Available online at www.sciencedirect.

com

ScienceDirect
Energy Procedia 63 (2014) 2289 – 2304

GHGT-12

Full-plant analysis of a PSA CO2 capture unit integrated in coal-

fired power plants: post- and pre-combustion scenarios
Luca Riboldia, *, Olav Bollanda, Jacob M. Ngoyb, Nicola Wagnerb
a
Energy and Process Engineering Department, the Norwegian University of Science and Technology, NO-7491Trondheim, Norway
b
School of Chemical and Metallurgical Engineering, University of the Witwatersrand, Johannesburg, South Africa

Abstract

It is well proven that a general policy to address the worldwide issue of global warming cannot disregard Carbon
dioxide Capture and Storage (CCS) in the portfolio of tools. Pressure Swing Adsorption (PSA) processes are
believed to be a promising option for achieving more energy and cost-effective capture of CO2 from large point
sources, especially coal-fired power plants. Nevertheless, there is a gap in knowledge with respect to information
and approaches for the integration of CO2 capture using PSA in power plants. The main contribution of this work is
to fill this gap, providing a plant-level comparison with other techniques of decarbonization (i.e., state-of-the-art
absorption processes) in terms of CO2 separation performance, energy efficiency and footprint of the technology.
Full-plant analyzes were developed based on a dynamic computational model representing coal-fired power plants
operating with a Pressure Swing Adsorption (PSA) cycle, both in a post- and pre-combustion configuration. The
resulting plant performance is compared to that of absorption-based systems using the same reference power plant
assumptions. The post-combustion scenario outputs reveal that the benchmark absorption process outperforms the
PSA alternative. Even though the CO2 separation requirements are met, the relatively large energy penalty and the
very large footprint seem to highlight the current unsuitability of PSA for post-combustion CO2 capture. Conversely,
the pre-combustion scenario analysis shows the PSA process as a promising alternative. The performance, in terms
of CO2 separation, energy efficiency and footprint of the technology, results just slightly lower than that of a plant
implementing absorption as a CO2 capture method. However, the novelty of the analysis and the non-maturity of the
technology in this application leaves the window open for future improvements.

© 2014
© 2013TheTheAuthors.
Authors. Published
Published by Elsevier
by Elsevier Ltd. is an open access article under the CC BY-NC-ND license
Ltd. This
Selection and peer-review under responsibility of GHGT.
(http://creativecommons.org/licenses/by-nc-nd/3.0/).
Peer-review under responsibility of the Organizing Committee of GHGT-12

* Corresponding author. Tel.: +47-73593559.

E-mail address: luca.riboldi@ntnu.no

1876-6102 © 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/3.0/).
Peer-review under responsibility of the Organizing Committee of GHGT-12
doi:10.1016/j.egypro.2014.11.248
2290 Luca Riboldi et al. / Energy Procedia 63 (2014) 2289 – 2304

Keywords: pressure swing adsorption; CO2 capture; full-plant analysis; system integration; post-combustion; pre-combustion.

1. Introduction

The atmospheric concentration of carbon dioxide has reached levels 40% higher than pre-industrial time [1]. The
influence of this rise on the climate system is widely accepted as clear, and the need for a strategy to tackle the
consequent environmental problem is an impelling issue. A portfolio of technologies need to be applied in order to
achieve the emission limits regarded as necessary to prevent major climatic consequences. In this context, Carbon
Capture and Storage (CCS) is an essential tool to reach CO2 concentration stabilization goals. CCS is a process
consisting of the separation of CO2 from industrial and energy-related sources, transport to a storage location and
long-term isolation from the atmosphere [2]. The capture step involves separating carbon dioxide (CO2) from other
gaseous products. Different technologies are available to fulfill this task [2-4]. The absorption of CO2 by means of
alkanolamine or physical solvents is regarded as the most mature technology. Nevertheless, some drawbacks related
to this method for CO2 separation has hampered its actual implementation. Particularly, the energy-intensive process
for the regeneration of the solvent gives a significant drop in the overall plant efficiency. Considering other negative
issues, like the possible corrosion of the equipment and the toxicity of the solvents, room is available for
investigating alternative capture methods. Pressure Swing Adsorption (PSA) process for separating CO2 from
process gas streams has gained some attention, especially considering its potentially low energy requirement. PSA is
a cyclic process, which exploits the phenomenon of attraction that the adsorbent surface produces on certain
molecules, like CO2. Before the adsorbent bed gets saturated, regeneration is carried out by decreasing the partial
pressure of the adsorbed component. PSA is a well-established gas separation technique in air separation, gas drying
and hydrogen purification separation [5]. More recently, PSA has been assessed for CO2 capture in power plants in
both post- [6-20] and pre-combustion [21, 22] applications. A significant lack of modern simulation tools for
understanding the dynamics of complex integrated systems, particularly with CCS, has been determined [23]. The
relevant few works present in the literature provide only preliminary analyzes based on a set of simplified
assumptions [24, 25]. In other words, the materials and processes assessed have still not demonstrated their
effectiveness when operating integrated in an actual full-scale power plant. The objective of the current work is to
address this issue, providing an assessment of the systems in terms of CO2 separation performance, energy
efficiency and footprint of the technology. Two scenarios have been considered:

x Post-combustion capture of CO2: CO2 is separated from the flue gas downstream the boiler of an Advanced
SuperCritical (ASC) pulverized bituminous coal fired power plant.
x Pre-combustion capture of CO2: CO2 is separated from the synthesis gas resulting from the gasification of
coal in an Integrated Gasification Combined Cycle (IGCC) plant.

A full-plant analysis is provided for both the scenarios. The resulting plant performance is compared to that of
absorption-based systems using the same reference power plant assumptions. Thus, the current competitiveness of
adsorption with respect to absorption can be stated, together with the potential advantages in the utilization of solid
sorbents.

Nomenclature

ai number of neighboring sites occupied by adsorbate molecule for species i

Ci gas concentration of species i, mol/m3
Cp,ads adsorbed phase specific heat at constant pressure, J/(kg • K)
Cp,g gas specific heat at constant pressure, J/(mol • K)
Cp,s particle specific heat at constant pressure, J/(kg • K)
Ctot total gas concentration, mol/m3
Dax,i axial dispersion coefficient of species i, m2/s
Dc,i micropore diffusivity of species i, m2/s
 Luca Riboldi et al. / Energy Procedia 63 (2014) 2289 – 2304 2291

dp particle diameter, m
ǻ+r,i heat of adsorption of species i, J/mol
ki equilibrium constant of species i, Pa-1
k’L adsorption constant at infinite temperature of species i, Pa-1
kLDF,i linear driving force coefficient, s-1
ৄ mole flow rate, mol/s
P pressure, Pa
Pr Prandtl number
qi* equilibrium adsorbed concentration of species i, mol/kg
Tࡄi averaged adsorbed concentration of species i, mol/kg
qm,i specific saturation adsorption capacity of species i, mol/kg
R universal gas constant, Pa • m3/(mol • K)
Re Reynolds number
T temperature, K
us superficial velocity, m/s
yi mole fraction of species i
z axial direction, m

Greek letters
Ȗ specific heat ratio
İ bed porosity
İp particle porosity
Ș isentropic efficiency
Ȝax axial thermal dispersion coefficient, J/(s • m • K)
ȝ dynamic viscosity, Pa • s
ȡg gas volumetric mass density, kg/m3
ȡp volumetric mass density of the particle, kg/m3
ȤLDF linear driving force geometrical factor (15 for zeolite 5A, 3 for activated carbon)

Subscripts
i species

Superscripts
NC number of components

2. Simulation of the PSA unit

In the systems to be analyzed, both in the post- and pre-combustion scenario, the CO2 separation process is
performed by a PSA process 1. Different packed beds working in parallel are used to synchronize and accommodate
steps in order to provide a continuous flow of the process gas streams. This coordinated set of packed beds
constitutes a PSA train. The steps considered for the PSA process include:

1
Since an under-atmospheric pressure level is applied for the regeneration process of the post-combustion scenario, technically the process
should be called Vacuum Pressure Swing Adsorption (VPSA). For simplicity the V will be dropped and the process will be referred as PSA.
2292 Luca Riboldi et al. / Energy Procedia 63 (2014) 2289 – 2304

x Feed (F): the high pressure gas stream is fed to the bottom of the adsorption column and the heavier
components start to be preferentially adsorbed and to saturate the bed.
x Rinse (R): a stream of CO2-rich gas is extracted from the product gas to be fed in the column. The purpose is to
displace the light components from the gas phase before starting the regeneration process.
x Pressure equalization – Depressurization (D): two columns at different pressure levels are put in contact. The
high-pressure column is depressurized to a medium level of pressure.
x Blowdown (BD): the bed is started to be regenerated by decreasing the total pressure.
x Purge (Pu): the regeneration is carried on by feeding a purging gas counter-current in the column.
x Pressure equalization –Pressurization (P): two columns at different pressure levels are put in contact. The low-
pressure column is pressurized to a medium level of pressure.
x Null (N): the bed is left to rest. This step may be introduced to guarantee the correct match between beds
undergoing different interconnected steps.
x Feed Pressurization (FP): the feed stream is co-currently admitted to the column in order to raise the pressure to
the feed level.

The different steps implemented in the PSA processes in post- and pre-combustion scenarios, and the way
different columns of a train interact, are shown in Figure 1 and Figure 2. The characteristics of the PSA processes
are also reported in Table 1 and Table 2. The behavior of each column in a PSA unit is described by a set of partial
differential and algebraic equations (PDAEs), representing the material, energy and momentum balances. Table 3
sums up the governing equations adopted. All the assumptions which are applied to the model are summarized
hereafter:

x The gases follow nearly the ideal gas law.

x Radial dispersion terms are negligible.
x Axial dispersed plug flow is considered.
x The adsorbent particles have a bidisperse structure.
x Micropore diffusion is the controlling mass transfer resistance. Thus, macropore and film mass transfer
resistances can be considered as negligible.
x The Linear Driving Force (LDF) approximation is applied for describing the kinetic of the process.
x No temperature gradients occur in the solid particle. The gas and solid phases are in thermal equilibrium.
x The heat of adsorption is independent of temperature and adsorbed phase loading
x The reactors are considered adiabatic. The heat transfer coefficients with the wall and with the environment are
set to zero.
x The Ergun equation applies for the momentum balance [26].

First PSA stage Second PSA stage

F R BD Pu FP
F D BD Pu P FP

F R BD Pu FP

BD Pu FP F R BD FP F D BD Pu P

R BD Pu FP F BD Pu P FP F D

Figure 1. PSA process configuration for the post-combustion scenario. First and second PSA stage.
 Luca Riboldi et al. / Energy Procedia 63 (2014) 2289 – 2304 2293

Table 1. Characteristics of the PSA cycles in the post-combustion scenario.

Step time (s) Mole flow rate (mol/s)

Stage Cycle time (s) F R D BD Pu P FP Feed Purge Rinse
1 2106 702 234 0 702 234 0 234 304,32 91,27 91,27
2 1548 543 0 50 543 181 50 181 336,00 5,00 0,00

F D X4 BD Pu PX4 N X4 FP

FP F D1 D2 D3 D4 BD Pu P4 N P3 N P2 N P1 N

P2 N P1 N FP F D1 D2 D3 D4 BD Pu P4 N P3 N P2

N P3 N P2 N P1 N FP F D1 D2 D3 D4 BD Pu P4

Pu P4 N P3 N P2 N P1 N FP F D1 D2 D3 D4 BD Pu

BD Pu P4 N P3 N P2 N P1 N FP F D1 D2 D3 D4

D2 D3 D4 BD Pu P4 N P3 N P2 N P1 N FP F D1 D2

F D1 D2 D3 D4 BD Pu P4 N P3 N P2 N P1 N FP F

Figure 2. PSA process configuration for the pre-combustion scenario.

Table 2. Characteristics of the PSA cycle in the pre-combustion scenario.

Step time (s) Mole flow rate (mol/s)

Cycle time (s) F DX4 BD Pu PX4 NX4 FP Feed Purge
630 90 41 80 59 41 8 41 3751,0 525,0

The adsorbent selected for the post and pre-combustion scenario are respectively a zeolite 5A [27] and an
activated carbon [28]. The uptake capacity of both the adsorbents is described by a multi-site Langmuir model:

ai
q*i ª NC § q* ·º § 'H r,i ·
a i k i Pi «1  ¦ ¨ i ¸¸ » , with k i k f ,i exp ¨  ¸ (1)
q m,i ¨
i © q m,i
«¬ ¹ »¼ © RT ¹

The physical properties and the equilibrium parameters of the zeolite and of the activated carbon, obtained
respectively from [27] and [28], are given in Table 4, together with the characteristics of the adsorption bed.
2294 Luca Riboldi et al. / Energy Procedia 63 (2014) 2289 – 2304

Table 3. Governing equations.

Component mass balance

wCi w  u s Ci  w § wy · wq
ªH  H p 1  H  ¼º   ¨ HDax,i Ci i ¸  Up 1  H  i (2)
wt ¬ wz wz © wz ¹ wt
LDF equation
wqi Dc,i
k LDF,i  q*i  qi  with k LDF,i F LDF (3)
wz rc2
Continuity equation
wC tot w  u s C tot  NC
w qi
ªH  H p 1  H  º¼   Up 1  H  ¦ (4)
wt ¬ wz i wz
Energy balance
NC
ª º wT wT
«HC p,G C tot  H p 1  H  C p,G C tot  1  H  C p,SUp  1  H  Up ¦ C p,ads,i qi » wt u s C p,G C tot
wz

¬ i ¼ (5)
w § wT · NC
w qi
 ¨ O ax ¸  Up 1  H  ¦  'H r,i 
wz © wz ¹ i wz

Momentum balance
wP ª150 (1  H) 2 1.75 (1  H) º
« 2 3
Pu s  3
UG u s u s » (6)
wz «¬ d p H dp H »¼

In order to reduce the computational time necessary to solve the model, a one-column approach has been applied.
This modeling strategy consists of simulating just one of the columns of the whole train [21, 29-31]. The
interactions between different columns are accounted for by virtual gas streams which are defined through the
information stored in the previous cycles.
The CO2 recovery and purity were selected as parameters to evaluate the CO2 separation performance of the
plant. They are defined as following:

³C
BD
CO2 ,out u s,out dt  ³ CCO2 ,out u s,out dt  ³ CCO2 ,in u s,in dt
Pu R
Re cov ery CO2 (7)
³ CCO2 ,in u s,in dt  ³ CCO2 ,in u s,in dt
F FP

³C
BD
CO2 ,out u s,out dt  ³ CCO2 ,out u s,out dt
Pu
Purity CO2 (8)
³ C tot,out u s,out dt  ³ C tot,out u s,out dt
BD Pu

The energy consumption associated to the PSA process is calculated as following:

J fan 1
ª º
1 J fan « § Pin · Jfan » n
Fan
power R
Tin ¨  1 (9)
Kfan J fan  1 « P ¸ » in
© out ¹
¬« »¼
 Luca Riboldi et al. / Energy Procedia 63 (2014) 2289 – 2304 2295

J compr 1
ª º
1 J compr «§ Pin · Jcompr »
Compressor
power R
5in «¨ ¸  1» n in (10)
Kcompr J compr 1 P
«¬© out ¹ »¼

J vacuum 1
ª º
1 J vacuum «§ Patm · vacuum  1» n
J
Vacuum
power R
5in ¨ ¸ (11)
Kvacuum J vacuum  1 « P » in
«¬© vacuum ¹ »¼

The expressions above correspond to the power consumption of: a fan overcoming the pressure drop in the bed; a
vacuum pump setting a regeneration pressure lower than atmospheric; a compressor raising the pressure of a CO2-
rich gas stream to be rinsed (this last applies only if a rinse step is implemented).

Table 4. Bed characteristics, physical properties and equilibrium data of the adsorbents.

Column Post-combustion Pre-combustion

Bed height (m) 10 10
Bed diameter (m) 8 6,6
Bed porosity 0,32 0,38
Bulk density (kg/m ) 3 735 522
Adsorbent Zeolite Activated carbon
Particle diameter (mm) 2,70 2,34
Particle porosity 0,300 0,566
Particle density (kg/m3) 1083 842
Particle specific heat (J/kg/K) 920 709

Multi-site Langmuir parameters Zeolite Activated carbon

aCO2 2,1 3,0
aN2 2,5 4,0
aH2 - 1,0
aCO - 2,6
-1
k’&2(Pa ) 1,468E-11 2,128E-11
-1
k’1(Pa ) 3,789E-11 2,343E-10
k’+(Pa-1) - 7,690E-11
-1
k’&2(Pa ) - 2,680E-11
qm,CO2 (mol/kg) 3,919 7,855
qm,N2 (mol/kg) 3,279 5,891
qm,H2 (mol/kg) - 23,570
qm,CO (mol/kg) - 9,063
ǻ+r,CO2 (J/mol) -37853 -29100
ǻ+r,N2 (J/mol) -19435 -16300
ǻ+r,H2 (J/mol) - -12800
ǻ+r,CO (J/mol) - -22600
2296 Luca Riboldi et al. / Energy Procedia 63 (2014) 2289 – 2304

The simulation is run until the Cycle Steady State (CSS) condition is achieved, meaning that the conditions in the
bed at the end of a cycle are identical to those at the beginning of the same cycle. The reported results refer to the
cycle at CSS.
All the simulations were implemented in gPROMS environment (Process System Enterprise, London, UK). The
discretization method for the spatial domain in the column is the Centered Finite Difference Method (CFDM) with
150 intervals in the whole column. The number of discretization intervals is a tradeoff between accuracy of the
simulation and computational time required to run it.

3. Simulation of the integrated power island

Two different types of thermal power plants are considered:

x an Advanced SuperCritical (ASC) pulverized bituminous coal fired power plant.

x an Integrated Gasification Combined Cycle (IGCC) plant.

The systems selected are meant to represent the current most diffused options for coal-based power generation,
respectively for the post- and pre-combustion scenario. The power islands - intended as the whole system except the
CO2 capture unit - are modeled by Thermoflow Inc. products: STEAM PRO, GT PRO and THERMOFLEX.
Steady-state analyzes were considered. In both the scenarios, the simulations were first run for a base case without
CO2 capture. The inputs for the simulations were defined in accordance to DECARBit’s recommendations [32].
They define assumptions and parameters that should be applied in CCS simulations such that a comparison of
different technologies is possible. The same approach has been used to simulate the two alternatives with CO2
capture by absorption. The relative outputs were thought to be useful for comparison with the most mature state-of-
the-art CO2 capture technology. Finally, the simulations of the power islands connected to the PSA unit were run
using THERMOFLEX. The PSA processes, simulated in gPROMS, were linked to Thermoflow through macros in
Excel. A total of six cases were performed:

x ASC plant without CO2 capture

x ASC plant with CO2 capture by absorption
x ASC plant with CO2 capture by PSA

x IGCC plant without CO2 capture

x IGCC plant with CO2 capture by absorption
x IGCC plant with CO2 capture by PSA

The plant layout of the ASC plant integrated with a PSA unit for CO2 capture is represented in Figure 3, with the
corresponding streams given in Table 5. Water, being detrimental to the CO2 adsorption process, needs to be
removed. A partial removal is carried out by condensation in a flash separator. A further section should bring water
presence down to trace levels, but this has not been included in the simulation. The water still present in the flue gas
has been neglected since no valid modeling data were found in the literature. It is not known to what extent this
assumption affects the results. The CO2 in the flue gas is separated in a double stage PSA process and the CO2-rich
stream is then compressed up to 110 bar for transport and storage. The PSA process is constituted by two stages,
since a single stage was not able to meet the purity specifications for the CO2-rich stream [12]. The gas stream
undergoes a recompression between the stages: the pressure is raised from 0.1 to 1.5 bar. The regeneration pressure
for the PSA process is set to 0.1 bar, a value which allows to carry out the desorption process satisfactorily.
 Luca Riboldi et al. / Energy Procedia 63 (2014) 2289 – 2304 2297

Gypsum Effluent
Limestone Water
Cooling tower ASC steam
Cooling water turbine cycle
Flue Gas 5 Water
IP HP Cold Feed Desulphurization removal
Air steam steam R/H water

2 4 6
Coal and ash 3 ElectroStatic PSA 8
handling ASC boiler island Precipitator CO2 capture
1 Coal
Fly ash
7
Bottom ash CO2 to transport
CO2 and storage
DeNOx plant
compression 9

Figure 3. ASC plant with CO2 capture by PSA.

Table 5. Stream table of the ACS plant with CO2 capture by PSA.

Stream Mass flow T P MW Composition

kg/s °C bar g/mol CO2 N2 O2 Ar SO2 H2O
1 66,2 25,0 1,01 7,74
2 619,6 15,0 1,01 28,86 0,03 77,29 20,74 0,93 0,000 1,01
3 800,9 117,5 1,00 29,82 13,63 74,37 4,38 0,89 0,040 6,69
4 800,9 117,5 0,98 29,82 13,63 74,37 4,38 0,89 0,040 6,69
5 822,7 62,3 1,02 29,34 4,23 71,42 4,23 0,86 0,003 10,40
6 781,0 20,0 1,00 30,36 14,27 77,84 4,61 0,94 0,004 2,33
7 150,8 22,8 0,10 43,04 92,98 6,62 0,39 0,02 0,000 0,00
8 616,3 45,5 1,00 28,61 1,55 91,89 5,45 1,55 0,000 0,00
9 150,8 28,0 110,0 43,04 92,98 6,62 0,39 0,02 0,000 0,00

The plant layout of the IGCC plant integrated with a PSA unit for CO2 capture is represented in Figure 4, with
the corresponding streams given in Table 6. A Water-Gas Shift (WGS) is necessary in order to convert as much CO
and H2O to CO2 and H2 as possible, thus facilitating the subsequent CO2 separation. After the H2S removal, the CO2
is separated from the syngas in a PSA unit [21]. The regeneration pressure is set to 1 bar. Since the CO2-rich gas
stream does not achieve the requirements for being processed and transported, a further purification step is
implemented. It consists of the removal of impurities by means of two flash units integrated in the CO2 compression
section [33, 34]. The schematic of this additional process is shown in Figure 5. Referring to [34], the temperatures
selected at the outlet of each heat exchanger are set respectively to -30°C and -54.5°C. The gas stream is compressed
up to 30 bar before entering the flash separation unit. The CO2-rich stream is finally compressed up to 110 bar for
transport and storage. The product syngas, rich in H2, is injected as fuel into the gas turbine.
2298 Luca Riboldi et al. / Energy Procedia 63 (2014) 2289 – 2304

6 Nitrogen to the gas turbine

5 Air from the gas turbine

H2-rich
Air CO2 to transport syngas 12
and storage CO2 compression
Gas turbine
and flash
2 13 separation 11 H2-rich
8
syngas 9
Air Separation 10
PSA
unit HRSG
CO2 capture
3 N2 4 O2
7
Syngas Convective Water Gas Steam
Gasifier cooler and H2S absorber
scrubber
Shift turbine cycle

1
Coal Slag

Figure 4. IGCC plant with CO2 capture by PSA.

Table 6. Stream table of the IGCC plant with CO2 capture by PSA.

Stream Mass flow T P MW Composition %

kg/s °C bar g/mol H2 CO2 CO CH4 N2 O2 Ar H2S H2O
1 38,5 25,0 1,01 7,74
2 64,6 15,0 1,01 28,86 0 0,03 0 0 77,29 20,74 0,93 0 1,01
3 8,5 82,5 1,01 8,49 0 0,00 0 0 100 0 0 0 0
4 31,2 123,9 44,9 31,98 0 0,00 0 0 3,50 95 1,50 0 0
5 64,6 351,9 10,6 28,86 0 0,03 0 0 77,29 20,74 0,93 0 1,01
6 87,5 116,2 24,1 28,02 0 0,00 0 0 100 0 0 0 0
7 107,6 64,9 38,8 20,18 53,54 37,89 1,50 0,06 6,72 0 0,27 0,0001 0,03
8 19,4 61,7 38,8 6,56 84,50 2,89 1,95 0,07 10,09 0 0,50 0 0
9 656,1 579,4 1,05 27,44 0 1,24 0,00 0,00 75,14 10,13 0,79 0 12,71
10 656,1 113,8 1,01 27,44 0 1,24 0,00 0,00 75,14 10,13 0,79 0 12,71
11 88,2 42,3 1,0 37,11 15,04 81,39 0,94 0,04 2,53 0 0 0 0,06
12 8,2 18,0 27,7 15,10 63,64 22,80 3,61 0,12 9,83 0 0 0 0
13 80,0 113,7 110,0 43,68 0,58 98,91 0,14 0,01 0,36 0 0 0 0
 Luca Riboldi et al. / Energy Procedia 63 (2014) 2289 – 2304 2299

Flash
separator
Multi-stream Multi-stream
CO2-lean heat exchanger heat exchanger
stream

Flash
Gas stream separator
from PSA

Throttling
valve
Throttling
CO2-rich valve
stream
Compressor Compressor CO2-rich stream
stage stage to transport and
storage
CO2-rich
stream Inter-cooling Inter-cooling

Figure 5. Double flash separation [33].

4. Results and discussions

4.1. Post-combustion scenario

The most relevant set of results, relative to the simulation of the ASC power plant integrated with PSA and
compression units, is presented in Table 7. For the sake of comparison, two more reference cases are reported in
Table 7: a plant without CO2 capture, and a plant with a state-of-the-art absorption unit, both defined following the
DECARBit’s specifications [32].
By comparing the effectiveness of the two CO2 capture options on the basis of their ability to efficiently separate
CO2, absorption displays a much smaller efficiency penalty. The process considered, which is a typical amine-based
chemical absorption process, succeeds to return a basically pure CO2 gas stream to be sent to compression with a
CO2 recovery factor of about 90%. Conversely, the PSA-based plant meets the requirement in terms of CO2
recovery (90.89%), but cannot reach the CO2 purity specification of 95%. The reported CO2 purity value of 93.0 %
is not far from the target level, and it could be increased by trading-off the CO2 recovery. Where current superiority
of the absorption process becomes evident is in the energy analysis of the integrated plants. The state-of-the-art ASC
power plant with absorption for CO2 capture has a net electrical efficiency of 33.6%, which means a drop of 11.9%
compared to the same plant without CO2 capture. When a PSA unit is considered, the efficiency drops to 28.8%
(16.5% points less than the reference plant). It is interesting to analyze how the CO2 capture related power losses are
distributed. Capturing CO2 through an absorption process results in two major sources of power losses. The first
consists of the need for some steam to comply with the reboiler heating duty for the regeneration of the solvent. The
steam is extracted from the steam turbine, resulting in a reduction of the gross power produced. The other important
power consumption is for the compression of the resulting CO2 stream. Some electrical power is also required for
the CO2 separation (i.e., for the solvent circulation pump and the cooling water pump), but this consumption is much
smaller. Using a PSA process completely modifies the picture. In this case the regeneration of the adsorption beds is
carried out by reducing the partial pressure. This operation does not need any steam to be bled from the steam
turbine. However, reducing the pressure to 0.1 bar has a consequence in the power demand of the CO2 compression
unit; the power spent for CO2 compression increases when dealing with a PSA process compared to an absorption
plant. Two other power consumption sources have an impact in the energy balance. One is the recompression of the
flue gas before entering the second PSA stage. It contributes to roughly one fourth of the overall energy
consumption. Second, the largest contribution comes from the CO2 separation process. The vacuum pumps, which
create the vacuum necessary for the saturated beds regeneration, are particularly energy demanding. Moreover, the
2300 Luca Riboldi et al. / Energy Procedia 63 (2014) 2289 – 2304

presence of a rinse step in the first PSA cycle implies a compression of part of the product gas to atmospheric level.
Those two contributions constitute almost the entire fraction of the so-called CO2 separation power consumption.

Table 7. Full-plant simulations results in the post-combustion scenario.

No Capture Absorption PSA

Coal flow rate [kg/s] 66,2 66,2 66,2

Coal LHV [MJ/kg] 25,2 25,2 25,2

Net fuel input [MWth] 1665,6 1665,6 1665,6

Steam turbine output [MW] 825,6 691,9 825,2

Gross electrical output [MW] 825,6 691,9 825,2

CO2 separation power consumption [MW] - 10,4 120,7

Flue gas compression power consumption [MW] - 0,00 69,7

CO2 compression power consumption [MW] - 47,4 84,5

Miscellaneous auxiliaries [MW] 71,2 78,6 71,5

Total auxillary power consumption [MW] 71,2 136,3 346,3

Net electrical output [MW] 754,4 555,6 479,0

Net electrical efficiency [%] 45,3 % 33,4 % 28,8 %

CO2 purity [%] - 100 % 93,0 %

CO2 recovery [%] - 90,0 % 90,9 %

It is worthwhile to also carry out an analysis on the footprint of the CO2 capture unit. In order to be able to handle
the entire flow rate of flue gas with a single PSA train, each column would need an excessively large diameter.
Thus, a reasonable maximum column diameter was set to 8 m and the flue gas flow rate was split between different
PSA trains. The output of this analysis reveals that a large number of trains are necessary (i.e., around 72 and 24
trains for the first and second PSA stage). Considering that a train is constituted by 3 or 2 columns, respectively for
the first or second PSA stage, and the significant size of a column, the total footprint of the CO2 capture unit seems
excessive to be considered feasible.

4.2. Pre-combustion scenario

The most relevant set of results, from the simulation of the IGCC plant integrated with PSA and compression
units, is presented in Table 8. For the sake of comparison two more reference cases are reported in Table 8: a plant
without CO2 capture, and a plant with a state-of-the-art absorption process, both defined following the DECARBit’s
specifications [32].
 Luca Riboldi et al. / Energy Procedia 63 (2014) 2289 – 2304 2301

Table 8. Full-plant simulations outputs in the pre-combustion scenario.

No Capture Absorption PSA + Flash

Coal flow rate [kg/s] 33,3 38,5 38,5

Coal LHV [MJ/kg] 25,2 25,2 25,2

Net fuel input [MWth] 837,3 968,1 968,2

Gas turbine output [MW] 253,1 290,3 287,0

Steam turbine output [MW] 192,6 163,7 167,4

Air expander output [MW] 4,5 5,9 5,4

Gross electrical output [MW] 450,2 459,9 459,8

CO2 separation power consumption [MW] - 16,5 0,05

CO2 compression power consumption [MW] - 18,8 41,3

Miscellaneous auxiliaries [MW] 54,4 68,6 68,3

Total auxiliary power consumption [MW] 54,4 103,8 109,6

Net electrical output [MW] 395,9 356,1 350,2

Net electrical efficiency [%] 47,3 % 36,8 % 36,2 %

CO2 purity [%] - 100 % 98,9 %

CO2 recovery - separation technology [%] - 94,6 % 89,7 %

CO2 recovery - overall plant [%] - 90,5 % 85,4 %

The main parameters describing the effectiveness of CO2 capture seem to suggest absorption as the best
alternative. The CO2 purity of the CO2-rich stream is well above the 95% suggested threshold in both cases. Where
the absorption displays a small advantage is in regard to CO2 recovery. While the PSA and double flash process
attains a recovery factor of 89.7%, the absorption technology reaches 94.6%. The syngas burnt in the gas turbine
also has a content of CO and CH4, which leads to the formation of some additional CO2. Taking this amount into
account, the overall CO2 balance yields final values of CO2 recovery for the overall plant of 85.4% and 90.5%. The
highest value corresponds to the absorption case. However, both technologies return acceptable outputs.
An energy performance analysis of the full-plant systems was also carried out. The net efficiency of the IGCC
plant without CO2 capture related unit is 47.3%. It drops to 36.8% when CO2 is separated by absorption through a
two-stage Selexol process. The integration of a PSA unit and a double flash separation into the IGCC plant leads to
a final net efficiency of 36.2%; the efficiency decreases are of 10.5% and 11.1%, respectively. In terms of net
electrical efficiency, the CO2 capture by absorption seems to perform slightly better than the PSA process. From a
breakdown of the power losses, it is clear that the main factor negatively affecting the PSA based case is the power
spent for the CO2 compression. The PSA process itself is directly responsible for a small power consumption,
related to the fan overcoming the pressure drop in the adsorption bed7KHFRQWULEXWLRQLVQHJOLJLEOH§0:
2302 Luca Riboldi et al. / Energy Procedia 63 (2014) 2289 – 2304

However, PSA is the cause of some other energy consumptions. Since the bed is regenerated by decreasing the
pressure, the CO2 compressors have to make up for this pressure reduction. The further CO2 flash separation also
contributes to an increase in the compression work by throttling the CO2-rich stream. The power spent for the CO2
compression is as high as 41.3 MW, more than twice that of the absorption case. In fact, thanks to the particular
configuration of the absorption/regeneration process, the CO2 compression duty is 18.77 MW. Even though the
absorption case displays a much higher CO2 separation power consumption (16.5 MW, mainly due to the pumps for
solvent circulation), the energy saved in the compression process is compensating for it. There are other sources of
power losses related to the presence of a CO2 capture unit. These power losses are connected either to the processes
undergone by the syngas (i.e., syngas Lower Heating Value (LHV) reduction during the WGS process and syngas
LHV reduction due to the H2 leaving with the CO2-rich stream) or to the extraction of steam from the steam turbine
to be fed to the WGS process. They cause a reduction in the gross power output of the plant, but are not significant
to our analysis since their effect is quite similar in the absorption and adsorption cases.
The footprint of the PSA process is rather small. Given the relatively high pressure of the syngas (i.e., 38.8 bar), a
single PSA train is able to process the entire syngas flow rate. The 7 columns constituting the train have been sized
accordingly.

5. Conclusions

A full-plant analysis of a PSA CO2 capture unit integrated in coal-fired power plants has been investigated. Both
a post- and pre-combustion scenario has been assessed. The simulation of the PSA process has been matched with,
respectively, the simulation of an Advanced SuperCritical (ASC) pulverized bituminous coal fired power plant, and
an Integrated Gasification Combined Cycle (IGCC) plant. For comparison two reference cases per scenario were
simulated, namely a plant without CO2 capture and a plant with a benchmark absorption unit for CO2 capture. The
suitability of PSA process implementation has been assessed on three levels: CO2 separation performance, energy
efficiency, and footprint of the separation unit.
The implementation of a PSA process for CO2 capture in a post-combustion scenario has demonstrated not to be
feasible at the current state of the technology. Whilst the CO2 separation performance is close to fulfill the
requirements, the net electrical efficiency is significantly reduced. The resultant 28.8% is far below the alternative
constituted by absorption, which reaches 33.4%. The net efficiency of the reference plant without CO2 capture is
45.3%. Another important issue concerns the footprint of the PSA unit. In order to process the whole volumetric
flow rate of flue gas, a large number of PSA trains would be required (up to 72 just for the first of the two PSA
stages considered). Each of these PSA trains is comprised of 3 columns working in parallel. Given the dimension of
a column (8 m diameter), the footprint would be very much larger compared to an absorption plant.
A PSA process in a pre-combustion scenario seems to perform better. It was determined that a PSA unit alone
was hardly able to achieve the required CO2 separation performance. However, the addition of a double flash
separation integrated in the CO2 compression process has positively addressed the problem. Eventually, the CO2
purity (98.9%) reached satisfactory levels. The CO2 recovery matches the requirement when considering the
technology in itself (89.7%) but it is still too low if considering the overall plant balance (85.4%). With regards to
energy efficiency, the plant achieves a promising level (36.2%), even though slightly lower than that of an
absorption-based plant (36.8 %). The net efficiency of the reference plant without CO2 capture is 47.3%. Finally the
footprint of the CO2 capture-related units is considered to be rather small, since a single train PSA unit is able to
process the whole volume of syngas. In conclusion, the pre-combustion implementation of a PSA process for CO2
capture displayed some encouraging results. Its general performance still ranks lower than that of a benchmark
absorption process. Nevertheless, there is margin to claim that further advancements in material technology, process
optimization and integration may fill this gap.

Acknowledgements

The authors gratefully acknowledge the financial support provided through the “EnPe – NORAD’s Programme
within the energy and petroleum sector”.
 Luca Riboldi et al. / Energy Procedia 63 (2014) 2289 – 2304 2303

References

[1] IPCC, Climate Change 2013: The Physical Science Basis Contribution of Working Group I to the Fifth Assessment Report of the
Intergovernmental Panel on Climate Change, T.F. Stocker, D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen , J. Boschung, A. Nauels, Y. Xia, V.
Bex and P.M. Midgley, Editor 2013: Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. p. 1535 pp.
[2] IPCC, IPCC Special Report on Carbon Dioxide Capture and Storage. Prepared by Working Group III of the Intergovernmental Panel on
Climate Change, B. Metz, O. Davidson, H. C. de Corninck, M. Loos, and L.A. Meyer, Editor 2005: Cambridge University Press, Cambridge,
United Kingdom and New York, NY, USA. p. 442 pp.
[3] Ebner, A.D. and J.A. Ritter, State-of-the-art Adsorption and Membrane Separation Processes for Carbon Dioxide Production from Carbon
Dioxide Emitting Industries. Separation Science and Technology, 2009. 44(6): p. 1273-1421.
[4] Herzog H, M.J., Hatton A., Advanced Post-Combustion CO2 Capture, 2009.
[5] Grande, C.A., Advances in Pressure Swing Adsorption for Gas Separation. ISRN Chemical Engineering, 2012. 2012: p. 13.
[6] Agarwal, A., L.T. Biegler, and S.E. Zitney, A superstructure-based optimal synthesis of PSA cycles for post-combustion CO2 capture. AIChE
Journal, 2010. 56(7): p. 1813-1828.
[7] Choi, W.-K., et al., Optimal operation of the pressure swing adsorption (PSA) process for CO2 recovery. Korean Journal of Chemical
Engineering, 2003. 20(4): p. 617-623.
[8] Chou, C.-T. and C.-Y. Chen, Carbon dioxide recovery by vacuum swing adsorption. Separation and Purification Technology, 2004. 39(1–2):
p. 51-65.
[9] Ishibashi, M., et al., Technology for removing carbon dioxide from power plant flue gas by the physical adsorption method. Energy
Conversion and Management, 1996. 37(6–8): p. 929-933.
[10] Kikkinides, E.S., R.T. Yang, and S.H. Cho, Concentration and recovery of carbon dioxide from flue gas by pressure swing adsorption.
Industrial & Engineering Chemistry Research, 1993. 32(11): p. 2714-2720.
[11] Ko, D., R. Siriwardane, and L.T. Biegler, Optimization of Pressure Swing Adsorption and Fractionated Vacuum Pressure Swing Adsorption
Processes for CO2 Capture. Industrial & Engineering Chemistry Research, 2005. 44(21): p. 8084-8094.
[12] Liu, Z., et al., Multi-bed Vacuum Pressure Swing Adsorption for carbon dioxide capture from flue gas. Separation and Purification
Technology, 2011. 81(3): p. 307-317.
[13] Mehrotra, A., A. Ebner, and J. Ritter, Arithmetic approach for complex PSA cycle scheduling. Adsorption, 2010. 16(3): p. 113-126.
[14] Na, B.-K., et al., CO2 recovery from flue gas by PSA process using activated carbon. Korean Journal of Chemical Engineering, 2001. 18(2):
p. 220-227.
[15] Na, B.-K., et al., Effect of Rinse and Recycle Methods on the Pressure Swing Adsorption Process To Recover CO2 from Power Plant Flue
Gas Using Activated Carbon. Industrial & Engineering Chemistry Research, 2002. 41(22): p. 5498-5503.
[16] Nikolic, D., et al., Generic Modeling Framework for Gas Separations Using Multibed Pressure Swing Adsorption Processes. Industrial &
Engineering Chemistry Research, 2008. 47(9): p. 3156-3169.
[17] Plaza, M.G., et al., Post-combustion CO2 capture with a commercial activated carbon: Comparison of different regeneration strategies.
Chemical Engineering Journal, 2010. 163(1–2): p. 41-47.
[18] Reynolds, S.P., A.D. Ebner, and J.A. Ritter, Stripping PSA Cycles for CO2 Recovery from Flue Gas at High Temperature Using a
Hydrotalcite-Like Adsorbent. Industrial & Engineering Chemistry Research, 2006. 45(12): p. 4278-4294.
[19] Takamura, Y., et al., Evaluation of dual-bed pressure swing adsorption for CO2 recovery from boiler exhaust gas. Separation and
Purification Technology, 2001. 24(3): p. 519-528.
[20] Tlili, N., G. Grévillot, and C. Vallières, Carbon dioxide capture and recovery by means of TSA and/or VSA. International Journal of
Greenhouse Gas Control, 2009. 3(5): p. 519-527.
[21] Casas, N., et al., A parametric study of a PSA process for pre-combustion CO2 capture. Separation and Purification Technology, 2013.
104(0): p. 183-192.
[22] Schell, J., et al., Precombustion CO2 Capture by Pressure Swing Adsorption (PSA): Comparison of Laboratory PSA Experiments and
Simulations. Industrial & Engineering Chemistry Research, 2013. 52(24): p. 8311-8322.
[23] The Future of Coal, 2007, Massachusetts Institute of Technology.
[24] Ho, M.T., G.W. Allinson, and D.E. Wiley, Reducing the Cost of CO2 Capture from Flue Gases Using Pressure Swing Adsorption. Industrial
& Engineering Chemistry Research, 2008. 47(14): p. 4883-4890.
[25] Panowski, M., R. Klainy, and K. Sztelder, Modelling of CO2 Adsorption from Exhaust Gases, in Proceedings of the 20th International
Conference on Fluidized Bed Combustion, G. Yue, et al., Editors. 2010, Springer Berlin Heidelberg. p. 889-894.
[26] Froment, G.F., K.B. Bischoff, and J. De Wilde, Chemical Reactor Analysis and Design, 3rd edition2010: John Wiley & Sons, Inc.
[27] Liu, Z., et al., Adsorption and Desorption of Carbon Dioxide and Nitrogen on Zeolite 5A. Separation Science and Technology, 2011. 46(3):
p. 434-451.
[28] Lopes, F.V.S., et al., Adsorption of H2, CO2, CH4, CO, N2 and H2O in Activated Carbon and Zeolite for Hydrogen Production. Separation
Science and Technology, 2009. 44(5): p. 1045-1073.
[29] Jiang, L., V.G. Fox, and L.T. Biegler, Simulation and optimal design of multiple-bed pressure swing adsorption systems. AIChE Journal,
2004. 50(11): p. 2904-2917.
[30] Park, J.-H., J.-N. Kim, and S.-H. Cho, Performance analysis of four-bed H2 PSA process using layered beds. AIChE Journal, 2000. 46(4): p.
790-802.
[31] Ribeiro, A.M., et al., Four beds pressure swing adsorption for hydrogen purification: Case of humid feed and activated carbon beds. AIChE
Journal, 2009. 55(9): p. 2292-2302.
[32] DECARBit: Enabling advanced pre-combustion capture techniques and plants. European best practice guidelines for assessment of CO2
capture technologies, 2011, European Benchmarking Task Force.
[33] Pipitone, G. and O. Bolland, Power generation with CO2 capture: Technology for CO2 purification. International Journal of Greenhouse
Gas Control, 2009. 3(5): p. 528-534.
2304 Luca Riboldi et al. / Energy Procedia 63 (2014) 2289 – 2304

[34] Posch, S. and M. Haider, Optimization of CO2 compression and purification units (CO2CPU) for CCS power plants. Fuel, 2012. 101(0): p.
254-263.