Article pubs.acs.org/EF
Integration Options and Economic Analysis of an Integrated Chemical Looping Air Separation Process for Oxy-fuel Combustion Cheng Zhou, Kalpit Shah, Hui Song, Jafar Zanganeh, Elham Doroodchi, and Behdad Moghtaderi* Priority Research Centre for Frontier Energy Technologies & Utilisation, Chemical Engineering, School of Engineering, Faculty of Engineering and Built Environment, The University of Newcastle, Callaghan, New South Wales 2308, Australia ABSTRACT: This paper is concerned about a detailed techno-economic assessment of a hypothetical 500 MWe coal-fired power plant in New South Wales, Australia, for oxy-fuel conversion using integrated chemical looping air separation (ICLAS) technology and cryogenic air separation unit (CASU). The key objectives of this study are to (i) investigate and compare the detailed integration options for oxy-fuel conversion using ICLAS and CASU and (ii) determine the technical merits of the above integration options and the conditions at which the technologies become economically feasible. The study produced scientific evidence that confirms the viability of the CLAS process from both technical and economic points of view under certain conditions. The detailed technical analysis revealed that ICLAS with natural gas integration is energy-efficient compared to CASU running on parasitic load. This is primarily due to the fact that ICLAS needs less auxiliary power compared to CASU. Despite the fact that ICLAS natural gas integration has resulted in higher efficiencies than CASU running on parasitic load, from a series of detailed economic analyses, it was observed that both ICLAS and CASU may not be viable under the present operating and economic conditions. Nevertheless, from sensitivity analysis, it was concluded that ICLAS can become feasible if economic conditions are improved, e.g., a low natural gas market price ($59/MWh), and/or a high carbon tax (>$33/tonne).
1. INTRODUCTION The oxy-fuel combustion is considered to be the most suitable option1,2 among the portfolio of low-emission technologies (LETs), such as pre- and post-combustion capture processes, mainly as a result of its ability to retrofit at a low cost. However, the oxy-fuel combustion process, similar to other LETs, faces many technological challenges. The major bottlenecks are the higher energy penalties associated with air separation unit (ASU) for oxygen production and CO2 processing unit (CPU) for CO2 purification and compression.2,3 Oxygen production has been considered to be one of the most critical bottlenecks.1,4 There are various technologies available for oxygen production. The three main technologies among the suite of conventional and emerging technologies are cryogenic distillation, adsorption, and membrane technology. To date, cryogenic distillation, namely, cryogenic air separation unit (CASU), is the only mature technology available for large-scale oxygen production for oxy-fuel power plant application. CASU works on a cryogenic distillation principle. It compresses the air to its liquefaction stage, followed by the fractional distillation of several components, such as N2, O2, and Ar. CASU can produce liquid or gaseous streams of N2 and O2 as per the specification of the end user. For oxy-fuel combustion, one needs gaseous oxygen, and therefore, compression energy requirement associated with liquefaction of the produce stream can be partially recovered with an advanced version of CASU specially designed for oxy-fuel combustion. Normally, CASU can consume between 10 and 40% of the gross power output. The energy penalty associated with CASU integration to the conventional coal-fired power plant is estimated to be as high as 6−10 percentage points. Also, if the presence of Ar in the oxygen product and lower purity of oxygen of about 95% is © 2015 American Chemical Society
acceptable for oxy-fuel combustion, then the energy requirement for oxygen production with CASU can be further reduced. The literature suggests that, with such advancement in CASU processes, specific energy requirement of CASU can be reduced from 350 to 200 kWh/tonne of O2.3,5 However, even with such reduced energy requirement, the energy penalty for oxy-fuel combustion still remains as high as ∼4−6%, which further questions the feasibility of the oxy-fuel combustion retrofit design. Under such a scenario, oxy-fuel combustion may not be feasible, and therefore, efforts have been made to reduce the energy requirements for the oxygen production by looking at advancement of the CASU process and also finding alternative options for oxygen production. The current study aims to develop a novel oxygen production process named (integrated) chemical looping air separation (CLAS/ICLAS) and compare it to the most mature CASU technology. To avoid repetition, the detailed concept description and operating principles of the CLAS and ICLAS systems can be referred to in our previous publications.1,2,6 A series of experimental and modeling studies around CLAS technology conducted at The University of Newcastle1,2,6−11 showed that CLAS may become a highly cost-effective alternative to CASU for oxygen production in oxy-fuel combustion and could produce oxygen at 40−70% lower cost than the CASU. The possible integration options for CLAS with oxy-firing and their assessment have been presented in our Special Issue: 5th Sino-Australian Symposium on Advanced Coal and Biomass Utilisation Technologies Received: September 25, 2015 Revised: December 6, 2015 Published: December 7, 2015 1741
DOI: 10.1021/acs.energyfuels.5b02209 Energy Fuels 2016, 30, 1741−1755
Article
Energy & Fuels
conducted by our group,2,8−11 which concluded that Cu, Mn, and Co oxides were the most suitable oxygen carriers for the CLAS system. Further to this, detailed kinetic experiments demonstrated that impregnated Cu oxide particles on the SiO2 support, as a result of their higher oxygen transport capacity (OTC) and reactivity, were the best candidate for the CLAS process. Therefore, CuO−SiO2 oxygen carriers have been selected for the techno-economic assessment. Moreover, a reducing environment in CLAS can be provided by introducing inert sweep gases, such as steam- and/or CO2-enriched recycled flue gas. In the current study, application of either steam- or CO2- enriched recycled flue gas in the reduction reactor for oxygen production has been considered. With the above selections, technoeconomic assessments of ASUs retrofitting in a large-scale coal-fired power plant were carried out following the methodology, as depicted in Figure 1. The details of the process flowsheets being studied are
published work.2 It was found in this study that CLAS in isothermal conditions can only provide an oxygen concentration of up to 12−14 vol % in the oxygen-enriched product stream to the furnace, which is significantly lower than the requirement of oxy-fuel combustion (up to 26−28 vol %). Therefore, as an alternative, the temperature swing option was found to be the most appropriate. In this approach, the reduction reactor temperature is kept at 80−100 °C higher than the oxidation reactor to reduce the equilibrium partial pressure of oxygen in the reduction reactor and, thus, enhance the oxygen release and concentration in the product stream. However, the reduction reaction is endothermic, and one needs additional energy to maintain a higher reduction reactor temperature. This has raised some important questions and concerns regarding the heat integration and management of CLAS with oxy-fuel combustion. Certainly, the energy required for the reduction reactor cannot be delivered by the heat released from the oxidation reactor because it operates at a lower temperature than the reduction reactor. Under these conditions, solar thermal energy or natural-gas-firing integration was considered, which may increase the exergy losses and/or total capital investment of the system. Also, there is a potential disadvantage of CASU using lowpurity oxygen (∼95%) in the oxy-fuel operation. In this case, the product gas stream from CASU has about 3% Ar and 2% N2, which will inevitably end up in the CPU. This may increase the energy requirement in the back end of the CPU given that the power demand for compression of CO2 in the CPU depends upon the purity of CO2. The ICLAS process studied in this report does not have this disadvantage because it is expected to produce 99% pure oxygen by shuttling the metal oxide oxygen carriers between the oxidation and reduction reactor and, thus, allowing for complete N2 and Ar separation from the oxygen product. Therefore, application of ICLAS would not create any efficiency drop in the backend of oxy-fuel combustion. For this reason, in this paper, the ICLAS comparison has been made against the CASU producing 99% pure oxygen. For the development of an innovative concept (ICLAS), however, there exist many uncertainties and knowledge gaps, for example, the heat integration, capital investment, auxiliary costs, and scale-up of the CLAS system with oxy-firing. To address these knowledge gaps, several flowsheets were developed in this study to answer the question of heat integration and provide the comparison of ICLAS to CASU system. The key objectives of this study was therefore to (i) carry out a techno-economic assessment of a hypothetical coalfired power station (with specifications similar to Liddell in New South Wales) for oxy-fuel conversion using CLAS and CASU and investigate and compare the detailed integration options for oxy-fuel conversion and (ii) determine the technical merits of the above integration options and the conditions at which the technologies become economically feasible.
Figure 1. Methodology for techno-economic assessment.
provided in the proceeding sections. It should be highlighted that, although a large number of process flowsheets were developed and examined, all possible combinations of interest could not be assessed. The assessment presented here was carried out using a combined Aspen Plus model and spreadsheet-based analysis toolkit developed inhouse specifically for the study. The techno-economic study summarized here was concerned with retrofitting of the hypothetical plant for oxy-firing operation. Several alternative retrofit configurations, differing only in ASU and heat integration but otherwise identical, were considered. The flowsheet options developed and studied in this project were (i) base case, “business as usual” operation scenario for the existing plant (air-fired) with no carbon capture and storage (CCS); (ii) cryogenic case, oxygen-fired retrofit with a cryogenic-type ASU and CCS: (a) conventional CASU system with 99% oxygen purity, (b) conventional CASU system with 95% oxygen purity, (c) advanced CASU system with 99% oxygen purity, and (d) advanced CASU system with 95% oxygen purity; (iii) ICLAS−FG case, oxygen-fired retrofit with an ICLAS (recycled flue gas reduction) and CCS: (a) direct/indirect CH4 firing in a reduction reactor and (b) solar thermal heating for a reduction reactor; and (iv) ICLAS−steam case, oxygen-fired retrofit with an ICLAS (steam reduction) and CCS: (a) solar thermal heating for a reduction reactor. Detailed process simulation models were developed in Aspen Plus, version 7.3. The technical analyses were carried out to investigate the retrofit impacts (including both ASU and CPU systems) on the performances of the plant for different integration options. The key technical performance indicators being evaluated are (i) power requirement for oxygen production, (ii) plant thermal efficiency, (iii) plant thermal efficiency penalty, and (iv) relative efficiency gain. In the economic assessment, net present value (NPV) calculations were performed on the basis of a number of cost data obtained from the literature. The key economic performance indicators being evaluated are (i) power plant retrofit costs (i.e., capital cost), (ii) operating costs, (iii) cost of CO2 avoided, and (iv) NPV.
2. METHODOLOGY 2.1. Overview. The in-depth assessment has been carried out mainly to evaluate the relative techno-economic merits of CLAS-based processes compared to the cryogenic distillation process for their integration with oxy-pulverized fuel (PF) firing. For the current study, it was assumed that space and layout were not restricted during the plant retrofit and CO2 storage can proceed technically and economically. The selection of the oxygen carrier in this study was based on the comprehensive thermodynamic assessments on CLAS technology 1742
DOI: 10.1021/acs.energyfuels.5b02209 Energy Fuels 2016, 30, 1741−1755
Article
Energy & Fuels Furthermore, a sensitivity analysis on several scenarios and factors was also performed by varying the key impact factors over a wide range of reasonable values. 2.2. Integration Options. The process flowsheets developed for CASU and ICLAS integration with oxy-fuel combustion are described here. The flowsheet for a conventional coal-fired power plant (airfiring), although not displayed here, was considered and studied as the base case. 2.2.1. CASU Integration. The schematic diagram of CASU integration with oxy-fuel is shown in Figure 2. The CASU produces
oxy-PF furnace in primary air heater (PAH) and secondary air heater (SAH) heat exchangers. The FG leaving the oxy-PF needs to be cooled to 150 °C before it enters the fabric filter (FF) for particulate removal. After that, it goes to a flue gas desulfurization (FGD) unit for sulfur removal. After the FGD unit, the FG stream will be split to recycled flue gas (RFG) stream and FG emission stream. The FG emission stream further goes to the CPU for compression, purification, and storage, while the RFG is recycled back to oxy-PF. 2.2.2. CLAS Integration. CLAS is a novel process where oxygen is produced in a chemical looping manner by shuttling the metal oxide oxygen carriers between an oxidation and a reduction reactor. CLAS has several technical challenges, and among them, heat integration is the most critical challenge. As noted earlier, in the CLAS process, a temperature difference must be maintained between the oxidation and reduction reactors because isothermal operation only lead to low oxygen production levels. One way of achieving the temperature difference is to operate the oxidation reactor at lower temperatures than the reduction reactor (∼100 °C). This is mainly to maintain a lower equilibrium partial pressure for oxygen in the oxidation reactor and a higher equilibrium partial pressure for oxygen in the reduction reactor. This will reduce the air and RFG requirements for the oxidation and reduction reactor, respectively, which further reduces the auxiliary power requirements for the respective air and RFG blowers for both reactors. However, this arrangement creates a problem of heat management because practically exothermic heat from the oxidation reactor cannot be used for the reduction reactor as a result of 100 °C negative temperature difference. To solve this dilemma, a new steam cycle is introduced to the oxidation reactor outlet to make use of its excess reaction heat, while the energy requirements for the reduction reaction are sufficed by an external heating/auxiliary fuel, such as methane firing or solar thermal integration. While this arrangement is expected to increase the capital investment, it would also increase the throughput of the power plant. As an alternative to the above configuration, we have also considered the feasibility of keeping the reduction reactor temperature lower than the oxidation reactor temperature. It was found that this arrangement will increase the air requirement and associated auxiliary power requirement of the oxidation reactor air blower. The sweep gas requirement in the reduction reactor will also increase significantly with such an arrangement. In addition, the reduction reactor would then produce a very dilute oxygen concentration (of a maximum of up to 10−14 vol %) in the oxygen-enriched stream to the boiler. With such a dilute oxygen concentration, the oxy-PF furnace may not achieve complete combustion. To improve the oxygen concentration
Figure 2. Oxy-firing integration with CASU. an oxygen stream with 95 or 99% purity, which is mixed with recycled flue gas and enters the oxy-PF furnace. The recycled flue gas (RFG) coming from the back end is generally split into two streams, namely, (1) primary or dry and (2) secondary or wet. The primary RFG after water removal goes to the coal mill to convey coal from the coal mill to the oxy-PF furnace. The secondary hot stream without water removal is being fed directly to the oxy-PF furnace. Bother primary and secondary streams are being preheated by the flue gas (FG) leaving the
Figure 3. Direct CH4 firing in the reduction reactor for the ICLAS−FG case. 1743
DOI: 10.1021/acs.energyfuels.5b02209 Energy Fuels 2016, 30, 1741−1755
Article
Energy & Fuels
Figure 4. Indirect CH4 firing in a reduction reactor for the ICLAS−FG case.
Figure 5. Solar integration with a reduction reactor for the ICLAS−FG case. in the oxygen product stream, the addition of steam together with flue gas was also investigated. However, it was found that the steam requirement was massive and could not be supported by the plant steam cycle. Given the above analysis and after weighing pros and cons of both options, the team decided to adopt the option of running the reduction reactor at higher temperatures than the oxidation reactor. One should also note that the current study is limited to CuO−SiO2 oxygen carriers. With other oxygen carrier systems, such as Co and Mn oxides, as a result of differences in their equilibrium oxygen partial pressure, ICLAS may require less air or flue gas/steam flow rates. This needs further investigations. Furthermore, the possibility of a gas turbine or combined cycle at the exit stream of the oxidation reactor to reduce the capital investment or improve the overall thermal efficiency of the power plant needs to be scrutinized in detail (note that this assumes that the oxidation reactor operates at high pressures). 2.2.2.1. CLAS Integration (RFG). Two cases have been considered for the recycled flue gas version of ICLAS. One is the direct/indirect
heating by natural gas firing, and the other is solar thermal integration. They are described as follows: 2.2.2.1.1. Direct/Indirect Methane Firing. Figure 3 shows the schematic of direct methane firing in the reduction reactor for the ICLAS−FG case. The RFG from oxy-PF is being preheated with the reduction reactor outlet stream before it enters the reduction reactor. Natural gas is directly fired in the reduction reactor to provide sufficient energy required for the reduction reaction. The oxidation reactor outlet stream is coupled with the steam cycle to produce additional electricity. Direct firing of methane dilutes the concentration of oxygen in the reduction reactor outlet stream up to 14 vol % as a result of H2O formation with natural gas combustion. However, 26 vol % O2 can be produced by condensing water from the oxygen product stream before entering the oxy-PF furnace. A total of 40% of the low-temperature energy extracted during the condensation process is assumed to be recovered for feedwater heating (FWH) and lowtemperature heating or cooling (HOC) applications for the power 1744
DOI: 10.1021/acs.energyfuels.5b02209 Energy Fuels 2016, 30, 1741−1755
Article
Energy & Fuels
Figure 6. Solar integration with a reduction reactor without the use of a new steam cycle for the ICLAS−steam case. (∗) Connections between two steam flows.
Figure 7. Solar integration with a reduction reactor with a new steam cycle for ICLAS−steam (∗) Connections between two steam flows. plant buildings. The remaining 60% of the low-temperature heat is considered as a net thermal loss. Figure 4 presents the flowsheet of the indirect methane-firing case. The major difference to the direct methane-firing case (as shown in Figure 3) is that the natural gas combustion occurred indirectly in the outer tube reactor, and therefore, the water removal process is not needed. The reduction reactor outlet stream containing the desired oxygen concentration is split into two streams: (1) for natural gas oxyfiring in the outer tube reactor and (2) for oxy-PF furnace. The reduction reactor product stream going to the oxy-PF is used for coal combustion. The outlet stream of oxy-PF (i.e., the main flue gas) is 100% recycled to the reduction reactor. The secondary flue gas resulting from natural gas combustion goes to the CPU. The advantage with this option is that the desired oxygen concentration of 26 vol % can be achieved directly at the reduction reactor outlet without the need for a steam condensation and removal process. 2.2.2.1.2. Solar Heating. The option of solar heating in the reduction reactor for the ICLAS−FG case is depicted in Figure 5. In the current study, high-temperature concentrated solar thermal
systems, such as parabolic trough, linear Fresnel, power tower, or parabolic dish, including high-temperature thermal storage, have been considered. However, to approach a reduction reactor temperature of 1040 °C, a solar thermal tower may be the only possible option. The solar heating option has several challenges, and the most important challenges are the capacity factor of solar thermal and hightemperature solar storage along with the high capital cost. If these problems are resolved, the solar heating option could provide better heat integration compared to the natural-gas-firing cases. 2.2.2.2. CLAS Integration (Steam). CLAS integration with oxy-PF using steam as the sweep gas has also been investigated. The advantage of steam integration is that it can produce high-purity oxygen stream of up to 99% at the outlet of the reduction reactor after steam condensation and water removal process. Two cases of solar heating integration with/without a new steam cycle have been considered in the current study (when no new steam cycle is used, the excess heat produced in the oxidation reactor will be used for feedwater preheating in the existing steam power cycle, which can also lead to increased power production), whereas a major difficulty exists with the use of 1745
DOI: 10.1021/acs.energyfuels.5b02209 Energy Fuels 2016, 30, 1741−1755
Article
Energy & Fuels Table 1. Process Simulation Cases case case case case case case case case case case
0 1 2 3 4 5 6 7 8
description hypothetical hypothetical hypothetical hypothetical hypothetical hypothetical hypothetical hypothetical hypothetical
power power power power power power power power power
station station station station station station station station station
(550 MWe), conventional retrofit to oxy-firing using retrofit to oxy-firing using retrofit to oxy-firing using retrofit to oxy-firing using retrofit to oxy-firing using retrofit to oxy-firing using retrofit to oxy-firing using retrofit to oxy-firing using
steam in the integration with direct natural gas firing. That is, when H2O is present in a large quantity, steam reforming of methane will occur instead of combustion. Indirect natural gas firing is still theoretically possible but was not examined in the current study. 2.2.2.2.1. Solar Heating without a New Steam Cycle. The option of solar heating integration without a new steam cycle is exhibited in Figure 6. It can be seen that the steam produced from the oxidation reactor is directly used in the reduction reactor after heating it to a desired temperature using the reduction reactor outlet stream in HE2. It should be noted that the excess heat generated in the oxidation reactor is high enough for high-temperature steam production, which can be used to both offset the turbine-bled steam, as required in the existing steam power cycle (i.e., via feedwater heating), and to produce and preheat steam for the reduction reactor. 2.2.2.2.2. Solar Heating with a New Steam Cycle. To avoid any significant exergy losses in the system, solar heating integration with a new steam cycle featured by a new steam turbine was considered, the schematic of which is presented in Figure 7. The arrangement is similar to that employed in the above case, i.e., Figure 6. The only difference is that the steam produced using the reaction heat of the oxidation reactor is used for power generation with a new steam turbine. 2.2.3. Process Modeling. Process modeling of the flowsheets described previously have been carried out using Aspen Plus, version 7.3. In total, nine models were generated. Their detailed descriptions are provided in Table 1. Table 2 shows the Liddell power plant data used in the process simulations for air-firing and retrofitting oxy-firing conditions. The process simulation model for oxy-fuel combustion with CASU is shown in Figure 8. The model for a conventional CASU system is shown in Figure 9. The process models for oxy-PF flue gas integration
Figure 8. Process model for oxy-PF combustion with CASU. with CLAS using natural gas firing are presented in Figure 10, with the detailed CLAS model presented in Figure 11.
Figure 9. Process model for conventional CASU.
Table 2. Power Plant Data (Adopted from Data for the Liddell Power Station) power plant performance gross output12 gross thermal efficiency12 reference capacity factor12 auxiliary power12 fuel composition (as-received basis) moisture ash carbon hydrogen nitrogen sulfur oxygen calorific value13 environmental emissions carbon dioxide sulfur
value
Table 3 provides information on the reactor type, process conditions, and function of the different unit operations used in the oxy-fuel process. The modeling of a combustion process was conducted using RYield and RGibbs models as instructed in the Aspen Plus tutorial of solids handling. Because coal is a nonconventional component according to the definition of Aspen Plus, it
unit
518 36 88 25
MW % % MW
8.8 30.3 49.7 3.3 1.1 0.5 6.3 20.90
% % % % % % % MJ/kg
1.822 0.010
kg/kg of fuel kg/kg of fuel
air firing conventional CASU (99% O2 purity) conventional CASU (95% O2 purity) advanced CASU (99% O2 purity) advanced CASU (95% O2 purity) ICLAS−FG−CH4−with a new steam cycle ICLAS−FG−solar−with a new steam cycle ICLAS−steam−solar−without a new steam cycle ICLAS−steam−solar−with a new steam cycle
Figure 10. Process model for oxy-PF with ICLAS. 1746
DOI: 10.1021/acs.energyfuels.5b02209 Energy Fuels 2016, 30, 1741−1755
Article
Energy & Fuels
2.3.1. ICLAS Reactors and Cyclone Cost. The capital cost of ICLAS reactors and cyclones were estimated on the basis of the cost of a large-scale fluid catalytic cracking (FCC) unit reported in the literature.19,20 The principle of economics of scale was applied for accurate estimation of reactors and cyclone costs. Besides, all costs reported in the literature were escalated to 2013 A$ (short for AUD$). The equation for the economics of scale is given as follows:
C 2 = C1(S2/S1)0.6
(1)
where C1 ($A) and S1 (m3) are the capital cost and volume of the FCC plant, respectively, and C2 ($A) and S2 (m3) are the cost and volume of the examined ICLAS plant, respectively. 2.3.2. New Steam Cycle Cost. The cost of NSC was determined by the capital cost per megawatt reported in the literature21,22 multiplied by year correction factor and the installed power capacity of the NSC. 2.3.3. Solar Thermal Heating Plant Cost. The solar thermal plant is needed only to provide heat for the reduction reactor. Therefore, its cost is only a fraction of conventional solar thermal power plants. In this study, the cost of the solar thermal plant was estimated considering a reduced capital cost at 2.3 million A$/MW and a capacity factor at 30%.21,22 2.3.4. Cost for Additional Cooling Tower Capacity. Revamping of the cooling tower is required in the ICLAS system cases with CH4 integration for the condensation and water-removal process as a result of methane combustion. The associated cost of this revamping cooling tower was estimated on the basis of the literature.23 2.3.5. Costs for Heat-Exchanger Banks, Piping, and Other Miscellaneous Items for Retrofit. For simplicity, this item was assumed at a fixed value of 15 million A$ for both ICLAS CH4 and solar cases based on the cost estimate provided in the National Energy Technology Laboratory (NETL) report.23,24 2.3.6. Conventional CASU Cost. The cost of CASU was obtained from the NETL report on oxy-fuel combustion.23 Similar to the ICLAS reactor cost estimate, economics of scale was employed as follows:
Figure 11. Process model for ICLAS.
shall be decomposed into constituent elements by the RYield block before it is sent to the RGibbs block. The downstream unit operations include preheaters, fabric filter, and FGD. For the CLAS process, oxidation and reduction reactors were considered as RGibbs. For the CASU model, Compr and RadFrac have been used. The ideal property method was applied, which uses both Raoult’s law and Henry’s law. The steam cycle was simulated in Aspen HYSYS. The user interface between Aspen Plus and Aspen HYSYS models was created to feed data in the steam cycle. 2.3. Economic Analysis. The cost estimates for CASU and CLAS systems along with any additional required components have been obtained from the literature. The economics of scale equation and standard currency conversion factors were considered and when required to estimate capital investments. Followed by the cost estimate, simple NPV calculations have been performed for cases 5 and 6, and they were compared to conventional CASU cases 3 (99% O2 purity) and 4 (95% O2 purity). Figure 12 illustrates the flowchart of the methodology used in the NPV analysis of different integration options. Table 4 provides the information on the basic cost data used for the current calculations. With regard to this, sensitivity analyses were performed for key economic parameters, including real discount factor, carbon tax, electricity price, and supplementary fuel price. From the capital cost estimate, the process simulation has identified that CLAS process integration requires the following additional components: (i) ICLAS reactors including cyclones, (ii) new steam cycle, (iii) solar thermal heating plant, (iv) additional cooling tower capacity, and (v) additional heat-exchanger banks, piping, and other miscellaneous items for retrofit. For each of the above items, the costs have been estimated as follows.
X 2 = X1(Y2/Y1)0.6
(2)
where X1 ($A) and Y1 (MW) are the cost and capacity of the reference power plant, respectively, and X2 ($A) and Y2 (MW) are the cost and capacity of the examined power plant, respectively. It should be noted that the present economic study only examined the cost of conventional CASU instead of the advanced CASU system considering that the cost of the advance CASU system is rarely available in the literature. However, one should consider that the advanced CASU system may need higher capital investment than the conventional system.
Table 3. Specifications of Different Unit Operation Blocks Used in Aspen Plus Relative Efficiency Gain unit operation
reactor type
decomposition burn combustion steam generator APH1 APH2 FF FGD
RYield RGibbs calculator heat exchanger heat exchanger heat exchanger FabFl RStoic
oxidation reactor reduction reactor
RGibbs RGibbs
compressor distillation columns
Compr RadFrac
condition Oxy-PF P = 1 bar, T = 75 °C P = 1 bar HSOT = 350−600 °C CSOT = 130 °C CSOT = 200 °C eff. = 99.9% P = 1 bar, T = 150 °C, eff. = 94% CLAS P = 1 bar, T = 950 °C P = 1 bar, T = 1040 °C CASU P = 6 and 7.5 bar T = from −177 to −193 °C 1747
function decomposition of coal to conventional components combustion calculate the combustion products and properties to generate steam to preheat the recycle stream to remove solids (fly ash) to remove SOx for oxidation for reduction for air compression for separation of oxygen DOI: 10.1021/acs.energyfuels.5b02209 Energy Fuels 2016, 30, 1741−1755
Article
Energy & Fuels
Figure 12. Flowchart of the methodology used for calculating NPV (O & M costs refer to operating and maintenance costs, and RECs refer to renewable energy certificates). avoidance of fuel costs plus incur costs as a result of the loss of power sales income. (xii) Allowance may be made for the decrease in main plant output and/or increase in the firing rate as a result of the new technology. (xiii) New technology auxiliary electrical energy demand is subtracted from net generation in calculating electricity sales income. In NPV calculations, we also considered the construction expenditure profile and period of O/S for installation for all investment options (Table 5). The cumulative net cash flows were then calculated for both CASU and ICLAS retrofit systems.
Table 4. Standard Economic and Operating Conditions basic cost data
value
unit
primary fuel cost12 electricity wholesale price14 renewable energy certificates15 carbon tax16,17 real discount factor fuel cost supplementary fuel (CH4)18 calorific value of supplementary fuel (CH4) supplementary fuel cost (CH4)
2 50 50 25 5 41.8 4 50.07 100.15
$/GJ $/MWh $/MWh $/tonne % $/tonne $/GJ MJ/kg $/tonne
Table 5. Construction Expenditure Profile and Period of O/ S for Installation construction expenditure profile
The NPV analysis is an advanced tool that enables the visualization of cash flows of varied investment options. The cumulative net cash flow, namely, the NPV at year T, is calculated using eq 3 T
NPV = − E0 +
∑ t=0
(R t − Et ) , (1 + i)t
year year year year
t = 1, ..., T
−3 −2 −1 0
percentage (%) 10 30 50 10
period O/S for installation year year year year
−3 −2 −1 0
weeks 0 6 26 13
(3) It should be noted that the process contingency considered in this study for CASU and ILCAS systems is determined according to their technology status, including mature, demo plant, pilot plant, and lab concept. The process contingency for those technology statuses were taken as 8, 15, 30, and 50%, respectively, on the basis of the NPV model of Allen Lowe.25
where NPV denotes the net present value of future accumulating costs (i.e., accumulative discounted cash flow), i is the discount rate, T is the operational time of the plant (in years), Rt is revenues per year, and Et is expenditures per year (i.e., costs for fixed and varied O & M, feedstock, and auxiliary power). The assumptions made for the current NPV analysis are summarized as follows: (i) The calculation considers only cash flow changes resulting from the installation of the new technology; as such, the standing and financing costs of operating the power plant are not considered. Cash flow changes include (a) allowance for loss in generation as a result of outages to install the equipment and for possible recurrent maintenance or refurbishment activities and (b) savings on fuel while the plant is out of service for technology installation or maintenance. (ii) Constant dollars in the year 2013 are assumed; i.e., no allowance is made for real growth in future costs or earnings. (iii) NPV is calculated on the basis of real discount factors; i.e., no allowance is made for expectations of future inflation. (iv) Base plant capacity factor is controlled by demand rather than plant availability; i.e., any outage to maintain/refurbish the new technology will incur a loss of capacity factor. (v) Carbon dioxide is ascribed a cost; installation of the new technology reduces the emissions and, therefore, provides a net benefit. (vi) Taxation and depreciation are not considered. (vii) The new technology installation starts up at the beginning of year 0. (viii) NPV of all cash flows is summed to the end of year 0. (ix) Project costs may be spread up to 3 years before new technology start data and also in the startup year (year 0). (x) Installation of the new technology will likely require allowance for one or more outages. (xi) Outages yield cost savings as a result of
3. RESULTS AND DISCUSSION Table 6 summarized the detailed technical analysis results for all nine cases being investigated. The impact of ASU retrofit on the base plant was evaluated and expressed in terms of power requirement for oxygen production, overall plant thermal efficiency, total energy penalty, and relative efficiency gain. 3.1. Power Requirement for Oxygen Production. Table 6 reveals that most of the thermal energy inputs from natural gas were converted into extra power in an indirect manner, namely, via the new steam cycle installed after the oxidation reactor. The unconverted part thus forms an energy penalty, called the NG-based penalty. This, in addition to the auxiliary power requirement, becomes the two major penalties for the oxygen production in ICLAS systems integrated with natural gas. Therefore, the power requirement for oxygen production in the ICLAS system can be obtained according to the following equation: EO2 = [Welec + (WNG − WNSC − WFWH)]/mO2 1748
(4)
DOI: 10.1021/acs.energyfuels.5b02209 Energy Fuels 2016, 30, 1741−1755
Article
Energy & Fuels
Table 6. Performance Comparison of the Base Case Scenario with CASU and ICLAS Systems Using Flue Gas as the Sweep Gas parameters
units
base case 0
case HP turbine IP turbine LP turbine power generation of the NSCb additional power recovery
MW MW MW MW
integration with conventional CASU 1 (refa)
air firing
99% O2 purity
144 213 172
144 213 172
integration with advanced CASU
2
3
95% O2 purity
99% O2 purity
integration with ICLAS using flue gas
4 95% O2 purity
Power Production 144 144 144 213 213 213 172 172 172
MW
5
CH4 integration (direct/indirect) 144 213 172 733
integration with ICLAS using steam
6
7
8
solar integration
solar integration (without NSCb)
solar integration (with NSCb)
144 213 172 332
144 249 236
144 213 172 332
86 Auxiliary Power 3.8 3.8 11.3 11.3
mill power consumption boiler feed pump (turbine-driven) condensate pump primary air fans forced draft fans induced draft fans spray dryer FGD miscellaneous
MW MW
3.8 11.3
3.8 11.3
3.8 11.3
3.8 11.3
3.8 11.3
3.8 10.4
3.8 11.3
kW kW kW kW kW MW
251 995 718 1881 2500 15.8
251 497 442 1131 2500 15.8
251 774 689 1483 2500 15.8
309 864 769 1550 2500 15.8
251 774 689 1483 2500 15.8
cooling tower duty cooler before the fabric filter steam condensation process waste heat recovery (40%) for heating/cooling application oxidation reactor reduction reactor thermal efficiency of the NSCb external heat input from solar/CH4 methane flow requirement
MWt MWt
659
251 251 251 251 831 831 831 831 741 741 741 741 1503 1503 1503 1503 2500 2500 2500 2500 15.8 15.8 15.8 15.8 Heating/Cooling Requirements 659 659 659 659 54 54 54 54
659 0
659 0
965 58
659 0
air flow pressure difference over the air and sweep gas blower oxidation and reduction reactor blowers power requirement oxidation reactor operating temperature equilibrium oxygen partial pressure (OXI) reduction reactor operating temperature equilibrium oxygen partial pressure (RED) oxygen carrier flow sweep gas flow sweep gas temperature sweep gas pressure specific power requirement3,5 specific power requirement for CPU15 CO2 for CPU plant
MWt
238
433
MWt
95
173
MWt MWt
1834 0 0.400
831 −803 0.400
0 −891
830 −849 0.400
MWt
2099
803
891
849
2115 50
561 50
835 50
561 50
MW
70
27
25
17
°C
950
950
950
950
%
4
4
4
4
°C
1042
1042
1040
1040
%
27
27
25
25
kg/s kg/s °C bar kWh/tonne of O2
415
330
250
200
2777 269 129 1.25 71
1098 371 147 1.25 69
1099 187 30 1.25 64
1099 187 150 1.25 43
kWh/tonne of CO2
135
CPU Plant 135 135
135
135
135
135
135
kg/s
123
123
123
234
124
124
124
kg/s kg/s kPa
41.9 479
Oxygen Plant 504 479
123
1749
504
DOI: 10.1021/acs.energyfuels.5b02209 Energy Fuels 2016, 30, 1741−1755
Article
Energy & Fuels Table 6. continued parameters
units
base case 0
air firing
case CPU power demandc
MW
coal consumption rate coal heating value gross power production coal plant parasitic loads oxygen requirement oxygen plant power requirement net power production plant thermal efficiency total energy penalty (reference to base case) efficiency gain (reference to case 1) minimum required efficiency of the NSC
kg/s MWt MW MW kg/s MW
70 1457 518 25
MW % %
492 33.8
%
integration with conventional CASU 1 (refa)
99% O2 purity
2
3
95% O2 purity
CPU 54 Overall Plant 70 70 1457 1457 518 518 25 25 110 116 165 138 54
274 18.8 15.0
integration with advanced CASU
99% O2 purity
4 95% O2 purity
Plant 54 54 Performances 70 70 1457 1457 518 518 25 25 110 116 99 84
integration with ICLAS using flue gas 5
CH4 integration (direct/indirect)
integration with ICLAS using steam
6
7
8
solar integration
solar integration (without NSCb)
solar integration (with NSCb)
102
54
54
54
70 1456 1338 24 275 70
70 1456 851 25 108 27
68 1457 618 26 110 25
70 1456 850 25 110 17
301 20.7 13.1
340 23.3 10.5
355 24.4 9.4
1,141 33.0 0.8
744 32.9 0.9
513 24.1 9.7
754 32.7 1.1
9.8
23.9
29.6
75.1
74.9
27.8
73.6
13.3
1.6
%
1.5
a
As mentioned in the Introduction, this reference case was chosen on the basis of the consideration that the ICLAS system can produce 99% purity oxygen, while in the case of 95% O2 purity, the 5% impurities, such as N2 and Ar, will have to be removed to achieve the same purity level, and this could incur additional costs at an increase of about 11%.24 bNSC = new steam cycle. cConsidering a 90% capture level of the plant CO2 emissions.
where EO2 denotes the power requirement for oxygen production (kWh/tonne), Welec denotes the direct electrical power requirement for the ICLAS system (kWe), WNG denotes the potential power production from natural gas (kWe), WNSC denotes the actual power production of the new steam cycle (kWe), WFWH denotes any other forms of power generation attributed by the use of natural gas [e.g., waste heat recovery for feedwater heating (FWH)], (kWe), and mO2 denotes the total oxygen production of the ICLAS system (tonne/h). The power requirements for oxygen production for all eight oxy-PF cases are presented in Figure 13a. It can be seen that most ICLAS with FG and steam cases have up to 67−85% reduction in energy requirement compared to conventional and advanced cryogenic cases. This is mainly due to the fact that, unlike CASU cases, ICLAS does not require energy to compress and liquefy the air. Of the four ICLAS cases studies, ICLAS−steam−solar−with a new steam cycle (i.e., case 8) was found to have the lowest oxygen production cost. This is mainly due to the lowest air, oxygen, and flue gas requirement compared to that of other cases. An exception is case 7, in which the sum of electrical and NG-based penalties reaches an extremely high value. This is because of the huge exergy losses as a direct result of not using a new steam cycle for converting (indirectly) natural gas thermal input into power generation. For a better understanding of ICLAS integrated with natural gas, Figure 13b shows a sample calculation of the oxygen production power requirement for ICLAS−FG−CH4 (i.e., case 5). As seen from Figure 12a, the direct electrical power requirement is 70 MWe for the ICLAS system. Now, if the thermal input from natural gas is considered as a penalty, the power generated from the new steam cycle together with any other forms of power production (e.g., via FWH) owing to the introduction of the external fuel should also be included as a
Figure 13. (a) Power requirement for oxygen production (note that the ICLAS−S case 7 is the steam case without installing a new steam cycle). (b) Sample calculation for ICLAS−FG case 5 (OXI, oxidation reactor; RED, reduction reactor; FWH, feedwater heating).
direct benefit. This results in a net NG-based penalty of 21 MWe in the oxygen production process (note that, in Figure 13b, we have 840 − 733 − 86 = 21) [note that the equivalent electrical energy of the thermal input from natural gas is obtained assuming that the existing ICLAS facility is used to combust natural gas (with air) and generate power]. Such penalty is then translated into the thermal part of the power 1750
DOI: 10.1021/acs.energyfuels.5b02209 Energy Fuels 2016, 30, 1741−1755
Article
Energy & Fuels requirement for oxygen production (see the red bars in Figure 13a). As is evident from Figure 13a, even when the thermal penalties are incorporated into the analysis, the ICLAS−FG options are still outperforming cryogenic-based systems and exhibit energy penalties that are about 21% of the conventional cryogenic systems (i.e., 79% reduction in energy penalty) and 33% of advanced cryogenic systems (i.e., 67% reduction in energy penalty). It should be highlighted though that the gas and solid flow rates of the ICLAS−FG−CH4 integration for a 500 MW retrofit seem to be considerably high (see Table 6). However, using a modular approach, one can achieve the same output using a series of multiple units, and hence, the ducting and reactors will be of reasonable size. The project team has indeed estimated the dimensions of reactor vessels and the size of fluid transport ducting and has found that, for case 5 (ICLAS−FG− CH4), seven ICLAS units each of 10.5 m in diameter (i.e., including both an oxidation reactor of 7 m in diameter and a reduction reactor of 2 m in diameter) and 30 m in height were required for retrofitting a 500 MW coal-fired power plant. It should be noted that these are estimated numbers calculated on the basis of the established scale-up rules for the specific system. 3.2. Plant Thermal Efficiency. The overall plant thermal efficiency was calculated for all nine cases using the net power production divided by the total energy input (see Table 6 for more detailed data). The total energy input included heat from both the primary and supplementary fuels as well as solar heat, while the net power production was calculated by the gross power production from the base plant and new steam cycle minus all auxiliaries. The efficiency results are presented in Figure 14. It can be seen that, for the CASU integration cases,
Figure 15. Total energy penalty.
the elimination of the compression and cooling step in ICLAS cases. However, it should be noted that heat integration in ICLAS is challenging, which needs an additional steam cycle, natural gas firing, or solar thermal integration that will increase the capital cost for ICLAS integration cases. 3.4. Relative Efficiency Gain. Relative efficiency gain, defined as the efficiency rise compared to that of case 1 (i.e., the CASU integration with a 95% O2 purity), was calculated and presented in Figure 16. It can be observed that ∼73−75% of
Figure 16. Relative efficiency gain.
the relative efficiency gain can be obtained for ICLAS cases 5, 6, and 8 compared to conventional CASU case 1. Even case 7 was found to be on par with advanced cryogenic cases 4 and 5. Some of the key conclusions derived from the process simulations are as follows: (i) The energy requirement for oxygen production in ICLAS remains 75−80% lower than CASU. (ii) The plant thermal efficiency can be revamped close to the air-firing case by providing CLAS integration. (iii) The energy penalty can be significantly reduced to $59/MWh), and/or a high carbon tax (>$33/tonne). However, such numbers for CASU appear to be reasonably high (e.g., carbon tax of >$35/tonne, natural gas, and electricity price are not applicable). For areas where natural gas is abundant, e.g., the northern U.S. and some places in Australia, ICLAS can certainly become an attractive oxy-firing technology option over CASU. It is also noteworthy that the exercise carried out here is limited to the cost and performance data used/obtained in the current study. The present investigation is also restricted to the retrofit study. Several ICLAS configurations, such as ICLAS−S−CH4 indirect firing, and the use of other metal oxide oxygen carriers, such as Co or Mn, were not studied in this paper, which should be considered in the future.
the NPV of the CASU system decreases as the electricity price increases, which is mainly due to the increased losses in revenues attributed to the large power losses in CASU retrofit. Figure 23 shows the NPV of the oxy-fuel technologies as a function of the supplementary fuel price (i.e., natural gas price).
■
AUTHOR INFORMATION
Corresponding Author
*Telephone: +61-2-4044-9062. Fax: +61-2-4033-9383. E-mail:
[email protected].
Figure 23. NPV of various oxy-fuel technologies as a function of the supplementary fuel price.
Notes
The authors declare no competing financial interest.
As Figure 23 shows, the feasibility of the ICLAS retrofit system is adversely affected by an increasing natural gas price, while CASU shows no impact. This is due to the substantial amount of natural gas being used in the ICLAS system to provide heat for the reduction reactor. As such, a high natural gas price will transfer into higher expenses in the cash flow and result in a decreased NPV. The critical natural gas price below which the ICLAS retrofit system becomes superior to the CASU system was found to be around $4.1/GJ. In addition, for an economically viable ICLAS system, a natural gas price at below $3.5/GJ is required. This is quite possible in areas where abundant natural gas resources exist, such as the northern U.S. and some regions of Australia.
■
ACKNOWLEDGMENTS
■
REFERENCES
The authors acknowledge financial assistance provided through the scheme of Coal Innovation NSW Fund under NSW Trade & Investment and Glencore (formerly Xstrata Coal Research Limited). The authors also thank Dr. Allen Lowe for his contribution to the techno-economic assessment of the CLAS process.
(1) Moghtaderi, B. Application of chemical looping concept for air separation at high temperatures. Energy Fuels 2010, 24 (1), 190−198. (2) Shah, K.; Moghtaderi, B.; Zanganeh, J.; Wall, T. Integration options for novel chemical looping air separation (ICLAS) process for oxygen production in oxy-fuel coal fired power plants. Fuel 2013, 107, 356−370. (3) Pfaff, I.; Kather, A. Comparative thermodynamic analysis and integration issues of CCS steam power plants based on oxycombustion with cryogenic or membrane based air separation. Energy Procedia 2009, 1 (1), 495−502. (4) Moghtaderi, B.; Wall, T. F. A Novel Chemical Looping Based Air Separation Technology for Oxy-fuel Combustion; NSW Clean Coal Council: Sydney, New South Wales, Australia, 2010−2013. (5) Khalel, A. M. Z.; Rabah, A. A.; Barakat, T. A. M. A new cryogenic air separation process with flash separator. ISRN Thermodyn. 2013, 2013, 1−4. (6) Zhou, C.; Shah, K.; Moghtaderi, B. Techno-economic assessment of integrated chemical looping air separation for oxy-fuel combustion: An Australian case study. Energy Fuels 2015, 29 (4), 2074−2088. (7) Shah, K.; Moghtaderi, B.; Wall, T. Selection of suitable oxygen carriers for chemical looping air separation: A thermodynamic approach. Energy Fuels 2012, 26 (4), 2038−2045. (8) Song, H.; Shah, K.; Doroodchi, E.; Wall, T.; Moghtaderi, B. Analysis on chemical reaction kinetics of CuO/SiO2 oxygen carriers for chemical looping air separation. Energy Fuels 2014, 28 (1), 173− 182. (9) Song, H.; Shah, K.; Doroodchi, E.; Moghtaderi, B. Development of a Cu−Mg-based oxygen carrier with SiO2 as a support for chemical looping air separation. Energy Fuels 2014, 28 (1), 163−172. (10) Song, H.; Shah, K.; Doroodchi, E.; Wall, T.; Moghtaderi, B. Reactivity of Al2O3- or SiO2-Supported Cu-, Mn-, and Co-based oxygen carriers for chemical looping air separation. Energy Fuels 2014, 28 (2), 1284−1294.
4. CONCLUSION A techno-economic study of a hypothetical coal-fired power plant retrofit to oxy-fuel combustion using CASU and ICLAS was completed. The technical analyses were performed using Aspen Plus. The key issue identified with ICLAS during technical analyses is the heat integration. It was found that running the oxidation reactor at a lower temperature than the reduction reactor creates a problem of heat management, because transferring the heat generated at a lower temperature from the oxidation reactor to the higher temperature reduction reactor is thermodynamically not possible. To solve this, ICLAS was integrated with the new steam cycle and methane firing or solar heating. The results indicate that ICLAS when integrated with natural gas firing in the reduction reactor is technically more efficient than CASU. The energy requirement for oxygen production for ICLAS is 75−80% lower than CASU. Such technical comparison, however, is insufficient to prove the ICLAS concept in the lack of a detailed economic analysis. In fact, the further economic analyses carried out based on the NPV approach found that ICLAS was not always feasible primarily as a result of the high cost associated with external fuel and/or solar plant. This thus leads the future research direction of ICLAS technology toward cost reduction and further energy efficiency improvement. Nonetheless, the sensitivity analysis implies that the oxyfiring technology using the ICLAS system with CH4 integration can become economically viable at certain economic conditions, such as a low natural gas price (
