9.1.4 Summary of Jet Fuel Characterization
Testing and analyses were performed to provide preliminary guidance regarding the potential suitability for using a synthetic fuel produced from coal and biomass-derived feedstocks in aviation applications. The approach followed recommended protocols for evaluation and certification of synthetic fuels for commercial and military applications and included evaluation of fuel specification and limited Fit-For-Purpose (FFP) properties of the neat synthetic fuel and a blend with petroleum-derived aviation fuel. It was determined the neat synthetic fuel could not be used as a direct ‘drop-in’ aviation fuel as several properties did not conform to required commercial and military fuel specification requirements. However, it was determined the potential exists for using the synthetic fuel as a blending feedstock with petroleum-derived fuels.
A 20% by volume blend of the synthetic fuel was prepared with a ‘nominal’ Jet A fuel and evaluated for potential suitability for use. Efforts included characterizing the chemical composition of the fuel blend, performing ASTM fuel property specification testing, and evaluating select FFP properties (due to limited available volume of the synthetic fuel blend). The composition of the fuel blend was similar to typical aviation fuels, with a slight increase in higher molecular weight cycloparaffins and cycloaromatics. The fuel blend satisfied all physical property requirements in current commercial and military aviation fuel specifications. Limited FFP testing demonstrated that the fuel blend had operational characteristics consistent with typical aviation fuels. Overall, the results indicate the potential exists for using the synthetic fuel as a synthetic blending feedstock for aviation applications. However, evaluation of other FFP properties identified in aviation fuel certification processes must be performed to provide improved guidance regarding suitability for use. These results would allow determination of subsequent required testing (e.g., Turbine Hot Section, Component-Level, Engine Testing) for the synthetic fuel blend to ultimately be consider for OEM and FAA approval.
9.2 Characterization of Battelle-CTL-Derived Neat Synthetic Fuel for Use as a Diesel Fuel
The distillate from Intertek Run #2 was fractionated into a synthetic diesel fuel. A portion of the gasoline-boiling portion of the distillate was removed to increase its flash point above 52°C. This report summarizes the evaluation of the Battelle synthetic fuel for diesel engine applications. This evaluation focused on testing the specification properties, as well as limited fit-for-purpose (FFP) properties, of the Battelle synthetic diesel fuel. The tests were performed on the unblended synthetic liquid; no conventional petroleum diesel was present.
9.2.1 Relevant Diesel Fuel Specifications
The ASTM standard specification for D975 S15 diesel fuel was chosen as the basis for comparison. This diesel fuel is a general-purpose, middle distillate fuel for use in diesel engine applications requiring a fuel with a maximum of 15 ppm sulfur. It is especially suitable for use in applications with conditions of varying speed and load.
9.2.2 Adjustment of Fuel Volatility
Because the presence of light hydrocarbon can affect other requirements such as flash point, we decided to remove by distillation the light fraction that has boiling point below 160 °C. This light fraction corresponds to 4.5 percent volume of the liquid product. Figure 71 is the profile of simulated distillation of total product by ASTM method D2887 and distillation profile (by ASTM method D86) of the product after removing the fraction with boiling point below 160°C.
9.2.3 Analysis of Battelle CTL Synthetic Fuel
According to ASTM D975 S15 specification for No. 2 diesel fuel, twelve requirements must be met for an acceptable diesel fuel. Table 62 reports the results of analysis performed on the Battelle synthetic fuel and compares these results with the ASTM D975 S15 requirements.
Figure 71. Profile of simulated distillation of total product (D2887 total product) and distillation of product after removing the fraction with boiling point below 160°C (D86 after distillation of light fraction).
Table 62. Comparison of Battelle Synthetic Diesel with ASTM Standard Specification for Diesel Fuel D975 S15
9.2.4 Discussion
The synthetic fuel meets all the ASTM D975 requirements except for conductivity, carbon residual, and T90.
9.2.4.1 Conductivity
The conductivity of Battelle synthetic is below the D975 requirement of 25 pS/m.
However, the conductivity can easily be increased though use of a conductivity additive.
9.2.4.2 Carbon Residual
The D975 requirement for carbon residue is 0.35 mass percent max according to ASTM D524. The ASTM D524 test method covers the determination of the amount of carbon residue left after evaporation and pyrolysis of an oil and it is intended to provide some indication of relative coke-forming propensity. The carbon residue of the Battelle synthetic fuel is 1.07 wt.% and therefore it does not meet the expectation. It is possible that further filtration of the Battelle Syncrude with a 0.2μm filter and/or blending with commercial diesel can drop the residual carbon to less than 0.35%.
9.2.4.3 Maximum T90 Temperature
The D975 requires the 90 percent volume recovered temperature (T90) to be 282-338°C. The T90 temperature for the Battelle synthetic fuel is 340.5 °C. which is 2.5 °C higher than the range allowed. A simple adjustment of the diesel by blending of commercial fuel or additional hydrocracking is projected to resolve this problem.
Alternatively, the carbon residual and T90 temperature requirements can be met by removing a heavy fraction from the fuel. For another similar synthetic fuel, we demonstrated that removing the fraction with boiling point above 337 °C eliminated the T90 temperature issue and also reduced the residual carbon concentration to 0.1 weight percent.
9.2.5 Additional Discussion of Battelle Synthetic Diesel Fuel Properties
In addition to the above requirements, several other analyses were performed.
In addition to the above requirements, several other analyses were performed.
9.2.5.1 Density
The density of the Battelle synthetic fuel is 0.890 kg/L. This relatively high density is a plus. Customers who normally buy fuel by volume (such as in gallons) will get more energy for their money. See also heat of combustion comment, below.
9.2.5.2 Heat of Combustion
The heat of combustion of the Battelle synthetic fuel is 44.9 MJ/kg. This relatively high heat of combustion is also a plus. Customers who buy fuel by weight will get more energy.
9.2.5.3 Molecular Structure Information Derived from GCxGC-MS
The GCxGC-MS data (see Table 63) show the synthetic fuel is comprised primarily of cycloparaffins (69.1%) and aromatics (19.4%). The concentrations of di-/tri-cycloparaffins of 50.7 wt. % are significantly higher than observed in typical diesel fuel 27 wt. % [C.A. Baldrich]. It is likely that the relatively high concentrations of di and tri-cycloparaffins are responsible for both the high fuel density and the high carbon residual. The concentration of di-/tri-cycloparaffins can be adjusted either by selective solvent extraction, by blending with commercial fuel or by selective hydrocracking of di-/tri-cycloparaffins.
Table 63. Hydrocarbon Type Analysis of Synthetic Fuel by GC x GC – MS
9.2.5.4 Carbon-Hydrogen Mole Ratio
The carbon-hydrogen mole ratio of a petroleum diesel is usually about 1.8. For example, the base fuel used in the Diesel Emission Control Sulfur Effects program, which was intended to be representative of diesel fuels used in the U.S., had a carbon-hydrogen ratio of 1.85. The ASTM D5291 shows lower H/C mole ratio of 1.71 for Battelle synthetic fuel. This probably related to high concentration of di-/tri-cycloparaffins.
9.2.6 Summary of Diesel-Fuel Characterization
Battelle has successfully demonstrated that using a novel biomass-derived hydrogen donor under relatively mild process conditions, a high proportion of the mass of coal (more than 85 percent) can be made into a synthetic diesel fuel that has the potential to meet all ASTM specifications.
10.0 CTL PLANT DESIGN AND ECONOMIC ANALYSIS
10.1 Introduction
Based on data on solvent refining of coal published by Longanbach and Chauhan of Battelle in late 1970s [1], and confirmed by Battelle project team member Quantex, coal can be dissolved quickly (˂10 minutes) at mild conditions (~400°C and 500 psi) with addition of only 0.3–0.5% hydrogen, by weight of coal, which increases the H/C molar ratio from about 0.80 to 0.86. In over 140 experiments at Battelle, it was proven that solvents can be engineered to alter the nature and quantities of the cyclic/aromatic species with desired hydrogen-donor properties. Based on the low hydrogen-addition requirements, it was demonstrated that as little as 10% of H-donor bio-solvent based on weight of coal is sufficient without requiring any gaseous H2.
The selected bio-solvent is important. Depending on the specific bio-solvent, the viscosity and conversion of coal can be affected. A chosen bio-solvent mixure was selected by Battelle. The selection is identified as a proprietary mixture and is called BS-41A hereafter. It has the following properties:
- The system allows solubilization of >85% of the moisture- and ash-free (MAF) composition of coal in less than 10 minutes at 400°C (>90% has been observed with this particular material and reaction systems for the plant were designed to give >90%).
- All components of the mixture are available, having both large (>100 MMT/yr available production in the US) and being available for purchase on the open market.
- The mixture can be economically used in a syncrude production process.
Coal, biomaterials, coal tar distillate, and recycled distilled syncrude product are mixed together. The product is a material that is solvated coal, undissolved coal and coal ash as solids, and a heavy-oil product. The recipe which was used in the engineering evaluation is shown in Table 64 below.
Table 64. Laboratory Formulation Used as Basis for Design
The products from the reaction are a gas and a solid/liquid mixture. Residual coal (MAF basis) was found to be in the 5-15 percent range (85-95% conversion). In the shorter runs, conversion was not reduced significantly, but gas production (probably decaroboxylation of the solvents) was reduced.
The heavy-oil product and solids are separated to remove solids. Liquid recovery from this run was 89% of the total mass fed. Gas produced was estimated at 2-3% of the total material fed after compensating for water. The reactor produces gases which are primarily H2, CH4, H2S, CO, and CO2. The gas contained less than 2 percent of hydrocarbon vapors.
Six pilot runs were made at Quantex to produce materials for use. The liquid from the final run, #6B, produced material which was closest to the products from laboratory reactions. The Quantex 6B product was made with limited recycle, so the material was heavier in start-up coal tar distillate (CTD) than expected from a final product. Hence, density is higher and C/H ratio is lower. That product has been used for characterization, hydrotreating, and distillation to final fuel products. These products have been blended to produce a fuel which can be blended with Jet A or used as a diesel product. Because the reactors were not optimized for these runs lower conversions than measured in the laboratory were achieved. Because material balances for the Quantex runs were not able to be closed adequately (for example, gaseous products were not collected, so could not be analyzed or quantified), the laboratory runs were used for the modeling results rather than the Quantex runs. The Quantex evaporators were also found to be unable to give a 500°C cut for the final product, which was the then-desired end point.
Intertek was contracted with and conducted re-evaporation, hydrotreating, and final distillation (see Section 8). The testing at Intertek provided feasibility that the <500°C material which is produced can be successfully hydrotreated and used as a blending stock for Jet-A or used as diesel fuel.
Process and economic modeling is based on Quantex Run #6B, Intertek Phase 2, as represented by small-scale run numbers 139 and 140 for syncrude production. The hydrotreating modeling is not done rigorously; it was scaled from a prior bio-based hydrotreating model and was done only to the extent necessary to produce energy and hydrogen estimates for the hydrotreater. Hydrogen production was done similarly from DOE references ([19], [20]). Block modeling was done for the overall material balance of the hydrotreater.
10.2 Process Flow Diagram
A block diagram of the process is shown in Figure 72. The nine components of the process are listed below. The direct process train elements (100, 300-600, and H100-H102) have been tested. The processes that are support blocks have not been tested (e.g., tank storage, process heater, scrubber) in the pilot testing. All syncrude and hydrotreater blocks are discussed below.
The reacted solids are sent from hydrocyclones to a centrifuge for further heavy oil recovery. The liquids are sent to a multi-effect evaporator and are separated into a heavy oil and a syncrude. These will be referred to by these names in the rest of this report.
The syncrude fraction can be varied somewhat base on desired products. If optimized to fuel, the top cut would be split at about 500°C. If, however, the bottom cut was desired for production of polyol, the cut would be made at about 380°C. This represents about 7.1% less oil recovered (Table 45, extrapolated).
Figure 72. Simplified flow diagram for the Battelle CTL process
The process is discussed in greater detail below. The paragraphs are numbered to represent the figure and the list of processes above. Drawings for Systems 100-700 are provided in Appendix B.
100.Coal receipt and preparation. Coal is received, piled, ground, and dried to about 2% moisture and ground to -8 mesh. Water from drying is condensed and further used as cooling water.
200.Process Tank Farm. B41A components are received and placed in heated tanks. Coal tar distillate (initial feed or makeup) is received and placed in heated tanks.
300.Mixing of coal, bio-solvent, and coal tar distillate. Recycle material, the components of BS-41A, makeup CTD, and dry crushed coal are mixed. The slurry is preheated to a maximum of 150°C to remove any additional free water. The heated material is then fed to the digester.
400.Digestion. The feed material is processed at 400°C. The coal is digested at about 28.6 Bar and residence time of 10-30 minutes. The slurry is partially depressurized and the lights are condensed from the digester exhaust. The exhaust gases (sulfur-containing gases, methane, CO2, and hydrogen are the principal constituents) are vented to the process furnace. Modeling was conducted to determine the size and number of reactors in process step 400 using the 10-minute and 30-minute residence time results. Batch reaction results were converted to continuous, stirred tank reactor (CSTR) assuming a first-order reaction rate. A series of four 3-reactor trains were defined to allow for three operating at a time and an extra train installed for maintenance to assure 90% operability.
500.Solid/liquid Separation. The hot, partially depressurized slurry is mixed with the evaporator heavy liquids and passed through two stages of hydrocyclones to remove the majority of solvent and light materials. The hydrocyclone system is operated with counter-current flow of >500°C product to recover most of the light material from the solids stream. A centrifuge is used to remove additional liquid from the cake. The liquid is passed to a fractionator and the solids are sent off-site for use.
600.Thermal separation. A vacuum evaporator is used to remove the low cut material. The highly volatile material is removed through the vacuum pump and also is fed to the process heater; lighter syncrude stream (ranges from 380 -500°C atmospheric boiling point [ABP]) is removed from the fractionator overhead, and a 500°C+ oil stream is removed from the bottoms of the evaporator. The digester exhaust is combined with syncrude evaporator lights from the vacuum pump and sent to the fired heater. The liquid light evaporator fraction is a syncrude. The heavy fraction is a separate product. The solids from the centrifuge and hydrocyclones will have all of the ash and unconverted coal as well as being rich
700.Product liquids are stored in tanks for shipment.
800.Utilities. The waste gas (non-condensables from the evaporator vacuum pump exhaust plus the digester gases) are burned in a heat-recovery furnace. In this analysis it is assumed they all are burned at an atmospheric-pressure furnace. Options such as O2 combustion at higher pressure were not evaluated. This furnace is used to provide all of the process heat energy during normal operating periods. The digester gas is supplemented by natural gas during start-up and other outages. The furnace will be fitted with a limestone scrubber to remove SO2 from the product gases. Hot oil (modeled preference) or molten salt is circulated through the heater. A cooling tower will be required to cool most process streams. Air cooling is used in selected areas where the high temperatures might cause equipment stress if water were used. Electrical and other miscellaneous utilities (e.g. compressed gas for instruments) are provided.
Hydrotreater. The syncrude is hydrotreated in a trickle-bed, fixed bed hydrotreater with hydrogen and naphtha is taken off in the hydrotreater off-gas. The hydrotreated product can then be used as a diesel stream. Alternately, Jet-A boiling-range material can be distilled off and the remainder used as a heavy oil (diesel, other fuel purposes are what is envisioned in the cost evaluation section later). Off-gases (H2, NH3, H2S, H2O, CO2, lights) are treated by PSA or Selexol-type process to recover H2. The off-gases are incinerated for energy recovery.
Hydrotreating, hydrocracking, and distillation are expected to be conducted either in a refinery or as a larger-scale process where the products from several liquefaction sites are put together. It is also entirely feasible to run liquefaction at a scale similar to the Fischer-Tropsch units that are currently used to make synthetic paraffinic distillates such as synthetic paraffinic kerosene (SPK).
The elements of the hydrotreater system are as follows:
H100. Reaction. Crude is fed with hydrogen through a mixed-fluid reactor. In the early part of the reactor, oxygen, sulfur, and nitrogen are converted into vapor, H2O, gaseous H2S, and NH3.
H101. Reforming and Compression. Hydrogen is reformed (preference) or received and is compressed to operating conditions.
H102. Gas/liquid separation. A high-pressure separator is used to separate the liquid and gas. The liquid flows to the distillation process. The gas is treated to remove sulfur, nitrogen, and water, and is mostly recycled. A bleed stream will be used to provide some economic benefit for process heating.
H103. Product distillation. The liquid products are distilled in a combination of atmospheric and vacuum distillation processes.
H200. Product and receiving storage. Distilled products are held for shipment.
The material would be received from the CTL “syncrude” process.
H300. Utilities. Water treatment, cooling water treatment, waste water, are all provided for the hydrotreatment process.
Process steps H100-H103 were tested at laboratory/pre-pilot scale and reported in prior sections of this report. The other process steps are virtually identical to synthetic petroleum distillates (SPDs) being produced and are standard steps used in many other chemical processes so they were not tested in the laboratory-scale testing.
10.3 Coal Liquefaction Process Modeling
10.3.1 Coal Liquefaction Balance
To size equipment and utilities for the coal liquefaction process, a mass and energy balance was constructed using ChemCAD process modeling software. This allows for a flexible and consistent mass and energy balance that can be used to explore process configurations, estimate utility requirements, and provide preliminary sizing parameters for process equipment. The model in this case was largely based upon laboratory and pilot scale testing and analytical results to calculate product yields and understand thermodynamic behavior.
For this project, the feed rate of coal was assumed to be 1,000 MT per day (MT/day), and the model includes systems to dry the coal, digest the coal, remove residual solids, and evaporation to split the fuel product into two boiling-point fractions. At this scale, the hydrotreatment size is about 4,000 barrels (bbl) per day (BPD) for the lighter fraction. The scales of 2,000 MT/day coal (8,000 BPD fuel) and 8,000 MT/day coal with 32,000 BPD fuel are also evaluated to give the economics on a larger coal and regional hydrotreatment system. This scale was used to demonstrate that the process is economical at a smaller scale than the 17,000-19,000 MT/day of indirect liquefaction plants with less than 50,000 BPD production.
The water cooling system, heating loop, a scrubber for sulfur oxides (SOx), and preliminary work for CO2 capture were also modeled.
A reformer model done for DOE/NETL by Molburg, et al [24] to evaluate H2 generation costs was used. Their baseline model, adjusted for size and using a 3/1 steam/NG ratio, was used as the reformer for the Battelle CTL cases. The reformer was scaled to 2011 costs using Ref [21] data.
At 90% on-line/year, 90% coal conversion, Table 65 shows the annual flows of syncrude inputs, hydrotreating inputs, and waste products.
Table 65. Syncrude Material Balance for 1,000 MT/Day Coal, 90% Conversion, 90% On-Stream
The process operates on these process inputs and produces waste products and the following estimated quantities of products and byproduct materials as shown in Table 66. The background on selection of yields, gases, etc., are provided in the ChemCAD modeling section.
Table 66. Syncrude and Products at 1,000 MT/Day Coal, 380°C Split, 90% On-Stream
10.3.2 Hydrotreating and Hydrotreating Balance
The hydrotreater balance is performed based upon data obtained from operations at the pilot scale and design data for a similar hydrotreater plus reformer. These operations yielded a hydrotreater carbon balance of 87 percent; the other 13 percent is assumed to be contained in the emissions and includes sulfur as H2S, nitrogen as NH3 and oxygen as H2O. Reformer product and CO2 yields are treated in Tables 67 and 68.
The hydrogen/carbon balance is based on taking the total H/C in the syncrude from 1.22 to a ratio of 1.86 which is typical of Jet-A products (our Jet products are slightly lower in H2 at 1.7 – 1.8 for Jet A distillate, so this H/C gives conservative H2 usage). Burning the hydrotreater tail gas plus natural gas is used for energy to operate the process. Removal of heteroatoms plus hydrogenation will require 2.712 scf/ bbl minimum; hydrogen requirements for the hydrotreating assume 120% of this value as fresh H2. Because Intertek uses single-pass hydrogen in their hydrotreater and because we used a 1.05 H/C crude (from Quantex Run #6B) at Intertek, this ratio could not be verified at the pre-pilot level.
Using the syncrude balance data discussed above gives the input data in Table 67. An efficiency of 70% of theoretical is assumed for converting natural gas to H2 in the reformer. This is consistent with references [19] and [20].
Table 67. Hydrotreater Material Balance at 4,000 BPD and 90% On-Stream
The products from the hydrotreater are listed in Table 68.
Table 68. Hydrotreater Products
The fuel yield is about 4.02 bbl/MT coal. The fuel split appears to be about 10% naphtha used as fuel, 60% jet, and 30% diesel from the hydrotreating of the syncrude. The weight yield of syncrude into hydrotreated and saleable product is estimated at 80.3%.
10.3.3 Options for the Heavy-Oil Product
Several applications are possible for the heavy-oil product. The economic approaches in order of preference are:
- Use as a feed for a high-value specialty product. A coal-to-polyol process is used here. Other uses are coker-feeds for high-value carbon coal tar distillate, as heavy fuel oil.
- Use as a high-value product such as an organic binder (binder pitch) for which both the solid by-product and the heavy-oil fraction can be used.
- Use as a direct fuel.
- Upgrade as a motor fuel, or as fuel oil.
The options used in the cost models were that of direct sale as a binder material and producing a specialty chemical from a significant fraction coming from a 1,000 tonne/day coal plant. In the former application, the world market is so large that sale was not expected to modify the world market. However, in the specialty chemical market, there is good growth but the current world capacity would be in the range of 6 million MT/year. Using all of the heavy oil from one plant would produce about 166,000 MT/year. Based on the considerations of flooding the market and depressing the price, the estimate uses about 3% of the total world market for production regardless of scale for the process. This application is proprietary and is not described in the following ChemCAD modeling section.
10.3.4 ChemCAD Modeling
10.3.4.1 Assumptions
The mass and energy balance generated in ChemCAD is based on a number of assumptions that were derived from laboratory testing and engineering judgement. These assumptions are outlined in the following based on whether they applied to the whole process or were specific to the liquefaction step or scrubbing step. Where appropriate, references are cited to support the assumptions.
10.3.4.1.1 Overall Assumptions
In order to estimate vapor liquid equilibria throughout the process, an appropriate thermodynamic model must be selected. For this model, Peng Robinson was selected as it is commonly applied for general hydrocarbon species at moderate to high pressures. Additional global assumptions are listed in Table 69. Pump efficiencies are taken to be 60%, while compressor efficiencies are taken to be 50%. Cooling water is assumed to be available at 30°C, and is limited to a heat exchanger outlet temperature of 40°C to prevent excessive scaling. In all heat exchangers, the minimum approach temperature is taken to be 10°C, although not all exchangers necessarily assume this aggressive of an approach. Liquid stream pressure drops through exchangers are taken to be on the order of 0.35 bar (~5 psi).
Table 69. Overall Assumptions Made in the ChemCAD Process Model
4.2.2.1.2 Coal Liquefaction Assumptions
Key mass balance assumptions for the coal liquefaction operation are provided in Table 70, and are based upon laboratory and pilot testing results. The coal tar distillate and bio-solvent feeds as a mass fraction of dry coal feed are about 10% and 40%, respectively. The centrifuge cake is taken to be 50% solids. Baseline coal conversion is assumed to be 90% of the organic component, although this conversion is changed in the model based upon the content and composition of the proprietary bio-solvent. Non-condensable gases from the reactor are taken to be about 7% by weight of the dry feed coal, and the recycle rates are adjusted to provide 3 mass units of liquid per mass unit of dry coal feed.
Table 70. Mass Balance Assumptions Made in the Coal Liquefaction Process
A few assumptions in the coal liquefaction step had to be made regarding the composition and properties of certain streams. For coal tar distillate, a number of surrogate compounds were selected, and their mass ratios assumed based on GC-MS data in literature sources[29]. The selected model compounds and their relative mass concentration in coal tar distillate is presented in Table 71. The organic portion of coal was modeled in ChemCAD as ovalene, which structure is shown in Figure 73 and was a user-added compound in ChemCAD built by Joback functional groups. The coal ash was modeled as silicon dioxide only, and the feed coal was taken to be 5% moisture, 9.5% ash, and the balance organic material. A distillation curve for the liquefaction product was created in laboratory testing and is shown Table 72. The final liquefaction product in the ChemCAD model was adjusted to match this distillation curve based upon model component boiling points.
To understand the heat of reaction for coal liquefaction, a number of surrogate compound reactions were run within the model. The first two reactions account for the processes of dehydrogenation and decarboxylation within the bio-solvent. The actual depolymerization of the coal was assumed to be cleavage of a phenolic ether. In this case, the reaction was simplified to cleavage of diphenyl ether into phenol and benzene. In this manner, the overall liquefaction reaction was estimated to be slightly endothermic, which is consistent with other depolymerization reactions. The evaporator was operated to obtain two products with a cut at approximately 500°C (later revised downward to 380°C) to allow for polyol processing. The 500°C (380°C+) product was assumed to be usable as a heat transfer fluid to provide heat in the process; heating the reactors, influent, and used to dry the incoming coal. This assumption will need validation, as the material will need to be recycled approximately 12 times through a fired heater without substantial degradation.
Table 71. Model Compounds and Relative Mass Concentrations of Each Used to Represent Coal Tar Distillate
Figure 73: Structure of the ovalene molecule, which was used to represent the organic portion of coal in the ChemCAD model.
Table 72. Distillation Curve of Coal Liquefaction Product
10.3.4.1.3 SOx Scrubbing Assumptions
The SOx scrubber in this model was assumed to be a limestone forced oxidation (LSFO) scrubber, and key mass balance parameters are shown in Table 73. The liquid to gas ratio was assumed to be 70 gallons per actual cubic foot of scrubbed gas effluent, and the oxygen required for sulfite oxidation was taken to be 3:1 O:SO2 removed. The entrance and exit temperatures were important for energy balances as well as the water balance to account for water vapor leaving the scrubber. SOx was estimated based on the analysis of gases from laboratory testing. Removal was taken to be 95%. Methane combustion was assumed to be 99.5% since emissions of greenhouse gases are tracked in this project. The excess limestone is assumed to be 1.05, and the gypsum cake moisture is taken to be 50%. This moisture does not account for the waters of hydration for gypsum, and in the model ‘gypsum’ is taken to be CaSO4⦁2H2O. It was assumed that the air sparge for oxidation would require 15 psi, and the pressure drop of the lime slurry going up the column and through the nozzles was 60 psi.
Table 73. Key SOx Scrubbing Mass Balance Assumptions Used in the ChemCAD Model
The non-condensable gas composition was measured in the lab by an HP online GC and by GC-MS. The gas entering the Gibbs reactor has the measured composition shown in Table 74.
Table 74. Non-Condensable Gas Composition used in the SOx Scrubber Model
10.3.4.1.2 CO2 Scrubbing Assumptions
Key CO2 scrubbing mass balance assumptions are shown below in Table 75. A monoethanol amine (MEA) scrubbing solution at 30% by weight in water is assumed for carbon capture, and the target is capture of 85% or more of the CO2 generated. The amine regeneration column is limited to a reboiler temperature of 125°C, and aggressive heat integration is used for cooling and heating of the MEA within the recycle loop. The captured CO2 is compressed to 2,200 psig, and the compression ratio in each compressor stage is limited to prevent compressor stage outlet temperatures above 250°C. One key assumption in the carbon capture modeling is that there are no heat stable salts formed, which would require an MEA reclaimer that uses caustic and temperature to recover MEA from the salts.
Table 75. CO2 Scrubbing Model Mass Balance Assumptions
10.3.4.1.3 Hydrotreatment Assumptions
Key assumptions in the simplified hydrotreatment model are shown in Table 76. Hydrogen consumption was calculated based upon C/H analysis of the liquid before and after hydrotreatment, while yields of methane, H2S, methane, and C2-C6 hydrocarbons were taken from analyses during the hydrotreatment runs. The reaction temperatures and pressures were similarly taken from the hydrotreatment runs. Hydrogen recovery through the pressure swing adsorber for hydrogen purification was assumed to be 80%, with the balance of the hydrogen and contaminants recovered in a blowdown stream used for fuel gas.
Table 76. Simplified Hydrotreatment Model Mass Balance Assumptions
10.3.4.2 Model Narrative
10.3.4.2.1 Base Case
Figure 74 shows a part of the ChemCAD flowsheet that covers the feed portion of the process. Moist, crushed coal is fed in stream 1, where it is then heated in HX 42 by 500°C+ product. Evaporated water is removed as stream 35 from a flash drum. The dry coal is mixed with coal tar distillate from stream 2 and bio-solvent in stream 4, before being mixed with recycle material in mixer number 36. This combined stream is pumped to 405 psig for feed to the reactors by pump 18. Figure 75 shows the reactor portion of the flowsheet. The combined feed enters heat exchanger 10, where it is heated to 400°C by heated 500°C+ product. This then enters Reactor #3, which is a stoichiometric reactor used to model dehydrogenation of the bio-solvent. This enters a feed-forward controller which provides inputs to a separate heat of reaction calculation. The next reactor is a stoichiometric reactor to model decarboxylation of the bio-solvent, and the feed forward controller 20 controls the coal conversion rate based upon the current state of the bio-solvent. Reactor #4 is the digester, where ovalene in the simulated coal is converted into hydrocarbon pseudocomponents with specific boiling points to match the distillation curve of the liquefaction product. This distillation curve matching is done in an external excel spreadsheet and fed into the stoichiometric reactor by data map functionality.
Controllers 17 and 23 feed forward heat duties to heat exchangers 11 and 24 to model the heat of reaction. The dehydrogenation and decarboxylation reactors are run isothermally, and the heat duty required to maintain constant temperature is fed into the heat exchangers. A separate model reaction of diphenyl ether converting to phenol and benzene is used to predict the heat of reaction of coal depolymerization. The overall heat of reaction is endothermic, and heat exchanger 25 adds heat from 500°C+ product to maintain reactor temperature at 400°C. Flash drum 5 removes any vapors from the reactor headspace at 405 psig and 400°C. This vapor is partially condensed with cooled 500°C+ product, with condensate reheated in heat exchanger 90 and returned to the reactor. A second condenser cools the vapor to 50°C, and this condensate is sent to the evaporator. Non-condensable gas is removed from the process and used as fuel in the fired heater.
Liquid from the reactors is let down in pressure through valve 9 before entering a flash drum (Figure 76). A train of hydrocyclones is then used to wash and concentrate the solids material by washing the solids with evaporator bottoms product. The flash liquid is mixed with solids product from the second-stage hydrocyclone in order to recover light material. The washed liquid and solids are then pumped to a second-stage hydrocyclone, where they are washed with evaporator heavy product (>500°C). The solids proceed to the centrifuge, while the liquid proceeds to the first-stage hydrocyclone. The solids are cooled to 75°C with water in heat exchanger 26. In this separation/washing scheme, light material is reduced in the centrifuge solids, making them better for binder applications. As a benefit, the product syncrude is also increased.
The hydrocyclone solids are then centrifuged. Solids leave the centrifuge at about 50% solids, and the centrate is mixed with the liquid from the hydrocyclone second stage. At divider 38, a recycle stream is removed to be taken back to the front of the process. The balance of material is mixed with light condensate from the reactor, lights from flash drum 64, and lights from the recycle stream. This is then fed through a valve to an evaporator operating at roughly 5.85 psia. Heat exchanger 30 preheats the feed at a 15°C approach temperature, and the evaporator operates at 370°C. After the preheater, the distillate is cooled with water in exchanger 31. Uncondensed vapors and gases are sent to the syncrude process furnace.
Figure 74. ChemCAD flowsheet for the feed portion of the coal liquefaction process.
Figure 75. Flowsheet for the liquefaction reactor portion of the coal liquefaction process.
Figure 76. Solids handling and evaporator portion of the coal liquefaction flowsheet.
Figure 77 provides the flowsheets for the product handling. Distillate oil from the evaporator is in stream 47, and enters a knockout pot. Lights from here are removed by the evaporator vacuum pump, and later used as fuel in the fired heater. Bottoms of the knockout pot are the crude fuel product, and are pumped out of the process. A portion of this stream is recycled back to the front of the process. The heavy oil from the evaporator enters feed forward controller 44, which controls how much of the 500°C+ product is removed from the heating loop. Pump 34 then delivers the heavy product to the fired heater process. Solids are cooled with water in heat exchanger 13, then removed from the process.
500°C+ liquid from the process is fed into the heating process (Figure 78), where it is mixed with recycle heating fluid, then pumped into fired heater 12. This heater is run on non-condensable gas from the process, with natural gas makeup as required. It heats the heavy oil to 475°C, and the heavy oil is used to heat the reactors, preheat the feed, and dry the feed coal. As mentioned previously, the product is used as both a washing and a heat-transfer oil so it is recycled approximately 12 times before it is removed from the system.
The recycle system, not pictured, mixes the two recycle streams and flashes them, with lights returning to the evaporator. It is then pumped to the front of the process. The cooling water system is not shown, and cools the water to 30°C in a cooling tower before pumping it to the required heat exchangers.
Figure 77. Products flowsheet from the coal liquefaction process.
Figure 78. Fired heater flowsheet from the coal liquefaction process.
The SOx scrubbing system is shown in Figure 79. Feeds to the process include non-condensable gas, combustion air, natural gas, and air for the limestone forced oxidation scrubber. The rate of non-condensable gas feed is equivalent to that produced from the reactors and evaporator in the process, and the composition is as shown Table 70 the assumptions section. Component separator 60 on the NC gas stream bypasses 0.005 mole percent of the methane to enforce the 99.5% methane conversion assumption. Reactor #61 is a Gibbs reactor that simulates the fired heater in the process. Air is added in excess to provide an exit temperature of about 900°C, and the Gibbs reactor automatically calculates the formation of SOx gases in the combustion reaction. Heat exchanger 65 represents the demands of the coal liquefaction process, and air cooler 48 reduces the temperature to about 140°C. Feed forward controller 50 controls the amount of limestone added to the scrubber and controller 63 the amount of air added for sulfite oxidation, both based upon the levels of SOx in the stream. Air for sulfite oxidation is fed in stream 91 and compressed to 15 psig in compressor 62.
Figure 79. SOx scrubbing flowsheet from the ChemCAD model.
The scrubber is modeled by Reactor #52, which is a stoichiometric reactor performing three reactions at set conversions. The reactions are shown below:
10.3.4.2.2 Other Cases
Several other process flow schematics were also tested. One notable case was changing the recycle stream to be composed entirely of syncrude product, rather than split 50:50 between product and centrate liquid. This case could have impacts on product quality that are not described by the ChemCAD model, but energy consumption is increased slightly due to increased flow through the evaporator. A second case is shown in Figure 80, and involved flowing through the evaporator and centrifuge in parallel rather than in series.
Here, the top stream from the cyclone goes to the evaporator, and bottoms from the evaporator are mixed with bottoms from the cyclone. This mixed stream is flashed, with lights recycled to the evaporator, and the balance being cooled and fed to the centrifuge. This layout was intended to improve energy efficiency, as the centrifuge feed for the selected centrifuge needs to be at low temperatures to avoid seal damage. Additionally, the total flow rate into the evaporator is reduced, lowering energy and capital demands. However, there are tradeoffs in reduced syncrude product recovery, and light components entering the hot oil system.
Figure 80. Flowsheet describing the case of the centrifuge and evaporator running in parallel rather than in series.
10.3.4.2.3 Amine Scrubbing
Figure 81 provides the ChemCAD flowsheet for the amine scrubbing process model. The fired heater is represented by Gibbs Reactor #61, which combusts non-condensable gas and natural gas to provide heat to the process. The heat demand from the process is accounted for in heat exchanger 65. There is still significant energy in this stream, so it is used to heat loaded MEA going into the regenerator, before being cooled in heat exchanger 66. This cools the flue gas prior to entering the scrubber. Component separator 63 removes trace compounds in the gas stream that interfered with convergence of the scrubbing column, and is not a part of the actual process. Flash drum 67 removes water from the cooled flue gas before the gas enters scrubber column 68. Flue gas flows counter flow to the MEA solution, which absorbs the CO2. Cleaned flue gas exits in stream 99 with about 85% of the CO2 removed.
The loaded MEA from the scrubber is pumped through a first heat exchanger that recovers heat from the regenerated MEA. The second heat exchanger uses the flue gas to heat the stream to about 118°C before entering the regenerator. In the regenerator, CO2 is released from the MEA by heating. A condenser recovers volatile MEA and water, and refluxes it to the column. The released CO2 is fed to a compressor. Regenerated MEA is pumped through a heat recovery exchanger, and mixed with some of the condensed MEA from the overhead condenser. Feed forward controller 73 maintains the mass of MEA within the recirculation loop, accounting for any MEA loss through volatilization or carryout. The stream is then mixed with makeup MEA and water. Water content in the recirculation loop is controlled by feedback controller 54, which feeds water in stream 98 to maintain solution concentrations. This stream is then pumped back to the scrubber column.
Figure 82 shows the CO2 compression portion of the ChemCAD flowsheet. CO2 from the regenerator is compressed to 2,200 psig in four stages. There is interstage cooling and a knockout pot for each stage to remove water. The compression ratio in each stage is adjusted to keep the compressor stage exit pressure below 250°C.
Figure 81. Amine scrubbing and regeneration flowsheet from ChemCAD.
Figure 82. CO2 compression flowsheet from ChemCAD.
10.3.4.2.4 Hydrotreating
Figure 83 below shows the ChemCAD flowsheet from the simplified hydrotreatment model. Stream 105 is feed hydrogen to replace what is consumed in the reaction along with lost in the pressure swing adsorption blowdown stream. The hydrogen mass feed is controlled by controller 82, which maintains 10,000 kg/hr of hydrogen in the recycle to ensure high hydrogen partial pressures. The feed hydrogen is compressed to 1,300 psig in compressor 73 before being mixed with recycle hydrogen and compressed to 1,500 psig in the recycle compressor number 74. Syncrude is fed in stream 101 then pumped to 1,500 psig in pump 71. For this simplified model, the syncrude is represented by cyclohexylbenzene, along with oxygen, nitrogen, and sulfur added per the syncrude’s compositional analysis. In the first hydrotreatment reactor, sulfur is converted to H2S, oxygen to water, and nitrogen to ammonia. This stream is cooled to 240°C, then the second reactor converts the cyclohexylbenzene to a pseudocomponent boiling at 450°C, methane, and n-pentane to represent the light hydrocarbons generated in the process. This stream is cooled to 50°C in flash 72, then split in divider 77. A portion of the gas is recycled directly, while a portion enters component separator 76 which models a pressure swing adsorption system recovering 80% of the incoming hydrogen. The blowdown from the pressure swing adsorption system is fed to a Gibbs reactor along with air to combust it for process heat.
Figure 83. ChemCAD flowsheet of the simplified hydrotreatment model.
10.3.5 ChemCAD Modeling Results
The primary outputs from the ChemCAD model are mass and energy balances that allow for estimations of the cost of manufacture for the fuel products. The mass balance covers all material inputs and outputs for the coal liquefaction and SOx scrubbing portions. The energy balance covers the heating and cooling inputs, as well as electrical inputs for pumps and compressors. It is important to note that electrical demands for centrifuges, filters, mixers, conveyors, etc. are not included in the model. Additionally, the SOx scrubbing and coal liquefaction models are independent models, primarily due to the complexity of matching the GC-MS data in the non-condensable gas through the digestion reactor stoichiometry. Accordingly, the scale of each process is very close, but they do not match exactly.
10.3.5.1 Liquefaction Reactor Heat and Mass Balances
Heating, cooling, and electrical requirements for the liquefaction process are shown in Table 77. Heating is primarily required for drying the coal, preheating the reactor charge, and running the evaporator. Cooling is required in the reactor condensers, and the cool the reactor effluent and product streams. Electricity is required to run feed and process pumps, including the vacuum pump for the evaporator. This electricity demand does not include any equipment for coal prep, nor conveyors, and excludes the centrifuge electricity demand.
Table 77: Heating, Cooling, and Electrical Demands for the Coal Liquefaction Process
Table 78 shows the process inputs and outlets for the coal liquefaction process. The primary feed is coal, with a coal tar distillate makeup stream and bio-solvent stream. These streams are processed, yielding a moisture stream from drying the coal, a non-condensable stream from the digester, solids stream from the centrifuge, lights stream from the evaporator, and an oil product and 500°C+ product stream. The overall mass balance closes to within 200 kg/hr, or 0.4% of the feed rate.
Table 78. Input and Outlet Mass Flow Rates for the Coal Liquefaction Process
10.3.5.2 SOx Scrubber Heat and Mass Balances
Table 79 provides the heating, cooling and electrical requirements for the SOx scrubbing process. The process requires no heat addition, as the fired heat effluent is the feed stream to this process. This stream is cooled prior to the spray tower. Electricity is required to run the recirculation pump(s) and the air compressor for forced oxidation. Note that this electricity demand does not include limestone milling, conveyors, and mixing, and excludes the gypsum solids separation process.
Table 79. Heating, Cooling, and Electrical Demands for the SOx Scrubbing Process
Inputs and outlets for the SOx scrubber operation are provided in Table 80 below. Fuel gas is a mixture of the non-condensable gas from the digester and the lights produced from the evaporator. Natural gas is included in the table, as it would be needed for process startup and potentially to make up any deficiencies in the fuel gas rate or heating value. Combustion air is used in the fired heater, while the forced oxidation air is used in the scrubber to convert sulfites into sulfates. Water is makeup for evaporative losses and losses with the gypsum cake, and limestone is the active scrubbing compound. The process outlets are solely the scrubbed gas and the gypsum cake. The mass balance for the SOx scrubbing process closes to within 5 kg/hr, or 0.004% of the total feed rate.
Table 80. Input and Outlet Mass Flow Rates for the SOx Scrubber Process
10.3.5.3 CO2 Scrubber Heat and Mass Balances
The amine scrubbing model was built for consideration early in the modeling task, and was ultimately not included in the final model, design, or costing for syncrude processing. Accordingly, the inputs are slightly different than the SOx scrubber or base case model. Table 81 provides the energy requirements for the CO2 scrubbing. Heating is necessary for the MEA regenerator reboiler. Cooling is required for the flue gas entering the process, cooling the recirculating MEA, and interstage cooling at the CO2 compressors. Electricity demand is primarily for CO2 compression from atmospheric to transport pressure.
Table 81. Energy Requirements for the CO2 Scrubber Portion of the ChemCAD Model
Inputs and outputs for the CO2 scrubbing operation are provided in Table 82 below. The model associated with this CO2 scrubber did not include as in-depth heat integration as the base case model, and so the fuel gas was augmented with natural gas. Water and MEA are required to make up for the losses in the scrubber column. The scrubbed gas still contains about 1,300 kg/hr of CO2, but the remaining 8,800 kg/hr is captured and compressed, potentially for subsurface injection or enhanced oil recovery projects. There is also a significant amount of condensed water recovered from the process. The mass balance for the CO2 scrubber closes to 71 kg/hr, or within 0.08% of the input mass.
Table 82. Mass Balance for the CO2 Scrubber Portion of the ChemCAD Model
10.3.5.4 Hydrotreater Heat and Mass Balances
The hydrotreater model was built only for a rough understanding of the heat and mass balances, and was not used in design or costing. The energy demands for the hydrotreater are shown in Table 83 Heating is required entering the first stage of hydrotreatment to bring the syncrude and hydrogen to 380°C. The cooling is required between stages and also in the product cooler. Electricity demand is primarily from the hydrogen feed compressor, but also from the feed pump and recycle compressor. Energy requirements from the pressure swing adsorber are not estimated in this model, but should be comparatively low due to the relatively small feed stream, and the feed pressure should be more than sufficient for adsorption of the contaminants. Only a small amount of heat can be realized from combustion of the PSA blowdown gas, and it would need to be augmented with natural gas or some other fuel. It is important to note that heat of reaction in the hydrotreatment reactors was not modeled in this exercise, and may reduce the heating requirements for the process.
Table 83. Energy Requirements for the Hydrotreater Portion of the ChemCAD Model.
Inputs and outputs for the hydrotreatment model are provided in Table 84. Roughly 1,200 kg/hr of hydrogen is used to upgrade the syncrude and maintain high hydrogen partial pressures. The syncrude feed is the same mass as exits the base case model, and combustion air is selected arbitrarily to keep the combustion reactor temperature below 900°C. The mass balance closes to within 62 kg, or 0.2% of the feed mass.
Table 84: Mass Balance for the CO2 Scrubber Portion of the ChemCAD Model.
10.4 CTL Process Design and Equipment Sizing
The process design is adapted from the CTL ChemCAD model using 47 MT/hr of coal as its basis and is used for the process conceptual development and equipment sizing. Flow rates from the ChemCAD model were used to estimate the throughput needed for each unit operation. Pumps and heat exchange surfaces were sized mostly from the values provided by the ChemCAD model. Tanks were sized for process inlet and outlet storage using maximum flow rates in the process and estimated duration of storage. The details are provided in Appendix B.
10.5 Capital Costs Estimation
Per the DOE requirements, the cost estimate is stated in June 2011 dollars. Various indexes are available to factor costs based on time. The Chemical Engineering Plant Cost Index (CEPCI) [21] (extracted) has been used to factor the capital costs obtained to June 2011 dollars. This index is more aligned to chemical equipment costs than a Marshall and Swift index or an RSMeans index. Indices used in this evaluation are shown in Table 85.
Table 85. Chemical Engineering Plant Cost Index
The data in Appendices B and C were used to generate capital costs for the total process. Capital costs were estimated at 1,000 MT/day (MTPD) and 4,000 BPD hydrotreater/reformer. Table 86 shows the breakdown of the total installed costs at the 1,000 TPD Syncrude, a 70 TPD Polyol Plant, and a 4,000 BPD Hydrotreater.
Table 86. Total Installed Cost for Syncrude, Polyol, and Hydrotreater at Likely Commercial Scale
Table 87. TASC for Two Process Size Options Without Polyol
The TASC including the polyol plant costs are shown in Table 88. These costs will be for a plant that produces fuel, polyol, and reduced quantities of heavy oil.
The costs at this scale were not easily compared to other costs by DOE primarily due to comparing the small-scale of production here to much larger DOE plants. To provide a summary of the costs that would be comparable to larger plants used to make Fischer-Tropsch fuels (Reference [25] is about 50,000 BPD), the costs were scaled for better economic comparability. To scale the results, the coal plant costs were scaled from 1,000 MTPD to 2,000 MTPD and assumed that 4 plants would feed one hydrotreater (scaled from 4,000 daily bbl to 32,000 daily bbl). The capital costs were scaled using a 0.62 scale factor. Figure 84 shows the effect on $/daily barrel capital cost for scaling. The costs were scaled at the Total Installed Cost level and then developed to Total As-Spent Costs because some factors are linear (e.g. raw materials use, electrical energy) which are contained in the Total As-Spent Cost as working capital costs.
These costs (8,000 MTPD coal and 32,000 BPD product) are used below and are those primarily discussed in the rest of this report.
Table 88. TASC for Two Process Size Options With Polyol
Figure 84. Effect of increasing scale of plant on TASC per daily bbl.
A Greenfield cost estimate is used as the likely maximum capital cost in all cases. It is more likely that a location on or nearby to a mine would be preferred which would reduce shipping and receiving facilities, potentially electrical and water infrastructure and more likely grading and other utilities as well. If a Brownfield were available (for example, a utility boiler facility or other coal plant) direct cost items such as coal grinding and drying, rail spurs, etc. may be already available and coal could be handled through these facilities. Site cost items such as a level area, transformers for power, water and wastewater treatment, and electrical substation, may also be available. On this basis, a Brownfield cost might be reduced by 10-20 percent considering only such facilities as the common ones, but would be reduced even more if the whole coal yard (about 5 percent of the installed cost) and combustion facilities were able to be used. This type of site would even more favorably affect economics, such as was done with F-T piloting at the Wabash facility. In an existing plant, many other decisions, such as using the solid stream for a boiler fuel, might also affect economics. To determine the impact of these items, an actual site-specific design and estimate would be required.
10.6 Operating Cost Estimation
The ChemCAD modeling and capital cost estimates above were also used to determine syncrude operating costs. The ChemCAD models were used to define flow rates of materials and products. The values in the model, which included factors such as coal, bio-solvent, and CTD flow rates, recycle rates, byproduct flow rates, syncrude flow rates, equipment sizing, and energy requirements all came from the ChemCAD flow sheet which had appropriate heat and mass balances. Limestone requirements for SO2 scrubbing and waste products were determined as discussed. CO2 scrubbing and compression energy were also modeled using ChemCAD. Yields, percent solids from centrifugation, and the yield by boiling point were all determined from the laboratory studies.
Hydrotreatment values were determined by calculating from the previously-discussed ChemCAD results using minimum hydrogen requirements, using prior ChemCAD modeling experience to determine hydrogen blow-down requirements, experience to determine energy requirements for hydrogen compression, hydrogen and liquid circulation, and prior experience to evaluate the hydrogen requirements and off-gases from the overall material balance. Rather than purchase H2 on the gas market, a package reformer cost was added to the hydrotreater estimate. The cost was derived from a DOE report for hydrogen generation as an alternate fuel [20].
While the composition of coal was used as measured from the laboratory samples, the cost of equivalent coal was calculated using DOE references which showed market costs for fuels by region and sector. Base prices were taken from Ohio data. Ranges are given in Table 89. The September 2011 EIA documentation was used [26]. Chemical costs for non-fuel components were obtained using www.icis.com data [27].
Table 89. Baseline Input Costs and Values Used for the Cost Model
Operating costs for the baseline system are provided in Table 90. Product values for the syncrude and for distillate products such as diesel and jet fuel were determined from the above EIA reference. For syncrude, a derating factor comparing Hardesty and WTI was developed in conjunction with Quantex and applied to the average crude price from the EIA reference to get the price expected for the CTL syncrude product. For the jet and diesel, a 13% reduction in selling price was applied to EIA data to account for taxes because EIA data include taxes. The 2011 price before tax was $3.09 for diesel and $3.10 for jet fuel [28]. A comparison of baseline cost for the CTL plant vs crude-derived products are listed in Table 89. The table contains the projected syncrude and hydrotreated product prices for both 1,000 MTPD/4,000 daily bbl and for four 2,000 MTPD plants and a 32,000 daily bbl hydrotreater. In both cases, polyol is not included.
Table 90. Product Prices Determined from Modeling, June 2011 Data(a-d)
The heavy oil and centrifuge cake that are produced are like other materials which have been treated as “binder” materials by Quantex. The values for these products which are used were provided by Battelle’s Quantex partner. Battelle included a 10% derating of that price into these factors [29]. Note that the values used for the hydrocarbon content are lower than the sale price for equivalent residual or bunker oil in the EIA data ($1.98 for our product vs $2.47 for residual oil in June 2011 [28]. The >500°C product is lower in ash and metals than residual oil and should go for a premium cost over resid. Because resid is not considered to be the final market, this cost basis was not used. Both of these materials could also potentially be used for gasification or petroleum feedstock purposes or possibly sold as a bunker or heavy material, all at a lesser value than they are worth as a binder or other feedstocks. Particularly for the >500°C liquid fraction, the solids are actually equivalent to most motor fuels, so delayed coking to recover lighter products and use of the carbon for a product like graphite or carbon fibers might be a reasonable use for this material. This was briefly evaluated but no interest could be obtained from Exxon in providing a quotation for a coker due to the small size.
DOE NETL had suggested in the October 2016 review that gasification of the centrifuge solid stream be considered and that hydrogen be produced in that manner. To evaluate this option, Battelle evaluated a “non-capital” economic value proposition to see what the operating costs might be, and to also see what the capital which might be available would be. A DOE Report described the comparison as well and is in general agreement with this rough calculation [20]. Assuming that a steam-blown gasifier would be used, that the steam/solid would be operated at 1.2/1 ratio to assure fluidization, and gasification efficiency is 60%, a cost of $3.94/kg of H2 was calculated for steam plus solid product. A similar cost for H2 reformed from natural gas shows a cost of about $2.21/kg H2. Typical costs for a gasifier plus all of the shift and membrane or PSA equipment being the same for a reformer and a gasifier would make capital costs for the equivalent amount of hydrogen substantially higher than a reformer. Hence, unless this product has a fair market value of half or less its assumed value, the gasification approach should not be considered unless other factors, like an existing gasifier already exists which could be used to gasify this material.
On the basis of total cost, a reformer was chosen. Reformer costs were scaled for this project from capital and operating costs [20]. Scaling was done from 2009 to 2011 costs and to the flow of H2 for this project using a 0.62 factor on the capital cost. The capital for reforming was added into the hydrotreater total budgetary cost. Operating costs for natural gas for reforming, catalyst, maintenance, and operating labor are also included in the hydrotreater operating costs.
Maintenance was assumed to be done by on-staff personnel and maintenance cost for materials was determined using 6 % of cost of equipment in addition to the on-line spares in the estimates. As mentioned previously, equipment spares were put in place in the syncrude plant where high-frequency repairs were expected as described in the capital equipment costing section previously. These included in the reactor area, in the pressure let-down valve area, in the cooling water circulation area, in the solids separation area, and in the evaporation area as the five prevalent areas in the Syncrude production. In the hydrotreater, an in-line spare reactor would be expected as would spare rotating equipment (pumps and compressors). Fired heaters were not spared for either case.
A labor estimate of 52 craft and 5 technical/supervisory persons were determined to be the appropriate number of staff for the 1,000 MT/day syncrude plant. This staffing level was scaled linearly for the 2,000 MT/day unit to 104 craft and 10 technical/supervisory persons. An integrated 70 MT/day polyol plant was assumed to require 32 craft and 5 technical/supervisory staff.
A separate 4,000 BPD hydrotreater would be staffed with 28 craft and 5 technical/supervisory staff. This staff was also scaled linearly for the 32,000 BPD hydrotreater to 224 craft and 40 technical/supervisory staff. It is likely that the larger plants could be staffed with lower numbers of staff, but this linear scaling assumption was used for conservatism. This derives a potential labor force for this project of approximately 1,000 persons for the larger facilities (8,000 MTPD/32,000 BPD fuel/70 TPD polyol).
The base case cost demonstrates that the syncrude could be sold for less than the cost of heavy crude without RIN credits even at 1,000 tonne/day of coal flow. The fuel from hydrotreating also could be sold for less than the then-current market cost of fuel at the 4,000 BPD size. If a 30% contingency were added to the calculated syncrude and fuel cost, syncrude would be more expensive than then-current heavy crude at 1,000 MTPD but less at 2,000 MTPD (still without RIN credit).
RINs were not promulgated in the U.S. until 2010 (Renewable Fuel Standards 2) but the concept, which allows a credit for biologically-derived carbon content is useful. With the RIN credit included, the value of the total proposition of direct liquefaction with a non-recovered bio-solvent will produce a valuable crude. The fuel produced, when inflated with a 30% contingency, would cost more at 4,000 BPD than then-current fuel price, but would cost less at 32,000 BPD than then-current fuel. The RIN credit would again make both scales of production economical.
The operating costs for including polyol production in the baseline plant were also considered. The advantage of polyol production is that a chemical product is produced from the heavy oil that can sell for a premium value. Polyol is valued for foam production. The costs below in Table 91assume that polyol will sell for about $0.90/lb. This “upcharge”, when taking into account all production costs and 12% capital cost, would yield additional revenue of $8.8 million. At 18% capital cost, it still would yield $5.5 million when compared with sale of heavy oil at the anticipated price.
Table 91. Product Prices Determined from Modeling, June 2011 Data Including Polyol(a-d)
10.7 Sensitivity Analysis for Costs
Cost sensitivity was run for varying cases of costs for the syncrude and the syncrude plus reformer cases discussed above. Sensitivity cases are provided in Appendix F. Cases evaluated included capital cost, electricity cost and carbon dioxide penalty, compression of CO2, and variability of byproduct value. A Monte-Carlo simulation of 10,000 iterations using Crystal Ball® was conducted on various process variables.
The standard capital cost factor was directed from Attachment 2 from the contract. By Battelle PM direction, capital was baselined at 12% and varied to 18% for costs. Costs were evaluated at 12% and 18% for the 1,000 MTPD/4,000 BPD and the 2,000 MTPD/32,000 BPD syncrude/polyol plus hydrotreatment cases.
Compression of CO2 was not considered at the CTL site and was evaluated only briefly at the hydrotreater site. The amount of CO2 produced from syncrude production is 73,000 MT/year for all syncrude sources at 1,000 MT/day (total coal conversion to CO2 would be about 1 million MT/year).
The cost avoidance for CO2 using compression is about $64/MT based on references [30].In the syncrude production process, a much higher per-MT cost is expected due to the much-smaller production of gas. This lower production of CO2 is because we are producing about 1% of the coal flow in equivalent CO2 and because both 1,000 MT/hr of coal are much less than the coal flow to this size pulverized coal facility equivalent production.
Compression of CO2 was considered as a sensitivity factor for the hydrotreater site, because the amount of CO2 produced is much larger than from liquefaction site, both because of hydrotreater operational energy and hydrogen which is made at the reformer plant at the hydrotreater site.
The reformer plant has two streams that are rich in CO2, one from the reformer tailgas itself and the other from the fired heater source. The tailgas is the only economical gas to compress; a cost avoidance of $47/MT for compression of the compressed gas plus several dollars/ton for shipping vs. a $60/MT emission charge are nearly equivalent and would give a net zero cost avoidance for compressed CO2 . The carbon dioxide from the reformer tailgas is about 40% of the total CO2 produced in the hydrotreater.
The syncrude reactor tailgas could potentially be a pressure source, but if it were then either oxygen would have to be supplied or feed air pressurized. This obviates the benefit of this gas unless it is steam shifted. The syncrude reactor bleed gas will have some H2 in it which could be recovered for hydrotreatment, but will mostly be CO2 and water. The gas is also mixed with the vacuum pump tailgas which is at atmospheric pressure. The mixed tail-gas is used as fuel rather than any other use. Similarly, the hydrotreater tailgas will have mostly hydrogen, hydrogen sulfide and ammonia in it. Hydrogen recovery has already been done to the extent feasible; on this basis, it is also used as a fuel gas. SO2 scrubbing with limestone was done on both streams after combustion. To recover a combustion stream at a pressure other than atmospheric, air will either need to be compressed or an oxygen plant will be required for combustion air. With these drawbacks it is likely that the only CO2 that might be compressed from the hydrotreater plant would be about 38,000 MT/year of the 146,000 MT/year of CO2 produced from the 4,000 BPD hydrotreater. This small quantity could not justify a pipeline, so it is more likely to be transported by rail than by pipeline (estimated 4 cars/day) which would add to the cost for CO2 recovery and further reduce the savings. However, this would be highly pure CO2 and may represent a higher-value product.
If there is also the possibility that the CO2 could be sold rather than disposed of, then economics should be revaluated. When it can be sold, the scenario changes in that a product value would be recovered for the CO2 sold. For our purposes, a market value of $150/MT is assumed with a compression cost of $64/MT. [20] By selling compressed CO2 at this price, the cost of fuel is reduced by $0.11 and $0.12 at the two scales of hydrotreater operation assuming a $64/MT compression cost for the CO2 (typical for reformer exhaust gas after shift and H2 recovery; reports on CO2 by DOE indicate that the cost of compressing CO2 from atmospheric gas is on the order of $75/MT based on the 2010 report previously cited.)
Other variables were also considered for hydrotreatment. The Crystal Ball® analysis for fuel sales cost is also evaluated here. Figure 85 shows the variation across the total suite of variables for both the syncrude and the fuel product. The analysis was generated using 20,000 Monte Carlo iterations and 10% variations in raw material and energy costs.
Table 89 (previously presented) contains the variations on inputs which were used for Crystal Ball. Both the 1,000 MTPD/4,000 BPD and the 8,000 MTPD/32,000 BPD cases were evaluated in Crystal Ball.
As stated above, a uniform 10%variation in inputs was used for the Crystal Ball analysis. The sensitivity results are shown in Figure 85. These results show that the same variables affect both the syncrude and the fuel break-even price. CO2 emissions is much more important for fuel because of the need to make H2 to produce the fuel. Natural gas is similarly important to fuel production because of the need to make H2 but is not a major variable in the cost of syncrude or fuel. Otherwise, the variables are quite similar in magnitude for syncrude and fuel.
Figure 85. Sensitivity of primary variables on crude and fuel break-even price (8,000 MTPD/32,000 BPD case).
Because the syncrude is highly dependent on CO2 emissions, so will be the hydrotreated product. It is important to the value of this project that the polyol be salable at more than $0.70/lb, which is break-even. The estimated of current market price for the equivalent product is $1.00. A 10% polyol variation produces a 29% variation in crude price and a 26% variation in fuel price.
Also important are capital charges (21% and 19% for crude and fuel) and CO2 penalties. Koppers solvent cost, coal cost, and bio-solvent cost all have a >1 ratio of effect to change in price.
10.8 Costs for Battelle CTL vs. FT Indirect CTL
Table 92 provides a summary comparison for CTL fuel costs ($/bbl) and the effect of capital charges. A 50% increase in capital charges will make a $13.00 (21%) increase in the break-even selling price of fuel without RIN credits, and a 28% increase in the break-even selling price for RIN-credited fuel. A 12% rate is representative of well-established, long-term technology. An 18% rate represents a more risky, shorter-horizon project. Either gives a better cost than other alternatives.
Table 92. Distillate Selling Price for Battelle’s CTL Process
Table 93 compares CTL operating at a 12% capital cost factor vs Gasification/FT operating at the same capital factor. This shows that the CTL project would be less costly (about 36% of a gasification/FT indirect liquefaction plant cost) and is also more economical at a smaller scale (about 65% of the plant product quantity). This demonstrates the ability of a CTL plant to produce fuel much more economically than an FT plant with indirect liquefaction.
Table 93. 2011 Capital and Operating Costs (at 12% CCF, no RIN credits)
11.0 GREENHOUSE GAS (GHG) EMISSIONS ANALYSIS
Emissions estimates were developed for the syncrude part of the process and the hydrogenation and distillation part of the process using ChemCAD runs for both the syncrude and hydrotreater, and the referenced Reformer report. Emissions estimates were developed without GHG control.
To demonstrate that GHG release can be limited to levels comparable to or lower than petroleum-based jet fuel, the approach of Life Cycle Assessment (LCA) [31] was applied on a commercial plant design proposed in this task. The GHG lifecycle emissions from the direct CTL jet fuel process is being assessed and compared to that for the production of petroleum-based jet fuels. The GHG accounting methodology developed is believed to be consistent with industry guidelines for performing LCA’s developed in ISO 14040 [31].The study uses applicable and publicly-available LCA evaluations contained in GREET [32]. OSU has developed a subset of this for the proprietary bio-solvent so that comparison of results with those from previous studies can be made. The boundary of the unique assessment is well-to-pump. Existing pump-to-wheels will be used so that raw material extraction through fuel use can be addressed for the fuel portion. The GHG assessment of the other products has been considered but is not addressed herein.
This effort was subcontracted to Dr. Bhavik R. Bakshi, a Professor in the Systems Analysis Department of The Ohio State University who specializes in GHG estimation work. Dr. Bakshi, supported by a graduate student, Mr. Kyuha Lee, prepared a GHG evaluation based upon data provided by the Battelle team. This work uses the GREET software as much as possible, and supplement it with data from other life cycle inventory sources. The overall model is developed in openLCA software [33].
The following data summarize the syncrude and fuel portion of the process as analyzed:
- Syncrude oil product from the liquefaction process, 220,095 MT/yr, employing 1,000 MT/day of coal and done at 1,000MT/day
- Jet fuel product from the hydrotreating process, 186,043 MT/yr
- Syncrude oil product from the conventional process, 220,095 MT/yr
- Jet fuel product from the conventional process, 186,043 MT/yr.
As mentioned earlier, there are by-products produced from this process also.
A network diagram for the process used for the GHG analysis is shown in Figure 86. Inputs to the GHG syncrude model are shown in Table 94. Outputs from the GHG syncrude model are shown in Table 95. The source for factors and their total quantities of impact are shown for a Well to Pump (WTP) analysis. References to data sources are shown in the Reference column. These data all assume that CO2 is not recovered for the synthetic crude portion of the process. To evaluate the emissions accurately for the hydrotreater if compression is assumed for the shift-gas CO2, the footprint for MEA was required. Because it could not be found in the GREET data base, a footprint was constructed. That footprint is shown in Table 96 and Table 97. Data for input and outputs for the hydrotreatment of the syncrude portion are provided in Table 98.
Figure 86. Block diagram for GHG analysis.
Table 94. Inputs from the GHG Syncrude Evaluation
Table 95. Outputs from the GHG Syncrude Evaluation
Table 96. Estimation of MEA GHG Input Footprint
Table 97. Estimation of MEA GHG Output Footprint
Table 98. Input and Output GHG Estimates for Hydroreating of Syncrude
The emissions burden of the proposed liquefaction process was divided into three products, synthetic crude oil, >500°C heavy oil, and centrifuge solids by a mass-based partitioning. The life cycle CO2 emissions were 97,401, 44,841, and 48,159 MT CO2/yr, respectively.
Figure 87 shows that the liquefaction process was the highest contributor to the emissions, followed by digester gas and bio-solvent. In case of the hydrotreating process, the life cycle CO2 emissions of jet fuel, switch engine fueling, ammonia, and compressed CO2, were 219,974, 250, 3039, and 43,748 MT CO2/yr, respectively. As shown in Figure 88, the hydrotreating process was the most dominant contributor. It should be noted that the shift-CO2 is not shown here. It is a water product (37,000 MT/yr).
The life cycle CO2 emissions of conventional alternatives to synthetic crude oil and jet fuel were obtained from corresponding GREET models. The CO2 emissions of synthetic crude oil, based on bitumen from oil sands as a conventional process were 224,497 MT CO2/yr, which are significantly higher than those from the liquefaction/hydrotreating process. The uncontrolled CO2 emissions of F-T jet fuel from coal were 852,079 MT CO2/yr, which are higher than those from the hydrotreating process.
The CO2 emissions of conventional jet fuel from crude oil and FT jet fuel from natural gas were lower than those from the hydrotreating process, showing 74,417 and 208,369 MT CO2/yr, respectively.
Figure 87. Syncrude relative emissions estimates.
Figure 88. Syncrude plus hydrotreating estimates of CO2 generation.
Figure 89 shows the life cycle CO2 emissions of products when different allocation methods were employed. The results of the displacement method show higher emissions that those of other partitioning methods. However, actual emissions of the displacement method may be lower because credits from some byproducts were not included in the analysis by the displacement method. In case of the partitioning method based on exergy, the CO2 emissions of synthetic crude oil from the liquefaction process were not calculated since the chemical composition of by-products, which are >500°C heavy oil and centrifuge solids, is unknown, and therefore, exergy values of those by-products are hard to calculate. Regardless of which allocation method was selected, the CO2 emissions of synthetic crude oil from the liquefaction process were much lower than those from the conventional process. That is, about 43.7% reduction of the CO2 emissions is expected to produce synthetic crude oil by employing the proposed liquefaction process. Also, regardless of the allocation methods, the CO2 emissions of jet fuel from the hydrotreating process were higher than those of conventional jet fuel from crude oil and crude oil, but 68.6% lower than those of F-T jet fuel from coal.
A mass basis was used for this analysis as previously described. Figure 89 shows that other choices could have a significant effect on emissions allocated to the fuel product.
Figure 89. Comparison of WTP GHG Footprints with Synthetic Crude, Petroleum and FT.
Table 99 shows the comparison of uncontrolled emissions from well-to-pump (WTP) and well-to-wheels (WTW) for the Battelle CTL process and coal gasification/FT process. As shown, the WTP emissions for the FT process are 2.75 times higher than for the Battelle CTL process. This is partly because the CTL has ~40% bio-content while the gasification/FT (and NG/FT) have none. On this basis, the pump-to-wheels (PTW) emissions for the Battelle CTL process are 40% less than either of these products on a MT-to-MT comparison; about 1.3 MT CO2/MT product. To be comparable, both FT processes would have to emit less than 0.6 MT CO2/MT product. This would represent 60% CO2 control on the NG to fuel indirect process and greater than 90% on the coal indirect liquefaction. The uncontrolled emissions are slightly lower (3.56 vs. 3.79 MT CO2/MT product) with those from conventional petroleum-to-jet fuel processes. Further, the 40% bio-content of CTL jet fuel from the Battelle process would help meet Section 526 of EISA 2007 goals without any CCS. If CCS were added to the highly-pure reformer shift-gas outlet, this could represent a further GHG emissions reduction of 0.2 MT CO2/MT product from the Battelle CTL process, adding to its benefit.
Table 99. Greenhouse Gas (GHG) Emissions (Petroleum Well-to-Wheels (WTW)
12.0 PRODUCTS DEVELOPED
A number of products were developed under this project, as discussed below.
12.1 Technologies Developed
In addition to advancing the Battelle CTL technology to TRL 5, the following technologies were created (patent application filed) [34]:
- Conversion of heavy-oil byproduct to a high-value carbon product
- Conversion of non-coal carbonaceous feedstocks to fuels and/or high-value carbon products.
12.2 Publications
The following papers were presented at international coal conferences:
- Chauhan, Satya P., “Direct Coal-to-Liquids (CTL) For Jet Fuel Using Biomass-Derived Solvents”, presented at the 2015 International Pittsburgh Coal Conference, held in Pittsburgh, PA, October 5-8, 2015.
- Chauhan, Satya P., “Scale-Up of Battelle’s Direct Coal-to-Liquids (CTL) Process For Jet Fuel Using Biomass-Derived Solvents”, presented at the 2016 International Pittsburgh Coal Conference, held in Cape Town, South Africa, August 8-12, 2016.
- Chauhan, Satya P., “Direct Coal-to-Liquids (CTL) For Jet Fuel and Diesel Using Biomass-Derived Solvents”, presented at the World CTX2017 Conference, held in Beijing, China, March 27-29, 2017.
- Chauhan, Satya P., “Direct Coal-to-Liquids (CTL) For Jet Fuel and Diesel Using Biomass-Derived Solvents”, presented at the 2017 International Pittsburgh Coal Conference, held in Pittsburgh, PA, October 5-8, 2017.
12.3 Post-Graduate Education Supported
This grant supported two Ph.D. students and two postdoctoral associates at three different universities: (a) Pennsylvania State University; (b) University of Dayton; and (c) the Ohio State University.
12.4 Technology Transfer
Several potential licensees around the world were identified. Also, the Battelle CTL process was nominated for an R&D 100 Award, since it became available for licensing in 2016/2017.
13.0 CONCLUSIONS
The following conclusions can be drawn from this project:
- Two bituminous and one subbituminous coals were successfully converted to syncrude to provide for the desired ≥80% coal solubility using one of many identified biomass-derived solvents without the use of H2 or catalyst. The typical coal conversion was over 85% on a moisture- and ash-free (MAF) basis.
- A total of 12 novel bio-solvents were identified that met the key liquefaction performance goals of the project, namely high coal solubility and low syncrude viscosity, both being indicative of excellent hydrogen transfer. The majority of these bio-solvents were superior than tetralin, which is the most researched and, until now, regarded the most effective hydrogen-donor solvent. The preferred bio-solvents are an order-of-magnitude lower in cost compared to tetralin.
- The coal liquefaction process was scaled up to 1 TPD, which represents TRL 5 targeted for this project. The pre-pilot-scale test results show a good correlation with laboratory-scale testing at Battelle. The pre-pilot test data was adequate to develop a conceptual, commercial-scale liquefaction-plant design.
- A 2-stage catalytic hydrotreatment/hydrogenation for upgrading the Battelle-CTL syncrude to jet and diesel was developed and demonstrated at pre-pilot scale. The Stage-1 catalyst removed more than 99.9% of nitrogen and sulfur removed and reduced oxygen below detection limit. In Stage-2, the hydrotreated syncrude was hydrocracked and further hydrogenated to obtain a distillate with high (60-70%) selectivity for jet-fuel boiling range. Nearly 100% of the distillate was in the diesel fuel range.
- A detailed characterization of the synthetic jet fuel from the Battelle CTL process indicated that:
- Up to about 30% of it could potentially be used for blending with a commercial jet fuel
- A 20% synthetic, 80% commercial fuel blend was tested to demonstrate it met all standard fuel specifications for Jet A/A-1 fuels.
- The synthetic diesel will likely not require any blending with a commercial diesel fuel.
- A conceptual plant design and an economic analysis following DOE/NETL methodology showed that the process is competitive for both syncrude and jet fuel (or diesel) applications at crude oil prices of less than $48/bbl. The selling price of jet fuel or diesel at the CTL plant is estimated to be $61/bbl or $1.46/gallon, compared to $95/bbl ($2.26/gal) for an indirect CTL plant using FT technology.
- The use of biomass-derived solvents in the Battelle CTL process brings about major process simplification, such as mild operating conditions and elimination of the need for gaseous hydrogen during liquefaction. As a result, the Battelle CTL process is economical at a much smaller scale for coal liquefaction (1000-2000 TPD) compared to FT-based plants. The resulting capital cost for a 32,000 BPD jet fuel/diesel plant 36% for the Battelle CTL process compared to that for FT based processes.
- A greenhouse gas (GHG) emissions analysis shows that the life-cycle GHG emissions for coal-mine-to-fuel combustion are 3.56 MT CO2/MT fuel, which is lower than the 3.79 MT CO2/MT fuel for petroleum based well-to-wheels (WTW) for jet fuel.
- The WTW emissions for the uncontrolled FT process is 7.77 MT CO2/MT fuel for FT based CTL, which is 2.2 times higher than for the Battelle CTL process.
This report demonstrates that the Battelle CTL meets the most important goal of this project, i.e., meeting the goal of Section 526 of EISA of 2007, for producing alternative jet fuel/diesel that has GHG emissions no worse than for petroleum-based fuels at a lower cost. The key reasons for this achievement for Battelle CTL process are (a) the ~40% reduction in hydrogen requirement for upgrading coal to jet fuel and (b) having a 40% bio-content in the fuel products. The complete elimination of the carbon capture and storage (CCS) requirements in the Battelle CTL process is a major achievement of this process since indirect CTL requires ~90% carbon capture for coal-to-jet fuel to have GHG emissions no worse than for petroleum-based jet fuel.
14.0 ACKNOWLEDGEMENTS
The authors greatly appreciate financial assistance provided by the following government agencies:
- The Department of Defense, especially the U.S. Air Force, which provided a budget to the U.S. Department of Energy for awarding a Cooperative Agreement.
- The Ohio Development Services Agency (ODSA) provided a grant under a cost-sharing arrangement.
- The authors also appreciate portions of the required cost-share contributed to several sub-grantees of Battelle: (a) Quantex Energy, Inc; (b) University of Dayton Research Institute (UDRI); (c) Pennsylvania State University (PSU); and (d) Applied Research Associates (ARA).
The key contributions from the following individuals are greatly appreciated:
- Gilbert Chalifoux and Elliot B. Kennel of Quantex for leading the pre-pilot plant coal liquefaction testing.
- Professor Caroline B. Clifford served as a PI for PSU’s bench-scale liquefaction efforts. Ron Wineck (a Ph.D. student), Dr. Xiaoxing Wang (a postdoctoral associate), and Sharon Miller made important contributions.
- Dr. Matthew DeWitt served as a PI for UDRI’s effort on catalyst testing and jet-fuel characterization. Dr. John J. Graham and Linda Shafer performed much of the testing and analysis.
- Dr. Tim Edwards of Air Force Research Lab guided the jet-fuel characterization efforts by UDRI and also provided a commercially-used jet fuel for blending with synthetic fuel for the Battelle CTL process.
- Ed Koppola and Sanjay Nana carried out efforts at ARA.
- John Brophy and Jared Gaydas led the efforts at Intertek on pre-pilot-scale upgrading of syncrudes from the Battelle CTL process.
- Jason Lewis of DOE’s NETL served as the technical monitor for the DOE-funded effort. Doug Archer also provided technical guidance.
- Gregory Payne of ODSA served as the technical Project Officer at ODSA. Bob Brown also provided very useful technical reviews of deliverables.
- Professor Bhavik Bakshi and Kyuha Lee (a Ph.D. student) of OSU carried out the GHG analysis.
Last, but not least, the authors greatly appreciate the efforts of the following Battelle research leaders, all of whom are now retired: (a) Dr. Herman Benecke; (b) Nick Conkle; (c) Russell K. Smith; and (d) Marty Zilka.
15.0 REFERENCES
[1]Longanbach, J. R., J. W. Droege, S. P. Chauhan, “Short-Residence-Time Coal Liquefaction”, ACS Symposium Series, No. 139, Coal Liquefaction Fundamentals, pp 165-177, 1980.
[2]Longanbach, J. R., J. W. Droege, S. P. Chauhan, “Short Residence Time Coal Liquefaction”, presented at ACS Meeting, Honolulu, HI, April 3, 1979.
[3]Kennel, E. B., Principal Investigator, “Enhanced Hydrogen Economics via Co-production of Fuels and Carbon Products”, Final Report to DOE/NETL from West Virginia University, Contract DE-FC26-06NT42761, March, 2011.
[4]Kennel, E. B., Principal Investigator, “Development of Continuous Solvent Extraction Processes for Coal Derived Carbon Products”, Final Report to DOE/NETL from West Virginia University, Contract DE-FC26-03NT41873, April, 2011.
[5]Burgess, C.E. “Direct Coal Liquefaction: A Direct Route to Thermally Stable Jet Fuel,” PhD Thesis, May, 1994.
[6]Mortensen, P.M., J.D. Grunwaldt, P.A. Jensen, K.G. Knudsen, A.D. Jensen, “A Review of Catalytic Upgrading of Bio-Oil to Engine Fuels,” Applied Catalysis A: General, 407 (2011), 1-19.
[7]Beshara. M., G. Rosinski, C. Olsen, B. Prins, G. Jacobs, A. Birch, E. Kohler, “Successful Implementation of State-of-the-art ULSD/Dewaxing Technology at Irving Oil, Saint John, NB,” Catalagram, 36-42, 2008
[8]Kagami, N., B.M. Vogelaar, A.D. van Langeveld, J.A. Moulijn, “Relationship Between Catalyst Structure and HDS Reaction Mechism,” Prep. Pap.-Am Chem Soc., Fuel Chem. 48 (2), 601, 2003.
[9]Burgess-Clifford, C, and H.H. Schobert, “Development of a Coal-Based Jet Fuel,” Prep. Pap.-Am Chem Soc., Fuel Chem. 52 (2), 403, 2007.
[10]Wang, H., J. Male, Y. Wang, “Recent Advances in Hydrotreating of Pyrolysis Bio-Oil and its Oxygen-Containing Model Compounds,” dx.doi.org/10.1021/cs400069z|ACS Catal. 1047-1070, 2013.
[11]Balster, L.M., et al, “Development of an Advanced, Thermally Stable, Coal-based Jet Fuel,” Fuel Processing Technology 89, 364-378, 2008.
[12]ASTM D1655-REV C, “Standard Specification for Aviation Turbine Fuels,” December 1, 2016.
[13]ASTM D7566-16b, “Standard Specification for Aviation Turbine Fuel Containing Synthesized Hydrocarbons,” July 1, 2016,
[14]MIL-DTL-83133J, “Detail Specification: Turbine Fuel, Aviation, Kerosene Type, JET A (NATO F-34), NATO F-35, and JET A+100 (NATO F-37),” 16 December 2015.
[15]ASTM D4054, “Standard Practice for Qualification and Approval of New Aviation Turbine Fuels and Additives,” August 1, 2016.
[16]MIL-HDBK-510A, “Department of Defense Handbook Aerospace Fuels Certification,” August 4, 2014.
[17]Striebich,R.; Shafer,L.; DeWitt, M.J.; West, Z.; Edwards, T.; Harrison III, W. “Dependence of Fuel Properties during Blending of Iso-Paraffinic Kerosene and Petroleum-Derived Jet Fuel,” AFRL-RZ-WP-TR-2009-2034. 2008.
[18]DeWitt, M.J.; West, Z.; Zabarnick, S.; Shafer, L.; Striebich, R.; Higgins, A.; Edwards, T. “Effect of Aromatics on the Thermal-Oxidative Stability of Synthetic Paraffinic Kerosene,” Energy & Fuels. 28(6), 3696–3703. DOI: 10.1021/ef500456e, 2014.
[19]Molburg, John C. and Doctor, Richard D. Hydrogen from Steam-Methane Reforming with CO2 Capture, 20th Annual International Pittsburgh Coal Conference, June 2003.
[20]“Assessment of Hydrogen Production with CO2 Capture: Volume 1: Baseline State-of-the-Art Plants”. US Department of Energy, 2010.
[21]Chemical Engineering, “Chemical Engineering Plant Cost Index (CEPCI), April 1998 – April 2017, McGraw-Hill, New York, NY.
[22]Maloney, T., “Railserve LEAF News Release”. Retrieved from Railserve LEAF: http://www.railserveleaf.biz/assets/leaf_info/RSI_WhtPap-fuel-emissions.pdf, Dec 2012.
[23]Srivastava, R. K., “Controlling of SO2 Emissions: A Review of Technologies”. U.S. Environmental Protection Agency, 2000.
[24]Sargent and Lundy, LLC, “Wet Flue Gas Desulfurization Technology Evaluation”. National Lime Association, 2003.
[25]Shah, V. E., “Cost and Performance Baseline for Fossil Energy Plants Volume 4: Coal-to-Liquids via Fischer-Tropsch Synthesis”. US Department of Energy, DOE/NETL-2011/1477, 2014.
[26]Energy Information Administration, “US Energy Information Administration Monthly Energy Review”. Retrieved from US Energy Information Administration: www.eia.gov, Sept 2011.
[27]ICIS, “Indicative Chemical Prices A-Z”. Retrieved from ICIS: http://www.icis.com/chemicals/channel-info-chemicals-a-z/, Aug 2016
[28]“Today in Energy”. Retrieved from Energy Information Agency: http://www.eia.gov/todayinenergy/detail.php?id=11671, June 2013
[29]Kennel, E., “Economic Value of Products” (R. K. Smith, Interviewer), Aug 2016.
[30]Haslbeck, J., “Cost and Performance Baseline for Fossil Energy Plants Volume 1: Bituminous Coal and Natural Gas to Electricity”, US Department of Energy DOE/NETL, 2010/1397, 2012.
[31] “Environmental management – Life Cycle Assessment – Principles and Framework (ISO 14040:2006), ISO. Geneva, Switzerland, 2016.
[32]“The Greenhouse Gases, Regulated Emissions, and Energy Use in Transportation (GREET) Model”, GREET 2016 version: 13098, Argonne National Laboratory, Argonne, IL, USA., 2016
[33]GreenDelta GmbH, “openLCA version 1.5.0”, Berlin, Germany, 2016.
[34]Chauhan, S. P. and D. Garbark, “Process of Providing Liquid Fuels and Polyols From Coal, Lignin, and Petroleum Residues Using Biomass-Derived Solvents”, Patent Application No. 62/532,801, July 2017.
LIST OF ACRONYMS AND ABBREVIATIONS
- ABP Atmospheric Boiling Point
- ACS American Chemical Society
- AIChE American Institute of Chemical Engineers
- AR As received
- ARA Advanced Research Associates
- Bbl barrel, 42 U.S. gallons
- BPD Barrel per day
- CBTL Coal-biomass to liquid
- CCS Carbon Capture and Storage
- CH Catalytic Hydrothermolysis
- CSTR Bench-Scale Liquefaction Reactors
- CTD Coal Tar Distillate
- CTL Coal-to-Liquids
- DBT Dibenzothiophene
- DCL Direct Coal Liquefaction
- DMF Dimethylformamide
- DOE Department of Energy
- EDS Exxon Donor Solvent
- FeS Troilite
- FSI Free Swelling Index
- FT Fischer Tropsch
- GC-MS Gas Chromatography–Mass Spectrometry
- GCxGC Comprehensive Two-dimensional Gas Chromatography
- GHG Greenhouse Gas
- GHSV Gas Hourly Space Velocity
- HCR Hydrocracking
- HDA Hydrodearomatization
- HDM Hydrodemetallization
- HDN Hydrodenitrogenation
- HDO Hydrodeoxygenation
- HDS Hydrodesulfurization
- HDT Hydrotreating
- LHSV Liquid Hourly Space Velocity
- MAF Moisture- and Ash-Free
- MF Moisture-Free
- MT Metric ton (tonne)
- MTPD Metric ton (tonne) per day
- N Nitrogen
- NDA Non-disclosure Agreement
- Ni Nickel
- NiO Nickel oxide
- NiW Nickel-Tungsten
- NMR Nuclear Magnetic Resonance
- O Oxygen
- ODSA Ohio Development Services Agency
- PAH Polynuclear Aromatic Hydrocarbons
- Pd Palladium
- PI Principal Investigator
- PMP Project Management Plan
- PSU Pennsylvania State University
- Pt Platinum
- PTW Pump-to-Wheels
- RCTD Recycle Coal Tar Distillate
- S Sulfur
- SBO Soybean Oil
- SiC Silicon carbide
- SIP Synthesized Iso-Paraffins
- SPK Synthetic Paraffinic Kerosene
- STP Standard Temperature and Pressure
- TEA Techno-economic Analysis
- THF Tetrahydrofuran
- TPD Ton per day
- TRL Technology readiness level
- UDRI University of Dayton Research Institute
- USAF United States Air Force
- W Tungsten
- WFE Wiped Film Evaporator
- WTW Well-to-Wheels
- WTP Well-to-Pump
- WV West Virginia
- WVU West Virginia University
APPENDIX A
LITERATURE REVIEW OF CATALYSTS FOR UPGRADING COAL-DERICED SYNCRUDE
Liquid fuels can be produced from coal through direct coal liquefaction (DCL) processes at 450-500 °C under 15-30 MPa hydrogen in a suitable solvent with appropriate catalysts. [1-5] In many processes, the solvent used can facilitate the heat and mass transfer during chemical reactions, and function as a hydrogen donor by shuttling hydrogen from the gas phase to the coal. Catalysts were often used to increase the rates of the desirable reactions such as the cracking, hydrogenation, and oxygen/nitrogen/sulfur removal reactions. Direct coal liquefaction was developed as a commercial process in Germany based on research pioneered by Friedrich Bergius in the 1910s (so-called the Bergius process). Table A-1 summarizes the DCL processed developed in different countries [2-5] and the operating parameters and experimental results of four major DCL processes are shown in Table A-2. [2-6]
Catalysts were employed in almost all of the DCL processes developed. Iron-based catalysts, such as pyrite (FeS2), troilite (FeS), pyrrhotite (Fe1-xS), iron oxide, iron sulfate, iron hydroxide and other iron resources, have been studied extensively due to their low costs and environmental tolerance. It is widely accepted that pyrrhotite is the most active form for iron sulfide catalysts. The iron-based catalysts could promote coal pyrolysis by markedly reducing the pyrolysis activation energy [7]. The major role of an iron-based catalyst in DCL is to promote the formation of activated hydrogen atoms and accelerate the secondary distribution of hydrogen atoms in the whole reaction system [8]. It was widely recognized that a highly dispersed catalyst can be superior to a supported catalyst, because the dispersed catalyst has an intimate contact with the surface of coal particles, which facilitates the activation and transfer of hydrogen to the coal-derived fragments and reactive sites. Accordingly, finer particles and a higher dispersion of the catalyst species would lead to a higher catalytic activity [8-11].
In addition to the Fe-based catalysts discussed above, Mo, Co and Ru were also tested as the catalysts for DCL. The results also implied that there were synergistic effects between the Ni and Mo catalysts [12-13] on promoting coal conversion and oil yield. Another type of novel catalysts for DCL is SO42−/MxOy solid acid, such as SO42−/Fe2O3 and SO4 2−/ZrO2 [14-15]. The solid acid could be a bi-functional catalyst for pyrolysis and hydrogenation. The use of SO42−/ZrO2 catalyst in the DCL of Shenhua coal at 400 °C and 4.0 MPa H2 resulted in a coal conversion and (gas + oil) yield, up to 76.3% and 62.5%, respectively, much higher than those obtained with FeS or (FeS + S) catalyst under the same conditions.
Table A-1. Summary of major DCL processes developed around the world [2-5].
Table A-2. The operating parameters and experiment results of some major DCL processes [2-6].
Catalytic hydrotreating (HDT) plays an essential role in the DCL processes for the conversion of heavy feedstocks and for improving the quality of oil products. Hydrotreating refers to a variety of catalytic hydrogenation processes, which saturate unsaturated hydrocarbons and remove S, N, O and metals from different petroleum streams in a refinery. Hydrotreatment usually implies only small changes in overall molecular structure but hydrocracking (HCR) reactions often occur simultaneously and may in fact be desired. Depending on the nature of the feed and the amount and type of the different heteroatoms (i.e., different reactivities compounds), specific hydrotreating processes have been developed. The reactions occurring during hydrotreating are hydrodesulfurization (HDS), hydrodenitrogenation (HDN), hydrodeoxygenation (HDO), hydrodearomatization (HDA), hydrodemetallization (HDM), and hydrodeasphaltenization (HDAs). For a long time, the most important hydrotreating reaction has been the removal of sulfur from various fuel fractions. Consequently, hydrotreating catalysts are also commonly referred to as hydrodesulfurization (HDS) catalysts. Typical hydrodesulfurization catalysts consist of molybdenum supported on an alumina carrier with either cobalt or nickel added as promoters for improving the catalytic activity.
The most common combinations of active elements in hydrotreating catalysts are the CoMo, NiMo and NiW families [16] such as Criterion NiMo SynCat-37, Grace NiMo AT-505 catalysts. The concentration by weight of the metals is usually 1-4% for Co and Ni, 8-16% for Mo and 12-25% for W. Typical support materials are alumina, silica-alumina, silica, zeolites, kieselguhur, and magnesia, with surface areas in the 100-300 m2/g range. CoMo catalysts are excellent HDS catalysts but are somewhat less active for HDN and hydrogenation of aromatics. As a result, the CoMo catalysts give relatively low hydrogen consumption. NiMo catalysts, on the other hand, are very good HDN and hydrogenation catalysts but give rise to higher hydrogen consumption. Consequently, NiMo catalysts are often preferred for treating unsaturated feeds. An increase in the selectivity of heteroatom removal vs. hydrogenation can be achieved for alumina-supported catalysts by addition of P [9-14]. Of the three combinations mentioned above, the NiW catalysts have the highest activity for aromatic hydrogenation at low hydrogen sulfide partial pressures [23-25] and are also active for HCR, but their use has been limited due to the higher cost.
In hydrotreating, reaction rates are often influenced by diffusion in the catalyst pores. Thus, the choice of catalyst particle size and shape, as well as the geometry of the pore system, is important [16]. This is especially true in the treatment of heavier feeds, where the reactions may be limited by diffusion of reactants and products in and out the pore system. The diffusion restrictions will in general become more severe during operation due to the deposition of metals and coke at the pore entrances. Therefore, for real application, there is a tendency to use small catalyst particles with relatively large pores and shapes, which expose appreciable external surface area [26-27]. For a given equivalent diameter, these shapes, especially rings, have the advantage of minimizing the pressure drop across the reactor [27]. Song’s Research Group at PSU has also explored novel dispersed (unsupported) sulfide catalysts for deep desulfurization of more refractory sulfur compounds in middle distillate fuels. High metal loaded NiMo and CoMo/MCM-41 catalyst show high HDS activity of 4,6-DMDBT at 300 and 325°C. Specifically, NiMo/MCM-41 has higher HDS activity than other catalysts and even higher than commercial NiMo catalyst (Cr424), which contains 14wt% MoO3 and 3wt% NiO on alumina before sulfidation.
Metal phosphides are a novel catalyst group for deep hydrotreating and have received much attention due to their high activity for HDS and HDN [28-36]. Transition metal phosphide catalysts have been studied in hydrogenation reactions [37-40] but research focusing on hydrotreating has been carried out only recently. There has been heightened interest in new supports for HDS catalysts in recent years, due to the need to improve catalytic activity, and the availability of new materials of high surface area with new properties. The alteration of catalytic activity by the support may arise as a result of changes in dispersion and morphology of the active component and possible metal–support interactions. Supported Ni phosphide catalysts were also tested by Song’s Group at the same conditions as the NiMo and CoMo catalysts were. Ni2P/MCM-41 had higher activity than the other phosphide catalysts and Ni2P/Al2O3 had a very high HYD/DDS ratio, while CeO2 supported Ni phosphide had a lower ratio [35].
It should be pointed out that the coal-derived liquids, especially through the process developed in this project using bio-oil as the H-donor solvent, may have relatively high concentrations of oxygen-containing molecules which may include ethers, furans, carboxylic acids and phenols [41]. Conventional metal sulfide catalysts [42-46] and noble-metal catalysts were employed for hydrodeoxygenation (HDO) processes of bio-oil [47-48]. The metal sulfide catalysts are cheaper than noble metal catalysts, and they have good hydrogenation activity. However, because of the low sulfur content of bio-oil, an additional sulfiding agent is needed to maintain the hydrogenation activity of the metal sulfide catalysts or the hydrogenation performance of the metal sulfide catalysts will degrade due to the graduate loss of sulfide. In addition, these typical hydrodesulfurization catalysts such as NiMoS/Al2O3 and CoMoS/Al2O3 were found to quickly deactivate by coke deposition in HDO reactions because of the acidity of the reactant [49]. To the contrary, transition metal phosphides supported on neutral silica could be a promising class of new hydroprocessing catalysts [50-51] for HDO reactions.
In a summary, for a specific hydrotreating application with given feed and product specifications, the choice of catalyst is seldom only related to catalyst activity. Many other features are important, such as catalyst life, activity toward side reactions, and pressure drop build-up. Ease of activation, regeneration and price should also be considered. Furthermore, for certain applications, the optimum solution may be to use different types of catalyst in the same reactors. Catalyst selection thus generally requires a detailed study of the specific situation. In addition, the process may require the use of mixed or multiple beds of catalysts.
REFERENCES FOR APPENDIX A
[1]Robinson, K.K. Reaction engineering of direct coal liquefaction. Energies 2009, 2, 976-1006.
[2]Whitehurst, D.D. Coal Liquefaction Fundamentals; ACS: Washington, DC, USA, 1980.
[3]Hirano, K. Outline of NEDOL coal liquefaction process development pilot plant program. Fuel Proc. Technol. 2000, 62, 109-118.
[4]Comolli, A.G.; Lee, T.L.K.; Popper, G.A.; Zhou, P. The Shenhua coal direct liquefaction plant. Fuel Proc. Technol. 1999, 59, 207–215.
[5]Kouzu, M.; Koyama, K.; Oneyama, M.; Aramaki, T.; Hayashi, T.; Kobayashi, M.; Itoh, H.; Hattori, H. Catalytic hydrogenation of recycle solvent in a 150 t/d pilot plant of the NEDOL coal liquefaction process. Fuel 2000, 79, 365–371.
[6]Shui, H.; Cai, Z.; Xu, C. Recent Advanced in Direct Coal Liquefaction. Energies 2010, 3, 155-170.
[7]Li, X.; Hu, S.; Jin, L.; Hu, H. Role of iron-based catalyst and hydrogen transfer in direct coal liquefaction. Energ. Fuel 2008, 22, 1126-1129.
[8]Hirano, K.; Kouzu, M.; Okada, T.; Kobayashi, M.; Ikenaga, N.; Suzuki, T. Catalytic activity of iron compounds for coal liquefaction. Fuel 1999, 78, 1867-1873.
[9]Ikenaga, N.; Kan-nan, S.; Sakoda, T.; Suzuki, T. Coal hydroliquefaction using highly dispersed catalyst precursors. Catal. Today 1997, 39, 99-109.
[10]Liu, Z.; Yang, J.; Zondlo, J.W.; Stiller, A.H.; Dadyburjor, D.B. In situ impregnated iron-based catalysts for direct coal liquefaction. Fuel 1996, 75, 51-57.
[11]Dadyburjor, D.B.; Fout, T.E.; Zondlo, J.W. Ferric-sulfide-based catalysts made using reverse micelles: Effect of preparation on performance in coal liquefaction. Catal. Today 2000, 63, 33-41.
[12]Song, C.; Parfitt, D.S.; Schobert, H.H. Bimetallic dispersed catalysts from molecular precursors containing Mo-Co-S for coal liquefaction. Energ. Fuel 1994, 8, 313-319.
[13]Hulston, C.K.J.; Redlich, P.J.; Jackson, W.R.; Larkins, F.P.; Marshall, M. Nickel molybdate-catalysed hydrogenation of brown coal without solvent or added sulfur. Fuel 1996, 75, 1387-1392.
[14]Wang, Z.; Shui, H.; Zhang, D.; Gao, J. A comparison of FeS, FeS + S and solid super-acid catalytic properties for coal hydro-liquefaction. Fuel 2007, 86, 835-842.
[15]Wang, Z.; Shui, H.; Zhu, Y.; Gao, J. Catalysis of SO42−/ZrO2 solid acid for the liquefaction of coal. Fuel 2009, 88, 885-889.
[16]Anderson, J.R.; Boudart, M. (Eds.) Hydrotreating Catalysis. Catalysis Science and Technology, Vol. 11, 1996, Springer-Verlag Berlin Heidelberg.
[17]Christensen H, Cooper BH (1990) Nat AIChE Meeting, March 18-22.
[18]Eijsbouts S, van Gruijthuijsen L, Volmer J, de Beer VHJ, Prins R (1988) Paper 67F, AIChE 88th Nat Meeting, Washington.
[19]Zeuthen P, Jacobsen AC, Nielsen IV (1990) 13th North American Meeting of the Catalysis Society, May 2-7.
[20]Eijsbouts S, van Gestel JNM, van Veen JAR, de Beer VHJ, Prins R (1991) J Catal. 131-412.
[21]Morales A, Prada Silvy R, Leon V (1992) Proc 10th Int Congr Catal, Guczi L et al. (eds) Elsevier, Amsterdam, p 1899.
[22]Cooper B H, Stanislaus A, Hannerup PN (1992) ACS Nat Meeting, San Francisco
[23]Stanislaus A, Cooper BH, Catal Rev-Sci Eng 1994, 36 (1), 75.
[24]Hannerup PN, Cooper BH (1994) Annual EWEFA ConfNoordwijk, Holland, June.
[25]Schott JW, Bridge AG (1971) Adv Chern Ser 103, Amer Chern Soc, Washington, DC.
[26]Bartholdy J, Cooper BH, ACS Prepr. Div. Petro. Chem., 1993, 38, 386.
[27]Cooper B H, Donnis B B L, Moyse BM, Technology Dec 8, Oil & Gas J., 1986, p 39.
[28]W.R.A.M. Robinson, J.N.M. van Gastel, T.I. Kora´nyi, S. Eijsbouts, J.A.R. van Veen, V.H.J. de Beer, J. Catal. 161 (1996) 539.
[29]W. Li, B. Dhandapani, S.T. Oyama, Chem. Lett. (1998) 207.
[30]C. Stinner, R. Prins, Th. Weber, J. Catal. 191 (2000) 438.
[31]D.C. Phillips, S.J. Sawhill, R. Self, M.E. Bussell, J. Catal. 207 (2002) 266.
[32]S.T. Oyama, J. Catal. 216 (2003) 343.
[33]Y. Shu, S.T. Oyama, Carbon 43 (2005) 1517.
[34]Y. Shu, Y.-K. Lee, S.T. Oyama, J. Catal. 236 (2005) 112.
[35]J.H. Kim, X. Ma, C. Song, Y.-K. Lee, S.T. Oyama, Energy Fuels 19 (2005) 353.
[36]Lee, Y.; Shu, Y.; Oyama, S.T. Appl. Catal. A: Gen. 2007, 322, 191-204.
[37]N.P. Sweeny, C.S. Rohrer, O.W. Brown, J. Am. Chem. Soc. 80 (1958) 799.
[38]E.L. Muetterties, J.C. Sauer, J. Am. Chem. Soc. 96 (1974) 3410.
[39]F. Nozaki, R. Adachi, J. Catal. 40 (1975) 166.
[40]F. Nozaki, M. Tokumi, J. Catal. 79 (1983) 207.
[41]Leyva, C.; Rana, M.S., Trejo, F.; Ancheyta, J. Ind. Eng. Chem. Res. 2007, 46, 7448-7466.
[42]A. Y. Bunch, U. S. Ozkan, J. Catal. 2002, 206, 177.
[43]O. I. Senol, T. R. Viljava, A. O. I. Krause, Catal. Today 2005, 100, 331.
[44]Y. Q. Yang, C. T. Tye, K. J. Smith, Catal. Commun. 2008, 9, 1364.
[45]R. Nava, B. Pawelec, P. Castaño, M. C. Álvarez-Galván, C. V. Loricera, J. L. G. Fierro, Appl. Catal., B 2009, 92, 154.
[46]A. Popov, E. Kondratieva, L. Mariey et al., J. Catal. 2013, 297, 176.
[47]E. L. Kunkes, D. A. Simonetti, R. M. West, J. C. Serrano-Ruiz, C. A. Gärtner, J. A. Dumesic, Science 2008, 322, 417.
[48]D. Y. Hong, S. J. Miller, P. K. Agrawal, C. W. Jones, Catal. Commun. 2010, 46, 1038.
[49]Laurent, E.; Centeno, A.; Delmon, B. in: Proc. 6th Inter. Symp. Catalyst Deactivation, Stud. Surf. Sci. Catal. 1994, 88, 573-578.
[50]Oyama, S.T. J. Catal. 2003, 216, 343.
[51]Oyama, S.T.; Gott, T.; Zhao, H.; Lee, Y.-K. Catal. Today 2009, 143, 94-107.
APPENDIX B: CTL Process Design and Equipment Sizing
The process design is adapted from the CTL Chemcad model using 47 tonnes/hr of coal as its basis and is used for the process conceptual development and equipment sizing. Flow rates from the ChemCAD model were used to estimate the throughput needed for each unit operation. Pumps and heat exchange surfaces were sized mostly from the values provided by the ChemCAD model. Tanks were sized for process inlet and outlet storage using maximum flow rates in the process and estimated duration of storage.
All Figures referenced below are inserted at the end of Appendix B.
Area 100 – Tank Farm
Figure B-1 shows Area 100, the raw material tank farm. Quotes were obtained for all equipment in the raw material tank farm area. The storage tanks were quoted as field erected with the erection cost included. Foundations were to be completed by others. The suction heaters were selected based on flow and condensing steam on the tubes. If we use a thermal oil for heating, these could be undersized. Conversely, the thermal oil could be used to generate steam for the suction heaters. Pumps were sized based on the material balance and a nominal head of 60 ft of fluid. Rail car unloading stations could be dedicated to a raw material or multi-use. There was no need for spare equipment in this area.
Area 200 – Coal Prep
The coal preparation system is shown in Figure B-2. The plant is assumed to be located near a coal mine, with the coal being delivered to the plant property fence line. The coal handling is sized for 100 tons per hour to allow for some down time in the coal yard for maintenance, break time, and refueling the loader without interrupting the reactor flow. The continuous usage is 47 tonness per hour. The distribution conveyor was sized based on 150 m of length for a conveyor to the coal pile. The pile collection conveyor was also assumed to be 150 m in length. A coal elevator would be used to lift the coal up to the storage silos, using buckets and running at 53 m per minute. Three raw coal storage silos are based on three days of inventory, with each silo holding 1200 tons of coal. A crusher feed conveyor is will feed the Impact Dryer System and is sized for a length of 61 m.
The Impact Dryer System was quoted by the Williams Patented Crusher and Pulverizer Company. The system consists of a feed system, crusher and classifier, cyclone separator, baghouse main circulating fan, heater with burner and an exhaust fan. Ground coal product is estimated to contain about 1% moisture with about 99% passing through a 40 mesh screen. This is a finer size than the 8 mesh coal used in the lab scale testing.
The crushed coal will be transported to two finished coal storage bins, each holding 1200 tons, using a 76 m belt conveyor with walkway. Two 30 m long crushed coal conveyor will feed the mix tank feed bins located above the mix tanks. The crushed coal will discharge from the feed bin through a rotary valve, onto a controlled flow feed bin conveyor. This crushed coal will then drop onto the coal distribution conveyor and be diverted to the mix tanks when making a batch of slurry to feed the reactors.
Conveyor costs were obtained from the Matches website (http://www.matche.com). Costs are in 2014 dollars, and are based on a 1 m wide conveyor running at 150 m per minute. Coal storage bin costs were also obtained from the Matches website.
Spare conveyors were not added to the capital costs. It is assumed that there would be spare parts purchased for needed repairs or replacements, such as motors, pulleys, bearings, and belts. The impact crusher would also have spare parts that would be expected to be stored on site, such as motor, bags, crusher bits, bearings.
Area 300 – Coal Slurry Mix Tanks
Figure B-3 shows the coal slurry mix tanks system. Four 38,750 l mix tanks will be required for the reactor feed system. They are batch filled will hold about 20 minutes of slurry each. It is assumed that the coal can be mixed with the coal tar distillate in 20 minutes. The four stages for the mix tanks will be:
- Fill with liquid (20 min)
- Start agitator, add powdered coal (20 min)
- Mix (20 min)
- Feed reactors (20 min)
The slurry will be mixed with the yellow grease, brown grease and tall oil in the suction line of the reactor feed pumps. Each reactor feed pump will pull from a common suction line to a reactor train. Feed to the reactor will be measured and controlled by the speed of the reactor feed pump. The reactor feed will be heated to the reaction temperature with reactor train feed heat exchanger, using a thermal fluid. It is assumed that we will be able to use the 500+ product from the reaction as the thermal fluid for the high temperature requirements.
Quotes were obtained for the mix tank, the slurry pump and the heat exchanger. The agitator was estimated using the matches website based on preliminary horsepower requirement. The reactor feed heat exchanger pump cost was also estimated using the matches website.
Installed spare parts include a feed mix tank with agitator, a reactor feed pump, reactor train feed heat exchanger, and its thermal fluid circulating pump.
Area 400 – Reactors
Figure B-4 and Figure B-5 show the reactor system process flow diagram. Overall single stage yield for 30 minutes in both pilot and laboratory was determined to be >85 percent, with values of >90 percent routinely obtained with the solvent combinations defined. A series of 3 trains of 4 CSTRs, each 100,000 L volume, was evaluated for operations. Based on first order kinetics, the reactors would give enough volume to react to about 93% based on an 85% single-stage yield. On this basis, an overall coal yield of 90% was taken as a reasonable value. Even though reaction times as low as 10 minutes were demonstrated in the laboratory with greater than 85% conversion, an average of 30 minutes of residence time was used for the reactor train. Two trains were sufficient for operation; a third train was used in costing to allow the process to achieve 90% on-stream time.
The reactor flow rate is 181,670 liters per hour. For a 30 minute residence time, the required reactor volume is 90835 liters. Using three reactor trains consisting of 4 reactors each, the required liquid reactor volume is 7570 liters each. The nominal reactor volume is 10,000 liters to allow for sufficient free board. Reactors are to be operated at 400 C and 405 psig. The first reactor will be operated at a higher pressure sufficient to overcome the pressure drop which allows for the flow to cascade from the first to the fourth reactor. Reactor pressure will ultimately be controlled by the valve on the slurry discharge line The reactors are connected by a pipe which Is split to feed both the top and bottom of the reactor. This allow the solids to be introduced to the bottom of the reactor and also allows the non-condensable gassed generated to pass from one reactor the next.
The fourth reactor in each reactor train will have a partial condenser, reflux vessel, reflux pump, reflux heater with a thermal fluid circulating pump, and a final condenser. The partial condenser will cool the vent stream using thermal fluid exiting the feed heater. The condensed fluid will then be heated and returned to the fourth reactor.
Further process development may indicate that condensate should be returned to the first reactor, if the components are to be incorporated into the reaction products and need more residence time. The slightly higher pressure in the first reactor may keep the reflux in the liquid phase longer to allow for the reactions to proceed.
The reactor products contain about 4.5% solids consisting of principally coal ash, unreacted coal solids, and silica. The product will flow from the reactors through a partial cooler to reduce the chance of vapor flashing through the pressure reducing valve. This high pressure letdown from 400 psig to about 150 psig is expected to wear and will need to be rebuilt periodically. An installed spare with isolation valving is estimated.
Quotes were obtained for the reactors, the partial condenser, and the reflux heater, product cooler, and reflux pump. The reflux tank cost was ratioed off of the reactor cost using the volume raised to the 0.6 power. The estimated final condenser surface area was within 10% of the partial condenser surface area, so the same cost was used for both. The agitator was estimated using the matches website based on preliminary horsepower requirement. The reactor jacket pump cost was also estimated using the matches website.
Installed spare equipment includes an additional pump for each pump in the reactor area, and two reactors with agitators.
Area 500 – Hydrocyclone and Centrifuge
A combined hydrocyclone/centrifuge system was chosen for the process. Figure B-6 shows the hydrocyclone process flow diagram. Figure B-7 shows the centrifuge process flow diagram. The reduced pressure product from the reactors will flow to the high pressure separator tank, allowing any flashed vapor to separate, before flowing to the hydrocyclone. The hydrocyclone will separate the solids in the product while still maintaining the high temperature of the product. The design is based on 90% of the solids contained in 25% of the total hydrocyclone flow in the hydrocyclone bottoms product. Quoted performance based on an assumed particle distribution is much better that the design and may simplify the process. The reported performance would need to be proved at the pilot plant level. The hydrocyclone lights will feed the first stage of the evaporator.
Evaporator bottoms from each train will be mixed with the hydrocyclone bottoms from each train in a common centrifuge feed tank. This heavy solids product will be pumped with a slurry pump through a cooler and through a centrifuge. The centrifuge centrate will be the 500+ product while the centrifuge bottoms will be the heavy solids product.
If the hydrocyclone proves to be as efficient as quoted, the hydrocyclone bottoms may be the heavy solids product directly. This would contain some portion of the product oil, as it was not driven off before the solids separation.
If the 500+ liquid can be sold with the 3% solids, the centrifuge may be eliminated. If not, the centrifuge may be required for just the 500+ fluid, reducing the size of the centrifuge from 53 m3/hr to 8 m3/hr.
Quotes were obtained for the high pressure separator tank, the hydrocyclone, centrifuge feed tank and the centrifuge feed pump. The centrifuge feed cooler was ratioed from a previous quote. The agitator was estimated using the matches website based on preliminary horsepower requirement. The centrate pump cost was also estimated using the matches website.
Installed spare equipment for this area include a complete hydrocyclone train and a redundant centrifuge system.
Area 600 – Evaporator
Figure B-8 shows the system used to separate the <500°C product for fuel processing from the >500°C product. The evaporator system is based on a triple effect evaporator, using vapor from two stages as the heating medium for two following stages. The selection of three evaporators will need to be proven and optimized in a further design effort. Latent heats at each pressure and temperature, along with any boiling point rise will need to be determined in further research.
Each effect is sized at 7570 liters which allow about 50% of the volume for vapor separation. Each effect will require an external heat exchanger and circulating pump, as a jacket would not provide sufficient heat to accomplish the evaporation load.
Quotes were obtained for the vessels and the vacuum pump. Heat exchangers surface areas were determined using a conservative overall heat transfer coefficient. The heat exchangers were estimated using the matches website based on surface area and pressure. The pump costs were also estimated using the matches website.
Installed spare equipment for this area include a second pump for each instance where a pump is required and a spare vacuum pump.
Area 700 – Product Tank Farm
The product tank farm (Figure B-9) is sized to accommodate seven days of production. The product oil is assumed to not need a suction heater to pump to the rail car loading station. Four rail loading stations are included in the estimate. Four 182 m3/hr load out pumps will be manifolded to the 15 – 380,000 liter product oil storage tanks. The 500+ liquid product will require a suction heater and local pump for each of the 4 - 380,000 liter storage tanks. The heavy solids product will also require a suction heater and local pump for each of the 4 380,000 liter storage tanks. The pump out rate for these two fluids will be dependent on the suction heaters’ ability to reach a sufficient pumping temperature. The maximum size suction heater offered by the vendor was selected. Recycle from the loading station to the tanks should be provided to prevent slurry settling and maintain an acceptable pumping temperature.
Quotes were obtained for the storage tanks, suction heaters, slurry pumps, and the load out station. 500+ liquid product pump costs were estimated using the matches.com website.
Area 800 - Process Utilities
A process furnace, scrubber, stack, and support equipment were defined and costed to provide energy to the syncrude process. A hot oil circulation and storage system was defined. The data from the ChemCAD flowsheet on process flows was used to size the items for cost.
A cooling tower and circulation system was incorporated to provide process cooling that was not provided by the air coolers. Air compression would likely be required for instrumentation operation and some other miscellaneous uses. A separate air compressor is not costed in direct equipment but would be provided through the service facilities allocations used.
Process Concerns and Potential Alternates
Thermal fluid
The process temperature requirement of 400oC is above is at the maximum use temperature of commercially available thermal fluids. The simulation uses a thermal fluid of 475oC to allow for a temperature difference to transfer the heat to the reactants. We are assuming that our 500+ product could be used as the thermal fluid for the reactors. There are cooling needs in the process that are suited to using a cooler thermal fluid, operating between 38oC and 60oC. The simulation indicates that the 500+ fluid would have a high viscosity at the cooling temperatures. Further evaluation of the reactor heating system should be addressed in the future. Thermal cooling could be by using air cooled fintubes directly, or using an evaporative cooler, which may reduce energy integration. The heating duty could be accomplished with an eutectic nitrate salt circulating system.
Centrifuge and hydrocyclone
A hydrocyclone was selected in the design to allow for solids separation without cooling to the centrifuge maximum temperature of 100°C. This allowed for the bulk of the fluid to pass to the evaporator with requiring the heat duty to bring it up to the evaporation temperature. The disadvantage is that the heavy solids product will have some product oil remaining in it, unless the heavy solids is stripped in a further processing step.
The evaporator bottoms will also carry a small amount of solids, and could be passed through a centrifuge or filter to separate the solids from the 500+ liquid or combined with the heavy solids product and sent to the centrifuge.
Start Up and Shutdown Requirements
The start-up and shutdown requirements have not been examined fully and may require additional tanks to hold flush out liquids and partially reacted fluids. Flush out with other materials, such as CTD, may also require additional material purchase which is not reflected in tankage. Tankage may not be required, and this flush may be able to be handled directly with 1-2 rail cars of material.
Area H – Hydrotreater
Because the goal of this study was to evaluate the cost of fuel and carbon emissions from this process, a hydrotreater area was defined. A detailed design was not performed; instead, verbal quotes on a similar hydrotreater (4500 bbl/day) and a natural gas reformer were defined for this process. As mentioned below, a gasifier for the solid waste was also briefly considered, and may make sense if there is not a market for the material. At this time, Quantex is evaluating the market for the material and has recommended a price for that material.
Figure B-1. Raw Material Tank Farm Process Flow Diagram
Figure B-2. Coal Preparation Process Flow Diagram
Figure B-3. Coal Slurry Mix Tanks Process Flow System
Figure B-4. Reactor Process Flow Diagram – Part 1
Figure B-5. Reactor Process Flow Diagram – Part 2
Figure B-6. Hydrocyclone Process Flow Diagram
Figure B-7 Centrifuge Process Flow Diagram
Figure B-8. Evaporator Process Flow Diagram
Figure B-9. Product Tank Farm Process Flow Diagram
APPENDIX C: Capital Costs Estimation Details
Per the RFP, the cost estimate is stated in June 2011 dollars. Various indexes are available to factor costs based on time. The Chemical Engineering Plant Cost Index (CEPCI) has been used to factor the capital costs obtained to June 2011 dollars. This index is more aligned to chemical equipment costs than a Marshall and Swift index or an RSMeans index. Indices used in this evaluation are shown in Table C-1.
Table C-1. Chemical Engineering Plant Cost Index.
The Chemical Engineering Plant Cost Index (CEPCI) was used to take the current costs from quotes and factor it to June 2011 cost by using the ratio.
Equipment cost x (June 2011 CEPCI/May 2016 CEPCI) = June 2011 cost
Equipment cost x (588.9/543.5) = June 2011 cost
Example: Reactor cost, 60" I.D. x 192" LG. Vessel, 450 PSIG @ 932° F, w/ jacket = $194,100
$194,100 x (588.9/543.3) = $210,314
The Matches website estimates equipment costs in 2014 dollars. Equipment that was estimated using the Matches website uses the July 2014 CEPCI to yield a June 2011 equipment cost.
Equipment cost x (June 2011 CEPCI/July 2014 CEPCI) = June 2011 cost
Equipment cost x (588.9/576.1) = June 2011 cost
Example: Reactor Train #1 Jacket Pump, 400 gpm, 60 ft TDH, 10 hp, 6” discharge, steel = $16,200
$16,200 x (588.9/576.1) = $16,560
This study estimate is based on generating the equipment cost and using specific factors based on historical plant cost to predict the total installed cost of the equipment. A study estimate of this type should be within the range of 30% of the final cost of the plant. Factors have been developed for different types of plants, based on the primary state of the chemicals used, such as solids, liquids, gases and slurries. Material and labor factors for foundations, structural steel, buildings, insulation, Instruments, electrical, piping, painting, and other miscellaneous expenses are applied to the capital equipment cost to estimate the total installed cost of the equipment.
Table C-2. Distributive Factors for Bulk Materials
Example of determining the total installed cost of the reactor from the escalated equipment cost is shown below.
Table C-3. Example of Determining the Total Installed Cost of the Reactor from the Escalated Equipment
Another factor applied to the capital costs are the labor factors for setting equipment. For this study, most labor factors for setting equipment were 20%, with a 30% factor used for the coal crusher system as the only exception. Pump costs were also factored by 20% for setting.
Table C-4. Distributive Labor Factors for Setting Equipment
The following is an example of applying the Distributive Labor Factors for Setting Equipment on the reactor cost in June 2011 dollars.
Reactor capital cost = $210,314
Distributive Labor Factors for Setting a Tank = 20%
Reactor setting in place cost = $210,314 x 20% = $42,063
Table C-5 summarizes both the Syncrude and Hydrotreater costs for a 1,000 MTPD syncrude/4,000 BPD hydrotreater system. The costs are provided in FY 2011 (June) dollars.
Process equipment was sized using ChemCAD results and quotes were requested. Where quotes were available, those were used for equipment costs. Where quotes were not received, estimates were obtained from the www.matches.com webpage. Where neither were available, other literature sources, and finally verbal quotes were relied upon.
Factors were applied to the total equipment installed cost for miscellaneous electrical (10% bare equipment), piping and ducting (30% bare equipment), and service facilities (35% bare equipment). With these factors added, the cost for the syncrude process is $109,472,000 and the cost for hydrotreatment at 4500 barrels/day is $40,420,000 (rounded totals, June 2011 costs).
Table C-5. Summary Equipment Costs for Syncrude and Hydrotreater Systems
While this is the estimated capital cost, it is not the final costs for financing this project. To these capital costs the normal project and owner’s costs were added; these factors are shown in Table C-6
Table C-6. Project and Owner's Costs
The TASC is used to apply the capital cost factor in the operating costs. Note that it also has operating cost factors included in categories such as All Labor, Non-fuel Consumables, Waste Disposal, Fuel Costs, and Consumables. Therefore, the capital cost will affect the operating cost which will affect the capital cost. To allow iteration of these costs, two spreadsheet books were used; one for capital costing, and one for operating economics, so that iteration could be controlled easily.
The hydrotreater cost was estimated at 4,500 bbl/day. This size was determined based upon availability of a verbal estimate for this size unit for a similar material. The cost was developed and was taken back from 2015 costs to 2011 costs. The hydrotreater was then scaled using the 0.62 exponential factor to 4,000 bbl/day. Normally, hydrotreaters are much larger and are probably closer to 40,000 bbl/day in most modern refineries which means that the capital becomes less of a factor in the operating cost. The hydrotreater size in the F-T example cost data appears to be is about this 40,000 bbl size. For this reason, the hydrotreater estimate developed was also scaled to 36,000 bbl/day to represent eight 1,000 tonne/day coal plants or four 2,000 tonne/day plants as feed to the hydrotreater.
It was assumed that the syncrude plant would require 4 years of construction (2008-2011) and the construction would proceed along a 20/50/80/100 percent completion schedule for the years of construction. Interest costs at 7% were assessed for the installed cost to convert it to a project cost. For the hydrotreater, a 2-year 40/100 schedule (2010-2011) was planned with the same 7% interest rate assessed.
Applying all of these factors and appropriate percentages, similar to those used for F-T plant capitalization (Shah, 2014) gives the following estimates of installed equipment cost and TASC. Note that the CO2 and the electrical costs are both spelled out in the tables, because these are the primary fluctuating parameters in the cases evaluated, and operating costs affect the TASC through the working capital costs.
Advertisement
The 10 largest coal producers and exporters in Indonesia:
- Bumi Resouces (BUMI)
- Adaro Energy (ADRO)
- Indo Tambangraya Megah (ITMG)
- Bukit Asam (PTBA)
- Baramulti Sukses Sarana (BSSR)
- Harum Energy (HRUM)
- Mitrabara Adiperdana (MBAP)
- Samindo Resources (MYOH)
- United Tractors (UNTR)
- Berau Coal
