Ioanna
Dimitriou
*a,
Pelayo
García-Gutiérrez
a,
Rachael H.
Elder
a,
Rosa M.
Cuéllar-Franca
b,
Adisa
Azapagic
b and
Ray W. K.
Allen
a
aUK Centre for Carbon Dioxide Utilization, Department of Chemical and Biological Engineering, University of Sheffield, Mappin Street, Sheffield, S1 3JD, UK. E-mail: i.dimitriou@sheffield.ac.uk; Fax: +44 (0)114 2227501; Tel: +44 (0)114 2227594
bSchool of Chemical Engineering and Analytical Science, The University of Manchester, The Mill, Room C16, Sackville Street, Manchester M13 9PL, UK
First published on 12th May 2015
Utilising CO2 as a feedstock for chemicals and fuels could help mitigate climate change and reduce dependence on fossil fuels. For this reason, there is an increasing world-wide interest in carbon capture and utilisation (CCU). As part of a broader project to identify key technical advances required for sustainable CCU, this work considers different process designs, each at a high level of technology readiness and suitable for large-scale conversion of CO2 into liquid hydrocarbon fuels, using biogas from sewage sludge as a source of CO2. The main objective of the paper is to estimate fuel production yields and costs of different CCU process configurations in order to establish whether the production of hydrocarbon fuels from commercially proven technologies is economically viable. Four process concepts are examined, developed and modelled using the process simulation software Aspen Plus® to determine raw materials, energy and utility requirements. Three design cases are based on typical biogas applications: (1) biogas upgrading using a monoethanolamine (MEA) unit to remove CO2, (2) combustion of raw biogas in a combined heat and power (CHP) plant and (3) combustion of upgraded biogas in a CHP plant which represents a combination of the first two options. The fourth case examines a post-combustion CO2 capture and utilisation system where the CO2 removal unit is placed right after the CHP plant to remove the excess air with the aim of improving the energy efficiency of the plant. All four concepts include conversion of CO2 to CO via a reverse water-gas-shift reaction process and subsequent conversion to diesel and gasoline via Fischer–Tropsch synthesis. The studied CCU options are compared in terms of liquid fuel yields, energy requirements, energy efficiencies, capital investment and production costs. The overall plant energy efficiency and production costs range from 12–17% and £15.8–29.6 per litre of liquid fuels, respectively. A sensitivity analysis is also carried out to examine the effect of different economic and technical parameters on the production costs of liquid fuels. The results indicate that the production of liquid hydrocarbon fuels using the existing CCU technology is not economically feasible mainly because of the low CO2 separation and conversion efficiencies as well as the high energy requirements. Therefore, future research in this area should aim at developing novel CCU technologies which should primarily focus on optimising the CO2 conversion rate and minimising the energy consumption of the plant.
Broader contextCarbon capture and utilisation (CCU) has recently become the focus of large scale international attention not only because it has the potential to reduce anthropogenic CO2 emissions which contribute to climate change but also because it could generate value from waste CO2 through the synthesis of fuels and chemicals. Globally, consumption of fuels is two orders of magnitude higher than that of chemicals; therefore, CO2 utilisation technologies should focus primarily on fuel synthesis to create significant economic value and to make a substantial contribution to the reduction of CO2 emissions. Successful market-entry of CO2-to-fuels technologies strongly depends on their economic competitiveness. In this article, a techno-economic assessment of the manufacture of transport hydrocarbon fuels from waste CO2 is performed through process simulation, cost modelling and sensitivity analysis. Unlike other studies, the present techno-economic assessment only employs the best currently available and proven CCU technologies. The aim is to support policy makers and businesses in their decision-making by establishing whether the production of liquid transport fuels from CO2 using current technology is economically feasible and identifying the modifications required to improve the economic competitiveness of CCU processes. |
Global CO2 emissions from fossil fuel combustion were approximately 31 Gt in 20119 and are likely to rise to 57 Gt in 2050.10 Currently, CO2 is only used for the production of chemicals, such as urea, salicylic acid and polycarbonates. However, in order to make a significant contribution to reducing CO2 emissions, its utilisation should focus primarily on the conversion to fuels since the market for chemicals is two orders of magnitude lower than that for fuels.5 Oxygenates and hydrocarbons can be produced via hydrogenation of CO2 and could offer feasible alternatives for the transportation sector, reducing its dependency on fossil fuels. CO2 hydrogenation for oxygenate production is at present the most intensively investigated area of CO2 utilisation with methanol synthesis from CO2 and H2 already being demonstrated at bench- and pilot-scale plants in Asia11,12 and Europe.13
Conversely, the production of hydrocarbon fuels from syngas (H2 and CO) produced from CO2 and H2 (e.g. via reverse water-gas-shift reaction) is yet to be demonstrated. This is mainly due to the fact that the production of hydrocarbons from CO2 and H2 requires a higher amount of hydrogen and energy than oxygenates.5 However, there is a noticeable lack of published techno-economic feasibility studies in this area that could potentially support this argument. It should also be noted that hydrocarbons produced from syngas via Fischer–Tropsch synthesis (the established industrial process for converting syngas to liquid fuels) are specifically attractive because of their unlimited compatibility with conventional fuels in any proportion and thus, unlike alcohols and ethers, can readily be incorporated and integrated with conventional markets and supply chains.
As highlighted above, the high hydrogen and energy requirements associated with CO2-to-fuels pathways is one of the main issues for the application of these technologies. In order to make such a process economically and environmentally sustainable, hydrogen should be made from a non-fossil resource or produced within the process itself. The latter would also decrease the overall operating costs, especially since fossil-derived hydrogen is still significantly cheaper than that produced from renewable technologies.14 One non-fossil source of hydrogen could be biogas produced by anaerobic digestion of wet waste, such as sewage sludge. Such biogas contains mainly CO2 and methane, the latter of which can be utilised to produce hydrogen (e.g. via steam reforming),15 thus avoiding the depletion of natural gas, currently used for hydrogen production. The CO2 from the biogas can be separated and utilised for the production of fuels. Within the EU alone, around 10 million tonnes (dry basis) of sewage sludge are generated per year and the digestion of each tonne could produce approximately 590 m3 of methane.16 Other wastes could result in even higher methane yields; for example, food waste can generate up to 3.5 times more methane per tonne than sewage sludge.17 Therefore, anaerobic digestion of waste could be an important source of hydrogen for a CO2 hydrogenation-to-fuels process. It could also be a suitable target process for CO2 utilisation technologies since it requires moderate capital investment.
This is the topic of this paper which examines different CO2 capture and utilisation process concepts for the conversion of sewage sludge to liquid hydrocarbon fuels. The aim of the study is to identify the most promising process configurations in terms of conversion efficiencies and costs. For these purposes, a comprehensive techno-economic assessment has been carried out to examine the technical and economic performance of four conceptual designs, considering only the best available and proven technologies: amine CO2 capture, steam reforming, reverse water-gas-shift (RWGS) process and Fischer–Tropsch synthesis.
An overview of the CCU process concepts developed in this work is presented in the next section. Section 3 outlines the methodology for process modelling and the economic assessment followed by the results in Section 4. A sensitivity analysis is carried out in Section 5 which examines the effect of key economic and technical parameters on the production costs of liquid fuels, including capital and energy costs as well as the CO2 conversion rate in the RWGS reactor.
Fig. 1 An overview of the CO2 utilisation system for production of synthetic fuels [MEA: monoethanolamine. Dashed lines represent steps which are not present in all design options; see Table 1 for details. In some design cases the heat and electricity generation plant is placed before CO2 capture]. |
Process sections | PD-MEA | PD-CHP1 | PD-CHP2 | PD-CHP3 |
---|---|---|---|---|
a Monoethanolamine (MEA) CO2 capture plant. b Combined heat and power (CHP). c Methane steam reforming. d Reverse water-gas-shift reaction. e Fischer–Tropsch synthesis. | ||||
Anaerobic digestion | ✓ | ✓ | ✓ | ✓ |
CO2 capturea | ✓ | ✓ | ✓ | |
Heat and power generationb | ✓ | ✓ | ✓ | |
Syngas productionc | ✓ | |||
CO2 conversion | ||||
RWGSd | ✓ | ✓ | ✓ | ✓ |
Hydrogen recovery | ✓ | ✓ | ✓ | ✓ |
Fuel synthesise | ✓ | ✓ | ✓ | ✓ |
Fig. 2 Process flow diagram for PD-MEA [the flue gas from the off-gas combustor in all design configurations contains N2 (68–88 vol%), CO2 (8–20 vol%), argon and oxygen]. |
In the packed absorption column, the biogas is fed counter-currently with an MEA aqueous solution (usually 15–35 wt%) which reacts with and absorbs CO2 in the biogas to form an MEA carbamate soluble salt. The gas stream lean in CO2 is released from the top of the absorber while the MEA solution rich in CO2 is pumped to a heat exchanger in which the solution is heated to about 120 °C and then fed into the stripping column. MEA is regenerated in the stripper and recycled to the absorber for re-use (lean MEA solution, Fig. 6). The regeneration conditions are maintained by the reboiler which uses low-pressure steam. Steam which acts as stripping gas in the column, is recovered in the condenser and fed back to the stripper, while the concentrated CO2 stream is released from the top of the stripper for downstream processing. Process conditions in the capture plant are summarised in Table 2.
Parameters | Value |
---|---|
CO2 removal efficiency (%) | 70 |
MEA solution (wt%) | 30 |
Absorber pressure (bar) | 1.013 |
Stripper pressure (bar) | 2.1 |
CH4 + H2O ⇄ CO + 3H2 | (1) |
CO2 + H2 ⇄ CO + H2O | (2) |
nCO + (2n + 1)H2O ⇄ CnH2n+2 + nH2O | (3) |
In this study, the FT reactor is operated at 30 bar and 220 °C and is assumed to be similar to the Sasol Slurry phase distillate reactor which is designed for LTFT synthesis.26,28 The reactor's gas effluent is passed to a three phase separator to remove water and heavy hydrocarbons from the residual vapour. The FT off-gas which mainly consists of light hydrocarbons (C1–C4) and unconverted syngas is combusted to generate low pressure steam for the anaerobic digesters, whereas the liquid fuels are sent to a central refinery plant for further upgrading.
As mentioned earlier, a wide range of products are obtained from the FT synthesis, therefore a quantitative approximation of product distribution is necessary. The most widely used approach to tackle this problem is the Anderson–Schulz–Flory (ASF) product distribution.31 According to this method, the adsorbed carbon chain can either undergo further addition of a –CH2– group or the chain can terminate.31 The ASF-product distribution model is represented by the following equation:
Cn = αn−1(1 − α) | (4) |
FT synthesis results in the production of various products, thus it is not a highly selective process. However, it offers the possibility to cover the entire range of petrochemical products so that gasoline, jet fuel and diesel can be produced with adequate process control. FT products are high quality and ultra clean fuels, free of sulphur and aromatic compounds and, unlike other fuels such as dimethyl ether and alcohols, they can be easily assimilated in the existing transport infrastructure, concerning both vehicle engines and distribution channels.
Various thermodynamic methods have been used to model the different unit operations considered in this study. The Aspen Physical Property System guide33 was used to ensure that the property methods are tailored to the different classes of compounds and operating conditions. The property method used for most unit operations is the Peng–Robinson with Boston–Mathias modifications (PR–BM) which is recommended for gas processing and refinery applications and provides accurate results for hydrocarbon mixtures and light gases, such as H2 and CO2.33 The non-random-two-liquid (NRTL) method with the Redlich–Kwong (RK) equation of state is used to simulate the anaerobic digestion process. The MEA gas treating unit is modelled using the electrolyte-NRTL based property method ENRTL-RK which is suitable for mixed electrolyte systems up to medium pressures. This method uses the RK equation of state for estimating the vapour phase properties.
The anaerobic digester is modelled using the Aspen Plus yield reactor block (RYield). The mass yields of CH4, CO2 and digestate were calculated separately considering a biogas production of 0.6 m3 per kg of volatile solids loading.19,34 It was assumed that neither NH3 nor H2S are present in biogas since the former is not produced when sewage sludge is used as feedstock and the latter is present in very low concentrations.35Table 3 shows the component mass yields calculated for the anaerobic digestion process.
Component | Mass yield (kg kg−1 sludge) |
---|---|
Digestate | 0.9792 |
CO2 | 0.0124 |
CH4 | 0.0084 |
In the CO2 capture plant, the absorber and stripper columns are simulated using the RadFrac block which is suitable for modelling all types of multistage vapour–liquid fractionation operations. As mentioned previously, the thermodynamic and physical properties are estimated using the ENRTL-RK method coupled with an electrolyte calculation option which models the electrolyte solution chemistry and consists of five equilibrium reactions.36 Design specifications are used to obtain the desired molar split fractions in both the absorber and the stripper. In the absorber, a design specification measured the CO2 flow rate in the stack stream and adjusted the lean MEA flow rate to ensure that a target recovery of 70% is achieved. In the stripper, a design specification measured the CO2 molar concentration in the CO2 product stream and adjusted the reflux ratio to achieve a 98 vol% purity target.21 The number of minimum equilibrium stages was 5 for the absorber and 10 for the stripper.
In the CHP unit, compressed biogas is mixed with steam and air (10% excess) and fed into the combustor. The combustor is simulated using a Gibbs reactor block (RGibbs) which models single-phase chemical equilibrium by minimizing the Gibbs free energy, subject to atom balance constraints. Steam is generated inside a network of heat exchangers while electricity is produced in steam and gas turbines by assuming common isentropic and mechanical efficiencies.37 The CHP plant simulation was based on a natural gas CHP model previously developed by AspenTech.38 The main modification made to this model was to replace the natural gas feed stream with the biogas outlet stream of the anaerobic digestion plant.
Both the steam reformer and the RWGS reactor are modelled using a stoichiometric reactor block (RStoic). Reaction stoichiometry was specified for each of them as well as the fractional conversion of relevant components (80% for CH4 and 65% for CO2, respectively).12,39
FT synthesis is modelled using a yield reactor block (RYield). The mass yields of the produced hydrocarbons were calculated in a separate spreadsheet using the ASF distribution model described in Section 2.6 with a chain growth probability of 0.85 which favours the production of middle distillates. The single-pass CO conversion was set to 80%.40,41 Even though such a CO conversion value is relatively high, it can be achieved in slurry phase reactors employing cobalt catalysts.26,42 The ASF hydrocarbon distribution was taken up to a carbon number of 100. Production of aromatics, oxygenates and olefins is assumed to be negligible in this study since the presence of these compounds is typically small for LTFT synthesis.25,43
a APEA default values (country base: UK). b Average sale price.47 c The quoted price is for hydrogen produced from water electrolysis which is a low-carbon technology.14 d Assumed. e Average sale price for molecular sieve 13X.48 f Electricity price for small UK industrial consumer.49 g Natural gas price for small UK industrial consumer.49 | |
---|---|
General economic parameters | |
Base year | 2013 |
Plant life | 20 years |
Plant annual operating hours | 8000 |
Loan interest rate | 10% |
Tax | 40%a |
Contingency | 18%a |
Working capital | 5% of TCIa |
Raw material prices | |
MEA | £1 per kgb |
Hydrogen | £4 per kgc |
Fischer–Tropsch catalyst | £22 per kg45 |
Reformer catalyst | £22 per kg45 |
RWGS catalyst | £22 per kgd |
PSA packing | £0.85 per kge |
Utility prices | |
Electricity | £0.1127 per kW hf |
Natural gas | £0.0319 per kW hg |
Cooling water | £0.21 per m346 |
The capital investment is comprised of installed equipment costs, indirect costs (e.g. contingency), tax and working capital.44 The capital investment required to establish the project is considered to be borrowed and repaid over the lifetime of the project (20 years) at a loan interest rate of 10% per annum. To estimate the fuel production costs of the CCU system, the annual amount required to pay back the loan on capital needs to be determined first:
(5) |
The total annual costs consist of capital annuities as well as operating costs: raw material, utilities, labour and maintenance costs. The fuel production costs are calculated by dividing the total annual costs by the amount of FT fuels (gasoline and diesel) produced in a year. The price inflation of equipment and raw materials is not considered for the ease of comparison between the evaluated CCU concepts. For the same reason, government subsidies, CO2 credits and by-product revenues are excluded from the economic analysis.
PD-MEA | PD-CHP1 | PD-CHP2 | PD-CHP3 | |
---|---|---|---|---|
a Values are given on a dry ash-free basis. b Total plant electricity requirements. c Electricity generation within the plant. | ||||
Plant inputs | ||||
Sewage sludge (kg h−1)a | 333.3 | 333.3 | 333.3 | 333.3 |
Sewage sludge (kW) (LHV)a | 1944.4 | 1944.4 | 1944.4 | 1944.4 |
Hydrogen (kg h−1) | — | 32.8 | 32.6 | 22.9 |
Hydrogen (kW) (LHV) | — | 1093 | 1086.3 | 763.9 |
Natural gas (kg h−1) | 50.7 | 33.7 | 45.3 | 75.7 |
Natural gas (kW) (LHV) | 688.1 | 456.6 | 572.5 | 655.1 |
Electricity (kW)b | 207 | 577 | 578 | 133 |
Plant outputs | ||||
Hydrogen (kg h−1) | 0.82 | — | — | — |
Hydrogen (kW) (LHV) | 27.3 | — | — | — |
Gasoline (kg h−1) | 12 | 4.6 | 4.5 | 7.9 |
Gasoline (kW) (LHV) | 143.8 | 54.3 | 53.9 | 95.4 |
Diesel (kg h−1) | 31 | 29.3 | 29.2 | 20.9 |
Diesel (kW) (LHV) | 370 | 350.6 | 349 | 249.1 |
Electricity (kW)c | 0 | 175 | 202 | 175 |
Efficienciesa | ||||
Fuel energy efficiency (%) | 26.4 | 20.8 | 20.7 | 17.7 |
Plant energy efficiency (%) | 17.1 | 11.7 | 11.9 | 14 |
The PD-MEA concept produces more fuels than the other three cases because of the higher amount of syngas processed in FT synthesis as a result of its upstream methane steam reforming unit which converts methane to syngas instead of simply burning it in a CHP plant. This also significantly affects the hydrogen requirements of the individual concepts. In the case of PD-MEA, the amount of hydrogen produced from the methane steam reformer is higher than the hydrogen requirements of the RWGS reactor; therefore, some hydrogen is produced as by-product (see also Fig. 2). This is not the case for the other three concepts where additional hydrogen is required from an external source. The impact of the hydrogen price on the fuel production costs of PD-CHP1 and PD-CHP2 is examined in Section 5.4.
The plant energy efficiency ηplant takes into account the total energy input (sludge, hydrogen, natural gas and electricity) and total energy output (fuels, hydrogen and electricity) and is calculated according to:
(6) |
As for the fuel efficiency, PD-MEA also has the highest plant energy efficiency, estimated at 17.1%; this is due to the higher fuel output as well as the excess hydrogen production from which the other process designs do not benefit. PD-CHP3 shows the highest efficiency (14%) of all three CHP-based cases despite the fact that the fuel production is approximately 18% lower than that of the other concepts. The primary reason for this is that PD-CHP3 produces more than enough electricity to cover all the power requirements of the plant so there is no need to provide electricity externally. PD-CHP1 and PD-CHP2 achieve similar efficiencies (11.7% and 11.9%, respectively) which suggests that combustion of upgraded biogas in a CHP plant does not significantly benefit the overall CCU plant performance.
PD-MEA | PD-CHP1 | PD-CHP2 | PD-CHP3 | |
---|---|---|---|---|
Heat (kW) | ||||
Anaerobic digestion | 84 | — | — | 110 |
MEA CO2 capture | 85 | — | 78 | 252 |
Steam reforming | 329 | — | — | — |
RWGS | 20 | 456 | 458 | 138 |
Total | 518 | 456 | 536 | 500 |
Power (kW) | ||||
MEA CO2 capture | 0.3 | — | 0.3 | — |
Steam reforming | 77 | — | — | — |
RWGS | 69 | — | — | — |
PSA | 57 | 29 | 29 | — |
FT synthesis | 4 | 373 | 347 | −42 |
Total | 207.3 | 402 | 376.3 | −42 |
Note that the FT synthesis is an exothermic process and thus requires cooling (water) rather than heating to maintain the operating temperature in the reactor. However, it is the most power consuming section for the PD-CHP1 and PD-CHP2 designs due to the compression of a large volume of processed syngas which contains excess air from the CHP plant. This is not an issue for PD-MEA which does not include a CHP unit. Similar is true for PD-CHP3 which employs an MEA unit right after the co-generation plant. Generally, it can be seen that none of the four CCU designs is energy self-sufficient with the exception of PD-CHP3 which produces surplus electricity but still requires additional heat. In this study, natural gas is used to cover the additional heating requirements of the plant, while electricity is bought from the grid, if needed.
The annual operating and maintenance (O&M) costs are shown in Fig. 9. The operating costs include expenditure for materials (e.g., catalysts, hydrogen), utilities (e.g., electricity, natural gas), labour, maintenance and other costs (e.g., overheads, insurance). The O&M costs range from £4.9–6.4 million with PD-MEA resulting in the lowest expenditure owing to the lower equipment and utilities requirements associated with this process configuration. Labour and maintenance costs are the largest contributor to O&M costs and represent 37–45% of the total O&M expenditure. Other costs represent 31–39%, including the expense for MEA and PSA packing. Hydrogen contributes 12–17% to the total operating costs of the CHP-based designs. Catalyst and electricity costs are higher for PD-CHP1 and PD-CHP2 because of the higher volume of processed gas compared to the other two cases. Finally, cooling and heating utilities represent a small fraction of the total operating costs (2–5%).
The production costs per litre of gasoline and diesel are presented in Table 7 for the four CCU configurations, along with the contribution of capital costs (as capital annuity) and O&M expenditure. The calculated production costs do not include tax, duties, producer and retailer profits, marketing expenditure and distribution costs. As can be seen in Table 7, O&M costs are a more important contributor to the production costs than the capital investment as they represent 58–62% of the total production costs. PD-MEA has the lowest production costs at £15.8 per litre because of its lower capital and operating costs as well as higher fuel production compared to the other three cases. The next best option is PD-CHP at £23.2 per litre which has the lowest production costs among the three CHP-based designs. PD-CHP3 is associated with the highest fuel production costs at £29.6 per litre which is approximately 87% higher than for PD-MEA. The main reason for this is that this concept produces a significantly lower amount of liquid fuels than PD-MEA, as discussed in Section 5.1. From these it is clear that the amount of fuel produced (and thus conversion efficiencies) is a very important element of the production costs; thus, its effect is investigated in the sensitivity analysis later in the paper (Section 5.5).
PD-MEA | PD-CHP1 | PD-CHP2 | PD-CHP3 | |
---|---|---|---|---|
Capital (£ per l) | 6.7 | 8.9 | 9.4 | 11.7 |
O&M (£ per l) | 9.1 | 14.3 | 15.2 | 17.9 |
Total (£ per l) | 15.8 | 23.2 | 24.6 | 29.6 |
Only the PD-MEA concept is considered as this process design has the lowest production costs. Twelve plant capacities are evaluated, ranging from 1 tonne (base case) to 1670 tonnes of liquid fuels produced per day. The latter capacity corresponds to the Bintulu gas-to-liquids (GTL) plant in Malaysia, one of the largest FT plants in the world, owned by Shell.51 The six-tenths factor rule52 was applied to estimate the investment costs of the scaled-up CCU plants as follows:
(7) |
Using the above approach, the capital investment for the PD-MEA plant of the largest capacity considered here (1670 tonnes per day) is estimated at £2.6 billion. This is three times higher than the capital investment of the Bintulu plant, which cost $1.3 billion51 or £831 million (2013 exchange rate: £1 = $1.5653). Therefore, it is highly unlikely that the industry would invest in a CCU plant when they could instead build a conventional fuel production plant of the same capacity at a much lower cost, while also reducing financial and other risks by relying on a commercially proven, rather than a new technology.
Fig. 10 shows the effect of scale on the costs of CCU fuels. For the largest plant capacity, the fuel production costs are almost 16 times lower than for the PD-MEA base case (£15.80 vs. £1.00 per litre). However, the effect of economy of scale levels off for capacities above 620 tonnes per day, with much smaller cost reductions thereafter. By comparison, the cost of producing conventional diesel in 2013 was £0.51 per litre and £0.47 per litre for gasoline54 (gate costs, excluding tax, duty, profits, marketing and distribution costs). This is around two times lower than the CCU fuel costs. Therefore, unless significant improvements are achieved in the conversion efficiencies of CCU technologies, along with introduction of government subsidies and incentives, it is highly unlikely that CCU fuels will be able to compete against conventional transport fuels, which despite their contribution to climate change, are still relatively cheap to produce.
The effect of economies of scale on the fuel production costs of the evaluated CCU process designs was also investigated using Aspen Plus and APEA. This also allows for comparisons with the costs estimated with the six-tenths rule above. Two scaled-up models of the PD-MEA concept were considered: a medium scale and a large scale plant at 850 and 1670 tonnes of fuel per day, respectively. Table 8 shows the capital investment and production costs of the two scaled-up designs calculated by APEA. For the medium plant capacity, the production cost drops to £2 per litre which is approximately eight times lower than that of the PD-MEA base case. As expected, the large scale plant with a capacity equal to the Bintulu plant has an even lower production cost at £1.2 per litre which is about 20% higher than the cost estimated with the six-tenths rule. The capital investment of the large scale process design estimated by APEA is £1.34 billion which is almost half of the equivalent six-tenths rule cost (£2.6 billion); however this is still 61% higher than the capital investment of the GTL Bintulu plant. Therefore, using either cost estimating method, it is clear that CCU fuels are currently significantly more expensive than conventional transport fuels.
Process design | Plant capacity (t d−1) | Capital cost (£ billion) | Production cost (£ per l) |
---|---|---|---|
Base case | 1 | 0.03 | 15.8 |
Medium capacity | 850 | 1.28 | 2.0 |
Large capacity | 1670 | 1.34 | 1.2 |
Fig. 11 Sensitivity of fuel production cost to the CO2 removal efficiency of the MEA CO2 capture plant. |
For the other technical and economic parameters, the sensitivity analysis was carried out by changing each parameter in turn by ±30% of its base-case value (see Table 4) with the exception of the plant operating hours which were changed by ±10% since they cannot exceed the maximum hours per year. The results for the four design concepts are shown in Fig. 12. The bars show deviations from the original values of the model parameters with longer bars indicating a higher degree of sensitivity to a particular parameter.
In the case of PD-MEA, production costs are most sensitive to the capital investment costs. If the total capital is decreased by 30%, the fuel production cost drops to £13.81 per litre or 14.6% below the base case cost. However, errors of ±30% for capital investment estimates are typical40,50 and increased accuracy can only be achieved through very detailed and expensive analysis of a real case. Another important factor that increases the inherent uncertainties in projecting CCU capital costs is the different level of development of some of the technologies considered in this study. For example, the RWGS process has only been proven on a small scale and therefore is still under development as opposed to steam reforming which is a mature and well understood technology. The high sensitivity to variations in capital costs also emphasises the importance of economies of scale, which the studied CCU options do not benefit from yet as opposed to conventional fuel production plants.
For all CHP based models, production costs are most sensitive to changes in the CO2 conversion rate in the RWGS reactor. This suggests that the lower the fuel output, the more sensitive production costs are to variations of the CO2 conversion efficiency. The production costs of the CHP based models can be reduced by 24–29% for a CO2 conversion rate of 84.5% (30% higher than the base case). Therefore, improving the performance of the CO2 hydrogenation technology should be an early priority.
The loan interest rate is the second most sensitive parameter for PD-MEA and third for all CHP cases. Production costs can be decreased by 8–9% when the interest rate is reduced from 10% to 7%. Interest rates influence the capital annuities and can be controlled by agreeing fixed rates with the lender throughout the life of the project, significantly reducing the uncertainty associated with this economic parameter. Finally, production costs are less sensitive to the project's lifetime, electricity and hydrogen prices as well as operating hours.
The overall plant energy efficiency and production costs of the evaluated designs range from 12–17% LHV and £15.8–29.6 per litre of produced fuels, respectively. The process configurations which incorporate a CHP plant result in significantly lower efficiencies and higher costs than the process design with MEA CO2 capture and steam reforming. The primary reasons for this are the higher syngas production in the steam reforming process and the high capital and operating costs of CHP. The sensitivity analysis reveals that the fuel production costs are mainly influenced by variations in capital costs, the CO2 removal efficiency of the CO2 capture plant and the rate of CO2 conversion. This emphasises the importance of optimising current CCU technology, as well as the significance of economies of scale which greatly benefit commercial plants. For example, for the best design case, the costs of fuel production for a larger capacity plant (1670 tonnes per day), are £1–1.2 per litre, down from £16 per litre for a plant producing 1 tonne per day. However, this is still twice as high as the cost of conventional transport fuels. Therefore, fuel production with current CCU technologies is not yet economically viable primarily due to the: (i) low CO2 conversion in the RWGS process, (ii) low selectivity of the Fischer–Tropsch synthesis and (iii) relatively low CO2 separation efficiency in the MEA absorber. This highlights the need for new CCU technologies, some of which are currently being developed (e.g. ionic liquids for CO2 capture, co-electrolysis of CO2 and water, dry methane reforming).
Further research will be carried out to complement the analysis presented here, including the assessment of other less developed technologies (e.g. dry methane reforming) and products (e.g. methanol, formic acid), life cycle assessment to examine the environmental impacts of the studied CCU designs and the possibility of additional revenue from sales of by-products and avoided greenhouse gas emissions.
This journal is © The Royal Society of Chemistry 2015 |