SIMS 63 Trondheim, Norway, September 20-21, 2022
Process Simulation and Automated Cost Optimization of CO2
Capture Using Aspen HYSYS
Lars Erik Øi*, Pouya Rahmani, Sumudu Karunarathne, Solomon Aromada
Department of Process, Energy and Environmental Technology, University of South-Eastern Norway
[Link]@[Link]
Abstract
A standard process for CO2 capture based on absorption in mono ethanolamine (MEA) has been simulated and
cost estimated with an equilibrium-based model in Aspen HYSYSTM. The aim has been to calculate cost optimum
process parameters and evaluate the possibility of automated cost optimization using a spreadsheet facility. An
Excel spreadsheet is used for dimensioning and cost estimation of the specified process. New in this work is that
Visual Basic for Application (VBA) was used to automatically update installation factors for next iteration based
on cost calculations in previous iteration. The equipment cost was obtained from the Aspen In-plant Cost
EstimatorTM, and an enhanced detailed factor (EDF) method was used to estimate the total investment cost. The
optimum process was found as the process with minimum calculated total cost. The cost optimum process
parameters for the standard process were calculated to 15 m absorber packing height, 9 °C minimum approach
temperature and 2.2 m/s superficial gas velocity through the absorber. With this approach, iterative cost estimation
and optimization of CO2 absorption and desorption processes can be performed automatically.
Keywords: Carbon capture, Aspen HYSYS, simulation, cost estimation, optimization
1. Introduction HYSYS. Some of the previous works are Kallevik
1.1. Aim (2010), Øi (2012), Park and Øi (2017), Aromada and
One aim of this work was to calculate cost optimum Øi (2017), and Øi et al. (2021).
process parameters for a traditional amine based The cost estimation part has been based on the
CO2 capture process. Another aim was to evaluate Enhanced Detailed Factor (EDF) method (Ali et al.,
the possibility of automated cost optimization using 2019; Aromada et al., 2021). While this method has
a spreadsheet facility. An Excel spreadsheet is several advantages, the time required to implement
traditionally used for dimensioning and cost the detailed installation factors in capital cost
estimation of the specified process with optimization estimation is a drawback. This becomes
performed by minimizing the total cost calculated in cumbersome when there is a need to run several
the spreadsheet. New in this work is that Visual process simulations by varying a process parameter
Basic for Application (VBA) was used to followed by cost estimation for each iteration.
automatically update installation factors for next Recently, the focus has been on automatic process
iteration based on cost calculations in previous simulation combined with cost estimation for fast
iteration. cost optimization of CO2 capture processes in Aspen
HYSYS (Haukås, 2020; Øi et al., 2021). The
1.2. Literature Iterative Detailed Factor (IDF) scheme was then
There is a large number of papers on cost estimation developed (Aromada et al., 2022a). The IDF scheme
of CO2 capture plants (Rao and Rubin, 2002; Rubin was applied for several minimum temperature
et al., 2013; van der Spek et al., 2019). Some of approach cost optimization studies in (Aromada et
these are based on a combination of process al., 2022a). However, there was yet a need for
simulation and cost estimation (Abu-Zahra et al., manual observation for implementing any change
2007; Amrollahi et al., 2012; Nwaoha et al., 2018). required in the detailed installation factors and
This work is a further development of earlier works subfactors whenever process parameters are varied
at the Telemark University College and the for subsequent simulation of the CO2 capture
University of South-Eastern Norway (USN). The process. Therefore, there is a need to make the entire
projects have been focused on process simulation, process simulations, equipment dimensioning and
equipment dimensioning, capital and operating cost cost estimation automatic, without requiring any
estimation, and cost optimization of CO2 capture manual input as done in the IDF scheme mode of
processes using the process simulation tool Aspen implementation (Aromada et al., 2022a). This was
SIMS 63 Trondheim, Norway, September 20-21, 2022
accomplished in this work by linking Aspen HYSYS
simulation spreadsheets with Microsoft Excel by a
VBA code. This has been discussed by Sharma and
Rangaiah (2016).
The Aspen HYSYS library was activated in
Microsoft Excel from the developer tab > visual
basic > tools> and preference (Rangaiah, 2016;
Rahmani, 2021). The Aspen HYSYS root needs to
be inserted into a Microsoft Excel sheet and it should
be updated for each model. The Aspen HYSYS
application should be closed during the process. The
VBA script was developed for coupling the Aspen
HYSYS and Microsoft Excel spreadsheets. The
code imports equipment prices from the Aspen
HYSYS spreadsheet into an Excel spreadsheet for
cost estimation and optimization. At the same time,
the right equipment units’ installation factors/sub- Figure 1: Process flow diagram of a standard amine-based
factors are automatically imported from the CO2 capture process (Aromada et al., 2020).
Microsoft Excel sheet into the Aspen HYSYS
spreadsheet. By this, the EDF method is simply and 2. Specifications and simulations
automatically implemented very fast without errors 2.1. Specifications and simulation of standard CO2
in the selection of the detailed installation capture process
factors/subfactors. The code and more details are Specifications for the base case are given in Tab. 1.
documented in (Rahmani, 2021). In the base case, 85 % CO2 removal efficiency and a
With this new approach of involving a VBA code, minimum approach temperature of 10 °C was
human errors in selecting EDF method installation achieved in the lean/rich heat exchanger. Aspen
factors and subfactors for different equipment are HYSYS version 11 was used with the Acid Gas
eliminated. And most importantly, each time a new package as the recommended equilibrium model by
process simulation is performed, the costs are Aspen HYSYS.
automatically available without requiring any form The calculation sequence is similar to earlier works
of manual inputs. (Øi and Haukås, 2021). The calculation strategy is
With this work, process simulation based CO2 based on a sequential modular approach (Kisala et
capture process parameter cost optimization studies al., 1987; Ishii and Otto, 2008). The calculation
and sensitivity analysis can be conducted quickly starts with the flue gas inlet stream and a guessed
and obtain reasonably accurate results. This paper amount of the lean amine stream. After calculation
documents cost optimization studies conducted with of the DCC and the absorption column, the
this new approach and comparison with other works. amine/amine heat exchanger is calculated based on
This work is based on the Master thesis work of a guessed (or specified) temperature in the stream
Rahmani (2021). from the desorber. The temperature can be adjusted
in an adjust block to obtain a specified minimum
1.3. Process Description temperature approach. After calculation of the
Prior to the CO2 capture process, the flue gas is desorber, the lean amine pump and the amine/amine
cooled in a direct contact cooler (DCC) with heat exchanger the lean amine cooler is calculated to
circulating water before it is sent to the absorption give an updated lean amine amount and
column. The amine with absorbed CO2 from the composition.
bottom of the absorption column is pumped through The updated lean amine amount and composition is
the rich/lean heat exchanger with the temperature checked in a recycle block with the amine stream
after the heat exchanger specified. The hot amine from the last iteration.
solution is entering the desorption column which Adjust and Recycle operations in the flowsheet are
separates the feed into the CO2 product at the top and used to get an automated simulation model. An
hot regenerated amine at the bottom. The Adjust block is adjusting the minimum approach
regenerated amine is pumped to a higher pressure in temperature in the lean/rich heat exchanger by
a pump, then passes through the lean/rich heat varying the temperature on the hot side outlet. A
exchanger and is further cooled in the lean cooler Recycle block is adjusting the removal efficiency by
before it again enters the absorption column. varying the lean amine mass flow. The Aspen
HYSYS flowsheet is shown in Fig. 2.
Water was added to the process (water make-up).
The make-up water can be calculated by a material
balance.
SIMS 63 Trondheim, Norway, September 20-21, 2022
3. Cost estimation procedures and assumptions
Table 1: Aspen HYSYS model parameters and
3.1. Equipment dimensioning and assumption
specifications for the base case alternative.
The equipment was dimensioned using the mass and
Value energy balances obtained from the process
Parameter
simulations as done in previous works (Øi et al.,
Inlet flue gas temperature [oC] 110/40.0 2021).
The absorber was specified to have 10 stages and the
Inlet flue gas pressure [kPa] 101/120 desorber 6 stages (Aromada and Øi, 2017). Each
Inlet flue gas flow rate [kmol/h] 10910 stage of both columns was assumed to be 1 meter
CO2 content in inlet gas [mole %] 3.3 (Aromada et al., 2020). A constant stage (Murphree)
efficiencies of 0.15 and 0.5 were specified for the
Water content in inlet gas [mole %] 6.9 absorber and the desorber respectively (Aromada et
o al., 2022b). Structured packing was specified for
Lean amine temperature [ C] 40.0
Lean amine pressure [kPa] 101.0 both the absorber and desorber. Superficial gas
velocity of 2.5 m/s was applied to estimate the
Lean amine rate [kg/h] 132100 diameter of the absorber (Øi et al., 2020). For the
MEA content in lean amine [mass %] 22.5 desorber, the desorber was evaluated using a
superficial gas velocity of 1 m/s (Park and Øi, 2017).
CO2 content in lean amine [mass %] 3.5
The tangent-to-tangent shell height of the absorber
Number of stages in absorber [-] 10 was specified to be 40 meter (Aromada et al.,
Murphree efficiency in absorber [m ] -1
0.25 2022a). A tangent-to-tangent shell height of 15 m
was used for the desorber. The height in both
Rich amine pump pressure [kPa] 200.0
columns were necessary to account for distributors,
o
Rich amine temp. out of HEX [ C] 103.6 water wash packing, demister, gas inlet, outlet and
Number of stages in desorber [-] 6 sump.
-1 The sizing of the direct contact cooler and the flash
Murphree efficiency in desorber [m ] 0.5
tank were based on Souders Brown’s equation with
Reflux ratio in stripper [-] 0.3 k-parameter 0.15 and 0.075 respectively. The heat
o
Reboiler temperature [ C] 120.0 exchange equipment units were dimensioned based
on the heat exchange areas calculated from the heat
Lean amine pump pressure [kPa] 500.0 duties. The overall heat transfer coefficient of 500
W/(m2K), 800 W/(m2K), 800 W/(m2K), and 1000
W/(m2K) were specified for the lean/rich heat
2.2. Parameter variations exchanger, lean amine cooler, reboiler and
10 stages, 85 % removal efficiency and 10 °C as condenser respectively (Aromada et al., 2020).
minimum approach temperature were specified in The pumps were specified as centrifugal pumps with
the base case simulation. For parametric studies, the 75 % adiabatic efficiency. They were sized based on
packing height and minimum approach temperature flow rate and duty.
were varied. The gas velocity through the absorber
column and the pressure drop were also varied. 3.2. Capital cost estimation methods
The Case study function in Aspen HYSYS was used The capital costs were estimated using the Enhanced
to perform a series of calculations automatically. An Detailed Factor (EDF) method (Ali et al., 2019;
important restriction is that when using the Case Aromada et al., 2021). The total capital cost is the
study function, it is not possible to perform other sum of all the equipment installed costs. The EDF
adjustments for each new parameter value. updated installation factors for 2020 by Nils Eldrup
was used (Aromada et al., 2021).
2.3. Process convergence and stability A traditional factor method for cost estimation is
A Recycle block and an Adjust function were used based on a table of factors multiplying the purchased
in the Flowsheet calculation. The Modified Hysim cost of each type of equipment unit. In a detailed
Inside-Out algorithm with adaptive damping was factor method, the total factor for each type is the
used according to a recommendation by Øi (2012). sum of contributions from e.g. installation,
This is the algorithm to solve the material, enthalpy electrical, instrumentation, administration etc. In the
and equilibrium equations in a column simulation. EDF method, these detailed factors are also
Stable convergence is especially important when dependent on the size and cost of the purchased
running a Case study in Aspen HYSYS for the equipment, so that the factors may change from one
purpose of optimization. iteration to the next in an optimization procedure.
SIMS 63 Trondheim, Norway, September 20-21, 2022
Figure 2: Aspen HYSYS flow-sheet of the base case simulation (Pouya, 2021).
The cost of each equipment was obtained from Currency exchange rate
9.21
Aspen In-plant Cost Estimator (v.10), based on the 2016
sizes estimated. The cost currency is in Euro (€) and Cost index 2016 103.6
the cost year is 2016. The default location for
Europe, Rotterdam was selected in this study. Cost index September 2020 111.3
Stainless steel (SS316) was specified for all
equipment. Welded equipment has a material factor 3.3. Operating cost estimation and assumption
of 1.75. The seamlessly manufactured equipment, The annual operating cost in this work was limited
the pumps and fan have a material factor of 1.3 (Øi to cost for consumption of steam, electricity, cooling
et al., 2021).
water and an annual maintenance cost as done in
Since this work is aimed at automatic cost
(Aromada and Øi, 2017). The electricity
estimation, the capital cost is initially estimated
consumption was based on the pump duties obtained
based on equipment costs obtained from Aspen In-
from Aspen HYSYS. Similarly, the steam
Plant Cost Estimator. Subsequent equipment cost is consumption was based on the reboiler steam duty
then estimated automatically based on the Power in kW. The annual hours of operation were assumed
law. The Power law is based on the assumption that
to be 8000 hours/year. The annual maintenance cost
the cost ratio of two sizes of a unit is proportional to
was specified as 4 %.
the dimension ratio raised to a power factor
(typically 0.65). The installation factors and
3.4. Net present value
subfactors are also automatically implemented in
each simulation iteration by the aid of the VBA code The cost metric in this work for cost optimization is
which connects Aspen HYSYS spreadsheets with negative net present value (NPV) as done in
Microsoft Excel spreadsheet. (Haukås, 2020). The NPV is the sum of investment
The economic assumptions used for the capital cost cost and the operation cost for each year in the plant
estimation are summarized in Tab. 2 (Rahmani, lifetime. The spreadsheet unit in Aspen HYSYS was
2021). used to calculate the detailed cost estimation of
CAPEX, OPEX and NPV (net present value). For
Table 2: Cost calculation specifications (Rahmani, 2021). the case of optimizing the temperature difference in
Parameter Value the main heat exchanger, the calculation could be
performed effectively by using the Case Study
Plant lifetime 20 years
option in Aspen HYSYS. The optimum solution can
Discount rate 8.5 % then be found by the simulation giving the lowest
4 % of installed NPV as shown in Fig. 3. For the case of optimizing
Maintenance cost the number of absorber stages, each calculation was
cost
performed independently by specifying the number
Electricity price 0.06 Euro/kWh of stages in each calculation.
Steam price 0.015 Euro/kWh
Annual operational time 8000 hours
Location Rotterdam
SIMS 63 Trondheim, Norway, September 20-21, 2022
4. Results and Discussion The minimum NPV when optimizing the absorber
packing height is 366 mill. EURO. This is a cost
4.1. Base case cost results reduction of 9 % compared to the base case.
For the base case, the total cost (or negative NPV) 4.4. Optimization of gas velocity
was estimated to 401 mill. EURO. This is the sum
of CAPEX and OPEX for a Plant lifetime of 20 When the superficial velocity through the absorption
years. column is increased, the cross section of the
absorber decreases and reduces the cost, while the
4.2. Optimization of minimum ΔT approach pressure drop increases and increases the cost. The
result when varying the superficial velocity is given
The minimum temperature approach optimization in Figure 5.
for the process is shown in Fig. 3. The absorber
packing height was 15 m in these optimizations. The
optimum value at 9 K can be found as the one with
minimum (negative) NPV. The cost optimum
minimum temperature approach has been calculated
in literature to values typically between 10 and 15 K.
The differences are due to different ratios between
cost of heat exchangers and cost of heat.
Figure 5: NPV as a function of superficial velocity with
85% capture rate, 20 years calculation period, and 8.5%
interest rate and constant packing volume.
Fig. 5 shows an optimum for a gas velocity of about
2.2 m/s. This is a value close to values from Park and
Øi (2017). There are not found other references in
the open literature showing similar results to
compare with.
5. Automation
Figure 3: NPV and energy consumption as a function of
ΔTmin with 85% capture rate, EM = 0.25, 20 years 5.1. Automation approaches
calculation period, and 8.5% interest rate. Automation of the simulations has been
investigated, and results have been compared with
4.3. Optimum absorber height manual simulations. Some of the input data should
The results from the optimization of the absorber be changed in the simulations manually, which is
packing height are given in Fig. 4. It shows an time-consuming. Connecting Excel and Aspen
optimum for 15 stages equivalent to 15 meter HYSYS to transfer the data is the first step toward
packing height. This is similar to results in earlier automating the process. In order to make the
work (Kallevik, 2010; Aromada and Øi, 2017; Øi et connection, one possibility is to use an Aspen
al., 2020). simulation workbook and programming in Visual
Basic. In addition, defining a case study in the Aspen
HYSYS can be useful for automating the
simulations.
The Aspen simulation workbook is an Excel feature
that can be activated through Excel's settings. The
Aspen HYSYS simulation model should be linked
to Excel, under the simulation tab in the Aspen
simulation workbook. Variables in the Aspen
HYSYS simulation can be copied to the Aspen
simulation workbook. In the scenario table, all of the
input data are collected once, and the simulation
Figure 4: NPV as a function of absorber packing height runs one at a time.
with removal efficiency 85%, EM=0.15, 20 years ΔTmin is considered as input in the lean-rich heat
calculation period and 8.5% interest rate. exchanger. The capture rate and NPV are considered
as outputs. In order to fix the capture efficiency at
about 85%, a controller is added to the simulation.
SIMS 63 Trondheim, Norway, September 20-21, 2022
Visual Basic for application (VBA) programming related to the limitations in Aspen HYSYS regarding
language in Excel is another method for automating possibilities for varying some specified parameters.
the process and cost estimation in Aspen HYSYS. In Examples are the number of stages in a column and
Aspen HYSYS, it can be activated in Excel from the the pressure drop for a column stage which have to
developer tab, visual basic, tools and preference. be specified manually.
One of the most time-consuming steps in cost
estimations is determining the correct installation
factor e.g. from a table. With this approach VBA 6. Conclusion
was used to automatically update installation factors The aim has been to calculate CO2 capture cost
for next iteration based on cost calculations in optimum process parameters and evaluate the
previous iteration. possibility of automated cost optimization using a
spreadsheet facility. The adjust and recycle blocks
5.2. Automatic optimization of ΔTmin are used to automate the energy and material balance
In Fig. 6, sensitivity analysis is performed for a specified simulation. An Excel spreadsheet is
comparing manual calculations and a case study used for dimensioning and cost estimation of the
including updated cost factors using the Visual specified process. New in this work is that Visual
Basic for Application approach as explained in Basic for Application (VBA) was used to
subsection 5.1. automatically update installation factors for next
iteration based on cost calculations in previous
iteration.
Equipment cost was obtained from Aspen In-plant
Cost EstimatorTM, and an enhanced detailed factor
method was used to estimate the total investment
cost. Parametric studies of the absorber packing
height and the minimum approach temperature in
the main heat exchanger were performed at 85 %
capture efficiency. The cost optimum process
parameters for the standard process were calculated
to 15 m absorber packing height, 9 °C minimum
Figure 6: NPV as a function of ΔTmin with removal approach temperature and 2 to 2.2 m/s superficial
efficiency 85%, EM=0.25, 20 years calculation period and gas velocity through the absorber.
8.5% interest rate for case study(automatic) and manually, With this approach, iterative cost estimation and
by using Aspen HYSYS model for ΔTmin=10°C. optimization of CO2 absorption and desorption
processes can be performed automatically.
The results are not exact equal for manual and Automatic optimization of some parameters like the
automatic calculations. The accuracy is however number of column stages is a challenge because they
reasonable, the difference in NPV is less than 1 %. must be specified manually.
The calculated optimum ΔTmin where 9 and 11 K,
respectively. The study shows that it is possible to References
calculate reasonable optimums automatically.
Abu-Zahra, M. R. et al. (2007) ‘CO2 capture from power plants:
5.3. Automatic optimization of column height and Part II. A parametric study of the economical performance based
gas velocity on mono-ethanolamine.’ International journal of greenhouse gas
control, 1(2), pp. 135-142.
Automatic optimization of column height (number Ali, H. (2019) Techno-economic analysis of CO2 capture
of absorber stages) is still a challenge. In Aspen concepts. PhD Thesis, University of South-Eastern Norway.
HYSYS, the number of stages is specified in the Amrollahi, Z. et al. (2012) ‘Optimized process configurations of
simulation input and can only be changed manually. post-combustion CO2 capture for natural-gas-fired power plant–
Power plant efficiency analysis.’ International Journal of
This is also the case for pressure drop in the absorber Greenhouse Gas Control, 8, pp. 1-11.
column which can be used for gas velocity Aromada, S. A. and Øi, L.E. (2017) ‘Energy and economic
optimizations. This limitation is not necessarily a analysis of improved absorption configurations for CO2
restriction in the future versions or in other tools, so capture.’ Energy Procedia, 114: pp. 1342-1351.
Aromada, S. A. et al. (2020) ‘Simulation and Cost Optimization
this is an interesting challenge for further work. of different Heat Exchangers for CO2 Capture’, Linköping
Electronic Conference Proceedings, SIMS 61, pp. 22-24.
5.4. Automatic optimization of other processes doi:10.3384/ecp20176318
The automatic optimization method in this work is Aromada, S. A. et al. (2021) ‘Capital cost estimation of CO2
capture plant using Enhanced Detailed Factor (EDF) method:
specific for a process simulation in Aspen HYSYS. Installation factors and plant construction characteristic
In principle, this approach could be used for any factors’ International Journal of Greenhouse Gas Control. 110,
process using any process simulation tool. However, pp.103394.
the specific challenges using Aspen HYSYS is
SIMS 63 Trondheim, Norway, September 20-21, 2022
Aromada, S. A. et al. (2022a) ‘Simulation-based Cost
Optimization tool for CO2 Absorption processes: Iterative
Detailed Factor (IDF) Scheme’, Scandinavian Simulation
Society, pp. 301-308.
Aromada, S. A. et al. (2022b) ‘Cost and Emissions Reduction in
CO2 Capture Plant Dependent on Heat Exchanger Type and
Different Process Configurations: Optimum Temperature
Approach Analysis’, Energies, 15(2), pp. 425.
Haukås, A. L. (2020) Process simulation and cost optimization of
CO2 capture Using Aspen HYSYS. Master’s Thesis, University of
South-Eastern Norway, Porsgrunn.
Ishii, Y. and Otto, F. D. (2008) ‘Novel and fundamental strategies
for equation-oriented process plowsheeting Part I: A basic
algorithm for inter-linked, multicolumn separation processes’
Computers and Chemical Engineering, 32, pp. 1842-1860.
Kallevik, O. B. (2010) Cost estimation of CO2 removal in HYSYS.
Master’s Thesis, Telemark University College, Porsgrunn.
Kisala, T. P. et al. (1987) ‘Sequential modular and simultaneous
modular strategies for process flowsheet optimization’,
Computers and Chemical Engineering, 11(6) pp. 567-579.
Nwaoha, C. et al. (2018), ‘Techno-economic analysis of CO2
capture from a 1.2 million MTPA cement plant using AMP-PZ-
MEA blend’, International Journal of Greenhouse Gas Control,
78 pp. 400-412.
Park, K. and Øi, L. E. (2017) ‘Optimization of gas velocity and
pressure drop in CO2 absorption column’, Linköping Electronic
Conference Proceedings SIMS 58, pp. 292-297. doi:
10.3384/ecp17138292
Rahmani, P. (2021) Process simulation and automated cost
optimization of CO2 capture using Aspen HYSYS. Master’s
Thesis, University of South-Eastern Norway, Porsgrunn.
Rao, A. B. and Rubin, E. S. (2002) ‘A technical, economic, and
environmental assessment of amine-based CO2 capture
technology for power plant greenhouse gas control’,
Environmental science & technology, 36(20) pp. 4467-4475.
Rubin, E. S. et al. (2013) ‘A proposed methodology for CO2
capture and storage cost estimates’, International Journal of
Greenhouse Gas Control, 17, pp. 488-503.
Sharma, S. and Rangaiah, G. P. (2016) ‘Mathematical Modeling,
Simulation and Optimization for Process Design. Chemical
Process Retrofitting and Revamping, Techniques for
Retrofitting and Revamping’, G. P. Rangaiah, Ed., 1. ed.
ProQuest Ebook Central: John Wiley & Sons, Incorporated, pp.
97-127.
Van der Spek, M. et al. (2019) ‘Best practices and recent
advances in CCS cost engineering and economic
analysis’, International Journal of Greenhouse Gas Control, 83,
pp. 91-104.
Øi, L. E. (2021) Removal of CO2 from exhaust gas. PhD Thesis,
Telemark University College, Porsgrunn.
Øi, L. E. et al. (2021) ‘Automated Cost Optimization of CO2
Capture Using Aspen HYSYS’, Linköping Electronic Conference
Proceedings SIMS 62, pp. 293-300. doi: 10.3384/ecp21185293
