Optimisation and evaluation of flexible operation strategies for coal- and gas-CCS power stations with a multi-period design approach

Optimisation and evaluation of flexible operation strategies for coal- and gas-CCS power stations with a multi-period design approach

International Journal of Greenhouse Gas Control 59 (2017) 24–39 Contents lists available at ScienceDirect International Journal of Greenhouse Gas Co...

4MB Sizes 0 Downloads 32 Views

International Journal of Greenhouse Gas Control 59 (2017) 24–39

Contents lists available at ScienceDirect

International Journal of Greenhouse Gas Control journal homepage: www.elsevier.com/locate/ijggc

Optimisation and evaluation of flexible operation strategies for coaland gas-CCS power stations with a multi-period design approach Evgenia Mechleri a,b , Paul S. Fennell c , Niall Mac Dowell a,b,∗ a b c

Centre for Process Systems Engineering, Imperial College London, South Kensington, London SW7 2AZ, UK Centre for Environmental Policy, Imperial College London, South Kensington, London SW7 1NA, UK Department of Chemical Engineering, Imperial College London, South Kensington, UK

a r t i c l e

i n f o

Article history: Received 13 July 2016 Accepted 28 September 2016 Keywords: Flexible CCS Dynamic optimisation Dynamic process modelling Multi-period design

a b s t r a c t Thermal power plants are increasingly required to balance power grids by compensating for the intermittent electricity supply from renewable energy resources. As CO2 capture and storage is integrated with both coal- and gas-fired power plants, it is vital that the emission mitigation technology does not compromise their ability to provide this high-value service. Therefore, developing optimal process operation strategies is vital to maximise both the value provided by and the profitability of these important assets. In this work, we present models of coal- and gas-fired power plants, integrated with a post-combustion CO2 capture process using a 30 wt% monoethanolamine (MEA) solvent. With the aim to decoupling the power and capture plants in order to facilitate profit maximising behaviour, a multi-period dynamic optimisation problem was formulated and solved using these models. Four distinct scenarios were evaluated: load following, solvent storage, exhaust gas by-pass and variable solvent regeneration (VSR). It was found that for both coal- and gas-fired power plants, the VSR strategy is consistently the most profitable option. The performance of the exhaust by-pass scenario is a strong function of the carbon prices and is only selected at very low carbon prices. The viability of the solvent storage strategy was found to be a strong function of the capital cost associated with the solvent storage infrastructure. When the cost of the solvent tanks has been paid off, then the solvent storage scenario is 3.3% and 8% more profitable than the baseline for the pulverised coal and gas-fired power plants, respectively. Sensitivity analyses showed that, for all strategies, the flexibility benefit declined with reduced carbon and fuel prices, while a “peakier” electricity market, characteristic of one with significant quantities of intermittent renewables deployment, more significantly rewarded flexible operation. © 2017 Elsevier Ltd. All rights reserved.

1. Introduction Carbon capture and storage (CCS) has been proposed as a means to enable a least-cost transition to a low carbon energy system and is also important for industrial sectors (Mac Dowell et al., 2010; Boot-Handford et al., 2014). Given the increasing penetration of intermittent renewable electricity generation and the inflexible nature of traditional nuclear power generation 1 , decarbonised power plants need to be designed for flexible operation in order to be able to promptly respond to variation in electricity demand

∗ Corresponding author at: Centre for Process Systems Engineering, Imperial College London, South Kensington, London SW7 2AZ, UK. E-mail address: [email protected] (N.M. Dowell). 1 It is recognised that small modular reactors (SMRs) have the potential to offer a flexible form of nuclear power but at the time of writing, this is a relatively immature technology, and has not been widely deployed. http://dx.doi.org/10.1016/j.ijggc.2016.09.018 1750-5836/© 2017 Elsevier Ltd. All rights reserved.

(Bui et al., 2014; Davison, 2011; Oates et al., 2014) and to exploit the associated variation of electricity prices, while maintaining the carbon intensity of the plant at low levels (Haines and Davison, 2009; Cohen et al., 2011; Dowell and Shah, 2015; Lucquiaud et al., 2007). Flexible capture can be achieved in a range of ways. At the level of an individual power plant, flexible operation can be achieved using measures such as adding a solvent storage tank, bypassing the capture facility for certain time periods or operating the capture facility at different capture rates according to electricity output requirements. To the best of our knowledge, the concept of flexible operation, was first introduced by Gibbins and Crane (2004) in 2004, noting that this study makes reference to private communication with Prof Rochelle 2 on this subject in 2002. In the 2004 study,

2

Prof G. T. Rochell, U. Texas at Austin.

E. Mechleri et al. / International Journal of Greenhouse Gas Control 59 (2017) 24–39

the concepts of solvent storage and exhaust gas venting (or capture bypass) were first introduced. It this study, it was concluded that solvent storage had the potential to reduce electricity costs by 6–7% and that exhaust gas venting was a viable strategy in the event that electricity prices ($/MWh) were 2–3 times greater than carbon costs ($/tCO2 ). Here, in the case of solvent storage, an approximation of the additional capital cost associated with the infrastructure required for solvent storage was provided, but a detailed design of that equipment was not performed. Following this study, several contributions focused on flexible operation of the capture process as a way to improve the economics of CCS power plants either by reducing the capture level through exhaust gas venting, by storing the solvent using rich and lean amine storage tanks or by varying the degree of solvent regeneration (Oates et al., 2014; Haines and Davison, 2009; Cohen et al., 2011, 2011, 2012; Dowell and Shah, 2015; Rao and Rubin, 2006; Scoping, 2009; Lucquiaud et al., 2009; Ziaii et al., 2009; Chalmers et al., 2009, 2009, 2011; Qixin et al., 2010; Wiley et al., 2011; Husebye et al., 2011; IEA, 2012; Delarue et al., 2012; Versteeg et al., 2013; Domenichini et al., 2013; van der Wijk et al., 2014; Zaman and Lee, 2015; Arce et al., 2012; Mac Dowell and Shah, 2014). With the exhaust gas venting option, the power plant operates with partial or no capture of the CO2 . Under this strategy, the energy required for solvent regeneration is anticipated to be reduced or eliminated by venting a portion of the exhaust gas directly to atmosphere. Thus, the steam that would have been used for solvent regeneration is instead not extracted, resulting in increased net power output. From a practical perspective however, it may not be the case that all of the steam could be redirected to the LP turbine. It is important to note that the duration of the periods for which exhaust gas would be vented in response to a peak in electricity prices would likely be relatively short – on the order of 2–5 h Dowell and Staffell (2016). During this time, there are likely two options for operating the capture plant: 1. Continue to circulate the solvent through the plant as normal and 2. stop the solvent circulation and allow the plant’s solvent inventory to accumulate in the sumps and pipework. Option 1. has the advantage that it is ready to begin scrubbing CO2 from the exhaust gas as the plant is essentially “idling”. However, as the solvent is circulated, it will likely cool relatively rapidly as it moves from the well-insulated desorption process to the absorption process which may be open to the atmosphere. This would likely lead to a rapid cooling of the solvent towards ambient temperature in addition to the potentially significant losses of volatile organic compounds (VOCs) to the atmosphere. This may mean that there will be a non-negligible delay in returning the capture plant to its normal set-point of capturing 90% of the CO2 – thus potentially incurring a substantial cost associated with emitting CO2 during periods of relatively low electricity prices. This may well undo much of the profitability benefit associated from venting the exhaust gas in the first place. Further to this point is the potential for increased emission of VOCs, which could potentially compromise a facility’s license to operate. Option 2. has the advantage that it avoids much of the solvent cooling effect and also the VOC emission. However, there will be a delay associated with bringing the solvent circulation back to a steady state of operation such that the capture plant is again ready to capture CO2 . Thus, this may also result in the imposition of increased costs associated with emitting CO2 during periods of reduced electricity prices. To the best of our knowledge, neither of these points have been addressed in the literature to date, and represent clear and important avenues for future research. In the solvent storage mode, the CO2 capture level is kept constant and solvent storage tanks (rich and lean) are used to shift the regeneration load to times when the electricity price (and thus the economic opportunity cost associated with solvent regeneration) is low. Following the work of Gibbins and Crane, Rao and Rubin (2006), identified

25

the most cost-effective level of CO2 capture using the exhaust gas venting option. They concluded that the optimal CO2 capture level is dependent on plant size and, if exhaust gas venting is considered, the cost-effectiveness of CO2 capture can be improved. The importance of electricity and CO2 price variations in determining the cost-optimal level of CO2 capture has since been shown by several authors (Haines and Davison, 2009; Cohen et al., 2011; Ziaii et al., 2009; Chalmers et al., 2009, 2011; Versteeg et al., 2013; Patino-Echeverri and Hoppock, 2012; Mac Dowell and Shah, 2013). In their work, Haines and Davison (2009), reviewed the ability of the main capture technologies (pre-, post- and oxy- combustion) to modify their operation and design to provide some economic peak power capability. To our knowledge, this contribution is unique in that it evaluates the potential of these three types of CCS to operate flexibly. A key conclusion of their analysis was that postcombustion systems offered the greatest possibility of operating flexibly. This makes intuitive sense, as given that between pre-, post- and oxy-combustion capture, the nominal electricity output penalty of post-combustion CO2 capture is typically considered to be the greatest (Mac Dowell et al., 2010; Boot-Handford et al., 2014), it therefore stands to make the largest relative gain by reducing this penalty at opportune times. An important caveat is that the majority, if not all, of these studies were performed using aqueous solutions of 30 wt% monoethanolamine (MEA) as a solvent. This solvent typically requires 3.5–4.2 GJ/tCO2 captured (Boot-Handford et al., 2014), and as such imposes a large electricity output penalty on the power plant. However, the current industrial state-of-the-art solvents include Shell’s Cansolv, Fluor’s Econamine or MHI’s KS-1 solvents which typically use higher concentrations of active ingredient (typically between 40 and 50%) and have an energy of regeneration of 2.33 GJ/tCO2 (Dixon et al., 2014, 2014) 2.8–3.0 GJ/tCO2 and 2.5–2.8 GJ/tCO2 respectively. Importantly, all of these solvents require a similar quality (temperature) of steam for solvent regeneration, therefore a lower energy of regeneration leads to a reduced electricity output penalty. Moreover, solvents offering further improvement are on the horizon, such as those reported by Ye et al. (2015) wherein materials requiring 2.0 GJ/tCO2 at temperatures of 80–100 ◦ C are reported. We can readily evaluate the impact that these advanced solvents have on process performance using the IECM tool IECM (2016). IECM indicates that the higher heating value (HHV) efficiency of an ultra supercritical (USC) power plant is 42.83%. Its worth noting at this point that IECM is a relatively conservative tool, and current USC plants in service today exhibit HHV efficiencies of 44% and above. So-called advanced ultra supercritical (AUSC) plants have the potential to operate with steam temperatures of above 700 ◦ C and with HHV efficiencies in the region of 47–48% (IEA Coal Industry Advisory Board, 2017). Then, applying amine-based CO2 capture to the IECM USC power plant reduces the HHV efficiency to 29.03%. Using Fluor’s FG+ solvent (as described above) results in an HHV efficiency of 33.25%, MHI’s KS-1 solvent gives an HHV efficiency of 33.73% and finally Shell’s Cansolv solvent gives an efficiency of 34.33%. Similar calculations using an oxy-combustion option gives an HHV efficiency of 36.58% – still greater than the post-combustion options, but the gap is reduced. At this point, it is worth noting that the average annual HHV efficiency of the existing US coal-fueled electricity generating fleet is approximately 32%, and this can be substantially lower in some parts of the world IEA Coal Industry Advisory Board (2017). In other words, through the deployment of state-of-the-art power and capture plant technology, it is conceivable that decarbonised coal-fired power generation could be more efficient than it is today. Confirming the results presented by Rao and Rubin (2006), stopping the solvent regeneration during peak hours increases electricity generation by 20%. However, owing to the range of CO2 and electricity prices assumed in their analysis, the additional

26

E. Mechleri et al. / International Journal of Greenhouse Gas Control 59 (2017) 24–39

revenue derived from selling the electricity was quite small; between 0 and 4% of additional revenue above the baseline scenario. A key limitation to the enhanced profitability that may be derived from flexible operation is the compromise between peak electricity prices and CO2 prices – the peak electricity price needs to be significant to offset the additional cost associated with the emission of additional CO2 . A potential limitation of this study is that it performed its analysis based on the UK’s electricity system in the first decade of the 21st century. In this period, the electricity system was composed of nuclear, coal- and gas-fired power stations, and – in line with their analysis – the electricity market would not be characterised by excessive peakiness. Going forward, as the UK experiences increased deployment of intermittent renewable power Dowell and Staffell (2016), an upwards pressure may be expected on electricity prices and the electricity market may be characterised by an increased peakiness. This link between peak electricity prices and costs associated with CO2 emission was also observed in the work of Ziaii et al. (2009), who reported that flexibility may improve the annual operating profits; however, the balance of the electricity and CO2 price needs to examined. In their work, Chalmers et al. (2009, 2009), presented an updated version of Gibbins and Crane’s 2004 analysis and discussed the flexible operation of coal fired power plants with post-combustion capture. They identified exhaust gas venting and solvent storage as two options. The conclusion of this work was that exhaust gas venting is economically valuable if the price per MWh was two to three times higher than the cost per tonne of CO2 emitted, and that solvent storage significantly reduces the CO2 price at which exhaust gas venting is economically attractive, repeating the earlier conclusions of Gibbins and Crane (2004). Whilst both Chalmers et al. (2009, 2009), Gibbins and Crane (2004) noted that solvent storage would come at an additional cost, a detailed design of the required solvent storage infrastructure was not performed in either of their analyses. In their work, Cohen et al. (2011), have created optimisation and rule-based models within the General Algebraic Modeling System (GAMS) GAMS (2013), to study profitmaximising operation of a coal fired power plant with flexible CO2 capture with and without solvent storage under varying degrees of electricity price foreknowledge. They concluded that the gas venting option is unprofitable at high CO2 prices (above $70/tCO2 ), while solvent storage maintains a 9–29% profit at any CO2 price, highlighting the value of flexible CCS. In their work, Chalmers et al. (2011), performed a first order techno-economic screening analysis to determine whether solvent storage could be an important factor to contribute to the economic performance of the power plant. They concluded, similarly to Haines and Davison (2009), that “the revenue increase which could be obtained in any one day by using solvent storage varies considerably depending greatly on the shape of the daily electricity price curve” and “could be an attractive option in some electricity networks”. When discussing flexible operation, it is important to bear in mind additional capital equipment costs, i.e., storage tanks, oversized power plant equipment, such as larger reboilers and so forth (Chalmers et al., 2009; Haines and Davison, 2009; Husebye et al., 2011; Brasington, 2012). In their work, Patino-Echeverri and Hoppock (2012), presented an analysis on the different electricity prices for an amine-storage to a CCS system. They examined two different plants (existing subcritical and new supercritical) and two design modes of the storage tank; two-mode and three-mode. In a two-mode system the solvent regeneration system has a binary mode of operation and either runs at 100% or 0% capacity. In the three-mode system, the solvent regeneration process might (1) run at 100% capacity to regenerate both the solvent flowing from the absorber and the stored solvent from the storage tank, (2) run at 0% capacity, or (3) operate so as to regenerate only the volume of solvent required for the absorber at that time. The study of Patino-Echeverri and Hoppock (2012) stands

out as one which does perform a detailed engineering design of the solvent storage tanks. Here, they assume that the additional volume of solvent will cost between $629 and 711/m3 and the total cost for the additional solvent and storage tanks will be $6.8M and $2.5M respectively. Here, carbon steel storage tanks were assumed, and, as noted by Haines and Davison (2009), solvent degradation effects would need to be properly taken into account. It may be that stainless steel storage tanks would we required, which could substantially increase the associated capital cost. They found that the required price differential was in fact a complex function of the cycling period, the capacity factor, the storage size, and whether the plant is a retrofit or new. The required price differential for two-mode operation ranged from $40 to 111/MWh for daily cycling and $92–677/MWh for weekly cycling. In the three-mode case the range was found to be $43–$141/MWh for daily cycling, and from $110/MWh to $285/MWh for weekly cycling. In general new plants require much higher price differentials to justify investment in solvent storage. This makes intuitive sense as the up-front capital expenditure of post-combustion CCS is already significant and current research efforts are prioritising its reduction as a means to reduce the $/MWh cost of CCS electricity. In their work, Versteeg et al. (2013), considered the profitability of coal and natural gas-fired power plants with amine and ammonia post-combustion CO2 capture for variable electricity prices. They have concluded that the solvent storage option increased profitability at low carbon prices (£40–60/tCO2 ). Husebye et al. (2011), developed an mixed integer linear programming (MILP) model to identify the optimum operating strategy of a coal fired power plant with post-combustion capture, and evaluated the potential value of flexible solvent regeneration and storage. The results showed that flexibility can lead to increased profits, particularly in volatile electricity markets. Finally, a correlation between profitability and cyclical (weekly, seasonally, etc.) demand patterns was observed. However, flexible operation is limited by case specific parameters, such as the size of the storage tanks or the maximum size of the desorber, which needs to be taken into account in the techno-economic analyses. In their work, Brasington (2012), presented an integrated coal fired power plant with a post-combustion CCS plant. They concluded that the operational complexity increases with solvent storage and due to increased operational and capital costs imposed, the profitability of the plant does not increase for long periods of storage (hours). They added that there might be a potential for short duration of solvent storage (i.e., less than 30 min), since this will not increase the operational complexity of the conventional coal fired power plant. The most recent contributions on the flexible operation of CO2 capture systems are reported by van der Wijk et al. (2014), Oates et al. (2014), Dowell and Shah (2015), Zaman and Lee (2015), Adams and Dowell (2016). In their work, van der Wijk et al. (2014) showed that the flexible options are not utilised. This is due to either the prevailing CO2 prices in Europe which do not favour the exhaust gas venting options or the regeneration constraints of the base load power plant for the solvent storage. However flexible CCS plants can increase the reserve capacity provision by 20–300% compared to non-flexible plants. The paper of Oates et al. (2014) is the first to discuss the flexibility options available to a natural gas fired power plant with post-combustion capture using MEA as a solvent. Their framework incorporated both a design and operating optimisation model to explore the exhaust gas venting and solvent storage as flexible options. The concluded that flexible CCS could result in an increased profit in the range 0–35% depending on the design of the regenerator and capacity of the solvent storage tanks. In the majority of the previous studies, the decision variables for flexible operation were primarily the capture level for exhaust gas venting scenarios and the regeneration rate for solvent storage options. These variables were not treated as optimisation variables but were varied according to different energy prices. Adams and

E. Mechleri et al. / International Journal of Greenhouse Gas Control 59 (2017) 24–39

Mac Dowell presented a detailed study of a CCGT integrated with a CO2 capture and compression process. Here, the performance of this system was evaluated under part-load conditions, with a key observation being that the whilst the cost structure of the integrated process remains approximately constant for off-design point operation, this will appreciably increase the levelised cost of electricity (LCOE) of these plants, which may have implications for the average price of electricity of the systems in which these processes are integrated. Two recent contributions discuss the rigorous optimisation of flexible CCS systems; that of Dowell and Shah (2015), Zaman and Lee (2015). In their work, Zaman and Lee (2015), presented an optimisation model for a post-combustion capture model for three flexible configurations: exhaust gas venting, solvent storage and combination. Compared to the base case, the three modes of operation showed 3.04%, 10.1% and 11.08% savings, respectively. In our previous work Dowell and Shah (2015), we presented a multi-period optimisation problem to evaluate the profitability of a load following coal fired power plant integrated with a postcombustion capture plant for three different operating strategies. As an addition to the literature on this subject, this paper introduced the concept of variable solvent regeneration (VSR) as another strategy of the flexible operation of the capture plant. In this study, it was shown that allowing CO2 to accumulate in the working solvent during periods of high electricity prices and the subsequent regeneration of the solvent during periods of low electricity prices offered substantially improved profitability over either venting exhaust gas or solvent storage. In the case of solvent storage, a key limiting factor was found to be the quantity of steam available from the power plant, and in the case of exhaust gas venting it was found that, in order to capture 90% of the CO2 produced by the power plant, it simply was not possible to vent a substantial portion of the CO2 , and similarly to previous work, the venting of CO2 was observed to incur a substantial cost. In this contribution we present a comprehensive study of the options for maximising profits via flexible operation whilst maintaining a low average carbon intensity (kgCO2 /MWh) for both coaland gas-fired power stations. A multi-period, dynamic optimisation problem is formulated and implemented in the gCCS toolkit3 and solved using the default solvers available within gPROMS.4 Using a load-following plant as the base case scenario, we consider three options for flexible operation of both coal- and gas-fired power plants: exhaust gas venting, solvent storage and time-varying solvent regeneration. In the case of the solvent storage option, we calculate the capital cost associated with the storage tanks and then mediate that as an increase in the short run operating cost of the plant. The remainder of this paper is laid out as follows: In Section 2, we present our approach for calculating electricity prices over the course of a 24 h period, the power plant and capture plant models and the optimisation problem. Section 3 presents the results and discussions for the different scenarios and in Section 3 we present the conclusions of our work. 2. Model development 2.1. Consideration of the electricity system in which CCS will operate Many techno-economic analyses of CCS in the literature assume steady state operations of the plant, consistent with baseload

3 Process Systems Enterprise. (2014). gCCS overview. Retrieved (09.09.14), from: http://www.psenterprise.com/power/ccs/gccs.html. 4 Process Systems Enterprise, gPROMS, http://www.psenterprise.com/gproms, 1997–2015.

27

power generation (Oates et al., 2014; Abu-Zahra et al., 2007, 2007; Dave et al., 2011; Khalilpour and Abbas, 2011). However, as noted in the introduction, it is quite unlikely that CCS power plants will operate in a baseload fashion in many electricity markets. Rather, they may be required to operate in an electricity system containing a large proportion of intermittent renewable energy and will consequently be required to provide a flexible, load-following service. A model price profile for a twenty four hour period was constructed by calculating the short run marginal cost (SRMC) for different types of plants (super-critical pulverised coal (SCPC), combined cycle gas turbines (CCGT) and open cycle turbines (OCGT) using the following equation:

£SRMC MWhr

=

£MWhr Fuel nplant



CO



Var

CO

2 2 2 +£ + £Tonne · CI TonnesCO O&M + £T &S MWhr

(1)

where £SRMC is the SRMC of the electricity generated by a given plant. In this calculation the variable operating and maintenance 2 ) for transport and storage are costs (£Var ) and fixed cost (£TCO &S O&M also considered. The data for this equation were obtained from the Department of Energy and Climate Change (DECC) for the fossil fuel prices and carbon prices and by the Joint Research Centre of the European Commission Vatapoulos et al. (2012) for the efficiencies and carbon intensities. These values are presented in Table 1. Following our previous work (Dowell and Shah, 2015), it is then assumed that over-night (off-peak) electricity prices will be set by SCPC plants, day-time prices will be set by CCGT plants with morning and evening peaks serviced by OCGT plants (Dowell and Shah, 2015). This is illustrated in Fig. 1, where the variable electricity profile for a 24 h period is presented. As can be observed, there are 6 distinct periods of operation: 2 peak periods (06:00–10:00 and 16:00–19:00) and 4 off-peak periods. In Fig. 1, the capacity factor of the load-following plant is also illustrated. 2.2. Pulverised coal-fired power plant A model of a supercritical pulverised coal power plant (SCPC) was developed using the SCPC model provided by the gCCS toolkit (PSE, 2015), illustrated in Fig. 2. The inputs of the model are the nominal power output, inlet and outlet steam conditions of the LP turbine, and flowrate of steam extracted as a function of the CO2 captured. Steam is extracted at the inlet of the LP turbine. The electricity output of the standalone power plant model is 500 MWe, while integrated with capture is 440 MWe. Compression is not considered in this work, since beyond recycling CO2 , it cannot provide flexibility. The efficiency of the standalone SCPC model is 44%, while integrated with the capture plant is 38 %. These values agree with the literature where 6% efficiency penalty is reported for post-combustion capture with MEA (Davison, 2007). The nominal power output along with the temperature and pressure of the steam at the extraction point is specified (T = 506 K, P = 3.9 bar), while the flowrate of the required steam is calculated by the capture plant by setting a value at the capture rate, taken to be 90% as a base case in this work. 2.3. Combined cycle gas-fired power plant A model of a Combined Cycle Gas Turbine plant was developed in order to specify the flowrate and composition of the flue gas stream supplied to the capture plant as well as the flowrate and thermodynamic state of the steam provided for the regeneration of the solvent. This model is based on the CCGT model provided by the gCCS toolkit (PSE, 2015) as it is illustrated in Fig. 3. For the CCGT model, we have used the Siemens SGT5-4000F with the usual configuration of one gas turbine and three steam turbines. The electricity output of the standalone power plant model

28

E. Mechleri et al. / International Journal of Greenhouse Gas Control 59 (2017) 24–39

Table 1 Values from DECC and the JRC for use in evaluating Eq. (1). The values reported here are for the central scenario by DECC. However, in this work a sensitivity analysis has been performed for fuel, carbon and electricity prices (more details in Sections 3.1.5 and 3.2.5) The efficiency values reported here are for plants with no CCS. In Eq. (1), we impose an 8–10% penalty on the power plant. The carbon intensity is for a capture plant operating at 90% capture for SCPC and CCGT while the OCGT operates in an unabated fashion. Fuel price (£/MWh)

nplant

CO2 price (£/tonneCO2 )

CI (tonneCO2 /MWh)

VarO&M (£/tonneCO2 )

T&S (£/tonneCO2 )

Electricity price (£/MWh)

Central scenario

SCPC CCGT OGCT

9.86 24.53 24.53

55 60 42

70 70 70

0.07 0.04 0.49

4.38 3.06 1.53

19.60 19.60 19.60

42.40 62.62 99.94

High scenario

SCPC CCGT OGCT

13.89 35.04 35.04

55 60 42

105 105 105

0.07 0.04 0.49

4.72 3.86 1.93

32.20 32.20 32.20

61.30 90.40 144.95

Low scenario

SCPC CCGT OGCT

7.17 14.02 14.02

55 60 42

35 35 35

0.07 0.04 0.49

4.02 2.46 1.23

8.20 8.20 8.20

27.15 35.40 55.04

Fig. 1. Illustration of the multi-period optimisation concept. In this graph the red line represents the electricity price for the base case. The black line represents the scenario with negative electricity prices. The dashed lines are the price differential (PD) between high and low electricity prices for the two cases. From this it can be observed that there are 6 distinct periods of operation denoted by the change of electricity prices within the 24 h period. The blue line represents the power plant capacity factor illustrating the load following profile of the power plant. The question we are addressing here is what is the optimal operation of the capture plant in order to maximise the profit during these periods. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

is 421 MWe, while integrated with capture is 395 MWe; 73% of the total power comes from the gas turbine while the remainder is provided by the steam turbines. The efficiency of the standalone CCGT model is 59%, based on the Siemens SGT5-4000F gas turbine Siemens (2017), while integrated with the capture plant is 53%. This is a 6% efficiency penalty for capture which agrees with the literature (Davison, 2007; Adams and Dowell, 2016). The gas turbine performance is changed by anything that affects the density and/or mass flow of the air intake to the compressor. As the ambient temperature increases, the air density reduces and so the mass flow to the compressor is reduced, reducing the system output and efficiency. The temperature effect is dependent on the turbine model, as it depends on the cycle parameters and component efficiencies. In order to account for the performance variation as a function of the ambient temperature we specify the gas turbine efficiency change (−0.1% ◦ C) and the exhaust flow change (−0.45 kg/s/◦ C) per degree ambient temperature change, following previous studies (Farouk et al., 2013; Ameri and Hejazi, 2004). This is very important, since in cases when the air temperature is very high, a pre-cooling system may need to be included. For example, for Australia, India or the GCC region where air temperatures and humidities (and thus densities) are substantially different to those

Fig. 2. This model is used to simulate a supercritical pulverised coal power plant and its integration with a capture plant. The SCPC high level model has three inlet and three outlet material ports. 1 – Coal inlet, 2 – Air inlet, 3 – Waste outlet, 4 – Flue gas outlet, 5 – Condensate inlet, 6 – Steam outlet. Port 7 is the cost port for connecting to the cost model. The inputs are the nominal power output, inlet and outlet steam conditions of the LP turbine and flowrate of the steam extracted.

of the UK and can vary significantly over the year. At the extraction point, the temperature and pressure of the steam should be specified (T = 500 K, P = 3.5 bar). Similarly to the coal plant, this allows the calculation of the electricity output penalty associated with the operation of the capture plant. The model also performs a check to determine whether the steam flow requested by the capture plant is available and flags a warning if the flow is likely to be grater than 90% of the total flow in the turbine, thus protecting the components of the CCGT from technical failure. Finally, the natural gas composition used in this work is: 95.2% CH4 , 3.26% C2 H6 , 1.03% C3 H8 and 0.51% C4 H1 0 NG (2006). 2.4. Capture plant model The model of post-combustion CO2 process has been modelled using the gCCS toolokit as illustrated in Fig. 4. The specifications for this model are the gas outlet temperatures for the heat balance and the CO2 capture rate which is set at 90%. From a list of different solvents (MEA (monoethanolamine), DEA (diethanolamine), DGA

E. Mechleri et al. / International Journal of Greenhouse Gas Control 59 (2017) 24–39

29

Table 2 Constraints and decision variables for the different scenarios. The first column are the scenarios for both the coal and gas-fired power plants with capture, the second column are the path constraints which must be satisfied at all times during operation, the third column are the end-point constraints which must be satisfied at the end of operation and the last column are the decision variables of the optimisation problem. Fsolv is the solvent flowrate in kg/s, CO2acc is the CO2 accumulated in kg/s, Frbp is the by pass fraction, Frst is the fraction to storage, and ˛, ˇ,  are the parameters used in the equation of the lean loading as a function of t for the solvent regeneration scenario.

Fig. 3. This model is used to simulate a combined cycle gas-fired power plant and its integration with a capture plant. The CCGT high level model has three inlet and two outlet ports. 1 – Fuel inlet, 2 – Air inlet, 3 – Flue gas outlet, 4 – Steam outlet, 5 – Condensate inlet. The inputs are the number of steam and gas turbines, the combined cycle efficiency (library or user-specified) and the temperature an pressure of the stem extracted.

Scenario

Path constraints

End-point constraints

Decision variables

Load following Solvent storage

DoC = 90% 0 ≤ DoC ≤ 100

– Frst

Exhaust gas venting

0 ≤ DoC ≤ 100

– 89.9 ≤ IDoC ≤ 90, CO2acc =0 89.9 ≤ IDoC ≤ 90

89.9 ≤ IDoC ≤ 90

˛, ˇ, 

-1 ≤ Frbp ≤ 1 0 ≤ Fsolv ≤ 1000 0 ≤ DoC ≤ 100

Time varying solvent regeneration

Frbp , Fsolv

the optimum design of multi-purpose chemical plants Grossman and Sargent (1979). The design and multi-period operation of the decarbonised power plant can be represented by the system of mixed differential and algebraic equations of the form: ˙ f (x(t), y(t), u(t), v, x(t)) = 0 ∀t ∈ [0, tf ]

(4)

where x(t) and y(t) are the differential and algebraic variables in ˙ are the time derivatives of the x(t). The conthe model, while x(t) trol variables, u(t), and the time invariant variables, v, are to be determined by the optimisation. In this study, the control variables u(t) include the bypass fraction to storage for the solvent storage scenario, the bypass fraction and the lean solvent flowrate for the exhaust bypass scenario and the lean solvent loading for the time varying solvent regeneration scenario. In some applications it is necessary to impose certain conditions that the system must satisfy at the end of the operation, i.e. the endpoint constraints. These can be equality or inequality end point constraints of type: Fig. 4. This model is used to simulate a post-combustion CO2 capture plant. It has two inlets, the flue gas inlet and the steam inlet and three outlets; the treated flue gas directed to the stack, the CO2 to compression and the condensate outlet from the reboiler.

w(tf ) = w∗ ,

wmin ≤ w(tf ) ≤ wmax

(5)

(Diglycolamine), MDEA (methyldiethanolamine) we have chosen a 30%wt MEA solvent with 0.23 lean loading and 0.5 rich loading. These specifications are in turn used to determine the required solvent flowrate. An illustration of the integrated coal- and gas-fired power and CO2 capture plants, are illustrated in Figs. 5 and 6, respectively.

where w is one of the system variables (x or y). In our case, the end-point constraints were that the IDoC would be in the range 89.9–90%, that there is no CO2 accumulation at the end of the optimisation period for the solvent storage scenario and that the lean loading is at the same value at the end of the optimisation period as that at the beginning for the variable solvent regeneration scenario. Our problem is also subject to path constraints in the case of the solvent storage and exhaust gas venting scenarios:

2.5. Optimisation problem

wmin ≤ w(t) ≤ wmax

In this study we have distinguished between the Degree of Capture (DoC) and the Integrated Degree of Capture (IDoC) as described in the following equations:



DoC = 100 ·

 IDoC =

COGenerated − COEmitted 2 2 COGenerated 2



(2)

tf

DoC dt

(3)

t0

where t0 and tf are the start and end periods of interest. The dynamic optimisation problem solved in this study is based upon the theory of Grossman and Sargent, originally intended for

∀t ∈ [0, tf ]

(6)

In the solvent storage scenario, the bypass fraction to storage should be between −1 and 1, the exhaust gas by pass fraction between 0 and 1, the lean solvent flowrate between 0 and 1000 kg/s (704 kg/s are required for 90% capture at full plant capacity) and the DoC to be in the range of 0–100%. The constraints and decision variables for each of the four scenarios are presented in Table 2. The dynamic optimisation seeks to determine the time variation of the control variables u(t) over the time horizon t ∈ [0, tf ] so as to maximise the final value of a single variable z subject to constraints (5)–(7): maxu(t),t ∈ [0,tf ] z(tf )

(7)

30

E. Mechleri et al. / International Journal of Greenhouse Gas Control 59 (2017) 24–39

Fig. 5. Integrated SCPC power plant and post combustion plant. There are three connection points between the power and capture plant. The exhaust gas flow rate from the power plant, going in the absorber column of the capture plant, the steam inlet to the reboiler of the capture plant for the solvent regeneration from the LP turbine and the condensate return to the power plant. A stack is also used for the treated flue gas.

Fig. 6. Integrated CCGT power plant and post combustion plant. There are three connection points between the power and capture plant. The exhaust gas flow rate from the power plant, going in the absorber column of the capture plant, the steam inlet to the reboiler of the capture plant for the solvent regeneration from the LP turbine and the condensate return to the power plant. A stack is also used for the treated flue gas. In this model we can also see an additional stack which is used for the exhaust gas venting scenario, where untreated flue gas is emitted to the atmosphere.

where z(tf ) is the short-run marginal cost (SRMC) profit of the plant.

the electricity price differentials in order to explore the difference in low and high electricity prices at which this scenario is profitable.

3. Results and discussion In this section we present the results of our study. We start with the SCPC, considering the load following, the solvent storage, the exhaust gas venting and the variable solvent regeneration scenarios and compare them by performing a sensitivity analysis on carbon and electricity price differentials. We then present the results for the same scenarios for the CCGT model. For the solvent storage scenario to be profitable, the benefits from the electricity price arbitrage need to exceed the significant capital costs associated with building the solvent storage infrastructure (Patino-Echeverri and Hoppock, 2012). In order to account for the capital cost influence on the storage scenario, we calculate the capital cost of the storage tank as an additional marginal cost and we include it in the profit function. This is discussed in detail in Section 3.1.2. We then perform a sensitivity analysis on

3.1. SCPC model 3.1.1. Load following In the first scenario, the capture plant operates in accordance with the power plant. As illustrated in Fig. 1 the power plant ramps up and down during the period of one day with variable electricity prices, while the DoC and the lean loading are kept at 90% and 0.23, respectively. In Fig. 7, a sensitivity analysis on the CO2 price is presented. A very low price of CO2 of £ 10/tCO2 , can lead to 7% increase in the total daily profit, as compared to the base case (black line) in the figure. For a capture rate kept at 90% and a very high carbon tax (£ 200/tCO2 ) the decrease in the profit from the base case is 18%. This shows that the carbon price is important to the overall profitability of this plant. A further implication of this observation is that CO2

E. Mechleri et al. / International Journal of Greenhouse Gas Control 59 (2017) 24–39

Fig. 7. Cumulative profit-SCPC-load following scenario. In this figure we illustrate the variation of CO2 prices as a function of the cumulative profit. The results of this sensitivity analysis show that the carbon price has a significant influence on the total profit and can lead to 7% increase for low CO2 price to 18% profit decrease if the CO2 prices are more than 150% of the base case (£ 70/tCO2 ).

prices should be sufficiently high to incentivise a high DoC, but no higher. Beyond a certain point, it simply becomes punitive.

31

Fig. 9. Storage volume of rich and lean tanks for a storage volume of 10,000 m3 . The blue line shows the electricity price variation within the day. The black dashed line is the solvent stored in the lean tank and the solid black line is the solvent stored in the rich tank. During high electricity prices the rich tank is filling up and the opposite during low electricity prices.

accumulated should be zero. The additional equations which describe this optimisation scenario are: Fsolvreb = Fsolv · (1 − SF)



tf

Rich CO2acc = Fsolvsto · sto · wsolvent · min(0, SF)dt

(8) (9)

t0

3.1.2. Solvent storage In the solvent storage scenario, two solvent storage tanks are added between the absorber and the stripper. If solvent storage is available then a portion of the rich solvent can temporarily stored, rather than being sent to the desorber for immediate regeneration. This stored rich solvent can then be subsequently regenerated by adding it to rich solvent generated by ongoing operations during a period of relatively low electricity prices. Previously-stored lean solvent from another tank is used to allow capture to continue. This scenario is illustrated in Fig. 8. The question in this scenario is how much solvent is sent to the storage tanks during the period of the simulation, subject to the constraint that at the end of the simulation the CO2

where, Fsolvreb is the solvent storage directed to the reboiler, Fsolv is the total solvent flowrate for 90% capture, Fsolvsto is the amount of Rich is the solvent stored, wsolvent is the mass fraction of the solvent, sto rich loading of the solvent stored and SF is the split fraction. If SF is less than zero, then the rich solvent is being regenerated whereas when SF is greater than zero, the rich solvent is being stored. The results of this optimisation problem are presented in Fig. 9. As can be observed, there are two periods of solvent storage during periods of high electricity prices (06:00–10:00 and 16:00–19:00) and four periods of solvent regeneration, with higher regeneration at the low electricity prices (£55/MWh) and lower regeneration at electricity prices of £70/MWh. Additional capital expenditure is required for the storage tanks which depend on a number of factors, such as the volume of solvent

Fig. 8. Solvent storage flowsheet. This figure presents the option of solvent storage in order to increase power output during peak times. The rich solvent flow is diverted from the absorber to a rich solvent storage tank instead of routing to the stripper. The lean solvent that is regenerated during low electricity price periods is stored to the lean solvent storage tank used for capture during peak electricity price periods.

32

E. Mechleri et al. / International Journal of Greenhouse Gas Control 59 (2017) 24–39

Fig. 10. Profit margin between the load following base case scenario and the solvent storage scenario with and without the capital cost of the storage tanks for different storage tank sizes. This figure shows a comparison between the base case scenario (black column) and the solvent storage scenario with capital cost with 3 years payback period (grey columns), 10 years payback period (blue columns) and without capital cost (red columns) for different solvent storage sizes. For the case where solvent storage cost is considered with 3 years payback period, as the capital cost of the storage tanks decreases with the size, the cumulative profit increases. The revenue from the electricity sold cannot outweigh the capital cost of the storage tank. When the payback period increases to 10 years then the cumulative profit increases with the storage tank size. After the capital cost has been paid off (red columns) there is an increase to the cumulative profit proportionally to the size of the tanks. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

required per kg of CO2 absorbed and the mass of CO2 absorbed. For a 30% MEA loading and a lean loading of 0.25 the required capacity for storage is 10,000 m3 for 90% capture in order to have a net power output increase of around 20% (Chalmers et al., 2009), and this is what we have used as a base case for this study. However, since the size of the storage tanks affects the profit, we have performed a sensitivity analysis to show this variation as presented in Fig. 10. At the start of the optimisation, the initial level at the rich tank is 0.3 m and at the lean tank is 0.7 m. When considering the solvent storage, the capital costs of the storage tanks need also to be considered. In this work we assumed that the storage tanks used are erected on site and are composed of stainless steel 304 Oates et al. (2014). We estimated the capital costs of the tanks using Couper’s Handbook Couper et al. (2010) as described in Eq (10). 2

CC st = 0.72 · 1.218 · FM · exp[11.662 − 0.6104 · (lnV ) + 0.04536 · (lnV ) ]

(10)

where CCst is the capital cost for the storage tank in £ and FM is 3.4 for a stainless steel tank 304. As Couper provides costs in 2003 US$, we then escalated these costs to 2015 using the Chemical Engineering plant cost index (CEPCI) (CEM, 2010), as described in Eq. (11), and converted them to £ GB, assuming a currency conversion of $1.5/£. CC st 2015 = CC st 2003 ·

CEPCI 2015 CEPCI 2003

(11)

where CEPCI2015 was 550.4 and CEPCI2003 was 402. Given that the addition of the solvent storage infrastructure is analogous to an efficiency improvement, it is reasonable to require a short payback period for this investment. As a base case in this analysis, we have selected a three year payback period. For the 10,000 m3 tank, using Eqs. (10) and (11) we calculate the capital cost to be £715k. We then divide this cost by 1095 in order to calculate the cost per day and then divide this by the integrated net power output of the power plant with solvent storage to transform

Fig. 11. Exhaust gas venting flowsheet. The flue gas from the power plant are redirected to the stack (black dashed line), mixed with the treated flue gases and vented. No modification and extra investment is needed for this option.

the capital cost into £/MWh. The SRMC of the plant is reformulated to include this cost: £

MWhr

SRMC

MWhr

=

£Fuel

nplant



CO

2 2 + £Tonne · CI TonnesCO MWhr



CO

+ £VarO&M + £T &S2 + £

SS

(12)

where £SS is the cost of the storage tank. We have also considered the case where the capital cost has been paid off (after 3 years) by solving the aforementioned optimisation problem without considering the cost of the storage tanks. In Fig. 10, we can observe the results on the total profit of the solvent storage scenario compared to the base case load following scenario and evaluate it for different solvent storage capacities, in each case accounting for the capital cost of the storage tanks. As can be observed, the cumulative profit of the base case scenario (black column) is always greater than the solvent storage scenario for any size of the solvent storage tanks (grey columns) when considering the capital cost of the tanks with payback time 3 years. Since the capital cost of the storage tanks decreases as the storage tank size decreases, the difference between the cumulative profit of the base case and the solvent storage cases decreases from 2.2% for a tank of 10,000 m3 to 0.4% for a very small storage tank of 500 m3 . This shows that, for the scenarios considered here, the additional electricity sold cannot outweigh the capital cost associated with the storage tanks. When the payback period is increased to 10 years (blue columns), then the solvent storage scenario is more profitable than the base case and this profit increases as the solvent storage size increases. As was mentioned in the introduction, many studies in the literature exclude the capital cost associated with the storage tanks, and conclude that the solvent storage scenario can have an additional profit of approximately 3.5% Zaman and Lee (2015). We have also examined this option and arrive at the same conclusion – we achieve an increase in profitability of 3.3% savings for the 10,000 m3 tank size relative to the base case, after the solvent storage capital cost has been paid off. This profitability decreases as the size of the storage tanks decreases Zaman and Lee (2015), which shows that even if the capital cost for the large storage tank gives less profit during the payback period, after this time a larger storage tank gives more profit and should be the one installed. 3.1.3. Exhaust gas venting In the exhaust gas venting scenario, as illustrated in Fig. 11 there is no modification to the plant’s original design. In this

E. Mechleri et al. / International Journal of Greenhouse Gas Control 59 (2017) 24–39

Fig. 12. Exhaust gas venting scenario for the SCPC for a constrained IDoC. The black line represents the fraction of the exhaust gas that is vented and the blue line presents the electricity price variation. 21% of the exhaust gas is vented during periods with high electricity prices while a DoC ∼100 % is chosen at all other times, leading to an IDoC of 90% at the end of the day. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

scenario, a portion of the exhaust gases are re-directed upstream of the absorber to the stack and vented directly to the atmosphere. In this scenario, the power plant ramps up and down as illustrated in Fig. 1 and in order to decouple the operation of the power and capture plants, we consider the option of venting part of the exhaust gas during periods of high electricity prices. Once again, the problem is solved subject to the end-point constraint of 89.9 ≤ IDoC ≤ 90. In this case, the vent fraction is 21% for both periods of high electricity prices as illustrated in Fig. 12. The reason for this is that the end point constraint of IDoC set to 90% requires the model to select venting in order to avoid solvent regeneration costs during periods of high electricity prices. During the time periods with lower electricity prices the DoC is ∼100% to make-up for what was emitted. However, we also solved this problem where the end-point constraint was relaxed such that 89.9 ≤ IDoC ≤ 100, or in other words the model was free to capture as much CO2 as was rational to maximise the profit. In this instance, for a CO2 price of £ 70/tCO2 , venting was not selected, and the model solved for an IDoC = 100%. It was only when the CO2 price was reduced to £10/tCO2 that venting was selected. This is a potentially interesting result. First, we must recall that in this problem, we are solving based on maximising the profit on a short run marginal cost basis. Whilst this is how power plants are typically dispatched within an electricity market, the SRMC does not include the capital or pre-development costs of the power and capture plant. Therefore, care must be taken not to interpret this result as a inferring that a CO2 price of £10/tCO2 is sufficient to incentivise the construction of the CCS power plant given the electricity prices discussed in this paper. However, the CCS plant would have an economic lifetime of 40 years but would typically aim for a payback period of 10–20 years, depending on rates of return required by investors. Therefore there is likely to be a significant period for which the power plant is operating on a SRMC basis. 3.1.4. Time varying solvent regeneration In this scenario, we use the working solvent as means to provide flexibility to the power plant. This is achieved by allowing CO2 to accumulate in the solvent during hours of peak electricity prices and regenerating the solvent during off-peak periods. The lean loading is therefore no longer a fixed variable as in the previous scenarios but can vary with time as expressed in the following equation: t Lean = ˛t · t 2 + ˇt · t +  t

(13)

33

Fig. 13. Solvent regeneration scenario as a function of electricity price and degree of capture (DoC) for the SCPC. As it is observed, instead of operating at a constant lean loading, the lean loading varies within the day in sympathy with the variable electricity prices. The solvent is regenerated during low electricity prices dropping down to 0.12 and CO2 is accumulated in the solvent during high electricity prices with increased loading up to 0.235. The DoC (red line) varies within the day, however the IDoC is 90% at the end of the optimisation framework. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.) Table 3 Integrated degree of capture (IDoC) and cumulative profit-SCPC. Scenario

IDoC (%)

Profit (k£)

Load following Exhaust gas venting (£70/tonCO2 ) Variable solvent regeneration Solvent storage (with CAPEX) Solvent storage (no CAPEX)

90 90 90 90 90

450 471 503 440 465

where the lean loading can vary in different time periods t, based on the quadratic expression above by varying the parameters ˛, ˇ and . The only constraints imposed are the IDoC should be 90% at the end of the optimisation and that the lean loading is bounded between 0.15 and 0.5 over the course of the optimisation. The variable time is set to zero at the beginning of each new period. The results of this optimisation problem are presented in Fig. 13. As it can be observed from Fig. 13, rather than operating at a fixed value, the lean loading varies with the electricity prices. With this operation, the plant redirects less steam for solvent regeneration when electricity prices are high, allowing the plant to increase profitability by selling more electricity. From the same figure we can observe that the degree of capture (DoC) varies within the day and drops down to 82% when the rate of solvent regeneration is low (or the lean loading is high). However, the cumulative capture at the end of the simulation is ∼ 90%. The total profit of this scenario is £503k, or 10.5% more profitable than the base case scenario. 3.1.5. SCPC Scenario comparison When comparing the various modes of flexible operation for the integrated pulverised coal power plant and amine-based capture plant, there are several things that need to be considered. In Figure, we present the cumulative costs for the different alternatives. For the solvent storage (SS) scenario, the capital cost of the storage tanks is an important factor that needs to be taken into account, as this contribution makes the difference between storage being profitable or not – particularly for large volumes. In addition, the rate at which the storage volume was discounted made a substantial difference to the profitability of this scenario. This is particularly evident in the case of a storage tank of 10,000 m3 , where an increase in profit of 3.3% relative to the base-case was observed, one the cost had been paid off. Interestingly, this distinction was reduced as storage volume was decreased. The data are presented graphically in Fig. 14 and tabulated in Table 3.

34

E. Mechleri et al. / International Journal of Greenhouse Gas Control 59 (2017) 24–39

Fig. 14. Cost comparison of the various modes for flexible operation for the SCPC. The grey column show the results for the exhaust gas venting scenario (EGV) for the constrained case (IDoC is considered as an end point equality constraint fixed to 90%), the red column is the varying solvent regeneration scenario (VSR) and the blue columns are the solvent storage scenarios for 10,000 m3 tanks with and without CAPEX. In all cases, we find that time varying solvent regeneration is the most profitable option for providing additional flexibility to the coal-fired power plant.

A key conclusion of our study so far is that none of the modes of flexible operation will compromise the carbon intensity of the power plant, and in fact have the potential to reduce it, depending on the control strategies employed. Further, the different options have the potential to enhance the profitability of the power plant, by allowing it to exploit price volatility in the electricity market.

3.1.6. Sensitivity analysis SCPC Traditionally, the main factors that affect the profits accrued by a power plant are the revenue from the increased power production during peak hours, the cost related to the carbon price and the fuel price. The increased deployment of intermittent renewable energy has two principle effects: to increase the volatility (or peakiness) of electricity market and to reduce the important of fossil fuel prices in setting wholesale electricity prices. As has already been observed in Europe, a high penetration of intermittent renewable energy has the potential to produce negative electricity prices in addition to very high electricity prices (Dowell and Staffell, 2016). It is therefore essential to explore the profit sensitivity to these price oscillations. In the ensuing sensitivity analysis, carbon prices vary from 0 to 150 £/tCO2 and the electricity price differential (difference between high and low electricity price (PD)) varies from negative (£−80/MWh) to extreme high values (£180/MWh) to take into account the potential volatility in a future electricity market. From Fig. 15, we can observe that for negative electricity price differential and up to £25/MWh, and regardless of the carbon price, there is a reduction in profit compared to the base scenario (PD = £45/MWh and CO2 price = £ 70/ton). As the electricity price differential increases then we observe an increase in profit which increases monotonically with the increase at the electricity price differential and carbon price. However, beyond a certain level of PD, the carbon price ceases to significantly influence the profitability. This trend is similar for all scenarios considered. We can also observe that the position of the “star”, which represents the base case can show the % difference in profit between the different scenarios (4.5%, 3.4% and 10.5% for the EGV, SS and VSR,

Fig. 15. Sensitivity analysis for the UK scenario with fuel price of £ 7.7/MWh. In this figure we illustrate the variation of CO2 prices and electricity price differential (PD) (difference between low and high electricity prices) as a function of the cumulative profit (k£) compared to the central scenario (PD = £ 45/MWh and CO2 price = £ 70/tonCO2 ). For high carbon prices and negative or low PD (less than £ 45/MWh) we observe negative profits. As the price differential increases then the gain increases and for high PD can reach more than £ 900k for the most profitable VSR scenario. The “star” represents the base case with CO2 price = £ 70 tonCO2 and PD = £ 45/MWh.

E. Mechleri et al. / International Journal of Greenhouse Gas Control 59 (2017) 24–39

Fig. 16. In this figure we illustrate impact that varying CO2 prices has on the cumulative profit. The results of this sensitivity analysis show that the carbon price is a significant influence on the total profit and can lead to 4% increase for low CO2 price to 16% profit decrease when the CO2 prices are increased to £ 200/tCO2 .

35

Fig. 17. Split fraction vs electricity prices for solvent storage scenario. The split fraction determines if the solvent is regenerated or stored. Negative split fraction shows regeneration during low electricity prices, while positive split fraction shows storage during low electricity prices. For intermediate prices (£ 70/ton CO2 ) we have ∼ 50 % regeneration

respectively). Moreover, for high electricity price differential, the increase in profit for the different scenarios becomes more obvious. It is, of course, important to note that we did not give the CCS plant the option of either shutting down or storing its energy during periods of negative electricity prices. This was done to mediate the effect of paying the intermittent renewable energy sources to “spill” their power – in other words, thermal power plants will have to accept a loss during these periods. In the case of a shut down, this is likely to incur a cost of approximately £250k per shutdown cycle. Therefore the thermal plant would need to evaluate the trade-off associated with accepting this one time cost in addition to the additional maintenance costs and reduced equipment lifetimes associated with more frequent shut-down cycles vs. the prospect of running at a loss during periods of negative prices. 3.2. CCGT model Globally, natural gas is becoming more important as an energy vector. Importantly, CCGT plants are often employed in a mid-merit role in the electricity system where they provide a peaking and load-following service. Thus, they may be well suited to flexible operation when combined with CCS. Indeed, here, they may enjoy two advantages over their coal-fired counterparts, namely their greater efficiency and lower carbon intensity. Thus, in this section we present the results of our optimisation problem, the CCGT-CCS plant presented in Sections 2.3 and 2.4. 3.2.1. Load following As for the coal-fired power plant, our base-case is a simple loadfollowing operation. We simulate the same behaviour in terms of load factor, ramp rates and electricity prices for both scenarios. In Fig. 16 we present the cumulative profit accrued by the CCGT for a range of CO2 prices. It is evident from Fig. 16 that the CCGT profit is approximately 50% of that of the coal plant for the central scenario. The primary driver for this is that we have assumed UK-type gas prices of £24.53/MWh which are approximately 3 times greater than the coal price of £7.7/MWh. However, as the carbon intensity of the CCGT is substantially less than that of the coal-fired power plant, both the costs associated with residual CO2 emissions and CO2 transport and storage are less on a per MWh basis. However, it is interesting to compare the sensitivity to carbon price exhibited in Fig. 7 to that presented in Fig. 16. In the case of the SCPP, increasing the CO2 price from £70/tCO2 resulted in a 20% reduction in profit whereas the same change in CO2 prices reduced the CCGT profits by 16%. In other words, CCGTs would appear to be less sensitive to CO2 prices than coal fired power plants.

Fig. 18. Storage tanks profile during the day for different CO2 prices. As can be observed from this figure, when the electricity prices are lowest then the lean solvent storage tank is filling up (black line). During higher electricity prices the solvent bypasses the regeneration process and is sent to the rich solvent storage tank which is filling up (black dashed line).

3.2.2. Solvent storage The results of this optimisation problem are presented in Fig. 17. As can be observed, there are two periods of solvent storage during periods of high electricity prices (06:00–10:00 and 16:00–19:00) and four periods of solvent regeneration, with higher regeneration at the low electricity prices (£55/MWh) and lower regeneration at electricity prices of £100/MWh. The solvent storage profiles are presented in Fig. 18. As can be observed from Fig. 18, for a price of £70/tCO2 (central scenario), during periods of high electricity prices the lean tank (black line) is emptying while the rich tank (black dashed line) is filling, as there is no regeneration. The opposite can be observed for periods with low electricity prices. Similarly to the SCPC plant, we have calculated the cumulative profit for different storage tank sizes for 3 and 10 years payback period. The results are presented in Fig. 19. 3.2.3. Exhaust gas venting For the exhaust gas venting scenario, we have performed two simulations with CO2 prices of £10/tCO2 and £70/tCO2 , as it is illustrated in Fig. 20. For a CO2 price of £70/tCO2 , venting is not selected, and the IDoC is 89.9 % (end point constraint of the optimisation). For a CO2 price of £10/tCO2 , during periods of high electricity prices (£100/MWh) between 6:00–10:00 and 16:00–19:00, 33 % of the CO2 is vented decreasing the DoC at these times at 57 %. However during the other time periods the DoC reaches values of 99.9 %, so

36

E. Mechleri et al. / International Journal of Greenhouse Gas Control 59 (2017) 24–39

Fig. 19. Profit margin between the load following base case scenario and the solvent storage scenario with and without the capital cost of the storage tanks for different storage tank sizes. This figure shows a comparison between the base case scenario (black column) and the solvent storage scenario with capital cost with 3 years payback period (grey columns), 10 years payback period (blue columns) and without capital cost (red columns) for different solvent storage sizes. For the case where solvent storage cost is considered, as the capital cost of the storage tanks decreases with the size, the cumulative profit increases for the 3 years payback period, while the opposite is observed for 10 years payback period. However, after the capital cost has been paid off (red columns) there is an increase to the cumulative profit proportionally to the size of the tanks. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

Fig. 21. Solvent regeneration as a function of time and electricity price for CCGT. As opposed to operating at a constant loading the lean loading varies with variable electricity prices. When the electricity prices are high, the lean loading increases and the DoC drops to 84% and the opposite with low electricity prices. This operation allows CO2 to be accumulated in the solvent during peak electricity prices and the profit increases due to higher electricity revenue.

Fig. 22. Cumulative profit for the three scenarios for CCGT plant. The solid black column presents the base case scenario. The grey column presents the results for the exhaust gas venting scenario for £70/tCO2 . The red column is the variable solvent regeneration scenario (VSR) and the blue column is the solvent storage scenario with and without CAPEX. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

Fig. 20. Results for the exhaust gas venting scenario for CCGT. The black line is the vent fraction for a CO2 price of £70/tCO2 . CO2 is vented during high electricity prices and extra regeneration is performed during lower electricity prices to keep the carbon intensity in a specific level.

at the end of the day the IDoC is 89.9 % (end point constraint of the optimisation). For the SCPC plant the CO2 price that venting is selected is £ 10/tCO2 , similarly to the CCGT plant. However, the amount that is vented is more for the CCGT (33%) compared to 21% for the SCPC, which is explained by the higher carbon intensity of the SCPC plants compared to CCGT plants. 3.2.4. Variable solvent regeneration As for the SCPC case, here the strategy is to allow CO2 to accumulate in the working solvent during periods of high electricity prices, and more deeply regenerate it at other times. The results of this optimisation problem are presented in Fig. 21. As can be observed from this figure, the value of the lean loading varies with the different electricity prices. During periods with low electricity prices (£55/MWh), the lean loading is reduced from 0.23 to 0.19 for the first period and in the last period it is set back to the starting point of 0.23. During periods with high electricity prices the lean loading increases up to 0.25, to allow the plant to direct less steam to solvent regeneration and increase the profit from selling

the additional electricity. The DoC varies through the course of the simulation and reaches a minimum value of 84% during periods with high electricity prices. However, the IDoC at the end of the simulation is 90%. 3.2.5. Scenario comparison In this section we compare the performance of the three scenarios based on the cumulative profit attained by the plant for each mode of operation, taking into account that the flexible operation. As can be observed from Fig. 22, the exhaust gas venting scenario is 6% more profitable than the conventional scenario at a carbon price of £70/tonCO2 and this profit increases as the carbon price is reduced. The time varying solvent regeneration scenario is the most profitable scenario, by 13% more than the load following scenario. The solvent storage scenario, is 11% less profitable than the base case scenario, since the revenue associated with the electricity selling cannot outweigh the capital cost of the storage tanks. After the capital payback, the solvent storage scenario has a profit of £232k, which makes it 8% more profitable than the base case scenario. If we compare these results with those from the SCPC we observe the same trend: VSR>EGV>Base case >Solvent storage. The results for the carbon intensity and cumulative profit for the different scenarios are summarised in Table . Comparing the behaviour of the CCGT and SCPC plants, we observe that the % profits for all the scenarios are higher for the

E. Mechleri et al. / International Journal of Greenhouse Gas Control 59 (2017) 24–39

37

Fig. 23. Sensitivity analysis for the UK scenario with fuel price = 24.53 £/MWh. In this figure we illustrate the variation of CO2 prices and electricity price differential (PD) (difference between low and high electricity prices) as a function of the profit compared to the central scenario (PD = £45/MWh and CO2 price = £70/ton). The “star” represents the base case with CO2 price = £70/tonCO2 / and PD = £45/MWh. For high carbon prices and negative or low PD (up to 45) we observe negative profits. As the price differential increases then the profit increases. For the VSR scenario the profit increases more than £420k for high electricity prices.

Table 4 Integrated degree of capture (IDoC) and cumulative profit-CCGT. Scenario

IDoC (%)

Profit (k£)

Load following Exhaust gas venting (£70/tonCO2 ) Variable solvent regeneration Solvent storage (with CAPEX) Solvent storage (no CAPEX)

90 90 90 90 90

213 230 245 190 232

CCGT compared to the SCPC plant, which shows that the proposed flexible strategies can provide more profit for the CCGT plants. 3.2.6. Sensitivity analysis CCGT Similarly to the SCPC plant, we have performed a sensitivity analysis on carbon and electricity prices, in order to show their impact on the results. As it can be observed from Fig. 23, negative electricity prices give negative profits for any carbon price, whereas for electricity price differential of £180/MWh, the profit increases more than £420k for the VSR scenario. The trend between the different scenarios is similar and follows similarly monotonic behaviour, while for each case the profit difference is (6%, 13% and 8% for the EGV, VSR and SS scenarios, respectively) 4. Conclusions We have presented integrated models of both SCPC and CCGT power plants, each integrated with a post-combustion aminebased CO2 capture plant, modelled using the gCCS toolkit. We have

used this model to evaluate the profitability of a decarbonised power plant operating flexibly by considering four different scenarios: conventional load following scenario, solvent storage, flue gas venting and variable solvent regeneration. We have formulated an optimisation problem in order to evaluate the profitability and carbon intensity for each of the scenarios. For the load following scenario,the SCPC exhibits higher profits due to lower fuel prices compared to the CCGT. When comparing the SCPC with the CCGT plant for the USA, wherein gas prices of £7.17/MWh and coal prices of £ 3.5/MWh, consistent with current low US gas and coal prices, we observe that the profit of the CCGT increases to £ 435k as opposed to £ 533k for the SCPC plants, which indicates that the CCGT plants become competitive with coal plants, with the potential to displace them from their position in the merit order. Similar trends were observed for all the scenarios. For the exhaust gas venting scenario, the IDoC constraint is very important and strongly affects the plant’s profitability. For the CCGT plant the % venting is higher due to the lower carbon intensity of the plant. For the solvent storage scenario, where the capital cost of the storage tanks is considered, the revenue from the electricity prices cannot outweigh the capital cost. However, after the payback period the solvent storage scenario has 3.4% and 8% extra profit for the SCPC and CCGT, respectively for the 10,000 m3 case. The effect of the capital cost on the results is very important, and needs to be considered when designing a new integrated power-capture system with flexible operation. Moreover, the payback period is another factor that needs to be taken into account when considering the solvent storage option, since the results showed that for 10 years

38

E. Mechleri et al. / International Journal of Greenhouse Gas Control 59 (2017) 24–39

payback period the solvent storage scenario is more profitable than the base case scenario. The most profitable option for both the SCPC and CCGT plants is the variable solvent regeneration scenario for the different carbon and fuel prices. This flexible operation can increase the profitability of both the SCPC and CCGT plant by 10.5% and 13%, respectively compared to the base case, while maintaining the carbon intensity at 90% respectively. This study has examined the various flexible operational modes of an integrated PCPP and CCGT plant with post-combustion capture for different electricity price differential (PD) and different CO2 prices. The SCPC plants seems to be more sensitive to increases in CO2 prices than gas and the CCGT plants exhibits higher increase in profits than the PCPP plant, showing that flexible operation is more profitable for CCGT plants. We can therefore with confidence draw the conclusion that for both SCPC and CCGT plants the variable solvent regeneration scenario is the most profitable and least carbon intensive option. Acknowledgements The authors gratefully acknowledge the financial support of the grant EP/M001369/1 MESMERISE-CCS. References Abu-Zahra, M.R., Schneiders, L.H., Niederer, J.P., Feron, P.H., Versteeg, G.F., 2007. CO2 capture from power plants. Part I: a parametric study of the technical performance based on monoethanolamine. Int. J. Greenh. Gas Control 1, 37–46. Abu-Zahra, M.R., Niederer, J.P., Feron, P.H., Versteeg, G.F., 2007. CO2 capture from power plants. Part II: a parametric study of the economical performance based on mono-ethanolamine. Int. J. Greenh. Gas Control 1, 135–142. Adams, T., Dowell, N.M., 2016. Off-design point modelling of a 420 MW CCGT power plant integrated with an amine-based post-combustion CO2 capture and compression process. Appl. Energy 178, 681–702. Ameri, M., Hejazi, S., 2004. The study of capacity enhancement of the Chabahar gas turbine installation using an absorption chiller. Appl. Thermal Eng. 24, 59–68. Arce, A., Dowell, N.M., Shah, N., Vega, L., 2012. Flexible operation of solvent regeneration systems for CO2 capture processes using advanced control techniques: towards operational cost minimisation. Int. J. Greenh. Gas Control 11, 236–250. Boot-Handford, M.E., Abanades, J.C., Anthony, E.J., Blunt, M.J., Brandani, S., Mac Dowell, N., Fernandez, J.R., Ferrari, M.-C., Gross, R., Hallett, J.P., Haszeldine, R.S., Heptonstall, P., Lyngfelt, A., Makuch, Z., Mangano, E., Porter, R.T.J., Pourkashanian, M., Rochelle, G.T., Shah, N., Yao, J.G., Fennell, P.S., 2014. Carbon capture and storage update. Energy Environ. Sci. 7, 130–189. Brasington, R., 2012. Integration and operation of post-combustion capture system on coal-fired power generation: load following and peak power. Massachusetts Institute of Technology (MSc thesis). Bui, M., Gunawan, I., Verheyen, V., Feron, P., Meuleman, E., Adeloju, S., 2014. Dynamic modelling and optimisation of flexible operation in post-combustion CO2 capture plants – A review. Comput. Chem. Eng. 61, 245–265. Chemical Engineering Magazine, http://www.chemengonline.com. Chalmers, H., Leach, M., Lucquiaud, M., Gibbins, J., 2009. Valuing flexible operation of power plants with CO2 capture. Energy Procedia 1, 4289–4296. Chalmers, H., Lucquiaud, M., Gibbins, J., Leach, M., 2009. Flexible operation of coal fired power plants with postcombustion capture of carbon dioxide. J. Environ. Eng. 135, 449–458. Chalmers, H., Gibbins, J., Leach, M., 2011. Valuing power plant flexibility with CCS: the case of post-combustion capture retrofits. Mitig. Adapt. Strateg. Glob. Change 17 (6), 621–649. Cohen, S.M., Chalmers, H.L., Webber, M.E., King, C.W., 2011. Comparing post-combustion CO2 capture operation at retrofitted coal-fired power plants in the Texas and Great Britain electric grids. Environ. Res. Lett. 6. Cohen, S.M., Rochelle, G.T., Webber, M.E., 2011. Optimal operation of flexible post-combustion CO2 capture in response to volatile electricity prices. Energy Procedia 4, 2604–2611. Cohen, S.M., Rochelle, G.T., Webber, M.E., 2012. Optimizing post-combustion CO2 capture in response to volatile electricity prices. Int. J. Greenh. Gas Control 8, 180–195. Couper, J.R., Penney, W.R., Fair, J.R., Walas, S.M., 2010. Chemical Process Equipment, 3rd ed. Dave, N., Do, T., Palfreyman, D., Feron, P., 2011. Impact of liquid absorption process development on the costs of post-combustion capture in Australian coal-fired power stations. Chem. Eng. Res. Des. 89 (9), 1625–1638. Davison, J., 2007. Performance and costs of power plants with capture and storage of CO2 . Energy 32, 1163–1176. Davison, J., 2011. Flexible CCS plants – a key to near-zero emission electricity systems. Energy Procedia 4, 2548–2555. 2012. DECC fossil fuel price projections. Dept. of Energy and Climate Change (DECC), Available from: https://www.gov.uk/government/uploads/system/

uploads/attachmentdata/file/212521/130718decc-fossil-fuel-priceprojections.pdf. Delarue, E., Martens, P., D’haeseleer, W., 2012. Market opportunities for power plants with post-combustion carbon capture. Int. J. Greenh. Gas Control 6, 12–20. Dixon, T., Herzog, H., Twinning, S., Campbell, M., 2014. 12th International Conference on Greenhouse Gas Control Technologies, GHGT-12 Technology Innovation & Advancements for Shell Cansolv CO2 capture solvents. Energy Procedia 63, 801–807. Dixon, T., Herzog, H., Twinning, S., Singh, A., Stéphenne, K., 2014. 12th International Conference on Greenhouse Gas Control Technologies, GHGT-12 Shell Cansolv CO2 capture technology: achievement from first commercial plant. Energy Procedia 63, 1678–1685. Domenichini, R., Mancuso, L., Ferrari, N., Davison, J., 2013. Operating flexibility of power plants with carbon capture and storage (CCS). Energy Procedia 37, 2727–2737. Dowell, N.M., Shah, N., 2015. The multi-period optimisation of an amine-based CO2 capture process integrated with a super-critical coal-fired power station for flexible operation. Comput. Chem. Eng. 74, 169–183. Dowell, N.M., Staffell, I., 2016. The role of flexible CCS in the UK’s future energy system. Int. J. Greenh. Gas Control 48 (2), 327–344. Farouk, N., Sheng, L., Hayat, Q., 2013. Effect of ambient temperature on the performance of gas turbines power plant. Int. J. Comput. Sci. Issues 10, 439–442. GAMS Development Corporation, 2013. General Algebraic Modeling System (GAMS) Release 24.2.1, Washington, DC, USA. Gibbins, J., Crane, R., 2004. Scope for reductions in the cost of CO2 capture using flue gas scrubbing with amine solvents. Proc. Inst. Mech. Eng. A: J. Power Energy 218, 231–239. Grossman, I.E., Sargent, R.W., 1979. Optimum design of multipurpose chemical plants. Ind. Eng. Chem. Process Des. Dev. 18, 343–348. Haines, M., Davison, J., 2009. Designing carbon capture power plants to assist in meeting peak power demand. Energy Procedia 1, 1457–1464. Husebye, J., Anantharaman, R., Fleten, S.-E., 2011. Techno-economic assessment of flexible solvent regeneration & storage for base load coal-fired power generation with post combustion CO2 capture. Energy Procedia 4, 2612–2619. IEA Coal Industry Advisory Board, Power Generation from Coal Measuring and Reporting Efficiency Performance and CO2 Emissions. Operating Flexibility of Power Plants with CC, IEA Greenhouse Gas R&D Programme IEA GHG. Report 2012/06, IEA-GHG. Department of Engineering, Public Policy, 2016. Integrated Environmental Control Model (IECM), v 9. 2. 1. Khalilpour, R., Abbas, A., 2011. HEN optimization for efficient retrofitting of coal-fired power plants with post-combustion carbon capture. Int. J. Greenh. Gas Control 5, 189–199. Lucquiaud, M., Chalmers, H., Gibbins, J., 2007. Potential for flexible operation of pulverised coal power plants with CO2 capture. Energy Mater. 2, 175–180. Lucquiaud, M., Chalmers, H., Gibbins, J., 2009. Capture-ready supercritical coal-fired power plants and flexible post-combustion CO2 capture. Energy Procedia 1, 1411–1418. Mac Dowell, N., Shah, N., 2013. Identification of the cost-optimal degree of CO2 capture: an optimisation study using dynamic process models. Int. J. Greenh. Gas Control 13, 44–58. Mac Dowell, N., Shah, N., 2014. Optimisation of post-combustion CO2 capture for flexible operation. Energy Procedia 63, 1525–1535. Mac Dowell, N., Florin, N., Buchard, A., Hallett, J., Galindo, A., Jackson, G., Adjiman, C.S., Williams, C.K., Shah, N., Fennell, P., 2010. An overview of CO2 capture technologies. Energy Environ. Sci. 3, 1645–1669. 2006, January. National Grid, National Grid, Safety Data Sheet, Natural Gas. Oates, D.L., Versteeg, P., Hittinger, E., Jaramillo, P., 2014. Profitability of CCS with flue gas bypass and solvent storage. Int. J. Greenh. Gas Control 27, 279–288. Patino-Echeverri, D., Hoppock, D.C., 2012. Reducing the energy penalty costs of postcombustion CCS systems with amine-storage. Environ. Sci. Technol. 46, 1243–1252. PSE, 2015. gCCS overview, Retrieved from: http://www.psenterprise.com/power/ ccs/gccs.html (01.01.15). Qixin, C., Kang, C., Xia, Q., 2010. Modeling flexible operation mechanism of CO2 capture power plant and its effects on power-system operation. IEEE Trans. Energy Convers. 25, 853–861. Rao, A.B., Rubin, E.S., 2006. Identifying cost–effective CO2 control levels for amine-based CO2 capture systems. Ind. Eng. Chem. Res. 45, 2421–2429. Scoping Study on Operating Flexibility of Power Plants with CO2 Capture, IEA Greenhouse Gas R & D Programme. Report 2008/T1, IEA-GHG. Siemens, Gas turbines, http://www.energy.siemens.com/hq/en/fossil-powergeneration/gas-turbines/. van der Wijk, P.C., Brouwer, A.S., van den Broek, M., Slot, T., Stienstra, G., van der Veen, W., Faaij, A.P., 2014. Benefits of coal-fired power generation with flexible CCS in a future northwest European power system with large scale wind power. Int. J. Greenh. Gas Control 28, 216–233. Vatapoulos, K., Andrews, D., Carlsson, J., Papaioannou, I., Zubi, G., 2012. Study on the state of play of energy efficiency of heat and electricity production technologies, Technical Report. Joint Research Centre of the European Commission. Versteeg, P., Oates, D.L., Hittinger, E., Rubin, E.S., 2013. Cycling coal and natural gas-fired power plants with CCS. Energy Procedia 37, 2676–2683.

E. Mechleri et al. / International Journal of Greenhouse Gas Control 59 (2017) 24–39 Wiley, D.E., Ho, M.T., Donde, L., 2011. Technical and economic opportunities for flexible CO2 capture at Australian black coal fired power plants. Energy Procedia 4, 1893–1900. Ye, Q., Wang, X., Lu, Y., 2015. Screening and evaluation of novel biphasic solvents for energy-efficient post-combustion CO2 capture. Int. J. Greenh. Gas Control 39, 205–214.

39

Zaman, M., Lee, J.H., 2015. Optimization of the various modes of flexible operation for post-combustion CO2 capture plant. Comput. Chem. Eng. 75, 14–27. Ziaii, S., Cohen, S., Rochelle, G.T., Edgar, T.F., Webber, M.E., 2009. Dynamic operation of amine scrubbing in response to electricity demand and pricing. Energy Procedia 1, 4047–4053.