This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
= 1 + 0.8447 x (0.913)1'2058 x (728.2)"0'0557 x (446y°M44 RI(20) =1.5105
Refractive Index (RI) @ 60°C (140°F) RI(60) = 1 + 0.8156 x (SG)L2392 x (VABP °K)'a0576 x (MW)'0'0007 RI(60) = 1 + 0.8156 x (0.913)L2392 x (728.2)"0'0576 x (446)"0'0007 Rl{60) = 1.4963
Hydrogen (H2) Content, wt% H2 = 52.825 - 14.26 x RI(20) - 21.329 x (SG) - 0.0024 x (MW) - 0.052 x (S) + 0.757 x In (v) H, = 52.825 - 14.26 x 1.5105 - (21.329 x 0.913) - (0.0024 x
FCC Feed Characterization
75
446) - (0.052 x 0.48} - (0.872 - In (7.37)) H2 = 12.23 wt%
Aromatic (CA) Content, wt% CA = -814.136 + (635.192 x RI(20)) - (129.266 x (SG)) + (0.013 x (MW)) - (0.34 x (S)) + (0.872 x In (v)) CA = -814.136 + (635.192 x 1.5105) - (129.266 x 0.913) + (0.013 x 446) - (0.34 x 0.48) + (0.872 x In 7.37) CA = 19.19 wt%
Where: SG AP °C VABP °C VABP °K S V
= Specific gravity at 20°C = Aniline Point, °C = Volumetric Average Boiling Point, °C = Volumetric Average Boiling Point, °K = Sulfur, wt% = Viscosity at 100°C
For FCC feeds, particularly the ones containing residue, the TOTAL correlation is more accurate at predicting aromatic carbon content than the n-d-M correlation. Table 2-9 illustrates this comparison. One option is to calculate MW, RI(20)» CA, and H2 from the TOTAL correlation, and use either the n-d-M or API method to calculate the wt% naphthene (CN) and wt% paraffin (Cp).
n-d-M Method The n-d-M correlation is an ASTM (D-3238) method that uses refractive index (n), density (d), average molecular weight (MW), and sulfur (S) to estimate the percentage of total carbon distribution in the aromatic ring structure (% CA), naphthenic ring structure (CN), and paraffin chains (% Cp). Both refractive index and density are either measured or estimated at 20°C (68°F). Appendix 4 shows formulas used to calculate carbon distribution. Note that the n-d-M method calculates, for example, the percent of carbon in the aromatic ring
76
Fluid Catalytic Cracking Handbook Table 2-9 Comparison of TOTAL Correlations with Other Methods
Correlation
Carbon Content (%C) n-d-M API TOTAL Hydrogen Content (%H) Linden Fein-Wilson-Winn Modified Winn TOTAL Molecular Weight (MW) API Maxwell Kesler-Lee TOTAL Refractive Index (RI) API @ 20°C Lindee-Whitter @ 20°C TOTAL @ 20°C TOTAL @ 60°C
Absolute Average Deviation
Bias Maximum Deviation
5.14 2.88 0.93
4.67 2.53 0.00
12.99 9.13 3.45
0.31 0.36 0.19 0.10
-0.05 0.19 0.07 0.00
1.57 1.43 0.86 0.42
Average Deviation
62.0 63.3 61.5 10.6
-62.0 -63.6 -61.1 -0.20
180.9 175.0 176.9 44.4
0.0368
-0.0367
0.0993
0.0315 0.0021 0.0021
-0.0131 0.0 0.0
0.0303 0.0074 0.0074
Source: Dhuleaia [1]
structure. For instance, if there was a toluene molecule in the feed, the n-d-M method predicts six aromatic carbons (86%) versus the actual seven carbons. ASTM D-2502 is one of the most accurate methods of determining molecular weight. The method uses viscosity measurements; in the absence of viscosity data, molecular weight can be estimated using the TOTAL correlation. The n-d-M method is very sensitive to both refractive index and density. It calls for measurement or estimation of the feed refractive index at 20°C (68°F). The problem is that the majority of FCC feeds are virtually solid at 20°C and the refractometer is unable to measure
FCC Feed Characterization
77
the refractive index at this temperature. To use the n-d-M method, refractive index at 20°C needs to be estimated using published correlations. For this reason the n-d-M method is usually employed in conjunction with other correlations such as TOTAL. Example 2-3 can be used to illustrate the use of the n-d-M correlations.
Example 2-3
Using the feed property data in Example 2-1, determine MW, CA, CN and Cp using the n-d-M method. Step 1: Molecular weight determination by ASTM method. 1. Obtain viscosity at 100°F (37.8°C) a. Plot viscosities at 130°F (54.4°C) and 210°F (98.89°C). b. Extrapolate to 100°F, VIS = 279 SSU. 2. Convert viscosities from SUS to centistoke (csT): a. From Appendix 6, viscosity @ 100°F = 60.0 cSt. b. Viscosity @ 210°F = 7.37 cSt. 3. Obtain molecular weight: a. From Appendix 5, H = 372 and MW = 430. Step 2: Calculate refractive index @ 20°C from the TOTAL correlation. RI(20) = 1 + 0.8447 x (SO)1'2056 x (VABP(deg C) + 273.16r°'0557 x (MW)-0'0044 RI(20) = 1 + 0.8447 x (0.913)!'2056 x 728.2^'0557 x 446^0044 RI,20) = 1.5046 Step 3: Calculate n-d-M Factors. V = 2.51 x (RI(20) - 1.4750) - (D20 - 0.8510) = 0.0271 V = 2.51 x (1.5046 - 1.4750) - (0.90 - 0.8510) = +0.0271 w = (D20 - 0.8510) - 1.11 (RI(20J - 1.4750) = +0.0226 w - (0.90 - 0.8510) - 1.11 x (1.5046 - 1.4750) = +0.0226
78
Fluid Catalytic Cracking Handbook Because V is positive: %CA = 20.16
Because w is positive: %CR =820xw-(3xS) + -
10,000
MW 10,000
%CR = 820 x 0.0226 - 3 x 0.48 +
430
The API method is a generalized method that predicts mole fraction of paraffinic, naphthenic, or aromatic compounds for an olefin-free hydrocarbon. The development of the equations is based on dividing the hydrocarbon into two molecular ranges: heavy fractions (200 < MW < 600) and light fractions (70 < MW 276x0"913) x(!276)-°-407x(0.913)-3-3333
RI = (1 + 2 x I/I- I) l/2 /
\l/2
_ 1 + 2x0.294 V R1(20) "~l 1-0.294 J RJ(20) = 1 .500
,,^ ,7. . ^ . „ SG - 0.24- 0.222 xlog(v210- 35.5) VG = Viscosity Gravity Constant = ™ 0.755 Q=
0.913 -0.24- 0.022 xlog(50- 35.5) 0.755
VG = 0.8575 XA = g + h(Ri) + i(VG)
XA = -403.8 + 265.7 xf 1.5000-^^1 + 161. Ox 0.8575 XA = 1 1 .5 rnol% XN = d + e(Ri) + f(VG)
XN = 246.4 - 367.0 x 1.5000 XN = 31.8 mol% Xp = a + b(Ri) + i(VG)
2
+ 196.3 x 0.8575
80
Fluid Catalytic Cracking Handbook Xp = 257.37 + 101.33(Ri) + 160.988(0.8575)
Xp = 56,7 mo.1% Where: Constants a b c d e f g h i
= = = = = = = =
+2.5737 +1.0133 -3.573 +2.464 -3.6701 +1.96312 4.0377 +2.6568 +1.60988
The findings from TOTAL, n-d-M, and API are summarized in Table 2-10. The comparison illustrates how sensitive the predicted feed composition is to the refractive index @ 20°C. For instance, using the TOTAL correlation, there is a 35% drop in the aromatic content in using RI(20) = 1.5000 instead of RI(20) = 1.5105. When using these correlations, every effort should be made to obtain accurate and consistent values for the refractive index at 20°C. With the refractive
Table 2-10 Comparison of the Findings Among the 3 Correlations API
n-d-M
TOTAL
Refractive Index @ 20°C Molecular Weight
1.5000
413
430
446
Carbon Content:
Mol%
Wt%
Wt%
Aromatic Naphthene Paraffin
11.5, (14.3)* 31.8, (27.9)* 56.7, (57.8)*
(20.2)*,(8.8)t (20.2)*,(41.1)f (59.6)*,(50.1)t
19.2, (12.5)*
1.5105
*Uses Rll2i)l from n-d-M correlation to determine composition, Uses RI^0ifrorn API correlation to determine composition.
T
FCC Feed Characterization
81
index at any given temperature, the RI(2o) can be calculated from the following equation. Example 2-5 illustrates the use of the equation. RI (2m = RI(t) + 6.25 x (t - 20) x 10"4 t = temp, °C
Example 2-5
With the refractive index @ 78°C = 1.4810, determine the refractive index @ 20°C. RI(2()) = 1.4810 + 6.25 x (67 - 20) x 10"4 RI(20) = 1.5104
(Note that the calculated RI(20) closely matches that using the TOTAL correlation.)
Pretreatment of FCC feedstock through hydroprocessing has a number of benefits including: • • • • •
Hydrodesulfurization (HDS) Hydrodenitrogenation (HDN) Hydrodemetallization (HDM) Aromatic Reduction Conradson Carbon Removal
Desulfurization of FCC feedstocks reduces the sulfur content of FCC products and SOX emissions. In the United States, road diesel sulfur can be 500 ppm (0.05 wt%). In some European countries, for example in Sweden, the sulfur of road diesel is 50 ppm or less. In California, the gasoline sulfur is required to be less than 40 ppm. The EPA's complex model uses sulfur as a controlling parameter to reduce toxic emissions. With hydroprocessed FCC feeds, about 5% of feed sulfur is in the FCC gasoline. For non-hydroprocessed feeds, the FCC gasoline sulfur is typically 10% of the feed sulfur.
82
Fluid Catalytic Cracking Handbook
The nitrogen compounds in the FCC feed deactivate the FCC catalyst activity resulting in an increase in coke and dry gas. Hydrodenitrogenation reduces nitrogen compounds in FCC feeds. In the regenerator, the nitrogen and the attached heterocyclic compounds add unwanted heat to the regenerator causing a low unit conversion. Hydrodemetallization reduces the amount of nickel and, to a lesser extent, vanadium in FCC feeds. Nickel dehydrogenates feed to molecular hydrogen and aromatics. Removing these metals allows heavier gas oil cut points. Polynuclear aromatics (PNA) do not react in the FCC and tend to remain in coke. Adding hydrogen to the outer ring clusters makes them more crackable and less likely to form coke on the catalyst. Hydroprocessing reduces the Conradson carbon residue of heavy oils, Conradson carbon residue becomes coke in the FCC reactor. This excess coke must be burned in the regenerator, increasing regenerator air requirements.
It is important to characterize FCC feeds as to their molecular structure. Once the molecular configuration is known, kinetic models can be developed to predict product yields. The simplified correlations above do a reasonable job of defining hydrocarbon type and distribution in FCC feeds. Each correlation provides satisfactory results within the range for which it was developed. Whichever correlation is used, the results should be trended and compared with unit operation. A clear understanding of feed physical properties is essential to successful work in the areas of troubleshooting, catalyst selection, unit optimization, and any planned revamp.
REFERENCES 1. Dhulesia, H., "New Correlations Predict FCC Feed Characterizing Parameters," Oil & Gas Journal, January 13, 1986, pp. 51-54 2. ASTM, "Standard Test Method for Calculation of Carbon Distribution and Structural Group Analysis of Petroleum Oils by the n-d-M Method," ASTM Standard D-3238-85, 1985. 3. Riazi, M. R., and Daubert, T. E., "Prediction of the Composition of Petroleum Fractions," Ind. Eng. Chem. Process Dev., Vol. 19, No. 2, 1982, pp. 289-294.
FCC Feed Characterization
83
4. ASTM, "Standard Test Method for Estimation of Molecular Weight (Relative Molecular Mass) of Petroleum Oils from Viscosity Measurements," ASTM Standard D-2502-92, 1992, 5. Flanders, R. L., Proceedings of the 35th Annual NPRA Q&A Session on Refining and Petrochemical Technology, Philadelphia, Pa., 1982, p. 59. 6. Wollaslon, E. G., Forsythe, W. L., and Vasalos, I. A., "Sulfur Distribution in FCC Products," Oil & Gas Journal, August 2, 1971, pp. 64-69. 7. Huling, G. P., McKinney, J. D., and Readal, T, C, "Feed-Sulfur Distribution in FCC Products," Oil & Gas Journal, May 19, 1975, pp. 73-79. 8. Campagna, R. J., Krishna, A. S., and Yanik, S. J., "Research and Development Directed at Resid Cracking," Oil and Gas Journal, October 31, 1983, pp. 129-134. 9. Davison Div., W. R. Grace & Co., "Questions Frequently Asked About Cracking Catalyst," Grace Davison Catalagram, No. 64, 1982, p. 29. 10. Andreasson, H. U. and Upson, L. L., "What Makes Octane," presented at Katalistiks' 6th Annual FCC Symposium, Munich, Germany, May 2223, 1985. 11. Van, K. B., Gevers, A., and Blum, A., "FCC Unit Monitoring and Technical Service," presented at 1986 Akzo Chemicals Symposium, Amsterdam, The Netherlands. 12. Scherzer, J., and McArthur, D. P., "Nitrogen Resistance of FCC Catalysts," presented at Katalistiks' 8th Annual FCC Symposium, Venice, Italy, 1986. 13. Dougan, T. J., Alkemade, V, Lakhampel, B., and Brock, L. T., "Advances in FCC Vanadium Tolerance," presented at NPRA Annual Meeting, San Antonio, Texas, March 20, 1994; reprinted in Grace Davison Catalagram No. 72, 1985.
CHAPTER 3
FCC Catalysts The introduction of zeolite in commercial FCC catalysts in the early 1960s was one of the most significant advances in the history of cat cracking. Zeolite catalysts provided a greater profit with little capital investment. Simply stated, zeolite catalysts were and still are the biggest bargain of all time for the refiner. Improvements in catalyst technology have continued, enabling refiners to meet the demands of their market with minimum capital investment. Compared to amorphous silica-alumina catalysts, the zeolite catalysts are more active and more selective. The higher activity and selectivity translate to more profitable liquid product yields and additional cracking capacity. To take full advantage of the zeolite catalyst, refiners have revamped older units to crack more of the heavier, lowervalue feedstocks. A complete discussion of FCC catalysts would fill another book. This chapter provides enough information to select the proper catalyst and to troubleshoot the unit's operation. The key topics discussed are: • • • • • • •
Catalyst Components Catalyst Manufacturing Techniques Fresh Catalyst Properties Equilibrium Catalyst Analysis Catalyst Management Catalyst Evaluation Additives
CATALYST COMPONENTS FCC catalysts are in the form of fine powders with an average particle size in the range of 75 microns. A modern cat cracking catalyst has four major components: • Zeolite • Matrix 84
FCC Catalysts
85
* Binder • Filler
Zeolite Zeolite, or more properly, faujasite, is the key ingredient of the FCC catalyst. It provides product selectivity and much of the catalytic activity. The catalyst's performance largely depends on the nature and quality of the zeolite. Understanding the zeolite structure, types, cracking mechanism, and properties is essential in choosing the "right" catalyst to produce the desired yields. Zeolite Structure Zeolite is sometimes called molecular sieve. It has a well defined lattice structure. Its basic building blocks are silica and alumina tetrahedra (pyramids). Each tetrahedron (Figure 3-1) consists of a silicon or aluminum atom at the center of the tetrahedron, with oxygen atoms at the four corners. Zeolite lattices have a network of very small pores. The pore diameter of nearly all of today's FCC zeolite is approximately 8.0 angstroms (°A). These small openings, with an internal surface area of roughly 600 square
Figure 3-1.
Silicon/aluminum-oxygen tetrahedron [15].
86
Fluid Catalytic Cracking Handbook
meters per gram, do not readily admit hydrocarbon molecules that have a molecular diameter greater than 8.0°A to 10°A. The elementary building block of the zeolite crystal is a unit cell. The unit cell size (UCS) is the distance between the repeating cells in the zeolite structure. One unit cell in a typical fresh Y-zeolite lattice contains 192 framework atomic positions: 55 atoms of aluminum and 137 atoms of silicon. This corresponds to a silica (SitX,) to alumina (A12O3) molal ratio (SAR) of 5. The UCS is an important parameter in characterizing the zeolite structure. Zeolite Chemistry As stated above, a typical zeolite consists of silicon and aluminum atoms that are tetrahedrally joined by four oxygen atoms. Silicon is in a +4 oxidation state; therefore, a tetrahedron containing silicon is neutral in charge. In contrast, aluminum is in a +3 oxidation state. This indicates that each tetrahedron containing aluminum has a net charge of -1, which must be balanced by a positive ion. Solutions containing sodium hydroxide are commonly used in synthesizing the zeolite. The sodium serves as the positive ion to balance the negative charge of aluminum tetrahedron. This zeolite is called soda Y or NaY. The NaY zeolite is not hydrothermally stable because of the high sodium content. The ammonium ion is frequently used to displace sodium. Upon drying the zeolite, ammonia is vaporized. The resulting acid sites are both the Bronsted and Lewis types. The Bronsted acid sites can be further exchanged with rare earth material, such as cerium and lanthanum to enhance their strengths. The zeolite activity comes from these acid sites. Zeolite Types Zeolites employed in the manufacture of the FCC catalyst are synthetic versions of naturally occurring zeolites called faujasites. There are about 40 known natural zeolites and over 150 zeolites that have been synthesized. Of this number, only a few have found commercial applications. Table 3-1 shows properties of the major synthetic zeolites. The zeolites with applications to FCC are Type X, Type Y, and ZSM-5. Both X and Y zeolites have essentially the same crystalline structure. The X zeolite has a lower silica-alumina ratio than the Y zeolite. The X zeolite also has a lower thermal and hydrothermal
FCC Catalysts
8?
Table 3-1 Properties of Major Synthetic Zeolites
Zeoiite Type
Pore Size Dimensions (°A)
Silica-toAlumina Ratio
Zeolite A Faujasite ZSM-5
4.1 7.4 5.2 x 5.8
2-5 3-6 30-200
Mordenite
6.7 x 7.0
10-12
Applications Detergent manufacturing Catalytic cracking and hydrocracking Xylene isomerization, benzene alkylation, catalytic cracking, catalyst dewaxing, and methanol conversion. Hydro-isomerization, dewaxing
stability than the Y zeolite. Some of the earlier FCC zeolite catalysts contained X zeolite; however, virtually all of today's catalysts contain Y zeolite or variations thereof (Figure 3-2). ZSM-5 is a versatile zeolite that increases olefin yields and octane. Its application is further discussed later in this chapter. Until the late 1970s, the NaY zeolite was mostly ion exchanged with rare earth components. Rare earth components, such as lanthanum and
USY Zeolite (~ 7 Al Atoms/u.c.)
Equilibrium REY (-23 Al Atoms/u.c.)
Unit Cell Dimension =24.25 A (SiO2/AI2O3=54)
Unit Cell Dimension = 24.39 A (SiO2/AI2O3 « 15)
Figure 3-2. Geometry of USY and REY zeolites [14].
88
Fluid Catalytic Cracking Handbook
cerium, were used to replace sodium in the crystal. The rare earth elements, being trivalent, simply form "bridges" between two to three acid sites in the zeolite framework. Bridging protects acid sites from being ejected and stabilizes the zeolite structure. Rare earth exchange adds to the zeolite activity and thermal and hydrothermal stability. The reduction of lead in motor gasoline in 1986 created the need for a higher FCC gasoline octane. Catalyst manufacturers responded by adjusting the zeolite formulations, an alteration that involved expelling a number of aluminum atoms from the zeolite framework. The removal of aluminum increased SAR, reduced UCS, and in the process, lowered the sodium level of the zeolite. These changes increased the gasoline octane by raising its olefinicity. This aluminumdeficient zeolite was called ultrastable Y, or simply USY, because of its higher stability than the conventional Y. Zeolite Properties
The properties of the zeolite play a significant role in the overall performance of the catalyst. Understanding these properties increases our ability to predict catalyst response to changes in unit operation. From its inception in the catalyst plant, the zeolite must retain its catalytic properties under the hostile conditions of the FCC operation. The reactor/regenerator environment can cause significant changes in chemical and structural composition of the zeolite. In the regenerator, for instance, the zeolite is subjected to thermal and hydrothermal treatments. In the reactor, it is exposed to feedstock contaminants such as vanadium and sodium. Various analytical tests determine zeolite properties. These tests supply information about the strength, type, number, and distribution of acid sites. Additional tests can also provide information about surface area and pore size distribution. The three most common parameters governing zeolite behavior are as follows: • Unit Cell Size • Rare Earth Level « Sodium Content Unit Cell Size (UCS). The UCS is a measure of aluminum sites or the total potential acidity per unit cell. The negatively-charged aluminum atoms are sources of active sites in the zeolite. Silicon atoms do not
FCC Catalysts
89
possess any activity. The UCS is related to the number of aluminum atoms per cell (N Af ) by [1]: NA, + 111 x (UCS - 24.215) The number of silicon atoms (Nsi) is; Nsi = 192 - NA,
The SAR of the zeolite can be determined either from the above two equations or from a correlation such as the one shown in Figure 3-3. The UCS is also an indicator of zeolite acidity. Because the aluminum ion is larger than the silicon ion, as the UCS decreases, the acid sites become farther apart. The strength of the acid sites is determined by the extent of their isolation from the neighboring acid sites. The close proximity of these acid sites causes destabilization of the zeolite structure. Acid distribution of the zeolite is a fundamental factor affecting zeolite activity and selectivity. Additionally, the UCS measurement can be used to indicate octane potential of the zeolite. A lower UCS presents fewer active sites per unit cell. The fewer acid sites are farther apart and, therefore, inhibit hydrogen transfer reactions, which in turn increase gasoline octane as well as the production of C3 and lighter components (Figure 3-4). The octane increase is due to a higher concentration of olefins in the gasoline. Zeolites with lower UCS are initially less active than the conventional rare earth exchanged zeolites (Figure 3-5). However, the lower UCS zeolites tend to retain a greater fraction of their activity under severe thermal and hydrothermal treatments, hence the name ultrastable Y. A freshly manufactured zeolite has a relatively high UCS in the range of 24,50°A to 24.75°A. The thermal and hydrothermal environment of the regenerator extracts alumina from the zeolite structure and, therefore, reduces its UCS. The final UCS level depends on the rare earth and sodium level of the zeolite. The lower the sodium and rare earth content of the fresh zeolite, the lower UCS of the equilibrium catalyst (E-cat). Rare Earth Level. Rare earth (RE) elements serve as a "bridge" to stabilize aluminum atoms in the zeolite structure. They prevent the
90
Fluid Catalytic Cracking Handbook
Figure 3-3.
Silica-alumina ratio versus zeolite unit cell size,
aluminum atoms from separating from the zeolite lattice when the catalyst is exposed to high temperature steam in the regenerator. A fully rare-earth-exchanged zeolite equilibrates at a high UCS, whereas a non-rare-earth zeolite equilibrates at a very low UCS in the range of 24.25 [3]. All intermediate levels of rare-earth-exchanged zeolite can be produced. The rare earth increases zeolite activity and
FCC Catalysts
24.24
24.28
91
24.32
24.36
24.32
24.36
Unit Cell Size, A
6.0 5.5
>s t
1 5.0 o 4.5 «*>
4.0
24.20
24.24
24.28 Unit Cell Size, A
Figure 3-4.
Effects of unit cell size on octane and C3-gas make [4].
92
Fluid Catalytic Cracking Handbook
90 80
1520°F,20% steam in air.
0
10
20
30
40
50
60
70
80
90 100
Time, hrs Figure 3-5. Comparison of activity retention between rare-earth-exchanged zeolites versus USY zeolites. (Source: Grace Davison Octane Handbook.)
gasoline selectivity with a loss in octane (Figure 3-6). The octane loss is due to promotion of hydrogen transfer reactions. The insertion of rare earth maintains more and closer acid sites, which promotes hydrogen transfer reactions. In addition, rare earth improves thermal and hydrothermal stability of the zeolite. To improve the activity of a USY zeolite, the catalyst suppliers frequently add some rare earth to the zeolite. Sodium Content. The sodium on the catalyst originates either from zeolite during its manufacture or from the FCC feedstock. It is important for the fresh zeolite to contain very low amounts of sodium. Sodium decreases the hydrothermal stability of the zeolite. It also reacts with the zeolite acid sites to reduce catalyst activity. In the regenerator, sodium is mobile. Sodium ions tend to neutralize the strongest acid sites. In a dealuminated zeolite, where the UCS is low (24.22°A to 24.25°A), the sodium can have an adverse affect on the gasoline octane (Figure 3-7). The loss of octane is attributed to the drop in the number of strong acid sites. FCC catalyst vendors are now able to manufacture catalysts with a sodium content of less than 0.20 wt%. Sodium is commonly reported as
FCC Catalysts
5
§3
Yield of Gasoline, % ~ I
4 3 2
Gasoline octane (R+MV2
1 0 0
2
4
6
8
10
12
Rare Earth, wt% Figure 3-6. Effects of rare earth on gasoline octane and yield.
the weight percent of sodium or soda (Na2O) on the catalyst. The proper way to compare sodium is the weight fraction of sodium in the zeolite, This is because FCC catalysts have different zeolite concentrations. UCS, rare earth, and sodium are just three of the parameters that are readily available to characterize the zeolite properties. They provide valuable information about catalyst behavior in the cat cracker. If required, additional tests can be conducted to examine other zeolite properties.
Matrix The term matrix has different meanings to different people. For some, matrix refers to components of the catalyst other than the zeolite. For others, matrix is a component of the catalyst aside from the zeolite having catalytic activity. Yet for others, matrix refers to the catalyst binder. In this chapter, matrix means components of the catalyst other than zeolite and the term active matrix means the component of the catalyst other than zeolite heaving catalytic activity.
§4
Fluid Catalytic Cracking Handbook MOTOR OCTANE VS. SODIUM OXIDE 81.5
-
,
;
0.3
0.4
0.5
81.0
O 5 80.5
80.0 0.2
0.6
Na2O, wt% on catalyst
RESEARCH OCTANE VS. SODIUM OXIDE
94
93
92
91
0
1
2
3
4
5
Na2O, wt% on zeolite Figure 3-7. Effects of soda on motor and research octanes: motor octane vs. sodium oxide [11]; research octane vs. sodium oxide [4].
FCC Catalysts
95
Alumina is the source for an active matrix. Most active matrices used in FCC catalysts are amorphous. However, some of the catalyst suppliers incorporate a form of alumina that also has a crystalline structure. Active matrix contributes significantly to the overall performance of the FCC catalyst. The zeolite pores are not suitable for cracking of large hydrocarbon molecules generally having an end point > 900°F (482°C); they are too small to allow diffusion of the large molecules to the cracking sites. An effective matrix must have a porous structure to allow diffusion of hydrocarbons into and out of the catalyst, An active matrix provides the primary cracking sites. The acid sites located in the catalyst matrix are not as selective as the zeolite sites, but are able to crack larger molecules that are hindered from entering the small zeolite pores. The active matrix precracks heavy feed molecules for further cracking at the internal zeolite sites. The result is a synergistic interaction between matrix and zeolite, in which the activity attained by their combined effects can be greater than the sum of their individual effects [2]. An active matrix can also serve as a trap to catch some of the vanadium and basic nitrogen. The high boiling fraction of the FCC feed usually contains metals and basic nitrogen that poison the zeolite. One of the advantages of an active matrix is that it guards the zeolite from becoming deactivated prematurely by these impurities.
Filler and Binder The filler is a clay incorporated into the catalyst to dilute its activity. Kaoline [Al2(OH)2, Si2O5] is the most common clay used in the FCC catalyst. One FCC catalyst manufacturer uses kaoline clay as a skeleton to grow the zeolite in situ. The binder serves as a glue to hold the zeolite, matrix, and filler together. Binder may or may not have catalytic activity. The importance of the binder becomes more prominent with catalysts that contain high concentrations of zeolite. The functions of the filler and the binder are to provide physical integrity (density, attrition resistance, particle size distribution, etc.), a heat transfer medium, and a fluidizing medium in which the more important and expensive zeolite component is incorporated. In summary, zeolite will effect activity, selectivity, and product quality. An active matrix can improve bottoms cracking and resist
96
Fluid Catalytic Cracking Handbook
vanadium and nitrogen attacks. But a matrix containing very small pores can suppress strippablity of the spent catalyst and increase hydrogen yield in the presence of nickel. Clay and binder provide physical integrity and mechanical strength.
The manufacturing process of modern FCC catalyst is divided into two general groups—incorporation and "in-situ" processes. All catalyst suppliers manufacture catalyst by an incorporation process that requires making zeolite and matrix independently and using a binder to hold them together. In addition to the incorporation process, Engelhard also manufactures FCC catalyst using an "in-situ" process in which the zeolite component is grown within the pre-formed miscrospheres. The following sections provide a general description of zeolite synthesis. Conventional Zeolite (KEY, REHY, HY) NaY zeolite is produced by digesting a mixture of silica, alumina, and caustic for several hours at a prescribed temperature until crystallization occurs (Figure 3-8). Typical sources of silica and alumina are sodium silicate and sodium aluminate. Crystallization of Y-zeolite typically takes 10 hours at about 210°F (100°C). Production of a quality zeolite requires proper control of temperature, time, and pH of the crystallization solution. NaY zeolite is separated after filtering and water-washing of the crystalline solution. A typical NaY zeolite contains approximately 13 wt% Na2O. To enhance activity and thermal and hydrothermal stability of NaY, the sodium level must be reduced. This is normally done by the ion exchanging of NaY with a medium containing rare earth cations and/ or hydrogen ions. Ammonium sulfate solutions are frequently employed as a source for hydrogen ions. At this state of the catalyst synthesis there are two approaches for further treatment of NaY. Depending on the particular catalyst and the catalyst supplier, further treatment (rare earth exchanged) of NaY can be accomplished either before or after its incorporation into the matrix. Post-treatment of the NaY zeolite is simpler, but may reduce ion exchange efficiency.
H2O
Na^aoHte Crystallization 200 F, 12-24 Hr
o n
Filtrate to waste treatment
Figure 3-8, Typical manufacturing steps to produce FCC catalyst.
i8
Fluid Catalytic Cracking Handbook
USY Zeolite An ultrastable or a dealuminated zeolite (USY) is produced by replacing some of the aluminum ions in the framework with silicon. The conventional technique (Figure 3-9) includes the use of a high temperature (1,300-1,500°F [704-816°C]) steam calcination of HY zeolite, (13%Na O,A,c
NAY
NHf4
2
24.68 A)
- EXCHANGES
NHY
I
(3%Na£ 90)
STEAM CALCINE 114OO DEG. F
USY
JVH.+
(3%Na O,A0* 24,50A!
. EXCHANGES
LOW-SODA USY
(
Figure 3-9. Synthesis of USY zeolite (NAY),
FCC Catalysts
9§
Acid leaching, chemical extraction, and chemical substitution are all forms of dealumination that have become popular in recent years. The main advantage of these processes over conventional dealumination is the removal of the nonframework or occluded alumina from the zeolite cage structure. A high level of occluded alumina residing in the crystal is thought to have an undesirable impact on product selectivities by yielding more light gas and LPG; however, this has not been proven commercially. In the manufacturing of USY catalyst, the zeolite, clay, and binder are slurried together. If the binder is not active, an alumina component having catalytic properties may also be added. The well-mixed slurry solution is then fed to a spray dryer. The function of a spray dryer is to form microspheres by evaporating the slurry solution, through the use of atomizers, in the presence of hot air. The type of spray dryer and the drying conditions determine the size and distribution of catalyst particles. Engelhard Process Engelhard's "in-situ" FCC catalyst technology is mainly based on growing zeolite within the kaolin-based particles as shown in Figure 3-9A. The aqueous solution of various kaolins is spray dried to form microspheres. The microspheres are hardened in a high-temperature (1,300°F/704°C) calcination process. The NaY zeolite is produced by digestion of the microspheres, which contain metakaolin, and mullite with caustic or sodium silicate. Simultaneously, an active matrix is formed with the microspheres. The crystallized microspheres are filtered and washed prior to ion exchange and any final treatment.
With each shipment of fresh catalyst, the catalyst suppliers typically mail refiners an inspection report that contains data on the catalyst's physical and chemical properties. This data is valuable and should be monitored closely to ensure that the catalyst received meets the agreed specifications. A number of refiners independently analyze random samples of the fresh catalyst to confirm the reported properties. In addition, quarterly review of the fresh catalyst properties with the catalyst vendor will ensure that the control targets are being achieved.
100
Fluid Catalytic Cracking Handbook
The particle size distribution (PSD), sodium (Na), rare earth (RE), and surface area (SA) are some of the parameters in the inspection sheet that require close attention.
Particle Size Distribution (PSD) The PSD is an indicator of the fluidization properties of the catalyst, In general, fluidization improves as the fraction of the 0-40 micron particles is increased; however, a higher percentage of 0-40 micron particles will also result in greater catalyst losses. The fluidization characteristics of an FCC catalyst largely depend on the unit's mechanical configuration. The percentage of less than 40 microns in the circulating inventory is a function of cyclone efficiency. In units with good catalyst circulation, it may be economical to minimize the fraction of less than 40-micron particles. This is because after a few cycles, most of the 0-40 microns will escape the unit via the cyclones, The catalyst manufacturers control PSD of the fresh catalyst, mainly through the spray-drying cycle. In the spray dryer, the catalyst slurry must be effectively atomized to achieve proper distribution. As illustrated in Figure 3-10, the PSD does not have a normal distribution shape. The average particle size (APS) is not actually the average size of the catalyst particles, but rather the median value.
Surface Area (SA), M2/g The reported surface area is the combined surface area of zeolite and matrix. In zeolite manufacturing, the measurement of the zeolite surface area is one of the procedures used by catalyst suppliers to control quality. The surface area is commonly determined by the amount of nitrogen adsorbed by the catalyst. The surface area correlates fairly well with the fresh catalyst activity. Upon request, catalyst suppliers can also report the zeolite surface area. This data is useful in that it is proportional to the zeolite content of the catalyst.
Sodium (Na), wt% Its
Sodium plays an intrinsic part in the manufacturing of FCC catalysts. effects are well known and, because it deactivates the
FCC Catalysts
Figure 3-10.
101
Particle size distribution of a typical FCC catalyst.
zeolite and reduces the gasoline octane, every effort should be made to minimize the amount of sodium in the fresh catalyst. The catalyst inspection sheet expresses sodium or soda (Na2O) as the weight percent on the catalyst. When comparing different grades of catalysts, it is more practical to express the sodium content on the zeolite.
Rare earth (RE) is a generic name for 14 metallic elements of the Ianthanide series. These elements have similar chemical properties and are usually supplied as a mixture of oxides extracted from ores such as bastnaesite or monazite. Rare earth improves the catalyst activity (Figure 3-11) and hydrothermal stability. Catalysts can have a wide range of rare earth levels.
102
Fluid Catalytic Cracking Handbook
0
1
2
3
Rare Earth, wt% Figure 3-11. Effect of rare earth on catalyst activity.
depending on the refiner's objectives. Similar to sodium, the inspection sheet shows rare earth or rare earth oxide (RE2O3) as the weight percent of the catalyst. Again, when comparing different catalysts, the concentration of RE on the zeolite should be used. EQUILIBRIUM CATALYST ANALYSIS Refiners send E-cat samples to catalyst manufacturers on a regular basis. As a service to the refiners, the catalyst suppliers provide analyses of the samples in a form similar to the one shown in Figure 3-12. Although the absolute E-cat results may differ from one vendor to another, the results are most useful as a trend indicator. The tests performed on E-cat samples provide refiners with valuable information on unit conditions. The data can be used to pinpoint potential operational, mechanical, and catalyst problems because the physical and chemical properties of the E-cat provide clues on the environment to which it has been exposed. The following discussion describes each test briefly and examines the significance of these data to the refiner. The E-cat results are divided into catalytic properties, physical properties, and chemical analysis.
E-Cat Data Sample Date 11/7/94
% MAT 69
11/10/94 11/14/94 11/21/94 11/24/94 11/28/94 12/1/94 12/5/94 12/12/94
69 70 69 68 69 69 67 70
Na ppm 11/7/94 4900 11/10/94 4800 11/14/94 4600 11/21/94 4600 11/24/94 4800 11/28/94" 4600 12/1/94 4800 12/5/94 4600 12/12/94 4500
C.F.
G.F.
1.3 1.4 1.3 1.2 1.4 1.2
2.2 1.9 3.1 2.6 3.2 2.6 2.3 2.8 2.9
Fe ppm 5600 5600 5600 5600 5600 5600 5600 5600 5600
C wt% 0.23 0.23 0.16 0.23 0.22 0.20 0.24 0.15 0.24
1.3 1.2 1.2
S.A. P.V. ABD m2/gm, cc/gm gm/cc 147 0.30 0.83 148 0.83 0.28 0.84 147 0.29 148 0.83 0.29 0.28 148 0.83 0,29 150 0.84 148 0.28 0.85 148 0.85 0.29 0.28 148 0.84 V Ni ppm ppm 1997 4106 4093 1948 1940 4051 4099 1974 4017 1942 3962 1910 3892 1893 3893 L_J885 1873 3875
0-20 wt%
0 0 0 2 0 0 2 0 4
0-40 wt% 10 7 8 9 6 9 10 7 10
0-80 wt% 63 61 67 69 65 67 71 64 67
Cu Sb UCS RE203 ppm ppm 24.27 1.79 25 416 1,80 23 446 24 1.79 440 ' 24.27 24 r™44i 1.79 24.25 24 445 23 [_ 420 1.80 179 24 24.27 458 25 432 1.79 24 24.27 409 1.76
Figure 3-12. Typical E-cat analysis.
r TM
APS
70 72 69 68 70 69 67 71 69 Z
130 130 130 130 130 132 131 130 130]
AI2O3 ppm 28.9 29.1 29.2 28.7 28.7 28.7 28.7 28.8 28.8 M
17 18 17 18 18 18 18 18 18
Sn ppm 902 909 910 932 939 931
932
n
104
Fluid Catalytic Cracking Handbook
Catalytic Properties The activity, coke, and gas factors are the tests that reflect the relative catalytic behavior of the catalyst. Conversion (Activity) The first step in E-cat testing is to burn the carbon off the sample. The sample is then placed in a MAT unit (Figure 3-13), the heart of which is a fixed bed reactor. A certain amount of a standard gas oil feedstock is injected into the hot bed of catalyst. The activity is reported as the conversion to 430°F (221°C) material. The feedstock's quality, reactor temperature, catalyst-to-oil ratio, and space velocity are four variables affecting MAT results. Each catalyst vendor uses slightly different operating variables to conduct microactivity testing, as indicated in Table 3-2. In commercial operations, catalyst activity is affected by operating conditions, feedstock quality, and catalyst characteristics. The MAT separates catalyst effects from feed and process changes. Feed contaminants, such as vanadium and sodium, reduce catalyst activity. E-cat activity is also affected by fresh catalyst makeup rate and regenerator conditions. Coke Factor (CF), Gas Factor (GF) The CF and GF represent the coke- and gas-forming tendencies of an E-cat compared to a standard steam-aged catalyst sample at the same conversion. The CF and GF are influenced by the type of fresh catalyst and the level of metals deposited on the E-cat. Both the coke and gas factors can be indicative of the dehydrogenation activity of the metals on the catalyst. The addition of amorphous alumina to the catalyst will tend to increase the nonselective cracking, which forms coke and gas.
Physical Properties The tests that reflect physical properties of the catalyst are surface area, average bulk density, pore volume, and particle size distribution. Surface Area (SA), M2/g For an identical fresh catalyst, the surface area of an E-cat is an indirect measurement of its activity. The SA is the sum of zeolite and
FCC Catalysts Star dard FCCl Feed n-Q Equilibrium Catalysts
!
L
105
Syringe Pump
J -"1^3 way Valve
@
Coke Burn Off -
"I I Reactor Furnace
control
Temp
Tempcontrol
ri i— is l~W
Preheat Zone Catalyst y
rJFlow Meter L^«—}50%
This method can also be used to calculate the catalyst retention factor. The above equations assume steady-state operation, constant unit inventory, and constant addition and loss rate.
FCC Catalysts
115
Catalyst management is a very important aspect of the FCC process. Selection and management of the catalyst, as well as how the unit is operated, are largely responsible for achieving the desired product. Proper choice of a catalyst will go a long way toward achieving a successful cat cracker operation. Catalyst change-out is a relatively simple process and allows a refiner to select the catalyst that maximizes the profit margin. Although catalyst change-out is physically simple, it requires a lot of homework as discussed later in this section. As many catalyst formulations are available, catalyst evaluation should be an ongoing process. However, it is not an easy task to evaluate the performance of an FCC catalyst in a commercial unit because of continual changes in feedstocks and operating conditions, in addition to inaccuracies in measurements. Because of these limitations, refiners sometimes switch catalyst without identifying the objectives and limitations of their cat crackers. To ensure that a proper catalyst is selected, each refiner should establish a methodology that allows identification of "real" objectives and constraints and ensures that the choice of the catalyst is based on well-thoughtout technical and business merits. In today's market, there are over 120 different formulations of FCC catalysts. Refiners should evaluate catalyst mainly to maximize profit opportunity and to minimize risk. The "right" catalyst for one refiner may not necessarily be "right" for another, A comprehensive catalyst selection methodology will have the following elements: 1. Optimize unit operation with current catalyst and vendor a. Conduct test run b. Incorporate the test run results into an FCC kinetic model c. Identify opportunities for operational improvements d. Identify unit's constraints e. Optimize incumbent catalyst with vendor 2. Issue technical inquiry to catalyst vendors a. Provide test run results b. Provide E-cat sample c. Provide processing objectives d. Provide unit limitations
116
Fluid Catalytic Cracking Handbook
3, Obtain vendor responses a. Obtain catalyst recommendation b. Obtain alternate recommendation c. Obtain comparative yield projection 4, Obtain current product price projections a. For present and future four-quarters 5, Perform economic evaluations on vendor yields a. Select catalysts for MAT evaluations 6, Conduct MAT of selected list a. Perform physical and chemical analyses b. Determine steam deactivation conditions c. Deactivate incumbent fresh catalyst to match incumbent E-cat d. Use same deactivation steps for each candidate catalyst 7, Perform economic analysis of alternatives a. Estimate commercial yield from MAT evaluations 8, Request commercial proposals a. Consult at least two vendors b. Obtain references c. Check references 9, Test the selected catalysts in a pilot plant a. Calibrate the pilot plant steaming conditions using incumbent E-cat b. Deactivate the incumbent and other candidate catalysts c. Collect at least two or three data points on each by varying the catalyst-to-oil ratio 10. Evaluate pilot plant results a. Translate the pilot plant data b. Use the kinetic model to heat-balance the data c. Identify limitations and constraints 11. Make the catalyst selection a. Perform economic evaluation b. Consider intangibles-research, quality control, price, steady supply, manufacturing location c. Make recommendations 12. Post selection a. Monitoring transition-% changeover b. Post transition test run c. Confirm computer model
FCC Catalysts
117
13. Issue the final report a. Analyze benefits b. Evaluate selection methodology There is a redundancy of flexibility in the design of FCC catalysts. Variation in the amount and type of zeolite, as well as the type of active matrix, provide a great deal of catalyst options that the refiner can employ to fit its needs. For smaller refiners, it may not be practical to employ pilot plant facilities to evaluate different catalysts. In this case, the above methodology can still be used with emphasis shifted toward using the MAT data to compare the candidate catalysts. It is important that MAT data are properly corrected for temperature, "soaking time," and catalyst strippability effects.
For many years, cat cracker operators have used additive compounds for enhancing cat cracker performance. The main benefits of these additives (catalyst and feed additives) are to alter the FCC yields and reduce the amount of pollutants emitted from the regenerator. The additives discussed in this section are CO promoter, SOX reduction, ZSM-5, and antimony.
CO Promoter The CO promoter is added to most FCC units to assist in the combustion of CO to CO2 in the regenerator. The promoter is added to accelerate the CO combustion in the dense phase and to minimize the higher temperature excursions that occur as a result of afterburning in the dilute phase. The promoter allows uniform burning of coke, particularly if there is uneven distribution between spent catalyst and combustion air. Regenerators operating in full or partial combustion can utilize the benefits of the CO promoter. The addition of the promoter tends to increase the regenerator temperature and NOx emission. The metallurgy of the regenerator internals should be checked for tolerance of the higher temperature. The active ingredients of the promoter are typically the platinum group metals. The platinum, in the concentration of 300 ppm to 800
118
Fluid Catalytic Cracking Handbook
ppm, is typically dispersed on a support. The effectiveness of the promoter largely depends on its activity and stability. Promoter is frequently added to the regenerator two to three times a day, normally at a rate of 3 to 5 pounds (1 to 2.3 kg) promoter per ton of fresh catalyst. The concentration of platinum required in a unit inventory is about 0.5 to 1.5 ppm. The promoter addition rate may be increased if antimony solution is being used to passivate the nickel. The use of CO promoter, particularly during unit start-ups, improves the stability of the regeneration operation. However, not every cat cracker can justify combustion-promoted operation. Heat balance, availability of combustion air, NOX emission metallurgical limits, and the presence of CO boiler are some of the factors that should be considered before using combustion promoter. For example, in units operating with low oxygen levels and partial combustion, a promoted system could increase carbon on regenerated catalyst (CRC). This is because CO combustion reaction competes with carbon burning reaction for the available oxygen. The combustion of CO to CO2 will also increase NOX emissions. This is largely due to the oxidation of intermediates such as ammonia and cyanide gases into nitrogen oxide (NO).
SOX Additive The coke on the spent catalyst entering the regenerator contains sulfur. In the regenerator, the sulfur in the coke is converted to SO2 and SO3. The mixture of SO2 and SO3 is commonly referred to as SOX, and approximately 80% to 90% of SOX is SO2, with the rest being SO3, The SOX leaves the regenerator with the flue gas and is eventually discharged to the atmosphere. Coke yield, thiophenic sulfur content of the feed, the regenerator operating condition, and the type of FCC catalyst are the major factors affecting SOX emissions. The environmental impact of SOX emissions has gained much attention over the past ten years. The United States Environmental Protection Agency (EPA) New Source Performance Standards (NSPS) went into effect in 1989. The ruling covers new, modified, and reconstructed FCC units since January 1994. It should be noted that the Southern California Air Quality Management District (SCAQMD) board has established a limit of 60 kilograms of SOX per 1,000 barrels of feed for the existing FCC units.
FCC Catalysts
119
There are three common methods for SOX abatement. These are flue gas scrubbing, feedstock desulfurization, and SOX additive. The SOX additive is often the least costly alternative, which is the approach practiced by many refiners. The SOX additive, usually a metal oxide, is added directly to the catalyst inventory. The additive works by adsorbing and chemically bonding with SO3 in the regenerator. This stable sulfate species is carried with the circulating catalyst to the riser, where it is reduced or "regenerated" by hydrogen or water to yield H2S and metal oxide. Table 3-3 shows the postulated chemistry of SOX reduction by a SOX agent. To achieve the highest efficiency of SOX additive, it is important that: * Excess oxygen be available; oxygen promotes the SO2 to SO3 reaction. SOX additive will only form a metal sulfate from SOV * The regenerator temperature be lower; lower temperature favors SO2 + 1/2 O2 -> SO3 * The capturing agent be physically compatible with the FCC catalyst and easily regenerated in the riser and stripper. * CO promoter be used, which oxidizes SO2 to SO3. * There be a uniform distribution of air and spent catalyst. Air/ catalyst mixing in the regenerator can significantly affect the SOX pick-up efficiency.
Table 3-3 Mechanism of Catalytic SOX Reduction
A. In the Regenerator Sulfur in Coke (S) + O2 SO2 + l/2 O2 MXO + SO,
—» —> —>
MXSO4
—> -> —>
MXS + 4H2O MXO + H2S + 3 H2O MXO + H2O
SO2 + SO3 SO3
B. In the Reactor and Stripper MXSO4 + 4H2
Mxso4 + 4H2 MXS + H2O Source: Thiel [9]
120
Fluid Catalytic Cracking Handbook
* Operation of the reactor stripper be efficient. The stripper efficiency is very important to allow the release of sulfate and the formation of H2S. Since most of the regenerators operating in full combustion mode usually operate with 1% to 3% excess oxygen, the capturing efficiency of SOX additive is often greater in full combustion than in partial combustion units.
ZSM-5 ZSM-5 is Mobil Oil's proprietary shape-selective zeolite that has a different pore structure from that of Y-zeolite. The pore size of ZSM5 is smaller than that of Y-zeolite (5.1°A to 5.6°A versus 8°A to 9°A), In addition, the pore arrangement of ZSM-5 is different from Y-zeolite, as shown in Figure 3-16. The shape selectivity of ZSM-5 allows
Figure 3-16.
Comparison of Y faujasite and ZSM-5 zeolites [13].
FCC Catalysts
121
preferential cracking of long-chain, low-octane normal paraffins, as well as some olefins, in the gasoline fraction. ZSM-5 additive is added to the unit to boost gasoline octane and to increase light olefin yields. ZSM-5 accomplishes this by upgrading low-octane components in the gasoline boiling range (C7 to C1O) into light olefins (C3, C4, C5), as well as isomerizing low-octane linear olefins to high-octane branched olefins, ZSM-5 inhibits paraffin hydrogenation by cracking the C7+ olefins. ZSM-5's effectiveness depends on several variables. The cat crackers that process highly paraffinic feedstock and have lower base octane will receive the greatest benefits of using ZSM-5. ZSM-5 will have little effect on improving gasoline octane in units that process naphthenic feedstock or operate at a high conversion level. When using ZSM-5, there is almost an even trade-off between FCC gasoline volume and LPG yield. For a one-number increase in the research octane of FCC gasoline, there is a 1 vol% to 1.5 vol% decrease in the gasoline and almost a corresponding increase in the LPG, This again depends on feed quality, operating parameters, and base octane. The decision to add ZSM-5 depends on the objectives and constraints of the unit. ZSM-5 application will increase load on the wet gas compressor, FCC gas plant, and other downstream units. Most refiners who add ZSM-5 do it on a seasonal basis, again depending on their octane need and unit limitations. The concentration of the ZSM-5 additive should be greater than 1 % of the catalyst inventory to see a noticeable increase in the octane. An octane boost of one research octane number (RON) will typically require a 2% to 5% ZSM-5 additive in the inventory. It should be noted that the proper way of quoting percentage should be by ZSM-5 concentration rather than the total additive because the activity and attrition rate can vary from one supplier to another. There are new generations of ZSM-5 additives that have nearly twice the activity of the earlier additives. In summary, ZSM-5 provides the refiner the flexibility to increase gasoline octane and light olefins. With the introduction of reformulated gasoline, ZSM-5 could play an important role in producing isobutylene, used as the feedstock for production of methyl tertiary butyl ether (MTBE).
122
Fluid Catalytic Cracking Handbook
Metal Passivation As discussed in Chapter 2, nickel, vanadium, and sodium are the metal compounds usually present in the FCC feedstock. These metals deposit on the catalyst, thus poisoning the catalyst active sites. Some of the options available to refiners for reducing the effect of metals on catalyst activity are as follows: • • • • • •
Increasing the fresh catalyst makeup rate Using outside E-cat Employing metal passivators Incorporating metal trap into the FCC catalyst Using demetalizing technology to remove the metals from the catalyst The MagnaCat separation process (demetalizing technology), which allows discarding the "older" catalyst particles containing higher metal levels
Metal passivation in general, and antimony in particular, are discussed in the following section. In recent years, several methods have been patented for chemical passivation of nickel and vanadium. Only some of the tin compounds have had limited commercial success in passivating vanadium. Although tin has been used by some refiners, it has not been proven or as widely accepted as antimony. In the case of nickel, antimony-based compounds have been most effective in reducing the detrimental effects of nickel poisoning. It should be noted that, although the existing antimony-based technology is the most effective method of reducing the deleterious effects of nickel, the antimony is fugitive and can be considered hazardous. In this case, a bismuth-based passivator may be a better choice. Antimony Antimony-based passivation was introduced by Phillips Petroleum in 1976 to passivate nickel compounds in the FCC feed. Antimony is injected into the fresh feed, usually with the help of a carrier such as light cycle oil. If there are feed preheaters in the unit, antimony should be injected downstream of the preheater to avoid thermal decomposition of the antimony solution in the heater tubes. The effects of antimony passivation are usually immediate. By forming an alloy with nickel, the dehydrogenation reactions that are
FCC Catalysts
123
caused by nickel are often reduced by 40% to 60%. This is evidenced by a sharp decline in dry gas and hydrogen yield. Nickel passivation is generally economically attractive when the nickel content of the E-cat is greater than 1,000 ppm. The Phillips Petroleum secondary antimony patent position is due to expire in late 1999, At that time, antimony passivation can become economically attractive at a lower nickel level than 1,000 ppm. The antimony solution should be added in proportion to the amount of nickel present in the feed. The optimum dosage normally corresponds to an antimony-to-nickel ratio of 0.3 to 0.5 on the E-cat. Antimony's retention efficiency on the catalyst is in the range of 75% to 85% without the recycling of slurry oil to the riser. If slurry recycle is being practiced, the retention efficiency is usually greater than 90%. Any antimony not deposited on the circulating catalyst ends up in the decanted oil and the catalyst fines from the regenerator. It is often a good practice to discontinue antimony injection about one month prior to a scheduled unit shutdown to make sure the exposure to catalyst dust containing antimony is reduced to a minimum when wearing a half-faced respirator. SUMMARY The introduction of zeolite into the FCC catalyst in the early 1960s was one of the most significant developments in the field of cat cracking. The zeolite greatly improved selectivity of the catalyst, resulting in higher gasoline yields and indirectly allowing refiners to process more feed to the unit. With the introduction of reformulated gasoline, new formulations in FCC catalyst will again help refiners meet new requirements in gasoline quality. Since there are over 120 different FCC catalyst formulations in the market today, it is important that the refinery personnel involved in cat cracker operations have some fundamental understanding of catalyst technology. This knowledge is useful in areas such as proper troubleshooting and customizing a catalyst that would match the refiner's needs. The additive technology will be expanding in coming years. The need to produce reformulated gasoline will increase demand for the shape-selective zeolite, such as ZSM-5. The pressure from environmental agencies to reduce SOX and NOX will further increase the demand for additives that reduce emissions.
124
Fluid Catalytic Cracking Handbook
REFERENCES 1. Breck, D. W., Zeolite Molecular Sieves: Structure, Chemistry, and Use, New York: Wiley Interscience, 1974. 2. Hayward, C. M. and Winkler, W. S., "FCC: Matrix/Zeolite," Hydrocarbon Processing, February 1990, pp. 55-56. 3. Upson, L. L., "What FCC Catalyst Tests Show," Hydrocarbon Processing November 1981, pp. 253-258. 4. Pine, L. A., Maher, P. J., and Wachter, W. A., "Prediction of Cracking Catalyst Behavior by a Zeolite Unit Cell Size Model," Journal of Catalysis, No. 85, 1984, pp. 466-476. 5. Magnusson, J. and Pudas, R., "Activity and Product Distribution Characteristics of the Currently Used FCC Catalyst Systems," presented at Katalistiks' 6th Annual FCC Symposium, Munich, Germany, May 2223, 1985. 6. John G. S. and Mikovsky, R. J., "Calculation of the Average Activity of Cracking Catalysts," Chemical Engineering Science, Vol. 15, 1961, pp. 172-175. 7. Gaughan, J. R., "Effect of Catalyst Retention on Inventory Replacement," Oil & Gas Journal, December 26, 1983, pp. 141-145. 8. Tamborski, G. A., Magnabosco, L. M., Powell, J. W., and Yoo, J. S., "Catalyst Technology Improvements Make SOX Emissions Control Affordable," presented at Katalistiks' 6th Annual FCC Symposium, Munich, Germany, May 22-23, 1985. 9. Thiel, P. G., Blazek, J. J., "Additive R," Grace Davison Catalagram, No. 71, 1985. 10. Engelhard Corporation, "Reduced Unit Cell Size Catalysts Offer Improved Octane for FCC Gasoline," The Catalyst Report, TI-762. 11. Engelhard Corporation, "Increasing Motor Octane by Catalytic Means Part 2," The Catalyst Report, EC6100P. 12. Engelhard Corporation, "The Chemistry of FCC Coke Formation," The Catalyst Report, Vol. 7, Issue 2. 13. Majon, R. J. and Spielman, J., "Increasing Gasoline Octane and Light Olefin Yields with ZSM-5," The Catalyst Report, Vol. 5, Issue 5, 1990. 14. Davison Div., W.R. Grace & Co., Grace Davison Catalagram, No. 72, 1985. 15. Humphries, Adrian P., "Zeolite Fundamentals and Synthesis," Akzo Chemicals, 1987. 16. Davison Octane Handbook. 17. G. Yaluris and A. W. Peters, "Studying the Chemistry of the FCCU Regenerator in the Laboratory Under Realistic Conditions," Grace Davison, Columbia, MD, 1998.
CHAPTER 4
Chemistry of FCC Reactions A complex series of reactions (Table 4-1) take place when a large gas-oil molecule comes in contact with a 1,200°F to 1,400°F (650°C to 760°C) FCC catalyst. The distribution of products depends on many factors, including the nature and strength of the catalyst acid sites. Although most cracking in the FCC is catalytic, thermal cracking reactions also occur. Thermal cracking is caused by factors such as non-ideal mixing in the riser and poor separation of cracked products in the reactor. The purpose of this chapter is to: • Provide a general discussion of the chemistry of cracking (both thermal and catalytic). • Highlight the role of the catalyst, and in particular, the influence of zeolites. • Explain how cracking reactions affect the unit's heat balance. Whether thermal or catalytic, cracking of a hydrocarbon means the breaking of a carbon to carbon bond. But catalytic and thermal cracking proceed via different routes. A clear understanding of the different mechanisms involved is beneficial in areas such as: • Selecting the "right" catalyst for a given operation • Troubleshooting unit operation • Developing a new catalyst formulation Topics discussed in this chapter are: • Thermal cracking • Catalytic cracking • Thermodynamic aspects
125
126
Fluid Catalytic Cracking Handbook Table 4-1 Important Reactions Occurring in FCC
1. Cracking: Paraffins cracked to olefins and smaller paraffins Olefins cracked to smaller olefins
C9Hl8 -> C4H8 + C5H10
Aromatic side-chain scission
ArC10H21
Naphthenes (cyclo-paraffins) cracked to olefins and smaller ring compounds
Cyclo-C1oH20 -> C6H12
CSH1
ArC5H9
C4H8
2. Isomerization: Olefin bond shift
1-C4H8 -^ trans-2-C4H8
Normal olefins to iso-olefin
n-C5H10 —> iso-C5H10
Normal paraffins to iso-paraffin
n-C4H10
Cyclo-hexane to cyclo-pentane
C6H12 + C5H9CH3
3. Hydrogen Transfer:
iso-C4Hlo
Naphthene + Olefin -» Aromatic + Paraffin
Cyclo-aromatization 4. Trans-alkylation/Alkyl-group Transfer
2C6H5CH3
C6H6
5. Cyclization of Olefins to Naphthenes
C7H14 -^ CH3-cyclo-C6H ii
6. Dehydrogenation
n-C8H18
7. Dealkylation
H6 lso-C3H 7-C6H5 -» C6H6 -\. C C3 H
8. Condensation
Ar-C3H == CH2 + R,CH = CHR2 v AT Ar + 2H
~^
C
8H16
+ H
2 3
7
6
jT\I
THERMAL CRACKING Before the advent of the catalytic cracking process, thermal cracking was the primary process available to convert low-value feedstocks into lighter products. Refiners still use thermal processes, such as delayed coking and visibreaking, for cracking of residual hydrocarbons.
Chemistry of FCC Reactions
127
Thermal cracking is a function of temperature and time. The reaction occurs when hydrocarbons in the absence of a catalyst are exposed to high temperatures in the range of 800°F to 1,200°F (425°C to 650°C). The initial step in the chemistry of thermal cracking is the formation of free radicals. They are formed upon splitting the C-C bond. A free radical is an uncharged molecule with an unpaired electron. The rupturing produces two uncharged species that share a pair of electrons. Equation 4-1 shows formation of a free radical when a paraffin molecule is thermally cracked.
R2 V "D IV,,
H
f*
7 IV
| H
I
*/"*
V-
T
t>
V- '
.IV
' H
(4-0
H
Free radicals are extremely reactive and short-lived. They can undergo alpha scission, beta scission, and polymerization. (Alphascission is a break one carbon away from the free radical; betascission, two carbons away.) Beta-scission produces an olefin (ethylene) and a primary free radical (Equation 4-2), which has two fewer carbon atoms [1]: J\
""""' V.'..il.^ —~ VvlT'} "•"*— V-, —
Lisy ""' "
°~~~
Vx
S~M.-y T IT'iV--- ~™ V_-ilo
\iT"1"jiii. )
The newly formed primary free radical can further undergo betascission to yield more ethylene. Alpha-scission is not favored thermodynamicaily but does occur. Alpha-scission produces a methyl radical, which can extract a hydrogen atom from a neutral hydrocarbon molecule. The hydrogen extraction produces methane and a secondary or tertiary free radical (Equation 4-3).
-» CH4 + R-CH2-CH2-CH2-CH2-'CH-CH2-CH3
(4-3)
This radical can undergo beta-scission. The products will be an alpha-olefin and a primary free radical (Equation 4-4).
128
Fluid Catalytic Cracking Handbook
R-CH2-CH2-CH2-CH2-'CH-CH2-CH3 -» R-CH2-CH2-'CH2 + H2C=CH-CH2-CH3
(4-4)
Similar to the methyl radical, the R-*CH2 radical can also extract a hydrogen atom from another paraffin to form a secondary free radical and a smaller paraffin (Equation 4-5). R,-'CH 2 + R-CH2-CH2-CH2-CH2-CH2-CH2-CH3 -> R,-CH 3 + R-CH2-CH2-CH2-CH2-CH2-*CH-CH3
R-*CH? is more stable than H3*C. Consequently, the hydrogen extraction rate of R-*CH2 is lower than that of the methyl radical. This sequence of reactions forms a product rich in C} and G,, and a fair amount of alpha-olefins. Free radicals undergo little branching (isomerization). One of the drawbacks of thermal cracking in an FCC is that a high percentage of the olefins formed during intermediate reactions polymerize and condense directly to coke. The product distribution from thermal cracking is different from catalytic cracking, as shown in Table 4-2. The shift in product distribution confirms the fact that these two processes proceed via different mechanisms,
CATALYTIC CRACKING Catalytic reactions can be classified into two broad categories: * Primary cracking of the gas oil molecules • Secondary rearrangement and re-cracking of cracked products Before discussing mechanisms of the reactions, it is appropriate to review FCC catalyst development and examine its cracking properties. An in-depth discussion of FCC catalyst was presented in Chapter 3.
FCC Catalyst Development The first commercial fluidized cracking catalyst was acid-treated natural clay. Later, synthetic silica-alumina materials containing 10 to
Chemistry of FCC Reactions
129
Table 4-2 Comparison of Products of Thermal and Catalytic Cracking Hydrocarbon Type
Thermal Cracking
Catalytic Cracking
n-Paraffms
C2 is major product, with C3 to C6 is major product; much C3 and C3, and C4 to few n-olefins above C4; C16 olefins; little branching much branching
Olefins
Slow double-bond shifts and little skeletal isomerization; H-transfer is minor and nonselective for tertiary olefins; only small amounts of aromatics formed from aliphatics at 932°F (500°C)
Rapid double-bond shifts, extensive skeletal isomerization, H-transfer is major and selective for tertiary olefins; large amounts of aromatics formed from aliphatics at 932°F (500°O
Naphthenes
Crack at slower rate than paraffins
If structural groups are equivalent, crack at about the same rate as paraffins
Alkyl-aromatics
Cracked within side chain
Crack next to ring
Source: Venuto [2]
15 percent alumina replaced the natural clay catalysts. The synthetic silica-alumina catalysts were more stable and yielded superior products. In the mid-1950s, alumina-silica catalysts, containing 25 percent alumina, came into use because of their higher stability. These synthetic catalysts were amorphous; their structure consisted of a random array of silica and alumina, tetrahedrally connected. Some minor improvements in yields and selectivity were achieved by switching to catalysts such as magnesia-silica and alumina-zirconia-silica. Impact of Zeolites The breakthrough in FCC catalyst was the use of X and Y zeolites during the early 1960s. The addition of these zeolites substantially increased catalyst activity and selectivity. Product distribution with a zeolite-containing catalyst is different from the distribution with an amorphous silica-alumina catalyst (Table 4-3). In addition, zeolites are 1,000 times more active than the amorphous silica alumina catalysts.
130
Fluid Catalytic Cracking Handbook Table 4-3 Comparison of Yield Structure for Fluid Catalytic Cracking of Waxy Gas Oil over Commercial Equilibrium Zeolite and Amorphous Catalysts
Yields, at 80 vol% Conversion Hydrogen, wt% C1's + C2's, wt%
Amorphous, High Alumina
Zeolite, XZ-25
Change from Amorphous
0.08 3.8
0.04 2.1
-0.04 -1.7
Propylene, vol% Propane, vol% Total C3's
16.1 1.5 17.6
11.8 1.3 13.1
-4.3 -0.02 -4.5
Butenes, vol% i-Butane, vol% n-Butane, vol% Total C4's
12.2 7.9 0,7 20.8
7.8 7.2 0.4 15.4
-4.4 -0.7 -0.3 -5.4
C5-390 at 90% ASTM gasoline, vol%
55.5
62.0
+6.5
Light Fuel Oil, vol% Heavy Fuel Oil, vol% Coke, wt%
4.2 15.8 5.6
6.1 13.9 4.1
+1.9 -1.9 -1.5
Gasoline Octane No.
94
89.8
-4.2
Source: Venuto [2]
The higher activity comes from greater strength and organization of the active sites in the zeolites. Zeolites are crystalline alumina-silicates having a regular pore structure. Their basic building blocks are silica and alumina tetrahedra. Each tetrahedron consists of silicon or aluminum atoms at the center of the tetrahedron with oxygen atoms at the corners. Because silicon and aluminum are in a +4 and 4-3 oxidation state, respectively, a net charge of -1 must be balanced by a cation to maintain electrical neutrality. The cations that replace the sodium ions determine the catalyst's activity and selectivity. Zeolites are synthesized in an alkaline environment such as sodium hydroxide, producing a soda-Y zeolite. These soda-Y zeolites have little stability but the sodium can be easily
Chemistry of FCC Reactions
131
exchanged. Ion exchanging sodium with cations, such as hydrogen or rare earth ions, enhances acidity and stability. The most widely used rare earth compounds are lanthanum (La3*) and cerium (Ce3+). The catalyst acid sites are both Bronsted and Lewis type. The catalyst can have either strong or weak Bronsted sites; or, strong or weak Lewis sites. A Bronsted-type acid is a substance capable of donating a proton. Hydrochloric and sulfuric acids are typical Bronsted acids. A Lewis-type acid is a substance that accepts a pair of electrons. Lewis acids may not have hydrogen in them but they are still acids. Aluminum chloride is the classic example of a Lewis acid. Dissolved in water, it will react with hydroxyl, causing a drop in solution pH. Catalyst acid properties depend on several parameters, including method of preparation, dehydration temperature, silica-to-alumina ratio, and the ratio of Bronsted to Lewis acid sites,
Mechanism of Catalytic Cracking Reactions When feed contacts the regenerated catalyst, the feed vaporizes. Then positive-charged atoms called carbocations are formed. Carbocation is a generic term for a positive-charged carbon ion. Carbocations can be either carbonium or carbenium ions. A carbonium ion, CH5+, is formed by adding a hydrogen ion (H+) to a paraffin molecule (Equation 4-6), This is accomplished via direct attack of a proton from the catalyst Bronsted site. The resulting molecule will have a positive charge with 5 bonds to it. R — CH2 — CH2 — CH2 — CH3 + H+ (proton attack) -» R — C+H — CH2 — CH2 — CH3 + H2
(4-6)
The carbonium ion's charge is not stable and the acid sites on the catalyst are not strong enough to form many carbonium ions. Nearly all the cat cracking chemistry is carbenium ion chemistry. A carbenium ion, R-CH2+, comes either from adding a positive charge to an olefin or from removing a hydrogen and two electrons from a paraffin (Equations 4-7 and 4-8). R — CH. = CH — CH2 — CH2 — CH3 + H* (a proton @ Bronsted site) —>_^ jp^ ~™_™. ^^ |"j
« v-'Jcin ——• v^JrJ'-j"""
v^in.'-) —"""— v^-ii-t
\£|.— / j
132
Fluid Catalytic Cracking Handbook
R — CH2 -— CH2 — CH2 — CH3 (removal of H~ @ Lewis site) _» R _ c+H — CH2 — CH2 — CH3
(4-8}
Both the Bronsted and Lewis acid sites on the catalyst generate carbenium ions. The Bronsted site donates a proton to an olefin molecule and the Lewis site removes electrons from a paraffin molecule. In commercial units, olefins come in with the feed or are produced through thermal cracking reactions. The stability of carbocations depends on the nature of alkyl groups attached to the positive charge. The relative stability of carbenium ions is as follows [2] with tertiary ions being the most stable: Tertiary R
.
C ""~ V.-C+
'"•"""" V--
> P
V_^
Secondary P
\*s
P+
V~"
P
V-"'
>
Primary R
JLX.
P
V-'
> Ethyl > Methyl P+
V_--
P
V--
P+
*—•
P* V,,'
c One of the benefits of catalytic cracking is that the primary and secondary ions tend to rearrange to form a tertiary ion (a carbon with three other carbon bonds attached). As will be discussed later, the increased stability of tertiary ions accounts for the high degree of branching associated with cat cracking. Once formed, carbenium ions can form a number of different reactions. The nature and strength of the catalyst acid sites influence the extent to which each of these reactions occur. The three dominant reactions of carbenium ions are: * The cracking of a carbon-carbon bond * Isomerization * Hydrogen transfer Cracking Reactions Cracking, or beta-scission, is a key feature of ionic cracking. Betascission is the splitting of the C-C bond two carbons away from the positive-charge carbon atom. Beta-scission is preferred because the energy required to break this bond is lower than that needed to break the adjacent C-C bond, the alpha bond. In addition, short-chain hydrocarbons are less reactive than long-chain hydrocarbons. The rate of
Chemistry of FCC Reactions
133
the cracking reactions decreases with decreasing chain length. With short chains, it is not possible to form stable carbenium ions. The initial products of beta-scission are an olefin and a new carbenium ion (Equation 4-9). The newly-formed carbenium ion will then continue a series of chain reactions. Small ions (four-carbon or five-carbon) can transfer the positive charge to a big molecule, and the big molecule can crack. Cracking does not eliminate the positive charge; it stays until two ions collide. The smaller ions are more stable and will not crack, They survive until they transfer their charge to a big molecule, R _ " V,' ri+"H 11 IV. "
CH V*' 1. !••->
PH V--' 1 !••) — PH V--J. .I')
"
CH \.~^ I to
-* CH3 — CH = CH2 + C+H2 — CH2 — CH2R
(4-9)
Because beta-scission is mono-molecular and cracking is endothermic, the cracking rate is favored by high temperatures and is not equilibrium-limited. Isomerization Reactions Isomerization reactions occur frequently in catalytic cracking, and infrequently in thermal cracking. In both, breaking of a bond is via beta-scission. However, in catalytic cracking, carbocations tend to rearrange to form tertiary ions. Tertiary ions are more stable than secondary and primary ions; they shift around and crack to produce branched molecules (Equation 4-10). (In thermal cracking, free radicals yield normal or straight chain compounds.) CH3 — CH, -— C+H — CH, — CH2R -» CH3 — C+ — CH — CH2R H
CR
or CH — CH2 — CH2R
(4-10) Some of the advantages of isomerization are:
134
Fluid Catalytic Cracking Handbook
* Higher octane in the gasoline fraction. Isoparaffins in the gasoline boiling range have higher octane than normal paraffins. * Higher-value chemical and oxygenate feedstocks in the C3/C4 fraction. Isobutylene and isoamylene are used for the production of methyl tertiary butyl ether (MTBE) and tertiary amyl methyl ether (TAME). MTBE and TAME can be blended into the gasoline to reduce auto emissions. * Lower cloud point in the diesel fuel. Isoparaffins in the light cycle oil boiling range improve the cloud point. Hydrogen Transfer Reactions Hydrogen transfer is more correctly called hydride transfer. It is a bimolecular reaction in which one reactant is an olefin. Two examples are the reaction of two olefins and the reaction of an olefin and a naphthene. In the reaction of two olefins, both olefins must be adsorbed on active sites that are close together. One of these olefins becomes a paraffin and the other becomes a cyclo-olefin as hydrogen is moved from one to the other. Cyclo-olefin is now hydrogen transferred with another olefin to yield a paraffin and a cyclodi-olefin. Cyclodi-olefin will then rearrange to form an aromatic. The chain ends because aromatics are extremely stable. Hydrogen transfer of olefins converts them to paraffins and aromatics (Equation 4-11). 4 CrlH2n -» 3 Cn H2n+2 + CnH2n^ olefins
—> paraffins
+ aromatic
(4-11)
In the reaction of naphthenes with olefins, naphthenic compounds are hydrogen donors. They can react with olefins to produce paraffins and aromatics (Equation 4-12). 3 C n H 2n + CmH2m
-» 3 Cn H 2n+2
olefins
—> paraffins
+ naphthene
+ Cm H2m^6 + aromatic
(4-12)
A rare-earth-exchanged zeolite increases hydrogen transfer reactions. In simple terms, rare earth forms bridges between two to three acid sites in the catalyst framework. In doing so, the rare earth protects
Chemistry of FCC Reactions
135
those acid sites. Because hydrogen transfer needs adjacent acid sites, bridging these sites with rare earth promotes hydrogen transfer reactions. Hydrogen transfer reactions usually increase gasoline yield and stability. The reactivity of the gasoline is reduced because hydrogen transfer produces fewer olefins. Olefins are the reactive species in gasoline for secondary reactions. Therefore, hydrogen transfer reactions indirectly reduce "overcracking'1 of the gasoline. Some of the drawbacks of hydrogen transfer reactions are: • * • *
Lower gasoline octane Lower light olefin in the LPG Higher aromatics in the gasoline and LCO Lower olefin in the front end of gasoline
Other Reactions Cracking, isomerization, and hydrogen transfer reactions account for the majority of cat cracking reactions. Other reactions play an important role in unit operation. Two prominent reactions are dehydrogenation and coking. Dehydrogenation. Under ideal conditions (i.e., a "clean" feedstock and a catalyst with no metals), cat cracking does not yield any appreciable amount of molecular hydrogen. Therefore, dehydrogenation reactions will proceed only if the catalyst is contaminated with metals such as nickel and vanadium. Coking. Cat cracking yields a residue called coke. The chemistry of coke formation is complex and not very well understood. Similar to hydrogen transfer reactions, catalytic coke is a "bimolecular" reaction. It proceeds via carbenium ions or free radicals. In theory, coke yield should increase as the hydrogen transfer rate is increased. It is postulated [4] that reactions producing unsaturates and multi-ring aromatics are the principal coke-forming compounds. Unsaturates such as olefins, diolefins, and multi-ring polycyclic olefins are very reactive and can polymerize to form coke. For a given catalyst and feedstock, catalytic coke yield is a direct function of conversion. However, an optimum riser temperature will minimize coke yield. For a typical cat cracker, this temperature is
136
Fluid Catalytic Cracking Handbook
about 950°F (510°C). Consider two riser temperatures, 850°F and 1,050°F (454°C and 566°C), at the extreme limits of operation. At 850°F, a large amount of coke is formed because the carbenium ions do not desorb at this low temperature. At 1,050°F (566°C), a large amount of coke is formed, largely due to olefin polymerization. The minimum coking temperature is within this range.
THERMODYNAMIC ASPECTS As stated earlier, catalytic cracking involves a series of simultaneous reactions. Some of these reactions are endothermic and some are exothermic. Each reaction has a heat of reaction associated with it (Table 4-4). The overall heat of reaction refers to the net or combined heat of reaction. Although there are a number of exothermic reactions, the net reaction is still endothermic. The regenerated catalyst supplies enough energy to heat the feed to the riser outlet temperature, to heat the combustion air to the flue gas temperature, to provide the endothermic heat of reaction, and to compensate for any heat losses to atmosphere. The source of this energy is the burning of coke produced from the reaction. It is apparent that the type and magnitude of these reactions have an impact on the heat balance of the unit. For example, a catalyst with less hydrogen transfer characteristics will cause the net heat of reaction to be more endothennic. Consequently this will require a higher catalyst circulation and, possibly, a higher coke yield to maintain the heat balance.
SUMMARY Although cat cracking reactions are predominantly catalytic, some nonselective thermal cracking reactions do take place. The two processes proceed via different chemistry. The distribution of products clearly confirms that both reactions take place, but that catalytic reactions predominate. The introduction of zeolites into the FCC catalyst in the early 1960s drastically improved the performance of the cat cracker reaction products. The catalyst acid sites, their nature, and strength have a major influence on the reaction chemistry. Catalytic cracking proceeds mainly via carbenium ion intermediates. The three dominant reactions are cracking, isomerization, and hydrogen
Table 4-4 Some Thermodynamic Data for Idealized Reactions of Importance in Catalytic Cracking Log KE (equilibrium constant) Reaction Class Cracking Hydrogen transfer Isomerization
Transalkylation Cyclization Dealkylation Dehydrogenation Polymerization Paraffin Alkylation Source: Venuto [2]
Specific Reaction n-C10H22 -> n-C7H16 + C3H6 1~C8H16 -> 2C4Hg 4C6H12 -» 3C6H14 + C6H6 cyclo-C6Hl2 + 3 1-C5H!0 -> 3n-C5H12 + C6H6 1-C4H8 -» trans-2-C4H8 n-C6H10 -» iso-C4H10 o-C6H4(CH3)2 -> m-C6H4(CH3)2 cyclo-C6H12 -» CH3-cyclo-C5H9 C6H6 + m-C6H4(CH3)2 -> 2C6H5CH3 1-C7H14 -» CH3-cyclo-C6H11 iso-C3H7-C6H5 -> C6H6 + C3H6 n-C6H14 ^ 1-C6H12 + H2 3C2H4 —> 1-C6H12 1-C4H8 + iso-C4H10 -> iso-C8H18
850°F
950°F
980°F
2.04 1.68 12.44 11.22 0.32 -0.20 0.33 1.00 0.65 2.11 0.41 -2.21 — —
2.46 2.10 11.09 10.35 0.25 -0.23 0.30 1.09 0.65 1.54 0.88 -1.52
— 2.23 —
— —
— 0.09 -0.36 — 1.10 0.65 — 1.05 — -1.2 3.3
Heat of Reaction BTU/moie 950°F 32,050 33,663 109,681 73,249 -4,874 -3,420 -1,310 6,264 -221 -37,980 40,602 56,008
— _
O
O O
138
Fluid Catalytic Cracking Handbook
transfer. Finally, the type and degree of reactions occurring will influence the unit heat balance.
REFERENCES* 1. Gates, B. C., Katzer, J. R., and Schuit, G. G., Chemistry of Catalytic Processes. New York: McGraw-Hill, 1979. 2. Venuto, P. B. and Habib, E. T., Fluid Catalytic Cracking with Zeolite Catalysts. New York: Marcel Dekker, Inc., 1979, 3. Broekhoven, E. V. and Wijngaards, H., "Investigation of the Acid Site Distribution of FCC Catalysts with Ortho-xylene as a Model Compound," 1988 Akzo Chemicals FCC Symposium, Amsterdam, The Netherlands, 4. Koerroer, G. and Deeba, M., "The Chemistry of FCC Coke Formation," Engelhard Corporation, The Catalyst Report, Vol. 7, Issue 2, 1991.
The author also expresses appreciation to Messrs. Terry Reid of Akzo Nobel and Tom Habib of Davison Div., W. R. Grace & Co., for their many helpful comments.
CHAPTER 5
Unit Monitoring and Control The only proper way to monitor the performance of a cat cracker is by periodic material and heat balance surveys on the unit. By carrying out these tests frequently, one can collect, trend, and evaluate the unit operating data. Additionally, meaningful technical service to optimize the unit operation should be based on regular test runs. Understanding the operation of a cat cracker also requires in-depth knowledge of the unit's heat balance. Any changes to feedstock quality, operating conditions, catalyst, or mechanical configuration will impact the heat balance. Heat balance is an important tool in predicting and evaluating the changes that will affect the quantity and the quality of FCC products. Finally, before the unit can produce one barrel of product, it must circulate catalyst smoothly. One must be familiar with the dynamics of pressure balance and key process controls. The main topics discussed in this chapter are: • • • •
Material Balance Heat Balance Pressure Balance Process Control Instrumentation
In the material and heat balance sections, the discussions include: • Two methods for performing test runs • Some practical steps for carrying out a successful test run • A step-by-step method for performing a material and heat balance survey • An actual case study
139
140
Fluid Catalytic Cracking Handbook
In the pressure balance section, the significance of the pressure balance in debottlenecking the unit is discussed. Finally, fundamentals of both "basic" and "advanced" process controls are presented. This chapter presents the entire procedure for performing heat and weight balances. The last section of the chapter discusses the use of the distributed control system and computer in automating the process,
MATERIAL BALANCE Complete data collection should be carried out weekly. Since changes in the unit are continuous, regular surveys permit distinction among the effects of feedstock, catalyst, and operating conditions. An accurate assessment of a cat cracker operation requires reliable plant data. A reasonable weight balance should have a 98% to 102% closure. In any weight balance exercise, the first step is to identify the input and output streams. This is usually done by drawing an envelope(s) around the input and output streams. Two examples of such envelopes are shown in Figure 5-1. One of the key pieces of data is the composition of products leaving the reactor. The reactor effluent vapor entering the main fractionator contains hydrocarbons, steam, and inert gases. By weight, the hydrocarbons in the reactor overhead stream are equal to the fresh feed plus recycle minus the portion of the feed that has been converted to coke. If the feed can contain water, it should be analyzed for and corrected. The sources of steam in the reactor vapor are: lift steam to the standpipe, atomization steam to the feed nozzles, dome steam, and stripping steam. Some units may have other streams and the feed may contain water. Depending on the reactor pressure, approximately 25% to 50% of the stripping steam is entrained with the spent catalyst flowing to the regenerator, which should be deducted. Inert gases such as nitrogen and carbon dioxide enter the riser entrained with the regenerated catalyst. The quantity of these inert gasses is directly related to catalyst circulation rate. These gases flow through the gas plant and leave the unit with the off-gas from the sponge oil absorber column. They are not significant for the weight balance, but they are usually the only source of inerts in the off-gas and should be deducted. FCC products are commonly reported, on an inert-free basis, as the volume and weight fractions of the fresh feed. In a rigorous weight
Unit Monitoring and Control
141
External Streams^-"
Figure 5-1. FCC unit input/output streams.
balance, gasoline and light cycle oil (LCO) yields and unit conversion are reported based on fixed end points. The common end points are 430°F (221 °C) TBP for gasoline and 700°F TBP for LCO, Other popular cut points are 430°F (221°C) ASTM D-86 for gasoline and 650°F (343°C) or 670°F (354°C) ASTM D-86 for LCO. Using fixed
142
Fluid Catalytic Cracking Handbook
cut points isolates the reactor system from the distillation system performance. Conversion is defined as the volume or weight percent of feedstock converted to gasoline and other lighter products, including coke. However, conversion is typically calculated by subtracting the volume percent or weight percent of liquid products heavier than gasoline from fresh feed, and dividing by the volume or weight of fresh feed. This is shown as follows: ~ , Feed - (light cycle oil + heavy cycle oil + decanted oil) . ,,„ Conversion m% = ^-^ —x 100
Depending on seasonal demands, the gasoline end point can range from 380°F to 450°F (193°C to 232°C). Undercutting of gasoline increases the LCO product and can appear as low conversion. Therefore, it is necessary to distinguish between the apparent and true conversion. The apparent conversion is calculated before the gasoline end point adjustment is made, and the true conversion is calculated after the adjustment.
Testing Methods The material balance around the riser requires the reactor effluent composition. Two techniques are used to obtain this composition. Both techniques require that the coke yield be calculated. The first technique is to draw an envelope with the reactor effluent as the inlet stream and the product flows as the outlet streams. Streams from other units must be included. The flow rates and compositions of the entering and leaving streams are then totaled. The net is the reactor effluent. This is the method practiced by most refiners. The second technique involves direct sampling of the reactor effluent (Figure 5-2). In this technique, a sample of reactor effluent is collected in an aluminized polyester bag for separation and analysis. There are several advantages and disadvantages to reactor effluent sampling; Advantages of Reaction Mix Sampling • Allows data gathering on different sets of conditions without waiting for the recovery side to equilibrate.
Sample probe
Gate and ball valves
00
Slop container
Figure 5-2,
Reaction mix sampling [2].
144
Fluid Catalytic Cracking Handbook
* Eliminates concern about rate and compositions of extraneous streams entering the gas plant because they are not included in the overall balance. * Eliminates concern about correcting for end points because the effluent sample is cut at the desired TBP end point. * Eliminates concern about obtaining a 100% weight balance. Disadvantages of Reaction Mix Sampling * Possible leaks during sampling. * Possible inaccurate measurement of volume of gas and weight of liquid. * Requires qualified individuals to perform the test. » Requires separate lab to perform analyses. * Can require special procedures and be expensive. Recommended Procedures for Conducting a Test Run A successful test run requires a clear definition of objectives, careful planning, and proper interpretation of the results. The following steps can be used as a guide to ensure a smooth and successful test run, Prior to the Test Run 1. Issue a memo to the involved departments: operations, laboratory, maintenance, and oil movement. Communicate the purpose, duration, and scope of the test run. Include a list of samples and the required analyses (Table 5-1). 2, Inform the units feeding the FCC. The composition of FCC feedstock should remain relatively constant during the test run. Flow meters should be zeroed and calibrated. Sample taps should be checked, particularly those that are not used regularly. 5, The sample bombs used to collect gas and LPG products should be purged, marked, and ready. Data Collection 1. The duration of a test run is usually 8 to 12 hours. 2. Operating parameters should be specified. It should be documented which constraints (i.e., blower, wet gas compressor, etc.) the unit is operating against.
Unit Monitoring and Control
145
Table 5-1 Typical Laboratory Analysis of FCC Streams
Tests °API D-86 D-1160 Gas Oil
Sulfur
Viscosity
Metals
/
/
GC
/
/
/
Slurry Recycle
/
/
/
Decanted Oil Product
/
/
/
LCO Product
/
/
/
Gasoline Product
/
/
/
/
/
/
Feedstock /
LPG
C.,\s and C4's
Tail Gas
/
3. The sample taps must be bled adequately before samples are collected. A reliable flue gas analysis is important; an extra sample can be collected. The laboratory should retain the unused samples until all analyses are verified. 4. Pertinent operating data must be collected. A form similar to the one shown in Table 5-2 can be used to gather the data. Mass Balance Calculations 1. The orifice plate meter factor should be adjusted for actual operating parameters. For liquid streams, the flow meters should be adjusted for °API gravity, temperature, and viscosity. For gas streams, the flow rate should be adjusted for the operating temperature, pressure, and molecular weight. 2. Chromatographs of each stream must be normalized to 100%. The GC of the off-gas must include accurate analysis of hydrogen, 3. The coke yield should be calculated using air rate and flue gas composition.
146
Fluid Catalytic Cracking Handbook Table 5-2 Operating Data
Feed and Product Rates Fresh Feed Rate Coker Off Gas FCC Tail Gas LPG Gasoline LCD DO
50,000 bpd (331 nrVhr) 3,000,000 scfd (3,540 m3/hr) 16,000,000 scfd (18,878 m3/hr) 11,565 bpd (77 mVhr) 30,000 bpd (199 m3/hr) 10,000 bpd (66 m3/hr) 3,000 bpd (20 mVhr)
Other Pertinent Flow Rates Dispersion Steam Reactor Stripping Steam Reactor Dome Steam Air to Regenerator
9,000 Ib/hr (4,082 kg/hr) 13,000 Ib/hr (5,897 kg/hr) 1,200 Ib/hr (544 kg/hr) 90,000 scf/min (152,912 m3/hr)
Temperature,°F/°C Riser Inlet Riser Outlet Blower Discharge Regen. Dense Phase Regen. Flue Gas Ambient
594/312 972/522 374/190 1,309/709 1,330/721 80/27
Pressure, psig/Kp Blower Discharge Regen. Dome Reactor Dome Regenerated Catalyst Slide Valve, AP Spent Catalyst Slide Valve, AP
43/296 34/234 33/227 5.8/40 6.0/41
Flue Gas Analysis, Mol% O, CO, CO SO2 N2 + AT Miscellaneous Data Relative Humidity Fresh Catalyst Makeup E-Cat MAT
1.5 15.4 0.0
500 ppm -> 0.05 mol% 83.05 80% 4 tons/day 68%
Unit Monitoring and Control
147
4. The flow rate of each stream should be converted to weight units. 5. The quantity of inert gases and extraneous streams should be subtracted from the FCC gas plant products. 6. The raw mass balance should be reported, including the error, Then the feed/products should be normalized to 100%. The error will be distributed in proportion to flow rates or a known inaccurate meter will be adjusted. 7. Gasoline and LCO rates will be adjusted to standard cut points. 8. The feed characterization correlations discussed in Chapter 2 should be used to determine the composition of fresh feed.
Analysis of Results 1. The yields and quality of the desired products should be reported and compared with the unit targets. 2. The results of this test run should be compared with the results of previous test runs; any significant changes in the yields and/ or operating parameters should be highlighted. 3. The final step is to perform simple economics of the unit operation and make recommendations that improve short- and longterm unit operation. The following case study demonstrates a step-by-step approach to performing a comprehensive material and heat balance.
A test run is conducted to evaluate the performance of a 50,000 bpd (331 m3/hr) FCC unit. The feed to the unit is gas oil from the vacuum unit. No recycle stream is processed; however, the off-gas from the delayed coker is sent to the gas recovery section. Products from the unit are fuel gas, LPG, gasoline, LCO, and decanted oil (DO). Tables 5-2 and 5-3 contain stream flow rates, operating data, and laboratory analyses. The meter factors have been adjusted for actual operating conditions. The mass balance is performed as follows: 1. Identification of the input and output streams used in the overall mass balance equation. 2. Calculation of the coke yield.
148
Fluid Catalytic Cracking Handbook Table 5-3 Feed and Product Inspections
Feed °API Gravity Sulfur, Wt% Analine Point, °F/ °C RI @ 67°C Viscosity, SSU @ 150°F (65.5°C) @ 210°F (98.9°C) Distillation, °F Vol% 10 30 50 70 90 EP
Decanted Oil
Gasoline
LCO
58.5
21.5
2.4
D-86
D-86
D-1160
125 160 213 285 369 433
477 514 547 576 627 666
646 687 720 771 846 1,055
25.2 0.5 208/97.8 1.4854 109 54 D-1160 682 766 835 901 1,001 1,060
Mole% Composition of FCC Gas Plant Streams Component
H, CH4 C, C2= C3
c,=
IC4 NC4 C4 C5+ H2S N2
CO2
Total Sp. Gravity
FCC Tail Gas
15.5 35.8 17.1 11.0 1.6 4.7 0.7 0.2 1.3 1.0 2.1 7.2 1.8 100.0 0.78
LPG
17.9 31.3 16.1 10.9 23.8
100.0 0.55
FCC Gasoline
0.4 2.0 4.4 93.2
100.0
Coker Off-Gas
8.0 47.2 14.9 2.5 8.4 4.4 0.9 3.2 3.4 4.9 2.0 0.2 100.0 0.96
Unit Monitoring and Control
3, 4, 5, 6,
149
Conversion of the flow rates to weight units (e.g., Ib/hr). Normalization of the data to obtain a 100% weight balance. Determination of the component yields. Adjustment of the gasoline, LCO, and decanted oil yields to standard cut points.
Input and Output Streams in the Overall Mass Balance As shown in Envelope 1 of Figure 5-1, the input hydrocarbon streams are fresh feed and coker off-gas. The output streams are FCC tail gas (minus inerts), LPG, gasoline, LCO, DO, and coke. Coke Yield Calculations As discussed in Chapter 1, a portion of the feed is converted to coke in the reactor. This coke is carried into the regenerator with the spent catalyst. The combustion of the coke produces H2O, CO, CO2, SO2, and traces of NOx. To determine coke yield, the amount of dry air to the regenerator and the analysis of flue gas are needed. It is essential to have an accurate analysis of the flue gas. The hydrogen content of coke relates to the amount of hydrocarbon vapors carried over with the spent catalyst into the regenerator, and is an indication of the reactor-stripper performance. Example 5-1 shows a step-by-step calculation of the coke yield. Example 5-1 Determination of the Unit's Coke Yield
Given: Wet air = 90,000 SCFM, Relative Humidity = 80%, Ambient Temperature = 80°F (26.7°C) Figure 5-3 can be used to obtain percent dry air as a function of ambient temperature and relative humidity. For this example, the percentage of dry air is 97.1% or: A- = Ami 90,000SCF x Imole x 60 Min = ,,,,., ,. • nDry Air 0.971 x — 13,817 moles/hr Min 379.5 SCF 1 hr Flue gas rate (dry basis) is calculated from the dry air rate using nitrogen and argon as tie elements.
150
Fluid Catalytic Cracking Handbook
„ ,, , . N (13,817 moles/hrx 0.7901) i a i _ , „ * Flue gas rate (dry = 13,145 moles/hr J basis)= 0.8305 0.7901 and 0.8305 are concentrations of (nitrogen + argon) in atmospheric dry air and flue gas (from analysis), respectively. The flow rates of each component in the flue gas stream are: « * * *
O2 out = 0.015 x 13,145 moles/hr = 197 moles/hr CO2 out = 0.154 x 13,145 moles/hr = 2,024 moles/hr SOj out = 0.0005 x 13,145 moles/hr = 7 moles/hr (N2~ + Ar) out = 0.8305 x 13,145 moles/hr = 10,917 moles/hr
An oxygen balance can be used to calculate water formed by the combustion of coke: * O2 out = 197 + 2,024 +7 = 2,228 moles/hr * 1)3 in = 0.2095 x 13,817 moles/hr = 2,895 moles/hr * O2 used for combustion of hydrogen = 2,895 - 2,228 = 667 raoles/hr Since for each mole of O2, two moles of water are formed, the amount of water is: * H2O formed = 667 x 2 = 1,334 moles/hr Components of coke are carbon, hydrogen, and sulfur. Their rates are calculated as follows: * * « *
Carbon = 2,024 moles/hr x 12 Ibs/mole = 24,288 Ibs/hr Hydrogen = 1,334 moles/hr x 2.02 Ibs/mole = 2,695 Ibs/hr Sulfur = 7 moles/hr x 32.1 Ibs/moles = 225 Ibs/hr Coke = 24,288 + 2,695 + 225 = 27,208 Ibs/hr
* H-, content of coke, wt% = —: — x 100 = 9.9 27,231 Ibs/hr (The hydrogen content of coke indicates the amount of hydrocarbon vapors carried through the stripper with the spent catalyst.
Conversion to Unit of Weight, Ibs/hr The next step is to convert the flow rate of each stream in the overall mass balance equation to the unit of weight (e.g., Ibs/hr). Example 5-2 shows these conversions for gas and liquid streams.
Dry Air versus Relative Humidity & Temperature
30
50
70
80
100
Temperature ,Deg F
Figure 5-3,
Dry air versus relative humidity and temperature.
1tO
152
Fluid Catalytic Cracking Handbook
Example 5-2 Conversion of Input and Output Streams to the Unit of Weight (Ib/hr)
„ , „ . 50,000bbl 1day 141.5 350.3 Ib • Fresh Feed = —-—-- x - - x • -- x - —— day 24 hr (131.5 + 25.2) bbl = 658,964 lb/hr »
„,
3,000,000 SCF A.V 1day 1mole 27.8 lbs ni-—«___Q« 1|«,„,,. , •* . *„._„,„_ _^ V _. ,. , . . . . u _.-_...-... Sf\ x1U/m/nr A. v ._.....,.._L..._, JvJ.O 111 day 24 hr 379.5 SCF 1mole
f^r\lrpir gas CTQG — v^UJVCI —*"_!_ " !_ ........
.™ .. 16,000,000 SCF Iday Imole • FCC tail gas = —-- -x - ^x day 24 hr 379.5 SCF
22.6 Ibs x Imole
= 39,701 Ib/hr The amount of inerts in the FCC tail gas is: 16,000,OOOSCF 1day _ _ _ „ Imole KT •N x - i- x 0.072 x 2 = —-- -day 24 hr 379.5 SCF „ 16,000,OOOSCF A n - t 1day 1mole . CO2 = —!- -x 0.021 x - ^-x - day 24 hr 379.5 SCF
281bs 3,542lb/hr x= 3,542 Ib/hr Imole 441bs , , » « , , „ x= 1,623 Ib/hr Imole
• Inert-free FCC tail gas = 39,701 - (3,542 + 1,623) = 34,537 Ib/hr .
LpG= H.565bbl x lday x
day
24 hr
141.5 X35031b = (131.5 + 123.5) bbl
„ r 30,000bbl day 141.5 350.31b • Gasoline = —-- x --x x day 24 hr (131.5 + 58.5) bbl = 326, 102 Ib/hr
day
24hr
141.5 (131.5 + 21.5)
f
bbl
3,000bbl 1day 141.5 350.31b ., «--,.,, = —x - i-x x -- = 46,2731b/hr day 24 hr (131.5 + 2.4) bbl
Unit Monitoring and Control
153
Normalization of the Data Because a preliminary weight balance seldom has a 100% closure, it is necessary to normalize the yield to obtain a 100% weight balance, Example 5-3 shows the preliminary overall weight balance.
Example 5-3 Preliminary Overall Weight Balance
Input = Fresh Feed + Coker Off-Gas Output = FCC tail gas + LPG + Gasoline + LCO + DO + Coke • Input = 658,814 + 9,182 = 667,996 Ib/hr • Output = 34,617 + 93,656 + 326,124 + 134,973 + 46,270 + 27,231 = 662,871 lb/hr « Difference = 667,996 - 662,871 = 5,125 lb/hr Error in mass balance = 0.8 wt% The products are adjusted upward in proportion to theilr rates to obtain a 100% weight balance. The normalized rates: • • • • « •
Tail gas LPG Gasoline LCO DO Coke
= 34,883 Ib/hr = 94,460 lb/hr = 328,766 Ib/hr = 136,054 lb/hr = 46,626 Ib/hr = 27,440 Ib/hr
= = = =
11,658 bpd 30,230 bpd 10,077 bpd 3,023 bpd
Component Yield The reactor yield is then determined by performing a component balance. The amount of C5+ in the gasoline boiling range is calculated by subtracting the C4 and lighter components from the total gas plant products. Example 5-4 shows the step-by-step calculation of the component yields. The summary of the results, normalized but unadjusted for the cut points is shown in Table 5-4.
154
Fluid Catalytic Cracking Handbook Example 5-4 Calculation of Individual Components = 0.155 x 16 MMSCFDx 2.02 ~~" 379.5x24
2
0.08 x 3 MMSCFDx 2.02 379,5x24
_„ 0.358x16 MMSCFDx 16 0.472x3.0 MMSCFDx 16 - C M 1 U r t CH 44 = --= 7,585 Ib/hr 379.5x24 379.5x24 C 2 ~~~ = 0-171 xl6MMSCFDx30 379.5x24 /„ _ 0.11x16 MMSCFDx 28 2
379.5x24 =
3
~~
0.149 x 3 MMSCFDx 30 379.5x24 0.025 x 3 MMSCFD x 28 379.5x24
0.016x!6MMSCFDx44 | 0.179x 11,65]8BPDx 175.3 + 379.5x24 " 24 0.084x3 MMSCFDx 44 = 15,262 lb/hr 379.5 x 24
^_ 0.047x16 MMSCFDx 42 0.313x11,658 BPDx 181.8 • Cr3 = + 379.5 x 24 24 0.044x3 MMSCFDx 42 O A C A / I I U / , = 30,504 Ib/hr 379.5 x 24 '
4
^0.002x16MMSCFDx58 "~ 379.5x24
| +
0.109x11,658BPDx204.6 24
0.02 x 30,230 x 204.6 MMSCFD x 42 0.032 x 3 MMSCFD x 58 24 379.5x24 =15,579 Ib/hr =
'4~
0.007x16MMSCFDx58 379.5x24
[ +
0.161x11658BPDx 197.2 24
0.004x30,230x204.6 BPDx 197.2 24 = 16,95 8 Ib/hr
0.009x3 MMSCFDx 58 379.5x24
Unit Monitoring and Control
155
0.013xl6MMSCFDx56 0.238x1 l,658BPDx213.4 + 379.5x24 24 0.044x30,230x213.4 0.034x3MMSCFDx56 = 37,150 Ib/hr 24 379.5x24
Table 5-4
Normalized FCC Weight Balance Summary with Coker Gas Subtracted
Stream Fresh Feed
bpd
ib/hr
Vol% of Feed
Wt% of Feed
50,000
658,814
100.00
100.00
Products
H7 C, C,
c; Total C2 and lighter
H2S C3
C? IC4 NC4
c;
Gasoline (Cs+)
LCO DO Coke Total Apparent Conversion Inerts
0.07 1.15 1.15 0.79 3.16
497 7,585 7,549 5,187 20,818
2,090 4,027 2,064 1,827 4,178
1,032 15,262 30,504 16,958 15,579 37,150
4.18 8.05 4.13 3.65 8.36
0.16 2.32 4.63 2.57 2.36 5.64
28,650 10,077
311,437 136,008
57.30 20.15
47.27 20.64
3,023
46,626 27,440 658,814
6.05
7.08 4.17 100.00 72.28
55,936
5,143
111.87 73.8
156
Fluid Catalytic Cracking Handbook
Adjustment of Gasoline and LCO Cut Points As discussed earlier in this chapter, gasoline and LCO yields are generally corrected to a constant boiling range basis. The most commonly used bases are 430°F TBP gasoline and 640°F TBP LCO end points. Since TBP distillations are not routinely performed, they are often estimated from the D-86 distillation data. The adjustments to the end points involve the following: * Adding to the raw LCO all the 430°F+ in the raw gasoline and subtracting the 430°F in the LCO stream. « Adding to the raw LCO all the 650°F~ in the raw decanted oil and subtracting the 650°F~ in the decant oil stream. * Adding to the raw gasoline all the 430°F~ in the raw LCO and subtracting the 430°F* in the gasoline stream. • Adding to the raw decanted oil all the 650°F+ in the raw LCO and subtracting the 650°F~ in the decant oil stream. Table 5-5 illustrates steps used to convert ASTM D-86 data to TBP. The laboratory usually converts D-1160 and reports the data as D-86, Extrapolation of the TBP data indicates the following: « « « •
The The The The 514
430°F+ content of the FCCU gasoline is 3 vol%, or 859 bpd. gasoline (430°F~) content of LCO is 8 vol%, or 806 bpd. 650°F+ content ofLCO is 12 vol%, or 1,209 bpd. LCO (650°F~) content of the decanted oil is 17 vol%, or bpd.
Therefore, the adjusted rates are as follows: Gasoline (C5+ to 430°F TBP end point) = 28,650 - 859 + 806 = 28,597 bpd LCO (430°F to 650°F TBP end point) = 10,077 + 514 - 1,209 - 806 + 859 = 9,435 bpd DO (650°F+) = 3,023 + 1,209 - 514 = 3,718 bpd
Table 5-6 shows the normalized FCC weight balance with the adjusted cut points.
Unit Monitoring and Control
Table 5-5 Conversion of ASTM Distillation to TBP Distillation for Gasoline, LCO, and Decanted Oil Gasoline TBP (From Appendix 9, TBP 50% point = 213°F) Given D-86 50% 30% 10% 70% 90% EP -
- 30% = - 10% = - IBP = - 50% = - 70% = 90% =
From Appendix 10 53°F 35°F 25°F 72°F 84°F 64°F
30% TBP = 10% TBP = IBP TBP = 70% TBP = 90% TBP = EP TBP =
140°F 77°F 26°F 297°F 383°F 501°F
LCO TBP (From Appendix 9: TBP 50% point = 561°F) Given D-86 50% - 30% = 30% - 10% = 10% - IBP = 70%-50% = 90% - 70% = EP-90% =
From Appendix 10 33°F 4FF 73°F 29°F 51°F 39°F
30%TBP = 511°F 10% TBP = 441°F IBP TBP = 343°F 70%TBP = 601°F 90% TBP = 660°F EPTBP = 712°F
Decanted Oil TBP (From Appendix 9: TBP 50% point = 744°F) Given D-86
From Appendix 10
50% 30% 10% 70% 90%
30% TBP = 694°F 10% TBP = 624°F IBP TBP = 425°F 70% TBP = 807°F 90% TBP = 886°F
-
30% = 33°F 10% = 41°F IBP = 236°F 50% = 51°F 70% = 75°F
157
158
Fluid Catalytic Cracking Handbook Table 5-6 Normalized and Adjusted FCC Weight Balance Summary
Stream Fresh Feed
bpd
ib/hr
Vol% of Feed
Wt% of Feed
50,000
658,814
100.00
100,00
Products
497 7,585 7,549 5,187 20,818
H, C, C,
c= Total C2 and lighter
0.07 1.15 1.15 0.79 3.16
2,090 4,027 2,064 1,827 4,178
1 ,032 15,262 30,504 16,958 15,579 37,150
4.18 8.05 4.13 3.65 8.36
0.1.6 2 32 4.63 2.57 2.36 5.64
28,597
312,073
57.19
47,37
LCO (430°F TBP to 650°F TBP)
9,435
1 26,004
18.87
19.13
DO (65Q°F+ TBP)
3,718
55,994
7.44
8.50
55,936
27,440 658,814
111.87
4.17 100.00
73.7
72.3
H2S C,=
C
3
IC4 NC4
c;
Gasoline (C5+ to 430°F TBP)
Coke Total True Conversion Inerts
5,143
HEAT BALANCE A cat cracker continually adjusts itself to stay in heat balance. This means that the reactor and regenerator heat flows must be equal (Figure 5-4). Simply stated, the unit produces and burns enough coke to provide energy to:
Unit Monitoring and Control
Steam
Steam Oil Feed Figure 5-4. Reactor-regenerator heat balance.
159
160
Fluid Catalytic Cracking Handbook
• Increase the temperature of the fresh feed, recycle, and atomizing steam from their preheated states to the reactor temperature « Provide the zendothermic heat of cracking • Increase the temperature of the combustion air from the blower discharge temperature to the regenerator flue gas temperature • Make up for heat losses from the reactor and regenerator to the surroundings • Provide for miscellaneous heat sinks, such as stripping steam and catalyst cooling A heat balance can be performed around the reactor, around the stripper-regenerator, and as an overall heat balance around the reactorregenerator. The stripper-regenerator heat balance can be used to calculate the catalyst circulation rate and the catalyst-to-oil ratio.
Heat Balance Around Stripper-Regenerator If a reliable spent catalyst temperature is not available, the stripper is included in the heat balance envelope (II) as shown in Figure 5-4, The combustion of coke in the regenerator satisfies the following heat requirements: « Heat to raise air from the blower discharge temperature to the regenerator dense phase temperature • Heat to desorb the coke from the spent catalyst • Heat to raise the temperature of the stripping steam to the reactor temperature • Heat to raise the coke on the catalyst from the reactor temperature to the regenerator dense phase temperature • Heat to raise the coke products from the regenerator dense temperature to flue gas temperature • Heat to compensate for regenerator heat losses • Heat to raise the spent catalyst from the reactor temperature to the regenerator dense phase temperature Using the operating data from the case study, Example 5-5 shows heat balance calculations around the stripper-regenerator. The results are used to determine the catalyst circulation rate and the delta coke. Delta coke is the difference between coke on the spent catalyst and coke on the regenerated catalyst.
Unit Monitoring and Control
161
Example 5-5 Stripper-Regenerator Heat Balance Calculations
I. Heat generated in the regenerator: C to CO2 = 24,288 Ib/hr x 14,087 Btu/lb = 342 x 106 Btu/hr H2 to H2O = 2,695 Ib/hr x 51,571 Btu/lb = 139 x 106 Btu/hr S to SO2 = 225 Ib/hr x 3,983 Btu/lb = 0.9 x 106 Btu/hr Total heat released in the regenerator: 342 + 139 + 0.9 = 482 x 106 Btu/hr II. Required heat to increase air temperature from blower discharge to the regenerator dense phase temperature: From Figure 5-5, enthalpies of air at 374°F and at 1,309°F are 90 Btu/lb and 355 Btu/lb. Therefore, the required heat is = 407,493 Ib/hr x (355 - 90) Btu/lb = 108.0 x 106 Btu/hr III. Energy to desorb coke from the spent catalyst: Desorption of coke = 27,208 Ib/hr x 1,450 Btu/lb = 39.5 x 106 Btu/hr IV. Energy to heat the stripping steam: Enthalpy of 50 psig-saturated steam = 1,179 Btu/lb Enthalpy of 50 psig at 972°F =1,519 Btu/lb Change of enthalpy = 13,000 Ib/hr x (1,519 - 1,179) Btu/lb = 4.4 x 106 Btu/hr V. Energy to heat the coke on the spent catalyst: 27,231 Ibs/hr x 0.4 Btu/lb-°F x (1,309 - 972)°F = 3.7 x 106 Btu/hr VI. Energy to heat the flue gas from regenerator dense phase to regenerator flue gas temperature: From Figure 5-5, enthalpy of flue gas at 1,309°F = 365 Btu/lb and at 1,330°F = 370 Btu/lb. The required heat is therefore = 433,445 Ib/hr x (370 - 355)°F = 2.6 x 106 Btu/hr VII, Heat loss to surroundings: Assume heat loss from the stripper-regenerator (due to radiation and convection) is 4% of total heat of combustion, i.e., 0.04 x 482.4 MM Btu/hr = 19.3 x 106 Btu/hr
162
Fluid Catalytic Cracking Handbook
VIII. Energy required to heat the spent catalyst from its reactor to the regenerator temperature = 481.9 - 108.0 - 39.5 - 4.4 - 3.7 - 2.6 - 19.3 = 304.4 x 106 Btti/hr IX. Calculation of catalyst circulation ^ , „. . . Catalyst Circulation =
304.4 xl0 6 Btu/hr (0.285 Btu/°F-lb) x (1,309 - 972)°F
= 3.169 x 106 Ibs/hr = 26.4 short tons/min. Where: 0.285 is the catalyst heat capacity (see Figure 5-6) Cat/oil ratio = 3.169 x 106/658,914 = 4.8 .„ , Coke Yield, wt% 4.2 A 0_ „ ACoke = = — = 0.87 wt% cat/oil ratio 4,8
Reactor Heat Balance The hot regenerated catalyst supplies the bulk of the heat required to vaporize the liquid feed (and any recycle) to provide the overall endothermic heat of cracking, and to raise the temperature of dispersion steam and inert gases to the reactor temperature. Heat In
Heat Out
Fresh Feed Recycle Air Steam
Reactor Vapors Flue Gas Losses
The calculation of heat balance around the reactor is illustrated in Example 5-6. As shown, the unknown is the heat of reaction. It is calculated as the net heat from the heat balance divided by the feed flow in weight units. This approach to determining the heat of reaction is acceptable for unit monitoring. However, in designing a new cat cracker, a correlation is needed to calculate the heat of reaction. The heat of reaction is needed to specify other operating parameters, such
Unit Monitoring and Control
W
I
f
0J
«M#
E
0,3
0.295
0.29
0.285
0.28
0.275
0.27
0.265
0.26
0.25S
to
20
30
40
50
60
70
Alumina Content, Wt.%
FIgyre
of the FCC
as a function of the
content,
Unit Monitoring and Control
165
as preheat temperature. Depending on conversion level, catalyst type, and feed quality, the heat of reaction can vary from 120 Btu/lb to 220 Btu/lb. In the unit, the heat of reaction is a useful tool. It is an indirect indication of heat balance accuracy. Trending the heat of reaction on a regular basis provides insight into reactions occurring in the riser and the effects of feedstock and catalyst changes. Example 5-6 Reactor Heat Balance
I. Heat into the reactor 1. Heat with regenerator catalyst = 3.169 x 106 Ib/hr x 0.285 Btu/lb-°F x 1,309°F = 1,182.4 x 106 Btu/hr = 1,182.4 x 106 Btu/hr 2. Heat with the fresh feed: At a feed temperature of 594°F, °API gravity = 25.2 and K factor = 12.08, the feed liquid enthalpy is 405 Btu/lb (see Figure 5-7), therefore, heat content of the feed is = 658,914 Ib/hr x 405 Btu/lb = 266.9 x 106 Btu/hr. 3. Heat with atomizing steam: From steam tables, enthalpy of 150 Ib saturated steam = 1,176 Btu/lb, therefore, heat with steam = 10,000 Ib/hr x 1,176 Btu/lb = 11.8 x 106 Btu/hr. 4 Heat of adsorption: The adsorption of coke on the catalyst is an exothermic process; the heat associated with this adsorption is assumed to be the same as desorption of coke in the regenerator (i.e., 35.3 x 106 Btu/hr). Total heat in = 1,182.4 + 266.9 + 11.8 + 35.3 = 1,496.4 x 106 Btu/hr. II. Heat out of the reactor 1, Heat with spent catalyst = 3,169 x 106 Ib/hr x 0.285 Btu/lb-°F x 972°F = 878 x 106 Btu/hr. 2, Heat required to vaporize feed: From Figure 5-8, enthalpy reactor vapors = 778 Btu/lb, therefore, heat content of the vaporized products = 658,814 Ib/hr x 778 Btu/lb = 512.6 x 106 Btu/hr. 3. Heat content of steam: Enthalpy of steam @ 972°F = 1,519 Btu/lb, therefore, heat content of steam = 10,000 Ib/hr x 1,519 Btu/lb = 15.2 x 106 Btu/hr. 4. Heat loss to surroundings: Assume heat loss due to radiant and convection to be 2% of heat with the regenerated catalyst (i.e., 0.02 x 304.4 = 6.1 x 106 Btu/hr)
161
Fluid Catalytic Cracking Handbook
III. Calculation of heat of reaction Total heat out = total heat in Total heat out = 878 x 106 + 512.6 x 106 + 15.2 x 106 + 6.1 x 1C)6 + overall heat of reaction = Total heat in = 1,499.6 x 106Btu/hr Overall endothermic heat of reaction = 84.5 x 106 Btu/hr or —» 128.2 Btu/lb of feed.
Analysis of Results Once the material and heat balances are complete, a report must be written. It will first present the data. It will then discuss factors affecting product quality and any abnormal results. It will then discuss the key findings and recommendations to improve unit operation. In the previous examples, the feed characterizing correlations in Chapter 2 are used to determine composition of the feedstock. The results show that the feedstock is predominantly paraffinic (i.e., 61.6% paraffins, 19.9% naphthenes, and 18.5% aromatics). Paraffinic feedstocks normally yield the most gasoline with the least octane. This confirms the relatively high FCC gasoline yield and low octane observed in the test run. This is the kind of information that should be included in the report. Of course, the effects of other factors, such as catalyst and operating parameters, will also affect the yield structure and will be discussed. The coke calculation showed the hydrogen content to be 9.9 wt%. As discussed in Chapter 1, every effort should be made to minimize the hydrogen content of the coke entering the regenerator. The hydrogen content of a well-stripped catalyst is in the range of 5 wt% to 6 wt%. A 9.9 wt% hydrogen in coke indicates either poor stripper operation and/or erroneous flue gas analysis.
PRESSURE BALANCE Pressure balance deals with the hydraulics of catalyst circulation in the reactor/regenerator circuit. The pressure balance starts with the static pressures and differential pressures that are measured. The various pressure increases and decreases in the circuit are then calculated. The object is to:
•5 I
Unit Monitoring and Control 167
1000
600 900
920
940
960
980
1000
1020
1040
1060
Deg. F -*-K«11
-*-K-t2
-*-K»13
Figyre 5-8. Hydrocarbon vapor enthalpies at various Watson K factors.
1080
1100
Unit Monitoring and Control
« * * *
169
Maximize catalyst circulation Ensure steady circulation Maximize the available pressure drop at the slide valves Minimize the loads on the blower and the wet gas compressor
A clear understanding of the pressure balance is extremely important in "squeezing" the most out of a unit. Incremental capacity can come from increased catalyst circulation or from altering the differential pressure between the reactor-regenerator to "free up" the wet gas compressor or air blower loads. One must know how to manipulate the pressure balance to identify the "true" constraints of the unit. Using the drawing(s) of the reactor-regenerator, the unit engineer must be able to go through the pressure balance and determine whether it makes sense. He or she needs to calculate and estimate pressures, densities, pressure buildup in the standpipes, etc. The potential for improvements can be substantial.
Basic Fluidization Principals A fluidized catalyst behaves like a liquid. Catalyst flow occurs in the direction of a lower pressure. The difference in pressure between any two points in a bed is equal to the static head of the bed between these points, multiplied by the fluidized catalyst density, but only if the catalyst is fluidized. FCC catalyst can be made to flow like a liquid, but only if the pressure force is transmitted through the catalyst particles and not the vessel wall. The catalyst must remain in a fluidized state as it makes a loop through the circuit. To illustrate the application of the above principals, the role of each major component of the circuit is discussed in the following sections, followed by an actual case study. As a reference, Appendix 8 contains fluidization terms and definitions commonly used in the FCC.
Major Components of the Reactor-Regenerator Circuit The major components of the reactor-regenerator circuit that either produce or consume pressure are as follows: * Regenerator catalyst hopper * Regenerated catalyst standpipe
170
• • • • •
Fluid Catalytic Cracking Handbook
Regenerated catalyst slide (or plug) valve Riser Reactor-stripper Spent catalyst standpipe Spent catalyst slide (or plug) valve
Regenerator Catalyst Hopper In some FCC units, the regenerated catalyst flows through a hopper prior to entering the standpipe. The hopper is usually internal to the regenerator and often of an inverted cone design. It provides sufficient time for the regenerated catalyst to be deaerated before entering the standpipe. This causes the catalyst entering the standpipe to have maximum flowing density. The higher the density, the greater the pressure buildup in the standpipe. In some FCC designs, the regenerated catalyst hopper is external with fluffing aeration to control the catalyst density entering the standpipe. Regenerated Catalyst Standpipe The standpipe's height provides the driving force for transferring the catalyst from the regenerator to the reactor. The elevation difference between the standpipe entrance and the slide valve is the source of this pressure buildup. For example, if the height difference is 30 feet (9.2 meters) and the catalyst density is 40 lb/ft3 (641 kg/m3), the pressure buildup is: 40 1h 1 ft2 Pressure Gain = 30 ft x -^3 x , = 8.3 psi (57Fkp) ft 144 in2
The key to obtaining maximum pressure gain is to keep the catalyst fluidized over the length of the standpipe. Longer standpipes will require external aeration. This compensates for compression of the entrained gas as it travels down the standpipe. Aeration should be added evenly along the length of the standpipe. In shorter standpipes sufficient flue gas is often carried down with the regenerated catalyst to keep it fluidized and supplemental aeration is unnecessary. Overaeration leads to unstable catalyst flow and must be avoided.
Unit Monitoring and Control
171
Aside from proper aeration, the flowing catalyst must contain sufficient 0-40 micron fines to avoid defluidization. Regenerated Catalyst Slide Valve The purpose of the regenerated catalyst slide valve is threefold: to regulate the flow of regenerated catalyst to the riser, to maintain pressure head in the standpipe, and to protect the regenerator from a flow reversal. Associated with this control and protection is usually a I psi to 8 psi (7 Kp to 55 Kp) pressure drop across the valve. Riser The hot-regenerated catalyst is transported up the riser and into the reactor-stripper. The driving force to carry this mixture of catalyst and vapors comes from a higher pressure at the base of the riser and the low density of the catalyst/vapor mix. The large density difference between the fluidized catalyst on the regenerator side (approximately 40 lb/ft3) and the mixture of cracked hydrocarbon vapors and catalyst on the riser side (approximately 1 lb/ft3) drives the system. As for the pressure balance, this transport of catalyst results in a pressure drop in a range of 5 psi to 9 psi (35 Kp to 62 Kp). This drop is due to static head and, to a lesser extent, friction and acceleration of the fluid. In an existing riser, operating changes, such as higher catalyst circulation or lower vapor velocity, can affect the density of reaction mixture and increase pressure drop. This will affect the slide valve differential and percent opening. Reactor-Stripper The catalyst bed in the reactor-stripper is important for three reasons: « to provide enough residence time for proper stripping of the entrained hydrocarbon vapors prior to entering the regenerator; • to provide adequate static head for flow of the spent catalyst to the regenerator; and • to provide sufficient backpressure to prevent reversal of hot flue gas into the reactor system. Assuming a stripper with a 20-ft bed level and a catalyst density of 40 lb/ft 3 , the static pressure is: 3 , 40 lbs/ft .. . 20n ft x f— r 2 2 = 5.5 psi
144 in /ft
172
Fluid Catalytic Cracking Handbook
Spent Catalyst Standpipe From the bottom of the stripper, the spent catalyst flows into the spent catalyst standpipe. Sometimes the catalyst is partially defluidized in the stripper cone. To counter this, "dry" steam is usually added (through a distributor) to fluidize the catalyst prior to its entering the standpipe. The loss of fluidization in the stripper cone can cause a buildup of dense phase catalyst along the cone walls. This buildup can restrict catalyst flow into the standpipe, causing erratic flow and reducing pressure buildup in the standpipe. Like the regenerated catalyst standpipe, the spent catalyst standpipe may require supplemental aeration to obtain optimum flow characteristics. "Dry" stearn is the usual aeration medium. Spent Catalyst Slide or Plug Valve The spent catalyst slide valve is located at the base of the standpipe. It controls the stripper bed level and regulates the flow of spent catalyst into the regenerator. As with the regenerated catalyst slide valve, the catalyst level in the stripper generates pressure as long as it is fluidized. The pressure differential across the slide valve will be at the expense of consuming a pressure differential in the range of 3 psi to 6 psi (20 kp to 40 kp). In earlier Model II and Model III FCC units, spent catalyst was transported into the regenerator using 50% to 100% of combustion air. This spent cat riser was designed for a minimum air velocity of 30 ft/sec (9.1 m/sec).
Case Study A survey of the reactor-regenerator circuit of a 50,000 bpd (331 m3/hr) cat cracker produced these results: Reactor dilute phase (dome) pressure Reactor catalyst dilute phase bed level Reactor-stripper catalyst bed level Reactor-stripper catalyst density Spent, catalyst standpipe elevation Pressure above the spent catalyst slide valve Spent catalyst slide valve AP (@ 55% opening)
= = = = = = =
19.0 psig/131 Kp 25.0 ft/7.6 m 18.0 ft/5.5 m 40 Ib/ft3/640 kg/m3 14.4 ft/4.4 m 26.1 psig/180 Kp 4.0 psi/27.6 Kp
Unit Monitoring and Control Regenerator dilute phase catalyst level Regenerator dense phase catalyst bed level Catalyst density in the regenerator dense phase Regenerated catalyst standpipe elevation Pressure above the regenerated catalyst slide valve Regenerated catalyst slide valve AP (@ 30% opening) Reactor-regenerator pressure AP
173
= 27.0 ft/8.2 m = 15.0 ft/4.6 m = 25 Ib/ft3/400 kg/m3 = 30.0 ft/9.1 m = 30.5 psig/210.3 Kp = 5.5 psi/37.9 Kp = 3 . 0 psi/20.7 Kp
Also, see Figure 5-9 for a graphical representation of the preliminary results. Starting with the reactor dilute pressure as the working point, the pressure head corresponding to 25 feet (7.6 m) of dilute catalyst fines is: (25 ft) x (0.6 lb/ft3) x (1 ft2/!44 in 2 ) = 0.1 psig (0.7 Kp) Therefore, the pressure at the top of the stripper bed is: 19.0 + 0.1 = 19.1 psig (131.7 Kp) The static-pressure head in the stripper is: (18 ft) x (40 lb/ft3) x (1 ft/144 in 2 ) = 5.0 psig (34.5 Kp) The pressure above the spent catalyst standpipe is: 19.1 + 5.0 = 24.1 psig (166.2 Kp) The pressure buildup in the spent catalyst standpipe is:
26.1 - 24.1 =2 psi (13.8 K p ) The pressure below the spent catalyst slide valve is: 26.1 -4.0 = 22.1 psig (152 Kp)
The pressure head corresponding to 28 feet (8.5 m) of dilute catalyst fines in the regenerator is: (28 ft) x (1 lb/ft3) x (1 ft2/144 in2) = 0.2 psig (1.4 Kp)
174
Fluid Catalytic Cracking Handbook
Rx Vapor
Reactor
Psi diff.
Oil Feed Figure 5-9. Preliminary pressure balance survey.
Unit Monitoring and Control
175
The pressure in the regenerator dome is: 22. J - 0,2 = 21.9 psig (151.0 KP) The static pressure head in the regenerator is: (18 ft) x (25 lb/ft3) x (1 ft2/!44 in2) = 3.1 psig (21.4 Kp) The pressure above the regenerated catalyst standpipe is: 22.1 + 3.1 = 25.2 psig (173.7 Kp) The pressure buildup in the regenerated catalyst standpipe is: 30.5 - 25.2 = 5.3 psi (36.5 Kp)
The pressure below the regenerated catalyst slide valve is: 30.5 - 5.5 = 25 psig (172.4 Kp) The pressure drop in the vertical riser is: 25 - 19 = 6 psi (41.4 Kp)
The catalyst density in the spent catalyst standpipe is: (2.0 lb/in2) x (144 in2/ft2)/(14.4 ft) = 20 lb/ft3 = 320 kg/m3 The catalyst density in the regenerated catalyst standpipe is: (5.3 lb/in 2 ) x (144 in2/ft2)/(30 ft) = 25.4 lb/ft3 = 407 kg/m3 Figure 5-10 shows the results of the above pressure balance survey. Analysis of the Findings The pressure balance survey indicates that neither the spent nor the regenerated catalyst standpipe is generating "optimum" pressure head. This is evidenced by the low catalyst densities of 20 lb/ft3 (320 kg/m3) and 25.4 lb/ft3 (407 kg/m3), respectively. As indicated in Chapter 8, several factors can cause low pressure, including "under" or "over"
176
Fluid Catalytic Cracking Handbook
Rx Vapor Reactor
Psi diff.
Oil Feed Figure 5-10. Pressure balance survey with calculated standpipe densities.
Unit Monitoring and Control
177
aeration of the standpipes. In a well-fluidized standpipe, the expected catalyst density is in the range of 35 - 45 lb/ft3 (561 kg/m 3 to 721 kg/m3). If the catalyst density in the spent catalyst standpipe was 40 lb/ft'* (640 kg/m3) instead of 20 lb/ft3 (320 kg/m3), the pressure buildup would have been 4.0 psi instead of 2.0 psi. The extra 2 psi (13.8 Kp) can be used to circulate more catalyst or to lower the reactor pressure. In the regenerated catalyst standpipe, a 40 lb/ft3 (640 kg/m3) catalyst ^ ' 3 density versus a 25.4 lb/ft (407 kg/m") density produces 3 psi (20,7 Kp) more pressure head, again allowing an increase in circulation or a reduction in the regenerator pressure (gaining more combustion air).
Process control instrumentation controls the FCC unit in a safe, monitored mode with limited operator intervention. Two levels of process control are used: • Basic supervisory control • Advanced process control (APC)
Basic Supervisory Control The primary controls in the reactor-regenerator section are flow, temperature, pressure, and catalyst level. The flow controllers are often used to set desired flows for the fresh feed, stripping steam, and dispersion steam. Each flow controller usually has three modes of control: manual, auto, and cascade. In manual mode, the operator manually opens or closes a valve to the desired percent opening. In auto mode, the operator enters the desired flow rate as a set-point. In cascade mode, the controller set-point is an input from another controller. The reactor temperature is controlled by a temperature controller that regulates the regenerated catalyst slide valve. The regenerator temperature is not automatically controlled but depends on its mode of operation. In partial combustion, the regenerator temperature is controlled by adjusting the flow of combustion air to the regenerator. In full burn, the regenerator temperature is a function of operating conditions such as reactor temperature and slurry recycle.
178
Fluid Catalytic Cracking Handbook
The reactor pressure is not directly controlled; instead, it floats on the main column overhead receiver. A pressure controller on the overhead receiver controls the wet gas compressor and indirectly controls the reactor pressure. The regenerator pressure is often controlled directly by regulating the flue gas slide or butterfly valve. In some cases, the flue gas slide or butterfly valve is used to control the differential pressure between the regenerator and reactor. The reactor or stripper catalyst level controller is controlled with a level controller that regulates the movement of the spent catalyst slide valve. The regenerator level is manually controlled to maintain catalyst inventory. Regenerated and Spent Catalyst Slide Valve Low Differential Pressure Override Normally, the reactor temperature and the stripper level controllers regulate the movement of the regenerated and spent catalyst slide valves. The algorithm of these controllers can drive the valves either fully open or fully closed if the controller set-point is unobtainable. It is extremely important that a positive and stable pressure differential be maintained across both the regenerated and spent catalyst slide valves. For safety, a low differential pressure controller overrides the temperature/level controllers should these valves open too much. The shutdown is usually set at 2 psi (14 Kp). The direction of catalyst flow must always be from the regenerator to the reactor and from the reactor back to the regenerator. A negative differential pressure across the regenerated catalyst slide valve can allow hydrocarbons to back-flow into the regenerator. This is called a flow reversal and can result in an uncontrolled afterburn and possible equipment damage. A negative pressure differential across the spent catalyst slide valve can allow air to back-flow from the regenerator into the reactor with equally disastrous consequences. To protect the reactor and the regenerator against a flow reversal, pressure differential controllers are used to monitor and control the differential pressures across the slide valves. If the differential pressure falls below a minimum set-point, the pressure differential controller (PDIC) overrides the process controller and closes the valve. Only after the PDIC is satisfied will the control of the slide valve return to the process.
Unit Monitoring and Control
179
To maximize the unit's profit, one must operate the unit simultaneously against as many constraints as possible. Examples of these constraints are limits on the air blower, the wet gas compressor, reactor/regenerator temperatures, slide valve differentials, etc. The conventional regulatory controllers work only one loop at a time and they do not talk to one another. A skilled operator can "push" the unit against more than one constraint at a time, but the constraints change often. To operate closer to multiple constraints, a number of refiners have installed an advanced process control (APC) package either within their DCS or in a host computer. The primary advantages of an APC are: * It provides more precise control of the operating variables against the unit's constraints and, therefore, obtains incremental throughput or cracking severity. * It is able to respond quickly to ambient disturbances, such as cold fronts or rainstorms. It can run a day/night operation, taking advantage of the cooler temperatures at night. * It pushes against two or more constraints rather than one single constraint. It can maximize the air blower and wet gas compressor capacities. As mentioned above, there are two options for installing an APC. One option is to install an APC within the DCS framework, and the other is to install a multivariable modeling/control package in a host computer. Each has advantages and disadvantages, as indicated below, Advantages of Multivariable Modeling and Control The multivariable modeling/control package is able to hold more tightly against constraints and recover more quickly from disturbances. This results in an incremental capacity used to justify multivariable control. An extensive test run is necessary to measure the response of unit variables. In APC on DCS framework, the control structure must be designed, configured, and programmed for each specific unit. Modifying the logic can be an agonizing process. Wiring may be necessary. It is difficult to both document the programming and to test.
180
Fluid Catalytic Cracking Handbook
With a host computer framework, the control package is all in the software. Changing the program can still be agonizing, but the program can be tested off-line. There is more flexibility in the computer system, which can be used for many other purposes, including on-line heat and weight balances. Disadvantages of Multivariable Modeling and Control A multivariable model is like a "black box." The constraints go in and the signals come out. Operators do not trust a system that takes the unit away from them. Successful installations require good training and continual communication. The operators must know the interconnections in the system. The model may need expensive work if changes are made during a turnaround. If the feed gets outside the range the unit was modeled for, results can be at best unpredictable. An upset can happen for which the system was not programmed. The DCS-based APC is installed in a modular form, meaning operators can understand what the controlled variable is tied to more easily. The host computer-based system may have its own problems, including computer-to-computer data links. In any APC, the operators must be educated and brought into it before they can use it. The control must be properly designed, meaning the model must be configured and properly "tuned." The operators should be involved early and all of them should be consulted since all four shifts may be running the unit differently. SUMMARY The only proper method to evaluate the performance of a cat cracker is by conducting a material and heat balance. One balance will tell where the unit is; a series of daily or weekly balances will tell where the unit is going. The heat and weight balance can be used to evaluate previous changes or predict the result of future changes. As discussed in the next chapter, material and heat balances are the foundation for determining the effects of operating variables. The material balance test run provides a standard and consistent approach for daily monitoring. It allows for accurate analysis of yields and trending of unit performance. The reactor effluent can be deter-
Unit Monitoring and Control
181
mined by direct sampling of the reactor overhead line or by conducting a unit test run, The heat balance exercise provides a tool for in-depth analysis of the unit operation. Heat balance surveys determine catalyst circulation rate, delta coke, and heat of reaction. The procedures described in this chapter can be easily programmed into a spreadsheet program to calculate the balances on a routine basis. The pressure balance provides an insight into the hydraulics of catalyst circulation. Performing pressure balance surveys will help the unit engineer identify "pinch points." It will also balance two common constraints: the air blower and the wet gas compressor. Finally, process control systems allow the unit to operate smoothly and safely. At the next level, an APC package (whether within the DCS framework or as a host-based multivariable control system) provides more precise control of operating variables against the unit's constraints. It will gain incremental throughput or cracking severity.
REFERENCES 1. Davison Div., W.R. Grace & Co., "Cat Cracker Heat and Material Balance Calculations," Grace Davison Catalagmm, No. 59, 1980. 2. Hsieh, C. R. and English, A. Ar., "Two Sampling Techniques Accurately Evaluate Fluid-Cat-Cracking Products," Oil & Gas Journal, June 23, 1986, pp. 38-43.
CHAPTER 6
Products and Economics The previous chapters explained the operation of a cat cracker. However, the purpose of the FCC unit is to maximize profitability for the refinery. The cat cracker provides the conversion capacity that every refinery needs to survive. All crudes have heavy gas oils and fuel oil; unfortunately, the market for these products has disappeared. FCC economics makes the refinery a viable entity. Over the years, refineries without cat crackers have been shut down because they were not profitable. Understanding the economics of the unit is as important as understanding the heat and pressure balance. The dynamics of FCC economics changes daily and seasonally. It is dependent on market conditions and the availability of feedstocks. The 1990 Clean Air Act Amendment (CAAA) has imposed greater restrictions on quality standards for gasoline and diesel. The FCC is the major contributor to the gasoline and diesel pool and is significantly affected by these new regulations. This chapter discusses the factors affecting yields and qualities of FCC product streams. The section on FCC economics describes several options that can be used to maximize FCC performance and the refinery's profit margin.
FCC PRODUCTS The cat cracker converts less valuable gas oils to more valuable products. A major objective of most FCC units is to maximize the conversion of gas oil to gasoline and LPG. The products from the cat cracker are: • Dry Gas • LPG • Gasoline 182
Products and Economics
• « • •
183
LCO HCO Decanted Oil Coke
Dry Gas The gas (C2 and lighter) leaving the sponge oil absorber is commonly referred to as dry gas. Its main components are hydrogen, methane, ethane, ethylene, and hydrogen sulfide (H2S). Once the gas is amine-treated for removal of H2S and other acid gases, it is blended into the refinery fuel gas system. Depending on the volume percent of hydrogen in the dry gas, some refiners recover hydrogen using processes such as cryogenics, pressure-swing absorption, or membrane separation. The recovered hydrogen is often used in hydrotreating. Dry gas is an undesirable by-product of the FCC unit; excessive yields load up the wet gas compressor (WGC) and are often a constraint. The dry gas yield is primarily due to thermal cracking, metals in the feed, and nonselective catalytic cracking. The main factors that contribute to the increase of dry gas are: • Increase in the concentration of metals (nickel, vanadium, etc.) on the catalyst • Increase in reactor or regenerator temperatures • Increase in the residence time of hydrocarbon vapors in the reactor • Decrease in the performance of the feed nozzles 8 Increase in the aromaticity of the feed
The overhead stream from the debutanizer or stabilizer is a mix of C3's and C4's, usually referred to as LPG (liquefied petroleum gas). It is rich in olefins, propylene, and butylene. These light olefins play an important role in the manufacture of reformulated gasoline (RFC), Depending on the refinery's configuration, the cat cracker's LPG is used in the following areas: * Chemical sale, where the LPG is separated into C3's and C4's. The C3's are sold as refinery or chemical grade propylene. The C4 olefins are polymerized or alkylated.
184
Fluid Catalytic Cracking Handbook
* Direct blending, where the C4's are blended into the refinery's gasoline pool to regulate vapor pressure and to enhance the octane number. However, new gasoline regulations require reduction of the vapor pressure, thus displacing a large volume of C4's for alternative uses. * Alkylation, where the olefins are reacted with isobutane to make a very desirable gasoline blending stock. A Iky late is an attractive blending component because it has no aromatics or sulfur, low vapor pressure, low end point, and high research and motor octane ratings. * MTBE, where isobutylene is reacted with methanol to produce an oxygenate gasoline additive called methyl tertiary butyl ether (MTBE). MTBE is added to gasoline to meet the minimum oxygen requirement for "reformulated" gasoline. The LPG yield and its olefinicity can be increased by: * Changing to a catalyst, which minimizes "hydrogen transfer" reactions » Increasing the conversion * Decreasing residence time, particularly the amount of time product vapors spend in the reactor housing before entering the main column » Adding ZSM-5 catalyst additive An FCC catalyst containing zeolite with a low hydrogen transfer rate reduces resaturation of the olefins in the riser. As stated in Chapter 4, primary cracking products in the riser are highly olefinic. Most of these olefins are in the gasoline boiling range; the rest appear in the LPG and LCO boiling range. The LPG olefins do not crack further, but they can become saturated by hydrogen transfer. The gasoline and LCO range olefins can be cracked again to form gasoline range olefins and LPG olefins. The olefins in the gasoline and LCO range can also cyclize to form cycloparaffins. The cycloparaffins can react through H2 transfer with olefins in the LPG and gasoline to produce aromatics and paraffins. Therefore, a catalyst that inhibits hydrogen transfer reactions will increase olefinicity of the LPG, The conversion increase is accomplished by manipulating the following operating conditions: * Increasing the reactor temperature. Increasing the reactor temperature beyond the peak gasoline yield results in overcracking
Products and Economics
185
of the gasoline and LCO fractions. The rate of production and olefinicity of the LPG will increase, Increasing feed/catalyst mix zone temperature. Conversion and LPG yield can be increased by injecting a portion of the feed, or naphtha, at an intermediate point in the riser (see Figure 6-1). Splitting or segregation of the feed results in a high-mix zone temperature, producing more LPG and more olefins. This practice
30% of Feed
70% of Feed
Figure 6-1.
A typical feed segregation scheme.
186
Fluid Catalytic Cracking Handbook
is particularly useful where the reactor temperature is already maximized due to a metallurgy constraint. » Increasing catalyst to oil ratio. The catalyst to oil ratio can be increased through several knobs including: reducing the FCC feed preheat temperature, optimizing the stripping and dispersion steam rate, and using a catalyst that deposits less coke on the catalyst, Reduction of the catalyst/hydrocarbon time in the riser, coupled with the elimination of post-riser cracking, reduces the saturation of the "already produced olefins" and allows the refiner to increase the reaction severity. The actions enhance the olefin yields and still operate within the wet gas compressor constraints. Elimination of post-riser residence time (direct connection of the reactor cyclones to the riser) or reducing the temperature in the dilute phase virtually eliminates undesired thermal and nonselective cracking. This reduces dry gas and diolefin yields. Adding ZSM-5 catalyst additive is another process available to the refiner to boost production of light olefins. ZSM-5 at a typical concentration of 0.5 to 3.0 wt% is used in a number of FCC units to increase the gasoline octane and light olefins. As part of the cracking of low octane components in the gasoline, ZSM-5 also makes C3, C4, and C5 olefins (see Figure 6-2). Paraffinic feedstocks respond the most to ZSM-5 catalyst additive.
Gasoline FCC gasoline has always been the most valuable product of a cat cracker unit. FCC gasoline accounts for about 35 vol% of the total U.S. gasoline pool. Historically, the FCC has been run for maximum gasoline yield with the highest octane. Gasoline Yield For a given feedstock, gasoline yield can be increased by: • Increasing catalyst-to-oil ratio by decreasing the feed preheat temperature • Increasing catalyst activity by increasing fresh catalyst addition or fresh catalyst activity • Increasing gasoline end point by reducing the main column top pumparound rate
Products and Economics
0
5
10
15
ZSM-5 wt% in Catalyst Inventory Figure 6-2. The effect of ZSM-5 on light-ends yield [5].
* Increasing reactor temperature (if the increase does not over-crack the already produced gasoline) Gasoline Quality The Clean Air Act Amendment (CAAA), passed in November 1990, has set new quality standards for U.S. gasoline. A complete discussion of the new gasoline formulation requirements can be found in Chapter 10.
188
Fluid Catalytic Cracking Handbook
The key components affecting FCC gasoline quality are octane, benzene, and sulfur and are discussed in the following sections. Octane. An octane number is a quantitative measure of a fuel mixture's resistance to "knocking." The octane number of a particular sample is measured against a standard blend of n-heptane, which has zero octane, and iso-octane, which has 100 octane. The percent of isooctane that produces the same "knock" intensity as the sample is reported as the octane number. Two octane numbers are routinely used to simulate engine performance: the research octane number (RON) simulates gasoline performance under low severity (@600 rpm and 120°F (49°C) air temperature), whereas the motor octane number (MON) reflects more severe conditions (@900 rpm and 300°F (149°C) air temperature). At the pump, road octane, which is the average of RON and MON, is reported. Factors affecting gasoline octane are: A. Operating Conditions 1. Reactor Temperature. As a rule, an increase of 18°F (10°C) in the reactor temperature increases the RON by 1.0 and MON by 0.4. However, the MON contribution comes from the aromatic content of the heavy end. Therefore, at high severity, the MON response to the reactor temperature can be greater than 0.4 number per 18°F. 2. Gasoline End Point, The effect of gasoline end point on its octane number depends on the feedstock quality and severity of the operation. At low severity, lowering the end point of a paraffinic feedstock may not impact the octane number; however, reducing gasoline end point produced from a naphthenic or an aromatic feedstock will lower the octane. 3. Gasoline Reid Vapor Pressure (RVP). The RVP of the gasoline is controlled by adding C4's, which increase octane. As a rule, the RON and MON gain 0.3 and 0.2 numbers for a 1.5 psi (10.3 Kp) increase in RVP. B. Feed Quality 1. °API Gravity, The higher the °API gravity, the more paraffins in the feed and the lower the octane (Figure 6-3). 2. K Factor. The higher the K factor, the lower the octane.
Products and Economics
93
92 U
Z
o 91
90 20
22
24
26
Feed Gravity, "API
82
81 Q Z O
s
80
79 20
22
24
26
Feed Gravity, "API Figure 6-3. Feed gravity comparisons (MON and RON) [7].
189
190
Fluid Catalytic Cracking Handbook
3, Aniline Point. Feeds with a higher aniline point are less aromatic and more paraffinic. The higher the aniline point, the lower the octane. 4. Sodium. Additive sodium reduces unit conversion and lowers octane (Figure 6-4). C. Catalyst 1. Rare Earth. Increasing the amount of rare earth oxide (REO) on the zeolite decreases the octane (Figure 6-5). 2. Unit Cell Size. Decreasing the unit cell size increases octane (Figure 6-6). 3. Matrix Activity. Increasing the catalyst matrix activity increases the octane. 4. Coke on the Regenerated Catalyst. Increasing the amount of coke on the regenerated catalyst lowers its activity and increases octane. Benzene. Most of the benzene in the gasoline pool comes from reformate. Reformate, the high-octane blending component from a reformer unit, comprises about 30 vol% of the gasoline pool. Depending on the reformer feedstock and severity, reformate contains 3 vol% to 5 vol% benzene. FCC gasoline contains 0.5 to 1.3 vol% benzene. Since it accounts for about 35 vol% of the gasoline pool, it is important to know what affects the cat cracker gasoline benzene levels. The benzene content in the FCC gasoline can be reduced by: • Short contact time in the riser and in the reactor dilute phase • Lower cat-to-oil ratio and lower reactor temperature • A catalyst with less hydrogen transfer Sulfur. The major source of sulfur in the gasoline pool comes from FCC gasoline. Sulfur in FCC gasoline is a strong function of the feed sulfur content (Figure 6-7). Hydrotreating the FCC feedstock reduces sulfur in the feedstock and, consequently, in the gasoline (Figure 6-8). Other factors that can lower sulfur content are: • Lower gasoline end point (see Figure 6-9) • Lower reactor temperature (see Figure 6-10) • Increased matrix activity of the catalyst (text continued on page 195)
Products and Economics RONC vs. SODIUM COMMERICAL DATA
0.40
0.60
EQUILIBRIUM CAT. SODIUM, WT.%
< 80.5 §80.0 -
79.5 79.0 78.5 _
78.0
0.20
0.40
0.80
EQUILIBRIUM CAT SODIUM, WT. %
Figure 6-4.
Effect of sodium on gasoline octane [8J.
191
192
Fluid Catalytic Cracking Handbook
84 PILOT PLANT DATA
83
C5-265T/C5-129"C 265-430"F/129-221*C
82
2; 81 o S
80 79
—8
78 77 0 .0
1.0
2.0
3.0
4.0
REO, WT. % Figure 6-5. Effect of fresh REO on MON [9].
82
95
94
81
93
UJ
z 80 O O O QC
1,200°F (650°C) and Grade H, \\% chrome for Q
0
S
2 a. CO
Troubleshooting
237
o o CO
.c
CO
< ^>
tss «M
QO
Low Pressure Upstream of the Slide Valve CO Q> Q C
Low Catalyst density in the Standpipe
Insufficient pressure build-up in the Standpipe
pQ>
'> at tf}
o
ID
De-fluidization of the Catalyst in the Standpipe
Too much, too little or no Aeration Gases
Improper placement of the Aeration taps
Solutions
m O
Restriction Orifices are either plugged or improperly sized
i Make sure instrument readings are correct
Check if the Catalyst properties have changed Figured-IB.
Verify Aeration Gas flow to maximize pressure build-up
Troubleshooting catalyst circulation.
Use Rotameters instead of Restriction Oriffces(RO's)
3
Ru
High Pressure Downstream of the Slide Valve
Causes:
r
High delta P across the Overhead Condensers
High Delta P across the Reactor Ovhd. vapor line
High delta P across the Main Fractionator r
J
\^
)r
Solutions:
f
\.
High Delta P across the Riser J
\ f
v
• Adjust the Pumparound rates • Add Top or Side P/A
^.
^/
Refer to 'Coking/Fouling' Troubleshooting Section V
Figure8-1 C. Troubleshooting catalyst circulation.
J
\f
\(
• Add Fins to the Trim Coolers -19 flns per inch * Water Wash the Condensers • Reduce No. of tube passes on the water side * Check pressure drop between Fin-Fans and Trim Coolers
V
J
• Increase Fluffing Gas or Steam to the base of Riser « Replace the Curved section of the Riser
V
J
3 OQ
240
Fluid Catalytic Cracking Handbook
(text continued from page 236)
reactor, slide valves are typically operated with a 25% to 60% opening and a pressure differential of 2 psi to 6 psi (15–40 Kp). Any increase in the throughput or the conversion will either increase the opening, reduce the differential, or both. Causes of Insufficient
Circulation
A lower pressure differential across the slide valve and/or a higher than "normal" slide valve opening are common evidence of a unit's reaching its circulation limitation. The causes of low differential are • Insufficient pressure buildup upstream of the slide valve * Excessive pressure drop downstream of the slide valve Catalyst circulation can also be limited by mechanical problems with the slide/plug valves. They may have limited travel or will not open fully. This is indicated by the lack of response when an adjustment is made. The differential pressure and the flow will not change. Troubleshooting the valves is a function of the valve design and the vendor will supply troubleshooting information, Low Upstream Pressure Insufficient pressure upstream of the slide valve can be caused by: • Partial plugging in the standpipe * Too little, too much, or no aeration gas either with the catalyst entering the standpipe or along the standpipe In a properly designed standpipe, the flow of the catalyst develops a smooth and uniform static head over the entire length of the standpipe, provided the catalyst entering the standpipe is properly fluidized. This buildup of pressure head provides the necessary driving force for catalyst circulation. However, as the catalyst travels down the standpipe, it can lose fluidity due to compression of interstitial gas being carried with the catalyst. This is particularly true in a long standpipe and in low-pressure regenerators operating at low pressure. To retain fluidity of the catalyst and to maintain catalyst densities in the 35 to 45 lb/ft3 (560–720 kg/m3) range (the fluid range), many standpipes require external aeration gas to be injected into the down-flowing
Troubleshooting
241
catalyst. The correct quantity of aeration medium and the correct location of aeration taps are essential in achieving optimum catalyst density (see Figure 8-2). External aeration is not ordinarily needed in short standpipes (less than 20 feet) (6 m) because sufficient gas is usually drawn into the standpipe to keep the catalyst fully fluidized. Troubleshooting Upstream Pressure Effective troubleshooting of the circulation system requires a methodology similar to the procedures outlined in Figure 8–1. Some of the key steps are as follows: • Obtain a pressure/density profile upstream and downstream of the slide valves. • Verify any changes in catalyst physical properties. • Ensure that the correct amount of aeration gas is injected along the standpipes. One procedure is to vary the aeration flow until the maximum slide valve differential is observed. Restriction orifices with upstream pressure regulators are frequently employed to distribute aeration gas into the standpipes. The orifices ..Slip Joint Rotameter or Orifice/Pressure Regulator
Steam or Nitrogen
Figure 8-2.
Typical standpipe aeration.
242
Fluid Catalytic Cracking Handbook
are sized for critical flow so that the constant flow of aeration gas is proportional to upstream pressure regardless of changes in the downstream pressure. Once the unit is running well, it is often assumed that the aeration system is sized properly, but changes in the catalyst physical properties and/or catalyst circulation rate may require a different purge rate. It should be noted that aeration rate is directly proportional to catalyst circulation rate. Trends of the E-cat properties can indicate changes in the particle size distribution, which may require changes in the aeration rate. Restriction orifices could be oversized, undersized, or plugged with catalyst, resulting in over-aeration, under-aeration, or no aeration. All these phenomena cause low pressure buildup and low slide valve differential.
High Downstream Pressure Sometimes insufficient differential across the regenerated catalyst slide valve is not due to inadequate pressure buildup upstream of the valve, but rather due to an increase in pressure downstream of the slide valve. Possible causes of this increased backpressure are an excessive pressure drop in the "Y" or "J-bend" section, riser, reactor cyclones, reactor overhead vapor line, main fractionator, and/or the main fractionator overhead condensing/cooling system. The pressure drop in the "Y" or "J-bend" section could be from improper fluidization or a flaw in the mechanical design. There are often fluffing gas distributors in the bottom of the "Y" or along the "J-bend" that are designed to promote uniform delivery of the catalyst into the feed nozzles. Mechanical damage to these distributors or too little or too much fluffing gas affect the catalyst density, causing pressure head downstream of the slide valve. The riser pressure drop is related mainly to the catalyst circulation rate and the slip factor. Catalyst circulation rate is largely a function of the oil feed rate, the reactor temperature, and the feed temperature. Increasing the feed rate, reactor temperature, or lowering the feed temperature will increase the pressure drop across the riser. Slip factor is defined as the ratio of catalyst residence time in the riser to the hydrocarbon vapor residence time. Some of the factors affecting the slip factor are circulation rate, riser diameter/geometry, and riser velocity.
Troubleshooting
243
High pressure in the riser could also be due to insufficient fluidization gas in the base of the riser. Fluffing gas will vary the catalyst density; more fluffing gas lowers the density in the system and the backpressure on the slide valve. The pressure drop across the reactor cyclones, reactor vapor line, main fractionator, and main column overhead condensing/cooling system can be too high. The pressure drop is primarily a function of vapor velocity. Any plugging can increase the pressure drop, Troubleshooting Downstream Pressure Use the same procedure as for upstream pressure.
Erratic Circulation Erratic circulation occurs when the catalyst is not developing a smooth and uniform static head over the entire length of the standpipe. When this happens, the catalyst packs and bridges across the standpipe. Symptoms of erratic circulation include: • • • • • • • •
Severe vibration and movement of the standpipes A noise similar to train chugging Sudden loss of the pressure above the slide valve Fluctuation in the slide valve delta P Ragged reactor temperature and/or stripper level control Pressure swings in the regenerator and the gas plant Cycling of the slide valve Other instrumentation problems
Causes of Erratic Circulation Several factors contribute to erratic circulation. Included are: • A foreign object, such as a piece of refractory, partially obstructing the flow of catalyst in the standpipe. • Improper aeration—either too much or too little. • Changing catalyst properties. This can be due to changes in the fresh catalyst's physical properties and/or malfunctioning of the cyclones.
244
Fluid Catalytic Cracking Handbook
Troubleshooting Erratic Circulation To troubleshoot erratic circulation, one must: » Verify that all the instrument readings are "telling the truth" « Verify all the aeration taps are open * Make sure that neither too much nor too little aeration gas is being applied * Verify aeration gas is not wet * Verify that the fresh catalyst properties have not changed * Verify any recent design changes in the standpipes and/or catalyst hopper * Check recycling of the regenerator fines and/or the slurry recycle Figure 8-3 shows a step-by-step approach to troubleshooting erratic circulation.
CATALYST LOSSES Catalyst losses will have adverse effects on the unit operation, the environment, and operating cost. Catalyst losses appear as excessive
Problem:
Catalyst is not developing a smooth and uniform static head over the entire length of the Standpipe
Ir
Symptom:
Evidence:
Figure 8-3A.
Catalyst packs and bridges across the Standpipe
Severe vibrations and movement of the Standpipes Fluctuation in the Slide Valve delta P A chugging noise similar to "Train" noise Ragged Reactor temperature and/or Stripper level control Sudden loss of pressure above the Slide Valve Pressure swings in the Regenerator and Gas Plant Troubleshooting erratic catalyst circulation.
Troubleshooting
245
Cause:
r Improper Aeration Aeration is not adequate to maintain fluidity due to compression
Solution: > Make sure that neither too much nor too little Aeration Gas is being used
Figure 8-3B.
Catalyst is too coarse: Catalyst fines content is too low
Recycling of Regenerator fines and/or Slurry recycle Consider another Catalyst with a different particle size distribution Check Catalyst Iron Content Check properties of recent Catalyst purchase
A foreign object has partially restricted flow of Catalyst in Standpipe
• Use Gamma Ray to confirm plugging • Use high pressure Steam or Nitrogen to dislodge material
Troubleshooting erratic catalyst circulation.
Cause: Catalyst is defluidized to its bulk density; Low Regenerator Pressure
Wet Aeration Air or Steam
Solution: • Increase Regenerator pressure
Figure 8-3C.
Ensure Aeration medium is dry
Recent mechanical revisions to the Unit
Verify the design or\ Hopper and/or Standpipe Plugged vent line on external Hopper Verify amount and orientation of Aeration injection
Troubleshooting erratic catalyst circulation.
24i
Fluid Catalytic Cracking Handbook
carryover to the main fractionator or losses from the regenerator. Evidences of catalyst losses are: * An increase in the ash and BS&W content of the slurry oil * An increase in the recovery of catalyst fines from the electrostatic precipitator or the tertiary separator » An increase in the opacity of the precipitator stack gases * A decrease in the 0 to 40 microns fraction of the equilibrium catalyst or an increase in average particle size * A gradual loss of the catalyst level in the reactor stripper and/or in the regenerator
Causes of Catalyst Losses Common causes of catalyst losses include: * * * *
Changes Changes Changes Changes
in in in in
catalyst properties operating conditions the mechanical condition of the unit operating practice
Changes in the fresh catalyst's physical properties may contribute to catalyst losses. The losses could be due to the fresh catalyst's being "soft." "Softness" is evidenced by the quality of the catalyst binder and the large amount of 0–40 microns. It will increase the attrition tendency of the catalyst and thus its losses. Changes in operating parameters also affect catalyst losses. Examples are: * An increase and/or decrease in catalyst loading to the cyclones * Overloading the cyclones, even at a constant/or higher efficiency, will result in higher catalyst losses * An increase in the feed atomizing and/or stripping steam, causing catalyst attrition and generating fines * An addition of a large amount of steam to the regenerator, particularly to the torch oil nozzles, again causing catalyst attrition Often, the main cause of catalyst losses is a change in the mechanical condition of the unit. Examples are: * Trickle valves are either stuck "closed" or "open," possibly due to the hinges being warped or bound. "Warpage" could be due to
Troubleshooting
• • • • • •
• »
247
exposure to high temperatures. Erosion could be due to excessive gas leakage into the diplegs. Trickle valves have fallen off due to inadequate bracing and/or high superficial gas velocity in the regenerator. Holes have formed in the diplegs because of high cyclone velocity or external impingement. Spalled coke or refractory is lodged in the diplegs. This can be caused by improper curing or inadequate refractory supports. Cracks can form in the internal plenum, possibly due to thermal stresses. The dipleg diameter is either too small or too large. Improperly designed, eroded, or even missing restriction orifices used for steam purge or aeration nozzles could cause catalyst attrition. Catalyst attrition is also caused by broken air and stripping steam distributors. Low catalyst level in the regenerator could uncover the diplegs and allow backflow. High catalyst level can prevent the primary cyclones from draining or prevent the trickle valves from operating properly.
Troubleshooting Catalyst Losses To stop excessive catalyst losses, it should be identified whether the loss is from the reactor or the regenerator. In either case, the following general guidelines should be helpful in troubleshooting catalyst losses: « Verify the catalyst bed levels in the stripper and regenerator vessels. • Conduct a single-gauge pressure survey of the reactor-regenerator circuit. Using the results, determine the catalyst density profile. • Plot the physical properties of the equilibrium catalyst. The plotted properties will include particle size distribution and apparent bulk density. The graph confirms any changes in catalyst properties. • Have the lab analyze the "lost" catalyst for particle size distribution. The analysis will provide clues as to the sources and causes of the losses. • Compare the cyclone loading with the design. If the vapor velocity into the reactor cyclones is low, consider adding supplemental steam to the riser. If the mass flow rate is high, consider increasing the feed preheat temperature to reduce catalyst circulation.
248
Fluid Catalytic Cracking Handbook
• Confirm that the restriction orifices used for instrument purges are in proper working condition and that the orifices are not missing. • Consider switching to a harder catalyst. For a short-term solution, if the losses are from the reactor side, consider recycling slurry to the riser. If the catalyst losses are from the regenerator, consider recycling catalyst fines to the unit. Figure 8–4 is a summary of the above discussions.
Nearly every cat cracker experiences some degree of coking/fouling. Coke has been found on the reactor walls, dome, cyclones, overhead vapor line, and the slurry bottoms pumparound circuit. Coking and fouling always occur, but they become a problem when they impact throughput or efficiency.
Evidence of Coking/Fouling Coking/fouling in the reactor and the main column can be detected by: * Cavitation and/or loss of the main column bottoms pumps * Fouling and subsequent loss of heat transfer coefficient in the bottoms pumparound exchange * High pressure drop across the reactor overhead vapor line * Excessive catalyst carryover to the main column
Causes of Coking/Fouling Coke forms in the reactor and main column circuit because of: * * * *
Changes Changes Changes Changes
in in in in
operating parameters catalyst properties feedstock properties mechanical condition of the equipment
Changes in Operating Parameters The operating conditions of the unit, particularly during startups and feed interruptions, will have a large influence on the formation of coke. Coke normally grows wherever there is a cold spot in the reactor system. When the temperature of the metal surfaces in the reactor
Troubleshooting
249
Problem: Excessive Catalyst loss can cause Unit Shutdown and possible State or Federal Environmental fines
* Increase in Ash and BS&W content of Slurry Bottoms Product * Increase in flue gas Opacity * Loss of Reactor/Regenerator levels * increase in recovery of fines from Electrostatic Precipitator(ESP) * Increase in Regenerator pressure * Decrease in 0–40 microns fractions of E-Cat * Increase in 80+ microns fractions of E-Cat * Change in Catalyst Average Particle Size (APS)
Evidence:
Figure 8-4A.
Changes in Fresh Catalyst properties
« O
Fresh Catalyst is "too soft" - has low concentration of I fines
L
(0 O
ys "5 tn
• Analyze make-up for PSD & Attrition 1 Plot E-Cat properties - Look for Abnormal Peaks ' Analyze Fines for PSD Consider using a "harder" Catalyst
Troubleshooting catalyst losses.
Changes in Operating conditions j j j i
j
' Increase in Catalyst circulation rate 1 Increase in Cyclone loading 1 Decrease in Cyclone Inlet velocity 1 Increase in Cyclone Gas Outlet velocity 1 Increase in use of Atomizing Steam 1 Increase in Reactor or Regenerator bed levels 1 Water in Steam x ^ Verify Cyclone loading ' Check for any missing RO's Verify accuracy of the Regen. & Rx Catalyst levels: raise or lower Bed level Recycle Slurry to Riser or recovered fines to Regen.
Figure 8-4B.
Changes in Mechanical conditions
• Trickle Valves are either stuck closed or opened 1 Trickle Valves are either warped or eroded ' Trickle Valves have fallen off Holes in the Oiplegs
Cracks in the Plenum 1
Diplegs diameter either "too small" or "too large" Holes in Catalyst Cooler/ Steam Generator "Chunk" of Coke has fallen into Dipleg
Bumping the Regenerator
Troubleshooting catalyst losses.
250
Fluid Catalytic Cracking Handbook
walls and/or the vapor line falls below the dew point of the vapors, condensation occurs. Condensation and subsequent coke buildup are due to cooling effects at the surface. A high fractionator bottoms level, a low riser temperature, and a high residence time in the reactor dome/vapor line are additional operating factors that increase coke buildup. If the main column level rises above the vapor line inlet nozzle, "donut" shaped coke can form at the nozzle entrance. A low reactor temperature may not fully vaporize the feed; unvaporized feed droplets will aggregate to form coke around the feed nozzles on the reactor walls and/or the transfer line. A long residence time in the reactor and transfer line also accelerate coke buildup. Insufficient bottoms pumparound to the main column heat-transfer zone can also form coke. Changes in Catalyst Properties Certain catalyst properties appear to increase coke formation. Catalysts with high rare earth content tend to promote hydrogen transfer reactions. Hydrogen transfer reactions are bimolecular reactions that can produce multi-ring aromatics. Changes in Feedstock Properties The quality of the FCC feed also impacts coke buildup in the reactor internals and vapor line and fouling/coking of the main column circuit. The asphaltene or the resid content of the feed, if not converted in the riser, can contribute to this coking. Changes in Mechanical Condition of the Equipment Damaged or partially plugged feed nozzles can contribute to coke formation due to poor feed atomization. Damaged shed-trays in the bottom section of the main cloumn can cause coke formation due to non-uniform contact between upflowing vapors and downflowing liquid.
Troubleshooting Steps The following are some of the steps that can be taken to minimize coking/fouling:
Troubleshooting
251
« Avoid dead spots. Coke grows wherever there is a cold spot in the system. Use "dry" dome steam to purge hydrocarbons from the stagnant area above the cyclones. Dead spots cause thermal cracking, * Minimize heat losses from the reactor plenum and the transfer line. Heat loss will cause condensation of heavy components of the reaction products. Insulate as much of the system as possible; when insulating flanges, verify that the studs are adequate for the higher temperature. * Improve the feed/catalyst mixing system and maintain a high conversion. A properly designed feed/catalyst injection system, combined with operating at a high conversion, will crack out highboiling feeds that otherwise could be the precursors for the formation of coke. * Follow proper start-up procedures. Introduce feed to the riser only when the reactor system is adequately heated up. Local cold spots cause coke to build up in the reactor cyclones, the plenum chamber, or the vapor line. * Keep the tube velocity in the bottoms pumparound exchanger(s) greater than 7 ft/sec. Putting the bundles in parallel for more heat recovery may lead to low velocity. * Hold the main column shed tray's liquid temperature under 700°F and minimize the level and residence time of the hot liquid. Ensure adequate wash to shed decks to minimize coking in the bottom of the main column. Some paraffinic feeds may require a lower temperature. * Utilize a continuous-cycle oil flush into the inlet of the bottoms exchanger. This keeps the asphaltenes in solution and increases tube velocity. * Verify that no fresh feed is entering the main column. Feed can enter the main column through emergency bypasses or through the feed surge tank vent line. Figure 8-5 is a summary of the above discussions. FLOW REVERSAL A stable pressure differential must be maintained across the slide valves. The direction of catalyst flow must always be from the regenerator (text continued on page 254)
Unscheduled Unit interruptions loss of profit and higher maintenance costs
Problem:
Evidences:
Cavitation and/or loss of Main Column Bottoms Pumps
Fouling and loss of Heat Transfer in Bottoms Exchanger
Higher pressure drop across the Reactor Overhead vapor line
£> fi
n
2 Changes in Operating Conditions
Changes in feedstock Properties
/• Cold Spots in the Reactor • Inadequate heat-up • High Main Column Bottoms Level • Low Reactor temperature • Long Residence Time in the Reactor and Main Column • High Bottoms temperature • Low Bottoms Pumparound Rate • Cold exchanger tube wall Vtemperature J
High Level of RareEarth in the Catalyst Low Catalyst Micro Activity Test (MAT)
Figure 8-5A.
High Mole weight Asphaltene & Resins, precipitate and bind to process equip. High Levels of
Cracked Feedstock
Troubleshooting coking/fouling.
Changes in Mechanical Conditions of the Equipment Damaged or partially plugged Feed Nozzles Loss of the Shed decks Feed leaking through Bottoms exchangers or Feed Diversion Vatve
Recommendations: Properly insulate RX Overhead piping and Main Column Inlet nozzle Keep the tube velocity > 7 ft/sec Keep Main Column Bottoms temperature < 700°F Use a "dry" Dome Steam System
Remember: When it comes to Coking, doubling Residence Time is the same as increasing Bottoms temp, by 25°F
* Increase Bottoms traffic * Inject a continuous Cycle Oil flush into inlet Bottoms PA Exchangers * Install duplex filters upstream of Bottoms Pumps * Install high efficiency feed nozzles * Use 1" or larger tube diameter * Keep C7 insolubles in Slurry System less than 5% wt * Use U-tube for Bottoms Exch. * Draw more Bottoms Product * Have a spare Bottoms Exchanger bundle
Figure 8-5B. Troubleshooting coking/fouling.
254
Fluid Catalytic Cracking Handbook
(text continued from page 251)
to the reactor and from the reactor-stripper back to the regenerator. A negative differential pressure across the regenerated catalyst slide valve can allow fresh feed and oil-soaked catalyst to backflow from the riser into the regenerator. This flow reversal can result in uncontrolled burning in the regenerator and potentially damage regenerator internals, costing a refiner several million dollars in production loss and maintenance expense. Similarly, a negative pressure differential across the spent catalyst slide valve can allow hot flue gas to backflow to the reactor and the main fractionator, severely damaging the mechanical integrity of these vessels. Some of the main causes of loss of pressure differential across the slide valves are as follows: • Loss of the air blower or the wet gas compressor • Presence of water in the feed • High catalyst circulation rates resulting in excessive slide valve opening and low differential • Loss of regenerator or stripper bed levels • Failure of the reactor temperature controller and reactor-stripper level controller • Bypass open around a shutdown valve Troubleshooting flow reversal is outlined in Figure 8-6. Reversal Prevention Philosophy The FCC process is very complex and many scenarios can upset operations. If the upset condition is not corrected or controlled, each scenario could lead to a reversal. Table 8-1 contains a cause/effect shutdown matrix indicating scenarios in which a shutdown (reversal) could take place. In most cases, a unit shutdown is not necessary if adequate warning (low alarms before low/low shutdowns) is provided. The operating staff must be trained to respond to these warnings. The shutdown system will have adequate interlocks to prevent inadvertent trips. The system must include two-out-of-three voting or backup instruments. The operators must trust the system for it to remain in service.
o T3
8 o. d g-8 '» (0 3 W
« ^ O —
e2»
.ts »
35 0) CD 50
4) o
0) |
§•» II •u ** « .--
.2J
o c 0
2 Ul
T3 OJ CO C Q TO
•Mil. 8 wS © «® »*$ o. >
0)
C 0)
•o
< o £ to
I ®f:
!t P O. 1 O •
3*' CO 111 •* «jk y cb *" O
33 S < m *° "O
to -o o
Troubleshooting
o < a: K J2
TJ C °
II
ii
£ o>
00
X5 3 O
C a
•- gr © o
il!I c So o
.
2
W "JT
F e C N - 4 + S-2
Normally, a layer of FeS scale, produced in the first reaction, protects the interior of the pipe or vessel. However, in the second reaction, cyanide removes the FeS protective scale exposing more free iron to react with H2S and releasing more hydrogen. Without protection, this cycle continues until blistering, cracking, and eventual total corrosion of equipment occurs.
Causes of Higher Levels of Cyanide The primary sources of higher cyanide production are: * * * »
Higher levels of nitrogen in the feed Higher reactor temperature Operating in partial combustion mode Higher matrix activity of the catalyst
Steps to Control Hydrogen Blistering The best way to minimize hydrogen blistering is to control the corrosion rate. Both corrosion and hydrogen blistering rates can be significantly reduced by implementing the following steps (also see Figure 8-9): 1 . Install a water-wash system for dilution and removal of cyanide from the unit. Cascading wash water from the high-pressure zone back to the main column overhead or to the first-stage wet gas compressor outlet is attractive, but it is better to use overhead water and pump it from low pressure to high pressure. 2. Add polysulfide solution to neutralize the cyanide. Air can be injected into the main column overhead to make polysulfide,
£
ffi
ffl
D e
Activity •Loss of Surface Area due to Thermal and Hydro-thermal Oeact. increase in CRC •Change in Catalyst ^formulation j
•addition rate
[
>r
^t
f
^
Reduce Gasoline Olefins Figure 8-11E
Loss of Revenue Off-Spec Products
Problem:
feed nozzles •Damaged Stripper Steam Distributor J
^Trend Reactor Temp, ^ cat/oil Ratio and Dispersion Steam Rate •Check recent temp, and/or press, excursions •Verify accuracy of VReactortemo J
V
f
J
\t
»
A
•Check pressure profile around Feed Nozzles •Track H2 in Coke •Survey ttie Stripper >J
Figure 8-11 A. Troubleshooting desired product quantity and quality.
Troubleshooting
267
The decreases in microactivity and surface area are strong functions of thermal deactivation in the regenerator and the presence of metals in the feed. Operating Variables The following operating parameters lower conversion: • • • •
Decrease Decrease Decrease Decrease
in in in in
the the the the
reactor temperature catalyst-to-oil ratio atomizing steam fresh catalyst addition rate
Mechanical Conditions Damaged or plugged feed nozzle(s) and/or damaged stripping steam distributor(s) are the common causes of mechanical failures that affect "true" conversion. Note that the "apparent" conversion, as discussed in Chapter 5, is affected by the distillation cut point and main column operations. Troubleshooting Steps • Trend the feedstock properties; look for changes in the K factor, 1,050°F+ (565°C+), aniline point, refractive index, and °API gravity. The feed endpoint may have been increased to fill the unit. The conversion penalty may be a small price to pay for the increased capacity, but the penalty can be minimized. Verify that the refinery LP reflects current data on yields and product quality. • Plot properties of the fresh and equilibrium catalysts; ensure that the catalyst vendor is meeting the agreed quality control specifications. Verify that the catalyst vendor has the latest data on feed properties, unit condition, and target products. Verify the fresh makeup rate. Check for recent temperature excursions in the regenerator or afterburning problems. • Trend the reactor temperature, cat-to-oil ratio, and atomizing steam rate. Verify the accuracy of the reactor temperature thermocouple and atomizing steam flow meter. • Perform a single-gauge pressure survey around the feed nozzles. Calculate the hydrogen content of the spent catalyst. Conduct a
268
Fluid Catalytic Cracking Handbook
gamma ray scan test to verify the mechanical condition of the stripping steam distributor.
Observing a High Dry Gas Yield Dry gas yield is affected by everything that affects conversion (Figure 8-11B). Changes to increase conversion can increase the dry gas yield. High gas yield shows up as higher speed on the compressor (if centrifugal). In many cases, lower molecular weight (due to higher hydrogen content) can have the same effect. Feedstock Quality The feed parameters that increase the dry gas yield are: * Increase in nickel and vanadium content * Increase in naphthene, olefin, and aromatic concentration, which is indicated by an increase in the refractive index and decreases in aniline point and K factor Catalyst Properties The E-cat properties that increase dry gas yield are: * Increase in the level of nickel, vanadium, and sodium * Decrease in E-cat activity, surface area, fresh catalyst activity, and rare earth content * Increase in the gas and coke factors of the E-cat Operating Variables Operating parameters that increase dry gas yield are: * Increase in the reactor temperature * Increase in the regenerator temperature « Decrease in the atomizing steam * Increase in slurry or HCO recycle Mechanical Conditions Mechanical conditions that can increase dry gas yield are: » A failing reactor temperature thermocouple * Partially plugged or damaged feed nozzles
•*•"-"•
"
Ti
>•,
mt (0 (O IK
0) O O xs
0) JC
•w »P"1
>,
to .2 3 •Q CO
•Increase in Ni, V, Is •Lower Activity, Sui Area and Re levels •Increase in Gas & < Factors
^ o 53
Troubleshooting
"C O IB 0} O 0) 0) O >< CC >
K
.£ (0
o
.w S£
CO z u.
•
I
4-1
g. < g1
C
08
0) CO (6
1 "8 (w 0)
^ u.
•Track F
in
II c u w ft
to
0
to O
f -Track c
Changes in Catalyst Condition;
2Si
w
CO CD 0) -C O) O)
-Q 3 O I-
r-
T"
ffi
oo g 3 CD
H
270
Fluid Catalytic Cracking Handbook
Troubleshooting Steps The following steps should be carried out: • Track changes in feed metals content, trend the aniline point, and refractive index. • Trend changes in catalyst activity, surface area, rare earth, and metals content. Consider adding/increasing metals inhibitor. • Trend changes in the molecular weight of the gas at the firststage suction. Verify that overhead cooling and wash systems are in order. • Verify the position of the wet gas compressor spillback. Determine if the compressor turbine needs water washing. Trend the level of inert gases in the dry gas. « Calibrate the reactor temperature controller. Conduct a pressure survey around the feed nozzle piping to verify its mechanical integrity. • If no significant problems are found other than feedstock changes, verify that the refinery LP team has current data on unit yields and product quality with this feedstock. The result of troubleshooting may be that increasing dry gas may be a necessary price for changes in the feed.
Observing a Lower Gasoline Yield The FCC "true" gasoline yield largely depends on changes in feed quality, catalyst properties, operating variables, and mechanical conditions (Figure 8–11C). Feedstock Quality Paraffinic feedstocks produce the most gasoline yield (but the lowest octane). The common indicators of any increase in feed paraffinicity are: • • • •
Increase in the K factor Increase in the aniline point Increase in the nickel-to-vanadium ratio Decrease in the fraction of "cracked" material
.£ 1 J2 » .2 § S c 5 o> « S
5*1 ©O
£
050
c .2
§ _c
^\
>•,
o S > m J» c* *•
«
0 J. -J _J CO Q O X
1
o
s .1 8 § .s>
Deterrioration of Feed Nozzl
J^ ^
fd o
> 0 U.
fil ® o Q)
2o ao
*t- 0>
(0
Troubleshooting
k.
O lt!0J
•o
271
2 ® ">. o !»
«C O)
o
OT CD
o o x:
00 3 £
272
Fluid Catalytic Cracking Handbook
Catalyst Properties The fresh catalyst properties that increase gasoline yield are: « Increase in the rare earth content • Increase in the zeolite content • Increase in the unit cell size Operating Conditions The operating parameters that increase gasoline yield are: • Decrease in the feed preheat temperature and subsequent increase in the catalyst-to-oil ratio • Decrease in the carbon content of the E-cat if the carbon is greater than 0.1 wt% • Increase in the reactor temperature if overcracking is not occurring • Decrease in the ZSM-5 additive—a shift in FCC gasoline at the expense of LPG Mechanical Conditions • Deterioration of the feed nozzles » Erroneous stripper level Troubleshooting Steps • Trend the feed °API gravity, K factor, and aniline point. Verify any changes in paraffin content of the feed. • Plot the catalyst's unit cell size, rare earth, and activity. Check if there is any fluctuation in catalyst properties. • Verify the gasoline end point, vapor pressure, and LCO distillation to ensure minimum undercutting of LCO.
Observing a Low Gasoline Octane In general, any parameter that increases the gasoline yield will also decrease its octane. One reason is that the high-octane components in the gasoline tend to be denser than the low-octane components. Therefore, any change that produces more gasoline will result in a lower octane. Again, feedstock, catalyst, operating variables, and mechanical conditions play important roles in affecting gasoline octane (Figure 8-1 ID).
I
.
•N
o
9
a
o
C *»? n? C
0)0
eO -M 09
0}
_ 0)
™_^
O
1
o 2?1
CO •C
J
§1
(0 U
iin Dnditions sin mditions
u.
0)
o o>
Ganges ir Properti
^w,
/"
I
f
t3 CD Q)
-N
7
1 1 1
•
0£
|f
*
•
o
A
T3 UJ O
a. >
«
ce i.
.1
O u
E C < 0(0
SI
W
•
CO
^ i® §• |
cLJJ £
J£ S z
M
X
o£
Q h- if _l Q O
*
•£ C
1^ V
A
J
_J
0 2 Q i u. £ (O i c:5 u.
S 2 m 2 S !3 2J «
^ w M W n* ^ O w "f» u SL i " "0 ^
(0 ^
J
1 1 SI I aw .S .S ' S O o
*• fc- i_ J?'i_ ® ® o 5 ,
A
Troubleshooting
111 J
^
273
03
c!
«? eg
8
o" SW
"^v
J3
—i
Q ••"
• I
«•
^ v»
a,
S
0 K
XT
oo
0 -
o
< . o
0} x fl) (8 2>
g>
y
A
>Oir
*ls, 1
*
« o J2 o
^""
V
(
v __
S 3
T3
1 1 5 81 C
S O < o to v • • y
--o-S a ou c o i. Z , >,^ «S T Bfl-gJ
*
HOTZO Q
i
V. *
(/) C
o "o CO
274
Fluid Catalytic Cracking Handbook
Solutions:
Higher Feed K Factor Increase in the Catalyst Rare Earth content Decrease in the Catalyst Matrix Activity Larger Zeolite Unit Cell Size Add ZSM-5 Additive Higher Cat/Oil ratio Higher Mix Zone temperature
Split Feed injection Riser Quench Figure 8-11E. Troubleshooting high gasoline olefins.
Feedstock Quality Gasoline octane is increased by: • Increases in the refractive index • Decrease in the K factor and aniline point • Increase in the bromine number Catalyst Properties The fresh catalyst's chemical properties also influence the FCC gasoline octane. Gasoline octane is increased by: • Decrease in rare earth and unit cell size • Decrease in sodium content • Increase in matrix activity Operating Conditions A number of operating variables can change the octane value. The factors that increase octane are: • Increase in the reactor temperature. In general, one research octane number increase per 17°F (10°C) increase in the reactor temperature. • Decrease in the catalyst-to-oil ratio (by increasing thermal reactions).
Troubleshooting
• • • • •
275
Increase in coke content of the regenerated catalyst. Increase in the regenerator temperature. Increase in the naphtha quench or HCO recycle. Decrease in the gasoline end point. Decrease in the gasoline vapor pressure.
Mechanical Conditions The main mechanical conditions that affect octane are the type and condition of the feed nozzles. Low-efficiency feed nozzles actually increase the gasoline octane due to promotion of thermal reactions in the mix zone. High-efficiency feed nozzles improve feed/catalyst mixing and increase the gasoline yield, but decrease gasoline octane. Troubleshooting Steps • Plot the feed refractive index, °API gravity, and aniline point. Determine any shift in the amount of cracked gas oil in the feed. • Track the unit cell size, matrix activity, and rare earth content of the catalyst. • Determine if coke on the catalyst has changed. « Verify accuracy of the reactor temperature. • Check for changes in the gasoline end point and vapor pressure. • Check the conditions of the feed nozzles and amount of atomizing steam. Gasoline Vapor Pressure/Light Olefln Yield Reformulated gasoline specifications require lower vapor pressure in the blended gasoline. It also requires maximum feed to the alkylation unit. This puts more pressure on the gas plant, particularly the debutanizer. Floating the tower pressure is often the best way to meet both constraints.
This chapter highlights the common problems, symptoms, and probable causes that may be encountered in troubleshooting FCC units. In addition, a systematic approach is outlined to provide solutions and corrective action. The suggested solutions are necessarily generic but apply to a wide variety of units.
CHAPTER 9
Debottlertecking and Optimization Troubleshooting, optimization, and debottlenecking are three steps in a continuous process. There is some overlap and gray area among them. Troubleshooting refers to the solution of short-term problems. The assignment is usually initiated by operations or maintenance. The solution usually involves something that can be done online. Troubleshooting was discussed in Chapter 8. Optimization refers to maximizing feed rate and/or conversion with the existing equipment while reaching as many constraints as possible. It can be the response to changes in the feed quality, ambient conditions, or the market demands. Although it is not discussed separately here, it is the incentive for most debottlenecking projects. Debottlenecking often refers to hardware changes, small or large. It is directed at the bottlenecks identified during optimization. It includes projects that cannot be completed online, such as installing new internals in a vessel. Debottlenecking is the main focus of this chapter.
INTRODUCTION Most FCC units are big profit makers. Therefore, they are operated to several constraints. Debottlenecking is the effort to locate and overcome these constraints. The profitability of an FCC operation is maximized when the unit is "pushed" simultaneously against multiple constraints. Debottlenecking means finding the constraint or combination of constraints that cost the refinery lost opportunities and arriving at the right fix. A properly configured advanced process control (APC) system could allow for on-line, continuous optimal unit operation and push the FCC operations to multiple constraints simultaneously. The main purpose of debottlenecking is to increase the refinery's profit margin. In the FCC, this usually means: 276
Debottlenecking and Optimization
27?
• Raising the feed rate » Processing lower quality feedstocks • Reducing dry gas and coke yields, therefore, increasing total liquid products As with troubleshooting, a proper debottlenecking exercise must consider the effects of feedstock, catalyst, operating conditions, mechanical hardware, environmental issues, and the ability of the rest of the refinery to handle the additional feed/product rates and quality. APPROACH TO DEBOTTLENECKING Debottlenecking requires a comprehensive test run to determine the operation's present status. Elements of a test run include: « • • • • • •
Overall and component material balance Reactor/regenerator heat balance Hydrogen balance Sulfur balance Reactor/regenerator pressure survey Utility balance Evaluation of the interaction among feed quality, catalyst properties, and operating conditions • Main fractionator and gas plant modeling If the object of debottlenecking is to run heavier feeds, multiple test runs may be needed with heavy feed added in stages. The next step is to identify the incremental value of: » • • •
Fresh feed rate Each FCC product Octane and cetane numbers Other product quality issues (sulfur, slurry ash level, etc.)
With this information, the constraints on operation can be identified and the value of addressing them can be evaluated. Improving FCC Profitability through Proven Technologies Once the performance of the FCC unit is optimized through the use of new catalyst and operating practices, the unit's profitability can be further improved by installing proven hardware technologies. The purpose of these technology upgrades is to enhance product selectivity
278
Fluid Catalytic Cracking Handbook
and unit reliability. Since the 1980s, mechanical upgrade of FCC units has proceeded at a fast pace. New feed/catalyst injection systems and elimination of post riser reactions have been the forefront of these mechanical upgrades.
Apparent Operating Constraints The unit operating philosophy and its apparent operating limits often dictate unit constraints. For example, limitations on the main column bottoms temperature, the flue gas excess oxygen, and the slide valve delta P often constrain the unit feed rate and/or conversion. Unfortunately, some of these limits may no longer be applicable and should be reexamined. Some of them may have resulted from one bad experience and should not have become part of the operating procedure.
Debottlenecking The remainder of this chapter contains suggested ways of addressing constraints in the following areas of the FCC unit: • Feed preheat section • Reactor-regenerator section • Main fractionator and gas plant Included are discussions regarding the feed/catalyst system, instrumentation, and off-sites. It should be noted that a change in one system usually affects others.
Feed Circuit Hydraulics Figure 1-5 shows a typical feed preheat configuration. A hydraulic limitation usually manifests itself when increasing fresh feed rate and/ or installing high efficiency feed injection nozzles.
Typical Feed Preheat Section The hydraulic pinch points in the feed preheat system are identified with a single-gauge pressure survey. The bottlenecks are often related to: • Feed pumps • Fresh feed control valve
Debottlenecking and Optimization
* • » »
279
Piping Preheat exchangers Preheat furnace Feed nozzles
The feed pump will be re-rated for the new conditions. With higher viscosity and higher gravity, the pump driver may need work. If the system is not adequate, heavier feed can be piped through a separate circuit in parallel with the existing circuit, preferably on flow ratio control. If the pump is the bottleneck, before changing it, consider: * Installing a larger impeller. • (Turbine:) Increasing turbine speed. Evaluate the steam level and consider adding an exhaust condenser. * (Motor:) Changing to a variable speed drive (VSD). VSD's make startup easier and most can support 10% overspeed. • Changing the driver. * Adding pumps in parallel. • Adding a booster pump downstream. As shown in Example 9–1, increasing the pump impeller size from 13 inches to 13.5 inches increases the flow by 3.8%, discharge pressure by 7.8%, and horsepower by 12%. Increasing the turbine speed from 3,300 rpm to 3,400 rpm increases the flow by 3%, the discharge pressure by 6.1%, and the horsepower by 9.4%. Example 9-1 * Q,, h,, bhp,, d,, n, = Initial Capacity, head, brake horsepower, diameter, and speed » Q2, h2, bhp2, d,,, n2 = New Capacity, head, brake horsepower, diameter, and speed
Diameter Change Only Speed Change Only Diameter & Speed Change Q2 = Q^d,) h2 = h^cyd,)2_ bhp2 = bhPl(d2/d,)3
Q2 = Q^yn,) H2 = h^/n,)2 Bhp2 = bhp^n/n,) 3
Q2 = Q,(d2/d, x n/n,) h2= h^d/d, x n2/n})2 bhp2= bhp.Cd/d, x n/n,) 3
280
Fluid Catalytic Cracking Handbook
Given: d1 =13 in. d? = 13.5 in.
n} = 3,300 rpm n2 = 3,400 rpm
• Flow Increase 3.8% (impeller only) 3.0% (speed only) 7% (impeller and speed) • Head Increase 7.8% (impeller only) 6.1% (speed only) 14.5% (impeller and speed) • Horsepower Increase 12.0% (impeller only) 9.4% (speed only) 22.5% (impeller and speed)
New internals in the control valve or a larger control valve can be the cheapest option if no piping needs to be changed. If the pressure drop in the feed piping is excessive, consider increasing the line size or installing a parallel line. Check the existing flange ratings if any changes are made in the pump or piping, or if the temperature is changed significantly. If diluent is being added to the feed, evaluate the optimum point for minimum pressure drop and maximum heat recovery. The preheat furnace can be a bottleneck. The first consideration is that it may not be needed in the new operation. With the increase in the FCC rate, the pressure drop will increase. Consider: • Using the furnace bypass. • Verifying the position of the inlet balancing valves. When balancing a heater, operators tend to pinch the valves. At least one of the valves should be wide open. • Decoking the heater. Consider hydraulic cleaning. • Increase the number of tube passes. Changing from a two-pass to a four-pass arrangement can reduce the pressure drop by over 75% (see Example 9-2). • Adding diluent downstream.
Debottlenecking and Optimization Example 9-2 Changing Piping in Furnace from Two-Pass to Four-Pass Case I: Two-Pass Furnace 50,000 BPD total charge (25,000 BPD to each pass) °API gravity of feed = 25 Furnace outlet temp. = 500°F Furnace tube diameter (I.D.) = 4.5 in. AP 100 = 0.0216 x Where: APi(K) f p Q d
= Pressure drop (psi) per = Friction factor = 0.017 = Flowing density = 47.4 = Actual flow rate = 864 = Tube inside diameter =
100 feet of pipe lb/ft3 GPM 4.5 in.
AP!00 = 7.0 psi Assuming a total 700 ft of equivalent pipe in the furnace, the total pressure drop is 49 psi » Case II: Switching to Four-Pass AP 1(X! = 1.9 psi Assuming a total 500 ft of equivalent pipe in the furnace, the total pressure drop is 9.4 psi Saving in pressure drop = 49.0 - 9.5 = 39.5 psi or an 81% reduction
This section addresses the following:
* Mechanical limitations • Riser termination device « Feed and catalyst injection system
281
282
Fluid Catalytic Cracking Handbook
* Spent catalyst stripper « Slide valves • Regeneration
Mechanical Limitations Mechanical limitations include the design temperature and pressure of the reactor and the regenerator. Dehottlenecking the Reactor Pressure/Temperature The FCC reactor pressure is usually controlled at the suction of the wet gas compressor. The reactor pressure is the wet gas compressor suction pressure plus pressure drop through the main fractionator system. Reactor temperature is usually directly controlled by adjusting the slide valve openings or changing the pressure differential between the regenerator and reactor. Mechanical design conditions of the reactor systems can limit operating at more severe conditions. To debottleneck these limitations: * The reactor vessel can be rerated based on actual metal thickness and corrosion history at the new operating temperature. • An external cyclone can be used to unload the vessel. • Internal lining can be added. * A reactor quench system can be used. • Split feed injection can be considered. * The riser and the reactor can be replaced with a cold-wall design. Debottlenecking the Regenerator Pressure/Temperature The regenerator is already a cold-wall vessel; re-rating is not often practical. High regenerator temperature typically requires installing either catalyst coolers, operating with partial combustion, or injecting a quench stream into the riser.
Riser Termination Device (RTD) Post-riser hydrocarbon residence time leads to thermal cracking and non-selective catalytic reactions. These reactions lead to degradation of valuable products, producing dry-gas and coke at the expense of
Debottlenecking and Optimization
283
gasoline and LPG. Improvements in FCC catalyst have eliminated any incentive for these reactions. Thermal reactions are a function of time and temperature; yields are proportional to (time)*(exp[~E/RT1). Figure 9-1 shows the typical effects of vapor residence time and temperature on dilute phase cracking. For example, at 5 seconds residence time, the dry-gas yield increases 8% when the reactor temperature increases from 960°F to 980°F. Increasing the residence time to 10 seconds increases the dry gas yield another 8%. Since the mid-1980s, FCC technology licensors and a number of oil companies have employed a number of RTD's to reduce nonselective post-riser cracking reactions. Two general approaches have been used to reduce post riser cracking. The most widely used approach is direct connection of the cyclones to the riser and on to the reactor vapor line. The second approach is quenching the reactor vapors downstream of the riser-cyclones (rough-cut cyclones). RTD's separate the catalyst and the oil vapor immediately at the end of the riser. The cyclone vapor usually discharges directly to the second-stage cyclones and then to the reactor vapor line. The catalyst is directly discharged into the stripper. The "reactor" is simply a vessel for holding the cyclones. Technologies are offered by:
0
10
20
30
40
50
Residence time, sec. Figure 9-1.
Liquid loss from thermal cracking.
60
284
• • » • •
Fluid Catalytic Cracking Handbook
ABB Lummus Global (Lummus) Exxon Research & Engineering (ER&E) Kellogg Brown & Root (KBR)Mobil Oil UOP Stone & Webster Engineering Corporation (SWEC)/IFP
ABB Lummus' DCC Features ABB Lummus's RTD consists of a two-stage reactor cyclone system (see Figure 9-2). The riser cyclones (the first stage) are hard-piped to the riser. Attached to the end of each riser cyclone dipleg is a conventional trickle valve as shown in Figure 9-3. Each trickle valve has a small opening to prevent catalyst defluidization, which can be a problem, especially during start-ups. At the vapor outlet of the first-stage cyclones, an opening allows entry of stripping steam/vapors and reactor dome steam. This opening is sized to allow the second stage cyclones to be operated at a negative pressure relevant to the reactor housing pressure. Attached to the end of the upper reactor cyclone diplegs are horizontal, counter-weighted flapper valves (Figure 9-4). These valves provide a tight seal between discharging catalyst and upflowing vapors in the reactor housing. ER&E's RTD offering is principally similar to the Lummus design. In the ER&E design, the riser cyclones are not hard-piped to the riser. However, the outlet of the riser-cyclones are directly connected to the inlet of the upper cyclones. KBR Closed Cyclone System In the KBR system, as with the ABB Lummus design, the riser cyclones are hard-piped to the riser. The diplegs of both the riser cyclone and the upper reactor cyclone are often sealed with catalyst. This minimizes the carry-under of reactor vapors into the reactor housing and maximizes the collection efficiency of the riser cyclones. No trickle or flapper valves are used on the first stage. The riser cyclone diplegs terminate with a splash plate (Figure 9-4A). The upper reactor cyclone diplegs use conventional trickle valves. Sealing the upper reactor cyclone diplegs with about two feet of catalyst provides
Debottlenecking and Optimization
285
Stripper Gas
(D
be
P1>P2>P3
Figure 9-2. Lummus direct-coupled cyclone design.
insurance in case the trickle valves stick open. In this design, the riser cyclones operate at a positive pressure and sealing the diplegs minimizes carry-under of reactor vapors into the reactor housing. The catalyst must be fluidized to provide an effective seal for the diplegs. Fluidization is critical; without it, the diplegs cannot discharge the catalyst and will plug, with possible massive carry-over to the main fractionator. To ensure this uniform fluidization, the system uses an additional steam distributor.
286
Fluid Catalytic Cracking Handbook
Cyclone Dipleg*
Restraint Figure 9–3. Typical trickle valve.
Cyclone Dipleg Pivot
Counterweight
\ Figure 94.
Typical flapper valve.
In KBR closed cyclone technology, each set of riser and upper reactor cyclones is connected via the use of a "slip joint" conduit. The stripper steam and hydrocarbons, as well as dome steam, exit the reactor housing by entering through this conduit as shown in Figure 9-5.
Debottlenecking and Optimization
287
Cyclone Dipleg
05
16 o Braces (as required)
Splash Plate Figure 9-4A.
Typical splash plate.
UOP VSS System UOP's current RTD offering is the vortex separation system (VSS), as shown in Figure 9-6. VSS is for FCC units having an internal riser and a similar design (VDS) is for external risers. The catalyst-vapor mixture travels up the riser through the chamber and exits through several arms. These arms generate a centrifugal flow pattern that separates the catalyst from the vapor inside the chamber. The catalyst accumulates in a dense phase at the base of chamber, where it is "pre~ stripped" prior to flowing into the reactor stripper. The stripped hydrocarbon vapors are fully contained in the chamber and exit with the rest of the riser effluent vapors to the secondary cyclones. The reactor vapors leave the VSS through an outlet pipe. Secondary cyclones are directly connected to this outlet pipe through an expansion
288
Fluid Catalytic Cracking Handbook
Product Dome
Riser
Splash Plate
Catalst Level
Figure 9-5. KBR closed cyclone system.
joint. The VSS outlet pipe contains several vent pipes in which the reactor dome steam and a portion of the stripping steam/hydrocarbon vapors leave the reactor through these vent pipes. Stone & Webster Engineering Corporation (SWEC) SWEC offers a reactor quench system rather than a closed cyclone system. Their typical RTD is an external, rough-cut cyclone (see Figure 9-7). The vapors from the rough-cut cyclone enter the reactor vessel.
Debottlenecking and Optimization
Expansion Joint
Flapper Valve
Spent Catalyst to Stripper
Figure 9-6.
UOP vortex separation system.
289
290
Fluid Catalytic Cracking Handbook
LCO Quench
Pre-Stripping Steam
Figure 9-7.
To Catalyst Stripper
SWEC external cyclone with quench.
Debottlenecking and Optimization
291
The recovered catalyst enters the reactor via an external dipleg. Aside from external rough-cut cyclones, SWEC also offers riser-cyclones, referred to as LD2 (Linear Disengaging Device), intended to separate catalyst from reactor vapors quicker than conventional cyclones (see Figure 9–7A). LCO quench is injected into the vapors leaving the rough-cut cyclone. The reactor temperature is usually reduced to 930°F where thermal cracking is minimal. This design often requires a pre-stripping ring at the outlet of each dipleg to ensure steady catalyst discharge from the external dipleg.
To Main Column
Upper Cyclones Vapor (Catalyst)
Stripper Figure 9-7A. Webster).
SWEC LD2 riser termination device (courtesy of Stone and
292
Fluid Catalytic Cracking Handbook
Feed Nozzles Important features of a feed injection system include: • • • • • • •
Fine atomization of feed High velocity coverage of riser cross-section Intimate mixing of catalyst and oil Rapid heat transfer from catalyst to oil Instantaneous vaporization of feed Minimum catalyst back-mixing Maximum catalytic reactions while minimizing thermal reactions
A good feed injection system will produce: • Small droplet size • Efficient mixing of oil and catalyst • Complete riser coverage The feed injection system has come a long way. The early designs featured open pipes with no consideration for feed vaporization or catalyst/vapor mixing. Currently, FCC technology licensors offer many versions of feed injection systems. Figure 9-8 is a typical modern feed nozzle. In general, these nozzles incorporate some of the following design features:
Oil Inlet
Diverging Dual Slot
Target Bolt
Figure 9-8. SWEC feed nozzle.
Debottlenecking and Optimization
2§3
* Steam is used to disperse and atomize the oil/residue feed • The spray pattern of the oil/steam leaving the nozzle tips tends to be flat (fan spray) • The assembly includes multiple nozzles in a radial pattern * The nozzles are designed for a "medium" oil-side pressure drop, generally in the order of 50 psi Some of the general criteria for choosing feed injection technology include: • Total installed cost * Dispersion steam and/or lift steam/gas requirements, including flow rate, temperature, and pressure * Oil pressure requirement • Proven track-record of operational reliability The choice of the feed injection system should be based on the vendor's experience in similar units with similar feeds and on his yield projection and/or performance guarantee. However, it may be difficult to substantiate the guarantee when other changes are being made in the unit.
Spent Catalyst Stripper Spent catalyst from the reactor/cyclones discharges into the stripper. Stripping steam displaces hydrocarbon vapors entrained with the catalyst and removes volatile hydrocarbons from the catalyst. As part of optimizing the unit, the stripping steam rate should be adjusted up or down by 5%. The regenerator temperature and/or CO2/ CO ratio will be the main indicator of insufficient stripping. The test ends when there is no significant response in the regenerator temperature. In the past several years, more attentions have been given to improving mechanical performance of the reactor stripper. Proprietary stripper designs are being offered by the FCC technology licensers in attempts to improve the catalyst/steam contact.
Debottlenecking Catalyst Circulation Any attempt to increase the unit feed rate will generally require an increase in catalyst circulation. The unit pressure balance and catalyst circulation were covered in Chapter 8.
294
Fluid Catalytic Cracking Handbook
The following should be considered when debottlenecking: * * * *
Differential pressure alarm/shutdown Increasing slide valve size Standpipes Catalyst selection
Differential Pressure Alarm/Shutdown Differential pressure shutdowns are a critical part of the unit's safety system. No attempt to lower the setting on the shutdown should be made without adequate consideration. On the other hand, pressure is lost across the slide valves and costs money. Multiple, independent differential pressure alarm/shutdown switches can be installed with two out of three voting. This can satisfy the safety requirement, increase the comfort factor, and gain valuable pressure drop. Radial feed nozzles also minimize the possibility of a reversal. New valve actuators can operate more quickly and reliably, also increasing the safety factor. The test run may indicate that the slide valve is open too far. Most operators prefer to keep the valve in the 40% to 60% range. They lose a major comfort zone if the valves open more than this. A larger valve or larger port can be installed in the existing valve.
Standpipes If the unit pressure balance indicates that either the pressure gain in the Standpipes is inadequate or the delta P across the slide valves is erratic, standpipe aeration and instrumentation should be examined. Redesigning the aeration systems or replacing the Standpipes can gain valuable pressure drop. Proper instrumentation can include independent aeration flow to each tap, flow indicators/controllers on each, and differential pressure indicators between the taps. Beyond the Standpipes, the available delta P across the valve is affected by the pressure drop in other circuits. For the regenerated catalyst slide valve, downstream pressure is affected by: * Feed injection system » Riser
Debottleneeking and Optimization
2S5
• Reactor cyclones • Reactor vapor line » Main fractionator and overhead system The regenerated catalyst slide valve upstream pressure is increased by: • Increasing the regenerator bed level « Increasing the regenerator pressure • Increasing the 0 to 40 micron content of the circulating catalyst Debottlenecking Combustion Air Many FCC units are constrained by the air blower, particularly during the summer months. Air blowers are commonly designed to deliver a given volume of air. However, the heat balance demands a given weight of air (oxygen). Therefore, the amount (by weight) of air pumped by an air blower decreases with: • Increasing air blower inlet temperature • Increasing ambient relative humidity • Decreasing suction pressure As the air rate is increased, low-cost items that can be implemented to increase the flow of air/oxygen into the regenerator include: • Ensuring the air blower suction filters are clean • Ensuring the pressure drop in the suction piping is not excessive • Ensuring the pressure in the air blower discharge piping system, particularly across the check valve and air preheater, is not excessive To deliver more air consider: • Lowering the regenerator pressure • Lowering the regenerator catalyst bed level Evaluate the trade-off between the air blower capacity and wet gas compressor capacity. Spare horsepower at one can be used to unload the other. Consider: • Cooling the inlet air through the use of a chiller or suction water spray • Using portable air blowers during the hottest months • Oxygen injection • A bypass around the air heater
296
Fluid Catalytic Cracking Handbook
Other more capital-intensive modifications include installing a dedicated air blower for the spent catalyst riser. The spent catalyst riser often requires a higher back-pressure than the main air blower to deliver the catalyst into the regenerator. Therefore, less total combustion air would be available if one common blower is used to transfer spent catalyst and provide combustion air to the air distributors. Taking higher-pressure services off the main air blower can allow it to run out on the curve and deliver more air. The main air blower can also be upgraded to provide added capacity. This includes reducing seal clearance, increasing the flow passing area, and increasing the wheel tip diameter. The original equipment manufacturer (OEM) can be contacted for feasibility of this upgrade.
Regeneration Regenerator designs have changed since most units were built. If the unit test run indicates high CRC or if the catalyst will benefit from a lower CRC, the regenerator internals should be reviewed. If the data indicates wide temperature differences across the bed or afterburning, or if the unit has had some excursions, it should be examined. The regenerator review will include spent catalyst distribution, air distribution, and cyclones. If the test run with heavy feed indicates a temperature limitation, catalyst coolers, partial combustion, or riser quench should be considered.
FLUE GAS SYSTEM The FCC is usually constrained by environmental permits. If the unit undergoes significant expansion, it may lose "grandfather" protection. The environmental limits include the amount of coke burned in the regenerator and emission rates of particulates, SOx, NOX, and gasoline sulfur. Increasing the feed rate or running heavier crude can increase all of these emissions. The technology for control is discussed in Chapter 10.
FCC CATALYST The FCC catalyst's physical and chemical properties dictate how much feed can be processed. Chemical properties, such as rare earth
Debottlenecking and Optimization
2§7
and unit cell size (UCS), affect the unit heat balance and wet gas compressor loading. Physical properties, such as particle size distribution and density, can limit catalyst circulation. Consider reformulating the catalyst—custom formulations are available. Increasing rare-earth content can reduce the wet gas rate. Catalyst is usually selected for properties other than its ability to flow. However, if it does not flow, it is not going to work well. Catalyst physical properties should be compared with those of catalysts that have circulated well. Evaluate the economics of using metal passivation additives and other catalyst enhancing additives.
Debottlenecking Main Column and Gas Plant Debotflenecking usually results in more feed. Both the main fractionator and the gas plant must be able to recover the incremental product. The main fractionator can be limited by several factors including: • Heat removal limitations » Tray flooding • Fouling and coking Heat removal can be limited by several factors including: • • • • •
Fixed reboiling duties in the gas plant Lack of heat exchanger in the pumparound circuits Jet or liquid flooding in one or more sections of the main fractionator High bottoms temperature leading to fouling or high LCO endpoint Overhead condensing capacity
Moving heat up the tower improves fractionation by increasing the vapor-liquid traffic. This is limited by flooding constraints and excessive temperature in the bottom. One method of maximizing the LCO end point is to control the main fractionator bottoms temperature independent of the bottoms pumparound. Bottoms quench ("pool quench") involves taking a slipstream from the slurry pumparound directly back to the bottom of the tower, thereby bypassing the wash section (see Figure 9-9). This controls the bottoms temperature independent of the pumparound system. Slurry is kept below coking temperature, usually about 690°F, while increasing the main column flash zone temperature. This will maximize the LCO endpoint and still protect the tower.
298
Fluid Catalytic Cracking Handbook
HCO
Reactor Vapors
P/A
Pool Quench
TC J-^
Prod
Slurry —> Recycle
Figure 9-9.
Pool quench to main column bottoms.
If the main fractionator bottoms temperature is limited to 690°F, adding a "pool quench" can provide additional LCO product recovery. Assuming there is no penalty for the bottoms product quality and there is available cooling capacity in the upper section of the fractionator, this incremental LCO yield is valuable. If flooding occurs in the main fractionator, increasing the bottoms pumparound rate reduces vapor loading, but can have a negative affect on fractionation. Normally, the economic incentive is to maximize the fresh feed rate and/or conversion, sacrificing the bottoms cut-point and rate. Increasing conversion by 1.5% (through increasing the riser top temperature by 10°F), provides an incremental profit even though LCO is lost to bottoms. Either high-capacity packing and/or high-efficiency, high-capacity trays can be installed. Trays in the bottoms wash-section can be replaced with grid or packing. The packing has greater capacity at lower pressure drop. The typical packed column has one or more beds, each consisting of packing, a support plate, a hold-down support plate, and a liquid distributor.
Debottlenecking and Optimization
299
In a packed column, liquid and vapor flow counter-currently and separation between the liquid and vapor phases takes place continuously. In contrast, in a column with trays, separation occurs in stages. In a packed column, vapor does not bubble through the liquid as in the columns with trays. For this reason, and due to the absence of the vapor-flow orifices, packed columns operate at a much lower pressure drop. In addition, because liquid and vapor contact in a packed column is less agitated than in a trayed column, packed columns are less likely to foam. Satisfactory operation must be between the upper and lower limits for both liquid and vapor flow rates. At liquid rates below 0.5 GPM per square foot of packing cross-section, liquid distribution is not uniform enough to ensure thorough wetting. At liquid rates between 25 GPM and 70 GPM per square foot of packing, the column is considered liquid-loaded and becomes very sensitive to additional liquid or vapor flow. An adequate vapor rate produces a pressure drop greater than 0.1 inch of liquid per foot of packing. Flooding occurs when the pressure drop exceeds 1.3 to 2.5 inches of liquid per foot of packing. At high vapor rates, the liquid cannot flow down the column. The liquid distributor is the most important internal structure of a packed column. The distributor strongly influences packing efficiency. It must spread the liquid uniformly, resist plugging/fouling, provide free space for gas flow, and allow operating flexibility. Packed columns can flood prematurely. Some of the reasons include: • Fouling (caused by precipitation, lodgment of loose material and debris damaged packing) • Foaming • Improper feed introduction • Restricted liquid outlet In addition to changing to packing or high-efficiency trays, the tower can be unloaded by: • Removing more heat from the pumparound returns, either by generating steam or adding coolers. This can decouple the fractionator from the reboilers in the gas concentration unit. • Reviewing the LCO product system. If some or all of the LCO is being hydrotreated, that portion can bypass the stripper if it is
300
• •
• «
Fluid Catalytic Cracking Handbook
direct-fed to the other unit through pressure vessels. Stripping is difficult to justify and sends wet feed to the unit. Changing the control system so stripping steam flow is proportional to LCO stripper product. Reviewing the overhead water wash: most overhead condensers are washed continually to minimize fouling. Since multiple bundles are common, solenoids and a PLC can be used to wash one bundle at a time, for approximately ten minutes each. This can lower the pressure drop and increase the available cooling with minimal impact. Advanced instrumentation can be used. If the rich oil is being returned from the secondary absorber, consider different processing.
Dehottlenecking Wet Gas Compressor A portion of liquid from the overhead receiver is refluxed back to the tower and the remainder is pumped on to the gas plant. The vapor from the receiver goes to the wet gas compressor. The pressure of the reactor/main fractionator system is usually controlled at the compressor suction. Improving overhead cooling will increase the wet gas compressor capacity. Excessive pressure drop or limited cooling in the overhead system decreases the capacity. This can result from: • • » «
Inadequate surface area Uneven distribution of hydrocarbon vapors and/or cooling water. Corrosion and salt deposition Limited water flow rate from elevated water coolers (consider adding a booster pump at grade) • Rapid fouling caused by water outlet temperature above 125°F The wet gas compressor is always run to a limit, therefore, increasing the available flow will always benefit the unit. The flow can be increased by: • Parallel cooling of the overhead vapor. The pressure drop across overhead cooling systems ranges from 2.0 psi to more than 10.0 psi; 5.0 psi is typical. • On-line solvent or water wash to minimize blade fouling on both the compressor and turbine.
Debottlenecking and Optimization
301
• Closing the spillback valves. • Removing external streams. If gas comes from another unit or vents from a column in the gas concentration unit, consider routing it to the interstage rather than the suction. The refinery needs to evaluate if external streams are worth recovering or whether they can be routed elsewhere. • Installing an advanced surge control system. • Verifying that the flow rates of corrosion inhibitor and antifoulant are adequate for the new operating conditions.
Improving Performance of Absorber and Stripper Columns The objective of the primary absorber/stripping towers is to maximize recovery of C3 and heavier components while rejecting C2 and lighter to fuel. C3 is first absorbed and then C, and lighter components are stripped. Although maximizing C3-C4 recovery for alkylate feed is very profitable, lower recoveries are often accepted to maximize the FCC conversion and/or feed rate. Propane/propylene recovery can be enhanced by: • Increasing the gas plant pressure. A 10 psi increase in absorber pressure increases C3 recovery by 2% (Figure 9-10). However, this can reduce the wet gas compressor capacity. Fractionation efficiency decreases as the column pressure increases. • Reducing the operating temperature. Consider adding an intercooler on the absorber. Minimize lean oil temperature. Consider the use of a chiller. Each 10°F reduction in lean oil temperature will increase C3's recovery about 0.8% (Figure 9–11). • Increasing lean oil rate. This rate is often limited by the debutanizer hydraulic and reboiling/cooling capacity. A 50% increase in lean oil/off-gas ratio increases C3's recovery about 2%. • Removing water from the lean oil. Installation of water draws and/ or a coalescer can improve recovery. Water can become trapped in the tower and cause poor tray efficiencies, foaming, and premature flooding. • Minimizing over-stripping. Over-stripping can start a wheel with the absorber. A 10% cut in stripping rate can increase C3's recovery by 0.8% (see Figure 9-12). (text continued on page 304)
302
Fluid Catalytic Cracking Handbook
96
10
20
30
40
50
Delta System Pressure, psig Figure 9-10.
C3 recovery vs. system pressure.
93.0
91.0
60
65
70
75
80
Lean Oil Temperature, °F Figure 9–11. C3 recovery vs. lean oil, °F.
85
90
1
gflg
3-
3
8-
% Recoveries
304
Fluid Catalytic Cracking Handbook
(text continued from page 301)
Debottlenecking Debutanizer Operation As the gasoline Reid vapor pressure (RVP) is reduced, the operation of the debutanizer becomes more critical. The allowable vapor pressure in gasoline makes it difficult to prevent heavy ends in the alkylation feed. This can limit the production of gasoline without sacrificing alkylation. This limitation is often from insufficient overhead cooling and rebelling: * Optimum debutanizer feed preheat temperature can optimize column loading. Increasing preheat temperature reduces reboiler duty and loading in the stripping section of the tower. Decreasing preheat temperature decreases overhead condensing duty and loading in the rectifying section. Adding an exchanger on the stripper bottoms can make this a controllable variable. * Delta P indicators should be installed on both the top and bottom section. * Optimize the operating pressure to balance reboiling, condensing, and loading. Consider floating pressure control. With tightening vapor pressure specifications, the debutanizer is an excellent candidate for this type of control. Floating pressure will unload the tower and provide better separation. * If slurry pumparound is the heat medium, consider HCO pumparound to minimize fouling. * Revamp the tower internals with high-capacity trays or packing. * If the receiver vent is in continuous service, route it back to the wet gas compressor interstage rather than to the suction. Consider adding a chiller on the vent gas. INSTRUMENTATION Additional analyzers should be considered. Temperature and pressure are no longer adequate to control distillation columns to tight specifications. Consider chromatographs on the overhead streams. One chromatograph with multiple sample streams can be adequate for most services. Ensure that qualified service is available locally.
Debottlenecking and Optimization
305
If the unit does not have a distributed control system (DCS), a debottlenecking project is the right time to justify it. If it does have a DCS, advanced control projects should be justified. A DCS: • Will provide better control of the unit and stay closer to constraints. Operating closer to constraints is what optimization and debottlenecking are all about. • Has trending and reporting ability. Data can be dumped to a spreadsheet program and variables plotted against one another. • Is a valuable troubleshooting tool. • With a host computer allows moving on to advanced control and rnulti-variable control. The unit is sensitive to day/night temperature swings and the multi-variable control can track ambient changes. Many case histories are available on converting to a DCS on the run or during a turnaround. Upgrading will pay off in the long ran.
Tankage/Blending Significant debottlenecking in the FCC will affect the tankfarm and blending system. They will handle increased product yields and changes in the quality. Blending needs maximum warning about changes in gasoline components. Steam/BFW Adding a catalyst cooler may back a boiler down, or it may require more BFW and a home for the steam. New feed nozzles may require more steam. A cogeneration unit can be an attractive option. Sour Water/Amine/Sulfur
Plant
Running heavier crude to the FCC will convert more of the sulfur in the refinery crude to H2S. Relief System Increasing the wet gas compressor capacity and increasing duties through the gas plant can impact the flare system.
306
Fluid Catalytic Cracking Handbook
Fuel System Both offgas rate and composition may change. Verify that increased hydrogen content will not impact any heaters. Depending on the header design, it could be a problem if it all goes to the same branch of the header. With the nation's projected natural gas shortfall and projected price increases, this may be a good time to consider gas export.
SUMMARY Cat cracking has been, and will continue to be, a big "money maker" for the refining industry. It is unlikely that any new cat crackers will be built (especially in the U.S.) in the near future. Therefore, emphasis will be placed on finding ways to improve the operational reliability and profitability of the existing FCC's. Performance of an FCC unit is often maximized when the unit is operated against multiple constraints simultaneously. It is essential that the specified constraints allow for minimum "comfort zones." An operator-friendly advanced control program, coupled with proper selection of catalyst formulation, would allow optimizing the performance of the unit on a daily basis. This chapter provided several cost recommendations that, once implemented, would provide cost-effective added value to the operation of the FCC. Examples of such items include tips on debottlenecking the air blower, wet gas compressor, and catalyst circulation. This chapter also discussed the latest technologies regarding the riser termination devices, as well as feed injection systems. Prior to implementing any new technologies, it is critical that the objectives and the limitations of the unit are clearly defined to ensure the expected benefits of the new technology are realized. The selected technology must match the mechanical limitations of a given cat cracker. All the technologies discussed in this chapter have been commercially proven, therefore the choice must include the total installed costs, as well as the projected benefits to the refinery.
CHAPTER 10
Emerging Trends in Fluidized Catalytic Cracking Although the demand for transportation motor fuels in North America is projected to be limited, economic growth in other parts of the world will require crude oil-based fuels. The Far East, Latin America, and the former Soviet Union are areas where there will be substantial demand for transportation fuels. The collapse of communism, the privatization of state-owned oil companies, and the global awareness of "environmentally clean fuels" will cause this growth. In the coming years, the refining industry will be experiencing major challenges. In the United States, refiners are faced with excess refining capacity, projected slow growth, and high capital and operating costs to comply with environmental health and safety regulations. The oil industry in general, and the refining industry in particular, are technologically sophisticated. They have a long history of innovations and proven track records in responding to challenge. It is likely that the reliable crude oil supply will not diminish any time soon. Petroleum-derived fuels will remain the primary source of transportation energy for well into the twenty-first century. Producers and refiners have been, and will be, environmentally responsible. The existing infrastructure of advanced product distribution systems can compete with alternative fuels readily. Future fuels will be competitive, both economically and environmentally. New global market conditions will dictate closure of inefficient facilities and investment in new technology. Larger and more efficient operations will survive and will focus on the "niche market." In the U.S., the crude processing capacity is expected to increase modestly, at a projected rate of 0.5 percent per year. No new refinery
307
308
Fluid Catalytic Cracking Handbook
is expected to be built in the U.S. The production of lighter, higher value products is expected to continue. Residual fuel will continue to decline. The demand for gasoline is projected to be stable with excess octane. Optimum performance and reliability of FCC units will play an important role in the competitiveness and survival of refineries. The FCC has proven to be a versatile process, changing to meet the needs and demands of refiners. As one of the most efficient conversion processes in the refinery, it will continue to play a key role in meeting future reformulated fuel demands. This chapter discusses: • « • •
Evolution of reformulated fuel and its impact on FCC operations Resid upgrading through the FCC Gaseous emissions from the FCC Emerging developments in catalyst, process, and hardware
REFORMULATED FUELS The passage of the Clean Air Act Amendment (CAAA) on November 15, 1990, started a process for regulating the composition and quality of gasoline and diesel fuels sold in the United States. The CAAA's intent was to improve the nation's air quality by reducing ozone and other air pollutants. Title II of the CAAA requires the manufacture and sale of "cleaner" fuels in order to reduce evaporative and combustible emission of: 1. Volatile organic compounds (VOCs) 2. Nitrogen oxides (NOX) 3. Toxins including benzene, formaldehyde, acetaldehyde, 1,3 butadiene, and polycyclic organic material (POM).
VOC Emissions VOCs can be emitted from the fuel system and from the exhaust system. Fuel system: Evaporative emissions of gasoline are mainly due to the presence of butane and the low-boiling light olefins (C4 and C5). Reducing gasoline vapor pressure and removing these olefins can limit the amount of evaporative emissions. Light olefins are photo-chemically reactive; removing them will improve ozone.
Emerging Trends in Fluidized Catalytic Cracking
309
Exhaust system: The engine operating mode controls the tailpipe emissions of hydrocarbons (HC) and carbon monoxide (CO). Over 80% of HC and CO emissions are generated during cold-start and warm-up due to incomplete combustion. Fuel vaporization and fuel/ air mixing are important factors in achieving thorough combustion of the hydrocarbons. Gasoline can be modified to vaporize quickly. This is accomplished by: • Decreasing the end point or 90% boiling point « Reducing the aromatic content * Adding oxygenates Since a gasoline engine burns vaporized fuel, the heavy end of the fuel contributes to its partial vaporization in a cold engine. Reducing the 90%-point or the 50%-point temperature will reduce HC emissions in the engine exhaust. Aromatic levels and carbon content of the gasoline also have a significant effect on the tailpipe emissions of HC and CO. Because of their high heat of vaporization and high boiling point (see Figure 10-1), aromatics do not vaporize readily. This is an incentive to minimize aromatics.
2.5 c •3 CO
1
X
'
1
i
.
1.0 ^Others
P
JP"p
.P
•« -Op..
1
0.5
0
-
\
\
\
Aromatics *A XV^ y •*»^ Vx "*^*.^__
.
I
I
\ Alcohols
8-3.1.5
<s
^x
2.0
•8o o* si o
!
___
A""**A"At». c r\
p. p.. op
_:
""*«*._
P
P
t 1 1 f i 100 200 300 Boiling point, °F
-
-
1 400
X = alcohols, A = aromatics, P = paraffins, O = olefins and E = ethers. Figure 10-1. Heat of vaporization versus boiling point [16].
310
Fluid Catalytic Cracking Handbook
Oxygenates reduce CO emissions by "enleaning" the fuel-to-air mixture. Enleanment of the fuel with oxygenates has the most impact on CO emissions. However, oxygenates, particularly ethers, are often used as a "substitute" and can replace aromatics in achieving octane specifications. Reducing aromatics further reduces CO and HC emissions, NOX Emission Direct exposure to NOX can cause respiratory problems. VOCs and NOX are catalyzed by sunlight to form ground level ozone, often referred to as smog. NOX can be generated from either stationary sources or mobile sources. In 1997, the EPA changed the ambient standard for ozone from 0.12 ppm to 0.08 ppm and the applicable test period was increased from 1 to 8 hours. The main stationary sources of NOX are gas turbines, fired heaters, power generation plants, and, of course, the FCC. The amount of NOX produced is a function of residence time and combustion temperature. Combustion temperature is influenced by fuel composition. The main mobile source of NOX is the combustion of a fuel in an internal combustion engine. Because aromatics have the highest combustion temperature among hydrocarbon types (see Figure 10-2), they
i
i
u- 4,200 -
£ I 4,000~-o
Aromatics Oiefins u
1 3,800
1 ig 3,600
'*~
Others x* Alcohols /x
j 0
""""""•* A«^,,Aii_ • ~ * " " "" • "o™*" * *"»O.
*
i
100 200 Boiling point, °F
300
Theoretical flame temperatures assuming adiabatic and stoichiometric air. Figure 10-2. Flame temperature of aromatics, olefins, paraffins, and alcohol [16].
Emerging Trends in Fluidized Catalytic Cracking
311
tend to produce higher amounts of NOX in the exhaust gases than olefins and paraffins. One might think that because oxygenates have lower combustion temperatures, they will generate less NO X . However, because the enleanment effect raises combustion temperature, oxygenates actually increase NOX emissions. Consequently, some compromise may be needed with respect to oxygen content of fuels and its effect on HC, CO, and NOX emissions.
Benzene Emission Benzene is a known carcinogen. The U.S. Environmental Protection Agency (EPA) has identified benzene as a toxic air pollutant (TAP). Benzene is present in automotive evaporation, refueling vapors, and exhaust. The control of benzene emissions from the fuel system includes limiting the amount of benzene in the fuel and vapor recovery at fuel stations. Exhaust benzene is a function of aromatics and benzene content. Exhaust benzene emission is calculated by: EXB = 1.884 + 0.949 x BZ + 0.113 x (A - BZ) Where: EXB = Exhaust benzene, milligram/mile BZ = Benzene, vol% A = Aromatics, vol%
For example, the exhaust benzene for a gasoline having 30 vol% aromatics and I vol% benzene is = 6.11 milligram/mile.
CAAA Regulations As of November 1, 1992, all gasoline sold in the 39 CO nonattainment areas contained 2.7 wt% oxygen during the winter months. Beginning January 1, 1995, regulations mandated that gasoline sold in the nine worst ozone non-attainment areas contain at least 2.0 wt% oxygen and not more than 1 vol% benzene and 25 vol% total aromatics. Other cities that have had mobile source emission problems can "optin" voluntarily to the use of reformulated fuels.
312
Fluid Catalytic Cracking Handbook
From 1995 through 2000, a 15% reduction in VOCs and other air toxins was expected in the non-attainment areas. Beginning in the year 2000, an additional reduction of 25% is required unless the EPA determines it too costly or not feasible. Even in that case, the reduction will not be less than 20%. Oxygen was added as oxygenated hydrocarbon components: methyl tert-butyl ether (MTBE), tert-amyl methyl ether (TAME), ethyl tertbutyl ether (ETBE), di-isopropyl ether (DIPE), ethanol, methanol, and tertiary butyl alcohol (TBA). The properties of oxygenates, as they relate to gasoline blending, are shown in Table 10–1. The key points of the regulations require: * The certification of fuels: Each refiner, blender, or importer of gasoline must ensure that per-gallon emissions levels of VOCs, NOX, CO, and toxic air pollutants do not exceed the gasoline sold in 1990. • Effective October 1993, highway diesel fuel was limited to a maximum sulfur content of 0.05 wt% (500 ppm) and a minimum cetane rating of 40. « Engine additives were required in all gasoline to prevent deposits in engines and fuel supply systems. » Lead and lead additives were eliminated by the end of 1995.
Table 10-t Oxygenates Properties MTBE
Blending octane (R + M)/2 Blending RVP, psi Boiling point, °F Density @ 60°F, Ib/gal Water solubility, wt% Max. concentration, vol% Max. oxygen, wt%
ETBE TAME TBA
Ethanol
Methanol
110
111
105
100
115
108
8 131 6.2
4 161 6.2
1 187 6.4
181 6.6
18 173 6.6
31+ 148 6.6
1.4 15.0
0.6 12.7
12.4 16.1
10
9.7
2.7
2.0
2.0
3.7
3.7
3.7
Emerging Trends in Fluidized Catalytic Cracking
313
The standards for conventional or non-RFG gasoline are shown in Table 10-2. Oil companies can choose to comply with the requirements on either a per-gallon baseline or the company's 1990 baseline. If there is an incremental volume of fuel above the 1990 production rate, the baseline will be adjusted using the industry's baseline data (see Table 10-3). The industry's baseline gasoline is an average of properties of all the U.S. gasoline marketed in 1990. The Simple Model RFG required the addition of oxygenates and it limited the amount of benzene, sulfur, olefins, and T9Q. The RVP was also lowered for six months during the summer period. Given these requirements, companies can choose to comply on a per-gallon basis (Table 10-4) or adopt the 1990 industry average basis (Table 10-5), Starting January 1998, the EPA's Complex Model went into effect. The Complex Model provides a set of equations that predict VOC, NOX, and toxic emissions, using eight gasoline properties. These properties are RVP, oxygen, aromatics, benzene, olefins, sulfur, E200,
Table 10-2 Conventional Gasoline Standards Specifications
Properties
Exhaust benzene, mg/mile Sulfur, ppmw—yearly average Olefins, vol%—yearly average T , °F—yearly average
Maximum Maximum Maximum Maximum
100% 125% 125% 125%
Table 10-3 U.S. Industry 1990 Baseline for Non-RFG Gasoline
Aromatics, vol% Olefins, vol% Benzene, vol% Sulfur, ppmw Exhaust benzene, mg/mile T_ °F
28.6 10.8 1.6 338 6.45 332
of of of of
baseline baseline baseline baseline
314
Fluid Catalytic Cracking Handbook Table 10-4 RFC Simple Model per Gallon Standards
RVP VOC Control Region 1 (south) VOC Control Region 2 (north) Oxygen content, wt% Toxics reduction Benzene, vol% Sulfur, ppmw—yearly average Olefins, vol%
1,2 psi, maximum 8,1 psi, maximum 2.0-2.7 15.0%, minimum 1.00%, maximum 100% baseline, maximum 100% baseline, maximum 100% baseline, maximum
Table 10-5 RFG Simple Model Average Gasoline Standards (Phase I)
Oxygen content, wt% Toxics reduction Benzene, vol% Sulfur, ppmw yearly average Olefins, vol% yearly average T90, °F, — yearly average
VOC—Control Region 1 (south) Standard: 7.1 psi, max. Per-gallon: 7.4 psi, max VOC—Control Region 2 (north) Standard: 8.0 psi, max. Per-gallon: 8.3 psi, max. Standard: 2.1-2.7 Per-gallon: 1.5-2.7 16.5%, min. Standard: 0.95, maximum Per-gallon: 1.30, maximum 100% baseline, maximum 100% baseline, maximum 100% baseline, maximum
and E300. E200 and E300 are the percent of gasoline evaporated at 200°F and 300°F, respectively. The Complex Model contains the following: • Seven exhaust emission equations for VOCs and NOX, and five for toxins (benzene, butadiene, formaldehyde, acetaldehyde, and polycyclic organic material (POM).
Emerging Trends in Fluidized Catalytic Cracking
315
» Four non-exhaust emission equations for VOCs (diurnal, hot soak, running loss, and refueling emissions). * Four corresponding non-exhaust emission equations for benzene. These non-linear equations can be embedded into the refinery's linear program (LP) to achieve compliance and optimize the gasoline blend. The key FCC gasoline components that influence RFG are: • Sulfur * Benzene and aromatics • Olefins In the year 2000, Phase II of reformulation begins. The Phase II standards are shown in Table 10-5A. Compliance with the standards is determined using the Complex Model. The regulations for reformulated gasoline were published in the federal register on February 16, 1994.
Sulfur Sulfur in gasoline contributes to the SOX air quality problem and deactivates the catalyst in the catalytic converter. Emissions from a poisoned converter contain higher levels of VOC, NOX, and CO. As stated earlier, VOC and NOX are catalyzed by sunlight to form smog. Table 10-5A Complex Model Phase II Per Gallon Standards (After Year 2000) VOCX emissions performance reduction
VOC control Region 1 VOC control Region 2
> 27.5 > 25.9
Toxic emissions reduction (%)
NOX emissions reduction (%) Gasoline designated as VOC controlled Gasoline not designated as VOC controlled
> 5.5 > 0.0
Oxygen (wt%)
> 2.0
Benzene (vol%)
> 1.0
316
Fluid Catalytic Cracking Handbook
The sulfur compounds in FCC gasoline consist of Cj-C4 mercaptan and various thiophenes. The California Air Resources Board (CARB) set an average sulfur specification of 40 ppm for 1996, with a maximum of 80 ppm. The CAAA's Complex Model also addresses sulfur issues in its set of equations. For the U.S., the new EPA rules will limit sulfur in gasoline to 30 ppm, phased between 2004 and 2006. The automobile industry has made a strong case for lower sulfur because of its effect on the catalytic converter. The converter has the same catalyst as the refinery reformer and it is poisoned just as easily by sulfur. Refiners can address the sulfur issue in stages, but decisions should be made that will leave the door open for further reductions. If hydrotreating is selected, the design can include oversized reactors, connections for a spare compressor, or connections for adding amine scrubbing inside the recycle loop. Some process or catalyst changes can buy time, some can solve the problem. Reducing gasoline sulfur specifications is not limited to the U.S. In Canada, gasoline sulfur will be reduced in two phases. The first phase is a reduction from the existing level of 360 ppm to 150 by 2002. Beginning in 2005, the gasoline sulfur level will be reduced to 30 ppm. In Europe, beginning in the year 2000, the gasoline sulfur in the 15 countries in the European Union is reduced to 150 ppm. By the year 2005, the gasoline sulfur will be lowered to 50 ppm. In Japan, typical gasoline sulfur is 35 ppm, which is below the current requirement of 100 ppm. Other countries are expected to follow the lead to enact regulations to reduce gasoline levels. FCC gasoline is by far the largest sulfur contributor (up to 90%) in the gasoline pool. Typical sulfur content of the FCC gasoline ranges from 150 ppm to 3,500 ppm. The amount of the FCC gasoline in the finished gasoline blend normally ranges from 35% to 45%. Controlling gasoline pool sulfur requires reducing the sulfur content of FCC gasoline. Several options are available: • • • • •
FCC feed hydrotreating Gasoline end point reduction FCC gasoline hydrotreating Catalyst additives Bio-catalytic desulfurization
Emerging Trends in Fluidized Catalytic Cracking
317
For nonhydrotreated feed, the gasoline sulfur level is about 10% of the feed sulfur level. For hydrotreated feed, it is about 5%. For example, if the sulfur content of a nonhydrotreated feed is 1.0% (10,000 pprn), the sulfur in FCC gasoline will be 1,000 ppm. Assuming 80% desulfurization, feed to the FCC unit will contain 0.2% (2000 ppm) sulfur, resulting in FCC gasoline containing 100 ppm sulfur. Sulfur compounds that survive the hydrotreater are in the heavy fraction and tend to end up in the LCO, decanted oil, and coke. Hydrotreating or moderate pressure hydrocracking of the FCC feed provides many benefits. Besides reducing sulfur in the FCC gasoline, FCC feed hydrotreating reduces NOX and SOX emissions from the regenerator flue gas, increases conversion, increases gasoline yield, and reduces catalyst consumption. Often around 95% desulfurization is required to achieve the desired gasoline sulfur. However, a number of refiners cannot justify the high capital cost of FCC feed hydrotreating. Reducing the gasoline end point can significantly decrease the FCC gasoline sulfur (Figure 10-3). As much as 50% of the sulfur can be contained in the last 10 vol%. With a high sulfur crude mix, this end point reduction may not be sufficient to meet sulfur specifications. The disposition of this high-sulfur, high-aromatic gasoline can be a problem. One option is to combine this heavy fraction with the LCO stream and desulfurize it in the diesel hydrotreater. After hydrotreating, the heavy gasoline can be separated and sent to the gasoline pool. This may require converting the stripper into a fractionator or adding a fractionator. Selective hydrogenation (HDS) of the FCC gasoline can be a positive choice for meeting the required sulfur levels. However, deep HDS of the FCC gasoline can saturate olefins and cause octane loss. The light FCC gasoline is rich in olefin while the heavy FCC gasoline is rich in aromatics and sulfur. The choice of a proper catalyst and operating conditions is important in maximizing sulfur reduction and minimizing octane loss. A number of commercial processes are proven in this service. Caustic extraction can remove mercaptan sulfur in light fractions, but not higher carbon number mercaptans or other types of sulfur molecules that are in the FCC gasoline. Catalyst additives can reduce FCC gasoline sulfur by about 15%, They work by converting mercaptan, thiophene, etc., to H7S. A secondary benefit of the additives is an approximate 10% reduction in the LCO sulfur.
0.1600
0.0000 380
400
410
420
430
Gasoline End Point, Deg. F Figure 10-3,
Gasoline sulfur versus its end point
440
450
Emerging Trends in Fluidized Catalytic Cracking
319
Biocatalytic desulfurization (BDS) technology employs the concept of enzymatic removal of sulfur compounds without changing the fundamental structure of hydrocarbons. This process uses bacteria to selectively remove sulfur from gasoline, diesel, and diesel blend stocks. In 1997, the Coordinating Research Council (CRC) commissioned a study to assess the applicability and feasibility of biodesulfurization technology for the removal of sulfur from gasoline. It is expected that this process will be ready for commercialization within the next four to six years. Energy BioSystems Corp, The Woodlands. Texas is developing a process to use bacteria for the removal of sulfur from diesel and diesel blend stocks. The first commercial BDS unit is expected to be operational in 2001. The choice for removal of sulfur from gasoline and diesel requires examining the whole refinery and understanding its total impact. For some refiners, post-treating, or pre-treating and post-treating may be the right decision. Factors to consider include available capital, availability of tankage/utilities, and the reliability of newly installed technology,
Aromatics and Benzene Tailpipe emissions of HC and CO are affected by the levels of heavy aromatics in gasoline. Like sulfur, the heavy aromatics are in the back end of the boiling range (Figure 10-4). As with sulfur, reduction of end point directly controls the concentration of heavy aromatics in finished gasoline. The benzene content of FCC gasoline is typically in the range of 0.6 vol% to 1.3 vol%. CAAA's Simple Model requires RFC to have a maximum of 1 vol% benzene. In California, the basic requirement is also 1 vol%; however, if refiners are to comply with averaging provisions, the maximum is 0.8 vol%. Operationally, the benzene content of FCC gasoline can be reduced by reducing catalyst-oil contact time and catalyst-to-oil ratio. Lower reactor temperature, lower rates of hydrogen transfer, and an "octane catalyst" will also reduce benzene levels. Most of the benzene in the gasoline pool comes from the reformer unit (reformate). To reduce the reformate's benzene, one must modify the feedstock quality and/or operating conditions. Benzene's precursors in the reformer feed (C5 and C6) can be prefractionated and sent to an isomerization unit. The reformer operating pressure can be reduced
320
Fluid Catalytic Cracking Handbook
%'|OA '
8 ti
ui
O
Emerging Trends in Fluidized Catalytic Cracking
321
to reduce benzene and aromatics. Post-fractionation of the re formate stream is another option but it requires the installation of a reformate splitter. The light aromatics (benzene, toluene, and mix xylene) in the "light" reformate can be extracted for petrochemical feedstock. Toluene can be converted to benzene through the hydrodealkylation process. Benzene can be saturated to cyclohexane and eventually isomerized. A combination of the benzene saturation system and paraffin isomerization will enable the refiner to control benzene while improving the gasoline octane pool.
Olefins In 1990, U.S. gasoline contained about 10 vol% olefins, the majority of which emanate from FCC gasoline. FCC gasoline has 25 vol% to 35 vol% olefins. Of these olefins, C5–C7 olefins account for about 85% of the total pool. For non-RFG gasoline, as with sulfur, the regulation allows the maximum olefin content to be 125% of the 1990 baseline values. Light olefins, particularly tertiary olefins, are very reactive in forming ozone and also increase gasoline pool RVP. However, the future trend of most FCC operations is projected to produce more olefin feed, but little will reach the gasoline pool. This is because olefins, particularly the C4 and C5 olefins, can either be "alkylated" and/or "etherified," or used for petrochemical feedstock. Commercial alkylation is the reaction of isobutane with C3 through C5 olefins in the presence of either sulfuric acid or hydrofluoric acid (see Example 10-1). Etherification is the reaction of a tertiary olefin with an alcohol or water in the presence of an acidic catalyst (see Example 10-2). Example 10-1 Alkylation of Propytene and Butylene
Propylene Alkylation CH 3 — CH = CH2 + CH3 — CH — CH2 -> CH 3 — CH — CH — CH 2 —CH 3 CH3
PROPYLENE Typical Yield:
+
ISOBUTANE
CH3
CH3
-> DIMETHYLPENTANE
322
Fluid Catalytic Cracking Handbook
1.0 Volume of propylene + 1.3 volume of isobutane -> 1.80 volume of alky late.
Butyiene Alk\ lation CH3 CH3 — C = CH2 + CH3 — CH — CH2 -> CH3 — CH — CH, — CH — CH3 CH,
PROPYLENE +
CH3
CH,
ISOBUTANE
-»
CH3
2,2,4 TRIMETHYLPENTANE
Typical Yield: 1.0 volume of butylene +1.2 volume of isobutane —> 1.70 volume of alkylate.
Example 10-2 Etherification of Isobutviene
CH3 — C — CH3 + CH3 — OH CH?
-> CH3 — C — O — CH3 CH3
ISOBUTYLENE + METHANOL -> METHYL TERTIARY BUTYL ETHER (MTBE)
Typical Yield: 1.0 volume of isobutane + 0.43 volume of methanol —> 1.27 volume of MTBE.
There are etherification processes, such as MTBE and TAME, aimed at producing ethers from C5, C6, and C7 tertiary olefins. Both alkylate and ether have excellent properties as gasoline blending components. They have a low RVP, a high road octane, no aromatics, and virtually zero sulfur. The emphasis on alkylation and etherification will continue in both the U.S. and the rest of the world.
Emerging Trends in Fluidized Catalytic Cracking
323
A conventional FCC unit can be an "olefin machine" with proper operating conditions and hardware. Catalysts with a low unit cell size and a high silica/alumina ratio favor olefins. Additionally, the addition of ZSM5, with its lower acid site density and very high framework silica-alumina ratio, converts C?+ gasoline into olefins. A high reactor temperature and elimination of the post-riser residence time will also produce more olefins. Mechanical modification of the FCC riser for "millisecond" cracking has shown potential for maximizing olefin yield.
Challenges Facing RFG RFG is a cost-effective fuel that improves air quality and is a mechanism through which the refining industry can be competitive. The Complex Model is most likely here to stay. The concentration of gasoline sulfur must be reduced and the gasoline RVP will most likely be limited to about 7 psi. Nevertheless, in the years to come, numerous issues regarding RFG will be facing refiners. Most are regulatory, political, and bureaucratic issues. Following are some of these issues: « Public perception of RFG regarding health effects of ethers, price increase, and engine performance complaints * EPA's ethanol mandate and the subsequent stay of that mandate by federal court * Complexity of testing, distribution, storage, handling, and blending facilities * Record-keeping and development of a uniform certification program, * Intel-changeability of MTBE to ETBE * Interpreting the baseline * The future of opt-in areas: the continual decline in air quality where RFG is not sold * Antidumping, credits, and trading * The program length of oxygenated fuels for CO nonattainment areas * The definition of "domestic supply"
RESIDUAL FLUIDIZED CATALYTIC CRACKING (RFCC) Deterioration in the worldwide crude oil supply (Table 10-6), continual decline in the demand for heavy fuel oil, and recent mechanical and catalyst advances have provided economic incentives to
324
Fluid Catalytic Cracking Handbook Table 10-6 U.S. Crude Characteristics
Year
°APi Gravity
Wt% Sulfur
1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993
32.92 32.96 32.46 32.33 32.22 31.93 32.14 31.86 31.64 31.32 31.30
0.88 0.94 0.91 0.96 0.99 1.04 1.06 1. 10 1.13 1.16 1.15
Source: Swain [24]
upgrade the atmospheric and/or vacuum bottoms in the residual fluidized catalytic cracking (RFCC) unit. Although residue upgrading in the United States is mostly delayed coker based, most new FCC units are either residue crackers or have in-place provisions to process residue at a later date. This is more pronounced in the new units built in the Far East, Europe, and Australia. The residue from their crude oils is more paraffinic and contains less metals than North Sea or Middle Eastern crude oils, which makes them more suitable for RFCC. An RFCC is distinguished from a conventional vacuum gas oil FCC in the quality of the feedstock. The residue feed has a high coking tendency and an elevated concentration of contaminants.
Coking Tendency Residue feedstocks have a higher coking tendency, which is indicated by higher levels of Conradson carbon and a higher boiling point. The common definition of residue is the fraction of the feed that boils above 1,050°F and Conradson carbon levels greater than 0.5 wt%. The residual portion of the feed contains hydrogen-deficient asphaltenes and polynuclear compounds. Some of these compounds will lay down on active catalyst sites as coke, reducing catalyst activity and selectivity.
Emerging Trends in Fluidized Catalytic Cracking
325
Feed Contaminants The residual portion of feedstocks contains a large concentration of contaminants. The major contaminants, mostly organic in nature, include nickel, vanadium, nitrogen, and sulfur. Nickel, vanadium, and sodium are deposited quantitatively on the catalyst. This deposition poisons the catalyst permanently, accelerating production of coke and light gases. Nickel in the feed is deposited on the surface of the catalyst, promoting undesirable dehydrogenation and condensation reactions. These nonselective reactions increase gas and coke production at the expense of gasoline and other valuable liquid products. The deleterious effects of nickel poisoning can be reduced by the use of antimony passivation. Vanadium in the feed poisons the FCC catalyst when it is deposited on the catalyst as coke by vanadyl porphydrine in the feed. During regeneration, this coke is burned off and vanadium is oxidized to a V*5 oxidation state. The vanadium oxide (V2O5) reacts with water vapor in the regenerator to vanadic acid, H3VO4. Vanadic acid is mobile and it destroys zeolite crystal through acid-catalyzed hydrolysis. Vanadic acid formation is related to the steam and oxygen concentration in the regenerator. Vanadium and sodium neutralize catalyst acid sites and can cause collapse of the zeolite structure. Figure 10-5 shows the deactivation of the catalyst activity as a function of vanadium concentration. Destruction of the zeolite by vanadium takes place in the regenerator where the combination of oxygen, steam, and high temperature forms vanadic acid according to the following equations: 4V + 5 O2 --» 2 V2O5 V2CL + 3 H2O -* 2 VO (OH)3
The produced vanadic acid, VO (OH)3, is mobile. Sodium tends to accelerate the migration of vanadium into the zeolite. This acid attacks the catalyst, causing collapse of the zeolite pore structure. The presence of increased basic nitrogen compounds, such as pyridines and quinoline in the FCC feedstock, also attack catalyst acid sites. The result is a temporary loss of catalyst activity and a subsequent increase
326
Fluid Catalytic Cracking Handbook
68
^>T~^ - ^ T
67
* *
i ^***^***«*^
"
I
C'C^L
^'rC? N „
>7, ^^^^S-^4
J? 65 '> 1 64
"j-. j Q $^ ^"""P"-^
ra
o 63 5 6?
^T*—^^
I
'~~\?°-o . I * ^ sit^O-.
i "''"i. f
h
t
61
~' ~ '"-*.' '"ft "
~~1
60
0
1000
2000
3000
4000
5000
6000
Vanadium, ppm Figure 10-5. Vanadium deactivation varies with regenerator severity {25],
in coke and gas yields. Additionally, in the regenerator, some of the adsorbed nitrogen is converted to nitrogen oxide (NOX). Although an increase in the sulfur content of the residue feedstock will have a minimal effect on unit yields, the sulfur content of the RFCC products and the flue gas is greater, requiring additional treating facilities.
Operational Impacts of Residue Feedstocks In the unit, residue feedstocks have the following effects: * Higher delta coke and coke yield, which are associated with residue feedstocks, will result in elevated regenerator temperature and higher combustion air requirements. * Exposure of the catalyst to a variety of feed contaminants and the higher regenerator temperature will reduce both selectivity and activity. * Greater levels of nitrogen and sulfur in the residue feed increase emissions of NOX and SOX from the regenerator.
Emerging Trends in Fluidized Catalytic Cracking
327
Minimizing Detrimental Effects of Processing Residual Feeds The proper choice of a feed injection system, regenerator, and catalyst are some of the key aspects of successful RFCC operation. An efficient feed injection system produces extremely small droplets that vaporize quickly. Rapid vaporization minimizes the amount of non-vaporized hydrocarbons that block the active sites. An effective feed nozzle system must instantaneously vaporize and crack asphaltenes and poly nuclear aromatics to lower boiling entities. The regenerator design, either single-stage or two-stage, should provide uniform catalyst regeneration, increase flexibility for processing a variety of feedstocks, and minimize thermal and hydrothermal deactivation of the catalyst. The catalyst design should be optimized to achieve the following objectives: » • • • •
Low coke and gas production Efficient bottoms cracking Improved metals resistance Improved thermal and hydrothermal stability An active matrix and a low hydrogen transfer activity to convert the bottoms and minimize delta coke
REDUCING FCC EMISSIONS The gaseous emissions from the FCC unit are CO, NOX, particulates, and SOX, All are either locally or nationally regulated. Table 10-7 shows the current allowable limits of the EPA New Source Performance Standards (NSPS) for the emissions of these airborne pollutants. NSPS levels can be triggered by one of the following conditions: • Construction of a new unit • Revamp of the regenerator, provided the modification costs are more than 50% of a comparable regenerator 8 Any capital modification of the unit that increases its emission rates There is no national requirement limiting NOX emissions from the FCC flue gas, but several state and regional agencies have imposed limits on their release. These emissions are directly proportional to
328
Fluid Catalytic Cracking Handbook Table 10-7 EPA's New Source Performance Standards (NSPS) for Gaseous Emissions from the FCC Regenerators
Source
Allowable Limits
Carbon monoxide (CO)*
Less than 500 ppmv in the flue gas
Nitrogen oxides (NOX)
None (local and regional only)
Participates**
A maximum of 1.0 pound of solids in the flue gas per 1,000 pounds of coke burned
Sulfur oxides (SO2 + SO3)*
Exempt if the feed sulfur is less than 0.30 wt% If there is no add-on control such as a wet gas scrubber, 9.8 kilograms of (SO2 + SO3) per 1,000 kilograms of coke burned. This is approximately equal to 500 ppmv. Add-on device: reduce (SO2 + SO3) by at least 90% or no more than 500 ppmv, whichever is less stringent.
*Effective January 1984 **Effective June 1973
the quality of FCC stocks, operating conditions, catalyst type, and mechanical condition of the unit. Processing feeds that contain a high concentration of residue, sulfur, nitrogen, and metals will release a greater amount of SOX, NOX, and particulates. Various technologies are available to reduce flue gas emissions.
Particulates Electrostatic precipitators (ESP) and wet gas scrubbers (WGS) are widely used to remove particulates from the FCC flue gas. Both can recover over 80% of filtrable solids. An ESP (Figure 10-6) is typically installed downstream of the flue gas heat recovery (prior to atmospheric discharge) to minimize particulate concentration. If both low particulate and low SOX requirements are to be met, a wet gas scrubber such as Belco's (Figure 10-7) should be considered. If SOX removal
BUS DUCTINSULATOR COMPARTMENT, ROOF
DISCHARGE ELECTRODE RAPPER
RAPPER INSULATOR HIGH-VOLTAGE SYSTEM SUPPORT INSULATOR
EN
COLLECTING SURFACE RAPPER
TRANSFORMER RECTIFIER
•SIDE
DOOR
Ci
o 2 DISCHARGE ELECTRODE HOPPER-
Figure 10-6. Typical electrostatic precipitator (ESP),
CAUSTIC SODA
Cd O
n 2
CIRCULATING PUMP
Figure 10-7. Schematic of Belco scrubbing system (courtesy of Belco Corporation).
Emerging Trends in Fluidized Catalytic Cracking
331
is not a prime objective, an ESP will be less expensive from the standpoints of both initial capital and operating costs. In some cases, a bag house system can be used instead of an ESP,
SOX Three methods are widely used to reduce SOX emissions from the FCC flue gas: FCC feed pretreatment Catalyst additives Flue gas desulfurization Feed hydrotreating or hydrocracking reduces SOX emissions and the sulfur content of FCC products. As discussed earlier in this chapter, many benefits are associated with FCC feed hydrotreating. It is important to note that most of the sulfur in a hydrotreated feed is in heavy organic compounds and will be concentrated in the decanted oil and coke. Consequently, for a given sulfur in the feed, more SOX will be produced with hydrotreated feed. For refiners having low to moderate levels of SOX in their FCC flue gas (less than 1,000 ppm), SOX additives are usually the most economical method of reducing SOX emissions. These additives are injected separately into the regenerator. They capture SO3 in the regenerator (oxidizing atmosphere) and release sulfur as H2S in the reactor (reducing atmosphere). A reliable on-line SO2 analyzer will ensure that a sufficient quantity of additive is injected. Operating conditions of the regenerator, especially partial versus full combustion and excess oxygen level, will greatly influence the additive's effectiveness. When processing high-sulfur feeds (greater than 1.0 wt%) or if the required SOX reduction levels are greater than 80%, other capitalintensive desulfurization technologies must be considered. Several flue gas desulfurization technologies are available. Haldor Topsoe's WSA, United Engineers' Mgo., Exxon, and Belco (Figure 10-7) wet gas scrubbing (WGS) are among the most widely used processes to remove SOX. The WGS process removes both SOX and particulates.
CO The CO levels released from the regenerator flue gas operating either in complete or partial combustion are normally less than 10
332
Fluid Catalytic Cracking Handbook
ppm. For units operating in partial combustion, the flue gas must be sent to a CO boiler. For units operating in complete combustion, the concentration of CO largely depends on the operating conditions of the regenerator (mainly temperature and excess oxygen), the CO promoter level, and the efficiency of the air/spent catalyst distribution system.
NOX NOX levels in the FCC flue gas typically range from 50-500 ppm. Nitrogen content of the feed, excess oxygen, regenerator residence time, dense phase temperature, and CO promoter all influence the concentration of NOX. In the regenerator, most of the NOX is formed as NO, with little N2O or NO2. About 90% of organic nitrogen in the spent catalyst is converted to inorganic nitrogen, and a very small amount becomes NO. NO can be lowered by reducing excess oxygen and CO promoter. The present platinum-based promoter oxidizes intermediates such as HCN and NH3 to NO and decreases the reducing agent such as CO. To reduce nitrogen oxide, thermal and catalytic processes are available. The thermal process is licensed by Exxon. NH3 or urea is injected into the flue gas at an elevated temperature (-1600°F, 870°C); NOX is reduced to nitrogen. This process is applicable to FCC units that have CO boilers. NOX can also be reduced over a catalyst at 500°F to 750°F (260°C to 400°C). EMERGING DEVELPMENTS IN CATALYSTS, PROCESSES, AND HARDWARE The FCC process has a long history of innovation and will continue to play a key role in the overall success of the refining industry. The continuing developments will primarily be in the areas of catalyst, process, and hardware technologies. Catalyst Since the mid-1960s, formulation of FCC catalysts has improved steadily. The focus of the research is in the following areas:
Emerging Trends in Fluidized Catalytic Cracking
* * • • * •
333
Improvement in zeolite quality Improvement in the catalyst's binder properties Increase in the quantity and choice of active matrix Customization of catalyst to the unit's objectives and constraints A widespread use of ZSM-5 or similar zeolite Improvements in the developments of catalyst additives for reducing gasoline sulfur and NOX emission
There has also been an ongoing trend to formulate a higher-quality zeolite. Higher quality has been reflected in: • Greater silica-to-alumina (SAR) of zeolite. Greater SAR results in a zeolite that is more stable, yields more olefins, improves octane, and increases product selectivity. * Improved crystallinity by producing more uniform zeolite crystals, FCC catalyst manufacturers have greater control over the zeolite acid site distribution. In addition, there is an upward trend in the quantity of zeolite being included in the catalyst. The selectivity and activity of the catalyst matrix will continue to improve. The emphasis on bottoms cracking and steady reduction in the reaction residence time demands an increase in the quantity of active matrix. The improvements in the catalyst's binder properties will reduce the catalyst attrition rate; thus, lowering the flue gas stack opacity. This improvement allows refiners to use a "harder" catalyst without adversely affecting the catalyst's fluidization properties. Future catalyst formulation will be customized to meet the individual refiner's needs. Catalyst manufacturers will be tailoring catalysts to meet each refiner's requirements. The demand for ZSM-5 additives will increase because of their inherent ability to crack low-octane, straight chain olefins to C3 and C4 olefins and also to isomerize low-octane linear olefins to higher octane branched olefins. Once ZSM-5's patent has expired, its use should increase. Further developments in the effectiveness of the FCC gasoline sulfur reduction additives will allow a number of refiners to meet the required reduction in gasoline sulfur without undertaking costly capital projects. Additionally, improvements in the CO promoter additives will reduce NOX emissions when the promoter is used. Finally, other
334
Fluid Catalytic Cracking Handbook
cost-effective additives will be developed to not only reduce NOX emissions, but also reduce catalyst related fouling in the regenerator flue gas heat recovery system.
Operating Conditions FCC will still play a dominant role in producing cleaner-burning fuels. The inherent flexibility of the process will allow refiners to meet the fuel reformulation requirements. With the anticipated growing demand for alky late and ethers, the FCC operating parameters will be adjusted to maximize production of propylene, isobutylene, and isoamylene. The projected trend in operating conditions will be to a higher reactor temperature, a higher catalyst-to-oil ratio, a higher reaction mix temperature, and shorter catalyst contact time.
Technology Development Since 1942, when the first FCC unit came onstream, new technologies have continuously evolved to maximize performance to meet the ever-changing product requirements and feedstock qualities. Future technology development will remain dynamic. Examples of the new and ongoing technologies aimed at enhancing the unit's operational and mechanical performance, as well as complying with environmental regulations, are: • Reducing sulfur and aromatics in gasoline and distillate. • Minimizing disposal of equilibrium catalyst. • Minimizing catalyst back-mixing in the riser to minimize production of undesirable products. Redesign of the conventional riser for a down-flow of catalyst and vapors could virtually eliminate back-mixing. • Achieving an ultra-short catalyst-hydrocarbon contact time, designed to maximize olefins and gasoline yields while minimizing the bottoms yields. • Eliminating long dilute-phase residence time downstream of the riser to prevent recracking of hydrocarbon vapors in the reactor housing. • Improving feed and catalyst injection systems. • Improving spent catalyst distribution.
Emerging Trends in Fluidized Catalytic Cracking
335
• Improving mechanical reliability of the FCC reactor-regenerator components. • Increasing use of feed segregation to maximize production of light olefins. « Increasing use of riser quench to maximize the reaction mix temperature and to promote maximum vaporization of the feedstock. • Increasing use of catalyst additives to reduce gaseous emissions and to maximize light olefins. These are just some of the many challenges facing FCC operations today. SUMMARY The United States refining industry is undergoing a restructuring phase. Refiners will continue to be under pressure and only the most efficient and profitable operations are going to survive. The survivors will be those who have some niche in the market place, have the versatility to handle low-cost crude, meet product demand, and conform to environmental regulations. FCC is one of the cheapest conversion processes. Its inherent flexibility can assist a refiner in meeting changing product requirements in spite of the steady decline in feedstock quality. The U.S. Federal RFG program has imposed new challenges for the FCC, particularly regarding the sulfur, aromatics, and olefin content of gasoline. Various commercially proven technologies, along with evolving technologies, will be available to comply with these new rules. The use of RFCC will continue to grow, particularly in regions of the world where atmospheric or vacuum residue contains low levels of contaminants. Careful regenerator and feed injection designs are important in ensuring a successful operation. Gaseous emissions (CO, NOX, SOX, particulates) have been regulated at local and national levels. The quantity of these emissions is directly related to the quality of the FCC stocks, operating conditions, catalyst type, and mechanical conditions of the unit. Processing heavy feeds will release a greater amount of SOX, NOX, and particulates. In conclusion, FCC has had a long history of innovations. New technological developments will continue to emerge, optimizing its performance. Its versatility and high degree of efficiency will continue to play a key role in meeting future market demands.
33§
Fluid Catalytic Cracking Handbook
REFERENCES 1. Mauleon, J. L. and Letzsch, W. S., "The Influence of Catalyst on the Resid FCCU Heat Balance," presented at Katalistik's 5th Annual FCC Symposium, Vienna, Austria, May 23-24, 1984. 2. A. W. Peters, G. Yaluris, G. D. Weatherbee, X. Zhao, "Origin and Control of NOX in the FCCU Regenerator," Grace Davison, Columbia, MD. 3. Davis, K., and Ritter, R. E., "FCC Catalyst Design Considerations for Resid Processing—Part 2," Grace Davison Catalagram, No. 78, 1988. 4. Hammershaimb, H. U., and Lomas, D. A., "Application of FCC Technology to Today's Refineries," presented at Katalistiks' 6th Annual FCC Symposium, Munich, Germany, May 22-23, 1985. 5. Kool, J. M., "Commercial Experience with Resid Cracking in Conventional FCC Units," presented at the 1984 Akzo Chemicals Symposium, 6. Hood, R., and Bonilla, J., "Residue Upgrading by Solvent Deasphalting and FCC," presented at the Stone & Webster 5th Annual Meeting, Dallas, Texas, October 12, 1993. 7. Dean, R. R., Kibble, P. W., and Brown, G. W., "Crude Oil Upgrading Utilizing Residual Oil Fluid Catalyst Cracking," presented at Katalistiks' 8th Annual FCC Symposium, Budapest, Hungary, June 1-4, 1987. 8. Johnson, T. E., "Resid FCC Regenerator Design," presented at the M.W. Kellogg Co. Refiing Technology seminar, Houston, Texas, February 9-10, 1995. 9. Letzsch, W., Mauleon, J. L. Jones, G., and Dean, R., "Advanced Residual Fluid Catalytic Cracking," presented at Katalistiks' 4th Annual FCC Symposium, Amsterdam, The Netherlands, May 18-19, 1983. 10. Elvin, F. J., and Krikorian, K. V., "The Key to Residue Cracking," presented at Katalistiks' 4th annual FCC Symposium, Amsterdam, The Netherlands, May 18-19, 1983. 11. Peeples, J. E., "The Clean Air Act, a Brave New World for Fuel Reformulation," Fuel Reformulation, Vol. 3, No. 6, November/December 1993. 12. Dharia, D., Brahn, M., and Letzsch, W., "Technologies for Reducing FCC Emissions," presented at Stone & Webster's 5th annual Refining Seminar, Dallas, Texas October 12, 1983. 13. Yergin, D. and Lindemer, K., "Refining Industry's Future," Fuel Reformulation, Vol. 3, No. 4, July/August 1993. 14. Perino, J. O., "Blending Control Upgrade Projects," Fuel Reformulation, Vol. 3, No. 4, July/August, 1993. 15. Clarke, R. H. and Ritz, G. P., "Method for the Analysis of Complex Mix of Oxygenates in Transportation Fuels," Fuel Reformulation, Vol. 3, No. 4, July/August, 1993.
Emerging Trends in Fluidized Catalytic Cracking
33?
16. Urizelman, G. H., "NOX," Fuel Reformulation, Vol. 1, No. 6, November/ December 1991. 17. Piel, W. J., and Thomas, R. X., "Oxygenates for Reformulated Gasoline," Hydrocarbon Processing, July 1990, pp. 68-73, 18. Hirshfeld, D. S. and Kolb, J., "Minimize the Cost of Producing Reformulated Gasoline," Fuel Reformulation, Vol. 4, No, 2 March/April 1994 19. Unzelman, G. H., "A Sticky Point for Refiners," Fuel Reformulation, Vol. 2, No. 4, July/August 1992. 20. Nocca, J. L., Forestiere, A., and Cosyns, J., "Diversify Process Strategies for Reformulated Gasoline," Fuel Reformulation, Vol. 4, No. 4, September/ October 1994. 21. Desai, P. H., Lee, S. L., Jonker, R. J., De Boer, ML, Vending, J., and Sarli, M. S., "Reduce Sulfur in FCC Gasoline," Fuel Reformulation, Vol. 4, No. 6, November/December 1994. 22. Sarathy, P. R., "Profit from Refinery Olefins," Fuel Reformulation, Vol. 3, No. 5, September/October 1993. 23. Hosteller, R. and Cain, M., BP Oil, private communication, 1995. 24. Reid, T. A., Akzo Nobel, private communication, 1995. 25. Swain, E. J., "U.S. Crude Slate Continues to Get Heavier, Higher in Sulfur," Oil & Gas Journal, January 9, 1995, pp. 37–42. 26. Dougan, T. J., Alkemade, V, Lakhampel, B., and Brock, L. T., "Advances in FCC Vanadium Tolerance," NPRA Annual Meeting, San Antonio, Texas, March 20, 1994, reprinted in Grace Davison Catalagram. 27. Cunic, J. D., Diener, R., and Ellis, E. G., Exxon Research and Engineering, "Scrubbing—Best Demonstrated Technology for FCC Emission Control," presented at NPRA Annual Meeting, San Antonio, Texas, 1990.
APPENDIX 1
Temperature Variation
of Liquid Vis
^HnnHHIHIIIHHIiniUIIIIIIMIHIIIIIIIIIIIIIIIIMIIIIIIIIIIIIIIIIHIIIIIIIIIIIIItllllilllilllUlllllilllllllllllllllMllfllliWIIItlll MnnitMIIIIIIIIIIIIIIIHIMIIIIIIIIIIItlllllllllllllllllllllllllllllllllllllillHieHlllllllillllltllllliSKUMIitHIUitilMfllHIIilllll
Source: U.S. Department of Commerce, adapted from ASTM D-342-39.
338
APPENDIX 2
Correction to Volumetric Average Boiling Point WABP C 80) F V
==• ""^ WABP O 60 3 F VABP
A8TM Diet, 10% - 90 % Slop*
339
APPENDIX 3
TOTAL Correlations Aromatic Carbon Content: CA = -814.136 + 635.192 x RI(20) - 129.266 x SG + 0.1013 x MW - 0.340 x S - 6.872 x ln(v) Hydrogen Content: H2 = 52.825 - 14.26 x RI(20) - 21.329 x SG - 0.0024 x MW - 0.052 x S + 0.757 x ln(v) Molecular Weight: MW = 7.8312 x 10-3 x SG-0-0976 x AP°C1238 Refractive Index @ 20°C: RI(20) = 1 + 0.8447 x SG1-2056 x (VABPoc+273.16r)0557 x Refractive Index @ 60°C: RK60) = 1 + 0.8156 x SG12392 x (VABP0(: + 273.16)-0.0576 x
Source: Dhulesia, H., "New Correlations Predict FCC Feed Characterization Parameters," Oil & Gas Journal, Jan. 13, 1986, pp. 51-54.
340
APPENDIX 4
n-d-M Correlations v = 2.5 x (RI20OC - 1.4750) - (d2()OC - 0.8510) 05 = (d2()OC - 0.8510) - 1.11 x (RI2fn, - 1.4750) If v is positive: %CA = 430 x v +
If v is negative: %CA = 670 x v +
3660
M
If 03 is positive: %CR = 8 2 0 x G J - 3 x S + 10,000/M 10,600 If 03 is negative: %CR = 1440 x 03 - 3S +
M
%CN = %CR — %CA %C_r = 100— %CRK
Average Number of Aromatic Rings per Molecule (RA): RA = 0.44 + 0.055 x M x v
If v is positive
R^ = 0.44 + 0.080 x M x v
If v is negative
Average Total Number of Rings per Molecule (RT): RT = 1.33 + 0.146 x M x (03 - 0.005 x S)
If 03 is positive
RN = RT — RA
RT = 1.33 + 0.180 x M x (03 - 0.005 x S)
If 05 is negative
Average Number of Napthene Rings per Molecule (RN): R
M
=
RT—RA
Source: ASTM Standard D-3238-80. Copyright ASTM. Used with permission.
341
APPENDIX 5
Estimation of Molecular Weight of Petroleum Ofts from Viscosity Measurements Tabulation of H Function H
40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190
334 355 372 386 398 408 416 424 431 437 443 448 453 457 461 465
336 357 374 387 399 409 417 425 432 438 443 449 453 458 462 466
339 359 375 388 400 410 418 425 432 438 444 449 454 458 462 466
341 361 377 390 401 410 419 426 433 439 444 450 454 459 463 466
343 363 378 391 402 411 420 427 433 439 445 450 455 459 463 467
342
345 364 380 392 403 412 420 428 434 440 446 450 455 460 463 467
347 366 381 393 404 413 421 428 435 441 446 451 456 460 464 468
349 368 382 394 405 414 422 429 435 441 447 451 456 460 464 468
352 369 384 395 406 415 423 430 436 442 447 452 456 461 465 468
354 371 385 397 407 415 423 430 437 442 448 452 457 461 465 469
Molecular Weight of Petroleum Oils Viscosity-Molecular Weight Chart LINES OF CONSTANT 210*F (98,89*C) VISCOSITY, cST
500
5
400
300
too
)0
j/
400
500
600
RELATIVE MOLECULAR MASS
Source: ASTM Standard D-2502-92. Copyright ASTM. Used with permission.
343
APPENDIX 6
Kinematic Viscosity to Saybolt Universal Viscosity Equivalent Saybolt Universal Viscosity, Sus
Kinematic Viscosity, cSt
1.81 2.71 4.26 7.37 10.33 13.08 15.66 18.12 20.54 43.0 64.6 86.2 108.0 129.5 139.8 151.0 172.6 194.2 215.8
At 100°F
At 210°F
32.0 35.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0 200.0 300.0 400.0 500.0 600.0 648.0 700.0 800.0 900.0 1000.0
32.2 35.2 40.3 50.3 60.4 70.5 80.5 90.6 100.7 202.0 302.0 402.0 504.0 604.0 652.0
Extracted from ASTM Method D-2161-87. Copyright ASTM. Used with permission.
344
APPENDIX 7
API Correlations Xr = a + b x (R.) + c x (VG) Xn = d + e x (R.) + f x (VG) Xn = g + h x (R.) + i x (VG)
Where constants vary with molecular weight range given below: Constants a b c d e f g h j
Heavy Fractions 200 < MW < 600 +2.5737 +1.0133 -3.573 +2.464 -3.6701 +1.96312 -4.0377 +2.6568 +1.60988
R. = Refractivity Intercept VGC = Viscosity Gravity Constant R, K
K ~-R i(20)
Where: R j(2()) = Refractive Index @ 20°C d' = Density @ 20°C Source: Riazi, M. R., and Daubert, T. E., "Prediction of the Composition of Petroleum Fractions," Ind. Eng. Chem. Process Dev., Vol. 19, No. 2, 1982, pp. 289-294.
345
346
Fluid Catalytic Cracking Handbook
VGC = SG ~ °-24 - °-022 x log(V210 ~ 35.5) 0.755 Where: V = Say bolt Universal Viscosity @ 210°F in seconds Refractive Index @ 20°C (68°F):
I = A x exp(B x MeABP + C x SG + D x MeABP x SG) x MeABPE x SGF Constants A B C D E F
2.341 * 10~2 6.464 x IQ"4 5.144 -3.289 x 10-4 -0.407 -3.333
MW = a x exp(b x MeABP + c x SG + d x MeABP x SG) x MeABP6 x SGf Where: Constants a b c d e f
20.486 1.165 x 10~4 -7.787 1.1582 x 10-3 1.26807 4.98308
APPENDIX 8
Definitions of Fluidization Terms Aeration. Any supplemental gas (air, steam, nitrogen, etc.) that increases fluidity of the catalyst. Angle of Internal Friction—a. Angle of internal friction, or angle of shear, is the angle of solid against solid. It is the angle at which a catalyst will flow on itself in the nonfluidized state. For an FCC catalyst, this is about 80°. Angle of Repose—p. The angle that the slope of a poured catalyst will make with the horizontal. For an FCC catalyst, this is typically 30°.
SoHdSurfaca
347
348
Fluid Catalytic Cracking Handbook
Apparent Bulk Density—ABD. The density of the catalyst at which it is shipped either in bulk volume or bags. It is density of the catalyst at minimum fluidization velocity. Bed Density—pb. The average density of a fluidized bed of solid particles and gas. Bed density is mainly a function of gas velocity and, to a lesser extent, the temperature. Minimum Bubbling Velocity (Umb). The velocity at which discrete bubbles begin to form. Typical minimum bubbling velocity for an FCC catalyst is 0.03 ft/sec. Minimum Fluidization Velocity (Umf). The lowest velocity at which the full weight of catalyst is supported by the fluidization gas. It is the minimum gas velocity at which a packed bed of solid particles will begin to expand and behave as a fluid. For an FCC catalyst, the minimum fluidization velocity is about 0.02 ft/sec. Particle Density—p . The actual density of the solid particles taking into account any volume due to voids (pores) within the structure of the solid particles. Particle density is calculated as follows:
Po =
Skeletal density (Skeletal density x PV) + 1
Pore Volume—PV. The volume of pores or voids in the catalyst particles. Ratio of Minimum Bubbling Velocity to Minimum Fluidization Velocity (Umb/Umf). This ratio can be calculated as follows: Umb
=
2300 x p°' 2< x n°"3 x exp° 7lteF
Umf= Where: pg = ji = F = dp = p = g =
d°8xg