The present invention relates to a direct coal liquefaction process capable of producing gasoline, jet fuel, or diesel from high inertinite content coal at high thermal efficiency.
Increases in the cost of petroleum and concerns about future shortages have led to increased interest in the use as a fuel source of the vast, easily accessible deposits of coal that exist in several parts of the world. Various processes have been proposed for converting coal to liquid and gaseous fuel products including gasoline, diesel fuel, aviation turbine fuel and heating oil, and, in some cases, to other products such as lubricants and chemicals. A number of problems have hampered widespread liquefaction of coal, however, including the relatively low thermal efficiency of indirect coal-to-liquids (CTL) conversion methods, such as Fischer Tropsch (FT) synthesis and methanol-to-liquids (MTL) conversion, and concerns about CO2 emissions. Direct coal liquefaction (DCL) methods have been developed for liquefying coal that have advantages in many applications relative to conversion by FT synthesis, including substantially higher thermal efficiency and lower CO2 emissions. Such direct liquefaction methods have typically involved heating the coal in the presence of a donor solvent, and optionally a catalyst, in a hydrogen containing atmosphere to a temperature in the range of about 700° to 850° F. to break down the coal structure into free radicals that are quenched to produce liquid products. The catalyst can typically be very finely divided iron or molybdenum or mixtures thereof. One source of the molybdenum catalyst is via in situ formation from a phosphomolybdic acid (PMA) precursor.
The reported DCL coal conversion units require the use of a hydrotreater for preparing the donor solvent that is fed back to the input of the DCL unit to act as a solvent for the coal being converted and to provide additional hydrogen to the liquefaction process. Solvent hydrotreating requires separation of the solvent fraction, additional equipment, additional hydrogen-rich treat gas, and as much as 50% of the heat of reaction is moved from the liquefaction reactor to the solvent hydrotreater. Thus, addition of an external solvent hydrotreater increases the required investment and decreases thermal efficiency. This is an important reason why low donor solvent-to-coal ratios (typically 1.2 to 1.5) are used in processes using a solvent hydrotreater. In addition, the hydrotreating reduces the viscosity and lowers the aromatics content of the solvent, which reduces its ability to suspend ash in the slurry and its compatibility with coal. The reduction in ability to suspend ash results in an increased likelihood of solids buildup, deposits, or plugging of high pressure feed pumps, transfer lines, heat exchangers, furnace tubes, and reactors. Therefore, use of such donor solvents results in higher gas hold-up in the liquefaction reactor in order to maintain the solids in the slurry in suspension, which in turn, requires a large reactor volume to achieve adequate coal residence time in the reactor. High recycle gas rates are also required because treat gas must be provided to both the solvent hydrotreater and the liquefaction reactor. Because hydrotreaters are expensive, require energy for recycling gas, and use the heat of reaction less efficiently, DCL reactor systems have been designed to minimize the amount of donor solvent recycle.
In order to reduce the required reactor volume and minimize solids build-up in the liquefaction reactors, Shenhua and Headwaters utilize ebullated bed reactors. Circulation of liquid through the reactors results in a decrease in gas hold-up. Because of the high liquid recycle in ebullated bed reactors, the reactors are fully back-mixed which results in an increase in reactor volume versus a plug flow reactor.
A further issue limiting the application of DCL methods is that lower quality coals having inertinite content much higher than about 12% have been considered unsuitable for use as a DCL feed stock. Such high inertinite coals are found in many parts of the world, including the United States and China. Many of these coals, such as that in the Ordos basin in China (1,2,3), have inertinite content of more than 25% and a low ratio of atomic hydrogen to carbon (H/C) and have historically been unacceptably more difficult to liquefy by DCL than higher quality coals that have a high vitrinite content.
In 1995, Okada (4) reported that oil yield from autoclave experiments were inversely proportional to the inertinite content of coal (
In 2001, Wasaka (5) (NEDOL) published results on 53 runs on 27 coals that were made in a 0.1 t/d pilot plant test program. The program specifically focused on identifying the preferred conditions for liquefying Chinese coals, including high inertinite coals. The study concluded that a combination of iron catalyst and an externally generated donor solvent produced the highest oil yield, particularly for a high inertinite content coal (
Based on these experiments, Wasaka concluded that obtaining higher oil yields was difficult in liquefaction of high inertinite content coals but that an iron catalyst, a donor solvent, temperatures up to 460° C., and a high recycle gas rate were preferred for maximizing oil yield. Ishibashi (6,7), et. al., published hydrodynamic information on a series of slurry liquefaction reactors at varying operating conditions. All of the operations utilized an externally hydrotreated donor solvent. Gas hold-up in the slurry reactors increased linearly with superficial velocity to about 8 cm/sec and then leveled out at approximately 70 percent. This line out is typically associated with a reactor that moves from bubbly flow to churned turbulent operation. This high gas hold-up and corresponding lower solid and liquid hold-up results in an increase in liquefaction reactor volume required to achieve a given coal residence time and coal conversion to liquid products.
In 2005, Shenhua (7,8) applied for a patent (issued in 2010) for a DCL process that utilized iron catalyst and an externally hydrotreated donor solvent. The donor solvent is produced in a suspended bed with a forced circulation reactor (ebullated bed). Suspended bed reactors with forced circulation were selected for liquefaction to provide lower gas hold-up (high utilization of reactor volume) and to avoid precipitation of mineral salts. Two suspended beds reactors with forced circulation are utilized for liquefaction with a treat gas rate (0.6-1.0 Nm3/kg slurry). The preferred embodiment operated both liquefaction reactors at 455° C., a reactor pressure of 19 MPa, a solvent to coal (S/C) ratio of 1.2, and an iron catalyst addition rate of 1 wt % Fe on dry coal. Oil yield for a low ranked bituminous coal is 57.17 wt % on a moisture and ash free (MAF) basis. In a separate publication, Shenhua reported Thermal Efficiency for DCL as 59.75%.
The Headwaters DCL Process(9,10) features include use of disperse, iron based catalyst, two-stage, back mixed, slurry phase reactors, mild hydrogenation to replenish the hydrogen content of the recycle solvent and to stabilize the raw distillate products, and deashed resid conversion to enhance net coal conversion. Reactor temperature in the liquefaction reactors is 435° C. to 460° C. at 17 Mpa.
In both the Shenhua and Headwaters processes, products would be further upgraded using conventional refining processes.
In accordance with the invention, it has been found that a dramatic increase in liquid product yield and thermal efficiency is achieved when the solvent to coal ratio is increased in a DCL process, particularly in microcatalytic coal liquefaction (MCL) processes that employ very finely divided molybdenum or iron catalysts, preferably a finely divided molybdenum catalyst, and a 600 to 700° F.+ non-donor recycle stream produced in liquefaction. The ratio of such stream to coal at the input to the reactor on a moisture free weight basis is greater than 1.6:1, preferably greater than 1.7:1, more preferably between 1.8:1 and 3.5:1, still more preferably between 2.0:1 and 3.5:1, and most preferably between 2.0:1 and 3.0:1. By “non-donor” is meant that the recycle stream has not been processed in a hydrotreater to partially hydrogenate multi-ring aromatic compounds in the stream to produce compounds that can donate hydrogen during liquefaction.
Surprisingly, increasing the ratio of the recycled 600 to 700° F.+ non-donor stream to coal in the MCL process does not increase the flow rate of recycled stream and fresh coal to liquefaction for a given rate of product generation. Instead, less coal is required and although the recycle stream increases relative to coal, the total feed to liquefaction remains essentially the same. The net impact of higher recycle and lower coal rate is a reduction of energy required in the slurry preheat furnace and the size of the vacuum fractionator. Hence, investment and energy requirements are reduced for the liquefaction section of the MCL plant.
Moreover, the MCL process of the invention has been found to efficiently liquefy even coals having inertinite contents much higher (e.g., more than 25%) than were previously thought to be suitable as feeds for direct coal liquefaction processes.
Because the viscosity and density of the coal slurry in the process of the present invention is much higher than is the case with slurry using a hydrotreated donor solvent, the maintenance of a stable slurry and the settling of ash in the reactor is not a concern. Therefore the process of the invention operates with a much lower gas hold-up than is required in hydrotreated donor solvent DCL systems. Finally, volume occupied by internals in the ebullated bed reactors is eliminated. As a result of the low gas hold-up and the preferred use of slurry reactors in series, the reactor volume can be significantly less than that required for a DCL system having the same output capacity operating with a donor solvent and high gas hold-up. Moreover, the use of a non-donor 600 to 700° F.+ stream eliminates the need for a hydrotreater for hydrotreating the donor solvent, which substantially reduces the complexity and capital cost of the system.
The MCL process of the invention has been found to efficiently liquefy even coals having inertinite contents much higher than were previously thought to be suitable as feeds for direct coal liquefaction processes. Moreover, it has been found that an MCL plant operating in accordance with the process of the invention can produce easily upgraded liquids and operate at a high thermal efficiency, in the range of approximately 65% to over 70%, even when processing high inertinite coal feeds.
The DCL process of the invention achieves a higher conversion of coal to liquid fuels and a higher thermal efficiency than those of other direct coal liquefaction processes, even with coals having high inertinite contents that were previously considered unsuitable as a feed for direct coal liquefaction processes. As illustrated in
Referring now to the embodiment of a DCL system illustrated in
Most of the remaining moisture in the coal is driven off in the mixing tank due to the hot atmospheric fractionator bottoms feeding to the mixing tanks. Residual moisture and any entrained volatiles are condensed out as sour water (not shown in
The coal slurry and hydrogen mixture is fed to the input of the first stage of the series-connected liquefaction reactors 209, 211 and 213 at between 600 to 700° F. (316 to 371° C.) and 2,000 to 3,000 psig (138 to 206 kg/cm2 g). The reactors 209, 211 and 213 are simple up-flow tubular vessels, the total length of the three reactors being 40 to 200 feet. The temperature rises from one reactor stage to the next as a result of the highly exothermic coal liquefaction reactions. In order to maintain the maximum temperature in each stage below about 800 to 900° F. (427 to 482° C.), a portion of the hydrogen based treat gas is preferably injected between reactor stages. The hydrogen partial pressure in each stage is preferably maintained at a minimum of about 1,000 to 2,000 psig (69 to 138 kg/cm2g).
The effluent from the last stage of liquefaction reactor is separated into a gas stream and a liquid/solid stream, and the liquid/solid stream let down in pressure, in the separation and cooling system 215. The gas stream is cooled to condense out the liquid vapors of H2O, naphtha, distillate, and solvent. The remaining gas is then processed to remove H2S and CO2
Most of the processed gas is then sent to a hydrogen recovery system, not shown, for further processing by conventional means to recover the hydrogen contained therein, which is then recycled to be mixed with the coal slurry. The remaining portion of the processed gas is purged to prevent buildup of light ends in the recycle loop. Hydrogen recovered therefrom can be used in the downstream hydro-processing upgrading system.
The depressurized liquid/solid stream and the hydrocarbons condensed during the gas cooling are sent to the atmospheric fractionator 219 where they are separated into light ends, naphtha, distillate, and 600 to 700° F.+ fractions. The light ends are processed to recover hydrogen and C1-C4 hydrocarbons that can be used for fuel gas and other purposes. The naphtha is hydrotreated to saturate olefins and other reactive hydrocarbon compounds. The 160° F.+ fraction of the naphtha can be hydrotreated and catalytically reformed to produce gasoline. The distillate fraction can be hydrotreated to produce products such as diesel and jet fuel.
A portion of the 600 to 700° F.+(316 to 371° C.+) is recycled to the slurry mix tank. The remaining 600 to 700° F.+ fraction produced from the atmospheric fractionator 219 is fed to the vacuum fractionator 221 (also referred to as a vacuum pipe still) wherein it is separated into a 1000° F.− fraction and a 1000° F.+ fraction. The 1000° F.− fraction is added to the 600 to 700° F.+ stream being recycled to the slurry mix tank 203.
In the illustrated embodiment, the 1000° F.+ fraction from the vacuum fractionator 221 is sent to be gasified by the partial oxidation system 223 to generate hydrogen for use in the liquefaction. A portion of the coal from the gas swept mill 201 is also fed to the partial oxidation system 223 to produce additional hydrogen. Alternatively, instead of the partial oxidation system 223, the 1000° F.+ bottoms from the vacuum fractionator 221 may be processed in a Circulating Fluid Bed boiler, a cement plant, or sold as a feed for asphalt paving or electrode manufacture. G.E., Shell, and others offer commercial processes for gasification (partial oxidation) of the 1000° F.+ bottoms and Circulating Fluid Bed boiler manufactures such as Foster-Wheeler and Alstom offer technology for combusting the 1000° F.+ bottoms.
Hydrogen for liquefaction and upgrading can also be produced by Steam Methane Reforming of a stream such as natural gas, shale gas, or coal mine methane. This technology is utilized worldwide in refineries and offered by many commercial vendors such as Haldor-Topsoe.
Catalysts useful in DCL processes also include those disclosed in U.S. Pat. Nos. 4,077,867, 4,196,072 and 4,561,964, the disclosures of which are hereby incorporated by reference in their entirety.
Other DCL reactor systems suitable for use in the process of the invention are disclosed in U.S. Pat. Nos. 4,485,008, 4,637,870, 5,200,063, 5,338,441, and 5,389,230, and U.S. patent application Ser. No. 13/657,087, the disclosures of which are hereby incorporated by reference in their entirety.
The preferred DCL Process combines several elements that contribute to maximum Premium Fuels Product production and maximum thermal efficiency. These include, very importantly, the recycle of a non-donor 600 to 700° F.+ stream, preferably including atmospheric fractionator bottoms, to maintain a ratio of the recycle stream to coal at the input to the reactors 209, 211, 213 that is greater than 1.6:1 on a moisture free weight basis, preferably greater than 1.7:1, more preferably between 1.8:1 and 3.5:1, still more preferably between 2.0:1 and 3.5:1, and most preferably between 2.0:1 and 3.0:1; the use of a microcatalyst in the form of finely divided molybdenum; and the use of a much lower treat gas rate than in previous systems. Also, use of bottoms recycle, and multiple slurry reactors in series contribute to the benefits of the process.
An exemplary process for upgrading the liquid product of the DCL reactors 209, 211, 213 is disclosed in U.S. Pat. No. 5,198,099, the disclosure of which is hereby incorporated by reference in its entirety. Other processes and systems suitable for upgrading the liquid products are commercially available from vendors such as UOP, Axens, Criterion and others.
The diesel product after upgrading will have a Cetane number of between approximately 42 and 47, depending upon cut points of the product and aromatics content. Specific gravity of the product will also vary between 0.83 and 0.90. A higher Cetane Number is required for Euro 4 diesel, thus a Fischer-Tropsch facility producing a 70-75 Cetane Number diesel blend stock may be added to the plant operating in accordance with the present invention. The gasoline produced by upgrading the relevant portion of product of the process of the present invention will meet all current gasoline specifications, or can be upgraded to a Research Octane of 106 if desired. This will permit the blending of the low octane naphtha into the gasoline pool while maintaining adequate octane for the blended fuel.
The upgrading process can also be operated to maximize the production of jet fuel or gasoline. The jet fuel produced will meet all Military JP-8 specifications.
The process of the invention was run in the same 0.1 t/d pilot plant used in the runs described by Wasaka. Modifications were made to the pilot plant to the slurry mix tank, pumps, slurry reactors, microcatalyst (moly) addition, and 600 to 700° F.+ recycle. Compared to Wasaka's pilot plant operations, slurry feed rate was increased to 10 to 12 kg/hour and gas rate was reduced by a factor of 2 to 4. During the pilot plant operation, the reactors did not experience solids buildup or reactor deposits. The coal used in this example is shown as the square symbol in the chart of
Oil yield at a 1.5/1 ratio of non-donor 650° F.+(343° C.) recycle stream to coal was 44.6 wt % (MAF basis) When the 650° F.+ stream rate was increased to 3/1 at a low H2 feed rate of 0.73 Nm3/kg of slurry, the C5+/371° C. (C5/650° F.) yield increased to 66.4 wt %. This is substantially higher than observed with an iron catalyst, donor solvent, and high recycle gas rate in a back mixed reactor. The oil yield is also significantly higher than reported by others (U.S. Pat. No. 7,763,167 B2.)
The net product from the high 650° F.+ recycle to coal run was essentially free of 650° F.+. The lower boiling product was upgraded to finished jet fuel and diesel in a conventional fixed bed upgrading unit and so avoiding the need for upgrading in an ebullated bed reactor in both a solvent hydrotreater and a conventional upgrader. This not only eliminates the solvent hydrotreater but also the need to recycle treat gas to this unit.
The coal used in the two runs described above was a high inertinite content coal from the Ordos Basin with an H/C of 0.72.
The following table presents operating conditions and oil yields for high inertinite coals from the Wasaka publication and
Even at a H2 gas rate of 24.8 Nm3/hour, the oil yield was only approximately 60 wt % (MAF basis) for the data presented in FIG. 8 of the Wasaka article for a coal with an inertinite content of 46.1 vol %. Also included is the highest data point for the 0.72 H/C Shangwan coal plotted in
It should also be noted that liquefaction yield in the Wasaka report is given only as oil yield. No information is provided on whether the oil yield is by distillation and/or extraction and the boiling range of the oil yield.
In comparison, operating conditions and oil yields are reported for the 29.9% inertinite content coal described in the Example above.
At a 343° C.+(650° F.+) fraction to coal ratio of 1.5, oil yield for the above data is comparable to the first column of the NEDOL data. It should be noted that donor solvent and iron catalyst have been replaced with 300 wppm of Microcatalyst. In addition, the gas rate to liquefaction has been reduced by almost a factor of 2.5. This excludes the additional H2 that would be required for the solvent hydrotreater in the NEDOL case.
The surprising result is presented in column 2 of the above results (operating in accordance with the process of the invention) versus the second and third columns of the NEDOL data. Increasing the 650° F.+ solvent fraction to coal ratio from 1.5/1 to 3.0/1 resulted in an increase in C5/343° C. (C5/650° F.) yield from 43.9 to 66.4 wt %. This is higher than either result reported in the Wasaka article and, as was reported earlier, the 66.4 wt % oil yield from liquefaction is for a product boiling between C5 and 650° F., whereas the makeup of the yield in Wasaka is not specified. More importantly, the H2 rate to liquefaction is only about 29% of the H2 rate required in the data presented by Wasaka in their
Hence, the recycle of solvent and hot bottoms from fractionation directly back to the slurry mix tank results in the highest recorded oil yield for a high inertinite coal while avoiding the need for compression, heating, and cooling of over three times the H2 rate required in the Wasaka operation.
The following cases were developed to demonstrate the impact of a high S/C ratio on the equipment in the liquefaction section of a commercial scale MCL DCL plant. The particular cases were developed for a 71 KB/D plant with 650° F.+ fraction to coal ratio of 1.5/1 and 2.5/1. Coal and hydrogen requirement for liquefaction were estimated based on the reported pilot plant data.
It is interesting to note that the total quantity of feed from the slurry mix tank is constant for the two cases because of the decreased amount of feed coal to liquefaction for the higher 650° F.+/Coal case.
Since the product and selectivity are similar, the hydrogen consumption and associated heat release are essentially the same.
Temperature rise in the liquefaction reactors is also equivalent in the two cases sense mass flow and heat release are approximately constant in the liquefaction reactors. Because of the higher boiling point of the 650° F.+ stream it remains liquid in the effluent from the liquefaction reactors.
After product separation, the liquid stream from the product separators flows to the Atmospheric Fractionator. 650° F.− product is distilled overhead and part of the 650° F.+ product is recycled to the slurry mix tank.
Recycle from APS to slurry mix tank, KT/D
Part of the 650° F.+ stream is sent to the Vacuum Fractionator where the 650/1000° F. solvent fraction is recovered and recycled to the solvent mix tank and the 1000° F.+ fraction leaves the liquefaction section. As mentioned previously, this stream can be used in partial oxidation (PDX), in a Circulating Fluid Bed boiler, production of cement, etc.
The fraction that is recycled from the Vacuum Fractionator is shown in the following table.
Recycle from VPS to slurry mix tank, KT/D
For this example, the temperature of the slurry mix increases for the high 650° F.+/Coal case and thus reduces the required duty for the slurry preheat furnace.
Thus the overall impact on equipment is shown in the following table.
Thus, the cost of equipment required for the high 650° F.+/Coal case is actually less than in the lower 650°+/Coal case. Thus investment should be lower for the high 650° F.+/Coal option.
As indicated in the table below, the calculated thermal efficiency of the MCL liquefaction process for the two cases is substantially higher than has been achievable in prior coal liquefaction processes.
The overall thermal efficiency for the a balanced MCL plant including hydrogen generation, coal liquefaction, upgrading, and 1000° F.+ bottoms disposal was calculated for plant's operating in accordance with the process of the invention with and without natural gas import. Coal conversion was reduced to 83% to maintain ash content to PDX of 25 wt %.
As shown, thermal efficiency for a balanced, microcatalytic direct coal liquefaction plant (no import or export of power or gas) producing finished gasoline, jet, or diesel from coal only is 66.7%. If natural gas (or equivalent) is available, the thermal efficiency increases to 73.6% This is a direct result of a higher solvent recycle, a major reduction in recycle treat gas rate, the elimination of the solvent hydrotreater, the decrease in the size of the slurry preheat furnace, and the release of the heat of reaction in the liquefaction reactors rather than in the solvent hydrotreater. This thermal efficiency is higher than any known thermal efficiency for a DCL plant operating on a high inertinite content coal such as present in the Ordos Basin.