optimisation and evaluation of exible operation strategies
TRANSCRIPT
Optimisation and evaluation of �exible operation
strategies for coal- and gas-CCS power stations with a
multi-period design approach
Evgenia Mechleria,b, Paul S. Fennellc, Niall Mac Dowella,b,∗
aCentre for Process Systems Engineering, Imperial College London, South Kensington,
London SW7 2AZ UKbCentre for Environmental Policy, Imperial College London, South Kensington, London
SW7 1NA UKcDepartment of Chemical Engineering, Imperial College London, South Kensington, UK
Abstract
Thermal power plants are increasingly required to balance power grids by com-
pensating for the intermittent electricity supply from renewable energy resources.
As CO2 capture and storage is integrated with both coal- and gas-�red 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 pro�tability of these important assets. In this work, we present mod-
els of coal- and gas-�red 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 pro�t
maximising behaviour, a multi-period dynamic optimisation problem was for-
mulated and solved using these models. Four distinct scenarios were evaluated:
load following, solvent storage, exhaust gas by-pass and variable solvent regen-
eration (VSR). It was found that for both coal- and gas-�red power plants, the
VSR strategy is consistently the most pro�table 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
∗Corresponding authorEmail address: [email protected] (Niall Mac Dowell)
Preprint submitted to International Journal of Greenhouse Gas Control February 4, 2017
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 o�,
then the solvent storage scenario is 3.3% and 8% more pro�table than the base-
line for the pulverised coal and gas-�red power plants, respectively. Sensitivity
analyses showed that, for all strategies, the �exibility bene�t declined with re-
duced carbon and fuel prices, while a �peakier� electricity market, characteristic
of one with signi�cant quantities of intermittent renewables deployment, more
signi�cantly rewarded �exible operation.
Keywords: Flexible CCS, Dynamic optimisation, Dynamic process
modelling, multi-period design, gCCS
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 [1], [2]. Given the increasing penetration of intermittent
renewable electricity generation and the in�exible nature of traditional nuclear5
power generation 1, decarbonised power plants need to be designed for �exible
operation in order to be able to promptly respond to variation in electricity
demand [3], [4], [5] and to exploit the associated variation of electricity prices,
while maintaining the carbon intensity of the plant at low levels[6], [7], [8], [9].
Flexible capture can be achieved in a range of ways. At the level of an individual10
power plant, �exible 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 di�erent capture rates according to electricity
output requirements.
To the best of our knowledge, the concept of �exible operation, was �rst15
introduced by Gibbins and Crane [10] in 2004, noting that this study makes
1It is recognised that small modular reactors (SMRs) have the potential to o�er a �exible
form of nuclear power but at the time of writing, this is a relatively immature technology, and
has not been widely deployed.
2
reference to private communication with Prof Rochelle 2 on this subject in 2002.
In the 2004 study, the concepts of solvent storage and exhaust gas venting (or
capture bypass) were �rst introduced. It this study, it was concluded that
solvent storage had the potential to reduce electricity costs by 6 -7% and that20
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, several25
contributions focused on �exible 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 [5], [6], [7], [8],
[11],[12], [13], [14], [15], [16], [17] [18], [19], [20], [21], [22], [23], [24], [25], [26],30
[27], [28], [29], [30].
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 been35
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 - on40
the order of 2 - 5 hours [31]. 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 gas45
2Prof G. T. Rochell, U. Texas at Austin.
3
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 signi�cant losses of volatile organic compounds50
(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 pro�tability bene�t associated from venting the exhaust gas in55
the �rst 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 e�ect
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 that60
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 CO265
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 [11], identi�ed the
most cost-e�ective level of CO2 capture using the exhaust gas venting option.70
They concluded that the optimal CO2 capture level is dependent on plant size
and, if exhaust gas venting is considered, the cost-e�ectiveness 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 [6], [7], [14], [16], [18], [25], [32], [33]. In their work, Haines and75
Davidson [6], reviewed the ability of the main capture technologies (pre-, post-
4
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 �exibly.
A key conclusion of their analysis was that post-combustion systems o�ered the80
greatest possibility of operating �exibly. 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 [1], [2], it therefore stands to make the largest relative gain by reducing
this penalty at opportune times.85
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/tCO2captured [2], 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's90
Econamine or MHI's KS-1 solvents which typically use higher concentrations
of active ingredient (typically between 40 - 50%) and have an energy of re-
generation of 2.33 GJ/tCO2 [34], [35] 2.8 - 3.0 GJ/tCO2 and 2.5 -2.8 GJ/tCO2
respectively. Importantly, all of these solvents require a similar quality (temper-
ature) of steam for solvent regeneration, therefore a lower energy of regeneration95
leads to a reduced electricity output penalty. Moreover, solvents o�ering fur-
ther improvement are on the horizon, such as those reported by Ye et al. [36]
wherein materials requiring 2.0 GJ/tCO2at temperatures of 80 - 100◦C are re-
ported. We can readily evaluate the impact that these advanced solvents have
on process performance using the IECM tool [37]. IECM indicates that the100
higher heating value (HHV) e�ciency of an ultra supercritical (USC) power
plant is 42.83%. Its worth noting at this point that IECM is a relatively conser-
vative tool, and current USC plants in service today exhibit HHV e�ciencies 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 HHV105
e�ciencies in the region of 47 - 48% [? ], [38]. Then, applying amine-based CO2
capture to the IECM USC power plant reduces the HHV e�ciency to 29.03%.
5
Using Fluor's FG+ solvent (as described above) results in an HHV e�ciency
of 33.25%, MHI's KS-1 solvent gives an HHV e�ciency of 33.73% and �nally
Shell's Cansolv solvent gives an e�ciency of 34.33%. Similar calculations using110
an oxy-combustion option gives an HHV e�ciency 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 e�ciency of the existing US coal-fueled
electricity generating �eet is approximately 32%, and this can be substantially
lower in some parts of the world [38]. In other words, through the deploy-115
ment of state-of-the-art power and capture plant technology, it is conceivable
that decarbonised coal-�red power generation could be more e�cient than it is
today.
Con�rming the results presented by Rao and Rubin [11], stopping the sol-
vent regeneration during peak hours increases electricity generation by 20%.120
However, owing to the range of CO2 and electricity prices assumed in their
analysis, the additional revenue derived from selling the electricity was quite
small; between 0 - 4% of additional revenue above the baseline scenario. A key
limitation to the enhanced pro�tability that may be derived from �exible op-
eration is the compromise between peak electricity prices and CO2 prices - the125
peak electricity price needs to be signi�cant to o�set the additional cost associ-
ated 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 �rst
decade of the 21st century. In this period, the electricity system was composed
of nuclear, coal- and gas-�red power stations, and - in line with their analysis -130
the electricity market would not be characterised by excessive peakiness. Going
forward, as the UK experiences increased deployment of intermittent renewable
power [31], 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 emis-135
sion was also observed in the work of Ziaii et al.[14], who reported that �exibility
may improve the annual operating pro�ts; however, the balance of the electricity
and CO2 price needs to examined. In their work, Chalmers et al.[15], [16], pre-
6
sented an updated version of Gibbins and Crane's 2004 analysis and discussed
the �exible operation of coal �red power plants with post-combustion capture.140
They identi�ed 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 signi�cantly reduces the CO2 price
at which exhaust gas venting is economically attractive, repeating the earlier145
conclusions of Gibbins and Crane [10]. Whilst both Chalmers et al [15], [16]
and Gibbins and Crane [10] noted that solvent storage would come at an addi-
tional cost, a detailed design of the required solvent storage infrastructure was
not performed in either of their analyses. In their work, Cohen et al. [7], have
created optimisation and rule-based models within the General Algebraic Mod-150
eling System (GAMS) [39], to study pro�t-maximising operation of a coal �red
power plant with �exible CO2 capture with and without solvent storage under
varying degrees of electricity price foreknowledge. They concluded that the gas
venting option is unpro�table at high CO2 prices (above $70/tCO2), while sol-
vent storage maintains a 9-29% pro�t at any CO2 price, highlighting the value of155
�exible CCS. In their work, Chalmers et al. [18], performed a �rst 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 et al. [6], that �the revenue increase which
could be obtained in any one day by using solvent storage varies considerably160
depending greatly on the shape of the daily electricity price curve� and �could be
an attractive option in some electricity networks�. When discussing �exible op-
eration, 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 [16], [6], [21] [40]. In their work, Patiño-Echeverri et al. [32], presented an165
analysis on the di�erent electricity prices for an amine-storage to a CCS system.
They examined two di�erent 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
7
and either runs at 100% or 0% capacity. In the three-mode system, the solvent170
regeneration process might (1) run at 100% capacity to regenerate both the sol-
vent �owing 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 Patiño-Echeverri et al. [32]
stands out as one which does perform a detailed engineering design of the solvent175
storage tanks. Here, they assume that the additional volume of solvent will cost
between $629-711/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 [6], solvent degradation
e�ects would need to be properly taken into account. It may be that stainless180
steel storage tanks would we required, which could substantially increase the
associated capital cost. They found that the required price di�erential was in
fact a complex function of the cycling period, the capacity factor, the storage
size, and whether the plant is a retro�t or new. The required price di�erential
for two-mode operation ranged from $40-111/MWh for daily cycling and $92-185
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 di�erentials to
justify investment in solvent storage. This makes intuitive sense as the up-front
capital expenditure of post-combustion CCS is already signi�cant and current190
research e�orts are prioritising its reduction as a means to reduce the $/MWh
cost of CCS electricity.
In their work, Versteeg et al. [25], considered the pro�tability of coal and
natural gas-�red power plants with amine and ammonia post-combustion CO2
capture for variable electricity prices. They have concluded that the solvent stor-195
age option increased pro�tability at low carbon prices (¿40-60/tCO2). Husebe
et al. [21], developed an mixed integer linear programming (MILP) model to
identify the optimum operating strategy of a coal �red power plant with post-
combustion capture, and evaluated the potential value of �exible solvent regen-
eration and storage. The results showed that �exibility can lead to increased200
8
pro�ts, particularly in volatile electricity markets. Finally, a correlation be-
tween pro�tability and cyclical (weekly, seasonally, etc.) demand patterns was
observed. However, �exible operation is limited by case speci�c paramters, 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,205
Brasington et al. [40], presented an integrated coal �red 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 pro�tability of the plant does not increase for long periods of stor-
age (hours). They added that there might be a potential for short duration of210
solvent storage (i.e., less than 30 min), since this will not increase the opera-
tional complexity of the conventional coal �red power plant. The most recent
contributions on the �exible operation of CO2 capture systems are reported by
Van Der Wijk et al. [27], Oates et al.[5], Mac Dowell and Shah [8], Zaman
et al.[28], and Adams and Mac Dowell [41]. In their work, Van Der Wijk et215
al. [27] showed that the �exible options are not utilised. This is due to ei-
ther 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 �exible CCS plants can increase the reserve
capacity provision by 20-300% compared to non �exible plants. The paper of220
Oates et al.[5] is the �rst to discuss the �exibility options available to a natural
gas �red 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 �exible options. The
concluded that �exible CCS could result in an increased pro�t in the range 0-225
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 �exible 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 di�erent230
energy prices. Adams and Mac Dowell presented a detailed study of a CCGT
9
integrated with a CO2 capture and compression process. Here, the performance
of this system was evaluated under part-load conditions, with a key observa-
tion being that the whilst the cost structure of the integrated process remains
approximately constant for o�-design point operation, this will appreciably in-235
crease 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 optimi-
sation of �exible CCS systems; that of Mac Dowell and Shah [8] and Zaman
and Lee [28]. In their work, Zaman [28], presented an optimisation model for240
a post-combustion capture model for three �exible con�gurations: 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 [8], we presented a multi-period optimisation problem
to evaluate the pro�tability of a load following coal �red power plant integrated245
with a post-combustion capture plant for three di�erent operating strategies.
As an addition to the literature on this subject, this paper introduced the con-
cept of variable solvent regeneration (VSR) as another strategy of the �exible
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 prices250
and the subsequent regeneration of the solvent during periods of low electricity
prices o�ered substantially improved pro�tability 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 the255
CO2 produced by the power plant, it simply was not possible to vent a substan-
tial 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 pro�ts via �exible operation whilst maintaining a low average car-260
bon intensity (kgCO2/MWh) for both coal- and gas-�red power stations. A
multi-period, dynamic optimisation problem is formulated and implemented in
10
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 op-
tions for �exible operation of both coal- and gas-�red power plants: exhaust265
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 our270
approach for calculating electricity prices over the course of a 24 hour period, the
power plant and capture plant models and the optimisation problem. Section 3
presents the results and discussions for the di�erent scenarios and in section 4
we present the conclusions of our work.
2. Model development275
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 power generation [5], [42], [43],
[44], [45]. 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 �exible, load-following service. A model price pro�le for a
twenty four hour period was constructed by calculating the short run marginal
cost (SRMC) for di�erent types of plants (super-critical pulverised coal (SCPC),
combined cycle gas turbines (CCGT) and open cycle turbines (OCGT) using
3 Process Systems Enterprise. (2014). gCCS overview. Retrieved September, 9, 2014,
from: http://www.psenterprise.com/power/ccs/gccs.html.4Process Systems Enterprise, gPROMS, www.psenterprise.com/gproms, 1997-2015.
11
the following equation:
£SRMC
MWhr=
£MWhrFuel
nplant+ (£CO2
Tonne · CITonnesCO2
MWhr ) +£V arO&M +£CO2
T&S (1)
where ¿SRMC is the SRMC of the electricity generated by a given plant. In this
calculation the variable operating and maintenance costs (¿V arO&M ) and �xed cost
(¿CO2
T&S) for transport and storage are also considered. The data for this equation
were obtained from the Department of Energy and Climate Change (DECC) [46]280
for the fossil fuel prices and carbon prices and by the Joint Research Centre of
the European Commission [47] for the e�ciencies and carbon intensities. These
values are presented in Table 1.
12
Table 1: Values from DECC and the JRC for use in evaluating Equation 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 section 3.1.5 and 3.2.5) The e�ciency
values reported here are for plants with no CCS. In Equation 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 9.86 55 70 0.07 4.38 19.60 42.40
CCGT 24.53 60 70 0.04 3.06 19.60 62.62
OGCT 24.53 42 70 0.49 1.53 19.60 99.94
High scenario SCPC 13.89 55 105 0.07 4.72 32.20 61.30
CCGT 35.04 60 105 0.04 3.86 32.20 90.40
OGCT 35.04 42 105 0.49 1.93 32.20 144.95
Low scenario SCPC 7.17 55 35 0.07 4.02 8.20 27.15
CCGT 14.02 60 35 0.04 2.46 8.20 35.40
OGCT 14.02 42 35 0.49 1.23 8.20 55.04
13
Following our previous work [8], it is then assumed that over-night (o�-
peak) electricity prices will be set by SCPC plants, day-time prices will be set285
by CCGT plants with morning and evening peaks serviced by OCGT plants
[8]. This is illustrated in Figure 1, where the variable electricity pro�le for a
24h 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 o�-peak periods.
In Figure 1, the capacity factor of the load-following plant is also illustrated.290
Figure 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 di�erential (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 24h period. The
blue line represents the power plant capacity factor illustrating the load following pro�le 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 pro�t during these periods.
2.2. Pulverised coal-�red power plant
A model of a supercritical pulverised coal power plant (SCPC) was developed
using the SCPC model provided by the gCCS toolkit [48], illustrated in Figure
2. The inputs of the model are the nominal power output, inlet and outlet steam
14
conditions of the LP turbine, and �owrate of steam extracted as a function of295
the CO2 captured. Steam is extracted at the inlet of the LP turbine.
gCCS 1.1.0
1 Model descrip on
The PCPP_high_level model is used to simulate the retrofit of a Pulverised Coal Power Plant and its integra on witha capture plant. This model is not suitable for part-load calcula ons. The main inputs of the model are the nominalpower output, inlet and outlet steam condi ons of the LP turbine and flowrate of steam extracted. Steam is extractedat the inlet of the LP turbine. This opera on can be carried out while keeping the coal flowrate or the power outputof the plant. If the former op on is chosen, the model calculates the power penalty of the LP turbine. This penalty isconsidered the same as the total power plant penalty. If the second op on is selected, the model calculates the coalflowrate required to produce the steam needed to keep the power output andmeet the steam demand of the captureplant.
1.1 Ports
The PCPP_high_level model has three inlet and three outlet material ports. All ports are shown on the model icon infigure 1 and described briefly in table 1.
1
2
3 4
5
6
7
Figure 1: Ports of the PCPP_high_level model.
2 Model specifica ons
The default view of the specifica on dialog for the PCPP_high_level model is shown in figure ??.
Commercial in confidence©Process Systems Enterprise Ltd (2016)
PCPP_high_level – Page 2
Figure 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 �owrate of
the steam extracted.
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 �exibility. The e�ciency
of the standalone SCPC model is 44%, while integrated with the capture plant300
is 38 %. These values agree with the literature where 6% e�ciency penalty
is reported for post-combustion capture with MEA [49]. The nominal power
output along with the temperature and pressure of the steam at the extraction
point is speci�ed (T=506 K, P=3.9 bar), while the �owrate of the required
steam is calculated by the capture plant by setting a value at the capture rate,305
taken to be 90% as a base case in this work.
15
2.3. Combined cycle gas-�red power plant
A model of a Combined Cycle Gas Turbine plant was developed in order
to specify the �owrate and composition of the �ue gas stream supplied to the
capture plant as well as the �owrate and thermodynamic state of the steam310
provided for the regeneration of the solvent. This model is based on the CCGT
model provided by the gCCS toolkit [48] as it is illustrated in Figure 3.
gCCS 1.0
CCGT high level
CCGT Power Plant(high level)
Contents1 Model descrip on 2
1.1 Ports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2 Model specifica ons 2
This document is the property of Process Systems Enterprise Ltd.
No part of the material contained herein may be copied, distributed, retransmi ed ormodified, in any way without the prior wri en consent of Process Systems Enterprise Ltd.
Commercial in confidence©Process Systems Enterprise Ltd (2015)
CCGT high level – Page 1
1
2
3
4
5
Figure 3: This model is used to simulate a combined cycle gas-�red 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 e�ciency (library or
user-speci�ed) and the temperature an pressure of the stem extracted.
For the CCGT model, we have used the Siemens SGT5-4000F with the usual
con�guration of one gas turbine and three steam turbines. The electricity output
of the standalone power plant model is 421 MWe, while integrated with capture315
is 395 MWe; 73% of the total power comes from the gas turbine while the
remainder is provided by the steam turbines. The e�ciency of the standalone
CCGT model is 59%, based on the Siemens SGT5-4000F gas turbine [50], while
16
integrated with the capture plant is 53%. This is a 6% e�ciency penalty for
capture which agrees with the literature [49], [41].320
The gas turbine performance is changed by anything that a�ects the density
and/or mass �ow of the air intake to the compressor. As the ambient temper-
ature increases, the air density reduces and so the mass �ow to the compressor
is reduced, reducing the system output and e�ciency. The temperature e�ect
is dependent on the turbine model, as it depends on the cycle parameters and325
component e�ciencies. In order to account for the performance variation as a
function of the ambient temperature we specify the gas turbine e�ciency change
(-0.1%/◦C) and the exhaust �ow change (-0.45 kg/s/◦C) per degree ambient
temperature change, following previous studies [51], [52]. This is very impor-
tant, since in cases when the air temperature is very high, a pre-cooling system330
may need to be included. For example, for Australia, India or the GCC region
where air temperatures and humidities (and thus densities) are substantially
di�erent to those of the UK and can vary signi�cantly over the year. At the
extraction point, the temperature and pressure of the steam should be speci�ed
(T=500K, P=3.5bar). Similarly to the coal plant, this allows the calculation335
of the electricity output penalty associated with the operation of the capture
plant. The model also performs a check to determine whether the steam �ow
requested by the capture plant is available and �ags a warning if the �ow is
likely to be grater than 90 % of the total �ow in the turbine, thus protecting
the components of the CCGT from technical failure. Finally, the natural gas340
composition used in this work is: 95.2% CH4, 3.26% C2H6, 1.03% C3H8 and
0.51% C4H10 [53].
2.4. Capture plant model
The model of post-combustion CO2 process has been modelled using the
gCCS toolokit as illustrated in Figure 4.345
17
gCCS 1.0
Amine_capture_high_levelgCCS Model Documenta on
CO Capture2
(high level)
Contents1 Model descrip on 2
1.1 Inlets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.2 Outlets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2 Model Specifica ons 2
This document is the property of Process Systems Enterprise Ltd.
No part of the material contained herein may be copied, distributed, retransmi ed or modified, in any waywithout the prior consent of Process Systems Enterprise Ltd.
Commercial in confidence©Process Systems Enterprise Ltd (2015)
Amine_capture_high_level – Page 1
CO2 to stack CO2 to compression
Flue gas inletSteam inlet
Condensate outlet
Figure 4: This model is used to simulate a post-combustion CO2 capture plant. It has two
inlets, the �ue gas inlet and the steam inlet and three outlets; the treated �ue gas directed to
the stack, the CO2 to compression and the condensate outlet from the reboiler.
The speci�cations 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 di�erent
solvents (MEA (monoethanolamine), DEA (diethanolamine), DGA (Diglyco-
lamine), MDEA (methyldiethanolamine) we have chosen a 30% wt MEA solvent
with 0.23 lean loading and 0.5 rich loading. These speci�cations are in turn used350
to determine the required solvent �owrate.
An illustration of the integrated coal- and gas-�red power and CO2 capture
plants, are illustrated in Figures 5 and 6, respectively.
18
Figure 5: Integrated SCPC power plant and post combustion plant. There are three connec-
tion points between the power and capture plant. The exhaust gas �ow 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 �ue gas.
19
Figure 6: Integrated CCGT power plant and post combustion plant. There are three connec-
tion points between the power and capture plant. The exhaust gas �ow 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 �ue gas. In this model we can also see
an additional stack which is used for the exhaust gas venting scenario, where untreated �ue
gas is emitted to the atmosphere.
2.5. Optimisation problem
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 · (COGenerated2 − COEmitted2
COGenerated2
) (2)
IDoC =
∫ tf
t0
DoC dt (3)
where t0 and tf are the start and end periods of interest.355
The dynamic optimisation problem solved in this study is based upon the
theory of Grossman and Sargent, originally intended for the optimum design
of multi-purpose chemical plants [54]. The design and multi-period operation
20
of the decarbonised power plant can be represented by the system of mixed
di�erential 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 di�erential and algebraic variables in the model,
while x(t) are the time derivatives of the x(t). The control 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 �owrate360
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 end-point constraints.
These can be equality or inequality end point constraints of type:
w(tf ) = w∗, wmin ≤ w(tf ) ≤ wmax (5)
where w is one of the sytem 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 optimisa-365
tion 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:370
wmin ≤ w(t) ≤ wmax ∀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 �owrate between 0 and 1000 kg/sec (704 kg/sec are required for 90%
capture at full plant capacity) and the DoC to be in the range of 0-100%.
21
The constraints and decision variables for each of the four scenarios are375
presented in Table 2.
The dynamic optimisation seeks to determine the time variation of the con-
trol variables u(t) over the time horizon t ∈ [0, tf ] so as to maximise the �nal
value of a single variable z subject to constraints (5)-(7):
maxu(t),t∈[0,tf ]z(tf ) (7)
where z(tf ) is the short-run marginal cost (SRMC) pro�t of the plant.
Table 2: Constraints and decision variables for the di�erent scenarios. The �rst column are
the scenarios for both the coal and gas-�red power plants with capture, the second column are
the path constraints which must be satis�ed at all times during operation, the third column
are the end-point constraints which must be satis�ed at the end of operation and the last
column are the decision variables of the optimisation problem. Fsolv is the solvent �owrate
in kg/sec, CO2acc is the CO2 accumulated in kg/sec, Frbp is the by pass fraction, Frst is the
fraction to storage, and α, β, γ are the parameters used in the equation of the lean loading
as a function of t for the solvent regeneration scenario
Scenario Path constraints End-point constraints Decision
variables
Load following DoC=90% - -
Solvent storage 0 ≤ DoC ≤ 100 89.9 ≤ IDoC ≤ 90,
CO2acc=0
Frst
Exhaust gas venting 0 ≤ DoC ≤ 100
-1 ≤ Frbp ≤ 1
0 ≤ Fsolv ≤ 1000
89.9 ≤ IDoC ≤ 90 Frbp, Fsolv
Time varying solvent re-
generation
0 ≤ DoC ≤ 100 89.9 ≤ IDoC ≤ 90 α, β, γ
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 and380
the variable solvent regeneration scenarios and compare them by performing a
sensitivity analysis on carbon and electricity price di�erentials. We then present
the results for the same scenarios for the CCGT model.
22
For the solvent storage scenario to be pro�table, the bene�ts from the elec-
tricity price arbitrage need to exceed the signi�cant capital costs associated with385
building the solvent storage infrastructure [32]. In order to account for the cap-
ital cost in�uence on the storage scenario, we calculate the capital cost of the
storage tank as an additional marginal cost and we include it in the pro�t func-
tion. This is discussed in detail in section 3.1.2. We then perform a sensitivity
analysis on the electricity price di�erentials in order to explore the di�erence in390
low and high electricity prices at which this scenario is pro�table.
3.1. SCPC model
3.1.1. Load following
In the �rst scenario, the capture plant operates in accordance with the power
plant. As illustrated in Figure 1 the power plant ramps up and down during the395
period of one day with variable electricity prices, while the DoC and the lean
loading are kept at 90% and 0.23, respectively. In Figure 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 pro�t, as compared to the base case (black
line) in the �gure. For a capture rate kept at 90% and a very high carbon tax400
(¿200/tCO2) the decrease in the pro�t from the base case is 18%. This shows
that the carbon price is important to the overall pro�tability of this plant. A
further implication of this observation is that CO2 prices should be su�ciently
high to incentivise a high DoC, but no higher. Beyond a certain point, it simply
becomes punitive.405
23
Figure 7: Cumulative pro�t-SCPC-load following scenario. In this �gure we illustrate the
variation of CO2 prices as a function of the cumulative pro�t. The results of this sensitivity
analysis show that the carbon price has a signi�cant in�uence on the total pro�t and can lead
to 7% increase for low CO2 price to 18% pro�t decrease if the CO2 prices are more than 150%
of the base case (¿70/tCO2)
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 subsequently410
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 Figure 8.
24
Flue gas
Rich Pump
Heat exchanger
Reboiler
Condenser
CO2 to compression
AbsorberDesorber
Lean Solvent
Cleaned flue gas
Steam
RICH SOLVENT STORAGE
LEAN SOLVENT STORAGE
Figure 8: Solvent storage �owsheet. This �gure presents the option of solvent storage in
order to increase power output during peak times. The rich solvent �ow 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
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 accumulated should be zero. The additional equations
which describe this optimisation scenario are:
Fsolvreb = Fsolv · (1− SF ) (8)∫ tf
t0
CO2acc = Fsolvsto · θRichsto · wsolvent ·min(0, SF )dt (9)
where, Fsolvreb is the solvent storage directed to the reboiler, Fsolv is the to-415
tal solvent �owrate for 90% capture, Fsolvsto is the amount of solvent stored,
wsolvent is the mass fraction of the solvent, θRichsto is the 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.420
The results of this optimisation problem are presented in Figure 9. As
25
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.425
Figure 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 �lling up and the opposite during low electricity prices
Additional capital expenditure is required for the storage tanks which depend
on a number of factors, such as the volume of solvent 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% [16], and this is what430
we have used as a base case for this study. However, since the size of the storage
tanks a�ects the pro�t, we have performed a sensitivity analysis to show this
variation as presented in Figure 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 tanks435
26
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 [5]. We estimated the
capital costs of the tanks using Couper's Handbook [55] as described in equation
(10).
CCst = 0.72 · 1.218 · FM · exp[11.662− 0.6104 · (lnV ) + 0.04536 · (lnV )2] (10)
where CCst is the capital cost for the storage tank in ¿and FM is 3.4 for a440
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)
[56], as described in equation (11), and converted them to ¿GB, assuming a
currency conversion of $1.5/¿.
CCst2015 = CCst2003 ·CEPCI2015CEPCI2003
(11)
where CEPCI2015 was 550.4 and CEPCI2003 was 402.445
Given that the addition of the solvent storage infrastructure is analogous to
an e�ciency 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 equations (10) and (11)
we calculate the capital cost to be ¿715k. We then divide this cost by 1095 in450
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 the capital
cost into ¿/MWh. The SRMC of the plant is reformulated to include this cost:
£SRMC
MWhr=
£MWhrFuel
nplant+ (£CO2
Tonne · CITonnesCO2
MWhr ) +£V arO&M +£CO2
T&S +£SS
(12)
where ¿SS is the cost of the storage tank.
We have also considered the case where the capital cost has been paid o�455
(after 3 years) by solving the aforementioned optimisation problem without
considering the cost of the storage tanks.
27
In Figure 10, we can observe the results on the total pro�t of the solvent
storage scenario compared to the base case load following scenario and evaluate
it for di�erent solvent storage capacities, in each case accounting for the capital460
cost of the storage tanks. As can be observed, the cumulative pro�t 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 di�erence between465
the cumulative pro�t 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 solvent470
storage scenario is more pro�table than the base case and this pro�t increases
as the solvent storage size increases.
28
Figure 10: Pro�t margin between the load following base case scenario and the solvent storage
scenario with and without the capital cost of the storage tanks for di�erent storage tank
sizes. This �gure 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 di�erent 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 pro�t 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 pro�t increases with the
storage tank size. After the capital cost has been paid o� (red columns) there is an increase
to the cumulative pro�t proportionally to the size of the tanks.
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 pro�t of approximately 3.5% [28]. We475
have also examined this option and arrive at the same conclusion - we achieve
an increase in pro�tability 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 o�. This
pro�tability decreases as the size of the storage tanks decreases [28], which shows
that even if the capital cost for the large storage tank gives less pro�t during480
the payback period, after this time a larger storage tank gives more pro�t and
29
should be the one installed.
3.1.3. Exhaust gas venting
In the exhaust gas venting scenario, as illustrated in Figure 11 there is no
modi�cation to the plant's original design. In this scenario, a portion of the485
exhaust gases are re-directed upstream of the absorber to the stack and vented
directly to the atmosphere.
Figure 11: Exhaust gas venting �owsheet. The �ue gas from the power plant are re-directed to
the stack (black dashed line), mixed with the treated �ue gases and vented. No modi�cation
and extra investment is needed for this option
In this scenario, the power plant ramps up and down as illustrated in Figure
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 high490
electricity prices.
Once again, the problem is solved subject to the end-point constraint of
30
89.9 ≤ IDoC ≤ 90. In this case, the vent fraction is 21% for both periods of
high electricity prices as illustrated in Figure12. The reason for this is that the
end point constraint of IDoC set to 90% requires the model to select venting in495
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 the500
pro�t. 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/tCO2that venting was selected. This is a potentially interesting
result. First, we must recall that in this problem, we are solving based on
maximising the pro�t on a short run marginal cost basis. Whilst this is how505
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 su�cient to incentivise the construction of the
CCS power plant given the electricity prices discussed in this paper. However,510
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 signi�cant period for which the
power plant is operating on a SRMC basis.
31
Figure 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
3.1.4. Time varying solvent regeneration515
In this scenario, we use the working solvent as means to provide �exibility 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 o�-
peak periods. The lean loading is therefore no longer a �xed variable as in the
previous scenarios but can vary with time as expressed in the following equation:
θtLean = αt · t2 + βt · t+ γt (13)
where the lean loading can vary in di�erent time periods t, based on the quadratic
expression above by varying the parameters α, β and γ. 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 opti-
misation. The variable time is set to zero at the beginning of each new period.520
The results of this optimisation problem are presented in Figure 13.
32
Figure 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.
As it can be observed from Figure 13, rather than operating at a �xed
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 pro�tability by selling more electricity.525
From the same �gure 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 pro�t of this scenario is ¿503k, or 10.5%
more pro�table than the base case scenario.530
3.1.5. SCPC Scenario comparison
When comparing the various modes of �exible operation for the integrated
pulverised coal power plant and amine-based capture plant, there are several
33
things that need to be considered. In Figure, we present the cumulative costs
for the di�erent alternatives.535
Figure 14: Cost comparison of the various modes for �exible 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 �xed 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 �nd that time varying
solvent regeneration is the most pro�table option for providing additional �exibility to the
coal-�red power plant.
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 di�erence between storage being pro�table or not - particularly for
large volumes. In addition, the rate at which the storage volume was discounted
made a substantial di�erence to the pro�tability of this scenario. This is partic-540
ularly evident in the case of a storage tank of 10,000 m3, where an increase in
pro�t of 3.3% relative to the base-case was observed, one the cost had been paid
o�. Interestingly, this distinction was reduced as storage volume was decreased.
The data are presented graphically in Figure 14 and tabulated in Table 3.
34
Table 3: Integrated Degree of Capture (IDoC) and cumulative pro�t-SCPC
Scenario IDoC (%) Pro�t (k¿)
Load following 90 450
Exhaust gas venting (¿70 /tonCO2) 90 471
Variable solvent regeneration 90 503
Solvent storage (with CAPEX) 90 440
Solvent storage (no CAPEX) 90 465
A key conclusion of our study so far is that none of the modes of �exible545
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 di�erent options have the potential to enhance the pro�tability
of the power plant, by allowing it to exploit price volatility in the electricity
market.550
3.1.6. Sensitivity analysis SCPC
Traditionally, the main factors that a�ect the pro�ts 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 deploy-
ment of intermittent renewable energy has two principle e�ects: to increase the555
volatility (or peakiness) of electricity market and to reduce the important of
fossil fuel prices in setting wholesale electricity prices. As has already been ob-
served in Europe, a high penetration of intermittent renewable energy has the
potential to produce negative electricity prices in addition to very high electric-
ity prices [31]. It is therefore essential to explore the pro�t sensitivity to these560
price oscillations. In the ensuing sensitivity analysis, carbon prices vary from 0
to 150 ¿/tCO2and the electricity price di�erential (di�erence 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 Figure 15, we can observe that for negative electric-565
35
ity price di�erential and up to ¿25/MWh, and regardless of the carbon price,
there is a reduction in pro�t compared to the base scenario (PD= ¿45/MWh
and CO2 price =¿70/ton). As the electricity price di�erential increases then we
observe an increase in pro�t which increases monotonically with the increase at
the electricity price di�erential and carbon price. However, beyond a certain570
level of PD, the carbon price ceases to signi�cantly in�uence the pro�tability.
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 % di�erence
in pro�t between the di�erent scenarios (4.5%, 3.4% and 10.5% for the EGV,
SS and VSR, respectively). Moreover, for high electricity price di�erential, the575
increase in pro�t for the di�erent scenarios becomes more obvious.
36
Figure 15: Sensitivity analysis for the UK scenario with fuel price of ¿7.7/MWh. In this �gure
we illustrate the variation of CO2 prices and electricity price di�erential (PD) (di�erence
between low and high electricity prices) as a function of the cumulative pro�t (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 pro�ts. As the
price di�erential increases then the gain increases and for high PD can reach more than ¿900k
for the most pro�table VSR scenario. The �star" represents the base case with CO2 price=¿70
tonCO2and PD= ¿45/MWh.
It is, of course, important to note that we didn't 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 e�ect of paying the intermittent
renewable energy sources to "spill" their power - in other words, thermal power580
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 shut-down cycle.
Therefore the thermal plant would need to evaluate the trade-o� associated with
37
accepting this one time cost in addition to the additional maintenance costs and
reduced equipment lifetimes associated with more frequent shut-down cycles vs.585
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. Im-
portantly, CCGT plants are often employed in a mid-merit role in the electric-
ity system where they provide a peaking and load-following service. Thus, they590
may be well suited to �exible operation when combined with CCS. Indeed, here,
they may enjoy two advantages over their coal-�red counterparts, namely their
greater e�ciency 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 2.4.595
3.2.1. Load following
As for the coal-�red power plant, our base-case is a simple load-following
operation. We simulate the same behaviour in terms of load factor, ramp rates
and electricity prices for both scenarios. In Figure 16 we present the cumulative
pro�t accrued by the CCGT for a range of CO2 prices. It is evident from600
Figure 16 that the CCGT pro�t 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-�red power plant, both the605
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 �gure 7 to that presented in �gure 16. In the case of
the SCPP, increasing the CO2 price from ¿70/tCO2 resulted in a 20% reduction
in pro�t whereas the same change in CO2 prices reduced the CCGT pro�ts by610
16%. In other words, CCGTs would appear to be less sensitive to CO2 prices
than coal �red power plants.
38
Figure 16: In this �gure we illustrate impact that varying CO2 prices has on the cumulative
pro�t. The results of this sensitivity analysis show that the carbon price is a signi�cant
in�uence on the total pro�t and can lead to 4% increase for low CO2 price to 16% pro�t
decrease when the CO2 prices are increased to ¿200/tCO2 .
3.2.2. Solvent storage
The results of this optimisation problem are presented in Figure 17. As
can be observed, there are two periods of solvent storage during periods of615
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.
39
Figure 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
The solvent storage pro�les are presented in 18. As can be observed from
Figure 18, for a price of ¿70/tCO2(central scenario), during periods of high620
electricity prices the lean tank (black line) is emptying while the rich tank
(black dashed line) is �lling, as there is no regeneration. The opposite can be
observed for periods with low electricity prices.
40
Figure 18: Storage tanks pro�le during the day for di�erent CO2 prices. As can be observed
from this �gure, when the electricity prices are lowest then the lean solvent storage tank is
�lling up (black line). During higher electricity prices the solvent bypasses the regenration
process and is sent to the rich solvent storage tank which is �lling up (black dashed line).
Similarly to the SCPC plant, we have calculated the cumulative pro�t for
di�erent storage tank sizes for 3 and 10 years payback period. The results are625
presented in �gure 19.
41
Figure 19: Pro�t margin between the load following base case scenario and the solvent storage
scenario with and without the capital cost of the storage tanks for di�erent storage tank
sizes. This �gure 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 di�erent
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 pro�t 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 o� (red columns) there is an increase to the cumulative pro�t
proportionally to the size of the tanks.
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 Figure 20.
For a CO2 price of ¿70/tCO2, venting is not selected, and the IDoC is 89.9630
% (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 at the end of the day the IDoC is 89.9 % (end point constraint of the635
optimisation). For the SCPC plant the CO2 price that venting is selected is
42
¿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.
Figure 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
speci�c level
3.2.4. Variable solvent regeneration640
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 regen-
erate it at other times. The results of this optimisation problem are presented
in Figure 21 .
43
Figure 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 pro�t increases due to higher electricity
revenue.
As can be observed from this �gure, the value of the lean loading varies645
with the di�erent electricity prices. During periods with low electricity prices
(¿55/MWh), the lean loading is reduced from 0.23 to 0.19 for the �rst 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 pro�t from selling650
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 on655
the cumulative pro�t attained by the plant for each mode of operation, taking
44
into account that the �exible operation.
Figure 22: Cumulative pro�t 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.
As can be observed from Figure 22, the exhaust gas venting scenario is 6 %
more pro�table than the conventional scenario at a carbon price of ¿70/tonCO2
and this pro�t increases as the carbon price is reduced. The time varying660
solvent regeneration scenario is the most pro�table scenario, by 13% more than
the load following scenario. The solvent storage scenario, is 11 % less pro�table
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 pro�t of ¿232k, which makes it 8665
% more pro�table 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 pro�t for
the di�erent scenarios are summarised in table 4.
45
Table 4: Integrated Degree of Capture (IDoC) and cumulative pro�t-CCGT
Scenario IDoC (%) Pro�t (k¿)
Load following 90 213
Exhaust gas venting (¿70 /tonCO2) 90 230
Variable solvent regeneration 90 245
Solvent storage (with CAPEX) 90 190
Solvent storage (no CAPEX) 90 232
Comparing the behaviour of the CCGT and SCPC plants, we observe that670
the % pro�ts for all the scenarios are higher for the CCGT compared to the
SCPC plant, which shows that the proposed �exible strategies can provide more
pro�t for the CCGT plants.
3.2.6. Sensitivity analysis CCGT
Similarly to the SCPC plant, we have performed a sensitivity analysis on675
carbon and electricity prices, in order to show their impact on the results. As it
can be observed from Figure 23, negative electricity prices give negative pro�ts
for any carbon price, whereas for electricity price di�erential of ¿180/MWh, the
pro�t increases more than ¿420k for the VSR scenario . The trend between the
di�erent scenarios is similar and follows similarly monotonic behaviour, while680
for each case the pro�t di�erence is (6%, 13% and 8% for the EGV, VSR and
SS scenarios, respectively)
46
Figure 23: Sensitivity analysis for the UK scenario with fuel price=24.53 £/MWh. In this
�gure we illustrate the variation of CO2 prices and electricity price di�erential (PD) (di�erence
between low and high electricity prices) as a function of the pro�t 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 pro�ts. As the price di�erential increases then the
pro�t increases. For the VSR scenario the pro�t increases more than ¿420k for high electricity
prices.
4. Conclusions
We have presented integrated models of both SCPC and CCGT power plants,
each integrated with a post-combustion amine-based CO2 capture plant, mod-685
elled using the gCCS toolkit. We have used this model to evaluate the pro�tabil-
ity of a decarbonised power plant operating �exibly by considering four di�erent
scenarios: conventional load following scenario, solvent storage, �ue gas venting
47
and variable solvent regeneration. We have formulated an optimisation prob-
lem in order to evaluate the pro�tability and carbon intensity for each of the690
scenarios. For the load following scenario,the SCPC exhibits higher pro�ts 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 pro�t of the CCGT increases to ¿435k as opposed to ¿533k for the695
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 a�ects the
plant's pro�tability. For the CCGT plant the % venting is higher due to the700
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 pro�t for the SCPC and CCGT,
respectively for the 10,000 m3 case. The e�ect of the capital cost on the results705
is very important, and needs to be considered when designing a new integrated
power-capture system with �exible 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 payback period the
solvent storage scenario is more pro�table than the base case scenario. The710
most pro�table option for both the SCPC and CCGT plants is the variable sol-
vent regeneration scenario for the di�erent carbon and fuel prices. This �exible
operation can increase the pro�tability 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 various715
�exible operational modes of an integrated PCPP and CCGT plant with post-
combustion capture for di�erent electricity price di�erential (PD) and di�erent
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 pro�ts than
48
the PCPP plant, showing that �exible operation is more pro�table for CCGT720
plants. We can therefore with con�dence draw the conclusion that for both
SCPC and CCGT plants the variable solvent regeneration scenario is the most
pro�table and least carbon intensive option.
5. Acknowledgements
The authors gratefully acknowledge the �nancial support of the grant EP/M001369/1725
MESMERISE-CCS.
References
[1] N. Mac Dowell, N. Florin, A. Buchard, J. Hallett, A. Galindo, G. Jackson,
C. S. Adjiman, C. K. Williams, N. Shah, P. Fennell, An overview of CO2
capture technologies, Energy and Environmental Science 3 (2010) 1645�730
1669.
[2] M. E. Boot-Handford, J. C. Abanades, E. J. Anthony, M. J. Blunt, S. Bran-
dani, N. Mac Dowell, J. R. Fernandez, M.-C. Ferrari, R. Gross, J. P. Hal-
lett, R. S. Haszeldine, P. Heptonstall, A. Lyngfelt, Z. Makuch, E. Mangano,
R. T. J. Porter, M. Pourkashanian, G. T. Rochelle, N. Shah, J. G. Yao, P. S.735
Fennell, Carbon capture and storage update, Energy and Environmental
Science 7 (2014) 130�189.
[3] M. Bui, I. Gunawan, V. Verheyen, P. Feron, E. Meuleman, S. Adeloju, Dy-
namic modelling and optimisation of �exible operation in post-combustion
CO2 capture plants - A review, Computers and Chemical Engineering 61740
(2014) 245 � 265.
[4] J. Davison, Flexible CCS plants - A key to near-zero emission electricity
systems, Energy Procedia 4 (2011) 2548 � 2555.
[5] D. L. Oates, P. Versteeg, E. Hittinger, P. Jaramillo, Pro�tability of CCS
with �ue gas bypass and solvent storage, International Journal of Green-745
house Gas Control 27 (2014) 279 � 288.
49
[6] M. Haines, J. Davison, Designing carbon capture power plants to assist in
meeting peak power demand, Energy Procedia 1 (2009) 1457 � 1464.
[7] S. M. Cohen, G. T. Rochelle, M. E. Webber, Optimal operation of �exi-
ble post-combustion CO2 capture in response to volatile electricity prices,750
Energy Procedia 4 (2011) 2604 � 2611.
[8] N. M. Dowell, N. Shah, The multi-period optimisation of an amine-based
CO2 capture process integrated with a super-critical coal-�red power sta-
tion for �exible operation, Computers and Chemical Engineering 74 (2015)
169 � 183.755
[9] M. Lucquiaud, H. Chalmers, J. Gibbins, Potential for �exible operation of
pulverised coal power plants with CO2 capture, Energy Materials 2 (2007)
175�180.
[10] J. Gibbins, R. Crane, Scope for reductions in the cost of CO2 capture using
�ue gas scrubbing with amine solvents, Proceedings of the Institution of760
Mechanical Engineers, Part A: Journal of Power and Energy 218 (2004)
231�239.
[11] A. B. Rao, , E. S. Rubin, Identifying Cost-E�ective CO2 Control Levels for
Amine-Based CO2 Capture Systems, Industrial & Engineering Chemistry
Research 45 (2006) 2421�2429.765
[12] Scoping Study on Operating Flexibility of Power Plants with CO2 Capture,
IEA Greenhouse Gas R & D Programme. Report 2008/T1, IEA-GHG.
[13] M. Lucquiaud, H. Chalmers, J. Gibbins, Capture-ready supercritical coal-
�red power plants and �exible post-combustion CO2 capture, Energy Pro-
cedia 1 (2009) 1411 � 1418.770
[14] S. Ziaii, S. Cohen, G. T. Rochelle, T. F. Edgar, M. E. Webber, Dynamic
operation of amine scrubbing in response to electricity demand and pricing,
Energy Procedia 1 (2009) 4047 � 4053.
50
[15] H. Chalmers, M. Leach, M. Lucquiaud, J. Gibbins, Valuing �exible opera-
tion of power plants with CO2 capture, Energy Procedia 1 (2009) 4289 �775
4296.
[16] H. Chalmers, M. Lucquiaud, J. Gibbins, M. Leach, Flexible Operation of
Coal Fired Power Plants with Postcombustion Capture of Carbon Dioxide,
Journal of Environmental Engineering 135 (2009) 449�458.
[17] C. Qixin, C. Kang, Q. Xia, Modeling Flexible Operation Mechanism of780
CO2 Capture Power Plant and Its E�ects on Power-System Operation,
IEEE Transactions on Energy Conversion 25 (2010) 853�861.
[18] H. Chalmers, J. Gibbins, M. Leach, Valuing power plant �exibility with
CCS: the case of post-combustion capture retro�ts, Mitigation and Adap-
tation Strategies for Global Change 17 (6) (2011) 621�649.785
[19] S. M. Cohen, H. L. Chalmers, M. E. Webber, C. W. King, Comparing post-
combustion CO2 capture operation at retro�tted coal-�red power plants in
the Texas and Great Britain electric grids, Environmental Research Letters
6.
[20] D. E. Wiley, M. T. Ho, L. Donde, Technical and economic opportunities for790
�exible CO2 capture at Australian black coal �red power plants, Energy
Procedia 4 (2011) 1893 � 1900.
[21] J. Husebye, R. Anantharaman, S.-E. Fleten, Techno-economic assessment
of �exible solvent regeneration & storage for base load coal-�red power
generation with post combustion CO2 capture, Energy Procedia 4 (2011)795
2612 � 2619.
[22] Operating Flexibility of Power Plants with CC, IEA Greenhouse Gas R&D
Programme IEA GHG. Report 2012/06, IEA-GHG.
[23] S. M. Cohen, G. T. Rochelle, M. E. Webber, Optimizing post-combustion
CO2 capture in response to volatile electricity prices, International Journal800
of Greenhouse Gas Control 8 (2012) 180 � 195.
51
[24] E. Delarue, P. Martens, W. D�haeseleer, Market opportunities for
power plants with post-combustion carbon capture, International Journal
of Greenhouse Gas Control 6 (2012) 12 � 20.
[25] P. Versteeg, D. L. Oates, E. Hittinger, E. S. Rubin, Cycling Coal and805
Natural Gas-�red Power Plants with CCS, Energy Procedia 37 (2013) 2676
� 2683.
[26] R. Domenichini, L. Mancuso, N. Ferrari, J. Davison, Operating Flexibility
of Power Plants with Carbon Capture and Storage (CCS), Energy Procedia
37 (2013) 2727 � 2737.810
[27] P. C. van der Wijk, A. S. Brouwer, M. van den Broek, T. Slot, G. Stienstra,
W. van der Veen, A. P. Faaij, Bene�ts of coal-�red power generation with
�exible CCS in a future northwest European power system with large scale
wind power, International Journal of Greenhouse Gas Control 28 (2014)
216 � 233.815
[28] M. Zaman, J. H. Lee, Optimization of the various modes of �exible op-
eration for post-combustion CO2 capture plant, Computers and Chemical
Engineering 75 (2015) 14 � 27.
[29] A. Arce, N. M. Dowell, N. Shah, L. Vega, Flexible operation of solvent
regeneration systems for CO2 capture processes using advanced control820
techniques: Towards operational cost minimisation, International Journal
of Greenhouse Gas Control 11 (2012) 236 � 250.
[30] N. Mac Dowell, N. Shah, Optimisation of Post-combustion CO2 capture
for Flexible Operation, Energy Procedia 63 (2014) 1525�1535.
[31] N. M. Dowell, I. Sta�ell, The role of �exible CCS in the UK's future energy825
system, International Journal of Greenhouse Gas Control 48, Part 2 (2016)
327 � 344.
52
[32] D. Patino-Echeverri, D. C. Hoppock, Reducing the Energy Penalty Costs of
Postcombustion CCS systems with Amine-Storage, Environmental Science
& Technology 46 (2012) 1243�1252.830
[33] N. Mac Dowell, N. Shah, Identi�cation of the cost-optimal degree of CO2
capture: An optimisation study using dynamic process models, Interna-
tional Journal of Greenhouse Gas Control 13 (2013) 44�58.
[34] T. Dixon, H. Herzog, S. Twinning, A. Singh, K. Stéphenne, 12th Inter-
national Conference on Greenhouse Gas Control Technologies, GHGT-12835
Shell Cansolv CO2 capture technology: Achievement from First Commer-
cial Plant, Energy Procedia 63 (2014) 1678 � 1685.
[35] T. Dixon, H. Herzog, S. Twinning, M. Campbell, 12th International Con-
ference on Greenhouse Gas Control Technologies, GHGT-12 Technology
Innovation & Advancements for Shell Cansolv CO2 capture solvents, En-840
ergy Procedia 63 (2014) 801 � 807.
[36] Q. Ye, X. Wang, Y. Lu, Screening and evaluation of novel biphasic solvents
for energy-e�cient post-combustion CO2 capture, International Journal of
Greenhouse Gas Control 39 (2015) 205 � 214.
[37] Department of Engineering and Public Policy, Integrated Environmental845
Control Model (IECM),v 9.2.1, 2016.
[38] IEA Coal Industry Advisory Board, Power Generation from Coal Measur-
ing and Reporting E�ciency Performance and CO2 Emissions.
[39] GAMS Development Corporation. General Algebraic Modeling System
(GAMS) Release 24.2.1. Washington, DC, USA, 2013.850
[40] R. Brasington, Integration and operation of post-combustion capture sys-
tem on coal-�red power generation: Load following and peak power., MSc
Thesis, Massachusetts Institute of Technology.
53
[41] T. Adams, N. M. Dowell, O�-design point modelling of a 420 MW CCGT
power plant integrated with an amine-based post-combustion CO2 capture855
and compression process, Applied Energy 178 (2016) 681�702.
[42] M. R. Abu-Zahra, L. H. Schneiders, J. P. Niederer, P. H. Feron, G. F.
Versteeg, CO2 capture from power plants: Part I. A parametric study
of the technical performance based on monoethanolamine, International
Journal of Greenhouse Gas Control 1 (2007) 37 � 46.860
[43] M. R. Abu-Zahra, J. P. Niederer, P. H. Feron, G. F. Versteeg, CO2 capture
from power plants: Part II. A parametric study of the economical perfor-
mance based on mono-ethanolamine, International Journal of Greenhouse
Gas Control 1 (2007) 135 � 142.
[44] N. Dave, T. Do, D. Palfreyman, P. Feron, Impact of liquid absorption865
process development on the costs of post-combustion capture in Australian
coal-�red power stations, Chemical Engineering Research and Design 89 (9)
(2011) 1625�1638.
[45] R. Khalilpour, A. Abbas, HEN optimization for e�cient retro�tting of
coal-�red power plants with post-combustion carbon capture, International870
Journal of Greenhouse Gas Control 5 (2011) 189�199.
[46] DECC, DECC fossil fuel price projections. Dept. of En-
ergy and Climate Change(DECC), 2012, Available from:
https://www.gov.uk/government/uploads/system/uploads/attachment
data/�le/212521/130718 decc-fossil-fuel-price-projections.pdf.875
[47] K. Vatapoulos, D. Andrews, J. Carlsson, I. Papaioannou, G. Zubi, Study
on the state of play of energy e�ciency of heat and electricity production
technologies, Technical Report. Joint Research Centre of the European
Commission, 2012.
[48] PSE, gCCS overview, Retrieved January, 1, 2015, from:880
http://www.psenterprise.com/power/ccs/gccs.html.
54
[49] J. Davison, Performance and costs of power plants with capture and storage
of CO2, Energy 32 (2007) 1163 � 1176.
[50] Siemens, Gas turbines, http://www.energy.siemens.com/hq/en/fossil-
power-generation/gas-turbines/.885
[51] N. Farouk, L. Sheng, Q. Hayat, E�ect of Ambient Temperature on the
Performance of Gas Turbines Power Plant, IJCSI International Journal of
Computer Science Issues 10 (2013) 439�442.
[52] M. Ameri, S. Hejazi, The study of capacity enhancement of the Chaba-
har gas turbine installation using an absorption chiller, Applied Thermal890
Engineering 24 (2004) 59 � 68.
[53] National Grid, Safety Data Sheet, Natural Gas, January 2006.
[54] I. E. Grossman, R. W. Sargent, Optimum design of multipurpose chemical
plants, Industrial and Engineering Chemical Process Design and Develop-
ment 18 (1979) 343�348.895
[55] J. R. Couper, W. R. Penney, J. R. Fair, S. M. Walas, Chemical Process
Equipment, 3rd edition.
[56] Chemical Engineering Magazine, www.chemengonline.com.
55