Ue/Uf 0.5 3 Ue/Uf 1.0 H0 /m 0.7 Uf / (m/s) 0.85 4 Ue/Uf 0.5 H0−static bed height, m; P−pressure, MPa; T−temperature, K; Ue−gas velocity in the entrained bed, m/s; Uf−gas velocity in the fluidized bed, m/s

The calculated domain was meshed by an advanced hybrid grid technology to build rational numerical cells. The total number of the elements was 20410. A commercial software package, Fulent 6.1, was employed in this study. The simulation was conducted in the transient mode. The time step and total time was selected to 0.0005s and100s respectively. The operating conditions used in the simulations are listed in Table 2. Five cases were conducted in simulation. Case 0 was only the fluidized bed. From case 1 to case 3, the entrained bed gas velocity gradually increased to investigate the solid movement and the interacting between the fluidized bed and entrained bed. Case 4 was to measure the influence of the dense phase bed height of the fluidized bed on the mixture of coarse and fine particles. 3. Simulation results and discussion 3.1 Gas-solid flow pattern The gas-solid hydrodynamics in the single fluidized bed and the entrained bed have been extensively studied respectively [7-9]. The hydrodynamic characteristics of the combined gasifier are the interactive behaviors of gas and solids between the two reactors. The effects of the gas velocity, bed height (hold-up) in the fluidized bed were investigated by the numerical simulation in the research works. Figure 2 shows the gas and particle velocity distributions for cases 1-3. From case 1 to case 3, the gas velocity is increased gradually. In the entrained bed, it can be seen that the flow patterns of gas and fine particle are similar due to the strong interaction in each operating cases. Fine particles entrained by high velocity gas from nozzle injects into the vessel. Swirling eddies appear near the wall along the axis direction. The rising gas velocity can decrease the swirling flow and make gas and particle flows tend toward the plug flow in the body of entrained bed although it may promote the swirling near the nozzle because of the high velocity jet causing the vacuum suction. In the connecting area of the entrained bed and fluidized bed, gas and fine particles form a strong up-flow. Meanwhile, the violent swirling flow along axial direction is produced due to the intensive momentum exchange among gases and particles from different beds. Higher gas velocity from the entrained bed may enhance the swirling flow. The flow pattern means that the inner circulation of gas and solid may occur in the joint area to promote the mixture efficiency, which is beneficial to the heat and mass transfer in the gasifier.

(1) Gas

(2) Fine particle (a) Case 1

(1) Gas

(2) Fine particle (b) Case 2

(1) Gas

(2) Fine particle (c) Case 3

Fig. 2. Speed vectors of gas and fine particle for case1-3

3.2 Fine Particle distribution The volume fractions of fine particle for case 1-3 are shown in Fig. 3. In the entrained bed, the fine particle distribution is not uniform because the injecting gas from nozzle leads the intensive swirling flow. Higher particle concentration appears near the nozzle area and the underside of the wall. In order to prevent coal ash from agglomerating and adhering to the wall, particle deposition should be as low as possible. From the comparison of different cases, it also can be seen that the higher gas velocity can reduce the particle deposit near the wall; however, it may promote the suction effect to increase the particle concentration near the nozzle. The higher particle concentration suggests a high temperature area formed near the nozzle area, which is harmful for the long-term operation of the nozzle. Therefore, the appropriate gas velocity is the key to the operation of the novel gasifier. Figure 4 and 5 shows the cross-sectional fine particle distribution and average volume fraction at the joint of the fluidized bed and the entrained flow bed gasifier, respectively. It can be seen that the fine particle concentration is higher on the side of the joint. Furthermore, lower gas velocity in the entrained bed leads the decrease of gas-solid drag force so that more fine particles enter the fluidized bed. In case 1, the fine particle volume fraction in the joint was much larger than that in the case 2 and case 3, as shown in Figure 5. More fine particles entered into the fluidized bed suggest that high temperature slag particles formed in the entrained bed can be mixed into the fluidized bed so that they are cooled and discharged as dry mode to improve the heat efficiency in the gasification process.

(a) Case 1

Case 2

Case 3

Fig. 3. Fine particle volume fraction distributions

(a) Case1

(b) Case2

(c) Case3

Fig. 4. Fine particle volume fraction distributions at the cross-section of the joint

Volume fraction of fine particle

0.0035 0.0030 0.0025 0.0020 0.0015 0.0010 0.0005 0.0000 Case 1

Case 2

Case 3

Fig.5. Fine particle volume fraction at the cross-section of the joint 3.3 Effects of the fluidized bed height The gas and solid interaction between the entrained bed and the fluidized bed is greatly influenced by the dense phase height in the fluidized bed. If the bed height is above cross-section of the joint zone, the gas and solid from different reactors can be extensively mixed, whereas the mixture may be lessened. Figure 6 and 7 show the fine particle and coarse particle volume fraction distributions and the cross-sectional fine particle volume at the joint for case 2 and case 4. The coarse particle volume fraction suggests the dense phase height in the fluidized bed. It can be seen that the fluidized bed height is below the joint zone in case 2, whereas that in case 4. The fluidized bed height is mainly influenced by the static bed height and the fluidizing gas velocity and less influenced by the gas velocity of the entrained bed. The simulation results show that the dense phase of the fluidized bed higher than the joint area is beneficial for

the mixture of fine particles and coarse particles. In case 4, more fine particles are entered into the fluidized bed because of the entrainment of fluidizing coarse particles.

Case 2 (a) Coarse particle

Case 2

Case 4

Case 4

(b) Fine particle Fig.6. Fine particle and coarse particle volume fraction distributions for case2 and case4 4. Conclusions A 3-D CFD simulation model by combining TFM (two-fluid model) and KTGF (kinetic theory of granular flow) based on commercial CFD package has been developed for the novel fluidized-entrained bed combined gasifier. A series of numerical simulations were performed under several operating conditions. The gas-solid mutual interaction between the entrained flow bed and the fluidized bed including the gas-solid flow fields and particle concentration distributions were predicted in detail using the model. It was concluded that the appropriate gas velocity in the entrained bed and the dense phase height in the fluidized bed are critical for the gasification process. Although only cold gas-solid flow was

Volume fraction of fine particle

0.008 0.007 0.006 0.005 0.004 0.003 0.002 0.001 0.000 Case 2

Case 4

Fig.6. Cross-sectional fine particle volume at the joint for case2, 4 considered and no more comparisons were conducted with experimental data, this work indeed will provide a base for the future comprehensive gasification model developing and give a better understanding to the combining gasifier. Acknowledgements This work was supported by the Special Funds for Major State Basic Research Projects Program (the 973 Program) (Grants 2005CB221200) and Knowledge Innivation Program of the Chinese Academy of Sciences (Grants KGCX2-YW-320). References 1. Y. T. Fang, The Study on Gasification Kinetics of Char Fines from Free-board of Fluidized Bed Gasifier, Master thesis, ICC, CAS, 1994. 2. Z. Tang, Simulation of Fluidized Bed Coal Gasification Process, PhD. thesis, ICC, CAS, 1997. 3. Jinhu Wu, Yitian Fang. A new integrated approach of coal gasification: the concept and preliminary experimental results, Fuel Processing Technology, 2004, 86: 261-266. 4. Matteo Chiesa, Vidar Mathiesen, Numerical simulation of particulate flow by the Eulirian-Lagrangian and the Eulerian-Eulerian approach with application to a fluidized bed, Computers&Chemical Engineering, 2005, 29: 291-304. 5. William Vicent, Salvador ochoa, An Eulerian model for the simulation of an entrained flow coal gasifier, Applied Thermal Engineering, 2003, 23: 1993-2008. 6. C. G. Speciale, S.Thangam, Analysis of an RNG based turbulence model for separated flows[J]. Int. J. Eng. Sci., 1992, 30(10): 1379-1388. 7. Jiantao Cao, Zhonghu Chen, Yitian Fang, et al, Simulation and experimental studies on fluidization properties, Powder Technology, 2008, 183: 127-132. 8. Kun Gao, Jinhu Wu, Yang Wang, et al, Bubble dynamics anf its effect on the performance of a jet fluidized bed gasifier simulated using CFD, Fuel, 2006,85:1221-1231. 9. H. Tominaga, T. Yamashita, Simulator development of entrained flow gasifier at high temperature and high pressure atmosphere, 9th Japan Australia Joint Meeting, Melbounre, Vic, Australia, 1999

2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010

NUMERICAL SIMULATION OF GASIFICATION PROCESS IN A CROSS-TYPE TWO-STAGE GASIFIER Yau-Pin Chyou, Chang-Bin Huang, Yan-Tsan Luan Nuclear Fuels and Materials Division Institute of Nuclear Energy Research Atomic Energy Council Longtan, Taoyuan, Taiwan (R.O.C.)

ABSTRACT Numerical simulation of the oxygen-blown coal gasification process inside a cross-type two-stage (E-Gas like) gasifier is studied with the commercial CFD solver ANSYS FLUENT. The purpose of this study is to use CFD simulation to improve understanding of the gasification processes in the E-Gas like gasifier. In this paper, chemical reaction time is assumed to be faster than the time scale of the turbulence eddies. All the species are assumed to mix in the intermolecular level. The 3-D Navier-Stokes equations and species transport equations are solved with the eddy-breakup reaction model (instantaneous gasification). The influences of coal slurry concentration and O2/coal ratio on the gasification process are investigated. Under the condition of feeding carbon being almost completely converted, low slurry concentration is preferred over high concentration if more H2 is wanted with lower syngas temperature; while higher slurry concentration is preferable for producing more CO with higher syngas temperature. The case of higher O2/Coal ratio results in more combustion and leads to lower syngas heating values and higher temperatures. Meanwhile, lower O2/Coal ratio involves more gasification reactions and results in higher CO concentration and lower temperature. The flow behavior in the gasifier, especially the single fuel injection design on the second stage, is examined and discussed. In summary, the trends of simulated results of coal combustion and gasification processes in the cross-type two-stage gasifier are reasonable and the developed simulation model in this study can be used as a tool for preliminary examination of the global effect of thermal-flow and turbulence on gasification process. Key words: clean coal technology, gasification modeling, integrated gasification combined cycle (IGCC). INTRODUCTION Fuel combustion is usually deemed as the main source of greenhouse gas emission inventories, and it typically contributes over 90% of CO2 emissions and 75% of total greenhouse gas emissions in developed countries (Garg et al., 2006). In Taiwan, 53.35% of electricity was generated from coal-fired power plants, 20.35% came from LNG-fired (Liquid Natural Gas), 18.10% came from Nuclear power, * Corresponding Author’s E-mail: [email protected]

Ting Wang Energy Conversion & Conservation Center University of New Orleans New Orleans, Louisiana, USA 3.31% came from oil-fired power plants, while renewable energy, including conventional and pumped storage hydro, generated only 4.9% of electricity in 2009 (BOE, 2010). Coalfired power plants generate the most electricity, but it also discharges the most CO2 emissions as contrasting to other electricity utilities. More than 45% of CO2 emissions were from coal-fired power plants. Therefore, to reduce CO2 emission, it is important to improve the efficiency of coalfired power plants or to capture CO2 from them. There are many ways to reduce carbon emissions and the associated carbon footprints. Technologies for employing renewable energy such as solar, wind, ocean, hydro, and biomass have been developed and are growing at a fast pace. However, it is undeniable that the most polluted fossil fuel, coal, is still the primary energy source throughout the world. Even in Taiwan, a country without domestic coal production, there is still 34% of energy consumption that comes from coal, in which, 77.62% is utilized for power generation, 1.29% for blast furnace, 8.52% for coke production, and 12.58% for industrial and others (BOE, 2010). The supply of coal is abundant, cheap, stable, and will last for 122 years (BP, 2009). The demand and consumption of natural gas and gasoline have increased during the past few decades. Due to the limited amount of petroleum reserves on the earth, oil prices will expectedly keep growing within the next forty to fifty years. Therefore, the importance of using coal fuel has been emphasized because coal fuels have higher stability and wider variety of sources. However, coal is not considered a clean fuel; thus, future efforts should focus on developing the usages of clean and affordable coal fuel. Gasification is the process of converting various carbonbased feedstocks to clean synthetic gas (syngas), which is primarily a mixture of hydrogen (H2) and carbon-monoxide (CO). This conversion is achieved through the reaction of the feedstock with oxygen and steam at a high temperature and pressure with less than 30% of the required oxygen for complete combustion being provided. The syngas produced can be used as a fuel, usually as a fuel for boilers or gas turbines to generate electricity, or it can be used as a source for manufacturing ammonia or hydrogenation applications in refineries to make methanol, hydrogen, or other chemical products. The gasification technology is applicable to any type of carbon-based feedstock, such as coal, natural gas,

burned to produce carbon dioxide, water, NOx, SOx and other trace elements. However, when air or oxygen is insufficient, incomplete or partial combustion occurs, and coal is gasified. In general, coal gasification is categorized for three main steps: (1) devolatilization, (2) volatile combustion and thermal cracking, (3) char combustion and gasification (or char oxidation). The char oxidation is the most important reaction mechanism to produce syngas, which contains complex chemical reactions interacting with the turbulent flow. Because coal combustion has a long history of being applied in industry, the combustion model study is relatively mature. Smoot (1984) and Eaton et al. (1999) comprehensively reviewed the development of coal combustion modeling in fixed, fluidized, and entrained bed. Considering the turbulence, particle dispersion and chemical reaction, Smoot and Smith (1985) and Hill and Smoot (1993) developed a two-dimensional (PGCG-2) and threedimensional (PGCG-3) pulverized coal combustion model at Brigham Young University. The model solved mass, momentum, and energy conservation equations to simulate coal gasification and combustion. Coal particles under high temperature stream, the heat transfer between dispersed and fluid phases, and the resulting chemical reaction and radiation are illustrated in detail by Smoot & Smith (1985). The previous studies of coal devolatilization and volatile combustion models developed in the coal combustion models are applied in coal gasification. Comprehensive research has been carried out to understand velocity and temperature distribution as well as syngas composition in the gasification chamber (Chen et al. 2000, Choi et al. 2001, Vicente et al. 2003, and Watanabe & Otaka 2006). Chen et al. (2000a) implemented a three-dimensional model to simulate a 200-ton two-stage air blown entrained type gasifier, developed for an IGCC process. The simulation showed that turbulent fluctuations affected the volatile and char-oxygen reaction and significantly influenced the temperature and gas composition. Chen et al. (2000b) investigated the coal gasification under different parameters such as air ratio and coal particle size. They found that carbon conversion is independent of the devolatilization rate and less sensible to coal particle sizes, but it is sensible to the heterogeneous char-oxygen, char-CO2 and char-steam reaction kinetics. Besides, air ratio had a significant effect on syngas composition. The research team at the Energy Conversion & Conservation Center (ECCC) at University of New Orleans has made a significant effort to develop a gasification numerical simulation through the commercial software Ansys/Fluent. Silaen & Wang (2005, 2006, 2009, and 2010) used the geometry and the operating conditions based on the information in Bockelie et al. (2002) and Chen et al. (2000a). They concluded that coal slurry feedstock produced more H2 than coal powder feedstock, while coal powder feedstock generated more CO. An instantaneous gasification model can provide an overall approach on gasifier performances in terms of temperature, heating value, and gasification efficiency but not adequately catch the local gasification process inside the gasifier. In 2010, they investigated the effects of different turbulence and coal particle size on coal gasification in an entrained-flow gasifier. They concluded that turbulence model significantly influences the gasification results, and only

heavy refinery residues, petroleum coke, biomass, and municipal wastes. Syngas can be employed in the Integrated Gasification Combined Cycle (IGCC) to produce electricity. Compared to regular or supercritical pulverized coal (SCPC) combustion power plants, IGCC plants can achieve higher efficiency and lower emissions. For power generation, gasification integrated with IGCC is considered a clean and efficient alternative to coal combustion.. The high-pressure and high-temperature syngas from the gasifier can take advantage of the new generation of advanced turbine systems (ATS) to a potential efficiency of more than 50% (LHV). Furthermore, the syngas stream can also be tapped to produce methanol and hydrogen. The gasification technology becomes more important when integrated with carbon capture and sequestration (CCS) technology. CCS is the technology that captures CO2 by a physical or chemical process and stores it. In a typical case, a fossil-fired power plant implemented with CCS technology can reduce 90% of CO2 emissions. Because IGCC+CCS capture CO2 before the syngas combustion, it is easier and cheaper to capture CO2 compared to PC power plants. In the next decade, coal-fired power generation will continue to be the main source of electric power in Taiwan; therefore, it is essential to improve the coal combustion efficiency or to develop new technologies such as coal gasification that can extend the usability of coal more cleanly. Before building an experimental gasification facility, this research team is interested in utilizing Computational Fluid Dynamics (CFD) to help understand the thermal-flow and gasification process in the gasifier and guide the design of an experimental gasifier. Hence, the objective of this study is to establish a preliminary coal gasification model to improve the understanding of the gasification processes in a cross-type (EGas like) two-stage gasifier as well as to investigate the effects of operating parameters such as slurry concentration and O2/coal ratio on gasification performance. LITERATURE REVIEW In an entrained type gasifier, coal particles are injected into the gasification chamber and mixed with an oxidant stream at a high speed. Under the assumptions of fast chemical reactions and local chemical equilibrium, the reaction and particle movement are dominated by turbulence. Therefore, turbulence models play an important role in coal combustion and gasification simulation, and it affects not only the flow field inside the gasifier but also the coal conversion rate of gasification. Launder & Spalding developed the k-ε model in 1972, and it has since been widely used due to its simple physics and robust feature. Over the past decades a considerable number of models, k-ε model, Reynolds stress model (RSM), large eddy simulation (LES), and direct numerical simulation (DNS), have been developed to study and simulate the turbulent flows.. Hunt & Savill (2004) gave an introduction of the turbulence models and the appropriate conditions to use these models. The chemical reactions between coal combustion and gasification are similar. Under high temperature, coal is first decomposed through pyrolysis to vaporized volatile substances. With abundant air or oxygen, coal is completely 2

standard k-ε model and RSM models gave the consistent result in their study. The Kobayashi devolatilization model produces a slower devolatilization rate than the other models. The single rate model and the chemical percolation model produces moderate and consistent devolatilization rate.

Mass of dry coal (mositure free) Mass of slurry (1) Mass of oxygen O 2 /Coal = Mass of coal (MAF, mositure and ash free) (2)

Slurry concentrat ion =

MODEL DESCRIPTION E-Gas (Destec) type Gasifier The E-Gas™ gasifier consists of two stages, a slagging first stage and an entrained-flow, non-slagging second stage, as shown in Fig. 1. The first stage is a horizontal, refractorylined vessel in which carbonaceous fuel is partially combusted with oxygen at an elevated temperature and pressure, 2500°F/420 psia (1400°C/29bar). Approximately 80% of preheated slurry and oxygen are fed to each of two opposing mixing nozzles, one on each end of the horizontal section of the gasifier. The geometry is designed to provide a means for thoroughly mixing the reactants and to disperse this mixture, so a high carbon conversion is realized. E-Gas™ has developed its own proprietary design for these slurry mixers. The oxygen feed rate to the mixers is carefully controlled to maintain the gasification temperature above the ash fusion point of 2400 – 2600 °F (1589 – 1700K) (NETL, 2000), ensuring good slag removal and high carbon conversion. The raw fuel gas flows upward into the second (upper) stage of the gasifier while the molten slag flows down the walls of the gasifier and passes into a slag quench bath. In this upper vertical cylindrical stage, the remaining coal slurry is fed to have additional gasification occur and cool down the syngas temperature. The fuel is almost totally gasified in this environment to form syngas consisting principally of hydrogen, carbon monoxide, carbon dioxide and water. Sulfur in the fuel is converted to primarily hydrogen sulfide (H2S) with a small portion converted to carbonyl sulfide (COS). With appropriate processing downstream, over 98-99% of the total sulfur can be removed from the feedstock prior to combustion in the combustion turbine (NETL, 2000; NETL, 2002). Descriptions of Baseline Case Since the exact dimension of the E-Gas gasifier is not known to the authors, a geometry of a cross-type, two-stage gasifier has been built and simulated based on Fig. 1 and the E-gas (Destec) gasification process with feedstock information published in the open literature (NETL, 2000 and 2002). The Fuel properties of the received Illinois #6 Coal are shown in Table 1 and Table 2. The conditions of coal/water slurry feed, prepared using Illinois #6 coal of the baseline case are shown in Table 3. Approximately 80% of the slurry is gasified / combusted with oxygen in the first (lower) stage using two burners positioned on the opposing ends of this horizontal cylindrical section. In this upper vertical cylindrical stage, only the remaining coal slurry is fed without oxygen. The slurry concentration is 0.67 and O2/coal ratio is 0.91 for the baseline case. The definition of slurry concentration and O2/coal ratio in this paper are as follows:

Figure 1 E-Gas ( Previous Destec ) two-stage entrained flow gasifier. (Source: Destec, 1996)

Table 1 Ultimate analysis of feedstock coal

Ultimate Analysis Moisture Carbon Hydrogen Nitrogen Chlorine Sulfur Ash Oxygen Total

3

Wt. % 11.12 63.75 4.50 1.25 0.29 2.51 9.70 6.88 100

Wt. %, dry 71.72 5.06 1.41 0.33 2.82 10.91 7.75 100

including the enthalpy formation from the chemical reaction of the species. The energy E is defined as p v2 E = h− + (7) ρ 2

Table 2 Proximate analysis of feedstock coal

Proximate Analysis Moisture Fixed Carbon Ash Volatiles Total

Wt. % 11.12 44.19 9.70 34.99 100

Wt. %, dry 49.72 10.91 39.37 100

where h is the sensible enthalpy and for incompressible flow and is given as p h= Y jh j + (8)

∑

ρ

j

Yj is the mass fraction of species j and

Table 3 The feedstock conditions of base case

T

h=

Feedstock conditions of base case Flow rate (tons/day) - Coal (Wet base): - Slurry Concentration: - ASU O2/Coal (MAF): 4627

Second Stage Flowrates (tons/day) - Coal (Dry base, MF): - H2O: - O2 - N2

711

Tref

Turbulence Models Standard k-ε Model – The standard k-ε model, based on the Boussinesq hypothesis, relates the Reynolds stresses to the mean velocity as ⎛ ∂u ∂u j ⎞ 2 ⎟ − ρkδ ij (10) − ρ ui′u′j = μt ⎜ i + ⎜ ∂x j ∂xi ⎟ 3 ⎝ ⎠

1793 896 1841 97

where k is the turbulent kinetic energy, and μt is the turbulent viscosity given by μ t = ρC μ k 2 / ε (11)

474 237 0 0

where Cμ is a constant and ε is the dissipation rate. The equations for the turbulent kinetic energy (k) and the dissipation rate (ε) are: ⎡⎛ ⎤ ⎞ ∂ (12) (ρui k ) = ∂ ⎢⎜⎜ μ + μt ⎟⎟ ∂k ⎥ + Gk − ρε ∂xi ∂xi ⎢⎣⎝ σ k ⎠ ∂xi ⎥⎦ 2 ⎡⎛ ⎤ ⎞ ∂ (13) (ρuiε ) = ∂ ⎢⎜⎜ μ + μt ⎟⎟ ∂ε ⎥ + C1ε Gk ε − C2ε ρ ε σ ε ⎠ ∂xi ⎦⎥ ∂xi ∂xi ⎣⎢⎝ k k

METHODOLOGY Governing Equations The equations for conservation of mass, conservation of momentum, and energy equations are given as: v ∇ ⋅ (ρv ) = S m (3) v vv v ∇ ⋅ (ρv v ) = −∇p + ∇ ⋅ ⎛⎜τ ⎞⎟ + ρg + F (4) ⎝ ⎠ ⎛ v v v ⎞ ∇ ⋅ (v (ρE + p )) = ∇ ⋅ ⎜ λ eff ∇T − h j J j + ⎛⎜τ eff ⋅ v ⎞⎟ ⎟ + S h ⎜ ⎝ ⎠⎟ j ⎝ ⎠ (5) where λeff is the effective conductivity (λ+λt, where λt is the turbulence conductivity) and Jj is the diffusion of species j.

The term Gk is the generation of turbulence kinetic energy due to the mean velocity gradients. The turbulent heat flux and mass flux can be modeled with the turbulent heat conductivity (λt) and the turbulent diffusion coefficient (Dt), respectively. μ ∂T ∂T ρc p ui′T ′ = −λt (14) = −c p t ∂xi Prt ∂xi

∑

The stress tensor τ is given by 2 v ⎤ v ⎡ v τ = μ ⎢ ∇v + ∇v T − ∇ ⋅ v I ⎥ 3 ⎣ ⎦

(

)

(9)

p , j dT

where Tref is 298.15 K.

2550 0.67 0.91

First Stage Flowrates (tons/day) - Coal (Dry base, MF): - H2O: - O2 - N2

∫c

ρ ui′C ′ = − ρDt

μ ∂C ∂C =− t ∂xi Sct ∂xi

(15)

The constants C1ε, C2ε, Cμ, σk, and σε used are: C1ε = 1.44, C2ε = 1.92, Cμ = 0.09, σk = 1.0, σε=1.3. The turbulence Prandtl number, Prt , is set to 0.85, and the turbulence Schmidt number, Sct , is set to 0.7.

(6)

where μ is the molecular dynamic viscosity, I is the unit tensor, and the second term on the right-hand side is the effect of volume dilatation. The first three terms on the right-hand side of equation (5) represent heat transfer due to conduction, species diffusion, and viscous dissipation. Sh is a source term

Reaction Model Coal gasification reactions occur when coal is heated with limited oxygen and steam in a gasification reaction

4

the solid-gas reaction process can be modeled as homogeneous combustion reactions. This approach is based on the locally-homogeneous flow (LHF) model proposed by Faeth (1987), implying infinitely-fast interphase transport rates. The instantaneous gasification model can effectively reveal the overall combustion process and results without dealing with the details of the otherwise complicated heterogeneous particle surface reactions, heat transfer, species transport, and particle tracking in turbulent reacting flow. The eddy-dissipation model is used to model the chemical reactions. The eddy-dissipation model assumes the chemical reactions are faster than the turbulence eddy transport, so the reaction rate is controlled by the flow motions.

chamber. The main global reactions in a gasification process are as follows: Heterogeneous (solid and gas) phase C(s) + 1 2 O 2 → CO, ΔH o R = −110.5 MJ/kmol (R1) C(s) + CO 2 → 2CO, ΔH o R = +172.0 MJ/kmol (Gasification, Boudouard reaction)

(R2) C(s) + H 2 O(g) → CO + H 2 , ΔH o R = +131.4 MJ/kmol

(Gasification) (R3) Homogenous gas phase CO + 1 2 O 2 → CO 2 , ΔH o R = −283.1 MJ/kmol

Solution Methodology The CFD solver used in this study is the commercial CFD code ANSYS FLUENT V.12.0. FLUENT is a finitevolume-based CFD solver written in C language and has the ability to solve fluid flow, heat transfer and chemical reactions in complex geometries, and supports both structured and unstructured meshes. Buoyancy induced flow is calculated. The density in the buoyancy term in the momentum equation is calculated using the ideal gas law. The segregated solution method is used. Segregated solution solves the governing equations of continuity, momentum, energy, and species transport sequentially (segregated from one another). The non-liner governing equations are linearized implicitly with respect to the dependent variables. The second-order discretization scheme is applied for the momentum, the turbulence kinetic energy, the turbulence kinetic dissipation, the energy, and all the species. The SIMPLE algorithm is used in this study as the algorithm for pressure-velocity coupling. The built-in standard k-ε turbulence model is used.

(R4) CO + H 2 O(g) → CO 2 + H 2 , ΔH o R = −41.0 MJ/kmol (Watershift)

(R5) CH 2.760 O 0.262 → 0.262CO + 1.011H 2 + 0.123C 6 H 6 (Volatiles cracking)

(R6) C 6 H 6 + 3O 2 → 6CO + 3H 2 (Volatiles gasificati on via C 6 H 6 )

(R7) Reactions given in R1 and R4 are two exothermic reactions that provide the complete energy for the gasification. Based on these global reactions, approximately 22% of the stoichiometric oxygen is required to provide sufficient energy for gasification reactions. In real applications, 25~30% of the stoichiometric oxygen is provided to ensure high-efficient carbon conversion. Partial combustion occurs when the coal mixes with oxygen (R1). The energy released from (R1) also heats up any coal that has not burned. When the coal is heated without oxygen, it undergoes pyrolysis during which phenols and hydrocarbon gases are released. At the same time, char gasification (R2) takes place and releases CO. If a significant amount of steam exists, gasification (R3) and water shift reaction (R5) occur and release H2. Reactions (R6) and (R7) involve volatiles. The volatiles are modeled to go through a thermal cracking (R6) and gasification processes (R7) via C6H6. The enthalpy of volatiles is calculated from the coal heating value and fully combustion of the carbon from the proximate analysis. The global reaction mechanism is modeled to involve the following chemical species: C, O2, N2, CO, CO2, H2O, H2 and volatiles (see reactions R1 through R7). All of the species are assumed to mix in the molecular level. In this study, the instantaneous gasification and equilibrium models will be established. The instantaneous model assumes that coal vaporizes very fast into gas without going through heterogeneous finite-rate reaction process. The equilibrium model will follow the conventional equilibrium concept by incorporating the equilibrium constants into the CFD simulation. The interphase exchange rates of mass, momentum and energy are assumed to be infinitely fast. Carbon particles are made to gasify instantaneously; therefore

Boundary Conditions The schematic diagram of investigated gasifier is shown in Fig. 2. The height of the whole gasifier is 12m, and the length of the horizontal section (1st stage) is 8m. The diameter of the horizontal section is 2m, and the one of the vertical section that includes a convergent-divergent section is 1.6m. There are two opposing inlets in the horizontal section with a height of 1m, while there is only one inlet in the vertical section with a height of 3.6m. The feedstock of the fuel is about 2550 ton/day with a 0.67 coal slurry concentration and a 0.91 O2/coal ratio. The inlet temperature of both coal slurry and oxidant is 425K, and the operating pressure is 28atm. The wall is assumed as no slip condition and adiabatic wall condition. The gridindependent studies have been tested in this study. However, even the mesh quantities grow to 2 million, the gridindependent is still not achieved. In order to save the computer resource, there are 1.05 million unstructured meshes used in the computational domain.

5

RESULTS AND DISCUSS

Figure 2

Baseline Case Figure 4 shows the velocity contours and vectors along xy plane at z=0 in the gasifier of the baseline case. The velocity vectors in each circle show the velocity field in a cuttingplane of the gasifier. The results show that the flow is injected strongly in the core of the 1st stage; however, there is backflow in the outer rim region of the bottom vessel. The four cutting-plane vectors of the bottom vessel show the backflow forms an annular recirculation region surrounding the jet core flow in the bottom cylinder on both ends. The slow-moving recirculation zones occupy a portion of the gasifier like blockages and reduce the effective area for the core flow to pass through. This leads to a longer residence time for the flow trapped in the recirculation zones but accelerates the core flow due to reduced effective throughflow area in the 1st stage. When the raw fuel gas produced in the 1st stage flows upward into the 2nd stage of the gasifier, it is accelerated through the converging throat region. Since there is an inlet feeding with coal slurry in this throat region, the flow field is deflected by the fuel jet and two separated-flow recirculation zones immediately form downstream of the second stage fuel jet locations. Each of these two recirculation zones contains a pair of counter-rotating vortices as shown in the third crosssectional velocity vector plot from top of Fig. 4. Due to the large include-angle of the divergent section immediately downstream of (or above) the throat, the recirculation zones sustain about three throat-diameters above the second fuel injection location and the flow reattaches to the wall at around 7 meter high of the gasifier. These recirculation zones increase the residence time of the trapped flow and provide well mixing between the unreacted coal slurry and the hot gases from the 1st stage. The lengthened residence time seems beneficial for allowing more thorough reactions, but the yield is low in this region because the produced syngas wastes time recirculating inside this region and does not effectively contribute to the syngas production rate at exit of the gasifier. Therefore, these recirculations actually exert a non-ideal condition because they lengthen only a small part of mass flow in these dead-flow regions and force the main-body flow move faster with less residence time by reducing the effective cross-section area of the flow passage. The accelerated main flow results in requiring a longer gasifier chamber to achieve satisfactory residence time. This uneven distribution of flow fields and residence times leads to the consideration of using multiple tangential jets in the second stage, which can provide a uniformly distributed spiraling flow field with evenly lengthened residence time. In this manner, the length of the gasifier chamber could be shortened and the cost reduced. Another option to reduce the recirculation zones volumes is to reduce the divergent section included angle by increasing the throat area.

Computational grid and boundary conditions

Grid Sensitivity study A grid sensitivity study is conducted by using four different grids: 120,000 meshes, 300,000 meshes, 680,000 meshes, and 105,000,000 meshes. Fig. 3 shows the temperature distribution along the gasifier height of the four different grids. In most part the temperature increases as the mesh number increases. The result of 680k grid almost coincides with that of 1.05M grid except in the region near the second fuel injection (height 4~6 m), where the largest temperature difference is about 40K or less than 3%. Considering the difference in most of the area is less than 1%, the 1.05M grid is then used in this study. 2300

Temperature (K)

2100 1900 1700 1500 1300 1100 0

2

4

6

8

10

12

Gasifier hight (m) 120k

300k

680k

1.05M

Figure 3 Grid sensitivity of vertical temperature distribution.

6

Velocity Field (m/s)

Figure 4 The velocity contours and vectors along x-y plane at z=0 in the gasifier of the baseline case.

7

Figure 5 shows the flow pathline in the gasifier. As the pathline diagram indicates, the gas injected into the gasifier from the opposing inlets in the horizontal section causes a strong upward flow in the middle of the 1st stage. Then most of the fluid turns an angle of 90 degrees and flows upward to the gasifier outlet. The residence time of the fluid from inlet to outlet ranges from 1.15 to 26.6 seconds. However, it can be seen that in the computational domain, the majority of the pathline colors are blue and turquoise ranging between 2 and 6 seconds, and few pathline showing longer residence time in green, yellow, and red. These are the fluid particles that have hung around recirculated for a couple of times before flowing to the outlet. Generally speaking, the residence time of an entrain-bed type gasifier is usually in the scale of seconds, and the result of the 2 – 6 seconds residence time seems to be reasonable. Figure 6 shows the 3-D temperature contours of the gasifier. Figure 7 shows the contours of temperature and mole fractions of CO2, CO and H2 along x-y plane in the gasifier. Combustion occurs when coal slurry and oxidant are injected into the first stage. The carbon and the oxygen react immediately and generate CO via R1 reaction. Then CO reacts with oxygen again to produce CO2 via R4 reaction. The oxidation and combustion reactions generate significant heat and the temperature in the middle of first stage is raised to over 2300K. Due to the feed of oxygen being insufficient for full combustion, some of the coal is consumed and converts into CO and H2 via gasification process by reacting with CO2 and H2O via R2 and R3 reactions. Because these two equations are endothermic reactions, the temperature in the exit of bottom vessel is lowered to around 2000K. When the hot raw syngas transports into the upper stage, it reacts with the coal slurry, which is fed from the upper jet. There is no additional oxidant being fed into the second stage, so the main reaction in this region is gasification. More CO and H2 are produced in this section. When syngas exits the gasifier, the temperature is cooled down to 1591K, due to endothermic gasification process and the mole fractions of H2 and CO are raised to 0.417 and 0.372, respectively. The exit temperature of E-gas gasifier is 1311K according to open literature [NETL, 2000]; however, the predicted value in the base case is 1591K. The difference between the simulation and the actual operation of E-gas gasifier is due mainly to the following causes: (1) The wall of gasifier is set to be adiabatic, which implies that there is no heat loss through the wall boundary; (2) The devolatilization process has not been implemented in this study; thus the energy needed to drive out the volatiles are not consumed. Due to the fast reaction in the eddy-dissipation model, there is no H2O remained in the exit syngas. The mole fraction of H2 in the simulation results is also higher than the counterpart documented in the NETL report (2000).

Pathline (Colored by time, sec)

Figure 5 Flow pathlines of the gasifier.

Temperature (K)

Figure 6

8

3-D temperature contours of the gasifier.

(a) Temperature (K)

(b) Mole fraction of CO2

(c) Mole fraction of CO

(d) Mole fraction of H2

Figure 7 The contours of temperature and mole fractions along x-y plane in the gasifier of baseline case.

9

Effects of Slurry Concentration Cases A1 to A5 in Table 4 are the cases showing the effects of slurry concentration of gasification process. In the coal slurry concentration series cases, the weight of feeding coal and oxidant are fixed; higher value of coal slurry concentration indicate less water in the slurry.

Table 5 The input conditions and simulation results showing effects of O2/coal ratio on gasification process.

Case A2

CaseA3

O2/Coal

0.91

0.91

Coal/Slurry

0.55

Temperature (K) Mole Fraction (%)

Case A1

Case A4

Case A5

0.91

0.91

0.91

0.6

0.67

0.75

0.8

1424

1494

1591

1675

1722

CO

25.0

30.7

37.2

45.4

50.1

H2

47.7

45.0

41.7

37.8

35.5

CO2

22.1

18.7

14.9

10.1

7.4

Baseline Case

Case B2

CaseA3 Baseline Case

Case B4

Case B5

O2/Coal

0.85

0.88

0.91

0.95

0.98

Coal/Slurry

0.67

0.67

0.67

0.67

0.67

Temperature 1448 (K)

1521

1591

1684

1750

CO

38.6

37.9

37.2

36.3

35.6

H2

42.1

41.9

41.7

41.5

41.3

CO2

13.2

14.1

14.9

16.0

16.9

Mole Fraction (%)

Table 4 The input conditions and simulation results showing effects of slurry concentration on gasification.

Case B1

Equations R1, R4 and R7 are the main exothermic reactions in the gasification process. When O2/coal ratio increases, there is more oxygen for R1, R4 and R7 reactions, and the temperature of the exit syngas becomes higher. The purpose of the gasification process is to reduce the combustion process and generate more syngas, i.e., CO and H2. When the input oxygen concentration is higher, there is more combustion reaction with less syngas production. However, the energy for devolatilizing and thermally cracking coal is from the combustion process. If there isn’t enough oxygen for the combustion process, there will be insufficient energy for gasification process, and the syngas production will be less. More oxygen is helpful for the combustion process and leads to higher temperatures in the gasifier. The exit temperature increases with elevated O2/coal ratio. In Case B1, due to the lack of oxygen input into the gasifier, the temperature is lower than in other cases. The temperature of exit syngas is 1448K in case B1 and 1750K in Case B5. The mole fractions of CO, H2 and CO2 are shown in Table 5. When there is more oxygen in the gasifier, there is also more CO2 generated from the combustion process via R4 reaction; meanwhile, the mole fraction of CO is less. The change of H2 from Case B1 to B5 is not distinct. The source of H2 is mainly from Char-H2O gasification. In the O2/Coal series cases, the feedstock of coal and water are fixed. The only changes in the series cases are oxygen inputs. So the change of H2 is not obvious in the O2/coal series cases.

The results in Table 4 show that when the value of coal slurry increases (i.e., less water), the mole fraction of CO increases, and on the contrary, the mole fraction of H2 becomes lower due to less water in the feedstock. However, when the value of coal slurry decreases (i.e., more water), the mole fraction of CO2 becomes higher, but the temperature decreases. When there is more water in the feedstock, more carbon reacts with H2O as well, which generates CO and H2 via R3 reaction. Then CO reacts with H2O and generates CO2 and H2 via R5 reaction. In the results, there are more R2, R3 and R5 reactions in lower slurry concentration cases comparing to higher concentration cases, especially more R3 and R5 reactions. Although carbon will react with CO2 to generate CO via R2 reaction, the concentration of carbon is less in low slurry case, there is also less carbon to react with CO2 via R2 reaction comparing with R3 and R5 reactions. This leads to more CO2, more H2 and less CO produced in the low slurry case. The results show that the temperature decreases with reduced slurry concentration. The temperature of exit syngas in Case A1 is 1424K and in Case A5 is 1722K. The equation R5 is an exothermic reaction, but there are more endothermic reactions, R2 and R3, which deplete more energy. This is the reason why the exit temperature of Case A1 is lower than that in other high slurry concentration cases. In summary, Table 4 shows lower slurry (or more water) produces more CO2 and provides the syngas with higher concentration of H2 and lower concentration of CO.

Comparison with the EPA data Table 6 shows that the operating parameters used in this paper include the operating conditions published in NETL (2000) and EPA (Nexant, 2006). The slurry concentration varies from 0.55 to 0.80, and the O2/Coal ratio varies from 0.85 to 0.98. It can be seen that the simulating parameters used in this paper are consistent with an actual working gasifiers.

Effects of Oxygen / Coal ratio Cases B1 to B5 in Table 5 show the results of the effects of O2/coal ratio. In the O2/coal series cases, the weight of feeding coal and water are fixed; higher value of O2/coal ratio indicates more oxidant in the feedstock.

10

higher slurry concentration is more preferable for producing more CO with higher syngas temperature and less CO2. Higher O2/coal ratio also leads to higher temperature. When the higher O2/coal ratio, there is more combustion reaction and generates more CO2. The effect of O2/coal ratio on H2 production is not obvious, especially for a high O2/coal ratio. The single jet injection in the second stage induces large recirculation regions, which introduces inefficiency and reduced syngas production by trapping a portion of the flow, reducing the effective flow passage area, speeding up the main flow, and decreasing the residence time of main flow. An alternate design of using multiple tangential jets injections in the second stage or enlarging the throat diameter can be considered. Although the simplified instantaneous gasification model is used in this study, the trends of simulated results of the coal combustion and gasification process in the cross-type, two-stage gasifier are reasonable. Future activities need to be considered in the finite-rate reaction mode, especially the water-shift reaction rate without catalysts, and investigate the effects of turbulent model on the gasification process.

Table 6 Operating parameters in the simulated cases

Coal(MF)/Slurry

0.85

0.88

0.55 0.60 0.64 0.67

x

0.75 0.80

x

O2/Coal(MAF) 0.91 0.95 x x x x(NETL)

x

0.98

x(EPA) x

x x

Table 7 shows the comparison of outlet species mole fraction with EPA results, which are computed by Aspen Plus. The major difference is the H2O mole fraction. The outlet species concentrations of H2 and CO2 in this study are higher than EPA results, while the values of CO and H2O in this study are lower than EPA results. And according to the R5 reaction, CO reacts with H2O and produces CO2 and H2. It can be judged that the major difference is caused by the water shift reaction rate of R5. It is clear that the water-shift reaction rate is too fast in the current model using the eddy-dissipation model. Implementation of a correct finite-rate model for water-shift reaction is essential for achieving a more accurate prediction. This will be left as a future task for this research.

ACKNOWLEDGEMENTS The authors are grateful to National Science Council of TAIWAN ROC for the financial support of the code number NSC 98-3114-Y-042A-005.

Table 7 The comparisons of simulated results with EPA's results.

Mole fraction (%) CO H2 CO2 H2O

Simulated results 32.2 42.8 19.1 0.0

REFERENCES BOE (Bureau of Energy), (2010). Energy Statistical Hand Book 2009. Bureau of Energy, R.O.C. (in Chinese). Bockelie, M.J., Denison, K.K., Chen, Z., Linjewile, T., Senior, C.L., Sarofim, A.F., (2002). CFD modeling for entrained flow gasifiers in vision 21 systems, Ninteenth International Pittsburgh Conference, Pittsgurgh, PA, USA. BP, (2009). BP Statistical Review of World Energy. BP. Chen, C., Horio, M. & Kojima, T., (2000a). Numerical simulation of entrained flow coal gasifiers. Part I: modeling of coal gasification in an entrained flow gasifier. Chem. Eng. Sci., 55, 3861-3874. Chen, C., Horio, M. & Kojima, T., (2000b). Numerical simulation of entrained flow coal gasifiers. Part II: effects of operating conditions on gasifier performance. Chem. Eng. Sci., 55, 3875-3883. Choi, Y. C., Li, X. Y., Park, T. J., Kim J. H. & Lee, J. H., (2001). Numerical study on the coal gasification characteristics in an entrained flow coal gasifier. Fuel, 80, 2193-2201. Destec Energy, Inc., (1996). The Wabash River Coal Gasification Repowering Project. TOPICAL REPORT NUMBER 7, NOVEMBER 1996. Eaton, A. M., Smoot, L. D., Hill S. C. & Eatough C. N., (1999). Components, formulations, solutions, evaluation and application of comprehensive combustion models. Prog. Energy Combust. Sci., 25, 387-436.

EPA’s results 42.3 30.7 9.6 14.9

CONCLUSION A three-dimensional computational model has been established to simulate the coal combustion and gasification in a cross-type two-stage gasifier. The instantaneous gasification model is adapted in this study to investigate the gasification processes. The instantaneous gasification model significantly reduces the computational time but can only provide a qualitative trend of gasification process for preliminary investigation. Although the instantaneous gasification model is simplified, it can adequately capture the global feature of the effect of the thermal-fluid field (including turbulence structure) on chemical reactions. The results show that when the value of slurry concentration (carbon to water ratio) increases, the temperature of exit syngas is also higher. When the slurry concentration is lower, there are more R2, R3 and R5 reactions and leads to a higher concentration of H2 and CO2; meanwhile, the concentration of CO and temperature is lower. Under the condition of feeding carbon being nearly completely converted, low slurry concentration is preferred over high concentration if more H2 is wanted with lower syngas temperature, while 11

Faeth, G.M., (1987). Mixing, Transport and Combustion in Sprays, Progress in Energy Combustion Science, Vol. 13, pp. 293-345. Garg, A., Kazunari, K., and Pulles, T., (2006). 2006 IPCC Guidelines for National Greenhouse gas Inventories, Chapter 1: Introduction. The Intergovernmental Panel on Climate Change (IPCC). Hill, S.C. and Smoot, L.D., (1993). A Comprehensive Three-Dimensional Model for Simulation of Combustion Systems: PCGC-3, Energy & Fuels, 7, 874-883. Hunt, J. C. R. & Savill, A. M., (2004). Guidelines and criteria for the use of turbulence models in complex flows. In Prediction of Turbulence Flows, eds. G. F. Hewitt and J. C. Vassilicos, Cambridge University Press, Cambridge Jones, W.P., and Lindstedt, R.P., (1988). Global Reaction Schemes for Hydrocarbon Combustion, Combustion and Flame, 73, 233-249. Launder, B. E. & Spalding, D. B., (1972). Lectures in mathematical models of turbulence. New York, Academic Press NETL, (2000). Destec Gasifier IGCC Base Cases, PEDIGCC-98-003. NETL, (2002). The Wabash River Coal Gasification Repowering Project: A DOE Assessment, DOE/NETL-2002/1164. Nexant, Inc., (2006). Environmental Footprints and Costs of Coal-Based Integrated Gasification Combined Cycle and Pulverized Coal Technologies, United States Environmental Protection Agency (EPA), EPA Contract No. 68-W-03-33, Work Assignment 2-02. Silaen, A. & Wang, T., (2005). Simulation of coal gasification inside a two-stage gasifier. TwentySecond International Pittsburgh Conference, Pittsgurgh, PA, USA. Silaen, A. & Wang, T., (2006). Effects of Fuel Injection Angles on Performance of A Two-Stage Coal Gasifier. Twenty-Third International Pittsburgh Conference, Pittsgurgh, PA, USA. Silaen, A. & Wang, T., (2009). Comparison of Instantaneous, Equilibrium, and Finite-rate Gasification Models in an Entrained-flow Coal Gasifier. Twenty-Sixth International Pittsburgh Conference, Pittsgurgh, PA, USA. Silaen, A. & Wang, T., (2010). Effect of turbulence and devolatilization models on coal gasification simulation in an entrained-flow gasifier. International Journal of Heat and Mass Transfer 53, 2074–2091. Smoot, L. D., (1984). Modeling of coal-combustion processes. Prog. Energy Combust. Sci., 10, 229-272 Smoot, L. D. & Smith, P. J., (1985). Coal combustion and gasification. Plenum Press, New York Vicente, W., Ochoa, S., Aguillon, J. & Barrios, E., (2003). An Eulerian model for the simulation of an entrained flow coal gasifier. Appl. Therm. Eng. 23, 1993-2008 Watanabe, H. & Otaka, M., (2006). Numerical simulation of coal gasification in entrained flow coal gasifier. Fuel, 85, 1935-1943 12

URANIUM AND SOME OTHER TRACE METAL ELEMENT CONCENTRATION OF SOME TURKISH COAL ASHES Isik Ozpeker, Dr., Fikret Suner, Dr., Mehmet Maral, MSc, Tahsin Aykan Kepekli, MSc. Istanbul Technical University, Dept. of Geological Engineering, Istanbul – Turkey [email protected], [email protected], [email protected], [email protected] ABSTRACT This study focuses on uranium and some other trace element concentrations and distributions of some coal occurrences that had been formed in Trachea and Anatolia, Turkey. Coal occurrences are in different age and rank. Most of them are young and their ranks are low. Ashes of coal samples have been picked up from different parts of Turkey and were analyzed and evaluated in terms of uranium and some trace element contents. Chemical investigations were performed on the coal ashes via fluorometric method for analyzing uranium concentration, some trace and major element concentrations were analyzed by AAS (Atomic Absorption Spectrometry) and FP (Flame Photometry) methods. The analyses results show that the uranium content in coal ashes change between 0 – 178 ppm, while the average of Turkish coals is 10 – 33 ppm. Ni, Co, Cu, Zn, Pb, Ag, Fe, Ca, Na and K concentrations were also detected. The mentioned trace element concentrations are over the world averages in most of coal ash samples. Uranium was enriched in the western Anatolia, especially in Mugla – Yatagan, Aydın – Soke, Kutahya – Gediz and Acemkiri coal fields. Also, an asphaltite sample from Sirnak includes noticeable amount of uranium concentration. Uranium accumulation of the coal samples probably depended on surrounding units as the source rocks. Keywords: low rank coal, ash, trace element, Anatolia

INTRODUCTION Coal is clearly important and Turkish lignite reserves are estimated to be over 8 billion metric tons beside 1.1 billion tons of hard coal reserves [1], [2]. Up to 15% of the total coal reserves in Turkey are of sub bituminous rank or higher [1]. Turkey has provided over half of electric generation from burning coal and natural gas [2]. Tons of coal ashes are produced every year. The trace elements' occurrence in coal is to be very important in terms of environmental, economic, by-product, and technological behavior of coals and their

1

effects on human health at present. Although, there are associated with the organic matter of the coal, usually their concentration in coal is related minerals, which show varieties both chemical composition and physical properties. While there are forming ash, their associated trace elements will be fallow [3]. The previous investigations have studied on trace elements in Turkish coals [4]. Furthermore, trace elements analyses were performed using instrumental neutron activation analysis [5] - [7] and recently, there has been more interest in trace elements in Turkish coal [8] - [14]. Our analysis and previous researches’ results were reevaluated and interpreted together. Therefore, samples were obtained from as many working mines as possible. The coals in Turkey are generally low rank (lignite or sub bituminous) formed in several different depositional environments at different geologic times and have differing chemical properties (Fig. 1). Eocene coals are limited to northern Turkey; Oligocene coals, found in the Thrace Basins of northwestern Turkey, are intercalated with marine sediments; Miocene coals are generally located in Western Turkey. The coal deposits, which have limnic characteristics, have relatively abundant reserves. Pliocene– Pleistocene coals are found in the eastern part of Turkey. Most of these coals have low calorific values, high moisture, and high ash contents [4].

Fig.1. Location map of the Turkish coal occurrences [15]

2

MATERIALS AND METHODS Different major coal occurrences were sampled and analyzed. Also previous researchers’ results were re-evaluated for trace elements. Chemical investigations were performed on the coal ashes via fluorometric method for analyzing uranium concentration, some trace and major element concentrations were analyzed by AAS (Atomic Absorption Spectrometry) and FP (Flame Photometry) methods.

RESULTS AND DISCUSSION Trace element contents of coals are related to their mineralogy and major elements [11], [16]. Element associations with coal ash have been widely reported, and information obtained on the degree of organic – inorganic association [8], [11], [16] - [20]. Significant positive correlations with ash content for As, Cd, Cr, Ni, Pb, Sb and U were recorded for the A1 lignite bed Paleocene, Calvert Bluff Formation of Wilcox Group, USA [11], [21]. Some trace element U, Ni, Pb, Zn, Cu and Co concentrations in ashes of the major coal fields are present in various concentrations (Table 1) in comparison with the range for most world coals and ashes [22], [23]. Table 1. Some trace elements analysis of the coal ash samples (*[23], **[24], +[25], ++ [26]) +

Sample No

Amasya-Suluova 17145 Ankara-Ayaş 16039 Ankara-Beypazarı-Nallıhan Ankara-Beypazarı Ankara-Beypazarı 7TKİ Aydın-Söke-Dedeman BüyükD Aydın-Söke-Dedeman KesmeliD Balıkesir-Gönen-Şaroluk 17617 Balıkesir-Gönen-Akas 12TKİ Bingöl-Karlıova 17110 Bingöl-Karlıova 17135 Bolu-Reşadiye Bolu-Kamil Bilgehan Bolu-Adasal No:1 Bursa-Orhaneli Bursa-Orhangazi Bursa-Keles Çanakkale-Çan 1

++

++

++

++

++

U ppm 9* 2 10 38 2

Ni ppm 76* 172 225 N/A 171

Co ppm 37* 56 11 N/A 17

Cu ppm 68* 102 172 N/A 17

Zn ppm 117* 216 101 N/A 30

Pb ppm 45* 68 29 N/A 5

10 136 15 14 N/A N/A N/A 0 40 4 3 N/A 17 N/A

310 165 127 90 52 49 117 101 N/A 223 N/A 102 N/A 68

6 38 23 33 19 6 21 2 N/A 30 N/A 16 N/A 56

64 80 72 2276 44 56 44 82 N/A 65 N/A 72 N/A 269

180 130 110 140 103 66 247 64 N/A 82 N/A 77 N/A 322

140 64 51 180 70 18 58 5 N/A 57 N/A 82 N/A 269

3

Çanakkale-Katar Ticaret Çanakkale-Örencik 6TKİ Çankırı-Ilgaz 17707 Çankırı-Orta 8TKİ Edirne-Keşan-Mandadere Edirne-Hacıömer 15940 Edirne-Hacıömer 15941 Erzurum-Oltu-Balkaya Isparta-Türk Civa TKİ276 İstanbul-Kilyos-Kısırkaya İstanbul-Ağaçlı İstanbul-Demirciköy İstanbul-Demirciköy Kastamonu-Azdavay Kastamonu-Azdavay Kastamonu-Azdavay (Artık) Kütahya-Gediz Büyük D. Kütahya-Acemkırı D. Manisa-Akhisar 16195 Manisa-Soma 16745 Manisa-Soma TKİ10 Manisa-Soma TKİ167 Maraş-Elbistan-Afşin 16195 Maraş-Elbistan-Afşin (gitya) Maraş-Elbistan-Afşin TKİ180 Muğla-Milas-Sekköy Muğla-Milas-Sekköy 17983 Muğla-Milas-Sekköy 17931 Muğla-Milas-Sekköy17999 Muğla-Milas-Sekköy17878 Muğla-Karakuyu-Yatağan Muğla-Karakuyu-Yatağan 17935 Muğla 17919 Muğla 17937 Muğla 17938 Muğla-Yatağan 17939 Sivas-Kangal 19808 Yozgat-Sorgun TKİ276 Zonguldak-İncirharmanı Zonguldak (Lavvar artık) Şırnak Asfaltit TKİ261 Yatagan** Soma** Seyitomer** Yenikoy** Catalagzi** Afsin-Elbistan** Tuncbilek**

5 1 2 6 0 N/A N/A 4 N/A 5 N/A 1 2 11 2 N/A 178 149 53 8 28 14 4 1 5 1 20 20 42 7 47 98 N/A N/A N/A 17 63 3 4 N/A 83 N/A N/A N/A N/A N/A N/A N/A

N/A 250 440 178 261 158 120 N/A 125 590 244 194 N/A N/A N/A 98 1302 N/A 105 89 103 N/A 200 N/A 170 N/A N/A 173 N/A 161 N/A N/A 151 189 151 113 144 77 125 69 2698 134.7 79.7 1975.9 79.6 118.5 119.4 1986.1

N/A 49 44 28 21 7 32 N/A 2 63 50 60 N/A N/A N/A 38 48 N/A 22 399 25 N/A 46 N/A 13 N/A N/A 26 N/A 23 N/A N/A 29 22 34 32 11 61 35 33 8 N/A N/A N/A N/A N/A N/A N/A

N/A 74 101 84 87 92 71 N/A 60 212 3360 368 N/A N/A N/A 198 247 N/A 52 79 26 N/A 87 N/A 40 N/A N/A 14 N/A 14 N/A N/A 50 86 80 58 42 40 1782 160 208 136.7 59.8 98.8 39.8 98.8 39.8 99.3

N/A 138 222 101 138 178 173 N/A 70 184 246 221 N/A N/A N/A 104 200 N/A 26 153 250 N/A 122 N/A 40 N/A N/A 42 N/A 72 N/A N/A 28 114 150 156 66 70 134 148 2800 404.1 195.2 112.6 71.6 138.3 79.6 137

N/A 60 86 12 60 66 52 N/A 80 138 180 250 N/A N/A N/A 88 60 N/A 2 124 32 N/A 78 N/A 14 N/A N/A 5 N/A 2 N/A N/A 2 52 30 20 5 82 176 100 36 57.7 79.7 79 39.8 118.5 79.6 99.3

The average U concentrations generally increased north western and south western of Anatolia. The U concentrations were insignificantly different from each other and low contents in Marmara and the Black Sea regions while moderate U concentrations were noticeable in Aegean and Central Anatolia regions. The concentration of Ni in the coal ash samples was high and changed between 49 – 1302 ppm. The average Ni values are

4

high in the South Marmara and western Anatolia. Analysis suggests that the central Anatolia has a significantly higher than does the eastern Anatolia. The coal ash samples generally have low and moderate Co contents. The maximum value of Co detected is 620 ppm. The average concentration of Co in Turkish coal ash is 27 ppm. The higher value of Cu detected was 3360 ppm. The average concentration of Cu in coal ash was 190 ppm. The average Cu values are high in the north western Anatolia. The average Zn concentrations changed from 30 to 322 ppm and, had insignificant changes in Anatolia and Trachea. Also Pb values showed similarities like Zn and changed between 2 – 269 ppm.

CONCLUSION The trace element concentration of coal ashes was higher, especially Ni and U values. Furthermore, Co was significant in Agean. Cu, Pb and Zn content of ash exhibited similarities. The mentioned elements were intense western Anatolia.

REFERENCES [1] Tuncali, E., Ciftci, B., Yavuz, N., Toprak, S., Koker, A., Acik, H., Gencer, Z., Tahin, N., 2002. Chemical and technological properties of Turkish Tertiary coals. General Directorate of Mineral Research & Exploration, Ankara, Turkey. [2] Lynch, R., 2003. An energy overview of the Republic of Turkey. U.S. Department of Energy. http://www.fe.doe.gov/international/turkover.html. [3] Davidson, R. M., 1996. Trace elements in coal. Energia 7, No. 3 [4] Palmer, C.A., Tuncali, E., Dennen, K.O., Coburn, T.C., Finkelman, R.B., 2004. Characterization of Turkish coals: a nationwide perspective. International Journal of Coal Geology 60, 85–115. [5] Ayanoglu, S.F., Gunduz, G. 1978a. Neutron activation analysis of Turkish coals; I. Elemental contents. Journal of Radioanalytical Chemistry 43, 155–157. [6] Ayanoglu, S.F., Gunduz, G. 1978b. Neutron activation analysis of Turkish coals; II. Analysis of ashes and the effects of burning condition on percent transference. Journal of Radioanalytical Chemistry 43, 159–164. [7] Ayanoglu, S.F., Gunduz, G. 1978c. Neutron activation analysis of Turkish coals; III. Relation between composition of coal and local earth crust. Journal of Radioanalytical Chemistry 43, 165–167.

5

[8] Karayigit, A. I., Akgun, F., Gayer, R. A., Temel, A. 1999. Quality, palynology, and paleoenvironmental interpretation of the Ilgin Lignite, Turkey. International Journal of Coal Geology 38, 219–236. [9] Karayigit, A. I., Gayer, R. A., Querol, X., Onocak, T. 2000a. Contents of major and trace elements in feed coals from Turkish coal-fired power plants. International Journal of Coal Geology 44, 169–184 [10] Karayigit, A. I., Spears, D. A., Booth, C. A. 2000b. Antimony and arsenic anomalies in the coal seams from the Gokler Coalfield, Gediz, Turkey. International Journal of Coal Geology 44, 1–17. [11] Karayigit, A. I., Spears, D. A., Booth, C. A. 2000c. Distribution of environmental sensitive trace elements in the Eocene Sorgun coals, Gediz, Turkey. International Journal of Coal Geology 42, 297–314. [12] Querol, X., Whateley, M. K. G., Fernandez-Turiel, J. L., Tuncali, E. 1997. Geochemical controls on the mineralogy and geochemistry of the Beypazari lignite, central Anatolia, Turkey. International Journal of Coal Geology 33, 225–271. [13] Querol, X., Alastuey, A., Plana, F. et al. 1999. Coal geology and coal quality of the Miocene Mugla basin, southwestern Anatolia, Turkey. International Journal of Coal Geology 41, 311–332. [14] Gurdal, G. 2008. Geochemistry of trace elements in Çan coal (Miocene), Çanakkale, Turkey. International Journal of Coal Geology 74, 28–40. [15] http://www.mta.gov.tr/v1.0/images/daire_baskanliklari/enerji/siteharitalar/4.jpg [16] Spears, D.A., Zheng, Y., 1999. Geochemistry and origin of elements in some UK coals. Int. J. Coal Geol. 38, 161–179. [17] Nicholls, G.D., 1968. The geochemistry of coal-bearing strata. In: Murchison, D.G., Westoll, T.S. _Eds.., Coal and Coal Bearing Strata. Oliver and Boyd, Edinburgh, pp. 269–307. [18] Goodarzi, F., 1988. Elemental distribution in coal seams at the Fording coal mine, British Columbia, Canada. Chemical Geology 68, 129–154. [19] Finkelman, R.B., 1994. Modes of occurrence of potentially hazardous elements in coal: levels of confidence. Fuel Processing Technology 39, 21–34. [20] Mukhopadhyay, P.K., Goodarzi, F., Crandlemire, A.L., Gillis, K.S., MacNeil, D.J., Smith, W.D., 1998. Comparison of coal composition and elemental distribution in selected seams of the Sydney and Stellarton Basins, Nova Scotia, Eastern Canada. Int. J. Coal Geol. 37, 113–141. [21] Crowley, S.S., Warwick, P.D., Ruppert, L.F., Pontolillo, J., 1997. The origin and distribution of HAPs elements in relation to maceral composition of the A1 lignite bed _Paleocene, Calvert Bluff Formation, Wilcox Group., Calvert mine area, east–central Texas. Int. J. Coal Geol. 34, 327–343. [22] Swaine, D.J., 1990. Trace Elements in Coal. Butterworths, London, 278 pp.

6

[23] Meij, R., 1995. The distribution of trace elements during the combustion of coal, Env. Aspects of Trace elements in coal, Ed: Swaine and Goodarzi. [24] Bayat, O., 1998. Characterisation of Turkish fly ashes. Fuel Vol. 77, pp. 10591066. [25] Ozpeker, I., Maytalman, D., Kural, O., Uranium concentration of some coal ashes. [26] Ozpeker, I., Maytalman, D., Trace element distribution in some coal ashes.

7

TRANSFORMATIONS OF KARAMAN -ERMENEK LIGNITES OF TURKEY UNDER ACCELERATED ELECTRONS IMPACT Islam Mustafayev* , Fethullah Chichek* , Guven Onal**

*Institute of Radiation Problems, Azerbaijan National Academy of Sciences. 31a H.Javid ave, Baku-Azerbaijan, [email protected] **Istanbul Technical University. Mining faculty. 80260 Ayazaga, Istanbul-Turkey Abstract. The regularities of transformation of lignites from Karaman-Ermenek deposits of Turkey under accelerated electron impact were studied. The absorbed doze in lignites changed within the limits of 1170-3120 kGy. As basic indexes of process rate of gas formation, decreasing of initial mass of lignite, the contents of sulfur in the solid have been defined. The gaseous products Н2, CO and СН4 were identified. The specific features of radiation-chemical decomposition of organic mass of lignite under accelerated electrons impact are discussed . Introduction. The study of radiation-chemical transformation of coal has great importance for creation of scientific basis of radiation-chemical technology of fossil fuels, establishment of the role of radiation in genesis and metamorphism of fossil fuels as well as determination of radiation stability of carbonaceous materials used in nuclear technology. Researches on radiation-chemical processes of transformation of fossil fuels have been spent for last 25 years in USA, Japan, Italy, Russia and other countries within the framework of the Programs on atomic-hydrogen energy and new methods of processing fossil fuels /1-3/. Radiation-stimulated processes in low-grade lignites of Turkey are not regularly investigated. The separate researches on radiation-chemical desulphurization and co- copyrolysis of lignites of Yatagan, Soma and Yenikoy has been conducted in the Istanbul Technical University and Dokuz Eylul University /3-5/. In this work the regularities of transformation of lignites from Karaman-Ermenek deposits of Turkey under accelerated electrons impact were studied.

EXPERIMENTAL The basic characteristics of the investigated samples are: ashes-19.26 %, moisture-15.6 %, sulphur-2.01 %, calorific value - 3775 kcal/kg. Processes were spent under the impact of accelerated electrons on linear electron accelerator (ELU-6) with power of electron beam P= 12 20 W. The rate of the absorbed doze was determined by combined method-cylinder of Faraday

and Calorimeter and equaled to 4680- 7800 kGy/h. The absorbed dose in lignites changed within the range of 1170-3120 kGy. Irradiation was carried out in semiflowing installation, gas products have been collected in gasometer and liquid- in receiver. As basic parameters of process rate of gas formation, losses of initial mass of lignite, the contents of sulfur in solid have been defined. The gases of Н2, CO and СН4 were identified by chromatography. Decrease of lignite organic weight was determined by gravimetric method. Definition of the contents of sulphur is based on spectroscopic determination of sulphur dioxide-products of oxidation of sulfur.

RESULTS AND DISCUSSION In figure 1.kinetics of formation of gases Н2, CO and СН4 at the radiation-chemical

x 1018 molec/gram

Concentration of gases,

decomposition of lignite sample was illustated.

16 14 12 10 8 6 4 2 0

H2 CO CH4

0

10

20

30

40

50

T, min

Fig 1.Kinetics of formation of gas Н2, СО, СН4 at decomposition of lignite under accelerated electron impacts The formation rate (W, 1015 molec/g sec) and radiation–chemical yield (G, molec/100 eV) of gases was determined. In our experiments these parameters are: W (H2) =3.6, W (CO) = 5.2, W (CH4) =0.98, radiation-chemical yield of these gases are: G (H2) =0.172, G (CO) =0.25 and G (CH4) =0.047. It should be mentioned that these values exceeds radiation-chemical yield of gases in case of γ-radiolysis. On table 1. the values of radiation-chemical yield of gases Н2, CO and СН4 from lignites of various deposit of Turkey at the γ-radiolysis were presented.

Table 1. Radiation-chemical yields (Gx103 molec/100 eV) of gases at the γ-

of

Turkish lignites

Lignite type

Н2

СО

СН4

С2Н6

С2H4

C3

C4

C5

Karaman –

24

-

0.7

-

-

-

-

-

22

-

1.7

-

-

-

-

-

Nevshehir, raw 24

-

4.9

2

-

-

-

-

21.7

0.7

0.8

1.0

0.6

0.3

-

-

-

-

Ermenek, cleaned Karaman – Ermenek, raw

Silopi, raw

62

Trakya, mixed

19

4

1.0

The radiation-chemical yield of these gases at the γ-radiolysis are changed: G (H2) =0.019-0.062, G (CO) =0.004 and G (CH4) =0.001-0.021 molec/100 eV. Comparison shows that in case of irradiation by accelerated electrons the radiation-chemical yield of gases 3-9 times for hydrogen, 62 times for CO, 2-47 times for СН4 exceeds the values in case of γ-radiolysis. The observable difference in the yield of gases at the impact of γ- radiation and accelerated electrons can be connected by the influence of high value of dose rate which can lead to qualitatively new processes, or change of physical condition (temperature, pressure, etc) of reaction zone. In our early researches /3-6 / it has been shown that hydrogen and methane from lignites are formed due to recombination reactions of radiolytic radicals of Н, СН3: M ® CH3, H, CO, M +, M* H+H ® H2 CH3+H ® CH4 and etc Carbon monoxide is formed at radiation splitting of oxygen-containing functional groups. Change of recombination mechanism to reaction of separation in the gas formation demands temperature more 200С which in our case is not observed. Small increase in gases yield with activation energy Е = 10-15 кJ/mole is possible due to increase of diffusion of particles in reactionary volume. It means that active particles formed at radiolysis can be stabilized in defective structure of lignite mass. Rise in temperature even above 500С can lead to simplification of their diffusion and by that increase in recombination rate of radical products with formation H2 and CH4. At braking accelerated electrons in the sample volume local rise in

temperature and increase in the yield of products may be observed. Such rise in temperature does not happen under the action of γ- radiation and yields of gases considerably decrease under the influence of accelerated electrons. The course of kinetic curves at rather high doses (increase of inclination) also testifies local heating of lignite mass. The dependence of content of volatile products in structure of lignite at absorbed doze at irradiation in the range of doze D=1170-3120 kGy was illustrated in fig. 2. The lignite mass during irradiation decreases up to 76.5 %. During irradiation volatile products decrease correspondingly from 35 % to 11.5 %, so approximately 23.5 % of volatile products are allocated in the form of gases and liquid. In the same figure laws of change of sulphur relative content in solid product, concerning initial value were shown too.

120

Content, %

100 80 Organic mass, %

60

Sulphur, %

40 20 0 0

10

20

30

40

50

T, min

Fig.2. Dependence of relative changes of lignite mass and sulphur content at the irradiation by electron beam It should be noted that rather deep decomposition of lignites at thermal decomposition below 400С temperatures is impossible. The observable phenomenon is connected with destructive effect of ionizing radiation. Material balance of this process it is resulted in the table 2.

Table 2. Material balance of process of lignite radiation-thermal decomposition (in mass %)

Initial mass

Solid product

Liquid product

Gases

100

74-78

20-24

1-2

It is evident that in semiflowing installation the products of reactions basically are in liquid phase. The changes of absolute content of sulphur in lignite sample at absorbed dose of accelerated electrons were adduced in fig.3. The contents of sulphur in solid decreases from 2.01 % up to 1.62 % in the course of irradiation of 40 minute in W=12 W .

S, abso lute, %

2,5 2 1,5 1 0,5 0 0

10

20

30

40

50

T, min

Fig.3. Influence of irradiation (absorbed doze) on absolute content of sulphur in solid product of radiation-chemical decomposition of lignite

If we consider that there is also decomposition of organic mass of lignite a degree of relative reduction the contents of sulfur reaches up to 57.1 % (fig 2-3). It shows that in spite of decays up to 23.5 % of organic weight, and thus leaves up to 43 % of sulphur. It testifies about selective desulphurization of lignite mass under electron beam impact. Last circumstance is connected with selective cleaning of sulphur from organic mass of lignites, which is connected with highest electronic density of atoms of sulphur than atoms of carbon and hydrogen. Radiation action on chemical bounds with more density of electrons is more effectively. The integrated influence of radiation and temperature can lead to increase of selectivity of radiation desulphurization due to stimulation of separation processes with participation of sulphur-content groups. Thus, at radiation-chemical decomposition of lignites under the impact of accelerated electrons a number of the specific features differing influence of γ-radiation was shown: 1)

At braking electron there is local heating of lignite mass, that stimulates diffusion processes in gas formation

2)

At pulse impact of accelerated electrons with frequency 50 Hz and duration of the impulse 2.5 sec happens high-speed (104-105 degree/s) heating of lignite mass. It leads to deep decomposition of organic mass with formation of liquid products.

These features of influence of accelerated electrons on lignites can be used by development of processes of radiation-chemical technology of processing fossil fuels.

REFERENCES 1. D. D’Anjou, R. Litman. Radiochemical and Radioanalytical letters, 50, No1, (1983), p.37 2. Yermakov A., Popov V.N., Dzantiyev V. Radiation-chemical influence of irradiation on thermoradiation processes of coal gasification/ Chemistry of High Energy, 1988, v.22, p.132136. 3. I.Mustafayev, H.Hajiyev, K.Yagubov, B.Dzantiyev. Hydrogen and Hydrogen-containing gas production from fossil fuels by thermoradiation. Proc. of the 7-th World Hydrogen Energy Conference. Moscow, USSR, 25 - 29 September, 1988. v - 2, p. 955 - 965 4.I.Mustafayev, A.Yamik, H.Mahmudov. Coal Desulphurization by Electron Beam. 5-th Inter. Mineral Processing Symposium Capadociya, Turkey, 6 - 8 September, 1994, Progress in Mineral Processing Technology Ed. H.Demirel , p. 343 - 346 5. I.Mustafayev, M.Kemal, Y.Tolgonay, A.Yamik. Radiation-stimulated Processes of Destruction and Polycondencation in Coals. J. Radioanalitical and Nuclear chemistry letter, 1994, 187 (5), 355 – 365 6. Mustafayev, F.Chichek. Gas Formation at Radiation-Chemical Decomposition of Turkish Lignites. XXIII International Mineral Processing Congress,.3-8 September, 2006, Proceedings V. 3, p.2554-2558, Promedadvertsing-2006.

Desulfurization and Kinetics of Removal of Sulfur from High Sulfur Coal under Hydrogen Atmosphere Guojie Zhang, Yongfa Zhang* , Fengbo Guo, Bingmo Zhang Key Laboratory of Coal Science and Technology of Shanxi Province and the Ministry of Education ,Taiyuan University of Technology , Taiyuan 030024 , China Abstract: The reaction between hydrogen and sulfur in high sulfur coal at high temperature was studied in this paper. Crashed and sieved high sulfur coal sample (with particle size of 0.6mm) was placed in batches in 23 mm I·D· differential reactor. The release of hydrogen sulfide at run temperature and under different hydroatmospheres was followed by a hydrogen sulfide detector. The desulfurization yield was obtained from elemental analysis of residual char. The hydrogen can greatly promote the effect of desulfurization and more than 65% sulfur in the coal can be removed. The releasing curves of H2S in hydropyrolysis process obviously showed two peaks. The desulfurization process in hydropyrolysis of high sulfur coal can be regarded as in two stages according to the evolution profiles of H2S. The first peak at 250~450 ℃ was from the desulfurization of aliphatic sulfide and the second peak at 450~650℃could be from both the sulfur in pyrite and aromatic thiophenic structure. Results show that the desulfurization of high sulfur coal could be described much better with the grain reaction model than with the random pore model. The random pore model is only adapted to the initial stage of sulfur removal from high sulfur coal under hydrogen atmosphere while the grain reaction model is adequate the whole stage. Key words：high sulfur coal, hydropyrolysis desulfurization, kinetics, grain reaction model, random pore model

1. Introduction China is the world's largest coal producer and consumer. Coal accounts for 75% of primary energy in China, high sulfur coal reserve is about 30% of total coal reserves，and the proportion of high sulfur coal mining has also been increasing year by year

[1]

. Along with the rapid

development of economy in China, a large amount of SO2 and NOx emissions owing to coal consumption in primary energy, have resulted in severe environmental pollution. It not only will result in the formation of photochemistry smoke, acid rain and the human body breath endanger, but also will destroy the ecological environment, as well as bring the serious influence on the mankind's life and production. According to the resources and the economic base of our country, in next 50 years, the primary energy sources are coal-based in china [1,2]. So, developing a cheap, easy operation coal desulfurization technology will have far-reaching economic and environmental significance. Many desulfurization for coal before combustion methods have been proposed [1]. But so far, an economical and effective method of desulfurization has not been found. Except liquefied *Corresponding author. Tel. : +86 0351 6018676 E-mail address: [email protected]

pyrolysis and gasification conversion, hydropyrolysis (Hypy) is the third ways conversion method, Hypy has got a great concern as a safe and clean alternative [3]. In this paper, hydrogen pretreatment on the impact of coal pyrolysis desulfurization and kinetics of removal of sulfur from high sulfur coal under hydrogen atmosphere at high temperature was studied.

2. Desulfurization reaction model Sulfur compounds in coal are divided into two categories: inorganic sulfur and organic sulfur. Inorganic sulfur is mainly pyrite sulfur and sulfate sulfur; organic sulfur includes aliphatic sulfur, aromatic and thiophene sulfur in coal. The reaction of sulfur compounds is very complexity in coal hydropyrolysis, The results were obtained by using the first order reaction model to treat the various parts of experimental data with different conversion levels [4]. (Ai－S)s(solid) + H2(gas) = (Ai) s(solid) + H2S(gas)

(1)

Taking into account structure characteristics of high-sulfur coal particles, the grain reaction model and the random pore model are used to define the mechanism of high sulfur coal desulfurization. [5-7] According to grain reaction model, the solid particles are made of a large number of tiny crystals. The diameter of tiny crystal is about nm, the particle size of 10-6 ~ 10-7. These crystals may be spherical, cylindrical or sheet of non-porous material. Anywhere these materials contacting with the gas reactants, the reaction cannot be completed immediately. In addition, the particles can also be seen as a pile of ceramic aggregate, closely together. The gas diffusion in small aggregate diameter is belonged to Knudsen`s type, which is much slower than the gas diffusion in the main hole. Although they are not absolute non-porous material, but have similar reaction behavior of non-porous material. The reaction step of the grain model is including the gas reactant diffusion from outer to inner, particle surface reaction, the diffusion of product gas from inside to outer. After block of coal grinded, the pore structure of coal particles destroyed, the coal particles make up a large number of small spherical dense granules. The reaction of each particle is subject to unreacted core model. The continuity equation of solid coal expressed by unreacted core model of a single particle is followed as:

∂rgc ∂t

Deg rgc ks C A

=

Deg rgc + ks CSgo (1-

rgc rgo

（2）

)

The boundary conditions can be expressed as rgc

t=0

（3）

=r g

The conversion scores of a single particle can be expressed as rgc 3 x =1-( ) , g rgo

Simultaneously

υS (or υA )=-

∂Cs ∂t

1

r =r (1-x ) gc go g

3

（4）

C Sg = C So (1 − xg ) =-Cso

∂ (1-x g ) ∂t

=Cso

∂x g ∂t

where 2

k s C A g C S O (1 -x g ) 3 3 υS = • rg o 1 + δ 2 [1 -(1 -x ) 1 3 ](1 -x ) 1 3 g g g ∂x g ∂t

2

=

k s C A g (1 -x g ) 3 3 • rg o 1 + δ 2 [1 -(1 -x ) 1 3 ](1 -x ) 1 3 g g g

where δ g2 =

k s rgo Deg

(5)

，which can be seen as the modulus of particles reaction.

Assumption coal particle used size uniformity in experiments, and material layer is very thin （ δ = 0 ）. Simultaneously, the size of coal particle is fine, internal mass transfer resistance of 2 g

single-particle can be negligible. The reaction rate of experiment coal samples can be expressed as: ∂x g ∂t

=

3k sC Ago r

go

2 (1-x ) 3 g

（6）

Let be β = t

3ksCAgo

（7）

rgo

where 1 1-(1-x) 3 =β t t s

（8）

Random pore model assumes that the fine grinding of coal particles have a porous structure, the desulfurization reaction meets the assumptions of the random pore model. The desulfurization rate of a single particle can be expressed as: E − dx R rT n =k o e g(PH )f(x)(1-x) dt 2

（9）

For this experiment, g(PH2)，f(x) can be considered constant, let be n=1, the integral equation: （10）

ln(1-x)=-β t 2

where β =k e 2 0

E R rT

g(P )f(x) H2

（11）

Because coal particles used in experiment are in the same conditions, similarly, the desulfurization reaction meets the assumptions of the grain model. The desulfurization conversion can be expressed by equation (10).

3. Experimental The apparatus is shown in Figure 1. The reactor was a stainless steel pipe, mounted in a vertical tube furnace which was controlled on the basis of bed temperature. The product gases were analyzed by gas Chromatography (GC)-950 with a FPD detector.. The gas (in coke oven gas) intake can be measured using mass flow meter. Furthermore, the output of product gas flow can be measured using soap liquid meter.

The reactor was brought up to temperature with nitrogen purge. Hydrogen was introduced into the system at the start of the run. Adjustments were then made to bring the flow to the desired levels. The gas After the run, the reactor was disassembled; the product char was weighed, and analyzed by means of KZDL--3C sulfur determinator and VG Scientific ESCALab220i-XL Electron Spectrometer for residual sulfur. All experiments with larger errors in repeated experiment were rejected. The proximate and ultimate analysis of the coal sample is presented in Table 1. Table.1 The proximate and ultimate analysis of the coal sample Proximate analysis/ %(mass) ,ar

Ultimate analysis/ %(mass) ,daf

Mad

Aad

Vad

Cad

Had

Nad

Sad

Oad*

0.92

22.48

29.84

70.02

4.38

0.95

2.92

21.73

*by difference

The formula of desulfurization rate (SR) of coal was as followed:

SR =

( M • S0 − m • s) × 100% M • S0

Fig.1

High sulfur coal desulfurization system

4. Results and discussion 4.1 Effects of flow on desulphurization rate

Fig .2 Effects of flow on desulphurization rate（Heating rate：15℃／min, T: 550℃） The results are presented in Figure 2 as per cent of the desulfurization rate after treatment vs. hydrogen flow. Inspection reveals that desulfurization rate increases as the hydrogen flow increasing. The 67.6 % of the sulfur is removed at 650ml/min of hydrogen flow. When the

hydrogen flow is continued to increase, desulfurization rate would decrease. At the beginning, desulfurization rate increasing, which are mainly due to hydrogen flows increasing to external diffusion power increase, coal macromolecular hydrogenation reaction probability increased. At the same time, the gas diffusion rate of volatile is also accelerated from coal particles interior to the surface, the secondary reaction of Sulfide gas and residue is reduced. Then low level of desulfurization rate may be a result of "air holes" caused by large hydrogen flow (>650ml/min) [8]. As the reaction temperature increased, the coal began to soften to form molten state. The flow of feed gas increases to a certain value, which is through the molten resin layer, the deflection is produced to form many small "air holes". Then the reaction gas is directly outflow from the "air holes", as a result, the sulfur in coal cannot fully contact reaction with hydrogen, some of the sulfur radicals further polymerize to form stable and difficult removal of thiophene compounds.

Relative volume,cm3/g

Figures 3 shows pore size distribution of Hypy char sample at 550℃. 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 -0.01 0

300ml/min 600ml/min 900ml/min

20

40 60 80 Raduis,pore rangs,nm

100

Fig.3 porous ranges of char sample after conversion Can be seen from figure 3, the porous amount of char residue increased with gas flow increased, decreased after gas flow exceeding a certain number of values. The results show that as the gas flow rate is less than a certain value, the reaction of hydrogen and coal samples is more fully in the Hypy process, the pore inducing effect is better. When the hydrogen flow is more than the certain value, hydrogen and coal samples cannot have full access to reaction caused by the deflection forming "air holes". Then the reaction gas is directly outflow from the "air holes", as a result, the sulfur in coal cannot fully contact reaction with hydrogen, the micropore volume of residues is significantly reduced.

4.2 H2S concentration in the gas under Hydropyrolysis

Fig.4 H2S concentration in the gas under Hydropyrolysis ( Q: 60ml/min T: 750℃)

As can be seen from Figure 4, the escape curve of hydrogen sulfide showed two main peaks: 250~450℃low-temperature peak; 450~650℃ high-temperature peak. Hydrogen sulfide formation reaction has three types in temperature range 250 to 450℃: (1) Pyrite hydrogenation reaction; (2) Aliphatic sulfur side chain fracture reaction; (3) Thiophene heterocycles hydrocracking reaction. Pyrite hydrogenation reaction was complete under the temperature of less than 450℃. In these three reactions, aliphatic sulfur side chain fracture was most complex, however, reaction temperature is low, aliphatic sulfur side chain fracture was completely reduced or decomposed under the temperature of less than 400℃. The contribution of thiophene heterocycles hydrocracking reaction was least compared to pyrite hydrogenation and aliphatic sulfur side chain fracture reaction. Hydrogen sulfide formation reaction was mainly from thiophene sulfur and FeS sulfur desulfurization in temperature range 450 to 650℃. The reduction temperature of thiophene sulfur was high, more than 450℃. As a result, increase of reaction temperature on hydrogenation pyrolysis desulfurization reaction was favorable. However, the reaction of organic sulfur cannot be completely in temperature range 450 to 650℃. FeS hydropyrolysis reaction is very difficult under the temperature of less than 800℃. The contribution of high peak is mainly thiophene sulfur reduced. a

b

c

e

d

f

(a)450℃Hypy char; (b)550℃Hypy char;(c)650℃Hypy char; (d)750℃Hypy char (e) Pyrolysis char; (f) coal sample Fig.5 XPS spectra of sulfur element Characteristic peak of sulfide is shown in Table 2. From the XPS spectra of the sulfur element (Fig. 5), it can be seen that desulfurization rate in hydrogen atmosphere was higher than direct pyrolysis the desulfurization. With the temperature increasing, the desulfurization rate

increased, then declined again, which is due to the presence of mineral matter in coal []. Comparison table 2 and Figure 5 can be seen, coal sample contains large amounts of inorganic sulfur (Fig. 5 (f)). Compared Pyrolysis char (Fig. 5 (f)) to coal sample (Fig. 5 (e)), Sulfur content in pyrolysis char residue obviously decreased. Under 450℃, there exists still a small amount of inorganic sulfur and organic sulfur in Hypy char. However, compared with the pyrolysis char, the sulfur content in char is lowered (Fig. 5 (a)). When the temperature increased to about 550℃ ~650℃,

The inorganic sulfur of residual char is almost removed completely, the amount of

organic sulfur is also removed (Fig. 5 (b,c)). However, the temperature is higher than 650 ℃ with sufficient hydrogen, the inorganic sulfur in residues increases caused by the minerals in coal [7]. Table 2 Characteristic peak of sulfide [9,10] Characteristic peak, eV

Sulfide

162.7~163.0

Sulfide sulfur, Sulfide, mercaptans or benzene disulfide and series Organic sulfur compounds（Non-aromatic sulfur，including alkyl -, aryl - sulfide and polysulfides） Thiophene (aromatic sulfur) Sulfoxide, Sulfone Sulfate sulfur Inorganic sulfur

163.1~163.5 164.1~164.5 169.8~166.2 167.8~166.2 169.1~169.5 > 170

4.3 Reaction model validation

Fig.6 Effect of temperature on sulfur removal in coal In the hydrogen conditions, desulfurization experimental data was shown in figure 6, under the temperature rang 300～700℃. Can be seen from Figure 2, in the early stages of reaction, increasing temperature, desulfurization rate was significantly increased, sulfur in the coal seemed easy to remove. However, sulfur content of residual char of was used to analyze, it was found that the desulfurization ratio in the 500 ℃ temperature was better than 700℃. This was mainly because that sulfur of coal easily was converted to a stable heterocyclic compounds under high temperature, such as thiophene forming by organic sulfur being very difficult to decompose [11], which is an endothermic reaction. Therefore, increasing reaction temperature was not only enhancing desulfurization rate, but also promoted organic sulfur conversion between different forms, made desulfurization rate decline. Figure 6 data was treated in accordance to the equation (8) and (10), the results shown in

Figure 7. Can see from the figures, the random pore model could be better desulfurization Kinetics than the grain reaction model. Early in the desulfurization, the random pore model data was fairly consistent with experimental data. While the declination of the particle reaction model and experimental data was relatively large. In reaction latter, the two models and were better consistent with experimental data. The random pore model was always consistent with the experimental data in total reaction process. 1.2

0

1.2

0

随机孔模型 0.05

1

粒子模型

1

0.002

粒子模型

0.2

0.4

0.25

0.2

0.3

0

0.8

0.004

0.6

0.006

0.4

0.008

0.2

0.01

0.35 50

100

150

200

0

250

0.012

0

t/min

50

100 150 t/min

(a) 0.01

0.008 0.007

0

7

0.1

6

随机孔模型

5

粒子模型

0.005

0.3

0.004 0.4

0.003 0.002

随机孔模型

0.001

粒子模型

0

100

150 200 t/min

250

Ln(1‐x)

Ln(1‐x)

0.006

1‐(1‐x)(1|3)

0.2

50

0 0.1 0.2 0.3 0.4

4

0.5 3

0.6 0.7

2

0.8

0.5

1

0.6

0

0.9 1 0

300

30

60

90

120

150

t/min

(b)

(d)

8

0

随机孔模型

7

0.2

粒子模型

6

0.4

5

Ln(1‐x)

250

(b)

0.009

0

200

1‐(1‐x)(1|3)

0

1‐(1‐x)(1|3)

0.15

0.6

4 0.8

3

1‐(1‐x)(1|3)

0.6

随机孔模型

Ln(1‐x)

Ln(1‐x)

0.8

1‐(1‐x)(1|3)

0.1

1

2

1.2

1 0

1.4 0

30

60

90

t/min

120

150

(e) Fig.7 grain reaction and random pore model (a) 300℃; (b) 4000℃; (c) 500℃; (d) 500℃(e) 600℃; (f) 300~600℃

5、Kinetics In order to investigate the desulfurization activation energy of high-sulfur coal, the followed methods were used to obtain activation energy in this paper, the advantage of this method is not considering the reaction mechanism change.

dx = kF ( x) dt

（12）

Let integral tx

∫0

dt =

dx

x

∫0 kF(x)

k = A exp −

E RT

Bexp

So t x =

A

Where： B=

∫

x

0

E R T ，as Lnt = E + b x

RT

dx B ，b=Ln A F ( x）

9

9.5

8.5

9

8

8.5

Lntx

Lnt x

7.5 7

8

7.5

6.5 7

6

x=0.1,   Ea=34.16KJ.moll‐1

5.5

6

5 0.0007

x=0.3,   Ea=28.58KJ.moll‐1

6.5

0.0012

0.0017

0.001

0.0022

0.0012

0.0014

10.5

9.5

10

9

9.5

Lnt x

Lntx

10

8.5

9

8

x=0.5,   Ea=23.31KJ.moll‐1

x=0.7,   Ea=20.75KJ.moll‐1

7.5

7 0.001

0.0018

8.5

8 7.5

0.0016

T‐1/K‐1

T‐1 /K‐1

0.0012

0.0014

T‐1 /K‐1

0.0016

0.0018

0.001

0.0012

0.0014

0.0016

0.0018

T‐1 /K‐1

Fig.8 Calculation of activation energy at designated reaction fraction Figure 8 was shown that the desulfurization activation energy of coal was low at hydrogen atmosphere. Overall, the sulphur of Goudi high-sulfur coal was easy to removal. Yergey 与文献 18 has reported the same results. With increasing of the reaction fraction(x), the activation energy was lower. Increasing reaction temperature was beneficial to desulfurization of coal hydropyrolysis. At the same time, as the reaction temperature increased, the desulfurization process changed: the inorganic sulfur removal was gradually changed into organic sulfur removal. The desulfurization activation energy of organic sulfur is relatively small [7].

6. Conclusion

1）The desulfurization rate of high sulfur coal is related with temperature. Because there is the conversion of organic sulfur forms in coal under high temperature, the higher reaction temperature is not more conducive to the removal of sulfur in coal. Suitable desulfurization temperature is the range from 500 to 600℃. When desulfurization reaction is controlled by chemical reaction, the desulfurization activation energy is in the range 20 ~ 40 kJ.mol – 1. 2) The hydrogen sulfide releasing curves of in Hypy process show two peaks: The first peak at 250~450℃, which is mainly due to pyrite hydrogenation and aliphatic sulfur side chain fracture reaction; and the second peak at 450~650℃, which is mainly due to thiophene sulfur reduced 3) Compared the grain reaction model with the random pore model, the random pore model is more consistent with experimental data: Early, the random pore model data was fairly consistent with experimental data, and the grain reaction model was relatively deviation large with experimental data; Latter, the two models and were better consistent with experimental data. The random pore model was always consistent with the experimental data in the total reaction process.

Acknowledgment Project supported by the National Basic Research Prograin of China (Grant No.2005CB221202), Innovation Team Funds Scheme, the Natural Science Foundation of China (Grant No. 21006066) and the youth found of School Funds of Taiyuan University of Technology (TYUT).

References [1] Tian Zhengshan, WangQuankun, Bai Suzhen. Present situation and future progress of combusting desulfurization technology in high sulfur coal. Chemical Industry Times, 2009, 23(7): 53-56. [2] Xu qiang, Qian jin. Investigation on development and application of clean coal technologies in China. Coal Ash China, 2003, 4: 35-39. [3] Fu Zhixin, Guo Zhancheng, Wang Shen-xiang. Desulfurization of high sulfur coal by recycling additional gases during coking process. The Chinese Journal of Process Engineering, 2007, 7(5): 910-915. [4] Chen Haohan, Li Baoqing, Yang Jili, et al. Study of Desulphurization and Its Kinetics in Hydropyolysis of Yanzhou Bituminous Coal. Journal of Fuel Chemistry and Technology.1997, 25(2):114-117 [5] Guo Hanxian. Applied Chemical Engineering Kinetics.Beijing: Chemical Industry Press, 2003 [6] Levenspiel O. Chemical Reaction Engineering.New York: John Willey & Sons Inc.1999.，570-577 [7] LIU Junli , TANG Huiqing, GUO Zhancheng, Kinetica of removal of organicsulfur from semicoke under hydrogen atmosphere. Journal of chemical Industry and Engineering (China), 2004, 155(8): 1335-1440. [8] Zhang Guojie, Zhang Yongfa, Xie Kechang. Study on desulphurization of high sulfur coal in hydropyrolysis. Chemical Engineering (China), 2006, l 34(4): 55-58. [9] Chen Peng. Application of XPS in study forms of organic sulfur in macerals of Yanzhou coal. Journal of Fuel Chemistry and technology, 1997, 25(3): 238-241. [10] Zhu Zhenping, Cui Hong, Li Yunmei, et al. investigation of sulfur and nitrogen forms in two high sulfur coals. Environment Chemistry, 1995, 14(6): 483-488. [11] Li Bin, CaoYan, Zhang Jianmin, et al. The progress of release from coal pyrolysis and gasification progress. Coal Conversion, 2001, 24(3): 6-11. [12] Yergey A L，Lampe F W，Vestal M L et al． Nonisothermal kinetics Studies of the Hydrodesulfurization of Coal． Ind Eng Chem Process Des Develop，1974，13(3)：233-240 [13] Cypres R，Mingels W，Lardlnois J P，et al. Feasibility Study of the Hydropyrolysis of Coal，UER 14110[R]. Luxembourg. Commission of the European Communites,1992.

2010 International Pittsburgh Coal Conference Istanbul, Turkey, October 11-14, 2010 PREPARATION OF ALTERNATIVE FUEL FROM COMPOST AND COAL SLURRIES ZAJONC Ondrej, RACLAVSKY Konstantin, SKROBANKOVA Hana, JUCHELKOVA Dagmar, RACLAVSKA Helena VSB – Technical University Ostrava, Czech Republic, [email protected]

Introduction New strategies in municipal solid waste (MSW) management, i.e., a separate collection of the organic fraction (EU Directive 1999/31/EC, EU Directive 2008/98/EC) and a reduction of the biodegradable MSW fraction allocated in landfills (EU Directive 2003/33/EC), have favored the development of composting as a useful biotechnology in transforming organic wastes into suitable agricultural products. In order to meet the requirements of the Directive in the Czech Republic in by the year 2013 it will be allowed only 697 tons for deposition on landfills and it will be necessary to find other utilization for 2477 tons of waste. The produced compost which does not meet the requirements for quality (concentrations of trace metals, C/N ratio and content of PHs) can be utilized for energy generation. It is supported by Directive 2009/28/EC of the European Parliament and the Council of 23 April 2009 on the promotion of the use of energy from renewable sources and amending and subsequently repealing Directives 2001/77/EC and 2003/30/EC. One of the most important steps in the production of pellets is an increase of pellet density and at the same time preservation of their sufficient mechanical durability and low consumption of energy. The aim of this paper is to describe the preparation of suitable fuel mixture which has required mechanical properties using waste products.

Material The pellets were prepared from compost produced in the facility of Kompostarna Bruzovice, operated by Frydecka skladka Ltd., located in Moravskoslezske Beskydy Mts., Czech Republic. In this plant, the compost is produced by classical method, in a land-farming hall, in aerated and drained heaps frequently turned over in order to achieve good aerobic conditions. The compost is prepared from following components: saw dust 5%, grass 60%, horse manure 20%, waste from processing of potatoes 10% and vegetable waste 5%. The gross calorific value of compost is varying around 13.78 MJ/kg d.m. (dry matter). The direct energetic utilization of compost is complicated by its high moisture which has value around 50 % even at the end of composting process, depending at the storage conditions. The decrease of moisture required six months when it reached 13.3 % (Vergnoux et al. 2009). This long time is difficult from the point of view of storage capacities. The preliminary drying for production of pellets resulted in moisture ranging from 20 to 30 %. Optimization of compost moisture can be achieved by means of suitable additives (waste paper – cardboard). The data in Table 2 show that addition of 20 % of dry cardboard decreased moisture approximately by 20 %. The graph of the thermal degradation of compost in oxidizing atmosphere (air) up to 1000 oC is in Fig.1. The weight losses were determined from thermogravimetric (TG) curve. The total loss of weight was 61.5 %. Three temperature maxima were read on DTG curves: up to 200 oC – loss of water (19.7 wt %), 200 – 400 oC – oxidation of organic matter (loss 28.3 wt. %), and 400-600 oC – oxidation of cellulose, hemicelulose and lignine (loss 20.5 wt.%). During the process of composting humic acids are formed by transformation of carbohydrates. From analysis of major components it results that compost contained 31wt.% of lignine, 8wt.% of hemicelulose, 24 wt.% of cellulose and 7wt.% of

humic acids. From the point of view of pelletization technology, the particle size of input components is very important. Grain size analysis of compost is listed in Table 1. The two grain-size classes (0.5 – 1 mm and 0.25 – 0.5 mm) have the highest percentages – both approximately 26 % of particles. Table 1 Grain size analysis of compost – content of particles in individual classes (%) before milling and after milling. Grain size (mm)

> 2,5

2,0-2,5

1,6-2,0

1,0-1,6

0,5-1,0

0,25-0,5

0,16 -0,25

0,063-0,16

0,045- 0,063

W > Ga > Be > Nb > Mo > Sc > Y > La > Zn > Pb. Elements accumulated in coal are: Li, Na, K, Rb, Mg, Ca, Sr, Ba, V, Cr, Ni, No, Pb, Ag, B and Ge. Average content of ash on the 3 series of samples was 23.91%, 22.92% and 25.43% respectively. Lignite hygroscopic moisture has a constant value on the three samples, around 5%. Calorific value of lignite It was determined on each sample the gross (heat power) calorific value and lower (net) calorific value was calculated. The results - the average values - were: - heat power value: 2466, 2500 and 2470 kcal/kg, respectively - lower calorific value: 2107.4, 2135.3 and 2107.7 kcal/kg respectively. Limits of variation of lower calorific values for the analyzed samples are between 1700 and 2600 kcal/kg.

4

3.2 Laboratory experimental research on auto-heating and auto-oxidation of coal in quarries 3.2.1 Determination of oxidation capacity – auto-heating of lignite in quarries Since the lignite exploited from the eight quarries has different characteristics, a prior research was designed to establish quarries with the most liable to qualitative depreciation through oxidation (Krausz and coworkers, 2003). With this purpose, to establish of oxidative and auto-heating capacity, we used a method based on recording the changes in temperature of the mixture of coal samples under 0.2 mm size, and a hydrogen peroxide solution (hydrogen peroxide) at a concentration of 20% (Ionescu and coworkers, 1993). It was measured the initial temperature of the mixture and then, changes were recorded every minute, until the temperature of 50oC is reached (first stage of heatingoxidation). After reaching this temperature, the time intervals when the temperature increases by 10 degrees have been pursued and noted until the maximum temperature is reached (phase II). Shortly after reaching maximum temperature, the reaction begins to decrease in intensity, a continued cooling of hydrogen peroxide solution – coal mixture being observed. The times when temperature decreased with 10 degrees are recorded, until temperature value reach 50oC. The results showed that the higher the sample’s calorific value, the heating duration is lower, and vice versa; the lower coal quality, the auto-oxidation duration is greater, so establishing which are the two quarries (among the eight taken in the analysis) that are more important to have in view for the investigation of qualitative degradation phenomenon during lignite storage, because these quarries contain the highest availability for oxidation – ignition coals. These careers were Lupoaia and Roşiuţa. Measurements for the two quarries (and not a single) were done for a greater certainty of obtained data. 3.2.2. Determination of lignite characteristic alteration depending on storage duration, in laboratory conditions Three identical samples, in terms of quantity (5 kg) and sizes (2.5 - 5 mm), but with different moisture content were taken from each Lupoaia and Roşiuţa quarries. The first sample is in anhydrous state (after drying) and the other two with larger moisture, are obtained by spraying the sample with water. Thus the Lupoaia samples have: average carbon content of 58.58%, the ash content of 29.7%, hygroscopic moisture of 4.91%, and moistures for samples II and III of 17.50% and 28.10%. Samples from Roşiuţa have: carbon content of 57.15%, ash content of 36.4%, hygroscopic moisture of 3.81% and moistures for samples II and III of 16.30% and 29.80%. The 2x3 samples were stored under identical conditions, in cylinders made of cardboard and fitted with 10 mm diameter holes to provide a natural draft of air through the lignite granules. Temporal variation of qualitative parameters was monitored and the results are presented in Table 2.

5

Table 2 Sample

I Wi=0 II Wi=17,5% III Wi=28,1%

Qualitative parameters variation during storage Storage Parameters DAF DAF N H VDAF CDAF duration, % % % % days LUPOAIA QUARRY initial after 5 days after 10 days after 20 days after 30 days initial after 5 days after 10 days after 20 days after 30 days initial after 5 days after 10 days after 20 days after 30 days

0,70 0,72 0,76 0,77 0,80 0,70 0,67 0,66 0,61 0,65 0,70 0,65 0,68 0,70 0,71

3,52 3,36 3,38 3,36 3,40 3,52 3,63 3,77 3,89 3,92 3,52 3,71 3,79 3,93 3,99

Qsup Kcal/kg

41,60 41,45 41,40 41,00 41,00 41,60 41,38 41,27 40,75 40,15 41,6 41,45 41,4 41,00 39,50

58,58 58,45 58,33 58,00 57,83 58,58 58,33 58,21 57,93 57,60 58,58 57,90 57,88 57,69 57,45

3814 3801 3785 3751 3740 3814 3768 3744 3742 3739 3814 3765 3744 3739 3698

38,62 38,48 38,52 38,45 38,45 38,62 38,59 38,75 38,81 38,76 38,62 39,43 39,89 40,09 39,21

57,15 57,21 57,18 57,11 57,10 57,15 57,83 56,91 56,78 56,75 57,15 56,73 56,68 56,65 56,61

3371 3364 3342 3328 3328 3371 3345 3338 3337 3324 3371 3321 3315 3308 3303

ROŞIUŢA QUARRY I Wi=0 II Wi=16,3% III Wi=29,8%

initial after 5 days after 10 days after 20 days after 30 days initial after 5 days after10 days after 20 days after 30 days initial after 5 days after 10 days after 20 days after 30 days DAF

0,59 0,63 0,64 0,62 0,64 0,59 0,58 0,61 0,57 0,58 0,59 0,61 0,63 0,58 0,62

DAF

3,25 3,12 3,15 3,10 3,15 3,25 3,41 3,28 3,35 3,38 3,25 3,61 3,52 3,59 3,51

DAF

DAF

Notation in table: N – nitrogen content; H - hydrogen content ; V - volatile matter content; C i h carbon content; W – initial moisture; W – hygroscopic moisture; Qsup – superior (heat) caloric value

   

Findings: nitrogen content has a slightly variation, even after the 30 days storage period; hydrogen content, has a slightly decreasing trend for dried coal samples and increase in wetted samples, for both quarries; volatile matter content slightly decreases over time for dried samples and a slightly increase is observed with the moisture increase, recording the same type of variation for both quarries, carbon content decreases over time, especially since the initial moisture content of coal is higher, ebbing is of 1.28% for sample I, 1.67% for sample II and 1.93% for sample III in the case of Lupoaia lignite and much weaker (from 0.1 to 0.8%) in the case of Roşiuţa.

6

Changes in heat power calorific value are presented graphically in figures 1 and 2, showing the qualitative depreciation that suits to carbon’s content reduction. This depreciation is more important as the sample has higher moisture content. The decrease in calorific value is 1.95% respectively 1.28% for dried lignite, 1.97% respectively 1.39% for the samples II and 3.05% respectively 2.02% for the samples III. These results confirm that oxidation process is more intense on higher coal quality.

calorific value(kcal/kg)

The caloric value variation on storage time for Lupoaia lignite

3820 3800 3780 3760 3740 3720 3700 3680 3660 3640

Initial

after 5 days

after 10 days

after 20 days

after 30 days

Wi=0

3814

3801

3785

3751

3740

Wi=17,5%

3814

3768

3744

3742

3739

Wi=28,1%

3814

3765

3744

3739

3698

Period (days)

Figure1 - The calorific value variation depending on storage time and moisture content for Lupoaia lignite

calorific value(kcal/kg)

The caloric value variation on storage time for Roșiuța lignite

3380

3360

3340

3320

3300

3280

3260

Initial

after 5 days

after 10 days

after 20 days

after 30 days

Wi=0

3371

3364

3342

3328

3328

Wi=16,3%

3371

3345

3338

3337

3324

Wi=29,8%

3371

3321

3315

3308

3303

Period (days)

7

Figure 2 - The calorific value variation depending on storage time and moisture content for Roşiuţa lignite 3.2.3 Research to determine the influence of industrial storage time on qualitative parameters of lignite from Lupoaia quarry Research in quarry was conducted over a period of six months, covering the months from January to June, to follow the progress of phenomenon depending on climate and weather changes within three seasons. The study was conducted for both quarries, based on monitoring of production and supply of coal stored at each loader and establishing lower calorific values as the main quality parameter. Table 3

Quarry

Lignite characteristics Lignite characteristics Layer no.

Total moisture W t, %

Ash content anh A ,%

Lower calorific value, Qlow,

V VI VII VIII X V-VI VII-VIII X-XII

40,2 41,9 40,5 40,7 40,0 42,0 42,3 42,5

28,9 28,7 30,1 28,1 40,21 27,9 28,7 35,8

2264 2192 2218 2282 1825 2210 2183 1923

Lupoaia

Roşiuţa

kcal/kg

Characteristics of directly sampled lignite from exploited layers are shown in Table 3 and constitute the basic parameters for daily reporting of the production features. Based on recorded data, were calculated the average monthly values by weighting, based on daily deliveries and the regression equations were obtained. The decreasing trend of lower calorific value of lignite during storage is clearly shown by these equations and is also presented graphically in figures 3 and 4. Delivered coal quality variation depending on the storage time on the careers LUPOAIA

2200

calorific value [Kcal/Kg]

2100

y = -13,702x + 2094,9 R 2 = 0,8721

2000 y = -13,96x + 2086,3 2 R = 0,7

1900

y = -11,835x + 2099,9 R2 = 0,7522

1800

1700

y = -16,563x + 2116,2 2 R = 0,9015

y = -10.397x + 2117 R² = 0.5822

1600

y = -13,255x + 2083,1 2

R = 0,8874 1500 0

10 January Liniară (January)

20 February Liniară (Marth)

30 Marth Liniară (April)

40 April Liniară (May)

50 60 storage time [days] May Liniară (June)

June Liniară (February)

8

Calorific value [Kcal/Kg]

Figure 3 Monthly variation of lower caloric value depending on storage time for Lupoaia quarry lignite Delivered coal quality variation depending on the storage time on the career ROSIUTA 2300 2200 2100 2000 1900 1800

y = -11,984x + 2075,2 R2 = 0,8631

y= -16,799x + 2108,9 R2 = 0,8388

1700

y = -16,009x + 2117,4 R2 = 0,7997

1600

y = -14,628x + 2105,5 2 R = 0,8744

y = -13,415x + 2093,8 R2 = 0,894 y = -13,523x + 2095,9 R2 = 0,9356

1500 0

5 January May Liniară (Marth)

10

15 February June Liniară (April)

20

25 Marth Liniară (January) Liniară (May)

30

35

April Liniară (February) Liniară (June)

40

storage time [days]

Figure 4 Monthly variation of lower caloric value depending on storage time for Roşiuţa quarry lignite Stationary period of coal deposits is variable from 5 to 50 days. The start of depreciation process occurs when coal is stocked in the deposit (when the calorific value is of approximately 2100 kcal/kg), and after an average of 30 days stocking, quality loss is over 10%. As the storage time of coal increases, the monthly regression straight line position changes, achieving a considerable difference between them. During probation period, the climatic factor changes were pursued: air and soil temperature, atmospheric relative humidity, atmospheric pressure, wind speed and specific weather phenomenon such as fog, rain, snow depth (Bacalu, Krausz, 2009). Variations in these parameters during the analysis are presented in Table 4. Table 4

Month January February March April May

Climatic factor values from January to June Air Soil Air relative temperature, temperature, humidity, o o C C % -1,74 -2,07 94,35 -3,97 -3,65 83,43 4,26 4,97 71,35 10,01 11,19 70,30 19,59 22,34 59,03

Atmospheric pressure, mBar 992,23 995,58 997,04 1289,34 991,54

Wind speed, m/s 0,14 0,41 0,56 0,73 5,43 9

June 22,79 27,12 66,13 991,02 0,55 Average 8,49 9,98 74,10 1042,79 1,30 The most influential factors on the storage behavior of lignite are air and soil temperature, relative humidity and wind speed. Air and soil temperature retains the same increasing course in January-June period, from -10 to 30oC, with negative values in winter months (January and February). The effect favoring the oxidation process occurs in the period with higher temperatures, during the summer months. This variation is the inverse of relative air humidity levels, which have high values in winter and low values in summer. Corroborating the evolution of the three parameters with their effect on the oxidation process, we have the explanation for the fact that lignite depreciation is not according to month and season. Thus, in January, when the negative temperature are recorded, relative humidity has a maximum value (94.35%) favoring oxidation; in June, the maximum temperature (23oC for air and 27oC for soil) favoring oxidation, even if relative humidity drops to 66.13%. The wind speed records increases almost constant on average 0.3 m/s every month, from January to April and higher in May (5.43 m/s). The correlation between calorific value decrease and wind speed increase is positive and accountable through its effect on increasing the oxidation process. Based on the obtained data it was determined a depreciation coefficient also due to these factors, representing the percentage reduction in calorific value from the initial value to that which is at delivery, values that are presented in Table 5 for the two quarries. Table 5 Quarry

Depreciation coefficient values Specification

Caloric initial value, delivery Lupoaia kcal/kg Storage time, days Depreciation coefficient, % Caloric initial value, delivery Roşiuţa kcal/kg Storage time, days Depreciation coefficient, %

Month January February March April 2155 2114 2134 2153

May 2151

June 2139

1834

1972

1918

1888

1974

1943

27

11

13

15

8

10

14.90 2176

6.72 2168

10.12 2210

12.31 2200

8.23 2210

9.16 2210

1880

1922

1950

1951

2034

2004

15

13

11

11

5

6

13.60

11.35

11.76

11.32

7.96

9.32

Depreciation coefficient value is increasing with the storage time of lignite, having a maximum value of 14.9% for 27 days of storage for Lupoaia lignite and 13.6% after 15 days storage lignite from Roşiuţa. Although not all climatic factors have an influence as strong, cumulative effect of their action is evident. 4. Conclusions 1. Investigation conducted under laboratory conditions revealed a high oxidation capacity of lignite. Also it was established a different behavior for the components of the 10

studied samples, distinguishing in all situations that there is a low carbon content decrease over 1% and accordingly a calorific value decrease of more than 2%. 2. The results for the two quarries of Oltenia basin are similar, validating the accuracy of determinations. Because they contain lignite with the highest oxidative capacity, the results from their study are even more justified for lignite from the other quarries, where oxidative capacity of lignite is lower. 3. Climatic factors influence the process of lignite depreciation, but their simultaneous influence determined a mutual compensation of these influences. 4. The long-time storage of lignite determines its unquestionable quality depreciation, which begins to appear even after five days; keeping stocks for a period exceeding 45 days leads to a sharp decrease in calorific value due to coal oxidation process. The start of the oxidation process occurs from excavation and dumping in storage place, the phenomenon is increasing day by day and leads to formation of nuclei of fire in different areas of the store, reaching finally to coal auto-firing. 5. Research results show the need for planning excavation process based on deliveries, in order to obtain duration of the material storage below the 5 days, after registering a qualitative depreciation of lignite. With this purpose, it is required a more careful planning in time of the production or consideration of special measures for the disposal of lignite, in conditions which decrease its oxidation.

References Bacalu I., Krausz S. (coord.), 2009 – Study of the dependence between the qualitative parameters of lignite and its exploitation and storage conditions, Doctoral thesis, University of Petrosani. Bowman R., 1980 – Low temperature oxidation of bituminous coal. Infrared spectroscopic study of samples from a coal pile, Fuel, 59. Davidson R., 1990 – Natural oxidation of coal, IEA Coal Research. Ghethner J.S., 1987 – The mechanism of the low temperature oxidation of coal by O2: observation and separation of simultaneous reactions in situ FT-IR difference spectroscopy, Applied Spectroscopy. Huggins, F.E., 1980 – Mossbauer detection of goethite in coal and its potential as indicator of coal oxidation, International Journal of Coal Geology. Huggins, F.E., 1989 – Chemistry of coal weathering, Elsevier Science Publishers, Amsterdam. Huidu E., 1996 – Coal storage systems, Technical Publishing House, Bucharest. Ionescu C., Matei A., Traistă E., 1993 – General Chemistry, University of Petrosani P.H. Krausz S., Trotea T., Bacalu I., 2003 – Behavior of the lignite from Oltenia National Company at oxidation and self heating, Mining Revue nr.12. Matei A., Toth I., Traistă E., Bacalu I., 2001 – Coal yard self combustion monitoring by gaseous emission control, Proceeding of Third International Symposium Mining and Environmental Protection, Vrdnik, Serbia. Parks, B.C.,1963 – Chemistry of coal utilization, New York – London, Lowry P.H. 11

Pearson D.E., 1979 – Mineral matter as a measure of oxidation of a coking coal, Fuel, 44. *** - 2009, Data of Statistic National Institute, Bucharest.

12

NATURAL-GAS-LEVEL EMISSIONS WHEN BURNING NAPHTHA (WITHOUT WATER INJECTION) IN A COMMERCIAL GAS TURBINE USING THE LPP TECHNOLOGY, CREATING A “CLEAN POWER” ALTERNATIVE FOR AN INTEGRATED GASIFICATION COMBINED CYCLE (IGCC) POLYGEN PLANT

Leo D. Eskin, Richard J. Roby, Michael S. Klassen, Richard G. Joklik and Maclain M. Holton LPP Combustion, LLC 8940 Old Annapolis Road, Suite K Columbia, MD 21045

Abstract Emissions test results from operation of a commercial dry, low-emissions Capstone gas turbine demonstrate the commercial feasibility of using the Lean, Premixed Prevaporized (LPP) Technology to burn a range of light liquid fuels in a power generation gas turbine, without water injection, while simultaneously achieving ultra-low, natural-gas-level emissions for NOx, CO and particulates. The presented test results have significant implications for future Integrated Gasification Combined Cycle (IGCC) plants. The Integrated Gasification Combined Cycle technology, as currently defined, couples a complex coal gasification process plant with a custom, coal syngasfired, combustion turbine combined cycle power plant. The IGCC process is a two-stage combustion design with gas cleanup between the stages. The first stage employs a gasifier where partial oxidation of the coal occurs. The second stage utilizes the gas turbine combustor to complete the combustion with the gas turbine/combined cycle (GT/CC) technology. Due to the impracticality of storing significant quantities of the coal syngas, it is necessary to ensure that the combustion turbine remains operational whenever the gasification plant is running. The shutdown of the combustion turbine requires immediate shutdown of the gasification plant. In addition, it is difficult to operate the gasification plant at part load; hence it is preferable to run the combustion turbine in a base load configuration. These are significant operating limitations. The current test results demonstrate the commercial viability of combining the gasto-liquid technology (GTL), e.g. Fischer-Tropsch synthesis or similar, and the Lean Premixed Prevaporized combustion technology to create a much more robust power generation system. The GTL process is a method whereby coal syngas is transformed into one or more forms of liquid fuel. These coal liquids can include diesel fuel, kerosene and naphtha (among others). The conversion of coal syngas to liquids is a well-known process and has been utilized for many years. The LPP process transforms a 27th Intl. Pittsburgh Coal Conference

1

October 2010 - Istanbul, Turkey

wide variety of liquid fuels into a substitute natural gas (or LPP gas) which the current results clearly show can be burned in conventional natural gas dry low emissions combustion hardware, precluding the use of water or steam to achieve low criteria pollutant (NOx, CO and PM) emissions levels. By combining the LPP combustion technology with the GTL process, IGCC operation is made much more flexible, dependable, and the overall economics are improved. Three alternative IGCC design scenarios are presented which would allow the IGCC plant to use a standard, natural gas fired combustion turbine together with an LPP skid to provide increased operating flexibility for the plant and to reduce the plant capital equipment cost. Emissions test results for a Capstone gas turbine are presented for both naphtha and bioethanol as a liquid fuel source for the LPP skid. The tests were conducted at the LPP Combustion facility located in Columbia, MD (just north of Washington, DC) and the generated power was sold to Baltimore Gas and Electric as part of a net metering agreement. Further gas turbine testing is in progress using additional liquid fuels such as diesel and biodiesel. Introduction The Integrated Gasification Combined Cycle (IGCC) technology has been recognized as an efficient and clean method of converting coal (or other carbonbased feedstock) into electricity, process heat, high-value liquid fuels and other chemicals. Several IGCC plants are now operating commercially both within the United States and internationally. While a number of IGCC power and polygeneration plants are currently under construction (notably in China), more widespread adoption of coal gasification and the IGCC technology has been hindered by overall plant economics, and by the inherent complexity associated with building and operating such facilities. The gasifier in a commercial IGCC plant is a very large device which generates syngas at high temperature and pressure (approx. 400 psia and 2600 F), and the gasifier has a large thermal mass. Hence, it is desirable to operate the plant at a stable base-load condition over a long period of time. Unfortunately, reduced demand for electrical power (at night and on weekends) leads to inefficient and less economical plant operation during these low-load time periods, due to decreased part-load performance of both the gasification block and the gas turbine power block. A typical coal-based IGCC plant is shown diagrammatically in Figure 1 below. Coal is partially oxidized by oxygen (or air) at high pressure and temperature, and the resulting syngas stream is cooled, cleaned of impurities (and potentially CO2) and burned in a combustion gas turbine. As shown in Figure 1, gas turbine modifications are normally required to accommodate the increased volumetric flow rate for the syngas, which occurs because of the reduced heating value of the syngas (125-250 BTU/scf LHV) as compared to the heating value of natural gas (800-1,000 BTU/scf LHV). These

27th Intl. Pittsburgh Coal Conference

2

October 2010 - Istanbul, Turkey

custom hardware changes increase the gas turbine capital and maintenance costs and restrict the turbine operation. As part of an overall effort to improve reliability, operability, and plant economics, a broad range of IGCC design and optimization studies have been conducted by the Electric Power Research Institute1,2 (EPRI) and by the United States Department of Energy3 (DOE), under the auspices of the Vision 21 program4,5,,among others. Common to all of the above studies, and common to the existing commercial IGCC facilities, is the requirement that the combustion turbine be fueled by the syngas produced in the gasification block. This design constraint couples the operation of the gasification block and the gas turbine power block, due to the impracticality of storing large quantities of syngas, and decreases the overall plant availability and reliability. Practical considerations of a large power generation facility, such as load following and part load operation, become much more difficult with the added complexity and operational requirements of the gasification plant. IGCC Plant Performance Enhancement Using the LPP Technology The process described herein proposes to combine the gas-to-liquid technology (GTL), e.g. Fischer-Tropsch synthesis or similar, and the Lean, Premixed Prevaporized (LPP) combustion technology with the coal gasification technology to create a much more robust power generation and co-production system. This combination would burn the coal-derived liquids using the same sophisticated combustor hardware used today. This technology has dramatically lowered the pollutant emission levels from natural gas power plants over the last 20 years. The emissions using the coal-derived fuels would be similar to those of natural gas. The GTL process is a method whereby syngas is transformed into one or more forms of liquid fuel. These liquid fuels, known generically as coal liquids or F-T liquids, may include diesel fuel, kerosene and naphtha, among others. The conversion of syngas to liquids is a proven technology that has been in

1

Evaluation of Alternative IGCC Plant Designs for High Availability and Near Zero Emissions: RAM Analysis and Impact of SCR. EPRI, Palo Alto, CA: 2005. 1010461. Integrated Gasification Combined Cycle (IGCC) Design Considerations for High Availability— Volume 1: Lessons from Existing Operations. EPRI, Palo Alto, CA: 2007 1012226. 2

3

http://www.fossil.energy.gov/programs/powersystems/vision21/

4

“Topical Report – Task 1 Topical Report, IGCC Plant Cost Optimization,” Gasification Plant Cost and Performance Optimization, United States Department of Energy, National Energy Technology Laboratory, Contract No. DE-AC26-99FT40342, May 2002. 5

“Topical Report – Task 2 Topical Report, IGCC Plant Cost Optimization,” Gasification Plant Cost and Performance Optimization, United States Department of Energy, National Energy Technology Laboratory, Contract No. DE-AC26-99FT40342, September 2003.

27th Intl. Pittsburgh Coal Conference

3

October 2010 - Istanbul, Turkey

commercial use around the world for many years6,7,8. The recently invented LPP process transforms a wide variety of liquid fuels into a vaporized fuel stream (or LPP Gas™) which may be burned in conventional, natural gas, dry low emissions combustion hardware, precluding the use of water or steam to achieve low criteria pollutant (NOx, CO and PM) emissions levels. By combining the LPP combustion technology with the GTL process, IGCC operation is made much more flexible, dependable, and the overall economics is improved.

Air or O2 Coal

Water/Steam from Syngas Cooler

Clean Gasification Syngas Slag

Sulfur, CO2

Ambient Air

Hot Exhaust Gas

Modified Gas Turbine

Cold Exhaust Gas to Stack

Heat Recovery Steam Generator

High Pressure Steam

Electrical Power

Electrical Power

Steam Turbine Low Pressure Steam to Condenser

Figure 1: Simplified Typical Coal-based IGCC Plant Process Flow Diagram LPP Technology and Liquid Fuel Combustion Overview Traditionally, spray diffusion combustors have been employed in gas turbines that operate on liquid fuels such as fuel oil #1 and fuel oil #2. However, this diffusion mode of operation tends to produce unacceptable levels of NOx emissions. The current technology for burning liquid fuels in gas turbines is to use water and/or 6

“Topical Report – Volume I, Process Design – Illinois No. 6 Coal Case with Conventional Refining”, Baseline Design/Economics for Advanced Fischer-Tropsch Technology, U.S. Department of Energy, Contract Number DE-AC22-91PC90027, October, 1994. 7

“Topical Report – Volume IV, Process Flowsheet (PFS) Models”, Baseline Design/Economics for Advanced Fischer-Tropsch Technology, U.S. Department of Energy, Contract Number DE-AC2291PC90027, October, 1994. 8

“Topical Report VI – Natural Gas Fischer-Tropsch Case, Volume II, Plant Design and ASPEN Process Simulation Model”, Baseline Design/Economics for Advanced Fischer-Tropsch Technology, U.S. Department of Energy, Contract Number DE-AC22-91PC90027, August, 1996.

27th Intl. Pittsburgh Coal Conference

4

October 2010 - Istanbul, Turkey

steam injection with conventional diffusion burners. Emissions levels for a typical “state of the art” gas turbine, such as a GE 7FA burning fuel oil #2 in diffusion mode with water/steam injection, are 42 ppm NOx and 20 ppm CO9. Water/steam injection has a dilution and cooling effect, lowering the combustion temperature and thus lowering NOx emissions. But at the same time, water/steam injection is likely to increase CO emissions as a result of local quenching effects. Thus, the “wet” diffusion type of combustion system for liquid fuels must trade off NOx emissions for CO emissions. In recent years, stringent emissions standards have made lean, premixed combustion more desirable in power generation and industrial applications than ever before, since this combustion mode provides low NOx and low CO emissions without water addition. Lean, premixed combustion of natural gas avoids the problems associated with diffusion combustion and water addition. Thus, lean, premixed combustion is the foundation for modern Dry Low Emissions (DLE) gas turbine combustion systems. When operated on natural gas, DLE combustion systems provide NOx and CO emissions of 25 ppm or less with no water addition. However, these systems cannot currently operate in premixed mode on liquid fuels because of autoignition and flashback within the premixing section. Plee and Mellor10 characterized autoignition of the fuel/air mixture in the premixer as an important factor that causes flashback in practical combustion devices. Autoignition of the fuel/air mixture occurs before the main combustion zone, when the ignition delay time of the fuel/air mixture is shorter than the mean residence time of the fuel in the premixer. Autoignition especially occurs with the higherorder hydrocarbon fuels, such as fuel oils, which have shorter ignition delay times compared to natural gas11. The short ignition delay times of vaporized higher hydrocarbons have proven difficult to overcome when burning in lean, premixed mode. Nevertheless, in order to overcome high NOx levels produced by spray combustion, gas turbine designers still desire to use lean, premixed, prevaporized (LPP) combustion. Several approaches have been reported in the 12,13,14,15,16,17,18,19 to overcome flashback and autoignition in the premixers literature

9

Davis, L. B. and Black, S. H., 2000, “Dry Low NOx Combustion Systems for GE Heavy-Duty Gas Turbines”, GER 3568G. 10

Plee, S. L. and Mellor, A. M., 1978, “Review of Flashback Term Reported in Prevaporizing/Premixing Combustors”, Combust. Flame, 32. pp. 193-203. 11

Oumejjoud, K., Stuttaford, P., Jennings, S., Rizkalla, H., Henriquez, J., Chen, Y., 2005, “Emission, LBO and Combustion Characterization for Several Alternative Fuels”, Proc. ASME Turbo Expo 2005, Paper# GT2005-68561. 12

Maier, G. and Wiitig, S., 1999, “Fuel Preparation and Emission Characteristics of a Pressure Loaded LPP Combustor”, 30th AMA Fluid Dynamics Conference, AIAA- 99-3774.

13

Imamura, A., Yoshida, M., Kawano, M., Aruga, N., Nagata, Y. and Kawagishi, M., 2001, “Research and Development of a LPP Combustor with Swirling Flow for Low NOx”, 37th Joint Propulsion Conference & Exhibit, AIAA-2001-3311.

27th Intl. Pittsburgh Coal Conference

5

October 2010 - Istanbul, Turkey

of LPP combustors. These approaches attempt to achieve low NOx emissions by designing premixers and combustors that permit rapid mixing and combustion before spontaneous ignition of the fuel can occur. In most of the work reported on LPP combustion systems in the literature, the fuel is sprayed directly into the premixer so that the liquid fuel droplets vaporize and mix with air at lean conditions. Typically, swirlers with multi-port liquid fuel injection systems are employed for better fuel/air mixing20. However, unlike these attempts to alter hardware, there has been no reported work on altering fuel combustion characteristics in order to delay the onset of ignition in lean, premixed combustion systems. In the present solution, vaporization of the liquid fuel in an inert environment has been shown to be a technically viable approach for LPP combustion. As described by Roby et al.21, a fuel vaporization and conditioning process has been developed and tested22 to achieve low emissions (NOx and CO) comparable to those of natural gas while operating on liquid fuels, without water or steam addition. In this approach, liquid fuel is vaporized in an inert environment to create a fuel vapor/inert gas mixture, LPP Gas™, with combustion properties similar to those of natural gas. Premature autoignition of the LPP Gas™ was controlled by the level of inert gas in the vaporization process. Tests conducted in both atmospheric and high pressure test rigs utilizing typical swirl-stabilized burners (designed for natural gas) found operation similar to that achieved when burning 14

Ikezaki, T., Hosoi, J. and Hidemi, T., 2001, “The Performance of the Low NOx Aero Gas Turbine Combustor Under High Pressure”, ASME paper 2001 -GT-0084.

15

Lin, Y., Peng, Y. and Liu, G., 2004, “Investigation on NOx of a Low Emission Combustor Design with Multihole Premixer-Prevaporizer”, Proc. ASME Turbo Expo 2004, Paper# GT2004- 53203. 16

Lee, C., Chun, K. S., Locke and R. J., 1995, “Fuel-Air Mixing Effect on NO X Emissions for a Lean Premixed-Prevaporized Combustion System”, 33rd AIAA Aerospace Sciences Meeting and Exhibit, Paper# AIAA-95-0729. 17

Michou, Y., Chauveau, C., Gijkalp, I. and Carvalho, I. S., 1999, “Experimental Study of Lean Premixed and Prevaporized Turbulent Spray Combustion”, 37th AIAA Aerospace Sciences Meeting and Exhibit, AIAA 99–0332. 18

Hoffmann, S., Judith, H. and Holm, C., 1998, “Further Development of the Siemens LPP Hybrid Burner”, ASME International Gas Turbine & Aeroengine Congress & Exhibition, ASME 98-GT-552. 19

Mansour, A., Benjamin, M., Straub, D. L. and Richards, G. A., 2001, “Application of Macrolamination Technology to Lean, Premixed Combustion”, ASME J. of Eng. Gas Turb. Power, 123, pp. 796–802. 20

Lin, Y., Peng, Y. and Liu, G., 2004, “Investigation on NOx of a Low Emission Combustor Design with Multihole Premixer-Prevaporizer”, Proc. ASME Turbo Expo 2004, Paper# GT2004- 53203. 21

Roby, R. J., Klassen, M. S. and C. F. Schemel, 2006, “System for Vaporization of Liquid Fuels for Combustion and Method of Use”, U.S. Patent, #7,089,745 B2. 22

P. Gokulakrishnan, M. J. Ramotowski, G. Gaines, C. Fuller, R. Joklik, L. D. Eskin, M. S. Klassen and R. J. Roby, 2008, "Experimental Study of NOx Formation in Lean, Premixed, Prevaporized Combustion of Fuel Oils at Elevated Pressures", Journal of Engineering for Gas Turbines and Power, Vol. 130.

27th Intl. Pittsburgh Coal Conference

6

October 2010 - Istanbul, Turkey

natural gas. Emissions levels were similar for both the LPP Gas™ fuels (fuel oil #1 and #2, Biodiesel and F-T synthetic JP-8) and natural gas, with any differences in NOx emissions ascribed to fuel-bound nitrogen present in the liquid fuel. Also, tests showed that the LPP combustion system helps to reduce the NOx emissions by facilitating stable combustion even at very lean conditions when using liquid fuels. Extended lean operation was observed for the liquid fuels due to the wider lean flammability range for these fuels compared with natural gas. An added advantage of the fuel vaporization and conditioning process is the ability to switch between LPP Gas™ and natural gas ‘on-the-fly’ in the same combustor, without significantly affecting the flame stability. Elevated pressure tests were conducted on a full temperature, full pressure combustor test stand capable of supplying combustor air at typical compressor discharge temperatures and pressures. During these high pressure gas turbine burner tests, the liquid fuel was supplied in gaseous form from the LPP liquid fuel vaporizer skid shown in Figure 2. The testing involved a study of emissions and combustion characteristics, such as flame stability and lean blow-out limits. The tests were performed at typical compressor discharge temperatures. For the high pressure tests, typical compressor discharge pressures were also used. The same fuel nozzle used for natural gas testing was also used for liquid fuel testing on LPP Gas™ without any modifications.

Figure 2: LPP liquid fuel vaporizer skid used for gas turbine burner testing at elevated pressures. Figure 3 shows NOx and CO emissions at full load conditions for both natural gas and fuel oil #2. During the testing, emissions and dynamics data were taken over a range of lean equivalence ratios from approximately 0.75 to the lean blow-off (LBO) limit. However, the emissions data is plotted against measured exhaust gas temperature in order to provide a common temperature reference. The lowest 27th Intl. Pittsburgh Coal Conference

7

October 2010 - Istanbul, Turkey

temperature data points shown in Figure 3 reflect the experimentally observed LBO limit. Figure 3 shows that fuel oil #2 LPP Gas™ has an extended LBO limit compared to natural gas and thus can achieve NOx emissions nearly as low as natural gas despite the fuel-bound nitrogen. Figure 3 also shows that the crossover point between NOx and CO emissions extends to lower temperatures (and therefore lower equivalence ratios) for fuel oil #2 LPP Gas™ as compared to natural gas. As can be seen from the figure, fuel oil #2 LPP Gas™ showed increased flame stability and an extended LBO limit at lower temperatures (equivalence ratio) compared to natural gas. Figures 4 and 5 show similar results (using the same burner at atmospheric-pressure) for a variety of fuels, including a F-T derived jet fuel (S-8).

NOx & CO - ppmvd (at 15% O2)

45 40

NOx (Fuel Oil #2) CO (Fuel Oil #2)

35

NOx (Natural Gas ) CO (Natural Gas )

30 25 20 15 10 5 0 1100

1200

1300

1400

Exhaust Temperature (K) Figure 3: Comparison of NOx & CO emissions measurements for fuel oil #2 and natural gas as a function of measured exhaust gas temperature for a single fuel nozzle at Solar Turbines Taurus 60 full load conditions (100%). Combustion air temperature was 648 K, combustor pressure was 12.6 atm, and fuel dilution was 5:1 (molar basis).

27th Intl. Pittsburgh Coal Conference

8

October 2010 - Istanbul, Turkey

CENTAUR 50 DATA (1 ATM) 50 NOx - ppmvd (at 15% O2)

Fuel Oil #2

40

Fuel Oil #1 Na tura l Ga s Biodie se l B100 (SME)

30

Etha nol (ASTM D-4806) Na phtha (Pe troleum) S-8 (FT-GTL)

20

JP-8 (4-12-07)

10 0 1500

1600

1700

1800

1900

2000

Exhaust Temperature (F)

Figure 4: Comparison of NOx emissions measurements for fuel oil #2, fuel oil #1, soy methyl ester biodiesel, ethanol, naphtha, synthetic JP-8 (S-8), JP-8 and natural gas as a function of measured exhaust gas temperature for a single fuel nozzle at Centaur 50 full load conditions (100%). Combustion air temperature was 627 K, combustor pressure was 1 atm CENTAUR 50 DATA (1 ATM) 25 CO - ppmvd (at 15% O2)

Fue l Oil #2

20

Fue l Oil #1 Na tura l Ga s Biodie se l B100 (SME)

15

Etha nol (ASTM D-4806) Na phtha (Pe trole um ) S-8 (FT-GTL)

10

JP-8

5 0 1500

1600

1700 1800 Exhaust Temperature (F)

1900

2000

Figure 5: Comparison of CO emissions measurements for fuel oil #2, fuel oil #1, soy methyl ester biodiesel, ethanol, naphtha, synthetic JP-8 (S-8), JP-8 and natural gas as a function of measured exhaust gas temperature for a single fuel nozzle at Centaur 50 full load conditions (100%). Combustion air temperature was 627 K, combustor pressure was 1 atm. 27th Intl. Pittsburgh Coal Conference

9

October 2010 - Istanbul, Turkey

Gas Turbine LPP Fuel Testing Testing of the LPP Combustion System was also performed using both naphtha and bioethanol to fuel a Capstone C30 gas turbine. The Capstone C30 microturbine is a 30 kW power generation system designed for use with natural gas. The C30 initially ignites and operates in diffusion mode until approximately 22 kW, when it switches to premixed mode. In premixed combustion mode, it produces nominally 9 ppm NOx and 30 ppm of CO. The combustion system operates at approximately 4 atm. An LPP system was developed that interfaces with the natural gas fuel feed, allowing for operation on LPP fuel, natural gas or a mixture of these fuels. Figure 6 shows the test facility developed for the C30 testing.

Figure 6: Capstone C30 Test Facility used to evaluate emissions for various fuels using the LPP Combustion Technology. NOx and CO emissions were measured for naphtha, pure ethanol and ethanol/water blends and compared to baseline operation on methane. NOx emissions when running on either naphtha or ethanol are comparable to or less than the emissions on methane (see Figure 7), and show a trend of decreasing emissions as the water content in the ethanol increases. At a water content greater than 35% in the ethanol, stable combustion could not be maintained. CO emissions with naphtha are equal to or lower than those with methane for both diffusion mode operation and for premixed mode operation (see Figure 8). CO emissions with ethanol are equal to or higher than those with methane for diffusion mode operation, and show a trend of lower emissions as the ethanol water content increases. CO emissions with ethanol are approximately equal to or lower than methane for premixed mode operation (see Figure 8). 27th Intl. Pittsburgh Coal Conference

10

October 2010 - Istanbul, Turkey

80 Baseline CH4

70

100% EtOH 95% EtOH

NOx [ppm] @15% O2

60

90% EtOH

50

65% EtOH Naphtha

40 30 20 10 0 0

5

10

15

20

25

30

Load [KW]

Figure 7: C30 NOx emissions for naphtha, pure bioethanol, blends of 95/5%, 90/10% and 65/35% bioethanol with water, and methane as a function of gas turbine load. The turbine operates in diffusion mode for loads below 25 KW, and operates in premixed mode at 25 KW and above. 250 Baseline CH4 100% EtOH

200

95% EtOH CO [ppm] @15% O2

90% EtOH

150

65% EtOH Naphtha

100

50

0 0

5

10

15 Load [KW]

20

25

30

Figure 8: C30 CO emissions for naphtha, pure bioethanol, blends of 95/5%, 90/10% and 65/35% bioethanol with water, and methane as a function of gas turbine load. The turbine operates in diffusion mode for loads below 25 KW, and operates in premixed mode at 25 KW and above. 27th Intl. Pittsburgh Coal Conference

11

October 2010 - Istanbul, Turkey

These tests demonstrate that the LPP Combustion System is capable of burning naphtha (a fuel readily available as a 25%-40% product stream from the FischerTropsch process) and ethanol (an important renewable liquid fuel), in a gas turbine combustor with NOx and CO emissions similar to those obtained from operation on natural gas. These results were obtained using a commercial DLE gas turbine combustion system designed for use with natural gas with no modifications to the combustion hardware. The pollutant emission levels achieved are much lower than can be obtained when using these fuels in conventional spray diffusion mode, which is how liquid fuels are burned today in gas turbines or reciprocating engines. The LPP Combustion System has demonstrated natural gas level emissions from naphtha without the need for expensive post-combustion cleanup such as selective catalytic reduction (SCR). An added advantage of the LPP fuel vaporization and conditioning process is the ability to achieve fuelinterchangeability of a natural gas-fired combustor with liquid fuels. Improved Plant Heat Rate While beneficially reducing NOx emissions, water injection used for NO control in traditional spray flame (diffusion) combustors incurs a substantial negative impact on the efficiency and maintenance of a liquid-fueled gas turbine. Water addition reduces the NOx emissions by reducing the flame temperature, but this in turn also reduces the Brayton Cycle thermodynamic efficiency of the gas turbine. In addition, the energy required to vaporize the injected water further reduces the gas turbine efficiency. Lastly, the reduced firing temperature reduces the exhaust temperature of the gas turbine for liquid-fueled operation, which in turn reduces the amount of steam production for a combined cycle plant. All of these effects serve to increase the net plant heat rate (and cost of operation) for a traditional liquid fueled gas turbine combined cycle power plant. In contrast, a combined cycle plant which is fueled by prevaporized fuel is not subject to the above losses. A gas turbine fueled in this way may be operated at the full natural gas firing temperature, thereby avoiding the above reductions in efficiency. Since there is no water injection with the LPP System, that vaporization loss is also avoided. Combined cycle performance calculations have shown that for a typical single pressure level heat recovery steam generator (HRSG) and GE Frame 7EA class gas turbine, one can expect to achieve at least a two percent (2%) improvement in the overall combined cycle plant heat rate when burning liquid fuel in LPP mode as compared against burning the same liquid fuel in traditional spray-flame diffusion combustors. These calculations were made using the GateCycle™ 23 power plant modeling software, and the losses associated with a fully integrated LPP system were included in the calculation of the net plant heat rate. This level of heat rate improvement is quite substantial, and represents an annual fuel savings of over five million dollars for base load operation of a GE Frame 7EA combined cycle plant (126 MW).

23

www.gepower.com/prod_serv/products/oc/en/opt_diagsw/gatecycle.htm

27th Intl. Pittsburgh Coal Conference

12

October 2010 - Istanbul, Turkey

LPP/IGCC Integration Scenarios Integration of the LPP and IGCC technologies can take place through several potential scenarios, as described below. In Scenario #1 (shown in Figure 9), it is proposed to enhance the coal gasification portion of the IGCC process plant with a GTL process plant to transform the coal syngas into coal liquids. These coal liquids could then be readily stored in tanks, for use as necessary. In this scenario, it is further proposed to replace the syngasfired combustion turbine used in a typical IGCC design with a conventional naturalgas fired combustion turbine, combined with an LPP skid to transform the coal liquid fuels into LPP Gas™ which will be burned by the conventional dry low emissions combustion turbine. By creating coal liquids, the gasification block would no longer require continuous operation of the combustion turbine, and the gasification block could maintain full base load operation, regardless of the power generated by the power block. If the combustion turbine load is reduced or removed altogether, the excess coal liquids produced would be stored as necessary in nearby tanks, or would be distributed via pipeline, truck or train. This configuration provides the widest load-following capability for the power block, with no minimum power generation requirement. Truck, Rail or Pipeline

Air or O2

Tanks

Clean Gasification Syngas

Coal

Slag

Gas to Liquid Coal Conversion Liquid

Sulfur CO2 LPP Gas

Hot Exhaust Gas Water/Steam from Syngas Cooler

Heat Recovery Steam Generator

Standard Gas Turbine

Cold Exhaust Gas to Stack

LPP Skid

Ambient Air Electrical Power

High Pressure Steam

Steam Turbine

Electrical Power

Low Pressure Steam Figure 9: Scenario #1: Integrated LPP Gas™ IGCC Plant

27th Intl. Pittsburgh Coal Conference

13

October 2010 - Istanbul, Turkey

In Scenario #2 (shown in Figure 10), the coal gasification portion of the IGCC process plant is enhanced with a GTL process plant to transform a portion (but not all) of the coal syngas into coal liquids. The syngas stream that was not converted to liquid would be burned in one or more syngas-fired combustion turbines, sized for base load operation during low plant load conditions (at night, etc.). The coal liquids would be consumed by standard combustion turbine hardware burning LPP Gas™, which would operate in “peaking mode” as necessary, and would allow the overall plant to respond to electrical load changes without having to change the rate of production of the syngas. In Scenario #3 (shown in Figure 11), it is proposed to completely decouple the gasification/GTL plant and the power plant. The coal liquids would be produced at the gasification/GTL plant and would be shipped to stand-alone combustion turbines that are equipped with the LPP technology. This would provide the added benefit of allowing the gasification/GTL plant to be sited at any location, including a location in close proximity to the coal (or other feedstock) source. A site within close proximity to the coal source would reduce the transportation cost for the coal, and would facilitate disposal of the slag waste product resulting from the gasification plant. Air or O2 Coal

Gasification

Clean Syngas

Truck, Rail or Pipeline

Coal Liquid Tanks

Gas to Liquid Conversion

LPP Skid

Slag Sulfur CO2 Ambient Air Ambient Air

Modified Gas Turbine

Hot Exhaust Gas

Standard Gas Turbine Water

Cold Exhaust Gas to Stack

Heat Recovery Steam Generator Water Heat Recovery Steam Generator Electrical Power

Steam Turbine

Cold Exhaust to Steam

High Pressure Steam

Electrical Power

Low Pressure Steam

Steam Turbine Low Pressure Steam

Figure 10: Scenario #2: Integrated LPP Gas™/Syngas IGCC Plant

27th Intl. Pittsburgh Coal Conference

14

October 2010 - Istanbul, Turkey

Air or O2 Coal

Gasification

Gasification/CTL Plant

Clean Syngas

Tanks

Gas to Liquid Coal Liquid Conversion

Slag Sulfur CO2

Power Generation Plant

Truck, Rail or Pipeline

LPP Skid

Heat Recovery Steam Generator

LPP Gas

Ambient Air

Standard Gas Turbine

Cold Exhaust Gas to Stack

Water

HP Steam Hot Exhaust Gas

Electrical Power

Electrical Power

Steam Turbine LP Steam

Figure 11: Scenario #3: Non-Integrated LPP Gas™ IGCC Plant Additional Benefits and Synergies of the LPP/IGCC Integration The combination of the coal gasification process with the LPP process also offers the following advantages, in addition to those described above: 1. Beneficial use of waste Nitrogen (N2) produced by the air separation unit portion of the coal gasification plant: Most coal gasifiers use high purity oxygen (O2) to partially oxidize the coal and create the coal syngas. This oxygen is produced by an air separation unit (ASU) that separates the oxygen and nitrogen from ambient air. The nitrogen produced by the ASU is considered to be a waste product stream, and is injected into the clean syngas burned by the combustion turbine in an attempt to reduce NOx emissions by the combustion turbine. However, nitrogen gas can be used by the LPP process to create the LPP Gas™. Hence, by using the waste nitrogen already available from the ASU, the energy requirements of the LPP process are substantially reduced. Note that the low NOx combustion hardware present in natural gas combustion turbines does not require the use of supplemental nitrogen as is the case for syngas combustion hardware. 2. Significant improvement in the IGCC value proposition may be realized by using lower value FT liquids (such as naphtha) as a feedstock for the LPP process, in place of the higher value coal liquid product streams such as kerosene or diesel fuel: Naphtha comprises a sizable portion (30-45%) of the total output stream of a Fischer27th Intl. Pittsburgh Coal Conference

15

October 2010 - Istanbul, Turkey

Tropsch based coal liquids plant. Unsuitable for use as a transportation fuel, the monetary value of naphtha is expected to significantly decline as commercial production of coal liquids increases in the future. Since naphtha represents a very good feedstock for the LPP process, the LPP technology will be able to convert the low value liquid stream into a high value LPP Gas™ stream, improving overall plant economics. 3. It may be possible to substantially reduce the plant capital cost if a spare gasifier is not needed for the coal gasification plant: As noted, the gasifier hardware portion of a coal gasification plant operates at a very high temperature and pressure. It has been found that the reliability of the gasifier hardware is such that traditional IGCC plant economics may require that a spare gasifier be built as a “hot standby” in case the primary gasifier fails or requires maintenance. The standby gasifier is needed because 1) there is a long lead time required to repair a gasifier, and 2) the syngas produced cannot be stored for use while the gasifier is being repaired. The gasifier hardware can cost tens or hundreds of millions of dollars in a typical IGCC plant. In the event of a gasifier failure, the integrated LPP/IGCC plant could continue to produce electricity by converting the stored coal liquid to LPP Gas™ and burning it in the gas turbine at the plant. 4. Ownership and operation of the CTL and power blocks may be separated: One of the operational issues of concern to IGCC plant owners is the fact that the coal gasification process is a complex chemical process for which the power industry does not have extensive experience. Scenarios #1 and #3 decouple the coal gasification/CTL plant from the power generation plant. This would allow, for example, a process plant company to own and operate the gasification/CTL plant, while a utility or independent power producer would operate a standard combustion turbine plant, along with the LPP skid. Summary and Conclusions The LPP Combustion technology presented in this paper allows the cleanest combustion of liquid fuels, achieving natural gas levels of criteria pollutants (NOx, CO, SOx & PM). In addition, no “net” carbon emissions are obtained when used with biofuels. This technology provides the capability for tremendous fuel flexibility and low emissions not previously attainable in modern DLE gas turbines with liquid fuels, and provides substantial operating cost savings for liquid fueled combined cycle power plants. The flexibility and economics of an IGCC plant can be considerably improved with the introduction the LPP technology, by cleanly burning coal-derived liquids as part of the plant configuration. An enhanced IGCC plant configuration has been 27th Intl. Pittsburgh Coal Conference

16

October 2010 - Istanbul, Turkey

proposed, which integrates the LPP technology with the traditional IGCC design. The combined LPP/IGCC concept overcomes several notable challenges associated with the operation and profitability of existing Integrated Gasification Combined Cycle facilities. By operating in a combined IGCC and CTL (polygen) configuration, the resulting plant has much more flexibility to handle turn-down due to reduced power demand. The resulting coal liquids are more readily stored than syngas, allowing the coal gasification block to continue to operate at design capacity even when the power production needs are reduced. Use of the LPP technology decouples the gasification and power blocks, allowing the gasification to take place at locations that may be more advantageous from an economic or environmental point of view. Other benefits include use of waste nitrogen for the LPP process, use of sidestream products such as naphtha as a feedstock for the LPP process and elimination of the need for a spare gasifier. It is anticipated that these enhanced configurations will significantly advance the acceptance and implementation of the IGCC concept in the future.

27th Intl. Pittsburgh Coal Conference

17

October 2010 - Istanbul, Turkey

Laminar Flame Speed Study of Syngas Mixtures (H2-CO) with Straight and Nozzle Burners Nicolas Bouvet, Christian Chauveau and Iskender Gökalp CNRS-Institut de Combustion, Aérothermique, Réactivité et Environnement, 1C Av. de la Recherche Scientifique,45071 Orléans Cedex 2, France Seong-Young Lee and Robert J. Santoro Propulsion Engineering Research Center and Department of Mechanical and Nuclear Engineering, The Pennsylvania State University, University Park, Pennsylvania 16802, U.S.A. Laminar flame speeds of undiluted syngas (H2/CO) mixtures have been studied at atmospheric conditions using chemiluminescence and schlieren techniques for straight cylindrical and nozzle burner apparatus. A wide range of mixture composition, from pure H2 to 1/99 % H2/CO, has been investigated for lean premixed syngas flames. To achieve a better flame stabilization and reduce flame flashback propensity, two nozzle burners of different sizes have been designed and fabricated and were further used to compare the flame cone angle and the total surface area of the flame techniques. Results are compared to predictions using recent H2/CO mechanisms developed for syngas combustion. I. Introduction Sythetic gas, better known as “syngas”, has been widely studied because of its numerous industrial applications: chemicals and gases synthesis, methanol production [1] alternative fuel for spark-ignition engine [2] etc. One of the most promising and active research areas at the moment is certainly the power generation with the development and optimization of Integrated Gasification Combined Cycle (IGCC) plants. Processes of gasification allow a wide range of solid combustibles, including coal, biomass and municipal solid wastes (MSW) to be converted into syngas mixtures that will be used in a gas turbine engine to generate electricity. In this paper, syngas is considered as an alternative fuel, as its production and combustion involve a considerably clean conversion of solid fuels into energy: gas clean up, sulphur removal, CO2 capture along with lean premixed combustion in modern gas turbine combustors are the key answers for demanding environmental standards [3,4]. One of the particular concerns, however, is the variability of syngas composition recorded in the industry. Some IGCC project shows H2/CO ratios between 0.33 and 2.36 by volume [5]. Studies on syngas production from coal have shown that parameters such as coal quality and origins, reaction temperatures in gasifiers, oxygen/coal ratio or steam/coal ratio have an important impact on the composition of the synthesised syngas [6]. Clearly, a strong flexibility is expected while designing the new generation of IGCC power plants, particularly for the hardware related to syngas combustion. Modern combustors, for example, will have to deal with various syngas and hydrogen-enriched fuels, which requires a better understanding of their combustion complexities [7] Phenomena, such as autoignition, flame flashback and combustion instabilities still have to be better understood in order to generalize the use of syngas by designing safer facilities [8] The knowledge of flame speed of such mixtures is therefore highly required.

This study proposes a systematic investigation of laminar flame speeds of pure H2/CO syngas mixtures at atmospheric conditions. Few studies related to laminar flame speeds of H2/CO mixtures are available in the literature. Furthermore, most of them have investigated a very limited number of syngas compositions for equivalence ratios not always representative of lean premixed combustion in a gas turbine engine. Studies of flame speed were performed by Scholte and Vaags [9,10] for rich H2/CO mixtures on a nozzle burner using the schlieren method. An identical diagnostic was used by Günther and Janisch [11] on a straight burner to study syngas mixtures at stoichiometric conditions. Yumlu [12] measured adiabatic burning velocities of various hydrogen/carbon monoxide mixtures for an equivalence ratio of 0.6, using a flat flame burner with direct cooling. Flame speeds were obtained by extrapolation to zero heat loss of burning velocities obtained for different gas flow rates. Vagelopoulos and Egolfopoulos [13] investigated laminar flame speed and extinction strain rates of hydrogen and carbon monoxide mixtures using the counterflow, twin-flame technique along with laser Doppler velocimetry. McLean et al. [14] and Brown et al. [15] reported flame speeds obtained with the constant-pressure expanding spherical flame method, followed very recently by Sun et al. [16] for pressures up to 40 atm. Laminar flame speed and stretch effects were studied by Hassan et al. for pressures between 0.5 and 4 atm [17]. Huang et al. [18] performed flame speed measurements of a pure reformer gas (28% H2, 25% CO and 47% N2 by volume) at atmospheric conditions using digital particle image velocimetry and the counterflow twin flame apparatus. Natarajan et al. [19] investigated the influence of CO2 addition and reactant preheating on laminar flame speeds of syngas mixtures obtained by the cone surface method based on broadband chemiluminescence. Flame speeds near extinction have been studied very recently by Love et al. [20] and Subramanya and Choudhuri [21] on a water-cooled nitrogenstabilized flat flame and a twin flame counterflow burner, respectively. Detailed chemical models for H2/CO combustion are also available in the recent literature. Three of them have been selected to perform flame speed computations for the present study. Sivaramakrishan et al. [22], proposed a H2/CO mechanism based on the works of Davis et al. [23], with updated parameters for the reaction OH + HO2 = H2O + O2 for combustion at elevated pressures. Sun et al. [16] presented a H2/CO mechanism to model their laminar flame speeds obtained by the constant-pressure spherical flame technique. A limited number of other experimental data has been used to validate their mechanism and a new reaction rate constant for the reaction CO + HO2 → CO2 + OH was determined using ab initio calculations. A mechanism for CO, CH2O and CH3OH combustion has recently been proposed by Li et al. [24] with new rate constant recommendations for the reactions CO + OH = CO2 + H and HCO + M = H + CO + M. These chemical models have been selected for their recent updates of reaction rates and thermodynamics data as far as their validation against various experimental results are related to syngas combustion. II. Experimental Apparatus A schematic of the experimental apparatus used is shown in Figure 1. Each gas of the H2/CO/Air mixture is initially stored in separated tanks. Degrees of purity are respectively 99.95% for the hydrogen and 99.995% for the carbon monoxide. Air used is of breathing quality (99.95%). Each gas flow rate is carefully controlled by calibrated mass flow meter with an accuracy of about 1% full scale. A mixing section allows a rapid mixing of the reactants prior to injection into the burner.

Figure 1. Schematic of the experimental apparatus.

Figure 2. Diagnostics set-up. Straight Burners A series of six stainless steel burners have been designed and fabricated. Tested burners diameters are 3, 4, 6, 8, 12, and 16 mm. A nominal burner length of 900 mm was chosen based on the 50 x diameter criteria, therefore ensuring a fully developed laminar flow at the burner rim. Particular care was taken to fabricate sharp-edged burner rims so as to avoid any irregularity that could affect the flame stabilization. The apparatus also integrates a pilot flame ring with a premixed mixture of methane and air as reactants. Tests were conducted to determine if the pilot flame could enhance the syngas flame stability at lower equivalence ratios. As flames obtained on the 3 mm diameter displayed ridges characteristic of cellular instabilities and the use of the 16 mm tube was severely restricted by the available flow rates, results presented in this study were exclusively obtained for the 4, 6, 8 and 12 mm diameter tubes. Diagnostics Set-up The Z-type two-mirror schlieren system is arranged as shown in Figure 2: the light beam emitted by a tungsten lamp (30W) goes first through a condenser lens. The diverging beam issued from an aperture located at the focal point of the condenser lens is used as a point source for the first spherical mirror (fl=1m). The top of the burner is placed in the middle of the test region defined by the parallel beam formed between two spherical mirrors. The

resulting image is formed on a digital camera with 5 Mega pixels. A vertical knife edge located at the focal point of the second mirror is used to block the white beam and therefore to form the schlieren images. OH* chemiluminescence images were recorded with an intensified CCD (ICCD) camera (PIMAX: 512´512 pixels) equipped with an UV lens (f#/4.5) and filters (WG305+UG11). OH* chemiluminescence images were captured by accumulation, for each case, of 50 instantaneous images. During the experiment, the ICCD camera was moved to achieve the best resolution possible for each series of measurements. Similarly to Natarajan et al. [19] an horizontal knife edge was used to reduce the intensity from the base of the flame and make its tip more visible. This method substantially improves the trace of the flame reaction zone boundaries and thus the laminar flame speed computations. Schlieren and OH* chemiluminescence images were simultaneously taken for every flame studied. III. Flame Speed Calculations Methods The syngas laminar flame speed (SL) calculations were performed using two different approaches. First, an averaging method has been used to determine SL. Assuming that the burning velocity is the same over the total surface area of the flame (A), the flame speed can be calculated as follows: SL = A / Q where Q is the total volumetric flow rate of the unburned mixture. This method requires knowledge of the total area of the flame, deduced by analysis of the chemiluminescence images. As shown in Figure 3, a FORTRAN program performs a three-point Abel inversion [25] of the ICCD images to deduce the 2-D boundaries of the flame, based on the maximum emission of OH*. Using the flame axisymmetry hypothesis, the flame speed is calculated from each half of the computed flame.

Figure 3. OH* chemiluminescence image processing: (Left) OH* chemiluminescence of the syngas flame and (Right) image after Abel inversion and maximum intensity trace. Two main concerns need to be considered when using this method: first of all, the flame speed is not roughly constant over the entire surface of the flame as shown by Lewis and Von Elbe [26]. Indeed, slower velocities due to heat sink effects have to be expected at the burner

rim and higher velocities due to intense burning rate at the flame tip. Furthermore, the flame area should be defined just upstream of the preheat zone where the properties of the unburned gas are not modified. Calculations based on the OH* chemiluminescence will consequently lead to higher calculated flame surfaces, implying an underestimation of the calculated flame speeds. Still this method yields good results compared to other methods found in the literature. Figure 4 displays a comparison of different data set points for the laminar flame speed of pure H2 mixtures. The data are taken over the 4 mm burner and compared to various published data. The model results by Li et al. [24] are also superimposed in the figure. Flame tips tend to be closed above equivalence ratios of ~0.8, while below equivalence ratios of ~0.8, the flame tip appears open. A more detailed description for the opened flame tips will be presented in a later section. Although laminar flame speeds from this study are, as expected, lower than predictions, present results show a very good agreement with recent data points obtained in spherical bomb devices by Lamoureux et al. [27]. The second method employed is the cone angle method. This method is particularly suitable for flames displaying a straight cone and therefore, for experiments using aerodynamically contoured nozzles. In this case, the velocity (Uo) of the unburned mixture at the nozzle exit can be considered as constant and the expression of the laminar flame speed (SL) is given by SL = Uo sin α, where Uo is the bulk velocity of the unburned mixture and α the half cone angle of the flame. A FORTRAN program was used to perform edge detection on acquired schlieren images and therefore determine the angle α.

Figure 4. Laminar flame speeds of pure H2/air mixtures. Model: Li et al. [24] Flame Tip Opening In the attempt to study ultra lean mixtures of syngas, the problem of flame tip opening arises. This phenomenon, mainly observed for rich heavy-hydrocarbon/air mixtures and lean hydrogen/air mixtures has been studied by various investigators, including experiments of Law et al. [32] for methane, propane and butane/air mixtures and also mathematical description of the phenomenon by Buckmaster [33]. Figure 5 shows chemiluminescence images of syngas flames with different H2/CO compositions for an equivalence ratio close to 0.6. It can be seen that for the first case of pure hydrogen flame, the tip is obviously open. This trend is due to the highly diffusive nature of hydrogen, characterized by a Lewis number (ratio of thermal diffusivity to mass diffusivity) lower than unity, combined with flame curvature effects. As the deficient species, hydrogen, approaches the flame tip, its concentration is further reduced due to the defocusing effect of the tip curvature. The rate of thermal diffusion being much lower than mass diffusion in this case, the burning intensity is considerably weakened at the tip and local extinction is consequently observed. By further adding CO, the tip opening is reduced (cases (b) and (c)) and a closed tip is finally obtained for the 40/60% H2/CO composition (d).

a b c d Figure 5. Open flame tip phenomenon – 4 mm burner – (a) pure H2, (b) 80/20% H2/CO, (c) 60/40% H2/CO, and (d) 40/60% H2/CO. The flame tip opening phenomenon described previously was mainly observed for mixtures with higher hydrogen content (pure hydrogen to 80/20% H2/CO) for which it was impossible to obtain closed tips for equivalence ratios below 0.8. As the influence of a possible mixture loss through the open tip on the laminar flame speed seems difficult to evaluate and would require complex modelling of the syngas flame, results herein presented only consider the case of closed laminar premixed cones. IV. Results and Discussion A wide range of syngas mixture compositions have been tested on four different straight burners with diameters varying from 4 to 12 mm. The PREMIX code [34] was used to systematically calculate laminar flame speed for equivalence ratios and mixtures compositions and of interest. In the PREMIX code, Soret effects and multicomponent diffusion were considered while performing the calculations. The continuation mode of the PREMIX code was used to gradually enlarge the computation domain and refine the resolution grid so as to improve the convergence and precision of the calculations. Mechanisms were computed with the thermodynamics and transport data selected by the different investigators. Figure 6 displays the results for a 50/50 % H2/CO mixture. Experimental results show a very good agreement with the works of Natarajan et al. [19] using Bunsen burner flames, and Mclean et al. [14] with constant-pressure expanding spherical flames. Model predictions are also closely describing the increase of laminar flame speed from 42 cm/s at an equivalence ratio of 0.62 to 107 cm/s at an equivalence ratio of 1. The overall agreement of the models with the available experimental data is not surprising as the tested mechanisms have been validated using Mclean et al. [14] data points for 50/50 and 5/95% H2/CO mixtures.

Figure 6. Laminar flame speeds of 50/50% H2/CO/Air mixture. Symbols represent experimental data and lines represent computations.

Figure 7. Laminar flame speeds of various syngas mixtures. Symbols represent experimental data (6 mm burner diameter), and dotted lines represent computations performed with the mechanism of Sun et al. [16] However, it is interesting to mention that mechanisms of Sun et al. [16] and Li et al. [24] presenting similar predictions over the range of investigated equivalence ratios, are closer to the experimental data points, while mechanisms from Sivaramakrishnan et al. [22] and Davis et al. [23] seem to predict lower flame speeds. Sun et al.16 mechanism has been selected to further compare experimental results and numerical predictions obtained for compositions varying from 10/90% to 70/30% H2/CO (see Figure 7). As expected, the laminar flame speed increases with increasing hydrogen content in the mixture. It is also worth mentioning that for a given mixture, the flame speed has a linear evolution for equivalence ratios between 0.6 and 1. In general, experimental results are slightly lower than calculated flame speeds. The choice of OH* chemiluminescence to determine the flame surface area and heat losses at the burner rim are seen to be the main reasons. Still, over the range of targeted compositions, agreement between experimental results and computed flame speeds is fully satisfactory. Influence of small hydrogen addition on the laminar flame speed of CO/air mixtures is shown in Figure 8. Results are here plotted against various experimental data points of different investigators along with different syngas mechanism predictions. Flame speed measurements of the present study show an excellent agreement with the numerical computations presented. It is interesting to notice that the before-mentioned offset between predictions of Sivaramakrishnan et al. [22] and predictions of other investigators is gradually reduced while the hydrogen content is further decreased. A more detailed analysis of the mechanisms will be performed to identify key reaction parameters responsible for the observed differences. It is also worth mentioning that the laminar flame speeds of CO/air mixtures are very sensitive to small hydrogen additions: at an equivalence ratio of 0.6, the 1/99% H2/CO mixture flame speed is approximately 10 cm/s, while 10/90% H2/CO mixtures display a laminar flame speed higher than 20 cm/s. This nonlinear increase in the flame speed for the first added 10% of hydrogen has been previously observed experimentally by Scholte and Vaags [10] and further explained by Vagelopoulos and Egolfopoulos [13], identifying the main CO oxidation reaction (CO + OH = CO2 + H) as a key reaction for such mixtures. On the other hand, flame speeds obtained for mixtures with high hydrogen contents show less sensitivity for small CO additions (Not shown here).

Figure 8. Laminar flame speeds of 10/90, 5/95 and 3/97 and 1/99% H2/CO mixtures, lines represent computations performed with the mechanism of Sun et al.[16], Sivaramakrishnan et al. [22] and Li et al. [24] solid symbols are experimental data from the present study, open symbols experimental data by various authors.

In order to study the laminar flame speed of ultra lean syngas mixtures, a premixed methane/air pilot flame was used to achieve a better flame stabilization. Results for three different mixture compositions are presented in Figure 9. Experimentally, the stability of the studied flames was greatly improved, reaching equivalence ratios close to 0.3 for mixtures with higher contents of CO. Cellular instabilities previously observed at an equivalence ratio of 0.6 close to flame blow-off conditions were in this case observed at much leaner conditions. However, it can be noticed that flame speeds obtained with the pilot flame seem to be systematically higher than the values suggested by the trend without a pilot flame. For instance, duplicated measurements for the 50/50% H2/CO mixture for an equivalence ratio of 0.62 show that the determined flame speed with the pilot flame is approximately 7 cm/s higher than the case without pilot flame. This indicates that heat transfer from the pilot flame to the unburned mixture is not negligible and consequently, cooling of the burner would be required.

Figure 9. Laminar flame speeds of 10/90, 50/50 and 70/30% H2/CO mixtures with (Solid symbols) and without (Open symbols) pilot flame. Lines represent computations performed with the mechanism of Sun et al. [16]

Nozzle Burner Development and Technique Comparison To achieve better flame stabilization for lean combustion and reduce the flame flashback propensity as far as reducing heat transfers from the pilot flame to the unburned mixture, two stainless steel nozzle burners, integrated with a cooling head, have been designed and fabricated. Both contractions have an inlet diameter of 50 mm and outlet diameter of 4 and 8 mm respectively. Their design is based on the works of Cohen et al. [35]. The cross sectional velocity at each burner exit has been measured using a hot-wire anemometer to ensure that the contraction nozzle allows a flow velocity nearly constant at the burner exit. Cold mixture velocity profiles for the 4 and 8 mm nozzle burners are displayed in figures 10 and 11. Both contractions show flat velocity profiles over about 63% and 71% of the nozzle diameter respectively. By elimination of the steep velocity gradient at the exit boundary, the fabricated nozzles were expected to enhance the flame stability and allow the formation of straight-sided flame cones required to accurately apply the flame cone angle method.

Figure 10. Velocity profile comparison for the 4 mm nozzle burner (3mm above burner tip - Total flow 1-D: 3184 sccm, straight burner: 3162 sccm and nozzle burner: 3153 sccm)

Figure 11. Velocity profile comparison for the 8 mm nozzle burner (3mm above burner tip - Total flow 1D: 3290 sccm, straight burner: 3222 sccm and nozzle burner: 3249 sccm) A considerable discrepancy for flame speeds extracted from two measurement techniques is shown in Figure 12 in case of 50/50% H2/CO mixture. Flame speeds determined from

chemiluminescence images are lower up to 30 % in comparison to flame speeds calculated from schlieren images. Results for the 40/60 H2/CO% mixture for the 8 mm nozzle burner appear to be more comprehensive. A maximum 15% decay between both methods can be noticed, and results obtained from the total area method for the nozzle burner closely reproduce those established for straight burners of different diameter sizes. In general, the cone angle method gives higher flame speed results than the flame surface area. Differences observed between the two set of measurements can be further explained by examining the schlieren and chemiluminescence images shown figure 12 and 13. While flames formed on the 8 mm nozzle burner display very straight-sided cones, flames stabilized on the 4 mm burner seem less affected by the presence of the nozzle and flanks of the flame are more curved. This would seem that the 4mm diameter nozzle was inappropriate to measure the flame speed for the range of flow rates considered. These observations agree with Andrews et al.36 review on the determination of burning velocities mentioning that the flame cone angle is strongly dependant on gas flow rates and burner design. In general, large burner diameters are selected for the cone angle method to minimize the influence of the curvature at the flame base and tip. In practice, this requirement is restrictive for the range of investigated composition as syngas mixtures with higher hydrogen contents cannot be properly stabilized on large burner diameters. Influence of the nozzle cooling, pilot flame and stabilisation of ultralean syngas flames on the nozzle burners will be considered in the next study.

Figure 12. Comparisons for the determination of the laminar flame speed of 50/50% H2/CO + Air mixture. NB = Nozzle Burner, SB = Straight Burner, computations performed with the mechanism of Sun et al. [16].

Figure 13. Comparisons for the determination of the laminar flame speed of 40/60% H2/CO + Air mixture. NB = Nozzle Burner, SB = Straight Burner, computations performed with the mechanism of Sun et al. [16] V. Summary and Future Work An extensive series of laminar flame speeds with various syngas mixtures have been performed using straight cylindrical burners. Results from the chemiluminescence technique show a good agreement with the experimental works of Natarajan19 et al. and McLean14 et al. and also with predictions calculated with recent H2/CO mechanisms. In the attempt to study mixtures in the ultra lean domain, flame flashback and flame tip opening phenomena strongly limited the equivalence ratio range of investigation, particularly for mixtures with higher hydrogen contents. Further measurements will be conducted using the designed contractions believed to improve the flame stability at leaner conditions. A parallel study of syngas flame speeds has been engaged at the CNRS Orléans using the counterflow twin flame apparatus enclosed in a high pressure chamber along with Particle Imaging Velocimetry. This configuration will allow a more accurate determination of laminar flame speeds by minimizing heat losses, flame stretch and curvature effects encountered during the present study.

Acknowledgments This work is supported by the CNRS and The Pennsylvania State University Propulsion Engineering Research Center. The authors thank Mr. Larry Horner for his help in fabrication and assembly of the different burner setups used for the present study. References 1Wender,

I., “Reactions of Synthesis Gas”, Fuel Processing Technology, Vol. 48, 1996, pp. 189-297. N.N., Miraglia, Y.C., Raine, R.R., Bansal, P.K., Elder, S.T., “Spark Ignition Engine Performance with ‘Powergas’ Fuel (Mixture of CO/H2): A Comparison with Gasoline and Natural Gas”, Fuel, Vol. 85, 2006, pp. 1605-1612. 3Moliere, M., “Benefiting from the Wide Fuel Capability of Gas Turbines: A Review of Application Opportunities”, 2Mustafi,

Proceedings of the ASME Turbo Expo, GT-2002-30017, Vol. 1, ASME, New York, 2002, pp.227-238. 4Richards, G.A., McMillian, M.M., Gemmen, R.S., Rogers, W.A., Cully, S.R., “Issues for Low-Emission, Fuel-Flexible Power Systems”, Progress in Energy and Combustion Science, Vol. 27, 2001, pp 141-169. 5Brdar, R.D., Jones, R.M., “GE IGCC Technology and Experience with Advanced Gas Turbines”, GE Power Systems, GER4207, 2000. 6Lee, J.G., Kim, J.H., Lee, H.J., Park, T.J., Kim, S.D., “Characteristics of Entrained Flow Coal Gasification in a Drop Tube Reactor”, Fuel, Vol. 75, 1996, pp. 1035-1042. 7Lieuwen, T., Richards, G., “Burning Questions”, Mechanical Engineering, Vol. 128, 2006, pp. 40-42. 8De Biasi, V., “Targeting Turbine Research Needed to Burn Syngas and Hydrogen Fuels”, Gas Turbine World, Vol. 35 2005, pp. 28-30. 9Scholte, T.G., Vaags, P.B., ”The Influence of Small Quantities of Hydrogen and Hydrogen Compounds on the Burning Velocity of Carbon Monoxide”, Combustion and Flame, Vol. 3, 1959, pp. 503-510. 10T.G. Scholte, P.B. Vaags, “Burning Velocities of Mixtures of Hydrogen, Carbon Monoxide and Methane with Air”, Combustion and Flame, Vol. 3, 1959, pp. 511-524. 11Günther, R., Janish, G., “Messwerte der Flammengeschwindigkeit von Gasen und Gasgemischen“, Chemie Ingenieur Technik, Vol. 43, N° 17, 1971, pp. 975-978. 12Yumlu, V.S., “Prediction of Burning Velocities of Carbon Monoxide-Hydrogen-Air Flames”, Combustion and Flame, Vol. 11, 1967, pp. 190-194. 13Vagelopoulos, C.M., Egolfopoulos, F.N., “Laminar Flame Speeds and Extinction Strain Rates of Mixtures of Carbon Monoxide with Hydrogen, Methane and Air”, Proceedings of the Combustion Institute, Vol. 25, 1994, pp. 1317-1323. 14McLean, I.C., Smith, D.B., Taylor, S.C., “The Use of Carbon Monoxide/Hydrogen Burning Velocities to Examine the Rate of the CO + OH Reaction”, Proceedings of the Combustion Institute, Vol. 25, 1994, pp. 749-757. 15Brown, M.J., McLean, I.C., Smith, D.B., Taylor, S.C., “Markstein Length of CO/H2/Air Flames, using Expanding Spherical Flames, Proceedings of the Combustion Institute, Vol. 26, 1996, pp.875-881. 16Sun, H., Yang, S.I., Jomaas, G., Law, C.K., “High Pressure Laminar Flame Speeds and Kinetics Modeling of Carbon Monoxide/Hydrogen Combustion”, Proceedings of the Combustion Institute, Vol. 31, 2007, pp. 439-446. 17Hassan, M.I., Aung, K.T., Faeth, G.M., “Properties of Laminar Premixed Flames at Various Pressures”, Journal of Propulsion and Power, Vol. 13, N° 2, 1997, pp. 239-245 18Huang, Y., Sung, C.J., Eng, J.A., “Laminar Flame Speed of Primary Reference Fuels and Reformer Gas Mixtures”, Combustion and Flame, Vol. 139, 2004, pp. 239-251. 19Natarajan, J., Nabdula, S., Lieuwen, T., Seitzman, J., “Laminar Flame Speeds of Synthetic Gas Fuel Mixtures”, Proceedings of GT2005, ASME Turbo Expo 2005, GT2005-68917, June 2005. 20Love, N.D., Periasamy, C., Gollahalli, S.R., Choudhuri, A.R., “Laminar Burning Velocity of Synthetic Gas Premixed Flames Near Extinction Conditions”, 4th International Energy Conversion Engineering Conference and Exhibit, AIAA 20064122, June 2006. 21Subramanya, M., Choudhuri, A., “Investigation on the Flame Extinction Limit of Fuel Blends”, 41st AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit, AIAA 2005-3586, July 2005. 22Sivaramakrishnan, R., Comandini, A., Tranter, R.S., Brezinsky, K., Davis, S.G., Wang, H., “Combustion of CO/H2 Mixtures at Elevated Pressures”, Proceedings of the Combustion Institute, Vol. 31, 2007, pp. 429-437. 23Davis, S.G., Joshi, A.V., Wang, H., Egolfopoulos, F., “An Optimized Kinetic Model of H2/CO Combustion”, Proceedings of the Combustion Institute, Vol. 30, 2005, pp. 1283-1292. 24Li, J., Zhao, Z., Kazakov, A., Chaos, M., Dryer, F.L., Scire, JR., J.J., International Journal of Chemical Kinetics, Vol. 39, 2007, pp. 109-136. 25Dash, C.J., “One-Dimensional Tomography: A Comparison of Abel, Onion-peeling, and Filtered Backprojection Methods”, Applied Optics, Vol. 31, N°8, 1992, pp. 1146-1152. 26Lewis, B., Von Elbe, G., Combustion, Flames and Explosions of Gases, third ed., Academic Press, Orlando, 1987, p. 283. 27Lamoureux, N., Djebaïli-Chaumeix, N., Paillard, C.-E., “Laminar Flame Determination for H2-Air-He-CO2 Mixtures Using the Spherical Bomb Method”, Experimental Thermal and Fluid Science, Vol. 27, 2003, pp. 385-393. 28Tse, S.D., Zhu, D.L., Law, C.K., “Morphology and Burning Rates of Expanding Spherical Flames in H2/O2/Inert Mixtures up to 60 Atmospheres, Proceedings of the Combustion Institute, Vol. 28, 2000, pp. 1793-1800. 29Kumar, K.S., “Laminar Burning Velocities of Lean Hydrogen-Air Mixtures”, Explosion Dynamics Laboratory Report, FM97-15, 1998. 30Dowdy, D.R.D., Smith, D.B., Taylor, S.C., “The Use of Expanding Spherical Flames to Determine Burning Velocities and Stretch Effects in Hydrogen/Air Mixtures, Proceedings of the Combustion Institute, Vol. 23, 1990, pp. 325-332. 31Egolfopoulos, F.N., Law, C.K., “An Experimental and Computational Study of the Burning Rates of Ultra-Lean to Moderately-Rich H2/O2/N2 Laminar Flames with Pressure Variations”, Proceedings of the Combustion Institute, Vol. 23, 1990, pp. 333-340. 32Law, C.K., Ishizuka, S., Cho, P., “On the Opening of Premixed Bunsen Flame Tips”, Combustion Science and Technology, Vol. 28, 1982, pp. 89-96. 33Buckmaster, J., “A Mathematical Description of Open and Closed Flame Tips”, Combustion Science and Technology, Vol. 20, 1979, pp. 33-40. 34Kee, R.J., Grcar, J.F., Smooke, M. D., Miller, J.A., Technical Report SAND85-8240, Sandia National Laboratories, Albuquerque, NM, 1987. 35Cohen, M.J., Ritchie N.J.B., “Low-Speed Three-Dimensional Contraction Design”, Journal of the Royal Aeronautical Society, Vol. 66, 1962, pp. 231-236. 36Andrews, G.E., Bradley, D., “Determination of Burning Velocity: A Critical Review”, Combustion and Flame, Vol. 18, 1972, pp. 133-153.

Manuscript Not AVAILABLE

2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission

PROGRAM TOPIC:

Coal-Derived Products

- Coal Utilization By Products (Ash, Fertilizers, Etc,)

Structural Changes in Bituminous Coal Fly Ash Due to Treatments with Aqueous Solutions Roy Nir Lieberman Department of Biological Chemistry, Ariel University Center of Samaria, Israel Contact Information: Ariel, 40700, Israel, Phone: 00972-547-699701, fax: 00972-9-7417429, email: [email protected] Haim Cohen1,2, Ariel Goldman3, Roy Nitzsche1,4 1. Department of Biological Chemistry, Ariel University Center at Samaria, Ariel, 40700 Israel, phone: 00972-52- 4306878, fax: 00972-8-9200749, email: [email protected] 2. Chemistry Department, Ben-Gurion University of the Negev, Beer Sheva, Israel 3. Department of Civil Engineering, Ariel University Center at Samaria, Ariel, 40700, Israel 4. TU Bergakademie Freiberg, Fakultät 4, Institut für Energieverfahrenstechnik und Chemieingenieurwesen 09599 Freiberg, Germany.

Abstract: Coal fly ash is produced in Israel via the combustion of class F bituminous coals. The bulk of coal fly ashes produced in Israel stems from South African and Colombian coals and therefore these ashes were the subject of the present study. It has been shown that during treatment of the flyash with aqueous solutions appreciable Structural Changes in the Matrix do occur. The flyash can be used as a scrubber and fixation reagent for acidic wastes. Recently it was reported that the scrubbing product can serve as a partial substitute to sand and cement in concrete. The bricks have proved to be strong enough according to the concrete standards. In order to have a better understanding of the fixation mechanism we have decided to treat the flyashes with water or acidic solution (0.1M HCl) thus changing the surface of the flyash particles. Surface analysis of the treated and untreated fly ashes have demonstrated that the treated flyashes have changed appreciably its' interactions with transition metal ions. Three possible modes of interactions were observed: cation exchange, chemical bonding and electrostatic adsorption of very fine precipitate at the flyash surface.

Influence Factors on Density and Specific Surface Area (Blaine Value) of Fly Ash from Pulverized Coal Combustion Hiromi Shirai, Michitaka Ikeda and Kenji Tanno Central Research Institute of Electric Power Industry Address: 2-6-1 Nagasaka, Yokosuka-shi, kanagawa-ken 240-0196, Japan Phone: +81 46 856 2121 Fax: +81 46 856 3346 Email: [email protected] Abstract In the Japanese electric power industry, it is desirable to reduce the cost of the treatment and expand the range of use of fly ash, such as in concrete admixtures. Therefore, it is necessary to form high-quality fly ash. To obtain high-quality fly ash, it is important to clarify the factors affecting its properties. In this study, the factors affecting the density and specific surface area (Blaine value) of coal fly ash were investigated on the basis of experimental results obtained using our combustion test facility and the ash data from the boiler of an actual electric power plant. The density was affected by the ash particle size, the true density of the component materials and aluminum content which is closely related to the fusibility. The specific surface area was affected by the particle size distribution and particle shape. The shape was affected not only by the ash particle size but also by the unburned carbon concentration, the ash fusibility and the ash content in the coal. It was also found that the specific surface area of the ash generated from our combustion test facility is higher than that of ash from actual boiler for ash with the same particle size. This result indicates that the shape of particles is affected by their heating and formation histories in the boiler. On the basis of the above findings, correlation equations were obtained for the density and Blaine value.

2. Experimental Section 2.1 Coal sample In the study using our combustion test facility, 18 types of

bituminous coal were used. The range of properties of each coal type is shown in Table 1. Coals with different combustibilities and fusibilities were selected to cover the range of properties of coal burned in Japanese pulverized coal-fired power plants, from which fly ash is produced.

Table 1 Coal properties Property

Our test furnace

Moisture [wt%]*

2 - 13

1-9

Ash [wt%]**

7 - 17

5 - 15

Volatile matter [wt%]**

31 - 47

28 - 45

Fixed carbon [wt%]**

41 - 63

42 - 58

Fuel ratio [-]

0.8 - 2.3

1.0 - 2.1

HHV [MJ/kg]** Ash composition [wt%]

1. Introduction The amount of fly ash discharged from existing pulverized coal-fired power plants is increasing and has reached over 10 million tons per year in Japan. The ratio of fly ash in Japan that is effectively used is more than 90%. However, over 60% of fly ash is treated so that it can be used as a material in cement. In the Japanese electric power industry, it is important to reduce the cost of the treatment and expand the range of the use of fly ash, such as in concrete admixtures. Therefore, it is necessary to form high-quality ash. The Japanese Industrial Standard (JIS) for the use of fly ash in cement classifies fly ash into three grades. The grade is determined by the properties of the fly ash (particle size, unburned carbon concentration, density, specific surface area obtained by the Blaine method) and the properties of the mortar (fluidity, solidity) to which the fly ash is added. To manage the ash quality, it is important to develop a method of predicting the ash properties. Various methods (1),(2),(3) of predicting particle size and unburned carbon concentration have been proposed. However, the study (4) of the prediction of ash density and specific surface area has been insufficient. The purpose of this study is to clarify influence factors on the density and specific surface area (Blaine value) from experimental results obtained using our combustion test facility and the ash data from the boiler of an actual electric power plant.

Actual boiler***

27 - 32

27 - 31

SiO2

45 - 73

47 - 75

Al2O3

18 - 41

15 - 30

Fe2O3

1 - 13

2 - 11

CaO

0.5 - 10

0.8 - 13

MgO TiO2

0.2 - 1.8

0.2 - 2.0

0.5 - 3.5

0.6 - 1.6

P2O5

0.05 - 2.1

0.07 - 1.3

Na2O

0.05- 1.7

0.1 - 1.7

K2O Unburned carbon

0.7 - 3.3

0.4 - 3.3

1.6 - 14.5

0.9 - 7.7

* At equilibrium humidity, ** Dry basis, ***Blended coal properties are included.

Temp. controller (Simulated GGH)

Stack

ESP

Fuel Supply Equipment

Lime-limestone

Pulverized Coal Refuse Biomass Furnace

Temp. controller (Simulated GGH)

AH

De-SOx Orimulsion Heavy oil

De-NOx

Water-cooled Gas Cooler Bag Filter

Scrubber with NaOH Soulution

Figure 1 Schematic diagram of the pulverized coal combustion facility – 50µm. Mixed ash collected from all parts of the facility, except the bottom of the furnace, was used as a sample. On the other hand, the two-stage combustion is being introduced in Japan, t. However, the two-stage combustion ratio at an actual boiler is unknown. The size of pulverized coal particles in two-stage combustion is similar to that in our furnace.

Air injection port for the two-stage combustion

900

400

7,991

1,900

No.3 Burner No.2 Burner No.1 Burner

1,000 1,000

1,600

Sampling port

Burner

Front view

Side view

Figure 2 Details of the furnace 2.2 Pulverized coal combustion facility A schematic diagram of the pulverized coal combustion facility is shown in Figure 1. The facility consists of a pulverizer, a furnace with three burners and a gas treatment system. The coal feed rate is determined by the thermal input of coal. The thermal input is 760 kW×3 burners. The coal feed rate is approximately 100 kg/h×3 when high-quality bituminous coal is burned. Details of the furnace are shown in Figure 2. The furnace has a width of 0.9 m and a depth of 1.9 m, a height of 9.5 m. The mean resident time of the combustion gas is approximately 4 s at 1200 oC. The wall has a water wall structure whose inert surface is covered with a refractory material. In this study, the outlet O2 concentration was set to 4%, which is equivalent to an excess air ratio of 1.24. The standard two-stage combustion ratio is 30%, and the median diameter of the pulverized coal particle size is 35

2.3 method of measuring ash properties The ash particle size is measured using a laser diffraction particle size analyzer. The particle density was determined by the displacement of ethanol in a Le Chatelier flask. The specific surface area was determined by the Blaine method, in which the specific surface area is derived from the resistance to the flow of air through a layer of ash particles using a known standard material. 3. Results and Discussion 3.1 Ash particle density The ash particle density ρash [g/cm3] depends on the true density without closed void ρt [g/cm3] and the ratio of closed void volume to a particle volume (the void ratio: ε [−] ). The particle density is given by

ρ ash = ρ t (1 − ε )

(1)

The factors affecting the true density and the void ratio are investigated. True density The true density is the average density of the composed materials. However, it is difficult to analyze the composition of materials because a large part of the materials is amorphous. Therefore, the true density was estimated by the chemical analysis (JIS M 8815) of oxides in the ash. The selected oxides were SiO2（density ρi: 2.2 g/cm3）, 3Al2O3・2SiO2（3.1）, Fe2O3 （5.2）, CaO（3.4）, MgO（3.7）, P3O4(2.4), TiO2（4.2）, Na2O (2.3) and K2O (2.4). In this study, all Al2O3 was assumed to be mullite (3Al2O3・2SiO2) because more than 50% of Al2O3

3.0

0.30

Data from our test facility 2.9

Original ash (Dv-ash:22-25µm)

2.8

Pulverized ash (Dv-ash:4.0-5.5µm)

‐2%

Dv-ash：20-25μm

Temperature at liquid phase ratio of 90 wt%

0.20

2.6

ε [-]

Measured density [g/cm3]

Melting point （JIS M 8801)

0.25

2.7

2.5

0.15

0.10

2.4 0.05

2.3

0.00 1,200

2.2

1,300

1,400

2.0

2.1

2.2

2.3

2.4

2.5

2.6

2.7

2.8

2.9

3.0

Figure 3 Comparison between estimated true density and the density after removing closed voids by grinding ash

1,800

Newlands two-stage combustion ratio:30% D v-coal:16μm Newlands two-stage combustion ratio:30% D v-coal:49μm Newlands two-stage combustion ratio:30% D v-coal:91μm Newlands two-stage combustion ratio:0% Dv-coal:49μm Newlands two-stage combustion ratio:15% D v-coal:49μm Newlands two-stage combustion ratio:40% D v-coal:49μm Wambo two-stage combustion ratio:30% Dv-coal:31μm Wambo two-stage combustion ratio:30% Dv-coal:48μm Wambo two-stage combustion ratio:30% Dv-coal:69μm

Data from our test facility Dv-ash：20-25µm

0.25

0.20

ε [-]

0.35

0.20

1,700

0.30

Estimated true density ρa [g/cm3]

0.25

1,600

Figure 5 Influence of fusibility of ash on void ratio

2.0

0.30

1,500

Temperatur [oC]

2.1

ε [-]

Data from our test facility

Fluid point (JIS M 8801)

+2%

0.15

0.10

0.05

0.00

0.15

0

5

10

15

20

25

30

35

40

45

50

55

60

WAl [wt%]

0.10

Figure 6 Influence of Al2O3 content on void ratio

0.05 Data from our test facility 0.00 10

15

20

25

30

Dv-ash [µm]

Figure 4 Influence of ash particle size on void ratio exists as mullite. The content ratio Wi [wt%] of each oxide was recalculated so that the total content ratio of the oxides was 100 wt%. The estimated true ash particle density ρa [g/cm3] was estimated using the following equation.

ρa =

∑

100 (Wi / ρ i )

(2)

The ash was also ground to a volume mean particle diameter DV-ash [μｍ] of less than 6 μm to remove as many of the closed voids as possible. DV-ash is defined as

DV − ash =

∑ ( XiDi ) ∑ Xi

(3)

where Di [μm] is the diameter and Xi [-] is the weight ratio of the particles. The comparison between ρa and the density of pulverized ash particle is shown in Figure 3. It was found that ρa is almost equal to the pulverized ash density. This result clarified that our method can approximately estimate the true density.

Closed voids The void ratio ε [−] was calculated from the measured ash particle density and ρa. The factors affecting the formation of closed voids were clarified using the void ratio. The relationship between ε and DV-ash is shown in Figure 4. ε was strongly affected by the ash particle size; ε increases with decreasing DV-ash. Small ash particles are formed from the small coal particles, which burn rapidly and whose temperature also increases rapidly during combustion. Additionally, the number of coalescent mineral particles is less than that of large coal particles. Small coal particles are advantageous for the coalescence and melting of mineral particles. Therefore, ε is lower in smaller ash particles. On the other hand, ε for Newlands coal was different from that of Wambo coal at the same particle size range. This indicates that there are other factors strongly affecting ε. Other possible factors were investigated. ε did not correlate with the fuel ratio, which is the index of combustibility. Moreover, ε did not correlate with the melting point of the ash. However, it was found that ε correlates with the temperature at which the liquid phase fraction is 90%, which is the weight ratio of the liquid phase of ash to the total of the solid and liquid phases of ash as shown in Figure 5. These phases were estimated by FactSage 5.5 which is a thermochemical software and database package. Moreover, it

12,000

2.8 Our test furnace (R：0.63) Boiler A (R：0.92) Boiler B (R：0.74) Boiler C (R：0.78) Boiler D (R：0.94)

2.6 2.5

+3% -3%

R: Correlation coefficient

ρash-est [g/cm3]

Our test furnace Boiler A Boiler B Boiler C Boiler D

10,000

SAb [cm2/g]

2.7

2.4

8,000

6,000

2.3

4,000

2.2

2,000

2.1 0

2.0

0

5

10

15

20

25

Ds-ash [μm]

1.9

Figure 8 Influence of ash particle size on specific surface area (Blaine value)

1.8 1.8

1.9

2.0

2.1

2.2

2.3

2.4

2.5

2.6

2.7

2.8 4.0

ρash-act[g/cm3]

Figure 7 Correlation between measured density and estimated density

3.5 3.0 2.5

SAb/SAd [-]

was found that the formation of mullite affects the above temperature. Figure 6 shows the correlation between ε and the content ratio of Al2O3 WAl [wt%] which comprise in the fly ash in the form of mullite. This correlation shows that the content of Al2O3 is an important factor affecting ε.

2.0 1.5 Our test furnace Boiler A Boiler B Boiler C Boiler D

1.0 0.5

Correlation equation for ash particle density The correlation equation for the ash particle density was set up as follows:

ρ ash − est = ρ a (1 − ( kd1 + kd 2 × (W Al / 100 ) + kd 3 × (W Al / 100 ) 2 + kd 4 × D p ))

(4)

where kdi is the coefficient calculated by the least-squares method. The correlation between the actual density ρash-act [g/cm3] and the estimated density ρash-est [g/cm3] is shown in Figure 7. It was clarified that the factors affecting density revealed in this study is correct, because the difference between two densities is within 5% of the actual density. 3.2 Specific surface area In JIS, the specific surface area of the ash, SAb [cm2/g-ash], is measured by the Blaine method. However, SAb is not the true specific surface area of the ash, though it is an important value for evaluating the outer surface area excluding the pore surface area. The specific surface area of a powder is affected by the particle size distribution and particle shape. Particle size The specific surface area is strongly affected by the surface area of fine particles. Therefore the influence of the ash particle size on SAb is investigated using the surface mean diameter DS-ash [µm], which is defined as D S − ash =

∑ Xi

∑ ( Xi Di )

(5)

0.0 0

5

10

15

20

25

30

35

Ds-ash [μm]

Figure 9 Influence of ash particle size on shape of ash particles The relationship between DS-ash [µm] and SAb is shown in Figure 8. If the shapes of the particles are the same, SAb increases as DS-ash decreases. However, no clear dependence of DS-ash on SAb was observed. This indicates that the influence of the particle size on SAb is weaker than that of the particle shape. Particle shape The specific surface area SAd [cm2/g-ash] when ash particles were assumed to be spherical was calculated from the ash particle size distribution and ash density. The relationship between DS-ash and the ratio of SAb to SAd, which is the reciprocal number of Carman’s shape factor, is shown in Figure 9. The ratio increases with the ash particle diameter. This indicates that the sphericity of the ash particles decreases as the diameter increases. It is estimated that the number of coalesced particles of minerals affects the particle shape similarly to its influence on density. Next, SAb/SAd in our furnace was compared with that in an actual boiler. For the same particle size range, SAb/SAd in our furnace was observed to be higher than that in the boiler. This result indicates that the sphericity of particles in our facility is lower than that in the boiler because the resident time of ash particles at the high-temperature zone of our furnace is shorter than that of the boiler.

5.0

4.0

Ds-ash：9-10µm

Data from our test furnace 3.5

Ds-ash：11-13µm

4.0

3.0

3.5

2.5

SAb/SAd [-]

SAb/SAd [-]

4.5

Data from our test furnace

3.0 2.5

1.5

2.0

1.0

1.5

0.5

1.0 0

2

4

6

8

10

12

0.0 1,200

14

Uc [wt%]

1,300

1,400

1,500

1,600

1,700

Melting point [oC]

Figure 12 Influence of fusibility of ash on shape of ash particles

Figure 10 Influence of unburned carbon concentration on shape of ash particles

1800

4.0 Uc:1-3wt%, Ds-ash:10-13µm

Data from our test furnace

Data from our test furnace

Uc:5-6wt%, Ds-ash:10-13µm

1700

Melting point [oC]

3.5

3.0

2.5

2.0

1.5

1600

1500

1400

1300

1.0 4

6

8

10

12

14

1200

16

0.0

Ash content [wt%]

1.0

2.0

3.0

4.0

5.0

Al2O3/(Fe2O3+CaO+MgO+Na2O+K2O) [-]

Figure 13 Relationship between index of ash fusibility and melting point

Figure 11 Influence of ash content on shape of ash particles

4.0

Furthermore, it was found that the unburned carbon concentration UC [wt%], which is affected by the fuel ratio and the combustion conditions such as the coal particle size and two-stage combustion ratio, affects SAb/SAd, as shown in Figure 10. The particles containing unburned carbon have a complex shape. In ash with a high unburned carbon concentration, SAb may not be correctly measured because the Blaine method cannot be efficiently applied to the particles with greatly different shapes. It was also found that the ash content affects the shape of ash particle as shown in Figure 11. The sphericity may decrease as the ash content Wash [wt%] increases because the number of coalesced particles of minerals affects the particle shape. On the other hand, SAb/SAd is affected by the fusibility of ash. As shown in Figure 12, it was confirmed that the melting point is correlated with SAb/SAd. In addition, to estimate SAb/SAd from the ash composition without measuring the melting point, an index that correlates with the melting point was investigated. It was also confirmed that Al2O3 / (Fe2O3+CaO+MgO+Na2O+K2O) (5) can be used for this purpose as shown in Figure 13. Moreover, it was found that the index is strongly correlated with SAb/SAd as shown in Figure 14.

Data from our test furnace 3.5 3.0

SAb/SAd [-]

SAb/SAd [-]

2.0

2.5 2.0 1.5 1.0 0.5 0.0 0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

Al2O3/(Fe2O3+CaO+MgO+Na2O+K2O) [-]

Figure 14 Relationship between index of ash fusibility and shape of ash particles

Correlation equation for SAb/SAd The correlation equation for SAb/SAd was set up as follows:

4.0 Our test furnace (R：0.95) Boiler A (R：0.93) Boiler B (R：0.93) Boiler C (R：0.89) Boiler D (R：0.99)

3.5

( S Ab S Ad ) est = ks1 ×(Wash

/ 100) ks2

× (U C

/ 100) ks3

3.0

Findex =

Al2O3 （Fe 2O3+CaO+MgO+Na2O+K2O）

Estimated SAb/SAd [-]

× Findex ks4 × Ds −ash ks5

(4)

where Ksi is the coefficient calculated by the least-squares method. The correlation between the actual SAb/SAd and the estimated SAb/SAd is shown in Figure 15. It was clarified that the factors affecting the shape in this study is correct, because of the strong correlation between the actual SAb/SAd and the estimated SAb/SAd. Moreover, the correlation between the measured SAb and the estimated SAb was calculated from the estimated SAb/SAd by equation (4) and SAd is shown in Figure 16. This indicates that SAb can be estimated from the estimated SAb/SAd and SAd .

From these results, it was found that the ash particle size distribution, ash composition and unburned carbon concentration are important factors in developing a method for predicting the density and specific surface area of fly ash. Reference (1)” Improvement of Pulverized Coal Combustion Technology for Power Generation”, CRIEPI Review No.46; Central Research Institute of Electric Power Industry: 2002. (2) Yan, L.; Gupta, P. R.; Wall, F. T. Fuel, 81, 337–344, 2002 (3) Barta, L. E.; Toqan, M. A.; Beer, J. M.; Sarofim, A. F. 24th Int. Conf. Comb., 1135–1144, 1992 (4) Shirai, H.; Tsuji, H.; Ikeda, M.; Kotsuji T., Energy & Fuels 2009, 23, 3406–3411 (5) NEDO report NEDO-C-9940, 2000

-15%

2.5 2.0 1.5 1.0 0.5 0.0 0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Actual SAb/SAd [-]

Figure 15 Correlation between actual SAb/SAd and estimated SAb/SAd 10,000 +15% 8,000 -15%

Estimated SAb [cm2/g]

4. Conclusion In this study, the influences of combustion conditions and coal properties on the ash particle size, density and specific surface area were investigated on the basis of experimental results obtained using our combustion test facility and the ash data from the boiler of an actual electric power plant. (1) Ash particle density The density was affected by the ash particle size, the true density of the component materials and aluminum content which is closely related to the fusibility. (2) Specific surface area of ash The specific surface area was strongly affected by the shape of the ash particles. The shape was affected not only by the ash particle size but also by the unburned carbon concentration in the ash, the ash fusibility and the ash content in the coal.

+15%

6,000

4,000 Our test furnace (R：0.84) Boiler A (R：0.91) Boiler B (R：077) Boiler C (R：0.75) Boiler D (R：0.92)

2,000

0 0

2,000

4,000

6,000

Measured SAb [cm2/g]

8,000

10,000

Figure 16 Correlation between measured SAb and estimated SAb

2010 International Pittsburgh Coal Conference

DEVELOPMENT OF COMMERCIAL CFBC BOILR FOR REFUSE DERIVED FULE Dowon Shun*, Dal-Hee Bae, Jaehyeon Park, Seung Yong Lee Korea Institute of Energy Research 102 Gajeong-ro, Yuseong-gu, Daejon 305-343, KOREA *[email protected] Introduction Korea officially started to promote waste fuel utilization by proclaiming the regulation regarding quality standards and use of recycled plastics fuel in year 2003. Since then continuing enactment of related law activated the business of waste energy recovery. Regulations defining refuse derived fuel(RDF) which is made of municipal solid waste, wood chip fuel(WCF), sludge derived fuel(SRF) are proclaimed successively. So far the regal background of waste fuel business is established and business sectors are about to start(MOE, 2006; 2009). The main structure of waste fuel business is in thermal utilization, i.e., combustion. Among various combustors circulating fluidized bed combustion (CFBC) is considered to be the optimum (Shun, et al., 2004; 2006). CFBC can handle the fuel with varieties of shape, heating value, moisture and mineral content. It also can control sulfur dioxide, NOx, HCl inside the combustor, thus it can alleviate the burden of flue gas cleaning system. The design concept of CFBC for RDF is different from conventional CFBC burning coal. The RDF burning CFBC should handle HCl instead of SO2 and above all it must show durability during the operation in the corrosive environment. The plant must be equipped with HCl removing system from the flue gas. Steam temperature seldom goes up above 723K to reduce the affect of high temperature corrosion. Often the flue gas temperature is lowered below 923K at the upstream of the final superheater. The density of fuel is much lighter than that of coal and its ash is finer. Cyclone collection efficiency must be adjusted according to the particle size of fly ash. A new design of CFBC is conceived for Korean RDF and its pilot plant was constructed with the rate of 8ton of steam per hour. Feasibility of the design is evaluated and design parameter was determined by burning RDF and RPFs. Emission level for the design of flue gas treating system was measured. With the evaluation of all those design and operating variables, a pilot scale (1MWe equivalent) CFBC for RDF was built and its performance was investigated. Experimental data collected from the plant were analyzed and a design of 10MWe CFBC for RDF was made. Design of a pilot scale CFBC boiler for RDF A circulating fluidized bed combustion (CFBC) boiler with 8ton/h steam rate was designed and constructed. The maximum steam quality is 450 deg.-C and 38ata. The boiler circuit was arranged as economizer → drum → wall evaporator → drum → super heater 1 → super heater 2. The system consists of a combustor with evaporator, a hot cyclone, a convection pass with superheater, economizer, and air heater. Flue gas was cleaned by dry absorber gas cleaner in series with a bag house. After the bag house there was a wet scrubber to remove HCl and cleaned gas passed through a stack to environment. Figure 1 shows the schematic of the facility. The produced steam was totally condensed and returned to the feed water tank again.

To Steam Generator

Drum

Cyclone

S/H3

S/H1

Combustor

Scrubber

RDF Hopper SDR

Econ. Water A/H

A ir

ID Fan Stack

FD Fan

Figure 1. The diagram of the CFBC pilot plant for RDF Operation and performance of the boiler Table 1 presents the characterization of RDF fuel. RDF from municipal waste is RDF 5 equivalent. It is fabricated in pillar type pellet with 30mm of diameter and 100mm of length. Table 1. Analysis of waste derived fuel Sample Moisture Volatile Dried Industrial Sludge 5.80 51.8 Agricultural PE film 0.6 97.2 RDF from municipal waste 6.7 69.2

Ash F.Carbon Sulfur Carbon HCl HHV(kcal/kg) 42.3 0.00 2.2 30.7 2773 0.14 2.1 0.2 10550 16.2 7.9 0.2 52.6 1.0 5710

Figure 2 shows variation of temperature of major combustor sector , feed rate, and steam rate. Combustor temperature varied according to feed rate. However the final steam rate could be maintained almost invariably even without de-superheating. The mission at the boiler exit and the plant exit was presented in table 2. The emission of HCl was further cleaned by flue gas treatment and particulates were captured in the bag filter. The emission level of pollutants was maintained below of the limitation of Korean regulation(MOE, 2009). Figure 3 shows fluidized bed pressure variation during the operation. Except start up and shut down period the pressure maintained stable values. Air box pressure was in between 700-800mmaq and combustor exit pressure was almost atmospheric. Table 2. Emission level during the plant operation Mean value at given O2 level At the combustor exit SO2, ppm 3.3 HCl, ppm 236 NO, ppm 74 O2, % 8.5 Particulates, mg/Sm3 12,832

At the stack 3.3 9 74 8.5 14.7

Regulation 30 20 70 12 40

1000 900

400 630

800

Temperature(C)

700

14130

600

Fuel

500

s team

400 300 200

400

100

630

14130

Steam

Fuel

0

Time(min) Figure 2. Combustor condition during operation

1,000 Air box

900

L400

L630

14130

800

Pressure(mmAq)

700

Start up

Air box Shut down

600 500 400 400

300 200

630

100

14130

0 -100

Time(min) Figure 3. Combustor pressure difference during operation

The combustion was very stable during operation. Since the fuel consists mostly of volatiles the combustion efficiency is always over 99%. Figure 4 shows the proximate analysis of fly ash. Combustible content in fly ash was too low and almost undetectable. 100%

80%

Weight percent

F.Carbon Ash Volatile

60%

Moisture 40%

20%

0% 1

2

3

4

Samples

Figure 4. Proximate analysis of the flyash Design of a 10MWe CFBC boiler for RDF A 10MWe scale CFBC for RDF was designed based on the collected data from pilot plant operation. The combustor consists of one combustor with water wall, one hot cyclone, and one convection pass. The combustor configuration is 4m x 4m x 24m. The boiler includes 2 super heater banks, one economizer and banks of air heater is located in the convection pass. Steam pressure and temperature are 45ata and 723K respectively. Maximum continuous steam rate of the boiler is designed as 60 ton/hour. Table 3 presents the design specifications of 10MWe CFBC for RDF. Table 3. The technical data of 10MWe CFBC Specifications Net heat output, MWth Steam rate, ton/h Output steam temperature, deg.-C Output steam pressure, ata Feed water temperature, deg.-C Boiler load variation, % MCR Fuel flow range, ton/h

Values 46 60 455 47 143 50-100 5.4-10.6

The 10MWe CFBC shares the flow diagram with the pilot plant thus it is same with the diagram of Figure 1. Figure 4 shows the front view of the boiler lay out. Considering that RDF is a low density fuel the combustor has 4 surge bins each has 4hours of service term. The project to build a 10MWe CFBC boiler

for RDF will be launched this year. This plant will be the first commercial scale CFBC co-generation boiler for RDF in Korea.

Figure 4. The layout of 10MWe CFBC for RDF Conclusion A pilot scale CFBC boiler with 8ton/h of steam rate is designed, constructed and performed the combustion experiment. The RDF in CFBC was burned stably and thoroughly. The temperature and pressure profiles in the combustor were stable. Combustion efficiency was over 99.5%. Emission was below the Korean emission regulation except HCl which required flue gas treatment. A 10MWe scale commercial CFBC for RDF is designed from this experience and the construction project has started. References Shun, D., Bae, D-H, Lee, S. Y, Kim, and M-S,: Development of CFB RDF combustion and co-generation demonstration plants, The ninth Asian Conference on Fluidized-Bed and Three-Phase Reactors, pp.79-84 (2004). MOE, Revision of the Regulations relative to the Application of the law of regarding Savings of Resources and Recycling Promotion, legislation of MOE (2006), MOE, Tables, Revision of the Law of Conservation of the Environment, Ministry of Environment, Korea (2009). Shun, D., Bae, D-H, Lee, S-Y, and Jo, S.,”Circulating Fluidized Bed Combustion of Refuse Derived Fuel,” J. of Korean Inst. of Resources Recycling, 15(1), PP.58-65 (2006).

SIMPLIFIED QUANTIFICATION OF TETRAFLUOROBORATE ION IN FLUE GAS DESULFURIZATION EFFLUENT FOR MANAGEMENT OF FLUORINE EMISSION Seiichi Ohyama, Hiroyuki Masaki, Shinji Yasuike, and Kazuo Sato Central Research Institute of Electric Power Industry, 1646 Abiko, Abiko-shi, Chiba 270 1194, Japan ABSTRACT Boron compounds in industrial wastewaters mainly exist as boric acid and tetrafluoroborate ion (BF4-). BF4contributes to the emission regulations of both boron and fluorine. In coal-fired power plants, BF4- is formed in a flue gas desulfurization (FGD) unit, depending on the type of FGD and coal. BF4- is quite stable and slow to decompose, which may lead to increased values of especially fluorine in discharged effluents. Thus, in some plants, a dedicated treatment to decompose BF4- is employed to manage BF4- emission. The conventional measurements of BF4-, i.e., spectrophotometry and ion chromatography requires a tedious pretreatment or a stationary equipment and are difficult to carry out onsite. The onsite measurement of BF4- is helpful for the efficient treatment of BF4-, which may lead to a reduced amount of sludge. We propose a simple and rapid analytical determination method of BF4- for properly managing the BF4- emission in FGD effluents. We build up a compact flow measurement system using BF4- ionselective electrode (ISE) mounted in a flow cell. By measuring samples in a flow mode, the stability and reproducibility of measurements is largely improved compared to the conventional static batch measurement. The system can measures BF4- in the range of 1-100 mg/L and the measurement completes in less than 10 min. The proposed method is an easier and more rapid determination compared with the conventional analytical methods. The application of the system to the FGD effluents is discussed. INTRODUCTION Boron compounds in industrial wastewaters mainly exist as boric acid and tetrafluoroborate ion (BF4-). In Japan, the Environmental Quality Standard (EQS) for water pollution, which is a target level to be achieved in public water to protect human health, is set at 1 mg/L for boron and at 0.8 mg/L for fluorine.1 Reflecting these establishments, the National Effluent Standard (NES) for boron is set at 10 mg/L and 230 mg/L for effluents discharged to terrestrial water bodies and to coastal water bodies, respectively.2 Similarly, NES for fluorine is 8 mg/L in noncoastal areas and 15 mg/L for coastal areas. There is a large difference in the coastal effluent standards of boron and fluorine. Thus, in coastal areas where almost all of Japanese power plants are located, more stringent regulation is enforced on fluorine emission compared to boron’s. Based on the elemental composition, BF4- contributes to the emission regulations for both boron and fluorine, but considering a gap between boron and fluorine in the coastal areas, it is more responsible for fluorine emission than for boron emission. In coal-fired power plants, BF4- is formed in a flue gas desulfurization unit depending on a type of FGD unit and coal. Once it is formed, BF4- is fairly stable and slow to decompose, which is difficult to precipitate in ordinary coagulation sedimentation treatment. (1) BF4- + 3H2O → H3BO3 + 4F- + 3H+ Thus, in some power plants, dedicated treatment to decompose BF4- is employed to manage BF4- emission. The conventional measurements of BF4-, i.e., spectrophotometry and ion chromatography requires a tedious pretreatment or a stationary equipment, which are impossible to carry out onsite. In addition, since F- is involved in the formation and decomposition of BF4-, the onsite and simultaneous measurement of BF4- and F- would be helpful for the efficient management and treatment of BF4-, which would lead to a reduced amount of sludge in the BF4- treatment. This study aims to propose a simple quantification method of BF4- in flue gas desulfurization (FGD) wastewater, which is carried out onsite with a commercial ion-selective electrode. EXPERIMENTAL We built up a compact flow measurement system that uses BF4- and F- ion-selective electrodes (ISE) mounted in the dedicated flow cells (FLC-12), as shown in Figure 1. In the measurement of BF4-, the flow cell housed a BF4- ISE (EL7464L) and a reference electrode (4401L). The BF4- ISE is a liquid membrane electrode, which disperses tetrafluoroborate trioctylammonium as an ionophore, in the polyvinyl chloride (PVC) membrane. The ISE can measure the concentration of BF4- anion in the range of 0.1-1000 mg/L as boron, but in practice, 1-100 mg/L in the case of high matrix samples such as process effluents. In contrast, the F- ISE (F2021) is a solid membrane electrode, which can measure 0.02-20000 mg/L as fluoride. Buffer solution (pH7AB) is added to the sample in 10% volume and

the sample flowrate was controlled at 5 mL/min with a peristaltic pump. The potentiometric data of ISEs were collected on a multifunction ion meter (MM-60R) and stored on a linked PC. All the apparatus, equipments and chemicals described here were obtained form DKK-TOA Corporation. Actual FGD wastewater was obtained from full-scale power plants.

Figure 1 Schematic diagram of compact flow measurement system for BF4- and F- ions. RESULTS AND DISCUSSION Easy and rapid determination of BF4-. We confirmed that measurement in a flow mode on this system achieves improved stability and reproducibility in data collection, compared to the conventional static batch measurement. The system can measure BF4- and F- in the range of 1-100 mg/L and the measurement completes in less than 10 min. Figure 2 shows calibration curves of BF4- and Fin simultaneous measurement. BF4- ISE. Each ISE gave a linear calibration to its anion, respectively. However, F- ISE also showed a potentiometric response in BF4- standard solution, which indicates the decomposition of BF4- solution according to reaction (1). This means that F- is formed in small amount immediately after the BF4- standard solution is prepared. We should take care of the decomposition of BF4- standard solution, especially in the case of long-term storage.

Figure 2 Simultaneous measurement of BF4- and F- in (a) BF4- standard solution and (b) F- standard solution. Table 1 summarize the results for actual FGD wastewater samples. Since we were not able to obtain samples containing BF4- (sample A in Table 1), we added BF4- and/or F- to the FGD sample, which are spiked samples B-D in

Table 1. The measured values were obtained with high reproducibility on the proposed system (compact flow), which were consistent with the conventional measurement data (ICP-AES and IC). In the F- spiked samples (sample C and D), the recovery of F- is significantly lower, which is due to the reaction with Ca2+ in the FGD samples to precipitate CaF2. Thus, the compact flow measurement system can be successfully applied to actual FGD wastewaters. The proposed method is easier and more rapid determination compared with the conventional analytical methods. Table 1 Compact flow measurement of BF4- and F- in FGD wastewater. Concentration (mg/L) Conventional Method

Compact Flow

Sample Target

Total Boron

Method

BF4-

F-

BF4-

F-

ICP-AES

IC

IC

ISE

ISE

A

FGD wastewater

89.4

ND

5.4

0.2 (14.6)

6.0 (0.1)

B

Sample A + BF4- (50 m-Bg/L) addition

138

47.7

8.2

47.2 (0.2)

9.2 (0.5)

C

Sample A + F- (50 mg-F/L) addition

85.7

ND

26.0

0.2 (0.7)

27.8 (1.8)

D

Sample A + BF4- (50 mg-B/L) + F- (50 mg-F/L) addition

131.5

50.2

35.7

49.9 (1.7)

24.9 (0.3)

The values by compact flow method are an average of 2 measurements. Parentheses indicate relative standard deviation in %. IC: Ion Chromatography; ISE: Ion Selective Electrode; ND: not detected.

Preparation of stable BF4- standard solution. As describe before, the BF4- standard solution decompose to generate F- immediately after its preparation. Thus, we calculated the chemical equilibrium of the BF4- decomposition according to the procedure reported by Mizoguchi et al.,3 considering the following fluoroborate-related reactions, (2) BF4- + H2O → BF3(OH)- + F- + H+, (3) BF3(OH)- + H2O → BF2(OH)2- + F- + H+, (4) BF2(OH)2- + H2O → BF(OH)3- + F- + H+, (5) BF(OH)3 → H3BO3 + F , and fluoride-related reactions, (6) HF → H+ + F- , (7) 2HF → H2F2 , (8) HF + F- → HF2- . Figure 3 (a) shows the equilibrium fraction of BF4- standard solution. The BF4- fraction depends on the pH and the initial concentration of BF4-. BF4- is stable and predominant at higher initial concentrations of BF4- and lower pH values, whereas it decomposes at the lower initial concentrations and at higher pHs. Thermodynamically, to suppress the decomposition of BF4-, an excess amount of F- should be added to a sample solution and control it at lower pHs. When F- is added to a sample that amounts to 0.1 mol/L and control the sample pH at below 3, the decomposition of BF4- in the sample would be below 1%, as shown in Figure 3 (b).

Figure 3 Equilibrium fraction of BF4- at different concentrations. (a) BF4- alone; (b) BF4- with 100 mmol/L F-.

Based on the findings, we prepare the BF4- standard solutions, in which 0.1 mol/L NaF is added and its pH is controlled at less than 2 with H2SO4, i.e., the modified BF4- standard. Figure 4 compares the BF4- concentration during storage between the ordinary and the modified BF4- standards. As we expected, the BF4- concentration in the ordinary standard showed a monotonous decline and then reached a constant value in the end. The decomposition and time to reach a constant concetration depends on the initial BF4- concentration. In contrast, the modified BF4- standard showed a stable concentration of BF4- during 120 day storage. Since we obtain stable BF4- standard solution, we can eliminate the standard sample preparation step that used to be carried out whenever we employ the compact flow measurement of BF4-.

Figure 4 Time courses of BF4- concentration in BF4- standard solution. Open circle: BF4- standard; closed circle: BF4standard with NaF and H2SO4. CONCLUSIONS We propose a simple and rapid determination method of BF4- in order to manage BF4- emission in FGD effluents. The proposed system employs a flow mode measurement, which achieves improved stability and reproducibility compared to the conventional static batch measurement. The system can measure BF4- in the range of 1-100 mg/L and the measurement completes in less than 10 min. The proposed method is an easier and more rapid determination compared with the conventional analytical methods. We also develop the modified BF4- standard solution that can be subjected to long-term storage. REFERENCES (1) http://www.env.go.jp/en/water/wq/wp.pdf (2) http://www.env.go.jp/en/water/wq/nes.html (3) Katagiri, J.; Yoshioka, T.; Mizoguchi, T. Basic study on the determination of total boron by conversion to tetrafluoroborate ion (BF4-) followed by ion chromatography, Anal. Chim. Acta, 2006, 570, 65-72.

ENHANCING THERMAL EFFICIENCIES IN STEAM POWER PLANTS BY UTILIZING THE “W2” PRIME MOVER AS AUXILIARY EQUIPMENT

Jerry F. Willis Admiral Air, Inc.

1

Abstract: Coal fired power plants are the primary source of electrical power today. There are huge reserves of relatively inexpensive coal resources around the world. The coal fired electrical plant is here to stay. The thermal efficiency of the typical coal fired electric plant is understood to be approximately 35%. This thermal efficiency leaves a lot of room for improvement. Thermal efficiency today is improved by incorporating new boiler and steam turbine technology that produces only minor improvements in thermal efficiency. Increasing thermal efficiencies in power plants will reduce emissions, create potential for carbon credits, and extend coal reserves well into the future. This paper describes a method to substantially improve the thermal efficiencies in existing power plants by adding the new “W2” Prime Mover as auxiliary equipment to steam turbines.

2

System Turbine Characteristics

•

Many variations of steam powered turbines have been used to produce work

•

They are all driven by the flow and expansion characteristics of steam under pressure

Schematic outlining the difference between an impulse and a reaction steam turbine.

3

System Turbine Characteristics

Generally, the thermal efficiency of a coal fired electrical generating plant is 35%. Approximately 15% of the electrical power generated in a typical power plant will be consumed by operating equipment required to run the plant. Adding an auxiliary W2 driven generator designed to operate utilizing the static steam pressure at the end of the last stage of the steam turbine assumed at 300 psi will make up the 15% loss (31.5 MW). In this case, let’s look at Plant Mitchell in the United States. Plant Mitchell has boilers capable of producing 1.4 million pounds of steam an hour, burning 71 tons of coal an hour. Production is 210 MW of electricity per hour.

4

W2 Prime Mover Characteristics

Thermal efficiency of the W2 Prime Mover is in excess of 90%. A W2 Prime Mover requires only the static pressure of steam to function. Steam is consumed at the rate that will keep the system in the desired temperature range and replace steam lost due to small mechanical clearances in the W2. Adding a W2 powered electric generating unit will substantially increase the thermal efficiency of the power plant.

5

W2 Prime Mover Characteristics

The operational action of the W2 Prime Mover is very simple. A wobble plate is attached to the crank arm of a drive shaft with bearings. A wobble plate rotates freely on the crank arm. The drive shaft with the crank arm is attached to a mounting plate with bearings. The drive shaft rotates freely on the mounting plate. The edge of the wobble plate engages the mounting plate at a point notes as “Fulcrum” in the front view of the illustration in Figure 1. Applying a force in the downward direction as shown in “F1” will result in a lateral force at “F2” against the crane arm. The wobble plate will roll down toward the mounting plate at the “F1” location. This action forces the crank arm and drive shaft to rotate in the direction of the wobble plate rotation. The W2 Prime Mover is designed so that high pressure is continually applied to the “F1” side of the wobble plate, resulting in a continuous rotation of the drive shaft.

6

[Figure 1] Drawing illustrating the operational action of the “W2 Prime Mover”.

7

W2 Prime Mover

8

Side View of W2 Prime Mover 3D Prototype

9

Front View of W2 Prime Mover 3D Prototype

10

There has never been a better time to reduce energy requirements at power generating plants. “A typical company spends about 65% of its revenue or more on fuel. While in part the problem is leadership that is pre-occupied with other issues, an even bigger share is tied to the fact that innovation in the energy sector has been lacking.” Editor-In-Chief, Ken Silverstein, noted this premise in an article in Energy Biz Insider (2006). Mr. Silverstein emphasized further, “Consider that for every BTU of coal consumed to drive a steam turbine only 35 percent is converted into useful power. The other 65 percent is lost forever in the form of waste heat that is discharged into the environment”. The steam circulating system for the W2 is much smaller than that required for a conventional steam turbine of the same output.

Advantages of the W2 Prime Mover 

less system heat loss



reduced fuel requirements



higher thermal efficiencies



lower carbon dioxide emissions

The “W2” Prime Mover will: •

Conserve revenue for electric utilities

•

Carbon footprint reductions

•

Generate carbon offsets

•

Replacing steam turbines with the “W2” Prime Mover will generate carbon offsets by directly and indirectly reducing the emissions of carbon to the atmosphere

11

COMPARING EFFICIENCIES OF THE STEAM TURBINE VERSUS THE “W2” PRIME MOVER

Jerry F. Willis Admiral Air, Inc.

1

Abstract: Coal fired power plants operate today using equipment which was designed during an era when fossil fuels were plentiful and relatively inexpensive. Greenhouse gas emissions and thermal efficiencies were of little concern when steam turbines were first developed. This paper compares the system characteristics of a typical coal fired power plant utilizing a steam turbine versus the new concept “W2” Prime Mover. The “W2” Prime Mover will be shown to increase the thermal efficiency of coal fired electric power plants from approximately 35% to values approaching 60%. This efficiency is achieved by utilizing static steam pressure to drive the “W2” Prime Mover whereas steam flow characteristics drive the steam turbine.

2

System Turbine Characteristics

•

Many variations of steam powered turbines have been used to produce work

•

They are all driven by the flow and expansion characteristics of steam under pressure

Schematic outlining the difference between an impulse and a reaction steam turbine.

3

System Turbine Characteristics

Generally, the thermal efficiency of a coal fired electrical generating plant is 35%. The thermal efficiency of a steam turbine can exceed 90%. What causes the thermal efficiency of a steam powered electric power plant to be 35% when the steam turbine is operating at 90%? The answer lies in the systems required to generate and deliver the massive volumes of steam required to run the steam turbines. In this case, let’s look at Plant Mitchell in the United States. Plant Mitchell has boilers capable of producing 1.4 million pounds of steam an hour, burning 71 tons of coal an hour. Production is 210 MW of electricity per hour.

4

W2 Prime Mover Characteristics

Thermal efficiency of the W2 Prime Mover is in excess of 90%. A W2 Prime Mover requires only the static pressure of steam to function. Steam is consumed at the rate that will keep the system at the desired temperature and replace steam lost due to small mechanical clearances in the W2. In this case, look at 210 MW production being driven by the W2 Prime Mover. What would be the estimated steam and coal requirements? .4 Million pounds of steam and twenty tons of coal would produce the 210 MW of power per hour if using the W2 Prime Mover.

5

W2 Prime Mover Characteristics

The operational action of the W2 Prime Mover is very simple. A wobble plate is attached to the crank arm of a drive shaft with bearings. A wobble plate rotates freely on the crank arm. The drive shaft with the crank arm is attached to a mounting plate with bearings. The drive shaft rotates freely on the mounting plate. The edge of the wobble plate engages the mounting plate at a point notes as “Fulcrum” in the front view of the illustration in Figure 1. Applying a force in the downward direction as shown in “F1” will result in a lateral force at “F2” against the crane arm. The wobble plate will roll down toward the mounting plate at the “F1” location. This action forces the crank arm and drive shaft to rotate in the direction of the wobble plate rotation. The W2 Prime Mover is designed so that high pressure is continually applied to the “F1” side of the wobble plate, resulting in a continuous rotation of the drive shaft.

6

[Figure 1] Drawing illustrating the operational action of the “W2 Prime Mover”.

7

W2 Prime Mover

8

Side View of W2 Prime Mover 3D Prototype

9

Front View of W2 Prime Mover 3D Prototype

10

There has never been a better time to reduce energy requirements at power generating plants. “A typical company spends about 65% of its revenue or more on fuel. While in part the problem is leadership that is pre-occupied with other issues, an even bigger share is tied to the fact that innovation in the energy sector has been lacking.” Editor-In-Chief, Ken Silverstein, noted this premise in an article in Energy Biz Insider (2006).

Mr. Silverstein emphasized further,

“Consider that for every BTU of coal consumed to drive a steam turbine only 35 percent is converted into useful power. The other 65 percent is lost forever in the form of waste heat that is discharged into the environment”. The steam circulating system for the W2 is much smaller than that required for a conventional steam turbine of the same output.

Advantages of the W2 Prime Mover •

less system heat loss

•

reduced fuel requirements

•

higher thermal efficiencies

•

lower carbon dioxide emissions

The “W2” Prime Mover will: •

Conserve revenue for electric utilities

•

Reduce carbon footprint

•

Generate carbon offsets

•

Replacing steam turbines with the “W2” Prime Mover will generate carbon offsets by directly and indirectly reducing the emissions of carbon to the atmosphere

11

2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 OPTIMIZATION OF FUEL PROPERTIES WITH UTILIZATION OF BIODEGRADABLE MUNICIPAL WASTES FOR COMBUSTION UNITS JUCHELKOVA Dagmar, HAJKOVA Martina, RACLAVSKA Helena, TARARIK Luboš1 VSB – Technical University Ostrava, 17. listopadu 15, Ostrava, Czech Republic [email protected] 1 Frydecka skladka, a.s., Panské Nové Dvory 3559, Frydek-Mistek, Czech Republic

Introduction Alternative fuel is created by combination of two or more various material compounds and their optimized weight shares. A suitable choice and structure of the compounds enable attainment of needed qualitative parameters under complying with limit parameters of contaminants, required calorific value under parallel utilization of energy content of the all suitable shares. As the first material, brown coal of lower quality from Most Lignite Coal Basin was used (Fig.1). The second material is represented by solid municipal waste (MSW) from small settlements up to 2,000 citizens with accent on utilization of biodegradable waste (BMW). Effort in utilization of BMW as a fuel is conditioned by need of minimization of waste landfilling. New strategies in municipal solid waste (MSW) management, i.e., a separate collection of the organic fraction (EU Directive 1999/31/EC, EU Directive 2008/98/EC) and a reduction of the biodegradable MSW fraction allocated in landfills (EU Directive 2003/33/EC), have favoured their energetic utilization and composting. Fig.1 Lignite Coal Basins in the Czech republic With Libous Open Pit Mine

The production of municipal waste in the Czech Republic currently represents merely 306 kg per an inhabitant and a year. From this amount 83 % will be deposited in landfills, only 13 % of waste is incinerated, 2 % recycled and 2 % are used for production of compost (Report EEA). The aim of the investigation was verification of possibilities of fuel preparation while meeting conditions of the Czech legislative (specific sulphur content and calorific value). The fuel will be prepared from lignite of lower quality and from biodegradable waste and at the same time, the fuel will meet required mechanical properties expressed by PDI index (pellet durability index). Material for alternative fuel preparation Solid Municipal Waste (MSW). Composition of the solid municipal waste keeps changing in different localities in all states (Amlinger at al. 2010) and it is dependent on many factors as regional

differences, climatic conditions, a frequency of separate waste collection, a season of the year, cultural customs (Tchobanoglous et al., 1997), development trends in community, living standards of citizens, settlement's size, built-up area character, and a kind of heating. A share of biodegradable waste in the solid municipal waste is 57% in FRG, 62% in Italy and 52% in Hungary (Herczeg M. at al., 2009). Average structure of the MSW in EU is the following: paper and cardboard - 35%; kitchen and garden waste - 25%; plastics - 11%; glass - 6%; metals - 3%; textiles 2%; others - 18% (Calabrò 2009). Estimation of individual fractions in the municipal waste according to Brebu et al. 2010 might be considered as a very realistic one. It specifies 20 % share of inorganic fraction, up to 65 % share of ligno-celluloses materials (paper, cardboard, cartons, dietary residue of vegetal and animal origin) and about 20% share of plastics. Table 1: Abundance of individual compounds in municipal waste (%) Waste 1. Plastics 2. Glass 3. Textiles 4. Metals 5. Inert Material 6. Paper & Cardboard 7. Garden & Assorted BMW Ʃ BMW (3+6+7) Number of Samples

Frycovice Annual Average 15,3 ± 7,0 6,2 ± 5,0 6,5 ± 4,0 5,0 ± 0,9 21,0 ± 9,9 17,8 ± 7,5 28,3 ± 2,7 52,56 6

FM Municipalities (Jul. – Oct. 2009) March 2010

20,0 ± 9,5 11,5 ± 7,7 5,9 ± 7,3 4,8 ± 3,2 8,5 ± 8,4 17,9 ± 9,1 31,0 ± 13,6 54,80 11

14,4 ± 7,2 4,7 ± 3,9 3,1 ± 2,4 3,5 ± 2,3 27,6 ± 7,7 11,5 ± 6,2 35,7 ± 5,6 50,27 6

EU Calabro, 2009 11 6 2 3 18 35 25 62

Results of municipal waste assortment in the municipalities up to 2,000 citizens with village buildings are stated in Table 1. Composition of the municipal waste in Frycovice municipality was monitored within a year (2 months cycle of samples collection). The samples were collected at the municipality waste landfill. Abundance of the individual fractions in the municipal waste from FM municipalities was calculated after separation of waste from containers of 110 liters capacity which were transported to Frýdek Místek AG Land field. Comparing results (Calabro et al. 2009), the waste from village built-up area has considerably lower content of paper fraction and considerably higher content (up to by 10 %) of inert material. The municipal waste collected in March 2010 had plastics depleted by materials with high calorific power (PET, PVC) due to the „winter season“. Especially flexible foils (HDPE, LDPE and PP) stood in for plastics. Also quite high amount of children napkins was a part of the MSW. Current disposable nappies consist of a pervious foil (made of polypropylene mainly), a moisture absorbing napkin (cellulose filled up by gel absorber on polyacrylate base by reason of improvement of the absorptivity), an impervious foil for cloth protection (made of polyethylene), adhesive strips or burfasteners and elastic bands. For purposes of BMW share calculation, the children napkins were classified from 50 % as biodegradable waste and from 50% as plastics. For the fuel production, the utilization of separated BMW from the solid municipal waste was supposed. From the proximate analysis (Table 2), it is obvious that separated part of BMW contains still considerable share of inorganic fractions which will decrease BMW's calorific value. For achievement of the specific sulphur content limit in the proposed fuel, in the case of brown coal of lower quality utilization, it is necessary to ensure in order that the second fraction was of higher calorific value. However, BMW does not meet this requirement. For achievement of the calorific power improvement, also separated flexible plastics were again added to the BMW sample (Table 3)

Table 2: Ash content and calorific value separated BMW in dry matter FM Municipalities Unit Jul. – Oct. 2009 March 2010 o Ash (815 C) % 15.52 ± 12.35 15.56 ± 7.88 Net Calorific Value MJ/kg d.m. 16.23 ± 5.24 13.68 ± 5.57 Gross Calorific Value MJ/kg d.m. 13.35 ± 4.28 11.75 ± 6.75

Frycovice Annual Average 48.44 ± 28.69 11.58 ± 6.54 10.87 ± 7.69

Fig.2 Percental abundance of fraction in municipal waste after elimination of inert material (metals and glass). Brown Coal. Coal sample was removed from Libouš Coal Pit. Libouš Coal Pit is the largest coal pit in Most Lignite Basin (the Czech Republic). The Most Lignite Basin forms a relic of tertiary sedimentary deposit. Libouš Coal Pit is controlled by the North Bohemian Lignite Mines, AG, Chomutov. The brown coal bed is mined at Tušimice-Libouš coal deposit with thickness from 25 to 35m and average ash content of 36.8 % in anhydrous state and with the sulphur content of 2.7 % and calorific power of 10.40 MJ/kg in the original state. Total exploitable coal deposit is about 286 million tons. The exploited coal is utilized for production of energy fuel mixtures. Table 3: Ultimate and proximate analysis of waste and fuels MW a

W Ad Std St Cd Hd Nd Qs d Qi r

% % % % % % % MJ/kg MJ/kg

As

mg/kg d.m.

Cd

mg/kg d.m.

Cr

mg/kg d.m.

Hg Ni

mg/kg d.m. mg/kg

3,38 14,30 0,14 0,13 47,73 6,95 1,00 25,38 23,86 0,461 ± 0,067 0,347 ± 0,046 19,7 ± 2,2 0,006 ± 0,001 3,88 ±

BMW 2,68 6,30 0,08 2,60 36,24 4,54 1,24 13,68 8,56 0,116 ± 0,021 0,116 ± 0,018 12,6 ± 1,21 0,003 ± 0,0015 2,14 ±

Coal 23,54 41,93 2,62 2,02 33,76 3,75 0,84 15,80 9,13 33,8

MW80+Coal20

MW60+Coal40

MW40+Coal60

19,826 0,636 0,51 44,936 6,31 0,968 23,46 20,91 7,12

25,352 1,132 0,886 42,142 5,67 0,936 21,54 17,96 13,79

30,878 1,628 1,264 39,348 5,03 0,904 19,632 15,02 20,46

2,25

0,72

1,10

1,48

126

40,96

62,22

83,48

2,08

0,42

0,83

46,8

12,46

21,04

1,25 29,63

d.m.

0,71 0,44 35,2 ± 24,8 ± 1040 236,16 638,08 mg/kg Pb 7,0 4,6 437,12 d.m. 2,15 ± 1,88 ± 65,6 14,84 27,53 mg/kg V 0,27 0,12 40,22 d.m. 1596 ± 564 ± 66 3150 1906,8 2217,6 mg/kg Chlorine 160 2528,4 d.m. Explanations: Ad – Ash in anhydrous state, Qsd – Calorific value in anhydrous state, Qir – Calorific value in original state, Std – Sulphur in anhydrous state

Fuel for middle-sized energy sources According to Direction No.13/2009, Definition of Requirements on Fuels Quality for Stationary Sources from Air Conservation Point of View, fuel with specific sulphur content < 0.95 g.kJ-1 (in original state) and calorific value higher than 10 MJ.kg-1 might be burnt off in big and small sources. The direction does not specify which fractions or their parts can be used for the fuel preparation. Meeting the condition of specific sulphur content will be ensured by preparation of the fuel with minimal content of 40 % MSW and 60 % coal. The original aim was creation of the fuel which would utilize only a separated fraction BMW from MSW for improvement of CO2 emission limits. Considering that separated fraction BMW from MSW in municipalities up to 2,000 citizens has too low calorific value of 4 – 8 MJ/kg (in supplied state) and coal from Libouš Coal Field has a high content of sulphur, the BMW fraction with additive of flexible plastics would have to be used from MSW for meeting the specific sulphur content condition. Ballast fractions were separated (glass, metals, inert material). The condition of specific sulphur content is satisfied also in the case of the fuel prepared from 60% of coal and 40% of MW. For production of the mixed fuel, the coal with size of 0 – 40 mm was grained to size lesser than 5 mm. This size appears to be optimal for usage of JGE 150 laboratory pelletizing press (Pest Control Corporation Company, Vlcnov, the Czech Republic). MSW was dried and further grained in 9FQ 50 hammer grinding mill with power of 800 kg/h (Pest Control Corporation Company, Vlcnov, the Czech Republic) to size by 35 % > 1 mm. For measurement of pellets mechanical durability (PDI index – pellet durability index) the Holmen tester was used. The PDI is calculated as the ratio of weight after tumbling and the weight before tumbling, multiplied by 100. A sample is circulated in the Holmen tester for 60 s. The remaining pellets are collected, sieved with sieve size of about 80 % of pellet diameter, weighted, and the PDI is calculated. In the case of humidity higher than 10 %, the PDI value might be influenced by the humidity (Arshadi et al.2008). The PDI value of all samples was measured at the same humidity (4 – 5%).

Table 4: Pellet durability index (PDI - %) MW BMW Coal MW80+Coal20 88 96 78 92

MW60+Coal40 86

MW40+Coal60 82

It is obvious from the table that PDI index value keeps decreasing with growing of coal content in pellets. Biodegradable waste has high PDI index, however, it might not create function of binding agent for the coal fraction. For solution of this problem CaO, Ca(OH)2, CaCO3 or another material with high content of lignine (compost) can be used as a plasticizing agent. At the same time, it supports self-desulphuring effect of the proposed fuel.

Conclusions Separated municipal waste (BMW and flexible plastics) might form an important energy fraction in preparation of the fuel for middle-sized energy sources that utilize brown coal of lower quality with a specific high content of the sulphur. The specific sulphur content parameter that is required by Direction No. 13/2009 is met in the fuel mixture consisting of 40% of separated MW and 60% of coal. However, the prepared fuel in such way does not comply with requirements on mechanical durability, a share of the biodegradable material is insufficient in order that so called "solid bridges", which bind quite heterogeneous material, were created among the individual grains. Successful utilization of the fuel on the base of municipal waste base and coal might be recommended not earlier than after suitable binding material was found.

Acknowledgement This work was supported by research projects of Ministry of Education, Youth and Sport of the Czech Republic: INTERVIRON No. 2B06068. References Arshadi M., Gref R., Geladi P., Dahlqvist S.A., Lestander T. (2008): The Influence of Raw Material Characteristics on the Industrial Pelletizing Process and Pellet Quality. Fuel Processing Technology, V.89, 1442-1447 Brebu M., Ucar S., Vasile C., Yanik J., 2010: Co-pyrolysis of Pine Cone with Synthetic Polymers, Fuel doi: 10.1016/j.fuel.2010.01.029 Diverting Waste from Landfill, Effectiveness of Waste-Management Policies in the European Union, EEA Report, Copenhagen, Denmark, 2009

Manuscript Not AVAILABLE

2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission

PROGRAM TOPIC: COMBUSTION SULFUR RETENTION IN THE ASH DURING COMBUSTION OF TUNCBILEK BRIQUETTES

Ayfer Parlak, Turkish Coal Enterprises (TKI), GLI, Tuncbilek, Kutahya, Turkey [email protected] Bekir Zühtü Uysal, Department of Chem. Eng., Faculty of Engineering and Clean Energy Research and Application Center, Gazi University, Ankara, Turkey Mustafa Ozdingis, Turkish Coal Enterprises (TKI), Ankara, Turkey H. Köksal Mucuk, Turkish Coal Enterprises (TKI), Ankara, Turkey Selahaddin Anac, Turkish Coal Enterprises (TKI), Ankara, Turkey

Abstract: In this study, it was aimed to briquette the mixture of Tuncbilek lignite in the size range of 0.5-10 mm and Tuncbilek Coal Washing Plant’s slurry waste involving 0.1-0.5 mm particles with molasses and dolomite bindings and to achieve retention of sulfur in the ash during combustion. In this way coal fines with high sulfur content was converted by briquetting into a strong and uniform fuel giving rise to low sulfur emission upon combustion. In addition, Coal Washing Plant’s slurry waste which is currently not being utilized has been converted into a usable form in industry by the addition to coal fines.

Different amounts of dolomite and waste were used in briquetting and their effects on the efficiency of retention of sulfur in the ash were examined. A pilot equipment with 100 kg capacity was used for briquetting. Briquettes were burned in a TGA analyzer and in a commercial stove, and sulfur analyses were made in the ash. A mixture of 67% lignite, 20% slurry, 7% molasses and 6% dolomite gave the best result. Sulfur retention in the ash was found as 74.6% and 60.2% by TGA and stove tests, respectively.

Types of sulfur in the original Tuncbilek lignite were also determined experimentally. Since original sulfate in the coal is not affected in combustion, the retention efficiency of sulfur in the ash was also evaluated on the basis of combustible sulfur and was found to be 90.8 % and 73.2 % by TGA and stove tests, respectively. Thus, briquetting provided a decrease in sulfur dioxide emission not less than 70%. Proximate analyses and heating values of the briquettes were also made and the results indicated that the briquettes obtained were marketable.

Development of an Analytical Solution for Jet Diffusion Flame Equations ´ Felipe Norte Pereira∗ , Alvaro Luiz de Bortoli∗ , Nilson Romeu Marcilio∗ ∗

Federal University of Rio Grande do Sul (UFRGS), Department of Chemical Engineering. Porto Alegre, Brazil. E-mail: [email protected], [email protected], [email protected]

Abstract—Exact solutions of nonlinear equations help us to understand the mechanism of nonlinear effects. Any exact solution for flame propagation must make use of basic equations of fluid dynamics modified to account for the liberation and conduction of heat and for charges of chemical species within the reaction zone. The present work develops an analytical solution for a jet diffusion flame. The model is based on the solution of the mixture fraction for turbulent fluid flow with no constant eddy viscosity and the results are found to compare favorably with data in the literature.

I. I NTRODUCTION Combustion theory is an important area of classical phenomenology, which treats a wide range of natural phenomena. In general, combustion is fast compared to molecular mixing, occurring in layers thinner than the typical length scales of turbulence, especially for high Damk¨ohler values. Flamelets are thin reactive diffusive layers embedded within an otherwise nonreacting flow field. The basic idea is to assume that a small instantaneous flame element embedded in a turbulent field has a structure of a laminar flame. The flamelet concept can be employed in most practical situations, when the chemistry is fast. To solve nonpremixed jet flames many approximate models are found in literature: Williams (1985), Li˜na´ n (1991), Peters and Donnerhack (1981), Peters (2000), Warnatz et al. (2001), Veynante and Vervish (2003), Fern´andez-Tarrazo et al. (2006). The analytical developments present in this paper were motivated by the following works: — Peters and Donnerhack (1981), which calculates the mixture fraction of an axisymmetric flame, considering the turbulent viscosity constant throughout the entire jet, using a similarity transformation. — Agrawal and Prasad (2003), which developed analytical expressions for the eddy viscosity for turbulent jets, plumes and wakes, using a semi-empirical equation for the axial velocity. The current work aims to determine an expression for the axial velocity and mixture fraction of a confined jet diffusion flame, using the similarity transformation applied by Peters and Donnerhack (1981), but considering the eddy viscosity, obtained by Agrawal and Prasad (2003), which varies along the length and radius of the jet. The expression obtained for the mixture fraction is then compared with experimental results for a hydrogen flame.

Fig. 1.

Schematic representation of a vertical jet flame

II. T URBULENT J ET D IFFUSION F LAMES INTO S TILL A IR It was considered a diffusion flame in which the fuel, delivered from a round nozzle, with diameter d and exit velocity u0 , mixes with the surrounding air by convection and diffusion. The jet flame is chosen because it seems to be a representative of the class of nonpremixed flames. The effects of buoyancy and pressure gradients in the flame have been neglected. The Fig. 1 shows the structure of a diffusion flame in cylindrical coordinates. This lead to a two-dimensional axisymmetric problem governed by equations of Continuity (Eq. (1)), Momentum in x-direction (Eq. (2)) and Mixture Fraction (Eq. (3)) in steady state. ρv˜r) ∂(¯ ρu ˜r) ∂(¯ + ∂x ∂r ∂u ˜ ∂u ˜ ρ¯u ˜r + ρ¯v˜r ∂x ∂r

=

0

(1)

=

  ∂ ∂u ˜ ρ¯ν˜t r ∂r ∂r

(2)

∂ Z˜ ∂ Z˜ ρ¯u ˜r + ρ¯v˜r ∂x ∂r

∂ ∂r

=



ρ¯ν˜t r ∂ u ˜ Sc ∂r

 (3)

The boundary layer assumption was considered in Eq. (2) and Eq. (3). Also, in Eq. (2), the viscous stress was neglected compared to the Reynolds stress component, which was modeled as show in Eq. (4). ∂u ˜ (4) ∂r The equations (1), (2) and (3) are subjected to the following boundary conditions (Eq. (5)).

The variable Uc corresponds to the centerline velocity, which for an axisymmetric jet varies as x−1 . The variable c corresponds to the spread rate of the jet. To simplify the calculation, it was considered that the mixture fraction is a product of two functions: the centerline mixture fraction Z˜cl and the variable ω, as shown in the Eq. (12).

00 00 −¯ ρug v = ρ¯ν˜t r

Z˜ = Zcl (5) ∂u ˜ r → ∞ u ˜ = 0, = 0, Z˜ = 0 ∂r The Continuity Equation (1) is automatically satisfied by introducing the stream function (ψ) (Eq. (6)). r

=

0

r¯2 = 2

Zr

ρ¯ rdr, ρ∞

ζ = x + x0

C=

C=

∂ ∂η ∂ r¯ ∂ = ∂r ∂ r¯ ∂r ∂η

(8)

Introducing the adimensional stream function F (Eq. (9)), the axial and radial velocity may be written as a function of F (η) (Eq. (10)), as follows: F =

ψ ρ∞ νtr ζ

(9)

  1 ∂ (νtr F ) ∂ (νtr F ) u ˜= , ρ¯v˜r = −ρ∞ νtr F − η ηζ ∂η ∂η (10) The variable νtr present in Eqs. (9) and (10) corresponds to the eddy viscosity. To quantify the variation of the eddy viscosity, it was employed a semi-empirical expression obtained by Agrawal and Prasad (2003) for an axisymmetric jet diffusion flame (Eq. (11)). νtr

Uc c4 = x 4



1 − exp(−ξ 2 ) ξ2

 ,

ξ=

r c·x

(11)

(13)

(ρ0 ρst )1/2 ρ∞

(14)

Substituing the equations (10) and (12) into equations (2) and (3), and applying the chain rule described in equation (8), results the equations (15) and (16)   ∂ (νtr F ) ∂(νtr F ) = ∂η η ∂η    ∂ ∂ 1 ∂(νtr F ) = Cνtr η ∂η ∂η η ∂η   ∂ C ∂ω ∂ ηνtr − ((νtr F )ω) = ∂η ∂η Sc ∂η −

0

∂ ∂ ∂η ∂ = + , ∂x ∂ζ ∂ζ ∂η

ρ¯2 νt r2 ρ¯2∞ νtr r¯2

According to Peters and Donnerhack (1981), the ChapmanRubesin parameter may be written as:

(7)

The Eq. (7) contains a density transformation leading to the density weighted radial coordinate r¯ . The new axial coordinate ζ corresponds to the sum between the axial coordinate x and the flame dislocation from the nozzle x0 . With the similarity transformation, the coordinate system (x, r) is transformed to (ζ, η) using the chain rule show in Eq. (8).

(12)

One introduces the Chapman-Rubesin parameter, C, in equation (13), and assumes it as a constant in the entire jet.

v˜ = 0,

∂ψ ∂ψ , ρ¯v˜r = − (6) ρ¯u ˜r = ∂r ∂x A similarity transformation is employed, following the work of Peters and Donnerhack (1981), in order to solve the equations analytically (Eq. (7)). r¯ η= , ζ

Z˜ = Z˜cl · ω

(15)

(16)

subject to the following bondary conditions (Eq. (17)): η r

= →

0: ∞:

F = 0,

ω=1 0

(νtr F ) = 0,

(17) 00

(νtr F ) = 0,

ω=0

To solve the two differential equations (Eq. (15) and (16)) analytically, the exponential term of equation (11) was expanded in a Taylor Series, truncated in the second term. III. C ALCULATION OF THE A XIAL V ELOCITY AND M IXTURE F RACTION Expanding the exponential term of equation (11) in a Taylor series truncated in the second term, results: exp(−ξ 2 ) ≈ 1 − ξ 2

(18)

With this consideration, the eddy viscosity can be written as shown in Eq. (19). Uc c4 x (19) 4 Now the eddy viscosity becomes independent of ζ and η, since Uc ∝ x−1 . Thereby, the equations (15) and (16) can be simplified, resulting in: νtr =

∂ − ∂η



   ∂ ∂ 1 ∂F = Cη ∂η ∂η η ∂η   ∂ ∂ C ∂ω − (F ω) = η ∂η ∂η Sc ∂η F ∂F η ∂η



(20)

(21)

Solving the equation (20), one obtains: F =

C(γη)2 1+

(γη)2 4

Z 0

∞

 −2 (γη)2 1+ 4

  ∞ ∂u ˜ ∞ ρ¯u ˜2 rdr + [¯ ρu ˜v˜r]0 = ρ¯νtr r ∂r 0

(23)

(24)

where ρ0 is the density of the fuel stream and R is the radius of the jet noozle. The substitution of equation (23) into equation (25), results in an expression to the constant γ (Eq. (26)). 3 64



ρ0 ρ∞ C 2



u ˜0 d νtr

2 (26)

In order to quantify the variation of the Schmidt Number (Sc), it was considered that Sc varies linearly with the mixture fraction, as shown in Eq. (27) Sc = (ScF − Scox )Z + Scox

(28) A2 = Scox

Substituiting the equations (22) and (28) into equation (21), it is possible to integrate the equation (21) to determinate ω (Eq. (29)). "

(γη)2 (A1 + A2 ) 1 + 4 

#−1

2A2 − A1

(29)

By combining the mixture fraction equation (Eq. (3)) with the continuity equation (Eq. (1)) and integrating from r = 0 to r → ∞, in a similar way as in Eq. (24), one finds that the integrated mass flow rate must be independent of x (Eq. (30)). Z ∞ 2 ˜ = ρ0 u0 R ρ¯u ˜Zdy (30) 2 0 The combination of equations (12), (29), (23) in (30) results in the mixture fraction along the jet centerline Zcl (Eq. (31)) as Z˜cl θ

=

1 32A2 Z ∞

=



u0 d νtr 2



(ζ/d)−1 θ 2A2

(µ ((A1 + A2 )µ

(31) −1

− A1 ))

−1 dµ

1

With the bondary conditions given by Eq. (5), the last two terms in Eq. (24) are zero. It follows that the first integral denoting the jet momentum is independent of x and therefore constant, equal to the momentum in x = 0. Assuming a constant exit velocity across the orifice with the radius R, the jet momentum can be written as in Eq. (25) Z ∞ ρ0 u0 2 R 2 ρ¯u ˜2 dy = (25) 2 0

γ2 =

A1

= (ScF − Scox )Z˜cl

ω = A2

The Eq. (23) shows that the axial velocity in the centerline varies as x−1 , which agrees with the previous consideration made for Uc in the equation (11). Combining the momentum equation with the continuity equation and integrating from r = 0 to r → ∞ (Eq. (24)), one obtains: ∂ ∂x

= A1 ω + A2

(22)



where the variable γ is a constant (concerning to η) obtained in the integration of Eq. (20). The substitution of Eq. (22) into Eq. (10) results in an expression for u ¯ in terms of (ζ, η), Eq. (23), as: 2νtr Cγ 2 u ˜= ζ

Sc

(27)

where ScF and Scox correspond to the Schmidt Number for the fuel and the oxidizer, respectively. To express the variation of the Schmidt Number as dependent of (Z˜cl , ω) (Eq. (28)), the equation (12) was inserted into equation (27).

where µ is the variable of integration in the equation for θ. It was considered that A1  A1 + A2 in order to simplify the equation (31). The difference between the Schmidt Number of fuel and oxidizer is small, and when this value is multiplied by Zcl , it became even smaller, which justifies the simplification. Thus the equation (31) can be written as (32). Z˜cl

  u0 d 1 ρ0 (ζ/d)−1 (32) 32 ρ∞ C νtr −1 −1 (2Scox + 1) − (ScF − ScOx ) Scox 

= Scox

The Mixture Fraction can now be obtained by multiplying the equations (29) and (32). IV. C OMPARISON WITH E XPERIMENTAL DATA The analytical results for the mixture fraction was compared with experimental results from Barlow (2010) for a turbulent 50/50% hydrogen/nitrogen jet flame (where ρ0 /ρst = 0.24, ScF = 0.45 and Scox = 0.90). The figure 2 shows the comparison of the analytical and the experimental mixture fractions along the flame centerline. The mixture fraction result corresponds to the axial decreasing behavior of a commom free jet. The figures 3, 4 and 5 show the comparison of the analytical and the experimental mixture fractions for x/d equal to 20, 40 and 60 respectively.

The analytical solution for the mixture fraction along the jet centerline is valid for x/d > 10 as seen in figure 2. The expression for the mixture fraction varies with x in a way similar to x−1 . Thereby the solution tends to infinity for small values of x/d, which makes the solution unable to represent the jet entrance. For the mixture fraction along the jet radius (Fig. 3, 4 and 5), the analytical solution has a behavior similar to the experiment. Overall, the results show a greater discrepancy at the beginning of the flame, which tend to decrease as the radius value increases. Fig. 2. Comparison between the analytical and experimental mixture fraction along the jet centerline

V. C ONCLUSION Overall, the results were satisfactory. The analytical expression obtained for the mixture fraction seems to represent well a hydrogen jet, except at the beggining of it (small values of x). In terms of the radial variation, the solution tends to have lower discrepancies as the radius value increases. The radial profile of the mixture fraction may be reasonably approximated using a Gaussian function (Agrawal and Prasad, 2003, De Bortoli, 2009). Therefore, the next step of this study is to improve the eddy viscosity formula to decrease the jet spreading along its streamwise direction. VI. N OMENCLATURE

Fig. 3. Comparison between the analytical and experimental mixture fraction at x = 20d

C

=

c =

Parameter in the eddy viscosity equation

d =

Nozzle diameter

F

=

Adimensional stream function

r

=

Coordinate along the jet radius

r¯ =

Fig. 4. Comparison between the analytical and experimental mixture fraction at x = 40d

Density weighted radial coordinate

ScF

=

Schmidt number of the combustible

Scox

=

Schmidt number of the oxidant

u0

=

Exit velocity from the nozzle

u ˜ =

Favre averaged axial velocity

v˜ =

Favre averaged radial velocity

00

u ,v

00

=

x = Z˜ = Z˜cl =

Fluctuating components of velocity Coordinate along the axis Favre averaged mixture fraction

ρ0

=

Favre averaged mixture fraction along the jet centerline Fuel density

ρ∞

=

Oxidant density

ρ¯ =

Fig. 5. Comparison between the analytical and experimental mixture fraction at x = 60d

Chapman-Rubesin parameter

Reynolds averaged density

νt

=

Viscosity

νtr

=

Eddy viscosity

ψ

=

Stream function

ζ, η

=

Coordinates used in the similarity transformation

ω

=

Normalized mixture fraction

ACKNOWLEDGMENT The authors gratefully acknowledge the finantial support from CNPq (Conselho Nacional de Desenvolvimento Cientifico e Tecnol´ogico) under process 303007/2009-5. R EFERENCES A. Agrawal and A. Prasad. Integral solution for the mean flow profiles of turbulent jets, plumes, and wakes. Journal of Fluids Engineering, 145, September 2003. R. Barlow. H2 /N2 jet flames, Consulted on March 2010 at www.ca.sandia.gov/TNF. A. L. De Bortoli. Analytical/numerical solutions for confined jet diffusion flame (Sandia Flame C). Latin American Applied Research, (34):157–163, 2009. E. Fern´andez-Tarrazo, A. L. S´anchez, A. Li˜na´ n, and F. A. Williams. A simple one-step chemistry model for partially premixed hydrocarbon combustion. Combustion and Flame 147, pages 32–38, 2006. A. Li˜na´ n. The structure of diffusion flames in fluid dynamical aspects of combustion theory. Longman Scientific and Technical, UK, 1991. N. Peters. Turbulent Combustion. Cambridge University, Press, 2000. N. Peters and S. Donnerhack. Structure and Similarity of Nitric Oxide Production in Turbulent Diffusion Flames. 18th Symp. (Int.) on Combustion, The Combustion Institute, 1981. D. Veynante and L. Vervish. Turbulent combustion modeling. Lecture Series du Von K´arman Institute, 2003. J. Warnatz, U. Maas, and R. W. Dibble. Combustion: Physical and Chemical Fundamentals, Modeling and Simulation, Experiments, Pollutant Formation. Springer-Verlog, 3 edition, 2001. F. A. Williams. Combustion Theory. Addison-Weskey, Redwood City, CA, 2 edition, 1985.

Manuscript Not AVAILABLE

2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission

PROGRAM TOPIC: COMBUSTION (FLUE GAS CLEAN UP) EXACT TITLE OF PAPER: IONIC LIQUIDS WITH AMINE FUNCTIONAL GROUP: A SHORCUT TO IMPROVE THE PERFORMANCE OF IONIC LIQUIDS FOR CO2 SCRUBBING Presenting Author: Jelliarko Palgunadi, Department of Chemistry and Research Institute of Basic Science, Kyung Hee University, 1 Hoegi-dong Dongdaemoon-gu, Seoul 130-701, Republic of Korea Contact Information: E-mail: [email protected]; Telp: +82-2-961-0431; Fax: +82-2-966-3701 Co-Authors: Jin Kyu Im; Antonius Indarto, Dr; Hoon Sik Kim, Prof.; and Minserk Cheong, Prof., Department of Chemistry and Research Institute of Basic Science, Kyung Hee University, 1-Hoegi-dong Dongdaemoon-gu, Seoul 130-701, Republic of Korea Contact Information: E-mail: [email protected] (H.S. Kim); [email protected] (M. Cheong); Telp: +82-2-961-0432; Fax: +82-2-966-3701

Abstract: One of the global environmental problems of today is the increase of the greenhouse gases concentrations in the atmosphere. This problem partly corresponds to the increase of carbon dioxide (CO2) emission from the burning of fossil fuels for power generations. To response this challenge, carbon capture and storage (CCS) using liquid scrubber receives great attentions because there is potential for retrofit to existing power plants without changing the existing process. Within this framework, ionic liquids (ILs) have been proposed as alternative media for scrubbing CO2 from post-combustion emission where SOx, NOx, and tiny particulates are also inevitably co-produced. Due to the ionic nature of these low-melting point salts, the problem associated with the solvent lost during cycled absorption-desorption processes might be minimized. Experimental results from our group combined with numerous published data demonstrated that the CO2 solubility in many conventional ILs at the pressure close to atmosphere is very low. Thus, these conventional ILs are not feasible compared to the molecular amine-based scrubbers, i.e. monoethanol amine (MEA) for capturing CO2 contained in a post-combustion stream. Regular solution theory and quantum chemical calculations demonstrate that the CO2 solubility in common ILs is merely controlled by weakly physical interactions (i.e. van der Waals interactions). To improve the performance of ILs, various task-specific ILs dissolved in a simple room temperature IL or in a non-volatile organic solvent were evaluated for CO2 capture at low pressures. Cost-effective ILs containing an amine moiety were prepared by quaternarization of commercially available diamines. Similar CO2 loading capacity as found in a molecular amine system was observed likely through the formation of a carbamate salt-like structure. The CO2 solubility measurements, the computational calculations of the CO2-IL complexes, and some factors associated with the optimum absorption conditions are discussed.

Manuscript Not AVAILABLE

2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission

PROGRAM TOPIC: COMBUSTION (FLUE GAS CLEAN UP) EXACT TITLE OF PAPER: ABSORPTION OF SULFUR DIOXIDE IN TASK SPECIFIC IONIC LIQUIDS CONTAINING SO2-PHILIC GROUPS ON THE CATION Presenting Author: Sung Yun Hong, Department of Chemistry and Research Institute of Basic Science, Kyung Hee University, 1 Hoegi-dong Dongdaemoon-gu, Seoul 130-701, Republic of Korea Contact Information: E-mail: [email protected]; Telp: +82-2-961-0431; Fax: +82-2-966-3701 Co-Authors: Jelliarko Palgunadi; Hoon Sik Kim, Prof.; and Minserk Cheong, Prof., Department of Chemistry and Research Institute of Basic Science, Kyung Hee University, 1-Hoegi-dong Dongdaemoon-gu, Seoul 130-701, Republic of Korea Contact Information: E-mail: [email protected] (H.S. Kim); [email protected] (M. Cheong); Telp: +82-2-961-0432; Fax: +82-2-966-3701

Abstract: Fossil fuel burning power plant is one of major producers of sulfur dioxide emissions worldwide. To mitigate the emissions of SO 2, scrubbing process employing liquid absorbents is considered as an alternative method in addition to the flue gas desulphurization (FGD). Room temperature ionic liquids (RTILs) have been demonstrated to absorb SO2 effectively. Tunable psychochemical properties derived from the combinations of tailored ionic components and non-volatility of RTILs resulted from the coulombic force stabilization are the key features making these materials more attractive than volatile organics for SO2 capture. Some literatures suggest that the formation of specific interactions such as Lewis acid-base interactions control the SO2 solubility. Thus, the presence of SO2-philic groups on the molecular structure of RTILs is required to achieve high SO2 solubility. In our group, imidazolium-based cations containing various pendant groups with capabilities to form specific yet reversible interactions with SO2 combined with [MeSO3]- as the anion were synthesized. The solubility measurements for SO2 are presented and the effects of the cation structure on the absorption-desorption processes are discussed.

2010 International Pittsburgh Coal Conference

REACTION CHARACTERISTICS OF NEW OXYGEN CARRIERS FOR CHEMICAL LOOPING COMBUSTION Ho-Jung Ryu1,*, Jaehyeon Park1, Gyoung-Tae Jin1, Moon-Hee Park2 1

Korea Institute of Energy Research, 102 Gajeong-ro, Yuseong-gu, Daejon 305-343, KOREA 2

Hoseo University, 185, Saechoolee, Baebangmyun, Asan, 336-795 Korea

*[email protected] Abstract In this paper, natural gas combustion characteristics of new oxygen carrier particles were investigated in a batch type fluidized bed reactor (0.052 m ID, 0.7 m high). Three particles, OCN703-1100, OCN7051100, and OCN708-1300 were used as oxygen carriers. Natural gas and air were used as reactants for reduction and oxidation, respectively. To check feasibility of good performance, inherent CO2 separation, and low-NOx emissions, CH4, CO, CO2, O2, H2, NO concentrations were measured by on-line gas analyzer. Moreover, the regeneration ability of the oxygen carrier particles was investigated by successive reduction–oxidation cyclic tests up to the 10th cycle. All three oxygen carrier particles showed high gas conversion, high CO2 selectivity, and low CO concentration during reduction and very low NO emission during oxidation. Moreover, all three particles showed good regeneration ability during successive reduction-oxidation cyclic tests up to the 10th cycle. These results indicate that inherent CO2 separation, NOx-free combustion, and long-term operation without reactivity decay of oxygen carrier particles are possible in the natural gas fueled chemical-looping combustion system. However, OCN708-1300 represented temperature and pressure fluctuations during reduction and slightly decay of oxidation reactivity with the number of cycles increased. Introduction Carbon dioxide, a major greenhouse gas, is produced in large quantities from combustion of fossil fuels, much of this related to electric power generation. In a conventional power generation system, fuel and air are directly mixed and burned; therefore it is not easy to separate CO2 from flue gas because CO2 is diluted by N2 in air. Chemical-looping combustion (CLC) is a novel combustion technology with inherent separation of the greenhouse gas CO2 and no NOx emission. The chemical-looping combustor consists of two reactors, an oxidation reactor and a reduction reactor. A fuel and an air go through the different reactors. Equations (1) and (2) explain a basic concept of the chemical-looping combustion system. In the reduction reactor, gaseous fuel (CH4, H2, CO or CnH2n+2) reacts with metal oxide according to the Eq. (2), and release water vapor and carbon dioxide from the top and metal particles (M) from the bottom. The solid products, metal particles, are transported to the oxidation reactor and react with oxygen in the air according to Eq. (1), and produce high-temperature flue gas and metal oxide particles. Metal oxide particles at high temperature are again introduced to the reduction reactor and supply the heat required for the reduction reaction. Between the two reactors, metal (or metal oxide) particles play an important role in transportation of oxygen and heat, therefore the looping material between the two reactors is named as an oxygen carrier particle. Oxidation: exothermic reaction, M + 0.5 O2 = MO

(1)

Reduction: endothermic reaction, CH4 + 4MO = CO2 + 2H2O + 4M

(2)

It is important for the exhaust gas from the reduction reactor to contain only highly concentrated CO2 and water vapor. Therefore, CO2 can be easily recovered by cooling the exhaust gas without any extra energy consumption (energy penalty) for CO2 separation. Another advantage of CLC is that NOx formation can be thoroughly eliminated because the oxidation reaction occurs at a considerably lower temperature (~900oC) without a flame; therefore there is no thermal NOx formation [1, 2]. Moreover, the efficiency of chemical-looping combustion system is very high. Wolf et al. [3] reported that natural gas fueled chemical-looping combustor achieves a thermal efficiency between 52–53% and is 5 percent points more efficient than an NGCC system with state-of-the-art technologies for CO2 capture. The previous results on chemical-looping combustion technology have concentrated on improvement of the oxygen carrier particles and most of the studies used methane or hydrogen as a reduction gas [4]. In order to understand the performance and the feasibility of the chemical-looping combustion system, it is necessary to check the performance of the oxygen carrier particles by using natural gas as a reduction gas because the NGCC, as a practical application of the CLC system, uses natural gas as a fuel. Moreover, confirmation of reduction of NOx emission in the oxidizer and high CO2 selectivity in the reducer are prerequisite for application of the chemical-looping combustion system in commercial plants. There are many reports on the reactivity of oxygen carriers for a chemical-looping combustor, but most of the previous works were performed with a TGA or fixed bed reactor and the data of gas concentrations from oxidizer and reducer were infrequent. Moreover, there is no report on NOx emissions during oxidation at all. In this paper, natural gas combustion characteristics of new oxygen carrier particles were investigated in a batch type fluidized bed reactor (0.052 m ID, 0.7 m high). Three particles, OCN703-1100, OCN7051100, and OCN708-1300 were used as oxygen carriers. Natural gas and air were used as reactants for reduction and oxidation, respectively. To check feasibility of good performance, inherent CO2 separation, and low-NOx emissions, CH4, CO, CO2, O2, H2, NO concentrations were measured by on-line gas analyzer. Moreover, the regeneration ability of the oxygen carrier particles was investigated by successive reduction–oxidation cyclic tests up to the 10th cycle. Experimental Multi-cycle tests were carried out in a bubbling fluidized bed reactor. A schematic of the reactor is shown in Figure 1. The major components consist of a gas input system, the fluidization column, a hot gas filter, a condenser, a gas cooler, and a gas sampling/analyzing unit. The fluidization column is 0.6 m high with an internal diameter of 0.05 m. A perforated gas distributor plate separates the fluidization column and the air box. Reactant gas was fed to the air box. An electric heater could be controlled by a thermocouple and a heater controller. Thermocouple measurements of temperature and pressure transducer data were recorded by a data acquisition system. The exit stream from the fluidized bed reactor was sampled at the outlet of the reactor. The CH4, CO, CO2, H2, NO and O2 concentrations were determined using an on-line gas analyzer and recorded by the data acquisition system. Further details of the reactor system are available elsewhere [5]. The three oxygen carrier particles tested were OCN703-1100, OCN705-1100, and OCN708-1300. Prior to the start of each experiment, the oxygen carrier particle was sieved to ensure that all particles were initially between 106 and 212 mm in size. The static bed height was 0.4 m in all cases, and the experiments were carried out batchwise for the particles, i.e. no particles were added during the run. The fluidized bed reactor operated with a total inlet gas flow of 2.0 Nl/min in all cases, corresponding to superficial gas velocities of 0.07 m/s at 900oC. Particles were oxidized in air as the bed temperature was increased from room temperature to 900oC. Once the particle was fully oxidized, the particle was exposed to a gas mixture. The gas mixture contained 10% natural gas and 90% N2 for all reduction tests. The inlet concentrations were measured by bypassing the reactor. Between the cyclic oxidation and reduction tests, nitrogen was used as a purge gas. A rapid decrease in the exit gas concentration marked the end of purge. For each particle, ten cycles of reduction–oxidation were carried out. The experimental conditions and the composition of natural gas are summarized in Tables 1 and 2, respectively.

Table 1 Summary of reaction conditions OCN703-1100 Particles OCN705-1100 OCN708-1300 Particle size range [mm] 106 – 212 Metal oxide content [%] 70 Temperature [oC] 900 Inert gas (purge) N2 (2 l/min) Reduction gas

NG (0.2 l/min) + N2 (1.8 l/min)

Oxidation gas

Air (2 l/min)

Table 2 Composition of natural gas

Figure 1. Schematic of BFB reactor.

Components CH4 C2H6 C3H8 i-C4H10 n-C4H10 i-C5H12 n-C5H12 N2

Content [vol%] 88.4857 6.8617 2.9631 0.6991 0.7222 0.0337 0.0089 0.2256

Results and discussion The effects of reduction–oxidation cycling on the fuel conversion of three oxygen carrier particles during reduction are shown in Figure 2. The fuel conversion is defined as (moles of reacted hydrocarbons)/(moles of input hydrocarbon). The moles of reacted hydrocarbon are calculated by measured CO and CO2 concentration. The input hydrocarbon concentration was measured before each cycle by bypassing the reactor. The fuel conversions for all three particles were maintained constantly during 10 cycles. The average values of fuel conversion during ten cycles for OCN703-1100, OCN7051100, and OCN708-1300 particles were high, the values being 95.67, 96.06, 96.11 %, respectively. OCN708-1300 particles showed slightly higher fuel conversion than other particles. The effects of reduction–oxidation cycling on the CO2 selectivity of three oxygen carrier particles during reduction are shown in Figure 3. CO2 selectivity indicates the portion of CO2 per carbon in consumed hydrocarbons ([CO2]/[Carbon in consumed hydrocarbon]). The CO2 selectivities for all three particles were maintained constantly during 10 cycles. The average values of fuel conversion during ten cycles for OCN703-1100, OCN705-1100, and OCN708-1300 particles were high, the values being 95.67, 95.29, 94.13 %, respectively. OCN703-1100 particles showed slightly higher CO2 selectivity than other particles. Figure 4 represents the effect of reduction-oxidation cycling on the NO emission of three oxygen carrier particles during oxidation. For all particles, NO concentrations were very low and the average values for OCN703-1100, OCN705-1100, and OCN708-1300 particles were 2.73, 0.11 and 0.07 ppm, respectively. Based on the results, we could conclude that NOx-free combustion is realizable in the natural gas fueled chemical-looping combustion system with OCN703-1100, OCN705-1100, and OCN708-1300 particles.

100

100

80

80

60

60

40

40

Fuel conversion [%]

95.67 96.06 96.11

OCN703-1100 OCN705-1100 OCN708-1300

20

0 1

2

3

4

5

6

7

8

9

Average

2 selectivity [%] CO

Average

10

OCN703-1100 95.67 OCN705-1100 95.29 OCN708-1300 94.13

20

0 1

2

3

4

5

6

7

8

9

10

Number of cycles [-]

Number of cycles [-]

Figure 2 Fuel conversion versus the number of cycles.

Figure 3 CO2 selectivity versus the number of cycles.

100

80

60

40 Average OCN703-1100 OCN705-1100 OCN708-1300

20

NO concentration [ppm]

2.73 0.11 0.07

0 1

2

3

4

5

6

7

8

9

10

Number of cycles [-]

Figure 4 NO concentration versus the number of cycles. Conclusion All three oxygen carrier particles showed high gas conversion, high CO2 selectivity, and low CO concentration during reduction and very low NO emission during oxidation. Moreover, all three particles showed good regeneration ability during successive reduction-oxidation cyclic tests up to the 10th cycle. These results indicate that inherent CO2 separation, NOx-free combustion, and long-term operation without reactivity decay of oxygen carrier particles are possible in the natural gas fueled chemical-looping combustion system. However, OCN708-1300 represented temperature and pressure fluctuations during reduction and slightly decay of oxidation reactivity with the number of cycles increased. Acknowledgement This work was supported by the Power Generation & Electricity Delivery program of the Korea Institute of Energy Technology Evaluation and Planning(KETEP) grant funded by the Korea government Ministry of Knowledge Economy

References 1. Ishida, M. and H. Jin; “CO2 Recovery in a Power Plant with Chemical Looping Combustion,” Energy Convers. Mgmt., 38, S187-S192 (1996). 2. Jin, H., T. Okamoto and M. Ishida; “Development of a Novel Chemical-Looping Combustion: Synthesis of a Looping Material with a Double Metal Oxide of CoO-NiO,” Energy & Fuels, 12, 12721277 (1998). 3. Wolf, J., M. Anheden and J. Yan; “ Comparison of Nickel- and Iron-based Oxygen Carriers in Chemical-Looping Combustion for CO2 Capture in Power Generation,” Fuel, 84, 993-1006 (2005). 4. Ryu, H. J., G. T. Jin, N. Y. Lim and S. Y. Bae; “Reaction Characteristics of Five Kinds of Oxygen Carrier Particles for Chemical-Looping Combustor,” Trans. Korean Hydrogen Energy Society, 14, 2434 (2003b). 5. Ryu, H. J., D. H. Bae and G. T. Jin; "LNG Combustion Characteristics of Oxygen Carrier Particles for Chemical-Looping Combustor", 31st KOSCO Symposium, pp. 141-147, Pusan Korea (2005).

COMBUSTION REACTIVITY OF CHAR DERIVED FROM SOLVENT EXTRACTED COAL Jeonghwan Lima, Hokyung Choia, Sangdo Kima, Youngjoon Rhima, Woosik Parkb, Sihyun Leea*, a

Clean Fossil Energy Research Center, Korea Institute of Energy Research, Republic of Korea b Dept. of Chemical & Biological Engineering, Hanyang University *E-mail: [email protected]

Abstract This study produced char from ash-free coals and investigated its reactivity with air. Ash-free coal was manufactured by using the solvent extraction technique. Three different ranks of coal were used as samples: Australian lignite coal, Indonesian subbituminous coal, and American bituminous coal. 1-MN and NMP were used as solvents for extraction. The FT-IR analysis showed that the 3490 cm-1 absorption band, which usually appears in high high-rank coal, appeared in the 1-MN-extracted coal regardless of the rank of the original coal. Furthermore, the 1011–1095 cm-1 band of the extracted coal decreased greatly due to the reduction of ash. The FT-IR data of the residual coal samples was not much different from that of the original coal. The 1-MN-extracted coals showed the same burning profile regardless of the rank of the original coal, and had a higher range of burning temperatures than the bituminous coal. On the other hand, the 1-MN-extracted coal had the same burning speed as that of the original coal. The NMPextracted coals all showed lower burning temperatures than the original coal, whereas the residual coals showed a similar range of burning temperature to that of the original coal. Key words: coal, ash-free, solvent extraction, 1-MN, NMP 1. Introduction The energy market is currently unstable. Oil and coal prices are fluctuating, and the predominant outlook is that both will increase in the long term. Furthermore, because coal is the primary source of CO2, which is a greenhouse gas, the general opinion is that CO2 emissions must be reduced to solve global environmental problems. To cope with the unstable energy supply, prepare for high oil and coal prices, and minimize CO2 emissions from the use of coal, many researchers are actively studying clean technology for using coal. Among the clean technologies for using coals, ash-free coal manufacturing technology manufactures solid fuels with a similar quality to oils by removing ash or extracting carbon from coal. Ash-free coal is now being developed as Ultra Clean Coal in Australia and as Hypercoal in Japan. Ultra Clean Coal, developed by the White Energy Co., is manufactured by leaching and removing ash from coal using alkalis such as NaOH and KOH [1,2]. Hypercoal, developed by Kobe Steel Ltd., extracts carbon from coal using solvents with a high affinity to coal [3,4]. Hypercoal produces higher quality coal than Ultra Clean Coal because of the high heating value of coal. Ash-free coals can be used as fuels in various applications such as thermal power generation, direct coalfired gas turbines, DCFC (direct carbon fuel cell), and IGFC (integrated gasification fuel cell). Interest in the use of ash-free coal as a means to solve global environmental problems is increasing because this increases power generation efficiency, leading to reduced CO2 emissions. As people’s expectations of

ash-free coal rise, interest in the reactivity of ash-free coal is also increasing. This study produced char from ash-free coal at 900 °C, using the solvent extraction technique, and investigated its reactivity with air. 2. Experimental 2.1. Sample preparation Ash-free coal was manufactured using the solvent extraction technique. The details of the manufacturing process for ash-free coal are described in other papers [5]. Three different ranks of coal were used as samples: Australian lignite coal, Indonesian subbituminous coal, and American bituminous coal. They were extracted at a temperature of 370 °C and a pressure of 10 atm using two solvents, 1-MN (1methylnaphthalene) and NMP (N-Methylpyrrolidone). The extraction time was 30 minutes. Char was produced from the extracted coal under the following conditions. In nitrogen atmosphere, a proximate analyzer was used to raise temperature from normal to 900 °C at a rate of 10 °C/min. Once they reached 900 °C, the samples were maintained at this temperature for 30 minutes before being cooled to normal temperature. 2.2. Analysis For the proximate analysis of the coal samples, the TGA-701 Thermogravimeter from LECO was used. The moisture, ash, volatile matter, and fixed carbon were measured for PC, EC, and RC in accordance with the ASTM standard. For ultimate analysis, the CHN-2000 Elemental Analyzer from LECO was used to measure carbon (C), hydrogen (H), and nitrogen (N). For analysis of sulfur (S), the SC-432DR Sulfur Analyzer from LECO was used. Among the ultimate analysis items, the oxygen (O) content was determined by subtracting the weights of carbon, hydrogen, nitrogen, sulfur, and ash contents from the total weight. For heating value analysis, the Parr 1261 Calorimeter from PARR was used. For FT-IR analysis, the Nicolet 6700 FT-IR spectrometer from Thermo Fisher Scientific was used. For DTG analysis, the SDT600 thermogravimetric analyzer from TA Instrument was used.

3. Results and discussion 3.1. Proximate analysis, ultimate analysis, and heating value Table 1 shows the results of proximate analysis, ultimate analysis, sulfur analysis, and heating value analysis of the original coals. Lignite coal is a low rank coal that has a low content of fixed carbons and a high content of volatile matters because it has high water content and low carbonization degree. Thus, it exhibits a low heating value. Bituminous coal is a high rank coal that has a high content of fixed carbon and a high heating value because it has low water content and high carbonization degree. Subbituminous coal is ranked in between lignite coal and bituminous coal, being accordingly characterized. Table 1 Proximate, ultimate analysis and high heating value of original coal samples. Proximate (wt%) Samples: Original coals Moisture Volatile matter

Ash

Fixed C carbon

H

N

O

S

Lignite 12.11 Subbituminous 6.68 Bituminous 2.88

10.06 3.6 12.32

28.19 36.8 52.19

4.65 5.36 5.08

0.69 0.65 1.55

28.31 21.73 6.25

0.11 0.06 0.53

49,64 52.92 32.61

Ultimate (wt%, dry and ash free)

55.98 68.76 73.5

High heating value (kcal/kg) 5,210 5,540 6,530

The anallysis results ffor the extraccted coal (EC C) and residuual coal (RC) by solvent aand coal typee are shown in Table 2. Residual coal refers to the coal remaining r in the solvent as solid thatt was not exttracted. As shown inn Table 2, thhe moisture content in tthe extractedd coal and reesidual coal was lower than t in the original coal. The ashh content in tthe extractedd coals was m much lower thhan in the oriiginal coal: itt was 0.05– 0.12 % w when extractted by 1-MN N and 0.76–0..16 % when extracted byy NMP. Thuss, the ash conntent in the NMP-exxtracted coal was higher than t that in thhe 1-NM-exttracted coal. The ash conttent in the reesidual coal was equaal to or a littlle higher thann that in the ooriginal coal.. Table 2 P Proximate, uultimate analyysis and highh heating valuue of extractiion products.. Samples: Extractionn products Liignite

Proximate (w wt%) Moisture

EC 1.49

U Ultimate (wt%,, dry and ash ffree)

Voolatile maatter

Ash

Fixed carbon

C

H

N

O

S

32.71

0.05

65.74

822.55

6.22

0.63

100.06

0.04

High heating Y Yield value (% %) (kcal/kg) 8,340 31

45.2

13.08 36.65

655.85

4.21

1.1

144.63

0.02

6,430

-

Suubbitum EC 4.43 1-MN innous RC 5.16

61.33

0.12

34.12

866.15

6.24

1.04

6.551

0.04

8,400

30

48.3

4.36

42.18

800

4.56

1.53

9.663

0.03

7,130

-

Biitumino EC 0.54 uss RC 1.58

42.71

0.06

56.69

855.75

6.68

2.15

4.997

0.25

8,460

34

37.9

12.24 48.28

71

4.33

2.08

100.09

0.26

6,890

-

EC 1.82

56.96

1.16

755.1

7.4

6.79

9.552

0.02

7,710

64

RC 2.16

43.88

24.28 29.68

600.45

4.45

4.04

6.443

0.35

5,920

-

Suubbitum EC 1.56 NMP innous RC 1.98

54.02

0.78

43.65

81.25

6.2

5.97

5.226

0.07

7,870

72

41.99

8.53

47.51

722.25

4.73

3.78

100.64

0.07

6,590

-

Biitumino EC 1.19 uss RC 1.46

50.59

0.76

47.47

788.3

6..74

6.71

7.224

0.22

7,860

68

32.78

23.26 42.4

622.3

3.82

2.76

5.112

0.28

6,090

-

Liignite

RC 5.08

40.05

The heatting value off the extracteed coal increeased greatly, but that of the residual coal was eqqual to or a little higher than thatt of the originnal coal. Thee heating valuue of the resiidual coal increased desppite the fact that the aash content w was equal to or higher thaan that in the original coaal because waater was remooved in the extractioon process. Fuurthermore, tthe oxygen aand sulfur conntents in the brown coal aand subbitum minous coal decreaseed. The meaan extraction yield was arround 30 % ffor 1-MN exttraction and aaround 65 % for NMP exxtraction. 1MN, whhich is a non--polar solvennt, only extraacts coal com mponents thaat are thermallly relaxed; in i contrast, NMP, w which is a polar solvent, uses u both theermal relaxaation and relaaxation by soolvent [6,7]. Therefore, the yieldd of the NM MP extractioon is higher than thatt of the 1-MN N. 3.2. FT-IIR analysis The FT--IR spectra ffor the originnal coal, 1MN exxtraction pproducts, aand NMP extractioon products aare comparedd in Fig. 1, Fig. 2 aand Fig. 3, rrespectively. FT-IR has been useed often for thhe analysis oof functional groups of o coal such as C-H, C-O O, and O-H. The smaall absorptioon band of 3050–3030

Fiig. 1 FT-IR sspectra for thhe original cooals

cm-1 indicates an aroomatic CH sttretch band, which show ws a cross-linnked structuree. Because oof this, it is generallyy reported thhat it does noot appear in thhe low-rank coals, and only appears in high-rank coals. The two absoorption bandss of 2920 cm m-1 and 2850 cm-1 represeent the aliphaatic CH3, CH H2, and CH ggroups. The absorptioon bands thaat appear aroound 1600 cm m-1 indicate aromatic a ringg vibration ddue to the binding with oxygen iin quinone, ether, and hheterocyclic ccompounds. The 1500 cm-1 absorptioon band represents the C=C binnding in the benzene b ringg. The absorpption band of 1450–14400 cm-1 repressents the CH H2 group, as -1 does thee 1380–1375 cm band. The 1370 cm m-1 and 14660 cm-1 absoorption bandss represent m methyl and methylenne, respectivvely. The 10000–1300 cm m-1 section caan be generaally divided iinto the alipphatic ether group att 1000–12000 cm-1 and thhe aromatic ether groupp in 1200–13300 cm-1. Thhe 900–700 cm-1 band represents the C-H ggroup combiined with thee aromatic compounds. c The sectionss 500–600 cm m-1, 1000– -1 -1 m , and 3600––3700 cm represent the bands of the minerals conntained in cooal [8,9]. 1100 cm As show wn in Fig. 2, 2 the absorpption band ccan be seen at 3490 cm m-1 for 1-MN N-extracted coals; this correspoonds to the aaromatic CH H stretch thaat appears inn high-rank coal. Another differencee from the original coals is thatt it is possiblle to clearly observe the CH2 or metthyl/methylenne group, aroomatic and d aliphaticc ether groupps, and the disappearanc e of the 10111–1095 cm--1 band due to the decrease in ash. Howeverr, the FT-IR of the residuual coal did not n show mucch difference from that off the original coal. In the caase of NMP extraction, aas shown in Fig. 3, the aabsorption baands of the extracted e coaal does not show muuch differencce from thosee of the originnal coal, exceept for the reemoval of ashh.

(a) 1-M MN extracted coals

(bb) 1-MN residdue coals

R spectra for tthe coals prooduced by 1-M MN extractioon. Fig. 2 FT-IR

(a) NM MP extracted ccoals

(bb) NMP residdue coals

Fig. 3 FT-IR R spectra for the coals prooduced by NM MP extractioon.

3.3. DTG G analysis Fig. 4 coompares the DTG burninng profiles off coal char saamples: the cchar of originnal coal, extrracted coal and residdual coal froom 1-MN annd NMP exttraction. As shown in Fiig. 4, the oriiginal coal sshowed the typical bburning charaacteristics thaat the higherr the rank, the higher the burning tem mperature. In the case of brown cooal, two peakks appeared because of thhe low fixedd carbon conttent and the hhigh content of volatile matters. That is, the volatile mattter burned firrst, and the ffixed carbons burned nexxt. Two peakks appeared because the contentss of volatilee matters andd fixed carbbons are sim milar. This phhenomenon sometimes appears in bituminouus coals as w well, in whichh case the buurning of the volatile mattters before tthe burning of the fixxed carbons appears as a small shoulder. In termss of DTG peaak height, orr the maximuum burning rate, the reactivity off the subbitum minous coal was the highhest, and the bituminous coal had a bbroad range of burninng temperatuures. The 1-M MN-extractedd coal show wed peculiar characteristiics. The 1-M MN-extractedd coals hadd the same burning profile, regarrdless of the rank of the ooriginal coal,, and had a w wider range oof burning tem mperatures G peak, the burning speeed of brownn coal was thhe highest, than bituuminous coaals. In terms of the DTG followedd by subbitum minous coal, and finally bituminous b cooal. On the other o hand, thhe 1-MN-exttracted coal had the ssame burningg speed as thaat of the origginal coal.

((a) Original coals c

(b) 1-M MN extractedd coals

(cc) 1-MN residue coals

(d) NM MP extractedd coals

(ee) NMP residdue coals

Fiig. 4 DTG buurning profilees of coal sam mples.

As show wn in the FT-IIR results abbove, 3490 cm m-1 was obseerved in the 11-MN-extractted coal, whiich appears in high-rrank coal; thhus, this indiccated an incrreased rank. The results of the burninng profile also indicate the samee trend becauuse all the burrning temperratures increaased. Unlike thhe 1-MN-exttracted coal, the NMP-exttracted coals all showed llower burninng temperaturres than the original coal, whereaas the residuual coals show wed a similaar range of bburning temperatures as tthe original milar burning temperatures to that of tthe original coal. Thuus, we can cconclude thatt residual coaals show sim coal regaardless of sollvent, but thee extracted cooals show low wer burning ttemperaturess than the original coal. Fig. 5 shhows the connversion rates of each coaal by temperrature. The reesidual coal conversion rrates of the coals exttracted by eaach solvent aare similar too that of the original coall. The subbittuminous coaals showed similar cconversion raates by tempperature, although the tem mperature vaaried slightly. In compariison, the 1MN-extrracted coals sshowed simillar conversioon rates, regaardless of rannk. In the casee of the NMP P-extracted coals, thhe subbituminnous coal shhowed an infflection pointt at conversiion 0.6, and the bituminoous coal at conversion 0.8, whicch seems to bbe due to chaanges in the burning b reacttion mechaniisms. This phhenomenon can be allso seen in thhe DTG burnning profile inn Fig. 4.

(aa) Original ccoals

(b) 1-M MN extractedd coals

(cc) 1-MN residdue coals

(d) NM MP extracted coals

(ee) NMP residdue coals

o coal samplles. Fig. 5 Conveersion rates of

4. Conclusions Ash-free coal was manufactured from coal using the solvent extraction technique. Char was produced from the extracted ash-free coal at 900°C in an inert atmosphere and the reactivity between the char and air was examined. The FT-IR analysis showed that the 3490 cm-1 absorption band, which usually appears in high-rank coal, appeared in the 1-MN-extracted coal regardless of the rank of the original coal. Furthermore, the 1011– 1095 cm-1 band of the extracted coal decreased greatly due to the reduction of ash. The FT-IR spectra of the residual coal samples were not much different from those of the original coals. The 1-MN-extracted coals showed the same burning profile regardless of the grade of the original coal, and had a higher range of burning temperatures than the bituminous coal. On the other hand, the 1-MNextracted coal had the same burning speed as that of the original coal. The NMP-extracted coals all showed lower burning temperatures than the original coal, whereas the residual coals showed a similar range of burning temperature to that of the original coal. References 1. Steel KM, Patrick JW, The production of ultra clean coal by chemical demineralization. Fuel 2001;80:2019-2023 2. http://www.det.csiro.au/science/lee_cc/ultra_clean_coal.htm 3. Okumura N, Komatsu N, Shigehisa T, Kaneko T, Tsuruya S, Hyper-coal process to produce the ashfree coal. Fuel Process Technol 2004;86:61-72 4. Takanohashi T, Shishido T, Kawashima H, Saito I, Characterization of hypercoal from coals of various ranks. Fuel 2008;87:592-598 5. Kim SD, Woo KJ, Jeong SK, Rhim YJ, Lee SH, Production of low ash coal from LRC (low rank coals) by thermal extraction with NMP (N-methyl-2-pyrrolidinone). KJChE 2008;25(4): 758-763 6. Li C, Takanohashi T, Saito I, H Aoki, K Mashino, Elucidation of mechanisms involved in acid pretreatment and thermal exstraction during ashless coal production. Energy Fuels 2004;18:97-101 7. Kashimura N, Takanohashi T, Saito I, Effect of noncovalent bonds on the thermal extraction of subbituminous coals. Energy Fuels 2006;20:1605-1608 8. Gupta R, Advanced coal characterization: a review. Energy Fuels 2007;21:451-460 9. Xuguang S, The investigation of chemical structure of coal macerals via transmitted-light FT-IR microspectroscopy. Spectrochimica Acta Part A 2005;62:557-564

th

27 International Pittsburgh Coal Conference October 11–14, 2010, Istanbul, Turkey

FORCED FLAME RESPONSE MEASUREMENT IN A GAS TURBINE COMBUSTOR WITH HIGH HYDROGEN FUEL K.T. Kim*, J.G. Lee, B.D. Quay, D.A. Santavicca Center for Advanced Power Generation Department of Mechanical and Nuclear Engineering The Pennsylvania State University, University Park, PA, 16802

ABSTRACT The forced response of swirl-stabilized lean-premixed turbulent flames to acoustic oscillations in a hydrogen enriched laboratory-scale gas turbine combustor was experimentally investigated. Nonlinear flame transfer function measurements were taken to investigate the flame’s heat release response to upstream acoustic perturbations. This analysis shows that the dynamics of natural gas-air premixed flames are characterized by several regimes: the linear, transition, and first and second nonlinear regimes, depending upon steady-state flame geometry, modulation frequency, and amplitude of excitation. The present results show that the flame geometry changes from a dihedral V flame to an enveloped M flame with an increase in hydrogen mole fraction, and the changes in steady-state flame structures have a significant impact on the flame’s response to acoustic modulations. The present results suggest that the M flame, unlike the V flame, has the unique dynamic characteristic of acting as a damper of flow perturbations. The response of the M flame remains in the linear regime, irrespective of the shedding of a vortex-ring structure, because the interaction between the large-scale structure and the flame is not strongly coupled. INTRODUCTION Combustion instabilities are characterized by large-amplitude pressure oscillations that are driven by unsteady heat release. It is by nature a self-excited oscillation, involving complicated physical phenomena such as unsteady combustion, acoustic fluctuations, heat transfer, and the vorticity field. These highamplitude pressure oscillations can substantially reduce the performance of a system and in extreme cases can cause structural damage to engines. Over the last two decades, a large number of studies using experimental [1-4], theoretical [5-8], and CFD simulation [9-11] approaches have been performed to investigate instability mechanisms, controlling parameters, and active or passive control methodologies to suppress the instability intensity level to acceptable ranges. Recently, Huang and Yang [12] have provided comprehensive reviews on this issue. Ultimately, to accurately predict instability characteristics at the development stage, a complete understanding of the nature of combustion instabilities will be required and sophisticated theoretical models must be developed to

incorporate the complex flame dynamics into the prediction model. Development of a theoretical model capable of predicting the formation of thermoacoustic pulsations is contingent on a physical understanding of the linear and nonlinear responses of a flame to periodic disturbances. When there are no equivalence ratio oscillations in the mixing plenum, the response of a premixed flame is primarily governed by acoustic velocity fluctuations. To quantitatively describe the response of a flame to acoustic forcing, the nonlinear flame transfer function (FTF) is introduced, defined as the normalized ratio of heat release and velocity fluctuations:

FTF ( f , A) =

Q '( f ) Q V '( f ) V

(1)

where Q is the time-averaged heat release rate, V is the mean velocity of the mixture in the mixing section, Q '( f ) and

V '( f ) are their corresponding amplitudes at the forcing frequency, f, and A is the magnitude of V '( f ) / V . The flame transfer function can be obtained by experimental [2-4, 13-14], theoretical [6-8], or numerical [5] methods. The flame transfer function provides an insight into the response of a flame to inlet disturbances, and it can be mathematically formulated and used in the thermoacoustic network modeling as a source term to predict self-induced combustion instability [14]. In the present article, the effects of fuel composition on the forced response of a swirled lean-premixed flame are experimentally investigated. Development of fuel-flexible gas turbine engines capable of operating with highly variable and potentially low quality fuels, such as coal-derived syngas fuels, while producing minimal air pollutants, is a key issue in the gas turbine combustion community [15-16]. Most previous studies on the effects of fuel composition on applications in gas turbines have dealt with static stabilities, pollutant emissions [17-20], flame propagation speed [21-23], and steady-state flame structures [24-25]. It was found that the addition of hydrogen to hydrocarbon fuels results in a significant increase in OH radical concentration, extending the lean stability limits and increasing flame propagation speeds. However, there have been relatively few experimental investigations on the influence

*

Corresponding author: University of Cambridge, Engineering Department, Trumpington Street, Cambridge, CB2 1PZ, UK. E-mail address: [email protected].

1

ICCD camera CH* inteferance filter, 432 nm

choked inlet for self-excited instability measurement

PT5 PT3 PT4

PT1 PT2

siren

forced air+fuel swirler

38.1

T2

T3

plug

centerbody

19.1

T1

1500

exhaust

762 ~ 1524

76.2

T4

109.2

PTc

333.5

334.8

mixing section

quartz combustor

steel combustor

CH*(432nm), OH*(307nm), CO2*(365nm) filters PMTs

Figure 1. Schematic of a swirl-stabilized, lean-premixed, model gas turbine combustor. Dimensions in millimeters.

of fuel composition (particularly, hydrogen-enrichment) on combustion dynamics [26-27]. It is not possible to make general statements about the impact of fuel composition variation upon a combustor’s propensity to become unstable– the effects can be either stabilizing or destabilizing. In the present paper, we focus on the linear/nonlinear dynamics of H2-enriched, swirl-stabilized, lean-premixed flames using a phase-resolved analysis and nonlinear flame transfer function measurements. Two key mechanisms of nonlinearity, i.e., shear layer rollup and unsteady flame liftoff for dihedral V flames, are examined. In particular, the dynamics of H2enriched enveloped M-shaped flames are compared with pure natural gas flames with inverted dihedral V flame geometry. It is shown that modification of steady-state flame structures significantly affects the linear and nonlinear dynamics of a laminar premixed flame [28]. The present results suggest that the M flames do not exhibit the strong overshoot behavior which is observed in the case of dihedral V flames. This study extends previous experimental and analytical results from laminar to turbulent premixed flames. EXPERIMENTAL METHODS Lean Premixed Gas Turbine Combustor A schematic of the experimental setup is shown in Figure 1. This facility consists of an air inlet section, a siren, a mixing section, an optically-accessible quartz combustor section, a steel combustor section, and an exhaust section. The air can be heated to a maximum temperature of 400 °C by a 30 kW electric heater. A siren-type modulation device (rotor + stator) is used to provide acoustic modulations. The siren is driven by a

variable-speed DC motor, providing capabilities for changing forcing frequency (~ 500 Hz). Two high-temperature globe valves are used to control the bypass flow rate, i.e., the inlet velocity fluctuation amplitudes. The mixing section is 0.333 m long and has an annular cross-section that is defined by a 19.1 mm O.D. centerbody and a 38.1 mm I.D. mixing tube. The centerbody is centered in the mixing tube, and it is positioned such that its downstream end is flush with the combustor dump plane. A 30o flat-vane axial swirler is mounted in the mixing tube 76.2 mm upstream of the combustor dump plane. At the entrance to the mixing section the flow is choked. This provides a well-defined acoustic boundary condition for self-excited instability measurements. For forced flame response measurements, the choking plate is removed and mounted upstream of the siren. Fuel (natural gas + H2) is injected and mixed upstream of the choked inlet in order to ensure that the reactant mixtures are spatially and temporally homogeneous before they enter the reaction zone. The combustor consists of a stainless steel dump plane, to which an optically accessible fused-silica combustor with a diameter of 109.2 mm and length of 334.8 mm is attached. The downstream end of the quartz combustor is connected to a stainless steel variable-length combustor section. The length of the combustor can be continuously varied between 762 mm and 1524 mm by moving a water-cooled plug along the length of the steel combustor section. For forced response measurements, the combustor length was maintained at the minimum to minimize the influence of system acoustics on upstream acoustic excitation.

2

3

360 V'/Vmean=9%(Linear)

(B)

V'/V mean=15%(Transition)

270

V'/V mean=25%(Nonlinear I)

Gain

Phase (degree)

2

1

180

90 V'/Vmean=59%(Nonlinear II)

(A)

0

0

0.2

0.4

0.6

0

0.8

0

0.2

0.4

0.6

0.8

V'/V mean

V'/V mean

Figure 2. (A) The gain and (B) phase of flame transfer function as a function of the magnitude of inlet velocity perturbation at a modulation frequency of 200 Hz. Inlet conditions: Tin = 200 oC, Vmean = 60 m/s, φ = 0.60, and X H 2 = 0.00. Instrumentation and Test Conditions High frequency-response, water-cooled, piezoelectric pressure transducers (PCB 112A04) were used to measure pressure perturbations in the mixing (denoted by PT1−PT4 in Figure 1) and the combustor sections (PTc and PT5). The pressure signals were conditioned by amplifiers, digitized by an analog-todigital converter, and stored in microcomputer memory for processing at a sampling rate of 8192 Hz. A total of 16,384 data points were taken during each test, resulting in a frequency resolution of 0.5 Hz and a time resolution of 0.122 msec. Spectral analysis of the signals was performed using the fast Fourier transform (FFT) technique. Two pressure transducers located at 12.7 mm (PT4) and 50.8 mm (PT3) upstream of the combustor dump plane were used to estimate the inlet velocity fluctuations using the two-microphone method. To calibrate the two-microphone method, direct measurements of velocity fluctuations were performed under cold flow conditions with a hot wire anemometer (TSI 1210-20). A photomultiplier tube (PMT, Hamamatsu model H7732-10) was used to measure the global CH* (432 ± 5 nm) chemiluminescence emission intensities from a whole flame. An ICCD camera (Princeton Instruments model 576G) with a CH* band pass filter centered at 430 nm (10 nm FWHM) was used to record the flame images. For phase-averaged imaging, the ICCD camera was synchronized with the combustor pressure signal and fifty averaged images were taken over a cycle of oscillation with a phase interval of 30°. The phase angle φ = 0° corresponds to the positive-to-negative zero transition, and φ = 270° to the maximum combustor pressure. Because CH* chemiluminescence images are line-of-sight integrated images,

a three-point Abel deconvolution scheme was used to extract two-dimensional information from the line-of-sight images. All tests were performed at a mean pressure of 1 atm and at mean equivalence ratio of 0.60. Mean velocity at the nozzle was 60 m/s and the inlet temperature was kept constant at 200 °C, giving a Reynolds number of approximately 33,000. Forcing frequencies were varied from 100 to 400 Hz ( ∆f = 25 Hz). The frequency range of 100−400 Hz includes two distinct frequencies at limit cycle oscillations observed in the lean premixed gas turbine combustor [27]. H2-blended (based on volume) natural gas fuels with X H 2 = 0.00 and 0.30 were used to examine the impacts of fuel composition variation upon the response of the flame to acoustic excitation. RESULTS AND DISCUSSION Linear and Nonlinear Dynamics of Inverted Dihedral V Flames Experiments were performed wherein the heat release response of hydrogen-enriched natural gas-air premixed flames to acoustic forcing was determined for various levels of inlet velocity oscillations. Under the conditions investigated here, the highest perturbation levels are not determined by flame blowoff, but by limitation of the modulating device. The magnitude of upstream forcing which can be achieved by the device is contingent on modulation frequency. It was observed that the maximum amplitude can be attained at a modulation frequency of near 200 Hz, inasmuch as the combustor mixing section with upstream plenum acts as a resonator with peak responses around that frequency. The frequency, f ≈ 200 Hz, also corresponds to one of the self-excited instability frequencies observed in the test rig [27]. In the present study, the flame

3

Figure 3. Phase-synchronized deconvoluted CH* chemiluminescence images at a modulation frequency of 200 Hz for (A) V ' Vmean = 9%(top) and (B) V ' Vmean = 15%(bottom). The intensity is displayed in linear pseudo color scale that white denotes the highest intensity and black means the lowest intensity. Inlet conditions: Tin = 200 oC, Vmean = 60 m/s, φ = 0.60, and X H 2 = 0.00.

transfer function measurements were performed at a modulation frequency of 200 Hz for different fuel compositions. It is believed that the response of the flame at f = 200 Hz will help to illuminate the flame dynamics under limit-cycle oscillations at a frequency near 200 Hz. Figure 2 presents the amplitude dependence of the gain and phase of the flame transfer function (see Eq. 1) at a modulation frequency of 200 Hz. These measurements were made at an inlet temperature of 200 oC, mean nozzle velocity of 60 m/s, mean equivalence ratio of 0.60, and hydrogen mole fraction of 0.00. The global CH* chemiluminescence intensities were used as an indicator of heat release rate oscillations, based on experimental observation that CH* chemiluminescence intensity increases linearly with fuel flow rate at a given overall equivalence ratio. The coherence function between inlet velocity and CH* chemiluminescence intensity was measured to close to unity at the forcing frequency of 200 Hz, enabling flame transfer functions to be accurately determined. Also, a high velocity perturbation magnitude of up to approximately 60% can be achieved at this forcing frequency. At these inlet conditions, the flame is stabilized in the inner shear layer and it is attached to the centerbody, exhibiting an inverted dihedral V structure [14]. The responses of the dihedral V flame to acoustic perturbations are divided into several regimes, as shown in Figure 2. In the linear regime ( V ' Vmean < 10%), the normalized heat release response increases linearly with the amplitude of excitation; thus, the gain is almost constant. When the magnitude of inlet velocity fluctuation is greater than 10%, the gain decreases with an increase in perturbation amplitude, representing typical nonlinear behavior. It is noteworthy that when the modulation amplitude is greater than 40% of the mean value, the normalized CH* chemiluminescence intensity levels

off again. Similar characteristics of acoustically forced swirlstabilized flames were also observed by Thumuluru and Lieuwen [29]. They state that the flame’s heat release response exhibits multiple saturating behaviors and a non-monotonic dependence upon amplitude. Phase-resolved flame imaging measurements in each regime will elucidate the amplitudedependent dynamic characteristics of acoustically forced swirl flames. Figure 2 (B) shows that the phase measured between the CH* chemiluminescence signal and the inlet velocity fluctuation is essentially independent of the modulation amplitude up to V ' Vmean = 0.60. This is consistent with previous studies on laminar premixed flames [28], and indicates that the frequency-dependent convection times for a velocity perturbation to reach the flame are not a function of the modulation amplitude, irrespective of linear and nonlinear regimes. To identify the key physical processes associated with the linear/nonlinear response of the dihedral V flame to flow forcing, phase-averaged chemiluminescence emission from the whole flame was recorded using an intensified CCD camera. Figure 3 presents a sequence of phase-synchronized CH* chemiluminescence images at a modulation frequency of 200 Hz during a period of oscillation for excitation amplitudes of (A) V ' Vmean = 9% and (B) 15%. They represent the response of the flame in the linear and transition regimes, respectively. Note that only the upper half of these deconvoluted images is shown, because the reconstructed images are axisymmetric. The direction of flow is from left to right. It can be observed from Figure 3 (A) that the flame moves back and forth due to the velocity oscillation in the annular jet region, and the reaction zone is located in the inner shear layer. Reaction occurs in the corner recirculation zone (CRZ) at φ = 90 ~ 180° when the

4

Figure 4. (A) Phase-synchronized deconvoluted CH* chemiluminescence images at modulation frequency of 200 Hz and V ' Vmean = 25% (top). (B) Line-of-sight integrated, background-corrected CH* chemiluminescence images at φ = 60, 150, 240, and 330° (bottom). Inlet conditions: Tin = 200 oC, Vmean = 60 m/s, φ = 0.60, and X H 2 = 0.00.

flame moves upstream. Evolution of flame surface area by shear layer rollup is not seen because the minimum excitation level for the formation of a large-scale coherent structure is not achieved. Figure 3 (B) shows the forced flame response in the transition regime. The flame shapes at each phase are similar to those of the linear flame behavior, displayed in Figure 3 (A). The formation and reaction of large-scale, coherent vortex structures are, however, observed at φ = 180 ~ 270°. They are convected by the mean flow, and the heat release reaches its maximum at φ = 270°. The vortices periodically entrain a large amount of combustible material into the reaction zone at the modulation frequency, and therefore the temporal variations of flame surface area are determined by the interplay between the coherent structure and the flame. Deconvoluted and line-of-sight integrated phase-averaged CH* chemiluminescence images in the first nonlinear regime are shown in Figures 4 (A) and (B), respectively. All inlet conditions are the same as for Figure 3, except for the modulation amplitude. The most distinct difference in flame behavior between the transition and the first nonlinear regimes is that in the first nonlinear regime, the flame front bends toward the inner recirculation zone at φ = 270°, as compared to the deconvoluted flame image at φ = 270° in Figure 3 (B). The interaction between a vortex-ring structure and the flame is significant at φ = 180° ~ 300°. The line-of-sight integrated images at φ = 150° and φ = 240° clearly show the vortex rollup and convection processes, respectively. It has been reported that the nonlinear response of premixed flames is related to the shear layer rollup [2, 4, 14, 29]. The shear layer rollup shortens the flame length, which in turn decreases the flame area in a nonlinear manner. Hence, the gain of the flame transfer function decreases in the nonlinear regime, as shown in Figure 2. Figure 5 shows phase-resolved flame images at a modulation amplitude of 59%, corresponding to the second nonlinear regime. Note that the normalized CH* chemiluminescence intensity is constant with respect to the

amplitude of excitation when V ' Vmean is greater than 40%. It is expected that the nonlinear dynamics of the flame is influenced by different controlling physics than in the first nonlinear response regime. The deconvoluted and line-of-sight integrated images at φ = 330° clearly show the unsteady flame liftoff from the attachment point. The high level of inlet velocity perturbation causes the flame attachment point to move off of the centerbody to the downstream inner recirculation zone. This indicates that when the inlet velocity perturbation magnitude is greater than 40% of the mean value, the dynamics of swirlstabilized premixed flames subjected to upstream acoustic forcing are less influenced by the flame-vortex interaction. The unsteady flame liftoff (flame holding) plays a dominant role in the flame dynamics, due to the high perturbation amplitude. At V ' Vmean > 40%, a vortex ring structure is shed (see φ = 150° in Figure 5), and it is convected with the mean flow in the annular jet region, but the main reaction region is located in the inner recirculation zone. Therefore, the response levels off when the magnitude of acoustic forcing is greater than 40%, as shown in Figure 2. To quantitatively describe the spatial distribution of the flame’s heat release depending on excitation amplitudes, coordinates of maximum CH* chemiluminescence intensity locations at φ = 270° for perturbation levels of V ' Vmean = 9, 15, 25, and 59% are plotted in Figure 6. It clearly shows that with increasing modulation amplitude, the intense reaction region moves toward the inner recirculation zone. In consequence, the key mechanism of nonlinearity is modified from shear layer rollup to unsteady flame liftoff, depending on perturbation amplitude. The amplitude-dependent dynamics of inverted V flames suggest that limit-cycle pressure oscillation magnitudes are also governed by different mechanisms of flame area modulation under self-sustained oscillations, depending on the magnitude of inlet velocity fluctuation. An experimental investigation of the dynamics of self excited flames and a

5

Figure 5. (A) Phase-synchronized deconvoluted CH* chemiluminescence images at modulation frequency of 200 Hz and V ' Vmean = 59% (top). (B) Line-of-sight integrated, background-corrected CH* chemiluminescence images at φ = 60, 150, 240, and 330° (bottom). Inlet conditions: Tin = 200 oC, Vmean = 60 m/s, φ = 0.60, and X H 2 = 0.00.

60

50

Y (mm)

40

30

20 V'/V mean = V'/V mean = V'/V mean = V'/V mean =

10

0

0

40

80

120

9% 15% 25% 59%

160

X (mm)

Figure 6. Coordinates of maximum CH* chemiluminescence intensity location at phase of φ = 270° (maximum CH* intensity) for inlet velocity magnitudes of V ' Vmean = 9, 15, 25, and 59%.

comparison between acoustically forced flames and flames under limit cycle oscillations are left for future studies. Forced Response of H2-Enriched Enveloped M Flames In previous work [27], the authors have provided details on the modification of stable flame configurations, depending upon H2 volume fraction in composite fuels. It has been shown that the flame geometry changes from an inverted dihedral V structure to an enveloped M geometry with increasing H2 mole fraction. The M flames have unique characteristics that allow them to damp flow perturbations, as compared to V flames. Thus, the

flame transfer function gain of enveloped M flames is much smaller than that of V flames for a given forcing frequency and amplitude. This is qualitatively consistent with theoretical and experimental studies on the response of laminar premixed flames by Schuller et al. [28, 30-31]. They described V flame behavior as an amplifier of flow perturbation in a certain range of frequencies, with the gain of laminar V flames being greater than that of conical and M-shaped flames. In this section, we will discuss the manner in which the combustion instability characteristics of a given gas turbine combustor are impacted by hydrogen levels in the fuel. The amplitude dependence of the normalized CH* chemiluminescence emission intensity at a modulation frequency of 200 Hz for a H2-enriched M flame with an enveloped geometry is shown in Figure 7. All inlet conditions are the same as for Figure 2, except for H2 mole fraction. Hydrogen blended natural gas fuel (30% H2 + 70% natural gas) was used and the overall equivalence ratio was kept constant at φoverall = 0.60. It can be observed from Figure 7 that the normalized CH* chemiluminescence intensity increases linearly with the amplitude of excitation up to approximately V ' Vmean = 50%, indicating that the flame behaves linearly with respect to flow perturbations. The fact that the response of the flame remains in the linear regime even at a high perturbation amplitude of 50% is an unexpected result. As previously shown in Figures 4 and 5, excitation of a large scale coherent structure plays an important role in inducing nonlinearity when the normalized acoustic velocity amplitude is greater than 25% at a frequency of 200 Hz. Considering that the formation and convection of the structure are governed by fluid dynamic conditions, not by fuel composition, it can be supposed that the H2-enriched M flame may feature significantly different response characteristics in comparison with the dihedral V flame. In order to elucidate the underlying physics, twelve successive phase-resolved CH* chemiluminescence images

6

1 X H2 = 0.00 X H2 = 0.30

CH*'/CH*mean

0.8

0.6

0.4

0.2

0

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

V'/V mean

Figure 7. The normalized heat release response as a function of modulation amplitude at a modulation frequency of 200 Hz. Inlet conditions: Tin = 200 oC, Vmean = 60 m/s, φ = 0.60, and X H 2 = 0.00, 0.30.

taken during a single cycle at a modulation frequency of 200 Hz and amplitude of V ' Vmean = 25% are illustrated in Figure 8. Line-of-sight integrated, background corrected images at the phases 90, 150, 210, and 270° are also presented for comparison. The figure shows that with H2-enrichment ( X H 2 = 0.30), the spatial distribution of the reaction zone becomes very compact in both axial and radial directions. It is obvious from Figure 8 (A) that the flame angle variation is substantial during a modulation cycle. In the case of V flames, however, the flame oscillates in the flow direction, which means that the overall flame length changes over a period of perturbations, as shown

in Figures 3 and 4. This result demonstrates that the dynamics of M flames is significantly different from that of V flames. The flame angle and CH* chemiluminescence intensity variations over one period of oscillation are shown in Figure 9. It can be seen that the variations in flame angle and CH* chemiluminescence intensity are out of phase. As the flame moves back (largest flame angle, see the phase 180°), the flame brush is almost vertical with respect to the flow direction and the CH* chemiluminescence intensity is low due to decreased flame area. These oscillations in the flame position (angle) result from fluctuations in the outer shear layer. It has been reported that if the shear layers in the swirl-stabilized gas turbine combustor are acoustically excited to form a coherent vortex ring structure, the inner shear layer rolls up inward and the outer shear layer rolls up outward [2, 32]. Note that even though periodic formation and convection of vortex structures occurs at φ = 210−270°, the flame response remains in the linear regime, as shown in Figure 7. The formation and convection of large scale structures are determined by fluid dynamic conditions, such as inlet velocity, forcing frequency, and amplitude. As already shown in Figures 2 (A) and 4, for a given set of fluid dynamic conditions the inverted V flame with long flame length exhibits strong interaction with a large scale structure, leading to highly nonlinear behavior. In contrast, an M flame is less influenced by the Kelvin-Helmholtz instability, since the relative ratio of convective time to the acoustic forcing period is very small, τ M − flame < τ V − flame . Furthermore, the M flame oscillates such that the flame angle varies significantly over a period of oscillation. The convection of a vortex structure is mainly determined by the mean flow, implying that the flame and the vortex structure exist in different paths for most of a given period. Hence there is not enough time for the coherent structure to interact with the flame. This further suggests that shedding of a large scale structure is not a sufficient condition for the nonlinear response of swirlstabilized flames to acoustic excitation; the interaction between the structure and the flame plays a more important role.

Figure 8. (A) Phase-synchronized deconvoluted CH* chemiluminescence images at a modulation frequency of f = 200 Hz and V ' Vmean = 25% (top), (B) line-of-sight integrated CH* chemiluminescence images at φ = 90°, 150°, 210°, and 270° (bottom). Inlet conditions: Tin = 200 oC, Vmean = 60 m/s, φ = 0.60, and X H 2 = 0.30.

7

0.6

0.6

50 CH* intensity Flame angle

100 150 200 250 300

0.5

0.4

30

0.4

CH*'/CH*mean

40

Flame angle (degree)

CH* intensity (a.u.)

0.5

Hz Hz Hz Hz Hz

0.3

0.2

0.3 0.1

0.2

0

120

240

0

20 360

0.1

0.2

0.3

0.4

0.5

0.6

V'/V mean

Phase (degree)

Figure 9. Time traces of global CH* chemiluminescence intensity and flame angle over one period of oscillation. Inlet conditions: Tin = 200 oC, Vmean = 60 m/s, φ = 0.60, and X H 2 = 0.30, f = 200 Hz, and V ' Vmean = 25%.

0

Figure 10. The normalized heat release response plotted as a function of forcing amplitude for a range of forcing frequency from 100 to 300 Hz. Inlet conditions: Tin = 200 oC, Vmean = 60 m/s, φ = 0.60, and X H 2 = 0.30.

inducing nonlinearity.

Figure 11. (A) Phase-synchronized deconvoluted CH* chemiluminescence imaging at a modulation frequency of f = 300 Hz and V ' Vmean = 30% (top). (B) Line-of-sight integrated, background-corrected CH* chemiluminescence images at φ = 90, 150, 210, and 270° (bottom). Inlet conditions: Tin = 200 oC, Vmean = 60 m/s, φ = 0.60, and X H 2 = 0.30.

Figure 10 shows the normalized CH* chemiluminescence intensity plotted as a function of forcing amplitude for the range of frequency 100−350 Hz. Unlike the cases without H2enrichment at the same inlet flow condition, the flame tends to remain in the linear regime even at high forcing frequency (f = 300 Hz) and amplitude ( V ' Vmean = 0.351). This result shows good agreement with data from a theoretical study by Lieuwen [33]. He reported that the response of the conical flame is nearly linear over the entire velocity disturbance amplitude,

while the V flames exhibit substantial nonlinearity, which is manifested as saturation. This evidence strongly suggests that the flames with enveloped M configuration are more stable than inverted V flames in terms of combustion instability. The modification of the steady-state flame configuration from a V to an M geometry by H2-enrichment was found to reduce the limitcycle pressure oscillation amplitude significantly [34]. It is interesting to note that in contrast with V flames, the normalized heat release response of M flames is insensitive to the forcing

8

6

3 X H2 =0.00(V structure) X H2 =0.30(M structure)

5

4

Gain

Phase (rad)

2

3

2

1

1

0

0

100

200

300

400

0

500

0

100

200

300

400

500

Frequency (Hz)

Frequency (Hz)

Figure 12. The gain and phase of the flame transfer function as a function of modulation frequency at a constant forcing amplitude, V ' Vmean = 0.100. Inlet conditions: Tin = 200 oC, Vmean = 60 m/s, φ = 0.60, and X H 2 = 0.00, 0.30.

frequencies, as shown in Figure 10. The measured CH* chemiluminescence intensity for a range of frequencies 100−300 Hz overlaps on the same straight line. This can be attributed to the fact that the dynamics of the M flame is less influenced by the formation and convection of large-scale structures, which is by nature dependent on perturbation frequency. The normalized CH* chemiluminescence intensity starts to decrease when the modulation frequency is greater than 350 Hz (not shown). Figure 11 shows phase-resolved CH* chemiluminescence images at a forcing frequency of 300 Hz and amplitude of 30% at the same inlet conditions as those shown in Figure 8. It can be concluded from Figures 8 and 11 that the dynamics of acoustically forced flames appear very similar, regardless of the forcing frequencies and amplitudes. The reason is that the response of the flames is linear in both cases and the forced response is independent of forcing frequencies. Under the conditions investigated here, shear layer rollup does not play an important role in inducing nonlinearity. The flame transfer functions in frequency domain are presented in Figure 12. The velocity modulation amplitude was maintained at a constant value of V ' Vmean = 10%. At this condition, the forced flame response is linear below the level of acoustic velocity perturbation of V ' Vmean = 10%. As reported by recent studies [2, 27, 28], the global flame response can be described as a low-pass filter, in that the gain decreases as the frequency increases. These results also show that with increasing X H 2 , the maximum gain of the flame transfer function decreases significantly and the phase decreases at a given forcing frequency, indicating that the gain depends strongly on flame length and steady-state flame configurations.

For the inverted V flames, the gain of the flame transfer function exceeds unity at the modulation frequency of 100 ≤ f ≤ 375 Hz, indicating that the flame amplifies flow perturbation at certain frequencies. In the case of M flames, however, the overshoot behavior is not observed, i.e., the gain is less than unity, suggesting that M flames damp upstream flow perturbations. Also, the gain is almost independent of forcing frequency when the frequency is less than 300 Hz. At high forcing frequencies, the gain asymptotically approaches zero. The phase of V flames is greater than that of M flames, due to the fact that the convection time for velocity disturbances to reach the flame increases with increasing the flame length. The phase increases almost linearly with frequency, confirming that the influence of flow disturbances on the flame’s heat release is purely convective. In the present article, the combustion response of turbulent premixed flames to acoustic oscillations was experimentally determined. The effects of spatio-temporal fluctuations of equivalence ratio nonuniformities were not considered, however. In a real gas turbine engine environment, mixture ratio oscillations can be induced in the mixing section, since fuel is generally injected at the swirl vanes or near the inlet of the combustion chamber to avoid auto-ignition and improve mixing of the fuel and air mixture. In this situation, the flame responds to both perturbations of acoustic velocity and equivalence ratio. From a scientific point of view, an experimental investigation of the response of a partially premixed flame raises many challenging questions. First, the linear/nonlinear response of a partially premixed flame subjected to acoustic velocity and fuel/air ratio modulations is expected to be appreciably different, both quantitatively and qualitatively, as compared to

9

the response of a premixed flame. The combined effects of two inlet disturbances may induce partial extinction of the reaction zone at certain inlet flow conditions, particularly when the combustor is operated near lean blowoff limits. Also, a systematic parametric study on the response of a partially premixed flame should be performed to illustrate the role of inlet disturbances in controlling the flame’s response characteristics and to elucidate the controlling physics. This work is left for future studies. CONCLUSIONS Nonlinear features of acoustically excited swirl-stabilized leanpremixed flames are characterized by nonlinear flame transfer function measurements and a phase-resolved analysis. Two distinct mechanisms of nonlinearity are investigated. The first mechanism is shear layer dynamics, which plays a key role in causing the nonlinearity of inverted dihedral V flames when the modulation amplitude is less than 40% at a modulation frequency of 200 Hz. H2-enriched enveloped M flames are found to be more stable in terms of combustion instability. Their interaction with a large-scale coherent structure is negligible, which leads to a linear response even at high forcing frequency and amplitude. This implies that steady-state flame configuration is one of the important controlling parameters determining linear and nonlinear dynamics. The second mechanism of nonlinearity is unsteady flame liftoff which causes rapid reduction in the flame surface area due to a merging of two flame branches. This mechanism is manifested when the forcing amplitude is greater than 40%. A systematic investigation of forced flame response measurements with special emphasis on the effects of fuel composition variation will be useful to develop and validate theoretical tools to predict combustion instability phenomena in fuel-flexible gas turbine engines. ACKNOWLEDGMENTS Funding for this research was provided by the Department of Energy University Coal Research Program through Contract # DE-FG26-07NT43069 and the National Science Foundation through Award #0625970. REFERENCES [1] Lieuwen, T. and Yang, V., 2005, Combustion Instabilities in Gas Turbine Engines. Progress in Astronautics and Aeronautics, 210, AIAA, Washington, DC. [2] Balachandran, R., Ayoola, B.O., Kaminski, C.F., Dowling, A.P., and Mastorakos, E., 2005, Experimental investigation of the nonlinear response of turbulent premixed flames to imposed inlet velocity oscillations. Combust. Flame, 143, 37−55. [3] Bellows, B.D., Neumeier, Y., and Lieuwen, T., 2006, Forced response of a swirling, premixed flame to flow disturbances. J. Propul. Power, 22, 1075−1084. [4] Kulsheimer, C. and Buchner, H., 2002, Combustion

dynamics of turbulent swirling flames. Combust. Flame, 131, 70−84. [5] Armitage, C.A., Balachandran, R., Mastorakos, E., Cant, R.S., 2006, Investigation of the nonlinear response of turbulent premixed flames to imposed inlet velocity oscillations. Combust. Flame, 146, 419−436. [6] You, D., Huang, Y., and Yang, Y., 2005, A generalized model of acoustic response of turbulent premixed flame and its application to gas-turbine combustion instability analysis. Combust. Sci. Tech., 177, 1109−1150. [7] Preetham, S.H. and Lieuwen, T., 2007, Response of turbulent premixed flames to harmonic acoustic forcing. Proc. Combust. Inst., 31, 1427−1434. [8] Fleifil, M., Annaswamy, A.M., Ghoneim, Z.A., and Ghoniem, A.F., 1996, Response of a laminar premixed flame to flow oscillations: a kinematic model and thermoacoustic instability results. Combust. Flame, 106, 487−510. [9] Roux, A., Gicquel, L.Y.M., Sommerer, Y., and Poinsot, T.J., 2008, Large eddy simulation of mean and oscillating flow in a side-dump ramjet combustor. Combust. Flame, 152, 154−176. [10] Sengissen, A.X., Van Kampen, J.F., Huls, R.A., Stoffels, G.G.M., Kok, J.B.W., and Poinsot, T.J., 2007, LES and experimental studies of cold and reacting flow in a swirled partially premixed burner with and without fuel modulation. Combust. Flame, 150, 40−53. [11] Staffelbach, G., Gicquel, L.Y.M., Boudier, G., and Poinsot, T., 2009, Large eddy simulation of self excited azimuthal modes in annular combustors. Proc. Combust. Inst., 32, 2909−2916. [12] Huang, Y. and Yang, V., 2009, Dynamics and stability of lean-premixed swirl-stabilized combustion. Prog. Energy Combust. Sci., 35, 293−364. [13] Kim, K.T., Lee, J.G., Quay, B.D., and Santavicca, D.A., 2010, Response of partially premixed flames to acoustic velocity and equivalence ratio perturbations. Combust. Flame, 157, 1731−1744. [14] Kim, K.T., Lee. J.G., Quay, B.D., and Santavicca, D.A., 2010, Spatially distributed flame transfer functions for predicting combustion dynamics in lean premixed gas turbine combustors. Combust. Flame, 157, 1718−1730. [15] Lieuwen, T., McDonell, V., Santavicca, D., and Sattelmayer, T., 2008, Burner development and operability issues associated with steady flowing syngas fired combustors. Combust. Sci. Technol., 180, 1167−1190. [16] Richards, G.A., McMillian, M.M., Gemmen, R.S., Rogers, W.A., and Cully, S.R., 2001, Issues for low-emission, fuelflexible power systems. Prog. Energy Combust. Sci., 27, 141−169. [17] Strakey, P., Sidwell, T., and Ontko, J., 2007, Investigation of the effects of hydrogen addition on lean extinction in a swirl stabilized combustor. Proc. Combust. Instit. 31, 3173−3180. [18] Schefer, R.W., 2003, Hydrogen enrichment for improved

10

lean flame stability. Int. J. Hydrogen Energ., 28, 1131−1141. [19] Jackson, G.S., Sai, R., Plaia, J.M., Boggs, C.M., and Kiger, K.T., 2003, Influence of H2 on the response of lean premixed CH4 flames to high strained flows. Combust. Flame, 132, 503−511. [20] Cozzi, F. and Coghe, A., 2006, Behavior of hydrogenenriched non-premixed swirled natural gas flames. Int. J. Hydrogen Energ., 31, 669−677. [21] Kido, H., Nakahara, M., Nakashima, K., and Hashimoto, J., 2002, Influence of local flame displacement velocity on turbulent burning velocity. Proc. Combust. Instit., 29, 1855−1861. [22] Sarli, V. and Benedetto, A., 2007, Laminar burning velocity of hydrogen-methane/air premixed flames. Int. J. Hydrogen Energ., 32, 637−646. [23] Mandilas, C., Ormsby, M.P., Sheppard, C.G.W., and Wooley, R., 2007, Effects of hydrogen addition on laminar and turbulent premixed methane and iso-octane-air flames. Proc. Combust. Instit., 31, 1443−1450. [24] Schefer, R.W., Wicksall, D.M., and Agrawal, A.K., 2002, Combustion of hydrogen-enriched methane in a lean premixed swirl-stabilized burner. Proc. Combust. Instit., 29, 843−851. [25] Wicksall, D.M., Agrawal, A.K., Schefer, R.W., and Keller, J.O., 2005, The interaction of flame and flow field in a lean premixed swirl-stabilized combustor operated on H2/CH4/air. Proc. Combust. Instit., 30, 2875−2883. [26] Wicksall, D.M. and Agrawal, A.K., 2007, Acoustics measurements in a lean premixed combustor operated on hydrogen/hydrocarbon fuel mixtures. Int. J. Hydrogen Energ., 32, 1103−1112.

[27] Kim, K.T., Lee, J.G., Lee, H.J., Quay, B.D., and Santavicca, D.A., 2010, Characterization of forced flame response of swirl-stabilized turbulent lean-premixed flames in a gas turbine combustor. J. Eng. Gas Turb. Power, 132, 04. [28] Durox, D., Schuller, T., Noiray, N. and Candel, S., 2009, Experimental analysis of nonlinear flame transfer functions for different flame geometries. Proc. Combust. Inst., 32, 1391−1398. [29] Thumuluru, S.K. and Lieuwen, T., 2009, Characterization of acoustically forced swirl flame dynamics. Proc. Combust. Inst., 32, 2893−2900. [30] Schuller, T., Durox, D., and Candel, S., 2003, Self-induced combustion oscillations of laminar premixed flames stabilized on annular burners. Combust. Flame, 135, 525−537. [31] Schuller, T., Durox, D., and Candel, S., 2003, A unified model for the prediction of laminar flame transfer functions: comparisons between conical and V-flame dynamics. Combust. Flame, 134, 21−34. [32] Cala, C.E.C, Fernandes, E.C., and Heitor, M.V., 2002, Analysis of oscillating shear layer. 11th Symposium on Application of Laser Techniques and Fluid Mechanics, Lisbon, Portugal. [33] Lieuwen, T., 2005, Nonlinear kinematic response of premixed flames to harmonic velocity disturbances. Proc. Combust. Inst., 30, 1725−1732. [34] Kim, K.T., Lee, H.J., Lee, J.G., Quay, B., and Santavicca, D., 2009, Flame transfer function measurement and instability frequency prediction using a thermoacoustic model. ASME paper GT2009−60026.

11

Manuscript Not AVAILABLE

SUITABILITY OF A SOUTH-AFRICAN HIGH ASH CONTENT AND HIGH ASH FLOW TEMPERATURE COAL SOURCE FOR ENTRAINED FLOW GASIFICATION JC van Dyk* and R Stemmer** * Sasol Technology, R&D, Sasolburg, South-Africa, [email protected], +27169604375 (tel), +27115224806 (fax) ** Corus Technology, RD&D, Ijmuiden, The Netherlands, [email protected], +31251492073 (tel), +31251470489 (fax) In slagging gasifiers the ash flows down the gasifier walls and drains from the gasifier as molten slag. Coals selected for slagging gasifiers should thus have an ash flow temperature (AFT) below the operating temperature of the gasifier, and in practice can be lowered by the addition of a flux, such as limestone. As the mineral matter start to melt and become a liquid (between the softening temperature and the flow temperature of the mineral matter in coal), it will have a specific composition and a related viscosity (ease of flow). During entrained flow gasification the viscosity has to be low enough for the slag to flow and drain from the gasifier. As a consequence, the viscosity of the slag, which depends on the slag composition, is one of the most critical factors in the operation of slagging gasifiers. The pilot gasifier at Corus in The Netherlands, as used to produce reduction gas for the steel industry, was used for the test. The objective of the experiment was to investigate the gasifiability of high ash coal in an oxygen blown, atmospheric pressure entrained flow pilot gasifier, and to quantify / qualify where possible and within the known limits of such a pilot test unit, the slag, fly-ash, water and gas characteristics. The coal was fed pneumatically with N2 at ±50kg/hr through the oxygen coal burner. The outlet velocity of the burner was set at 110m.s-1 (oxygen flow rate of 20Nm3hr-1). The header temperature was controlled at maximum 1600oC, resulting in a temperature at the bottom part of the gasifier between 1200 and 1300oC. Due to this rapid heat loss over the gasifier length, the coal was over-fluxed in order to maintain a running slag and avoid blockages at the bottom temperature of 1200oC, which is almost 150oC lower than normal slagging operating conditions. The refractory hot face showed limited wear or penetration of minerals into the refractory. The slag viscosity was thus fluxed adequately for the specific setup to flow gently over the refractory and protecting the SiC. The measured and equilibrium simulated carbon balances corresponded well and could be closed with a high degree of accuracy. During stable and high load periods, the CO content varied between 50 and 60 vol%, H2 between 10 and 20 vol% and CO2 350oC) at the same steam/CO ratio. The effect of steam/CO ratio on CO conversion showed different results for three WGS catalysts. For MDC-7 and RSM catalysts, CO conversion increased slightly as the steam/CO ratio increased up to 2.0, and maintained. However, CO conversion of PC catalyst increased continuously as the steam/CO ratio increased up to 5.0. The reactivity of MDC-7 catalyst was maintained more than 8 hours but that of PC catalyst decreased as the reaction time increased. As a conclusion, MDC-7 and RSM catalysts showed better reactivity and PC catalyst should be improved its reactivity. However, MDC-7 catalyst generated much fines during operation, and therefore, attrition resistance should be improved.

Introduction Hydrogen production is the most fundamental part of the hydrogen energy system, and has always been the object of intense and vigorous research and development. World hydrogen production has been growing rapidly at 8-10% per annum for many years [1]. At present, hydrogen is produced mainly from fossil fuels, water and biomass. However, more than 90% of the hydrogen is produced from fossil fuels [2]. Series of gasification of coal, water gas shift, and CO2 separation is the predominant production route to hydrogen from coal for commercial scale application. However, this process needs multiple-steps such as high- and low-temperature water gas shift reaction as shown in eq. (1) to improve hydrogen yield, and CO2 separation process to separate almost pure hydrogen from the gas mixture of CO, CO2, and H2. To separate CO2 from the exhaust gas, additional energy and equipments are required. More than 22% of hydrogen generation cost comes from CO2 separation process for purifying hydrogen[3]. Although the

previous process has been used for many years, there are some areas for improvement. The previous process requires many reactors and many kinds of catalysts and/or sorbents. Therefore, it will be extremely desirable if new concepts can be developed which can reduce the capital and operating cost of the conventional process[4]. To overcome these disadvantages, SEWGS (Sorption Enhanced Water Gas Shift) system has been developed. Equation (2) and (3) explain concept of SEWGS system. The thermodynamic equilibrium in the shift reaction can be enhanced to give more hydrogen yield by adding a CO2 absorbent into the shift reactor. Carbon dioxide is then captured as a solid carbonate as soon as it formed, shifting the reversible water-gas shift reactions beyond their conventional thermodynamic limits as shown in eq. (2). Regeneration of the sorbent releases pure CO2 suitable for sequestration as shown in eq. (3). It is important that the gas composition of the exhaust gas from the SEWGS reactor contains only highly concentrated H2 and excess water vapor. Therefore, H2 can be easily recovered by cooling the exhaust gas without any extra energy consumption for H2 separation. Moreover, the exhaust gas from the regeneration reactor contains only carbon dioxide and water vapor if we use steam as fluidization gas. After water condensation, almost pure carbon dioxide can be obtained with little energy loss for component separation. Water gas shift reaction CO + H2O  H2 + CO2 SEWGS reaction CO + H2O + MO  H2 + MCO3 Regeneration reaction MCO3  MO + CO2 where, MO: metal oxide, MCO3: metal carbonate

(1) (2) (3)

In this study, the reaction characteristics of three WGS catalysts for SEWGS have been investigated in a bubbling fluidized bed reactor, as a preliminary research. The commercial low temperature WGS catalyst (MDC-7) produced by Süd-chemie and new WGS catalysts (PC and RSM) produced by KEPRI (Korea Electric Power Research Institute) by means of spray-drying were used as bed materials. Reaction temperature, steam/CO ratio, and gas velocity were considered as experimental variables. Moreover, long-term operation results of WGS catalysts were compared as well.

Experimental The reactivity tests were carried out in a bubbling fluidized bed reactor. A schematic of the reactor is provided in Figure 1. The major components consist of a gas input system, a fluidized bed, a condenser, a hot gas filter, a gas sampling/analyzing unit, and water feeding pump. The fluidization column is 0.7 m high with an internal diameter of 0.05 m. A perforated gas distributor plate separates the fluidization column and air box. Reactant gas was fed to the air box. An electric heater could be controlled by a thermocouple and a heater controller. Temperature and pressure data were recorded by a data acquisition system. The exit stream from the fluidized bed reactor was sampled at the outlet of the reactor. The CH4, CO, CO2, H2, NO, and O2 concentrations were monitored using an on-line gas analyzer and recorded by a data acquisition system. Further details of the reactor system are available in our previous paper[5]. Three water gas shift catalysts, the commercial low temperature WGS catalyst (MDC-7) produced by Süd-chemie and new WGS catalyst (PC, RSM) produced by KEPRI (Korea Electric Power Research Institute) were used. MDC-7 catalyst had pellet shape and we crushed the pellets to 106~212 m. However, PC and RSM catalyst has spherical shape and the same particle size range was prepared. Figure 2 shows photos of three WGS catalysts. The PC and RSM catalysts show spherical shape and the MDC-7 catalyst shows irregular shape. The static bed height was 0.4 m in all cases, and initial solid masses were 0.57 kg for PC catalyst (b=724.7 kg/m3), 0.64 kg for RSM(b=747.3 kg/m3), and 0.88 kg for MDC-7 catalyst (b=1117 kg/m3), respectively. The fluidized bed reactor operated with a total inlet gas flow of 2.0 Nl/min in all cases, except for tests to check effect of gas velocity. The total inlet gas contained 50% of syngas and 50% of nitrogen. The syngas composition was 60.5% of CO, 27.2% of H2, 9.9% of CO2 and N2 as a balance.

Figure 1. Schematic of a bubbling fluidized bed reactor.

Figure 2. Photos of PC, RSM and MDC-7 catalysts. Results and discussion Prior to the start of each experiment, catalysts were reduced by H2 gas (57%, N2 balance) at 400oC. Figure 3 shows breakthrough curves of hydrogen concentrations during pretreatment (reduction) of catalysts. A breakthrough of hydrogen concentration marked the end of reduction. As shown in Figure 3, since MDC-7 and RSM catalysts showed sharper breakthrough curve than PC catalyst, and therefore, we could expect that the MDC-7 and RSM catalysts would show better reactivity than PC catalyst.

70 PC RSM (reformed) MDC-7

H2 concentration [%]

60 50 40 30 20 10 0 0

20

40

60

80

Time [min]

Figure 3. H2 breakthrough curves during pretreatment of catalysts. Figure 4 shows effect of temperature on CO conversion of catalysts. The CO conversion to H2 and CO conversion to CH4 were calculated by mass balance based on the output gas concentration and the tie component (N2). For MDC-7 catalyst, high CO conversion up to 99.4% was observed in the range of 220~240oC at 4.0 of steam/CO ratio. For RSM catalyst, lower CO conversion than MDC-7 catalyst was observed. Moreover, for PC catalyst, lower CO conversion than MDC-7 catalyst was observed even at higher temperature (380~400oC).

100

CO conversion [%]

90

100

PC catalyst

90

100

RSM

MDC-7

90

80

80

80

70

70

70

60

60

60

50

50

50

40

40

40

30

30

30

20

20

20

Total CO conversion CO to H2 conversion

10

10

10

CO to CH4 conversion

0 200

250

300

350

400 o

Temperature [ C]

Steam/CO ratio = 4

0 0 450 140 160 180 200 220 240 260 280 300 320 140 160 180 200 220 240 260 280 300 320

Temperature [oC]

Temperature [oC]

Figure 4. CO conversion versus reaction temperature. Figure 5 shows effect of steam/CO ratio on CO conversion of two catalysts. For MDC-7 catalyst, CO conversion increased slightly as the steam/CO ratio increased up to 2.0, and maintained at high level thereafter. For RSM catalyst, CO conversion increased as the steam/CO ratio increased up to 2.0, and maintained thereafter. However, CO conversion of PC catalyst increased continuously as the steam/CO ratio increased up to 5.0. MDC-7 catalyst showed higher CO conversion at the same steam/CO ratio and at lower temperature. Moreover, PC catalyst generated higher CH4 at lower steam/CO ratio.

CO conversion [%]

100

100

100

90

90

90

80

80

80

70

70

70

60

60

PC series o 360 C

50

60

RSM 200 oC

50

MDC-7 200 oC

50

40

40

40

30

30

30

20

20

20

Total CO conversion CO to H2 conversion

10

10

10

CO to CH4 conversion

0

0 1

2

3

4

5

6

0 1

2

Steam/CO ratio [-]

3

4

5

6

1

Steam/CO ratio [-]

2

3

4

5

Steam/CO ratio [-]

Figure 5. CO conversion versus steam/CO ratio. Figure 6 shows long-term test results of two catalysts. The reactivity of MDC-7 catalyst was maintained more than 8 hours but that of PC catalyst decreased as the reaction time increased. As a conclusion, MDC-7 catalyst showed better reactivity than PC and RSM catalysts from the viewpoints of reaction temperature, seam/CO ratio, CO conversion, and long-term durability. 100 90

CO conversion [%]

100

gas analyzer trouble

90

80

80

70

70

60

60

50

50 (a) PC catalyst at 380oC steam/CO ratio = 5

40 30

(b) MDC-7 catalyst o at 200 C steam/CO ratio = 4

40 30

20

Total CO conversion CO to H2 conversion

20

Total CO conversion CO to H2 conversion

10

CO to CH4 conversion

10

CO to CH4 conversion

0 0

100

200

300

Time [min]

400

0 500 0

100

200

300

400

500

Time [min]

Figure 6. CO conversion versus time. To check effects of syngas concentration and gas velocity, supplementary tests were performed using MDC-7 catalyst in the same reactor and the results are provided in Figure 7. The CO conversion of MDC7 catalyst decreased slightly as the syngas concentration increased, but increased as the gas velocity decreased and steam/CO ration increased. However, these values are much higher than the results from the fixed bed with the same catalyst[6].

100 99

CO conversion [%]

98 97 96 95 94

Steam/CO ratio=2, U=0.05 m/s Steam/CO ratio=4, U=0.05 m/s Steam/CO ratio=4, U=0.025 m/s Steam/CO ratio=3, U=0.078 m/s Steam/CO ratio=3, U=0.039 m/s

93 92 91 90 0

20

40

60

80

100

Syngas concentration [%]

Figure 7. CO conversion to H2 versus syngas concentration (MDC-7 catalyst). Conclusion The best operating temperature and steam/CO ratio showed different results depend on the WGS catalysts. For MDC-7 catalyst, high CO conversion up to 99.4% was observed in the range of 220~240oC at 4.0 of steam/CO ratio. However, for PC catalyst, 90% of CO conversion achieved even at higher temperature (>350oC) at the same steam/CO ratio. For MDC-7 and RSM catalysts, CO conversion increased slightly as the steam/CO ratio increased up to 2.0, and maintained thereafter. However, CO conversion of PC catalyst increased continuously as the steam/CO ratio increased up to 5.0. The reactivity of MDC-7 catalyst was maintained more than 8 hours but that of PC catalyst decreased as the reaction time increased. For RSM catalyst showed better reactivity than PC catalyst but worse reactivity than MDC-7 catalyst. As a conclusion, MDC-7 catalyst showed better reactivity than PC and RSM catalysts from the viewpoints of reaction temperature, seam/CO ratio, CO conversion, and long-term durability. However, MDC-7 catalyst generated much fines during operation, and therefore, attrition resistance should be improved. Acknowledgement This work was supported by the Energy Efficiency and Resources program of the Korea Institute of Energy Technology Evaluation and Planning(KETEP) grant funded by the Korea government Ministry of Knowledge Economy References 1. Kothari, R., Buddhi, D. and Sawhney, R. L. (2004). “Sources and Technology for Hydrogen Production: a Review”, Int. J. Global Energy Issues, 21(1/2), 154-178. 2. IEA report (2005). “Prospects for Hydrogen and Fuel Cells”, IEA Books, 49-55. 3. Maurstad O. (2008), “An Overview of Coal Based Integrated Gasification Combined Cycle (IGCC)”, MIT report, Publication No. LFEE 2005-002 WP, 1-43. 4. Ryu, H. J. (2009). “Selection of Process Configuration and Operating Conditions for SEWGS System”, Trans. of the Korean Hydrogen and New Energy Society, 20(2), 168-178. 5. Ryu, H. J., Shun, D., Bae, D. H., and Park, M. H. (2009). “Syngas Combustion Characteristics of Four Oxygen Carrier Particles for Chemical Looping Combustion in a Batch Fluidized Bed Reactor”, Korean J. of Chem. Eng., 26(2), 523-527. 6. Chen, W. S. and Jheng, J. G. (2007). “Characterization of Water Gas Shift Reaction in Association with Carbon Dioxide Sequestration”, J. of Power Sources, 172, 368-375.

CARBON DIOXIDE CAPTURE OF FLUE GASES FROM COAL-FIRED POWER PLANT USING ENZYMES ORIGINATED MARINE LIFE Soonkwan Jeonga, Kyungsoo Lima, Jeongwhan Lima, Daehoon Kimb, Mari Vinobaa,c, S.H. Leea,* a

Climate Change Technology Division, Korea Institute of Energy Research 102, Gajeong-ro, Yuseong-gu, Daejeon, 305-343, Korea b Department of Chemical and Biological Engineering, Korea University, Seoul 136-713, Korea c Department of Chemical Engineering, Anna University-Chennai, Chennai 600 025, India

*E-mail: [email protected]

ABSTRACT The focus of this study is the separation and storage of green house gas, CO2, and the use of enzymes from marine life in development of technology to provide novel method for CO2 capture. Carbonic anhydrase has recently been used as a biocatalyst to accelerate an aqueous processing route to carbonate formation. In this study, we compared soluble proteins of HDS (Hemocyte from Diseased Shell) and EPF (Extrapallial Fluid) extracted from crassostrea gigas with HCA (Human Carbonic Anhydrase) and BCA (Bovine Carbonic Anhydrase) on their ability to promote CO2 hydration and the production of calcium precipitates. HCA, BCA, HDS, and EPF have shown promising results for use as promoter to accelerate CO2 hydration and increase the rate of precipitation of carbonate mineral with Ca2+ ions. The ideal temperature and pH of operation was found to be 40℃ and 6~7, respectively. Lineweaver-Burk relationship was employed to estimate Michaelis-Menten kinetic parameters for the enzymes. BCA showed the fastest rate constant and followed HCA, HDS and EPF. In previous efforts to use CO2 mineralization as a method for CO2 sequestration the slow rate of hydration of CO2 to carbonic acid has been limiting factor of CO2 mineralization. In the presence of enzymes, rate determining step is eliminated, and therefore the overall reaction rate is enhanced dramatically. Precipitated CaCO3 was all calcite and the particles size was below 100nm. These nano-particles could be use in other industrial processes such as paper, ink, or building materials. Therefore it can be reduced the operating cost of CO2 capture, which increase feasibility to install CO2 capture process in coal-fired power plant. This result suggests that enzymes mentioned above may be involved not only in CO2 hydration but also in CO2 mineralization.

1. INTRODUCTION The contribution of greenhouse gases, especially carbon dioxide, to global warming is well recognized [1]. An energy source option for the humankind energy system has become too numerous, and the thirst for more and more energy targets on usage of fossil fuels in coal-fired power plants, steel industries, and cement contributes to an increasing amount of CO2 emission. The flue gases from conventional coal-fired power plants in Korea typically contain 12~14% CO2 with the balance consisting mainly of nitrogen and small quantities of oxygen and impurities. The carbon dioxide concentration of flue gases is insufficient for direct compression

to transpport and stoorage of carbbon dioxidee. In order to t get conceentrated CO2 from flue gases, we have to install CO O2 capture pprocess suchh as amine or ammonnia absorptioon. The am mine-based solventss are widely used for the CO2 ccapture proccess. Despiite of the ggood CO2 aabsorption capacityy and fast absorption ab raate, they haave some drrawbacks suuch as corroosion, loss oof solvent, generatiion of heat stable salts,, and high ppenalty for eenergy conssumption duuring regeneeration [24]. If ann alternativee solvent syystem havingg less energgy for regenneration couuld be develloped, it is very atttractive opption to innstall CO2 capture faacility. Sincce recentlyy, emergingg ex-vivo applicattions of carbbonic anhyddrase for its potential usse in CO2 caapture technnologies are attracting attentionns [5-6]. The obbjective of tthe present study was tto investigatte the feasibbility of usinng enzymess extracted from maarine life ass a biocatalyyst for hydrration of CO O2, as well as a its precippitation in thhe form of calcium m carbonate. We have bbeen investigated a kineetic analysiss and the eff ffects of conncentration of enzyyme, pH, teemperature oon hydratioon of CO2. In additionn, an enzym matic precippitation of calcium m carbonate was w studiedd.

2. EX XPERIMENTAL O2 capture aand sequesttration procedures 2-1. CO The eexperiment w was conducteed in a glass reactor (1 L)) with a conttrol system, aas shown in F Fig. 1. The reaction temperature was controllled by a wateer bath (JEIO OTECH CW-10G) and C CO2 was introoduced into the bottoom of the reactor. The sttirring rate and a temperatture in the reeactor were ccontrolled byy a control system, and the pH was measurred using a ppH meter. C CaCl2 (Sigmaa C1016) waas used as thhe calcium source. T The bovine ccarbonic anhhydrase II ussed in this stuudy was obttained from S Sigma (St. L Louis, MO, U.S.A.). HDS and EP PF were extraacted from Crassostrea C ggigas (C. gigaas). microscope The miicrostructure of precipitattes was studied using a JJEOL JSM-8840A scanninng electron m with AN N-10000/85S (LINK systeem) energy diispersive speectrometer (E EDS) capabiliity.

Figure1. Experimenttal illustratioon for CO2 sequestratioon using byy enzyme. T The experim ments were approxim mately divideed into two parts (dotteed line), reacctor part (a)) (i: reactor, ii: electric stirrer, iii: thermal couple, iv: ppH meter, v: CO2 line) annd control syystem part (bb). (i: controller of tempeerature and

stirring rate, ii: flow mass). And the other parts were equipped with computer (c), water bath (d), and CO2 gas (e).

2-2. Separation and purification of enzymes The diseased area in oyster shell, Crassostrea gigas, was defined in previous study. The diseased area was separated using a knife and ground to a fine powder in an electric mill. The powder (25 g) was extracted in 150 ml of 0.2 M EDTA (pH 8.0) at 4 oC with continuous stirring. After 2 days the soluble extract was obtained by centrifugation at 25,000 × g for 20 min. The supernatant was diluted with an equal volume of deionized water. The diluted solution was concentrated by ultrafiltration using a minimodule. The concentrated fraction was dialyzed against 1 L of deionized water at 4 C for 3 days. The dialyzed fraction was concentrated to a volume of 20 ml and subjected to ethanol precipitation at 25 C for 1 week. The precipitate was dissolved with continuous stirring in 3 ml of 50 mM NaHPO2·H2O (pH 7.18) at 5 C, and dialyzed for 3 days against deionized water at 4 C. The dialyzed solution was lyophilized. The soluble protein extracted from the diseased area was termed HDS (hemocyte from diseased shell). Specimens of C. gigas were collected (Namhae, Korea) and the EPF was removed immediately by prying the valves open, inserting the needle of a sterile hypodermic syringe under the mantle edge into the central extrapallial space, and withdrawing the fluid. The EPF samples were maintained on ice in a refrigerator at 4 C.

3. RESULTS and DISCUSSION Fig. 2 demonstrates the pH change at room temperature with buffer and without buffer system. Carbonic anhydrase catalyzes the hydration of CO2 and consequently hydrogen ions are generated. This results in a change of pH. Therefore, measuring pH is a viable method to monitor the progress of this enzymatic reaction [6]. Addition of CO2 into the reactor at time zero tended to drop immediately without buffer solution, but the presence of buffer solution pH value leveled off. BCA was widely used as a biocatalyst for hydration of CO2 and showed very fast reaction rate. HDS and EPF extracted from C. gigas showed almost same activity. This result suggests that HDS and EPF can apply a biocatalyst for CO2 capture. 7.0

7.2 o

Temp : 25 C

6.5

7.0

Blank(Buffer) BCA (Buffer) Blank(Water) BCA (Water)

5.5

6.8

pH

6.0

pH

blank BCA HDS EPF

6.6

5.0

6.4

4.5

6.2

4.0

6.0

0

200

400

Time (sec)

600

800

0

200

400

600

Time (sec)

Figure2. pH experiments at T = 25oC with or without buffer.

800

Though enzyme shows high catalytic activity for hydration of CO2, it has certain limitation in its application due to the short lifetime of enzyme. There are different methods to improve catalytic stability of enzyme such as enzyme immobilization, enzyme modification and genetic modification. Among the different methods, immobilization of the enzyme into the solid support is used because of its proper applications. We made several kinds of solid supports such as SBA15, SBA-16, and KIT-6. Pore expanded SBA-15 that has long chain pore shows high activity. SBA-15 has a 2D hexagonal structure and 14nm of pore diameter. Enzyme was immobilized over large pore expended meso-porous silica through cross linked enzyme aggregation with covalent bonding. Fig. 3 shows comparison of free and immobilized enzyme on hydration of CO2. Bio-catalytic activity of free enzyme demonstrates slightly high due to loss of enzyme during immobilization. The catalytic activity of immobilized enzyme did not change during cyclic experiment. The stability of immobilized enzyme indicated the enzyme attached as cross linked aggregates with covalent bond within the porous silica did not lose its activity by either denaturation or detachment from the structure. 7.0

7.0

CO2 in

CO2 in

CO2 in

Immoblized BCA

Free BCA 6.8

6.8

6.6

pH

pH

6.6

6.4

6.4

6.2

6.2

0

100

200

300

Time(sec)

400

500

600

CO2 out N2 in

CO2 out N2 in

6.0

6.0 0

2000

4000

BCA Immobilization 6000

8000

Time(sec)

Figure3. Comparison of the catalytic activity for free and immobilized enzyme. Determination of Km and Kcat values of enzyme was carried out by measuring the activity in the presence of p-nitrophenyl acetate. Km and Kcat were determined using the Lineweaver-Burk double reciprocal plot, in which the reciprocals of the initial velocities of the enzyme activity were plotted against the reciprocals of the concentration of p-nitro phenyl acetate used. Km and Kcat value of BCA were 16.8mM and 2.1s-1 with p-NPA as substrate. However, Km and Kcat value of EPF were 8.8mM and 0.1s-1. HDS and EPF are soluble biopolymer. They consist of active site for hydration of CO2 and inert site, and then their activities are lower than those of BCA. As bivalve shells are composed mainly of calcium carbonate, they provide a good model for the study of CO2 sequestration because the shell is derived from the calcium ions and CO2 in seawater. In the precipitation reaction, the onset time for precipitation with enzyme is approximately 4 times faster as compared to the carbonation reaction in the absence of enzyme. It means that enzyme accelerates the formation of bicarbonate or carbonates ions. Precipitated calcium carbonate has calcite structure by XRD and size of it is below 100nm. In previous efforts to use CO2 mineralization as a method for CO2 sequestration, the slow rate of hydration of CO2 has been limiting factor of CO2 mineralization [7]. A calcium precipitate was identified in experiments using enzyme such as BCA, HCA, HDS and EPF. This result suggests that all

enzymes in this study may be involved not only in CO2 hydration but also in CO2 mineral carbonate conversion. Calcium carbonate is a common and thermodynamically stable mineral found in rocks worldwide, and is the main component of shell of marine organism, snails and eggs. If the widespread transformation of CO2 to CaCO3 is possible, it will represent a stable process for long-term CO2 capture and storage. The study presented here is only the beginning for the capture and storage of CO2 using enzyme, especially HDS and EPF. Additional studies on various conditions which include cloning of enzyme, operating conditions, and scale-up factor are underway.

COMCLUSIONS Enzyme extracted from marine life has been successfully applied for the hydration and mineralization of carbon dioxide. HDS and EPF accelerate the rate of CO2 hydration and mineralization. Enzyme immobilization over pore expanded silica enhances the stability of enzyme as compared to free enzyme. Kinetic parameters of enzyme were also evaluated from Lineweaver-Burk plot. The Km was 16.8mM and Kcat was 2.1s-1 for BCA, whereas of the EPF Km was 8.8mM and Kcat was 0.1s-1. Therefore, HDS and EPF need to purify to improve catalytic activity. HCA, BCA, HDS, and EPF have shown promising results for use as promoter to accelerate CO2 hydration and increase the rate of precipitation of carbonate mineral with Ca2+ ions.

Reference 1. IPCC, Special Report on Carbon Dioxide Capture & Storage, (2005). 2. B.P. Mandal, M. Guha, A.K. Biswas, S.S. Bandyopadhyay, Chem. Eng. Sci., 56 (2001) 6217. 3. P.Y. Chung, A.N. Soriano, R.B. Leron, M.H. Li, J. of Chemical Thermodaynamics, 42 (2010) 802. 4. G.Gao, G.H. Liang, H. Wang, Corrosion Science, 49 (2007) 1833. 5. G.M. Bond, J. Stringer, D.K. Brandvold, F.A. Simsek, M.G. Medina, G.E. Egeland, Energy and Fuels 15 (2001) 309. 6. P. Mirjafari, K. Asghari, N. Mahinpey, Ind. Eng. Chem. Res., 46 (2007) 921. 7. S.W. Lee, S.B. Park, S.K. Jeong, K.S. Lim, S.H. Lee, M.C. Trachtenberg, Micron, 42 (2010) 273.

Manuscript Not AVAILABLE

2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission

CARBON MANAGEMENT PREPARATION AND CHARACTERISTICS OF FORMED ACTIVE CARBONS FOR NATURAL GAS STORAGE Grzegorz Łabojko, Ph.D., Institute for Chemical Processing of Coal Zamkowa stree,t 1, 41-803 Zabrze, Poland [email protected] Tel: + 48 32 271-00-41 int. 133, Fax. + 48 32 271-08-09

Aleksander Sobolewski, Ph.D., Institute for Chemical Processing of Coal Zamkowa street, 1, 41-803 Zabrze, Poland [email protected] Tel: + 48 32 271-00-41 int. 220, Fax. + 48 32 271-08-09 Leszek Czepirski, prof. , AGH – University of Science and Technology, Faculty of Fuels and Energy, al. Mickiewicza 30, 30-059 Cracow, Poland [email protected] Tel.+48 12 617-46-36, 617-20-66; Fax. +48 12 617-45-47

Abstract: Natural gas as an automotive fuel presents various advantages versus petrol and gasoline including reduced vehicle emissions, lower maintenance and savings in fuel costs. Main problem in using natural gas as fuel for vehicles is low volumetric energy density of methane at room temperature. Compressed Natural Gas (CNG) is a solution used worldwide (more than 4 millions vehicles), but heavy and expensive cylinders have to be used for its storage at pressures about 200 bar. Adsorbed Natural Gas (ANG) can overcome issues with heavy storage systems by lowering operating pressure down to 35-40 bar. Key to the success of adsorptive storage is the choice of suitable adsorbent and operating conditions. Porous carbonaceous materials are well known as one of the best adsorbents for gases. Preparation procedure of carbonaceous adsorbents for natural gas storage is described. Based on physicochemical parameters Polish hard coal type 34.1 from KWK Marcel was selected as precursor material for adsorbents for gas storage. Four grindings were performed on selected coal and particle size distribution was determined. Fullers distribution was calculated as providing best packing of precursor material. Comparison between optimal and obtained distributions allowed to select the best way of grinding coal providing best packing of precursor material. Thermo prepared coal tar and phenol-formaldehyde resol type resin were selected as binders, thermoreological study of binders and coal-binder mixtures were performed. Granules and pastilles from coal-binders mixtures were prepared and pyrolysed. Influence of binder amount in mixtures on mechanical resistance of granules were studied. Influence of steam activation time of granules on sorption parameters (Iodine number) of obtained active carbon monoliths was studied. High pressure adsorption of methane and low pressure adsorption of nitrogen was performed on selected samples of obtained active carbons monoliths.

The kinetics of the CO2 reforming of CH4 over carbonaceous catalyst Fengbo Guo, Yongfa Zhang* , Guojie Zhang, Bingmo Zhang Key Laboratory of Coal Science and Technology of Shanxi Province and the Ministry of Education ,Taiyuan University of Technology , Taiyuan 030024 , China Abstract: The CO2 reforming of CH4 on carbonaceous catalyst was performed in a Plug Flow Reactor (PFR) at temperatures range from 1223 K to 1323K, the ratio of CH4/CO2=1 and residence time 3 ～ 30s under normal pressure. The outlet gas was analyzed by Gas Chromatogram (GC-960TCD and GC-950TCD), the carbonaceous catalyst was analyzed with element analyzer and specific surface area analyzer. The experimental results show that the conversion of CH4 and CO2 increases with the increase of the reaction temperature and residence time. The reaction temperature remains the chief influential factors on the conversion of CH4 and CO2. Under the conditions of the temperature of 1323K and residence time over 20s, the conversion of CH4 and CO2 can be expected over 90%. In addition, the conversion of CO2 was significantly higher than the conversion of CH4, which indicate that the gasification reaction of the carbonaceous catalyst and carbon dioxide was occurred during the reforming process. A mechanism of the CO2 reforming of CH4 on carbonaceous catalyst has been proposed based on the experimental results. Based on the mechanism, a kinetic model was developed. The kinetics of the CO2 reforming of CH4 on carbonaceous catalyst was described in a rate law, the experimental data was analysed in non-linear regression. The apparent activation energy Ea and the pre-exponential factor A were solved. The rate constant was as asfollows: k

= 1765

exp （

−

109.9

± 10.1KJ RT

）

A comparison is made between calculation data and experimental data of the CH4 conversion, which illustrates the rationality of the kinetic model. Keywords:

Kinetic; Carbon dioxide reforming of methane; carbonaceous catalyst

1. Introduction Methane is the main composition of natural gas , coal bed gas and coke oven gas. The CO2 reforming of CH4 provides a practical method for effectively combines CH4 utilization with CO2

*Corresponding author. Tel. : +86 0351 6018676 E-mail address: [email protected]

transforming, since both methane and carbon dioxide are the main greenhouse gas, and the reaction gets raw gases for Fischer-Tropsch, methanol and carbonyl synthesis [1-3]. In recent years, the CO2 reforming of CH4 has received considerable attention word wide. There are four major kinds of catalysts, which are transitional metals catalysts, Composite metal oxide catalysts, supported metal sulfide catalysts and supported noble metal catalysts[4-6], but carbonaceous catalyst has attracted the attention of most researchers who are studied CO2 reforming of CH4 reaction , carbonaceous catalyst is both cheap and abundantly available in the word, and which cannot lead to catalyst poisoning . Li Yan-bing[7] studied the influence of different chars and operation parameters on reactant gas conversion, and the results show that Tongchuan char with the lowest ash concentration exhibits the highest activity, and all the char samples show similar behaviors. A higher conversion is always achieved at the beginning, and then it decreases and levels off at a stable value after 30 min. Guojie Zhang [8] studied the effects of the coal char catalyst pretreatment and the ratio of CO2/CH4, experimental results showed that the coal char was an effective catalyst for production of syngas, and the product gas ratio of H2/CO is strongly influenced by the feed ratio of CO2/CH4. Zhang Hua wei[9] studied the effect of coke on the steam and carbon dioxide reform ing reaction s of methane at 700℃-1300℃, the results show that the conversion of methane to produce syngas obviously increased with the present of coke in the reactor. In the present work, the CO2 reforming of CH4 over the DaTong carbonaceous catalyst was performed in a Plug Flow Reactor. The effect of reaction parameters on the carbonaceous catalyst activity for CO2 reforming of CH4 were studied , and the kinetic behavior was investigated as functions of temperature and residence time .

2. Experimental 2.1 Experimental Equipment The CO2 reforming of CH4 over the DaTong carbonaceous catalyst was performed in a Plug Flow Reactor at temperatures range from 1223 K to 1323K,

the ratio of CH4/CO2=1 and residence

time 3～30s under normal pressure. The reactor (internal diameter 20mm; length 100mm)was horizontal heated in a furnace , A weighed amount (15g) of carbonaceous catalyst was loaded in the middle of the reactor . The CH4 and CO2 gas flow rates were measured and controlled by Mass flow

controllers (flow control range 0～200mL/min).The outlet gas was analyzed by Gas Chromatogram (GC-960TCD and GC-950TCD), the carbonaceous catalyst was analyzed with element analyzer and specific surface area analyzer. The schematic diagram of the Experimental equipment is given in Fig. 1.

Fig. 1 Schematic diagram of the experiment 1.2. cylinder 3. N2 cylinder 4. Mass flow controllers 5. quartz reactor6. carbonaceous catalyst 7. Temperature controller 8. Heat oven 9. Sample sampler 10. Gas chromatograph

2.2 Catalyst preparation and characterization The carbonaceous catalyst was prepared by pyrolysis of Da tong coal at 800℃ for 2h , crushing the catalyst mass to 80～100 mesh-size particles . The carbonaceous catalyst was dried in muffle furnace for 12h at 900℃ in nitrogen . The Proximate analysis and ultimate analysis of carbonaceous catalyst were shown in table 1. Table 1 The Proximate analysis and ultimate analysis of carbonaceous catalyst proximate analysis/ w %, ad

Ultimate analysis/ w %, daf

Sample M

A

V

C

H

N

S

O(diff)

DT Coal

3.10

12.20

29.00

87.70

4.96

1.27

0.42

5.36

C- catalyst

1.20

13.30

4.50

94.60

1.47

0.99

0.17

2.36

3. Result and discussion 3.1 Effect of residence time and reaction temperature on CH4 and CO2 conversion The carbonaceous catalyst activity was investigated at different temperatures range from 1223

K to 1323K with CH4/CO2=1. The residence time was varied by changing the feed gases flow. The effect of residence time on CH4 and CO2 conversion at different temperatures is shown in Fig. 2. 100

100

a

90

CO2 Conversion (%)

CH4 Conversion (%)

90

1223K 1273K 1323K

b

1223K 1273K 1323K

80

70

60

50

80

70

60

40 50 0

5

10

15

20

25

30

35

0

5

10

Residence time (s)

Fig.2

15

20

25

30

35

Residence time(s)

Effect of residence time and reaction temperature on CH4 and CO2 conversion

As can be seen from Figure 2, with the increase of residence time the conversion of CH4 and CO2 have increased rapidly, then leveled off. When the residence time from 3s to 30s in reaction temperature 1223K, the conversion of CH4 can be increased from 40% to 83% , but the conversion of CO2 can be increased from 60% to 93% . The result showed that the conversion of CO2 was significantly higher than the conversion of CH4 at the same condition of residence time and reaction temperature, which indicate that the gasification reaction of the carbonaceous catalyst and carbon dioxide was occurred during the reforming process. Increase temperature is in favor of CH4-CO2 reforming reaction, which promotes the conversion of CH4 and CO2.When the reaction temperature arrived 1323K under the residence time 30s, the conversion of CH4 and CO2 are about 92.8% and 98%. And the temperature is main factor affecting reaction rate 2.2.

Mechanism The CO2 reforming of CH4 on carbonaceous catalyst is not a independent reaction, duing

reforming of CO2 - CH4 the following reaction may take place The Carbon dioxide reforming of methane Carbon dioxide gasification

C + CO 2 → 2 CO

Methane decomposition

CH

4

CH

→ C + 2H

4

2

+ CO

2

→ 2 CO + 2 H 2

Water-gas shift reaction

H

2

+ CO

→ CO + H 2 O

2

The Carbon dioxide reforming of methane is the dominating reaction, carbon dioxide gasification, carbon dioxide gasification and methane decomposition are Subordinate reaction。 On the condition of carbonaceous catalyst，the oxygen-bearing functional group is the active center of CO2 reforming of CH4 , which could change the reaction course and reduce the activation energy . After a thorough analysis of experimental data and literature data [10-11], reaction mechanism of CO2 reforming of CH4 over carbonaceous catalyst as follows: CH

→ C

* 4

x [ CH

(1 +

+ * →

4

y )[ C O

2

+ 2H 2]

（2）

+ * → C O *2 ]

（3）

+ * →

y[C

*

*

C

（4）

]

y [ CO

* 2

+ C

2 x[ H

2

+ 2* → 2 H

* 2

CO

+ H

(1

−

CO

*

→

→ CO

*

+ OH

*

x

)[ C

*

→ CO + *

H

* 4

*

2 CO

*

( 4 x − 1 )[ H

（1）

* 4

CH

+

*

*

（5）

]

]

+ OH

（6） *

（7）

→ H 2O + 2 * ] O H

*

→

（8）

C O

*

+

5 2

H

2

+ * ]

（9） （10）

Where * denotes the active site on the carbonaceous catalyst Which corresponds to the over reaction stoichiometry： CH

4

+ (1 + y

)C O

2

+ yC →

4 -x -z H 2

2

x ⎞ ⎛ + 2 ⎜1 + y⎟ C O + zH 2O + xC 2 ⎠ ⎝

2.3. Kinetic model The CO2 reforming of CH4 is very complex reaction course, reaction mechanism is not the same with the different types of catalyst, the kinetic model was created in conversion of CH4 . The basic assumption of Kinetic model are as follows： ⅰ

The CO2 reforming of CH4 is the dominating reaction;

ⅱ The reaction temperature and content in feed of CH4 and CO2 were considered, nothing to do with the intermediate . Molar content in feed and concentration distribution of CH4 and CO2 are same in CO2 reforming

ⅲ

of CH4； The conversion of CH4 defined as: x

CH

4

= 1 −

[ CH [ CH

]

4

]

4

（11）

0

Where [CH 4 ]0 is content of CH4 in feed， [CH 4 ] is content of CH4 in outlet . The global rate for the transformation of CO2 reforming of CH4 can be expressed as follows: d [ CH dt

−

4

]

= k [ CH

4

] a [ CO

2

（12）

]b

where k is reaction rate constant; a and b are the reaction orders for CH4 and CO2, respectively.； Taking into account Eqs. (11) and (12) the global rate for CH4 conversion is represented in Eq. (13) d[CH4 ]0 (1 − xCH4 ) dt

(

= k[CH4 ]0a+b 1 − xCH4

)

( a +b )

（13）

Eq. (13) integral obtained： x CH

4

= 1 − [1 +（ n - 1） k [ CH

4

]

n −1 0

t]

1 1− n

Where，n=a+b is the overall reaction order. Using some basic mathematic models, after adjusting the experiment data, the kinetic equations are created in origin multi-parameter curve fitting. Fig.3 shows curve fitting of different reaction temperatures. The result as Fig.3 shown，the coefficient of multiple determination (R2)was above 0.92 , indicating the goodness of fit . Assuming an Arrhenius temperature dependence, the reaction rate constant can be expressed as：

k = A exp（ −

Ea ） RT

where A is the pre-exponential factor, Ea is the apparent activation energy (J mol−1), R is the universal gas constant (8.314 J mol−1 K−1) and T is the temperature (K).

1.0

0.8

C H 4 C on version

0.6

k=0.03481 n=2.04865 2

R =0.92809

0.4

k=0.05394 n=2.11935

0.6

2

R =0.98792 0.4

0.2 10

20

0

30

10

20

Resident time (s)

Resident time (s)

a T=1223k

b T=1273k 1.0

CH4 Conversion

0

0.8

k=0.07866 n=2.06985

0.6

2

R =0.98242

0.4 0

10

20

30

Resident time (s)

c Fig.3

T=1323k

kinetic curves of methane conversion of different temperatures

-2.5

lnK

C H 4 C o n v e r sio n

0.8

-3.0

-3.5

0.000090

0.000093

0.000096

1/RT

Fig.4 The lnk vs.1/RT relation

0.00009

30

The kinetic parameters were estimated by linear regression of the experimental data using the least squared method , the pre-exponential factor A and the apparent activation energy Ea are obtained , Fig.4 shows the lnk vs.1/RT relation . As can be seen from Fig.4，the activation energy and the pre-exponential factor were 109.9kJ/mol and 1765min·L/mol respectively.

k = 1765 exp（ −

109.9 ± 10.1KJ ） RT

2.4 Comparison between calculation data and experimental data In order to compare between calculation data and experimental data , the conversion of CH4 and CO2 were calculated in different residence time and reaction temperatures. Fig.5 shows a comparison between calculation data and experimental data of conversion of CH4 and CO2. 90

Measured Calculated

80

Experiment Calculation

90

CH4 Conversion(%)

CH4 Conversion(%)

80 70

60

50

40

30

70

60

50

40 0

5

10

15

20

25

30

35

0

5

Residence time(s)

15

20

25

30

35

25

30

35

Residence time(s) 100

100

Experiment

Measured Calculated

Calculation

90

CO2 Conversion (%)

90

CH4 Conversion(%)

10

80

70

60

80

70

60

50 50 0

5

10

15

20

Residence time(s)

25

30

40 0

5

10

15

20

Residence time (s)

100

100

Experiment Calculation

Measured calculated

90

CO2 Conversion (%)

CO2 Conversion (%)

90

80

70

60

50

80

70

60

40

50 30 0

5

10

15

20

25

30

35

Residence time (s)

0

5

10

15

20

25

30

Residence time (s)

Fig.5 Comparison between calculation and experiment of CO2 reforming of CH4 in different temperature

The result as Fig.5 shown，the calculation data of CH4 conversion are close to the experimental data, the maximum error of which is 7.3%. which illustrates the rationality of the kinetic model. However, the calculation data of CO2 conversion of the simulation had a gap when compared to the experimental data and the maximum error of which is up to 20%, one of the reasons for the error is caused by dynamic equation which is the established in an ideal condition, and the experiment conditions were assumed, which ignored the effect of single particle in the diffusion and mass transfer on reforming reaction, the second reasons for the error is caused by the gasification reaction between carbonaceous catalyst and carbon dioxide.

4. Conclusion （1）With the residence time and reaction temperatures increasing , the conversion of CH4 and CO2 increases ,

and the temperature is main factor affecting reaction rate .

（2）the conversion of CO2 was significantly higher than the conversion of CH4, which indicate that the gasification reaction of the carbonaceous catalyst and carbon dioxide was occurred during the reforming process. （3）The kinetics of the CO2 reforming of CH4 on carbonaceous catalyst was described in a rate law, the experimental data was analyzed in non-linear regression. The apparent activation energy Ea and the pre-exponential factor A were solved. The rate constant was as as follows:

k = 1765 exp（ −

109.9 ± 10.1KJ ） A comparison is made between calculation data RT

and experimental data of the CH4 conversion, which illustrates the rationality of the kinetic model.

Acknowledgements This work was supported by the National Basic Research Program of China (2005CB221202) and Shanxi Provincial Natural Science Foundation (2010011014-1).

References [1] Zhang Yongfa. A new technology of three products from coal based on a L T integrated apparatus [C]. Proceedings of Workshop on Coal Gasification for Clean and Secure Energy for China. Beijing, 2003: 235-246. [2] Zhang Yongfa et al. patent , ZL 031527973, 2005. [3] I. Suelves, M.J. Lázaro, R. Moliner et al. Hydrogen production by methane decarbonization: Carbonaceous catalysts[J]. Hydrogen Energy. 32 (2007) 3320 – 3326 [4] Seok Seung-Ho, Han Sung Hwan, Lee Jae Sung. The role of MnO in Ni/MnO-Al2O3 catalysts for carbon dioxide reforming of methane[J]. Applied Catalysis A: General, 2001, 215 (1 /2) : 31-38. [5] NowosielskaM, JozwiakW K, Rynkowski J. Physicochemical characterization of Al2O3 supported Ni– Rh systems and their catalytic performance in CH4 /CO2 reforming [J]. Catal. Lett, 2009, 128: 83-93. [6] Guo Hai - jun, Yang Min, Zhou Le et al. Progress in methane catalytic reforming with carbon diox ide to synthesis gas[J]. Guangzhou Chemical Industry. 2009, 37(5): 2-5. [7] Li Yan-bing , Xiao Rui , Jin Bao-sheng et al. Carbon dioxide reforming of Methane to produce syngas with Coal Char[J]. Journal of Combustion Science and Technology. 2009, 15(3): 238-242 [8] Guojie Zhang, Yue Dong , Meirong Feng et al. CO2 reforming of CH4 in coke oven gas to syngas over coal char catalyst[J]. Chemical Engineering Journal. 2010, 156: 519-523. [9] Zhang Huawei. Experimentalstudy ofoven gas reforming to producesynthesis gas over the coal char holding in the reactor [D]. Taiyuan university of technology , 2005. [10]Nandini A, Pant K K, Dhingra S C. Kinetic study of the catalytic carbon dioxide reforming of methane to synthesis gas over Ni - K/CeO2 -Al2O3 catalyst[J]. App l. Catal. A, 2006, 308: 119 - 127. [11]A. Nandini, K.K. Pant , S.C. Dhingra. Kinetic study of the catalytic carbon dioxide reforming of methane to synthesis gas over Ni-K/CeO2-Al2O3 catalyst[J]. Applied Catalysis A: General, 2006,308: 119–127.

2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 CARBON MANAGEMENT:

A Study on the absorption characteristics of CO2 with a vortex tube type absorber KEUN-HEE HAN, WOO-JUNG RYU, JONG-HO PARK*, WON-KIL CHOI, JONG-SUB LEE, BYOUNG-MOO MIN Korea Institute of Energy Research, Jungang Vacuum Co. Ltd* 102, Gajung-ro, Yuseong-gu, Daejeon 305-343, KOREA [email protected] Abstracts : In this study, the CO2 removal characteristics of the vortex tube type absorption apparatus were investigated to enhance the compactness of CO2 absorption process and to reduce the amount of absorbing solution for the process. The vortex tube with the diameter of 12 mm and the length of 200 mm was introduced in the experimental apparatus 3

to treat 20 Nm /hr of CO2 containing flue gas. The flue gases for the experiments containing 11-15 vol% of CO2 were supplied from the coal-firing CFBC power plant with 12 ton/hr of steam producing capacity. EDA and DETA based on MEA were used as the absorbing solution. The absorption experiments were executed under the various conditions like the absorbing solution concentrations in the 30 wt%, the flow rate of CO2 containing 3

3

flue gases in the range of 6 to 12 m /hr and the flow rate of absorbing solution in the 0.18m /hr and the range of operation pressure in the 1.4 to 5.5 kgf/cm2. As a results, the CO2 removal efficiency increases with the operation pressure but deceases with the flow rate of flue gas. However, the development of an additional process to improve the efficiency of CO2 absorption is required. Key words: Vortex Tube; Absorption; Flue Gas; Carbon Dioxide; Alkanol Amine; 1. Introduction One of the major known causes of global warming is the CO2 emitted by the use of fossil energy. CO2 is mostly emitted by the steel making, chemical, cement and power plant industries, which form the mainstay of the national economy. In particular, coal-fired thermal power plants generate a significant amount of CO2 and have a major impact on the global environment. Measures aimed at restricting the emission of CO2 include reducing energy consumption, increasing the efficiency of equipment, using more natural sources of energy, and retrieving emitted CO2. Many technologies and processes are being developed to remove the emitted CO2 while ensuring the stable operation of existing energy production facilities. All such measures pursue the effectiveness, low cost and high efficiency. The CO2 separation processes can be divided into absorption, adsorption, cryogenics (distillation), oxygen enriched combustion, and membrane separation, and etc. The most suitable process for recovering the emitted CO2 from large-scale sources such as thermal power plants is the chemical absorption method, which is appropriate when the CO2 concentration and pressure in the flue gas are relatively low1,2). Alkanol amines are popularly used as the absorbent. A vortex tube with a simple structure can separate the compressed gas into low temperature portion and high temperature portion using the vortex generated when high pressure gas is sprayed into a tube

3,4)

. When a gas/liquid

mixture is passed through a tube at high speed, the gas and liquid can be separated by using the difference in gravity

- 1 -

5)

caused by the centrifugal force of the mixture . The mixture is formed of micro droplets as it is moved at high speed. Its principal benefit is the processing treatable 6)

a large volume of mixed gas from a small device . Using such characteristics, it can be applied to CO2 absorption from the flue gas. The most important factor of CO2 absorption from the flue gas is the contact time between the flue gas and the absorbent. This study used MEA (Monoethanolamine), EDA (Ethylenediamine), and DETA (Diethylenetriamine), which are alkanol amine type absorbents. The absorption characteristics of CO2 in the flue gas generated by the burning of coal were invetigated by varying the flue gas flow rate, the aqueous solution flow rate and the operating pressure of the vortex tube type absorber. 2. Experimental equipment and procedure Fig.1 shows the equipment used for this study. In a vortex tube, the compressed gas is needed to create a vortex. 2

3

For this purpose, the flue gases were charged to the flue gas storage tank (pressure: 5~6 kgf/m , volume:0.75m ) using a compressor. Flue gases were generated by combustion of coal in the circulating fluidized bed combustor (CFBC). The vortex tube for CO2 absorption composed of 12 mm in diameter (ID) and 200 mm in length(L). Mixing of the flue gas and the aqueous solution was controlled in the spray chamber before supplying to the Gas

vortex tube. As the flue gas and the aqueous

CO2 Analyzer Condenser

solution were sprayed into the vortex tube, the

T P

T P

Regulrator

MFC

CO2 in the flue gas was absorbed through the vortex flow. A CO2 analyzer (Horiba, VA-3000) P

was installed in the gas exit line of the vortex

Vortex Tube

tube

Liquid

reactor

to

measure

the

change

in

the

concentration of CO2, which was monitored and

Pump

recorded in a personal computer. Flue gas Rich Absorbent

Fresh Absorbent

Flue Gas Storage Tank

The experimental conditions are described in Table 1. MEA, EDA and DTEA were used as the

Compressor

absorbent. The aqueous solution was supplied at Fig. 1 Schematic Diagram of Vortex Tube System for CO2

3.0 liter/min as a fixed value, while the flue gas

Absorption

flow rate and pressure of the tube in which the absorption occurs were varied. The flue gas and the aqueous solution were flowed through the tube at the end of which the liquid/gas separation

Table 1. Experimental Conditions

occurs by centrifugal force due to the difference in Parameter

Conditions

gravity. The concentration of CO2 was measured

A kind of Absorbent

MEA, EDA, DETA

for monitoring the CO2 removal at the exit of the

Concentration of Absorbent (wt %)

30

flue gas. The gas analyzer measured the reduction

3

Flow Rate of Absorbent (m /hr) 3

6.0~12.0

2

1.4~5.5

Flow Rate of Flue Gas (m /hr) Operation Pressure (kgf/cm )

in the concentration of CO2 as the absorption took

0.18

place. The equilibrium was assumed when there was no further change in the concentration of CO2. The removal efficiency and CO2 loading were calculated based on the equilibrium state

assumed.

- 2 -

3. Results and Discussion The removal efficiency and CO2 loading of the vortex tube equipment can be expressed by the following equations.

             ×  

              

  

 ×    × 

 ×     × 

            ×    ×         

             ×    ×         



here  : CO2 removal efficiency, %

 : CO2 loading, mole/mole_solution

    : mole of CO2 in flue gas at inlet, -

   : mole of CO2 in flue gas at outlet, -

  : mole of aqueous solution,  : flow rate of inlet flue gas, ℓ/min,

  : flow rate of out gas, ℓ/min

   : CO2 mole fraction of flue gas, -

    : CO2 mole fraction of out gas, -

2

  : absorber pressure, kgf/cm

  : outlet pressure, kgf/cm2

  : temperature of absorber, ℃

  : temperature of out gas, ℃

Absorbents Fig. 2 and 3 shows the effect of G/L ratio on CO2 loading and the removal efficiency. The concentrations of aqueous solutions are 30 wt% in the water. The gas/liquid ratio was varied to observe CO2 loading and removal 2

efficiency at the operating pressure of 1.4 kgf/cm . As the G/L ratio increased, CO2 loading also increased while the removal efficiency decreased. The reason for the increase of CO2 loading was because the vortex velocity increased with the G/L ratio. The higher vortex velocity makes the droplets of the aqueous solution smaller which increases the 7)

interfacial area to react with CO2 in the flue gas per unit volume . Furthermore, the reason for the removal efficiency decrease was the higher gas flow per unit volume which increases the concentration of CO2 to be absorbed. Different CO2 loadings were observed for different aqueous solutions. In the molecular structure of alkanol amine, the capacity to absorb CO2 depends on the number of amine groups formed. In the case of DETA, the structure of 2-NH2 and 1-NH has more amine groups to absorb CO2 than that of MEA.

Fig. 2 Effect of G/L ratio on CO2 loading for each

Fig. 3 Effect of G/L ratio on CO2 removal efficiency

absorbent

for each absorbent

- 3 -

Operating Pressure Fig. 4 and 5 show the effect of operation pressure on CO2 removal efficiency and the CO2 loading. The operating pressure was adjusted by using the opening of the throttle valve at the end of the vortex tube absorber. The pressure of the flue gas ranged from 1.4~5.5 kgf/cm2. The aqueous solution was 30 wt% MEA in the water. As the operating pressure increases, the CO2 removal efficiency and CO2 loading per mole of absorbent increased. The higher pressure 7)

means higher CO2 partial pressure and higher CO2 partial pressure increases the absorbing capability . In addition, at higher operating pressure, the absorbent may produce the physical absorption by the forced hydrolysis of CO2 in the water and amine.

Fig. 4 Effect of Operation Pressure on CO2 Removal

Fig. 5 Effect of Operation Pressure on CO2 Loading at

Efficiency at G/L Ratio.

G/L Ratio.

Absorption rate Fig. 6 shows the absorption rates of three absorbents (MEA: 4.91mol/ℓ, EDA: 4.90mol/ℓ, DETA: 2.90mol/ℓ) with the same concentration 30 wt% at 30℃. The absorption rate is measured by the time between the CO2 in the flue gas gets in contact with the sprayed aqueous solution to begin absorption and the CO2 does not increase any longer. The difference in absorption rates is due to the number of amine groups in the absorbent. EDA, which has 2-NH2 that can quickly absorb CO2 was the fastest. It also showed the highest CO2 removal. If the absorbents with the same number of moles are used, DETA which has one more amine group (NH) than EDA would have shown the highest removal efficiency. Fig. 6 Effect of CO2 Concentration Change on Operation time

4. Conclusion The absorption rate was in the order of EDA, DETA and

MEA. EDA, which had 2-NH2, showed higher rate than MEA, which had 1-NH2. DETA, which had 2-NH2 and 1-NH, showed similar rate to EDA but had higher CO2 loading.

- 4 -

As the flue gas flow rate increased, CO2 loading increased and removal efficiency decreased. That is because the higher flue gas flow rate caused the droplets of the absorbent to be smaller as the vortex velocity increased, and the smaller droplets increased the contact surface. As the operating pressure increased, the CO2 loading and removal efficiency increased. Higher pressure means higher CO2 partial pressure and that increased the number of CO2 moles that can be absorbed per unit mole of absorbent. Furthermore, increased pressure enables the physical absorption due to the forced hydrolysis of the water not reacting with CO2 and amine. Reference 1. Chakma A., and Tontiwachwuthikul P., “Designer Solvents for Energy Efficient CO2. Separation from Flue Gas Streams”, Greenhouse Gas Control Technologies, 35~42(1999). 2. Han K. H. et al., "A Study on the CO2 Removal Efficiency with Aqueous MEA and Blended Solutions in a Vortex Tube Type Absorber", Korean Chem. Eng. Res., 47(6), 197~202(2009). 3. Rangue, G. J., "Method and Apparatus for Obtaining from Fluid under Pressure Two Currents of Fluids at Different Temperatures", US Patent No. 1,952,281 (1934). 4. Hilsch R., “The Use of Expansion of Gases in a Centrifugal Field as a Cooling Process”, The Review of Scientific Instruments, Vol. 18, No. 2, pp. 108~113(1947). 5. Eiamsa, S. and Promvonge, P., "Review of Ranque-Hilsch Effect in Vortex Tube", Renew. Sust. Energ. Rev., 12, 1822~1842(2008). 6. Han K. H. et al., "Absorption Equilibrium of CO2 in the Sterical Hindered Amine, AMP Aqueous Solution", Korean Chem. Eng. Res., 45(2), 795~800(2007). 7. A. Chakma et al, "Absorption of CO2 by aqueous triethanolamine(TEA) solutions in a high shear jet absorber", Gas Separation & Purification, Vol. 3, June (1989).

- 5 -

Manuscript Not AVAILABLE

2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission

PROGRAM TOPIC: CARBON MANAGEMENT Pd-FREE COMPOSITE MEMBRANE FOR PRE-COMBUSTION CAPTURE Jung Hoon Park, PhD., Korea Institute of Energy Research [email protected], Tel +82-42-860-3766, Fax +82-42-860-3134 71–2, Jang–dong, Yuseong–gu, Daejeon, 305–343, Republic of Korea Sung Il Jeon, Master, Korea Institute of Energy Research [email protected], Tel +82-42-860-3796, Fax +82-42-860-3134 71–2, Jang–dong, Yuseong–gu, Daejeon, 305–343, Republic of Korea Young Jong Choi, Master, Innowill Corporation [email protected], Tel +82-42-862-7500, Fax +82-42-862-6500 533,Yongsan–dong, Yuseong–gu, Daejeon, 305–343, Republic of Korea

Abstract: Pre-combustion CO2 capture technology is recently focused on one of reduction methods of carbon dioxide from power generation system in view of environmental (carbon management) and sustainable (Hydrogen economy) point. Separation of hydrogen from water-gas shift reactors through dense hydrogen transport membranes, while retaining CO 2 produces essentially pure hydrogen in the permeate and CO2 at high pressure and high concentration in retentate, which is ideal for efficient sequestration of CO2. Moreover, the combination of a hydrogen selective membrane with a water-gas shift catalyst in a single reactor would allow a high degree of CO conversion, despite a low equilibrium constant at high temperature, due to the continuous depletion of H 2. This equilibrium shift can provide more hydrogen productivity, higher concentration of CO 2 and lower impurity such as CO. Nowadays, many researches have been studied on various membranes for low cost separation of hydrogen. Pd and Pd alloy membranes have been studied for separating hydrogen from pre-combustion capture process but there is limitation for large scale application without reduction of Pd layer and improvement of membrane stability, due to the price and embrittlement of noble metal. Recently, Pd-free membranes using V, Ta, ans Nb which has higher permeability have been researched to improve its high cost and low stability. In this work, metal alloy and composite membranes have been developed to separate hydrogen from mixed model gases, particularly product streams generated during coal gasification and/or water gas shift reaction. The powder mixture for fabricating the cermet membranes was prepared by mechanically mixing 60 vol.% vanadium with Y 2O3-stabilized ZrO2 (YSZ). The powder mixture was pressed into disks, which were then sintered in vacuum at 1600 ℃ for 2 h. As-sintered membrane was dense and mounted to a stainless steel ring with brazing filler. Hydrogen fluxes of V/YSZ membrane have been measured in the range of 200~350℃ with 100% H2. The crack was formed in the both sides of membrane at 350 ℃ and pressure of 0.5 bar. During permeation experiment, vanadium of V/YSZ membrane reacted with hydrogen to form V2H which was the origin of crack formation. To improve the membrane stability, we prepared metal alloy membranes with vanadium. Hydrogen fluxes of metal alloy membranes have been measured in the range of 350~500℃ with 100% H2 and the mixture gas of H2 and He (or CO2). In addition, the stability of membrane was investigated according to operating temperature and hydrogen partial pressure.

Manuscript Not AVAILABLE

2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission

PROGRAM TOPIC: CARBON MANAGEMENT COMPOSITE CERAMIC MEMBRANE FOR OXYGEN SEPARATION Jung Hoon Park, PhD., Korea Institute of Energy Research [email protected], Tel +82-42-860-3766, Fax +82-42-860-3134 71–2, Jang–dong, Yuseong–gu, Daejeon, 305–343, Republic of Korea Jong Pyo Kim, Master, Chungnam National University [email protected], Tel +82-42-862-4200, Fax +82-42-862-6500 220, Gung-dong, Yuseong-gu, Daejeon 305-764, Republic of Korea Soo Hwan Son, Master, Korea Institute of Energy Research [email protected], Tel +82-42-860-3796, Fax +82-42-860-3134 71–2, Jang–dong, Yuseong–gu, Daejeon, 305–343, Republic of Korea

Abstract: Oxygen and nitrogen are used in many industrial processes. Pure oxygen is used in the production of metals and in the integrated gasification combined cycle (IGCC) for partial oxidation. Recently, oxy-fuel combustion CO2 capture process as one of carbon capture technologies has been developed and this process also needs large scale oxygen separation unit. It has been predicted that the total market would grow significantly if pure oxygen could be produced at lower cost. The technology used for commercial separation of oxygen varies according to the scale and requirements for oxygen purity. For example, the cryogenic distillation method that was started in 1902 is used for large-scale production of pure oxygen, and the simultaneous production of nitrogen, argon and helium. However, high investment costs and energy consumption make it difficult to integrate this process with other power generation. Over the last decade, membrane technology for gas separation has developed rapidly. The interest in dense ceramic membranes for the transport of oxygen has grown considerably. As a result, the knowledge of the intrinsic properties of the membrane materials is now overwhelming, and new reports are being published frequently. The use of a dense mixed-conducting, perovskite-type ceramic membrane is a new technology for the production of pure oxygen. An obvious advantage of perovskite membranes is their 100% selectivity for the permeation of oxygen. However, ceramic membrane process has some drawbacks; the crack formation of membrane under high pressure and temperature condition, low stability of perovskite structure under ambient air, the difficulty of heat exchange integration with flue gas. In this work, the effect of carbon dioxide in ambient air was studied using the composite membrane with La0.6Sr0.4Ti0.3Fe0.7O3-δ (coating layer, denoted as LSTF-6437) and Ba0.5Sr0.5Co0.8Fe0.2O3−δ (bulk permeation body, denoted as BSCF-5582). BSCF-5582 and LSTF-6437 powder have been synthesized using polymerized complex method. BSCF-5582 powders were compressed into disks of 20 mm in diameter and 1.0 ~ 2.0 mm of thickness in a stainless steel mold under a hydraulic load by unilateral press. The green disk sintered at 1353 K for 5 hr. The sintered disk was polished to smooth the surface and to control the thickness of disk with 600 grit SiC. The surface of membrane was modified by coating of LSTF-6437 slurry. The optimum coating condition was evaluated according to coating time, rate and number. The coating membrane was sintered again to obtain composite dense membrane. The phase of the powder and the disk before and after sintering was characterized with an X-ray diffraction. Prior to oxygen permeation test, the cell part is purged with He gas to remove the air in permeation cell tube and to confirm sealing of the assembly for 20 hr. The leakage through membrane during oxygen permeation test was also measured for all runs at each temperature and the oxygen permeation fluxes were corrected on the basis of the measured leakage. Permeation study was performed in the temperature range of 750~950 ℃

and pressure of 1~3 atm with synthetic air (21 vol.% O2+79 vol.% N2) and ambient air model gases (CO2 300~700 ppm). Oxygen permeation flux was increased as temperature increased irrespective of membrane coating. In the case of LSTF coating membrane, it reached 1.9 ml/cm2·min at 950 ℃ exposed to flowing air (Ph =0.21 atm, feed side of membrane) and helium (Pl =10 -5 atm, permeated side of membrane). The oxygen permeation of BSCF-5582 membrane in the condition of air and CO2 (300~700 ppm) in feed stream decreased more than 43% in comparison with air feed stream while that of LSTF coating membrane maintained almost same flux irrespective of CO2 concentration.

Reactions of Coal Structures with Polymers Leading to Hydrogen Production P. Straka Institute of Rock Structure and Mechanics, v.v.i., Academy of Sciences of the Czech Republic, V Holešovičkách 41, 18209 Praha 8, Czech Republic E-mail address: [email protected]

Abstract Thermal reactions of polystyrene, acrylonitrile-butadiene-styrene and styrene-butadiene rubber with chosen coal fraction were investigated from point of view of hydrogen production. As method a two-stage copyrolysis of coal/polymer mixtures was selected. Experiments were carried out on pyrolysis laboratory unit with a vertical quartz reactor (the first stage) and a horizontal cracking oven (the second stage). Thus, coal/polymer mixtures contained 30 wt.% of polymer were heated and products further cracked. From the results, the process conditions leading to maximum hydrogen production were defined and the yields of products determined. If the mixtures with 30 wt.% of polymers are considered, the heating rate of 5 K/min and final temperature of 900oC at vertical reactor and 1200oC at horizontal cracking oven are sufficient for achievement of a hydrogen-rich gas with 77–79 vol.% H2. On the basis of solid-state NMR, FTIR and GC analyses of coal fraction and cokes, obtained tar, and obtained gas, respectively, reactions of coal structures with polymers leading to hydrogen production were described and, using the results of isoconversional analysis, discussed. Introduction Other than standard technologies, hydrogen can be obtained by non-traditional methods using waste polymers, in particular their mixtures. Studies have been published dealing with the transformation of polymer structures into gaseous and liquid products and solid carbon during the carbonization process [1], while others have been published on the possibility of utilizing the liquid products of pyrolysis as a substitute for fuels [2-4]. Possibility of hydrogen obtaining from the coal/waste-tyre mixture including the use of solid carbonaceous residue is discussed in the work [5]. Our preliminary study on the hydrogen distribution during copyrolyses of coal/polymers mixtures containing 15– 60 wt.% of polymer showed that a significant portion of hydrogen is distributed just in tar. Subsequently, the tar as a part of raw gas can be led into cracking oven where it will be decomposed into hydrogen and pyrolytic carbon. Therefore, a two-stage process involving the thermal decomposition of the mixture in the first stage and the subsequent splitting of liquid and gaseous products in the second stage can be used to obtain hydrogen. To control such a process should be identified the ongoing chemical reactions and determined the process conditions. The aim of this study is on the basis of solid-state NMR, FTIR and GC analyses of the used coal fraction and obtained products and kinetic evaluation describe reactions of coal structures with polystyrene polymers leading to hydrogen production, and the optimal process conditions determine. Experimental Experiments were carried out on pyrolysis laboratory unit with a vertical quartz reactor (an inner diameter of 6 cm and a length of 50 cm) as the first stage, and a horizontal oven for cracking of raw gas components in a quartz tube (with an inner diameter of 4.4 cm and a length of 60 cm) which worked as the second stage. 100 g of sample was pyrolysed/copyrolysed; coal/polymer mixtures contained 30 wt.% of polymer. As polymers, polystyrene (PS), grain size 3–4 mm, acrylonitrile-butadiene-styrene (ABS), grain size 2–3 mm, and styrene-butadiene rubber (SBR) with grain size -3 mm were tested, further, coal

1

fraction from the An Tai Bao coal (Shanxi, China) was used (apparent density of 1.30–1.35 g/cm3; for size analysis and proximate and ultimate analyses see Table 1). For description of Table 1 Parameters of the coal fraction used. Size analysis Proximate and ultimate analyses (wt.%) Grain size Share (wt.%) water 2.85 C (daf) 75.00 +3 0.45 ash (d.b.) 10.17 H (daf) 6.46 2–3 19.23 total S (d.b.) 1.70 N (daf) 0.45 1–2 44.81 VM (daf) 37.48 S+O(daf) 18.09 -1 35.50 Qs (daf) 32.20 (MJ/kg) reactions in question the coal fraction and obtained cokes were characterized by the solid-state 13 C CP/MAS NMR structural parameters; tars were analyzed by FTIR spectroscopy; gas components were during copyrolyses continuously controlled by infrared analyzers and finally determined by GC method. Further, kinetic measurements by TG/DSC method were carried out and parameters of Arrhenius plot determined. Finally, an isoconversional analysis of thermal degradation process of coal fraction and the mixtures in question by the Friedman, Ozawa, and Flynn and Wall methods [6] was performed. Results and Discussion In all the cases, the yield of tar from one-stage copyrolyses was significantly higher in comparison with that from pyrolysis of coal alone (Table 2). Moreover, at the one-stage Table 2 The yields of the products from one-stage pyrolysis of coal fraction alone and copyrolysis with 30 % of polymer (wt.%). Tar Reaction water Gas Losses Initial sample Coke Coal 67.5 6.2 11.8 11.9 2.6 Coal+PS 48.4 30.0 9.4 8.3 3.9 Coal+ABS 50.0 31.5 5.6 8.9 4.0 Coal+SBR 58.4 19.4 8.3 8.5 5.4 process the content of non-methanic hydrocarbons in obtained gas was higher with copyrolyses (4 vol.%, Table 3, coal+ABS and coal+SBR) as against pyrolysis (3 vol.%, Table 3). Due to this, the gas from two-stage copyrolyses contained always quite high amount of hydrogen (Table 4). Table 3 Parameters of gas from the one-stage pyrolysis and copyrolyses with polymers. Coal Coal+PS Coal+ABS Coal+SBR 3 3 3 vol.% Ndm vol.% Ndm vol.% Ndm vol.% Ndm3 H2 54.85 12.38 55.13 8.88 53.80 8.98 55.53 9.22 CH4 25.43 5.75 25.82 4.16 26.65 4.45 25.12 4.17 C2H4 0.43 0.10 0.52 0.08 0.77 0.13 0.78 0.13 C2H6 2.02 0.46 1.98 0.32 2.37 0.40 2.33 0.39 C3H6 0.30 0.07 0.30 0.05 0.48 0.08 0.49 0.08 C3H8 0.48 0.11 0.44 0.07 0.55 0.09 0.65 0.11 0.09 0.02 0.06 0.01 0.10 0.02 0.10 0.02 ΣC4 N2 1.76 0.40 1.57 0.25 1.65 0.28 2.16 0.36 CO 8.51 1.92 8.65 1.39 8.39 1.40 7.83 1.30 CO2 6.13 1.39 5.53 0.89 5.24 0.88 5.01 0.83

2

Table 4 Parameters of gas from the two-stage pyrolysis and copyrolyses with polymers. Coal Coal+PS Coal+ABS Coal+SBR 3 3 3 vol.% Ndm vol.% Ndm vol.% Ndm vol.% Ndm3 H2 69.43 30.83 77.32 40.52 78.71 42.66 77.82 41.09 CH4 3.64 1.62 4.80 2.52 5.52 2.99 5.34 2.82 C2H4 0.02 0.01 0.05 0.03 0.06 0.03 0.07 0.04 C2H6 0.01 0.00 0.01 0.00 0.02 0.01 0.01 0.01 C3H6 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 C3H8 0.00 0.00 0.00 0.00 0.01 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 ΣC4 N2 3.46 1.54 3.15 1.65 1.99 1.08 1.97 1.04 CO 18.83 8.36 12.83 6.72 12.00 6.50 13.16 6.95 CO2 4.61 2.05 1.84 0.96 1.69 0.92 1.62 0.90 From comparison with one-stage process it follows that the products and their yields from these two processes differ substantially. Due to cracking of raw gas components in the second stage as products of two-stage pyrolysis or copyrolysis the coke, glossy and amorphous carbons, and gas were formed, further, the reaction water was found as a minor product. Quantitatively, if coke, glossy and amorphous carbon are considered as a carbon material, this process is a two-products one as the yield of reaction water is low (Table 5), often lower than losses. Table 5 The yields of the products from two-stage pyrolysis of coal fraction alone and copyrolysis of coal with 30 % of polymer (wt.%). Initial sample Coke Glossy and amorphous Gas Reaction water Losses carbons Coal 69.7 4.1 20.2 5.6 0.4 Coal+PS 48.8 22.4 17.8 3.1 7.9 Coal+ABS 51.4 20.7 17.2 4.1 6.6 Coal+SBR 59.7 13.9 17.4 4.3 4.7 Presented results were achieved under the process conditions leading to maximum hydrogen production. If the mixtures with 30 wt.% of polymer are considered, the heating rate of 5 K/min and final temperature of 900oC at the first stage reactor and 1200oC at cracking oven in the second stage are sufficient for achievement of a hydrogen-rich gas with 77–79 vol.% H2 and the yields of the gas 17–18 wt.%. Probably, these conditions can be applied even if the polymers content is 15–60 wt.% [7]. For exploring of thermal decomposition of mixtures in question the dependences of both effective activation energy of decomposition (Eα) and the pre-exponencial factor on the extent of conversion (α) were followed. These dependences for coal alone and coal/polymer mixtures showed that in the case of coal alone the Eα is substantially higher than that of decomposition of coal/polymer mixtures. Generally, on heating, both polymers and coal degrade forming low molecular products; the degradation involves breaking of C–C bonds whose bond energy is around 350 kJ/mol and occurs above 200–300oC with polymers in question and above 300oC with coal. The effective activation energy varies throughout the process as at earlier stages the weak links are initiated and cleaved, but at later stages the stronger bonds are splitted. From TG/DSC curves in the key temperature range of 300–550oC the dependences of Eα on α were calculated and variability of Eα evaluated by applying an advanced isoconversional method [6]. The results are shown in Figs. 1–5. In the case of coal

3

alone, Eα 300–350 kJ/mol was found in the range of α 10–90 % (Fig. 1). Contrary, for coal/PS mixture only 200–250 kJ/mol (Fig. 2) and 180–290 kJ/mol for coal/ABS mixture (Fig. 3) was registered in the same range of α. It means that thermal degradation of mixtures with PS and ABS proceeds more easily in comparison with coal alone. The reason is that the mixtures contain substantially more aliphatic C–H bonds that are easier to cleave.

Fig. 1 The dependences of activation energy of decomposition (red line) and the preexponential factor (blue line) on the extent of conversion for thermal decomposition of coal fraction in the range of 300–550oC.

Fig. 2 The dependences of activation energy of decomposition (red line) and the preexponential factor (blue line) on the extent of conversion for thermal decomposition of coal/PS mixture in the range of 300–550oC. 4

Fig. 3 The dependences of activation energy of decomposition (red line) and the preexponential factor (blue line) on the extent of conversion for thermal decomposition of coal/ABS mixture in the range of 300–550oC. Another case is the coal/SBR mixture. Unlike PS and ABS, this mixture is decomposed in two temperature intervals, at 300–400oC and further at 400–550oC (Fig. 4).

Fig. 4 DTG curves of coal/SBR mixture at 3 heating rates. Due to this, with coal/SBR mixture two dependencies of Eα on α were followed. The first, for interval of 300–400oC, the second, 400–550oC were considered. In the first interval the weak bonds were splitted, thus, Eα varied between 150–100 kJ/mol and decreased with an increasing α (considered again between 10–90 %) (Fig. 5). In the second interval of 400– 550oC, the stronger bonds were cleaved. Therefore, in the range of α 10–70 %, Eα varied between 200–300 kJ/mol and increased with an increasing α (Fig. 6). In the range of α 70–90 % the Eα increased up to 700 kJ/mol as strong bonds were cleaved. In all the cases,

5

simultaneously with degradation the polycondensation reactions leading to new aromatic formations occurred. Therefore, in the range of α 90–100 % the Eα increased up to 420 or 700 or 2,400 kJ/mol together with pre-exponential factor. On the whole, in the temperature range of 300–550oC the prevailing Eα values were lower than those at decomposition of coal fraction alone. It means that splitting reactions in the mixtures proceed more easily than those with coal alone; due to this the hydrogen production is higher.

Fig. 5 The dependences of activation energy of decomposition (red line) and the preexponential factor (blue line) on the extent of conversion for thermal decomposition of coal/SBR mixture in the range of 300–400oC.

Fig. 6 The dependences of activation energy of decomposition (red line) and the preexponential factor (blue line) on the extent of conversion for thermal decomposition of coal/SBR mixture in the range of 400–550oC. 6

Further, tars from pyrolysis and copyrolyses were analyzed by FTIR method. No significant differences between tar from pyrolysis of coal fraction alone and copyrolyses with considered polymers were found (Fig. 7), but the aromaticity index of the tar from pyrolysis of coal (0.136) was higher in comparison with those for tars from copyrolyses (0.095 and 0.089 at PS/coal and SBR/coal, respectively). It means that the tars from copyrolyses contain more aliphatic C–H bonds than pyrolysis coal tar. This feature is favorable for hydrogen production in the second stage. A

B

C

Fig. 7 FTIR spectra of the tars from pyrolysis of coal fraction (A) and copyrolyses with SBR (B) and PS (C). Further, coal fraction and cokes resulting from pyrolysis and copyrolyses were investigated by solid-state 13C CP/MAS NMR method. The initial coal fraction had lower aromaticity (0.6259) and worse coking properties. Its 13C CP/MAS NMR spectrum is shown in Fig. 8. From spectrum analysis was found that a share of phenolic carbons from total carbons was of 8.22 %, alkylated aromatic carbons 9.3 %, protonated aromatic carbons 36.91 % and bridgehead carbons 6.92 %; further, the share of carbons in CH2 and CH3 groups was 26.90 % and quaternary carbons 6.69 %. From these data it follows that together with quite high content of hydrogen (Table 1) the coal fraction has the good preconditions for hydrogen production because of high amounts of protonated aromatic carbons and CH2 and CH3 groups. 13 C CP/MAS NMR parameters of resulting cokes are summarized in Table 6, spectra of cokes are shown in Figs. 9 (coke from coal fraction) and 10 (coke from coal/PS mixture). Spectra of cokes from coal/ABS and coal/SBR were quite similar to that from the coal/PS mixture. Table 6

13

C CP/MAS NMR parameters of cokes from pyrolysis of coal fraction alone and copyrolyses (%). protonated Bridgehead C Coke from phenolic C alkylated aromatic C aromatic C Coal fraction 3.66 3.98 46.43 23.76 Coal/PS 2.37 4.76 41.06 24.48 Coal/ABS 4.32 7.73 51.28 22.69 Coal/SBR 0.97 3.29 49.34 24.92

7

Fig. 8

13

C CP/MAS NMR spectrum of the coal fraction used.

Fig. 9

13

C CP/MAS NMR spectrum of coke from pyrolysis of coal fraction.

8

Fig. 10

13

C CP/MAS NMR spectrum of coke from copyrolysis of coal fraction with 30% PS.

From the data in Table 6 is obvious that in the case of cokes from coal/ABS and coal/SBR mixtures the share of protonated aromatic carbons is higher (with the coal/ABS mixture significantly higher) as compared with coke from coal fraction alone. It can be deduced that during copyrolysis the similar structures as at pyrolysis are formed but with a higher content of bound hydrogen. On the whole, during copyrolysis, polymers promote the creation of hydrogen-rich structures, consequently, together with coals with worse coking properties the waste polymers can be used for production of hydrogen. From presented data a following picture of reactions of polymers with coal can be outlined. In the one-stage copyrolysis, hydrogen is released during thermal splitting of C–H bonds, mainly from alicyclic rings, methylene and ethylene bridges and polymer chains, further, from aromatic rings during their polycondensation. Decomposition of styrene polymers yields e.g. styrene, ethyl benzene, toluene, isopropyl benzene, 1,3-diphenyl propane and α-methyl styrene which enrich the coal tar. Simultaneously, the new aromates arise by aromates-aromates reactions as alicyclic parts of coal clusters are cleaved during copyrolysis and formed aromatic structures react with aromatic products of decomposition of polystyrene polymers. Thus, essential constituents of copyrolysis tar are formed. In this way, a significant portion of hydrogen from coal and polymers is transferred into tar. Formed H2 molecules and aliphatic C1–C4 hydrocarbons are transferred into the gas. Nitrogen containing compounds are predominantly converted to ammonia, and sulfur containing structures to hydrogen sulfide. In the two-stage copyrolysis, the tar components, and gaseous hydrocarbons are thermally decomposed up to pyrolytic carbon and hydrogen. Oxygen containing compounds are splitted and elemental oxygen is generated, which reacts with carbon and hydrogen to carbon monoxide and reaction water. Through endothermic gasifying reactions between water and carbon, and through water-shift reaction with carbon monoxide, again hydrogen is produced. Finally, ammonia is splitted to hydrogen and nitrogen. On the whole, through twostage copyrolysis the polystyrene polymers with coal are converted to hydrogen-rich gas and pyrolytic carbon.

9

Acknowledgements The work was supported by the Czech Science Foundation (GAČR) as the project No. 105/07/1407 and conducted under the Institutional Research Plan AVOZ30460519, Academy of Sciences of the Czech Republic. References 1. Kidena K., Murata S., Nomura M.: Studies on the chemical structural change during carbonization process, Energy and Fuels 10, 1996, 672-678. 2. Ucar S., Karagoz S., Yanik J., Saglam M. Yuksel M.: Copyrolysis of scrap tires with waste lubricant oil, Fuel Processing Technology 87, 2005, 53-58. 3. Walendziewski J.: Engine fuel derived from waste plastics by thermal treatment, Fuel 81, 2002, 473-481. 4. Rodriguez I.M., Laresgoiti M.F., Cabrero M.A., Torres A., Chomon M.J., Caballero B.M.: Pyrolysis of scrap tyres, Fuel Processing Technology 72, 2001, 9-22. 5. Kříž V., Brožová Z., Přibyl O., Sýkorová I.: Possibility of obtaining hydrogen from coal/waste-tyre mixture, Fuel Processing Technology 89, 2008, 1069-1075. 6. Vyazovkin S., Sbirrazzuoli N.: Isoconversional Kinetic Analysis of Thermally Stimulated Processes in Polymers, Macromolecular Rapid Communications 27, 2006, 1515-1532. 7. Bičáková O., Straka P.: The resources and methods of hydrogen production, Acta Geodynamica et Geomaterialia 7, No.2(158), 2010, 175-188.

10

Manuscript Not AVAILABLE

2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Abstract Submission

COAL-DERIVED PRODUCTS:

COAL SUPPLY AGREEMENTS AND COMPETITION: Değer Boden Akalın, LLM Contact Information: Abdülhakhamit Cad. No:70 K:4 D:11Taksim,İstanbul, Turkey Tel: +90 212 253 09 28 Fax: +90 212 253 09 29 [email protected] www.boden-law.com

Abstract: Coal is used for many different purposes. It is primarily used as a solid fuel to produce electricity and heat. It can also be converted into liquid fuels such as gasoline or diesel. It can be used to produce synthetic natural gas, synthesis gas and hydrogen through gasification. Coal supply agreement is critical for the usage of coal for different purposes. Under a coal supply agreement, the supplier agrees to sell coals to a purchaser at pre-determined prices with minimum/maximum annual quantities. Duration, quality of coal, quantities, delivery, risk and ownership, price determination, termination and force majeure clauses are important clauses of a coal supply agreement. In drafting coal supply agreements antitrust laws should be reviewed. This paper aims to analyze how a coal supply agreement may violate competition.

CHEMICALS FROM TURKISH LIGNITES Vedat Mihladız TKI HISTORY Coal was used in Turkey to produce chemicals (ammonia, nitric acid and ammonium nitrate) since year of 1961 to 2000. For this purpose two type of gasifiers (both atmospheric) were used successfully; - fludize bed (Winkler) - entrained bed (Koppers-Totzek) TEST RESULTS Various coal types were tested in these gasifiers and the data collected from these tests were carried to date. Soma coal gasification test results are given in table 1 and table 2. Seyitömer lignites were used in both gasifiers and operation results are given for K-T gasifiers in table 3. Table 1. Soma coal gasification test results K-T GASIFIER

WINKLER GASIFIER

Consumption

Consumption 10.000

Kg/h

Coal

7.700

Kg/h

C

43,7

%

C

40.1

%

H

3,2

%

H

2.9

%

O

15.0

%

O

13.7

%

N

0,8

%

N

0.8

%

Coal

S

0,3

%

S

0.3

%

Ash

34.7

%

Ash

36.6

%

H2O

2.2

%

H2O

5.6

%

Oxygen (95%)

3.950

Nm3/h

Oxygen (95%)

2.300

Nm3/h

Quench water

2.500

kg/h

Steam

2.600

kg/h

8.000

Nm3/h

Products Gas

Products 11.750

Nm3/h

Gas

CO2

11.4

%

23.5

%

CO

57.8

%

33.0

%

H2

27.8

%

37.8

%

N2

2.8

%

2.7

%

O2

0.2

%

0.2

%

CH4

NA

2.8

%

70.4

%

CO+H2 Slag C in slag Flue ash C in flue ash

85.6 2.430 0 1.140 8.0

% kg/h % w/w Kg/h % w/w

CO+H2 Slag C in slag Flue ash C in flue ash

1.400 8,5 2.000 18.0

kg/h % w/w kg/h % w/w

Table 2. Seyitömer coal gasification test results Depending on these tests, formulations were created and now conceptual design of gasifiers and following process stages are being calculated (Table 3) to utilize in feasibility studies.

Table 3. Soma coal gasification calculation results K-T GASIFIER Consumption Coal

10.000

Kg/h

C

43.76

%

H

3.19

%

O

14.96

%

N

0.84

%

S

0.3

%

Ash

37.74

%

H2O

2.21

%

Oxygen (98%)

4,060

Nm3/h

Quench water

3,195

kg/h

10,123

Nm3/h

CO2

11.76

%

CO

64.90

%

H2

22.48

%

N2

0.2

%

H2S

0.66

%

CO+H2

87.38

%

Products Gas

K-T GASIFIER Consumption Coal

11.959

Kg/h

C

39.5

%

H

3.8

%

O

15.4

%

N

0.5

%

S

3.2

%

Ash

29.6

%

H2O

8.0

%

Oxygen (98%)

4.329

Nm3/h

Quench water

3.710

kg/h

Products 12.689

Nm3/h

CO2

12.10

%

CO

55.27

%

H2

28.87

%

N2

1.07

%

H2S

1.91

%

COS

0.20

CO+H2

85.6

%

1.238

kg/h

Gas

Slag

GASIFIER SELECTION Most important criteria in selection of gasifier type is ash melting point. This should be made under reducing atmosphere (CO+CO2). Atmospheric gasification can easily be made and problems of pressure gasification can be avoided. CONCEPTUAL DESIGN A block scheme for SNG production is given in figure 1 and consumption and production figures are given in Table 3. and same for the production of ammonia is given in figure 2 and table 4.

Figure 1. SNG production block scheme

COAL

GASIFIERS

HRSG

AIR SEPERATION

AIR

METHANE SYNTHESIS

ASH

METHANE

Table 4. SNG from coal Consumptions Coal Electric Fresh water

: 600 t/h : 132,000 kwh/h : 4,700 m3/h

Production Ammonia

: 800,000,000 Nm3/year

GAS WASHING COOLING

GAS PURIFICATIO N

CO2

H2S REMOVAL

CO CONVERSION

Figure 2. Ammonia production block scheme

COAL

GASIFIERS

HRSG

AIR SEPERATION

AIR

AMMONIA SYNTHESIS

ASH

AMMONIA

GAS WASHING COOLING

GAS PURIFICATIO N

H2S REMOVAL

CO CONVERSION

CO2

Table 5. Ammonia from coal Consumptions Coal Electric Fresh water Steam

: 100 t/h : 66,000 kwh/h : 1,030 m3/h : 60,000 kg/h

Production Ammonia

: 1,000 t/d

CONCLUSIONS Production of chemicals from coal is nowadays economic when it is compared with natural gas as feedstock. Ammonia production from coal was stopped 10 years ago because of high cost of production when compared to imported ammonia and to ammonia from natural gas. Now Turkey has no ammonia production ability with his natural resources. Care should be taken urgently to build an coal based ammonia plant to produce fertilizer and explosives.

Methane Cracking over De-ashed Coal Chars and the Effect of the De-ashing Conditions †‡

†

†

†

†

Ling Wei , Yisheng Tan* , Yizhuo Han , Hongjuan Xie , Jiantao Zhao , §

Jinhu Wu*, and Dongke Zhang †

Institute of Coal Chemistry, Chinese Academy of Sciences, Taiyuan 030001, China ‡

Graduate School of Chinese Academy of Sciences, Beijing 100039, China

* Qingdao Institute of Bioenergy and Bioprocessing Technology, Chinese Academy of Sciences, Qingdao 266101, PR China §

Centre for Energy (M473), The University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Australia

Abstract: Methane cracking over different de-ashed coal chars derived from the same parent coal (Xiao-long-tan lignite) was studied using a fixed-bed reactor operating at atmospheric pressure and 1123 K. The first set of samples of Xiao-long-tan lignite chars were prepared by heating the coal 1173 K in nitrogen for 30 minutes and then washing the resulting char with 5 N HCl (char 1), 5 N HCl and 29 N HF (char 2), respectively. The second set of samples of Xiao-long-tan lignite chars were prepared by washing the coal with 5 N HCl, 5 N HCl and 29 N HF, respectively and then pyrolyzing the de-ashed coal samples in nitrogen, also at 1173 K for 30 min (char 3 and char 4). Comparing to blank experiments using quartz particles, the chars were shown to have a significant catalytic effect on methane cracking. Hydrogen was the primary gas-phase product of methane cracking. Different chars also showed different catalytic activities in methane cracking. Chars 1 and 2 were shown to have higher catalytic activities than chars 3 and 4. These observations were also confirmed by additional experiments using a temperature-programmed desorption coupled with mass spectroscopy (TPD-MS), using methane, instead of nitrogen. Chars 1 and 2 were found to be more porous than chars 3 and 4, which is speculated as the major cause of the difference in their catalytic effect in methane cracking.

Keywords: methane cracking; coal char; de-ash; hydrogen

* Corresponding author: Prof. Tan, Tel.: +86-351-4044287; E-mail address: [email protected]

1. Introduction Methane cracking to produce hydrogen over a bed of coal char may offer an effective means for utilization of coal-bed methane (CBM) as coal and coal-bed methane are co-located and coal char has been observed to exert a catalytic effect on methane cracking 1

. There are several conventional methods of hydrogen production from methane such as

steam reforming of methane (SRM) and partial oxidation reforming (POR). SRM is a highly endothermic process and requires substantial external heating. The POR is an exothermic reaction and requires pure oxygen supply. Both SMR and POR produce large amounts of CO (and CO2) which has to be converted to CO2 and removed using the water-gas shift. The direct catalytic decomposition of methane has been considered as an attractive alternative. Compared with other reforming of methane, it is simple, energy-efficient and without the need of CO2 and CO removal. In methane cracking, high purity hydrogen can be produced and the produced clean carbon may also have a reasonable commercial value.8, 9 Methane can undergo thermal decomposition without a catalyst but this process requires a high operating temperature in the order of 1500-2000 K8, 10 . Catalytic decomposition can reduce the required temperature of methane cracking. Transition metals (such as Ni, Fe, Co, Pb) showed superior catalytic activity in decomposition of methane, but they can be rapidly deactivated owing to carbon from methane cracking accumulating on the catalyst surface. The catalysts have to be re-generated by burning off the deposited carbon.4, 8, 11, 12 The catalytic decomposition of methane and other hydrocarbons over carbon has been proposed as a viable method compared with the other methods, offering advantages including (i) high temperature resistance, (ii) no metal carbides formed, (iii) production of a marketable byproduct carbon which could substantially reduce the net cost of hydrogen production, and (iv) no need of regeneration of the carbon catalyst.10,

13

Muradov

investigated various kinds of carbon, such as activated carbon, carbon blacks, glassy carbon, graphite, diamond, carbon fibers and carbon nanotubes as the catalyzers for methane cracking9, 10, 13-15 . Our previous work has also found that the coal chars, rather than the ash, have a profound catalytic effect on methane cracking. In the reaction, methane decomposition over coal chars showed a reaction order of 0.5 and the activation energies from 89 to 105 kJ mol-1 which is much lower than the methane C-H bond

dissociation energy about 440 kJ mol-1,16, 17. In order to ascertain the catalytic effect of the different de-ashed coal chars and the process of methane decomposition, the present work studied methane cracking over de-ashed coal chars prepared by different de-ashing processes.

2. Experimental A Chinese coal (Xiao-long-tan lignite) was employed in this work. The Xiao-long-tan lignite char was prepared by devolatilising the air-dried Xiao-long-tan lignite in nitrogen at 1173 K for 30 minutes. In order to study the effect of ash removal on the effect of char on methane cracking, two sets of de-ashed char samples were prepared. The first set of samples of chars were prepared by heating the coal 1173 K in nitrogen for 30 minutes and then washing the resulting char with 5 N HCl (char 1), 5 N HCl and 29 N HF (char 2), respectively. The second set of samples of chars were prepared by washing the coal with 5 N HCl, 5 N HCl and 29 N HF, respectively and then pyrolyzing the de-ashed coal samples in nitrogen, also at 1173 K for 30 min (char 3 and char 4). The chars prepared were sieved to a size fraction of 0.355-0.63 mm for the experiment. The methane cracking experiments were carried out in a vertical quartz tube fixed bed reactor as shown in Figure 1. The reactor has a diameter of 25 mm and a height of 620 mm, housed in an electrically heated furnace. The flow rates of methane and nitrogen were controlled using mass flow controllers. The methane and nitrogen gases used in the experimentation were of high purity analytical grade gases. In a typical experimental run, a char of about 10 g was weighed and placed into the quartz tube and heated in nitrogen. The nitrogen was continuously flowed through the reactor at 160 ml min-1 while the furnace was heated to a desired temperature of 1123 K. When the temperature of the reactor reached the desired value and stabilized, the gas was switched to the reactant gas mixture of methane and nitrogen at a volumetric ratio of 1:4 (CH4:N2). The flow rate of nitrogen was also 160 ml min-1, but the flow rate of methane was 40 ml min-1. Allowing 3 min after the gas switching, gaseous samples of the exit stream were taken periodically and analyzed using a gas chromatograph (GC-14C) fitted with a molecular sieve column and a thermal conductivity detector (TCD), using Ar as the carrier gas for H2 measurement and H2 for hydrocarbons, respectively.

The methane adsorption and desorption on the char was also investigated using a temperature-programmed desorption coupled with mass spectroscopy (TPD-MS). The amount of char sample used in each experiment was ~0.1 g. The furnace was heated to 1073 K and then reduced to 373 K with Ar used as the protective gas. After 30 min, the Ar was replaced by mixture gas (CH4:N2=4:1). Then the char was heated in the gas mixture to 1073 K at 7 K min-1. The MS

measured the composition of the gases desorbed from

the char, which may contain H2, CH4, H2S, CO2, SO2 and so on. In addition, the superficial morphologies of the coal chars before and after methane decomposition were examined using a NOVA NANO, SEM 430 scanning electronic microscope (SEM), the samples were observed under an acceleration voltage of 20.0 kV. The amplificatory multiple is 5000. And the specific surface area and pore structure properties are determined using a TriStar 3000 physical adsorption apparatus using N2 sdsorption at 77 K. The conversion of methane can be calculated according to the following equation:

X CH 4 

VinCCH 4 ,in  Vout CCH 4 ,out VinCCH 4 ,in

and the hydrogen yield is expressed as:

YH 2 

Vout CH 2 ,out 2VinCCH 4 ,in

where “C” is the conversion of CH4 or H2 and the subscripts “in” and “out” refer to reactor inlet and outlet conditions, respectively.

Figure 1. A schematic diagram of the experimental system: (1) mass flow controllers, (2) mixing chamber, (3) temperature controller, (4) electrically heated furnace, (5) quartz tube reactor.

3. Results and discussion 3.1 Methane Decomposition over Different Coal Chars It is well known that methane molecules are highly stable and could only crack or decompose at high temperatures, usually greater than 1303 K.18 In a blank experiment, methane cracking on a bed of quartz particles of the same size fraction as the char was studied. The results are shown in Figure 2. The highest value of methane conversion achieved over quartz was less than 4.3 % at 1123 K. However, the methane conversion over the chars was observed to be higher than 70%, as also shown in Figure 2. Figure 2 also presented the results of methane cracking over different de-ashed coal char. It can be clearly seen that the methane conversions were different on the different coal char. The best coal chars were char 1and char 2 at the same temperatures. It proved that the ash of the coal char did not have a catalytic effect on methane cracking. Figure 3 displayed the results of hydrogen yield over the different de-ashed coal chars at the same temperatures. It can be seen that the hydrogen yield decreases with increasing the time in the stream. The trend of hydrogen yield was similar to that of methane conversion. The highest hydrogen yield over quartz was 0.13 %, but over the coal chars, about 52 %.

100

Xiao-long-tan lignite char char 1 char 2 char 3 char 4 quartz

CH4 conversion (%)

80

60

40

20

0 0

20

40

60

80

100

120

140

Time (min)

Figure 2. Methane conversion on the different coal char and quartz(T =1123 K, CH4:N2 = 1:4, total flow rate 200 ml/min)

100

Xiao-long-tan lignite char char 1 char 2 char 3 char 4 quartz

H2 yield (%)

80

60

40

20

0 0

20

40

60

80

100

120

140

Time (min)

Figure 3. Hydrogen yield on the different coal char and quartz(T =1123 K, CH4:N2 = 1:4, total flow rate 200 ml/min)

The methane decomposition on the coal char was also investigated using the temperature-programmed desorption coupled with mass spectroscopy. Figure 4 shows the results of methane cracking over the quartz in TPD-MS. The figure displays that the signals of H2, CO2, H2S and SO2 were about zero. The content of methane was still stable. The methane cannot crack over the quartz below 1073 K. The hydrogen signal on the MS was the same as that in methane decomposition in the fixed bed experiments. The methane started cracking from 850 K over the chars, which is much lower than the temperature of the methane cracking without a catalyst. So the coal chars are effective catalysts for methane decomposition. The signals of the hydrogen increased with increasing reaction temperature and the chars 1 and 2 provoked greater hydrogen yields. It proved that the best catalysts were char 1 and char 2. The results of the methane decomposition on the different de-ashed coal chars were shown in the Figure 5.

2.00E-009

H2

Gas singal

1.50E-009

CH4 H2S 1.00E-009

CO2 SO2

5.00E-010

0.00E+000 300

400

500

600

700

800

900

1000

1100

Temperature (K)

Figure 4. Gas signals of the methane cracking over the quartz on TPD-MS

1.20E-010

char 1 char 2 char 3 char 4

1.00E-010 8.00E-011

H2

6.00E-011 4.00E-011 2.00E-011 0.00E+000 300

400

500

600

700

800

900

1000

1100

Tempreature (K)

Figure 5. Hydrogen signals of methane cracking over the different coal chars in TPD-MS

Figure 6 shows the CO2 signals of methane cracking on the different coal chars. It can be seen that the carbon dioxide was produced at temperature above 500 K. With the temperature increasing, the content of carbon dioxide was firstly enhancive, and then moderative. After 1073 K, the content of carbon dioxide was about zero over the bed of the chars 1 and 2. But it was much more on the chars 3 and 4.

8.00E-010

CO2

6.00E-010

char 1 char 2 char 3 char 4

4.00E-010

2.00E-010

0.00E+000 300

400

500

600

700

800

900

1000

1100

Temperature (K)

Figure 6. Carbon dioxide signals of the methane caracking over the different coal char on TPD-MS

3.2 Char Characterization The above experimental results have demonstrated that the coal chars possess the catalytic effect, promoting methane cracking. At the same time, the results presented so far consistently show a decreasing trend with increasing reaction time. The phenomena indicated that the chars rapidly lose the catalytic activity in methane cracking. In order to speculate the mechanisms of the char deactivation, a series of scanning electron microscopy image analyses were performed on the chars before and after the reaction. Figure 7 shows the SEM images of different chars before and after the methane cracking. It is evident that the fresh chars have a very clear surface and clean porous structure with clear edges, but the deactivated char after methane cracking is covered with carbon deposits. The surfaces of the char 1 and char 2 have more porous structures than the surfaces of char 3 and char 4. The BET surface areas and pore characteristics of the fresh coal chars are detailed in Table 1. It can be seen that the total surface area decreased in the order from char 1 to char 4. The charge in the char surface area is consistent with the methane conversion over the char. It appears that the catalytic effect increases with increasing total surface area of the char and has no correlation with the micropore area. It is speculated that the total area surface is the main reason of the catalytic effect on methane cracking.

Figure 7. SEM images of chars before (A, B, C and D) and after (a, b, c and d) 123 min being subjected to methane cracking on the chars from the different de-ashed processes (N2:CH4 = 4:1, total flow rate 200 ml min-1, A and a: char1, B and b: char 2, C and c: char 3, and D and d: char 4) Table 1. Variations in the surface properties of the different coal chars

Sample BET Area, m2/g Micropore Area m2/g Total Volume, cm3/g Micropore Volume cm3/g Average Pore Diameter nm

Char 1

Char 2

Char 3

Char 4

Before reaction

Before reaction

Before reaction

Before reaction

44.6

38.5

26.6

14.9

16.5

12.5

22.15

4.95

0.048

0.050

0.0225

0.0205

0.007

0.006

0.01

0.002

4.2

5.2

3.3

5.4

4. Conclusions Methane cracking experiments have been performed over coal chars prepared using different de-ashing methods and quartz in a fixed-bed reactor at 1123 K and temperature-programmed desorption coupled with mass spectroscopy. It has been clearly demonstrated that the chars has a profound catalytic effect and methane does not crack on quartz. Methane conversion and hydrogen yield decrease with increasing reaction time, but increase with the reaction temperature. The best catalyst is char 1. The ash in the coal char is not responsible for the catalytic effect. Carbon deposition on the char surface

during methane cracking is responsible for the loss of catalytic activity of the char. The catalytic activity of the char in methane cracking is consistent with the total surface area of the char.

References 1.

Sun, Z.; Wu, J.; Zhang, D., CO2 and H2O gasification kinetics of a coal char in the presence of methane.

Energy & Fules 2008, 22, (4), 2160-2165. 2.

Haghighi, M.; Sun, Z.; Wu, J.; Bromly, J.; Wee, H.; Ng, E.; Wang, Y.; Zhang, D., On the reaction

mechanism of CO2 reforming of methane over a bed of coal char. Proceedings of the Combustion Institute 2007, 31, 1983-1990. 3.

Li, Y.; Xiao, R.; Jin, B.; Zhang, H., Experimental study of the reforming of methane with carbon dioxide

over coal char. International Journal of Chemical Reactor Engineering 2008, 6, 15. 4.

Chen, J.; He, M.; Wang, G.; Li, Y.; Zhu, Z., Production of hydrogen from methane decomposition using

nanosized carbon black as catalyst in a fluidized-bed reactor. Int. J. Hydrog. Energy 2009, 34, (24), 9730-9736. 5.

Song, Q.; Xiao, R.; Li, Y.; Shen, L., Catalytic carbon dioxide reforming of methane to synthesis gas over

activated carbon catalyst. Industrial & Engineering Chemistry Research 2008, 47, (13), 4349-4357. 6.

Xu, Z; Wu, J.; Wang, Y.; Zhang, D., Methane cracking over lignite char. Journal of Fuel Chemistry and

Technology 2009, 37, (3), 277-281. 7.

Sun, Z.; Wu, J.; Haghighi, M.; Bromly, J.; Ng, E.; Wee, H.; Wang, Y.; Zhang, D., Methane cracking over a

bituminous coal char. Energy & Fules 2007, 21, (3), 1601-1605. 8.

Yoon, S. H.; Park, N. K.; Lee, T. J.; Yoon, K. J.; Han, G. Y., Hydrogen production by thermocatalytic

decomposition of butane over a carbon black catalyst. Catal. Today 2009, 146, (1-2), 202-208. 9.

Muradov, N. Z.; Veziroglu, T. N., From hydrocarbon to hydrogen-carbon to hydrogen economy. Int. J.

Hydrog. Energy 2005, 30, (3), 225-237. 10. Lee, S. Y.; Ryu, B. H.; Han, G. Y.; Lee, T. J.; Yoon, K. J., Catalytic characteristics of specialty carbon blacks in decomposition of methane for hydrogen production. Carbon 2008, 46, (14), 1978-1986. 11. Gac, W.; Denis, A.; Borowiecki, T.; Kepinski, L., Methane decomposition over Ni-MgO-Al2O3 catalysts. Appl. Catal. A-Gen. 2009, 357, (2), 236-243. 12. Kim, M. H.; Lee, E. K.; Jun, J. H.; Kong, S. J.; Han, G. Y.; Lee, B. K.; Lee, T. J.; Yoon, K. J., Hydrogen production by catalytic decomposition of methane over activated carbons: kinetic study. Int. J. Hydrog. Energy 2004, 29, (2), 187-193. 13. Muradov, N.; Smith, F.; T-Raissi, A., Catalytic activity of carbons for methane decomposition reaction. Catal. Today 2005, 102, 225-233. 14. Muradov, N., Catalysis of methane decomposition over elemental carbon. Catal. Commun. 2001, 2, 89-94. 15. Muradov, N., Hydrogen via methane decomposition: an application for decarbonization of fossil fuels. Int. J. Hydrog. Energy 2001, 26, (11), 1165-1175. 16. Bai, Z.; Chen, H.; Li, W.; Li, B., Hydrogen production by methane decomposition over coal char. Int. J. Hydrog. Energy 2006, 31, (7), 899-905. 17. Zhang, Y.; Wu, J.; Zhang, D., Cracking of simulated oil refinery off-gas over a coal char, petroleum coke, and quartz. Energy & Fules 2008, 22, (2), 1142-1147. 18. Gueret, C.; Daroux, M.; Billaud, F., Methane pyrolysis: Thermodynamics. Chem. Eng. Sci. 1997, 52, (5), 815-827.

CeO2-K2O promoted Co-Mo sulfur-tolerant shift catalyst for the shift reaction of CO in coke oven gas Yuqiong ZHAO, Yongfa ZHANG () , Guojie ZHANG Key Laboratory of Coal Science and Technology of Shanxi Province and Ministry of Education, Taiyuan University of Technology, Taiyuan 030024, China Abstract To avoid energy and H2 consumption in the process of CH4-CO2 reforming (CH4 in coke oven gas and CO2 in gasification gas), a method was proposed, which through CO shift reaction makes CO in coke oven gas converted and simplifies the separation of CH4 and H2. In this work, CeO2-K2O promoted Co-Mo sulfur-tolerant shift catalyst for the shift reaction of CO in coke oven gas is investigated. The results indicate Ce and K have a synergy effect on promoting the catalytic activity, and the Co-Mo-Ce-K/γ-Al2O3 catalyst with 3.0 wt-% CeO2 and 6.0 wt-% K2O exhibits the highest activity. Moreover, CeO2 contained Co-Mo catalysts have higher catalytic activity with low steam/COG ratio, demonstrating that introduced CeO2 increases the water adsorb ability of catalyst. Keywords coke oven gas, water gas shift, sulfur-tolerant catalyst, cerium dioxide

1．Introduction In the “dual-gas resources” polygeneration system based on coke oven gas (CH4) reforming of gasification gas (CO2) [1, 2], a large number of H2 in coke oven gas (COG) is heated from ambient temperature to above 950℃, which consumes substantial energy of compression, transportation and heating [3]. Moreover, a quantity of H2 is wasted due to the oxidation of H2 [4]. Through CO shift reaction (CO+H2O→CO2+H2) converting the CO in COG into H2 and CO2, the product gas will be composed mainly of H2 and CH4, which can be easily separated into H2 and CH4-rich gas. The CH4-rich gas reforming of CO2 would save the energy and reduce H2 consumption. Based on the ideas above, the CO shift catalysts and shift properties of CO in COG are investigated. Coke oven gas contains high concentration of tar and multiple forms of sulfur, which requires the CO shift catalyst should be more active, more E-mail: [email protected]

sulfur-tolerant and more anagotoxic. Compared to the Fe-Cr and Cu-Zn catalysts, Co-Mo catalysts are more active and can be applied to a wider temperature range without sulfur poisoning [5]. CeO2 with a unique electronic structure has excellent oxygen-storage capacity (OSC). Gorte R. J. et al. [6] found a close relationship between CO shift activity and OSC. To further increase the conversion of CO in COG, CeO2 promoted Co-Mo sulfur-tolerant shift catalyst is investigated.

2. Experimental 2.1.

Catalyst preparation CoMo-CemKn, CoMo-Cem and CoMo-Kn catalysts were prepared by incipient

wetness impregnation of γ-Al2O3 with aqueous solution of metal salts according to certain sequence, where m and n represent mass ratio of promoter/catalyst respectively. The impregnation was followed by drying at 120℃ for 2 h and calcinating at 450℃ for 2 h. All obtained catalysts contain 3.0 wt-% of CoO and 10.0 wt-% of MoO3. FB301 is a commercial catalyst. 2.2．Catalytic activity measurement Catalytic activity tests were carried out with 15.0 ml of catalyst (diameter 3.0~ 4.0 mm) in a fixed-bed reactor. Prior to activity tests, the samples were sulfided in situ at 300℃ for 4h with a mixture of CO/H2/H2S (45.0/50.0/5.0, v/v/v). The reaction was conducted under the conditions: feed gas composition CO/CH4/H2/H2S = 9.7/30.0/60.0/0.3 (vol-%), steam-to-COG ratio = 0.5, COG space velocity = 2000 h−1, reaction pressure = 0.1 MPa. CO content was analyzed by using a gas chromatograph equipped with a thermal conductivity detector (a 13X molecular sieve column, H2 as the carrier gas). The conversion of carbon monoxide XCO is given by: XCO = (Vco-Vco')/Vco(1+Vco')×100% Where Vco and Vco' are the volume fraction of CO in the gas at the reactor inlet and outlet, respectively.

3．Results and discussion 3.1．Effect of CeO2 content on the catalytic activity of Co-Mo-Ce/γ-Al2O3

Conversion of CO/%

80

250OC 350OC

70 60 50 40 30

0

1

2

3

4

CeO2/%

5

6

7

8

Fig.1 Influence of CeO2 content on the catalytic activity of Co-Mo-Ce/γ-Al2O3

Figure 1 shows the catalytic performances of Co-Mo-Ce/γ-Al2O3 catalysts. It can be seen that CeO2 has a promoting effect on the activity of CoMo-Cem catalysts. When the CeO2 content is raised to 3.0 wt. % at 350℃, the conversion of CO reaches a maximum value, 77.6%. Hereafter, the catalytic activities decrease with CeO2 content increasing. As shown in Table 1, when the content of CeO2 on the CoMo-Cem catalyst is replaced by the same content of K2O, the obtained CoMo-Kn catalysts exhibit lower activities than that of the CoMo-Cem catalysts (m and n≦3). Furthermore, the Ce/K atom ratio in CoMo-Cem and CoMo-Kn catalysts is only 0.265, indicating that adding a little CeO2 can increase the catalytic activity of CoMo catalyst. This phenomenon may be due to valence variability and excellent oxygen-storage capacity of CeO2 [6, 8] which strengthens the transfer process of S-O and promotes CO shift reaction. Although the role of CeO2 is significant, the promotion of CeO2 is limited. The fact suggests that CeO2 can not serve as the main additive used in the catalyst, and should coordinate with other additives to further increase the conversion of CO in COG. Table 1 Comparison of CoMo-Cem and CoMo-Kn catalytic activity Content (wt%) Sample

CO conversion

(%)

K2O

CeO2

250℃

350℃

CoMo-Ce1

0

1

45.80

62.60

CoMo-K1 CoMo-Ce3

1 0

0 3

40.80 57.87

58.86 77.60

CoMo-K3

3

0

54.27

75.80

3.2 Effect of CeO2-K2O on the catalytic activity of Co-Mo-Ce-K/γ-Al2O3 90

Conversion of CO /%

80 70 60 50

CoMo-Ce3K6 CoMo-Ce1K8 CoMo-K9 CoMo-K8 CoMo-K6 CoMo-Ce3

40 30 20 200

250

300

350

400

Temperature/C O

Fig. 2

Figure

Influence of CeO2-K2O on the catalytic activity of Co-Mo-Ce-K /γ-Al2O3

2

shows

the

influence

of

CeO2-K2O

on

the

activity

of

Co-Mo-Ce-K/γ-Al2O3 catalysts under different reaction temperatures. It can be seen that the activities of all CoMo-CemKn catalysts increase rapidly with reaction temperature increasing，and CO conversion reaches an optimum value at 300℃. Hence, the catalytic activities decrease with the increase of reaction temperature, which may be limited by thermodynamic equilibrium. CeO2 and K2O promote the catalytic activity better comparing with each along, especially at the lower temperature. When CeO2 is added into CoMo-K6 and CoMo-K8 catalysts respectively, the conversion of CO greatly increases. CoMo-Ce3K6, CoMo-Ce1K8 and CoMo-K9 catalysts have the same content of total promoter (m + n = 9). However, CoMo-Ce3K6 and CoMo-Ce1K8 catalysts exhibit higher activities than that of CoMo-K9 catalyst. The fact indicates that there is a synergy effect between Ce and K in the Co-Mo-Ce-K/γ-Al2O3 catalyst, which improves the catalytic activity obviously. The specific synergy mechanism is not clear yet now and the further research is needed. In addition, the optimum values of total promoter content and CeO2/K2O mass ratio are found. When the total mount of promoter CeO2-K2O reaches 9.0 wt-% and the mass ratio of CeO2/K2O reaches 0.5, the conversion of CO reaches 91.72% at 300℃, being 15.51% higher than that of CoMo-K9 catalyst.

Besides, no methane is detected in the product gas when the feed gas dose not contain CH4, demonstrating that CO methanation reaction dose not occur and CoMo-CemKn catalysts have good selectivity. 3.3 catalytic activity of CoMo-Ce3K6 under different atmosphere 90

Conversion of CO/%

80 70 60 50

a b c d

40 30

200

250

300

0 Temperature/ C

350

400

Fig.3 The catalytic activity of CoMo-Ce3K6 under different atmosphere a. CoMo-Ce3K6：H2O/Gas~0.25； b. CoMo-Ce3K6：Semi-water gas c.FB301：H2O/Gas~0.25； d. FB301：Semi-water gas

The activities of CoMo-Ce3K6 and FB301 catalysts with low steam/COG ratio and semi-water gas are presented in Figure 3. It shows that CoMo-Ce3K6 catalyst has better activity than FB301 catalyst at the two kinds of atmosphere, which demonstrates that Co-Mo-K-Ce/γ-Al2O3 shift catalyst used in energy-saving technology and ammonia synthesis also obtain good results. With the ratio of steam/COG decreasing, the activity of Ce3K6-CoMo and K9-CoMo catalysts have a slip. However, the descend degree of Ce3K6-CoMo catalytic activity is lower than that of K9-CoMo catalyst. X.M.Миначев [9] reported that a large number of oxygen which can combine with hydrogen at a certain temperature adsorbed on the surface of CeO2. Therefore, CeO2 contained Co-Mo catalyst may have higher adsorption capacity of H+ in H2O, which makes OH- enrich on the surface of catalyst and maintain a high conversion of CO with low steam/COG ratio.

Table 2 Influence of CeO2 on activity of the catalyst with a low steam gas ration Sample

CO conversion/%(H2O/Gas~0.25)

CO conversion/%(H2O/Gas~0.5)

300℃

400℃

300℃

400℃

CoMo-K9

70.27

64.47

76.60

81.37

CoMo-Ce3K6

86.22

77.63

91.72

84.91

4. Conclusion CeO2 has a promoting effect on the activity of Co-Mo/γ-Al2O3 catalyst. However, the promotion of CeO2 is limited, indicating that CeO2 coordinate with other additives would promote the catalytic activity better. In the Co-Mo-Ce-K/γ-Al2O3 catalyst, Ce and K have a synergy effect on promoting the shift reaction of CO in COG. The catalytic activity reaches a maximum when the total promoter mount of CeO2-K2O reaches 9.0 wt-% and the mass ratio of CeO2/K2O reaches 0.5. Moreover, with the ratio of steam/COG decreasing, the descend degree of CoMo-Ce3K6 catalytic activity is lower than that of CoMo-K9 catalyst, which can be interpreted by that the introduced CeO2 increases the water adsorb ability of catalyst.

Acknowledgements This work was supported by the National Basic Research Program of China (2005CB 221202) and Shanxi Provincial Natural Science Foundation (2010011014-1).

References [1] Xie K C, Zhang Y F. CN Patent, 1974732A, 2007-06-06. [2] Xie K C, Zhang Y F, Zhao W. The basic research of dual-gas resources multi-generation coke oven gas to syngas. Shanxi Energy and Conservation, 2008, 49: 10-12. [3] Tian C S, Zhang Y F. Optimization of process conditions for syngas production by non-catalytic partial oxidation of methane. In: 17th International Symposium on Alcohol Fuels. Taiyuan: 2008, 350. [4] Wang J C, Zhang S Y, Bai T Z. Technical evaluation on the process of methanol

syngas production based on coke oven gas. Coal Chemical Industry, 2006, 126: 48-51. [5] Li S Y, Zhou X Q. Latest progress in CO shift catalysts. Coal Chemical Industry, 2007, 129: 31-34. [6] Gorte R J, Zhao S. Studies of the water-gas-shift reaction with ceria-supported precious metals. Catalysis Today, 2005, 104: 18-24. [7] Yang Y Q, Fang W P. CN Patent, 1559679A, 2005-01-05. [8] Ye B H, Jiang L L. Effects of CeO2 content on structure and properties of Ni-Mn-K/Al2O3 multiple-metal catalysts for water-gas shift reaction. Journal of The Chinese Rare Earth Society, 2006, 24: 661-664. [9] X.M.Миначев et al．《Application of rare earth in catalysts》. Beijing: Science Press, 1987: 1-25.

Effects of preparation conditions on Ru/Al2O3 catalyst for coal-based syngas methanation reaction Liping Wang, Yongfa Zhang* ,Yaling Sun, Xianglan, Li Key Laboratory of Coal Science and Technology of Shanxi Province and the Ministry of Education ,Taiyuan University of Technology , Taiyuan 030024 , China Abstract: Effects of preparation conditions on Ru/Al2O3 catalyst for coal-based syngas methanation reaction have been studied. The catalysts have been characterized by XRD technique. Using of ultrasound impregnation method can significantly decrease impregnated time and improve catalytic activity of the catalyst. The catalyst prioritizing preparation conditions are:Ru loading of about 2%, calcination temperature of 500℃, and H2 reduction temperature of 400℃. Under the optimum catalysts conditions, the conversion of CO and the CH4 selectivity reach 97.18% and 83.29%, respectively. Besides, the catalyst washed with deionized water and diluted ammonia can remove chlorine ions and increase catalytic activity. Key words：Ru catalyst；coal to methane；methanation；preparation conditions

1. Introduction Since Sabatier and Senderens first discovered methane formation by the reaction of carbon monoxide and hydrogen over a nickel catalyst in 1902[1], the methanation reaction has been developed and widely used for the purification of hydrogen in ammonia and hydrogen plants. In recent years, the interest for the reaction has grown significantly as a result of recent advancements in clean coal technology and the demand for development of converting coal source into substitute natural gas（SNG）. This was stimulated by a shortfall of natural gas supplies. The traditional catalyst for methanation is Ni supported on aluminum oxide[2,3] and Ru-based catalysts, they are also smaller loadings

[4–8]

. Many studies

[9,10]

have reported Ru or Ni catalyst

research for CO and CO2 methanation for purifying fuel cell or ammonia synthesis feed gas, but very few results for Ru-based catalyst for high CO concentrations about *Corresponding author. Tel. : +86 0351

6018676

E-mail address: [email protected]

coal-based syngas to natural gas. In the present work, a number of Ru catalyst with different metal loading supported on γ-Al2O3 were prepared and studied in the CO methanation reaction. Ru/γ-Al2O3 catalysts were prepared by wet incipient and ultrasonic impregnation method from RuCl3.nH2O precursor and their performance in CO methanation was investigated. From the catalysis test results, the best preparation conditions are found.

2. Experimental 2.1. Catalyst preparation The γ-Al2O3（Shanxi Fenyang Catalyst Plant produce）used as Ru carrier was prepared through crushed and sieved to 40-60 mesh, the material was then calcined in static air by heating to 500℃. The temperature was maintained for tree hour before cooling to room temperature. Two different methods of preparation catalysts were applied by using RuCl3.nH2O as precursor. One was that the materials xRu- Al2O3, with x = 1%, 1.5%, 2.0% or 2.5%, were prepared by impregnating Al2O3 with the aqueous solutions of RuCl3.nH2O of required concentrations for 24 hour. The other was prepared by the ultrasonic impregnation method. The xRu-Al2O3 was treated with a constant ultrasonic frequency of 40 kHz for 20 min, and then impregnated 12 hour. All solids were again dried and calcined under air at 500℃ for 3 hour. 2.2. Activity measurements A known mass of catalyst was loaded into a quartz tube reactor and capped with a quartz wool pulg. The catalyst was heated to 500℃ under flowing N2 (1000 ml. min -1) and H2 (50 ml. min -1) with the consumption of hydrogen monitored by an on-line gas chromatograph (GC950). Once hydrogen consumption was complete, signifying reduction of the ruthenium particles, the reaction was initiated by passing CO (75 ml. min -1) and H2 (25 ml. min -1) over the catalyst in the temperature range of 180～500℃. The mixtures of reactant gases and products were periodically analysed on-line using a TCD chromatography which contained a 1m column packed with 5A molecular sieve. Prior to analysis, the effluent was passed through a water-trap in

order to remove reaction water. The calculation equation of CO conversion、CH4 and CO2 selectivity was followed: X CO =

S CO2 =

FCOin × YCOin − FCOout × YCOout × 100% FCOin × YCOin FCO2 out × YCO 2out FCOin × YCOin − FCOout ×Y COout

S CH 4 =

FCH 4 out × YCH 4out FCOin × YCOin − FCOout × YCOout

× 100%

× 100%

where X CO : CO conversion, S : the selectivity of the CH4 or CO2, F : the gas flow rate of in and out, ml. min -1, Y : different fraction volume percentage, Superscript in : the inlet, superscript out : the outlet.

2.3 Catalyst characterization The crystalline structure of mixed oxides was determined by powder X-ray diffraction using Shimadzu XRD-6000 apparatus equipped with a monochromator for the Cu radiation at 40 kV and 30 mA. The 2θ scans covered the range from 20°to 80°.

3. Results and discussion 3.1 Effect of different impregnation methods on the catalysts activities

CH4 Selectivity, %

60 50

1# 2#

40 30 20 10 0

400

440

Tempreture, °C

480

Fig. 1 Catalytic activity of Ru/Al2O3 catalysts prepared with different impregnation methods 1#-0.5%Ru/Al2O3(ultrasonic impregnation)，2#-0.5%Ru/Al2O3(wet impregnation)

As can be seen from Fig.1, in the same experimental conditions, the catalyst catalytic activity prepared by ultrasonic impregnation method was higher than that prepared by wetness impregnation. CH4 selectivity was 53.9% over the 2#-0.5%Ru/Al2O3 catalyst at 480 ℃ , CH4 selectivity was reached 63.1% over 1#-0.5%Ru/Al2O3 catalyst, increased by 9.2%. The results indicate that the ultrasonic wave has an important influence on the catalyst activity. It may be the effects of

ultrasonic cavitation of ultrasonic wave to impregnating solution. The shock waves generated by cavitation form a certain pressures in the catalyst surface and internal. These pressures act on catalyst carrier surface, which make Ru metal uniformly and quickly distributed in the matrix. The ultrasonic impregnation can not only shorten the impregnation time but also can improve the catalytic activity which has also been reported by other authors [11, 12] 3.2 Effect of calcination temperature on the catalysts activities 100

(B) 100

80

80

CH4 selectivity, %

Convertion of CO, %

(A)

60 40 20 0 200

250

300

350

300°C 400°C 500°C 600°C

calcination calcination calcination calcination

400

450

60

40

300°C 400°C 500°C 600°C

20

0

500

200

250

Tempreture,℃

300

350

calcination calcination calcination calcination

400

450

500

Tempreture,℃

Fig.2 Effects of calcination temperature of Ru/γ-Al2O3 on catalyst performance

RuO2 γ-Al2O3

300℃

400℃

500℃

600℃ 0

10

20

30

2θ, 40

50

　 60

ã

70

80

90

2θ, ° Fig. 3 XRD patterns of Ru/Al2O3 catalysts under different calcined temperatures

Figure 2 shows that the different calcination temperature has effect on the Ru/γ-Al2O3 catalyst performance. It can be seen from Fig. 2 (A) that CO conversion increases and then decreases rapidly with the calcination temperature increasing. When the reaction temperature was increased from 300℃ to 600℃, CO conversion was decreased rapidly from 98.4% to 26.5%, decreased by 71.9%. This may be that the higher calcination temperature easily leads to the catalyst pore volume reduced. As

a result, the metal grain growth, Ru dispersion was decreased, the catalytic activity was decreased. It also can be seen the XRD spectra of the different calcination temperature Ru/γ-Al2O3 catalyst, Fig. 3. With the increasing of calcination temperature, the characteristic peaks of γ-Al2O3 gradually increase. It indicates that more γ-Al2O3 explored in surface area, that is, the surface active sites amount is reduced. Because of high temperature calcinations, the metal catalysts have a certain degree of sintering, the catalysts active sites and catalytic activity decreased. The effect of calcination treatment on catalysts had been previously reported in the literature [13]. Figure 2 (B) shows the effect of calcination temperature on the CH4 selectivity. When the calcinated temperature was increased from 300℃ to 500℃, the CH4 selectivity increased significantly from 75.6% to 98.3%, with a growth of 22.7%. In the lower calcination temperature, the catalyst can not form a suitable active phase, detriment of reaction. It also may be due to RuCl3 decomposition incomplete, the residual Cl-1 affected the catalyst activity, which was proved by PH test paper after experiment. However, when the temperature was further up to 500℃, the CH4 selectivity decreased. Comprehensive consideration CO conversion and methane selectivity, the optimum calcination temperature is 500℃ which is different from literature reported [14]. 3.3 Effect of reduction temperature on the catalysts activities (B) 100 80

80

60

40 300°C reduction 400°C reduction 500°C reduction

20

0

200

250

300

350

400

Tempreture,℃

450

500

CH4 selectivity, %

Convertion of CO, %

(A) 100

60

40 300°C reduction 400°C reduction 500°C reduction

20

0

200

250

300

350

400

450

500

Tempreture,℃

Fig. 4 Effects of reduction temperature on Ru/Al2O3 catalyst for CO methanation

Four different Ru/Al2O3 catalysts samples were reduced by hydrogen at 300、400 and 500℃, respectively. The reduction temperature effecting on the catalytic performance can be seen from Fig. 4(A) and (B). As shown in Fig. 4, the reduction temperature has obvious influence on catalytic performance of catalysts. The reduction temperature was heated from 300 to 400℃, the conversion of CO increased from 66.8% to 90.2%, while the selectivity of CH4 gradually increased from 60.6% to 80.3%. However, the reduction temperature further increased to 500℃, CO conversion gradually decreased to 54.5%. It also found that, with the reduction temperature increasing, CO2 concentration in export gas increases. The optimum reduction temperature is 400℃. 3.4 Effects of Ru metal loading on catalytic performance The effects of metal loading on catalytic activity have been investigated over catalysts with variable Ru content (1.0～2.5wt%), the results are shown in Fig. 5. 100

80

CH4

Selectivity, %

80

60

60

40

40 20

20

CO2

RuO2 γ-Al2O3

Convertion of CO, %

100

CO

5.0％Ru 2.5％Ru 2.0％Ru 1.5％Ru 1.0％Ru 0.5％Ru

0

0

1

2

3

4

5

0

Ru loading, %

Fig. 5 CO methanation over Ru/Al2O3 catalysts with different Ru loading

10

20

30

40

50

60

70

80

90

2θ， °

Fig. 6 XRD patterns of Ru/Al2O3 catalysts with different Ru loadings

As can be seen from Fig. 5, with the Ru loading increases, CO conversion and methane selectivity increased significantly. When the Ru content arrived at 2.0%, CO conversion and methane selectivity were 98.33% and 81.30% respectively. As the Ru content increased continuously, CO conversion did not change significantly. While methane selectivity decreased slightly with by-product of the selectivity of CO2 nearly unchanged. It is believed this phenomenon is due to the excess of ruthenium content,

uneven distribution on the carrier surface, and thus the catalytic activity decreased caused by metal crystal pile-up [15]. Figure 6 shows the Ru/γ-Al2O3 catalyst XRD spectra of different ruthenium loadings. It can be seen from Fig. 6, increasing the ruthenium metal content from 0.5% to 5.0 wt%, the intensity of the RuO2 peak is increased. When the ruthenium loadings of γ-Al2O3 was 5%, from the XRD curve, two peaks were found to at 57.3° and at 59.2 ° respectively, which could be assigned to the overload RuO2 accumulation in the carrier surface. Based on these results, the 2% Ru/Al2O3 catalyst is thought to be optimal when the thermal stability and metal dispersion are comprehensively considered, which can attractive more CO molecules to the catalyst active center by adsorption, dissociation and activation from all directions [16]. A result was slightly different from that previously published [17]. 3.5 Effect of Cl- on the catalysts activities The influence of washing in different ways on residual chlorine on catalyst activity has been studied. Sample 1 is named Ru/Al2O3(L), the sample without any treatment to elucidate the effect of residual chlorine; Sample 2 is named Ru/Al2O3(M), the sample washed with deionized water to remove chlorine ions; Sample 3 is named Ru/Al2O3(P), the sample washed with deionized water and diluted ammonia to remove chlorine ions. Then all catalysts samples were dried at 110℃ for 12h and reduced in H2 flow at 400℃ for 2h. Tab. 1 the effect of Cl- on Ru /Al2O3 catalyst activity Output concentration of CH4 (%) Sample

washing agent 300℃

Ru/Al2O3(L) Ru/Al2O3(M)

none deionized water

Ru/Al2O3(P) deionized water and diluted ammonia

350℃

400℃

4.0

5.1

18.4

3.4

11.7

41.1

2.7

46.6

50.0

As seen from tab. 1, the catalyst Ru/Al2O3(P) exhibits the highest activity and selectivity, and the catalyst Ru/Al2O3(L) shows lowest catalytic activity. The results indicate that the existence of Cl- did not favor the improvement of the activity of the

catalyst, which was in agreement with previously reported in the literature [18, 19].

4. Conclusion The ultrasonic impregnation method not only promote the catalytic activities of Ru/Al2O3, but also speed up the proceeding of impregnation. By comprehensive consideration, calcination temperature 500℃, H2 reduction temperature 400℃, and Ru loading 2%, an ideal Ru/Al2O3 catalysts for CO methanation can be obtained. The catalyst leached and washed with distilled water and diluted ammonia is an effective way to remove chloride ions and improve Ru /Al2O3 catalysts activity.

Acknowledgements This work was supported by the National Basic Research Program of China (2005CB221202) and Shanxi Provincial Natural Science Foundation (2010011014-1).

References [1] Jan Kopyscinski, T. J. Schildhauer, Frédéric Vogel, et al. Applying spatially resolved concentration and temperature measurements in a catalytic plate reactor for the kinetic study of CO methanation [J]. Journal of Catalysis, 2010, 271(2): 262-279. [2] Geoffrey Colin Bond. Metal-catalysed reactions of hydrocarbons [M]. American, spring science, 2005, 441-448. [3] A.L. Kustova, A.M. Freya, K.E. Larsena, et al. CO methanation over supported bimetallic Ni–Fe catalysts: From computational studies towards catalyst optimization [J]. Applied Catalysis A: General, 2007, 320: 98-104. [4] Zhang Cheng. Research progress of methanation of carbon monoxide and carbon dioxide [J]. Chemical industry and engineering process, 2007, 26(9): 12169-1273. [5] M A. Vannice. The catalytic synthesis of hydrocarbons from H2/CO mixtures over the Group VIII metals(V):The catalytic behavior of silica supported metals[J]. Journal of Catalysis, 1977, 50 (2): 282-236.

[6] Takashi Amano, Atsusshi Takumi, Shugou Zhang, et al. Carbon monoxide removing catalyst and production process for the same as well as carbon monoxide removing apparatus [P].US:20060160697 A1,2006. [7] Karin Yaccato, Ray Carhart, Alfred Hagemeyer, et al. Competitive CO and CO2 methanation over supported noble metal catalysts in high through put scanning mass spectrometer [J]. Applied Catalysis A: General, 2005, 296: 30-48. [8] Huang Zhongtao. Handbook of industrial catalysts [M]. Beijing China: Chemical Industry Press, 2001, 705-707. [9] S Takenaka, T Shimizu, K Otsuka. Complete removal of carbon monoxide in hydrogen-rich gas stream through methanation over supported metal catalysts [J]. Inter J Hydrogen Energy, 2004, 29 (10) : 1065-1073 [10] V P Londhe, V S Kamble, N M Gupta. Effect of hydrogen reduction on the CO adsorption and methanation reaction over Ru/TiO2 and Ru/Al2O3 catalysts[J]. J Mol Catal A, 1997, 121 (1): 33-44. [11] Kenneth Suslick, Taeghwan Hyeon, Mingming Fang, et al. Cichowlas. Sonochemical synthesis of nanostructured catalysts [J]. Materials Science and Engineering A: General, 1995, 2004 (1/2): 186-192. [12] T. J. Mason, A. Newman, J. P. Lorimer, et al. Hutt. Ultrasonically assisted catalytic decomposition of aqueous sodium hypochlorite [J]. Ultrasonics Sonochemistry, 1996, 3(1) : 53-55. [13] V. Ragaini, R. Carli, C.L. Bianchi, et al. Fischer-Tropsch synthesis on alumina-supported ruthenium catalysts II. Influence of morphological factors [J]. Applied Catalysis A: General, 1996, 139(1/2) : 31-42 [14] LU Hong-xuan, QIN Bang-hui, SUN kun-peng,et al. Effect of pretreatment on catalytic performance of Ru / Al2O3 for methanation [J]. Natural Gas Chemical Industry (Chinese), 2004, 29(4) : 1-9. [15] Di Li, Nobuyuki Ichikuni, Shogo Shimazu, et al. Catalytic properties of sprayed Ru/Al2O3 and promoter effects of alkali metals in CO2 hydrogenation [J]. Applied Catalysis A: General, 1998, 172(2): 351-358. [16] C H Bartholomew, R B Pannell. The stoichiometry of hydrogen and carbon

monoxide chemisorption on alumina and silica-supported nickel [J]. Catalysis, 1980, 65(2): 390-401. [17] P Panagiotopoulou, D.I. Kondarides, X. E. Verykios. Selective methanation of CO over supported Ru catalysts [J]. Applied Catalysis B: Environmental, 2009, 88(3/4): 470 - 478. [18] T Narita, H Miura, M Ohira,et al. The effect of reduction temperature on the chemisorptive properties of Ru/Al2O3: Effect of chlorine [J]. Applied Catalysis A: General, 1987, 32:185-190. [19] Shuzo Murata, Ken-Ichi Aika. Removal of chlorine ions from Ru/MgO catalysts for ammonia synthesis [J]. Applied Catalysis A: General, 1992, 82(1/2): 1-12.

Manuscript Not AVAILABLE

THE PRODUCTION OF ORGANIC FERTILIZERS FROM GÖYNÜK, ILGIN AND ELBİSTAN LIGNITES WİTH H2SO4 OXİDATİON M.Çöteli, A.Güntürk, A.Yavuz, A.Köker, G.Yıldırım, S.Atlıhan General Directorate of Mineral Research and Exploration ,Üniversiteler Mahallesi Dumlupınar Bulvarı No:139 06800- Çankaya/ANKARA

[email protected] Abstract Various organic fertilizer production processes related with alkali and oxidation of coals, with HNO 3 has been developed. In spite of this, these products are controversial for agriculture due to their high pH content of alkali oxidation products, high temperature and pressure comprising control difficulties. It is known that production with HNO 3 is difficult and has not found any application in time. Except pH, due to similar reasons and chemical structures of the produced products are not clearly known, therefore, nitro humates have not found any application areas. In this study, partial molecular disintegration of high humus, with H 2SO4 oxidation, comprising of Ilgın, Göynük and Elbistan lignite were studied, the neutralization reaction was carried out with ammonia by increasing acid ratios and a new type of organic fertilizer production with a total chemical work base and a new molecular structure was produced. While the used material was chosen randomly from the material’s beds and Ilgın contains of 42.02%, Göynük of 33.19 % and Elbistan of 53.00 % humic compounds. The samples cleaned with “acid leaching” in case of metal pollution and stoichiometric amounts of H2SO4, H3PO4 and Mardin phosphate were added through these compounds to substantiate of required amount for theoretical N-P2O5-K2O. Moreover, rested oxidized coal was neutralized with liquid ammonia and KOH. As a final product, fertilizer was produced in theoretical form of (8-6-1) + S. Humic acid ratio in produced fertilizers varied between 33.43% and 37.40 %; nitrogen amount, between 7.35 % and 8.46; P 2O5 amount, between 5.84 % and 6.46 %; K2O ratio, between 0.96 % and 1.12 %. Organic material changes between 70.92 % and 74.58 %. This is a cheap fertilizer which may be evaluated in the framework of the National Organic Fertilizer Regulation. Cheap raw material resource is a type of fertilizer which fixation of direct industrial chemicals’ usage and “industrial humification”, being used as an alternative of long time taking natural humidification as well as theoretical N, P 2O5, K2O like macro and micro nutrition elements, can be adjusted as desired ratios. Keywords: Humic acid fertilizers, Organomineral fertilizer, Organic sourced fertilizers

Modeling, Scaleup and Optimization of Slurry Bubble Column Reactors for Fischer-Tropsch Synthesis Laurent Sehabiague and Badie I. Morsi Chemical and Petroleum Engineering Department, University of Pittsburgh, 1249 Benedum Hall, 3700 O'Hara Street, Pittsburgh, PA 15261

1. ABSTRACT The gas holdup and mass transfer coefficients for syngas in actual Fischer-Tropsch wax were measured in a pilot-scale (0.3 m ID and 3-m height) Slurry Bubble Column Reactor (SBCR) operating under conditions typifying those of the actual Fischer-Tropsch (F-T) synthesis (T up to 530K, P up to 30 bar, UG up to 0.3 m/s, CS up to 20 vol.% ). Two new correlations were developed to predict the gas hold up and mass transfer coefficients experimental data with an Absolute Average Relative Errors (AARE) of 15.7% and 17.6%, respectively. These new correlations along with four different kinetic expressions, taken from the literature, were used in our simulator to predict the performance of a conceptual commercial-scale (7m ID, 30-m height) SBCR operating with iron catalyst. The simulator was used to study the effects of various operating variables (H2/CO ratio, CS, and UG) on the syngas conversion, the space time yield (STY) and catalyst productivity; and to delineate the “optimal” operating conditions for this a conceptual reactor. The simulation results led to the following conditions which could be considered “optimal” for operating this conceptual SBCR: H2/CO ratio =1, superficial gas velocity = 0.3 m/s and iron catalyst concentration = 20 vol%. Thus, this conceptual SBCR operating with iron catalyst could be integrated with a Coal-To-Liquid (CTL) process. It should be remembered, however, that the severe problems associated with iron catalyst attrition and separation from viscous F-T products could hamper the performance of such a conceptual SBCR and therefore a cobalt-supported catalyst could be used instead.

2. INTRODUCTION The Fischer-Tropsch (F-T) synthesis (originally known as “Synthol”) was developed in the 1920’s in Germany at the Kaiser Wilhelm Institute by Franz Fischer and Hans Tropsch, with the intent of producing synthetic hydrocarbons [1, 2]. Their work was based on the 1902 discovery of Sabatier and Senderens [3] that methane can be produced from H2 and CO in the presence of nickel catalyst. In the Fischer-Tropsch process, the syngas (H2+CO) reacts in presence of a solid catalyst to produce a wide spectrum of hydrocarbon, such as paraffins, olefins, and oxygenated chemicals (alcohols, aldehydes, acids, ketones, etc…). The F-T synthesis is a combination of oligomerization reactions [4], which can be summarized as follows: •

n-Paraffins synthesis:

+ 2 +1 •

+

(1)

1-Olefins synthesis:

+2 •

→

→

(2)

+

Alcohols synthesis:

+2

→

+

−1

(3)

A simplified expression of the above F-T overall reactions can be written as: + 1+

1 2

+

(4)

It has been reported that the main products of the F-T synthesis are linear paraffins with a typical H2/CO usage ratio ranging from 2.06 to 2.16 [4, 5]. The syngas is conventionally produced through steam reforming of natural gas (methane) or gasification of coal and/or biomass. Considering the huge quantities of coal and natural gas available worldwide (Figure 1), the F-T synthesis appears to be a promising technology which could help alleviating the dependence on oil imports and securing more environmentally friendly and sustainable alternative energy sources. The F-T process could play an important role in national security not only in the US, but also in developing countries, such as China and India.

The F-T reactiions has b been carrie ed out for o over 50 yea ars mainlyy in fixed-bed reactorrs (FBRs); and Sasoll (Sou uth Africa) pioneered d low-tempe erature F-T T synthesiss in slurry bubble column reacttors (SBCR Rs) in earlyy 1993 3. In these e reactors, the solid-p phase conssists of miccron-size ccatalytic pa articles whiich are susspended in n the lliquid-phasse (F-T pro oducts) by the synga as usually ssparged fro om the botttom of the e reactor th hrough the e slurrry-phase (ssolid catalyyst+ F-T liq quid products). The a advantagess of SBCR Rs over FBR Rs are [4, 6 6-8]: • • • • • •

Betterr temperatu ure control//removal Lowerr capital co ost (~25% o of that of a multi-tubu ular reactor) due to th heir relative ely simple design Low pressure drrop (4 times less than n in fixed b bed reactorr) Ability of using fine catallyst particlles (50%) in the case of CO for solid concentration >26 vol.% while in the case of H2 the resistance to mass transfer

does not become significant over the range of operating conditions used, except when using Ledakowicz et al. [20] rate expression at a catalyst concentration of 35 vol.%. Thus, CO appears to be the limitingreactant and could explain why in Figure 8 at a solid concentration of about 26 vol.% the syngas conversion reaches a maximum value, indicating that the reactor behavior is changing from a kineticcontrolled to a mass transfer-controlled regime. During the kinetics-controlled regime, the reactor performance greatly increases with the addition of catalyst, then it level off quickly; whereas in the mass transfer-controlled region, the addition of solids leads to a dramatic decrease of the reactor performance.

Deckwer et al.

100%

Huff and Satterfield

mass transfer controlled regime

90%

Ledakowicz et al.

80%

Zimmerman and Bukur

70%

β CO , %

60% 50% 40% 30%

kinetics controlled regime

20% 10% 0% 0%

10%

20%

30%

40%

50%

Cs , vol.%

Figure 10: Effect of Catalyst Concentration on Mass Transfer Resistance of CO (H2/CO ratio = 1, and UG = 0.2 m/s)

Deckwer et al.

100%

Huff and Satterfield

mass transfer controlled regime

90%

Ledakowicz et al.

80%

Zimmerman and Bukur

70%

β H2 , %

60% 50% 40% kinetics controlled regime

30% 20% 10% 0% 0%

10%

20%

30%

40%

50%

Cs , vol.%

Figure 11: Effect of Catalyst Concentration on Mass Transfer Resistance of H2 (H2/CO ratio = 1, and UG = 0.2 m/s) Figure 8 also shows that increasing the solid concentration from about 20 to 26 vol%, does not lead to significant increase in the syngas conversion for the four kinetic rate expressions used. Considering that the highest syngas conversion occurred at catalyst concentration of about 20 vol%, and a H2/CO ratio of 1, these values were used in the following simulations.

6.3 Effect of superficiial gas veelocity The effect of tthe superficial gas vvelocity on the synga as converssion and STY is pressented in F Figures 12 2 bserved, in ncreasing tthe gas ve elocity was found to d decrease tthe syngass and 13, respecctively. As can be ob nd increase e the hydrocarbons yield, exp pressed in n kg of hyd drocarbons produce ed per unitt convversion an reacctor volume per unitt time. Un nder the operating cconditions used in th he simulattion, the lo owest gass velo ocities allow w the highe est syngass conversio on (see Figure 9) du ue primarilyy to the lon ng residen nce time off the gaseous re eactants in n the reacttor; whereas the highest gas vvelocities a allow the g greatest hyydrocarbon n d (see Figure 10). Itt can also be observved that th he hydroca arbon yield d seems tto slowly level off att yield supe erficial gass velocitiess greater th han 0.3 m//s. Increassing the su uperficial ga as velocityy from 0.1 m/s to 0.3 3 m/s appears to o more tha an double the STY w while furthe er increase es from 0.3 m/s to 0 0.5 m/s only leads to o ut 30% inccrease in th he STY forr all the kin netics rate e expression n used. Thus, considering the lo ow syngass abou convversion and high ope erating cosst associate ed with run nning the S SBCR at high superfficial gas ve elocities, itt coulld be concluded that an optima al superficial gas velo ocity would d be aboutt 0.3 m/s fo or such a cconceptuall mmercial-sccale SBCR R operating with iron ccatalyst. com

Deckw wer et al.

100%

Huff and Satterfieeld

90%

Ledako owicz et al. 80%

Zimmeerman and Bukur

Syngas Conversion , %

70% 60% 50% 40% 30% 20% 10% 0% 0.1

0.2

0.3

0.4

0.5

UG, m/s

Figure 12: Effectt of Superfficial Gas Velocity o on Syngas s Convers sion (Catalys st concenttration = 2 20 vol%, H2/CO ratio o = 1)

Figure 13 3: Effect o of Superfic cial Gas V Velocity on n STY (Catalys st concenttration = 2 20 vol%, H2/CO ratio o = 1)

7. CONCLUDING REMARKS

The simulations results of a conceptual commercial-scale F-T SBCR (7-m ID, 30-m height) operating with iron catalyst considering mass transfer and using different kinetic rate expressions, taken from the literature, led to the following remarks: • • • •

•

•

At a superficial gas velocity of 0.2 m/s and a catalyst concentration of 20 vol%, the maximum syngas conversion occurred when using inlet H2/CO ratio of about 0.65. At H2/CO ratio of 1, CO is the limiting reactant and the regime in which the SBCR is operating is therefore dependent on the β ratio of CO. At a H2/CO ratio of 1 and a superficial gas velocity of 0.2 m/s, the maximum syngas yield occurred at a catalyst concentration of about 26 vol%. At a H2/CO ratio of 1 and a catalyst concentration of 20 vol%, no optimal superficial gas velocity could be clearly found as increasing the gas velocity led to the decrease of the syngas conversion and to the increase of the space time yield. Under the operating conditions studied, a superficial gas velocity about 0.3 m/s appears to be a good compromise. These above findings indicate that the following conditions which could be considered “optimal” for operating the conceptual SBCR: H2/CO ratio =1, superficial gas velocity = 0.3 m/s and iron catalyst concentration = 20 vol%. Thus, this conceptual SBCR operating with iron catalyst could be fully integrated with a CTL process. It should be remembered, however, that considering the sever attrition problems and the difficulties associated with separating the micron-size catalyst from the viscous F-T products inherent to the iron catalyst, a cobalt-supported catalyst could be used instead for such a conceptual reactor size.

8. NOMENCLATURE a

coefficient in kinetic rate expressions, dimensionless or Pa

aL

Gas-liquid interfacial area per unit liquid volume, m-1

Ci,L

Component i concentration in the liquid phase, mol/m3

CS

Solid concentration in the slurry phase by volume, vol %

dR

Reactor diameter, m

dorf.

Diameter of the orifices of the gas distributor, m

dP

Diameter of the solid particles, m

kLaL

Volumetric liquid-side mass transfer coefficient (based on liquid volume), s-1

L

Reactor length, m

MW

Molecular weight, kg.kmol-1

NO

Number of orifices of the gas distributor, -

P

Pressure, Pa

PS

Saturation vapor pressure, Pa

r

Reaction rate, mol.kg-1 catalyst.s-1

U

Superficial velocity, m s-1

X

foaming coefficient, -

xS

Solid concentration in the slurry phase, kg/m3

Greek symbols ε

holdup, -

Γ

Gas sparger coefficient, -

μ

Viscosity, kg m-1 s-1

ρ

Density, kg m-3

σ

Surface tension, Nm-1

Acronyms R

Ring

SBCR Slurry Bubble Column Reactor

Subscripts FT

Refer to the Fischer-Tropsch reaction

G

Refer to the gas phase

L

Refer to the liquid phase

P

Refer to the solid particles

S

Refer to the solid particles suspension

SL

Refer to the slurry phase

9. REFERENCES

1. 2. 3. 4. 5. 6. 7.

8. 9.

10. 11. 12. 13.

14.

15. 16. 17. 18.

19.

20.

21.

22. 23.

24.

25.

Dry, M.E., Commercial conversion of carbon monoxide to fuels and chemicals. Journal of Organometallic Chemistry, 1989. 372(1): p. 117-127. Stranges, A.N. Germany’s Synthetic Fuel Industry 1927-45. in AIChE 2003 Spring National Meeting. 2003. New Orleans, LA. Sabatier, P. and J.D. Senderens, Nouvelles Syntheses du Methane. Comptes Rendus, 1902. 134: p. 514. Dry, M.E., The Fischer-Tropsch process: 1950-2000. Catalysis Today, 2002. 71(3-4): p. 227-241. Steynberg, A. and M. Dry, Fischer-Tropsch Technology. Studies in Surface Science and Catalysis, ed. G. Centi. Vol. 152. 2004: Elsevier Science. Nigam, K.D.P. and A. Schumpe, Three-Phase Sparged Reactors. 1996, Amsterdam, The Netherlands: Gordon and Breach Publishers. Satterfield, C.N. and G.A. Huff, Product Distribution from Iron Catalyst in Fischer-Tropsch Slurry Reactors. Industrial & Engineering Chemistry Process Design and Development, 1982. 21(3): p. 465-470. Zimmerman, W.H. and D.B. Bukur, Reaction kinetics over iron catalysts used for the FischerTropsch synthesis. Canadian Journal of Chemical Engineering, 1990. 68(2): p. 292-301. Sehabiague, L., et al., Modeling and optimization of a large-scale slurry bubble column reactor for producing 10,000 bbl/day of Fischer-Tropsch liquid hydrocarbons. Journal of the Chinese Institute of Chemical Engineers, 2008. 39(2): p. 169-179. WEC, 2004 SURVEY OF ENERGY RESOURCES. 2004, Amsterdam, The Netherlands: Elsevier. WEC, 2007 SURVEY OF ENERGY RESOURCES. 2007, London, United Kingdom: World Energy Council. Sehabiague, L. and B.I. Morsi. Scaleup of SBCRs for Fischer-Tropsch Synthesis. in AIChE Annual Meeting. 2009. Nashville, TN. Behkish, A., et al., Novel Correlations for Gas Holdup in Large-Scale Slurry Bubble Column Reactors Operating under Elevated Pressures and Temperatures. Chemical Engineering Journal, 2006. 115(3): p. 157-171. Lemoine, R., et al., An Algorithm for Predicting the Hydrodynamic and Mass Transfer Parameters in Bubble Column and Slurry Bubble Column Reactors. Fuel Processing Technology, 2008. 89(4): p. 322-343. Schultz, H., Short history and present trends of Fischer-Tropsch synthesis. Applied Catalysis A: General, 1999. 186: p. 3-12. van der Laan, G.P., Kinetics, Selectivity and Scale Up of the Fischer-Tropsch Synthesis. 1999, University of Groningen: Groningen. Satterfield, C.N., et al., Effect of water on the iron-catalyzed Fischer-Tropsch synthesis. Industrial & Engineering Chemistry Product Research and Development, 1986. 25(3): p. 407-414. Callaghan, C.A., Kinetics and Catalysis of the Water-Gas-Shift Reaction: A Microkinetic and Graph Theoretic Approach, in Department of Chemical Engineering. 2006, Worcester Polytechnic Institute. p. 400. Huff, G.A. and C.N. Satterfield, Intrinsic kinetics of the Fischer-Tropsch synthesis on a reduced fused-magnetite catalyst. Industrial & Engineering Chemistry Process Design and Development, 1984. 23(4): p. 696-705. Ledakowicz, S., et al., Kinetics of the Fischer-Tropsch Synthesis in the Slurry Phase on a Potassium-Promoted Iron Catalyst. Industrial & Engineering Chemistry Process Design and Development, 1985. 24(4): p. 1043-1049. Deckwer, W.D., et al., Kinetic studies of Fischer-Tropsch synthesis on suspended iron/potassium catalyst - rate inhibition by carbon dioxide and water. Industrial & Engineering Chemistry Process Design and Development, 1986. 25(3): p. 643-649. de Swart, J.W.A., Scale-Up of a Fischer-Tropsch Slurry Reactor. 1996, University of Amsterdam: Amsterdam, Netherlands. de Swart, J.W.A. and R. Krishna, Simulation of the transient and steady state behavior of a bubble column slurry reactor for Fisher-Tropsch synthesis. Chemical Engineering and Processing, 2002. 41(1): p. 35-47. de Swart, J.W.A., R.E. van Vliet, and R. Krishna, Size, Structure and Dynamics of "Large" Bubbles in a Two-Dimensional Slurry Bubble Column. Chemical Engineering Science, 1996. 51(20): p. 4619-4629. Lemoine, R., A. Behkish, and B.I. Morsi, Hydrodynamic and Mass Transfer Characteristics in Organic Liquid Mixtures in a Large-Scale Bubble Column Reactor for the Toluene Oxidation Process. Industrial & Engineering Chemistry Process Design and Development, 2004. 43(19): p. 6195-6212.

26. 27.

28. 29.

30.

31.

32. 33. 34. 35. 36. 37.

Rados, N., M.H. Al-Dahhan, and M.P. Dudukovic, Modeling of the Fischer–Tropsch synthesis in slurry bubble column reactors. Catalysis Today, 2003. 79-80: p. 211-218. Behkish, A., et al., Gas holdup and bubble size behavior in a large-scale slurry bubble column reactor operating with an organic liquid under elevated pressures and temperatures. Chemical Engineering Journal, 2007. 128(2-3): p. 69-84. Grund, G., A. Schumpe, and W.D. Deckwer, Gas-Liquid mass transfer in a bubble column with organic liquids. Chemical Engineering Science, 1992. 47(13-14): p. 3509-3516. Krishna, R., M.I. Urseanu, and A.J. Dreher, Gas holdup in bubble columns: influence of alcohol addition versus operation at elevated pressures. Chemical Engineering and Processing, 2000. 39(4): p. 371-378. Vermeer, D. and R. Krishna, Hydrodynamics and mass transfer in bubble columns in operating in the churn-turbulent regime. Industrial & Engineering Chemistry Process Design and Development, 1981. 20(3): p. 475-482. Behkish, A., Hydrodynamic and Mass Transfer Parameters in Large-Scale Slurry Bubble Column Reactors, in Chemical and Petroleum Engineering Department. 2004, University of Pittsburgh: Pittsburgh, PA. p. 331. Kato, Y., et al., The behavior of suspended solid particles and liquid in bubble columns. Journal of Chemical Engineering of Japan, 1972. 5(2): p. 112-118. Kojima, H., H. Anjyo, and Y. Mochizuki, Axial mixing in bubble column with suspended solid particles. Journal of Chemical Engineering of Japan, 1986. 19(3): p. 232-234. O’Dowd, W., et al., Gas and solids behavior in a baffled and unbaffled slurry bubble column. American Institute of Chemical Engineering Journal, 1987. 33(12): p. 1959-1970. Smith, D.N. and J.A. Ruether, Dispersed solid dynamics in a slurry bubble column. Chemical Engineering Science, 1985. 40(5): p. 741-754. Deckwer, W.-D., et al., Chemical Engineering Science, 1981. 36: p. 765. Deckwer, W.D., et al., Chemical Engineering Science, 1981. 36: p. 765.

BIOGASIFICATION OF SOMA LIGNITE (A Preliminary Study) Mustafa Baysala, Sedat İnanb, Yuda Yürüma a

Faculty of Engineering and Natural Sciences, Sabanci University, Orhanli, Tuzla, Istanbul 34956, Turkey TÜBİTAK Marmara Research Centre, Earth and Marine Sciences Institute, Gebze-Kocaeli, Turkey Corresponding author: e-mail: [email protected] b

ABSTRACT In this project, the bacterial gasification on the coal samples which were evacuated from Soma basin in Turkey and gas adsorption mechanism of these samples were analyzed. It is known that coal can be solubilized chemically (alkaline solutions) and biologically by using wood-rotting fungi species. Chemical solubilization of coal samples was investigated. For this purpose, coal samples were solubilized in the different Lewis base solutions. For biogasification process, solubilization at moderate pH (9≥ pH ≥5) level is an important factor for the conserve bioactivity of the microorganisms. We found that carbonate and oxalate systems can be solubilized coal at moderate pH and also these Lewis bases was used in biogasification process to solubilized coal samples and increase gasification efficiency. To understand gas adsorption on the coal surface, high pressure gas adsorption experiments were conducted. Keywords:

Coal, biogasification, solubilization, adsorption

1. INTRODUCTION Within the increasing interest of the renewable energy source and clean energy, coal utilization has become a more important subject in today’s world. Natural gas is a relatively clean energy with respect to the combustion of the coal. We know that coal bed methane is a considerable large energy source for most of the countries. For example in America, coal bed methane covers more than %10 of the natural gas demand [1]. Coal bed methane (CBM) can be arise from both thermogenic and biogenic activity on the coal beds and can be adsorb on the porous structure of the coal. Biogenic methane gas, due to the bacterial activity of the coal beds, can be called primary biogenic methane, and more than %20 of the coal bed methane is originated from these activity. Coal can be utilized by using biological processes. Last two decades, lots of coal utilization methods have been developed. After the first report published by Cohen & Gabrielle in 1982 [2], lots of biological methods have been developed by scientists like microbial coal liquefaction, methane production. Methane can be produced by using laboratory incubation experiment, called secondary biogenic methane. Methane production due to the biological treatment is a relatively clean process compared to the coal combustion. Biological treatment of the coal takes place under mild conditions at low temperature and pressure unlike the classic processes. Production rate changes from the rank of the coal, usually low rank coals can be converted to the methane more easily under the proper environmental conditions [3]. Although the mechanism of the process have not been fully discovered, the pathways of the microbial treatment of the coal are common in most of the studies. Isbiste and Baric [4] reported two main approaches for the synthesis of the methane from coal. First one is the direct gasification of the coal. In this process, solid coal samples are treated with the methanogenic bacteria to produce methane from coal without any solubilization/depolymerization step. In the second approach, which is called indirect gasification, first, lignite is depolymerized/solubilized biologically or chemically (using aqueous alkali). By this way, complex polymer structure of the coal is broken down to smaller organic molecules. Then, acetogenic

bacteria ferment these smaller organic molecules (fatty acids, alcohols and aromatic acids) to acetate, hydrogen and carbon dioxide. After that, methanogens used to convert acetate, hydrogen and carbon dioxide to methane. The goal of this study is the investigation of the biogenic natural gas potential and microbial gasification process of Soma lignite in Turkey by laboratory experiments. We tried to improve the efficiency of gasification process and investigate biogasification suitability of Soma lignite. Also, gas potential of lignite was explored by using surface adsorption and desorption isotherms.

2. EXPERIMENT AND RESULTS In this study, solubilization of the coal samples was investigated as a beginning. Studies show that, coal can be solubilized biologically or chemically. In biological methods, wood-rotting fungi species are able to solubilize/depolymerize low rank coals by secreting oxalate ions. Bumpus et al. shows that chemically, coal macromolecules are solubilized in an alkaline medium, like aqueous solution of oxalate ions, instead of using any fungi species [5]. In the light of this point, we used different Lewis bases to reach maximum solubilization capacity for our coal samples. Effect of pH on the ability of Lewis bases to solubilize coal was investigated. Also, characterization of the coal samples which were excavated in different depths from Soma Basin in Turkey was performed. Coal samples which come from Soma basin were broken and grinded into small particles, and then passed through the sieve to reduce particle size to 80 µm. Resulting samples were preserved at nitrogen atmosphere. For the determination of the nitrogen, sulfur and carbon content in the coal, ultimate and proximate analysis of the samples were performed. Table1. Ultimate and proximate analysis of the Soma lignites Samples Sample Depth No 690.45 K1 626,20 K2 726,20 K3 728,20 K4 779,60 K5

Original sample

Dry sample

C%

H%

N%

S%

C%

H%

N%

S%

O%

44,5350 24,8250 59,4530 56,1470 68,1460

3,9394 2,9924 5,7984 4,7999 5,6492

1,2000 0,4089 1,8942 0,9614 1,6289

0,6966 2,0001 1,2361 1,4146 1,3910

47,5490 26,1840 63,6270 59,8460 73,1650

3,4486 2,5437 5,4198 4,3790 5,2411

1,2812 0,4313 2,0272 1,0247 1,7489

0,7438 2,1095 1,3229 1,5078 1,4934

9,4874 10,4815 12,6931 11,2225 10,5216

Samples

Original sample

Sample No

Depth

Moisture %

Volatile Matter %

Ash %

K1 K2 K3 K4 K5

690.45 626,20 726, 20 728,20 779,60

6,34±0,03 5,19±0,02 6,56±0,03 6,18±0,03 6,86±0,03

26,21±0,03 29,58±0,04 37,63±0,05 35,12±0,05 37,94±0,05

35,12±0,04 55,23±0,06 13,93±0,01 20,66±0,02 7,29±0,01

Dry sample Fixed Carbon % 32,33 10,00 41,88 38,04 47,91

Volatile Matter % 27,98 31,2 40,27 37,43 40,73

Ash % 37,49 58,25 14,91 22,02 7,83

Fixed Carbon % 34,53 10,55 44,82 40,55 51,44

Furthermore, aromatic groups of the coal samples are degraded into aliphatic chains by microorganisms during the biogasification process. Therefore, the functional groups of the sifted coal samples were determined by the Bruker Equinox 55 FTIR spectrometer.

140

T R 120 A N S 100 M I T 80 T A N 60 C E [%] 40 20 0 4000

3500

3000

2500

K1

2000

1500

1000

500

Wavenumber cm-1 K2 K3 K4 K5

Fig.1. FTIR spectrometry for Soma lignites

Figure1 shows that FT-IR spectra of the lignite samples which were evacuated from different depths of Soma basin. Most of the peaks are common for all samples. The broad band at 3400 cm-1 is due to the O-H and N-H groups. The peaks at 2900-2800 represent C-H stretching vibration of aliphatic and alicyclic methyl, methylene and alkyl groups. The intensity of the peak at 2900 cm -1 is greater than intensity of the peak at 2800 cm-1. This shows the presence of long aliphatic chains in the samples. Also, the peak at 1600 cm-1 and the shoulder at 1700 cm-1 indicate the high concentration of conjugated carbonyl structures and conjugated aromatic structures. Relatively strong band at 1400 cm-1 is due to the C-H bend in methyl, methylene or aromatic C-H bending. Usually, peaks between 1100 cm-1 - 400 cm-1 belong to inorganic mineral matter in the coal. For the chemical solubilization experiments, sodium oxalate, sodium carbonate and sodium phosphate were used as Lewis bases. 10 ml, 0.1 M aqueous solutions of sodium oxalate, sodium carbonate and sodium phosphate at varying pH values (4-12) were prepared. Then 50 mg of coal was added to these solutions. Mixtures were shaken at 200 rpm, for 24 hours on a rotary shaker. After that time, aliquots were centrifuged at 5000 rpm for 20 min and absorbance intensities of final solutions were determined by using UV-visible spectrometer.

Fig.2. Absorbance vs. pH dependence chart

This figure shows the comparison of the absorbance intensity at 275 nm of the solubilized coal in the aqueous solution of the different Lewis bases at varying pH. At high pH values (pH ≥10), solubility of the coal is more in carbonate and phosphate anions systems than oxalate anions. On the other hand, at lower pH levels between 9≥ pH ≥5, oxalate anion system solved coal more effectively. This result is important, since in the microbial gasification processes of the coal, harsh conditions reduce microorganism’s activity. Therefore, higher solubilization of the coal at moderate pH levels is a very important parameter to increase the efficiency of the biogasification processes due to the decomposition of the complex structure of the coal into the smaller organic substances which are more easily converted to methane by microorganism. For the future prospect of this study, Lewis bases will be used for the direct and indirect biogasification of the lignites samples to increase efficiency of these processes. Gas adsorption experiments were performed by using a Hiden isochema intelligent gravimetric analyzer IGA-001. Mechanism of the system is based on measuring the weight change of the sample against reference as a function of time, pressure and temperature. High pressure sensor (vacuum to 10bar) and temperature controller allow us to make a wide variety of sorption measurements and gives information about kinetics of the sorption processes. The precision of the measurement can be controlled by PC with long term stability of microbalance was 0.1µg with a weighting resolution 0.2 µg. To understand the mechanism of the coal adsorption, we start with nitrogen adsorption isotherm at room temperature with temperature stability is 0.1oC. Coal sample was outgased at 105oC until the constant weight was achieved at 10-6 mbar. For the biogasification experiments, anaerobic media which contain minerals and vitamins were prepared to isolate microorganisms from coal in order to understand the mechanism of the methane generation due to the endemic population. After this procedure, optimization of the methane generation will be done by adjusting pH, temperature and coal amount.

Acknowledgements We would like to thank Fırat Duygun from TUBITAK Marmara Research Center, Earth and Marine Sciences Institute for ultimate and proximate analyses.

References [1] S.H. Harris et al, International Journal of Coal Geology 76, 46–51(2008) [2] Cohen, M.S., Gabriele, P.D., Applied and Environmental Microbiology 44, 23-27 (1982) [3] Rice, D.D., Claypool, G.E., American Association of Petroleum Geologists Bulletin 65, 5–25 (1981) [4] Isbister, J. D., Barik, S., Crawford, D. L., Ed.; CRC Press: Boca Raton, FL,139-156 (1993) [5] Bumpus et al, Energy & Fuels, Vol. 12, No. 4, 664-671 (1998)

2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 PROGRAM TOPIC: COAL-DERIVED PRODUCTS (FISCHER-TROPSCH) CATALYTIC PERFORMANCE IN FIXED-BED AND BUBBLING FLUIDIZED-BED REACTOR DURING FISCHER-TROPSCH SYNTHESIS ON THE IRON-BASED CATALYSTS Jong Wook Bae* Petroleum Displacement Technology Research Center, Korea Research Institute of Chemical Technology (KRICT), P.O. Box 107, Sinseongno 19, Yuseong, Daejeon, 305-600, Korea *Presenting author: Tel.: +82-42-860-7383 Fax: +82-42-860-7388; E-mail address: [email protected]. Ki-Won Jun, Yun-Jo Lee and Kyoung Su Ha Petroleum Displacement Technology Research Center, Korea Research Institute of Chemical Technology (KRICT), P.O. Box 107, Sinseongno 19, Yuseong, Daejeon, 305-600, Korea Abstract: Fischer-Tropsch synthesis (FTS) for olefin production from syngas was investigated on the four different iron-based catalysts in a fixed-bed and a bubbling fluidized-bed reactor. The catalysts were prepared by wet-impregnation using Al2O3, SiO2 and iron ore (FeOx) with active components of Fe, K and (or) Cu, and K/FeCuAlOx catalyst was prepared by co-precipitation method. The impregnated K/FeOx catalyst is found to be one of the promising catalysts to be applied in bubbling fluidized-bed reactor for high temperature FTS reaction due to its high resistance to catalyst attrition with high catalytic performance. Keywords: Fischer-Tropsch synthesis, iron-based catalyst, fixed-bed reactor, bubbling fluidizedbed reactor. 1. Introduction FTS is an important technology in the production of liquid fuels and valuable chemicals from syngas derived from the gasification of coal or biomass and the reforming of natural gas or other carbon-containing components.1-3 In FTS process, the reactor is chosen appropriately according to the targeted products such as light olefin, gasoline, diesel or heavy hydrocarbons including wax etc., and it also requires the appropriate selection of catalyst preparation method depending on the types of reactor such as plug flow (fixed-bed or supercritical), fluidized-bed or slurry bubble column reactor.4-6 In the present investigation, FTS reaction with the various iron-based catalysts for the production of light hydrocarbons from syngas was carried out in fixed-bed reactor (FBR) and bubbling fluidized bed reactor (BFBR). 2. Experimental Iron-based FTS catalysts adopted in the present study were prepared by using the conventional coprecipitation for K/FeCuAlOx catalyst and wet-impregnation method for FeCuK/γ-Al2O3,

FeCuK/SiO2, and K/FeOx catalyst. Catalytic performance on the four different iron-based FTS catalysts was studies in FBR (I.D. = 10.7 mm) with a catalyst of 0.3 g possessing a particle size of 53 - 120 µm. In addition, catalytic performance was also carried out in BFBR (0.034 m I.D. × 1.5 m height) equipped with the perforated type of gas distributor. The catalyst possessing same particle size distribution adopted in FBR were loaded with 100 g in BFBR equipped with a perforated plate containing 11 evenly spaced holes of 1.0 mm diameter, which served as a reactant distributor. Prior to the reaction in both reactors, the catalyst was reduced at 450 oC for 24 h in a flow of 50 % H2 balanced with helium. After reduction, the synthesis gas (H2/CO = 2) was fed into the reactor. The FTS reaction was carried out subsequently under the following reaction conditions; T = 300 oC, P =1.0 MPa and Ug (linear velocity of syngas) = 0.2, 0.4, 1.0, 2.0, 4.0, 8.0 and 10.0 cm/s for BFBR and T = 300 oC, P =1.0 MPa and space velocity (SV) = 2000 ml/gcat/h for FBR. 3. Results and discussion The textural properties of four different catalysts prepared using different preparation methods are summarized in Table 1. The larger surface area was observed on the impregnated catalysts such as FeCuK/Al2O3 and FeCuK/SiO2 around 192.1 and 157.8 m2/g separately. The lower average pore diameter on the impregnated FeCuK/Al2O3 and co-precipitated K/FeCuAlOx catalysts was observed around 4.4 and 5.9 nm respectively. The lower average pore diameter on FeCuAl/Al2O3 catalyst with high surface area is mainly attributed to the presence of bimodal pore structure with the co-existence of micro and mesopores in the regions of 3-5 nm and 10-20 nm. The apparent and bulk density of four different catalysts is in the following order; K/FeOx > K/FeCuAlOx > FeCuK/Al2O3 > FeCuK/SiO2. Table 1. Textural properties of the iron-based FTS catalysts catalysts

a

FeCuK/Al2O3 FeCuK/SiO2 K/FeCuAlOx K/FeOx

preparation method (wt. ratio of catalyst) Impregnation (20/2/4/100) Impregnation (20/2/4/100) Coprecipitation (4/100/6.6/15.7) Impregnation (6/94)

surface area (m2/g)

pore volum e (cm3/g)

average pore diameter (nm)

apparent density (g/cm3)

192.1

0.264

4.4

1.53

157.8

0.678

13.0

0.84

89.5

0.171

5.9

2.43

0.6

0.010

14.4

3.81

a

FeCuK/Al2O3 and FeCuK/SiO2 catalysts were prepared by conventional impregnation method at a fixed weight ratio of Fe/Cu/K as 1/0.1/0.2 on γ-Al2O3 abd SiO2 support with a fixed amount of Fe/support at 20/100. K/FeCuAlOx catalyst was prepared by coprecipitation method using K2CO3 (Fe:K = 100:4 wt%) and K/FeOx catalyst was made by wet-impregnation of K2CO3 on iron ore (FeOx) with K/Fe weight ratio of 6/94.

The summarized results of C2 - C4 yield with respect to superficial gas velocity in FBR and BFBR are shown in Figure 1. The hydrocarbon selectivity in BFBR is found to be shows a superior than that of FBR. The beneficial effects are the main advantages of BFBR for the production of C2 - C4 olefins. The higher formation rate of active iron carbide species with facile reducibility of iron oxides and appropriate acidity plays an important role to obtain high catalytic performance in FBR and BFBR.

Figure 1. The yield to C2 - C4 olefin with respect to the superficial gas velocity in FBR and BFBR 4. Conclusions FTS reaction for the production of C2 - C4 light olefins from syngas was carried out in FBR and BFBR for HTFT reaction with four different iron-based catalysts. Under similar operating conditions such as temperature and pressure, the C2 - C4 selectivity and olefin selectivity in BFBR are higher than that in FBR on the K/FeCuAlOx and K/FeOx catalysts. K/FeOx catalyst which is simply impregnated with potassium promoter shows a high resistance for catalyst attrition with high catalytic performance and this catalyst can be one of the best catalysts for applying HTFT reaction using BFBR. References 1. 2. 3. 4. 5. 6.

A.Y. Khodakov, W. Chu, P. Fongarland, Chem. Rev. 107 (2007) 1692-1744. B.H. Davis, Top. Catal. 32 (2005) 143-168. M.E. Dry, Catal. Today 71 (2002) 227-241. D.B. Bukur, X. Lang, L. Nowicki, Ind. Eng. Chem. Res. 44 (2005) 6038-6044. S.H. Kang, J.W. Bae, K.J. Woo, P.S. Sai Prasad, K.W. Jun, Fuel Process. Technol. 91(4) (2010) 399-403. S.H. Kang, K.J. Woo, J.W. Bae, K.W. Jun, Y. Kang, Korean J. Chem. Eng., 26(6) (2009) 1533-1538.

Manuscript Not AVAILABLE

2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010

Abstract Submission

COAL-DERIVED PRODUCTS: (COAL-TO LIQUIDS)

Operation of slurry reactor for Fischer-Tropsch synthesis

Ho-Tae Lee, Korea Institute of Energy Research, Daejeon 305-343, Korea [email protected], Fax) 82-42-860-3134, Tel) 82-42-860-3662

Jung-Il Yang, Korea Institute of Energy Research, Daejeon 305-343, Korea [email protected], Fax) 82-42-860-3134, Tel) 82-42-860-3795 Jung Hoon Yang, Korea Institute of Energy Research, Daejeon 305-343, Korea [email protected], Fax) 82-42-860-3134, Tel) 82-42-860-3695 Dong-Hyun Chun, Korea Institute of Energy Research, Daejeon 305-343, Korea [email protected], Fax) 82-42-860-3134, Tel) 82-42-860-3071 Hak-Joo Kim, Korea Institute of Energy Research, Daejeon 305-343, Korea [email protected], Fax) 82-42-860-3134, Tel) 82-42-860-3654 Heon Jung, Korea Institute of Energy Research, Daejeon 305-343, Korea [email protected], Fax) 82-42-860-3134, Tel) 82-42-860-3663

ABSTRACT A slurry bubble column reactor with a capacity of 0.03 bbl/day, was designed and operated for Fischer-Tropsch reaction. Active Fe based catalysts were prepared and tested at 2.5 MPa and H2/CO=1. The average CO conversion higher than 80% was observed. The effects of the reaction temperature and the superficial velocity of synthesis gas on the conversion, the product selectivity and the oil productivity were investigated. The liquid oil productivities increased with the increasing superficial velocity and the increasing temperature.

Coal-Derived Products Process Simulation of Steam Hydrogasification to Produce F-T Products and Electricity Xiaoming Lu, Chan S Park*, Joseph M Norbeck Bourns College of Engineering Center for Environmental Research and Technology Department of Chemical and Environmental Engineering, University of California, Riverside *Corresponding author: Tel: 1-951-781-5771; Fax: 1-951-781-5790; E-mail address: [email protected] Abstract: A new process for the co-production of synthetic fuels and electricity based on steam hydrogasification is being developed at the College of Engineering – Center for Environmental Research and Technology (CE-CERT) at the University of California, Riverside. One of the key benefits of this new process is the enhanced conversion of carbonaceous material to synthesis gas compared to other thermochemical processes. Another benefit is that it does not require the use of oxygen from a cryogenic air separation unit or ASU, thus reducing capital costs. One current coal-based application being considered for this technology (called the CE-CERT process) is to co-produce Fischer Tropsch products and electricity. The results of a process simulation model of an integrated conceptual design of a 4000 TPD Sub-Bituminous coal conversion to Fischer-Tropsch liquids and electricity is presented in this paper. A circulating fluidized bed with a regenerator setup for providing the heat to the Steam Hydrogasification Reactor (SHR) is modeled. The process simulation model involves the major steps of: 1) simulation of the steam hydrogasification reactor using gasification units based on built-in Aspen reactor blocks with the determination of the equilibrium composition of the gaseous components in the reactor by means of Gibbs free energy minimization; 2) empirical simulation of a warm gas cleanup system; 3) a steam methane reforming process simulated by using a built-in REQUIL equilibrium block; 4) empirical simulation of a hydrogen separation process for syngas ratio adjustment and excess hydrogen recycle to the steam hydrogasifier; and 5) a Fischer-Tropsch diesel synthesis by means of an empirical expression. The regenerator is essentially a combustor where residue char is burnt in the presence of air to heat circulating sand. The hot sand flows back into the SHR to provide the process heat for the main reactor. The material and energy balance of the whole process was developed using Aspen Plus. The overall process performance and the optimum F-T liquids/electricity output from the 4000 TPD coal plant is determined. The optimum process thermal efficiency is 50.25% with 37.1% of coal carbon in F-T liquids. 1.

Introduction

Petroleum was widely used in the 20th century and dominated the energy production and chemical industries. Fuels from crude fossil oil supply about 96–98% of the worldwide energy for transportation (cars, ships, airplanes). More than 55% of the petroleum extracted is refined to produce fuels. Estimates of remaining petroleum availability at the present rate of consumption span from 40 to 60 years [1]. Not surprising the price of oil rose from $12/barrel to more than $130/barrel during the period from 1945 to 2008 [2]. The price is expected to increase in future years. Conversely, the currently known reserves of coal exceed those of crude oil by a factor of 25 [3]. Production of syngas from coal and the successive conversion to a range of fuels and chemicals become of increasing interest as the reserves of crude oil are depleted and the global price per barrel increases. The University of California Riverside’s College of Engineering Center for Environmental Research and Technology (CE-CERT) is developing a new process for the production of F-T products and electricity based on steam hydrogasification. This process is also called the CE-CERT process. The details of this technology have been published elsewhere [4-7] but a short description of the important unit operations are presented below for completeness. Initially, the feedstock is made into a slurry with water and is fed

along with H2 into the Steam Hydrogasification Reactor (SHR). The SHR produces methane, along with carbon monoxide and carbon dioxide. A regenerator is setup to provide the process heat for SHR by combusting the leftover char using circulating sand [8]. The methane rich gas from the SHR is then subjected to cleanup in order to remove contaminants, primarily sulphur species. The methane is then converted into a synthesis gas, a mixture of H2 and CO, in the Steam Methane Reformer (SMR) [9]. The SMR is followed by a fuel synthesis reactor, usually a F-T reactor. Alternatively, the syngas can be used for power generation if desired. The excess H2 from the syngas is separated and is fed into the SHR as feed thus eliminating the need for a separate source of hydrogen. The F-T reactor operating parameters and catalyst can be varied to obtain a maximum yield of diesel, gasoline or kerosene or other fuels such as jet fuel. The basic chemical reactions taking place during the different stages of the process are given below. Steam Hydrogasification (SHR): C + H2O + 2H2 → CH4 + H2O + others (CO, CO2, C2+, etc)

(1)

Steam Methane Reforming (SMR): CH4+ H2O → 3H2+ CO

(2)

F-T Reaction (FTR): CO + 2H2 → -(CH2)- + H2O 2.

(3)

Features of CE-CERT process

A schematic of the CE-CERT process is shown in Fig.1. The main features are summarized below. 1.

Enhanced conversion of carbonaceous material to synthesis gas compared to other thermochemical processes.

2.

No need for oxygen from a cryogenic air separation unit or ASU and a moderate operation temperature and pressure compared with partial oxidation gasification, thus reducing capital costs.

3.

Slurry feed, wet feedstock can be used directly which reduces cost of drying the feedstock and offers the potential of more efficient handling of feedstock.

4.

Closed-loop H2 cycle, no external H2 is needed and make the process self sustainable.

5.

A high rate of methane is produced in SHR and be used as a source of clean synthetic natural gas if liquid fuel is not desired.

6.

A flexible control, optimum H2 to CO ratio for efficient downstream production of fuel products can be achieved by controlling the input H2O/feed ratio in the SHR.

Fig.1 Schematic of the CE-CERT Process

The technologies for the commercial production of F-T diesel using synthesis gas are currently considered to be mature. As mentioned earlier, the versatile nature of the CE-CERT process allows the product syngas H2/CO ratio to be controlled in a relatively simple manner by varying the H2 to carbon and H2O to coal ratios of the SHR feed. This enables the production of F-T products in an efficient manner irrespective of the moisture content of the coal and catalyst used in F-T reactor. 3.

Aspen Plus Simulation Details and Assumptions

Peng–Robinson equation (PR) is used to estimate all physical properties for the gasification and downstream unit operations. The SOLIDS property option is used for the coal crushing and screening section. The enthalpy model for both COAL and ASH, the nonconventional components, is HCOALGEN and the density model is DCOALIGT. The HCOALGEN model includes a number of empirical correlations for heat of combustion, heat of formation and heat capacity. All other values used were retrieved from the Aspen Plus database. For these simulations the H2/CO ratio of the synthesis gas feed to the F-T reactor will be set at 2.0 since the F-T model reactor is assumed to use a cobalt catalyst. The temperature of the SHR is set to be 750 ºC and the SMR is at 850 ºC for the simulation results shown below. The F-T reactor is operated at 220 ºC. The operating pressure is set at 400 psi for all the reactors. The average operation condition for the model is a H2O/Coal mass ratio of 2.1and a H2/C molar ratio of 1.0. This was selected to be consistent with the previous simulation results of this process [6]. Material and energy balances were accounted for and solved for every process unit, although no detailed chemical kinetic models were considered in the simulation. Detailed operation units used in the simulation of the CE-CERT process is given in Table 1. Table.1 Representive unit operations used in the simulation

Unit operation

Aspen plus model

Specifications

Feedstock crushing

Crusher

Rigorous simulation of particle size distribution

Feedstock particles screening

Screen

Rigorous simulation of the separation efficiency of the screen

Feedstock gasification

RGIBS

Specification of the possible products: H2O, H2 ,Cl2, HCl, C, CO, CO2, CH4, COS, H2S, CS2

Regenerator

Rstoic

Rigorous simulation of char combustion

Solid removing

Sep

Simplified simulation of gas/solid separation

Warm gas cleanup

Absorber, Rstoic

Rigorous simulation of the H2S, trace metal chloride removal

SMR furnance

Rstoic

Rigorous simulation of fuel gas combustion

H2 separation

Sep, Split

Simplified simulation of gas separation and split

FT reactor

RYield

Empirical simulation of F-T production distribution

HRSG

Counter current multiple

Rigorous simulation of the steam cycle with heat

stream heat exchanger

recovery

CO2 Recovery

Absorber, Rstoic

Rigorous simulation of the CO2 amine unit

Gas and steam turbines

Compressor

Calculate power produced

and

The main assumptions made in the process simulation are listed as below. 1.

No tar generated in the steam hydrogasification process.

2.

Process is steady state and isothermal, no heat and mass loss during the whole process.

3.

The product chain growth rate in F-T reaction is 0.9 with one single syngas pass.

4.

The byproducts such as sulfur and ammonia are not considered as credits.

In F-T reactor, the -(CH2)- is a basic structure for hydrocarbons with long chain. A main characteristic regarding the performance of the F-T synthesis is the process liquid selectivity. The liquid selectivity is determined by chain growth probability. This is the chance that a hydrocarbon chain grows with another -(CH2)- group instead of terminating. A high liquid selectivity (or C5+ selectivity: SC5+) is necessary to obtain a maximum amount of long hydrocarbon chains. The yield in the C1–C4 range decreases with increasing SC5+ and any C1–C4 in the outlet may efficiently be used for power generation. The F-T reactor block used an external model called through a FORTRAN module. This external model was empirically developed by Hamelinck et al [10], to predict the selectivity of the F-T process and can be expressed as below.

S C5+ = 1.7 + 0.0024T + 0.088

[H 2 ] + 0.18([ H 2 ] + [CO]) + 0.0078 pTotal [CO]

(4)

Where, Sc5+ – Mass fraction of hydrocarbons in the product with 5 or more carbon atoms [H2] and [CO] - Concentrations of H2 and CO expressed as fraction of the feed gas T – Temperature (K) P – Pressure (bar) According to Hamelinck et al, a least sum of squares fit of the above model with proprietary data resulted in the following equation, which was also found to be in accord with experimental results on cobalt catalysts reported [3]. This equation has been used to simulate the F-T reaction in this paper. The F-T liquids produced in the reactor is naphtha, diesel and wax. Waxes are finally converted into high quality diesel by hydrocracking. The products of hydrocracking are diesel (80%), naphtha (15%) and gaseous compounds, such as methane, ethane, propane and butane (5%) by weight [11]. The whole simulation is controlled using FORTRAN routines (calculator blocks) and design specifications to reduce the number of independent specifications and to adjust automatically those associated variables, i.e. The dependent variables are automatically adjusted when independent input variables are modified by the a calculator block or a design specification. The main functional relationships of the simulation are: the amount of H2O input as a function of the feedstock mass input, the amount of H2 input as a function of the carbon in the coal, the recycled H2 depends on the H2O input. The coal used in the simulation is Utah bituminous coal. The physical properties of the coal are given in Table 2. A heat exchange net is built based on maximization of heat recovery [12]. The results from Aspen Plus are directly imported into an excel worksheet and the mass and energy balance are calculated using excel. 4.

Simulation results

The simulations are based on a 4000 TPD coal plant with 65% CO2 capture to product F-T products and electricity using CE-CERT steam hydrogasification technology with H2/CO ratio of 2.0. The plant is operated under a capacity factor of 90%. The schematic diagram of the simulation process is shown in Fig 2. The total plant performance is compared with the simulation results in

Table.2 Feedstock analysis of Utah bituminous coal Ultimate Analysis Carbon

Dry, %

As Received, %

68.85

58.40

Hydrogen

4.74

4.02

Nitrogen

1.04

0.88

Sulfur

1.18

1.39

Ash

10.57

9.94

Oxygen

11.39

13.43

Proximate

Dry Basis, %

Moisture

As Received, %

--

15.18

Ash

10.55

8.95

Volatile Matter

40.00

33.93

Fixed Carbon

49.45

41.94

3.07

2.89

12,077

10,244

Sulfur Btu Content

National Energy Technology Laboratory (NETL) report within the Department of Energy (DOE). NETL has reviewed the CE-CERT process through a Cooperative Research And Development Agreement (CRADA) by validating the equilibrium model, process flow sheet for producing F-T products using CE-CERT technology with H2/CO ratio of 1.0 in 2008. The performance comparison of the two plants is given in Table 3. Table.3 Performance of 4000 TPD coal plants using different H2/CO ratio H2/CO ratio

2.0

1.0

Plant input (TPD)

4000

4000

Net Plant Power (MW)

74

107

9,990

7,143

50.25%

53.40%

CO2 Recovered (lb/hr)

321,742

306,507

Total Carbon fixed in the F-T liquids

37.10%

28.80%

F-T Production (BPD) Net Plant Effective Thermal Efficiency (HHV)*

*-ETE=(Higher Heating Value of F-T products + Electrical equivalent)/Thermal input

Fig.2 Schematic diagram of CE-CERT simulation process 5.

Conclusion A process simulation based on CE-CERT steam hydrogasification technology for co-production of F-T products and

electricity was developed using Aspen Plus simulation. The process includes an internal heat regenerative step where char from the gasifier is sent to a combustor and recirculation of sand is used to heat the main reactor. Detailed simulation methods and operation conditions were introduced. H2O/Coal mass ratio of 2.1and H2/C molar ratio of 1.0 was used as optimal operation condition in the process with Utah bituminous as feedstock. The overall process performance for a 4000 TPD coal plant with 65% CO2 capture were determined. Based on the simulation results, the plant had a F-T liquids production of 9,990 BPD along with 74MW electricity available for export to the grid with a process thermal efficiency of 50.25%. The total carbon fixed in the F-T liquids was 37.1% of coal carbon.

Acknowledgement Thanks to the NETL for their dedicated CRADA support and partial funding support from Viresco Energy, LLC.

Reference 1.

Ifp,

Innovation

Energy

Environment.

Crude

oil

supply

and

oil

availability

estimations.

See

also:

http://www.ifp.fr/IFP/en/files/cinfo/IFP_Panorama05_06-CarburantsalternatifsVA.pdf 2.

Federal Reserve bank of St. Louis. Advancing economic knowledge through research & data. Oil price available.

See also:

http://research.stlouisfed.org/fred2/series/OILPRICE/98/10yrs. 3.

JR. Anderson and M. Boudart, Berlin, Dry ME, The Fischer–Tropsch synthesis, Catalysis: science and technology, Germany, Springer, 160-253, 1981.

4.

JM. Norbeck and CE. Hackett, United States patent number 7, 208,530 B2

5.

SK. Jeon, CS. Park, CE. Hackett and JM. Norbeck, Characteristics of steam hydrogasification of wood using a micro-batch reactor, Fuel, 86, 2007, 2817-2823.

6.

CS. Park, SP. Singh and JM. Norbeck, Steam hydrogasification of carbonaceous matter to liquid fuels, 24th Annual International Pittsburgh Coal Conference, Johannesburg, South Africa, 2007

7.

ASK. Raju, CS. Park and JM. Norbeck, Baseline technical and economic assessment of a small scale steam hydrogaisification process with Fischer-Tropsch liquids facility. 25th Annual International Pittsburgh Coal Conference, PA, USA, 2008.

8.

Masuo Hasegawa and Junzo Fukuda, Gasification of solid waste in a fluidized bed reactor with circulating sand, Conservation & Recycling, 3, 1979, 143-153.

9.

ASK. Raju, CS. Park, JM. Norbeck, Synthesis gas production using steam hydrogasification and steam reforming, Fuel Processing Technology, 90, 2009, 330-336.

10. CN. Hamelinck, APC. Faaij, H. Uil and H. Boerrigter, Production of FT transportation fuels from biomass; technical options, process analysis and optimization, and development potential, Energy, 29, 2004 11. Tijmensen MJA, et al. Exploration of the possibilities for production of Fischer-Tropsch liquids and power via biomass gasification. Biomass and Bioenergy, 2002, 23:129-52. 12. The Cost and Performance Baseline for Fossil Energy Power Plants, Volume 1: Bituminous Coal and Natural Gas to Electricity” DOE/NETL 2007/1281 (May 2007).

Further development of the PSRK model for the prediction of the Vapor-Liquid-Equilibria of Direct Coal Liquefaction System at High Temperatures and High Pressures Xue-feng Mao*, Shi-dong Shi , Wen-bo Li, Zhen-nan Gao （China Coal Research Institute, 100013, Beijing, China） Abstract： Since Huron and Vidal (1979) developed the basic idea of so called G E mixing rules, similar models have been proposed by different authors. The aim of all recent developments of G E mixing rules is to combine the successful GE models or group contribution methods with equations of state to enable the description of Vapor-liquid equilibria at high temperatures and pressures including supercritical compounds. The system involved in direct coal liquefaction was a three-phase state of solid, liquid, and gas phases at high temperatures and high pressures. One of the challenges in the analysis of the Coal liquefaction process was due to the fact that the contents of both hydrogen and methane are very high and both of them were in the supercritical state. Besides, when small, spherical hydrogen and methane molecules were mixed with liquefaction oil, the huge difference between their molecule weights causes extreme asymmetry of the system, bring about the difficulty in analysis and modeling of the liquefaction process. To Predict the vapor-liquid equilibria of direct coal liquefaction system at high temperatures and high pressures, a group contribution equation of state called PSRK was proposed , wherein the group contribution equation of state PSRK(predictive Soave-Redich-Kwrong) as suggested by Holderbaum and Gmehling (1991) combines the mod UNIFAC model (Hansen et al.1991) with the SRK equation of state. In this work, the range of applicability of the PSRK method was extended by the introduction of additional gases and the determination of missing interaction parameters between the following gases: CH4, CO2, CO, H2S, H2, H2O and the original UNIFAC structural groups. A computational method to solve the flash model for coal liquefaction system at high temperatures and high pressures has been developed. The numerical code has two-cycle iteration, including the inner cycle of β iteration and the outer cycle of K iteration, and the results of fast convergence can be achieved. Generally, after about 5 times of iteration (outer cycle), the program can reach the convergence precision. Based on the model of PSRK, the composition of 24 components of the gas-liquid phase and the gas-liquid equilibrium constant has been calculated. Calculated results were consistent with the literature results and the vapor-liquid equilibria for the coal liquefaction reactor were calculated. The comparison of the results showed that the capability of the PSRK model to describe and predict the vapor-liquid equilibria of coal liquefaction systems at high temperatures, high pressures, strong asymmetry and strong polar.

1. Introdcution Since the pioneer work of Vidal[1],who matched the excess Gibbs free energy GE,from an equation of state (EOS) at infinite pressure with that from an existing GE model,the so-called EOS/ GE method has received great attention. One of them is the PSRK model proposed by Holderbaum and Gmethling [2],wherein the GE-mixing rule the UNIFAC model[3] is combined with the SRK equation of state, the PSRK method is one of the most successful models,whose wide practical applicability has identified by many investigations due to its good predictive

accuracy and the large parameter matrix.However,as pointed out by many researches[4,5-7]. A lot of attempts have been made to improve it to allow PSRK model to describe complex systems accurately over wide temperature and pressure ranges[8-12], Among the various efforts, it does not work well for highly asymmetric systems,For example, The system involved in direct coal liquefaction was a three-phase state of solid, liquid, and gas phases at high temperatures and high pressures. One of the challenges in the analysis of the Coal liquefaction process was due to the fact that the contents of both hydrogen and methane are very high and both of them were in the supercritical state. Besides, when small, spherical hydrogen and methane molecules were mixed with liquefaction oil, the huge difference between their molecule weights causes extreme asymmetry of the system, bring about the difficulty in analysis and modeling of the liquefaction process. To Predict the vapor-liquid equilibria of direct coal liquefaction system at high temperatures and high pressures, a group contribution equation of state called PSRK was proposed. In this work, the range of applicability of the PSRK method was extended by the introduction of additional gases and the determination of missing interaction parameters between the following gases: CH4, CO2, CO, H2S, H2, H2O and the original UNIFAC structural groups. 2. The Research Content The direct liquefaction was carried out at 0.01t/d at BSU at CCRI using Shenhua Shangwan bituminous coal. In order to get the vapor-liquid equilibrium constants under the real liquefaction conditions, the Pseudo-components and existing gas,such as H2 and C2H6 were combined together ot establish the coal lquefaction oil flash distillation system. By using the flash-distillation system,the Vapor-liquid equilibrium composition and distribution were calculated at different reaction temperature and pressure. 24 components were studyed as the research object,include 14 narrow boiling-range fractions and 10 Light Hydrocarbons gas. The resulted liquid product was separated as 14 narrow boiling-range fractions (IBP～110℃、110～150℃、150～180℃、180～200℃、200～220℃、 220～240℃、240～260℃、260～280℃、280～300℃、300～320℃、320～340℃、340～360℃、 360～380℃、380～450℃) successively by atmosphere distillation. 10 Light Hydrocarbons gas : H2、CH4、CO、CO2、C2H4、 C2H6、H2S、C3H8、C4、H2O 3. Thermodynamic model 3.1 The PSRK model The PSRK model was proposed by Holderbaum and Gmehling[2],which is a predictive SRK EOS with the MHV1 mixing rule[13] coupled with the UNIFAC model. And the constant A0=-0.64663 in the PSRK mixing rule is basically calculated using experimental liquid volumes in the saturated state of a large number of substances at atmospheric pressure. Details of the PSRK model are as follows. 3.1.1 The EOS The SRK EOS[14] is adopted in the PSRK model in the PSRK model:

p

RT a  v  b v (v  b )

(1)

With the pure component parameters ai and bi obtained from the critical data Tc and Pc of the pure components

ai  0.42748

R 2Tci2 f (T ) pci

(2)

bi  0.08664

RTci pci

(3)

Enables the correct reproduction of pure component vapor pressures with the help of the Mathias –Copeman parameters c1,i c2,i and c3 i fitted to experimental vapor pressure data: 1

f (T )  [1  (0.48  1.574i  0.176i 2 )(1  Tr 2 )]2

(4)

or 1

1

1

f (T )  [1  c1,i (1  Tr ,i 2 )  c2,i (1  Tr ,i 2 )2  c3,i (1  Tr ,i 2 )3 ]2 1

f (T )  [1  c1,i (1  Tr ,i 2 )]2

Tr ,i  1

Tr ,i  1

(5) (6)

For nonpolar substances,Eq(4) is good enough,however,for polar substances,Eqs.(5) and (6) proposed by Mathias and Copeman[15] are much more accurate. Eqs.(5) and (6) are used in this work except for those large molecules,whose c1,c2,c3 cannot be found in the papers of Holderbaum and Gmehling and Fischer and Gmehling. In this case, Eq(4) is adopted.The critical properties and the acentric factors for those molecules whose values cannot be found in the book of Reid et al.[16] are listed in Table 1. Table 1 critical parameters and basic physical properties of the narrow boiling-range fractions Component

Tc（K） Pc（MPa）

i

Component

Tc（K）

Pc（MPa）

i

H2

33.2

1.3

0

150～180

649.9

31.718

0.342

CH4

111.6

4.6

0.008

180～200

685.1

32.201

0.370

CO

132.9

3.5

0.049

200～220

708.6

30.915

0.381

CO2

304.1

7.38

0.225

220～240

728.4

28.506

0.392

C2H4

282.4

5.04

0.095

240～260

747.9

26.372

0.417

C2H6

305.4

4.88

0.098

260～280

768.7

24.842

0.454

H2S

373.2

8.94

0.100

280～300

792.7

24.341

0.489

C3H8

369.8

4.25

0.154

300～320

816.0

23.707

0.535

C4

408.2

3.65

0.195

320～340

840.3

23.371

0.582

H2O

647.3

22.12

0.344

340～360

863.8

22.888

0.645

IBP～110

524.9

30.834

0.217

360～380

889.2

22.748

0.714

110～150

590.1

26.739

0.284

380～450

934.2

20.457

1.034

Applying Eos to mixtures,the parameters f (T ) and b can be calculated using the PSRK mixing rule. Therefore, the pure component parameters ai , bi and the excess Gibbs energy at a reference state( G0E ) are required. At the reference state (the liquid at atmospheric pressure) an optimized ratio of the inverse packing fraction ,the following relation is obtained. 3.1.2 The Mixing rule

The MHV1 mixing rule[14]，so-called EOS/GE mixing rule，is used in the PSRK model:

GE a RT a  b  0   xi i  bi A1 i  A1

 b 

 x ln  b  i

i



i

(5)



b   xi bi

(6)

i

Where a and b are the mixture co-energy and co-volume paramters，resperctively and GE is the molar excess Gibbs free energy,which can be calculated from a GE model.The constant A1 is set to be -0.64663，which is little different from the value in the original MHV1 mixing rule. The fugacity coefficient is given by

ln i 



bi b  b  ( Z  1)  ln  Z (1  )    ln(1  ) b V  V 

(7)

ai 1 b bi   ln  i  ln   1  A1  bi b  bi RT

(8)

3.1.3 The Modified UNIFAC model For the modified UNIFAC model, interaction parameters between 45 subgroups and 21 main solvent groups have been determined by Lasen et al.(1987). In the present work we have added 10 new groups for systems involving gases. The group definitions for these new groups together with the structural parameters Rv and Qv (Sander et al.,1983) are listed in Table 2. Table 2 molecular Rv and Qv values and the main-group (MG) Definitions for the Gases Component

MG

Rv

Qv

H2

22

0.8320

1.1410

CH4

28

2.2440

2.3120

CO

25

2.0940

2.1200

CO2

26

2.5920

2.5220

C2H4

30

3.1482

2.9700

C2H6

31

3.6044

3.3920

H2S

27

2.3330

2.3260

C3H8

33

4.9532

4.4720

C4

34

6.3020

5.5520

H2O

35

0.9200

1.400

The modified UNIFAC parameters (aij)d etermined in this work for the interactions between the modified UNIFAC main groups 1-7, 9-10 (Larsen et al., 19871, and the 13 gases (main group 22-34) are given in Table III. In the expression used to describe the temperature dependence of the interaction parameters in modified UNIFAC we use only two terms,i.e.

aij  aij ,1  aij ,2 (T  T0 )

(9)

where To is a reference temperature, To = 298.15 K, and i and j represent the main groups. Further, all gas-gas interaction parameters have been assigned a value of zero. In Table 3 (a) we have given supplementary values of the interaction parameters between the CH2 group of an alcohol and the gases H2 and N2. It was not possible to describe the VLE behavior for N2 and H2 by using the same values for the interaction parameters (a CH3 of the CH2

groups resent in alkanes and in alcohols. The same problem appeared with the GC-EOS model for which special "water-soluble" CH2 groups in alcohols and ketones were introduced. The CH2 group present in an alcohol is here denoted CH2′.The a

CH2,gas values

in Table 3 should be used

when the CH2 group is in an alcohol molecule. Table 3 (a) Modified UNIFAC interaction Paramaters ( aij First Row Gives aij ,1 and Second Row aij ,2 (Eq ) i

j 22

25

26

27

28

30

31

33

34

1

509.0

58.78

123.9

188.3

-109.4

-133

-86.55

35.48

-22.08

CH2

1.440

0.000

-0.4065

0.0251

-0.8640

-0.8723

-0.7601

-0.0620

-0.5395

2

527.6

na

-28.36

na

na

-115.9

-173.9

-140.1

171.1

C=C

0.3671

na

1.169

na

na

0.0569

9.200

0.7277

0.5866

3

188.3

141

134.1

326.9

-3.522

214.9

1.951

2.323

2.323

ACH

3.334

-1.770

1.738

0.805

-0.4008

0.00

-0.2326

-0.6311

-0.6311

4

432.1

446.4

261.7

425.8

503.1

802

514.5

469.3

524.7

OH

-0.1257

-2.385

5.381

6.850

0.000

-5.435

-3.574

-4.203

-1.664

5

198.9

-19.85

-126.6

-132.8

-55.3

-70.84

-37.64

-67.57

-53.35

CH3OH

-2.138

-0.9319

-0.2024

0.000

-1.104

-0.8463

-0.8812

-0.7372

-0.8700

6

949.9

494

226.6

253.9

499.2

346.5

405

361.8

331.8

H2O

-0.3100

0.1390

-0.2410

-0.3050

-0.2550

-0.3326

0.0930

0.1271

-0.2550

7

273.7

na

81.51

na

769.5

216.2

-53.51

-9.088

na

CH2CO

-6.293

na

0.000

na

0.00

0.000

0.000

-1.490

na

9

319

159.0

-80.55

-0.1893

na

na

-80.97

na

na

CCOO

-5.3620

-1.039

-0.5914

0.1456

na

na

0.000

na

na

10

341.3

57.04

117.7

152.4

177.5

na

308.5

na

na

CH2O

-7.130

-3.694

5.759

-2.750

-1.502

na

0.000

na

na

9

10

(b) Modified UNIFAC interacton Parameters ( aij ) i

j 1

2

3

4

5

6

7

22

-228.1

-227

-54.49

1746

1540.90

1586

273.7

319

341.3

H2

-1.304

-1.305

-1.545

5.153

-5.567

3.924

-6.293

-5.362

-7.130

25

4.278

17.54

141.0

1170

1368

1455

na

159.0

57.04

CO

0.000

0.000

-1.770

-9.344

-2.928

-2.906

na

-1.039

-3.694

26

-55.69

-47.16

-16.56

583.3

727.9

1067

-0.0172

153.0

82.87

CO2

-0.4904

0.6526

-1.928

-4.940

-1.331

-0.4180

0.00

0.4509

-2.877

27

-93.27

na

-196.2

594.7

800.3

753

na

-0.1893

152.4

H2S

-0.6816

na

-0.7045

3.480

0.000

-0.8809

na

0.1453

-2.750

28

118.0

na

148.7

979.7

1532

1608

-11.9

na

198.8

CH4

0.3910

na

-0.2084

0.000

-4.137

-2.059

0.000

na

0.5265

30

149.1

152.7

-144.5

802

672.6

1354

76.87

na

na

C2H4

0.5241

0.2693

0.000

-5.453

-2.160

-1.542

0.000

na

na

31

99.22

185

70.69

663.1

760.9

1529

579.2

501.6

138.6

C2H6

0.4350

-5.864

-0.1656

0.0349

1.374

-3.081

0.000

0.000

0.000

33

-27.08

234.9

84.60

561.1

1138.0

1644

115.1

na

na

C3H8

-0.1935

-1.321

0.3180

-0.9400

1.18

-3.55

3.014

na

na

34

31.77

-88.39

84.60

524.7

689.8

1606

na

na

na

C4

0.4364

-0.4392

0.3180

-1.664

4.078

-2.221

na

na

na

The original UNIFAC model was adopted as the required GE model in the PSRK model,however, for the interaction parameters of gas-containin group pairs temperature dependent interaction parameters were intorduced,and the UNIFAC expression

nm  exp(

anm ) T

(10)

anm  bnmT  cnmT 2 ) T

(11)

Is replaced by

nm  exp(

Eqs.(1)～(11) constitute the PSRK model,and a variety of group interaction parameters are available,which makes the model have a good applicability. However, this temperature dependence is only used if necessary,e,g. for interaction parameters between gases and water,where a large temperature range is covered,or if a strong temperature dependence of the phase equilibrium behavior (e.g. for the Henry coefficients) is observed. With PSRK ,the results of the UVIFAC group contribution method can be reproduced,and furthermore ,the model can be used at higher temperatures and pressurres even at super-critical conditions. The optimization of a large number of interaction parameters between UNIFAC and new PSRK main groups was performed using the vapor-liquid equilibrium ,gas solubility and critical data sotored in the Dortmund Data Bank (DDB).These developments enable the calculation of the phase equilibrium behavior and other thermodynamic properties for various of systems[17～18]. Up to now,parameters for the 50 original UNIFAC main groups and 31 new PSRK groups (exoxy group and 30 different gases,such as CO2,CO,O2,N2,CH4,etc.)were established.As for the pure component values,the supplemtary material contains all available interactions parameters for the PSRK/UNIFAC method. The current status of the parameter matrix is shown in Fig.1.

Fig.1 Current status of the PSRK group interaction parameter matrix. 4. Results and discussion In most cases only two binary interaction parameters(aij,aji) were fitted. In a few cases ,linear(bij,bji) or even a quadratic(cij,cji) temperature dependent paameters were required to describle the correct temperature dependence of the phase equilibrium behavior,e.g. for the system CH4+H2O.the obtained parameters for 34 group pairs are given in Table 2 and Table 3. By using the flash-distillation system,the Vapor-liquid equilibrium composition and distribution were calculated at reaction temperature and pressure (T=675K /P=19MPa) ,the result was shown in Table 4. Table 4 The Vapor-liquid equilibria Data of high temperature Separator (T=675K/ P=19MPa)

yi

Ki

xi

H2

0.1508

0.7796

5.1689

150～180

0.0112

0.0022

0.2000

CH4

0.0228

0.0734

3.2171

180～200

0.0119

0.0017

0.1465

CO

0.0024

0.0092

3.803

200～220

0.0163

0.0019

0.1194

CO2

0.0072

0.0064

0.8858

220～240

0.0503

0.0046

0.0914

C2H4

0.0003

0.0007

2.6133

240～260

0.04

0.0027

0.0681

C2H6

0.0098

0.0274

2.7777

260～280

0.0446

0.0023

0.0508

H2S

0.002

0.0036

1.7452

280～300

0.0565

0.0022

0.0381

C3H8

0.0085

0.018

2.1105

300～320

0.0878

0.0024

0.0276

Component

xi

yi

Ki

Component

C4

0.0031

0.0054

1.7551

320～340

0.1175

0.0023

0.0194

H2O

0.0038

0.0487

12.7421

340～360

0.0651

0.0009

0.0135

IBP～110

0.002

0.0011

0.5436

360～380

0.0482

0.0005

0.0096

110～150

0.0072

0.0022

0.3007

380～450

0.2306

0.0008

0.0033

So through the change of temperature, The Vapor-liquid equilibria Data of coal liquefaction reactor were shown in Table 5. Table 5 The Vapor-liquid equilibria Data of coal liquefaction reactor (T=728K/ P=19MPa)

xi

Component

Ki

yi

xi

Component

yi

H2

0.1905

0.7679

4.0307

150～180

0.0084

0.0025

0.2977

CH4

0.0249

0.0725

2.9059

180～200

0.0088

0.002

0.2311

CO

0.0026

0.0091

3.4733

200～220

0.0121

0.0024

0.1951

CO2

0.0101

0.0063

0.6258

220～240

0.0376

0.0059

0.156

C2H4

0.0003

0.0007

2.4942

240～260

0.0305

0.0037

0.1217

C2H6

0.0101

0.0271

2.6882

260～280

0.0348

0.0033

0.0961

H2S

0.002

0.0035

1.7662

280～300

0.0457

0.0035

0.0761

C3H8

0.0082

0.0178

2.1643

300～320

0.0744

0.0044

0.0586

C4

0.0029

0.0054

1.8771

320～340

0.1057

0.0046

0.0439

H2O

0.004

0.048

12.038

340～360

0.062

0.002

0.033

IBP～110

0.0016

0.0011

0.6747

360～380

0.0482

0.0012

0.0257

110～150

0.0056

0.0023

0.4132

380～450

0.2693

0.0029

0.0107

0.74 0.72 H2 Solubility，mol/kg

Ki

0.70 0.68 0.66 0.64 0.62 13.0

13.5

14.0

14.5

15.0

15.5

16.0

16.5

17.0

H2 partial pressure，MPa Fig. 2 H2 partial pressure -dependence curves of the hydrogen solubility at high temperature

0.95

H2 Solubility，mol/kg

0.90

0.85

0.80

0.75

0.70 280

300

320

340

360

380

400

420

440

460

480

Temperaturre，℃ Fig. 3 Temperature-dependence curves of the hydrogen solubility in high temperature Separator

Also, the solubility of hydrogen in oil of coal liquefaction at high temperature and high pressure has been obtained. Further the solubility of hydrogen linearly increases with the rise of partial pressure of hydrogen as well as the temperature. But the linear fitting in the curve of hydrogen solubility-partial pressure departs from the origin of coordinate, which demonstrates that the solubility of hydrogen in coal liquefaction oil does not strictly comply with the Henry’s law. 5. Conclusions A computational method to solve the flash model for a coal liquefaction system at high temperature and high pressure has been developed. The numerical code has two-cycle iteration, including the inner cycle of β iteration and the outer cycle of K iteration, and the results of fast convergence can be achieved. Generally, after about 5 times of iteration (outer cycle), the program can reach the convergence precision. Based on the model of PSRK, the composition of 24 components of the gas-liquid phase and the gas-liquid equilibrium constant has been calculated.

References [1] M.Huron, J. Vidal,New mixing rule in simple equations of state for representing vapour-liquid equilibrim of strongly non-ideal mixtures. Fluid Phase Equilibria 3 (1979) 255–272. [2] T. Holderbaum, J. Gehling, PSRK: A group contribution equation of state based on UNIFAC, Fluid Phase Eqnilibria 70 (1991) 251-265. [3] H.K. Hansen, P. Rasmussen, A.a. Fredenslund, M. Schiller, J. Gmehling, Vapor-liquid equilibria by UNIFAC group contribution: 5. revision and extension, Ind. Eng. Chem. Res. 30 (1991)2352-2355. [4] K.Fischer,J,Gmehling. Further development status and results of the PSRK method for the prediction of vapor-liquid equilibria and gas solubilities. Fluid Phase Equilibria 121 (1996) 185–206. [5] M.L. Michelsen, A modified Huron-Vidal mixing rule for cubic equation of state.Fluid phase Equilibria 60(1990) 213-219. [6] S.Dahl.M.L. Michelsen. High pressure vapour-liquid equilibrium with A UNIFAC Based equation of state. ALCHE J (1990) 1829-1836. [7] D.S.H.Wong,S.I.Sandler,A theoretically Correct mixing rule for cubic equations of state, ALCHE J .38(1992) 671-680. [8] Qingyuan. Yang, Chongli,A modified PSRK model for the prediction of the vapor-liquid equilibria of asymmetric systems, Fluid Phase Equilib. 192 (2001)103–120. [9] Jurgen. Gmehling, Jiding.Li, Kai.Fischer,Further development of the PSRK model for the prediction of gas solubilities and vapor-liquid-equilibria at low and high pressures Ⅱ, Fluid Phase Equilib. 141 (1997)113–127. [10] J. Li, K. Fischer, J. Gmehling, Prediction of Vapor-Liquid Equilibria for Asymmetric Systems at Low and High Pressures with the PSRK Model, Fluid Phase Equilibria submitted, 1997. [11] C. Boukouvalas, N. Spoliotis, P. Coutsikos, N. Tzouvaras, D. Tassio, Fluid Phase Equilib. 92 (1994) 75–106. [12] Jiding Li,Kai Fischer,Jurgen Gmehling. Prediction of vapro-liquid equilibria for asymmetric systems at low and high pressures with the PSRK model , Fluid Phase Equilibria 143 (1998) 71–82. [13] M.L. Michelsen, A modified Huron-Vidal mixing rule for cubic equations of state ,Fluid Phase Equilib. 60 (1990) 213–219. [14] Soave G．Equilibrium constants from a modified Redlich-Kwong equation of state[J]．Chem Eng Sic，1972， 42：381～387 [15] P.M. Mathias, T.W. Copeman, Extension of the Peng-Robinson equation of state to complex mixtures: Evaluation of the various forms of the local composition concept ,Fluid Phase Equilib. 13(1983) 91–108. [16] R.C. Reid, J.M. Prausnitz, B.E. Poling, The Properties of Gases and Liquids, 4th Edition, McGraw-Hill, Singapore, 1987. [17] S. Horstmann, K. Fischer, J. Gmehling, PSRK group contribution equation of state: revision and extension III, Fluid Phase Equilib. 167 (2000) 173–186. [18] J. Gmehling, K. Jiding Li, Fischer, Further development of the PSRK model for the prediction of gas solubilities and vapor-liquid -equilibria at low and high pressures II ,Fluid Phase Equilib. 141 (1997)113–127.

Arsenic and Mercury Removal by Using Iron Humate Prepared from Turkish Coal Based Humic Acid Hacer Dogan1, Murat Koral1, Tulay Inan1, Selahattin Anaç2, Zeki Olgun2 1

TUBITAK Marmara Research Center, Chemistry Institute, 41470 Gebze, Kocaeli TURKEY 2 TKI(Turkish Coal Enterprises)Hipodrom Cad. No:12 Yenimahalle 06330 Ankara

Abstract: Humic acid produced by TKI (Turkish Coal Enterprises) was used in the production of iron humate. Different iron sources (iron (II) sulfate, iron (III) sulfate and iron (III) chloride) were used in the ion exchange reactions. Iron (II) sulfate (FeSO4) was found to be most suitable iron source in terms of iron content and adsorption capacity for As, Pb, Ni and Cd. The Fourier transform infrared (FTIR) spectroscopy, thermal gravimetric analyzer (TGA) and scanning electron microscopy (SEM) were used to characterize the humic acid and iron humate samples. The efficiency of iron humate as adsorbent has been studied as a function of amount, contact time and initial arsenic (As) and mercury (Hg) concentration by a series of batch experiments. The adsorption capacity of iron humate for As and Hg was above 90 % and higher than that of humic acid. It was concluded that iron humate can be used as an effective sorbent for the removal of As and Hg. Its application on the sorption of cadmium (Cd), cobalt (Co) and nickel (Ni) was not successful.

1. Introduction Humic acids (HAs) are present in soils, natural waters, river, lake and sea sediments, peat, brown and brown-black coals and other natural materials as a product of chemical and biological transformations of animal and plant residues [1]. The principal properties of HAs and their subsequent potential applications depend strongly on their origin (source) as well as on the isolation procedure. HA is a chemically heterogeneous compound having different types of functional groups in different proportions and configurations. It contains carboxyl (-COOH), amine(-NH2), hydroxyl(-OH), and phenol(Ar-OH) functional groups [3]. Carboxyl and hydroxyl groups are capable of substituting their hydrogen atoms for ions of metals. When humic acids interact with multi valency metals, such as iron, zinc, copper and others they form new type of insoluble compounds [4,5]. Metal humates can be prepared in a relatively simple way by the precipitation of HAs with suitable metal compounds. Recent years, low cost and highly effective adsorban materials have been developed for heavy metal adsorption. Some of the reported low-cost sorbents include bark/tannin-rich materials, lignin, peat moss, iron-oxide-coated sand, leaf mould, seaweed/algae/alginate, dead biomass etc. Humic acid based materials are also shown as one of the best examples for low cost sorbents. Iron humate that is markedly less soluble in aqueous solutions has been used as a new low-cost sorbent for inorganic and organic pollutants [6-10].

Humic acid extracted from Turkish coal in Ilgın Konya by TKI was used without any purification in the ion exchange reactions. The effect of the iron resources was examined on the ion-exchange reaction and the ability of the reaction was evaluated in terms of the iron content in the final product and adsorption of the heavy metal mixture containing As, Pb, Cd and Ni in the amount of each 10 ppm. Adsorption experiments have been performed for arsenic (As) and mercury (Hg) removal and the kinetics of the metal adsorption has been investigated. 2. Materials and Methods 2.1. Humic acid characterization Humic acid sent by TKI was analyzed and the results are given in Table 1. Carboxyl groups, total acidity and the coefficient E4/E6 were then determined. The amount of carboxylic groups and the total acidity were determined using the calcium acetate and the barium hydroxide methods, respectively. The quantity E4/E6 is the ratio of absorbances at 465 and 665nm of humic acid solutions. Elemental analyses of carbon, hydrogen, nitrogen and sulfur were obtained with a Thermo Finnigan Flash 1112 Series EA elemental analysis instrument. E4/E6 ratio was found as 3,26 of humic acid obtained from coal. This value shows that humic acid has high aromatic groups.

2.2. Adsorbent preparation Humic acid solution (1 % wt.) in the glass jacketed reactor was prepared and converted to sodium humate with NaOH granules at 60°C and stirred for two hours. pH value of the solution was measured as 11.2. Then iron salt (iron III chloride FeCl3.6H20, iron II sulfate-FeSO4.7H2O, and iron III sulfate- Fe2(SO4)3.7H2O) was added to the solution at 60°C and it was stirred for 24 h. Resulting suspensions of Fe-HA (iron humate) complexes were separated by centrifugation and rinsed with distilled water until removal of chlorine and sulfate ions. Finally the solid complexes were dried at 70°C. The content of Fe in the dry Fe-HA was analyzed by ICP-AES. The solubility of adsorbent was studied as a function of pH.

Table 1. Properties of humic acid TKI Humic Reference Acid Humic Acid* Content (Wt. %) Moisture Ash (at 950°C) Organic matter Inorganic matter Humic acid Hexane extraction residue Methanol extraction residue Petroleum ether extraction residue Elemental Analysis (wt. %) C H N S O Elemental ratio N/C H/C O/C Functional Groups (meq/g) COOH Phe-OH Total acidity E4/E6

11.8 38.2 56.6 43.4 31.3 0.12 3.63 0.018 55.4 6.4 1.4 0.8 36

54.72 4.04 1.47 0.36 38.54

0.022 1.38 0.49

0.023 0.88 0.53

2.16 6.84 9.0 3,26

5.0 1.16 6.16

* International Humic Substances- IHSS- (http://ihss.gatech.edu/ihss2/index.html) Pahokee Peat

2.3. Batch adsorption experiments Batch adsorption experiments were carried out by shaking 1 g of the adsorbent with 25 ml of solution of As and Hg with desired concentration in PE bottles. The metal solution together with the adsorbent was agitated magnetically in the closed PE bottle for different mixing times. Then the solid phase was separated by centrifugation, the concentration of the metal in the solution was determined immediately by ICP-AES. The adsorption capacity was calculated using the following equations:

q(mg / g ) =

(Ci − Ct ) V m

where Ci and Ct are the concentrations of the metal ion in initial and final solutions respectively. V is the volume of the aqueous phase (mL), m is the weight of Fe-HA (g). The adsorption degree, AD, as a function of time was also determined from the experimental data using the following relationship: AD % = 1 − C(t)/C(i) *100

3. Results and Discussion

3.1. Iron humate Humic acid containing 43.4 wt. % inorganic matters was used without any purification in the production of humate. The potential acidity of the iron salt added to the humate solution lowers the pH, and a precipitate of chiefly iron humate. The change in the pH value with the addition of FeCl3.6H2O caused different iron content in Fe-HA samples (Table 2). Lower pH values gave higher iron content. The increase in the HA concentration decreased iron content in the final product with the addition of FeSO4.7H2O. Fe-HA samples obtained with the addition of FeSO4.7H2O and Fe2(SO4)3.7H2O have about 11 wt. % iron content. It can be concluded that FeSO4.7H2O and Fe2(SO4)3.7H2O salts are more appropriate in the formation of complexes. However, iron humate obtained from FeSO4.7H2O showed higher adsorption capacity for As, Ni, Pb and Cd mixture than others. FTIR studies (Fig. 1) also exhibited the differences between FeCl3.6H2O (Fe-HA1) and FeSO4.7H2O (Fe-HA 2) compared to TKI humic acid. Thermogravimetric analysis of Fe-HA sample showed less weight loss less than 200°C compared to TKI humic acid (Fig. 2). Table 2. Iron humate production conditions

FeHA-1 FeHA-2 FeHA-3 FeHA-4 FeHA-5 FeHA-6

HA Conc. % 1 1 1 1 1 1

FeHa-7 FeHA-8 FeHA-9

5 5 1

Sample Code

Fe source FeCl3.6H2O FeSO4.7H2O FeCl3.6H2O FeCl3.6H2O FeSO4.7H2O K4P2O7 and FeSO4.7H2O FeSO4.7H2O* FeSO4.7H2O Fe2(SO4)3.7H2O

* Amount of FeSO4.7H2O was taken two fold.

pH value

Fe amount in Fe-HA, wt.%

2.3 6.42 6.42 3.4 4.5 7.05

4.91 11.08 2.29 4,61 9.74 9.87

6.42 6.42 2.4

5.31 4.72 11.89

The solubility of FeHA-2 and the calcined form of the same sample at pH:11 were determined as 1.2 and 0.4 wt. %, respectively. Because of the low solubility at high pH, calcined iron humate samples were used in the adsorption studies. SEM photograph of FeHA-2 sample is depicted in Fig. 3.

Fig. 1. FTIR spectra of Fe-HA samples and TKI humic acid

Fig. 2. Thermogravimetric analysis of Fe-HA sample and TKI humic acid

Element CK OK Al K Si K KK Ca K Ti K Fe K

Weight % 26.68 33.66 8.17 11.00 0.81 0.36 0.32 19.00

Totals

100.00

Atomic% 41.16 38.99 5.61 7.25 0.38 0.17 0.12 6.30

Fig. 3. SEM micrograph of Fe-HA sample

3.2. Adsorption studies 3.2.1. Effect of contact time The effect of contact time on the adsorption of As and Hg ions on Fe-HA was investigated over time intervals from 2 up to 48 hours. Fig. 4 shows the adsorption capacity as a function of contact time. As seen from Fig. 4, it reached to a maximum at about 2 hours for Hg and about 24 hours for As. About 90% of the total metal ion sorption was achieved within 2 and 24 hours for Hg and As, respectively (Fig. 5).

4,5

Adsorption capacity, q (mg/g)

4 3,5 3 2,5 2 1,5 1

As

0,5

Hg

0 0

10

20

30

40

50

60

Tim e (hour)

Fig 4. Variation of adsorption capacity with contact time

120 2h

Adsorption Degree (%)

100

6 h 10 h 16 h 20 h 24 h

24 h 48 h 10 h

80

16 h

20 h

6h 2h

60

40

20

0 As

Fig. 5. Variation of adsorption percentage y with contact time

Hg

3.2.2. Adsorption isotherms Several equilibrium models have been developed to describe adsorption isotherm relationships, the two main isotherm models used in this work are the Langmuir and Freundlich models. The Langmuir isotherm described the monolayer coverage of adsorbate over specific homogeneous sites within an adsorbent. Linear form of the Langmuir model could be described by the following equation:

1 ⎛ 1 ⎞ 1 1 ⎟⎟ + = ⎜⎜ q e ⎝ K L qm ⎠ Ce qm Where Ce was the equilibrium concentration (mg/L), qe was the amount of metal ion adsorbed at specified equilibrium (mg/g), qm and KL were the Langmuir constants related to adsorption capacity and adsorption energy. In this study, Langmuir isotherm was applied to analyze relationship between As and Hg concentration and adsorption capacity of iron humate. As shown in Fig. 6, experimental data of iron humate were well fitted for As and Hg by the Langmuir plots. The Langmuir model effectively described the sorption data with all R2 values >0.99 (Table 1). Freundlich isotherm suggested that sorption energy exponentially decreased on the completion of sorptional centers of an adsorbent and described heterogeneous systems. The Freundlich adsorption isotherm was tested in the following linearized form:

1 ln qe = ln K f + ln Ce n Where Ce was the equilibrium metal ion concentration in solution (mg/L), qe was the amount of metal ion adsorbed at specified equilibrium (mg/g). Kf and n were the Freundlich constants characteristics of the system, indicating the adsorption capacity and adsorption intensity, respectively. As can be seen in Fig. 7, the Freundlich model well fitted for As and Hg. Because of the correlation coefficient, Langmuir isotherm was more applicable for As. Both isotherms are applicable for Hg.

4,5 4 3,5

1/qe (g/mg)

3 2,5 2 y = 40,653x + 0,0538 R2 = 0,9996

1,5 1 0,5 0 0

0,02

0,04

0,06

0,08

0,1

0,12

1/Ce (L/m g)

a) 0,45 0,4 0,35

1/qe (g/mg)

0,3 0,25 0,2 0,15 0,1

y = 40,69x - 0,0011 R 2 = 0,9998

0,05 0 0

0,002

0,004

0,006

0,008

0,01

0,012

1/Ce (L/mg)

b) Fig. 6. Langmuir adsorption isotherm of arsenic (a) and mercury (Hg) on Fe-HA

2 1,5 1

Lnqe

0,5 0 -0,5 -1

y = 0,9102x - 3,4589 R 2 = 0,9867

-1,5 -2 0

1

2

3

4

5

6

LnCe

a) 2,5

2

Lnq e

1,5

1

0,5

y = 0,9987x - 3,6946 R2 = 0,9998

0 3

3,5

4

4,5

5

5,5

6

LnCe

b) Fig. 7. Freundlich adsorption isotherm of arsenic (a) and mercury (Hg) on Fe-HA

Table 3. Isotherm parameters of As and Hg ion adsorption on Fe-HA Adsorption model Langmuir isotherm qm KL R2 Freundlich isotherm Kf n R2

isotherm Arsenic

Mercury

adsorption 18.59 0.01156 0.9996

-909.09 0.000027 0.9998

31.78 1.0986 0.9867

0.0248 1.0013 0.9998

adsorption

4. Conclusion Different iron sources (iron (II) sulfate, iron (III) sulfate and iron (III) chloride) were used in the ion exchange reactions. Iron (II) sulfate (FeSO4) was found to be most suitable iron source in terms of iron content and adsorption capacity for As, Pb, Ni and Cd. The solubility of the iron humate sample at high pH values (pH:11) was obtained to be as low as 0.4 wt. % as a result of the calcination treatments applied after obtaining Fe-HA. Calcined Fe-HA sample was successfully used for selective As and Hg removal and adsorption capacities were obtained to be over 90 % for both Hg and As in a period of 2 and 24 hours, respectively.

References 1. F.J. Stevenson, Humus Chemistry. Genesis, Composition, Reactions (2nd Edition ed.), Wiley, New York (1994). 2. J. Novák, J. Kozler, P. Jano , J. Cezíková, V.Tokarová and L. Madronová, Humic acids from coals of the North-Bohemian coal field: I. Preparation and characterisation Reactive and Functional Polymers. 47 (2001) 101-109. 3. S. Erdogan, A. Baysal, O. Akba, C. Hamamci, Interaction of Metals with Humic Acid Isolated from Oxidized Coal, Polish J. of Environ. Stud. 16 (2007) 671675. 4. Humates and Humic Acids. How do they work?http://www.simplicitea.com/humates_how_they_work.doc 5. E. Tipping, Cation binding by humic substances, Cambridege environmental chemistry series -12, Cambridge University Pres, (2002). 6. P. Janos, J. Fedorovic; P. Stanková; S. Grötschelová; J. Rejnek; P. Stopka, Iron Humate as a Low-Cost Sorbent for Metal Ions, Environmental Technology, 27 (2006)169-181. 7. P. Janoš, Iron humate as a multifunctional sorbent for inorganic and organic pollutants, Environ. Chem. 2 (2005) 31-34. 8. P. Janoš, L. Herzogová, J. Rejnek, J. Hodslavská, Assessment of heavy metals leachability from metallo-organic sorbent—iron humate—with the aid of sequential extraction test, Talanta 62 (2004) 497-501. 9. P. Janos, Sorption of Basic Dyes onto Iron Humate Environ. Sci. Technol. 37 (2003) 5792-5798. 10. P. Janos, V. Šmídová, Effects of surfactants on the adsorptive removal of basic dyes from water using an organomineral sorbent—iron humate, Journal of Colloid and Interface Science 291 (2005) 19–27.

Heavy Metal Adsorption of Turkish Coal Based Humic Acid/Epoxy Composites Emel Yıldız1, Hacer Dogan1, Murat Koral1, Tulay Inan1, Selahattin Anaç2, Zeki Olgun2 1

TUBITAK Marmara Research Center, Chemistry Institute, 41470 Gebze, Kocaeli TURKEY 2 TKI(Turkish Coal Enterprises)Hipodrom Cad. No:12 Yenimahalle 06330 Ankara Abstract: The objective of this study was to investigate the adsorption capability of Humic acid/epoxy based composites. Humic acid produced by TKI (Turkish Coal Enterprises) was used as a co-curing agent for epoxy resin system based on Bisphenol F. The stoichiometrical amount of humic acid as co-curing agent, Diethylene triamine (DETA) as curing agent and Bisphenol F based epoxy resin were mixed. The homogenous mixtures were cured at 190 °C into the preheated molds. The curing agent/humic acid compositions were optimized to investigate the effect of humic acid concentration to removal of As, Pb, Ni and Cd. The Fourier Transform Infrared (FTIR) Spectroscopy, Thermal Gravimetric Analyzer (TGA), Differential Scanning Calorimeter (DSC) and Scanning Electron Microscopy (SEM) were used to characterize the humic acid/epoxy based composite samples. 1. Introduction Humic acids (HAs) are present in soils, natural waters, river, lake and sea sediments, peat, brown and brown-black coals and other natural materials as a product of chemical and biological transformations of animal and plant residues [1]. The principal properties of HAs and their subsequent potential applications depend strongly on their origin (source) as well as on the isolation procedure. HA is a chemically heterogeneous compound having different types of functional groups in different proportions and configurations. It contains carboxyl (-COOH), amine(-NH2), hydroxyl(-OH), and phenol(Ar-OH) functional groups [3]. Carboxyl and hydroxyl groups are capable of substituting their hydrogen atoms for ions of metals. When humic acids interact with multi valency metals, such as iron, zinc, copper and others they form new type of insoluble compounds [4,5]. Recent years, low cost and highly effective adsorbents have been developed for heavy metal adsorption. Some of the reported low-cost sorbents include bark/tanninrich materials, lignin, peat moss, iron-oxide-coated sand, leaf mould, seaweed/algae/alginate, dead biomass etc. Humic acid based materials are also shown as one of the best examples for low cost sorbents. The immobilization of HAs on a support is a promising approach in order to investigation of adsorption characteristics [6-9].

The aim of the study is to prepare the immobilized HAs and to investigate the adsorption capacity of Humic acid/epoxy based composites. 2. Materials and Methods 2.1 Materials Diglycidylether of Bisphenol F (DER 354, Dow Chemical Company) was used as epoxy resin. Diethylene triamine (DETA) (Dow Chemical Company) was used as curing agent without further purification. The chemical structures of epoxy resin and DETA are shown in Scheme 1. The epoxy equivalent weight of DER 354 is 167-174 and amine value of DETA is 1626 mgKOH/g. The calculated ratio of DETA to DER 354 is 11.84 parts curing agent per hundred parts epoxy resin (parts curing agent per 100 parts epoxy resin will be abbreviated as pph amine). Humic acid was supplied by TKI and was used as co-curing agent. 2.2. Humic acid characterization Humic acid sent by TKI was analyzed and the results are given in Table 1. Carboxyl groups, total acidity and the coefficient E4/E6 were then determined. The amount of carboxylic groups and the total acidity were determined using the calcium acetate and the barium hydroxide methods, respectively. The quantity E4/E6 is the ratio of absorbances at 465 and 665nm of humic acid solutions. Elemental analyses of carbon, hydrogen, nitrogen and sulfur were obtained with a Thermo Finnigan Flash 1112 Series EA elemental analysis instrument. E4/E6 ratio was found as 3,26 of humic acid obtained from coal. This value shows that humic acid has high aromatic groups. Humic acid was dried at 70ºC for 8 hours before used. 2.3. The preparation of humic acid/epoxy based composites The epoxy resin used in this work is DER 354 , a bifunctional epoxy resin formulated from the reaction of epichlorohydrin and Bisphenol F. For the composite sample preparation, the stoichiometric calculations were made on the basis of the amount of amine hydrogen present in DETA. The humic acid was added, in stoichiometric proportions (0-100 w/w %) relative to the total amount of amine hydrogen in DETA to the mixtures and all formulations were based on 100 parts of epoxy resin. Epoxy resin, curing agent (DETA) and humic acid were mixed homogenously at room temperature. The resulting homogenous mixture was poured into a mold. The curing cycle was 2 hours at room temperature (25°C) and then 190°C for 4 hours. The cured samples were grounded and then extracted by Soxhlet extraction method by using chloroform in order to remove the uncured epoxy resin. The crosslinked densities of humic acid/epoxy based composite samples were given in Table 2.

Table 1. Properties of humic acid

Content (Wt. %) Moisture Ash (at 950°C) Organic matter Inorganic matter Humic acid Hexane extraction residue Methanol extraction residue Petroleum ether extraction residue Elemental Analysis (wt. %) C H N S O Elemental ratio N/C H/C O/C Functional Groups (meq/g) COOH Phe-OH Total acidity E4/E6

TKI Humic Reference Acid Humic Acid* 11.8 38.2 56.6 43.4 31.3 0.12 3.63 0.018 55.4 6.4 1.4 0.8 36

54.72 4.04 1.47 0.36 38.54

0.022 1.38 0.49

0.023 0.88 0.53

2.16 6.84 9.0 3,26

5.0 1.16 6.16

* International Humic Substances- IHSS- (http://ihss.gatech.edu/ihss2/index.html) Pahokee Peat

O O

OH O

O

O O

F3 C F3 C

CF3 n (a)

CF3

H2N

N H

NH2

(b) Scheme 1 Chemical structures of a- DER 354 epoxy resin, b- DETA as curing agent Table 2. The crosslinked densities of humic acid/epoxy based composite samples Sample 100 DER 354/0 HA/100 DETA

Crosslinked density (% by w/w) 98.56

100 DER 354/20 HA/80 DETA

98.79

100 DER 354/40 HA/60 DETA

97.24

100 DER 354/60 HA/40 DETA

95.88

100 DER 354/80 HA/20 DETA

92.14

100 DER 354/100 HA/0 DETA

91.56

2.4. The Characterization of the Composite Samples

Heat of Reaction Measurements of Composite Samples The heat of curing reaction of the Humic acid/epoxy based composite sample was determined using a Perkin Elmer Jade Differential Scanning Calorimeter (DSC) operated in a nitrogen atmosphere according to ASTM E 2160. All the measurements were performed in alumina pans with sample weights ranging from 4 to 10 mg. The sample was heated along with an empty sample holder at a heating rate of 10°C/min from room temperature to 400°C with a nitrogen flow rate of 30 mL/min. The total heat of curing reaction was estimated by the area of an exothermic peak in temperature scanning mode calculated on Perkin Elmer Jade DSC Software.

Glass Transition Temperature Glass transition temperatures of composite samples were determined by using a Perkin Elmer Jade DSC according to ASTM D 3417. DSC scans were run at a heating rate of 10°C/min from room temperature to 200°C under nitrogen purge at a rate of 30 cm3/min. The glass transition temperatures (Tg) of composites were obtained from the second heating after a quick cooling. Tg value was taken from the second scan as midpoint of the change in slope of the baseline. The analysis of the resulting curve was performed on Perkin Elmer Pyris 1 DSC Software. Thermal Gravimetric Analysis (TGA) TGA of the humic acid/epoxy based composite samples was performed using a Perkim Elmer Pyris 1 TGA according to ASTM D3850. A small sample (8-20 mg) of composite sample was placed into a ceramic sample holder and loaded into the heating chamber. The heating chamber was kept at room temperature for approximately five minutes to purge the chamber of ambient air. The sample was heated at a rate of 10°C/min from ambient temperature to 900°C in N2 atmosphere. The test equipment measured and recorded the sample weight over the course of the analysis. 2.5. Batch adsorption experiments Batch adsorption experiments were carried out by shaking 1 g of the adsorbent with 25 ml of solution of As, Cd, Ni and Pb with desired concentration in PE bottles. The metal solution together with the adsorbent was agitated magnetically in the closed PE bottle for different mixing times. Then the solid phase was separated by centrifugation, the concentration of the metal in the solution was determined immediately by ICP-AES. 3. Results and Discussion In order to investigate the effect of humic acid on the thermal degradation behaviour and glass transition temperatures of humic acid/epoxy networks, the parent and composite networks were subjected to TGA and DSC, respectively. Thermal decomposition behaviour of the Humic acid, neat epoxy resin and composite samples determined by TGA under N2 and are reported in Table 3. The thermal decomposition temperatures of composite samples are higher than neat epoxy network and Humic acid. As espected, the decomposition temperatures at 50% weight loss, the weight losses at 500ºC and residue values at 900ºC of composite samples increased as the increasing amount of Humic acid as co-curing agent. TGA thermograms are given in Figure 1.

Table 3. Thermal properties of humic acid/epoxy based composite samples. Sample Humic acid

5% Weight 50% Weight Residue at loss (°C) weight loss loss at 900°C (%) (°C) 500°C (%) 184.8 539.7 56.0 41.8

100 DER 354/0 HA/100 DETA

205.3

392.8

20.3

14.8

100 DER 354/20 HA/80 DETA

327.6

402.9

26.9

20.4

100 DER 354/40 HA/60 DETA

321.7

409.1

32.0

25.0

100 DER 354/60 HA/40 DETA

323.1

427.2

41.0

33.0

100 DER 354/80 HA/20 DETA

325.3

459.5

47.2

38.3

100 DER 354/100 HA/0 DETA

329.7

680.4

56.9

46.0

Figure 1. TGA thermograms of Humic acid, Neat epoxy resin and composite samples

The effect of humic acid concentration on the glass transition temperatures of composite samples were investigated using dynamic DSC experiments. All the samples have one glass transition temperature. The DSC results are presented in Table 4 and DSC thermograms are given in Figure 2. Table 4. The Glass Transition Temperatures of composite samples Sample

Tg (°C)

100 DER 354/0 HA/100 DETA

100.7

100 DER 354/20 HA/80 DETA

93.9

100 DER 354/40 HA/60 DETA

92.2

100 DER 354/60 HA/40 DETA

77.4

100 DER 354/80 HA/20 DETA

69.6

100 DER 354/100 HA/0 DETA

79.1

Figure 2. The glass transition temperature curves of neat epoxy resin and composite samples

As seen in Table 4 and Figure 2, the glass transition temperature values of composite samples decreased as the increasing amount of Humic acid. It is well known that the glass transition temperature of a polymer is closely correlated by the rigidity of the polymer backbone. Therefore, the glass transition temperature of neat resin cured by DETA only is higher than composite samples including Humic acid as co-curing agent. The dynamic DSC scans at 10 ºC/min heating rate for the curing reaction of neat epoxy resin and composite samples under N2 atmosphere are shown in Figure 3. From DSC thermograms, the onset temperature of curing (T 0), the exothermic peak maximum (Tmax), the final temperature of curing (Tf), and the heat of curing reaction are determined. As seen in Figure 3, there are two curing peak maximum, the first peak area between 50-150 ºC decreased as the decreasing amount of DETA, but the second peak area between 250-350 ºC increased as the decreasing amount of DETA. The structural characterizations of neat epoxy resin cured by DETA only and composite samples including Humic acid used as co-curing agent were performed by FTIR spectroscopy technique. The FTIR spectrum of epoxy resin cured by DETA and by Humic acid are given in Figure 4. The strong absorptive peak at 3435 cm-1 is due mainly to the stretching vibration of –OH group and there is no stretching absorption peak belongs to –NH group at 3304 cm-1 in FTIR spectrum of neat epoxy resin cured by DETA only. FTIR spectrum of epoxy resin cured by Humic acid shows a considerably smaller peak of epoxy ring at 910 cm-1 but there is no peak belongs to oxirane group in the FTIR spectrum of epoxy resin cured by DETA. It is verified by crosslinked density values of cured samples. And also there is no peak belongs to carbonyl group between 1725-1700 cm-1 in both of FTIR spectrum.

Figure 3. DSC thermograms of Neat epoxy resin and composite samples

100.0

Cured by DETA

95 90 85 80 75 70 65 60 55 %T 50 45 40 35 30

Cured by HA

25 20 15 10 3.2 4000.0

3600

3200

2800

2400

2000

1800 cm-1

1600

1400

1200

1000

800

600

450.0

Figure 4. FTIR spectra of epoxy resins cured by DETA and Humic acid

The effect of Humic acid content of composite samples on the removal of As, Cd, Ni and Pb ions are given at Figure 5. The uptake capacity for As, Cd, Ni and Pb ions depended on the Humic acid content of composite samples. As seen in Figure 5, the Humic acid/Bis F epoxy based composites are not effective sorbent for As removal. But they are effective for Pb removal and adsorption power depends on directly to the Humic acid content of composite samples.

Heavy metal content (ppm)

14 12 10

As

8

Cd Ni

6

Pb

4 2 0 0

20

40

60

80

100

Humic acid content (% w/w )

Figure 5. Removal of As, Cd, Ni, Pb by Humic acid/Bis F epoxy based composite 4. Conclusion Humic acid/epoxy based composites were used as sorbent in order to As, Cd, Ni and Pb removal. Humic acid produced by TKI (Turkish Coal Enterprises) was used as a co-curing agent for epoxy resin system based on Bisphenol F. DETA/humic acid compositions were optimized to investigate the effect of humic acid concentration on the adsorption capability for As, Pb, Ni and Cd. The uptake capacity for As, Cd, Ni and Pb ions depended on the Humic acid content of composite samples. Humic acid/Bis F epoxy based composites are effective for Pb removal and adsorption power depends on the Humic acid content of composite samples.

References 1. F.J. Stevenson, Humus Chemistry. Genesis, Composition, Reactions (2nd Edition ed.), Wiley, New York (1994). 2. J. Novák, J. Kozler, P. Jano , J. Cezíková, V.Tokarová and L. Madronová, Humic acids from coals of the North-Bohemian coal field: I. Preparation and characterisation Reactive and Functional Polymers. 47 (2001) 101-109. 3. S. Erdogan, A. Baysal, O. Akba, C. Hamamci, Interaction of Metals with Humic Acid Isolated from Oxidized Coal, Polish J. of Environ. Stud. 16 (2007) 671675. 4. Humates and Humic Acids. How do they work?http://www.simplicitea.com/humates_how_they_work.doc 5. E. Tipping, Cation binding by humic substances, Cambridege environmental chemistry series -12, Cambridge University Pres, (2002). 6. M. Klavins, L. Eglite, Immobilization of Humic Substances, Colloids and Surfaces A: Physicochemical and Engineering Aspects 203 (2002) 47-54 7. X. Wang, X. Shan, L. Luo, S. Zhang, B. Wen, Sorption of 2, 4, 6Trichlorophenol in Model Humic Acid-Clay Systems, Journal of Agricultural and Food Chemistry, 2005, 53, 3548-3555. 8. M. Klavins, L. Eglite, A. Zicmanis, Immobilized Humic Substances as Sorbents, Chemosphere 62 (2006) 1500-1506. 9. A. H. Rosa, A. A. Vicente, J. C. Rocha, H. C. Trevisan, A new Application of Humic Substances: Activation of Supports for Invertase Immobilization, Fresenius J. Anal. Chem. (2000) 368: 730-733.

Manuscript Not AVAILABLE

NUMERICAL SIMULATION OF SYNGAS PRODUCTION BY PARTIAL OXIDATION OF COKE OVEN GAS UNDER NON-PREMIXED CONDITION

Honggang Chen, Prof . ,National Engineering Laboratory for Biomass Power Generation Equipment, North China Electric Power University, Beijing 102206, China Email: [email protected] Kai Zhang, Prof. , National Engineering Laboratory for Biomass Power Generation Equipment, North China Electric Power University, Beijing 102206, China Email: [email protected] Hui Zhao Dr. ,State Key Laboratory of Heavy Oil Processing China University of Petroleum, Dongying 257061, China E-mail：[email protected] Yongfa Zhang, Prof. Key Laboratory of Coal Science and Technology of Shanxi Province and Ministry of Education Taiyuan University of Technology,Taiyuan 030024,China Email: [email protected]

Abstract: China is the largest coke producer in the world. The yield of coke oven gas (COG), a byproduct of the coking process, reached 1190 billion Nm3 per year in 2007. COG is a good feedstock for many chemical processes e.g. FT synthesis, methanol synthesis and ammonia synthesis, instead of only as a heating fuel. The main component of COG is hydrogen (~56-60 vol. %) but there are also other compounds such as methane (~25-30 vol. %), carbon monoxide, carbon dioxide, and nitrogen. There are several methods to convert the methane into hydrogen in order to fully make use of the COG resources to the larger extent. The non-catalytic partial oxidation of COG with oxygen is one of important routes for COG comprehensive utilization. In the present work non-premixed combustion of coke oven gas under low oxygen condition was investigated through computational fluid dynamic (CFD) simulation coupled with radiation heat transfer. The mathematical model was formulated to describe the fluid flow, heat transfer, mass transfer , and gas phase chemical reactions. The obtained model was numerical solved with finite volume method using commercial software FLUENT TM. Temperature profile and product distribution in the reactor were obtained from simulation. Calculated temperature profile indicates that oxygen react strongly with coke oven gas in region near the oxygen nozzle and the temperature in this region increase sharply. Significant influence of reactor wall temperature on product distribution could be found in simulation results. Higher selectivity of syngas could be archived at higher reactor wall temperature, while conversion rate of methane in coke oven gas was suppressed simultaneously. Reactor wall temperature should be treated as a key parameter and optimized well in this process. The simulation results are helpful the development and design of the COG non-catalytic partial oxidation reactor and process.

Modeling of Coal Drying in a Pneumatic Dryer Kyoungsoo Lim, Sangdo Kim, Soonkwan Jeong, Youngjun Rhim, Sihyun Lee*,

Clean Fossil Energy Research Center, Korea Institute of Energy Research Daejeon 305-343 South Korea *E-mal: [email protected]

ABSTRACT Operation of the pneumatic conveying system for drying coal was influenced by many parameters, in particular, gas velocity, gas temperature, coal residence time, and coal particle size. In this study, the pneumatic conveying drying system was simulated for drying wet coal. Numerical studies were conducted to examine the effect of these parameters on drying coal. As the gas temperature, the drying rate of wet coal was increased. In addition, coal drying increased with an increase of coal residence time. Particle size is one of the most important parameter on moisture removal efficiency of the pneumatic conveying dryer. The moisture removal increased in short drying time with a decrease in particle size due to the large surface area. KEY WORDS: Drying; coal; hot gas, moisture; residence time; gas temperature; particle size.

INTRODUCTION Pneumatic conveying drying is a combination of heat and mass transfer between hot gas and wet particles. This technique has been widely used in various industry processes, due to the reasons of maximum efficiency, minimum gas flows and power consumption, improved product quality and increased workplace safety. In pneumatic dryer system, a wet particle is rapidly dried as it comes into contact with hot gas, and the atomized feed sample travels concurrently with the drying-gas. Operation of the pneumatic conveying system for drying wet particles was influenced by many parameters, in particular, gas velocity, gas temperature, particle residence time, and particle size. Many experimental and numerical studies of a pneumatic conveying dryer have focused on an effect of these parameters on drying rate of wet particles. As for experimental approach, Yamazawa et al. (1972) presented an improvement of rice physical properties using a pneumatic conveying dryer. Tandano et al (1981) investigated pneumatic conveying dryer operation and clarified the drying time experimentally. Baeyens et al. (1995) presented a design procedure to predict moisture and temperature profiles in a pneumatic dryer. Rocha and Paixao (1996) developed a pseudo two-dimensional model for pneumatic drying and predicted axial and radial profile for a gas solid velocity and solids moisture content. Fyhr and Rasmuson (1997) presented a complex model for a pneumatic dryer considering a distribution of particle sizes for steam drying of wool chips. Levy and Borde (1999) studies the drying of wet PVC particle in a large-scale pneumatic dryer and compared the results with prediction of a numerical simulation taking into account particle shrinkage. More recently, many models for pneumatic conveying dryer have been developed and used to predict the behavior of a batch (Pelegrina and Crapiste, 2001; Narimatsu et al., 2004; Skuratovsky et al., 2004; Tanaka et al., 2006; Narimatsu et al., 2007). Pelegrian and Crapiste (2001) developed the model to analyze the effect of different variables as solid to air mass flowrates ratio, velocity and temperature of air and, size and shape of potato particles. Tanaka et al. (2008) investigated the influence of initial rice

particle size and moisture content distributions on the moisture propagation in a pneumatic conveying dryer. However, most previous studies had focused on drying the food particles. Although a few studies developed the models to analyze the coal drying in the pneumatic conveying dryer, the variables were limited. Levy et al. (1998) simulated a dispersed gas-coal solids flow, without mass or heat transfer in one-dimensional transport system. Ross et al. (2005) compared the experimental data for the final moisture content and gas temperature profile for a pressurized entrained flash dried at 800 oC and 10 bar to a one-dimensional mathematical model. This study primarily seeks to present the drying characteristics of wet coal particles in a pneumatic conveying dryer by exploring the effects of several variables numerically, such as coal particle diameter, gas temperature and dyer length, and so on. Numerical solutions are derived employing energy, momentum and mass balance using FLUENT program.

NUMERICAL METHOD The geometry and dimensions of calculation domain to study the influence of several important parameters on coal drying are shown in Fig. 1 and Table 1. There are three parts construing the calculation domain i.e., the vertical upstream pipe, U-bend and cyclone. The length of the vertical upstream pipe is 5.0 m for all variations. The radius of U-bend is 0.3 m and the horizontal pipe between the vertical upstream pipe and cyclone is 0.7 m. The diameter of cyclone body is 0.24 m and total height of cyclone is 0.8 m. The computational grids of the pipe and cyclone are approximately 61,500 cells. The grid was generated using Gambit, and the commercial FLUENT software was used for solving the set of governing equation. FLUENT provides flexibility in choosing discretization schemes for each governing equation. The discretized equation, along with the initial condition and boundary conditions, were solved using the segregated solution method to obtain a numerical solution. Using the segregated solver, the conservation of mass and momentum were solved sequentially and a pressure correction equation was used to ensure the conservation of momentum and the conservation of mass. The RNG k- model for turbulence model was employed to predict flow behavior in the pipes and cyclone, and wet combustion model was also employed for predict the evaporation and drying from wet coal particles.

Fiig. 1. Geomettry of the callculation dom main.

Table 1. Dimensions of the pneum matic dryer. Param meters Length of the verticcal upstream m pipe, Lv Diametter of the vvertical upsstream pipe, Dp Radius of U-bbend, Ru Inner R Length of the horizoontal pipe, Lhh D Diametter of the cyclone body, Dc Length of cyclone ccylinder, h1 Length of cyclone ccone, h2 Diametter of cyclonee exit, dc

Diimensions (mm) 5000 40 300 700 240 300 500 40

In the prresent investiigation, the steady state fl flow of wet soolid particless and a gas thhrough the vertical gassolid sysstem was described. The ssimulation w was conductedd consideringg the followinng assumptioons; 1) Gas and a solid phaase flow co-ccurrent in thhe upward diirection. 2) M Mass, momenntum, and heeat transfer occur onnly between tthe phases annd not within the particless. 3) Theree is no radiantt energy trannsfer betweenn the gas- sollid suspensionn and the surrroundings. 4) The soolid feed is uuniform in sizze and shape. The solid particles are sspherical.

5) The gas phase is considered as a mixture of gas and liquid vapor

RESULTS AND DISCUSSION The simulation of pneumatic conveying dryer was carried out at different drying conditions. Solid feeding was coal particles of 5.0 kg/hr at temperature Tcoal = 25oC and water content of coal X0 = 0.274 kgwater/kg-coal in most cases. The gas absolute humidity (Y0) was 0.007 kg-water/ kg-dry gas and gas velocity was 20 m/s. Effect of gas temperature, particle size and initial moisture content is discussed in this section. Gas temperature and absolute humidity profiles along the dryer for a typical drying simulation are shown in Figs. 2 and 3. Usually, after the entrance region, the velocity of coal particles is accelerated rapidly in the up-stream gas flow according to the momentum balance equation, and then reached almost constant velocity. During the residence of these coal particles in dryer, there are heat and mass exchange between particles and hot gas. The temperature of coal particles rises faster from its initial value up to the weltbulb temperature of the drying gas, while gas temperature decreases as shown in Fig. 3. As the initial gas temperature increases, the gas temperature drop is large and absolute gas humidity increases. It means that the drying rate of coal particles increases as the gas temperature increases because the water content of coal particles decreases and gas moisture content increases at the same rate. 200 oC 300 oC 400 oC 500 oC 600 oC 700 oC

Gas Temperature (oC)

800

600

400

200 Initial absolute humidity : 0.007 kg/kgdry air Initial coal particle diameter : 500 m Initial coal moisture content : 0.274 kg/kg coal 0

1

2

3

4

5

Dryer length (m)

Fig. 2. Gas temperature profiles at different initial gas temperature.

Absolute Humidity (kg/kgdry air)

0.05 200 oC 300 oC 400 oC 500 oC 600 oC o 700 C

0.04

Initial absolute humidity : 0.007 kg/kgdry air Inital coal particle diameter : 500 m Initial coal moisture content : 0.274 kg/kg coal

0.03

0.02

0.01 0

1

2

3

4

5

Dryer length (m)

Fig. 3. Absolute humidity profiles at different initial gas temperature. Fig. 4 shows the final moisture content of the dried coal product and moisture removal efficiency by evaporation. The moisture removal efficiency of coal particles in 500 m size is about 40% at gas temperature of 400 oC, and more than 90 % at 900 oC. The evaporation rate continuously increases and the moisture removal efficiency also linearly increases as gas temperature increases because of the higher driving force for heat transfer. However, the maximum temperature that the dried product can reach without having thermal damage limits initial gas temperature.

100 Initial moisture content : 0.35 kg water/kg coal Coal particle diameter : 500 m Dryer length : 5 m

0.25

80

0.20 60 Moisture content Moisture removal

0.15

40 0.10

Moisture Removal (%)

Moisture Content (kg water/kg coal)

0.30

20

0.05

0.00

0 100

200

300

400

500

600

700

800

o

Gas Temperature ( C)

Fig. 4. Moisture content and moisture removal efficiency at different initial gas temperature. The size of particles conveyed using a pneumatic system varies greatly, so it is important to study the influence of particle diameter on heat transfer behavior. The predicted change of gas temperature and absolute temperature at different particle size as function of dryer length is shown in Figs. 5 and 6. Simulation were carried out for particles in 100, 300, 500, 1000 m size As particle size decreases, the decline rate of gas temperature with dryer length is higher, with the corresponding absolute humidity increasing. The particles of 300, 500, 1000 m size continuously dry as they travel in dryer column. However, the particle of 100 m size dry up within 1.5 m in dryer length. In this simulation the initial moisture content of coal particles was set as constant value of 0.273 kg-water/kg-coal. Fig. 7 shows the final moisture content and moisture removal efficiency at different particle sizes. As particle size decreases, the moisture removal efficiency dramatically increases from 20% to 100%. An increase in particle diameter is the most dominant variable in the inter-phase momentum exchange coefficient. The

decrease in the inter-phase momentum exchange coefficient will result in an increase in slip velocity between gas an solid phases. Smaller particles have a higher surface area by unit weight of solid than larger particles, favouring the rate of heat and mass transfer between solid and gas. 500

Initial absolute humidity : 0.007 kg/kgdry air Inital gas temperature : 400 oC Initial coal moisture content : 0.274 kg/kg coal

Gas Temperature (oC)

450

100 m 300 m 500 m 1000 m

400

350

300

250

200 0

1

2

3

4

5

Dryer length (m)

Fig. 5. Gas temperature profiles at different particle sizes. 0.040

Initial absolute humidity : 0.007 kg/kgdry air o

Inital gas temperature : 400 C Initial coal moisture content : 0.274 kg/kg coal

Absolute Humidity (kg/kg dry air)

0.035

100 m 300 m 500 m 1000 m

0.030

0.025

0.020

0.015

0.010

0

1

2

3

4

5

Dryer length (m)

25

100

20

80

15

60 Moisture content Moisture removal

10

40

5

Moisture Removal (%)

Moisture Content (kg water/kg coal)

Fig. 6. Absolute humidity profiles at different particle sizes.

20 Initial moisture content : 0.274 kg water/kg coal Initial gas temperature : 400 oC Dryer length : 5 m

0 0

200

400

600

800

1000

0 1200

Particle Diameter (m)

Fig. 7. Moisture content and moisture removal efficiency at different particle sizes.

This result is consistent with observances made for other studies. Saravanan et al. (2007) carried out the simulation with two resin particle size to verify the effect of particle size on drying rate. They also described that the outlet solid moisture content increased with an increase in the solid particle size because of large particle surface area per unit mass of solid. Ross et al. (2005) compared the experimental data for the final moisture content and gas temperature profile for pressurized entrained flash drier at 800oC and 10 bar to mathematical model developed previously for the brown coal mill process. They conducted the simulation with five Yallourn lignite particle size. They presented that the small particles dried almost instantaneously (C=O and ether. The functional groups –OH and –COOH are available at the site of phenanthrene nucleus either at 9 position or at any other place. Position 10 is available for linkage with other unit of coal. >C=O group is present in a cyclic ring. Assuming this model structure of coal to be correct, oxidation of such a coal should lead to carboxylic acid formation up to a maximum extent of 5-6% because oxidation of methyl group and carbonyl group are expected to yield of –COOH groups. The actual experimental results obtained on oxidation of the same coal by 3N dilute nitric acid yielded 18% of –COOH groups. Hence this model cannot explain this experimental fact. If one assumes that besides these two positions, oxidation occurs at several other sites also then more solubility of coal in alkali / organic solvents on oxidation is expected to occur but this would lead to the formation of low molecular weight products on oxidation, which is not actually obtained. Hence a modification is needed in this coal structure so that the new model structure of coal can explain these experimental results also. Thus a modified model structure of coal has been proposed. Mr. Mazumdar has not spelt out anything about model structure of high sulphur coals of North Eastern Region. With the help of the data generated by Dr. Srivastava’s group on disposition pattern of sulphur in NE region coal structure, a model structure of high sulphur NE coal has been proposed.

INTRODUCTION A number of model structures of coal have been proposed word wide. Based on specific reactions yielding known products, some group of workers considered aromatic cluster as the nucleus of the coal.

Yet another group of researchers found out the presence of aliphatic and hydro-

aromatic structures associated with the aromatic cluster of coal. Since coal is not a exactly a polymer but constitute a number of macromolecule attached to each other by certain linkage, the same was also studied by various researchers. Keeping all the above researches in mind, it was concluded that coal is a condensed poly aromatic mass associated with aliphatic & hydroaromatic groups and these units are linked with ether and methylene bridges. In eighties, Indian scientist Mr. B.K. Mazumdar proposed the coal systematics from Lignite to Anthracite covering all the physical and chemical properties of different ranks of coal. It is well known that the elementary carbon and hydrogen are distributed in coal into three basic parts viz. aromatic, aliphatic and hydroaromatic. Similarly oxygen is distributed in the form of functional groups such as hydroxyl, carboxyl, carbonyl and ether. The other elementary part nitrogen exists in his coal model as aromatic, but sulphur have not been covered by the author in his proposed coal model. It has been found recently [1] that this model is not able to explain some of the experimental results and hence there is an urgent need to modify this model structure of coal. The present authors have taken an attempt to fill these gaps.

EXPERIMENTAL Oxidation of Samla Coal by 3N Nitric Acid : Proximate and Ultimate analyses of raw Samla coal and the demineralised Samla coal were done following IS methods. The methods for determination of functional groups such as –OH, --COOH and -CH3 have been described in an earlier publication made from this laboratory [1]. Oxidation of coal was done as follows. Samla coal was taken in a flat bottom flask and to this was added 3N Ntric Acid as an oxidizing agent for carrying out oxidation reaction at desired temperature for desired time. After completion of the reaction, the oxy coal was washed with distilled water and filtered. The residual mass i.e. oxy coal was extracted using polar organic solvent. After solvent removal, the remaining solid was taken out followed by its characterization. The above data are summarized in table-1.

Disposition Pattern of Sulphur in N.E.Region Coals : N.E. Region coal and resorcinol (hydrogen-donating compound) mixture was heated from room temperature at the rate of 10°C / min in the presence of hydrogen gas in a reactor up to 620°C. The H2S evolved at different temperature ranges were absorbed in cadmium acetate solution to estimate sulphur gravimetrically for each temperature range[2]. The coal particle size used for this purpose was 212 µ. The total sulphur in the coal was determined using a Leco Sulphur Determinator (Model SC-132 from Leco Corporation, USA). The pyritic sulphur was estimated [3] by refluxing coal with HNO3 and determining the iron content while the sulphate sulphur was determined [4] by refluxing coal with HCI and estimating the sulphur as barium sulphate.

RESULTS AND DISCUSSION

Table – 1 Summarizes the proximate, elemental and functional groups analyses of raw and demineralised Samla coal as well as of oxy coal and soluble part of oxy coal in polar organic solvent. On careful examination of the above table, it is observed that the ash content has been drastically reduced from 14.6% to 1% and the moisture content of demineralised coal is reduced to 1.7% from 8.6%. Except these, practically no change in Carbon, Hydrogen and nitrogen percentages (dmmf basis) have occurred and hence in functional groups also. It is observed from Table-1 that after oxidation of samla coal, the carboxylic acid functional groups have enhanced from 1.2% to 18.3%. This order of enhancement in carboxylic acid functional groups at the periphery of aromatic rings / have come mainly from 9-10 position [1,5] in angle ring aromatic nucleus like phenanthrene. Methylene and methyl groups present in this coal is only about 3-4% and the reduction of methyl group from this level to 1% level alone cannot explain [1] the generation of carboxylic acid functional groups of the order of 18.3%. The model structure of samla coal (80-83 %C) as proposed by Mazumdar [6] is as follows and shown in (fig-1).

On careful examination of the above model structure of coal, it is crystal clear that 9,10 position of each phenanthrene unit is blocked either by functional groups or by other linkages. It is known that the mechanism of oxidation and hydrolysis of phenanthrene in which 9,10 position are available , gives a product generating carboxylic and hydroxyl groups as follows [7,8].

In some cases, it is known that oxidation facilitates at benzylic and 9,10 position of

phenanthrene units yielding a product containing more number of carboxylic acid functional groups as follows [9].

In the present case, as also appeared in an earlier publication [10], it was observed that samla coal was easily oxidized by oxidizing agents. In order to explain the experimental result that – COOH group has enhanced from 1.2% to 18.8%, it is concluded that 9,10 position of phenanthrene of each unit is not blocked. It is either vacant in coal structure or present in the form

of 9,10 dihydrophenanthrene. Many of the model structure proposed such as that of Lander’s [11] and Gibson [12], are ‘’two dimensional” although the presence of folded ring such as 9-10 dihydrophenanthrene means that the structure are not flat as shown in fig-2. Pitt [13] has proposed models of coal structure based on 9,10- dihydrophenanthrene structures (fig-3) and that these molecules exist in a tangled state in coal. Heredy and Wender (fig-4) proposed model structure of coal molecules (83-84%C) indicating availability of 9,10 dihydrophenanthrene and phenanthrene in coal units which are linked with each other by ether and methylene bridges. The angled aromatic rings like phenanthrene is the mother nucleus of coal. Based on its rank, association of different functional groups in different proportions with other units of coal by methylene and by ether linkages [14 -16] are reported in literature. It is also known that this linkage is less reactive towards oxidation hence it cannot explain the above result. By virtue of vast literature available on oxidation of condensed aromatic ring, it is obvious that the 9-10 position of phenanthrene is more reactive towards oxidation rather than other sites [14,15,17,20]. Thus, during oxidation preferably dissociation of this bond takes place. In a previous work [1], it was observed that not only the carboxylic acid functional groups are generated on oxidation of coal but subsequently the nitro groups are also substituted in the periphery of aromatic structure on coal matrix. Thus, creation of carboxylic acid functional groups in the coal matrix takes place first and then nitration occurs, so that nitro groups substitute at meta position with respect to carboxylic acid functional groups which is meta-directing as shown in Reaction- 1. The quantity of OH functional group on oxidation of demineralised coal to oxy-coal [18,19] remained practically the same (Table-1).. The above data clearly indicates the availability of 9,10 position of phenanthrene type of aromatic ring in coal which are responsible for generating large amount of carboxylic acid functional groups.

NO2 COOH

HNO3

-------(1)

OXIDATION

COOH

NO2

Table-2 Summarizes the proximate, ultimate, functional groups analysis of Tipong coal

and

distribution of carbon in different form.. Based on the experimental results the hydroxyl group recorded from 3--4.7% which is within the range. It is the trend the percent hydroxyl group gradually decreases with increasing the rank and caking behavior developed. In case of high sulphur coal the percent of oxygenated function groups replaced by sulphur in some extent shown in the unit-1 and unit-2. It is also observed the total sulphur present in coal exist in different form shown in above model structure proposed by two units. The factor of aromatic and hydroaromatic also reflect the behavior of coal within rang of caking even the sum of hydroxyl and thiol cover the same value as compare to original coal of this range. Proximate, Ultimate & functional groups analyses and distribution of carbon in different forms of Tipong coal are summarized in Table 2. It is observed from functional group analysis of several N.E.region coals that hydroxyl group ranged between 3 – 4.7 % which is within the range for caking variety of coal. It is well known that percent hydroxyl group gradually decreases with increase in rank and develop caking properties. In case of high sulphur N.E.region abnormal coal [21-24], it is known that part of the oxygen in coal is replaced by sulphur. It has already been established [2] that the organic sulphur in high sulphur N.E.region coals exist in five forms viz. disulphide, thiol, thioether, thioketone and thiophene. It is evident from table2 that aromaticity, hydroaromaticity,H / C aromatic and Non aromatic values are in the range of caking variety coals. Also the sum of –OH and –SH groups is equal to that of –OH group present in normal coals of same carbon rank. Temperature programmed reduction (TPR) studies on high sulphur coals of N.E.region of India [ ] to determine the sulphur functional groups have been reported by Attar and Dupuis [ ] and Majchrowicz et al. [ ] using l - 75 mg of coal in each experiment with continuous heating at a certain rate with gravimetric analysis of sulphur evolved as H2S and absorbed to yield precipitate. Table 3 presents the data on TPR studies on Tipong Coal. Our approach was - the C-S bond is broken and then the sulphur is removed as H2S. If sulphur is present in coal in more than one

form depending upon the nature of sulphur functionality / bond strength, the weakest bond will break first while the strongest at the last. The quantity of H2S evolved at five different temperature ranges indicate that the sulphur present in coal is not in one form rather it is in five forms. The evolution of H2S at different temperature ranges depends on the nature of C-S bond. The studies were done on Tipong coal [2] under different reaction conditions. Based on the experimental work it was concluded that the organic sulphur present in this coal are in five different forms viz. disulphide, thiol, thioketone, thioether and thiophene. It was also concluded that the Sulphur is interchangeable with oxygen in N.E. region coal structures at all places. The following were some of the findings of the above work. 1. During reduction, mercaptan SH will react with Hydrogen to form H2S which will be evolved in the temperature range 190-220°C. The disulphide will yield mercaptan on hydropyrolysis which will further react with Hydrogen to yield H2S in the same temperature range. If the disulphide is converted into sulphide and H2S, then the aliphatic sulphide will be transformed into complex thiophenic sulphur below the temperature at which H2S is evolved’. 2. The thiol group reacts with Hydrogen to form H2S in the temperature range 260- 2900C. 3. The thioether linkage or the aromatic sulphide reacts with Hydrogen to give H2S, directly or via -SH, which will be evolved in the temperature range 360-390°C. 4. Reaction of simple thiophene with Hydrogen will break the ring giving H2S in the temperature range 460-490°C. 5. During continuous pyrolysis, any C=S groups will cyclize in the presence of aliphatic chains which will lead to the formation of complex/ condensed thiophenic rings, which are difficult to decompose but small amounts of which may be removed as H2S above 500°C. 6. It is well known that the reaction: FeS2 + H2 = FeS + H2S is complete at 530°C and the reaction FeS + H2 = Fe + H2S starts well above 900°C. These two reactions are believed to be responsible the peak of H2S evolutions at 530°C and the shoulder at 620°C. On careful examination of Table 2 (Upper Part), it is evident that the molecular formula / molecular weight and the ultimate analysis data based on experimental results, when compared with the

corresponding data based on proposed model structure are in excellent agreement. Thus we can assume that the proposed model structure of high sulphur NE region Tipong coal is correct. But to prove this, other supporting evidences are a must. The proposed model structure of NE region of coal must explain the formation of the products on various coal conversion processes. It must explain the existence of aromatic, aliphatic, hydroaromatic and hetero-atoms present in NE region coals. The gaseous products obtained on pyrolysis of coals from aliphatic groups are [25] Carbon di oxide, carbon monoxide, methane, hydrogen, hydrogen sulphide, ammonia, other hydrocarbons etc. can be very well be explained by the proposed model structure of NE region coal supporting our model. The heteroatoms present in coal are oxygen, sulphur and nitrogen. Nitrogen appears to be evenly distributed and has no significant effect on coal reactivity [25], but oxygen functional groups play a very important role in coal conversion. Some of the heteroatoms containing units identified [17] are fluorenone, Anthraquinone, Benzanthraquinone and/or Naphthacenequinone, Dibenzofuran,

Benzonaphathofuran,

Xanthone,

Benzoxanthone,

Dibenzo-p-dioxin,

Benzothrophene, Dibenzothiophene, Pyridine, Quinoline and/or Isoquinolene, Carbazole, Acridone. These products were obtained on heating coal [17] at 2500C in 0.4 m sodium dichromate, and the aromatic acids produced were isolated. The proposed model structure of coal on its oxidation can easily explain the above formations. The products of pyrolysis also include aromatics like Benzene, Pyrene, Benzopyrene, naphthalene, phenanthrene etc. production of which can also be explained from our proposed model structure of coal. Thus these validate our proposed model structure. Similarly during solvent refining of coal, the hydroaromatics obtained [25]

are

Decahydropyrenes,

Hexahydropyrenes,

Dihydropyrenes,

Pyrenes,

Octahydrophenanthrenes, Tetrahydrophenanthrenes, Phenanthrenes, Tetralins, Napthalenes, Tetrahydroacenaphthenes, Acenaphthenes, Our proposed model structure of coal can very well explain their formations. Thus it can be concluded that the proposed model structure of high sulphur NE region coal is very much valid and can explain all the Physico-Chemical properties of coal.

REFERENCES

1. D.P. Choudhury, S.S.Choudhury, Raja Sen, J.Mukherjee, G.Ghosh and S.K.Srivastava, Energy and Fuels, 21,1006, 2007. 2. A.Kumar and S.K.Srivastava, Fuel, 71, 718, 1992. 3. IS : 6444 : 1979 4. IS : 1350 Part III 1969 5. G.Ghosh, A.Banerjee, B.K.Mazumdar, Fuel, 54, 294, 1975. 6. Atma Ram Singh, D.P.Choudhury, G.Ghosh and S.K.Srivastava, The Open Fuels and Energy Science Journal, 3, 8, 2010. 7. A.R.Singh, D.P.Choudhury and S.K.Srivastava, Composite Polymeric Materials from Coal, Unpublished work. 8. D.P.Choudhury, Raja Sen, G. Ghosh and S.K.Srivastava, Direct Sourcing of Coal: Solubilization of coal from N.E.region of India, Presented at 23 rd Annual International Pittsburg Coal Conference, Pittsburg, USA, 25-28 September, 2006 and its references. 9. P.H Given, Fuel, 39, 147, 1960. 10. R. Hayatsu, R.E.Winans, R.G Scott, L.P.Moore, M.H. Studier, Oxidation degradation study of coal and solvent refined coal, ACS Symp. Ser.,71, 108,1978. 11. J.N.Bhowmik, P.N.Mukherjee, A.Lahiri, Fuel, 38. 21, 1959. 12. L.A.Heredy,I.Wender, Flammability , Ignition and Electrostatic Properties, ACS Div. Fuel Chem. Preprints, 25, 1980.

13. D.W. Vankrevelen, Fuel, 29, 269, 1950. 14. W.R.Lander and C.E. Snape, Fuel, 57, 658, 1978. 15. J. Gibson, J. Inst. Fuel, 51, 61, 1978. 16. G.J. Pitt, G.R.Millward, in Coal and Modern Coal Processing; an introduction, structural analysis of coal, Academic Press, London, UK, 27, 1979. 17. R.Hayatsu, R.G.Scott, L.P.Moore, M.H.Studier, Nature, 257, 378, 1975. 18. R.Hayatsu, R.G.Scott, L.P.Moore, M.H.Studier, Nature, 77, 261, 1976. 19. R.Hayatsu, R.E.Winans, R.G.Scott, L.P.Moore, M.H.Studier, Nature, 275, 116, 1978. 20. R.Hayatsu,R.G.Scott, L.P.Moore, M.H.Studier, Nature, 57, 541, 1978. 21. A.Ali, S.K.Srivastava, R.Haque, Fuel, 71, 835, 1992. 22. S.K.Srivastava, A.Kumar, R.Haque, Fuel Sci. Tech., 11, 111, 1992. 23. S.K.Srivastava, A. Kumar , R.Haque, Proceedings of 5th Australian Coal Science Conference, Melbourne, Australia, 30 November-2 December,1992, p. 344. 24. A.Kumar, S.K.Srivastava, K.K.Singh, Fuel Sci. Tech., 13, 71, 1994. 25. Coal Science and Technology 2 : Fundamentals of Coal Beneficiation and Utilization, Edited by S.C.Tsai, Elsevier Scientific Publishing Company, Printed in The Netherlands, 1982.

Table – 1 Characterization of Raw and Demineralised Samla coal, Oxy coal & soluble part of oxy coal in polar organic solvent Analyses Moisture % *

8.6

Demineralised coal 1.7

Ash%*

14.6

1.0

6.0

0.0

Carbon % (dmf)

80.6

80.5

62.3

59.7

Hydrogen % (dmf)

4.5

4.6

2.9

4.5

Nitrogen % (dmf)

2.1

2.0

4.4

4.4

OOH % (dmf)

4.8

5.2

5.0

2.0

OCOOH % (dmf)

0.96

1.2

18.3

12.6

CCH3 % (dmf)

3.1

2.4

1.0

0.8

* (as received basis)

Raw coal

oxy coal 9.1

Organic soluble oxy coal 7.9

Table-2 Analysis of Tipong Coal on as received basis _________________________________________________________________________ | | | Experimental data | Data based on | | Proposed Structure ------------------------------------------------------------------------------------------------------------------------Proximate( %) Moisture 2.8 Ash 5.3 Molecular formula C87H77O7N1.5S2.5 Volatile matter 42.0 Fixed Carbon 49.9 Ultimate Analysis (%) (dmf) Carbon 71.19 78.3 78.3 Hydrogen 5.3 5.8 5.8 Nitrogen 1.22 1.4 1.6 Sulphur 5.7 6.3 6.0 Oxygen(By diff.) 8.49 8.3 8.4 _________________________________________________________________________ _______________________________________________________________________ Unit-1

Unit-2

AV

BKM

fa 0.62 0.64 0.63 0.67 fhar 0.29 0.27 0.28 0.22 fcCH3 0.045 0.034 0.039 0.047 Har/C 0.21 0.21 0. 21 0.22 H/C(non aromatic)` 1.7 1.8 1.75 1.66 %OOH 3.5 4.7 4.1 5.8 % SH 0.05 0.02 0.035 --_________________________________________________________________________

Table-3 Sulphur evolved from Tipong coal during TPR studies _________________________________________________________________________ Temperature Sulphur evolved Sulphur functionality 0 Range ( C) (%) assigned _________________________________________________________________________ 1 190—220 0.36 Mercaptain/disulphide 2 260—290 1.7 Thiol 3 360--390 0.51 Aromatic sulphide 4 460--490 0.88 Simple thiophene 5 510—540 0.35 Pyritic and part of complex thiophene 6 590--620 0.24 Part of complex thiophene _________________________________________________________________________

CH3

CH3

co

CH3

co o

0

CH2

OH

CH3

CH2

N

N H

H

COOH

0 HO

OH

OH

OH

OH

CH3

co

Fig-1

Unit-1

CH3

CO

CH3

CH3

O

CS

CH

O

CH3

OH

OH

SH

N H

O OH

Fig-2

Unit-2 CH3

CH2

CH3

COOH

CH3

O

N H

OH

CS

SH

OH

Fig-3

OH

CH2

N H

S OH

Influence of the Surface Treatment with O3 and NH3 on the Physical and Chemical Characteristics of Dried Low Rank Coal Gi Bo Han, Yongseung Yun, Changsik Choi*

Plant Engineering Center, Institute for Advanced Engineering, 633-2 Goan-ri, Baekam-myon, Cheoin-gu, Yongin-si, Gyeonggi-do Gyeonggi-do, Republic of Korea

Abstract In this study, the dried coal with the surface treatment was characterized to investigate the effect of the surface treatment on the chemical and physical properties of dried coal. The surface treatment was conducted by the dry method with 5 vol% NH3 at 200 oC and 15 g/m3 O3 at room temperature. As compared with the fresh dried coal, it was found that the dried coal obtained after the surface treatment was changed in aspect of chemical and physical properties such as the increase in the content of carbon, hydrogen and fixed carbon. From the change of the chemical and physical properties, the various effects were obtained as follows: 1) Removal of oxygen-contained functional group, 2) Elevation of calorific value, 3) Inhibition of spontaneous ignition potential, 4) Suppression of H2O adsorption of dried coal. 1. Introduction After 1980’s, the green house effect was caused with depletion of high rank coal and the consumption amount of coal was diminished in the whole world. In 2005, coal industry was dropped; however, it was in charge of electricity of 41% and the energy consumption of 28% all over the world. Recently, gas price is raised up to higher than $100/bbl, the consumption amount and price of coal steeply rising. In energy crisis of the 1970’s, the coal technologies such as an integrated gasification combined cycle (IGCC) and coal liquefaction process have been developed to utilize and convert coal to clean energy. Coal conversion technology such as IGCC system, coal liquefaction and coal-water mixture production process is an issue as a one of the coal technologies but, has not yet been commercialized due to huge expenses for development [1-13]. Recently, the concerns of environments have been increased to coal industry due to the air pollutant emission with coal utilization. The green house gas emission especially increased global warming. The development of the various clean coal technologies to reduce green house gas emissions and energy consumption is necessary for the sustainability of world economy. Upgrading of low rank coals is more focused due to the increase in energy consumption and demand. Coal is a heterogeneous natural polymer material which is composed of moisture, volatile matter, ash and fixed carbon. According to the degree of carbonization, the coal rank is classified into a lignite (or brown coal), (Sub-) bituminous coal. Lignite is not

coking, quickly burned and it is mainly used as a resource for the electric power generation. On the other hand, lignite coal contains the low content of sulfur and ash and can be converted into an alternative energy of petroleum with the development of technology for the elevation of calorific value and the inhibition of CO2 generation. To upgrade the coal rank of lignite, it is necessary to elevate the calorific value, inhibit the spontaneous ignition and drop the CO 2 generation. The various low rank coals including lignite, brown coal and sub-bituminous coal have generally characteristics of high moisture content which plays an important role in combustion, gasification and liquefaction. Drying process of low rank coals is very important for combustion technologies and boilers using low rank coals. The high content of moisture of low rank coal causes many problems in coal processes including storage and transportation. The high moisture content in low rank coals causes high energy consumption and emission of CO2. Spontaneous ignition associates with moisture adsorption. Upgrading process of low rank coal for the pretreatment of coal which results in the decrease in the water content, the increase in the calorific value and the prevention of self-heating and spontaneous combustion during transportation and storage has been presented in recent. In this paper, the surface treatment of dried coal by NH3 and O3 was conducted for upgrading the coal rank of lignite. Also, the chemical and physical properties of coal were investigated for the characterization of coal such as composition, calorific value and ignition temperature. 2. Experimental 2-1. Coal Sample and Drying Process The raw coal was an Indonesian lignite coal having high moisture content as shown in Table 1 and its particle size range was less than 3 mm. To reduce the H2O content of raw coal with high moisture content, the raw coal was dried at 150 oC under the air atmosphere. 2-2. Surface Treatment of Dried Coal The fixed bed reactor was horizontal-typed quartz tube with OD of 1/2 inch and length of 30 cm. The dried coal with the appropriate amount was packed and fixed in the reactor. The gas for the treatment was flowed into the reactor and then the flow rate was 100 ml/min. The surface treatment conditions were listed in Table 1. The treatment temperatures of the dried coal with O3 and NH3 were room temperature and 200 oC, respectively and then their concentrations were 15 g/m3 and 5 vol%, respectively. 2-3. Characterization of Coal with the Surface Treatment The composition of carbon, hydrogen, nitrogen, oxygen and Sulfur was analyzed by elemental analysis. Also, the content of fixed carbon, moisture, volatiles and ash was found by the proximate analysis. The relative comparison of oxygen-contained functional group like

carbonyl and carboxyl group was conducted by FT-IR spectra. Calorific value was measured with surface treatment. Under N2 atmosphere, thermal gravimetric analysis (TGA) for the raw and dried coals was conducted with raising the temperature at the ramping rate of 5 oC/min from room temperature to 800 oC. Ignition temperature of raw and dried coal before and after the surface treatment was measured by raising the temperature from room temperature to 900 oC at the ramping rate of 5 oC/min under air atmosphere after the coal sample was packed in ignition temperature tester (Krupp Flash Point Tester, Model: KRS-RG-9000). H2O adsorption for raw and dried coals before and after the surface treatment was conducted as follows: H2O content of coal sample absorbing H2O was measured by proximate analysis after the vapor of water of 200 ml boiled in the round flask was passed through the packed coal sample of about 2 g for 4 h. 3. Results and discussion 3-1. Drying Process for Dewatering Raw Coal Figure 1 shows the thermal gravimetric analysis result of raw and dried brown coals to investigate the temperature for the removal of water of raw coal. O2. The amount of raw and dried of coal was 8 mg and it was packed into bowl and then the temperature increased from room temperature to 800 oC at the ramping rate of 5 oC/min under the atmosphere of N2 of 100 ml/min. As compared with Figure 1(b), it was known that the water of raw coal was removed under 120 oC in Figure 1(a). Therefore, the temperature for the water removal of raw coal was fixed at 110 oC. Figure 2 shows the proximate analysis of raw and dried coals. In Figure 2, the moisture, volatiles, ash and fixed carbon were contained in the raw coal and then their contents were 34.4, 32.2, 1.8 and 31.6%, respectively. However, in the coal dried at 107 oC, the moisture of 1.4%, volatiles of 48.4%, ash of 2.7% and fixed carbon of 47.5% were contained and then it was known that the water content decreased and the contents of other components increased because the water was removed from the coal through drying process as compared to the raw coal. 60

0.008

100

Weight loss (%) dw/dt 80

Raw coal Dried coal

50

(b) Dried coal

48.4

47.5

0.004 40

Content (%)

60

dw/dt

Weight loss (%)

0.006

40 34.4

32.2

31.6

30

20

0.002 20

10 1.8

1.4 0

0.000 100

200

300

400

500

Temperature (oC)

600

700

800

2.7

0

Moisture

Volatiles

Ash

Component

Fixed carbon

Figure 1. Thermal gravimetric analysis of raw Figure 2. Proximate analysis of the raw and dried coals.

and dried coals.

3-2. FT-IR Spectra of Dried Coal with Surface Treatment To investigate the characteristics of dried coal Figure 3 shows the FT-IR spectra of the surface-treated coals with O3 and NH3 after drying at 107 oC. As shown in Figure 3, it was known that the coal dried at 150 oC under air atmosphere have oxygen-contained functional groups such as carbonyl and carboxyl croups corresponding to peaks observed in the range of 1200-1700 cm-1 and hydroxyl group corresponding to peak observed in the range of 2800-3000 cm-1. The various analysis results and molecular models of coal structures such as have been presented in the previous report and then it was presented that low rank coals such as brown coal have relatively high content of oxygen-contained functional groups like carbonyl and carboxyl groups as compared to high rank coals; therefore, their energy efficiency such as calorific vale were low [14]. In general, it was well-known that high oxygen content results in the low calorific value according to the various theoretical equations for the correlation between elemental analysis result and calorific value. The calorific value of the coal of low oxygen content was higher than that of high oxygen content because the oxygen content decreased and the carbon and hydrogen content relatively increased. As shown in Figure 3, in case of the dried coal treated with O3, the peak intensity of 1200-1700 cm-1 increased in comparison with the fresh dried coal and it might be due to the generation of carbonyl and carboxyl groups with partial oxidation of carbon structure surface by O3. On the other side, in case of dried coal treated by NH3, the peak intensity of carbonyl and carboxyl groups decreased because the carbonyl and carboxyl having acidic properties contained in dried coal might be decomposed by NH3 having basic properties.

In addition, the peak intensity of hydroxyl group corresponding -1

to the range of 2800-3000 cm decreased after the treatment with NH3 but increased after the treatment with O3 through the surface reaction between either NH3 or O3 and the hydrogen matrix contained in dried coal. It was known that oxygen-contained function group plays an role of precursor for the CO2 generation in combustion [15]. Therefore, the removal of oxygencontained functional group with the surface treatment of dried coal by NH3 leads to the inhibition of CO2 generation.

1.2

46.89 Fixed carbon

Transmittance (%)

80

Content (%)

B

A

C

4000

3500

0.96 Fuel ratio

49.33 Fixed carbon

1.03 Fuel ratio

60 2.9

2.8 Ash

2.89 Ash

Ash

1.0

44.83 Fixed carbon 0.85 Fuel ratio

0.6

40 48.39 Volatile

48.78 Volatile

1.82 Mositure

0.02 Mositure

0.4

52.37 Volatile

20

3000

2500

2000

1500

1000

0.8

0

0.2

A

-1

2.8 Mositure

B

Fuel ratio (FC(%)/VM(%))

100

0.0

C

Wavelength (cm )

Figure 3. FT-IR spectra of dried coal with the Figure 4. Proximate analysis of dried coals surface treatment (A: Fresh dried coal, B: with the surface treatment (A: Fresh dried NH3, C: O3).

coal, B: NH3, C: O3, VM: Volatile matter, FC: Fixed carbon).

3-4. Composition of Dried Coal with Surface Treatment Figure 4 shows the proximate analysis result of surface-treated coal after dried process. It was known that the rank of coal is related to the proximate analysis and fuel ratio (Fuel ratio = Fixed carbon (FC, %)/Volatile matter (VM, %)) and the rank of coal increases with increasing the fuel ratio [14]. In this section, the effect of the surface treatment with NH 3 and O3 on the proximate and fuel ratio for the dried coals was described. The fixed carbon of 49.33%, ash of 2.89%, volatile of 48.78% and moisture of 0.02% were contained in the dried coal treated by NH 3 and then the fuel ratio was about 1.03. In comparison to the fresh dried coal, both the content of the fixed carbon and fuel ratio increased because the oxygen-contained functional group carbonyl and carboxyl group might be decomposed by the surface reaction with NH3. On the other hands, the dried coal treated by O3 was composed of 44.83% fixed carbon, 2.8% ash, 52.37% volatile and 2.8% moisture and then its fuel ratio was about 0.85. The surface treatment with NH3 raised Fuel ratio but that with O3 did not. In a viewpoint of fuel ratio which is related to the energy resource efficiency, the rank of coal was improved by the surface treatment with NH 3 as compared with the fresh dried coal. 3-6. Ignition Temperature of Dried Coal with Surface Treatment Figure 5 shows the ignition temperature of the dried coal before and after the surface treatment by NH3 and O3. In this section, the effect of the surface treatment on the ignition temperature was investigated to compare the simultaneous ignition potential according to surface treatment method, relatively. The ignition temperature of raw coal was about 284 oC and that of dried coal was about 312 oC. This result was caused by decrease in the moisture content.

Therefore, it was estimated that the spontaneous ignition potential was dropped slightly. In the previous report, it was described that, in fact, it is not yet well understood how the moisture affects the tendency of a coal to selfheating and what is the mechanism of adsorption and desorption. In general, it has been found that a coal reacts with oxygen more rapidly when the coal is wet than when it is dry [16-17]. The ignition temperature goes up to 363 oC after the surface treatment with NH3; however, the ignition temperature was dropped to 293 in case of the dried coal after the surface treatment by O3. It was thought that this result comes from the variation of the content of volatile matter. It was previously referred that the content of volatile matter has influenced on the spontaneous ignition temperature and the higher volatile matter content, the higher spontaneous ignition potential [18]. Therefore, as compared with the dried coal, the ignition temperature increased from 312 to 363 oC in case of dried coal treated by NH3 and decreased from 312 to 293 oC in case of dried coal treated by O3. This result will contribute toward the safety coal stockpile and transportation. 400

Ignition temperature (oC)

363

300

312

293

284

200

100

Weight (g) & H2O content (wt%)

40 H2O content 35.0 wt%

30

20 H2O content 14.3 wt%

H2O content 11.7 wt%

10 W2 W1 2.2901 g ¥ÄW 2.0113 g 0.2788 g

W2 W1 2.2414 g ¥ÄW 2.0101 g 0.2313 g

W1 2.0108 g

W2 2.9607 g

¥ÄW 0.9499 g

0

0

A

B

C

A

D

B

C

Figure 5. Ignition temperature of dried coals Figure 6. Variation of weight and H2O content with the surface treatment (A: Raw coal, B: of the dried coal treated by NH3 and O3 with Fresh dried coal, C: NH3, D: O3).

H2O adsorption (A: Fresh dried coal, B: NH3, C: O3).

3-7. H2O Adsorption of Dried Coal with Surface Treatment Figure 6 shows the relative weight variation of the dried coal treated by NH3 and O3 with H2O adsorption to investigate the relationship between H2O adsorption and surface treatment. After adsorbing H2O, the weight of dried coal increased from 2.0113 to 2.2901 g with the H2O content of 14.3 wt%. In case of the dried coal treated by NH3, the weight increased from 2.0101 to 2.2414 g with H2O content of 11.7 wt%. In case of the dried coal treated by O3, the weight increased from 2.0108 to 2.9607 g with H2O content of 35 wt%. It was thought that these results are related to the removal and generation of oxygen-contained functional group such as carbonyl and carboxyl groups. In general, it was well known that carbonyl and carboxyl groups

are hydrophilic. As described in the previous section involving FT-IR spectra, the peak intensity corresponding to the oxygen-contained functional group decreased with the surface treatment by NH3 but increased with the surface treatment by O3. In case of the surface treatment by NH3, and then H2O adsorption capacity decreased with the removal of the oxygen-contained functional group. On the other hand, the H2O adsorption capacity increased with the generation of the oxygen-contained functional group. Therefore, it was expected that the surface treatment of dried coal by NH3 to prevent the H2O adsorption contributes toward the improvement of the coal stockpile and transportation. 4. Conclusion In this study, the influence of the surface treatment with NH3 and O3 on dried coal was investigated by characterizing the chemical and properties of the coal such as functional group, composition, H2O adsorption capacity, calorific value and ignition temperature. In case of the surface treatment with O3, as compared with the fresh dried coal, the content of carbon, hydrogen and fixed carbon decreased and that of oxygen and volatile matter increased with the treatment by O3. And then the H2O adsorption capacity relatively increased and the ignition temperature and the calorific value decreased with the variation of composition. As compared with fresh dried coal, the dried coal treated by NH3 has the low content of oxygen and volatile matter and the high content of the carbon, oxygen and fixed carbon. With this composition variation, the ignition temperature and calorific value increased and the H2O adsorption capacity decreased. Therefore, the surface treatment with NH3 after the drying process of raw lignite coal having the high content of moisture is the effective method for the safety storage and transportation for the long periods. Reference 1. Sakaguchi, M., Laursen, K., Nakagawa, H., Miura, K.: “Hydrothermal Upgrading of Loy Yang Brown Coal: Effect of Upgrading Conditions on the Characteristics of the Products,” Fuel Processing Technology, 89, 391 (2008). 2. Mahidin, Ogaki, Y., Usui, H., Okuma, O.: “The Advantages of Vacuum-Treatment in the Thermal Upgrading of Low-Rank Coals on the Improvement of Dewatering and Devolatilization,” Fuel Processing Technology, 84, 147 (2003). 3. Sato, Y., Kushiyama, S., Tatsumoto, K., Yamaguchi, H.: “Upgrading of Low Rank Coal with Solvent,” Fuel Processing Technology, 85, 1551 (2004).

4. Elliott, D.C.: “Decarboxylation as a Means of Upgrading the Heating Value of Low-Rank Coals,” Fuel, 59, 805 (1980). 5. Umar, D.F., Usui, H., Daulay, B.: “Change of Combustion Characteristics of Indonesian Low Rank Coal due to Upgraded Brown Coal Process,” Fuel Processing Technology, 87, 1007 (2006). 6. Friesen, W.I., Ogunsola, O.I.: “Mercury Porosimetry of Upgraded Western Canadian Coals,” Fuel, 74, 604 (1995). 7. Ruyter, H.P., Raam, L., Poel, H.: “Upgrading of Wakefield Browncoal from South Australia,” Fuel Processing Technology, 9, 163 (1984). 8. Ogunsola, O.I.: “Thermal Upgrading Effect on Oxygen Distribution in Lignite. Fuel Processing Technology,” 34, 73 (1993). 9. Ogunsola, O.I., Mikula, R.J.: “Effect of Thermal Upgrading on Spontaneous Combustion Characteristics of Western Canadian Low Rank Coals,” Fuel, 71, 3 (1992). 10. Petela, R., Petela, G.: “Indices for Coal Desulfurization and De-Ashing Processes,” Fuel, 75, 1259 (1996). 11. Friesen, W.I., Ogunsola, O.I.: “Principal Component Analysis of Upgraded Western Canadian Coals,” Fuel Processing Technology, 38, 139 (1994). 12. Lockhart, N.C.: “Dry Beneficiation of Coal,” Powder Technology, 40, 17 (1984). 13. Nandi, B.N., Macphee, J.A., Ciavaglia, L.A., Chornet, E., Arsenault, R.: “The Role of Mesophase in Upgrading Inert-Rich Oxidized Coal for Combustion,” Carbon, 20, 148 (1982). 14. Kabe, T., Ishihara, A., Qian, E.W., Sutrisna, I.P., Kabe, Y.: “Coal and Coal-Related Compounds,” Studies in Surface Science and Catalysis, 150, 85 (2004). 15. Sato, Y., Kushiyama, S., Tatsumoto, K., Yamaguchi, H.: “Upgrading of Low Rank Coal with Solvent,” Fuel Processing Technology, 85, 1551 (2004).

16. Küçük, A., Kadıoğlu, Y., Gülaboğlu, M.Ş.: “A Study of Spontaneous Combustion Characteristics of a Turkish Lignite: Particle Size, Moisture of Coal, Humidity of Air,” Combustion and Flame, 133, 255 (2003). 17. Bhattacharyya, K.K., Hodges, D.J., Hinsley, F.B.: The Mining Engineer. 126, 274 (1968). 18. Hesketh, R.P.; DAVIDSON, J. F. The Effect of Volatiles on the Combustion of Char in a Fluidised Bed. Chemrcal Engineering Science, 46, 12, 3101-3113, (1991).

Manuscript Not AVAILABLE

Energy, Natural Gas, Türkiye & Ankara İbrahim Halil KIRŞAN Başkent Doğalgaz Dağıtım A.Ş. [email protected] ABSTRACT Energy, Natural Gas, Türkiye & Ankara Natural gas; environment friendly, economic, confortable, safe as long as it is used properly, clean, harmless to nature, is an “environmentalist” energy source. Is a combustible gas formed in the lower layers of the earth and primarily consisting of methane and ethan hydrocarbons. Has been formed by decaying organic substances millions of years ago. It is a primary energy source ie: it can be used immediately after drilling out. Being lighter than air it can readily effuse into the atmosphere. Natural gas; when burned does not produce environmentally harmfull wastes such as sulphurdioxide and carbon particles. Natural gas; is a serious insurance to nature, environment hence the future of mankind. As a fossil energy source similar to petrol, however does not leave behind ash or slag when burnt and has no need for storage during use. Total natural gas reserves of the world is 178,7 trillion cubic meters and 3 Trillion cubic meters natural gaz is consumed annually throughout the world. May be all natural gas reserves will last 50 or 60 years. Our country in 2006, 19.6 billion cubic meters of natural gas consumed imported from Russia. This corresponds to 63% of total consumption. The amount of LNG imported from Nigeria and Algeria, 5.3 billion cubic meters. Our total reserves our calculated to be less than our annual consumption amount. Türkiye relies on foriegn countries for the supply of natural gas. Majority from Russia and Iran(through pipes), remainder from Algeria and Nigeria (as LNG). Başkentgaz is the second largest company with 2 billion cubic meters consumption per year and oldest distribution company in Turkiye. Keywords: Energy, Natural Gas, Reserves, Başkentgaz.

1 Characterization of Chars Made of Solvent Extracted Coals Hokyung Choia, Jiho Yooa, Sangdo Kima, Jeongwhan Lima, Thiruppatha Rajaa, Wantaek Job, Sihyun Lee a* a

Clean Fossil Energy Research Center, Korea Institute of Energy Research 102 Gajeong-ro, Yuseong-gu, Daejeon, 305-343, Korea b Department of Chemical Engineering, Yonsei University, Seodaemun-gu, Seoul, Korea *Email : [email protected]

ABSTRACT Of contaminants contained in coal, ash was blamed for the serious issues in the power sector; a decrease in the power efficiency and discharge of fly ash into the air. Thermal extraction of coals with solvents has produced ash-free coals successfully, potentially solving the ash problem and bringing new applications such as direct coal feeding into gas turbine. Whereas the properties of the ash-free coals are well known, chars made of the extracted coals have not yet been characterized. In this study, the organic portion of a sub-bituminous coal (Roto south) was extracted at 370 °C using 1-methyl-naphthalene solvent. The extract/residue coal as well as its parent coal were pyrolyzed at 300 − 900 °C. The carbonized products were then characterized. Proximate and ultimate analysis were performed to study the compositional change. The difference in chemical structure was investigated using a solid-state 13C-NMR. Calorific value was also determined. The discussion focused on how the physical/chemical properties of the chars varied depending on the kind of the coals and the temperature pyrolyzed.

1. INTRODUCTION Coal is one of the most important energy sources, currently accounting for ~25% of worldwide energy consumption [1,2]. The use of coal is expected to increase, thanks to its abundance and economic advantage. However, there are various problems coupled with the combustion of coal such as ash deposition, emission of pollution and greenhouse gas. Coal ash decreases the power efficiency and discharges fly ash as an air pollutant [3−5]. Therefore, a lot of works have concentrated on the development of efficient methods to prepare ash-free coals [6−18]. Among them, thermal extraction with organic solvents has produced ashfree coals, namely “Hypercoal”, most successfully, potentially solving the ash problem and also bringing new applications such as direct coal feeding into gas turbine [19,20]. Once the direct combustion of ashless coal in gas turbines is achieved, an increase in the power output and reduction of CO2 emissions is expected. Economically, ash-free coals derived from low rank coals are even more attractive [21].

2 Recently, ash-free coals have found new applications; fuel for direct carbon fuel cell (DCFC) and catalytic gasification, binder/additive of metallurgical coke, and aluminum anode coke, mainly owing to their ashless character and high thermoplasticity [22−25]. The devolatilization of relatively small organic molecules happens in the initial stage of coal utilization. This early devolatilization is influential on the subsequent processes, normally homogeneous/heterogeneous reactions of char. Utilizing ash-free coals for the new applications, therefore, necessitates better understanding of the changes associated with conversion of the extracted coal into char. The devolatilization behavior of raw coals, including the structural/compositional transformation and associated thermoproperty, was studied previously [26−30]. Chemical and structural change of char, the size of aromatic rings and the nature of functional groups and cross-links, were determined by the spectroscopic methods [26]. Alonso et al. investigated the changes in char reactivity as a function of pyrolysis temperature and found a strong dependence of char reactivity on the pyrolysis temperature [27]. Partially carbonized samples at the intermediate temperature (290 − 630 °C) were also characterized, where the transformations taking place in the thermoplastic stage of coking were described [28]. However, only limited information is available on the carbonization characteristics of the solvent extracted coal and its insoluble by-product. In this study the solvent extracted coal originated from an Indonesian Roto south coal (a subbituminous coal), its insoluble by-product (residual coal), as well as its parent coals, were pyrolyzed at the temperature range, 300 − 900 °C. The effect of pyrolysis temperature was then analyzed by comparing with one another, based on the data obtained using proximate/ultimate analysis and solid-state 13C-NMR.

2. EXPERIMENTAL

2-1. Preparation of ashless coal (solvent extraction) The organic portion of a subbituminous coal (Indonesian Roto south coal) was thermally extracted using non-polar 1-methylnaphthalene (1-MN) solvent. As shown in Fig. 1, the extraction was processed via extraction, filtration, and drying step. A batch-type autoclave (0.5 liter volume) was adapted as an extractor. The raw coal was ground, meshed to < 74 μm (200 mesh), and then dried in a vacuum oven. Coal slurry was prepared by combining 20 g of the dried coal with 200 g of 1-MN. The mixture was added to the autoclave, which was then purged with N2 gas. While stirring with the magnetic agitator, the coal/solvent mixture was heated to 370 °C and held for an hour under 30 bar. The thermally treated sample was transferred to the filtering unit (glass or stainless steel filter) to separate the solvent extract from the residual matter. The solvent in both the extract and the residue was removed by being kept in a vacuum oven (~300 °C) for 3 − 4 hr under N2 atmosphere. After drying, solid materials were obtained, namely, extracted coal (EC) and residual coal (RC). Proximate/ultimate analysis, calorific value, BET surface area of the parent coal (PC), EC, and RC are tabulated in Table 1. Based on ASTM D3172 standard, the proximate analysis determined the content of moisture, ash, volatile matter, and fixed carbon, using TGA-701 Thermogravimeter (LECO Co., USA). Elemental composition (C, H, N, and O) was obtained using CHN-2000 Elemental Analyzer (LECO Co., USA), calorific value using Parr 1261

3 Calorimeter (PARR Co., USA), and BET surface area using ASAP 2420 (Micromeritics Co., USA). Agitator TC

Vent

Extractor

PG

N2 Filter Heater To dryer

Fig. 1. Schematic diagram of solvent extraction. (PG: pressure gauge, TC: thermocouple) Table 1. Proximate/ultimate analysis, calorific value, BET surface area of the parent coal (PC), extracted coal (EC), and residual coal (RC). Proximate analysis (wt%)

Sample

Moisture

Volatile matter Ash (dry) (daf)

Ultimate analysis (wt%, daf)

Fixed carbon

C

H

N

O

(daf)

BET (m2/g)

Calorific value (kcal/kg)

Parent coal (PC)

7.4

59.0

2.6

41.0

69.0

4.3

0.3

26.4

3.9

5,950

Extracted coal (EC)

-

33.6

< 0.1

66.4

87.9

4.3

0.0

7.9

0.2

8,160

Residual coal (RC)

-

39.5

4.6

60.5

87.7

4.2

0.0

8.0

4.1

7,490

2-2. Preparation of char The pyrolysis of the samples (PC, EC, and RC) was carried out at 300, 400, 500, 600, 700, 800, and 900 °C. The sample (~ 5 g) was heated at 10 °C/min ramp rate and kept at the predetermined peak temperature for 30 min under N2 sweep gas, which prevented coal oxidation and removed the pyrolyzed from the reaction zone. The product was then naturally cooled down to room temperature. The chars were named with the combination of the char origin (PC, EC, and RC) and the pyrolysis temperature; for example, ‘EC-600’ corresponds to the char made of EC at 600 °C.

4 2-3. Solid-state 13C CP/MAS NMR Solid-state 13C nuclear magnetic resonance (NMR) was performed using the CP/MAS NMR spectrometer (400 MHz, Bruker II+ model, Avance co.) at KBSI Daegue center. The measurements were made at the carbon frequency = 100.62 MHz. ~70 mg of sample was loaded in a sample rotor. All the spectra were acquired, employing contact time of 2 ms, repetition time of 3 sec, and 90° 1H pulse width of 3.8 μs. This was combined with a magic angle-spinning frequency of 13.5 kHz. The chemical shifts were calibrated with respect to tetramethylsilane (TMS) using the peak of the methyl group on hexamethyl-benzene at 0 ppm as the external standard. 3. RESULTS and DISCUSSION 3-1. Proximate/ultimate analysis and calorific values The proximate analysis data of the pyrolyzed PC, EC, and RC are given in Table 2. The ash content of PC and RC increased as the pyrolysis temperature increased. The ash content of PC changed from 2.6 wt.% for the untreated (RT) to 5.9 wt.% after the carbonization at 900 °C. For residual coal (RC), the ash content increased from 4.6 wt.% at RT to 7.9 wt.% after heat treatment at 900 °C. This increase resulted from the enrichment of ash due to the loss of volatile matters by raising temperature. During the solvent extraction, most of ashes in PC were retained in RC and the ash content in EC chars was less than 0.1 wt.%, regardless of the temperature. With the rise in the carbonization temperature, a marked decrease in volatile matter along with an increase in fixed carbon content was observed for all three samples, leading to an increase of fuel ratio. The change is most significant for the chars treated at 500 − 700 °C. Higher content of volatile matter in RC than in EC seemed to be attributed to devolatilization, which happened during the solvent drying step at 300 °C. EC is more susceptible to devolatilization than RC, since EC has more light components (low molecular weight) readily volatized or decomposed at the elevated temperatures [21]. Much higher fuel ratio (Fixed carbon wt.% / volatile matter wt.%) of EC than PC and RC agree with the above idea. More than 65% of the volatiles for PC/EC and ~50% for RC, were released by heating at 900 °C, compared to the chars made at 300 °C. Table 2. Results of proximate analysis and calorific values of pyrolyzed PC, EC, and RC. Ash (wt.%, dry)

Volatile matter (wt.%, daf*)

Fixed carbon (wt.%, daf)

Calorific value (kcal / Kg)

Fuel ratio

T (°C)

PC

EC

RC

PC

EC

RC

PC

EC

RC

PC

EC

RC

PC

EC

RC

300

2.0

0.0

4.7

52.6

32.0

34.5

47.4

68.0

65.5

0.9

2.1

1.9

6020

7610

7330

400

2.9

0.0

4.9

50.6

24.8

32.1

49.4

75.2

67.9

1.0

3.0

2.1

6150

7820

7350

500

3.7

0.0

5.3

49.0

24.3

32.5

51.0

75.7

67.5

1.0

3.1

2.1

6120

7940

7390

600

3.9

0.0

6.7

37.5

16.4

26.4

62.5

83.6

73.6

1.7

5.1

2.8

6910

8020

7410

700

4.8

0.0

7.2

23.3

16.4

22.2

76.7

83.6

77.8

3.3

5.1

3.5

7490

8120

7430

800

4.9

0.0

7.8

19.4

11.4

18.7

80.5

88.6

81.3

4.2

7.8

4.4

7520

7960

7480

900

5.9

0.0

7.9

17.7

8.9

15.3

82.3

91.1

84.7

4.6

10.2

5.6

7430

7980

7520

*daf: dry, ash free

5

The calorific values of PC increased notably by heating at 500 − 700 °C, from 6120 kcal/Kg for PC-500 to 7490 kcal/Kg for PC-700. Negligible changes were, however, shown at the other temperatures. The abrupt increase seemed to be due to a decrease of oxygen content coupled with an increase in carbon content. As shown in Table 3, data of the elemental analysis indicated that oxygen content was reduced from 23.7 wt.% at PC-500 to 5.2 wt.% at PC-700. It has been known that the calorific values of the oxidized coals are lower than those of the counterpart [31]. On the contrary, temperature dependence was not observed for EC and RC, showing 7610 − 8120 kcal/Kg for EC and 7330 − 7520 kcal/Kg for RC. These values are as high as those for usual bituminous coals. The heating values of EC were higher than those of PC and RC, especially for the chars prepared below 600 °C, partly as a result of its lower ash and oxygen content [31, 32]. Table 3. Results of ultimate analysis of pyrolyzed PC, EC, and RC. C (wt.%, daf)

H (wt.%, daf)

N (wt.%, daf)

H/C, O/C

O (wt.%, daf)

T

PC

EC

RC

PC

EC

RC

PC

EC

RC

PC

EC

RC

PC

EC

RC

300

70.8

84.1

84.2

3.3

3.9

3.4

0.2

0.0

0.1

25.6

12.0

12.3

0.05, 0.36

0.05, 0.14

0.04, 0.15

400

72.6

87.0

85.1

3.6

3.8

3.3

0.2

0.0

0.3

23.5

9.3

11.3

0.05, 0.32

0.04, 0.11

0.04, 0.13

500

73.4

87.3

85.8

2.4

2.2

2.0

0.5

0.0

0.4

23.7

10.5

10.7

0.03, 0.32

0.03, 0.12

0.02, 0.12

600

82.1

91.5

89.0

2.5

2.0

2.2

0.8

0.0

0.7

14.6

6.5

8.1

0.03, 0.18

0.02, 0.07

0.02, 0.09

700

92.7

93.9

91.7

1.3

1.2

1.4

0.8

0.0

0.7

5.2

4.9

6.1

0.01, 0.06

0.01, 0.05

0.01, 0.07

800

93.6

93.2

93.0

0.8

0.5

0.8

0.5

0.3

0.4

5.1

5.9

5.7

0.01, 0.05

0.01, 0.06

0.01, 0.06

900

93.7

94.6

95.5

0.6

0.6

0.6

0.2

0.2

0.1

5.4

4.6

3.8

0.01, 0.06

0.01, 0.05

0.01, 0.04

*daf: dry, ash free

As expected, the carbon content increased by increasing the heating temperature (Table 3), while the hydrogen and oxygen content decreased. No difference in the nitrogen content was observed on temperature variation. In addition, 0.01 − 0.1 wt.% of sulfur remained in the chars and its content was not affected by the different pyrolysis conditions (data is not shown). A decrease of the hydrogen content according to the temperature occurred either by devolatilization of alkyl groups or condensation of the aromatic molecules [32]. The oxygen was most likely released as water or CO2 by decomposition of oxygen containing functional group (such as carboxylic and etheric group) at the higher temperature. The carbonized products were generally of more condensed aromatic form than the raw coals and the oxygen contents of the chars pyrolyzed above 700 °C were 4 − 6%, which corresponded to the oxygen level of typical bituminous coal. The H/C ratio of the three chars and its temperature dependence were very similar to each other in spite of a large difference in their nature. In common, the H/C ratio varied from 0.05 at 300 °C to 0.01 at 900 °C, decreasing gradually with the increase of the pyrolysis temperature. Again, this resulted from devolatilizing low molecular weight fragments and dehydrogenation of the coals [33]. The O/C ratios of PC obtained at ≤ 600 °C were at least twice as large as those of

6 EC and RC, indicating the difference in the chemical compositions between PC and EC/RC. The PC-originated chars at ≤ 600 °C were richer in oxygen functional groups as well as aliphatic hydrocarbons [19]. The pyrolysis at the higher temperature (600 − 700 °C) devolatilized the majority of the oxygenated groups in PC, giving about the same level of O/C ratio at ≥ 700 °C as in EC/RC. 3-2. Solid-state 13C-NMR The 13C-NMR spectra of the chars made of PC, EC, and RC are shown in Fig. 2. The NMR spectra revealed the presence of a variety of carbon types and also estimated the ratio of aromatic and aliphatic carbon [34−37]. A weak peak appeared at chemical shift (δ) = 190 − 175 ppm, a characteristic peak of carboxylic carbons, for PC-300 and PC-400 in Fig. 2(a), but disappeared when heat-treated at the higher temperatures. That peak was absent for both EC (Fig. 2(b)) and RC (Fig. 2(c)) at all the temperatures.

7 0 0

5 9 5

(a) PC

(b) EC PC-800

EC-700

6 0 0

PC-700

Intensity (arb.)

Intensity (arb.)

4 9 5

5 0 0

PC-600 4 0 0

PC-500 3 0 0

EC-600 3 9 5

EC-500 2 9 5

EC-400

PC-400

1 9 5

2 0 0

PC-300 1 0 0

5 9

200

5 9 0

150

100

ppm

50

0

200

150

100

EC-300 50

ppm

(c) RC RC-700

Intensity (arb.)

4 9 0

RC-600 3 9 0

RC-500 2 9 0

RC-400 1 9 0

RC-300 0 9

200

150

100

ppm

50

0

Fig. 2. 13C NMR spectra of chars made of (a) PC, (b) EC, and (c) RC.

0

7 A phenolic carbon peak showed up at δ = 165 − 150 ppm. A phenolic peak was stronger in PC than in EC/RC. For all the samples the intensity became weaker with an increase in the temperature, such that it disappeared when the carbonization was done at ≥ 700 °C. The oxygencontaining functional groups were removed from the aromatic rings during the high temperature carbonization. Broad peaks covering 150 − 100 ppm was the strongest with the maximum positioned at ~128 ppm. This peak corresponds to the overlap of non-protonated (148 − 129 ppm) and protonated (129 − 93 ppm) aromatic carbon peaks [35, 36]. The protonated peaks in PC were relatively strong, compared to those in EC/RC. Intense peaks at δ = 50 − 0 ppm were shown for PC-300 and PC-400, whose intensity was comparable to the intensity of their aromatic components (Fig. 2(a)). Much smaller peaks were detected for EC and RC at the same condition (Fig. 2(b) and (c)). The peaks at 50 − 25 ppm can be assigned to unsaturated aliphatic carbons. The peaks in the 25 − 0 ppm region are due to the resonance of methyl groups attached either to straight-chain aliphatic groups or aromatic/alicyclic structures [34]. The strongest peak in PC-300 and PC-400 was found at ~31 ppm and known to be related to the methylenic carbon [34]. A lack of aliphatic peaks in EC/RC was most likely caused by the severe solvent drying (300 °C for 3 − 4 hr), indicating that the aliphatic molecules were devolatilized at the lower temperature than any other aromatic species [28]. Therefore, significant loss of volatile matters observed in Table 2 at the higher temperature seemed to come from small aromatic molecules. The aromaticity of the three components became higher with increasing temperature, in agreement with the ultimate analysis results.

CONCLUSION Chars made of three different coals (sub-bituminous coal, its solvent extracted and residual portion) have been characterized using proximate/ultimate analysis and solid-state 13C-NMR. The chemical analysis data agreed with the expected trends, which were an increase in carbon content/fuel ratio, a decrease in volatile matter/hydrogen/oxygen contents with increasing pyrolysis temperature. The change was more significant at 500 − 700 °C. An increase of the calorific value in PC was observed, which was due to a decrease of oxygen content and a concomitant increase in carbon content. The hydrogen content decreased with increasing carbonization temperature, as a result of devolatilization of alkyl groups and condensation of the aromatic molecules. The changes in the chemical structures were examined using solid-state 13CNMR results. With an increase of the pyrolysis temperature, aliphatic hydrocarbons were removed, such that only a negligible amount of the aliphatic groups existed above 500 °C.

8 Reference 1. International Energy Agency (IEA), World Energy Outlook 2007 China and India Insights, IEA, Paris, 2007, p663. 2. R. Kurose, M. Ikeda, H. Makino, M. Kimoto, T. Miyazaki, Fuel 83 (2004) 1777. 3. K. Steel, J. Besida, T. O’Donnell, D. Wood, Fuel Processing Technol. 70 (2001) 171. 4. K. Steel, J. Besida, T. O’Donnell, D. Wood, Fuel Processing Technol. 70 (2001) 193. 5. K. Steel, J. Patrick, Fuel 80 (2001) 2019. 6. Kershaw Jr., Fuel 76 (1997) 453. 7. K. Bartle, D. Jones, H. Pakdel, C. Snape, A. Calimli, A. Olcay, T. Tugrul, Nature 277 (1979) 284. 8. T. Takanohashi, T. Yanagida, M. Iino, Enegy Fuel 10 (1996) 1128. 9. L. Pang, A. Vassallo, M. Wilson, Org. Geochem. 16 (1990) 853. 10. B. Van Bodegom, J. Van Veen, G. Van Kessel, M. Sinnige-Nijssen, H. Stuiver, Fuel 63 (1984) 346. 11. N. Kashimura, J. Hayashi, T. Chiba, Fuel 83 (2004) 353. 12. K. Miura, M. Shimada, K. Mae, Y. Huan., Fuel 80 (2001) 1573. 13. K. Renganathan, J. Zondlo, E. Mintz, P. Kneisl, A. Stiller, Fuel Process Technol. 18 (1988) 273. 14. K. Miura, M. Shimada, K. Mae, S. Yoo, Fuel 80 (2001) 1573. 15. H. Shui, Z. Wang, G. Wang, Fuel 85 (2006) 1798. 16. C. Li, T. Takanohashi, T. Yoshida, Fuel 83 (2004) 727. 17. D. L. Khoury, "Coal Cleaning Technology", Noyes Data Corporation, Park Ridge, New Jersey (1981). 18. G. Andrews, M. Dorroch, T. Hansson, Biotechnol. Bioeng. 32 (1988) 81. 19. R. Ashida, M. Morimoto, Y. Makino, S. Umemoto, H. Nakagawa, K. Miura, K. Saito, K. Kato, Fuel 80 (2009) 1485. 20. N. Okuyama, N. Komatsu, T. Shigehisa, T. Kaneko, S. Tsuruya, Fuel Processing Technol. 85 (2004) 947. 21. T. Takanohashi, T. Shishido, H. Kawashima, I. Saito, Fuel 87 (2008) 592. 22. X. Li, Z. Zhua, R. De Marcob, J. Bradleya, A. Dicksa, J. Power Sources 195 (2010) 4051. 23. M. Casal, A. González, C. Canga, C. Barriocanal, J. Pis, R. Alvarez, M.A. Díez, Fuel Processing Technol. 84 (2003) 47. 24. A. Sharma, T. Takanohashi, I. Saito, Fuel 87 (2008) 2686. 25. R. Andrews, T. Rantell, D. Jacques, J. Hower, J. Steven Gardner, M. Amick, Fuel 89 (2010) 2640. 26. X. Li, J. Hayashi, C. Li, Fuel 85 (2006) 1700. 27. M. Alonso, A.Borrego, D. Alvarez, J. Parra, R. Menéndez, J. Anal. Appl. Pyrolysis 58–59 (2001) 887. 28. M. Maroto-Valer, C. Atkinson, R. Willmers, C. Snape, Energy Fuel 12 (1998) 833. 29. J. Yua, J. Lucasb, Terry F. Wall, Prog. Energy Combustion Sci. 33 (2007) 135. 30. P. Nelson, I. Smith, R. Tyler, J. Mackie, Energy Fuel 2 (1988) 391. 31. M. Sakaguchi, K. Laursen, H. Nakagawa, K. Miura, Fuel Processing Technol. 89 (2008) 391. 32. Y. Sato, S. Kushiyama, K. Tatsumoto, H. Yamaguchi, Fuel Processing Technol. 85 (2004) 1551. 33. H. Wachowska, M. Kozlowski, Fuel 75 (1996) 517. 34. S. Supaluknari, F. Larkins, P. Redlich, W. Jackson, Fuel Processing Technol. 23 (1989) 47. 35. M. Solum, R. Pugmire, D. Grant, Energy Fuels 3 (1989) 187. 36. P. Straka, J. Brus, J. Endrýsová, Chem. Pap. 56 (2002) 182. 37. B. Erdenetsogt, I. Lee, S. Lee, Y. Ko, D. Erdene, Intl. J. Coal Geology 82 (2010) 37.

An Alternative Application to the Centrifugal Dryer at a Coal Preparation Plant Ahmet Gitmez, Mustafa Yılmaz Western Lignite Establishment (GLI), Tavsanli-Kütahya, TURKEY

ABSTRACT In this study, compared with centrifugal dryer and dewatering screens substitution instead of centrifugal dryer used for drying of the fine clean coal product at the Omerler Coal Preparation Plant. This comparison was made to both rates of humidity of the product initial and investment and management costs. In the Omerler Coal Preparation Plant, fine clean coal washing at primary heavy medium cyclone and then dewatering. Fine clean coal had been feeding a centrifugal dryer after classification. The centrifugal conical sieve required change once a month and mechanical malfunction had increased, so it was thought that a different solution. The centrifugal dryer instead mounted two units dewatering screens manufactured by Omerler Coal Preparation Plant’s repair and maintenance team. The average moisture content in the range 18,5-19,5% at fine clean coal were obtained using centrifugal dryers. After the use dewatering screen, moisture content in the range 16,5-18% at fine clean coal were obtained. Initial investment cost of centrifugal dryer was $100,000. Initial investment cost of founded two units dewatering screen instead of centrifugal dryers was $50,000. These dewatering screens were mounted with machinery spare parts of Omerler Coal Preparation Plant. The annual operating cost of the centrifugal dryer has been $75,000. But the annual operating cost of two units dewatering screens have been $20,000. At the same time these dewatering screens was in operation a major contribution for dewatering process of fine clean coal. 1 INTRODUCTION Tavsanli is a town with a population of 62.000 people by Kütahya in the western part of Turkey in the Western zone. Tavsanli, the biggest town of Kütahya, continues its development as a lignite town. The town’s economy depends on coal mining which is run by the Western Lignite Establishment, a private companies and the Thermal Power Plant. Tavsanli’s general view and map are shown on Figure 1.

Figure 1. Tavsanli-Kütahya The lignite business in Turkey via the state administration began on 16/02/1938 with the Establishment of Degirmisaz enterprise of Etibank. Later, on 18.05.1939 Tunçbilek and on 23.09.1939 Soma Enterprises are operational. These three businesses combined on 01.01.1940 and established as "Western Lignite Establishment (GLI)" a subsidiary of Etibank and took place in Turkey Coal Enterprises (TKI) Authority as of 15.09.2007 which is established with the law numbered 6974. Approximately 4.6% of the lignite reserves of Turkey are in Tunçbilek. The total annual production of saleable lignite in Turkey is 50 million tons, whereas 7.5% are produced in Western Lignite Establishment (GLI) out of this. The Head Office of GLI enterprise has been operating in the open pit and underground facilities in the concession field numbered IR 4364 which is located in Kütahya province and district boundaries of Tavsanli. Approximately 80% of the scale of 7,000,000 tons 7 year run-of-mine lignite coal produced in opencasts and the rest is produced in underground businesses. The Western Lignite Establishment has removed 46 million m3 overburden, recovering 6.4 million tones of raw coal yielding 3.7 million tones coal in 2009. The Western Lignite Establishment has 30 million tones of reserves at open pit mine; 245 million tones at underground mine totaling over 275 million tones of coal reserve. The Western Lignite Establishment sells the coal to Thermal Power Station (365 MW) and individual domestic buyers.

2 DESCRIPTION of the OMERLER COAL PREPARATION PLANT The first performance tests for a 600 t/h capacity plant were completed at the end of 1993, and the final acceptance of this plant was confirmed by the end of 1994. One operator runs the processing plant with the SCADA PLC automation system. It consists of 3 sections; sizing of the run-of-mine (ROM) coal, preparation and stockpiling. The processing plant can be divided into 3 groups according to size and the required equipment. ROM coal sizes and their processing equipments can be summarized as: -150x18mm (heavy medium vessels), -18x0.5mm (heavy medium cyclones) and -0.5x0.1mm (spirals). 3 CENTRIFUGAL DRYER Centrifugal dryers in coal preparation are especially used to continuously separating mixtures of coal solids and liquids. Typical applications include coal fines, refuse fines, coarse slurry and middling. Centrifugal dryer at Omerler Coal Preparation Plant had been used to dewater the coal particles, from 10mm down to 0.5mm. Feeding capacity of centrifugal dryer had been 300 tons per hour. The photo of a centrifugal dryer was showed in Figure 2.

Figure 2. Centrifugal Dryer 4 DEWATERING SCREEN Most of the modern processes for mineral dressing and classification consume large amounts of water. Different types of machinery and equipment have been developed to recover the water used for processing and to have a final product that is easy to store. Considering its low investment and operating costs, its high capacity and ease of use, dewatering screen is the most efficient solution for recovering the water used in processing of aggregates like sand, gravel, salt and coal.

In a process, dewatering screens are mostly placed just after the hydro cyclones, monosizers, hydrosizers or spiral classifiers for a drip-free product. 4.1 Principle of Operation The screen incorporates a 45° sloping back deck section, fitted with slotted apertures across the direction of flow. Incoming slurry is fed uniformly along the top of this back section, which acts as a vibrating drainage panel. The main deck of the screen slopes upwards at 5° and is fitted with smaller slotted apertures (again slots across flow). At the lowest point of the screen, where the sloping back and main deck meet, a pool of partially dewatered slurry forms. Here solid particles bridge over the apertures and form a cake. This cake then acts as a filtration bed, only allowing very fine particles to pass through. The vibrating action conveys the cake along the screen and out of the pool where further dewatering takes place depending on the porosity of the cake. The cake is finally discharged over the adjustable weir into the product chute. Vibration is produced by two exciter motors operating at 980rpm. Alternatively geared exciters with external drive motor can be fitted to the larger screens. Easy adjustment of the amplitude of vibration, deck inclination, as well as the discharge weir plate are features incorporated to suit changes in process requirements. A high solids recovery is achieved when the screen underflow is kept in closed circuit and the only solid losses occurring would be the very fine material exiting in the cyclone overflow. With the benefits of high efficiency, high capacity, low headroom and reduced operating and maintenance costs, the screens are ideally suited for heavy-duty applications. The initial dewatered cut flows onto the top deck of a high frequency dewatering screen with a 0.5mm profile wire of Cr-Ni screen. Fine coal is taken the sieve over and bottom decks can be independently recovered. Moisture content varies from 15 to 18% depending on particle size. Screens are designed for high conductance and abrasion resistance. The operation principle of the dewatering screen was showed in Figure 3.

Figure 3. The operation principle of the dewatering screen

4.2 Operational Advantages and Gains of Application The -18x0.5 mm fine clean coals of primary heavy medium cyclone have been further classified to produce two size fractions a -18 mm by 10 mm stream and -10 mm by 0.5 mm stream. The fine clean coal has been drained, rinsed, and classified by three fine clean coal screens. The -10x0.5 mm fine clean coal has been collected on a common conveyor belt and dewatered by the centrifugal dryer until 2008. The basket screen it belongs to the centrifugal dryer has been amended once a month. At the same time this centrifugal dryer has been break downed and it was standstill in 2007. The centrifugal conical sieve required change once a month and mechanical malfunction had increased, so it was thought that a different solution. The centrifugal dryer instead mounted two units dewatering screens manufactured by Omerler Coal Preparation Plant’s repair and maintenance team in 2008. The average surface moisture content in the range 18,5-19,5% at fine clean coal were obtained using centrifugal dryers. After the use dewatering screen, surface moisture content in the range 16,5-18% at fine clean coal were obtained. These surface moisture contents belongs to the samples of the -18+0.5 mm fine clean coal it taken every shift. 5 RESULTS The centrifugal dryer instead used two units dewatering screens can be reduced the surface moisture content of the -18+0.5 mm fine clean coal from 19,5% to 18%. Initial investment cost of centrifugal dryer was $100,000. Initial investment cost of founded two units dewatering screens instead of centrifugal dryers were $50,000. These dewatering screens were mounted with machinery spare parts of Omerler Coal Preparation Plant. The annual operating cost of the centrifugal dryer has been $75,000. But the annual operating cost of the two units dewatering screens have been $20,000. Accordingly, the annual earnings are $55,000. At the same time, these dewatering screens were in operation a major contribution for dewatering process of fine clean coal.

REFERENCES CLI-Tekfen Consortium, 1991. Technical Specification of TKI Omerler Coal Washing Plant, Ankara, V-1/V-7, 1.1-4.34 http://www.linatex.com/en/process-equipment/dewatering-and-sizing-screens.html http://www.dermak.com.tr/EN/dewatering.htm http://www.kemtron.com/Products http://www.metso.com/miningandconstruction/mm_bulk.nsf http://www.elginindustries.com/sitecore/content/Solution/Industrial

The Evaluation of the Contributions to the Productivity of the Process Changes at Tuncbilek Coal Preparation Plant Ahmet Gitmez, F. Zehra Taksuk, Fatih Albayrak Western Lignite Establishment (GLI), Tavsanli-Kütahya, TURKEY ABSTRACT The most important business mission in coal washing plants, just as in all mineral processing plants, is to implement the processes and working methods in order to get access to extract the saleable products in targeted and desired quality which is theoretically possible. Coal is a non-recyclable energy source as all other underground sources. Therefore, the coal should be made best of it to ensure maximum benefit since it is produces in very difficult conditions. The process changes for the acquisiton of products for all of the additional revenue derived from the run-of-mine coal are identified in this study. At Tunçbilek Coal Preparation Plant under the direction of Western Lignite Establishment of (TKI) Turkish Coal Enterprise, the implementation regarding process changes and engineering application is discussed and an evaluation is made on the increase of production and productivity and decrease in operating costs and increase in profitability. In this assessment, the decreases in operating costs as well as increase in profitability are achieved as a result four years of implementation with respect to productivity. The approximate total income increase was annual $15,600,000, through process changes and engineering applications. 1 INTRODUCTION Tavsanli is a town with a population of 62.000 people by Kütahya in the western part of Turkey in the Western zone. Tavsanli, the biggest town of Kütahya, continues its development as a lignite town. The town’s economy depends on coal mining which is run by the Western Lignite Establishment, a private companies and the Thermal Power Plant. Tavsanli’s general view and map are shown on Figure 1.

Figure 1. Tavsanli-Kütahya

The lignite business in Turkey via the state administration began on 16/02/1938 with the Establishment of Degirmisaz enterprise of Etibank. Later, on 18.05.1939 Tunçbilek and on 23.09.1939 Soma Enterprises are operational. These three businesses combined on 01.01.1940 and established as "Western Lignite Establishment (GLI)" a subsidiary of Etibank and took place in Turkey Coal Enterprises (TKI) Authority as of 15.09.2007 which is established with the law numbered 6974. Approximately 4.6% of the lignite reserves of Turkey are in Tunçbilek. The total annual production of saleable lignite in Turkey is 50 million tons, whereas 7.5% are produced in Western Lignite Establishment (GLI) out of this. The Head Office of GLI enterprise has been operating in the open pit and underground facilities in the concession field numbered IR 4364 which is located in Kütahya province and district boundaries of Tavsanli. Approximately 80% of the scale of 7,000,000 tons 7 year run-of-mine lignite coal produced in opencasts and the rest is produced in underground businesses. The Western Lignite Establishment has removed 46 million m3 overburden, recovering 6.4 million tones of raw coal yielding 3.7 million tones coal in 2009. The Western Lignite Establishment has 30 million tones of reserves at open pit mine; 245 million tones at underground mine totaling over 275 million tones of coal reserve. The Western Lignite Establishment sells the coal to Thermal Power Station (365 MW) and individual domestic buyers. 2 DESCRIPTION of the TUNCBILEK COAL PREPARATION PLANT Coal preparation plant has two parallel circuits with the capacity of 350 tph each. Tunçbilek coal preparation plant general view is shown on Figure 2. Raw coal is fed from 600 X 300 mm grate and screened on a 150 mm opening screen. Oversize raw coal is crushed with a rotary breaker below 150 mm and combined with undersize raw coal. This prepared material is divided into two parallel circuits. Minus150 mm size material is fed to 18 mm screens.

Figure 2. A view of Tunçbilek preparation plant

Coal preparation plant section is considered in three sections: -150+18mm heavy medium drums as two parallel units each of capacity 200 ton/hour; -18+0,5mm primary heavy medium cyclones and secondary heavy medium cyclones; -0,5+0,1mm spirals. 3 The DEVELOPMENT PROCESS of TUNCBILEK COAL PREPARATION PLANT Tuncbilek Coal Preparation Plant was established in years 1952-1954 as a Jig Lavvar with the capacity 286 t/h. In 1958, Jig group was established to the fine coal part. Jig has a capacity of 150 t/h and used in 0.5-18 mm. coal dust enrichment work until 2005. During 1967-1968 revision, the Jig with the capacity 286 t/h, working with compressed air and water was dismantled due to the many failures and replaced by two pieces of magnetite working with a heavy media (2 * 200 t / h capacity, 3600 * 4700 mm in size) washing drum mounted. In 1968, the heavy medium cyclone circuit (Roberts and Schaefer) has been established to the fine coal part. Its capacity is 150 t/h and is used for enrichment of 0.5-18 mm until March 2006. In 1985 the treatment plants were established. Thereby, coals in 1000 t/h waste water previously flowing to Adranos Stream , are acquired and pumped to open pit ponds within the coal plants by the cyclones and so the solids sinked and overflowed clean water can be reused in lavvar. In 1996 thickener pool was established. Thus, 1000t/h waste water chemical precipitants (flocculants) were cleaned with 800t/h by feeding back part plant is used as wash water. 200 t/h waste water collapsed into the thickener floor is pumped to the open pit pond. In April 2006, the Jig circuit which was established in 1958 was dismantled, and then 1st heavy medium cyclone circuit to enrich 0.5 to 18 mm. coal was established instead of it. In April 2006, the current heavy medium cyclone circuit which was established in 1968 has been adapted as 2nd heavy medium cyclone circuit in order to obtain intermediate product from the separated part of 1st heavy medium cyclone circuit. In April 2006, the spiral circuit was established to enrich the 0.1 - 0.5 mm. slime coal 4 PROCESS CHANGES DONE and ITS ECONOMICAL YIELDS 4.1 Establishment of 1st Heavy Medium Cyclone Circuit by Dismantling of Current Jig Circuit Before 2006, 50% of the 0.5-18 coal obtained from the Lavvar was enriched in Jig circuit. It is determined by studies at different times, the ratio of clean coal in the Jig Schist is 10% of run-of-mine coal fed to Jig. This determination shows that 0.5 - 18 clean coal recovery as of April 2006 is 110.122 tones and the monetary value of it is approximately USD 7,693,733. The amount of washed coal and its monetary value between the years 2006 -2009 are shown in the following table 1.

Table 1. The amount of washed coal and its monetary value AMOUNT OF (0.5/18) COAL SOLD (TON)

AMOUNT OF 0.5/18 CLEAN COAL RECOVERY* %50*%10 (TON)

UNIT PRICE (USD/TON)

AMOUNT (USD)

2006

335,098

16,755

46,50

779,108

2007

579,574

28,979

66,32

1,921,887

2008

771,386

38,569

80,20

3,093,234

2009

516,376

25,819

73,57

1,899,504

2,202,434

110,122

69,87

7,693,733

YEARS

TOTAL

4.2 Adaptation of Current Heavy Media Cyclone Circuit as 2nd Heavy Fluid Cyclone Circuit In 2006, the current Heavy Media Cyclone Circuit in the fine coal circuit has been adapted as 2nd Heavy media cyclone circuit and by this revision the thin mixed product has been obtained. Since April 2006, obtained from the total amount of 245.915 tonnes of mixed product and the monetary value of approximately is 8,901,328 USD. The production and monetary value between the years are shown in table 2: Table 2. The production and monetary value RUN-OF-MINE COAL FED (TON)

AMOUNT OF GAINED MIXTURE

UNIT PRICE (USD/TON)

AMOUNT (USD)

2006

2.519.851

62.000

31,29

1.939.980

2007

2.730.464

61.365

36,22

2.222.640

2008

3.323.009

60.725

37,82

2.296.620

2009

2.838.541

61.825

39,50

2.442.088

TOTAL

11,411,865

245,915

36,20

8,901,328

YEARS

4.3 Establishment of Spiral Circuit By the establishment of Spiral Circuit in April 2006, 0.1-0.5 mm was begun to enriched, and approximately 2% of run-of-mine coal fed as clean coal has been obtained. This obtained product is offered to sale within the coal dust. Before the establishment of spiral cycle, the product obtained by classification only was given to the thermal power plants. The amount of coal is at a rate of 3% of the amount of run-of-mine coal fed in this case. The acquisition that would be obtained is shown in Table 4 if the coal production had been continued in this system.

Table 4. – Material gain if 0.5+0.1 mm coal were classified and sold YEAR 2006

RUN-OF-MINE COAL FED CLASSIFIED RUN-OF-MINE (TON) COAL * 3%

UNIT PRICE (USD/TON)

AMOUNT (USD)

2006

1.828.610

54.858

31,29

1,716,507

2007

2.730.464

81.914

36,22

2,966,925

2008

3.323.009

99.690

37,82

3,770,276

2009

2.838.541

85.156

39,50

3,363,662

TOTAL

11,411,865

342,356

34,52

11,817,370

After the establishment of spiral unit, material gains from the sales to the industrial sector instead of thermal power plant are shown in Table 5. Table 5. – Material gains from the sale of 0.5+0.1 mm coal after enrichment in the spiral: RUN-OF-MINE COAL FED (TON)

AMOUNT OBTAINED: RUN-OF-MINE COAL* %2

UNIT PRICE (USD/TON)

2006

1.828.610

36,572

46,50

2007

2.730.464

54,609

66,32

2008

3.323.009

66,460

80,20

2009

2.838.541

56,771

73,57

TOTAL

11,411,865

214,412

69,87

YEAR 2006

AMOUNT (USD)

14,980,966

By running the spiral enrichment unit, it could be possible to produce clean coal from 0.5+0.1 mm sized raw coal. Therefore more amount of clean coal could be sold to industrial consumers instead of selling raw coal to Thermal Power Plant. The difference is the additional income. 4.4 The Establishment of Coarse Mixture Product Circuit In 2010, in order to produce big mixture, the second heavy medium roller circuit will be established. With a second washing process, it would be possible to produce big mixture from the mixture and the schist which are removed from big coal washing unit. By running the system it is estimated that an average of 175,000 tons of big mixture will be produced per year. At least $7,000,000 income will be expected by selling this mixture to Thermal Power Plant with a price of $40 per ton. 4.5 The Income Obtained from Regaining of Coal from the Material which Goes to the Thickener In current system, the coal inside of the slime coal which is smaller than 0.1 mm can not be taken and therefore this slime coal is sent to the thickener then goes to waste dam. In order to regain the coal inside of the 0.1 mm sized slime coal a facility will be established in Tunçbilek Lavvar in 2010. By establishing this facility, %3 of the run of mine coal will be regained and 90000 tons of merchantable coal will be produced. It is

estimated that at least $3,600,000 income will be gained by selling this product to Thermal Power Plant with a price of $40 per ton. 5 RESULTS Western Lignite Establishment has made many changes and additions due to technological developments and changing market conditions since it was established. In this study, it is seen that the process changing lead to significant increments on incoming. After removing Jig circuit and establishing the heavy medium cyclone circuit, 110,122 tons of 0,5-18 clean coal has been regained since April, 2006. Its value is approximately 7,693,733 USD. Heavy medium cyclone circuit 245,915 tons of thin mixture has been regained since 2006. Its financial value is 8,901,328 USD. Establishment of the spiral circuit led to $3,163,596 of additional income between 2006 and 2009. The total additional income resulting from the process changing in a period of four years is $19,758,657. This means an annual average of $5,000,000 additional income. In 2010 with the establishment of second heavy medium roller system it is estimated that 175000 tons of big mixture will be produced and at least $7,000,000 income will be gained. Also 90000 tons of coal will be regained from thickener waste and this will provide $3,600,000 additional income. With these new systems, an annual total of $10,600,000 will be saved. The process changing in Tunçbilek Lavvar and the additions which will be made in 2010 will provide an annual average of $15,600,000 additional income. Like all mineral preparation plants, the most important goal of coal washing plants is realizing the appropriate process and working techniques in order to reach the desired and targeted quality. In this study the importance of regaining the merchantable coal from run of mine coal is clearly showed by the help of additional incomes. REFERENCES The Evaluation of the Performance's Tuncbilek Coal Preparation, The Evaluation of the Waste’s Plant, and the Final Report of the Process Assessment Project for Omerler Coal Preparation Plant, METU, The Applied Research Project Code Nr: 87.03.05.01.06, April 1989, Ankara, sh:2-5 GLI Annual Reports, 2006,2007,2008,2009, Tavsanli, Not published.

THE DUMP TRUCK REQUIREMENT PLANNING STUDİES OF TURKİSH COAL CORPORATİON

Mustafa ZIYPAK, Turkish Coal Corporation, Machine and Supply Department, Ankara

ABSTRACT Operation field of Turkish Coal Corporation (TKİ) is lignite production , most of which is achieved from open pits. The Corporation is realized important portion of overburden removal and coal production activities carried out at open pits by means of its own facilities (personnel and equipment). For overburden removal workings carried out at TKİ’s open pits, generally electrical excavator – dump truck combinations have been aplied. Economical life of electrical excavators, whose operation costs are low, has been going on. In addition, these excavators have idle capacities due to dump truck inefficiency. On the contrary, economical life of The Corporation’s dump trucks is expired to a great extent owing to high working hours and obsolete technology. Because of this, by scraping most of existing dump trucks of Turkish Coal Corporation, instead new dump truck investment is inevitable. In this study, by taking capacities and loading heights of existing excavators as a base, classes of trucks to be purchased have been determined; by considering estimated overburden removal to be performed by TKİ facilities, total capacity of excavators which is working at TKİ site, annual capacity of dump trucks whose economical life is not completed, annual capacity of dragline excavators that are working at TKİ sites, number of dump trucks to be invested have been calculated. Furthermore, comparison of operating costs of advanced technology dump trucks (electrical) with old technology mechanical trucks have been made; calculation of pay back period of trucks to be replaced has been made.

1. GENERAL INFORMATION Turkish Coal Enterprise (TKİ) which carries out activities of production, enrichment and marketing of lignite, that is one of the important energy sources of Turkey, is among the leader mining establishments in Turkey. TKİ, which was founded in 1957, is a state owned organization. For the year of 2009, sales income of the Enterprise is approximately $1.500.000.000, profit of that is $280.000.000. As end of 2009, TKİ recruited 8832 personnel in total. Lignite production activities of the Corporation have been conducted by 8 limited autorised companies which are spread in various regions. These are as followings: - ELİ: Aegean Lignite Establishment (Soma/Manisa) - GLİ: Western Lignite Establishment (Tavşanlı/Kütahya) - SLİ: Seyitömer Lignite Establishment (Seyitömer/Kütahya) - GELİ: South Aegean Lignite Establishment (Yatağan/Muğla) - YLİ: Yeniköy Lignite Establishment (Milas/Muğla) - ÇLİ: Çan Lignite Establishment (Çan/Çanakkale) - BLİ: Bursa Lignite Establishment (Orhaneli/Bursa) - ILİ: Ilgın Lignite Establishment (Ilgın/Konya) While 20% of total known lignite reserves in Turkey belongs to TKİ Corporation, the company realises 50-55% of lignite production in the country. 2009 lignite production of the corporation is 43 million tonnes.

80% of lignite produced by TKİ is used for power plant, remaning is used for heating and industry. TKİ suplies 8 power plant at 4209 MW which corresponds to approximately 10% of Turkey’s total electricity production capacity. TKİ Corporation produces Lignite by surface mining as well as by underground mining. For 2009 production 80% lignite from open pits, remaining by underground pits was obtained. However, it is estimated that underground rate will remain around last achieved level and open pit production will preserve its importance. 2. SURFACE MINING AT TKİ As known, overburden removal is an important activity of surface mining which is most costly. TKİ Corporation performs work of overburden removal by means of its own facilities (machinery, personnel) as well as outsourcing application. At recent years, The Corporation carries out 90-100 million m3 overburden removal (on average 42%) annually by its facilities, between 200-275 million m3 in total together with outsourcing. For overburden removal at TKİ’s open pits which is carried out by its facilities, generally electrical ekscavator (shovel) – dump truck method have been applied. In this method, firstly the area to be excavated and loaded are softened by means of drilling and blasting, after trucks are loaded by electrical excavators with 2-6 passes depending on compatibility. Trucks are unloaded at dumping site in 1-8 km distance according to projects. In the case of lacking of electrical excavators or owing to platform conditions, sometimes high capacity hydraulic excavators (backhoe type) may be used. In landslided areas, in addition to electrical excavators, high capacity loaders (12-14 yd3) and hydraulic excavators have also been used. For lignite production mostly hydraulic excavators and loaders, at intermittent coal seams fully hydraulic excavators are operated. Moreover, at some of TKİ’s open pits dragline machines that work integrated with excavator – truck combinations have been used. Draglines are usually used for digging overburden at first stage and dumping it to area where coal production has allready been completed. Since it does not require transportation, it is the most economical method at surface mining. 3. A GENERAL VIEW TO OPEN PIT MACHINERY OF TURKISH COAL CORPORATION Prior to 1980 open pit machinery of TKİ Corporation was limited, therefore coal production from surface mining was quite low. The eighties were TKİ’s progress years; huge machinery investments were realised (majority of coal fired power plant investment coincided during same period). Existing mining machinery of TKİ are comprised mostly the machines which purchased during the eighties which can be seen in table 3.1. For the works of TKİ’s overburden removal and coal production, various type, mark, model approximately 950 mining machinery (excavator, dump truck, loder, buldozer, grader, drilling machine etc.) have been used. In addition, The Corporation owns approximately 800 auxiliary machinery (fuel tanker, road truck, watering truck, forklift, mobile vinch etc.) From the nineties machinery investments have been carried out as replacement at low volumes. Outstanding investments from 1990 up to today have been 12-14 yd3 capacity loaders (18 pieces), 170 short ton electrical dump truck (20 pieces), 6.5 yd3 capacity hydraulic excavators (22 pieces). Purchasing of the mentioned loaders and excavators have been realised gradually. During last ten years, nearly 8-9% of mining machinery have been replaced; these are excavators, loaders, bulldoders, graders, auxiliary machines etc. As firstly dump trucks, due to high age and working hours, performances of majority of TKİ’s mining machinery have dropped and operating cost of those increased; therefore economical lives of the machinery have expired. However economical lives of the machines procured during nineties and millennium years and electrical machinery namely excavators, draglines have still been going on.

Table 3.1 According to Commissioning Years, Distribution of TKİ Machinery Machine Dragline Electrical exc. Hyraulic exc. Dump truck Bulldozer Grader Loader Drilling machine Total %

2006 0 0 0 0 9 17 0 0 7 1 10 4 16 9

Total 8 68 40 466 173 50 71

0

24

28

1

0

2

0

55

12 1,3

404 43,1

354 37,7

38 4,0

55 5,9

44 4,7

31 3,3

931 100,0

4. DUMP TRUCKS AND EXCAVATORS As mentioned above, excavators and dump trucks are main machinery for overburden removel and coal production. As 170 short ton 20 electical, 85 short ton 372 rigid mechanical, 50 short ton 50 rigid mechanical, 120-150 short ton 24 trailer type, there are totally 462 dump trucks in TKİ’s Machinery park. Total capacity of these trucks is 40850 short ton. Transportation of almost all TKİ’s own overburden has been accomplished by means of 85 and 170 short ton dump trucks. 50 short ton dump trucks are usually used for coal transportation. 120-150 short ton trailer type dump trucks are also specially designed for coal transportation. Experiences show that transportation cost for a surface coal mining is approximately 50% of total overburden removal cost. Thus it is very important performance of dump trucks. Performance level of dump trucks are mainly monitored by working hours and transported material. In this context, operation and maintenance costs are also important parameters on which more attention should be paid. As shown in previous section, total working hours of 85 short ton dump trucks have increased considerably due to high age (table 4.1). In addition to high age, as being products of old tecnology causes high operating cost and therefore ending economical life. Athough working hours of 170 short ton electrical dump trucks have reached at certain level, they still fulfill high performance and thus have not completed their economical life. Table 4.1 Working Hours Distribution of TKİ Dump Trucks

Dump Truck Tonnage 85 s.ton ORAN (%) 170 s.ton ORAN (%)

Toplam 0(adet) 10.000

Total 10.000 20.000 30.000 40.000 50.000 60.000 > averega 70.000 working 20.000 30.000 40.000 50.000 60.000 70.000 hours

372

0

8

30

60

130

89

37

18

100

0

2,2

8,1

16,1

34,9

23,9

9,9

4,8

8

2

10

40,0

10,0

50,0

0,0

0,0

0,0

20 100

0,0

0,0

47317

35.925

Electrical excavators were put into operation during the eighties in the scope of the open pit projects that TKİ started. Movements of bom and bucket of electrical excavators, which are shovel type, are achieved by means of wire ropes. Total number of excavators, whose bucket capacities ranges from 10 to 20 yd3, are 68 and total capacity of those are 985 yd3. Average working hour of TKİ’s shovels is around 47500. Since electrical excavators are built from durable units and parts, reliabilities and lives of them are high. In fact, it is argued by experts that economical life of electrical excavators can reach at 100.000 hours. Therefore, it can be stated that TKİ’s excavators are at half life. Due to explained reasons, operating cost of electrical excavators is quite lower than that of diesel powered hydraulic excavators. Moveover, although ages of these machines are quite high, there is no any problem in procuring spare parts. 5. EXAMİNATION OF DUMP TRUCK OPERATION COSTS Direct cost of overburden removal and coal production activities carried out in the open pits consists of drilling and blasting, excavating-loading, transpotation and auxialliry machinery expenditures. Among these, dump truck, used for transportation, cost comes first; because truck costs constitute 45-50% of total overburden removal cost. Main items of dump truck operation cost (including maintenace cost) are fuel, spare part, tire, oil, driver and maintenace labour expences. For comparison, operation costs of 85 short ton mechanical and 170 short ton electrical dump trucks have been calculated by taking ELİ ve GLİ operating data as base. The costs heve been determined as $/hour and $/m3, and showed in table5.1 and 5.2 respectively. It is important to stress that there are similarities and differences in working conditions of ELİ and GLİ. Table 5.1 Dump Truck Operation Costs ($/hour) Driver

Maintenace

Tonnage

Fuel

85 S.ton

68,9

21,67

16,28

21,09

4,36

2,79

135,09

170 S.ton

91,5

17,42

11,39

11,98

12,79

1,48

146,56

Spare part Labour

Tire

Oil

Total

Table 5.2 Dump Truck Operation Costs ($/m3) Tonnage

Maintenance

Fuel

Driver

85 S.ton

0,7

0,23

0,16

0,24

0,05

0,03

1,41

170 S.ton

0,48

0,11

0,05

0,07

0,07

0,01

0,79

Spare part Labour

Tire

Oil

Total

Examining $/hour costs in table 5.1, although double capacity, it is understood that total expenditure of 170 short ton electrical trucks is only 8,5% higher than that of 85 short ton mechanical trucks; fuel expenditure of 170 short ton electrical trucks is only 32,8% higher than that of 85 short ton mechanical trucks. Looking at $/m3 costs in table 5.2, it is calculated that total expenditure of 170 short ton electrical trucks is 56,0% of that of 85 short ton mechanical trucks; fuel expenditure of 170 short ton electrical trucks is 68,6% of that of 85 short ton mechanical trucks.

6. TKİ’S DUMP TRUCK NECESSITY DETERMINATION 6.1 Importance of New Dump Truck Investment It can easily be stated that TKİ is urgently in need of dump trucks when considering TKİ’s open pit reverves, life of these reserves, envisaging of TKİ’s own operation continuation and necessity of this, expiring of most of the existing dump trucks, being more economical of new tecnology product dump trucks (electrical), continuation of economical life of existing electrical excavators and being idle capacity of these etc.; some of which have already been mentioned in previous sections. 6.2 Determination of Dump Truck Capacity As mentioned in section 4, economical life of electrical excavators has still been going on; thus when new truck investment is realised, existing excavators will be continued to use. Because of this, in determination of the dump truck capacity, capacities and dimensions (loading height) of existing excavators should be considered. In addition, for economical operation, 3 matters should be taken into consideration in surface mining. Firstly, it should be benefited from “scale of economics”, in other words as the capacity increases, it should be known that unit cost decreases. Secondly, capacities of excavators and trucks are fitted at integer multiples as close as poosible (e.g. excavator capacity: 20 m3, truck capacity: 61 m3, 3 pases). Lastly, in order to optimize total idle cost of excavator plus truck, trucks sould be loaded certain number of pases; this is determined as 3-6 pases in practice. The excavators with capacities of 10, 15, 17 and 20 yd3 exist in TKİ’s machinery park though there are differences from site to site. Moreover, there are 6 dump truck manufacturers in the world; they manufacture classes of 100, 150, 190, 240, 300 and 360 short ton standart trucks. Therefore by considering TKİ’s excavators and other matters mentioned above, it is necessery to select from these trucks. As a result of investigation that we made, the maximum capacity trucks which TKİ’s excavators can load are 190 short ton class trucks. What is more, dimensions of existing workshops provide maintenace of maximum 190 short ton class trucks. Because of these, trucks to be purchased must be 190 short ton or lower class. As a result of analysis carried out in relation to above explanations, it has been come out that 190 short ton class truck for ELİ, GLİ and ÇLİ mining sites, 150 short ton class for YLİ, 100 short ton class for SLİ are suitable. 6.3 Criterions for Dump Truck Need Determination and Quantity Calculation For determining truck need, TKİ’s own overburden removal programme has taken into consideration, envisaging on the assumption of performing coal production by existing trucks, coal production hes been neglected in this study. In this context, following paramaters have been taken as base in determinig number of trucks whose classes are chosen for every open pit site: 1. Annual planned quantity of TKİ’s own overburden removal 2. Annual capacity of the highest capacity excavators of every open pit site 3. Annual capacity of TKİ’s dump trucks whose economical life has not completed 4. Annual capacity of TKİ’s dragline machines As seen table 6.1 in determining dump truck need, 85 short ton mechanical trucks have not been considered. As dump truck investment is realised, It is envisaged that these trucks will be scrapped. Furthermore, these trucks may be used for coal production. In the context to these conditions, firstly dump truck which has economical life and dragline capacities have been deducted from planned overburden removal quantity, then the obtained value has been compared with excavator capacities, eventually small value has been used for calculating

Table 6.1 Calculation of Truck Requirement

20 yd3

5

2.250.000

20 yd3

5

Ann.o.burden Selected truck's Determination of annual truck capacity Cal.truck.num. removal Nominal D.body Average Undigged Annual prog. for Bucket capacity volume dumping density capacity Real Interger truck need number (s.ton) (m3) distance(m) (tonne/m3) (m3/truck) (m3) 20.089.880 16.000.000 190 77-107 3200 6 2,45 756.703 21,1 21 7.400.000 190 77-107 2750 6 2,2 830.413 20.536.320 8,9 9

11.000.000

6.000.000

15 yd3

9

28.140.939

5.000.000

SLİ

10.000.000

3.000.000

10 yd3

10

22.003.200

7.000.000

100

40-60

2000

6

2

636.127

11,0

11

ÇLİ BLİ

8.600.000 7.000.000

20 yd3

5

20.536.320

8.600.000

190

77-107

2000

6

2,2

950.353

9,0

9

1.000.000

15 yd3

3

7.677.384

6.000.000

YLİ

20.000.000

4.000.000

15 yd3

4

18.482.688

16.000.000

915.430 17,5

18

Ann. capacity of trucks that have econ. Life (m3)

ELİ

25.000.000

9.000.000

GLİ

15.650.000

6.000.000

GELİ

Production company

Annual own overborden removal prog. (m3)

Excavators' annual capacity of Nominal Total ann. draglines capacity Number capacity (m3) (m3) (m3)

New truck investment not needed

New truck investment not needed 150

Maneuver time (loading)………: 190 s.ton: 32 sn, 150 s.ton: 31 sn, 100 s.ton: 30 sn Maneuver time (unloading)……: 190 s.ton: 38 sn, 150 s.ton: 37 sn, 100 s.ton: 36 sn Unloading time…………...........: 190s.ton: 94 sn, 150 s.ton: 92 sn, 100 s.ton: 90 sn Average truck speed…………...: 190 s.ton: 25 km/hour, 150 s.ton: 26 km/hour, 100 ston: 27 km/hour Excavator cycle time…..………: 20 yd3: 30 sn, 15 yd3: 30 sn, 10 yd3: 28 sn Average dumping site distance..: GLİ-SLİ-YLİ: 2000m, ELİ: 2400m, ÇLİ: 2250m Undigged overburden density....: SLİ-YLİ: 2 t/m3, GLİ-ÇLİ: 2,2 t/m3, ELİ: 2,3 t/m3 Annual working time………….: 4800 hour Assumed productivity………....: GLİ-SLİ-ÇLİ-YLİ: %70, ELİ: %80

55-88

2000

6

2

truck number. The number of needed truck has been calculated by dividing the idendified value with the selected truck capacity. Excavator cycle time, dump truck maneuver time (for loading amd unloading), truck unloading time, average truck speed, average dumping site distance, undigged overburden density, annual working time and workplace productivity which are used in the calculation of capacity of the truck chosen for every open pit site have been given under table 6.1. As shown in table 6.1, according to given parameters and assumptions, number of necessary dump truck for each mining site has been calculated, i.e. - 190 short ton 21 dump trucks for ELİ, - 190 short ton 9 dump trucks for GLİ, - 190 short ton 9 dump trucks for ÇLİ, - 150 short ton 18 dump trucks for YLİ, - 100 short ton 11 dump trucks for SLİ. 6.4 Economical Evaluation of Dump Truck Investment In section 5, operation costs of TKİ’s two major dump trucks, namely 85 short ton mechanical, 170 short ton electrical have been compared in terms of $/hour and $/m3. It has been understood that operation cost of 170 short ton electrical truck is much lower and therefore more economical. For simplicity, dump truck investments for every mining site have been evaluated with payback period method. Regarding this, operation cost difference between 85 short ton mechanical truck and 170 short ton electrical truck of TKİ which is 0,62 $/hour has been taken as base. As shown in table 6.2, it has been calculated that payback period is 6.2 years for ELİ, 5.6 years for GLİ, 3.8 years for SLİ, 4.9 years for ÇLİ, 4.2 years for YLİ. It can be argued that calculated payback periods are reasonable; since economical life of new trucks is at least 20 years. Table 6.2 Economical Evaluation of TKİ Truck Investment Selected Production truck company tonnage

Neccesary Price Truck unit overburden difference price ($) removal for ($/m3) payback (m3)

Annual truck capacity (m3)

Payback period (year)

ELİ

190

2.900.000

0,62

4.677.419

756.703

6,2

GLİ

190

2.900.000

0,62

4.677.419

830.413

5,6

SLİ

100

1.500.000

0,62

2.338.710

636.127

3,8

ÇLİ YLİ

190

2.900.000

0,62

4.677.419

950.353

4,9

150

2.400.000

0,62

3.870.968

915.430

4,2

7. CONCLUSION 466 dump trucks are operated at TKİ’s open pits for overburden removal and coal production. Economical life of the most trucks has expired, operation cost of them increased. Therefore they must be replaced as soon as possible. TKİ’s dump trucks must be scrupped in short or medium term whether they are repaced or not. Consequently, as taking a strategical decision, direction of the corporation shoul be made clear regarding its own operation. If new dump truck investment will not be done, TKİ have to decrease own operations to minimum level by reducing machinery with respect to a certain Schedule.

If a positive decision is taken for new dump truck investment, according to the result achieved from this study, as 190 short ton 39 pieces, 150 short ton 18 pieces and 100 short ton 11 pieces, totally 68 dump trucks should be purchased. Total investment cost of these trucks is approximately 173.000.000$ and procurement of them can be realised step by step over years. Analysis conducted in this study indicates that considerable reduction in operation cost of trucks can be obtained and investment can pay back within 4-6 yeras. Owing to low operation cost, high operation productivity that is provided by high safety and confort, easiness of management and organization enabled by few trucks, it is estimated that cost of TKİ’s own overburden removal can compete with that of contractor easily.

REFERENCES TKİ INVESTİGATİON, PLANNİNG AND COORDİNATİON DEPARTMENT (2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010), Activity Reports of TKİ Enterprise for 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008 and 2009, Ankara/Turkey TKİ MACHINERY AND SUPPLY DEPARTMENT (2010), The book of TKİ Machinery Park, Ankara/Turkey ZIYPAK, M. (2007), Dump trucks, Examination of Their Structures and Performances, Proceedings of the 1st Mining Machinery Symposium and Exibition, Kütahya/Turkey, p.335-346 ZIYPAK, M. (2009), A General View to Open Pit Machinery of Turkish Coal Corporation, Proceedings of the 2nd Mining Machinery Symposium and Exibition, Zonguldak/Turkey, p.117-127

2010 International Pittsburgh Coal Conference Istanbul, Turkey October 11 – 14, 2010 Coal Science: Coal Chemistry Investigation of Radioactive Contents Soma Coals İsmail DEMİR, İlgin KURŞUN Istanbul University Engineering Faculty Mining Engineering Department, Istanbul, Turkey [email protected], [email protected] Abstract: Coal, world's the most abundant, the most accessible and the most versatile source of fossil energy was brought to the forefront of the global energy scene by the industrial revolution of the 18th century. Like any fossil fuel, coal is associated with naturally occurring radioactive materials. This is due to their U, Th, and K contents. This certainly has radiological implications not only for the miners but also for the populace in the immediate environment of the mines and the users. In this study, the radioactive elements in Manisa-Soma coals and their ashes were carried out. In the experimental section, the coal and thermal power plant ashes which were taken from Manisa – Soma were used. Sieve, moisture, ash, calorific value, volatile amount, total carbon, total sulphur, major element and radioactive element analysis of the samples were carried out. The float and sink analysis and flotation tests of the samples which were taken from Manisa-Soma were carried out. Thus, radioactive elements changes and moving mechanisms were investigated with coal preparation and burning methods. Furthermore, the pre-investigation of the assessment of the thermal power plant ashes was carried out with the experiments on the ash samples which were taken from Soma thermal power plant. 1. General Information Social and technological development changes in direct proportion to the amount of energy that is consumed. As a result of the fast growth of world population, consumed energy naturally increases alongside. Especially, the fact that the population growth of Turkey is higher than the worldwide average means that the requirement for energy will increase day by day. In 2008, petroleum has the highest share in energy consumption in Turkey with 32.8%, which is followed by natural gas with 30.4%, coal with 28% and the remaining 10% is occupied by renewable resources including hydraulic (TEİAŞ 2008). A clean environment is needed for a healthy life and energy is needed for a comfortable life, which requires utilisation of resources by minimising their impact. The fact that even ashes of burned coal is usable is an important point both for economic benefit and environmental impact, and this may only be possible by fulfilment of the features of existing coal. When coal is combusted in thermal power stations, toxic trace elements in the coal like As, Cd, Ga, Ge, Pb, Sb, Se, Sn, Mo, Ti and Zn which have the potential of contaminating transfer to the waste products (cinder, ash and gas). Volatile ashes containing many poisonous elements may be collected in ash collection pools under furnaces or as piles. Because resoluble metal ions and compounds that leak from the ash pools or piles have the potential to contaminate soil, and surface and underground water, severe environmental problems may occur (Karayiğit and others 2000, Perçinel 2000, Esenlik 2005, Tuna and others 2005). When coal is combusted, toxic trace elements like arsenic (As), cadmium (Cd), lead (Pb), antimony (Sb), selenium (Se), stannum (Sn) and zinc (Zn) inside that are contaminants are transferred to waste products like cinder, ash and gases. When waste products are disposed of, contained poisonous (toxic) trace elements may be conveyed to the atmosphere, earth surface and oceans. These elements may be seriously threatening for living organisms by creating environmental, area use and health problems when the waste products are washed under rain and these elements are carried away with underground water to the soil, surface waters and underground waters (Baba 2001, Ateşok 2004). Some of the diseases occurring in time on people, who live near thermal power plants, due to the toxic elements spread in the neighbourhood are given below (Perçinel 2000): As: Anaemia, nausea, renal symptoms, ulcer, skin and pulmonary cancer, defective births. Be: Malfunction of respiration and lymph, lungs, spleen and kidneys, carcinogenic effects. Cd: Lung emphysema and fibrosis, kidney diseases, cardiovascular effects, carcinogenic effects. Hg: Nervous and kidney damages, cardiovascular effects, birth problems. Mn: Respiratory problems. Ni: Skin and intestinal diseases, carcinogenic effects. Pb: Anaemia, nervous and cardiovascular problems, delayed growth, gastric and intestinal problems, carcinogenic effects, birth problems. Se: Gastric and intestinal nausea, pulmonary and splenic damages, anaemia, cancer, teratogenic effects. V: Acute and chronic respiratory malfunction (Perçinel 2000). Radon gas forms in the area in which ashes of the thermal power plant collect (ash chambers) reach the air. Even if these ashes are buried in soil, radon gas infiltrates through the pores of the soil and blends in the air. Radon gas may transform into polonium and active lead in 3.8 days. Therefore, piles of ash emit radioactivity. Perhaps the most critical material that is disposed through the chimneys is uranium that is contained in lignite and revealed during combustion to spread around. Uranium is also a serious problem (Özyurt 2006).

1.1 Major and Trace Elements Contained in Coal and Coal Ash C, H, O, N and S contained in the structure of coal, contents of which are generally higher than 1000ppm, form the organic matrix and they are called as major elements. Al, Fe, Mg, As, Zn, Cu, F, Th, V, etc with a concentration that is generally less than 1000ppm are called as trace elements in coal (Ateşok 2004, Özyurt 2006). There are some elements in coal which are inorganic based, which may form inorganic or organometalic compounds, and which may be produced if they are at an economic level. Kin the sediments containing coal layers and in coal formations, Ge, Ga, U and Cu may be found at economic levels. Other than these, coal contains toxic trace elements like Be, Mo, V, Zn, W, Co, Cd, As, Pb, Se, and Cr which are contaminants (Kural 1998, Özyurt 2006). After combustion of coal, trace elements Pb, Cu, Zn, V, As and Th become volatile and concentrated in the furnace ash (Özyurt 2006, Riley 2008). When coals are combusted in high temperatures, their molecular structure demolished, and an important portion of Cl and F is disposed into air as gases together with smoke (Özyurt 2006). When dust coal is combusted in thermal power plants, carbon, nitrogen and sulphur contained in the coal structure oxidises and transforms into carbon oxide (CO x), nitrogen oxide (NOx) and sulphur oxide (SOx). Some water vapour forms during this transformation, too. Whereas cinder is collected under combustion furnaces, volatile ashes are caught by electro-filters and some are transported with the chimney gas. Researches show that trace elements mostly collect on volatile ashes (Karayiğit and others 2000, Özyurt 2006). In thermal power plants that use coal, combustion in the furnaces occurs at around 900-1400°C depending on the type of coal. Coal pieces heat up in the furnace, vaporizable materials convert into gases and combustion occurs. Minerals disintegrate and melt under heat, start decomposing and agglomerate (Kural 1998, Özyurt 2006). Hg, As, Se, Ni, Pb, Ce are Zn mostly related to sulphide minerals and organic substances. Combination (formation) of coal minerals or organic substances with trace elements may seriously affect vaporisation limit and consequently its ratio in the chimney gas disposed by the plant. Trace elements detected in chimney gases are mostly associated to sulphide minerals (Riley 2008, Shah and others 2008). Behaviour of trace elements in coal during and after combustion is shown in the Figure. During combustion of coal, some trace elements contained in coal like As, Cd, Ga, Ge, Pb, Sr, Mo, Zn, Ba transfer to the waste products (cinder, ash and gases). Especially volatile ashes of such wastes produce very convenient media for adhesion of elements in liquids and gases because they have clayish structure, endure high temperatures and have large surface area (Özyurt 2006). 2. Findings 2.1. Results of Particle Size Analysis Coal samples collected from the site were crashed in the laboratory type jaw crusher and roll crusher in Istanbul University, Department of Mining Engineering, Ore Preparation and Concentration Laboratory, and then dry sieve analysis were applied to determine the particle size distribution. Undersize and oversize curves were drawn according to the sieve analysis performed using 8mm, 4mm, 2mm, 1mm and 0.5mm laboratory type Retsch brand stainless steel sieves of square section and d 80 size were calculated.

Quantity (%)

Figure 1 shows undersize and oversize curves drawn according to the results of sieve analysis made after crashing of coal samples collected from Manisa-Soma region in roll crusher. The curves show that d80 size of roll crusher output of Manisa – Soma coal is 5.4mm.

Undersize Elek Altı Oversize Elek Üstü

Sieve Size (mm)

Figure 1: Manisa – Soma Sieve Analysis of the Roll Crusher Output

Quantity (%)

Figure 2 shows undersize and oversize curves of the coal samples collected from Manisa – Soma Region, which were fed to the float and sink test, according to the results of sieve analysis. The curves show that d80 size of Manisa – Soma coal that is used in float and sink test is 2100µ.

ea Undersize eü Oversize

Sieve Size (µ) Figure 2: Undersize and Oversize curves of Manisa – Soma float and sink test sample 2.2. Results of Chemical Analysis Ash samples and coal samples were brought to the Ore Preparation and Concentration Laboratory of Istanbul University, and humidity, ash, density, volatile substances, total sulphur and thermal value analyses were carried out on these samples. Evaluation and interpretation of the chemical analysis results of the coal and ash samples are summarised below: Humidity: Coal samples collected from the Turkish Coal Enterprise (TKI) Ege Lignite Enterprise (Manisa-Soma) were brought to the Ore Preparation and Concentration Laboratory of Istanbul University, and total humidity analysis was made on these samples. Analysis was made according to standard TS 690 ISO 598 (Method-C). Total humidity analysis was made after bringing samples to the laboratory in closed nylon bags without losing time. Due to the high temperature at electro-filters, humidity values of ash samples are mostly almost zero.

Humidity (%)

Figure 3 shows the percentage curve of the humidity that is lost in time when coal samples collected from Manisa-Soma region are heated in drying oven at 105ºC. It has been calculated that coal samples collected from Manisa-Soma contain 15.49% humidity. The analyses show that almost entire humidity contained in the samples may be removed in 3 hours.

0 2

1 3

Humidity loss (%)

Time (h) Figure 3: Manisa-Soma Coal – Humidity Loss % vs. Time Curve Volatile Amount: It has been observed that volatile substance content of the ash samples collected in the scope of the study is lower than the coals’. Whereas Manisa-Soma coal sample contains 31.32% volatile, it is 1.2% in the ashes of the thermal power plant. The ashes have volatile substance content because the unburned coal pieces. Ash: Ash analysis of the coal samples within the study were realised in Istanbul University, Department of Mining Engineering, Ore Preparation and Concentration Laboratory. The results of this analysis show that the coals collected from Manisa – Soma region have high ash ratio.

Theoretically, ash content of waste products formed as the result of combustion in thermal power plants (fly ash and bottom ash) must be 100%. However, depending on characteristics of combusted coal and conditions of combustion systems, it is always possible to find some unburned coal remnants in these wastes. As a matter of fact, Manisa – Soma thermal power plant ash contains 1.2% unburned pieces. Total Sulphur: Total sulphur content analysis in dry base of the the coal samples were carried out in Acme Analytical Laboratories Ltd.in Canada. Leco carbon sulphur device was used in total sulphur analysis. When coals used in Manisa-Soma thermal power plant and ashes formed by combustion are examined for total sulphur contents, it is observed that a little portion is disposed into the air by burning and the rest is concentrated in ash. Thermal Value: Upper and lower thermal values of the coal samples were analysed in the Environment and Fuel Analysis Laboratory approved by TÜRKAK belonging to Istanbul Metropolitan Municipality, Department of Environmental Protection & Development. IKA C7000 device was used for thermal value analyses. Density Analysis: Density analysis with a pycnometer was performed to determine the densities of the coal samples collected for the study and to compare optimum sorting density used in float and sink experiments. Density Analysis was performed in Istanbul University, Department of Mine Engineering, Ore Preparation and Concentration Laboratory. TS ISO 5072 Brown Coal and Lignite – Assessment of Real and Apparent Relative Density standard was utilized in the analysis method. It has been concluded that Manisa – Soma coals have high density. When density analysis and optimum sorting density used in float and sink experiments are combined, it is found that it would be fit to sort Manisa – Soma coals in the heavy medium adjusted to 1.6g/cm3. Table 1: Dry base chemical analysis results of Manisa – Soma coal samples and Manisa – Soma power plant ash samples in air

Sample Manisa-Soma Coal Manisa-Soma Ash

Amount of Volatile Substances (%) 31.32 1.2

Ash Content (%)

Total Sulphur Content (%)

Upper Thermal Value (Kcal/kg)

Lower Thermal Value (Kcal/kg)

Density (g/cm3)

35.88 98.80

0.67 1.26

2942

2761

1.553

Elementary Analyses: Elementary Analysis of the coal samples collected for the study were performed in Advanced Analysis Laboratory of Istanbul University. Dry base elementary analysis results of the analysed samples in air are given in Table 2. Table 2: Dry base elementary analysis of Manisa – Soma coal samples in air Manisa – Soma

C% 43.13

H% 2.75

N% 0.42

S% 0.28

Methods used in elementary analysis are explained below: 1DX Analysis: In 1DX analysis, 0.5g of the sample is leached with Aqua Regia heated up to 95ºC (Aqua Regia is generally obtained by mixing one third concentrated hydrochloric acid and nitric acid) and the solution placed in ICP-MS device to read the values. Elements detected by 1DX analysis are: Mo, Cu, Pb, Zn, Ni, As, Cd, Sb, Bi, Ag, Au, Hg, Tl, Se Leco TOT/C and TOT/S analysis: Total C and Total S analyses are performed with Leco carbon sulphur device. 4A Analysis: In 4A analysis, 0.2g coal and ash samples are applied lithium metaborate/tetraborate fusion and decomposed with diluted nitric acid, and then major oxides they contained were detected with the ICP-ES device. Minerals that are analysed with 4A are: SiO2, Al2O3, Fe2O3, MgO, CaO, Na2O, K2O, TiO2, P2O5, MnO, Cr2O3 4B Analysis: In 4B analysis, 0.2g coal and ash samples are applied lithium metaborate/tetraborate fusion and decomposed with diluted nitric acid, and then rare soil elements and refractor elements they contained were detected with the ICP-MS device. In addition, 0.5g samples were decomposed in royal water, and precious metals and base metals were detected with ICP-MS. Elements that are analysed with 4B are (nitric acid and ICP-MS): Ba, Be Co, Cs, Ga, Hf, Nb, Rb, Sc, Sn, Sr, Ta, Th, U, V, W, Y, Zr, La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu

Elements that are analysed with 4B are (Royal Water and ICP-MS): Au, Ag, As, Bi, Cd, Cu, Hg, Mo, Ni, Pb, Sb, Se, Tl, Zn Combustible Efficiency Analysis: It is used in interpretation of combustible efficiency washing performance. Combustible efficiency has been calculated with the formulae as below: (1) Where; t: waste schist ash % f: raw coal ash % c: clean coal ash %. Float and Sink Experiments: Trace element analysis results of the products obtained by float and sink experiments are given in Attachment 1. Combustible efficiencies of samples according to the float and sink experiment results are given in Table 3. Table 1: Combustible efficiencies of the coal samples that floated and sank in float and sink experiments Float and Sink (-4mm+0,5mm) Soma +1,6 Floating Soma -1,6 Sinking Soma Fed

Quantity % 47.51 52.49 100.00

ΣC Content % 63.35 15.37 38.17

Ash % 9.90 59.40 35.88

Combustible Efficiency % 66.76 33.24 100.00

Combustible efficiencies of the flotation experiment results are given in Table 4. Table 4: Combustible efficiencies of the floating and sinking coal samples in flotation results Flotation (-0,5mm) Soma Floating Soma Sinking Soma Fed

Quantity % 61.21 38.79 100.00

ΣC Content % 31.44 30.81 31.20

Ash % 45.70 45.50 45.62

Combustible Efficiency % 61.12 38.88 100.00

Combustible efficiencies according to total coal dressing works, in which results of float and sink experiments and flotation experiments are evaluated together, are given in Table 5. Flotation Experiments: Trace element analysis results of the products obtained by flotation experiments performed in the study are given in Attachment 1 as tables. Combustible efficiencies of Manisa – Soma coal is low because it is high ash coal and lignite flotation is difficult. Table 5: Combustible efficiencies of floating and sinking coals in total after coal dressing processes Total Coal Dressing (-4mm) Soma ΣFloating Soma ΣSinking Soma ΣFed

Quantity % 50.92 49.08 100.00

ΣC Content % 53.80 18.41 36.43

Σ Ash % 45.70 45.50 45.60

Σ Combustible Efficiency % 50.83 49.17 100.00

Trace element analysis results of Manisa Soma Coal, Ash and Thermal Power Plant Ash, are given in Attachment 2.

Attachment: All Analysis Results And Averages Of Trace Elements

Float and Sink Coal Float and Sink Ashes Unsorted Coal Unsorted Coal Ashes Flotation Flotation Ashes

Soma -4+0.5mm +1.60 Floating Soma -4+0.5mm -1.60 Sinking Soma -4+0.5mm +1.60 Floating Ash Soma -4+0.5mm -1.60 Sinking Ash Soma Unsorted Coal

Method Analyte Unit MDL Coal Coal Ash Ash Coal

Soma Unsorted Coal Ashes

Ash

21.06

8.7

2.51

2.47

58.14

0.23

0.98

0.3

0.19

0.06

0.007

7.00

425.00