Accepted Manuscript

Amine based CO2 capture - CFD simulation of absorber performance

Paweł Niegodajew , Dariusz Asendrych

PII: S0307-904X(16)30384-5
DOI: 10.1016/j.apm.2016.07.003
Reference: APM 11265

To appear in: Applied Mathematical Modelling

Received date: 3 December 2015

Revised date: 3 July 2016
Accepted date: 21 July 2016

Please cite this article as: Paweł Niegodajew , Dariusz Asendrych , Amine based CO2 cap-
ture - CFD simulation of absorber performance, Applied Mathematical Modelling (2016), doi:
10.1016/j.apm.2016.07.003

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service
to our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and
all legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT

Highlights

• A CFD absorber model linking flow hydrodynamics with CO2 absorption was proposed

• Good correspondence between the model and a reference lab installation was found

• A parametric study of the absorber was performed, showing its performance

• Capture efficiency is slightly influenced by varying CO2 content in flue gases

T
IP
CR
US
AN
M
ED
PT
CE
AC

1
 ACCEPTED MANUSCRIPT

Amine based CO2 capture - CFD simulation of absorber performance

Paweł Niegodajew, Dariusz Asendrych*

Institute of Thermal Machinery, Częstochowa University of Technology

al. Armii Krajowej 21, 42-200 Częstochowa, Poland

* corresponding author, email address: [email protected]

T
ABSTRACT

IP
The paper deals with the numerical modelling of CO2 removal from flue gases by amine solvent in an

CR
absorber column, i.e. the first stage of a post-combustion capture installation. The complex flow
system including multi-phase gas-liquid countercurrent streams with chemical reaction and heat
transfer is considered to resolve the CO2 absorption. A Computational Fluid Dynamics (CFD) model
simulates a small laboratory test rig which has provided the reference data for model validation. A

US
30% aqueous monoethanolamine (MEA) solution was employed as a chemical solvent. A 2-fluid
Eulerian model has been employed to resolve the flow in a random packed bed. The calculations were
carried out for varying lean loading (0.375-0.425 mol CO2/mol MEA), solvent to flue gases ratio
AN
(4.46-6.92kg/kg) and CO2 content (8-16%), with the results showing reasonable absorber behaviour
and satisfactory consistency with experimental data. It was found that the removal efficiency is weakly
dependent on CO2 content in exhaust gases as long as the amine solution loading is kept at a
M

sufficiently high level.

Keywords: carbon dioxide capture, chemical absorption, monoethanolamine (MEA), packed bed,
computational fluid dynamics (CFD)
ED

SYMBOLS

C molar concentration, kmol/m3

PT

Er enthalpy source term, W/m3

G mass flux of gas phase, kg/h
CE

J stream of species diffused, kg/(m3s)

L mass flux of liquid phase, kg/h
AC

M molecular weight, kg/kmol

MB momentum source term in porous zone, N/m3
Q heat flux exchanged between phases, W/m3
R heterogeneous reaction rate, kg/(m3s)
S mass source term, kg/(m3s)
T temperature, K
X molar fraction, mol/mol
Y mass fraction, kg/kg

2
 ACCEPTED MANUSCRIPT

h specific enthalpy, J/kg

h holdup, %
k reaction rate constant, m3/(kmols)
p static pressure, Pa
t time, s
u velocity, m/s
y axial coordinate, m
 volume fraction, m3/m3

T
 CO2 loading, mol CO2/mol MEA


IP
thermal conductivity, W/(mK)
 stoichiometric coefficient, -

CR
 dynamic viscosity, Pas
 CO2 capture efficiency, %
 density, kg/m3

SUBSCRIPTS
US
AN
CO2 denotes carbon dioxide
MEA denotes monoethanolamine
MEAC denotes MEA carbamate
M

MEAH denotes protonated MEA

N denotes nominal value
ED

in/out denotes inlet to/outlet from the absorber

l/g denotes liquid/gas phase
i/k denotes species/phase
PT
CE

1. Introduction
Concern about the environment is forcing reduction of greenhouse gases emission, in particular
AC

significant cuts in carbon dioxide being released from the power industry. Carbon Capture and Storage
(CCS) is regarded as the most efficient solution for the near future allowing CO2 emission targets to be
achieved. Among the available CCS technologies the "post-combustion" capture, in forms of
absorption, adsorption [53], membrane [22] and cryogenic, is the most promising for implementation
into industrial practice in the short term [30,47]. The remaining two CCS methods, i.e. "pre-
combustion capture" and "oxy-fuel combustion" require much more research effort and higher
investment before their commercial introduction into the power engineering industry [21,47].

The present paper is devoted to the absorption-based post combustion capture (PCC) which can be
easily adopted by existing power plant installations operated on various kinds of fossil fuels emitting
different contents of acid gases [3]. Its implementation requires building a CCS unit and integrating it

3
 ACCEPTED MANUSCRIPT

with a power plant cycle. The principle of operation of a capture unit is presented in Fig. 1. Flue gases
are sent through the absorber where carbon dioxide is captured by chemical absorption with the use of
an aqueous monoethanolamine (MEA) solution supplied to the column at its top. Both phases pass
through the packed bed where the most of the chemical reaction occurs due to the enlarged contact
area between phases. The loaded (or rich) MEA (containing absorbed CO2) solution is pre-heated
before entering the stripper where, with the help of heat (coming from the power plant cycle), the
carbon dioxide is released, then sent to the compression unit and transported to the storage location.
The recycled (lean) MEA then flows through the heat exchanger (improving process efficiency) and
finally enters the absorber being ready for the next capture cycle. MEA is the most common choice of
chemical solvent used for CO2 absorption thanks to its high reactivity [31], relatively low cost [47] and
well recognized behaviour in other installations of the chemical industry [25,37].

T
As the CCS technology is still under development and extensive research studies, the source literature

IP
is rapidly growing. Two main approaches to the subject include experimental investigations with the
use of laboratory or pilot CC installations [15,32,37,41,45,50] as well as numerical modelling by

CR
means of 1D models with the use of commercial software like ASPEN Plus [3,4,7] or gPROMS
[29,30] as well as in-house codes (e.g. [13,25,28,48]). Two modelling strategies have been used to
simulate MEA-H2O-CO2 system, namely the equilibrium-based approach and the rate-based approach

US
[51,58]. These approaches including their versions allow to model the CO2 absorption process at
various levels of complexity. In the present paper a simplified strategy is used and all the mass transfer
and chemical reactions aspects are represented by the reaction rate constants. Moreover, the
AN
Computational Fluid Dynamics (CFD) has been employed, a powerful numerical tool allowing for an
insight into the details of the process including its complex hydrodynamics in 3D space. Considering
the packed beds with the use of 1D models does not allow to take into account the actual
characteristics of the porous media regardless their structured or random pattern. Modelling two-phase
M

gas-liquid flow in a complex geometry is a challenging issue [52], however, it is a crucial point to
adequately simulate the heat and mass transfer processes taking place in the absorber column. A
number of papers has been published in recent years considering the liquid flow patterns and pressure
ED

drop of non-reacting gas-liquid systems in various geometries of the packed beds, however, mostly of
the structured character. Xu and co-workers [54-56] using 3D model based on volume of fluid (VOF)
method considered the liquid film flow on flat plate for various flow conditions and liquid properties.
PT

The main finding of their simulations was the variety of liquid flow pattern types including rivulets
and droplet flows leading to the incomplete wetting of the wall surface. Haroun and co-workers [18]
simulated the liquid film flow in a structured packing represented by the 2D wavy surface. They have
CE

found that the liquid follows the surface shape creating a uniform film only for very low Reynolds
numbers. For increasing Re, liquid started to agglomerate in surface cavities increasing significantly
liquid static holdup and reducing interface area. Developing their model Haroun and co-workers [17]
AC

performed a 3D CFD-VOF simulation of a liquid flow in a commercially available packing Mellapack

250X. The simulation results have shown the influence of liquid flow rate and the contact angle on a
flow structure taking different patterns, i.e. from the droplet flow through the rivulets to the film flow,
and consequently affecting the wall wetted area. CFD modelling has been also applied to simulate the
liquid flow hydrodynamics in structured packed beds coupled with physical mass transfer between gas
and liquid phases in 2D [19] and 3D [44] geometries. Haroun and co-workers [16] considered
additionally the reactive absorption in a gas-liquid flow using Direct Numerical Simulation (DNS)
approach showing the importance of the liquid flow pattern on the intensity of mass transfer.

Modelling the flow hydrodynamics in a packed bed gains additional importance when liquid spreading
has to be taken into account. Effective operation of the absorption column requires highly efficient

4
 ACCEPTED MANUSCRIPT

liquid transfer across the porous zone from the locations where the liquid is supplied from the
distributors. CFD modelling may help in optimisation of the packing columns by the comparative
analyses of different designs of distributing systems as well as types and sizes of packing elements.
Fourati and co-workers [12] performed 2D CFD simulation of the liquid redistribution in a Mellapack
250X packed bed by adding to the momentum equation the source term responsible for the spreading
effect. The liquid volume fraction distribution has been found to be in a qualitative agreement with the
reference experimental data [11], however, quantitative discrepancies indicate the need for further
model development. An alternative approach to model liquid dispersion in porous media composed of
the Raschig rings was presented in [40], who developed a 3D two-fluid Eulerian CFD model. The
liquid spreading was simulated by implementing into the momentum equation the statistical
description of the liquid velocity vector orientation. The corresponding probability density function

T
was determined by the VOF simulation of the liquid flow through the real-geometry random packed

IP
bed generated numerically [34]. The simulation has shown that for the point liquid supply the liquid
volume fraction radial profiles are highly non-uniform even several column diameters downstream the
top of the packing section. It shows the importance of the correct flow hydrodynamics reconstruction

CR
in a packed bed, especially when the complex carbon dioxide absorption process has to be modelled.

In the present paper a 2D axisymmetric CFD model of the CO2 capture process taking place in a 1.1m

US
high packing column has been developed and employed to perform a parametric study allowing the
determination of some optimization guidelines, which are extremely important from the viewpoint of
CCS cost reduction. The model is an extension of the previous work [6] where the authors used a
AN
simplified mechanism of the interaction between countercurrently flowing gas-liquid phases in porous
zone. The correlation for the interphase interaction force proposed by [33] allowed correct prediction
of liquid holdup (see [39]) and consequently all the processes built on the flow hydrodynamics. It
should be noted that these simulations revealed practical independence of the resulting flowfield of the
M

radial position, thus allowing for its description as a 1D system with sufficient accuracy. The choice of
a 2D modelling approach (instead of a 1D strategy, being much more computationally efficient) in the
reported work resulted from the authors' intention to further develop the CFD model in a way enabling
ED

a combined simulation of a carbon dioxide capture process with a nonuniform flowfield being a
consequence of a spatial variation of packed bed porosity and/or a nonuniform supply of a column
with a liquid.
PT

The reported research work was carried out in collaboration with the Institute for Chemical Processing
of Coal (IChPW) in Zabrze (Poland) within the frames of the Strategic Research Programme -
CE

Advanced Technologies for Energy Generation: Development of a technology for highly efficient
zero-emission coal-fired power units integrated with CO2 capture. IChPW was responsible in the
project for the experimental studies of CCS with the use of a small-scale laboratory installation, being
the reference for testing and validation of the numerical model.
AC

2. Model description

2.1. Object of the simulation

Numerical simulation of the entire PCC unit is a challenging task as it is necessary to achieve the
thermo-chemical equilibrium in the system. It means that all the parameters characterising the working
media have to reach steady-state levels, which are unknown before the simulation is started. To
achieve this, lengthy computations are required. That is why it was decided in the reported work to
split the PCC cycle into two parts and to consider absorber and stripper as separate partial CFD
models. The mutual interactions between the units of CO2 capture installation are replaced by the

5
 ACCEPTED MANUSCRIPT

carefully chosen boundary conditions. The model formulated in such a way, although simplified, may
still be regarded as a numerical tool allowing for adequate and much more computationally efficient
analysis and optimisation of the PCC cycle. The present paper is devoted to the simulation of CO2
absorption process only. It should be noted that the authors in parallel have developed a CFD model of
the stripper column and the preliminary results can be found in [38].
The model under development reflects the geometry and dimensions of the CCS installation at IChPW
being schematically shown in Fig. 2. The absorber column is of cylindrical shape with the following
main dimensions: diameter of 0.1m, height of 1.5m and the height of the working section (packed bed)
equal to 1.1m. Standard 6mm Raschig rings were used as a packing material. The nominal volumetric
fluxes of media in the column are: 8m3/h and 0.054m3/h of flue gases and MEA solution, respectively,
corresponding to the mass fluxes ratio L/G = 6.92kg/kg. Details of the experimental arrangement and

T
its characteristics can be found in [6,27]. The operational parameters of the column including media

IP
composition are collected in Tab. 1. It should be noted that at this stage of the project the mixture of
CO2 and air was used instead of actual flue gases outgoing from combustion chambers. In addition, the
outlet temperatures of both phases and the amount of CO2 not captured are collected in Tab. 2. They

CR
were used for model validation.

Table 1. Operating parameters of media at their inlets to the column used in experiment and simulation

mass flux [kg/h]

3
volumetric flux [m /h]
flue gas
8
7.8
US amine solvent
0.054
54
AN
air 88 H2O 70
mass fraction Y [%]
CO2 12 MEA 30
temperature T [K] 298.5 313
M

In previous works the authors applied arbitrary lean loadings, i.e. in = 0-0.25mol CO2/mol MEA [6]
or just in = 0 [39]. In the present paper the base value of CO2 loading at the absorber inlet in was
chosen to be 0.4 mol CO2/mol MEA. The authors' intention was to reflect the experimental conditions
ED

of the reference installation operated at IChPW [27]. Although the solvent composition was not
measured in the lab installation, the lean loading range could be estimated by fitting simulation results
to experimental data, i.e. temperatures and capture efficiency. It should be noted that the laboratory
PT

installation was working at relatively low stripper temperatures (105-110C), i.e. yielding high loading
values, so the in = 0.4 mol CO2/mol MEA may be regarded as justified. It should be also noted that
high loadings required relatively high L/G ratios in order to reach sufficiently high capture efficiency.
CE

Although the value in = 0.4 mol CO2/mol MEA seems rather high (majority of pilot installations are
operated at lower levels, say 0.2-0.25), there are a number of works applying loading values
comparable to that chosen in the reported work. For instance, Mores and co-workers [36], when
AC

modelling stripper performance, applied in a part of a simulation series relatively high loading
conditions included in the range 0.35-0.41 mol CO2/mol MEA. Dinca and Badea [7] in their
optimization study have arrived at the conclusion that the lowest heat required for solvent regeneration
corresponds to the following lean loading values: 0.42 mol CO2/mol MEA and 0.44 mol CO2/mol
MEA for 40 wt.% and 30 wt.% aqueous MEA solution, respectively. The lean loadings around 0.4
mol CO2/mol MEA were also applied by Gáspár and Cormoş [14], Tobiesen and co-workers [48] and
Plaza and co-workers [43].
For the assumed base lean loading value, the corresponding mass fractions of amine species in liquid
bulk equal to:
- YMEA = 5.52%
- YMEAH = 11.23%

6
 ACCEPTED MANUSCRIPT

- YMEAC = 18.83%

Table 2. Outlet parameters of the absorber column - experiment

parameter value
temperature of the gas [K] 311
temperature of the solvent [K] 313.2
CO2 mass fraction [%] 1.2
In order to achieve stable and convergent solution the simulations of the countercurrent flow of gas-
liquid mixture were carried out with the use of the following boundary conditions (see Fig. 2):

T
- the pressure inlet (at the column bottom) and outlet (at its top) boundary conditions for the gas
phase ensuring the prescribed flue gases flux,

IP
- liquid source located just above the porous zone.
It should be noted that due to complex physio-chemical interactions between exothermically reacting

CR
phases counterflowing in a porous zone, it was not possible to predict the flow characteristics
including the pressure drop coefficient. Thus in order to establish the required gas flow rate the inlet-
outlet pressure difference was determined iteratively in few steps. Such a system of boundary

US
conditions allows the phases freely leave the domain, i.e. aqueous MEA solution flows down (driven
by gravity) reaching bottom of the column, while gas quits the domain at its top. Concerning the
thermal boundary conditions the heat flux exchanged with the surroundings was described by the fixed
AN
ambient temperature and the heat transfer coefficient through the absorber wall.

2.2. Governing equations

The model of the absorber column was developed as a 2-dimensional axi-symmetric, laminar and
M

unsteady multiphase gas-liquid flow. The 2-fluid Euler-Euler approach was implemented to describe
the two-phase countercurrent flow. In this approach fluid phases are treated as interpenetrating
ED

continua. In order to describe the coexistence of different phases in the flow the concept of volumetric
fraction is introduced. The volume fractions for all the phases sum up to unity

 k 1 (1)
PT

Conservation equations are formed for each phase separately. The equation of mass conservation for
kth phase takes the form
CE

 
 k  k    k  k u k   S k (2)
t

AC

where u k is the velocity vector, k the phase density and Sk is the mass source term corresponding to
species production/destruction due to chemical reaction.
The momentum equation for kth phase in Eulerian multiphase model (assuming flow incompressibility)
has the following form:
 

t
  
  
 
 k  k u k    k  k u k u k    k p      k  k u k  u kT   k  k g  M B,k (3)


where p is the static pressure shared by all phases, k stands for the dynamic viscosity, g is the gravity

vector and M B,k describes the interaction force between phases in porous zone according to [33] and
its implementation into the code is described in detail in [39].

7
 ACCEPTED MANUSCRIPT

For the reacting system additional governing equations have to be included in the model. The general
transport equation of ith species in multiphase flow is given as

  
( k  k Yi ,k )  ( k  k u k Yi ,k )   k J i ,k  Ri ,k (4)
t

where Yi,k is the mass fraction, Ri stands for the heterogeneous reaction rate and J i ,k is the flux of the ith
species due to diffusion. In laminar flow the diffusion flux can be described by Fick’s law.

To describe the energy transfer in Eulerian model, a separate enthalpy equation is solved for kth phase


 k  k hk    k  k uk hk     k p   k Tk   Qk  Er ,k


T
(5)
t t

IP
where hk is the specific enthalpy, k stands for the thermal conductivity, Qk is the intensity of heat
exchange between phases and Er,k is the enthalpy source term due to chemical reaction.

CR
The chemistry of CO2 absorption by aqueous monoethanolamine solution is regarded as a non-
reversible reaction [8,9,20,49] of the form:

US
CO 2  2RNH 2  RNHCOO   RNH 3

resulting in formation of the MEA carbamate (first term on the r.h.s.) and the protonated MEA (second
term on the r.h.s.) regarded together as loaded MEA or rich MEA. In the above reaction, R represents
(6)
AN
the alcanol group (CH2)2OH–. It should be noted that some researchers treat the chemical absorption of
carbon dioxide as a system of several chemical reactions (see [2,4,36]). Depending on the chemical
+ -
mechanisms various rare ions may appear in the system like bicarbonate HCO3 , hydroxide OH and
M

-
carbonate CO32 . However, their concentrations contents for typical lean MEA temperatures and
loading values are negligible and the CO2 capture process may be represented by Eq. 6 with sufficient
accuracy [2,9,35]. The molar rate of ith chemical species produced due to the heterogeneous second-
ED

order reaction is given by:

Ri  k  i  CMEA  CCO 2 (7)

PT

where k is the reaction rate constant, i represents the stoichiometric coefficient and CMEA and CCO2
stand for the molar concentrations of reacting media, i.e. MEA and carbon dioxide, respectively. It
should be noted that fluxes of species described by Eq. 7 sum up for the kth phase to the mass source
CE

term appearing in continuity equation (Eq. 2) as follows:

Sk   M k ,i  Rk ,i  (8)
AC

where Mk,i is the molecular weight. The reaction rate constant as a function of temperature may be
approximated by the following expression:
log(k )  10.99  2152/ T (9)
which was adopted from [14].

2.3. Numerical tools

The ANSYS FLUENT v.13 was employed to develop the model of the absorber column. The
Eulerian/Eulerian approach has been chosen to describe the two-phase countercurrent flow. The 2nd

8
 ACCEPTED MANUSCRIPT

order discretization schemes were applied ensuring sufficient accuracy of computations. The SIMPLE
scheme was used for the pressure correction.

The ANSYS GAMBIT (FLUENT's pre-processor) was used for geometry and numerical grid
generation. The optimal grid was produced as a result of series of tests concerning various variants of
meshing strategies (including structured and unstructured meshes) as well as several grid sizes. A
number of test calculations have shown that the mesh size of appr. 13,000 cells provided the best
compromise between high accuracy and computational speed. More detailed information about the
mesh and its generation can be found in [6].

3. Simulation results

T
IP
3.1. Carbon dioxide absorption
The CFD model presented in previous sections has been applied to conduct a series of simulations of

CR
the carbon dioxide absorption by monoethanolamine. At first the calculations followed the set of
parameters (see Tab. 1) used during the experimental test trials being performed at IChPW with the
use of the laboratory PCC installation.

US
Before the use of the complete CCS model, a series of simulations was conducted focussing on the
hydrodynamics of the countercurrent flow through the packed bed (i.e. excluding chemistry and
thermal effects). The results presented in [6] showed the reasonable behaviour of the countercurrent
AN
gas-liquid streams, however, the predicted holdup levels in porous zone were significantly too low,
affecting the overall picture of the CO2 capture process. In the reported work the model was updated
by the adequate phase interaction force term according to [33].
M

The presentation of simulation results is started from the demonstration of the liquid holdup through
the packing section as well as distributions of chemical species taking part in the reaction. Fig. 3
presents contour maps of liquid holdup (Fig. 3a) and mass fractions of reactants (Fig. 3b,c). One can
ED

easily notice practically constant level of liquid holdup inside the packed bed (see Fig. 3a) which
results from a dominating character of the flow resistance in a porous zone (represented by MB term in
momentum equation Eq. 3) over the other forces. The mass fraction contour maps (Fig. 3b,c) show
PT

that the gradual change of MEA content (Fig. 3b), when passing vertically through the porous zone,
leads to a corresponding decrease of CO2 share in flue gases (see Fig. 3c). Distributions shown in Fig.
3 prove the practical absence of radial gradients of all the quantities under consideration. Therefore, it
CE

was decided to present hereafter the results just as the axial distributions, which are easier for analysis
and interpretation. Consequently the liquid holdup and the mass fractions of all species taking part in
the reaction (Eq. 6), i.e. including also the absorption product, are presented in Fig. 4, along the
AC

column axis.

The contour maps presented in Fig. 3 provided the qualitative information only, while the axial
profiles allow for the quantitative analysis of the results. As can be seen from Fig. 4a the liquid holdup
reaches its maximum level immediately downstream the top of the porous zone and becomes constant
(as previously discussed) within entire packing section. One can notice that nonlinear character of
mass fraction profiles presented in Fig. 4b. It is especially evident for the carbon dioxide distribution,
and much less pronounced for MEA and its relatives. The inlet CO2 content was 12% of flue gases,
while at the outlet it fell down to 0.8% (see Fig. 4b), which means the capture efficiency reaches
93.6%. The predicted carbon dioxide distribution has been compared with the available experimental
data [27], which are shown in Fig. 4b as empty circles. CO2 content values are the same at the inlet as

9
 ACCEPTED MANUSCRIPT

the boundary condition reflects experimental configuration. The outlet CO2 mass fraction was
underpredicted when compared with the experimental value equal to 1.2% corresponding to the 90%
capture efficiency. It should be remarked that the CO2 removal efficiency is dependent on the lean
MEA loading, being assumed arbitrarily in the simulation at the level of in=0.4 though not known
exactly in experiment. Discussion of the influence of the lean amine loading on the CO2 capture
efficiency is presented in further section of the present paper.

In addition to carbon dioxide the distributions of MEA and absorption products are presented in Fig.
4b. Their profiles are fully consistent with CO2 distribution and they result from the stoichiometry of
the absorption reaction. The previously discussed decrease in CO2 mass fraction is accompanied by the
corresponding reduction of MEA content and increase in MEA carbamate (denoted as MEAC) and

T
protonated MEA (denoted as MEAH). It should be noted, that the level of MEA decreased from its
inlet level equal to 5.52%, reaching the output value of 2.17% (see Fig. 4b), which means that about

IP
60% of MEA reacted with CO2 for the above mentioned capture efficiency. The molar concentrations
of both reaction products must be the same as it implies from the stoichiometry of absorption reaction

CR
(see Eq. 6). However, the mass fraction profiles (shown in Fig. 4b) differ quantitatively as the molar
mass of carbamate is greater than the one of protonated MEA. It is worth noting that for the lean
amine loading in=0.4 a significant excess of absorption products (appr. 85% of circulating solution)

US
enters the absorber column not taking part in the reaction. Such an excessive amount of amine solvent
is necessary to keep the amine loading at the level corresponding to the equilibrium conditions
established as a result of the mutual interactions between absorber and stripper [7].
AN
As the reaction of CO2 absorption by MEA is an exothermic process, it is of great importance to
control the temperature field in the column. Its distributions are shown in Fig. 5a for both phases
together, and additionally the reaction rate profile is presented in Fig. 5b. As can be seen from Fig. 5a
M

gas inlet temperature is noticeably lower than liquid temperature. Downstream the absorber inlet gas is
warmed up and nearly reaches the liquid temperature at the top of the absorber. One can note that gas
temperature rises appr. 14K while liquid temperature is constant along the entire column, which can be
ED

explained with the immensely (appr. two orders of magnitude) greater thermal capacity of the liquid
phase.

As a supplement to the simulation results, temperature values taken from the experiment [27] have
PT

been put in Fig. 5a. As can be seen the actual outlet temperature of the rich MEA (see Fig. 5a, open
square for y=0) exceeds the simulation result by approximately 0.2K, while the outlet temperature of
decarbonised flue gas obtained in the simulation slightly overpredicts the experiment by 0.4K. The
CE

slight discrepancies between experimental and simulation data with respect to the temperature and the
CO2 mass fraction data do not exceed the corresponding measurement uncertainties.
AC

The reaction rate profile shown in Fig. 5b indicates significant variation along the column with its
maximum located at the bottom of the porous zone exceeding its level at the top part by appr. one
order of magnitude. Such a shape of R may be explained with the use of Eq. 7 showing the linear
dependence of the reaction rate on the molar concentrations (consequently on mass fractions) of
reacting species.

As can be seen from Fig. 4b, moving down from the top of the column:
- the mass content of MEA decreases two times,
- the share of CO2 increases by more than one order of magnitude.
Therefore the mass fraction of CO2 plays a key role in determining the reaction rate profile.

10
 ACCEPTED MANUSCRIPT

In order to validate the CFD model in a more convincing way temperatures of both working media
were analysed in detail and compared with the reference experimental data [27]. Temperature changes
(i.e. rise in gas temperature and drop in liquid one) within the absorber column are presented in Fig. 6
as a function of the L/G. One may easily notice a very good correspondence between simulated and
measured data. Both sets of results indicate the same trends, i.e. increasing gas temperature rise and
decreasing liquid temperature drop with increasing amine flux. Moreover, the data are quantitatively
consistent as the discrepancies do not exceed 0.5K for the gas phase and 0.3K for the liquid one.
Summing up the results presented in Figs. 4-6, one may conclude that the CFD model developed in the
paper is able to reflect the complex physio-chemical CO2 absorption process with good accuracy and
thus it may be treated as an adequate numerical tool.

T
3.2. Influence of solvent composition

IP
The CFD model of carbon dioxide removal in the PCC technology, presented in the previous section,
was used to perform a parametric study allowing to show the performance of absorption column under

CR
varying conditions. In the present chapter the level of solvent regeneration in stripper is analysed as
one of key parameters determining CO2 capture performance. In real CCS installations, the mass
fractions of MEA and its absorption products (carbamate and protonated MEA) result from the
thermo-chemical equilibrium achieved in the system. This means that with the present CO2 absorption

US
model (stripper neglected) it is not possible to capture the interactions between both parts of the cycle.
Thus the inlet composition of amine solution had to be substituted with the arbitrarily set inlet
boundary condition and the level of CO2 loading was adopted from the source literature. The extensive
AN
literature studies including reports from pilot CCS installations [10,14,41] as well as simulation results
[3,7] reveal a large scatter of information. For instance Gáspár and Cormoş [14], using a simulation
model developed within the MATLAB/Simulink environment, applied αin=0.18 in the incoming
M

solvent flux. Experimental investigation of [5] showed that the loading at the absorber inlet reaches
αin=0.27, whereas the analysis with the use of pilot plant data [7] revealed that the lean loading value
could even exceed the level of αin=0.5. It should be remarked that such a high values of CO2 loading
ED

can be reached only when the more complex chemical model (including several reactions involving
additionally water as a reaction species) of carbon dioxide absorption is considered [2,4]. As discussed
in section 2.1 the value of αin=0.4 was regarded as a nominal level in the current paper. For the
purpose of present section the following three loading values and the corresponding mass fractions of
PT

MEA were applied:

 αin = 0.375 - YMEA,in = 6.94%,
 αin = 0.40
CE

- YMEA,in = 5.52%,
 αin = 0.425 - YMEA,in = 4.12%.
The mass content of MEA in aqueous solution as well as flue gases composition were unchanged.
Simulations were conducted for the above cases and the results are presented in Fig. 7 as the axial
AC

distributions of mass fractions of absorption reactants and products. As can be seen from Fig. 7a the
increase of inlet MEA loading (related to the decrease of amount of MEA) leads to reduced CO2
capture efficiency and for the analysed flue gases composition (12% CO2 mass content) the outlet CO2
level increases from appr. 0.29% to 2.0%. It is worth noticing that the change of CO2 content is
nonlinearly related to the loaded solvent fraction. The change of loading from αin= 0.375 to αin = 0.40
leads to a corresponding increase in CO2 from 0.29% to 0.78%, while its growth from αin = 0.40 to αin
= 0.425 (the same increase) gives more than double the increase in CO2 (from 0.78% to 2.0%). The
axial distributions along the column of the mass fraction of MEA are shown in Fig. 7b. The YMEA
profiles corresponding to varying content of loaded amine (represented by ) are nearly parallel and
they reveal the gradual decrease in the downstream direction. Consequently, the distributions of the

11
 ACCEPTED MANUSCRIPT

mass fractions of the reaction products MEAH and MEAC presented in Figs. 7c and 7d indicate the
corresponding growth resulted from the stoichiometry of the reaction (6).
The qualitative influence of lean loading on the species mass fraction distributions is complemented
quantitatively in Tab. 3, including the loading values at absorber inlet and outlet as well as their
change within the column. One can easily notice that the change of inlet loading leads to its nonlinear
change at the outlet which results from the progressively decreasing MEA content in the solution and
its absorption potential.

Tab. 3. CO2 loading at the absorber inlet and outlet

αin αout α

T
[mol CO2/mol MEA]

IP
0.375 0.439 0.064
0.4 0.46 0.060

CR
0.425 0.478 0.053

In addition to the already presented results, the profiles of the reaction rate and temperatures of both

US
phases are collected in Fig. 8. As can be seen from Fig. 8a for the lowest lean loading value the
reaction rate distribution has a local maximum located at the bottom of the packing section. When
moving up the column, the reaction rate significantly falls down reaching its minimum at the top of the
AN
porous zone. In order to explain the shape of the reaction rate distribution one should note that the
absorption is a 2nd order reaction, which means that the molar fractions of reagents (MEA and CO2)
contribute linearly to its rate (see Eq. 7). Accordingly the location of the maximum is associated with
the highest content of CO2, which has a major impact on the reaction rate as YCO2 changes in much
M

wider range than YMEA (see Fig. 7a,b). When moving up, the CO2 mass content significantly decreases
while the MEA content increases in limited extent leading finally to the deceleration of absorption
process. For the increasing lean loading the reaction rate:
ED

- decreases at the absorber bottom which is mainly caused by significant drop (several times) in
YMEA (see Fig. 7b) and the limited (20-30%) rise in YCO2 (see Fig. 7a),
- increases at the upper part of the column which results from substantial CO2 content growth
PT

(approximately one order of magnitude - see Fig. 7a) and much less pronounced YMEA decrease
(about two times - see Fig. 7b).
Consequently with increase of lean loading, the reaction rate is inhibited in the bottom part of the
CE

packing section, while it is intensified in its top part. Therefore the maximum of the distribution
becomes less and less pronounced and moves upward, from about y=0.12m for αin=0.375 through
about y=0.26m for αin= 0.40 to about y=0.52m for αin=0.425. It is interesting to note that all the
AC

reaction rate curves intersect in close vicinity of the absorber centre. The demarcation between upper
and bottom parts of the column (determined by the intersection region of the reaction rate profiles)
corresponds to the location where both influencing factors (i.e. CO2 and MEA fractions) are in
balance. Outside the porous region the reaction rate is abruptly decelerated (see Fig. 8a) and does not
contribute to the CO2 capture process.

In Figs. 8b,c, influence of solvent composition on temperatures of both media is presented. It can be
easily seen that lean loading has no influence on gas temperature as all the curves in Fig. 8b collapse
into one distribution. Similarly liquid temperature is hardly influenced by amine solution composition
as can be seen from Fig. 8c. Although the curves presented here indicate noticeable discrepancies, they
are of minor importance when the range of temperatures T0.1K is taken into account. This justifies

12
 ACCEPTED MANUSCRIPT

the neglect of the effect of liquid temperature in the discussion concerning the evolution of the
reaction rate distribution. Joint analysis of temperature profiles collected in Fig. 8b,c shows that the
most intensive interphase heat exchange takes place at the bottom part of the packing section where
the liquid-gas temperature difference and the reaction rate take the maximum values.

3.3. Influence of CO2 content in flue gases

Carbon dioxide content in flue gases is never constant and it is strongly dependent on the fuel type and
the combustion conditions [3,24,57] and may vary from 3% to as high as 20%. Thus it is important to
evaluate the absorber performance, especially in terms of capture efficiency under varying flue gas
composition. To make the post combustion process more efficient, the varying content of CO2 in flue

T
gases should be considered to show the behaviour of the installation. For this purpose three CO2 levels
were investigated, in particular 8%, 12% (regarded as a nominal mass fraction YCO2,in,N corresponding

IP
to the coal-fired power plants) and 16%, with air as the complement. The results are presented in Fig.
9 as the axial distributions of mass fractions of reacting species. Distributions shown in Fig. 9a show

CR
that the rising CO2 inlet content results in the increasing concentration of CO2 at the outlet. This
parameter changes from 0.36% for 8% CO2 content in flue gases, through 0.78% for 12% CO2 content
and reaches 1.66% for 16% CO2 inlet concentration. This quantitative comparison shows that with a
linear change of CO2 content in flue gases the concentration of the CO2 at the column outlet grows

US
nonlinearly, which is attributed to the gradually decreasing CO2 absorption potential of the solvent.
Fig. 9b shows how the changing flue gases composition affects the MEA distribution along the
column. The increasing carbon dioxide mass fraction leads to the acceleration of the reaction rate and
AN
in turn to the intensified consumption of monoethanolamine. As a consequence the outlet MEA
content decreases from 3.3% for 8% CO2 inlet content to 1.16% for 16% CO2 inlet concentration.
As a complement to already presented data the profiles of the mass fractions of reaction products (both
M

components of loaded MEA) are presented in Figs. 9c and 9d.

The mass fractions of both species rise with increasing CO2 content. It is worth noticing that all the
ED

curves, i.e. corresponding to CO2, MEAH, MEAC, are of identical shape reflecting again the
stoichiometry of absorption reaction.
The axial distributions of the reaction rate and temperatures of both working media are collected in
PT

Fig. 10. Reaction rate profile for the lowest carbon dioxide content (see Fig. 10a) reaches the
maximum value just above the entrance to the packing section and decreases in nonlinear manner
reflecting the profile of CO2 mass fraction (see Fig. 9a). With a rise in CO2 concentration in flue gases
CE

the reaction rate is:

- intensified by the higher CO2 molar concentration,
- inhibited by the reduced MEA molar content.
AC

Consequently the reaction rate at the top part of the packing section significantly increases and the
maximum of its profile is shifted upward. It should be also noted that increase in CO2 content in flue
gases leads to more intensive reaction in the entire absorber indicating that carbon dioxide
concentration is the factor that has the most influence on the reaction rate. Temperature profiles for gas
and liquid phases (see Fig. 10b,c) show their practical independence of the carbon dioxide content in
flue gases. Slight influence of YCO2 on Tl=f(YCO2) can be observed, however temperature range
Tl=0.25K is of minor importance in terms of its impact on reaction rate (see Eq. 9).

As a complement to the already presented results, the capture efficiency η as a function of both CO2
content and lean solvent loading is presented in Fig. 11. The capture efficiency was defined as a ratio
of the absorbed CO2 mass flux to its flux entering the column. The efficiency decreases with the

13
 ACCEPTED MANUSCRIPT

increasing CO2 content from 95.6% for 8% CO2 inlet content through 93.5% for 12% CO2 inlet level
down to 89.6% for 16% CO2 concentration. One should note very weak influence of the carbon
dioxide content change (although doubled) in flue gases on the capture efficiency. This influence can
be explained by the intensified reaction rate due to higher CO2 (as reagent) concentration which is
manifested by the growing area under the reaction rate curve (see Fig. 10a). In addition, the impact of
varying lean MEA loading within the range αin  <0.375; 0.425> on capture efficiency is shown in
Fig. 11 with the use of square symbols. Taking in mind the relative changes around the nominal values
of CO2 mass fraction (YCO2,in=(1±0.33)YCO2,in,N) and lean loading (in=(1±0.06)in,N), one can easily
conclude that the CO2 removal efficiency is much more strongly dependent on the solvent
regeneration level than on the flue gas composition. In particular the efficiency decreases with a rise in
lean loading from 97.7% for αin=0.375, through 93.7% for αin=0.4 down to 83.2% for αin=0.425. It is

T
worth noticing that the slope of =f(αin) significantly increases when approaching the limit level

IP
αin=0.5, resulting from the simplified absorption chemistry assumed in the present paper (see Eq. 6).
The simulation results agree qualitatively well with the experimental data [1] obtained from an

CR
extensive analysis of absorption process in a pilot plant with absorber column 2.4m high. Similarly the
removal efficiency is noticeably less sensitive to the variation of CO2 content in flue gases (changed
within the range 5-20%) than to lean MEA loading. It should be also noted that the experiment [1] was
carefully performed and well instrumented, providing detailed information about the variation of

US
process parameters inside the absorber column. In particular the CO2 concentration profiles were
measured for various process configurations indicating very good correspondence to the present CFD
results presented in Figs. 7a, 9a.
AN
3.4. Influence of the liquid to gas ratio
The liquid-gas ratio is one of the key parameters governing CO2 capture efficiency as well as the
M

running costs of the CCS process [15,23,32,35,46]. In order to show its influence on the absorber
performance, additional simulations were conducted for varying L/G value, which was changed by the
liquid flux with the flue gases stream kept constant. Three L/G values were analysed, i.e.
ED

L/G=6.92kg/kg, being the nominal flux ratio applied in the simulations so far, and additionally
L/G=4.46kg/kg and L/G=5.71kg/kg. The remaining key process parameters, i.e. carbon dioxide mass
fraction in flue gases and lean MEA loading were maintained at their nominal levels, i.e. YCO2=12%,
in=0.4. The first set of results in the form of axial distributions of species mass fraction is presented
PT

in Fig. 12. CO2 content at the absorber outlet presented in Fig. 12a decreases with the rise of L/G from
about 1.92% for L/G=4.46kg/kg through 1.17% for L/G=5.71kg/kg down to 0.78% for
CE

L/G=6.92kg/kg. As could be expected, with the gradual increase in L/G, the amount of CO2 captured
decreases as the removal efficiency approaches 100%. This trend shows that the excessive growth in
the solvent flux, although improving capture efficiency, may lead to the rise in CCS process costs paid
AC

for the heating in the stripper. Identical influence of varying L/G on the CO2 content profiles were
found experimentally by Aboudheir and co-workers [1], proving the relevance of the simulations
performed with the present CFD model. In Figs. 12b-d the distributions of remaining species are
presented for the varying L/G. One could notice that the rise in solvent flux results in reduced relative
MEA consumption (see its increasing content at the absorber outlet for the growing L/G in Fig. 12b)
and the corresponding evolutions of absorption products (see Fig. 12c,d). It is worth noting that the
L/G has the opposite impact on MEA and absorption products than the CO2 content in flue gases
(compare Figs. 9b-d with 12b-d).

As a supplement to the mass fraction profiles the axial distributions of the reaction rate as well as
temperatures of both countercurrently flowing media are collected in Fig. 13. As can be seen from Fig.

14
 ACCEPTED MANUSCRIPT

13a the increase in L/G slows down the reaction rate in the upper part of the column, while it supports
the CO2 absorption at the bottom of packing section. It can be explained with the reduced CO2 content
at the absorber top (due to higher capture efficiency) and the increased mass fraction of MEA at the
bottom (resulting from increased solvent flux). As a consequence all curves intersect in the vicinity of
the column centre. Similarly to the simulation results presented in the previous sections the reaction
rate is influenced by the mass fractions of reagents and slightly dependent on the liquid temperature
which varies in a limited extent of less than 2K for the applied L/G range (see Fig. 13c). The gas
temperature profiles presented in Fig. 13b indicate qualitatively similar shape to those for the previous
cases, increasing in the upward direction asymptotically to the liquid temperature, however not
reaching it. Variation of L/G, in contrast to CO2 content in flue gases and inlet MEA loading, leads to
quantitative modification of gas temperature profiles. In particular the change of L/G in the range

T
4.46-6.92kg/kg results in a temperature difference at the absorber top equal to appr. 2K.

IP
As a complement to already presented results, the CO2 capture efficiency as a function of L/G for
three levels of lean loading (in=0.375, 0.4, 0.425) and constant carbon dioxide level in flue gases, are

CR
collected in Fig. 14. The simulation results are compared with the corresponding data acquired at the
reference laboratory installation located in IChPW Zabrze [26,27]. Unfortunately the information
about the solvent composition (and solvent CO2 loading) were not available. The simulation data

US
points corresponding to the constant in levels were approximated with second order polynomials. As
can be seen, the capture efficiency indicates an increasing tendency with growing solvent flux as well
as with decreasing inlet MEA loading. The slope of  profiles, reflecting the change of capture
AN
efficiency, increases with the reduction of L/G and for the rising lean MEA loading. Similar
tendencies are observed in the experimental data available from the CCS pilot plants [5]. The
simulation data seem not to be fully consistent with the reference capture efficiencies obtained from
the IChPW pilot installation [27] as far as the slopes of  profiles are concerned. It is quite evident that
M

for the simulation (open symbols in Fig. 14) the change of efficiency within the analysed range of L/G
is noticeably overpredicted when compared to the experiment (see solid symbols in Fig. 14). In order
to explain these discrepancies one should note the following:
ED

- as is known from source literature [36,42], with the rise in solvent flux the inlet MEA loading is
growing as a result of thermo-chemical equilibrium achieved by the PCC cycle at new operating
point,
PT

- in present simulations the assumption was made about the constant  level.
Therefore, by progressively increasing the value of lean loading for higher amine solvent fluxes the
resulting capture efficiency would go down modifying the slope of the =f(L/G) profile towards the
CE

experimental distribution. Bearing in mind the above discussion the CFD model prediction in Fig. 14
may be regarded as consistent with the experimental data.
AC

4. Conclusions
A Computational Fluid Dynamics (CFD) model has been developed to investigate carbon dioxide
capture by chemical absorption in an absorber column. The model combines the complex
hydrodynamics of 2-phase gas-liquid flow in a random packed bed with absorption chemistry and
accompanying thermodynamics. An Euler-Euler 2-fluid multiphase approach was able to resolve the
countercurrently flowing streams of flue gases and 30% aqueous monoethanolamine (MEA) solution
in a packed bed composed of 6mm Raschig rings. The model has been validated with the use of
reference experimental data provided by laboratory CCS installation operated at the Institute for
Chemical Processing of Coal.

15
 ACCEPTED MANUSCRIPT

The process parameters applied during simulations were rather unusual (flue gas temperature =
25C, lean loading = 0.375-0.425 mol CO2/mol MEA, L/G = 4.46-6.92kg/kg), however, they have
been chosen to follow the experimental conditions. Good correspondence has been found between
model predictions and the reference data. The parametric simulations for the varying process
parameters have confirmed the tendencies known from the literature proving the model relevance. The
process has been found to be quite complex due to mutual interactions between particular partial
processes (convection, diffusion, heat release, species production/destruction), which makes the
prediction of the absorber performance challenging. An interesting finding of the paper is that the
absorber performance is slightly influenced by varying the CO2 content in flue gases, over a wide
range. This result shows the flexibility of the CCS installation which may effectively function under
non-stationary working conditions (change of fuel, unsteady phases of power plant operation, etc.)

T
without noticeable reduction of capture efficiency.

IP
It should be noted that the proposed 2D CFD model is relatively computationally expensive when
compared to a 1D modelling strategy, commonly used to simulate absorption/desorption columns.

CR
However, it has a great potential as it can be extended to combine a carbon dioxide capture process
with nonuniform fluid flow hydrodynamics. It is planned in further work to implement liquid
spreading model, spatial variation of packed bed porosity and partial catalyst wetting and thus

US
enabling to investigate flow maldistribution phenomena including wall effects.
AN
Acknowledgements
The results presented in this paper were obtained from research work co-financed by the National
Centre of Research and Development in the framework of Contract SP/E/1/67484/10 - Strategic
M

Research Programme - Advanced technologies for energy generation: Development of a technology

for highly efficient zero-emission coal-fired power units integrated with CO2 capture.
ED

Also the Statutory Funds under the contract BS/PB-1-103-3010/2011/P are gratefully acknowledged.

References
PT

[1] A. Aboudheir, D. DeMontigny, P. Tontiwachwuthikul, Important Factors Affecting Carbon

Dioxide Removal Efficiency By Using Extra-high Concentrated Monoethanolamine Solutions
CE

and High-Capacity Packings, in: SPE (Society of Petroleum Engineers) Gas Technology
Symposium, Calgary, Alberta, Canada, 15-18 March 1998, pp. 1-4.
[2] A. Aboudheir, P. Tontiwachwuthikul, A. Chakma, R. Idem, Kinetics of the reactive absorption
AC

of carbon dioxide in high CO2-loaded, concentrated aqueous monoethanolamine solutions,

Chem. Eng. Sci. 58 (2003) 5195-5210.
[3] C. Alie, L. Backham, E. Croiset, P.L. Douglas, Simulation of CO2 capture using MEA
scrubbing: a flowsheet decomposition method, Energy Convers. Manag. 46 (2005) 475-487.
[4] U.S. Arachchige, M.C. Melaaen, Aspen plus simulation of CO2 removal from coal and gas fired
power plants, Energy Procedia 23 (2012) 391-399.
[5] Y.Artanto, J. Jansen, P. Pearson, T. Do, A. Cottrell, E. Meuleman, P. Feron, Performance of
MEA and amine-blends in the CSIRO PCC pilot plant at Loy Yang Power in Australia, Fuel
101 (2012) 264-275.
[6] D. Asendrych, P. Niegodajew, S. Drobniak, CFD Modelling of CO2 Capture in a Packed Bed by
Chemical Absorption, Chem. Process Eng. 34 (2013) 269-282.

16
 ACCEPTED MANUSCRIPT

[7] C. Dinca, A. Badea, The parameters optimization for a CFBC pilot plant experimental study of
post-combustion CO2 capture by reactive absorption with MEA, Int. J. Greenh. Gas Control 12
(2013) 269-279.
[8] R. Faiz, M. Al-Marzouqi, Mathematical modelling for the simultaneous absorption of CO2 and
H2S using MEA in hollow fiber membrane contractors, J. Memb. Sci. 324 (2009) 269-278.
[9] L. Faramarzi, G.M. Kontogeorgis, M.L. Michelsen, K. Thomsen, E.H. Stenby, Absorber Model
for CO2 Capture by Monoethanolamine, Ind. Eng. Chem. Res. 49 (2010) 3751-3759.
[10] F. Ferrara, G. Calì, C. Frau, A. Pettinau, Experimental and numerical assessment of the CO2
absorption process in the Sotacarbo pilot platform, in: 1st Int. Conf. on Sustainable Fossil Fuels
for Future Energy, Rome, 2009, pp. 1-12.

T
[11] M. Fourati, V. Roig, L. Raynal, Experimental study of liquid spreading in structured packings,
Chem. Eng. Sci. 80 (2012) 1-15.

IP
[12] M. Fourati, V. Roig, L. Raynal, Liquid dispersion in packed columns: Experiments and
numerical modeling, Chem. Eng. Sci. 100 (2013) 266-278.

CR
[13] J. Gabrielsen, M.L. Michelsen, E.H. Stenby, G.M. Kontogeorgis, Modeling of CO2 Absorber
Using an AMP Solution, AIChE J. 52 (2006) 3443-3451.
[14] J. Gáspár, A-M. Cormoş, Dynamic modeling and validation of absorber and desorber columns

US
for post-combustion CO2 capture, Comput. Chem. Eng. 35 (2011) 2044-2052.
[15] H.R. Godini, D. Mowla, Selectivity study of H2S and CO2 absorption from gaseous mixtures by
MEA in packed beds, Chem. Eng. Res. Des. 86 (2008) 401-409.
AN
[16] Y. Haroun, D. Legendre, L. Raynal, Direct numerical simulation of reactive absorption in gas-
liquid flow on structured packing using interface capturing method, Chem. Eng. Sci. 65 (2010)
351-356.
[17] Y. Haroun, L. Raynal, P. Alix, Prediction of effective area and liquid hold-up in structured
M

packings by CFD, Chem. Eng. Res. Des. 92 (2014) 2247-2254.

[18] Y. Haroun, L. Raynal, D. Legendre, Mass transfer and liquid hold-up determination in
ED

structured packing by CFD, Chem. Eng. Sci. 75 (2012) 342-348.

[19] J. Hu, X. Yang, J. Yu, G. Dai, Numerical simulation of carbon dioxide (CO2) absorption and
interfacial mass transfer across vertically wavy falling film, Chem. Eng. Sci. 116 (2014) 243-
253.
PT

[20] S. Jayarathna, B. Lie, M.C. Melaaen, Development of a Dynamic Model of a Post Combustion
CO2 Capture Process, Energy Procedia 37 (2013) 1760-1769.
CE

[21] M. Kanniche, R. Gros-Bonnivard, P. Jaud, J. Valle-Marcos, J-M. Amann, C. Bouallou, Pre-

combustion, post-combustion and oxy-combustion in thermal power plant for CO2 capture,
Appl. Therm. Eng. 30 (2010) 53-62.
AC

[22] S. Khaisri, D. DeMontigny, P. Tontiwachwuthikul, R. Jiraratananon, CO2 stripping from

monoethanolamine using a membrane contactor, J. Memb. Sci. 376 (2011) 110-118.
[23] J.N. Knudsen, J.N. Jensen, P-J. Vilhelmsen, O. Biede, Experience with CO2 capture from coal
flue gas in pilot-scale: Testing of different amine solvents, Energy Procedia 1 (2009) 783-790.
[24] M. Kosowska-Golachowska, K. Kłos, T. Musiał, The fragmentation of coal particles during
oxy-fuel combustion in a circulating fluidized bed, Arch. Combust. 32 (2012) 131-139.
[25] A. Kothandaraman, L. Nord, O. Bolland, H.J. Herzog, G.J. McRael, Comparison of solvents for
post-combustion capture of CO2 by chemical absorption, Energy Procedia 1 (2009) 1373-1380.
[26] A. Krótki, D. Śpiewak, L. Więcław-Solny, A. Wilk, A. Tatarczuk, Wpływ natężenia przepływu
gazu i cieczy na sprawność absorpcji CO2 w 30% roztworze monoetanoloaminy, Inżynieria i
Apar. Chem. 53 (2014) 23-26.

17
 ACCEPTED MANUSCRIPT

[27] A. Krótki, L. Więcław-Solny, A. Tatarczuk, A. Wilk, D. Śpiewak, Laboratory studies of CO2

absorption with the use of 30% aqueous monoethanolamine solution, Arch. Combust. 12 (2012)
195-203.
[28] H.M. Kvamsdal, A. Chikukwa, M. Hillestad, A. Zakeri, A. Einbu, A comparison of different
parameter correlation models and the validation of an MEA-based absorber model, Energy
Procedia 4 (2011) 1526-1533.
[29] H.M. Kvamsdal, J.P. Jakobsen, K.A. Hoff, Dynamic modeling and simulation of a CO2 absorber
column for post-combustion CO2 capture, Chem. Eng. Process. 48 (2009) 135-144.
[30] A. Lawal, M. Wang, P. Stephenson, G. Koumpouras, H. Yeung,. Dynamic modelling and
analysis of post-combustion CO2 chemical absorption process for coal-fired power plants, Fuel
89 (2010) 2791-2801.

T
[31] F. Lecomte, P. Broutin, E. Lebas, CO2 Capture: Technologies to Reduce Greenhouse Gas

IP
Emissions, IFP Public. ed. Paris.
[32] S. Lee, S. Maken, J-W. Park, H-J. Song, J.J. Park, J-G. Shim, J-H. Kim, H-M. Eum, A study on

CR
the carbon dioxide recovery from 2ton-CO2/day pilot plant at LNG based power plant, Fuel 87
(2008) 1734-1739.
[33] J. Maćkowiak, Fluid Dynamics of Packed Columns, Principles of the Fluid Dynamic Design of
Columns for Gas/Liquid and Liquid/Liquid Systems, Springer-Verlag, 2010.

US
[34] M. Marek, Numerical generation of a fixed bed structure, Chem. Process Eng. 34 (2013) 347-
359.
AN
[35] P. Mores, N. Scenna, S. Mussati, Post-combustion CO2 capture process: Equilibrium stage
mathematical model of the chemical absorption of CO2 into monoethanolamine (MEA) aqueous
solution, Chem. Eng. Res. Des. 89 (2011) 1587-1599.
[36] P. Mores, N. Scenna, S. Mussati, CO2 capture using monoethanolamine (MEA) aqueous
M

solution: Modeling and optimization of the solvent regeneration and CO2 desorption process,
Energy 45 (2012) 1042-1058.
[37] P. Moser, S. Schmidt, G. Sieder, H. Garcia, T. Stoffregen, Performance of MEA in a long-term
ED

test at the post-combustion capture pilot plant in Niederaussem, Int. J. Greenh. Gas Control 5
(2011) 620-627.
[38] P. Niegodajew, D. Asendrych, S. Drobniak, Numerical modelling of CO2 desorption process
PT

coupled with phase transformation and heat transfer in CCS installation, J. Power Technol. 93
(2013) 354-362.
[39] P. Niegodajew, D. Asendrych, S. Drobniak, Numerical analysis of CO2 capture efficiency in
CE

post combustion CCS technology in terms of varying flow conditions, Arch. Thermodyn. 34
(2013) 123-136.
[40] P. Niegodajew, D. Asendrych, M. Marek, S. Drobniak, Modelling liquid redistribution in a
AC

packed bed. J. Phys. Conf. Ser. 530 (2014) 1-8.

[41] R. Notz, N. Asprion, I. Clausen, H. Hasse, Selection and Pilot Plant Tests of New Absorbents
for Post-Combustion Carbon Dioxide Capture, Chem. Eng. Res. Des. 85 (2007) 510-515.
[42] R. Notz, H.P. Mangalapally, H. Hasse, Post combustion CO2 capture by reactive absorption:
Pilot plant description and results of systematic studies with MEA, Int. J. Greenh. Gas Control 6
(2012) 84-112.
[43] J.M. Plaza, D. van Wagener, G.T. Rochelle, Modeling CO2 capture with aqueous
monoethanolamine, Int. J. Greenh. Gas Control 4 (2010) 161-166.
[44] D. Sebastia-Saez, S. Gu, P. Ranganathan, K. Papadikis, 3D modeling of hydrodynamics and
physical mass transfer characteristics of liquid film flows in structured packing elements, Int. J.
Greenh. Gas Control 19 (2013) 492-502.

18
 ACCEPTED MANUSCRIPT

[45] L.L. Simon, Y. Elias, G. Puxty, Y. Artanto, K. Hungerbuhler, Rate based modeling and
validation of a carbon-dioxide pilot plant absorbtion column operating on monoethanolamine,
Chem. Eng. Res. Des. 89 (2011) 1684-1692.
[46] T.L. Sønderby, K.B. Carlsen, P.L. Fosbøl, L.G. Kiørboe, N. von Solms, A new pilot absorber
for CO2 capture from flue gases: Measuring and modelling capture with MEA solution, Int. J.
Greenh. Gas Control 12 (2013) 181-192.
[47] B.P. Spigarelli, S.K. Kawatra, Opportunities and challenges in carbon dioxide capture, J. CO2
Util. 1 (2013) 69-87.
[48] F.A. Tobiesen, H.F. Svendsen, O. Juliussen, Experimental Validation of a Rigorous Absorber
Model for CO2 Postcombustion Capture, AIChE J. 53 (2007) 846-865.

T
[49] P.D. Vaidya, E.Y. Kenig, CO2-Alkanolamine Reaction Kinetics: A Review of Recent Studies,
Chem. Eng. Technol 30 (2007) 1467-1474.

IP
[50] I. von Harbou, H.P. Mangalapally, H. Hasse, Pilot plant experiments for two new amine
solvents for post-combustion carbon dioxide capture, Int. J. Greenh. Gas Control 18 (2013) 305-

CR
314.
[51] M. Wang, A. Lawal, P. Stephenson, J. Sidders, C. Ramshaw, Post-combustion CO2 capture with
chemical absorption: A state-of-the-art review, Chem. Eng. Res. Des. 89 (2011) 1609-1624.

US
[52] Y. Wang, J. Chen, F. Larachi, Modelling and simulation of trickle-bed reactors using
computational fluid dynamics: A state-of-the-art review, Can. J. Chem. Eng. 91 (2013) 136-180.
[53] D. Wawrzyńczak, D. Bukalak, I. Majchrzak-Kucęba, W. Nowak, Effect of Desorption Pressure
AN
on CO2 Separation from Combustion Gas by Means of Zeolite 13X and Activated Carbon,
Polish J. Environ. Stud. 23 (2014) 1437-1440.
[54] Y.Y. Xu, S. Paschke, J-U. Repke, J.Q. Yuan, G. Wozny, Portraying the Countercurrent Flow on
Packings by Three-Dimensional Computational Fluid Dynamics Simulations, Chem. Eng.
M

Technol. 31 (2008) 1445-1452.

[55] Y.Y. Xu, S. Paschke, J-U. Repke, J.Q. Yuan, G. Wozny, Computational Approach to
Characterize the Mass Transfer between the Counter-Current Gas-Liquid Flow, Chem. Eng.
ED

Technol. 32 (2009) 1227-1235.

[56] Y.Y. Xu, J.Q. Yuan, J-U. Repke, G. Wozny, CFD study on liquid flow behavior on inclined flat
plate focusing on effect of flow rate, Eng. Appl. Comput. Fluid Mech. 6/2 (2012) 186-194.
PT

[57] M. Zajemska, D. Musiał, A. Poskart, Application of Chemkin and Comsol programs in the
chemical composition calculations of natural gas combustion products. Combust. Sci. Technol.
186 (2014) 153-172.
CE

[58] M. Zaman, J.H. Lee, Carbon capture from stationary power generation sources: A review of the
current status of the technologies, Korean J. Chem. Eng. 30 (2013) 1497-1526.
AC

19
 ACCEPTED MANUSCRIPT

gure 1
ick here to download high resolution image

T
IP
CR
US
AN
M
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT
igure 2
lick here to download high resolution image

T
IP
CR
US
AN
M
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

gure 3
ick here to download high resolution image

T
IP
CR
US
AN
M
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

gure 3 color
ick here to download high resolution image

T
IP
CR
US
AN
M
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

gure 4
ick here to download high resolution image

T
IP
CR
US
AN
M
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

gure 5
ick here to download high resolution image

T
IP
CR
US
AN
M
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

gure 6
ick here to download high resolution image

T
IP
CR
US
AN
M
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

gure 7
ick here to download high resolution image

T
IP
CR
US
AN
M
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

gure 8
ick here to download high resolution image

T
IP
CR
US
AN
M
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

gure 9
ick here to download high resolution image

T
IP
CR
US
AN
M
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

gure 10
ick here to download high resolution image

T
IP
CR
US
AN
M
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

gure 11
ick here to download high resolution image

T
IP
CR
US
AN
M
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

gure 12
ick here to download high resolution image

T
IP
CR
US
AN
M
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

gure 13
ick here to download high resolution image

T
IP
CR
US
AN
M
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

gure 14
ick here to download high resolution image

T
IP
CR
US
AN
M
ED
PT
CE
AC