Energy & Fuels 2004, 18, 160-171

Study of Canadian FCC Feeds from Various Origins and Treatments. 1. Ranking of Feedstocks Based on Feed Quality and Product Distribution
Siauw H. Ng,*,† Jinsheng Wang,‡ Craig Fairbridge,† Yuxia Zhu,⊥ Liying Yang,# Fuchen Ding,¶ and Sok Yui§
National Centre for Upgrading Technology, 1 Oil Patch Drive, Suite A202, Devon, Alberta T9G 1A8, Canada, CANMET Energy Technology CentresOttawa, 1 Haanel Dr., Ottawa, Ontario K1A 1M1, Canada, Research Institute of Petroleum Processing, 18 Xue Yuen Road, P.O. Box 914, Beijing 100083, China, Beijing University of Chemical Technology, 15 Bei San Huan East Road, Chao Yang District, Beijing 100029, China, Beijing Institute of Petrochemical Technology, Daxing, Beijing 102600, China, and Syncrude Research Centre, 9421-17 Avenue, Edmonton, Alberta T6N 1H4, Canada Received April 24, 2003

Accurate ranking of fluid catalytic cracking (FCC) feeds will help oil producers price their commodities. The value of an FCC feed depends on its properties, which, in turn, contribute toward producing high-value products with good qualities. Feed ranking based on general analyses is the simplest way to determine the best feeds but can be misleading for some. In this study, the authors propose a feed grading method with consideration of only concentrations of gasoline precursors, total nitrogen, and microcarbon residue (MCR), assuming that gasoline is the most desirable product. By assigning the merit and discount values of the three elements, the relative gasoline yield of a feed can be calculated and its apparent feed rank established. The method was tested on 10 feeds that were catalytically cracked in both a microactivity test (MAT) unit and a riser pilot plant. With some exceptions, the apparent feed ranks indicated an agreeable order with those based on individual high- and low-value MAT product yields at both constant test severity (catalyst/oil (C/O) ratio of 7) and conversion (65 wt %). The apparent feed ranks were also validated with the exact feed ranks determined by comparing the individual MAT yields at an achievably high conversion, using the maximum gasoline yields as a guide. The two feed ranks also showed good conformity, with respect to the sequence. Verification of apparent feed ranks against riser pilot plant yields in this study was difficult, because of limited test data for comparison. MAT yields of the 10 feeds compared better with riser pilot plant yields at the same conversion than with yields at the same C/O ratio. Among the six cracked products, dry gas gave the best comparison, in terms of absolute yields, whereas coke showed the worst results.

1. Introduction The Canadian oil-sands resource is spread across 77 000 km2, mostly in northern Alberta. The oil sands in Canada hold 1.7-2.5 trillion barrels of bitumen. Of this amount, ∼300 billion barrels are considered recoverable, with currently known surface mining and in situ technologies. This scenario puts the oil-sands resource on par with Saudi Arabia’s conventional oil reserves. In 1999, oil sands accounted for ∼26% of the total Canadian crude oil production. Based on industry-wide expansion programs announced or underway, this amount could increase to 50% beyond 2005.1 Syncrude Canada, Ltd., which is the greatest oilsands bitumen upgrader in the world, operates a surface
* Author to whom correspondence should be addressed. E-mail: sng@nrcan.gc.ca. † National Centre for Upgrading Technology. ‡ CANMET Energy Technology Centre-Ottawa. ⊥ Research Institute of Petroleum Processing. # Beijing University of Chemical Technology. ¶ Beijing Institute of Petrochemical Technology. § Syncrude Research Centre. (1) Newell, E. P. Oil Gas J. 1999, 97, (26), 44-53.

mining plant in northern Alberta and produces a synthetic crude oil (SCO) called “Syncrude Sweet Blend”, or “SSB”. The extracted bitumen is currently upgraded in two fluid cokers and an ebullated-bed LC-Finer hydrocracker, followed by hydrotreatment of naphtha, light gas oil (LGO), and heavy gas oil (HGO). The hydrotreated products are combined as SSB and shipped to refineries by pipelines. Refiners use the naphtha, LGO, and HGO fractions in SSB as, respectively, catalytic reformer feed, diesel or jet fuel blending stock, and feed to the hydrocracker or fluid catalytic cracking (FCC) unit. The disadvantages of using nonconventional bitumenderived HGOs or vacuum gas oils (VGOs) as FCC feeds have been described in previous publications.2,3 Over the
(2) Ng, S. H.; Ring, Z.; Humphries, A.; Yui, S. In Proceedings of the Third Joint China/USA Chemical Engineering Conference; Sponsored by Chemical Industry and Engineering Society of China, American Institute of Chemical Engineers, and National Natural Science Foundation of China: Beijing, China, 2000; Vol. 1, scope 01, pp 01-5101-58. (3) Ng, S. H.; Zhu, Y.; Humphries, A.; Zheng, L.; Ding, F.; Gentzis, T.; Charland, J. P.; Yui, S. Energy Fuels 2002, 16, (5), 1196-1208.

10.1021/ef0300996 CCC: $27.50 © 2004 American Chemical Society Published on Web 12/11/2003

Oil Gas J. Wang. 1 Energy & Fuels. S. F. S. Thus. Because the reaction temperature range was rather small ((10 °C at 510 °C). with varying weight hourly space velocity (WHSV. (4). C. 2. which is a modified version of ASTM D 3907. and DA-LCF were used as received. it was assumed that the temperature effect on MAT yields was negligible and that one regression could be applied to the same set of data points obtained at three different temperatures. S. Catalyst.. under typical conditions in a pilot plant. some runs were repeated later... 343-525 °C) was obtained from batch distillation of an entire bitumen to simulate a commercial product from Syncrude’s newly installed vacuum tower. where WHSV ) 1/τ ) 3600/(C/O × t) ) 3600/[C/(O/ t)]. given in units of g/s) was varied to obtain different amounts of oil (O.1.9 and 8. Yui.6.. Yang. respectively. 343 °C+). (3) Feeds from a bitumen source: (i) The untreated virgin VGO (VIR.1. S. 343 °C+). and it is important to assess their performances in catalytic cracking. In 1998. L.1 and 25. These feeds. Zhu. given in seconds) constant. Fairbridge. (iii) The atmospheric topped bitumen from Syncrude’s diluent recovery unit4 was deasphalted. thereby ranking them as FCC feeds.3. (ii) VIR was further hydrotreated in a fixed-bed catalyst testing unit (CTU) at Syncrude Research. Riser results were reported fully by Yui et al. given in grams). Y.. was used in this study. (9) Ng. (3). S.1. Feedstock Properties. 2001..9 The FCC pilot plant data were obtained from a modified ARCO-type riser reactor. F.1. 2. For each feed. H. which is a typical Alberta light crude mix. three runs were conducted at 510 °C and one and two runs were done at 500 and 520 °C. However. 2004 161 (4) Feeds from Syncrude’s LC-Fining hydrocracker: (i) The untreated LC-Finer VGO (LCF.5 years. (4). Dimension 60. 343-525 °C) was obtained by blending the LGO. and bottoms from the LC-Finer fractionator in a representative ratio. 783-785. The objective is to investigate and compare the catalytic cracking behavior of the 10 VGOs.3 wt %. (iv) A mixture of deasphalted oils from bitumen under different extraction conditions was hydrotreated in a mini-ebullated bed pilot plant at Syncrude Research under typical LC-Finer operating conditions. (5) Feed from Syncrude’s fluid coker: The blend of LGO and HGO from the coker fractionator was distilled in batch to IBP343 °C and 343 °C+ fractions.. Yang. Chung. (iii) The deasphalted oil from LC-Finer resid (DA-LCF. Y. the effects of progressive treatments of feeds on their yield and product quality improvements. HCB. for brief comparisons of cracking yield2. The TLP was distilled to yield the hydrotreated deasphalted VGO from bitumen (HT-DA..5-7 Feed analyses. 99. given in hours). The 10 feeds supplied by the Syncrude Research Centre can be classified into the following categories: (1) Feed from a conventional source: The Rainbow Zama VGO (RZ. Prior to the riser experiments. HT-DA and HT-C were mostly 343 °C+ but relatively high in 525 °C+ (36. The TLP was distilled to yield the hydrotreated coker VGO (HT-C. The deasphalted oil (DAO) from bitumen (DA-BIT. the company conducted a series of riser pilot plant tests that included 10 feeds. 96. Fairbridge. This VGO was used to simulate its counterpart in Syncrude’s commercial plant. A refinery-generated equilibrium catalyst. Z. as received) was obtained from Petro-Canada’s Edmonton refinery when it processed 100% SSB. 14. The fixed-bed MAT unit used was an automated Zeton Automat IV. 343-525 °C). Detailed description of the MAT reaction system and operation have been reported elsewhere. Yui. The omission of the temperature effect can be justified by the rather low sensitivity of the MAT yields to temperature in this study (Table 3).. J.. . S. Zhu.. As mentioned previously. Yui. The distillation range can obviously affect the feedstock quality. using a typical commercial NiMo catalyst at 380 °C. was used as a reference in this study. and the comprehensive comparison of the product yields between the riser reactor and the MAT unit. This paper supplements previous publications2. given in units of h-1) or space time (τ. (7) Ng. DA-BIT. No. 18. Energy Fuels 2001. 172-187. Distillation Range. Energy Fuels 1995. C. HGO. except one from conventional crude. H. 945-946. H. Note that the distillation data were obtained by simulated distillation (simdist). Syncrude has been engaged in research activities to improve the quality of HGO from SCO. 1276 psig. H. followed by a batch distillation. of which the operation and methodology have been described by Yui et al. Energy Fuels 2004. Results and Discussion 3.2. Oil Gas J. Table 2 gives some properties of this catalyst. Vol. the oil injection rate (O/t. The latter was hydrotreated under the same conditions as those for VIR. therefore.. In this study. To obtain different conversions for each feed.6 by providing detailed MAT results. microactivity (MAT) tests were performed to establish pilot-plant operating conditions..6 and product quality7 between the riser reactor and the MAT unit. This work is complemented by a companion paper (Part 2)8 that focuses on specific observations in this study. X. obtained by distillation of Rainbow Zama crude. All feeds were characterized using ASTM and other supplementary methods at the Syncrude Research Centre. (6) Ng.. and were related to Syncrude’s existing4 or future potential processing schemes that involved a hydrocracker and a solvent deasphalter. Ring. Fairbridge. 2.2. (8) Ng. Catalytic Cracking of Feedstocks. Ding. H. (4) Yui. H.fractions (19.. and an LHSV value of 1 h-1. as received) was used directly for catalytic cracking. 1. C. X.. Wang.5 and partially by Ng et al. 1998. Future Syncrude operations may produce various VGOs for downstream FCC use. equipped with collection systems for gas and liquid products. The resulting total liquid product (TLP) was distilled to produce the hydrotreated virgin VGO (HT-VIR. respectively). Y. Y. 343 °C+). (ii) LCF was also hydrotreated under the same conditions as those for VIR. using a mixture of n-butane and i-butane as the solvent. they contained relatively higher 343 °C. 343 °C+). Feedstocks. Y. compared with other feeds.. N. 216-224. (5) Yui. S. 18. J. Briker.. Experimental Section 2. C ) 4 g and t ) 30 s. Ding. Matsumoto. Gentzis. Primary feed properties were reported previously. Feed properties in Table 1 are discussed in several aspects. (2) Feed from a refinery hydrocracker: The hydrocracker bottoms (HCB. Y. 9. 15. with some addition and correction. S. Sasaki..Canadian FCC Feeds from Various Origins.6 wt % for HCB and ∼11 wt % for the two DAOs) and 525 °C+ fractions (38. Zhu. (17). K. J. the catalyst/oil (C/ O) ratio was changed while keeping the catalyst contact time (t. 3. HCB was included in the test program in consideration of the potential process development.. T. are summarized in Table 1 for ease of discussion. as received) was obtained by deasphalting the atmospheric resid from LC-Finer in the same manner as that for the topped bitumen.0 wt % for DA-BIT and DA-LCF.. Zhu. Energy Fuels 2000. 4351. were all derived from oil-sands bitumen. 46-53. S. 3.. FCC feed quality is wellknown to be the most important factor affecting catalytic cracking yields and product quality. The resulting TLP was distilled to produce the hydrotreated LC-Finer VGO (HT-LCF.

4 99. Vol.9 35.67 wt % 291 wppm 314 wppm 1.7 29.4 1.7 wt % 1.3 6.1 0.5 2 HT-VIR 0.7 68.5 13.2 71.6 89.8 0.1 28.9284 12.1 3.5 3.7 7.4 16.6 1.9 15. Density or API gravity reflects the “lightness” of an oil and is considered to be a measure of feed quality.8 20.1 10. Catal. nitrogen (10) Fisher.0 22. 3.2.1 64.2 30.892 0 <1 <10 0.2 45.3 58.9776 11.5 5.5269 1.2 2.0 1.1 17.4 1.56 13.6 2. the retained core compounds would likely boil in the LCO range.7 0.4 20.562 2150 439 4290 0. 18.2 1.8 1.8 1.162 Energy & Fuels.638 1090 148 880 0. 3.9 4. by MSD (wt %) saturates paraffins cycloparaffins aromatics monoditritetra.6 25.9 12.3 0. Appl.0 24.0 2.0 5.4 50.4 6.b relative to 5.0 2.3 8.1 23.5 1.9 0.1 2. 189-210.5 8.2 0. including dry gas.40 4.3 0.50 24.8 24.9 65.1 4.61 4.8 26.8 1.9 1.10 and less light cycle oil (LCO.9 75.9712 11.0 1.0 2.4 wt % 37.4 62. 65.6 10.and greater aromatic sulfur 2-ring compounds 3-ring compounds 4-ring compounds unidentified polar compounds gasoline precursorsb LCO precursorsc a 1 HCB 0.1 2.8 2.4 5.5 C/O (wt %) yielda at 520 °C and 5.1 21.1 µm 72 wt % dry gas LPG gasoline LCO HCO coke conversion a Table 3.9511 11. P.9 14.9430 11.9 1.3. Feed impurities are destructive in the catalytic cracking process.5 29. Some Properties of Equilibrium Catalyst Dimension 60 property X-ray diffraction unit-cell size zeolite content nitrogen adsorption-desorption total surface area zeolite surface area matrix surface area zeolite/matrix (Z/M) micropore volume zeolite content mercury porosimetry pore volume pore area average pore diameter water absorption pore volume chemical analyses SiO2 Al2O3 RE2O3 (on catalyst) RE2O3 (on zeolite) Na2O TiO2 Fe2O3 nickel vanadium particle size distribution 0-20 µm 20-40 µm 40-80 µm 80-120 µm 120-150 µm >150 µm average particle size MAT activity value 24. Feedstock Propertiesa Ng et al.1 1.9 9.5 5 HT-C 0. .4 7.5 2.1 1.4 90.1 0.5 1.8 1.556 3050 815 35400 5.9 0.8 38.0 2.1 3.7 6.00 0.4747 19. c Diaromatics + two-ring aromatic sulfur + 1/2 × three-ring aromatic sulfur.7 3.5323 4.1 0.9 4. and some catalytic coke.8 59.9 3.8 38.5 1.0 0.31 mL/g 55.7 51.0 73.1.1 26.4967 2.7 1.6 wt % 0.1 16.9642 11.3 46.4 3.5 61.0 10.1 1.8988 12.5 8.5397 6.1 4.4 wt % 82. 2004 Table 1.7 0.5153 5.05 mL/g 15.5 44.3 3 RZ 0.8643 13.5324 7.6 1.2 1.2 2.2 0.6 7.5 14.8 1.3 13.0 wt % 33.24 0 0 30.4 1.7 14.0 0.3 34.0 5.3 24.7 5. more convertible components or gasoline precursors (sometimes also called “conversion precursors”. Among them.9252 12. two-ring aromatic sulfur.7 61.1 4.5 3.4 16.4 1.1 4 HT-LCF 0.536 4020 1212 15200 3.8 4. respectively).619 2450 613 7040 2.1 42.5370 11.6 0.5 1.6 6.1 6.9 9 DA-LCF 0.8 5.9 81.6 10 DA-BIT 0.8 29.1 6. an unconverted product) precursors (defined as diaromatics + two-ring aromatic sulfur + 1/ × three-ring aromatic sulfur).8 1.7 1.21 wt % 7.8 wt % 0. and (ii) after removing the side chains of diaromatics.5 wt % 8.6 1.4 6.4 9. would predominantly form gasoline and its derived products (also the converted products). b Saturates + monoaromatics.3 13.0 25.4 57.1.1 6. these are defined as saturates + monoaromatics).8 1.19 wt % 1.1 18.3 27.4 38.7 1.9562 11.6 61.5 12.764 800 240 9170 0.2 1.3 1.7 wt % 9.0 30.6 11.0 1.3 3.by simdist (wt %) 524 °C+ by simdist (wt %) hydrocarbon type.37 11. feed number feed name density at 15 °C (g/mL) hydrogen content (wt %) H/C atomic ratio total nitrogen content (wppm) basic nitrogen content (wppm) total sulfur content (wppm) MCR content (wt %) nickel content (wppm) vanadium content (wppm) aromatic carbon content (%) aniline point (°C) refractive index at 20 °C 343 °C.7 2. and some three-ring aromatic sulfur.1 56.0 15.7 3. b Based on yields at 500 and 520 °C.6 7.4 21.8 0.08 0 0 14.7 2.5 Mostly taken from ref 5.5 0.2 8.0 0.7 wt % 39. 1.2 21. Table 2.549 1930 610 32500 0.04 4.02 20.1 58.0 wt % 10.7 20.0 1.7 97.8 6 HT-DA 0.1 25.2 36.7 11.3 52.8 55. The rest of the feeds were essentially in the 343-525 °C range.2 47. Sensitivity of MAT Yields to Temperature yielda at 520 °C and absolute difference.2 For an average of 10 feeds. Impurity Levels. Density and Precursors Concentrations.0 6.2 0.4 4.9 77.28 Å 13 wt % 150 m2/g 100 m2/g 50 m2/g 2.0 0.10 Note that the precur2 sors are based on the assumptions that (i) saturates and monoaromatics.532 3370 991 13500 0.6 13.8 65.7 1. upon catalytic cracking.6 5.5 C/O (%) 2.7 1.4 1.4 7.5393 10.22 mL/g 61 m2/g 109 Å 0.7 52.3 6.4 35.4 1.2 1.3 29. Low-density or high-API-gravity feeds usually contain less impurities.2 1.5 36.9 3.9 7 VIR 0.7 15.5130 7.53 65.4 5.7 3. No.7 63.0 29.3 9.7 8 LCF 0.6 74.669 460 51 981 0 17.8 63.62 wt % 0.7 5.0 28.6 1.7 6. I. 1990. liquefied petroleum gas (LPG).33 0 0.0 4.6 22.4 5.

and 38.12 Similarly.5. Z. are emphasized. No. for positive contribution. the amounts of which were greater in heavier feeds and less in lighter feeds. (2) individual product yields at a common severity and a common conversion. High metals (especially nickel and vanadium) in feeds deposit on the catalyst surface. However. and DABIT): These feeds were characterized by high contents of nitrogen (3000-4000 wppm. 1999.9 wppm for HT-DA). HT-VIR. the two deasphalted oils (DAOs) had much greater MCR values (3. Canada. respectively.4 versus ∼0. (11) Hou. Paper No. respectively. we assume that gasoline is a premium product with the highest value. only three elements in feed analyses are considered: gasoline precursors. Ranking Based on Feed Analyses. Table 1 also indicates that basic nitrogen content was approximately one-third to one-fourth of the total nitrogen. and (2) a 0. Common sense dictates that a feed is good when it is low in density and impurities but high in saturates. The relative gasoline yield of a feed can be explicitly expressed as follows: relative gasoline yield (wt %) ) (0. San Antonio. HT-VIR ≈ RZ > HT-LCF . The value of an FCC feed depends on its processability to produce desirable products with good qualities. because of the following reasons: (i) the long-term negative effects of nickel + vanadium cannot be shown in MAT runs. In Europe and the United States.5 × MCR) (given in units of wt %). they contained <53 wt % aromatics. because of the dependence of product value on geographical location. 43. Our in-house study provides estimates on the positive and negative impacts on gasoline yield by the three elements: (i) a 0.70 wt % for HT-DA).3 and 25. 3. With respect to chemical structure. 2004 163 (specifically basic nitrogen) and microcarbon residue (MCR) are poisons to catalysts. (13) Ritter. sulfur (<0. Note that the relative positions of feeds with similar ranks (those connected by a double approximation sign.95-0. polars (5.. and (3) individual product yields at a practically high conversion. which had 1930 wppm).4 wt %. greater aromatics contents (58-62 wt %). followed by LPG. HT-C and HT-DA had higher densities (0. low-valued).95 g/mL).005 × total nitrogen) (given in units of wppm) . “≈”) may be interchangeable. because of its virgin nature). The definition of “desirable products” is controversial.Canadian FCC Feeds from Various Origins. Their relative gasoline yields were estimated at 20. The 10 feeds from different origins and treatments were lumped into the following categories: (1) Good-Quality Feeds (HCB. Paper No. AM88-57. 18.7.2. in North America. In this study. one might conclude that hydrotreating substantially removed the impurities and improved the feed quality. which were then compared with those based on (i) conversion.3. In this paper. compared with gases and LCO.1 wppm for the two non-DAOs). The sequence of these ranks is as follows: HCB .(0. With proper design and operation. which are not shown in limited MAT runs. HT-C was lower in MCR (0. and LCO precursors (1722 wt %). C. Feed Ranking Protocol.. It is thus necessary to establish a general formula for feed ranking. However. aromatics (58-66 wt %).(0. and (ii) sulfur affects only the liquid product quality.11. 3.5 wt % penalty each for every 100 wppm total nitrogen and for every weight percent of MCR.93 g/mL. Other analyses are excluded. resulting in both positive and negative impacts on cracking. In comparison with good-quality feeds. For simplicity. because of experimental error. RZ. 148. and total nitogen and MCR. Petrochem.7-15.6 × gasoline precursors) (given in units of wt %) . Presented at the 52nd Canadian Chemical Engineering Conference. 16.4. based on the formula described previously. based on relative gasoline yields. their gasoline yields were estimated at 58. Presented at the 1988 National Petroleum Refiners Association (NPRA) Annual Meeting. D. E.3 wt % for the two non-DAOs) and nickel + vanadium contents (13-37 wppm versus ∼0. Vol.92 wt %. The gasoline yields thus calculated for the 10 feeds were used to establish the apparent feed ranks.4-3.. and nickel + vanadium (undetectable versus 10. in some countries where gasoline is not in high demand. (3) Low-Quality Feeds (VIR. the 10 feeds were ranked in descending order from feed 1 to feed 10. 1 Energy & Fuels. 1 (diesel quality) and No. In Table 1. Wang. sulfur (0. and 10. heavy cycle oil (HCO. and coke (unwanted or least-valued). as well as high densities (0.2. propylene.2 wt %). and nickel + vanadium (below detection limit). Technol. except RZ. Upon cata- . X.6 wt % for HT-DA). the FCC unit. LCO (medium-valued). 4345.6 wt % contribution for every weight percentage of gasoline precursors. Although sulfur causes no harm in catalytic cracking.2.5 wt %). For example. HT-VIR’s versus VIR’s and HTDA’s versus DA-BIT’s). because of its low capital cost (one-fourth that of a hydrocracker).5 wt % versus 2. Process. 42. Among the four feeds in this category.e.5 wt %. R. DA-LCF. By comparing the analyses of a hydrotreated feed against its predecessor’s (i. commercial processes exist to maximize the yields of ethylene. only yields. 11. Their densities were <0. respectively. and <14 wt % LCO precursors. The untreated feeds (RZ and VIR) and nonhydroprocessed feeds (DALCF and DA-BIT) were high in sulfur and nitrogen. is used as a major producer of middle distillate.1 wt %. uncertainties exist when feed properties vary toward opposite directions. 1-2. dry gas.4-5. (12) Li. an FCC unit can produce 40-45 wt % LCO. (2) Medium-Quality Feeds (HT-C and HT-DA): These feeds had average levels of nitrogen (2300 ( 150 wppm).98 g/mL). British Columbia. TX. which had 0.13 which can be upgraded to No.94-0. low in nitrogen (<1100 wppm). in FCC operation. >73 wt % gasoline precursors. gasoline is generally considered the most desired product. 2002. many refiners seek seasonal means of increasing LCO yields to meet greater heating-oil demands in the winter.1 wt %). Vancouver. 2 fuel oils that generally have a lower initial boiling point (∼182 °C). Xie. for negative contribution to gasoline yield. and a larger amount of LCO precursors (∼15 wt %) but much lower gasoline precursors (∼64 wt %). it deteriorates product quality and may induce operational problems. and HTLCF): These feeds were. On a relative scale. in China and Thailand.7 wt %. Between the two feeds. MCR (<0. LCF. sulfur (1. Their relative gasoline yields were estimated at 27. but low amounts of gasoline precursors (48-56 wt %). Table 1 shows that the 10 feeds contained a wide range of impurities.43 wt % versus 0. China Pet. resulting in negative impacts. 1. HT-C ≈ HT-DA > VIR > LCF > DA-LCF ≈ DA-BIT. It is simple to rank a feed based on its quality. except VIR. rather than product qualities. and other light olefins for use in petrochemical industries.9. in general.

A.4.70 wt % for HT-DA.. pilot-plant results were not used to evaluate the apparent feed ranks. The combination of Figures 1 and 2a gives the relationship between dry gas yield and conversion (Figure 2b). D. such as (i) RZ was inferior to HT-LCF. Conversion. for ease of interpretation. with the trend lines established for MAT data only. Figure 1 shows the relationships between conversion (from both MAT and riser data) and the C/O ratio for 10 feeds.5 wt % for VIR and DA-BIT) and MCR (2. In this study. M. Vol. RZ > HT-VIR > HT-DA > DA-BIT ≈ HT-LCF . Y. AM-92-17. 3. 3. Gilman. One should remember that a plot of yield versus the C/O ratio. and test and analytical methods). 1. New Orleans. At a constant C/O ratio of 7. 2004 Ng et al. This order did not quite agree with the apparent feed ranks. In FCC. promoting nonselective cracking to produce dry gas. . 3. and 3. DA-BIT.1. 0. LA. it was expected that the individual product yields. e.. Conversion and yield data on common bases will be used to verify the validity of apparent feed ranks that have been established.products. In the following sections of this paper. At constant C/O ratio (C/O ) 7).4 wt % for DA-BIT). closed symbols) and a microactivity test (MAT) unit (open symbols). reaction severity.g. hydrodynamics.g. and the catalyst weight and the catalyst contact time were all fixed for a given feed. Paper No. reflects the feedstock quality and is affected by the catalyst properties. 18. 3. HT-C ≈ DA-LCF ≈ VIR . and C1-C2 hydrocarbons. A. the conversions of 10 feeds were in the following descending order: HCB . because of different degrees of deactivation during reaction.4. the MAT unit gave slightly higher conversions for the seven non-DAO feeds but much higher conversions for the three DAO feeds.e. Asim. and (iii) VIR produced more dry gas than DA-LCF. Conversion is defined as the portion of the feed converted to 216 °C. T..g. whereas ∼40% feed sulfur might contribute to dry gas as H2S14 (i. this established sequence is compared with the MAT cracking results. dry gas and coke) following the order shown previously. Cracking results are graphically presented in Figures 1-7. such as Figure 2a. high MCR and nitrogen contents caused deviation from the predicted order. For a given feed at a fixed severity. gasoline) or increase (for lowvalue products. The subsequent discussions on individual product yields will shed more light in this respect. No. provides a comparison of product yields for different feeds at a chosen C/O ratio when the feeds are in contact with the same catalyst which exhibits different activities and selectivities to feeds. H. lytic cracking at a constant severity or conversion.43 wt % H2S). This plot provides a different comparison basis. For DAO feeds. and operating system used (e.5 wt % for LCF and DA-LCF. the yield-conversion relationship is interfered with less by the operating system. (ii) HT-DA also behaved worse than LCF. Presented at the 1992 National Petroleum Refiners Association (NPRA) Annual Meeting. because (14) Keyworth. which affected the weight hourly space velocity through the relationship WHSV ) 3600/[(C/O)t]. This order was generally consistent with the apparent feed ranks for non-DAO feeds. therefore.92 wt % for RZ. which were inter-related to some degree. The nitrogen poisoning effect on the catalyst probably could not be observed clearly in Figure 2a because (i) dry gas was mostly produced thermally rather than catalytically and (ii) other than the four good feeds. R. LCF. when compared with the corresponding conver- sions obtained from the riser pilot plant. and 5. The low-value dry gas is composed of H2. Validation of Apparent Feed Ranks with Conversion and Individual Yields. MCR might block the catalyst pores. would decrease (for high-value products. This phenomenon was explained in our earlier publication. the rest were all high in nitrogen with only moderate variations in concentration (1900-4000 wppm). 1 wt % feed sulfur might produce 0. the only variable was the catalyst-to-oil ratio (C/O). 1.. It demonstrated clearly that the dry gas yield was strongly influenced. similar to the product yield.4.2. yields of dry gas for the 10 feeds varied in the following order: HCB < HTLCF ≈ HT-VIR < RZ ≈ HT-C < LCF < HT-DA ≈ DALCF < VIR .3-3. Dry Gas. The conversion of a feed is a key parameter closely monitored in a reaction study. e. which is more commonly adopted. The conversion. H2S.6 Because of limited data. including gas and coke. the MAT reaction temperature was assumed to be constant. Figure 2a shows the increase in dry gas yield with the increase in the C/O ratio.164 Energy & Fuels. reactor configuration. The feed sulfur and MCR could explain some of the unexpected observations on the aforementioned order. with each product yield plotted against both C/O ratio and conversion.. Relationship between conversion and the catalystto-oil (C/O) ratio for 10 feeds cracked in a riser pilot plant reactor (R.4-1. This masked the nitrogen effect to some degree. other than the hydrocarbon composition.. by the feed sulfur (0..4 wt % for DA-LCF. Reid. At the same C/O ratio. Figure 1.6 wt % for HT-DA.

with greater differences in yield for feeds with higher MCR. among the 10 feeds. In contrast.4. Careful examination of Table 1 revealed that DA-BIT. as conversion increased. the order of dry gas yields appeared as follows: HCB . but in a slightly concave shape. LPG (C3 + C4 hydrocarbons) is considered a valuable product. Generally. Liquefied Petroleum Gas. closed symbols) and a MAT unit (open symbols). LPG accelerated its net increase between its formation from the precursors and its decomposition to dry gas.2 wt %) that were not counted as gasoline precursors. i. but in a convex shape. No. similar to the dry gas. this produced the following order: HCB < HT-VIR < HT-DA ≈ DA-BIT ≈ DA-LCF . less its decomposition to dry gas). In FCC operation. At a given C/O ratio (C/O ) 7). At 65 wt % conversion (assuming that some trend lines can be extrapolated). except for VIR. at the same C/O ratio. It is known that long contact time favors the secondary cracking of gasoline to produce more gases. which had an extremely large amount of sulfur. In this study. Assuming that both trend lines of HCB and LCF could be extended for a comparison of LPG yields at a common conversion of 65 wt %. LPG yields decreased in the following order: HCB . This indicates that. 18. 2004 165 Figure 2. HT-VIR > HT-DA > HT-LCF ≈ DA-BIT > HT-C ≈ VIR > DA-LCF ≈ LCF. HT-C ≈ HT-DA < LCF ≈ DA-LCF < VIR < DA-BIT. (a) Relationship between dry gas yield and C/O ratio for 10 feeds cracked in a riser pilot plant reactor (R. (b) Relationship between dry gas yield and conversion for 10 feeds cracked in a riser pilot plant reactor (R. 3. 1. with the increase in conversion. or the lack of sufficient experimental data.e. Vol. Figure 2b depicts the exponential increase in dry gas yield with conversion. Figure 3b shows the increase in LPG. Figure 2a shows that. 1 Energy & Fuels.Canadian FCC Feeds from Various Origins. performed better than expected. except that DA-BIT outperformed HT-C and three other feeds. rendering the extrapolation of the trend line impossible or unreliable. as explained previously. This order generally agreed with the apparent feed ranks. both the yield and conversion are affected to the same degree (the yield of a converted product is part of the conversion). This order was consistent with that determined by the feed quality. This also helps explain why HTDA and DA-LCF. This was due to the differences in reaction systems. This implies that as the C/O ratio increased. and high nitrogen content. LPG also increased as the C/O ratio increased (Figure 3a). This seems to suggest that some of the polar compounds in DAOs might be crackable. despite having the lowest gasoline precursors content. resulting from feedstock characteristics. LPG gradually relaxed its rate of increase (resulting from its formation from the precursors. because of some limitations. which showed more effect on yield-severity than on yieldconversion relationships. highest MCR content. the MAT unit gave a greater dry-gas yield than the riser reactor did for the same feed. HT-VIR < RZ ≈ HT-LCF .3. components in gasoline and the feed. at the same C/O. The drawback of the yield-conversion plot is that one may not be able to compare all feeds at the same conversion. This is probably attributed to the much longer contact time in MAT than in the riser reactor. because it consists of components that can be used as alkylation and petrochemical feedstocks. the MAT unit gives a higher dry-gas yield than the riser reactor. including dry gas. closed symbols) and a MAT unit (open symbols).. Figure 2b gives a much better comparison in dry-gas yield between the MAT unit and the riser reactor. had the highest content of polar compounds (15. RZ . which also have greater polars contents.

No. similar to the case of LPG. at the same conversion. based on feed analyses. Figure 3. Thus. the better the feed quality. the higher MCR in the feed. Assuming that all yields could be compared at 65 wt % conversion. Figure 4a illustrates that gasoline yield increased as the C/O ratio increased and tended to level off or decrease at high C/O ratios. Gasoline (C5 plus heavier hydrocarbons which boil at or below 216 °C) is the major and the most desirable product in FCC operation. Gasoline. One could only conclude from Figure 3b that. (b) Relationship between LPG yield and conversion for 10 feeds cracked in a riser pilot plant reactor (R. Figure 4b gave better comparison between the MAT data and the riser data. 1. at the same C/O ratio for the same feed.5. closed symbols) and a MAT unit (open symbols). In Figure 4a. respectively) at the same C/O ratio (Figure 1). Overcracking was observed for some feeds in Figure 4b. However. the riser reactor gave lower LPG yields than the MAT unit did. and between HT-DA and HTC. Vol. DA-BIT and DA-LCF (both rich in polar compounds) were slightly higher in the order than LCF. The result is given in Table 4. which shows the general increase in gasoline yield (prior to overcracking) with conversion. The aforementioned order is the closest to the pre-established ranking. HT-VIR > RZ > HT-LCF > HT-DA > HT-C > VIR > DA-BIT ≈ DA-LCF ≈ LCF. the 10 feeds could be ranked as follows: HCB . 18. The approximate maximum gasoline yield of a feed could be mathematically determined from Figure 4b. generally. with the former showing lower gasoline yields. (a) Relationship between LPG yield and C/O ratio for 10 feeds cracked in a riser pilot plant reactor (R. the greater the difference. Similar to the case of dry gas. the yield-conversion plot also gave a better comparison of the LPG yields between the MAT unit and the riser reactor than the yield-C/O ratio plot did. ≈ HT-LCF ≈ RZ ≈ HT-C < VIR < LCF. MAT gasoline yields compared favorably with those from the riser reactor.166 Energy & Fuels. . In contrast. relative to HT-LCF and HT-C. 3. At a constant C/O ratio of 7. The order was even closer to the expected trend. assuming that the yield was a parabolic function of conversion. were due to RZ and HT-DA having much-higher conversions (because they contained more MCR and/or sulfur. the approximate order of the yields was estimated as follows: HCB > HT-VIR > HT-LCF ≈ RZ > HT-DA ≈ HT-C > VIR > LCF ≈ DA-LCF ≈ DA-BIT.4. which showed much-higher MAT gasoline yields than those of the riser reactor at a C/O ratio of ∼5. closed symbols) and a MAT unit (open symbols). with the exception of DA-BIT and HT-DA. except that. the greater the LPG yield at high conversion. 2004 Ng et al. because of an unexpectedly high gasoline yield (probably related to the “easy-to-crack” polars) at low C/O ratios. The yield profile of DA-BIT was relatively flat. This order did not follow the expected trend that was based on feed analyses. This order agreed well with the apparent feed ranks. based on analyses. and with reference to Figure 4b. The maximum gasoline yield was probably a better indicator to reflect the feed rank. Based on this table.4. In Figure 3a. HT-VIR ≈ RZ > HT-LCF > HT-DA ≈ HT-C > VIR > LCF ≈ DA-LCF ≈ DA-BIT. the gasoline yields of the 10 feeds decreased in the following order: HCB . Note that the observed crossovers of yield curves between RZ and HT-LCF. generally. Again. except for the three DAOs. the maximum gasoline yield is more difficult to determine. based on feed analyses. gasoline yield is an important indicator to judge the feed rank.

which should increase at low conversion but decrease at high conversion.6 Figure 4. 1 Energy & Fuels. 3. Vol.5 48. Light cycle oil (LCO. Light Cycle Oil. the lower the LCO yield at high conversion. which outperformed HT-C. with a boiling point of 216-343 °C) is classified as an unconverted product that is not highly desirable in North America.3 51. 3. However. LCF.6. based on feed analyses. HTDA < HT-VIR < DA-BIT ≈ HT-LCF < DA-LCF < HT-C < LCF < VIR.1 37. HCO yield should be reduced to a minimum. At a constant C/O ratio of 7. closed symbols) and a MAT unit (open symbols). if the DAO feeds are excluded. LCF. and VIR.6 44. prob- . closed symbols) and a MAT unit (open symbols). the LCO yields of the feeds varied in the following order: HCB . The convex yield curves in Figure 6a show the gradual decreases in HCO yield as the C/O ratio increases. At a high C/O ratio (C/O ) 7). Heavy cycle oil (HCO. This order generally agreed with the feed quality. The cetane number of LCO from FCC is generally low and its sulfur content is high.5 43. HCO is usually burned as heavy fuel oil. especially when derived from an aromaticsrich feed. 1. (b) Relationship between gasoline yield and conversion for 10 feeds cracked in a riser pilot plant reactor (R. with the exception of the three DAOs. which currently has limited markets. 18.6 36. A yield-conversion plot (Figure 5b) gave a slightly better comparison of LCO yields between the MAT unit and the riser reactor than a yield-C/O ratio plot (Figure 5a). Thus. It is interesting to note the concave yield curves of HCB and HT-VIR as the C/O ratio increased from 4 to 7. this was consistent with the established feed rank with a few exceptions. unconverted product with comparatively high aromatics and sulfur. VIR ≈ HT-C < DA-LCF . the riser reactor gave lower LCO yields than the MAT unit. the following order was produced: RZ ≈ HCB . (a) Relationship between gasoline yield and C/O ratio for 10 feeds cracked in a riser pilot plant reactor (R. One could only conclude from Figure 5b that. in Figure 5b.Canadian FCC Feeds from Various Origins. with a boiling point of 343-525 °C) is an unwanted. RZ . as the C/O ratio increases. Generally.5 39.2 53. because of the enforcement of environmental regulations. generally.5. This order did not quite follow the expected trend. because of the enrichment of aromatics at higher conversion. DA-LCF ≈ HT-DA ≈ HT-C ≈ HT-LCF ≈ DA-BIT ≈ LCF ≈ HT-VIR < VIR. the better the feed quality. these DAO feeds performed better than expected. the poorer its quality.4. Assuming that both trend lines of HCB and LCF could be extended for comparison of LCO yields at a common conversion of 65 wt %. HT-VIR < RZ < HT-DA ≈ DA-BIT ≈ HT-LCF . 2004 167 Table 4. for the same feed.4. the variation of HCO yield was as follows: HCB . Maximum Gasoline Yield feed HCB HT-VIR RZ HT-LCF HT-C HT-DA VIR LCF DA-LCF DA-BIT maximum gasoline yield (wt %) 63. Heavy Cycle Oil. The lower the LCO yield. Figure 5a shows the general decrease in LCO for most feeds (except for VIR and LCF). There seemed to be no definite relationship between the riser and MAT data for the same feed in Figure 5a. No.8 37. Figure 5b shows the familiar convex yield curves. especially the two DAOs.

No. (b) Relationship between HCO yield and conversion for 10 feeds cracked in a riser pilot plant reactor (R. 18. (a) Relationship between LCO yield and C/O ratio for 10 feeds cracked in a riser pilot plant reactor (R. closed symbols) and a MAT unit (open symbols). Vol. closed symbols) and a MAT unit (open symbols). 1. (b) Relationship between LCO yield and conversion for 10 feeds cracked in a riser pilot plant reactor (R.168 Energy & Fuels. 2004 Ng et al. Figure 6. (a) Relationship between HCO yield and C/O ratio for 10 feeds cracked in a riser pilot plant reactor (R. Figure 5. . closed symbols) and a MAT unit (open symbols). closed symbols) and a MAT unit (open symbols).

except for HT-DA. too much coke can seriously poison the catalyst and overload the air blower during catalyst regeneration. Assuming that trend lines could be extrapolated. (iii) possible experimental error. (15) Voorhies.9 wt % higher in gasoline. LCO. based on its maximum gasoline yield. and (ii) the apparent feed rank was designed for gasoline yield by taking the relevant feed properties into consideration. feed qualities would be further considered. Some feeds did not show this expected observation because of (i) varying reaction temperatures. some unrecovered heavy product (HCO) that remained at the “cold” spot near the exit of the reactor was picked up later as coke at a higher reactor temperature (600 °C) during catalyst in situ regeneration). In FCC. The aforementioned discussions showed the exceptions in ranking for some feeds where the product yields did not follow the sequence established by the apparent feed ranks. the MAT gave higher coke yields than the riser at a given C/O ratio or conversion.6 wt % higher than that of RZ in gasoline and LCO. because of its high MCR content (2.15 (ii) insufficient data points in the low C/O ratio (<4) region.8 wt % lower in dry gas. causing excessively high temperature in the regenerator. notably because of the influences of MCR and nitrogen contents on the other feeds. an attempt was made to compare yields of selected feeds (in lieu of all feeds) on a practical and flexible basis. coke is necessary to supply heat for feed preheating and cracking. this order did not quite follow the expected trend that was based on feed analyses. because higher conversion required a higher C/O ratio or less feed per unit weight of catalyst. HT-VIR was just marginally better than RZ. This order generally agreed with the feed quality.9 wt % higher than that of RZ. At a constant C/O ratio of 7. 3. No. which shows consistently higher riser yields than those of the MAT at a given C/O ratio for the same feed. the maximum gasoline yields in Table 4 were used as a guide to group feeds for comparison at a high conversion.5. the order of delta coke yields was as follows: HCB . 18.2.9 wt % . Comparison of Apparent Feed Ranks with Exact Feed Ranks.5. RZ was more flexible to be cracked at conversions of >75 wt % (Figure 1) to reduce the yield difference in HCO. the lower the HCO yield at high conversion. A. Yield curves of feeds appear in a narrow band. 2. 2004 169 ably because of the conversion of heavy aromatics to coke. However. in regard to coke yield. for all feeds. with the exception of the three DAOs. Ind. the straight lines representing each feed intersected the y-axis (C/O ) 0). Again. LPG. because its maximum gasoline yield was 3.LCO.g. 1. Figure 7b shows the exponential increases in coke yield with conversion. DA-LCF . This order was only consistent with the ranking of the good-quality feeds. Similar to the LCO yields. In Figure 7a. and 2. Chem. the riser gave much-smaller delta coke yields than the MAT for each feed. DA-BIT. respectively. respectively. HT-C < VIR < HT-DA < LCF . HTDA .e. and 2. 0.0 and 2. which was 1. which poisoned the catalyst. Figure 7a shows the linear increase in coke yield as the C/O ratio increases. H2S on dry gas and MCR on coke) had been overlooked. Therefore. and 0. Yields at a conversion of 70 wt % also favored RZ.7. and 0. and 0. DA-LCF ≈ HT-DA ≈ HT-C ≈ HT-LCF ≈ DA-BIT ≈ LCF ≈ HT-VIR > VIR. HT-VIR ≈ HT-LCF < RZ ≈ LCF < HT-C ≈ VIR .9. At 65 wt % conversion. Thus. 318-322. Figure 6b shows the monotonic decrease in HCO with conversion. especially the DAOs. The following illustrates how the exact feed ranks were established: (1) HCB was indisputably the top feed. the better the feed quality. 1. If uncertainty in ranking still persisted after yield comparison. Given this information. These might be attributed to the following: (i) the common comparison bases. which is a low-value product. In both cases. because of its much-higher maximum gasoline and LPG yields and its lowest undesirable product yields. HT-C ≈ LCF ≈ VIR . HCO yields from the riser compared reasonably well with their corresponding counterparts in Figure 6b but not in Figure 6a. a C/O ratio of 7 and a conversion of 65 wt %. were not adequately chosen for these feeds. as decribed previously3 (i. at a common conversion of 65 wt %. In FCC operation.. based on the simple relationship HCO ) 100 . being a conventional crude VGO. Coke. which was realistic and achievable.e. it was more beneficial to crack RZ at conversions of >70 wt % to reduce the production of LCO and HCO while increasing the yields of gasoline and LPG. 3. LPG.4. Vol. DA-BIT. the conclusion reached from Figure 6b is that. Figure 7c shows the general decline of delta coke.. This order was also consistent with the feed quality. the coke yields varied in the following order: HCB . and HCO. However. giving intercepts that should be close to their MCR values.conversion . the order of coke yields at 65 wt % conversion was as follows: HCB . (4) HT-LCF was ranked the fourth-best feed. 1945. (3) RZ was the third-best feed. The magnitude of HCO yield for all feeds at a given conversion should be in the reverse order to that of LCO. and HCO. except for those of RZ and HCB. 1 Energy & Fuels. Figure 7b also had a better comparison of coke yields between the MAT unit and the riser than Figure 7a.0 wt % higher than that of HTLCF. There was no difference between HT-VIR and RZ. which affected empirical constants in Voorhie’s equation. which was 4. and coke. compared with other feeds at a given high conversion. Again. 37. The bad performance of LCF might be related to a high nitrogen content. Delta coke is calculated from the coke yield divided by the C/O ratio. HT-VIR < RZ ≈ HTLCF . (2) HT-VIR was ranked second. because of its maximum gasoline yield. delta coke (the amount of coke per unit weight of spent catalyst) is an important parameter to watch. relative to HT-LCF: 1.Canadian FCC Feeds from Various Origins. Properties with strong influences on other yields (e. the order of HCO yields of the 10 feeds was as follows: RZ ≈ HCB . except for HT-DA. Because gasoline was considered the most valuable product. HT-VIR ≈ HT-LCF ≈ RZ . i. This was further supported by other yields at a conversion of 75 wt %: the yield for HT-VIR was 2. with greater differences evident in Figure 7a. This is understandable. Again. with the increase in conversion.. DA-LCF ≈ HT-DA . DA-BIT ≈ DALCF. Eng.6 wt %).1. respectively. respectively.4.6 wt % lower than that in dry gas. because of its high MCR content. because its magnitude affects the heat balance.7.

(b) Relationship between coke yield and conversion for 10 feeds cracked in a riser pilot plant reactor (R. 2004 Ng et al.0 and 0. closed symbols) and a MAT unit (open symbols). higher than that of HT-C. Figure 7. Vol. which is the next-best feed (to be discussed later). closed symbols) and a MAT unit (open symbols).6. .0.5 wt % higher in gasoline and LCO. Additional yields at a conversion of 65 wt % were also considered: 5. closed symbols) and a MAT unit (open symbols). 18.170 Energy & Fuels. (a) Relationship between coke yield and C/O ratio for 10 feeds cracked in a riser pilot plant reactor (R. (c) Relationship between delta coke yield and conversion for 10 feeds cracked in a riser pilot plant reactor (R. 1. respectively. 0. No. and 1.

7 wt % higher in gasoline. respectively. 2. the yield of LPG would increase (see Figure 3b). and 1.5. HT-VIR ≈ RZ > HT-LCF > HT-C ≈ HT-DA > VIR > LCF > DA-LCF ≈ DA-BIT.1.2. LPG. it was more convincing to rank HT-C marginally higher.9 wt % higher in gasoline. and microcarbon residue (MCR). 0.5 wt % lower in dry gas. Microactivity test (MAT) yields of the 10 feeds compared better with riser pilot plant reactor yields at the same conversion than at the same C/O ratio. respectively. HCO. HCO. Partial funding for this research has been provided by Syncrude Research. respectively.1.4 wt % higher in LPG and LCO. the sequence of feed ranks would need to be reexamined. which was 1. The supporting evidence was based on yields at a conversion of 60 wt %: 2.0 wt % higher in maximum gasoline yield than the latter. respectively.1.4 wt % lower in LPG. dry gas gave the best comparison. it had the lowest gasoline yield and the highest dry gas and coke yields at conversions of >65 wt %. This compares favorably with the sequence that is based on apparent feed ranks. HCO. because its maximum gasoline yield is 5. (10) Consequently. 6 feed. 18. the Canadian Program for Energy Research and Development (PERD). in addition to significant increases in drygas and coke yields (see Figures 2b and 7b).6 and 1.7. total nitrogen. and 2. and nickel + vanadium content. Vol. HCO.0.5. and 0. respectively. Note that HT-C could be readily cracked at higher conversions (>65 wt %) for further reduction of LCO and HCO (see Figures 5b and 6b). and dry gas.7 wt % lower in dry gas.5. and 0. (8) LCF was classified as the next-best feed after VIR. and 1. 1. HCO. The validity of the apparent feed ranks thus established could be verified by cracking experiments. Minor discrepancies existed if the comparison was against the sequence that was established by dry gas or coke yields at a conversion of 65 wt %. and 0. This difference is large enough to ignore some of the unfavorable yields at a conversion of 65 wt %: 4.1. and 1. (9) The ranking of DA-LCF and DA-BIT could also be controversial.3. the apparent feed ranks were compared with the exact feed ranks that were based on individual product yields at a common realistically high conversion that was achievable. 4. 0. respectively. yields at a conversion of 65 wt % were slightly in favor of HT-C: 1.5. For final validation.4. and 1. which is tentatively the next-best feed after VIR. The results indicate a good conformity in sequence between the two sets of feed ranks. 2. and increased in gasoline and LPG (see Figures 4b and 3b). that HT-DA has much higher sulfur. the remaining DA-BIT was the worst feed. (7) VIR was ranked No. 1. and 3. and 0. 2.5. and LCO.2.1. and coke. and 2. LCO.3 wt % higher than that of LCF. in terms of absolute yields. EF0300996 .9 wt % less than the latter. 2004 171 0.2 wt % higher than that of DA-LCF. although its maximum gasoline yield was only 0. 0. this would also increase dry gas and coke yields significantly (see Figures 2b and 7b).7. 0. it was justifiable to rank DA-LCF slightly higher than DA-BIT as the ninth-best feed. After further consideration on feed qualities.7 wt % lower in dry gas. No. HCO. whereas coke showed the worst. and LCO. The authors wish to thank the Analytical Laboratory of the National Centre for Upgrading Technology (NCUT) for their technical support. (5) The ranking of HT-C and HT-DA could be controversial. Note that DA-BIT could be readily cracked at higher conversions (>65 wt %). Although the former had a maximum gasoline yield that was 0. This would reduce LCO and HCO (Figures 5b and 6b) but would not increase gasoline (see Figure 4b). MCR. (6) HT-DA was rated the No.1. and 2. Although the former was 1.5 wt % lower in dry gas and LCO. because some feeds were rich in sulfur or MCR. the order of apparent feed ranks for 10 feeds compared best with those based on either the maximum gasoline yields or gasoline yields at a conversion of 65 wt %. and nickel + vanadium contents than HT-C.Canadian FCC Feeds from Various Origins.4 wt % higher than that of VIR. MCR. gasoline. Among the six cracked products. and coke. 1. If the choice of process conditions or catalysts significantly altered the product distributions. considering its maximum gasoline yield. however. In view of their feed qualities. respectively. respectively. Conclusions Fluid catalytic cracking (FCC) feeds could be reasonably ranked based on their analyses. LPG. The justification was largely based on yields at a conversion of 60 wt %: 1. respectively. 7. In addition.8 wt % higher in gasoline. 0. 1 Energy & Fuels. and coke. In this study. respectively. with an emphasis on the concentrations of gasoline precursors. and 1. Note that the exact feed ranks were strictly based on MAT yields that were obtained in this study.4. Acknowledgment. Based on the aforementioned summary. and coke. LPG. LPG. the sequence of the exact rankings of the 10 feeds could be symbolized as follows: HCB . in terms of total sulfur. and the Alberta Energy Research Institute.4 wt % lower in dry gas. and 0. and coke. respectively. yields at a conversion of 65 wt % were somewhat less supportive: 0. the Alberta Research Council.3 and 0.3 wt % higher in gasoline. and coke.

Sign up to vote on this title
UsefulNot useful

Master Your Semester with Scribd & The New York Times

Special offer for students: Only $4.99/month.

Master Your Semester with a Special Offer from Scribd & The New York Times

Cancel anytime.