Maximization of dimethyl ether production from synthesis gas by obtaining optimum temperature profile and water removal

Maximization of dimethyl ether production from synthesis gas by obtaining optimum temperature profile and water removal

Fuel 190 (2017) 386–395 Contents lists available at ScienceDirect Fuel journal homepage: Full Length Article Maximiza...

2MB Sizes 0 Downloads 19 Views

Fuel 190 (2017) 386–395

Contents lists available at ScienceDirect

Fuel journal homepage:

Full Length Article

Maximization of dimethyl ether production from synthesis gas by obtaining optimum temperature profile and water removal Davood Iranshahi ⇑, Reza Saeedi, Kolsoom Azizi Department of Chemical Engineering, Amirkabir University of Technology (Tehran Polytechnic), No. 424, Hafez Avenue, Tehran 15914, Iran

h i g h l i g h t s  The reactor is optimized by applying differential evolution (DE) algorithm as an effective and robust optimization method.  In case 1, inlet temperature of each segment has been optimized via differential evolution (DE).  Same approach has been applied in the second case in order to achieve optimum water permeation rate.  In the third case, the optimum profiles of temperature and water removal have been obtained.  1.5%, 55% and 70% enhancement in the dimethyl ether production have been obtained in case 1, case 2, case 3 respectively.

a r t i c l e

i n f o

Article history: Received 8 February 2016 Received in revised form 7 April 2016 Accepted 27 October 2016 Available online 10 November 2016 Keywords: Fixed bed reactor Direct synthesis dimethyl ether (DME) Synthesis gas Optimization Water removal Differential evolution (DE)

a b s t r a c t In this study, a novel method optimizing direct dimethyl ether generation from syngas is proposed. A bifunctional catalyst is used containing commercial catalyst for direct DME synthesis (CuO/ZnO/Al2O3). The length of reactor has been discretized into twenty segments and optimum temperature profile in the inlet of each segment and the amount of water removal from the main stream are calculated. This optimization is done via differential evolution (DE) method, in which the inlet temperature and water removal flux in each segment are considered as the decision variables and mole fraction of dimethyl ether in the reactor outlet is the objective function. Three different cases are considered. In the first case, inlet temperature of each segment has been optimized. Then, flux of water removal is considered. And, finally, a combination of two previous cases is investigated. Obtained results of this novel theoretical study showed that water removal results in more methanol production; and Consequently, dimethyl ether production increased. By applying the proposed optimization methods, 1.5%, 55% and 70% enhancement in the dimethyl ether production have been obtained in comparison with the conventional reactor of direct dimethyl ether production. Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction 1.1. Dimethyl ether (DME) Because of excessive demand for energy, deprivation fuel resources and environmental concerns, the global community is seeking new alternative fuels. DME could be considered to be a perspective of future fuel due to its unique characteristics. Its physical properties are similar to those of liquefied petroleum gas (LPG) [1–5]. It burns with the discharge of no sulfur oxides (SOx), less nitrogen oxides (NOx) and less carbon monoxide (CO). Also, it could

⇑ Corresponding author. E-mail address: [email protected] (D. Iranshahi). 0016-2361/Ó 2016 Elsevier Ltd. All rights reserved.

be used as hydrogen source in fuel cells. DME could be used as an intermediate to produce chemicals such as methyl acetate [6]. Basically, there are two routes to synthesize DME: direct synthesis and indirect synthesis. In the indirect method, DME is produced via a two- step process. In the first step, syngas is converted to methanol in the presence of catalyst. Then, methanol is dehydrated over the alumina or zeolite based acidic materials [7]. On the other hand, in the direct method, known as the single step process, the feedstock (syngas) converts directly to DME. In this method, a bifunctional catalyst is used which consist of metallic side for methanol synthesis and acidic side for methanol dehydration. It should be mentioned that all of the reactions take place in one reactor simultaneously and the produced methanol is not separated during the process [8]. Therefore, the direct method is more economical in comparison with the other one.

D. Iranshahi et al. / Fuel 190 (2017) 386–395


Nomenclature Ac asp Ci Cp De dp fi F hf Kfi Keff Kpi Mi P Q r1 r2 r3

cross section area (m2) specific surface area of catalyst pellet (m2 m3) concentration of each component (mol m3) specific heat of the gas at constant pressure (J mol1) effective diffusivity (m2 s1) particle diameter (m) partial fugacity of component i (bar) molar flow rate (mol s1) heat transfer coefficient of fluid (W m2 K1) rate constant for the rate of reaction conductivity of fluid phase (W m1 K1) equilibrium constant based on partial pressure for component i in DME synthesis reaction molecular weight of component i (g mol1) total pressure (bar) volumetric flow rate (m3 s1) rate of reaction for hydrogenation of CO (mol kg cat1 s1) rate of reaction for hydrogenation of CO2 (mol kg cat1 s1) rate of reaction for DME (mol kg cat1 s1)

1.1.1. Experiment studies In recent years, many studies have been done to investigate various reactor configurations for one-step DME production. Sun et al. studied on a series of bifunctional catalysts CuO/ZnO/ZrO2/HZSM-5 with different ZrO2. These catalysts were characterized by surface area, XRD and XPS analysis [9]. Cai et al. studied the effect of Tin addition to catalyst of direct synthesis of DME and concluded that there is an optimized concentration of Tin which the yield of DME generation is the highest [2]. García-Trenco and Martínez considered the effect of different metallic oxides ratios on the CO conversion and DME generation [10]. Raoof et al. conducted an experimental study on a fixed bed reactor of DME production [11]. They investigated the effect of methanol purity on the DME production yield in the indirect synthesis method. 1.1.2. Modeling studies Hu et al. presented a successfully mathematical model in relation to DME reactor in direct synthesis method [12]. Vakili et al. designed an optimal industrial scale dual-type reactor for direct dimethyl ether production. The results of their study showed that the proposed configuration led to an increase in DME production capacity, which was estimated to be about 60 ton/day in comparison to conventional industrial DME reactor [5]. In other study, Vakili et al. proposed a thermally coupled heat exchanger reactor for direct dimethyl ether (DME) synthesis. In this novel configuration, DME production increases about 600 ton/year [13]. Nasehi et al. simulated an adiabatic fixed bed DME reactor at steady state condition [14]. They used indirect synthesis method in their simulation and showed that obtained results for one dimensional and two dimensional reactors have a little difference. The effect of heat transfer boundary condition on the process was studied by Farsi et al. [15]. They designed and simulated an isothermal reactor and showed that isothermal reactor has a positive effect on the yield of reaction in comparison with adiabatic reactor. Coupling of endothermic and exothermic reaction has been done by Khademi et al. [16]. In their proposed configuration, the exothermic reaction of DME generation was coupled with the endothermic reaction of cyclohexane dehydrogenation. Omata et al. [17] studied DME production from syngas in a temperature gradient reactor to overcome both the equilibrium limitations at high temperatures

R T Tref Z

gas constant (kJ kmol1 K1) temperature (K) reference temperature (K) length of reactor coordinate

Acronyms CR conventional packed bed reactor NP Number of Population OF objective function SC Schmidt number Re Reynolds number Greek letters M viscosity of fluid phase (kg m1 s1) P density of gas phase (kg m3) qb density of catalytic bed (kg m3) e porosity DHi heat of reaction ith (kJ kg1) ti stoichiometric coefficient of component ith in reaction /s sphericity

and low catalyst activity related to low temperatures. Then, they optimized the operating conditions of the reactor for higher CO conversion by combining genetic algorithms and artificial neural networks. Iranshahi et al. [18] suggested a reactor configuration for naphtha reforming process, in which hydrogen and aromatics productions were increased by obtaining optimum temperature profile and hydrogen removal rate. The main advantage of their research compared with other studies about naphtha reforming is the application of the optimum temperature profile and hydrogen removal along the reactors. The length of naphtha reforming reactor has been discretized into twenty segments. In order to find the optimum values of the temperature and hydrogen removal rate, differential evolution (DE) method was used, which is a simple heuristic approach. The inlet optimum values for each segment were found and they were joined on the figures along the reactor. Arabpour et al. [19] also evaluated maximum gasoline production of FischereTropsch synthesis reactions in GTL technology by a discretized approach. In this way, the conventional synthesis reactor has been discretized into some elements. For each element, the inlet temperature, the injected hydrogen and the removed water are considered as decision variables. Then, the optimum amount of these decision variables is defined by using the differential evolution (DE) algorithm as an optimization method. 1.1.3. Objectives In the present study, the effect of temperature and concentration on the process yield is studied to obtain the maximum production of dimethyl ether. Regarding this, the length of reactor has been discretized into twenty segments. Three cases are considered to obtain the inlet optimum values for each section of reactor. In the first case, temperature has been optimized. Then if second case, water removal has been optimized. And finally, in the third case, combination of optimized temperature and water removal has been investigated. Obtained results compared with the results of Hu et al. [12] to validate the accuracy of the method. 2. Reaction scheme and kinetics DME synthesis from syngas is an exothermic reaction. In this study, a bi-functional catalyst is used containing commercial


D. Iranshahi et al. / Fuel 190 (2017) 386–395

catalyst for methanol synthesis (CuO/ZnO/Al2O3) and methanol dehydration (c  Al2 O3 Þ with equal mass ratio [20]. DME is synthesized from syngas gas according to the following reactions [21]:

wall. The specification of catalysts and operational conditions are showed in Table 2.

1. Methanol synthesis from CO:

In the new proposed model, optimum temperature profile and optimum separated water are determined and the effects of these optimized parameters on the mole fraction of DME in the output of reactor are considered. Three cases are investigated during the optimization process. The reactor length is divided into twenty segments and material and energy balances are evaluated to obtain optimum temperature profile and separated water. In the first case, the inlet temperature of each segment has been optimized to achieve the final temperature profile. In the second case, same procedure has been applied to optimize the amount of removed water and the maximum amount of DME in the output of reactor has been calculated. In the third case, combination of previous cases has been applied to achieve the maximum mole fraction of DME. In Fig. 2 schematic diagram of the optimized reactor is shown. Water is one of the products of direct DME synthesis, and therefore, its separation from the reaction zone would shift the equilibrium towards the production side. Indeed, according to Le Chatelier’s principle, the reaction shifts toward more water and induces a dramatic increase in DME production. Water could be removed by membrane or water adsorbent. Zeolite 4A is a solid particle with the composition of Na12(Si12Al12O48)27H2O and high water adsorption affinity that makes it desirable for water removal or separation [23].

CO þ 2H2 () CH3 OH


2. Methanol synthesis from CO2:

CO2 þ 3H2 () CH3 OH þ H2 O


3. Methanol dehydration:

2CH3 OH () CH3 OCH3 þ H2 O


The rate expressions have been selected from [22] and are represented here as follow: 2

k1 f CO f H2 ð1  £1 Þ r CO ¼  3 1 þ K CO f CO þ K CO2 f CO2 þ K H2 f H2



k2 f CO2 f H2 ð1  £2 Þ r CO2 ¼  4 1 þ K CO f CO þ K CO2 f CO2 þ K H2 f H2 r DME ¼ 

£1 ¼

£2 ¼

£3 ¼


k3 f CH3 OH ð1  £3 Þ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2 1 þ K CH3 OH f CH3 OH


3.2. Novel model

3.3. Mathematical model

f CH3 OH



K f 1 f CO f H2 f CH3 OH f H2 O



K f 2 f CO2 f H2 f DME f H2 O K f 3 f CH3 OH


where f i is fugacity of component i and kf i is equilibrium constant of reaction i. The reaction rate constants are presented in Table 1.

In order to simulate the discussed reactor in this study, a one dimensional model has been used according to the following assumptions:      

Steady state condition has been considered in this study. Axial dispersion is assumed negligible. In each pipe, plug flow pattern is considered. Heat loss is neglected. Radial diffusion is neglected. The model has been considered to be heterogeneous.

The application of these assumptions can simplify the calculation procedure and may not affect the accuracy of the model. The following mass and energy balance equations can be used for the proposed model [24]. Fluid phase:

3. Model description 3.1. Conventional reactor for DME synthesis Direct synthesis of DME in a pipe-shell fixed bed reactor has been proposed by Hu et al. [12]. In this reactor, catalysts are placed in the pipes and boiling water, which aims to absorb the generated heat of the reaction side, is flowed through the shell side. Fig. 1 shows schematic diagram of a conventional reactor of DME synthesis. Generated heat is transferred to the boiling water through pipe


  1 @ @C j 1 @  Ac ðAcuC j Þ þ kjg asp ðC js  C j Þ  J H2 O ¼ 0; Ac @z Ac @z @z

j ¼ 1; 2; . . . ; n keff


  1 @ @T 1 @ Ac  ðqAcuC p ðT  T ref ÞÞ þ asp hf ðT s  TÞ ¼ 0 Ac @z @z Ac @z ð8Þ

Table 1 Reaction rates constants [27]. K = A exp (B/RT) K1 K2 K3 KCO K H2 K CO2 K CH3 OH

Solid phase: A

B 3

1.828  10 0.4195  102 1.939  102 8.252  103 2.1  104 0.1035 1.726  104

43,723 30,253 24,984 30,275 31,846 11,139 60,126

kjg asp ðC j  C js Þ þ qB

m X

mij ri ¼ 0


mij DHi ri ¼ 0



asp hf ðT  T s Þ þ qB

m X i¼1

where j ¼ 1; 2; . . . ; n.


D. Iranshahi et al. / Fuel 190 (2017) 386–395

Fig. 1. Configuration of pipes and shell reactor of direct DME synthesis.

Table 2 Operational conditions and catalyst features for one step conventional, DME reactor. [13]. Parameter



Feed composition (mole fraction) CO CO2 DME CH3OH H2O H2 N2 CH4 Inlet temperature Inlet pressure Number of pipes Diameter of pipes Volumetric flow rate of raw gas Length of reactor Temperature of boiling water Thermal conductivity of wall

0.1716 0.0409 0.0018 0.003 0.0002 0..4325 0.316 0.044 493 50 4177 /38  2 2.04  105 5.8 513 48

– – – – – – – – K Bar – mm N m3 h1 m K W m1 K1

Typical properties of catalyst Particle diameter Density of catalyst bed Porosity

/5  5 1200 0.455

mm kg m3 –

Synthesis gas

Synthesis gas

Removed water


The following boundary conditions have been applied for solving these equations. Boundary and initial conditions

r ¼ R0 ;

C j ¼ C j0 ;

T ¼ T0


r ¼ Ri ;

@C j ¼ 0; @r

@T ¼0 @r



Synthesis gas

Physical properties and correlations which should be used to solve mentioned equations are listed in Table 3. Pressure drop in reactor is also calculated by Ergun equation [25].

dP 150l ð1  eÞ2 Q 1:75q ð1  eÞ Q 2 ¼ 2 2 þ dz e3 Ac /s dp e3 Ac2 /s dp


3.4. Optimization As it was mentioned, in this study, the length of reactor is discretized into twenty segments. Differential evolution (DE) method has been used to obtain the optimized temperature profile and

Fig. 2. Schematic diagram of the optimized reactor.

water removal. Differential Evolution (DE) is a parallel direct search method which utilizes NP (Number of Population) D-dimensional parameter vectors. NP does not change during the minimization process. The initial vector population is chosen randomly and should cover the entire parameter space. DE method contains four stages: (1) Choosing an initial vector population


D. Iranshahi et al. / Fuel 190 (2017) 386–395

The results are compared with the performance of conventional DME reactor. Therefore, the objective function is defined as follows:

Table 3 Physical properties and correlations. Parameter


Ref. 3

Component capacity

cp ¼ a þ bT þ cT þ dT

Gas density

P q ¼ ZRT ð15Þ   l ¼ exp C 1 þ C 2 =T þ C 3 ln T þ C 4 T c5 ð16Þ


Average molecular weight Mass transfer coefficient between gas and solid phases Reynolds number



kgi ¼ 1:17Re0:42 sci0:67  ug  103

Sci ¼

Diffusivity of component i in the gas mixture The binary diffusivities

ð18Þ l l ð19Þ qDim 104

Dim ¼

1yi P

107 T 3=2

Dij ¼ P

[29] [30]


yi Dij



1 1 M j þM i


v 2ci þv 2cj


because of sintering of catalyst, the temperature of system should be controlled. In the temperature profile in the DME production process, temperature should not exceed 533 K or 260 °C [27]. Also, pressure drop in the reactor should not exceed the conventional condition. The amount of water at the end of reactor should not reach to zero. Finally, the following restrictions have been applied to the optimization method.



Out put Mole of DME Total molar flow rate


2Rp ug

Re ¼

Schmit number

OF ¼ yDME ¼

[30] ð21Þ

493 6 T0 6 533 K


Pex P 49:6 bar


FH2O P 0


To eliminate the inappropriate answers, penalty function has been applied in the optimization procedure. So, by using mentioned function and its parameters, the defined restrictions are implemented in the procedure and desired results will be obtained. Here, the objective function is as follows:

randomly. (2) Generating a mutant vector for each point (mutation/perturbation). (3) Using the crossover process for a trial vector generation for each point. (4) Using objective function evaluations for comparing each point with its associated trial vector and selecting the better vector for next generation step. By following this procedure, new populations are generated which approach the global optimum through iteration to iteration. The basic strategy of DE algorithm is illustrated in Fig. 3 [26].

Result ¼ OF þ 108

4 X G2i




A Generation of Parameter Vectors M1








M63 M64 M65 M66 M67 M68 M69 M70

















No. of objectives


















































2-Two random selected vector


Pa 1-Target vector selection





3-Generate the scaled difference vector


4- Add a third randomly selected vector se


Crossover Mt Rt

Trial Vector


St Ot




6- SELECTION, the vector with the lowest cost survives and passes to the next generation M'1 R'1 No. of objectives


S'1 O'1 55

New Fitter Generation Fig. 3. The basic strategy of DE algorithm.

D. Iranshahi et al. / Fuel 190 (2017) 386–395

G1 ¼ maxð0; ðT  533ÞÞ


G2 ¼ maxð0; ð493  TÞÞ


G3 ¼ maxð0; ð49:6  TÞÞ


G4 ¼ maxð0; F H2O Þ




In order to reach this aim, for case (I) twenty, case (II) twenty and case (III) forty decision variables, which affect the production rate, have been chosen. These decision variables are consist of temperature and water penetration of each segment ðJ H2 O Þ. By applying of differential evolution method, this objective function will be minimized and the term defined for mole fraction of DME will be maximized. 4. Results and discussion 4.1. Model validation The results of this study have been compared with that of the Hu et al., which showed good agreement. This comparison is tabulated in Table 4.


4.2. Discussion As it was mentioned, in this study, the reactor has been divided into 20 equal segments and three cases have been studied for each segment. Temperature profile has been optimized in the first case. Then, the maximum amount of water removal has been investigated. In the last case, the combination of these two cases has been considered. Fig. 4(a) shows temperature profile of conventional reactor and first case. Since, in conventional reactors cooling water is used and the reactions are exothermic, the reaction side temperature increases and then decreases. This is because in the beginning the reaction (A) which has the higher reaction heat is dominant and then because of methanol, reaction (C) which has lower heat of reaction and leads to production of DME will be dominant. Then by production of methanol, reaction C, which has the lowest heat of reaction, will be dominant. Reaction A has the highest activation energy. Hence, an increase in the temperature has the most significant effect on the rate of reaction A. On the other hand, rising of temperature increases rate of all reactions. Since reactions are exothermic, therefore the temperature enhancement should be sufficient to prevent the production of unwanted reactants. Temperature profiles of second and third case are presented in Fig. 4(b). In the second case, removal of water from the reactor, heat capacity and total flow rate reduction cause rising of temperature in the output of reactor which is not a significant problem. In the third case, since in each segment of reactor, water is removed from main stream, in addition to reaction (A), reactions (B) and (C) are also performed. They have more positive heat than reaction (A), which cause an increase in temperature in the beginning of reactor. Suggested fluxes for water removal along the reactor are shown for second and third cases in Fig. 5. It should be noticed that water

Fig. 4. Temperature profile of (a) the conventional reactor and case 1 (b) case 2 and case 3.

Table 4 Comparison between predicted production rate and plant data.

DME (ton/day) Methanol (ton/day) Hot spot (K) Outlet temp. (K)

Hu et al. [12]

Model prediction

Relative error

342.7 49.3 531.9 516.6

362.82 56.3 523.9 517.15

5.87% 14.2% 1.50% 0.11%

Fig. 5. Water removal profiles for case 1, 2 and 3.


D. Iranshahi et al. / Fuel 190 (2017) 386–395

removal depends on its production. Due to temperature variation during the third case, production of water and so separation of water is greater than the second case. As it was previously considered, at the beginning of reactor, generation of water is low and then its generation gradually increases. Also, at the end of the reactor, because of reduction in water content of main stream, separation of water will be decreased. The rate of the first reaction is demonstrated in Fig. 6(a). Two factors affecting the reaction rate are temperature and concentration of reactants. During the first and third cases, because of higher temperature and higher reactants concentration, the reaction rate is high which leads to greater production of methanol. It should be mentioned that methanol generation will be decreased as the reactants concentration decreases. Although temperature rising in the second case is lower than that in the first and third cases, separation of water from the reactor forces the reaction to produce more water. Therefore B1 decreases because of conversion of methanol to DME. Fig. 6(b) demonstrates the changes of the rate of reaction B in the first and third case. In these reactions, as the water is separated, the concentration of H2 which depends on partial pressure of methanol and water vapor decreases. As a result, despite of the reduction of reactants concentration, the rate of this reaction increases. Reaction rate of DME is displayed in Fig. 6(c). The effects of temperature are clearly illustrated in this figure. In the first and third cases, reaction rate increases with increasing temperature. Then it decreases because of reduction of methanol concentration. However, in the second and third cases reactions (B) and (C) reactions B and C lead to higher methanol, water and DME production. Even with the consumption of reactant due to methanol production, reaction (C) will shift toward forward product generation. Since activation energy of this reaction is lower than that of the other reactions, temperature has minor effects on the rate of reaction. Fig. 7 shows variation of mole fraction of water. As it can be clearly seen, water generation is nearly high (its initial mole fraction is zero whereas its mole fraction at the end of reactor is 0.035 which is approximately 70 mol/s of water). Such an occurrence is not surprising because of production of water in the reactions (B) and (C). So, water removal has a significant effect on the DME production. As the water is separated, the rate of reaction B is enhancing, which results in an increase in methanol generation. Therefore by increasing methanol and decreasing water, reaction (C) will produce more DME. Since temperature is higher in the third case, its water production is slightly more than that in the second case. Despite of water removal in this case, mole fraction of water will be increased. Conversion of hydrogen and carbon monoxide are illustrated in Figs. 8 and 9. According to Le Chatelier’s principle, whenever a system in equilibrium is disturbed the system will adjust itself in such a way that the effect of the changes will be neutralized. As it can be seen from the figures, water separation decreases its partial pressure, shifting the reaction towards the production side. Consequently, conversion of hydrogen and carbon monoxide will be increased. Temperature affects the reaction rate, and thus it has a significant influence on the conversion. The effect of temperature on the reaction (A) is more than the others. Hence, carbon monoxide conversion is more considerable in the first case in comparison with the second one. Fig. 10(a) demonstrates mole fraction of DME in the reactor. As it was expected, by optimizing the temperature, reaction rates of DME has been increased in the first case. In the second and third cases, water reduces and reaction rates shifts towards the production side to compensate the changes. Therefore conversion of DME increases in these cases.



Fig. 6. Rate of (a) Reaction (A) and (b) Reaction (B) and (c) Reaction (C) for each case.

Fig. 10(b) shows the amount of methanol in the reactor. As it is clearly shown in this figure, in the second and third cases mole fraction of methanol is increased initially because of its production. Then, methanol converts to DME and its mole fraction will be

D. Iranshahi et al. / Fuel 190 (2017) 386–395


Fig. 7. Mole fraction of water in each case.

Fig. 10. Mole fraction and production of (a) DME (b) methanol for each case.

Fig. 8. Hydrogen conversion in reactor for each case.

Fig. 9. Conversion of carbon monooxide for each case.

Fig. 11. Comparison of DME production between CR and 3 cases.


D. Iranshahi et al. / Fuel 190 (2017) 386–395

decreased. In the first case, reaction (C) is dominant and so the amount of methanol is lower in comparison with the conventional reactor. Therefore, DME will be increased. Production of DME in the conventional reactor and three considered cases are shown in Fig. 11. As it was previously shown, maximum production of DME can be obtained in the third case. This type of reactor is completely ideal and as a real-world reactor design is closer to this case, the rate of DME production will be higher. As mentioned initially, the reactor is divided into twenty equal segments. However, from an engineering point of view, this method is not cost-effective. Fig. 12 depicts the difference between DME production in the proposed and conventional reactors. As it is shown, with increasing the segmentation DME production is increased. Difference between DME production in reactors with 5, 7 and 10 segments is approximately equal. It appears for operating units a reactor with 5 segments is practical and economical. One of the optimizing restriction is Pext > 49.6 bar. This constrain does not let the reactor pressure drop exceed that of the conventional reactor. As it is illustrated in Fig. 13 pressure drop in second and third case is slightly lower compared with conventional reactor which is because of removal of water from the main stream.

Fig. 12. Comparison of DME production in reactors with different segments to remove water.

Fig. 13. Pressure drop in reactor for CR and each case.

5. Conclusion In this novel theoretical study, the effect of temperature, water removal and combination of these two parameters were investigated for DME reactor. As the reactions of this process are exothermic, it is necessary to optimize the temperature profile to prevent catalyst deactivation and to enhance the reaction rates. Also, it was concluded that separation of water has direct effect on the production of DME and its production will be increased up to 55% in case 2 and 70% in case 3. Optimization of temperature profile in reactor increased DME production. Water could be removed by membrane or in any other ways. Another way to remove water is using water adsorbent. Zeolite 4A is a solid particle with high water adsorption affinity that makes it desirable for water removal or separation. Since this study is a theoretical, it should be considered economically because of importance of DME as an alternative fuel in the future. Such applications of DME make it economical. References [1] Bonura G, Cordaro M, Cannilla C, Mezzapica A, Spadaro L, Arena F, et al. Catalytic behaviour of a bifunctional system for the one step synthesis of DME by CO2 hydrogenation. Catal Today 2014;228:51–7. [2] Cai M, Subramanian V, Sushkevich V, Ordomsky V, Khodakov A. Effect of Sn additives on the CuZnAl–HZSM-5 hybrid catalysts for the direct DME synthesis from syngas. Appl Catal A 2015;502:370–9. [3] Ghavipour M, Behbahani RM. Fixed-bed reactor modeling for methanol to dimethyl ether (DME) reaction over c-alumina using a new practical reaction rate model. J Ind Eng Chem 2014;20:1942–51. [4] Kang Y, Lu T, Lu X, Wang Q, Huang X, Peng S, et al. Study on combustion characteristics of dimethyl ether under the moderate or intense low-oxygen dilution condition. Energy Convers Manage 2016;108:549–65. [5] Vakili R, Eslamloueyan R. Optimal design of an industrial scale dual-type reactor for direct dimethyl ether (DME) production from syngas. Chem Eng Process 2012;62:78–88. [6] Takeishi K, Akaike Y. Hydrogen production by dimethyl ether steam reforming over copper alumina catalysts prepared using the sol–gel method. Appl Catal A 2016;510:20–6. [7] Asthana S, Samanta C, Bhaumik A, Banerjee B, Voolapalli RK, Saha B. Direct synthesis of dimethyl ether from syngas over Cu-based catalysts: enhanced selectivity in the presence of MgO. J Catal 2016;334:89–101. [8] Naik SP, Ryu T, Bui V, Miller JD, Drinnan NB, Zmierczak W. Synthesis of DME from CO2/H2 gas mixture. Chem Eng J 2011;167:362–8. [9] Sun K, Lu W, Qiu F, Liu S, Xu X. Direct synthesis of DME over bifunctional catalyst: surface properties and catalytic performance. Appl Catal A 2003;252:243–9. [10] García-Trenco A, Martínez A. Direct synthesis of DME from syngas on hybrid CuZnAl/ZSM-5 catalysts: new insights into the role of zeolite acidity. Appl Catal A 2012;411:170–9. [11] Raoof F, Taghizadeh M, Eliassi A, Yaripour F. Effects of temperature and feed composition on catalytic dehydration of methanol to dimethyl ether over calumina. Fuel 2008;87:2967–71. [12] Hu Y, Nie Z, Fang D. Simulation and model design of pipe-shell reactor for the direct synthesis of dimethyl ether from syngas. J Nat Gas Chem 2008;17:195–200. [13] Vakili R, Pourazadi E, Setoodeh P, Eslamloueyan R, Rahimpour M. Direct dimethyl ether (DME) synthesis through a thermally coupled heat exchanger reactor. Appl Energy 2011;88:1211–23. [14] Nasehi M, Eslamloueyan R, Jahanmiri A. Simulation of DME synthesis reactor from methanol. In: 11th National Iranian chemical engineering congress, Tehran, Iran, 2006. [15] Farsi M, Jahanmiri A, Eslamloueyan R. Modeling and optimization of MeOH to DME in isothermal fixed-bed reactor. Int J Chem React Eng 2010;8. [16] Khademi M, Farsi M, Rahimpour M, Jahanmiri A. DME synthesis and cyclohexane dehydrogenation reaction in an optimized thermally coupled reactor. Chem Eng Process 2011;50:113–23. [17] Omata K, Ozaki T, Umegaki T, Watanabe Y, Nukui N, Yamada M. Optimization of the temperature profile of a temperature gradient reactor for DME synthesis using a simple genetic algorithm assisted by a neural network. Energy Fuels 2003;17:836–41. [18] Iranshahi D, Bahmanpour A, Paymooni K, Rahimpour M, Shariati A. Simultaneous hydrogen and aromatics enhancement by obtaining optimum temperature profile and hydrogen removal in naphtha reforming process; a novel theoretical study. Int J Hydrogen Energy 2011;36:8316–26. [19] Arabpour M, Rahimpour M, Iranshahi D, Raeissi S. Evaluation of maximum gasoline production of Fischer-Tropsch synthesis reactions in GTL technology: a discretized approach. J Nat Gas Sci Eng 2012;9:209–19. [20] Iliuta I, Larachi F, Fongarland P. Dimethyl ether synthesis with in situ H2O removal in fixed-bed membrane reactor: model and simulationsy. Ind Eng Chem Res 2010;49:6870–7.

D. Iranshahi et al. / Fuel 190 (2017) 386–395 [21] Kim J-H, Park MJ, Kim SJ, Joo O-S, Jung K-D. DME synthesis from synthesis gas on the admixed catalysts of Cu/ZnO/Al2O3 and ZSM-5. Appl Catal A 2004;264:37–41. [22] Vakili R, Rahmanifard H, Maroufi P, Eslamloueyan R, Rahimpour M. The effect of flow type patterns in a novel thermally coupled reactor for simultaneous direct dimethyl ether (DME) and hydrogen production. Int J Hydrogen Energy 2011;36:4354–65. [23] Zhu W, Gora L, Van den Berg A, Kapteijn F, Jansen J, Moulijn J. Water vapour separation from permanent gases by a zeolite-4A membrane. J Membr Sci 2005;253:57–66. [24] Bird RB. Transport phenomena. Appl Mech Rev 2002;55:R1–4. [25] Fogler HS. Elements of Chemical Reaction Engineering. 2nd ed. Englewood Cliffs: Prentice-Hall; 1992.


[26] Storn R, Price K. Differential evolution – a simple and efficient heuristic for global optimization over continuous spaces. J Glob Optim 1997;11:341–59. [27] Vakili R, Setoodeh P, Pourazadi E, Iranshahi D, Rahimpour M. Utilizing differential evolution (DE) technique to optimize operating conditions of an integrated thermally coupled direct DME synthesis reactor. Chem Eng J 2011;168:321–32. [28] Green DW. Perry’s chemical engineers’ handbook. New York: McGraw-hill; 2008. [29] Cussler EL. Diffusion: mass transfer in fluid systems. Cambridge University Press; 2009. [30] Reid R, Prausnitz J, Sherwood TK. The properties of gases and liquids. New York: McGraw-Hill; 1977.