Carbon Capture From Natural Gas Combined Cycle (NGCC) Power Plants: Solvent Performance Comparison at An Industrial Scale

View metadata, citation and similar papers at core.ac.
uk brought to you by CORE

provided by UCL Discovery
Please cite the following paper as:

Mahdi Sharifzadeh, Nilay Shah. Carbon capture from natural gas combined cycle power plants:
Solvent performance comparison at an industrial scale. AIChE Journal, 2016, Volume 62, Issue 1,
Pages 166–179,
Carbon capture from natural gas combined cycle (NGCC) power
plants: solvent performance comparison at an industrial scale
Sharifzadeh, Mahdi1; Shah, Nilay
Centre for Process Systems Engineering (CPSE), Department of Chemical Engineering, Imperial College London,
London, United Kingdom, SW7 2AZ.
Abstract
Natural gas is an important source of energy. This paper addresses the problem of integrating an existing natural
gas combined cycle (NGCC) power plant with a carbon capture process using various solvents. The power plant
and capture process have mutual interactions in terms of the flue gas flowrate and composition versus the
extracted steam required for solvent regeneration. Therefore, evaluating solvent performance at a single
(nominal) operating point is not indicative and solvent performance should be considered subject to the overall
process operability and over a wide range of operating conditions. In the present research, a novel optimization
framework was developed in which design and operation of the capture process are optimized simultaneously
and their interactions with the upstream power plant are fully captured. The developed framework was applied
for solvent comparison which demonstrated that GCCmax, a newly developed solvent, features superior
performances compared to the MEA baseline solvents.
Key words
1
Corresponding Author: Dr Mahdi Sharifzadeh; Room C603, Roderic Hill Building, South Kensington Campus, Imperial
College London, UK. SW7 2AZ. E-mail: mahdi@imperial.ac.uk ; Tel: +44(0)7517853422.
This contribution was identified by the AIChE Session Chair, Dr Athanasios Papadopoulos (Centre for Research and
Technology Hellas), as the Best Presentation in the session of Design of CO2 Capture Systems during the 2014 AIChE
Annual Meeting in Atlanta, GA, November 16-21, 2014.
1|P a g e
Pages 166–179,
CO2, Carbon capture, natural gas combined cycle (NGCC) power plant, energy efficiency, integrated process
design and control, GCCmax.
2|P a g e
Pages 166–179,
Introduction
Increasing energy demand and associated pollution have posed an important challenge around the security of
energy supply and environmental protection. Among various prospective scenarios, the International Energy
Agency (IEA) asserts that fossil fuels are most likely to remain the dominant sources of energy for a foreseeable
future1. Therefore, carbon capture from existing fossil-fuel-driven energy infrastructure will be a major pathway
for sustainability and environmental protection. Despite such clarity, there are various barriers against
commercialization of carbon capture technologies. Firstly, the current energy infrastructure is relatively mature
and the number of existing processes is significantly larger than the number of processes under construction.
Therefore, enhancing energy efficiency and mitigating the emissions should require minimal process retrofit.
Secondly, seamless integration of energy conversion processes with carbon capture technology requires the
latter process to be at least as flexible as the former (standalone) process. Finally, in order to justify the process
retrofit and overcome financial barriers, the energetic implications of the carbon capture process should be
minimal.
Solvent-based CO2 removal using aqueous amines is the most promising technology for carbon capture, as this
technology is an end-of-pipe treatment and has been in use since the 1930s for natural gas sweetening 2.
However, adaptation of this technology for post-combustion carbon capture is nontrivial as gas processing is
significantly different from power generation in several aspects. Natural gas is often produced at an elevated
pressure and does not contain any oxygen. Furthermore, in natural gas sweetening, a higher degree of CO2
removal is required and the separated CO2 is emitted to the atmosphere. Finally, gas processing is a relatively
steady-state process. By comparison, post-combustion carbon removal is conducted at near atmosphere
pressure and deep CO2 separation is often uneconomic. In addition, power plants are subject to drastic variations
in electricity demand and for the capture plant to remain integrated with the power plant, it should feature a
high degree of flexibility.
Retrofitting the existing power generation processes with carbon capture technologies has been the focus of
academic and industrial researchers. The International Energy Agency (IEA) conducted a comparative study3 on
pre-combustion and post-combustion CO2 capture in natural gas combined cycle (NGCC) power plants, as well
3|P a g e
Pages 166–179,
as pre-combustion carbon capture from coal gasification plants. Post-combustion carbon capture from natural
gas was identified as the lowest cost retrofit. Later, the National Energy Technology Laboratory (NETL), in a
comprehensive study 4, evaluated the implications of carbon capture from natural gas combined cycle (NGCC),
integrated gasification combined cycle (IGCC), and pulverized coal (PC) power plants. The observation was that
the energetic penalties associated with carbon capture from NGCC power plants are less than PC and IGCC power
plants, mainly due to the lower carbon intensity of the natural gas and the higher conversion efficiency of NGCC
plants. Large increases in the boiler water withdrawal and cooling water were observed for the scenario of NGCC
power plant, integrated with carbon capture process. Recently, a group of European researchers 5 conducted a
comprehensive study on the carbon capture from a supercritical pulverised coal power plant and a natural gas
combined cycle (NGCC) power plant. They employed two economic analysis methods; a top-down method in
which the historical data from previous projects and similar studies were used, and a bottom-up method that
was based on mass and energy analysis and detailed equipment costing. Significant difference between the
results of two studies was reported (table 7 of reference5), illustrating the difficulties associated with
reproducible and comparative economic analysis.
Furthermore, researchers have focused on the method of process integration from a thermodynamic point of
view. The heat integration schemes investigated include steam extraction and condensate recycling 6,
integrating compressor inter-coolers to the low pressure section of the steam cycle7 or stripper reboiler8,
preheating combustion air using waste heat from the capture plant9, and application of pressurized hot water
instead of steam for solvent regeneration10,11. Furthermore, the CO2 concentration of the flue gas can be
increased by recirculation of exhaust gases 10,12,13 or using a supplementary burner placed in the duct connecting
the turbine exhaust and heat recovery steam generation (HRSG) system12-14. Other researchers have explored
the implications of the process configuration on the capital investment and energy costs.
It was shown that depending on the solvent heat of desorption, either a multi-pressure or vacuum desorber
could be the optimal configuration15. Other configurations include the absorber with intercooling, condensate
heating, evacuation using water ejector, stripper overhead compression, lean amine flash, split-amine flow to
absorber and desorber, and their combinations. Le Moullec, et al. 16 classified these configurations into three
4|P a g e
Pages 166–179,
categories of (1) absorption enhancement, (2) heat integration and (3) heat pump applications. They
enumerated twenty process configurations from the open literature and patents. In general, up to 37% energy
17
saving in terms of the required reboiler steam was reported . Damartzis et al.18 applied a module-based
generalized design framework in order to optimize process flow diagram including the stream topologies, the
heat redistribution and the cascades of desorption columns. They reported significant economic improvement
(15%-35%) and reductions in the reboiler duty (up to 55%). However, as discussed by Karimi 19, a high degree of
energy integration may result in poor dynamic behaviour, because in energy integrated processes, disturbances
propagate in several paths. Therefore, a trade-off between energy saving and process controllability should be
established 20.
Nevertheless, integrated operation of carbon capture processes may not be realizable without considering the
main operational characteristics of the upstream power plant. Power plants are subjects to drastic variations in
the electricity demand. Examples of such variations include regular daily and hourly variations in the consumer
demand or stochastic variations such as extreme weather conditions or local events. It is expected that by the
introduction of renewable energy resources the fluctuations in the electricity grid will also increase on the supply
side, as some of these new resources such as solar or wind have intermittent generation characteristics. Thus,
it is for the fossil-based power plants to operate flexibly and balance the supply deficit in order to meet the
demand. Therefore, commercialization of new CO2 capture technologies strongly depends on their adaptability
in order to remain integrated as the upstream power plant experiences variations in the electricity demand.
Recently, the flexibility of solvent-based carbon capture processes has been the focus of various research groups.
21
Shah and Mac Dowell studied the multi-period operation of a coal-fired power plant. They adapted a time-
22
varying solvent regeneration strategy in order to minimize the costs of CO2 capture. Delarue, et al. had a
similar observation that flexible operation of capture plant would offer a better economy. Lawal, et al. 23 studied
the dynamic performance of carbon capture from a coal-fired sub-critical power plant. They concluded that the
capture plant has a slower dynamic response than the power plant, which can prolong the power plant start-up
or load-change due to steam extraction. In addition, it was observed that the interactions between the control
5|P a g e
Pages 166–179,
loops in the power plant and capture plant limit the overall process controllability. Bypassing the flue gas, solvent
storage and stripper-bypass can potentially offer flexibility and economic savings 24,25.
In the present paper, we explore model development and validation, scale up, NGCC power plant integration
and flexible operation of the capture processes. The research questions also include the interactions between
the power plant and carbon capture plant in terms of the flow and composition of the flue gas and the required
steam for solvent regeneration, which have implications for the overall energy efficiency and operational
flexibility. The performances of GCCmax and the MEA solvents for carbon capture from an NGCC power plant
are studied. GCCmax is recently developed by Carbon Clean Solutions Limited (CCSL) and belongs to the class of
amine-promoted buffer salt (APBS) solvents. MEA is chosen as the reference solvent for comparison.
Nevertheless, the research methodology is general in nature and can offer effective standards for carbon
capture solvent development and comparison.
In the following sections first, the overall process block diagram and the process flow diagram of the sub-
processes are presented and discussed. Then, the capture process model is briefly discussed and justified. The
discussions continue with the implications of various operating modes of the combined cycle gas turbine (CCGT)
for the flue gas flowrate and composition. These enable the application of an optimization framework for the
problem of retrofitting an existing NGCC power plant using solvent-based capture process. Finally conclusions
are made with respect to the implications GCCmax and MEA reference solvent in terms of technical and
energetic performance measures.
Overall process block diagram
The overall process block diagram is shown in Figure 1. As shown in this figure, the natural gas combined cycle
(NGCC) power plant comprises two trains of combined cycle gas turbines (CCGT), heat recovery and steam
generation (HRSG) systems which are integrated to the steam turbines at three high, medium and low pressure
levels. The advantages of parallel trains are due to the fact that in the presence of large variations in the
electricity demand, it is possible to shut down a gas turbine and operate the other train close to its nominal
operating conditions, i.e., at a high conversion efficiency. Based on a similar justification, in the present research,
separate carbon capture and compression trains are considered in the downstream process, in order to enable
6|P a g e
Pages 166–179,
flexible operation of the overall integrated process. Figure 1 shows that the NGCC power plant integrates with
the carbon capture plant at three points. The flue gas is sent from the power plant to the capture plant for CO 2
separation. In addition, the capture plant relies on the steam from the power plant for regeneration of the
solvent and it returns the condensates to the power plant for reuse and further steam generation.
Simulation Steam
Cleaned Flue
Compressor train
gas CO2
Condensates
Simulation
Electricity power
Natural
gas Heat recovery steam generator
Gas Turbine (HRSG)
Combustion
Air Carbon
Capture
Optimization
HP, MP, and LP Steam Electricity
power
Turbines
Carbon
Electricity power
Natural
gas Heat recovery steam generator Capture
Gas Turbine (HRSG)
Combustion
Air
Condensates
Simulation
Steam Cleaned Flue CO2

gas
Natural gas combined cycle (NGCC) power plants Compressor train
Figure 1. The block diagram for a natural gas combined cycle (NGCC) power plant integrated with CO 2 capture and CO2
compression processes.
Process flow diagram of natural gas combined cycle (NGCC) power plant
Figure 2 shows the process flow diagram of the natural gas combined cycle (NGCC) power plant, in more detail.
This process consists of three sub-processes, combined cycle gas turbine (CCGT), heat recovery and steam
generation (HRSG) system and stream turbines at high pressure (HP), medium pressure (MP) and low pressure
(LP) levels. Firstly, air is compressed and fed to the combustor where natural gas feed is burned in order to
release heat. Hot exhaust gases are expanded in the gas turbine in order to produce electricity. Then, the hot
gases are exploited in the HRSG for generating steam at the three pressure levels. The steam from the high
pressure steam drum is sent to the HP steam turbine. The discharge of HP turbine is mixed with the steam from
the medium pressure steam drum and is superheated in the economisers before entering the MP steam turbine.
The discharge of the MP steam turbine is mixed with the steam generated at the low pressure steam drum. A
7|P a g e
Pages 166–179,
fraction of this steam stream is extracted and sent to the reboiler of the CO2 capture plant for solvent
regeneration. The pressure and temperature of the extracted steam is adjusted in the desuperheater using the
condensates returning from the capture plant. In addition, a valve before the LP steam turbine ensures that the
extracted steam is at the desired pressure. Both HP and MP turbines are of the back-pressure type. However,
the LP turbine is a condensing turbine. The justification of using a condensing turbine is that the produced power
is proportional to the pressure ratio between suction and discharge. Therefore, it is possible to enhance the
produced work by creating vacuum conditions at the turbine discharge using a surface condenser. The
condensates from the surface condenser and the condensates returning from the carbon capture plant are
mixed, deaerated, pressurized and recycled to the steam drums for further steam generation.
Process flow diagram of CO2 capture and compression sections
Figure 3 shows the process flow diagram of the CO2 capture and compression sections. In the first column, the
flue gas from the power plant comes into direct contact with cooling water in order to reduce its temperature
and remove any entrained particles. In the next column, absorber, the CO 2 is chemisorbed and removed by the
solvent. The CO2-rich solvent leaves from the absorber bottom, and the cleaned flue gas exits from the absorber
top and is sent to the water wash column. The aim of the water wash column is minimizing the solvent loss by
absorbing the solvent spilled from the top of the absorber. The CO2-rich solvent from the bottom of the absorber
is sent to the top of the desorber for CO2 stripping and solvent regeneration. The CO2-lean solvent from the
desorber reboiler is recycled to the absorber for reuse and CO2 separation. The absorption reactions are
exothermic and favour low temperatures. By comparison, the desorption reactions are endothermic and favour
high temperatures. Therefore, there is an opportunity for heat integration between hot CO2-lean and cold CO2-
rich streams. The separated CO2 from the desorber condenser is sent to the compression section. The
compression section consists of seven compression stages. In each compression stage, due to pressure
enhancement, the temperature of the CO2 gas is increased, and needs to be cooled in the subsequent inter-
stage cooler. As a result of sequential compression and cooling, most of the water content is condensed in the
early stages. The remaining water is removed using an adsorption process in the dehydrators. The compressed
CO2 is sent from the last stage for storage and sequestration.
8|P a g e
Pages 166–179,
9|P a g e
Mahdi Sharifzadeh, Nilay Shah. Carbon capture from natural gas combined cycle power plants: Solvent performance comparison at an industrial scale.
AIChE Journal, 2016, Volume 62, Issue 1, Pages 166–179,
Hot flue gas
Natural gas CS-100

Combustion air
Combustor Exhaust
gases
P-105
M
Heat recovery steam generator (HRSG)
C-101 T-101 Electric Driver P-104
Air compressor Gas turbine
Flue Gas to
E-105 E-106 E-107 E-108 E-109 E-110 E-111 E-112 E-113 Carbon
Economizer Economizer Economizer Economizer Economizer Economizer Economizer Economizer Economizer
capture plant
E-101 E-102 E-103 E-104 D-101 D-102 D-103
Economizer Economizer Economizer Economizer High pressure Medium pressure Low pressure
steam drum steam drum steam drum
Multipliers
represent the
M M M M M
existence of
two parallel
trains
T-102 T-103 T-104

HP Steam Turbine MP Steam Turbine HP Steam Turbine Electricity generator
E-114
DS-100 Surface condenser
Desuperheater CWin
CWout
P-103
P-101 P-102
Condensate
Steam extraction
return
10 | P a g e
Figure 2. The process flow diagram of the natural gas combined cycle (NGCC) power plant.
11 | P a g e
Electric Driver
C-301 C-302 C-303 C-304 C-305

1st stage 2nd stage 3rd stage 4th stage 5th stage
compressor compressor compressor compressor compressor
CWin CWin CWin CWin CWin
D-300 D-301 D-302 D-303 D-304 D-305

Knock out 1st stage 2nd stage 3rd stage 4th stage 5th stage
drum separator separator separator separator separator
CWout CWout CWout CWout CWout

E-301 E-302 E-303 E-304 E-305 DH-305a-c
Drain
1st stage intercooler Condensates
2nd stage intercooler Condensates
3rd stage intercooler Condensates
4th stage intercooler Condensates
5th stage intercooler Condensates Dehydrators
purge
Cleaned C-306 C-307
flue gas 6th stage 7th stage
Water make-up compressor compressor
CWin CWin
T-203 CWin E-205 D-306 D-307

Water wash CWout Intercooler 6th stage 7th stage
separator separator
Captured CWout CWout

Solvent E-306 E-307
recovery CO2
6th stage intercooler 7th stage intercooler
E-202 CWin Condensates CO2 product
Desorber
Condenser
E-204 CWout
Lean Solvent
T-201 Cooler
Absorber
CWin CWout
T-202
Cleaned flue gas Solvent Desorber
Water Make-up
make-up
Water make-up
T-200 E-200 CWin
CWout TK-201
Direct contact Intercooler
cooler (DCC) Buffer Tank Steam
E-201
Fan Lean Solvent- Rich Solvent
Flue gas from F-201 Heat Exchanger Condensates
pre-scrubber E-203
To waste water
Desorber Reboiler
treatment
12 | P a g e
Figure 3. The process flow diagram of the CO2 capture and CO2 compression processes
13 | P a g e
Pages 166–179,
Research methodology
For the capture plant, the detailed of model development, pilot plant trials and model validation were reported
in a previous contribution 34. The present research builds upon this initial results and aims at evaluating the
performance of the GCCmax solvent in comparison with the MEA benchmark solvent, at an industrial scale and
when integrated to an NGCC power plant. In the following, firstly the problem statement for retrofitting an NGCC
power plant with carbon capture and compression processes is presented. Then, model development and
validation are briefly discussed. Then, the capture process model is scaled up and integrated to the power plant
model. The main feature of interest is uncertainties in the power plant electricity demand that require flexible
operation of the capture process in order to realize seamless process integration and retrofit. A novel
optimization framework is proposed to address the posed retrofit problem. In the proposed optimization
framework, the design and operation of the capture process are optimized simultaneously, and the interactions
of the down steam capture process with the upstream in terms of the flue gas flowrate and composition and
the required steam for solvent regeneration, as various electricity load scenarios are fully considered. Then, the
implementation software tools are elaborated upon. Finally, the results are reported and discussed.
Problem statement
The present research addresses the problem of retrofitting an existing natural gas combined cycle (NGCC) power
plant using solvent-based carbon capture and compression processes. The specifications of an existing NGCC
power plant including the nominal operating conditions and the performance curves of process equipment
under various partial load scenarios are given. It is intended to retrofit the power plant with CO2 capture and
compression plants, so that 90% of the CO2 from natural gas combustion is captured and compressed to 111 bar
for subsequent storage and sequestration. In addition, it is desired to ensure that the capture plant and its
compression network remain operable at a wide range (i.e., 25%-100%) of electricity power demands.
Capture plant model development and validation
The accurate modelling of the solvent-based CO2 capture processes for the purpose of solvent benchmarking
and comparison requires a thorough understanding of the underlying physical and chemical phenomena
involved, as discussed in the following.
14 | P a g e
Pages 166–179,
Rate-based modelling of gas-liquid contactors
The rate-based model of the gas-liquid contactor is founded on the two-film theory. In this method,
thermodynamic equilibrium is assumed only at the interface of vapour and liquid phases. Unlike equilibrium-
based models, the exiting vapour phase is superheated and the exiting liquid phase is subcooled and they have
different temperatures. The exchanged mass and energy between phases depend on the driving forces,
transport coefficients, and the interfacial area. Often, both convective and diffusive transport phenomena are
involved and component-coupling effects need to be considered 26. Various empirical correlations for calculating
the mass transfer coefficient are proposed by researchers for random 27-29 and structured packing 29-31. Finally,
the bulk liquid and gas phases may have different flow configurations such as plug or mixed flows.
Reaction kinetics and thermodynamics
In the present research, the statistical associating fluid theory (SAFT) was adopted for modelling the chemical
and phase equilibria. In this approach, the rate of reactions, the concentration of intermediate ionization
species, and their thermophysical properties are not formulated directly. Instead, CO2 and solvent are
represented as molecule chains with associating sites. The concentration of CO2 in association with the solvent
molecules represents the actual CO2 loading at different temperatures and pressures.
Justification of the modelling strategy
The combination of rate-based modelling and representation of chemical reactions using statistical associating
fluid theory (SAFT) provides a consistent modelling approach. The justification is that for solvents such as MEA
and GCCmax, the rate of reaction is significantly faster than the heat and mass transfer rates. Therefore, the
knowledge of the reaction kinetics is unnecessary and chemical equilibrium sufficiently describes the actual
system behaviour at the gas-liquid interface. This modelling approach offers several advantages; firstly unlike
activity-based models, the same equation of state is used to describe both liquid and vapour phases. Secondly,
the chemical equilibria are treated at the same level as phase equilibria. Furthermore, this approach results in
significant model reduction because the speciation of intermediate ions is not included in the mathematical
formulation and the uncertainties associated with their thermophysical parameters are disentangled from
problem formulation. Most of all, the aforementioned approach establishes a connection between the chemical
15 | P a g e
Pages 166–179,
and physical behaviour of the mixture and the molecular structure of the involved materials. This is of particular
importance to modelling new solvents as the required information can be acquired from the available data for
the molecular segments of associating sites.
In the present research, the applied software tools were advanced modelling library gas-liquid contactors (AML-
GLC) and gSAFT toolboxes developed by Process System Enterprise (PSE). The parametric values of
thermodynamic models for the GCCmax solvent are obscured in order to respect the confidentiality agreements
with Carbon Clean Solutions Limited (CCSL) and Process Systems Enterprise Ltd (PSE). The modelling equations
32
of the gas-liquid contactors and the underlying assumptions are described in . More details on the
thermodynamic model are given in reference 33.
Pilot plant studies and model validation
As discussed in a previous contribution 34, In order to ensure effective benchmarking and model validation, two
sets of pilot plant runs were conducted using the monoethanolamine (MEA) and GCCmax solvents. MEA served
as the baseline reference solvent. The pilot plant studies were conducted in the US National Carbon Capture
Center (NCCC) located in Alabama, USA. Table 1 shows the results of model validation for MEA reference solvent
34
. Since the US NCCC pilot plant was not previously operated under natural gas exhaust conditions, the model
validation was conducted based on historical data for a scenario of coal-fired exhausts. The last column in Table
1 reports the prediction of the model, when the system is operated for natural gas exhaust conditions. Table 1
shows a very good agreement between pilot plant data and simulation results, in terms of the captured CO2 and
the solvent composition. Minor discrepancies in the consumed steam are deemed to be associated with heat
losses and temperature indicator errors. Table 2 reports the results of the GCCmax solvent model validation
under natural-gas-fired conditions 34. Two sets of pilot plant data were used, which are different with respect to
the lean solvent temperature entering the absorber top. The justification was due to the fact that in different
parts of the world, cooling water may be supplied at different temperatures. Again the model predictions are
in good agreement with the pilot plant data with respect to the captured CO 2 and the solvent concentrations,
giving confidence in the model’s predictive capabilities. The discrepancies in the required steam and
temperatures were attributed to lack of insulation or temperature measurement errors.
16 | P a g e
Pages 166–179,
17 | P a g e
Pages 166–179,
Table 1. Model validation for MEA baseline solvent 34.

Alabama-coal Alabama-coal Alabama NG
Pilot Plant Simulation Simulation
Flue gas to the absorber
Nitrogen + Oxygen Mass Fraction 0.786 0.786 0.909
Carbon Dioxide Mass Fraction 0.165 0.165 0.067
Water Mass Fraction 0.049 0.049 0.024
Total flowrate Kg/s 0.6279 0.6280 0.6279
Flue Gas Temperature - Absorber Inlet K 316.56 316.56 316.56
Lean Solvent - Absorber Inlet
Amine Mass Fraction 0.297 0.298 0.296
CO2 Mass Fraction 0.063 0.059 0.059
Water Mass Fraction 0.640 0.643 0.645
Total Kg/s 2.5200 2.5100 1.3000
Lean solvent temperature K 316.15 316.15 316.15
Intercoolers outlet temperature K 316.15 316.15 -
Reboiler Steam
Steam pressure bar 2.92 2.92 2.92
Steam temperature K 405.60 405.60 405.60
Steam flowrate Kg/s 0.18 0.15 0.07
Lean-Rich Heat Exchanger
Lean in K 388.87 388.9 389
Lean out K 331.4 338.9 330.8
Rich in K 327.91 330.4 321.7
Reboiler temperature K 385.6 388.0 389.0
Absorber bottom pressure bar 1.1 1.17 1.16
Absorber top pressure bar 1.04 1.04 1.04
Desorber bottom (reboiler) pressure bar 1.71 1.71 1.16
Desorber top pressure bar 1.69 1.70 1.70
General specifications
CO2 capture target % 91.84 91.85 90
Inter-stage Cooling Yes Yes No
18 | P a g e
Pages 166–179,
Table 2. Model validation for GCCmax solvent 34.

Data Set 1 Data Set 1 Data Set 2 Data Set 2
Absorber inlet gas stream Pilot Plant Simulation Pilot Plant Simulation
Nitrogen + Oxygen Mass Fraction 0.895 0.895 0.896 0.896
Carbon dioxide Mass Fraction 0.065 0.065 0.065 0.065
Water Mass Fraction 0.040 0.040 0.039 0.039
Temperature K 313.1 313.1 312.8 312.8
Total flowrate kg/s 0.995 0.995 0.9951 0.9951
Absorber
Absorber top pressure bara 1.160 1.160 1.160 1.160
Absorber bottom pressure bara unavailable 1.224 unavailable 1.221
Absorber outlet CO2 concentration Mass Fraction 0.0053 0.0056 0.004 0.005
Lean solvent - absorber inlet temperature K 304.1 304.1 325.4 325.4
Rich solvent - absorber outlet temperature K 318.1 316.5 318.7 316.1
Lean solvent flowrate kg/s 0.857 0.857 0.756 0.756
Desorber (regenerator)
Desorber Bottom Temperature K 388.6 388.6 395.4 395
Desorber Top Pressure bara 1.701 1.708 2.031 2.03
CO2 stream kg/s 0.0578 0.0611 0.060 0.0605
Reboiler Steam
Steam pressure bar 3.606 3.605 4.075 4.075
Steam temperature K 402.8 402.8 408.5 408.5
Steam condensate Temperature K 401.9 401.9 407.8 407.5
Steam flowrate kg/s 0.091 0.078 0.086 0.0823
Lean-Rich Heat Exchanger Temperatures
Lean solvent in K 387.4 388.6 394.2 395.0
Lean solvent out K 325.0 322.6 324.7 327.4
Rich solvent in K 318.9 316.5 319.9 316.1
Rich solvent out K 380.4 379.9 384.0 379.9
Lean solvent concentration
GCCmax Solvent Mass Fraction 0.410 0.410 0.439 0.439
Water Mass Fraction 0.536 0.541 0.504 0.515
CO2 Mass Fraction 0.054 0.049 0.057 0.046
General specifications
CO2 Capture target % 89.1 91.7 92.50 92.13
Inter-stage Cooling No No No No
19 | P a g e
Pages 166–179,
20 | P a g e
Pages 166–179,
The validated pilot plant model was used to extract several technical key process indicators (KPIs) which are
important measures that quantify the difficulties associated with CO2 separation from the flue gas in terms of
the required heating and cooling duties, required packing, and solvent circulation. These measures are scaled
with respect to the amount of pure CO2 captured, to become independent of the pilot plant throughput and
enable comparisons. Table 3 shows the key process indicators for the baseline solvent (MEA) and the GCCmax
solvent 34. The first indicator is the heating duty, in terms of the required energy needed in the desorber reboiler
for separating 1 ton of CO2. Around a 25.4-29.4% reduction in heating duty was observed. Furthermore, a
comparison between the values of the second KPI, suggest significant reductions (73-84.4%) in the cooling
duties. The third key process indicator is concerned with the volume of packing in the absorber and desorber
columns showing improvements in the case of GCCmax solvent. The last KPI is concerned with the required
solvent circulations, and is an indicator of the electricity power needed for pumping. The observed
improvements are between 58.8-64.7%.
Table 3. Key process indicators (KPIs) for the GCCmax solvent and baseline MEA solvent 34.
Key Process Indicators (KPIs) Unit MEA GCCmax (Data1) GCCmax (Data2)
Heating duty (MJ/ton CO2) 3986 2813 2975
Cooling duty (MJ/ton CO2) 5644 1524 884
Volume of packing (m3/ton CO2 hr-1 ) 46.619 45.64 45.91
Solvent circulation flowrate (ton solvent/ton CO2) 34 14 12
21 | P a g e
Pages 166–179,
Process Scale-up
The validated model was applied for analysis at the large scale corresponding to the retrofitted power plant. The
assumptions behind process scale up are summarized in the following. The bulk liquid and gas phases are
assumed to be well-mixed at each stage. Phase equilibrium was assumed only at the vapour–liquid interface. It
was assumed that the reaction kinetics is significantly faster than the heat and mass transfer rates and therefore,
equilibrium chemical reactions sufficiently represent the species composition at the gas-liquid interface. In the
present study, the effects of solvent degradation and heat losses were not considered. In practice, for large-
scale CO2 capture processes, achieving the aforementioned performances will require effective gas and liquid
distributors. In addition, the process should be carefully insulated and the composition of the solvent should be
tightly controlled using make-up.
Solution algorithm: Simulation-optimization framework
The aforementioned problem statement falls into the category of Integrated Process Design and Control (IPDC).
The motivation of the integrated approach, as opposed to sequential process design and control design, is due
to the fact that when the process design is fixed, there is little room left to improve its operational performance.
Therefore, it is highly recommended that operational characteristics should be considered at the early design
stages (i.e., process retrofit in the context of this research). A comprehensive review of the methods for
integrated process design and control is provided by Sharifzadeh 20.
The challenge is that the full-space formulation of integrated process and control design for large scale industrial
problems such as the abovementioned retrofit problem results in numerically intractable optimization
problems. Therefore, an objective of the present research was to identify critical process variables and ensure
process operability at the plant-wide level, with a reasonable computational complexity. To this end, a novel
simulation-optimization framework was developed and tailored for the above-mentioned retrofit problem, as
shown in Figure 1 and discussed in the following.
The proposed optimization framework is shown in Figure 1. Here, the overall process is decomposed into three
parts; simulation is used for the power plant and compression plant; optimization is applied to the capture
process. These three parts are linked together through flow of materials and energy. As shown in Figure 1, the
22 | P a g e
Pages 166–179,
CO2 capture process receives the flue gas from the NGCC power plant and depends on the steam supply for
regeneration of the solvent. The flowrate of flue gas depends on the electricity power demand and changes as
the NGCC power plant experiences variations. In addition, the composition of the flue gas depends on the ratio
of the combustion air and natural gas. However, the ratio of combustion air and natural gas is subject to
constraints on (1) the maximum allowable temperature of the turbine suction area, and (2) maximum allowable
temperature of the turbine discharge gases. Therefore, the operation of CO2 capture plant is highly entangled
with the operational procedure followed for power plant load reduction. As will be shown in the first part of the
Results section, the second constraint (i.e., maximum allowable temperature of the turbine discharge gases)
becomes active first and by its satisfaction, the first constraint is automatically met.
The variables involved in the optimal design of the CO2 capture process can be classified as (i) process design
variables and (ii) process control variables. The differentiation is necessary as process design variables (such as
the dimensions of process equipment) have a physical realization. After the process is commissioned, they are
fixed and cannot be changed without costly process modifications. By contrast the control variables (such as
the flowrate of the reboiler steam or the circulation rate of the solvent) are available during the process
operation in order to adapt the capture process to the variations in the upstream power plant.
In the proposed optimization framework, without loss of generality, we focus on optimizing the capture plant in
order to manage the numerical size of the problem. The solution algorithm for the optimization framework is as
follows:
Algorithm I:
Step (1) The power plant model is run for a series of steady-state electricity load reduction (100%, 75%, and
50%) scenarios, and a series of default values for the extracted steam and condensate recycle rates. The
results of the simulation will determine the flowrate and composition of the flue gas in each scenario.
Step (2) Given the flowrate and composition of the flue gas at various load reduction scenarios, the design and
control variables of the capture plant are optimized (as discussed in the following).
Step (3) The results of the optimization determine the optimal values of the extracted steam and recycled
condensates. These values are compared to the previous values of the extracted steam and recycled
23 | P a g e
Pages 166–179,
condensates and if the differences are less than the tolerance, the solution is found. Otherwise, the value
of the extracted steam and recycled condensates are updated in the power plant model and the algorithm
is repeated from Step (1).
Note the compression section does not have mutual interaction with the power plant and capture process. The
required energy for CO2 compression is calculated once the above iterative calculation (Steps 1-3) is converged.
In the present study, the economic analysis was concerned only with the capture process. However, the
energetic study also studied the interactions between the power plant and capture plant in terms of the required
steam and flowrate and composition of the flue gas in addition to the electricity power required for CO 2
compression.
The abstract formulation of the proposed optimization program (blue envelope in Figure 1) is as follows:
𝐍𝒔
𝐎𝐛𝐣𝐞𝐜𝐭𝐢𝐯𝐞 = 𝐄(𝐓𝐀𝐂𝐬 ) = ∑ 𝛍𝒔 × 𝐓𝐀𝐂𝐬 𝐏𝐫𝐨𝐛𝐥𝐞𝐦 − 𝟏

𝒔=𝟏
Subject to
Constraints associated with first principles (transport phenomena, thermodynamics)
Technical Constraints: maximum reboiler temperature (limited by the possibility of solvent degradation)
Control Constraints: 90% CO2 Capture, maximum turbine discharge temperature
Disturbances: Composition and flowrate of flue gas for various power load reduction scenarios
Process design decision variables: The dimensions of absorber, desorber, and heat exchangers
Control (recourse) decision variables: Circulation flowrate, Reboiler steam flowrate
In the above formulation, 𝐄 is the expected value, 𝒔 is the index of the load reduction scenarios, 𝛍𝒔 is the
likelihood of each scenario and 𝐍𝒔 is the total number of scenarios. 𝐓𝐀𝐂 refers to the total annualized cost (TAC)
of the capture plant, and was calculated as:
𝐹𝑖𝑥𝑒𝑑 𝐶𝑎𝑝𝑖𝑡𝑎𝑙 𝑖𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡

𝑇𝑜𝑡𝑎𝑙 𝐴𝑛𝑛𝑢𝑎𝑙𝑖𝑧𝑒𝑑 𝐶𝑜𝑠𝑡𝑠 = + 𝑇𝑜𝑡𝑎𝑙 𝐴𝑛𝑛𝑢𝑎𝑙 𝐸𝑛𝑒𝑟𝑔𝑦 𝑐𝑜𝑠𝑡𝑠 (1)
𝑃𝑙𝑎𝑛𝑡 𝑒𝑓𝑓𝑒𝑐𝑡𝑖𝑣𝑒 𝐿𝑖𝑓𝑒
where the value of 5 years was considered for the capture plant effective life, in order to combine the plant life
and the time value of money. The costs of process equipment were calculated according to the costing
correlations provided in 35. A Lang factor of 6 was considered for estimating the total capital investment 36. The
24 | P a g e
Pages 166–179,
utility costs considered were 65 $/MWh for electricity 37, 0.048 $/tonne for cooling water 38, and 14.5 $/tonne
for steam. The MEA solvent loss is around 1400 mg/m3 of flue gas. Equivalent value for the GCCmax is around
28 mg/m3. However, since the GCCmax is not priced yet, the costs of solvent losses is not included in the
objective function. Solvent degradation was not considered in this study. The considered load reduction
scenarios were 100%, 75% and 50% and were assumed to be equally likely. Since the overall process (Figure 1)
consists of two parallel trains, the 50% load reduction in each train will be sufficient to realize a large range of
potential operational part-load scenarios (25-100%). The part-load operation of power plants is limited to their
turndown ratio (approximately 50%). The turndown ratio is dictated by the technical limitations such as
excessive pressure drops across the power plant, or the surge margins of the compressor and turbines. In the
present research, three operational scenarios were considered; in the first scenario both gas turbines are
operated at full-load (100%). This scenario refers to the highest conversion efficiency, i.e., the highest CO2
concentration and smallest steam demand per ton of captured CO2. In the third scenario, both gas turbines are
operated at 50%, which refers to the worst conversion efficiency and hence, the lowest CO2 concentration and
the largest steam demand per ton of captured CO2. The second scenario is intermediate, where both gas
turbines are operated at 75%. These scenarios cover all the operating regions thoroughly. It is notable that there
are other operating scenarios where the gas turbines could be operated at different loads (e.g., operating one
of trains at full load and shutting down the other), which could be more energy efficient. However, the
aforementioned scenarios are more comprehensive with respect to CO2 concentration and flowrate.
From the optimization programming point of view, the above formulation conforms to a two-stage recourse-
39
based optimization under uncertainty . From the Control Engineering point of view the above formulation
conforms to a steady-state inversely controlled process model (ICPM) 40, 41. Here the treatment is based on the
property that the inverse solution of process model can be applied in order to evaluate the best achievable
41
control performance. The idea is shown in Figure 4, adapted from . In a steady-state inversely controlled
process model, the values for the manipulated variables (MVs) required for maintaining the controlled variables
(CVs) at constant setpoints are calculated using the inverse of process model. As discussed by Sharifzadeh 20, 41
using this strategy, it is possible to ensure that the process remains operable under various disturbance scenarios
25 | P a g e
Pages 166–179,
(i.e., electricity load reduction). It is notable that application of a dynamic inversely controlled process model 42
also enables studying the process controllability during transient states. However, we defer such detailed
analysis to our future research. In the context of the present study concerning carbon capture from the NGCC
power plant, two model inversions were conducted. Firstly, the temperature of the turbine discharge gases (as
discussed and justified later in the result section) is chosen as the controlled variable (CV). The corresponding
manipulated variable (MV) is the flowrate of the combustion air which is varied in order to maintain the
temperature of the turbine discharge gases constant at its maximum allowable value. The second controlled
variable was the CO2 capture target. Here, the corresponding manipulated variables are the reboiler steam
flowrate and the solvent circulation rate which are optimally varied in order to keep the controlled variable at
the 90% CO2 capture target. It is notable that in the context of present study, the NGCC power plant model is
applied in order to derive realistic disturbance scenarios (red envelop in Figure 4) in terms of the flowrate and
composition of the flue gas for the carbon capture process. The challenge is that there are mutual interactions
between the capture process and the power plant, shown by arrays in Figure 4. While the steam needed for
solvent regeneration depends on the flowrate and composition of the flue gas, the overall fuel consumption and
hence the flue gas itself, also depends on the required steam in the desorber reboiler. In the present study such
mutual interactions is captured using the iterative steps in Algorithm I, as outlined earlier.
The value of the

Carbon capture target (i.e., 90%)
objective function
(i.e., TAC) Capture
process
Optimization The required values of

algorithm the manipulated
variables (solvent A steady-state inversely controlled
circulation flowrate, process model
reboiler steam
The values of process design flowrate)
(column dimensions, heat-
exchanger size) and operational Condensates Steam Flue gas
(column pressures and
temperatures) optimization NGCC Power
plant model Disturbance
variables
scenarios
Figure 4. Optimizing a steady-state inversely controlled process model, adapted from 40.
26 | P a g e
Pages 166–179,
Model development and implementation software tools
The NGCC power plant and compression process were modelled in gCCS 43, a software tool developed by Process
Systems Enterprise Ltd (PSE). The specification of the NGCC power plant model was received from PSE from one
of their earlier industrial projects. The important characteristics of the developed model were calculation of the
efficiency of the compressors and turbines using performance curves and calculation of material flowrates based
on pressure differences. The capture plant model was developed using the Advanced Model Library for Gas-
44
Liquid Contactors (AML:GLC) and gSAFT 45. As described extensively earlier, the main characteristics of the
capture process model were rate-base modelling of mass and heat transfer phenomena and representation of
chemisorption reactions using SAFT equation of state. The heat-exchangers were modelled using gCCS in the
operational mode. The implication is that the surface area was an optimization variable, and given the heat
transfer coefficient, the temperatures of the hot and cold streams were calculated. In the present study, the
gPROMS default values for the solution parameters were used (e.g., 10-5 for absolute tolerance). Similar to other
NLP algorithms, the solution time depends on the initial guess for the optimization variables, and typically takes
1-2 days to converge.
Results of optimization programming
The Results Section is organized as follows. Firstly, it is investigated how the CCGT control strategy influences
the flue gas composition and flowrate. These discussions enable underpinning the interactions between the
power plant and capture process during electricity load reduction scenarios. Then, the results of the optimization
Problem 1 are reported and discussed. Finally, the implications of NGCC power plant retrofit and integration
with capture plant for the overall energy conversion are evaluated and discussed.
Control strategy for combined cycle gas turbine (CCGT)
This section discusses the operation of CCGT at steady-state which has profound implications for the flowrate
and composition of the flue gas. When the power plant is operated at full load, the ratio of the combustion air
and natural gas flowrates is adjusted in order to maximize energy conversion. However, as the electricity power
demand is reduced, the flowrate of natural gas is reduced accordingly and maintaining a constant ratio with
combustion air flowrate would increase the temperature of the combustor exhaust gases and turbine discharge
27 | P a g e
Pages 166–179,
gases which could potentially damage the process equipment. Therefore, a control strategy is needed that
systematically safeguards the process equipment.
Setpoint Setpoint
(a) (b)
FI FC FI FC
Natural Natural
gas gas
Exhaust Discharge Exhaust Discharge
Combustion air gases gases Combustion air gases gases
TI TC
TI TC
Air compressor Gas turbine Air compressor Gas turbine
Figure 5. The Control structure for combined cycle gas turbine (CCGT) during power plant load reduction: (a) the
temperature of combustor exhaust gases is controlled (b) the temperature of turbine discharge gases is controlled.
28 | P a g e
Pages 166–179,
In practice, there are two control structures in use 46, shown in Figures 5a and 5b. Both control structures have
a similar control loop in that, the setpoint of the natural gas flow controller is adjusted according to the electricity
power demand. However, the two control structures differ in the selected controlled variable in the second
control loop. In the first control structure (Figure 5.a), the temperature of the combustor exhaust gases is
controlled. By comparison, in the second control structure (Figure 5.b), the temperature of the turbine discharge
gases is controlled. The first control strategy requires a recurrent procedure. The reason is that maintaining the
temperature of the combustor exhaust gases at a constant value results in an increase in the temperature of the
turbine discharge gases which can damage the downstream HRSG section. Therefore, the operational strategy
in the first control structure consists of two iterative control modes:
Mode (i): The flowrate of the natural gas is reduced while the flowrate of the combustion air is maintained
constant (dotted line in Figure 6a). This results in a reduction in the temperature of the turbine discharge
gases (descending dotted line in Figure 6c).
Mode (ii): The temperature of combustion exhaust gases is controlled using the combustion air flowrate
(constant dotted line in Figure 6d) as the flowrate of the natural gas is further reduced. This results in an
increase (ascending dotted line in Figure 6c) in the temperature of the turbine discharge gases until it
reaches a limit where there is a risk of thermal shock to the downstream equipment. The control system
switches to Mode (i).
Unlike the first control strategy, the second control strategy requires only one operational mode. The reason is
that by controlling the temperature of the turbine discharge gases (solid line in Figure 6c) the temperature of
combustion exhaust gases decreases (solid line in Figure 6d). In other words, the constraint on the turbine
discharge temperature becomes active first and automatically satisfies the constraint on the combustion
exhaust temperature. Figures 6a and 6b suggest that the second control strategy is optimal with respect to the
CO2 separation as it produces less flue gas with a higher CO2 content, i.e., easier carbon capture task.
29 | P a g e
Pages 166–179,
(a) (b)
0.06
620
CO2 mass fraction in the flue gas

600
0.055
580
Flow rate of flue gas (kg/s)
560
0.05
540
520 0.045
500 Control structure (a) Control structure (a)
480 0.04
Control structure (b) Control structure (b)
460
440 0.035
7 8 9 10 11 12 13 14 7 8 9 10 11 12 13 14
Flow gas of natural gas (kg/s) Flow gas of natural gas (kg/s)
(c) 1500 (d)

900 Combustion exhaust temperature (oC)
Turbine discharge temperature (oC)
890 1450
880
1400
870
860 1350
850 1300
Control structure (a) Control structure (a)
840
1250 Control structure (b)
830 Control structure (b)
820 1200
7 8 9 10 11 12 13 14 7 8 9 10 11 12 13 14
Flow gas of natural gas (kg/s) Flow gas of natural gas (kg/s)
Figure 6. The flowrate of flue gas (a), The CO2 mass fraction of flue gas (b), the temperature of turbine discharge gases (c)
and the temperature of combustion exhaust gases (d) for the control structures (a) and (b) in Figure 5.
Overall energy conversion efficiency and implications of carbon capture and compression
Table 4 reports the results summary for the scenario in which the capture process is operate with the GCCmax
solvent. The features of interest include the flowrate of natural gas feed, the flowrate and composition of the
flue gas, the generated power, the required steam for solvent regeneration, the power needed for CO 2
compression, the cost of produced electricity and the overall energy efficiency. Similar results are reported in
Table 5 where the MEA reference solvent is used. In both scenarios, the flowrate of natural gas is gradually
reduced from the nominal value of 26.87 kg/s by almost 50% and the design and operation of the capture plant
are optimized according to the simulation-optimization framework shown in Figure 1. These Tables exhibit
common observations regarding the implications of electricity load reduction for power generation with CO2
capture. In all scenarios, CO2 capture and compression impose energetic penalties in terms of the required steam
for solvent regeneration and the electric power needed for CO2 compression. In addition, as the electricity load
30 | P a g e
Pages 166–179,
is decreased, the energy conversion efficiency is reduced, due to the reduced efficiency of process equipment
such as turbines and compressors. The combination of these penalties reduces the net produced electricity and
decreases the overall energy efficiency.
Table 4. The results of flexible operation of NGCC power plant for various electricity load, with and without
CO2 capture and compression plants: GCCmax solvent.
Nominal a 100% load 75% load 50% load
NG flowrate b kg/s 26.87 26.87 21.08 15.25
Flue gas flowrate kg/s 1214.8 1214.8 1022.6 801.5
Flue gas composition: N2 Mass fraction 0.7601 0.7601 0.7611 0.7623
Flue gas composition: O2 Mass fraction 0.1169 0.1169 0.1230 0.1294
Flue gas composition: H2O Mass fraction 0.0647 0.0647 0.0615 0.0581
Flue gas composition: CO2 Mass fraction 0.0583 0.0583 0.0544 0.0502
Generated power in NGCC b MW 747.18 698.78 510.43 341.75
Extracted steam b kg/s - 68.24 52.4 36.76
Power consumed in compressors b MW - 20.68 15.88 11.64
Net produced electricity b MW 747.18 678.1 494.55 330.1
Energy content of feed (HHV) b MW 1292.62 1292.62 1014.07 733.66
CO2 captured Kg/s 63.74 63.74 50.07 36.20
Electricity costs $/MWh 65.00 71.62 77.04 83.50
Overall conversion efficiency % 57.8 52.46 48.77 44.99
Notes: a Nominal refers to the standalone scenario where the power plant is operated at its nominal operating point without
CO2 capture and compression plants. b the reported flowrates and power values are for the overall process and include the
two trains of CCGT, HRSG, CO2 capture and compression sections.
Table 5. The results of flexible operation of NGCC power plant for various electricity load, with and without
CO2 capture and compression plants: MEA baseline solvent.
Nominal a 100% load 75% load 50% load

NG flowrate b kg/s 26.87 26.87 21.08 15.25
Flue gas flowrate kg/s 1214.8 1214.8 1022.6 801.5
Flue gas composition: N2 Mass fraction 0.7601 0.7601 0.7611 0.7623
Flue gas composition: O2 Mass fraction 0.1169 0.1169 0.1230 0.1294
Flue gas composition: H2O Mass fraction 0.0647 0.0647 0.0615 0.0581
Flue gas composition: CO2 Mass fraction 0.0583 0.0583 0.0544 0.0502
Generated power in NGCC b MW 747.18 679.94 495.89 331.09
Extracted steam b kg/s - 98.12 77.16 54.59
Power consumed in compressors
b MW - 20.68 15.88 11.64
Net produced electricity b MW 747.18 659.26 480.01 319.45

CO2 captured Kg/s 63.74 63.74 50.07 36.20
Energy content of feed (HHV) b MW 1292.45 1292.45 1013.95 733.53
Electricity costs $/MWh 65.00 73.67 79.38 86.29
31 | P a g e
Pages 166–179,
Overall conversion efficiency % 57.8 51.01 47.34 43.55

Notes: a Nominal refers to the standalone scenario where the power plant is operated at its nominal operating point without
CO2 capture and compression plants. b the reported flowrates and power values are for the overall process and include the
two trains of CCGT, HRSG, CO2 capture and compression sections.
The implications of load reduction for operation of the capture plant are more convoluted. To enable the
discussions more details are provided in Tables 6 and 7 which report the design and operational specifications
for the load reduction scenarios, in the case of GCCmax and MEA solvents, respectively. As the electricity load is
reduced, the concentration of CO2 in the flue gas (Tables 4 and 5) decreases, which suggests a more difficult
separation task. On the contrary, more contact area (shown by packing volume KPI in Tables 6 and 7) becomes
available between the gas and liquid phases. Then, it is for the optimization algorithm to adjust the solvent
circulation rate and reboiler steam for each electricity load scenario and establish a trade-off between the capital
investment and the energy costs. Overall a minor decrease in the heating and cooling indicators and solvent
circulation indicators are observed for load reduction scenarios. Another important feature of interest is the
design and operation of the absorber column. The absorber experiences the largest variations during load
reduction due to drastic variations in the flue gas flowrates. While the desired extent of CO2 capture constrains
the required gas-liquid contact area, a tall/thin column would result in very high pressure drops at full load
operation and a short/fat column would result in channelling during part-load operation. Therefore, it was for
the optimization algorithm to find a compromise design which satisfies the CO2 capture constraint and ensures
process operability in all load reduction scenarios. Tables 6 and 7 suggest that the optimized columns were
neither fat nor thin but almost square. The justification for the large heat transfer areas is the fact that the
overall economy is governed by the required reboiler steam. Such a large heat transfer area may require special
equipment such as plate heat exchangers. Since the heat-transfer area was the same in all scenarios, the
approach temperatures are smaller in part-load scenarios as less solvent is circulated. Finally, a comparison
between the KPIs in Tables 6 and 7 suggests that GCCmax features superior performance as it required 45% less
column packing, 30% less steam, 54% less cooling water, and 7% less pumping energy (shown by solvent
circulation rate), per unit mass of captured CO2. The estimated total annualized costs of the capture process at
full-load were 3.15×107 $/year and 3.91×107$/year for GCCmax and MEA solvents, respectively.
32 | P a g e
Pages 166–179,
33 | P a g e
Pages 166–179,
Table 6. The results of GCCmax solvent for various load reduction scenarios (all the results are reported for
one train)
50% load 75% load 100% load
Absorber
Diameter m 13.58 13.58 13.58
Length m 12.47 12.47 12.47
Absorber top pressure Pa 1.29E+05 1.18E+05 1.03E+05
Absorber bottom pressure Pa 1.35E+05 1.35E+05 1.35E+05
Lean Solvent to absorber
Flowrate kg/s 840.622 1139.86 1383
Temperature K 313.15 313.15 313.15
Water Mass fraction 0.5036 0.5036 0.5036
CO2 Mass fraction 0.0864 0.0864 0.0864
GCCmax Mass fraction 0.4100 0.4100 0.4100
Area m2 69398 69398 69398
Lean inlet temperature K 384.3 384.3 384.3
Lean outlet temperature K 328.1 329.5 330.8
Rich inlet temperature K 327.0 327.9 328.4
Rich outlet temperature K 383.9 383.4 382.8
Desorber
Diameter m 6.59 6.59 6.59
Length m 6.11 6.11 6.11
Reboiler
Reboiler pressure Pa 2.21×105 2.21×105 2.21×105
Stream flowrate kg/s 18.38 26.20 34.12
Steam inlet pressure Pa 3.61×105 3.61×105 3.61×105
Steam inlet temperature K 402.8 402.8 402.8
Condenser temperature K 313.15 313.15 313.15
Lean solvent cooler temperature K 313.15 313.15 313.15
Carbon capture target % 90.0 90.0 90.0
Key process indicators (KPIs)
Packing volume m3/ (tonne CO2 hr) 30.9 22.4 17.5
Heating duty MJ/tonne CO2 2166 2251 2348
Cooling duty MJ/ tonne CO2 1990 2179 2305
Circulation rate ton solvent/ tonne CO2 46.7 46.5 45.7
Total Purchased Equipment costs $ 8.47×106 8.47×106 8.47×106
Annualized Energy Costs $/year 1.13×107 1.62×107 2.13×107
Total Annualized Costs (TACs) $/year 2.15×107 2.63×107 3.15×107
34 | P a g e
Pages 166–179,
Table 7. The results of MEA baseline solvent for various load reduction scenarios (all the results are reported
for one train)
50% load 75% load 100% load
Absorber
Diameter m 14.99 14.99 14.99
Length m 14.75 14.75 14.75
Absorber top pressure Pa 1.28E+05 1.20E+05 1.03E+05
Absorber bottom pressure Pa 1.35E+05 1.35E+05 1.35E+05
Lean Solvent to absorber
Flowrate kg/s
Temperature K 313.9 313.9 313.9
Concentration
Water Mass fraction 0.6504 0.6504 0.6504
MEA Mass fraction 0.2820 0.2820 0.2820
CO2 Mass fraction 0.0676 0.0676 0.0676
Area m2 60174.5 60174.5 60174.5
Lean inlet temperature K 388.6 388.6 388.7
Lean outlet temperature K 332.7 334.3 334.3
Rich inlet temperature K 326.8 327.7 327.9
Rich outlet temperature K 384.3 382.7 381.6
Desorber
Diameter m 11.47 11.47 11.47
Length m 10.20 10.20 10.20
Reboiler
Reboiler pressure Pa 1.85×105 1.85×105 1.85×105
Stream flowrate kg/s 27.30 38.58 49.06
Steam inlet pressure Pa 3.05×105 3.05×105 3.05×105
Steam inlet temperature K 400.6 400.6 400.6
Condenser temperature K 313.9 313.9 313.9
Lean solvent cooler temperature K 313.9 313.9 313.9
Carbon capture target % 90.0 90.0 90.0
Key process indicators (KPIs)
Packing volume m3/ (tonne CO2 ×hr) 55.95 40.7 32.2
Heating duty MJ/tonne CO2 3241.2 3329.9 3348.2
Cooling duty MJ/ tonne CO2 4754.6 4991.7 4998.2
Circulation rate ton solvent/ tonne CO2 53.1 51.6 49.3
Total Purchased Equipment costs $ 8.28×106 8.28×106 8.28×106
Annualized Energy Costs $/year 1.54×107 2.18×107 2.76×107
Total Annualized Costs (TACs) $/year 2.64×107 3.31×107 3.91×107
35 | P a g e
Pages 166–179,
Conclusions
The present research studied scale up and integration of a solvent-based carbon capture process into a natural
gas combined cycle (NGCC) power plant for a novel solvent, GCCmax, and the MEA reference solvent. The aim
was to establish and quantify the superior performance of the new solvent at an industrial scale. Furthermore,
the present research provided in-depth insights into retrofit and flexible operation of NGCC power plants. It was
observed that the control strategy for the combined cycle gas turbine (CCGT) during load reduction, has
profound implications for the flowrate and composition of flue gas, and hence affects carbon capture costs. It
was also observed that NGCC power plants are less efficient at part-load operational scenarios. In the present
research, the method of integrated process design and control was adapted and solved. The proposed
optimization algorithm successfully established a trade-off between the design and operational criteria. The
overall total annual costs in terms of capital investment and energy costs were minimized while the process
operability was ensured under all load reduction scenarios.
Since comparison between various economic analysis available in open literature is challenging due to different
scope of system analysis, modelling details and the economic estimation methods, and in the absence of
economic data from industrial-scale demonstration plants, the present study chose to apply a set of key process
indicators (KPIs) enabling objective and reproducible comparisons. In all scenarios the GCCmax performed better
KPIs than the MEA reference solvent. GCCmax belongs to the family of the amine-promoted buffer salt (APBS)
solvents. It features a lower heat of absorption compared to MEA and its kinetics is enhanced by a buffer salt.
This combination enables GCCmax to require less regeneration energy and to feature a higher CO2 loading,
resulting in a superior performance compared to the MEA benchmarks. While the comparative study was
tailored to the aforementioned solvents, the research methodology is generic and provides effective standards
and benchmarking criteria for new solvent development.
36 | P a g e
Pages 166–179,
Acknowledgements
The authors would like to acknowledge the financial support by Carbon Clean Solutions (CCSL) under UK-
Department of Energy & Climate Change (DECC) grant. We are also thankful to Process Systems Enterprise Ltd
(PSE) for technical support and providing modelling libraries.
37 | P a g e
Pages 166–179,
References
1. International Energy Agency (IEA). World energy outlook. Paris: International Energy Agency (IEA), 2014.
2. Bottoms RR. Process for separating acidic gases. US patent no. 1783901, 1930.
3. International Energy Agency (IEA). Greenhouse gas R&D Programme. Retrofit of CO2 capture to natural gas
combined cycle power plants. Paris: report number 2005/1.
4. National Energy Technology Laboratory (NETL). Cost and performance baseline for fossil energy plants,
Volume 1: Bituminous coal and natural gas to electricity. Revision 2. November 2010, report number
DOE/NETL-2010/1397.
5. Manzolini G, Sanchez Fernandez E, Rezvani S, Macchi E, Goetheer ELV, Vlugt TJH. Economic assessment of
novel amine based CO2 capture technologies integrated in power plants based on European
Benchmarking Task Force methodology. Applied Energy. 2015; 138: 546–558.
6. Lucquiaud M, Gibbins J. On the integration of CO2 capture with coal-fired power plants: A methodology to
assess and optimise solvent-based post-combustion capture systems. Chemical Engineering Research and
Design. 2011; 89: 1553-1571.
7. Romeo LM, Bolea I, Y Lara, Escosa JM. Optimization of intercooling compression in CO 2 capture systems.
Applied Thermal Engineering. 2009; 29: 1744-1751.
8. Karimi M, Hillestad M, Svendsen HF. Capital costs and energy considerations of different alternative stripper
configurations for post combustion CO2 capture. Chemical Engineering Research and Design. 2011; 89:
1229-1236.
9. Pfaff I, Oexmann J, Kather A. Optimised integration of post-combustion CO2 capture process in greenfield
power plants. Energy. 2010; 35: 4030-4041.
10. Jonshagen K, Sipöcz N, Genrup M. A novel approach of retrofitting a combined cycle with post combustion
CO2 capture. Journal of Engineering for Gas Turbines and Power. 2010; 133: 011703-1-011703-7.
11. Khalilpour R, Abbas A. HEN optimization for efficient retrofitting of coal-fired power plants with post-
combustion carbon capture. International Journal of Greenhouse Gas Control. 2011; 5: 189-199.
38 | P a g e
Pages 166–179,
12. Biliyok C, Yeung H. Evaluation of natural gas combined cycle power plant for post-combustion CO2 capture
integration. International Journal of Greenhouse Gas Control. 2013; 19: 396-405.
13. Luo X, Wang M, Chen J. Heat integration of natural gas combined cycle power plant integrated with post-
combustion CO2 capture and compression. Fuel. 2015; 151: 110-117.
14. Li H, Ditaranto M, Berstad D. Technologies for increasing CO2 concentration in exhaust gas from natural
gas-fired power production with post-combustion, amine-based CO2 capture. Energy. 2011; 36: 1124-
1133.
15. Oyenekan BA, Rochelle GT. Energy performance of stripper configurations for CO2 capture by aqueous
amines. Industrial & Engineering Chemistry Research. 2006; 45: 2457-2464.
16. Le Moullec Y, Neveux T, Al Azki A, Chikukwa A, Hoff KA. Process modifications for solvent-based post-
combustion CO2 capture. International Journal of Greenhouse Gas Control. 2014; 31: 96-112.
17. Ahn H, Luberti M, Liu Z, Brandani S. Process configuration studies of the amine capture process for coal-
fired power plants. International Journal of Greenhouse Gas Control. 2013: 16; 29-40,
18. Damartzis, T., Papadopoulos, A.I., Seferlis, P., Process flowsheet design optimization for various amine-
based solvents in post-combustion CO2 capture plants. Journal of Cleaner Production. 2015: In press.
(doi:10.1016/j.jclepro.2015.04.129).
19. Karimi M, Hillestad M, Svendsen HF. Investigation of the dynamic behavior of different stripper
configurations for post-combustion CO2 capture. International Journal of Greenhouse Gas Control. 2012;
7: 230-239.
20. Sharifzadeh M. Integration of process design and control: A review. Chemical Engineering Research and
Design. 2013; 91: 2515-2549.
21. Mac Dowell N, Shah N. The multi-period optimisation of an amine-based CO2 capture process integrated
with a super-critical coal-fired power station for flexible operation. Computers & Chemical Engineering.
2015; 74: 169-183.
22. Delarue E, Martens P, D’haeseleer W. Market opportunities for power plants with post-combustion carbon
capture. International Journal of Greenhouse Gas Control. 2012; 6: 12-20.
39 | P a g e
Pages 166–179,
23. Lawal A, Wang M, Stephenson P, Obi O. Demonstrating full-scale post-combustion CO2 capture for coal-fired
power plants through dynamic modelling and simulation. Fuel. 2012; 101: 115-128.
24. Chalmers H, Gibbins J. Initial evaluation of the impact of post-combustion capture of carbon dioxide on
supercritical pulverised coal power plant part load performance. Fuel. 2007; 86: 2109-2123.
25. Cohen SM, Rochelle GT, Webber ME. Optimal operation of flexible post-combustion CO2 capture in
response to volatile electricity prices. Energy Procedia. 2011; 4: 2604-2611.
26. Seader JD, Henley EJ, Roper DK. Separation Process Principles (3rd edition). Hoboken, John Wiley & Sons,
Inc., 2013.
27. Onda K, Takeuchi H, Okumoto Y. Mass transfer coefficients between gas and liquid phases in packed
columns. Journal of Chemical Engineering of Japan. 1968; 1: 56-62.
28. Bravo JL, Fair JR. Generalized correlation for mass transfer in packed distillation columns. Industrial &
Engineering Chemistry Process Design and Development. 1982; 21: 162-170.
29. Billet R, Schultes M. Predicting mass transfer in packed columns. Chemical Engineering & Technology.
1993; 16: 1-9.
30. Bravo JL, Rocha JA, Fair JR. Mass transfer in Gauze Packings. Hydrocarbon Processing. 1985; 64: 56-60.
31. Bravo JL, Rocha JA, Fair JR. A comprehensive model for the performance of columns containing structured
packings. Institution of Chemical Engineers Symposium Series. 1992; 128: A489-A507.
32. Mac Dowell N, Samsatli NJ, Shah N. Dynamic modelling and analysis of an amine-based post-combustion
CO2 capture absorption column. International Journal of Greenhouse Gas Control. 2013; 12: 247–258.
33. Mac Dowell N, Llovell F, Adjiman CS, Jackson G, Galindo A. Modeling the fluid phase behavior of carbon
dioxide in aqueous solutions of monoethanolamine using transferable parameters with the saft-vr
approach. Industrial & Engineering Chemistry Research. 2010; 49: 1883-1899.
34. Sharifzadeh M, Shah, N,. Comparative studies of CO2 capture solvents for gas-fired power plants:
Integrated modelling and pilot plant assessments. Accepted for publication at International Journal of
Greenhouse Gas Control.
40 | P a g e
Pages 166–179,
35. Couper JR, Penney WR, Fair JR. Chemical process equipment selection and design (3rd edition). Burlington:
Elsevier Science, 2012.
36. Peters MS, Timmerhaus KD, West RE. Plant design and economics for chemical engineers (5th edition).
London: McGraw-Hill Chemical Engineering Series, 2003.
37. Electricity wholesale market [Internet]. Washington: US Energy Information Administration; US
Department of Energy, c2001-15. [cited 2015 April 8]. Available from:
http://www.eia.gov/electricity/wholesale/
38. Ulrich GD, Vasudevan PT. How to estimate utility costs. Chemical Engineering. 2006; April: 66-69.
39. Grossmann IE. Advances in mathematical programming models for enterprise-wide optimization.
Computers & Chemical Engineering. 2012; 47: 2-18.
40. Sharifzadeh M, Thornhill NF. Optimal selection of control structures using a steady-state inversely
controlled process model. Computers & Chemical Engineering. 2012; 38: 126-138.
41. Sharifzadeh M. Implementation of a steady-state inversely controlled process model for integrated design
and control of an ETBE reactive distillation. Chemical Engineering Science. 2013; 92: 21-39.
42. Sharifzadeh M, Thornhill NF. Integrated design and control using a dynamic inversely controlled process
model. Computers & Chemical Engineering. 2013: 48: 121–134.
43. gCCS: Whole chain CCS systems modelling software tool. Process Systems Enterprise. [Internet] [cited 2015
April 8], Available from: http://www.psenterprise.com/power/ccs/gccs.html).
44. Advanced Model Library - Gas-Liquid Contactors (AML:GLC) software tool. Process Systems Enterprise.
[Internet] [cited 2015 April 8], Available from:
http://www.psenterprise.com/processbuilder/libraries/aml_glc.html)
45. gSAFT software tool. Process Systems Enterprise. [Internet] [cited 2015 April 8], Available from:
http://www.psenterprise.com/gproms/options/physprops/saft/).
46. Weber PT. Modeling gas turbine engine performance at part-load. A Thesis for UTSR Gas Turbine Industrial
Fellowship Program, Electric Power Research Institute, Southwest Research Institute, University of
41 | P a g e
Pages 166–179,
Wyoming. [Internet] [cited 2015 April 8], Available from:
http://www.swri.org/utsr/presentations/WeberPatrickRept.pdf
42 | P a g e
Pages 166–179,
43 | P a g e
Pages 166–179,
44 | P a g e
Pages 166–179,
45 | P a g e
Pages 166–179,
46 | P a g e
Pages 166–179,
47 | P a g e
Pages 166–179,
48 | P a g e

Carbon Capture From Natural Gas Combined Cycle (NGCC) Power Plants: Solvent Performance Comparison at An Industrial Scale

Uploaded by

Copyright:

Available Formats

Carbon Capture From Natural Gas Combined Cycle (NGCC) Power Plants: Solvent Performance Comparison at An Industrial Scale

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

Carbon Capture From Natural Gas Combined Cycle (NGCC) Power Plants: Solvent Performance Comparison at An Industrial Scale

Uploaded by

Copyright:

Available Formats

View metadata, citation and similar papers at core.ac.

uk brought to you by CORE

Please cite the following paper as:

Carbon capture from natural gas combined cycle (NGCC) power

plants: solvent performance comparison at an industrial scale

Sharifzadeh, Mahdi1; Shah, Nilay

London, United Kingdom, SW7 2AZ.

performances compared to the MEA baseline solvents.

design and control, GCCmax.

steady-state process. By comparison, post-combustion carbon removal is conducted at near atmosphere

high degree of flexibility.

reproducible and comparative economic analysis.

capture solvent development and comparison.

energetic performance measures.

Overall process block diagram

Steam Cleaned Flue CO2

Process flow diagram of CO2 capture and compression sections

Hot flue gas

Natural gas CS-100

T-102 T-103 T-104

C-301 C-302 C-303 C-304 C-305

D-300 D-301 D-302 D-303 D-304 D-305

CWout CWout CWout CWout CWout

T-203 CWin E-205 D-306 D-307

Captured CWout CWout

Capture plant model development and validation

involved, as discussed in the following.

Rate-based modelling of gas-liquid contactors

Reaction kinetics and thermodynamics

Justification of the modelling strategy

the molecular segments of associating sites.

thermodynamic model are given in reference 33.

Pilot plant studies and model validation

temperatures were attributed to lack of insulation or temperature measurement errors.

Table 1. Model validation for MEA baseline solvent 34.

Table 2. Model validation for GCCmax solvent 34.

improvements are between 58.8-64.7%.

tightly controlled using make-up.

Solution algorithm: Simulation-optimization framework

integrated process design and control is provided by Sharifzadeh 20.

shown in Figure 1 and discussed in the following.

is repeated from Step (1).

𝐎𝐛𝐣𝐞𝐜𝐭𝐢𝐯𝐞 = 𝐄(𝐓𝐀𝐂𝐬 ) = ∑ 𝛍𝒔 × 𝐓𝐀𝐂𝐬 𝐏𝐫𝐨𝐛𝐥𝐞𝐦 − 𝟏

Constraints associated with first principles (transport phenomena, thermodynamics)

Control Constraints: 90% CO2 Capture, maximum turbine discharge temperature

Control (recourse) decision variables: Circulation flowrate, Reboiler steam flowrate

of the capture plant, and was calculated as:

𝐹𝑖𝑥𝑒𝑑 𝐶𝑎𝑝𝑖𝑡𝑎𝑙 𝑖𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡

The value of the

Optimization The required values of

Model development and implementation software tools

1-2 days to converge.

Results of optimization programming

Control strategy for combined cycle gas turbine (CCGT)

systematically safeguards the process equipment.

Air compressor Gas turbine Air compressor Gas turbine

in the first control structure consists of two iterative control modes:

gases (descending dotted line in Figure 6c).

switches to Mode (i).