8/3/2019 41_3_ORLANDO_08-96_1079

1/5

FISCHER-TROPSCH lNDIRECT COA L LIQUEFACTION DESlGNlECO NOM lCS -MILD HYDROCRACKING VS. FLUID CATALYTIC CRACKINGGerald N. Choi, Sheldon J. Kramer and Samuel S. Ta m(Bechtel Corporation, San Francisco, CA)Joseph M. Fox 111 (Consultant)William J. Reagan (Amoco Oil Company, Naperville. IL)

Keywords: Fischer-Tropsch, Economics, Fluid Catalytic CrackingIn order to evaluate the economics of Fischer-Tropsch (F-T) indirect coal liquefaction, conceptual plantdesigns and detailed cost estimates were developed for plants producing environmentally acceptable, high-quality. liquid transportation fuels meeting the Clean Air Act requiremen ts. The designs incorpo rate thelatest developments in coal gasification technology and advanced (F-T) slurry reactor design. In addition,an AS PEN Plus process simulation mo del was developed to predict plant material and en ergy balances,utility requirements, operating and capital costs at varying design conditions. This paper co mp ares mildhydrocracking and fluid catalytic cracking as alternative methods for upgrading the F-T w ax.FISCHER-TROPSCH PLANT DESIGNPlant ConfigurationsFigure I is a block flow diagram showing the overall process configuration for the original design usingmild liydrocracking. The plant contains three main processing areas. Area 100generates a clean syngasfrom Illinois No. 6 coal from the B urning Star mine. Area 200 is the Fischer-Tropsch (F-T) synth esisarea, and Area 3 00 is the product upgrading and refining area. Areas 100and 200 are identical for boththe mild hydrocrack ing and fluid catalytic cracking (FCC) cases. Utility plants and storage requirem entsi n the offsites were estimated, hut they are not detailed here.I . Area 100- Svn aas Production -- Synthesis gas is generated in Shell gasifiers from ground, dried coal.Processing of the raw synthesis ga s from the gasifiers is conventional, with Wet Scrubbing followedby single-stage COS /HCN Hydrolysis and Cooling. Acid Gas Removal by inhibited amin e solutionand Sulfur Polishing. Sour Water Stripping and Sulfur Recovery units are included in this area.2.PD- This area includes the Fischer-Tropsch Synthesis,C 0 2Removal, Recycle Gas Compression and Dehydration, Hydrocarbon Recovery by deeprefrigeration, Hydrogen R ecovery and Autothermal Reforming. The Hyd rocarbon Recovery unit

includes deethanization, depentanization. fractionation and an oxygenates wash colu mn . At lowHd CO ratios, COz is the prinlary byproduct of the F-T reaction (using a iron based cataly st) so a largeCO2 removal unit is required. The A utothermal Reformer converts the unrecovered lighthydrocarbons to additional syngas which is recycled back to the F-T synthesis reactors.3. Area 300 - Product UugradingHydrocracking Design -_Figure 2 is a block flow diagram of Area 300 of the mild hydrocrackingdesign. This area contains eight processing steps; 1 ) wax hydrocracking, 2) distillate hy drotreatin g3) naphtha hydrotreating, 4) naphtha reforming, 5) C4 isomerization, 6 ) C5/C 6 isomerization,7) C3/C4/C5 alkylation, and 8 ) saturated gas processing and product blending. Th e hydrocracked andhydroueated naphthas are catalytically reformed to produce an aromatic gasoline blendingcomponent. Th e lighter materials are isomerized and alkylated to produce a high quality gasolineblending stock. Purchased butan es are required to alkylate all the available C3/C4 /C5 o lefins.

Fluid Caralyric Cracking Cases -- Figure 3 is a block flow diagram of Area 300 for the FluidCatalytic Cracking (FCC) cases. Two FC C upgrading cases are cons ide red one uses a beta zeolitecracking catalyst, and the other uses an equilibrium U SY cracking catalyst. In these cases , the waxhydrocracker is replaced by a FCC unit, a MTBE (methyl-tertiary-butyl ether) plant, and aNExTAM E (m ixed C5/C6/C7 ethers) plant. The other seven processing plants are unchanged.The F-T slurry reactor is a bubble column reactor in w hich the slurry phase is a mixture of liquid hydro-carbons (molten wax) and catalyst. Synth esis gas provides the agitation necessary for good mixing andmass transfer of the reactants and products between the two phases. The slurry bed reactor design waschosen over a fixed bed reactor design based on an earlier DOE sponsored Bechtel study'.'. Th e reactordesign is based on M obil's two -stage, slurry reactor pilot plant studies'. Th ese results were the basis forthe yield correlations contained in the F-T slurry bed reactor computer model used in this study4. Detailsconcerning the overall design basis, process selection, and costs have been reponed'.Product and Bvuroduct YieldsA Fischer-Tropsch liquefaction facility can produce a wide variety of products of various qualitiesdepending on the m ethod used to upgrade the F-T ax.In the mild hyd rocracking case, the facility produces C3 LPG , an upgraded C5-350 "F naphtha and a350-850 "F distillate. Liquid sulfur also is produced by the syng as production area. Th e hydrocarbonproducts have no measurable sulfur or niuogen contents. Oxygen is removed to less than 30 ppmv.There are virtually no aromatics in the distillate. Both the naphtha and distillate products have low

1079

8/3/2019 41_3_ORLANDO_08-96_1079

2/5

residual olefin concentrations. Th e diesel fraction has a very high cetane number. Th e jet fuel and heavydistillate fractions have low smok e points. The naphtha product is a mixture of C3/C4 /C5 alkylate, C5/C6isomerate and catalytic reformate. It is basically a raw gaso line with a clear (R+M)/2 octane of about 88.In the FCC u pgra ding cases, the facility produces a propylene product in addition to those produced in thehydrocracking case. Methanol is purchased and reacted with the tertiary C4, C5, C6 and C7 olefins toproduce M TBE and a mixed C5/C6/C7 ethers stream from the NExTAME etherification plant. Th e etherstreams are mixed with the 0 - 3 5 0T naphtha to form an oxygenated gasoline blending com ponent whichcontains significant amounts of olefins and has higher octane numbers than in the mild hydrocarackingcase. Except for a low er pour point, the distillate fraction has about the same properties as that producedin the hydrocracking case.PROCESS SIMULATION MODELThe ASPEN Plus process flowsheet simulation model predicts the effects of key process variables on theove rall material and utility balances, operating requirements and capital costs. It was developed as aplanninghesearch guidance to ol for use by the DOE and its subcontractors to explore, evaluate and defin eadditional prom ising are as for Future research in the production of liquid transportation fuels. Th e modelis not a detailed plant design tool although it contains some design features. Th e F-T synthesis loopdesign is modeled in some detail, and Bechtels slurry bed F-T reactor sizing and yield models have beenincorporated into the AS PE N model. For the other plants, only overall yield, utility requirements andcapital costs are estimate d. Individual plant costs are prorated on capacity using cost-capacity exponentsin conjunction with minimum and maximum single train capacity limits.All plants in the three ma in processing sections are simulated either by stand-alone user Fortran blocks ora combination of ASPEN Plus process blocks and user Fortran blocks. Material balances, as well asutility consumptions, operating personnel requirements and lSB L costs for each plant are generated. T heoffsites, eng ineering and co ntingenc y costs are estimated as a percentage of the processing plant cos ts togenerate the total installed cost of the facility. Detailed discussions of the ASP EN mod el developmentand simu lation results have been p resented in three separate papers; the last of which discusses thebeneficial effcct of treating the F-T reactor vapor products in a close-coupled ZSM-5 reactor as analternative product upgrading scheme.A linear programming (LP) model of a typical PADD I1 refinery was developed using BechtelCorporations proprietary Process Industry Modeling System (PIMS) to assess the values of the F-Tproducts from the mild hydrocracking case9. With the ASPEN and LP model results, a discounted-cash-flow analysis was carried out under a given set of financial assumptions to calculate the cost of F-Tproduction fo r a 15% eturn on investm ent. Results are presented in terms of a Crude Oil Equivalent(COE) price which is defined a s the hypothetical break-even crude oil price at which the liquefactionproducts are competitive with those produced from crude oil. Table I summaries the overall simulationmodel results for the mild hydrocracking case and the two FC C upgrading cases; one of which uses a betazeolite cra cking catalyst, and the oth er uses an equilibrium USY cracking catalyst.FLUIDIZED-BED C ATA LYTIC CRACKING OF F-T WAXFischer-Tropsch wax can be readily cracked in a FCC unit under normal petroleum feedstock operatingconditions, as demonstrated by the Amoco Oil Company. The product is rich in C4 to C 7 reactiveolefins which are valuable for oxygen ates production. The hydrocracking design and the ASPEN Plussimulation m odel were mo dified to use FC C instead of mild hydrocracking for upgrading the F-T wax. Inaddition, both a M TB E plant and a NExTA ME (mixed C5/C6/C7 ethers) plant were included in A rea 300.Both ether plants contain an associated selective hydrogenation unit to saturate diolefins i n the feed.Based on the Amoco d ata, two F CC cases were considered; a beta zeolite catalyst case and an equilibriumUSY ca talyst case. Although not in widespread comm ercial use, beta zeolite catalyst was selected forcomparison with an equilibrium USY catalyst since it produces more.olefins which can be converted toethers for use as reformulated ga soline blending components. The propylene is purified and sold. Thebutenes are sent to the MTBE (methyl-tertiary-butyl ether) plant in which the isobutene is converted toMTBE , and the normal butenes are passed through to the alkylation unit. T he C5, C6 and C 7 olefins aresent to a NExT AM E unit which con vert s most of the C5 olefins, less of the C 6 olefins, and still less of theC7 olef ins to ethers. This design produces significantly more gasoline blending components at theexpense of distillate production than the hydrocracking design.Amoco found that the F-T wax cra cks so easily i n a FCC unit that it does not produce enough coke tomaintain the unit in heat balance. On e solution to this problem is to supply additional fuel, sometimescalled torch oil, to the regenerator to he at the regenerated catalyst to a high enough tem perature to sustainthe cracking operation. In a conventional petroleum refinery wheneve r torch oil is required, a low-valueheavy material is used. How ever, in this stand-alone situation, no high-boiling low-valued streams areavailable from sources othe r than the FC C unit. Therefore, the heaviest portion of the potential distillateproduct is used as torch oil and burned in the regenerator to maintain the FCC unit heat balance.Table 1 com pares the model results for the wax hydrocracking case with those for the two FCC w axupgrading cases. In the hydrocracking case, approximately 3100 bbldday of butanes are purchased and

1080

8/3/2019 41_3_ORLANDO_08-96_1079

3/5

isomerized for alkylation unit feed since the F-T reaction does not produc e a sufficient amou nt to alkylateall the C3/C4/C5 olefins. The two FCC cases produce more olefins for alkylation unit feed, and as aresult, still more butanes have to be purch ased. Both FCC cases produce significantly more gasoline thanthe hydrocracking case. This additional gasoline is produced at the expens e of distillate production whichis reduce d by over 60% esulting in gaso line to distillate ratios of over 4/1 co mpared to a 0.97/1 ratio forthe wax hydrocrac kingcase. In addition, the gasoline produc ts are of betrer quality. They now containoxyg enates and ha ve highe r octanes, lower Reid vapor pressures, and contain less arom atics.

\

\

In addition, there is another significant difference between the beta zeolite and eq uilibrium USY catalystcases which can influence the choice between them. Both FCC u pgrading cases consume m ethanol andprodu ce a prop ylene product; whereas neither component is present in the hydrocrackingcase. The betazeolite catalyst case cons umes about 54% more methanol and produ ces about 57% more prop ylene thanthe equilibrium US Y catalyst case. For 1995, propylene prices have ranged betwe en 350 and 495 $/s-ton.and m ethanol prices have ranged at least between 0.47 and 1.35 $/gal (141 to 40 6 $/s-ton)12. The aboveexpected COE prices were calculated using the average propylene price of 422.5 $/s-ton and the averagemethanol price of 0.91 $/gal (273 $/wo n). The following table shows the effect of these variations in themethanol and propy lene prices on the CO E price for the two FCC upgrad ing cases.

Most MostOotimistic Averape PessimisticPropylen e price, $/ton 495 422 350Metha nol price, $/ton 141 273 40 6COE Prices. $/bblBeta zeolite catalyst FCCEquilibriumUSY atalystupgrad ing case 33.2 34.5 35.8FCC upgrad ing case 33.7 34.5 35.3

Thus, the COE price for the beta zeolite catalyst FCC upgrading case can vary betw een 33.2 in d 35.8$hb l , a r ange of 2.6$/bbl. For the equilibrium USY atalyst case, the COE price can vary between 33.7and 3 5.3 $/bbl, a smaller range of only 1.6 $/bbl. This leads to the conclusion that with all other factorsbeing the same, the equilibrium USY atalyst FCC u pgrading case is less risky based on rece nt pricessince it minim izes the effect of price variations on the expected COE price.In conclusion, FCC upgrad ing of the F-T wax appe ars o be preferable to upgrad ing it by mildhydro cracking. Recent wide variations in methanol and propylene prices have been shown to causesignificant fluctuations in the COE price of the F-T liquefaction products. Add itional petroleum refinerymodeling studies are needed to d etermine more accurately the values of the FC C up graded products and todefine a more reliable COE price for both of the FCC indirect F-T liquefaction cases. It is expected thatthis will impro ve the econom ics of u pgrading the F-T wax by fluid catalytic cracking.ACKNOWLEDGMENTBechtel, along with Am mo who w as the main subco ntractor for a major portion of this study, expressesour appreciation to the D OYP ittsburgh Energy Technology Center for both technical guidance andfinancial funding u nder Con tractNo. DE-AC22-91PC90027

1081

8/3/2019 41_3_ORLANDO_08-96_1079

4/5

REFERENCESI . Fox, I. M., Slurry vs. Fixed-Bed Reactors for Fischer-Tropsch and M ethanol. DOE C ontract No.DE-AC2 2-89PC89876, Final Report, June 1990.2. Fox, I., M., Fischer-Tropsch Reactor Selection, CatalysisLetters1 1990) 281-292.3. Kuo, . C., Slurry Fischer-Tropschhlobil Two-Stage Process of Converting Syngas to High OctaneGasoline, DOE C ontract No., DE-AC22-80P C30022, Final Report, June 1983; and Tw o-Sta geProcess for Conversion of SynthesisGas o High Quality Transportation Fuels, DOE Contract No.DE-AC 22-83K60 019. Final Report, October. 1985.4. Fox, J. M. and Tam . S. S., Correlation of Slurry Reactor Fischer-Tropsch Yield Data, Topics in

Catalysis2,(1-4),285-300 (1995).5. Choi, G. N., Tam, S.S.,Fox, J. M., Kramer, S. J. and Marano, J. J., B aseline Design/ Econom ics forAdv anced Fis cher-Trop sch Techno logy, DOEYProceedings of The Coal Liquefa ction and Ga sConversion Contractors Review Conference, September 27-29, 1993.6. Choi, G. N., Tam , S. S .,Fox, J. M.. Kram er, S. J. and Rogers, S.,Process Simulation Model forIndirect Coal L iquefaction using S lurry Reactor Fischer-Tropsch Technology. Symposium onAlternative Routes for the Production of F uels, ACS National Meeting, Washington, D. C., August21-26, 1994.7. Cho i. G. N., Tam , S. S., Fox, J. M., Kramer, S I. and Marano, J. J., Process Design SimulationModels for Advanced Fischer-Tropsch Technology, DOElProceedings of The Coal Liquefaction andGas Conversion C ontractors Review Conference, September 7-9, 1994.8. Choi, G. N., Kramer, S. J. Tam, S . S . and Fox, J. M., Simulation Models and Design for AdvancedFischer-Tropsch Technology, DOFJP roceedings of The Coal Liquefaction and GasConversionContractors Review Conference, August 29-3 I , 19959. Marano, J. J., Cho i, G. N. and Kramer, S. I. , Product Valuation of Fischer-Tropsch Derived Fuels,Symposium on Alternative Routes for the Productionof Fuels, ACS National M eeting, Washington,D. C., August 21-26, 1994.

of Fischer-Tropsch Liquids to H igh Value Transportation Fuels, DOE C ontract No. DE-AC22-91PC90057 , Final Report, August 1994.January 1995.Fuel New s, Vol. VII, No. 6, Sept. 4, 1995

IO . Reagan, W. 1..Nicholas, J. I., Hug hes, R. D., and M. M. S chwartz, The Selective Catalytic Cracking

I I . U S . Department of Energy, Energy Information Administration, Annual Energy Outlook 1995.12. Propylene prices are from the Chemical Marketing Reporter. Methanol prices are from H arts Oxy-

TABLE IComparison of Mild Hydrocracking and FluidCatalytic Cracking Upgrading of Fischer-Tropsch Wax

Wax U pgrading ModeFCC C atalyst TypePlant Inout:ROM Coal, TSD (MF)Methanol, TSDMixed B utanes, BSDElectric Power, Mwatts

Gasoline, BSDDistillate, BS DPropylene, BSDLiquid Propane, BSDSulfur, TS DSlag, TSD (MF)

Plant Ournut:

Total Net C5+ Production, BSDTotal Net C4+ Production, BSDBlended Gasoline Prouerties:Research Octane Num berMotor Octane Number(R+M)/2 Octane NumberReid V apor Pressure, psiBenzene, wt%Aromatics, wt%Olefins, wt%Oxygen, wt%Blended Distillate Prouerties:Pour Point. TCetane IndexInstalled Plant Cost, MM$ in mid-I993

HvdrocrackingHydrocrackingMethod of U omadinn the F-T Wax

Feel FCC2FCC FCC

18575031 IO54

23943246860192256022444862945519

. 90.986.188.55.00.328.10.00.0

- 28742964

1 0 8 2

BetaZeolite1857532 15204583972397645060157356 022444948744283

96.888.992.84.70. I11.012.73.3- 4 074

2987

EquilibriumUSY1857520 94327563995093413215158456 02244492974497095.887.891.84.70.I13.915.52.1- 4074

2978

P

c\JI

8/3/2019 41_3_ORLANDO_08-96_1079

5/5

\

I

1083