ORIGINAL RESEARCH

published: 06 September 2019

doi: 10.3389/fenrg.2019.00081

Methanol Production via CO2

Hydrogenation: Sensitivity Analysis
and Simulation—Based Optimization
Prapatsorn Borisut and Aroonsri Nuchitprasittichai*

Department of Chemical Engineering, Suranaree University of Technology, Nakhon Ratchasima, Thailand

Carbon dioxide (CO2 ) is one of greenhouse gases, which can cause global warming.
One of studies to mitigate CO2 emissions to the atmosphere is to convert CO2 to
valuable products (i.e., methanol). To make methanol production via CO2 hydrogenation
a competitive process, the optimal operating conditions with minimum production
cost need to be considered. This paper studied an application of response surface
methodology (RSM) in optimization of methanol production via CO2 hydrogenation.
The objective of this optimization was to minimize the methanol production cost per
tons produced methanol. The sensitivity analysis was performed to determine the
Edited by: parameters that show significant impacts on the methanol production cost. Response
José Carlos Netto-Ferreira,
Universidade Federal Rural do
surface methodology coupled with non-linear programming solver were used as the
Rio de Janeiro, Brazil optimization tool. The results showed RSM was successfully applied to the methanol
Reviewed by: production via CO2 hydrogenation process. The obtained minimum methanol production
Eduardo René Perez Gonzalez,
cost was $565.54 per ton produced methanol with the optimal operating conditions as
São Paulo State University, Brazil
Wei Liu, follows. Inlet pressure to the first reactor: 57.8 bar, Inlet temperature to the first reactor:
Molecule Works Inc., United States 183.6◦ C, Inlet pressure to the second reactor: 102.6 bar, Outlet temperature of the liquid
Gabriel Segovia,
University of Guanajuato, Mexico
stream cooler after the second reactor: 63.5◦ C, Inlet temperature to the first distillation
*Correspondence:
column: 51.8◦ C.
Aroonsri Nuchitprasittichai Keywords: methanol production, CO2 hydrogenation, response surface methodology, simulation based
[email protected] optimization, sensitivity analysis

Specialty section:
This article was submitted to INTRODUCTION
Carbon Capture, Storage, and
Utilization, Most of energy in the world is currently from combustion of carbonaceous fuels, which are coal,
a section of the journal
oil, and natural gas. CO2 emission from this combustion is considered as the second contribution
Frontiers in Energy Research
to the greenhouse effect (9–26%), after the water vapor and clouds (36–72%). The recent attempts
Received: 07 April 2019 predict that CO2 will show stronger greenhouse effect when its amount in the atmosphere is double
Accepted: 31 July 2019
(Jaworowski et al., 1992). The net CO2 emissions could increase at around 5.4% over the next few
Published: 06 September 2019
decades (Radhi, 2009). Due to this concern, many applications and researches on CO2 conversion
Citation:
and utilization have been studied to control amount of CO2 releasing to the atmosphere. CO2 can
Borisut P and Nuchitprasittichai A
(2019) Methanol Production via CO2
either be used directly or as feedstock to produce useful chemicals and materials. For the direct use,
Hydrogenation: Sensitivity Analysis CO2 is utilized in many different applications such as food preservation, beverage carbonation, fire
and Simulation—Based Optimization. extinguisher, supercritical extraction, dry ice, etc (Song et al., 2002). For conversion of CO2 to other
Front. Energy Res. 7:81. products, urea synthesis is the largest production while the methanol synthesis is the second largest
doi: 10.3389/fenrg.2019.00081 production in this area (Naims, 2016).

Frontiers in Energy Research | www.frontiersin.org 1 September 2019 | Volume 7 | Article 81

Borisut and Nuchitprasittichai Optimization of Methanol Production

Methanol is commonly used as both solvent and reactant cooling water temperature. The results showed that the methanol
in chemical industry. Uses of methanol is found in many production rate can be increased by 7% with higher feed
household products, including paints, varnishes, cleaning pressure and lower feed temperature (Hoseiny et al., 2016).
products (Dasgupta and Klein, 2014). Moreover, methanol can For methanol synthesis via direct CO2 hydrogenation, Grazia
use as a motor fuel or gasoline blending component (Ingamells Leonzio developed mathematical model of the reactor used
and Lindquist, 1975). The methanol—fueled vehicles use a blend in methanol production. The impacts of reaction temperature,
of 85 percent methanol with 15 percent unleaded gasoline (M85) reaction pressure, H2 /CO2 ratio, and the recycle factor on
(Ingamells and Lindquist, 1975; EPA, 2002; Bukhtiyarova et al., methanol production rate and reactor volume were studied
2017). From laboratory and road tests indicate that adding of 10% (Leonzio, 2017).
methanol can raise octane 2–3 numbers. In the research of Gabele However, optimization of methanol production via CO2
(1990), the results showed that increase of methanol content does hydrogenation, which considers all possible operating
not affect the emission rate of exhaust gas. parameters, for the minimum production cost has not been
In the conventional methanol production, methanol studied. This paper studied an application of response surface
is produced from petroleum product (synthesis gas) via methodology (RSM) in optimization of methanol production
hydrogenation of CO and CO2 , and reversed water—gas via CO2 hydrogenation process. To be able to apply the RSM
shift reaction (María et al., 2013). The commercial methanol in optimization, the response surface has to be in the form
productions from synthesis gas use CuO/ZnO/Al2 O3 as catalyst of the second order model. The sensitivity analysis was used
(Sayah et al., 2010; Jadhav et al., 2014). The commercial to determine the significant operating parameters. The RSM
CuO/ZnO/Al2 O3 coupled with a zeolite membrane reactor can coupled with non-linear solver was employed to obtain the
provide higher CO2 conversion, methanol yield, and selectivity local optimal operating conditions and the minimum methanol
compared with a traditional reactor (Gallucci et al., 2004). production cost.
Copper based catalyst is mostly used for CO2 hydrogenation to This paper is organized as follows: section Process simulation
methanol due to its cheap and higher catalyst activity (Ali et al., and economic evaluation gives details of process simulation
2015). Saito and Murata studied of Al2 O3 supported Cu-based and economic evaluation. Section Methodology provides
catalyst with Co-precipitation technique. The catalyst shows methodology used in this work including details of sensitivity
high activity and the stability was improved by adding colloidal analysis and simulation—optimization framework. Section
silica (Saito and Murata, 2004). Results and discussion discusses results of process optimization.
Several developments of catalysts used in catalytic conversion Section Conclusion provides conclusion.
to improve the performance of catalyst were studied. Mg and Mn
was promoted on CuZnZr catalyst. Adding of MgO and MnO PROCESS SIMULATION AND ECONOMIC
lead to increase in catalytic activity of the catalyst (Sloczyński
EVALUATION
et al., 2003). Doping of Mn to Cu/Zn/Zr catalyst can increase
the methanol production rate. The zirconium indicates the In the study of methanol production process via CO2
advantageous influence on the catalyst activity (Lachowska and hydrogenation, the process simulation was combined with
Skrzypek, 2004). The Pd/ZnO catalysts over multi-walled carbon the economic analysis to evaluate the objective function (the
nanotubes was found to have the turnover frequency reached methanol production cost) corresponding to the decision
0.015 per second (Liang et al., 2009). variables. In what follows, details of process simulation and
Recently, the production of methanol from direct CO2 economic evaluation are explained.
hydrogenation is of interest (Samimi et al., 2017; Marlin et al.,
2018). The process has potential to mitigate the CO2 emission Process Simulation
to the atmosphere. The methanol production from direct CO2 The methanol production process via CO2 hydrogenation was
(using pure sources of CO2 and H2 ) has several advantages simulated using Aspen Hysys version 8.8 process simulator.
over the conventional process—it results in significantly less Peng-Robinson was used as the thermodynamic property
byproducts, and requires less energy in product purification package. The physical properties were predicted using the
(Marlin et al., 2018). However, the methanol production cost via thermodynamics based equation as shown in Equations (1) and
direct CO2 hydrogenation is 2–2.5 times higher than the cost (2). Where HID is the Ideal Gas Enthalpy and SID is the Ideal
of conventional process (Atsonics et al., 2015). The process of Gas Entropy.
methanol production via CO2 hydrogenation consumes more
utilities than the conventional methanol production (Machado ZV   
H − H ID 1

∂P
et al., 2014). =Z−1+ T − P dV (1)
RT RT ∂T V
There are some studies on simulation—based optimization ∞
of methanol synthesis to increase the methanol production V
S − SID 1 1
Z    
rate (Hoseiny et al., 2016; Leonzio, 2017). For methanol o P ∂P
= ln Z − ln o + − dV (2)
production from synthesis gas, the effects of changes in operating RT P R ∂T V V
∞
conditions on the production rate were studied to maximize
the production rate. The studied operating parameters were In the process simulation, two reactors were employed due to
feed flow rate, pressure and temperature of feed, and the low conversion of the CO2 hydrogenation reaction. Figure 1

Frontiers in Energy Research | www.frontiersin.org 2 September 2019 | Volume 7 | Article 81

Borisut and Nuchitprasittichai Optimization of Methanol Production

FIGURE 1 | Simulation of methanol production via CO2 hydrogenation process.

represents the process flow diagram of methanol production cost involved all major equipment, except pump, and piping.
via CO2 hydrogenation (Wiesberg et al., 2016). In this process, The capital cost was calculated using equations and data from
the feed of 1,000 kmoles per hour of carbon dioxide at 40◦ C the capital equipment-costing program (Turton et al., 2003). The
and 20 bar was mixed with the 3,000 kmoles per hour of data was adjusted for inflation from year 2001–2017 by using
hydrogen (at the same conditions). The mixture was then values of the Chemical Engineering Plant Cost Index, CEPCI.
compressed, heated and sent to the first equilibrium reactor. The CEPCI value in 2001 is 297, and the CEPCI value in 2017
The first reactor partially converted CO2 to methanol as liquid is 541.7 (Jenkins, 2018).
product, as shown in Equations (3)–(5) (Tidona et al., 2013). The equipment costs were estimated base on total module
The unreacted CO2 and H2 then entered the second equilibrium costs (CTM ), shown in Equation (6), where n represents the total
reactor to produce more methanol product. The pressure of number of pieces of equipment. CBM is the bare module cost,
gas phase leaving the second reactor was reduced to recover which can be estimated from Equation (7)
methanol as liquid phase. All liquid methanol products were sent n n
to the first distillation column, where the light components (CO,
X X
CTM = CTM,i = 1.18 CBM,i (6)
CO2 , and H2 ) leaved at the top of the column. The mixture i=1 i=1
of methanol and water leaved the column at the bottom and CBM = CP o (B1 + B2 FM FP ) (7)
entered the second distillation column, where the methanol
product with purity of 99.5%mole was obtained at the top of
the column. Where CP o is the purchased cost for base conditions, which can
be determined from Equation (8)

CO2 + H2 ↔ CO + H2 O 1H300 K = 41.2 kJ per mol CO2 (3)

CO + 2H2 ↔ CH3 OH 1H300 K = − 90.4 kJ per mol CO2 (4)
CO2 + 3H2 ↔ CH3 OH + 1H2 O H300 K = − 49.2 kJ per mol CO2 (5)

For equipment specifications, both reactors were simulated using FP is the pressure factor
equilibrium model. The sets of reaction used in each reactors FM is the material factor
are Equations (3)–(5). The efficiency of the pump was assumed
at 75% (adiabatic). The efficiencies of all compressors were log10 Cop = K1 + K2 log10 (A) + K3 [log10 (A)]2 (8)
assumed at 75% (adiabatic). The specifications of column T−100
were condenser temperature at 40.1◦ C and reflux ratio at 0.5. In Equation (8), parameters A is the capacity or size parameter
The specifications of column T−101 were 99.5%mole methanol for the equipment, K1 , K2 , K3 are the maximum and minimum
product purity, and reboiler temperature at 143.2◦ C. values used in the correlation.
The detailed process simulation of methanol production via The description of major equipment, and K1 , K2 , and K3 used
CO2 hydrogenation is provided in Appendix A. in this studied are shown in Table 1.
For cost modeling of reactors, the dimension of reactors were
Economic Evaluation fixed at height of 5.8674 m and diameter of 1.068 m. The values
In an economic analysis, capital, and operating costs were of K1 is 3.4974, K2 is 0.4485, K3 is 0.1074, B1 is 2.25, B2 is 1.82,
included in calculation of methanol production cost. The capital and FM is 3.1. The pressure factors (Fp) for the reactors were

Frontiers in Energy Research | www.frontiersin.org 3 September 2019 | Volume 7 | Article 81

Borisut and Nuchitprasittichai Optimization of Methanol Production

TABLE 1 | Equipment description and parameter values. TABLE 2 | Base case conditions used for sensitivity analysis.

Equipment Equipment K1 K2 K3 Capacity, Parameters Base value

type description units
Inlet pressure to the 1st reactor (bar) 60
Compressors Centrifugal, 2.2897 1.3604 −0.1027 Fluid power, Inlet temperature to the 1st reactor (◦ C) 210
axial, and kW
Inlet temperature to the 1st separator (◦ C) 60
reciprocating
Inlet pressure to the 2nd reactor (bar) 120
Heat Floating head 4.8306 −0.8509 0.3187 Area, m2
exchangers Spiral tube 3.9912 0.0668 0.2430 Area, m2 Outlet temperature of the liquid stream cooler after the 2nd 80
reactor (◦ C)
Heaters Steam boiler 6.9617 −1.4800 0.3161 Duty, kW
Outlet temperature of the vapor stream cooler after the 2nd 80
Packing Loose (for 2.4493 0.9744 0.0055 Volume, m3
reactor (◦ C)
towers)
Outlet pressure at the valve VLV-102 for recovering methanol from 8
Process Vertical 3.4974 0.4485 0.1074 Volume, m3
gas phase mixture (bar)
vessels
Inlet temperature to the 1st distillation column (◦ C) 50
Pumps Centrifugal 3.3892 0.0536 0.1538 Shaft power,
kW
Towers Tray and 3.4974 0.4485 0.1074 Volume, m3
packed the significant operating parameters of methanol production
Trays Sieve 2.9949 0.4465 0.3961 Area, m2 process for the minimum methanol production cost. Details of
Sensitivity analysis and Simulation—Optimization are described
in the following sections.
determine using Equation (9), where P is in barg.
Sensitivity Analysis
(P+1)D Sensitivity analysis is an efficient tool, which can be used to
2[850−0.6(P+1)] + 0.00315
FP,vessel = for tvessel > 0.0063m (9) determine the significant parameters. In this section, the impacts
0.0063 of eight parameters on the methanol production cost were
Assumption used in the economic analysis are listed below. determined using sensitivity analysis. An increase in pressure
and decrease in temperature leads to increase of the equilibrium
- The plant operates for 8,400 h per year. conversion of CO2 to methanol (Witoon et al., 2015). The eight
- The plant is expected to have a 10—year plant life with on studied parameters are (1 and 2) inlet pressure and temperature
salvage value. to the first reactor, ERV-100, for CO2 and H2 conversion
- The labor cost is $40,000 per operator per year. conditions, (3) inlet temperature to the first separator for
- The working capital is 15% of fixed capital investment. methanol product separation condition, (4) inlet pressure to the
- Total capital investment including fixed capital investment second reactor, ERV-101, for CO2 and H2 conversion conditions,
and working capital. (5 and 6) outlet temperature of both coolers, located after
- The maintenance and repairs are 5% of fixed the second reactor, for methanol product separation condition
capital investment. from gas and liquid mixture, (7) outlet pressure at the valve
- The operating supplies are 15% of maintenance and repairs. VLV-102 for recovering methanol from gas phase mixture, (8)
- The laboratory charge is 15% of labor. Inlet temperature to the first distillation column for separation
- The administrative expenses is 50% of labor cost. between liquid methanol and water products, and gaseous
- The maintenance and repairs, the laboratory charge and the unreacted reactant separation.
administrative expenses increase annually 3%. For all parameters, except inlet temperature to the first
- The local taxes and insurance are 4% of fixed reactor, we performed the sensitivity analysis by varying value
capital investment. of interested parameter within the range ± 25% of its base
- The plant overhead cost is 60% of labor. value, and fixing other parameters at their base values. For
- The cost of carbon dioxide is $12.10 per ton (Wiesberg et al., inlet temperature to the first reactor parameter, we varied the
2016). value of temperature within the range ± 10% of its base value.
- The cost of hydrogen is $1,250 per ton (Wiesberg et al., 2016) Table 2 shows the base value used in sensitivity analysis of
- The cost of process water at 25◦ C is $0.0259 per cubic meter. each parameter.
- The cost of saturated steam at 6.9 bar is $55 per ton.
- The cost of electricity is $0.127 per kW h. Sensitivity Analysis: Results
The results from sensitivity analysis show that there are
METHODOLOGY five parameters that show significant impacts on methanol
production cost. Figure 2 shows inversely proportional
The methodology in this work consists of two sections. The relationship between operating parameters and methanol
first section is “Sensitivity analysis” to determine the parameters production cost (the methanol production cost decreases with
that show significant impacts on the methanol production cost. an increase in the value of parameter). In the figure, the inlet
The second section is “Simulation—Optimization” to optimize pressure to the first reactor and the inlet temperature to the

Frontiers in Energy Research | www.frontiersin.org 4 September 2019 | Volume 7 | Article 81

Borisut and Nuchitprasittichai Optimization of Methanol Production

FIGURE 2 | Inversely proportional relationship between parameters and methanol production cost.

first distillation column show significant inversely proportional determined. In this work, the ranges of decision variables
relationship with the methanol production cost. were obtained from literature (Shen et al., 2000; Witoon
On the other hand, Figure 3 shows proportional relationship et al., 2015). Factorial design with two levels (2k Factorial
between operating parameters and methanol production cost. design, section 2k Factorial design) and a center point were
In the figure, the inlet temperature to the first reactor, the inlet used as a data set of the decision variables. Then, process
pressure to the second reactor, and the outlet temperature of the simulation, coupled with cost analysis, was run corresponding
liquid stream cooler after the second reactor show significant to operating conditions in the data set to obtained methanol
proportional relationship to the methanol production cost. production costs. The data of operating conditions and the
Five significant variables obtained from this sensitivity corresponding methanol production costs were combined and
analysis will be used as decision variables to the fit the first order model by regression analysis (using Design
optimization problem. Expert 11 software). If the data fit the first order model
(R—squared > 0.7), it means that the region of operating
Simulation—Optimization conditions and methanol production costs is in linear region.
In this section, RSM (Nuchitprasittichai and Cremaschi, 2011) The steepest descent (section Steepest descent) was performed
coupled with non-linear solver were employed to optimize to move operating conditions to the region of lower methanol
the methanol production via CO2 hydrogenation process. The production cost. On the other hand, if the data do not fit
objective is to minimize the methanol production cost per the first order model, it means that the data are in non-
tons produced methanol ($/tons produced methanol). Response linear region. The Box—Behnken Design [BBD, section Box—
surface methodology is a statistical tool used to represent Behnken design (BBD)] was performed to collect more data
the relationship between independent variables and dependent in non-linear region. All data were then combined to fit the
variable(s). In this work, RSM was employed to represent the second order (non-linear) model. The non-linear model was
relationship between significant operating parameters and the then used as the objective function in optimization problem.
methanol production cost per tons produced methanol. The Microsoft Excel (non-linear programing, NLP) solver was used to
significant operating parameters (decision variables) are the solve the non-linear optimization problem (section Optimization
significant parameters obtained from sensitivity analysis (in formulation) for the minimum methanol production cost. The
section Sensitivity analysis) which are x1 : inlet pressure to the 1st optimal solution obtained from this RSM is considered as a local
reactor, x2 : inlet temperature to the 1st reactor, x3 : inlet pressure optimal solution.
to the 2nd reactor, x4 : outlet temperature of the liquid stream
cooler after the 2nd reactor, x5 : inlet temperature to the 1st 2k Factorial Design
distillation column. The 2k factorial design is a design for k factors. Each factor is at
Figure 4 represents simulation—optimization algorithm. In only two levels, which are values at the lower and upper bounds of
the algorithm, the range of each decision variables was first the parameter range. The design is generated in coded variables.

Frontiers in Energy Research | www.frontiersin.org 5 September 2019 | Volume 7 | Article 81

Borisut and Nuchitprasittichai Optimization of Methanol Production

FIGURE 3 | Proportional relationship between parameters and methanol production cost.

FIGURE 4 | Optimization algorithm.

Frontiers in Energy Research | www.frontiersin.org 6 September 2019 | Volume 7 | Article 81

Borisut and Nuchitprasittichai Optimization of Methanol Production

FIGURE 5 | Experimental design: (A) 2k factorial design for three independent factors, (B) Box – Benhken design for three independent factors.

The coded variable −1 represents the lower bound value, and Experiments were conducted along the path of steepest descent
the coded variable 1 represents the upper bound value. The until no further decrease in response.
design is widely used in factor screening experiments. The design
observes at each corner of the cube (corner points). Figure 5A Box—Behnken Design (BBD)
demonstrates 2k factorial design for three independent factors. The Box—Behnken design is the three—level design for fitting
response surface (the second order model). The design consists
Steepest Descent of the midpoint of each edge of the space, and a center
The steepest descent was performed when the data obtained from point. Figure 5B demonstrates Box—Benhnken design for three
2k factorial design (including one center point) fits the first order independent factors.
model. The steepest descent moves the independent variables to
the direction of maximum decrease in the response (Douglas, Optimization Formulation
2013). To perform the steepest descent, the independent variables This section gives detailed optimization formulation. In the
will be coded to the (−1, 1) interval as shown in Equation (10). optimization formulation, a second order model was used to
Where xi is the coded variable, and ξi is the natural variable. represent the response surface. The form of a full second—order
model is shown in Equation (14). All five decision variables
ξi − Mid point studied in this work are continuous variables. The decision
xi = (10)
1 variables were coded in the ranges between −1 and 1. Therefore,
2 Range
all the constraints are in the boundary of coded variables.
The regression analysis was performed to fit the data set of 2k
k k
factorial design and a center point with the first order model.
βii xi2 +
X X XX
MIN : y = β0 + βi xi + βij xi xj (14)
Equation (11) represents example of the first order model with
i=1 i=1 i<j
five independent factors
Where y is the methanol production cost ($ per tons
y = b0 + b1 x1 + b2 x2 + b3 x3 + b4 x4 + b5 x5 (11)
produced methanol)
A step size (1xi ) in each independent variable was determined Decision variables:x1 : inlet pressure to the 1st reactor
by Equation (12). Where bi, highest is the highest value among all x2 : inlet temperature to the 1st reactor
coefficients bi , except bo . x3 : inlet pressure to the 2nd reactor
x4 : outlet temperature of the liquid stream cooler after the
bi 2nd reactor
1xi = (12) x5 : inlet temperature to the 1st distillation column
bi, highest
Subject to : −1 ≤ xi ≤ 1; i = 1, 2, 3, 4, 5
The path of steepest descent for xi was determined by using
Equation (13). Where j is point along the path. RESULTS AND DISCUSSION
1
 
xi condition along the path = xi, mid point + j1xi Range of xi ; We analyzed the impacts of eight operating parameters on
2 the methanol production cost. Then, optimized the value of
j = 1, 2, 3, . . . (13) significant parameters for the minimum methanol production

Frontiers in Energy Research | www.frontiersin.org 7 September 2019 | Volume 7 | Article 81

Borisut and Nuchitprasittichai Optimization of Methanol Production

TABLE 3 | Ranges of decision variables. 2nd reactor decreased from 74◦ C to around 65◦ C yielded
an increase in liquid methanol separated from the separator.
Decision variables Ranges
The inlet temperature to the first distillation column increased
Data set 1 Data set 2 from 48.0 to 52.0◦ C yielded higher efficiency to separate most
of remaining CO2 and lighter components from methanol
x1 : Inlet pressure to the 1st reactor (bar) 56.0–58.0 57.84–59.84
and water.
x2 : Inlet temperature to the 1st reactor (◦ C) 192.0–196.0 182.0–186.0 The steepest descent moves the operating conditions to the
x3 : Inlet pressure to the 2nd reactor (bar) 110.0–118.0 102.6–110.6 new conditions as follows: inlet pressure to the first reactor:
x4 : Outlet temperature of the liquid stream 74.5–78.5 63.5–67.5 58.84 bar, inlet temperature to the first reactor: 184.0◦ C, inlet
cooler after the 2nd reactor (◦ C)
pressure to the second reactor: 106.64 bar, outlet temperature
x5 : Inlet temperature to the 1st distillation 48.0–50.0 51.8–53.8
of the liquid stream cooler after the second reactor: 65.54◦ C,
column (◦ C)
and inlet temperature to the first distillation column 52.75◦ C.
We then used these new operating conditions as the middle
value of the range of decision variables. The ranges of the
new operating conditions are shown in Table 3, Data set
2 column.
The new data set (Data set 2) was then constructed with
2k factorial design. The corresponding methanol production
costs were collected, as shown in Figure 6. Most of the
data in data set 2 shows lower methanol production cost
than data set 1, except the data at low inlet pressure to
the first reactor. The results agree with the results from
sensitivity analysis that the inlet pressure to the first reactor
has inversely proportional relationship with the methanol
production cost.
By fitting data set 2 with the first—order regression
model, the data did not appropriately fit the model as R—
squared value is 0.6353 (<0.7). This means that there is
high possibility to find the non-linear region at this step.
We then used theses ranges of operating condition with
FIGURE 6 | The steepest descent results. BBD in collecting more data for constructing the non-
linear model.

The Optimal Conditions

cost. The results obtained from the sensitivity analysis, The reduced non-linear model (the model with only significant
the steepest descent, and the optimization are discussed in terms) was constructed with 73 sample points (32 sample
this section. points from 2k factorial design, and 41 sample points from
BBD) as shown in Equation (15). The model represents the
Simulation—Optimization: Steepest relationship between the operating conditions and the methanol
production cost.
Descent
In performing simulation—optimization algorithm, the first data y = 563.37 + 7.98 x1 − 8.22x2 − 11.02x1 x2 + 4.45x12 + 7.87x22 (15)
set (data set 1) from 2k factorial design with initial range of each
decision variable (shown in Table 3, Data set 1 column) was fitted Where y is the methanol production cost per tons produced
with the first order regression model. The obtained R—squared methanol ($ per ton produced methanol), x1 is inlet pressure
value is 0.9995. This means the data still fit the linear model. to the first reactor, and x2 is inlet temperature to the first
Therefore, we performed steepest descent to search for the region reactor. Please be noted that the obtained values of x1 and
of lower methanol production cost. Figure 6 shows steepest x2 in Equation (15) are in code variables (−1 to +1).
descent in moving the operating conditions to lower methanol The obtained results have to be converged to the actual
production cost region. operating condition values by using the ranges in Table 3,
From the steepest descent results, for the first reactor, data set 2 column and Equation (10), where −1 represents
an increase in inlet pressure from 56.0 to 58.0 bar, and a the lowest value and +1 represents the highest values in
decrease in inlet temperature from 192.0 to 183.0◦ C yielded the range.
an increase in methanol product at equilibrium condition Table 4 shows the local optimal operating conditions (xi )
(%conversion increased from 40 to 45%). For the second of methanol production via CO2 hydrogenation with the
reactor, a decrease in inlet pressure from 118.0 to 102.0 minimum production cost per tons produced methanol (y).
bar yielded an increase in methanol product in the reactor. Since three parameters, which are inlet pressure to the 2nd
The outlet temperature of the liquid stream cooler after the reactor (x3 ), outlet temperature of the liquid stream cooler

Frontiers in Energy Research | www.frontiersin.org 8 September 2019 | Volume 7 | Article 81

Borisut and Nuchitprasittichai Optimization of Methanol Production

TABLE 4 | The optimal operating conditions obtained from the reduced 2nd CONCLUSION
order model.

Parameters Optimal values

In the optimization of methanol production via CO2
hydrogenation, the sensitivity analysis coupled with RSM
x1 : Inlet pressure to the 1st reactor (bar) 57.8 was successfully represent the relationship between methanol
x2 : Inlet temperature to the 1st reactor (◦ C) 183.6 production operating conditions and the methanol production
x3 : Inlet pressure to the 2nd reactor (bar) 102.6 cost per tons produced methanol ($ per ton produced methanol).
x4 : Outlet temperature of the liquid stream cooler after 63.5 The non-linear solver was applied to optimize the non-linear
the 2nd reactor (◦ C) relationship model for the minimum methanol production cost.
x5 : Inlet temperature to the 1st distillation column (◦ C) 51.8 Two reactors were employed in the process. In the optimal
Predicted methanol production cost ($/ton) 559.59 region of study, the inlet pressure and temperature to the first
Actual methanol production cost ($/ton) 565.54 reactor show significant impacts on the methanol production
cost. The optimal operating conditions of inlet pressure to the
first reactor is 57.8 bar, inlet temperature to the first reactor
after the second reactor (x4 ), and inlet temperature to the is 183.6◦ C, and the other three insignificant parameters were
first distillation column (x5 ), do not show significant impacts set at their lower bounds. The obtained minimum methanol
on the production cost in the non-linear model (Equation production cost is $565.54 per ton produced methanol.
15), the values of these three parameters were set at their
lower bounds. DATA AVAILABILITY
The prediction accuracy of the model was determined by the
percent error. The percent error of the model was estimated All datasets generated and analyzed for this study are included in
by comparing between the predicted methanol production the manuscript/Supplementary Files.
cost with the actual production cost. We obtained the actual
production cost by running the process simulation with the AUTHOR CONTRIBUTIONS
corresponding optimal operating conditions. The percent error
of this non-linear model is 1.05. For regression analysis of this PB simulated the process simulation, collected data, analyzed
model, the R—squared value is 0.8705, and the adjusted R— results, and wrote the first draft of the manuscript. AN designed
squared value is 0.8610. Therefore, with the small number of the methodology of the work, and revised the manuscript for the
percent error, it can be concluded that this methanol production final version.
via CO2 hydrogenation process can be represented by the
second order model. The process is successfully optimized FUNDING
using RSM.
The optimization formulation was solved in the Intel Suranaree University of Technology funding on Aspen
Pentium 4 processor, core i3. The local optimal solutions Hysys simulator.
were obtained in 2 s. The obtained minimum production
cost ($ per tons produced methanol) is an offset between ACKNOWLEDGMENTS
the methanol production cost and the amount of produced
methanol. The methanol production cost includes energy The financial support from Suranaree University of Technology
consumption, utilities, and capital cost costs. For the optimal is greatly acknowledged.
solution, the amount of produced methanol increased so that
the minimum production cost per tons of produced methanol SUPPLEMENTARY MATERIAL
was obtained. The optimal methanol production rate is 964
kmoles per hours, which produced from 1,000 kmoles per The Supplementary Material for this article can be found
hour of feed carbon dioxide and 3,000 kmoles per hour of online at: https://www.frontiersin.org/articles/10.3389/fenrg.
feed hydrogen. 2019.00081/full#supplementary-material

REFERENCES Bukhtiyarova, M., Lunkenbein, T., Kähler, K., and Schlögl, R. (2017).
Methanol synthesis from industrial CO2 sources: a contribution to
Ali, K. A., Abdullah, A. Z., and Mohamed, A. R. (2015). Recent development in chemical energy conversion. Catal. Lett. 147:416. doi: 10.1007/s10562-01
catalytic technologies for methanol synthesis from renewable sources: a critical 6-1960-x
review. Renew. Sustain. Energ. Rev. 44, 508–518. doi: 10.1016/j.rser.2015. Dasgupta, A., and Klein K. (2014). “Chapter 5: Oxidative stress induced by
01.010 household chemicals,” in Antioxidants in Food, Vitamins and Supplements, eds
Atsonics, K., Panopoulos, K. D., and Kakaras, E. (2015). Thermocatalytic CO2 A. Dasgupta and K. Klein (San Diego, CA: Elsevier), 77–95.
hydrogenation for methanol and ethanol production: process improvements. Douglas, C. M. (2013). Design and Analysis of Experiments, 8th Edn. New York,
Int. J. Hydrog. Energy 41, 792–806. doi: 10.1016/j.ijhydene.2015.12.001 NY: John Wiley & Sons, Inc.

Frontiers in Energy Research | www.frontiersin.org 9 September 2019 | Volume 7 | Article 81

Borisut and Nuchitprasittichai Optimization of Methanol Production

EPA (2002). US Environmental Protection Agency. Clean Alternative Fuels: Radhi, H. (2009). Evaluating the potential impact of global warming on the
Methanol. Technical Report EPA 420-F-00-040. Washington DC, USA: EPA; UAE residentialbuildings – a contribution to reduce the CO2 emissions. Build.
2002. Available online at: www.epa.gov. Environ. 44, 2451–2462. doi: 10.1016/j.buildenv.2009.04.006
Gabele, P. A. (1990). Characterization of emissions from Saito, M., and Murata, K. (2004). Development of high performance Cu/ZnO-
a variable gasoline/methanol fueled car. J. Air Waste based catalysts for methanol synthesis and the water-gas shift reaction. Catal.
Manage. Assoc. 40, 296–304. doi: 10.1080/10473289.1990.104 Surv. Asia 8, 285–294. doi: 10.1007/s10563-004-9119-y
66685 Samimi, F., Rahimpour, M. R., and Shariati, A. (2017). Development of an
Gallucci, F., Paturzo, L., and Basile, A. (2004). An experimental study of CO2 efficient methanol production process for direct CO2 hydrogenation over a
hydrogenation into methanol involving a zeolite membrane reactor. Chem. Cu/ZnO/Al2 O3 catalyst. Catalysts. 7:332. doi: 10.3390/catal7110332
Eng. Proc. 43, 1029–36. doi: 10.1016/j.cep.2003.10.005 Sayah, A. K., Hosseinabadi, S.h, and Farazar, M. (2010). CO2 abatement by
Hoseiny, S., Zare, Z., Mirvakili, A., Setoodeh, P., and Rahimpour, M. R. (2016). methanol production from flue-gas in methanol plant. World Acad. Sci. Eng.
Simulation–based optimization of operating parameters for methanol synthesis Technol. Int. J. Chem. Molecul. Eng. 4:9. doi: 10.5281/zenodo.1072503
process: application of response surface methodology for statistical analysis. J. Shen, W., Jun, K., and Choi, H. (2000). Thermodynamic investigation of methanol
Nat. Gas. Sci. Eng. 34. 439–48. doi: 10.1016/j.jngse.2016.06.075 and dimethyl ether synthesis from CO2 hydrogenation. Kor. J. Chem. Eng. 17,
Ingamells, J. C., and Lindquist, R. H. (1975). Methanol as a motor fuel or a gasoline 210–216. doi: 10.1007/BF02707145
blending component. JSTOR 84, 582–568. doi: 10.4271/750123 Sloczyński, J., Grabowski, R., Kozlowska, A., Olszewski, P., Lachowska, M.,
Jadhav, S. G., Vaidya, P. D., Bhanage, B. M., and Joshi, J. B. (2014). Catalyticcarbon Skrzypek, J., et al. (2003). Effect of Mg and Mn oxide additions on structural
dioxide hydrogenation to methanol: a review of recent studies. Chem Eng Res and adsorptive properties of Cu/ZnO/ZrO2 catalysts for the methanol synthesis
Design. 92, 2557–2567. doi: 10.1016/j.cherd.2014.03.005 from CO2 . Appl. Catal. 249, 129–138. doi: 10.1016/S0926-860X(03)00191-1
Jaworowski, Z., Segalstad, T. V., and Hisdal, V. (1992). Atmospheric CO2 and Song, C., Gaffney, A. F., and Fujimoto, K. (2002). CO2 Conversion and Utilization:
Global Warming: A Critical Review, 2nd Edn. Oslo: Norsk Polarinstitutt. an Overview in CO2 Conversion and Utilization. Washington, DC: American
Jenkins, S. (2018). CEPCI Updates: January 2018 (PRELIM.) and December 2017. Chemical Society, 2–30.
Available online at: http://www.chemengonline.com/cepci-updates-january- Tidona, B., Koppold, C., Bansode, A., Urakawa, A., and Rudolf, V. R. P. (2013).
2018-prelim-and-december-2017-final/?printmode=2011. CO2 hydrogenation to methanol at pressures up to 950 bar. J. Supercrit. Fluid.
Lachowska, M., and Skrzypek, J. (2004). Methanol synthesis from carbon dioxide 78, 70–77. doi: 10.1016/j.supflu.2013.03.027
and hydrogen over Mn-promoted copper/zinc/zirconia catalysts. React. Kinet. Turton, R., Bailie, R. C., Whiting, W. B., and Shaeiwitz, J. A. (2003). Analysis,
Catal. Lett. 83, 269–273. doi: 10.1023/B:REAC.0000046086.93121.36 Synthesis and Design of Chemical Processes, 2nd Edn. Upper Saddle River, NJ:
Leonzio, G. (2017). Optimization through response surface methodology of a Prentice Hall.
reactor producing methanol by the hydrogenation of carbon dioxide. Processes Wiesberg, I. L., Medeiros, J. L., Alves, R. M. B., Coutinho, P. L. A., and Araújo, O.
5:62. doi: 10.3390/pr5040062 Q. F. (2016). Carbon dioxide management by chemical conversion to methanol:
Liang, X.-L., Dong, X., Lin, G.-D., and Zhang, H.-B. (2009). Carbon nanotube- Hydrogenation and Bi – reforming. Energy Convers. Manage. 125, 320–335.
supported Pd–ZnO catalyst for hydrogenation of CO2 to methanol. Appl. Catal. doi: 10.1016/j.enconman.2016.04.041
B. 88, 315–322. doi: 10.1016/j.apcatb.2008.11.018 Witoon, T., Bumrungsalee, S., Chareonpanich, M., and Limtrakul, J.
Machado, C. F. R., Medeiros JLd, O. F. Q., and Araujo, O. F. Q. (2014). “A (2015). Effect of hierarchical meso macroporous alumina-supported
comparative analysis of methanol production routes: synthesis gas versus copper catalyst for methanol synthesis from CO2 hydrogenation.
CO2 hydrogenation,” in Proceedings of the 2014 International Conference on Energy Conver. Manage. 103, 886–894. doi: 10.1016/j.enconman.2015.
Industrial Engineering and Operations Management, Bali, Indonesia. 07.033
María, R. D., Díaz, I., Rodríguez, M., and Sáiz, A. (2013). Industrial methanol
from syngas: kinetic study and process simulation. Int. J. Chem. React. Eng. 11, Conflict of Interest Statement: The authors declare that the research was
469–477. doi: 10.1515/ijcre-2013-0061 conducted in the absence of any commercial or financial relationships that could
Marlin, D. S., Sarron, E., and Sigurbjörnsson, Ó. (2018). Process be construed as a potential conflict of interest.
advantages of direct CO2 to methanol synthesis. Front. Chem. 6:446.
doi: 10.3389/fchem.2018.00446 Copyright © 2019 Borisut and Nuchitprasittichai. This is an open-access article
Naims, H. (2016). Economics of carbon dioxide capture and utilization – a distributed under the terms of the Creative Commons Attribution License (CC BY).
supply and demand perspective. Environ. Sci. Pollut. Res. Int. 23, 22226–22241. The use, distribution or reproduction in other forums is permitted, provided the
doi: 10.1007/s11356-016-6810-2 original author(s) and the copyright owner(s) are credited and that the original
Nuchitprasittichai, A., and Cremaschi, S. (2011). Optimization of CO2 capture publication in this journal is cited, in accordance with accepted academic practice.
process with aqueous amines using response surface methodology. Comput. No use, distribution or reproduction is permitted which does not comply with these
Chem. Eng. 35, 1521–1531. doi: 10.1016/j.compchemeng.2011.03.016 terms.

Frontiers in Energy Research | www.frontiersin.org 10 September 2019 | Volume 7 | Article 81