1 s2.0 S0306261916308625 Main

Applied Energy 178 (2016) 681–702
Contents lists available at ScienceDirect
Applied Energy
journal homepage: www.elsevier.com/locate/apenergy
Off-design point modelling of a 420 MW CCGT power plant integrated

with an amine-based post-combustion CO2 capture and compression
process
T. Adams a, N. Mac Dowell a,b,⇑
a
Centre for Process Systems Engineering, Imperial College London, South Kensington Campus, London SW7 2AZ, UK
b
Centre for Environmental Policy, Imperial College London, South Kensington Campus, London SW7 1NA, UK
h i g h l i g h t s
Multi-scale modelling of CCGT with CO2 capture and compression.

Evaluation of full- and part-load behaviour.
Impact of exhaust gas recycle on process performance quantified.
Financial evaluation performed for full- and part-load operation.
a r t i c l e i n f o a b s t r a c t
Article history: The use of natural gas for power generation is becoming increasingly important in many regions in the
Received 17 January 2016 world. Given that the combined cycle gas turbine (CCGT) power stations are lower in capital cost and
Received in revised form 6 June 2016 carbon intensity than their coal-fired counterparts, natural gas fired power stations are considered a vital
Accepted 18 June 2016
part of the transition to a low carbon economy. However, CCGTs are not themselves ‘‘low carbon” and in
Available online 29 June 2016
order to reach a carbon intensity of less than 50 kgCO2 /MWh, it will be necessary to decarbonise them via
CCS, with post-combustion CCS currently regarded as being a promising technology for this application.
Keywords:
In this study, we present a detailed model of a 420 MW triple-pressure reheat CCGT and evaluate its tech-
CCGT
Gas-fired power plant
nical and economic performance under full and part load conditions. We evaluate the technical perfor-
CO2 capture mance of our CCGT model by comparison to an equivalent model implemented in Thermoflow
CCS THERMOFLEX and observe agreement of power output and efficiency to within 4.1% and the temperature
SAFT-VR profile within the HRSG within 2.9%. We further integrate the CCGT with a dynamic model of an amine
Off-design point modelling based CCS process, and observe a reduction in the base plant efficiency from 51.84% at full-load and
50.23% at 60% load by 8.64% points at full-load and 7.93% points at 60% load. A core conclusion of this
paper is that CCGT power plants equipped with post-combustion CCS technologies are well suited to
dynamic operation, as might be required in an energy system characterised by high penetrations of
intermittent renewable power generation.
Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction However, fossil fuels are expected to play an important role in

the global energy system for the foreseeable future, with natural
The evidence base for anthropogenic CO2 emissions causing cli- gas possibly becoming increasingly important as part of a transi-
mate change is unequivocal [1]. CO2 emissions associated with the tion towards a low carbon energy economy [2]. One key advantage
combustion of fossil fuels are rising, both in absolute terms and as of power generation via combined cycle gas turbine (CCGT) is their
a proportion of total anthropogenic greenhouse gases (GHGs). high efficiency relative to their coal-fired counterparts. Whilst
modern CCGTs have an efficiency of approximately 51%, CCGTs
with an efficiency of approximately 60% are considered feasible
in the near-to-medium term [3]. That said, whilst natural gas is a
⇑ Corresponding author at: Centre for Process Systems Engineering, Imperial less carbon intense fuel than coal, producing in the range of
College London, South Kensington Campus, London SW7 2AZ, UK. 450 kgCO2 /MW h as opposed to 800–1000 kgCO2 /MW h [4] it is
E-mail address: niall@imperial.ac.uk (N. Mac Dowell).
http://dx.doi.org/10.1016/j.apenergy.2016.06.087
0306-2619/Ó 2016 Elsevier Ltd. All rights reserved.
682 T. Adams, N. Mac Dowell / Applied Energy 178 (2016) 681–702
Nomenclature
V specific volume (m3 kg1)

W_ power input/output (MW)
Abbreviations
A heat transfer area (m) W specific work input/output (kJ kmol1)
C heat capacity flow rate (kW K1) xd vapour quality
C heat capacity flow rate ratio y molar composition
C max maximum heat capacity flow rate (kW K1) c heat capacity ratio
C min minimum heat capacity flow rate (kW K1) Df ho standard heat of formation (kg kmol1)
cp specific heat capacity (kJ kmol1 K1) Dhwater;fg specific latent heat of vaporisation of water (kJ kg1)
CAP capital cost of the project DP pressure change (MPa)
CCGT combined cycle gas turbine DT AM arithmetic mean temperature difference (T)
CF capacity factor (average plant load) DT GM geometric mean temperature difference (T)
CRF capital recovery factor DT NM new mean temperature difference (T)
EGR exhaust gas recycle e heat exchanger effectiveness
f CO2 fraction of the CO2 in the flue gas that is captured f excess air ratio
FGC fixed geometry compressor g efficiency
g i;j value of variable i at load j tk;i stoichiometric coefficient relating the number of moles
h specific enthalpy (kJ kmol1) of each species, i, that is produced or consumed from
HP high pressure one mole of species, k
i interest rate (%) Uc ; Ut compressor and turbine constants (respectively)
IP intermediate pressure %AD percentage absolute deviation (%)
LCOE levelised cost of electricity %AAD percentage average absolute deviation (%)
LF load factor
LHV lower heating value Subscripts
LP low pressure c compressor
m _ mass flow rate (kg s1) calc as calculated from the gPROMS model
MEA monoethanolamine cw cooling water
n molar flow rate (kmol s1) fg flue gas
nLF number of load factor operating points isen isentropic
nv ar number of variables mech mechanical
NC number of components ss steady state
NG natural gas t turbine
NTU number of transfer units TF as calculated by the Thermoflow THERMOFLEX model
O&M operation and maintenance water,f water in the liquid phase
P pressure (MPa) water,fg water in its liquid-gas transition phase
Psat saturation pressure of water (MPa) water,g water in the gaseous phase
PCC post combustion capture
Q_ heat transfer rate (kW) Superscripts
r compression/pressure ratio design denotes values calculated at full-load operation of the
R ideal gas constant (kJ kmol1 K1) CCGT
s specific entropy (kJ kg1 K1) ideal as calculated under ideal isentropic conditions
ST steam turbine in denotes an inlet stream to a process unit
T temperature (T) n plant lifetime
T average temperature (T) out denotes an outlet stream from a process unit
T sat saturation temperature of water (T) a; b exponential constants
U overall heat transfer coefficient (kWm2K1) e-NTU as calculated by the e-NTU method
not in itself ‘‘low carbon” - defined by the Committee on Climate It has been noted by, e:g., Gerbelová et al. [9], that CCGT-CCS
Change1 to be less than 100 kgCO2 /MW h. Indeed, it has been noted plants have an increased exposure to fuel prices relative to coal-
that energy systems with large portfolios of CCGT plants, without fired power plants, CCGT-CCS plants are less costly per MW h of
CO2 capture and storage (CCS), cannot deliver sufficient reductions low carbon energy generated [10]. Thus, CCGT-CCS plants can be
in CO2 emissions in the medium to long term to meaningfully miti- deployed in response to a relatively low CO2 price [11,12,9]. Whilst
gate climate change [5–7]. Thus, for natural gas to play a meaningful novel technologies, such as molten carbonate fuel cells and cal-
role in providing low carbon energy, CCGT power plants will need to cium looping processes [13], have been proposed for decarbonising
be decarbonised via the addition of CCS technology. This is particu- CCGT power plants [14], the vast majority of current academic lit-
larly important for countries like the UK where we expect to see fur- erature focuses on the utilisation of post-combustion amine scrub-
ther deployment of unabated CCGTs over the coming decade. For this bing as the technology of choice for this purpose [15,16]. In this
reason, the development of design methodologies for the optimal context, an aqueous 30 wt% solution of monoethanolamine
retrofit of CCS technology to combined cycle gas turbine (CCGT) (MEA) is still considered the benchmark solvent for the majority
power plants, such as those proposed recently by Pan et al. [8] are of academic studies [17].
of increasing importance. Over the course of the 21st century, the energy system is
expected to integrate increasing quantities of intermittent renew-
able energy ([18,19]). In the absence of suitable solutions for grid
1
The Committee on Climate Change: https://www.theccc.org.uk/. scale, inter-seasonal energy storage, the flexible and dispatchable
T. Adams, N. Mac Dowell / Applied Energy 178 (2016) 681–702 683
nature of fossil-based power generation will become increasingly part of a CCS process in the event of part load operation of the
valued [19–22]. Thus, decarbonised CCGTs will be required to power plant leading to a substantially reduced flow of CO2 from
operate in a flexible load following manner – providing peaking the power plant. Importantly, it has been observed that the appli-
capacity in addition to their current role as mid-merit provider in cation of EGR to a CCGR would result in negligible variation in tur-
the energy system. This has been noted in a number of previous bine performance [35,36], implying that this could be readily
contributions to this literature, such as that of Ceccarelli et al. incorporated as part of a CCS-retrofit operation.
[16] wherein a study detailing the likely start-up/shut-down beha- In this study, we present a detailed mathematical model of a
viour of a CCGT-CCS plant was presented. For this reason, models 420 MW triple-pressure combined cycle gas turbine (CCGT) power
that describe the dynamic and off-design point of these systems plant integrated with a post-combustion CO2 capture process. We
will be of increasing value. evaluate the performance of the CCGT with exhaust gas recycle
As has been noted by Brouwer et al. [18] and Mac Dowell and (EGR) and CCS under range off-design point operational modes.
Staffell [19], increased penetration of intermittent renewable All modelling work in this study was performed using the gPROMS
power will tend to decrease the efficiency of thermal power plants, platform.
and will likely exert an upward influence on wholesale electricity The remainder of this paper is laid out as follows. First, mathe-
prices. However, it is also important that the potential role of matical descriptions of the individual sub-models, which comprise
CCS in the provision of baseload power generation should not be the CCGT and compression system, are presented. Then, the model
dismissed, and in regions with limited availability of renewable validation process is described and the adequacy of our model
energy resources, CCS can play a role in the provision of baseload across a range of operating points is quantified. We go on to
power [12]. Importantly, as has been noted by Kang et al. [23] describe the integration of the CCGT, CO2 capture and compression
the likely operating mode of the CCGT-CCS plant (base load or flex- system, and validate the technical and economic performance of
ible operation) will significantly influence the profitability of the our model by comparison with the current literature. Finally, we
plant. conclude a perspective on the implications of this work on the role
Building on this point and in light of the increasing role that that CCS-equipped CCGTs may play in the future energy system
natural gas is playing in the global energy system, Middleton and and some thoughts for future work.
Eccles [24] presented an insightful analysis highlighting the link
between variations in generating patterns of CCGT-CCS plants 2. Model presentation
and their CO2 emissions. They found that the prices required to
incentivise the capture of the CO2 was a function of the generating The physical properties package Multiflash 4.1 was used to pre-
patterns required of the plants – in other words, a single CO2 price dict the physical properties of the air, flue gas, water, steam and
(or tax) may not be sufficient to deliver low carbon electricity. For fuel streams in the CCGT and compression train models that were
this reason, advanced operating strategies for operation under created. The SAFT-VR equation-of-state [37,38] was used to
load-following conditions are important to maximise profitability describe the thermophysical properties of the MEA-solvent used
[25,26]. However, a recent study by Benato et al. [27], highlighted in the PCC plant model using the approach described in our previ-
the fact that increased frequency of power plant start-up/shut- ous work [39–41].
down cycles and ramping of the power plant can result in a reduc-
tion in the operational lifetime of the power plant owing to 2.1. CCGT power plant
thermo-mechanical fatigue, creep and corrosion. This serves to
underscore the requirement to develop detailed models describing A detailed steady-state model of a 420 MW triple-pressure
the behaviour of decarbonised CCGT power plants operating at a reheat gas-fired CCGT was created based on the parameters and
range of off-design point and part-load operating conditions. process configuration of a Thermoflow THERMOFLEX model of a
One of the key challenges in capturing CO2 from the exhaust gas triple-pressure reheat CCGT with a state-of-the-art Siemens
of a CCGT is the relatively dilute nature of the exhaust gas of a SGT5-4000F gas turbine. A process flow diagram of the standalone
CCGT relative to a coal-fired power plant. Where a coal-fired power CCGT model is shown in Fig. 1.
plant will typically have an exhaust gas with a CO2 concentration Our model has two operational modes: calibration and opera-
of 10–15 mol%, a CCGT will have an exhaust gas with a CO2 concen- tion. The model of the CCGT plant was first simulated in calibration
tration in the range of 3–5 mol%. Similarly, the volumetric flow of mode in order to size the units of the plant so that the outputs of
exhaust gas per MW h from a CCGT is greater than the correspond- the gPROMS model could be calibrated against those of the THER-
ing flow from a coal-fired power station, typically by a factor of MOFLEX model. This is, in effect, a model parameterisation step.
1.4–1.6 [28]. An unavoidable consequence of this is that the min- This allows us to assign values to parameters such as compressor
imum thermodynamic work required to perform this separation efficiency or accurately size heat exchangers. Subsequently, the
will inevitably increase from 7 kJ/molCO2 in the case of a coal operation mode model of the CCGT was simulated in which the
plant to 10 kJ/molCO2 in the case of the CCGT. A further challenge sizes of the process units were assigned to the values determined
associated with the application of amine-based post-combustion in the calibration mode model and the unit models were modified
CO2 capture to CCGTs is the relatively high O2 content in the in order to predict their behaviour at off-design conditions.
exhaust gas. Where in a coal-fired power plant, the exhaust gas
will typically contain 4–6 mol% O2, the exhaust gas of a CCGT will 2.1.1. Compressor
potentially contain up to 10–12 mol%. This is a problem, as O2 is 2.1.1.1. Calibration mode. A compressor model was used to calcu-
well known to degrade alkanolamines [29,30], thus adding to pro- late the unit’s power requirement and discharge temperature.
cess operating cost. Thus, exhaust gas recirculation (or recycling), The model assumes that the air behaves as an ideal gas, the com-
known as EGR, is considered as a useful method to address these pression is reversible and adiabatic and there is a constant mass
challenges [31] and reduce process costs [32]. EGR involves the flow rate through the unit. The power requirement of the compres-
recycling of a portion of the exhaust gas exiting the heat recovery sor was calculated using the following equations [42]:
steam generator (HRSG) to the compressor of the gas turbine [33]. in
The concept of EGR has previously been applied to gas compressors _ in ¼ nair W in
W ð1Þ
[34], and indeed may well be applied to compressors operating as 1000
LP Pump Flow Steam

Controller Condenser LP Turbine IP Turbine HP Turbine
IP Pump HP Pump
S
S S
LP SD IP SD HP SD
M
Flue gas
to stack
LPE LPB IPE IPB HPE1 IPS1 LPS HPE2 IPS2 HPB HPS1 RH1 HPS2 RH2 HPS3
Fuel
Fuel
Air Preheater
Combuson
Compressor Gas Turbine
Chamber
Equipment Key:
HRSG Units:
LPE – LP Economiser LPB – LP Evaporator IPE – IP Economiser IPB – IP Evaporator HPE1 – HP Economiser 1 IPS1 – IP Superheater 1
LPS – LP Superheater HPE2 – HP Economiser 2 IPS2 – IP Superheater 2 HPB – HP Evaporator HPS1 – HP Superheater 1 RH1 – Reheater 1
HPS2 – HP Superheater 2 RH2 – Reheater 2 HPS3 – HP Superheater 3 LP SD – LP Steam Drum IP SD – IP Steam Drum HP SD – HP Steam Drum
Miscellaneous units:
M – Stream Mixer (water/steam) S – Stream Splier (water/steam)
Fig. 1. Process flow diagram of the standalone CCGT power plant model.
W in;ideal In practice the air compressor will have variable inlet guide
W in ¼ ð2Þ
gc vanes (VIGVs) that allow for more efficient operation at part loads.
The following empirical correlation is therefore used to correct the
2 !c1 3
c outlet pressures predicted by the above law to the outlet pressures
c P out
W in;ideal ¼ Pin V in 4 air 15 ð3Þ from the compressor in the Siemens SGT5-4000F gas turbine in the
c 1 air air P inair THERMOFLEX model:

where the specific volume of the inlet air is: Pout out;FGC
air ¼ P air 0:36 nin
air þ 1:8316 ð7Þ
RT in Eqs. (1)–(5) are used again to determine the work requirements

air ¼
V in ð4Þ
air
Pin and the temperature of the outlet stream from the compressor. It
air
was observed from THERMOFLEX simulations that the value of gc
The temperature of the outlet stream from the compressor is does not vary by more than 4% between 40% and 100% load. There-
[43]: fore, a constant value of 0.86 was therefore assumed.
2 !c1 3
c
T in out
air 4 P air 2.1.2. Combustion chamber
T out ¼ 15 þ T in ð5Þ
air
gc P in
air
air
It is important to note that typical "natural gas" is not composed
of 100% methane [45], and could potentially contain as little as 89%
The isentropic efficiency of the compressor, gc , was assigned the CH4, with a typical value is approximately 97% CH4. Therefore, the
value of 0.88. following reactions were included in our model in order to provide
sufficient flexibility to consider the impact of a range of potential
2.1.1.2. Operation mode. In operation mode, the outlet pressure is natural gas compositions:
no longer assigned. Instead, Eq. (6) [44] is used to predict the outlet
CH4 þ 2O2 ! CO2 þ 2H2 O ð8Þ
pressure of a fixed geometry compressor (FGC) at off-design condi-
tions. We first calculate the value of the compressor constant, Uc , 7
by substituting in the values of the other parameters as obtained C2 H6 þ O2 ! 2CO2 þ 3H2 O ð9Þ
2
from the simulation of the compressor in calibration mode.
qffiffiffiffiffiffiffiffi2 C3 H8 þ 5O2 ! 3CO2 þ 4H2 O ð10Þ
Uc nin
air T in
air
Pout;FGC ¼ þ Pin ð6Þ 1
air
Pin
air H2 þ O2 ! H2 O ð11Þ
air 2
3 Table 1
H2 S þ O2 ! H2 O þ SO2 ð12Þ Standard heats of formation and lower heating values of the constituent species of the
2
air, fuel and flue gas components.
1 Species Standard heat of formation, Df h° Lower heating value, LHV

CO þ O2 ! CO2 ð13Þ (kJ kmol1) (MJ kmol1)
2
O2 0 0
N2 0 0
3 1 3 CO2 393,510 0
NH3 þ O2 ! N2 þ H2 O ð14Þ
4 2 2 H2O 241,826 0
SO2 296,810 0
The following matrix of coefficients, mk;i , relates the number of Ar 0 0
moles of each species, i, that is produced or consumed following CH4 74,870 800.992
the introduction of one mole of species, k, in the fuel to the com- C2H6 84,000 1425.96
C3H8 104,700 2041.6
bustion chamber. The fuel species, k, are either inert or undergo
H2 0 240.136
complete combustion. H2S 20,600 No data
CO 110,530 283.164
O2 N2 CO2 H2 O SO2 Ar NH3 45,940 No data
tO2 ;i ¼ ½1; 0; 0; 0; 0; 0
tN2 ;i ¼ ½0; 1; 0; 0; 0; 0 There will be a pressure drop in the air stream across the com-
bustion chamber of 3% of the compressor delivery pressure at all
tCO2 ;i ¼ ½0; 0; 1; 0; 0; 0
loads [43]:
tCH4 ;i ¼ ½2; 0; 1; 2; 0; 0
tC2 H6 ;i ¼ ½ 72 ; 0; 2; 3; 0; 0 Pout in
fg ¼ P air ð1 DP loss Þ ð19Þ
tC3 H8 ;i ¼ ½5; 0; 3; 4; 0; 0
DPloss ¼ 0:03 ð20Þ
tH2 ;i ¼ ½ 12 ; 0; 0; 1; 0; 0
tH2 S;i ¼ ½ 32 ; 0; 0; 1; 1; 0
2.1.3. Gas turbine
tCO;i ¼ ½ 12 ; 0; 1; 0; 0; 0
2.1.3.1. Calibration mode. The turbine model is based on the same
tNH2 ;i ¼ ½ 34 ; 1
2
; 0; 3
2
; 0; 0 thermodynamic principles and assumptions that were used to
The molar flow rate of each species in the flue gas, nout model the compressor in Section 2.1.1, and the same approach
fg;i , is calcu-
was taken during calibration and operation modes. The power out-
lated by mass balance:
put of the turbine is calculated using the following equations [42]:
X
NC fuel
nin
fg W out
_ out ¼
fg;i ¼ nair yair;i þ nfuel
nout fuel;k tk;i : ð15Þ
in in in
yin W ð21Þ
k¼1
1000
The total molar flow rate of the flue gas stream is given by: W out ¼ gt W out;ideal ð22Þ
X
NC fg 2 !c1 3
c
fg ¼
nout fg;i :
nout ð16Þ c in in 4
Pout
15
fg
W out;ideal ¼ P V ð23Þ
i¼1 c 1 fg fg Pinfg
The excess air ratio is calculated by comparing the number of
moles of O2 present in the air supplied to the net number of moles where the specific volume of the inlet flue gas is:
of O2 required for complete combustion of the fuel: RT in
fg
fg ¼
V in ð24Þ
nin in Pin
air yair;O2 fg
fair ¼ ð17Þ
X
NC fuel
The temperature of the outlet stream from the gas turbine is
nin
fuel yin t
fuel;k k;O2 [43]:
k¼1
2 !c1 3
c
The temperature of the flue gas leaving the combustion cham- in 4 P out
T out ¼ T in g
fg 5 ð25Þ
ber is calculated via an energy balance on the streams entering fg fg t T fg 1
Pin
fg
and leaving the unit, with the combustion process assumed to be
adiabatic: The isentropic efficiency of the gas turbine, gt , was assigned a value
" NC # of 0.9.
X fg
o out o
nout
fg fg;i Df hfg;i þ c p;fg T fg T
yout
2.1.3.2. Operation mode. In the operation mode gas turbine model,
i¼1
"NC # the outlet pressure of the turbine is related to the inlet conditions
X air
o in o through Stodola’s law as shown in Eq. (26). The value of the turbine
¼ nin
air air;i Df hair;i þ c p;air T air T
yin
i¼1
constant, Ut , must be calculated and this is done in the same man-
"NC # ner in which C c was calculated in Section 2.1.1.
X fuel
o
T in o 0 0 qffiffiffiffiffiffi12 1
þ nin
fuel fuel;k Df hfuel;k
yin þ cp;fuel fuel T ð18Þ
2 nin T in
k¼1 B @ fg fg
A C
Pout
fg ¼ P in
fg @ 1 A ð26Þ
The standard heats of formation of air, fuel and flue gas compo- Pinfg Ut
nents were taken from the NIST Webbook database and from Perry
[46]. For completeness, both heats of formation at 25 °C and lower The outlet pressure predicted by Stodola’s law is very sensitive to
heating values, LHV, of each component are reported in Table 1. changes of the temperature, pressure and flow rate of the inlet flue
gas to the turbine as indicated by the squared term in the right- 2.1.5.2. Operation mode. Once the heat transfer areas of each econ-
hand side of Eq. (26). For this reason, Stodola’s law may predict neg- omiser section have been set, the e-NTU method can be used to
ative outlet conditions depending on how the inlet conditions are predict the outlet temperatures of the water and flue gas streams.
varied relative to each other. Therefore, the outlet pressure of the The rate of heat transfer in each economiser section is given by the
turbine, Pout
fg , was set to a constant value of 0.1035 MPa at all loads
following equations [49]:
and the gas turbine model therefore sets the flow rate of air into the
compressor (which then determines the compressor discharge pres- Q_ ¼ eC min T in in
fg T water ð34Þ
sure and combustion chamber outlet pressure) that will produce
this value of P out
fg .
1 exp ½NTU ð1 C Þ
e¼ ð35Þ
Eqs. (21)–(25) are used again to determine the work output and 1 C exp ½1 C
the temperature of the outlet stream from the turbine. The THER-
MOFLEX model showed that the value of gt does not vary by more UA
NTU ¼ ð36Þ
than 2% between power plant load factor of 40% and 100% so that a C min
constant value of 0.89 could be assumed.
C min
C ¼ ð37Þ
2.1.4. HRSG (flue gas side) C max
The heat recovery steam generator (HRSG) is a distributed
model with 15 discretisation points where flue gas flows down a where the heat capacity flow rates of the water and flue streams are
series of economiser, evaporator and superheater tube sections. given as:
There will be a slight pressure drop as the flue gas flows through _ in
C water ¼ m ð38Þ
water cp;water
the HRSG and the flue gas exits at approximately 0.1033 MPa. A
simple linear correlation was implemented to produce a flue gas
C fg ¼ nin
fg c p;fg ð39Þ
with this outlet pressure:
The higher and lower of these two heat capacity flow rates can
Pin;stage
fg
1
0:1033
Pout;stage
fg
i
¼ Pin;stage
fg
i
ð27Þ be identified as:
number of stagesð15Þ
C max ¼ max C water ; C fg ð40Þ
The flue gas exhaust temperature is calculated by energy bal-
ance that considers the quantity of heat transferred from the gas
to the steam cycle along the HRSG. The system is assumed to be C min ¼ min C water ; C fg ð41Þ
adiabatic with perfect radial distribution of heat. The value of the overall heat transfer coefficient will change
in out_ during off-design operation. In 1995, Dechamps and co-workers
fg hfg ¼ nfg hfg þ Q
nin ð28Þ
out
[50] proposed a simple formula for relating the overall heat trans-
fer coefficient in the HRSG sections to the off-design gas-side
2.1.5. Economiser properties:
2.1.5.1. Calibration mode. A calibration mode economiser model !a !b
was created to size the heat transfer area of every economiser sec- design
nin
fg T
U¼U ð42Þ
tion in the HRSG. The model assumes counter-current heat nin;design
fg
T design
exchange, no radial temperature profiles in the tubes, no pressure
drop along the water side and the heat capacity of the water is This is used in its simplified form, as shown in Eq. (43), as the aver-
evaluated at the average temperature along the economiser sec- age flue gas temperatures across each economiser section do not
tion. The outlet water temperatures in each economiser section change drastically from their respective full-load values, thus the
are using the THERMOFLEX model at full load and heat exchange second quotient in Eq. (42) can be approximated as unity, reducing
areas sized accordingly. Economiser sections that send water the numerical complexity in solving this model.
directly to steam drums will be set to produce outlet water streams !a
with an approach temperature of 5 K (i.e. 5 K below the saturation design
nin
fg
U¼U ð43Þ
temperature at the operating pressure of the economiser section). nin;design
fg
The heat transfer rate required in each economiser section is there-
fore calculated as: The value of the exponential term a was assigned a value of a ¼ 0:7.
The outlet temperature of the flue gas is given as:
Q_ ¼ m
_ in out in
water c p;water T water T water ð29Þ
Q_
T out in
fg ¼ T fg ð44Þ
The heat transfer area of each economiser can then be calculated C fg
from [47]:
The temperature of the outlet water stream as predicted by the
Q_ ¼ UADT NM ð30Þ e-NTU equations is:
2 1 eNTU Q_
DT NM ¼ DT GM þ DT AM ð31Þ T out;
water ¼ T in
water þ ð45Þ
3 3 C water
The economiser sections normally operate with water exit tem-
ðDT GM Þ2 ¼ T in out out in
fg T water T fg T water ð32Þ peratures very close to the saturation temperature of water and the
e-NTU equations do not account for the possibility that partial

evaporation may occur and the temperature of the outlet water
2DT AM ¼ T in out out in
fg T water þ T fg T water ð33Þ eNTU
predicted by Eq. (45), T out;
water , may be higher than the saturation
sat
The value of the overall heat transfer coefficient, U, was taken as temperature, T water . Once the saturation temperature is reached,
0.375 kWm2K1 [48]. any additional heat transfer will go towards the latent heating of
2.1.6.2. Operation mode. In the operation mode, the perfect control

law equations are modified to account for cases where the inlet
water stream to the steam drum is partially evaporated. The heat
transfer rate required to completely convert the inlet water to sat-
urated steam to maintain the condition of steady-state is calcu-
lated using the following set of logical conditions which
considers the cases where (1) the inlet water is sub-cooled or (2)
it is partially evaporated:
d;water ¼ 0 then
If xin

_ _ in sat in
Q ss ¼ m water c p;water T water T water þ Dhwater;fg
in
ð51Þ
ElSE

Q_ ss ¼ m
_ in
water 1 xd;water Dhwater;fg
in
Fig. 2. Temperature profile graph illustrating the difference between the actual The required flow rate of water circulating through the evaporator
economiser outlet temperature and the temperature predicted by the e-NTU
tubes can be calculated using the obtained value of Q_ ss and the
method.
assumption that there will be perfect evaporation with no super-
heating of the steam in the evaporator tubes.
the water and no further temperature rise will occur until the full
latent heat of vaporisation of the water is delivered. This concept is Q_ ss ¼ m
_ out
water Dhwater;fg ð52Þ
illustrated Fig. 2.
The actual temperature, T out out
water , and vapour quality, xd;water , of the 2.1.7. Evaporator
outlet water is therefore determined by: 2.1.7.1. Calibration mode. This model is used to determine the heat
out;e-NTU transfer area of every evaporator section in the HRSG. As before,
If T water 6 T sat
water then
out;e-NTU
perfect evaporation in every evaporator section is assumed. This
T out
water ¼ T water in turn sets the heat transfer rate from the flue gas in the corre-
d;water ¼ 0
xout sponding stage of the HRSG. The operational assumptions for this
ELSE ð46; 47Þ model are the same as those for the economiser model.
The temperature of the outlet steam stream will therefore be
T out sat
water ¼ T water
equal to the saturation temperature and the heat transfer duty of
out;e-NTU
_ in
m water c p;water T water T sat
water
each evaporator section can be calculated as:
d;water ¼
xout
_ in
m water Dhwater;fg Q_ ¼ m
_ in
water Dhwater;fg ð53Þ
Whilst this model cannot account for cases where superheating
of the exit stream occurs, it is noted that this is not likely to occur The heat transfer area of each evaporator can then be calculated
under normal power plant operating conditions. from Eqs. (30)–(33) which were used to size the economiser sec-
tions. The value of the overall heat transfer coefficient, U, was taken
2.1.6. Steam drum as 0.7 kWm2K1 [48].
2.1.6.1. Calibration mode. The steam drum is an inherently dynamic
unit and requires a controller to set the flow rate of water circulat- 2.1.7.2. Operation mode. Once the heat transfer areas of each evap-
ing through the evaporator tubes in order to maintain a steady orator section have been set, the e-NTU method can be used to pre-
level of water in the drum. Our model assumes perfect control dict the heat transfer rate and outlet temperatures of the water and
and that the water level in the drum remains constant under all flue gas streams at off-design conditions [49]:
conditions. This means that the mass flow rate of the steam leaving
the drum must be the same as the flow rate of the water entering Q_ ¼ eC min T in in
fg T water ð54Þ
the drum:
_ out _ in
e ¼ 1 eNTU ð55Þ
m steam ¼ mwater ð48Þ
It is also assumed that the pressure inside the drum is equal to UA
NTU ¼ ð56Þ
the inlet water pressure and the contents of the drum are main- C min
tained at the saturation temperature of water at the drum’s oper-
where the heat capacity flow rate of the flue gas stream and C min are
ating pressure at all times – which means that the dryness
given as:
fractions of the outlet water and steam streams will be 0 and 1
respectively. C fg ¼ nin
fg c p;fg ð57Þ
The mass flow rate of water going to the evaporator, m _ out
water ,
must be sufficient to provide the energy needed to heat the steam C min ¼ C fg ð58Þ
drum’s inlet water stream to its saturation temperature and then
Eq. (43), with a value of the exponential term a of 0.7, will also be
completely evaporate it, Q_ ss . The mass flow rate of water to the
used to determine the value of U at off-design operation of the
evaporator is therefore set by the following perfect-control law:
evaporator sections.

Q_ ss ¼ m
_ in in sat in
water c p;water T water T water þ Dhwater;fg ð49Þ The outlet temperature of the flue gas is given as:
Q_
T out in
fg ¼ T fg ð59Þ
Q_ ss ¼ m
_ out
water Dhwater;fg ð50Þ C fg
The temperature of the outlet steam requires consideration of the

phase change boundary. If the value of Q_ is less than the amount
of heat required for perfect evaporation of the water, the water will
leave as a saturated water-steam mixture and its temperature will
be equal to the saturation temperature. If the value of Q_ is greater
than the amount of heat required for perfect evaporation, the water
will leave as superheated water stream are calculated via:
If Q_ 6 m
_ in
water Dhwater;fg then
T out sat
water ¼ T water
Q_
d;water ¼
xout
_ in
m water Dhwater;fg
ð60; 61Þ
ELSE
Q_ m_ in
water Dhwater;fg
T out sat
water ¼ T water þ
_
min
water c p;steam
d;water ¼ 1
xout
2.1.8. Superheater/reheater Fig. 3. Steam expansion scenarios in the steam turbine.
The same model is used for the superheater and reheater sec-
tions as they both describe the superheating of dry steam. The
2. Dry steam (point 1II) enters the turbine and the outlet stream
operation mode superheater/reheater model contains the same
leaves as a partially condensed water-steam mixture under
equations as the operation mode economiser model apart from
isentropic expansion (point 2sII) but as superheated steam
Eqs. (45)–(47). Unlike the economiser and evaporator sections
under actual expansion (point 2II).
there are no phase change boundaries to be considered at the out-
3. Dry steam (point 1III) enters the turbine and the outlet stream
let of the superheater/reheater sections and the temperature of the
leaves as superheated steam under both isentropic and actual
outlet steam is given as:
expansion (points 2sIII and 2III respectively).
Q_
T out in
water ¼ T water þ ð62Þ For isentropic expansion, the entropy of the stream leaving the tur-
C water
bine is equal to the entropy of the stream entering the turbine:
The value of the overall heat transfer coefficient, U, at full-load was
taken as 0.375 kWm2K1 [48]. Eq. (43), with a value of the expo- sout;ideal
water ¼ sin
water ð64Þ
nential term a of 0.7, will also be used to determine the value of U at
We know that the outlet stream will either be partially con-
off-design operation of the superheaters/reheaters.
densed or superheated and its properties must lie on the Pout
water iso-
2.1.9. Fuel preheater bar. The temperature and vapour quality of the outlet stream will
The natural gas feed is heated from its ambient temperature to be calculated by first assuming that the outlet stream is partially
approximately 210 °C using some of the water leaving the IP econo- condensed and then checking whether this assumption holds true.
miser in the HRSG. A calibration mode model was first implemented The dryness fraction of the outlet stream under conditions of isen-
to size the preheater to deliver outlet fuel and water temperatures tropic expansion may be first estimated as xout;1
d;water :
equal to the corresponding values in the full-load THERMOFLEX
sout;ideal
water ¼ swater;f T sat out
water ; P water
model. Eqs. (29)–(47) used in the calibration and operation mode sat sat
out out
economiser models are applicable here. The value of the overall heat þ xout;1
d;water swater;g T water ; P water swater;f T water ; P water
transfer coefficient, U, was taken as 0.3 kWm2K1, both at full- ð65Þ
and part-loads, which the average value reported for organic sol-
vents and water in shell and tube heat exchangers [48]. If xout;1
d;water > 1 this means that the steam is superheated. The follow-
ing IF/ELSE condition can be used to determine the temperature,
2.1.10. Steam turbine vapour quality and specific enthalpy of the outlet stream under con-
2.1.10.1. Calibration mode. The steam turbine model serves the pur- ditions of isentropic expansion:
pose of predicting the unit’s output power and the temperature,
pressure and vapour quality of the outlet steam. The outlet pres- If xout;1
d;water 6 1 then
out
sure of each steam turbine will be specified by the user to match T water ¼ T sat
out;ideal
water P water
the corresponding values in the full-load THERMOFLEX model.
The pressure ratio across each turbine is given as:
xout;ideal out;1
d;water ¼ xd;water
h
out;ideal
Pin hwater ¼ hwater;f T out;ideal out out;1 out;ideal out
water ; P water þ xd;water hwater;g T water ; P water
r t ¼ water ð63Þ i
Pout
water hwater;f T out;ideal out
water ; P water
There are three different scenarios that can occur during normal ELSE
operation of the turbine and these are illustrated in Fig. 3:
sout;ideal
water ¼ swater;g T out;ideal out
water ; P water
1. Dry steam (point 1I) enters the turbine and the outlet stream xout;ideal
d;water ¼ 1
leaves as a partially condensed water-steam mixture under
out;ideal
both isentropic and actual expansion (points 2sI and 2I hwater ¼ hwater;g T out;ideal out
water ; P water ð66; 67; 68Þ
respectively).
We can now use these values and the isentropic efficiency of the tropic efficiency of the HP turbine varied by 25.6% so the following
turbine to calculate the actual temperature, vapour quality and correlation was derived to relate the isentropic efficiency of the HP
enthalpy of the steam leaving the turbine. The actual enthalpy of turbine to the mass flow rate of steam flowing through it:
the stream leaving the turbine is calculated via:
gisen;HP ST ¼ 0:7388m_ in;HP
water
ST
þ 33:947 ð75Þ
in out
hwater hwater
gisen ¼ in out;ideal
ð69Þ
hwater hwater 2.1.11. Water pump
A pump model was created to predict the power requirements
The isentropic efficiency of each steam turbine was set to the
of increasing the pressure of water in the LP, IP and HP circuits of
corresponding value given in the full-load THERMOFLEX model
the steam cycle. The same model is suitable for both the initial
(nisen;HP ST = 0.86, nisen;IP ST = 0.90, gisen;LP ST = 0.91).
design and part-load operation of the plant. The power require-
We first assume that the actual steam leaving the turbine is par-
ment of each pump is given as [51]:
tially condensed and the dryness fraction of the outlet stream
1 1
under conditions of actual expansion may be first estimated _ in ¼ m
_ in V in out in
W water P water P water ð76Þ
as,xout;2
d;water :
water
gisen gmech
out Çengel [51] has indicated that the temperature of the outlet
hwater ¼ hwater;f T sat out
water ; P water water stream will be no more than 1–2 K greater than the temper-
sat out
sat out

þ xout;2
d;water hwater;g T water ; P water swater;f T water ; P water ð70Þ ature of the inlet stream and so the outlet is assumed to be equal to
the inlet temperature.
The actual temperature and vapour quality of the stream leav-
ing the turbine can be calculated using the following IF/ELSE 2.1.12. Steam condenser
condition: The calibration mode condenser model is used to determine the
heat transfer area and cooling water flow rate required to perfectly
If xout;2
d;water 6 1 then condense the steam leaving the LP steam turbine (with no subcool-
out
T water ¼ T sat
out
water P water ing). The same approaches used in the calibration and operation
out;2 mode economiser models can be applied here. However, the cool-
d;water ¼ xd;water
xout
ð71; 72Þ ing water outlet temperature and specified hot side temperature
ELSE difference, DT hot side , (10 K for a reasonable heat transfer area)
out
hwater ¼ hwater;g T out out
water ; P water
set the minimum pressure that condensation of the LP turbine out-
let steam can condense at (and hence the minimum discharge
d;water ¼ 1
xout
pressure from the LP turbine). The minimum condensation tem-
The power output of each steam turbine is given as: perature and pressure of the outlet stream from the LP turbine
are given respectively as:
_ t ¼g in out
W mech hwater hwater ð73Þ
T min out
water ¼ T cw þ DT hot side ð77Þ
The mechanical efficiency of each steam turbine was set to the cor-
responding value given in the full-load THERMOFLEX model Pmin
water ¼ P sat min
water T water ð78Þ
(gmech ¼ 0:997 for the HP, IP and LP steam turbines).
The value of the overall heat transfer coefficient, U, was taken as
1.25 kWm2K1 which the average value reported for water-
2.1.10.2. Operation mode. The operation mode steam turbine model cooled condensers with aqueous condensates [48].
will contain the same equations as the calibration mode model. The model is also used to determine the electrical power required
However, new equations are required as the outlet pressures and to return the hot cooling water leaving the condenser to its original
isentropic and mechanical efficiencies of the turbine will vary with inlet temperature. A simple correlation between the cooling water
load factor. power requirements (in MW) and the mass flow rate of cooling
It was first attempted to use Stodola’s law to predict the outlet water was made based on data from the THERMOFLEX model:
pressure of each steam turbine. However, it was found that the law
was very sensitive to changes in the temperature, pressure and _ cw
_ in ¼ ð1640=3580Þ m
W ð79Þ
flow rate of the inlet stream and this lead to the prediction of outlet 1000
pressures that were significantly different to the ones in the THER-
MOFLEX model. Thus, in the interest of computational stability, the 2.1.13. CCGT inlet stream properties and control variables
outlet pressures of the IP and LP steam turbines were set to con- The CCGT model is controlled by manipulating the fuel flow rate
stant values (Pout;IP
water
ST
¼ 0:36 MPa and P out;LP
water
ST
¼ 0:008149 MPa). to the combustion chamber, the total water flow rate in the steam
The outlet pressure of the IP steam turbine was chosen to be suffi- cycle, the water flow rates going to the LP and IP sections of the
ciently high to be suitable to provide steam to the reboiler in the steam cycle and to the fuel preheater and the outlet pressures of
capture plant. A correlation between the pressure ratio across the the LP, IP and HP water pumps. Data from the THERMOFLEX model
HP turbine and the mass flow rate of steam through this unit was used to develop simple correlations between these control
was created based on THERMOFLEX data: variables and the plant’s load factor (LF):
2 Fuel flow rate (kg s1):
r t;HP ST ¼ 0:001 m_ in;HP
water
ST
_ in;HP
0:1901 m water
ST
þ 12:885 ð74Þ _ fuel ¼ 0:1385LF þ 2:3617
m ð80Þ
The values of the mechanical efficiencies of the three steam tur- Total water flow rate in the steam cycle (kg s1):
bines and the isentropic efficiencies of the LP and IP turbines in the _ water;total ¼ 0:5928LF þ 41:328
m ð81Þ
THERMOFLEX model did not vary by more than 0.4% and 0.92%
respectively between 40% and 100% load. Hence these values were LP splitter outlet 2 flow rate (water diverted to the LP section)
taken to be constant at all load factors. On the other hand, the isen- (kg s1):
_ water;LP ¼ 5 105 LF 3 þ 0:0117LF 2 0:8329LF

m Table 3
Fuel stream properties.
þ 0:0151e 0:04022LF
þ 26:631 ð82Þ
Molar composition (%) Temperature Pressure
IP splitter outlet 2 flow rate (water diverted to the IP section of (K) (MPa)
CH4 C2H6 C2H6 CO2 N2
the plant) (kg s1):
gPROMS and 93.1 3.2 1.1 1.0 1.6 298.1 2.44
_ water;IP ¼ 0:0017LF 2 0:1005LF þ 29:445
m ð83Þ THERMOFLEX values
Flow rate of water diverted to the fuel preheater (kg s1):

_ water;FP ¼ 0:0008LF 2 þ 0:2274LF þ 3:2886
m ð84Þ model. The capture plant model that was implemented is limited
to operating only with an inlet exhaust gas from the CCGT with a
LP pump outlet pressure (MPa): CO2 composition of 7 mol%. The absorption and desorption col-
Pout ¼ 6 106 LF 2 þ 0:0015LF þ 0:20868 ð85Þ umns were sized for 70% of flooding at 100% of power plant name-
LP pump
plate capacity. All heat exchangers were sized for a 10 K approach
IP pump outlet pressure (MPa): temperature; the reboiler was sized for a residence time of 100 s in
the first instance and thereafter controlled to operate at a pressure
Pout
IP pump ¼ 0:02012LF þ 1:3582 ð86Þ
of approximately 0.19 MPa via monitoring of the solvent lean load-
HP pump outlet pressure (MPa): ing and manipulation of the steam supply. Finally, the overhead
condenser was controlled to operate at 308 K. The mathematical
Pout
HP pump ¼ 0:01354LF þ 13:15 ð87Þ models of the units in the PCC plant are not presented in this paper
but the key equipment dimensions and operating parameters are
The use of these correlations in the gPROMS model allow the
reported in Table 4.
user to control the plant by simply specifying the load factor and
It is observed that the dimensions presented in Table 4 are in
further allow an equitable comparison of both models. The air
good agreement with other studies in the literature, e.g., the recent
and fuel inlet streams in the gPROMS model were set to have the
study of Rezazadeh et al. [20].
same compositions, temperatures and pressures as the corre-
sponding streams in the THERMOFLEX model and are presented
in Tables 2 and 3 respectively. 2.4. CO2 Compression train
A five-stage intercooled compression train model was created

2.2. CCGT modifications for PCC integration
and implemented in gPROMS. The configuration follows the sug-
gestions of Biliyok and Yeung [54] and Moullec and Kanniche
The standalone CCGT plant model must be modified to include a
[55]. A process flow diagram of the compression train model is
steam extraction scheme and an exhaust gas recycling (EGR)
shown in Fig. 6.
scheme in order to be integrated with the capture plant. The pro-
The final discharge pressure will be 110 bar and the five com-
cess flow diagram of the modified CCGT is shown in Fig. 4.
pression stages will operate with equal pressure ratios and the
The purpose of the steam extraction scheme is to supply steam
compression ratio in each stage, rp;stage , will be [42]:
to the reboiler of the desorber in the capture plant. The steam is
extracted from the LP/IP cross-over and is then sent through a 1=no: of stages
Final pressure ð110 barÞ
desuperheater which uses the condensate leaving the reboiler to rp;stage ¼ ð88Þ
Pressure into first stage
remove most of the superheat from the steam. The remaining
superheat is removed using a trim cooler so that the steam enter- The intercoolers in every stage of the compression train will
ing the reboiler is always at its saturation temperature. The con- cool the gas stream down to 40 °C using cooling water supplied
densate is re-introduced into the steam cycle after the LP pump at 17 °C. The first three stages contain water knock-out drums
as the operating pressure in this section is always high enough to where condensed water is drawn off from the gas stream. The
ensure that the condensate will be in its sub-cooled state. A con- fourth stage contains an adsorptive drying unit which is modelled
densate pump is needed to increase the pressure of the condensate as a black-box that simply reduces the water content in the CO2
to the outlet pressure of the LP pump. stream to its required specification of 480 mg m3 [56]. This model
In the EGR scheme, flue gas leaving the cold end of the HRSG is does not consider the energy requirements of regenerating the
further cooled in a direct contact cooler before it is split into two adsorption beds which would be done in practice by heating the
streams. One stream is sent to the absorber of the capture plant beds with some of the hot flue gas from the CCGT.
and the other is mixed with the fresh air stream to the compressor
in the gas turbine, after displacing an equal mass flow rate of the 2.5. Plant cost model
fresh air.
One of the key objectives of this work is to determine the
2.3. Post-combustion capture plant model impact that applying PCC has on the levelised cost of electricity
(LCOE) from the 420 MW CCGT when considering natural gas
An existing dynamic model of an amine-based PCC plant [52,53] prices predicted by DECC for the year 2030. The LCOE is calculated
was re-sized so that it was suitable for use with the CCGT plant. using the following equations [57]. These equations are used to cal-
Fig. 5 shows the configuration of the process units in the PCC plant culate the LCOE from the 420 MW CCGT power plant model in the
Table 2
Air stream properties.
Molar composition (%) Temperature (K) Pressure (MPa)

N2 O2 CO2 H2O SO2 Ar
gPROMS and THERMOFLEX values 77.292 20.738 0.030 1.009 0.000 0.931 288.2 0.10132
DSS Trim
Reboiler Cooler
(steam side)
Condensate Desuperheater
Pump
M
M
Control
Valve
Flow
LP Pump Steam
Controller LP Turbine HP Turbine
Condenser IP Turbine
IP Pump HP Pump
M
S S S
Flue Gas HP SD
LP SD M IP SD
Cooler
Flue Gas FGS

to Capture
Plant LPE LPB IPE IPB HPE1 IPS1 LPS HPE2 IPS2 HPB HPS1 RH1 HPS2 RH2 HPS3
Water to Fuel
drain Fuel
Preheater
Air FGM
Combuson
Compressor Chamber Gas Turbine
Equipment Key:
HRSG Units:
LPE – LP Economiser LPB – LP Evaporator IPE – IP Economiser IPB – IP Evaporator HPE1 – HP Economiser 1 IPS1 – IP Superheater 1
LPS – LP Superheater HPE2 – HP Economiser 2 IPS2 – IP Superheater 2 HPB – HP Evaporator HPS1 – HP Superheater 1 RH1 – Reheater 1
HPS2 – HP Superheater 2 RH2 – Reheater 2 HPS3 – HP Superheater 3 LP SD – LP Steam Drum IP SD – IP Steam Drum HP SD – HP Steam Drum
Miscellaneous units:
M – Stream Mixer (water/steam) S – Stream Splier (water/steam) FGS – Flue Gas Splier FGM – Flue Gas Mixer
Fig. 4. Process flow diagram of the modified CCGT power plant showing the steam extraction and exhaust gas recycle schemes.
CO 2 stripped
Vent
Flue Gas
To atmosphere
To atmosphere MEA Water
Trim Cooler Condenser

CO 2
Flue Gas FGC FGS To compression
From CCGT train
Solvent
Make-up
RC
Reflux
To drain
Rich-Lean
Heat Exchanger
Blower Absorber Desorber
Reboiler
Steam
Sump FC
To/from CCGT
FC
Condensate
Rich
Solvent
To drain
Equipment Key:
Miscellaneous Units:
FGC – Flue Gas Connector FGS – Flue Gas Splier FC – Flow Connector RC – Reflex Connector
Fig. 5. Process flow diagram of the post-combustion capture plant.

Table 4 Table 5
Key PCC plant equipment dimensions and operating parameters. Parameter values required to calculate the levelised cost of electricity.
Equipment dimension or operating parameter Value Cost or parameter Value Source

Absorber packed height (m) 40 CAPCCGT £229.5 million IECM software version 8.0.2
Absorber internal diameter (m) 21 CAPCCGT+PCC £352.6 million IECM software version 8.0.2
Desorber packed height (m) 25 MEA unit price £1601/tonne Biliyok and Yeung (2013)
Desorber internal diameter (m) 12 MEA
Liquid-gas ratio (mol/mol) 2.3 Water unit price £0.181/tonne Sinnott [48]
Solvent lean-loading 0.3 water
CO2 capture efficiency (%) 90 CO2 transportation and £4/tonne CO2 For transportation via pipeline and
storage cost per unit storage in depleted gas fields [56]
Plant life, n 25 years Mathieu and Bolland [57]
year 2030. Two cases are considered – one where the CCGT is a Average capacity factor 70% Assumed
over the plant’s life,
standalone plant and one where the CCGT is integrated with PCC.
CG
The case with PCC will consider the costs of transporting the com- Interest rate, i 10% Assumed
pressed CO2 via pipeline and storing it in depleted gas fields. Cost CO2 tax £0/tonne CO2 Assumed
data and the values of the other parameters that are required for Fraction of CO2 0.9 Assumed
use in the equations below are presented in Tables 5 and 6. captured, f CO2
CAP CRF
LCOEð£=MW hÞ ¼ þ ðO & M costsÞ
CF Net power Table 6
Natural gas price predictions for the UK in 2030 [66].
þ CO2 taxes paid ð89Þ
Low (£/GJ (LHV)) Central (£/GJ (LHV)) High (£/GJ (LHV))
n
ið1 þ iÞ
CRF ¼ n ð90Þ 3.90 6.82 9.74
ð1 þ iÞ 1
CAP ¼ CAPCCGT þ CAPCCGTþPCC ð91Þ CO2 emitted ðtonne CO2 =MW hÞ

nfg xCO2 1 f CO2 44 kg CO2
O & M costs ¼ NG costs þ Water costs þ MEA costs ¼
Plant net power output ðMW hÞ 1 kmol CO2
þ CO2 transportation and storage costs ð92Þ
1 tonne CO2
3600 s ð98Þ
1000 kg CO2
NG unit price ð£=GJÞ Net fuel input ðLHVÞ ðMWÞ
NG costs ð£=MW hÞ ¼
Plant net power output ðMWÞ
CO2 captured ðtonne CO2 =MW hÞ
3600 s
nfg xCO2 f CO2 44 kg CO2
ð93Þ ¼
Plant net power output ðMW hÞ 1 kmol CO2
Water unit price ð£=tonneÞ Make-up water rate ðkg s1 Þ 1 tonne CO2
Water costs ð£=MW hÞ ¼
3600 s ð99Þ
1000 kg CO2
3600 s
ð94Þ where
MEA unit price ð£=tonneÞ Make-up MEA rate ðkg s1 Þ CAP – capital cost of the project,
MEA costs ð£=MW hÞ ¼ CAP CCGT – capital cost of the standalone CCGT power plant,
3600 s CAP CCGTþPCC – capital cost of the CCGT with integrated PCC,
CRF – capital recovery factor,
ð95Þ
CF – capacity factor (average plant load),
O&M costs – operation and maintenance costs,
CO2 transportation and storage costs ð£=MW hÞ
i – interest rate,
¼ CO2 transportation and storage unit cost ð£=tonne CO2 Þ n – plant lifetime,
CO2 captured ðtonne CO2 =MW h ð96Þ nfg – molar flow rate of flue gas going to the PCC plant,
xCO2 – mole fraction of CO2 in the flue gas going to the PCC plant,
CO2 taxes paid ð£=MW hÞ ¼ CO2 tax ð£=tonne CO2 Þ f CO2 – fraction of the CO2 in the flue gas that is captured.
CO2 emitted ðtonne CO2 =MW hÞ

It should be noted that only the natural gas prices used in this
ð97Þ work are specific to the year 2030. All of the other prices quoted
Stage 1 Stages Stage 4 Stage 5

2-3
Knock-out
Drum
CO2 CO2
From capture To transportaon
plant CO2 Intercooler CO2 Intercooler CO2 Aercooler pipeline
Compressor Compressor Compressor
Adsorpve
CO2 Drying Unit
Fig. 6. Process flow diagram of the CO2 compression train.

are current and therefore may not reflect the prices in the year where
2030. It has also been assumed that there are no CO2 emissions
taxes in effect. Although this will not be the case in 2030, this nv ar – number of variables,
was purposely done in order to calculate the tax rate that will be nLF – number of load factor operating points,
needed to make the LCOE of the plant without PCC equal to that g i;j;calc – calculated value of variable i at load j,
of the plant with PCC. In addition, for the purposes of simplicity, g i;j;TF – THERMOFLEX value of variable i at load j.
the calculations do not include site operating taxes, maintenance
costs and labour costs. Comparisons between the net power outputs and plant efficien-
Power producers will only have an incentive to apply carbon cies (LHV) (gplant ) of the gPROMS and THERMOFLEX models are
capture if CO2 taxes are put into legislation that will make the LCOE given in Table 7.
for a CCGT without CCS higher than that for a plant with PCC. The The percentage absolute deviations of the plant power output
CO2 tax rate that will make the LCOE of the two cases equal can be and plant efficiency values increase as the load factor is decreased.
calculated as the cost of CO2 avoided [58]: The results obtained below load factors of 50% are considered
½LCOEwith CCS ½LCOEwithout CCS

Cost of CO2 Avoided ¼ ð100Þ
½tonne CO2 Emitted=MWwithout CCS ½tonne CO2 Emitted=MWwith CCS
unreliable as indicated by the %ADs of these parameters exceeding

5%. This trend is primarily influenced by the growing deviation in
3. Model validation the net power output predicted by the gas turbine model with
decreasing load and to a lesser extent by the deviations in the
The standalone 420 MW CCGT power plant model developed in power outputs predicted by the HP, IP and LP steam turbine mod-
this work was validated using the Thermoflow THERMOFLEX els. We note that the model calibration was performed at 100%
model of a triple-pressure reheat CCGT with a Siemens SGT5- power plant load. Thus all calculations performed at off-design
4000F gas turbine. We have selected power plant efficiency, net conditions can be regarded as predictions, with total model accu-
power output and the HRSG temperature profile at full and part racy within 4.1%, noting that the performance of our model reduces
load as key variables against which to validate the performance at load factors below 50%. However, based on some of our recent
of our model. work [19], we have found that thermal power plants will likely
Plant efficiency (LHV) is defined as follows: operate at greater than 50% load whilst contributing to the electric-
ity system. Once the demand requires the power plant to con-
W _ plant tribute less than 50% of its nameplate capacity, it will likely be
nplant ¼ 100 ð101Þ outcompeted by a more efficient generator operating at full load
_ fuel LHV fuel
m
and will therefore be displaced from the market. For this reason,
we propose that our current model is sufficient for the task of
where the net power output of the plant ðW_ plant Þ is given as the sum
describing the part-load behaviour of decarbonised CCGT power
of the power produced by the gas turbine, LP, IP and HP steam tur- plants. Finally, the gPROMS model describing the thermal beha-
bines minus the power consumed by the compressor in the gas tur- viour in the HRSG at design point was observed to reproduce that
bine, the LP, IP and HP pumps and the power needed to cool the of the THERMOFLEX model 2.860%.
water used in the steam condenser. The net power output predicted by the gPROMS gas turbine
The differences are expressed in terms of the percentage abso- model (i.e. the power output of the turbine minus the power con-
lute deviations (%AD) of the values predicted by the gPROMS sumption of the compressor) decreased significantly less with
model compared to the corresponding values in the THERMOFLEX decreasing load factor compared to the corresponding THERMO-
model. The %AD of variable i at load j calculated as: FLEX value. It was not possible to compare the power consumption
of the gPROMS compressor model and the power output of the
g i;j;calc g i;j;TF
%ADi;j ¼ 100 ð102Þ gPROMS turbine model to their corresponding THERMOFLEX values
g i;j;TF as these values were not reported individually in the THERMOFLEX
where Table 7
Comparison of the net plant power output (MW) and plant efficiency (LHV) predicted
g i;j;calc – calculated value of variable i at load j, by the gPROMS and THERMOFLEX models. The Absolute-relative Deviation (%AD)
between the gPROMS model and the THERMOFLEX model are presented in the
g i;j;TF – THERMOFLEX value of variable i at load j.
columns and the Average Absolute-relative Deviation (%AAD) is presented in the final
column for each performance indicator.
The overall deviation of the gPROMS CCGT from the THERMO-
Load Plant power output (MW) Plant efficiency (LHV) (%)
FLEX equivalent was measured by comparing the deviations of
the temperatures, pressures and vapour qualities or compositions TF gPROMS %AD TF gPROMS %AD
(where applicable) of every stream in the two models over all load 100 415.5 420.8 1.275 51.47 51.84 0.724
factors. The following equation was used to calculate the percent- 90 383.2 383.4 0.047 51.06 51.65 1.152
80 340.9 345.6 1.355 50.46 51.35 1.763
age average absolute deviation (%AAD) of the CCGT model as a
70 298.7 307.1 2.819 49.67 50.88 2.431
whole: 60 256.4 268.3 4.652 48.58 50.23 3.391
50 214.1 229.5 7.213 46.48 49.36 6.211
!
1 X nv ar Xload g i;j;calc g i;j;TF
1 100% 40 171.6 190.4 10.959 42.63 48.13 12.902
%AAD ¼ 100 ð103Þ
nv ar i¼1 nLF j¼40% g i;j;TF %AAD 4.046 4.082
model. It is suggested that the theoretical design equations that exhaust gas recycle (EGR) has on the performance of the plant in
were used to calculate the power input required by the compressor terms of power output and exhaust gas composition. In the case
at part-load (Eqs. (1)–(3)) are less accurate in this application as with PCC, the CCGT plant will be set to deliver an exhaust gas with
they do not consider the mechanical aspects of the compressor a CO2 concentration of 7% (made possible by the use of EGR) to the
and hence underestimate the power input required during part- PCC plant at all load factors. The PCC model was specified so as to
load operation. Similarly, Eqs. (21)–(23) that were used to calculate operate with a constant CO2 capture efficiency of 90%, i.e., 90% of
the power output of the turbine are believed to be inaccurate for the the CO2 in the flue gas stream is captured.
same reason.
The power outputs predicted by the gPROMS LP, IP and HP 4.1. Impact of EGR on the performance of the CCGT
steam turbines decreased more with decreasing load factor com-
pared to the corresponding THERMOFLEX values. These deviations This section will evaluate how varying the level of exhaust gas
were exacerbated by two assumptions that were made in the recycle (EGR) between 0% and 45% affects the composition of the
gPROMS model: (1) there were no pressure drops in the water/ flue gas going to the capture plant, the composition of the air being
steam side of the HRSG thus affecting the inlet pressures to the fed to the combustion chamber and its effect on plant power out-
HP and IP turbines and (2) the outlet pressures of the IP and LP tur- put. The maximum EGR ratio possible at each load factor will also
bines were assumed to be constant. be examined.
A comparison of the HRSG temperature profiles in the gPROMS
and THERMOFLEX models running at full load is shown in Fig. 7.
Whilst excellent agreement with the Thermoflow models is 4.1.1. Effect of EGR ratio on the compositions of the flue gas and air to
observed, those dissimilarities that are observed are attributed to the combustion chamber
the differences in the heat capacities of the water and steam pre- Fig. 8 illustrates how the compositions of the flue gas leaving
dicted by the gPROMS and THERMOFLEX models and the fact that the CCGT and the air to the combustion chamber change respec-
the flue gas enters the HRSG in the gPROMS model at a slightly tively as the EGR ratio is varied between 0% and 45% while the
higher temperature than in the THERMOFLEX model. CCGT is operated at full load.
It can be seen from Fig. 9 that increasing the level of EGR from
0% to 45% when the CCGT is operating at full-load increases the
4. Results and discussion molar concentration of CO2 in the exhaust gas from 4.32% to
9.04% whilst decreasing the molar concentration of O2 in the air
This section focuses on the impact that the application of post to the combustion chamber from 20.74% to 13.03%. It is gratifying
combustion carbon capture plant with CO2 compression to to note that the results presented in Fig. 8 are consistent with those
110 bar has on the performance of the 420 MW natural gas-fired presented by Canepa et al. [35].
CCGT power plant in terms of its power output and levelised cost The effects of varying the EGR ratio between 0% and 45% on the
of electricity (LCOE). This work will also examine the effects that CO2 and O2 concentrations in the flue gas and in the air to the com-
Fig. 7. Comparison of the full-load HRSG profiles in the gPROMS and THERMOFLEX models.
Fig. 8. Effect of EGR ratio on the composition of the exhaust gas leaving the CCGT at full-load.
Fig. 9. Effect of EGR ratio on the composition of the inlet air to the combustion chamber at full-load.
Fig. 10. Effect of EGR ratio on the CO2 concentration in the exhaust gas leaving the CCGT whilst operating at load factors between 40% and 100%.
bustion chamber respectively at plant load factors between 40% 4.2. Effect of EGR ratio on power plant output
and 100% are shown in Figs. 10 and 11.
As plant load drops, increased EGR has a reduced influence on It was found that increasing the EGR ratio from 0% to 45% causes
the CO2 concentration in the flue gas and on the O2 concentration a small decrease in the plant’s net power output by approximately
in the combustion chamber’s inlet air. The reason for this is that 1 MW whilst the plant is operated at any load factor between 40%
the CCGT’s controls increase the excess air ratio to the combustion and 100% as shown in Fig. 12.
chamber to reduce the plant’s output power. This leads to the pro- As the EGR ratio increases, the temperature of the inlet air to the
duction of flue gas that is more dilute in CO2 at lower loads. compressor will increase as a greater proportion of warm flue gas
Fig. 11. Effect of EGR ratio on the O2 concentration in the inlet air to the combustion chamber when the CCGT is operating at load factors between 40% and 100%.
at 40 °C is mixed with fresh air which is at 15 °C. An increase in the the capture plant which makes the capture process less energy
inlet temperature to the compressor, with molar flow rate held intensive. However, the EGR ratio is limited to 35–40% in practice
constant, might be expected increase the work requirement of [59,60] as increased an increasing EGR ratio reduces the level of
the compressor (see Eqs. (1)–(7)). However, the exhaust gas enter- oxygen that is available in the air to the combustion chamber.
ing the gas turbine will be at a slightly higher temperature and The O2 concentration in the inlet air to the combustion chamber
pressure and this will in turn increase the power output of the tur- should be greater than 16% to prevent the production of significant
bine (see Eqs. (21)–(26)) thus counteracting the increased power quantities of unburned hydrocarbons and carbon monoxide [59].
requirement of the compressor. However, the outlet pressure of The maximum EGR ratio is therefore the value that leads to an
the gas turbine in the model implemented in this work was set O2 concentration of 16% in the inlet air to the combustion chamber.
to a constant value of 0.1035 MPa due to the highly sensitive nat- The maximum EGR ratio increases as the plant’s load factor
ure of Stodola’s law, which was used to predict the part-load beha- decreases as can be seen in Fig. 13. This effect results from the
viour of the turbine. This means that the gas turbine equations set increase in the air-fuel ratio in the combustion chamber as load
the flow rate of air that must enter the compressor at each load fac- factor decreases.
tor so that the exhaust gas leaves the turbine at 0.1035 MPa. An In the remainder of this paper, we continue with an EGR of
increase in the temperature of the inlet stream to the compressor approximately 35% – consistent with the maximum EGR possible
causes the system to reduce the molar flowrate of the inlet stream. for practical applications, as illustrated in Fig. 13. This gives us a
This in turn reduces the power requirement of the compressor and CO2 concentration of approximately 7 mol% at the inlet to the
the power output of the gas turbine is also reduced but by a CO2 capture process, and is consistent with the conclusions of
slightly greater magnitude which results in the very slight decrease Pan et al. [8].
in the plant’s net power output that was observed when the EGR
ratio was increased from 0% to 45%. 4.4. Plant power output with and without PCC
4.3. Maximum EGR ratios This section presents an evaluation of the impacts that applying
PCC has on the gross and net power outputs and efficiency of the
Operating the CCGT with the highest possible EGR ratio is desir- 420 MW CCGT model. The analysis examines how the contribu-
able because it increases the concentration of CO2 in the flue gas to tions of the gas and steam turbines to the gross power output of
Fig. 12. Effect of EGR ratio on net power output of the CCGT.
Table 8
Gross and net power outputs from the CCGT with and without PCC.
Load Plant without CCS Plant with CCS

factor (%)
Gross power Net power Gross power Net power
output (MW) output (MW) output (MW) output (MW)
100 703.4 420.8 625.5 350.6
90 634.0 383.4 567.8 319.7
80 569.9 345.6 512.1 287.9
70 507.8 307.1 458.3 257.4
60 446.9 268.3 404.9 226.0
50 386.8 229.5 351.6 194.1
40 326.7 190.4 297.7 161.6
Fig. 13. Maximum EGR ratios at plant load factors between 40% and 100%.
the plant change when PCC is applied. The relative contributions of

the various components of the PCC plant to the associated energy
penalty will also be examined as well as the changes in reboiler
duty and the steam extraction rate over part-load.
4.4.1. Gross and net power outputs with and without PCC
As might be expected, the gross and net power outputs of the
CCGT power plant were reduced when PCC was applied as shown
in Fig. 14 and Table 8.
The application of PCC led to a 16.7% drop in the net power out-
put of the plant at full-load. This is in good agreement with similar
studies in the literature [61–64]. Fig. 15. Net plant LHV efficiency with and without PCC.
It was further observed that the percentage drop in net plant
power output decreased slightly with decreasing load factor to
Table 9
15.8% at 60% load, except for an increase at 80% load.
Contributions of the turbines to the gross power output (no PCC).
Load factor Contribution to gross plant output (%)

4.4.2. Net plant efficiency (%)
Gas HP steam IP steam LP steam
The standalone CCGT was found to have a net plant efficiency of turbine turbine turbine turbine
51.84% at full-load, dropping to an efficiency of 50.23% at 60% load
100 79.337 4.081 7.884 8.698
as shown in Fig. 15. This is a satisfactory result as it is representa- 90 78.511 4.254 8.091 9.144
tive of the state-of-the-art CCGT power plants that are currently 80 77.869 4.370 8.198 9.563
deployed [22] although it is also noted that is possible for CCGT 70 77.393 4.432 8.269 9.905
power plants to approach a net efficiency of 60%. 60 77.031 4.416 8.300 10.253
50 76.778 4.274 8.249 10.699
The application of PCC caused the net plant efficiency to drop by
40 76.714 3.941 7.980 11.365
8.64% points at full-load. This value agrees with the 6.2–9.6% point
drop in net plant efficiency resulting from the implementation of
PCC as found in nine studies previous studies [62,65]. The reduced 4.4.3. Contributions to the gross power output
impact of PCC at lower load factors was again evident as the net The power generated by the CCGT arises from the gas turbine
plant efficiency penalty dropped to 7.93% points at 60% load. and the LP, IP and HP steam turbines. Table 9 and Fig. 16 show
Fig. 14. Gross and net power outputs from the CCGT with and without PCC. The red curves represent the calculated gross (diamonds) and net (crosses) power generated
without CCS and the blue curves represent the calculated gross (diamonds) and net (crosses) power generated with CCS. (For interpretation of the references to color in this
figure legend, the reader is referred to the web version of this article.)
Table 11
Contributions of the PCC plant units to the energy penalty.
Load Percentage contribution to energy penalty (%)

factor (%)
Reboiler CO2 Trim cooler Miscellaneous
Compressors (capture plant)
100 74.309 18.891 5.642 1.158
90 74.108 19.138 5.598 1.156
80 73.786 19.477 5.579 1.159
70 73.033 20.162 5.604 1.201
60 72.079 21.005 5.657 1.259
50 70.913 22.029 5.720 1.338
40 69.514 23.260 5.779 1.447
Fig. 16. Contributions of the turbines to the gross power output (no PCC).
Table 10
Contributions of the turbines to the gross power output (with PCC).
Load factor Contribution to gross plant output (%)

(%)
Gas HP steam IP steam LP steam
turbine turbine turbine turbine
100 84.282 4.870 9.351 1.496
90 83.489 5.004 9.463 2.045
80 82.827 5.106 9.525 2.542
70 81.932 5.161 9.569 3.337
60 81.164 5.132 9.577 4.127
50 80.459 4.966 9.502 5.073
Fig. 18. Contribution of the PCC plant units to the energy penalty.
40 79.927 4.587 9.191 6.295
Fig. 17. Contributions of the turbines to the gross power output (with PCC).
Fig. 19. Steam extraction rate from the IP/LP cross-over and the amount of steam
extracted as a percentage of the total steam flow in steam cycle.
the percentage contributions of the gas and steam turbines in the
case without CCS.
The gas turbine provides the bulk of the CCGT’s power output by the CCGT must be used to drive the five CO2 compressors, the
regardless of whether PCC is implemented or not. The LP turbine flue gas blower and the equipment needed to provide cooling
contributes 8.70% to the standalone CCGT’s gross power output water to the flue gas cooler, the trim cooler in the steam extraction
at full-load but this contribution drops significantly to 1.50% when scheme, the trim cooler in the capture plant and to the intercoolers
PCC is implemented due to the large amount of steam extracted to in the compression train. The percentage contributions of these
provide the reboiler duty. This is a key factor in CCGT plants suffer- units to the total energy penalty imposed when the CCGT is inte-
ing a lower efficiency penalty than an equivalent coal-fired power grated with the PCC plant are shown in Table 11 and Fig. 18.
plant on addition of post-combustion CCS – the coal plants gener- At full-load, the extraction of steam from the IP/LP cross-over to
ate a much greater fraction of their overall power from their LP tur- provide the reboiler duty accounts for 74.31% of the energy pen-
bine. Table 10 and Fig. 17 show the percentage contributions of alty, the electricity required to run the CO2 compressors accounts
these units in the case with PCC. for 18.89%, the trim cooler in the PCC plant constitutes 5.64% and
the remaining units are grouped as miscellaneous equipment and
4.4.4. Contributions to the PCC energy penalty account for 1.16%. It is interesting to note that the energy penalty
When PCC is implemented, energy is lost because steam is associated with operating the compressors increases at part load to
diverted from the LP steam turbine to provide the required for sol- 23.26% of the total energy penalty, whilst that associated with
vent regeneration. Further, and some of the electricity generated solvent regeneration reduces to 69.5%. The power requirement
Fig. 20. Levelised cost of electricity from the 420 MW CCGT model with and without PCC.
associated with operating the trim cooler in the PCC plant is rela-
tively large. This arose from the fact that the cooling water was
set to have a 5 K temperature rise in the gPROMS model and this
resulted in a required cooling water flowrate of approximately
8,000 kg s1.
4.4.5. Reboiler duty and steam extraction rate

The reboiler’s (heat) duty decreases linearly with decreasing
load factor as the total steam extracted from the IP/LP cross-over
decreases more-or-less linearly with load factor as shown in
Fig. 19. Fig. 19 also shows that the amount of steam extracted as
a percentage of the total water flow rate in the steam cycle also
decreases in a roughly linear manner with load factor. At full-
load 85.25% of the total steam flow is extracted whilst at 40% load
this figure is 51.17%. This factor is responsible for the observed
reduction in the energy penalty at lower loads. Fig. 21. Contributions to the LCOE from the CCGT without PCC under central
natural gas price predictions.
4.5. Plant economics
Carbon capture and storage is expected to see deployment to

the UK’s energy system in the 2030–2040s [19]. We therefore
examine the impact that applying PCC to the 420 MW CCGT has
on the cost of electricity generated when considering the natural
gas prices predicted by DECC for the 2030s. A break-down of the
total levelised cost of electricity (LCOE) is presented and the CO2
tax rate that will be needed to make the LCOE of the plant without
CCS equal to that of the plant with CCS is also calculated.
4.5.1. Levelised cost of electricity with and without PCC

The graph in Fig. 20 shows the LCOE for the two power genera-
tion cases (with and without PCC) over the plant’s operating range.
The LCOE values predicted will be lower than what they would be
in reality because the calculations that were used do not include
CO2 taxes, site operating taxes, maintenance costs, labour costs,
etc. Applying PCC caused the LCOE from the CCGT to increase by Fig. 22. Contributions to the LCOE from the CCGT with PCC under central natural
34.7% (from £38/MW h to £51/MW h), 29.6% (from £58/MW h to gas price predictions.
£75/MW h) and 27.1% (from £78/MW h to £100/MW h) under
low, central and high natural gas price predictions respectively
when operating at full load. and LCOE of CCS applied to CCGT power stations. An increase in
The UK’s Department of Energy and Climate Change (DECC) pre- natural gas prices of £1/GJ will cause the LCOE in the CCGT without
dicts that the LCOE from a CCGT commissioned in 2030 will PCC to increase by £6.94/MW h and by £8.33/MW h in the case
increase by 20.9%, from £86/MW h to £104/MW h following imple- with PCC. The LCOE also increased marginally in both cases as
mentation of PCC, under central capital cost predictions [66]. We the plant is ramped down due to the decrease in the plant’s
note that our results are in good agreement with those of DECC. efficiency.
Further, these results also show good agreement with the recent
work of Canepa and Wang [67]. 4.5.2. Contributions to the levelised cost of electricity
It is clear that changes in the price of natural gas have a very A break-down of the LCOE from the two CCGT cases (with and
significant impact on both the short run marginal cost (SRMC) without PCC) under central natural gas price predictions can be seen
Further, Fig. 24 quite conclusively shows that, in the absence of

CCS, gas-fired power plants cannot be considered to be low carbon.
5. Conclusions
We have presented a detailed mathematical model of a triple-

pressure CCGT power station integrated with an amine-based
post-combustion CO2 capture process and subsequent multi-stage
CO2 compression train. We have validated the performance of the
CCGT model by comparison with the commercial Thermoflow
THERMOFLEX simulation package and observed that our model
accurately reproduces the performance of the commercial simula-
tor over a range of full- and part-load conditions. We have found
Fig. 23. Cost of CO2 avoided under low, central and high natural gas price
that our model reproduces the net power output and efficiency of
predictions.
the Thermoflow model to within 4.1% and the full load temperature
profile within the HRSG to within 2.9%. Performance suffers at load
factors less than 50% owing to increasing phase change in the heat
exchangers and a lack of detail in the compressor, gas turbine and
steam turbine sub-models. However, as CCGTs might be expected
to operate between 50% and 100% load under normal operation,
with operation below 50% of nameplate capacity as part of a shut-
down operation, this model is deemed appropriate for simulation
of normal operation. This is of particular importance as CCGT power
stations, with and without CCS will increasingly be required to
operate in a highly flexible and load following manner, in sympathy
with the availability of intermittent renewable energy sources.
In integrating the CCGT and CO2 capture process, steam for the
solvent reboiler was extracted from the IP-LP crossover point. Here,
it was first necessary to de-superheat this steam, and whilst this
Fig. 24. Rates of CO2 emitted by the CCGT plant with and without PCC respectively.
heat was subsequently recovered and added to the condensate
train, it follows that it may be preferable to extract steam from
in Fig. 21 and in Fig. 22. Fuel costs contribute approximately 82% of an inter-stage bleed within the LP turbine, rather than prior to
the LCOE in the case without PCC and 76% in the case with PCC. the turbine. This would have the benefit of reducing the efficiency
The costs of transporting the captured CO2 via pipeline and stor- penalty associated with the operation of the CO2 capture process.
ing it in an appropriate geological reservoir account for approxi- One of the primary challenges associated with the application of
mately 2.6% of the total LCOE from the CCGT with PCC. Another post-combustion CCS to CCGTs is the dilute nature of the exhaust
interesting observation is that the proportions of the various con- gas. To mitigate against this, we evaluated the ability of exhaust
tributions remain more or less constant across the range of plant gas recycle (EGR) to increase the concentration of CO2. We found
loads evaluated in this study. that the CO2 concentration varied between 3 and 5 mol% as a func-
tion of load factor in the range 40–100% in the absence of EGR, but
4.5.3. Cost of CO2 avoided with EGR it was possible to increase the concentration of the CO2 to
The cost of CO2 avoided is strongly influenced by the price of between 6 and 9 mol%. An important side-effect of EGR was also to
natural gas but is relatively weakly influenced by the plant’s load reduce the O2 concentration in the exhaust gas from approximately
factor as can be seen in Fig. 23. This is primarily driven by the 20–21 mol% to 13–15 mol%, which may be important in contribut-
relatively small capital intensity of CCGTs relative to other forms ing to a reduction in solvent degradation. However, the require-
of power plants, for example, coal-fired power plants. The CO2 ment for greater than 16 mol% O2 in the combustion chamber
tax rate will need to be £59/tonne CO2, £48/tonne CO2 and £36.50/- acts to limit the amount of EGR as a function of load factor –
tonne CO2 under high, central and low natural gas price predictions increasing from 33.65% at 100% load factor to 40.4% at 40% load
respectively. These calculated values are higher than the penalty of factor.
£27/tonne CO2 that was proposed by the International Energy We simulated the dynamic behaviour of the integrated power-
Agency to cover the costs of CCS [56]. capture plant, and noted that the relaxation times of the individual
Fig. 23 shows that the cost of CO2 avoided increases at load fac- units within this system mean that the is a non-negligible delay for
tors of 80% and 90%. This is an anomaly caused by fluctuations in the system to reach a steady state after each simulation. A key area
the CO2 capture efficiency in the capture plant model. of future work will be to include the individual dynamics of the
The effect of the fluctuations in the CO2 capture efficiency is sub-units within the CCGT model. The calculated net efficiency of
illustrated in Fig. 24, which show graphs of the CO2 emitted to the integrated power-capture plant was observed to be in excellent
the atmosphere and captured in the two power plant scenarios. agreement with similar studies in the literature for both full and
The slightly non-monotonic behaviour that is observed in Fig. 24 part load scenarios.
is an artefact of the CO2 capture plant model. The CCS plant An evaluation of the LCOE associated with our model was also
dynamics lag those of the steady state CCGT model as the system found to be in good agreement with literature values. Importantly,
is ramped from full to part load. This leads to a surplus of solvent it was observed that whilst the absolute costs and efficiency of the
in the capture plant between load factors of 75–95%. Once the CCGT-CCS varied appreciably as a function of power plant load fac-
power plant is ramped below 75% of nameplate capacity, the sol- tor, the cost structure did not. The implication is that once an opti-
vent circulation rate in the CO2 capture plant has also reduced, mal process configuration and operating strategy has been
and therefore responds in line with the power plant perturbations. determined, it will not be necessary to revisit this for the actual
dispatch profiles that are likely to be experienced by plants operat- [10] Alhajaj A, Mac Dowell N, Shah N. A techno-economic analysis of post-
combustion CO2 capture and compression applied to a combined cycle gas
ing in a real energy system.
turbine: Part I. A parametric study of the key technical performance indicators.
Taken together, this gives us confidence that this tool is suitable Int J Greenhouse Gas Contr 2016;44:26–41.
for use in a range of design and systems modelling purposes. [11] Nogueira L Pinheiro Pupo, de Lucena A Frossard Pereira, Rathmann R, Rochedo
Importantly, the general trend that sustained off-design point P Rua Rodriguez, Szklo A, Schaeffer R. Will thermal power plants with CCS play
a role in Brazil’s future electric power generation? Int J Greenhouse Gas Contr
operation will have the effect of increasing the LCOE of these plants 2014;24:115–23.
implies that energy systems should aim to maximise their utilisa- [12] Johnsson F, Odenberger M, Göransson L. Challenges to integrate CCS into low
tion of CCS assets wherever possible. However, it is important to carbon electricity markets. Energy Proc 2014;63:7485–93.
[13] Cormos AM, Simon A. Dynamic modelling of CO2 capture by calcium-looping
note that a simple metric such as LCOE does not accurately reflect cycle. Chem Eng Trans 2013;35:421–6.
value provided to a decarbonised energy system by a dispatchable, [14] Campanari S, Chiesa P, Manzolini G, Bedogni S. Economic analysis of CO2
low carbon power station such as a decarbonised CCGT power capture from natural gas combined cycles using Molten carbonate fuel cells.
Appl Energy 2014;130:562–73.
plant. [15] Spence B, Horan D, Tucker O. The Peterhead–Goldeneye gas post-combustion
It was interesting to note that the energy cost associated with CCS project. Energy Proc 2014;63:6258–66.
operating the compression train became relatively greater under [16] Ceccarelli N, van Leeuwen M, Wolf T, van Leeuwen P, van der Vaart R, Maasa
W, et al. Flexibility of low-CO2 gas power plants: integration of the CO2 capture
part-load scenarios in contrast to the relative reduction in cost unit with CCGT operation. Energy Proc 2014;63:1703–26.
associated with solvent regeneration. Given that CCS-CCGTs are [17] Boot-handford M, Abanades JC, Anthony EJ, Blunt MJ, Brandani S, Mac Dowell
expected to be required to operate in part load with increasing fre- N, et al. Carbon capture and storage update. Energy Environ Sci 2013;7
(1):130–89.
quency, the identification of improved designs and operating pro-
[18] Brouwer AS, van den Broek M, Seebregts A, Faaij A. Operational flexibility and
cedures under realistic conditions is of increasing importance. economics of power plants in future low-carbon power systems. Appl Energy
Simply recycling CO2 in the same manner as EGR for increasing 2015;156:107–28.
the concentration of CO2 is unlikely to be an optimal solution. It [19] Mac Dowell N, Staffell I. The role of flexible CCS in the UK’s future energy
system. Int J Greenhouse Gas Contr 2016;48:327–44.
will also be necessary to operate this aspect of the process such [20] Rezazadeh F, Gale WF, Hughes KJ, Pourkashanian M. Performance viability of a
that the impact on the reliable and safe operation of the transport natural gas fired combined cycle power plant integrated with post-
and sequestration elements of the CCS chain is minimised. combustion CO2 capture at part-load and temporary non-capture operations.
Int J Greenhouse Gas Contr 2015;39:397–406.
Owing to reduced efficiency under off-design point operation, [21] Flø NM, Kvamsdal HM, Hillestad M. Dynamic simulation of post-combustion
the carbon intensity of the decarbonised CCGT increased. Thus, in CO2 capture for flexible operation of the Brindisi pilot plant. Int J Greenhouse
order to ensure that CCS-CCGTs actually provide reliably low car- Gas Contr 2016;48(Part 2):204–15.
[22] Turconi R, O’Dwyer C, Flynn D, Astrup T. Emissions from cycling of thermal
bon power, it may be necessary to increase the fraction of CO2 that power plants in electricity systems with high penetration of wind power: life
is removed at part-load conditions to greater to 95% under part- cycle assessment for Ireland. Appl Energy 2014;131:1–8.
load conditions. A key implication is that this would serve to [23] Kang CA, Brandt AR, Durlofsky LJ. A new carbon capture proxy model for
optimizing the design and time-varying operation of a coal-natural gas power
increase both the average levelised and short run marginal cost station. Int J Greenhouse Gas Contr 2016;48(Part 2):234–52.
of gas-CCS power plants, potentially further reducing access to [24] Middleton RS, Eccles JK. The complex future of CO2 capture and storage:
the electricity market. This serves to emphasise the importance variable electricity generation and fossil fuel power. Appl Energy
2013;108:66–73.
of identifying reduced cost approaches for delivering gas-CCS
[25] Mac Dowell N, Shah N. Optimisation of post-combustion CO2 capture for
designs and operating procedures optimised for part load. How- flexible operation. Energy Proc 2014;63:1525–35.
ever, it is also noted that the cost of gas-CCS is a strong function [26] Mac Dowell N, Shah N. The multi-period optimisation of an amine-based CO2
of the prevailing gas price. In this context, the increasing fungibility capture process integrated with a super-critical coal-fired power station for
flexible operation. Comput Chem Eng 2015;74:169–83.
of the gas market may serve to reduce gas prices, which, in turn, [27] Benato A, Stoppato A, Mirandola A. Dynamic behaviour analysis of a three
could reduce the cost of gas-CCS. pressure level heat recovery steam generator during transient operation.
Energy 2015;90(2):1595–605.
[28] Sipöcz N, Jonshagen K, Assadi M, Genrup M. Novel high-performing single-
Acknowledgements pressure combined cycle with CO2 capture. J Eng Gas Turb Power 2010;133
(4):041701–8.
[29] Mertens J, Lepaumier H, Desagher D, Thielens ML. Understanding
The authors thank the Engineering and Physical Sciences
ethanolamine (MEA) and ammonia emissions from amine based post
Research Council (EPSRC) of the UK (Grant No. EP/M001369/1) combustion carbon capture: lessons learned from field tests. Int J
for funding required to carry out this work. Greenhouse Gas Contr 2013;13:72–7.
[30] Lepaumier H, Picq D, Carrette PL. New amines for CO2 capture. II. Oxidative
degradation mechanisms. Ind Eng Chem Res 2009;48:9068–75.
References [31] ElKady AM, Evulet A, Brand A, Ursin TP, Lynghjem A. Application of exhaust gas
recirculation in a DLN F-class combustion system for post-combustion carbon
[1] Intergovernmental Panel on Climate Change. Climate change 2013: the capture. J Eng Gas Turb Power 2009;131(3):034505–11.
physical science basis. New York (USA): Cambridge University Press; 2013. [32] Sipöcz N, Tobiesen FA. Natural gas combined cycle power plants with CO2
[2] Tian R, Zhang Q, Wang G. Market analysis of natural gas power generation in capture – opportunities to reduce cost. Int J Greenhouse Gas Contr
China. Energy Proc 2015;75:2718–23. 2012;7:98–106.
[3] Ehyaei MA, Tahani M, Ahmadi P, Esfandiari M. Optimization of fog inlet air [33] Herraiz L, Hogg D, Cooper J, Gibbins J, Lucquiaud. Reducing water usage with
cooling system for combined cycle power plants using genetic algorithm. Appl rotary regenerative gas/gas heat exchangers in natural gas-fired power plants
Therm Eng 2015;76:449–61. with post-combustion carbon capture. Energy 2015;90:1994–2005.
[4] Hemmer F, Klassen E. International comparison of fossil power efficiency and [34] Loud RL, Slaterpryce AA. Gas turbine inlet air treatment. In: GE power
CO2 intensity. ECOFYS; 2012. generation, GER-3419A; 1991.
[5] Riesz J, Vithayasrichareon P, MacGill I. Assessing ‘‘Gas Transition” pathways to [35] Canepa R, Wang M, Biliyok C, Satta A. Thermodynamic analysis of combined
low carbon electricity – an Australian case study. Appl Energy 2015;154:794–804. cycle gas turbine power plant with post-combustion CO2 capture and exhaust
[6] Mac Dowell N, Florin N, Buchard A, Hallett J, Galindo A, Jackson G, et al. An gas recirculation. Proc Inst Mech Eng Part E: J Process Mech Eng
overview of CO2 capture technologies. Energy Environ Sci 2010;3 2013;227:89–105.
(11):1645–69. [36] Jonshagen K, Sipöcz N, Genrup M. Optimal combined cycle for CO2 capture
[7] Duan HB, Fan Y, Zhu L. What’s the most cost-effective policy of CO2 targeted with EGR. In: Proceedings of the ASME turbo expo, Glasgow, UK, 14–18 June
reduction: an application of aggregated economic technological model with 2010. New York: ASME; 2010. p. 867–75 [GT2010-23420].
CCS? Appl Energy 2012;112:868–75. [37] Gil-Villegas A, Galindo A, Whitehead PJ, Mills SJ, Jackson G, Burgess AN.
[8] Pan M, Aziz F, Li B, Perry S, Zhang N, Bulatov I, et al. Application of optimal Statistical associating fluid theory for chain molecules with attractive
design methodologies in retrofitting natural gas combined cycle power plants potentials of variable range. J Chem Phys 1997;106:4168–86.
with CO2 capture. Appl Energy 2016;161:695–706. [38] Galindo A, Davies LA, Gil-Villegas A, Jackson G. The thermodynamics of
[9] Gerbelová H, Versteeg P, Ioakimidis CS, Ferrão P. The effect of retrofitting mixtures and the corresponding mixing rules in the SAFT-VR approach for
Portuguese fossil fuel power plants with CCS. Appl Energy 2013;101:280–7. potentials of variable range. Mol Phys 1998;93(2):241–52.
[39] Mac Dowell N, Llovell F, Adjiman CS, Jackson G, Galindo A. Modeling the fluid [54] Biliyok C, Yeung H. Evaluation of natural gas combined cycle power plant for
phase behaviour of carbon dioxide in aqueous solutions of monoethanolamine post-combustion CO2 capture integration. Int J Greenhouse Gas Contr
using transferable parameters with the SAFT-VR approach. Ind Eng Chem Res 2013;19:396–405.
2009;49(4):1883–99. [55] Le Moullec Y, Kanniche M. Screening of flowsheet modifications for an efficient
[40] Mac Dowell N, Pereira FE, Llovell F, Blas FJ, Adjiman CS, Jackson G, et al. monoethanolamine (MEA) based post- combustion CO2 capture. Int J
Transferable SAFT-VR models for the calculation of the fluid phase equilibria in Greenhouse Gas Contr 2011;5(4):727–40.
reactive mixtures of carbon dioxide, water, and n-alkylamines in the context of [56] Hammond GP, Akwe SSO, Williams S. Techno-economic appraisal of fossil-
carbon capture. J Phys Chem B 2012;115(25):8155–68. fuelled power generation systems with carbon dioxide capture and storage.
[41] Rodriguez J, Mac Dowell N, Llovell F, Adjiman CS, Jackson G, Galindo A. Energy 2011;36(2):975–84.
Modelling the fluid phase behaviour of aqueous mixtures of multifunctional [57] Mathieu P, Bolland O. Comparison of costs for natural gas power generation
alkanolamines and carbon dioxide using transferable parameters with the with CO2 capture. Energy Proc 2013;37(0):2406–19.
SAFT-VR approach. Mol Phys 2012;110(11–12):1325–48. [58] Wilcox J. Carbon capture. 1st ed. New York: Springer-Verlag; 2012.
[42] Pritchard PJ. Fox and McDonald’s introduction to fluid mechanics. 8th [59] Li H, Haugen G, Ditaranto M, Berstad D, Jordal K. Impacts of exhaust gas
ed. Hoboken (NJ): Wiley; 2011. recirculation (EGR) on the natural gas combined cycle integrated with
[43] Horlock JH. Advanced gas turbine cycles. Amsterdam (London): Elsevier; 2003. chemical absorption CO2 capture technology. Energy Proc 2011;4:1411–8.
[44] Moller BF, Genrup M, Assadi M. On the off-design of a natural gas-fired [60] Merkel TC, Wei XT, He ZJ, White LS, Wijmans JG, Baker RW. Selective exhaust
combined cycle with CO2 capture. Energy 2007;32(4):353–9. gas recycle with membranes for CO2 capture from natural gas combined cycle
[45] Mansouri MT, Ahmadi P, Kaviri AG, Nazri M, Jaafar M. Exergetic and economic power plants. Ind Eng Chem Res 2013;52(3):1150–9.
evaluation of the effect of HRSG configurations on the performance of [61] Amrollahi Z, Ertesvag IS, Bolland O. Thermodynamic analysis on post-
combined cycle power plants. Energy Convers Manage 2012;58:47–58. combustion CO2 capture of natural-gas-fired power plant. Int J Greenhouse
[46] Perry R. Physical and chemical data. In: Anonymous Perry’s chemical Gas Contr 2011;5:422–6.
engineers’ handbook. McGraw Hill; 1997. p. 155. [62] Jordal K, Ystad PAM, Anantharaman R, Chikukwa A, Bollan O. Design-point and
[47] Paterson WR. A replacement for the logarithmic mean. Chem Eng Sci 1984;39 part-load considerations for natural gas combined cycle plants with post
(11):1635–6. combustion capture. Int J Greenhouse Gas Contr 2012;11:271–82.
[48] Sinnott RK. Coulson & Richardson’s chemical engineering. Chemical [63] Dillon D, Grace D, Maxson A, Børter K, Augeli J, Woodhouse S, et al. Post-
engineering design, 4th ed., vol. 6. Oxford: Butterworth-Heinemann; 2005. combustion capture on natural gas combined cycle plants: a technical and
[49] Shah RK. Fundamentals of heat exchanger design. New York (Chichester): John economical evaluation of retrofit, new build and the application of exhaust gas
Wiley & Sons; 2003. recycle. Energy Proc 2013;37:2397–405.
[50] Dechamps PJ, Pirard N, Mathieu P. Part-load operation of combined-cycle [64] Kuramochi T, Ramírez A, Faaij A, Turkenburg W. Post-combustion CO2 capture
plants with and without supplementary firing. J Eng Gas Turb Power 1995;117 from part-load industrial NGCCCHPs: selected results. Energy Proc
(3):475–83. 2009;1:1395–402.
[51] Çengel YA. Thermodynamics: an engineering approach. 6th ed. New York [65] Lindqvist K, Jordal K, Haugen G, Hoff KA, Anantharaman R. Integration aspects
(London): McGraw-Hill; 2007. p. 567–89. of reactive absorption for post-combustion CO2 capture from NGCC (natural
[52] Mac Dowell N, Samsatli NJ, Shah N. Dynamic modelling and analysis of an gas combined cycle) power plants. Energy 2014;78:758–67.
amine-based post-combustion CO2 capture absorption column. Int J [66] DECC. Electricity generation costs (December 2013). Department of Energy
Greenhouse Gas Contr 2013;12:247–58. and Climate Change; 2013. p. 25.
[53] Mac Dowell N, Shah N. Identification of the cost-optimal degree of CO2 [67] Canepa R, Wang M. Techno-economic analysis of a CO2 capture plant
capture: an optimisation study using dynamic process models. Int J integrated with a commercial scale combined cycle gas turbine (CCGT)
Greenhouse Gas Contr 2013;13:44–58. power plant. Appl Therm Eng 2015;74:10–9.

1 s2.0 S0306261916308625 Main

Uploaded by

Copyright:

Available Formats

1 s2.0 S0306261916308625 Main

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1 s2.0 S0306261916308625 Main

Uploaded by

Copyright:

Available Formats

Applied Energy 178 (2016) 681–702

Contents lists available at ScienceDirect

Off-design point modelling of a 420 MW CCGT power plant integrated

Multi-scale modelling of CCGT with CO2 capture and compression.

1. Introduction However, fossil fuels are expected to play an important role in

V specific volume (m3 kg1)

LP Pump Flow Steam

RT in Eqs. (1)–(5) are used again to determine the work requirements

1 Species Standard heat of formation, Df h° Lower heating value, LHV

2.1.6.2. Operation mode. In the operation mode, the perfect control

The temperature of the outlet steam requires consideration of the

2.1.8. Superheater/reheater Fig. 3. Steam expansion scenarios in the steam turbine.

_ water;LP ¼ 5  105 LF 3 þ 0:0117LF 2 0:8329LF

Flow rate of water diverted to the fuel preheater (kg s1):

A five-stage intercooled compression train model was created

Molar composition (%) Temperature (K) Pressure (MPa)

Flue Gas FGS

Trim Cooler Condenser

Blower Absorber Desorber

Fig. 5. Process flow diagram of the post-combustion capture plant.

Equipment dimension or operating parameter Value Cost or parameter Value Source

CAP ¼ CAPCCGT þ CAPCCGTþPCC ð91Þ CO2 emitted ðtonne CO2 =MW hÞ

 CO2 emitted ðtonne CO2 =MW hÞ

Stage 1 Stages Stage 4 Stage 5

Fig. 6. Process flow diagram of the CO2 compression train.

½LCOEwith CCS ½LCOEwithout CCS

unreliable as indicated by the %ADs of these parameters exceeding

Load Plant without CCS Plant with CCS

the plant change when PCC is applied. The relative contributions of

Load factor Contribution to gross plant output (%)

Load Percentage contribution to energy penalty (%)

Load factor Contribution to gross plant output (%)

4.4.5. Reboiler duty and steam extraction rate

4.5. Plant economics

Carbon capture and storage is expected to see deployment to

4.5.1. Levelised cost of electricity with and without PCC

Further, Fig. 24 quite conclusively shows that, in the absence of

We have presented a detailed mathematical model of a triple-

You might also like

_ water;LP ¼ 5 105 LF 3 þ 0:0117LF 2 0:8329LF

CO2 emitted ðtonne CO2 =MW hÞ

½LCOEwith CCS ½LCOEwithout CCS