- Email: [email protected]
monolithic two-stage catalyst bed reactor for oxidative coupling of methane
chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 390–399
Contents lists available at ScienceDirect
Chemical Engineering Research and Design journal homepage: www.elsevier.com/locate/cherd
Numerical simulation of particle/monolithic two-stage catalyst bed reactor for oxidative coupling of methane Zhang Zhao, Guo Ziqi, Ji Shengfu ∗ State Key Laboratory of Chemical Resource Engineering, Beijing University of Chemical Technology, 100029 Beijing, China
a r t i c l e
i n f o
a b s t r a c t
Article history:
Three-dimensional models were set up for the oxidative coupling of methane (OCM)
Received 20 March 2015
two-stage packed bed reactors loaded with Na2 WO4 –Mn/SiO2 particle catalyst and
Received in revised form 25 August
Na3 PO4 –Mn/SiO2 /cordierite monolithic catalyst using the computational fluid dynamics
2015
simulation. Firstly, the reactor with particle and monolithic catalyst bed heights of 10 mm
Accepted 2 September 2015
and 50 mm was simulated. Secondly, the effects of particle and monolithic catalyst bed
Available online 8 September 2015
heights on reactor performance were investigated. The results showed that the simulation values matched well with the experimental values on the conversion of CH4 and the selec-
Keywords:
tivity of products (C2 H6 , C2 H4 , CO, CO2 ) in the reactor outlet with an error range of ±10%. The
Oxidative coupling of methane
monolithic catalyst bed had a higher C2 (C2 H4 and C2 H6 ) selectivity and less O2 consumed
Monolithic catalyst
than the particle catalyst bed did. When the heights of particle and monolithic catalyst bed
Packed bed reactor
were 10 mm and 50 mm, respectively, the best values of C2 yield were 21.8% obtained due to
Computational fluid dynamics
the effects of residence time and the CO2 blocking in the oxidation reaction.
Numerical simulation
© 2015 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
1.
Introduction
The oxidative coupling of methane (OCM) is a promising way for the conversion of methane to C2 hydrocarbons (ethylene and ethane) and is considered to be the most concise way of natural gas utilization because that OCM is a direct conversion methods and it can avoid the syngas step (Gesser and Hunter, 1998; Lunsford, 2000). A large number of catalysts, mainly metal oxide based catalysts, have been reviewed by Khammona et al. (2012). Over the past thirty years, one of the most effective catalysts is believed to be Na–W–Mn/SiO2 catalyst which was first reported by Fang et al. (1992). Then, its structure, property for OCM reaction and addition of rare earth oxides were reported by substantial researchers in the literatures (Ji et al., 2003; Jiang et al., 1992; Malekzadeh et al., 2007; Wu et al., 2007). However, one major challenge of the commercialization of the OCM process is that the C2 yield is still not high enough.
∗
Another challenge of commercialization is that the OCM reaction is highly exothermic and hot spots are easily formed in the reactor (Pak and Lunsford, 1998; Schweer et al., 1994; Taniewski et al., 1997). Apart from the exploration of highly selective catalysts, the reactor type and alternated the contact mode between reactant and catalysts to inhibit the formation of hot spots and improve the selectivity of reactant was been researched (Liu et al., 2008a,b; Talebizadeh et al., 2009; Taniewski et al., 1997). Monolithic catalyst, which is prepared by coating the active components onto a regular structure support with appropriate channels, has lower pressure drop, smaller diffusion resistance, and more excellent mass and heat transfer than the traditional catalyst mode, i.e. in the particle form (Groppi et al., 2001). Liu et al. (2008b) and Tang et al. (2009) reported that the hot spot effect was effectively depressed and the selectivity of C2 was improved over the monolithic catalyst. However, the monolithic catalyst channels are relatively large, leading to
Corresponding author. Tel.: +86 10 64419619; fax: +86 10 64419619. E-mail address: [email protected] (S. Ji). http://dx.doi.org/10.1016/j.cherd.2015.09.001 0263-8762/© 2015 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 390–399
a low methane conversion because of the very short contact time between the reactants and active component. It is probably favorable to construct a two-stage reactor by combination of particle and monolithic catalysts in view of their disadvantages and advantages for OCM reaction. Pan et al. (2010) study the OCM reaction in a twostage reactor which is filled with Na2 WO4 –Mn/SiO2 particle catalyst and Ce–Na2 WO4 –Mn/SiO2 /cordierite monolithic catalyst. When particle and monolithic catalyst bed heights are 10 mm and 50 mm, respectively, and the raw gas goes through the particle catalyst first, C2 yield reaches its maximum value of 23.6%. Wang et al. (2011) confirm that similar results can be obtained if the monolithic catalyst is replaced by Na3 PO4 –Mn/SiO2 /cordierite. Since the raw gases first go through the particle catalyst bed, the OCM reaction in monolithic catalyst bed is depressed by the low partial pressure of O2 . A flow of supplementary O2 between the two beds is better to activate the monolithic catalyst (Ji and Wang, 2012). Finally, they improve the C2 yield to 24.3% when the flow rate of supplementary O2 is 3 ml/min. Meanwhile, the flow rate of CH4 and O2 in raw gas is 60 ml/min and 20 ml/min. Computational fluid dynamics (CFD) can accurately predict the effect of reactor flow field on heat transfer and chemical reaction. It has been applied to simulate the OCM reactor performance (Nakisa and Reza, 2008, 2009; Zhang et al., 2015a,b). Reaction kinetic model is the main accuracy factor of reactor model for CFD simulation. The blocking effect of CO2 and computational time must be considered if we plan to select a suitable reaction kinetics model. A kinetic model is built by Stansch et al. (1997), includes 10 step reactions and 8 species, is applied to the OCM reaction catalyzed by La2 O3 /CaO particles first and meet with our selected subjects. The model has been improved for simulation of reactors filled with Na2 WO4 –Mn/SiO2 particle catalyst (Zhang et al., 2015b) and Na3 PO4 –Mn/SiO2 /cordierite monolithic catalyst (Zhang et al., 2015a) in our previous work. In this work, the improved Stansch reaction kinetics models were used again and three-dimensional numerical models were established for the OCM tubular packed two-stage reactors filled with Na2 WO4 –Mn/SiO2 particle catalyst and Na3 PO4 –Mn/SiO2 /cordierite monolithic catalyst. The FLUENT commercial code was used to solve the Navier–Stokes equations and species transport equations. The reaction kinetics models were added by the user-defined function (UDF) of FLUENT software. The advantages of two-stage reactor were illustrated using the simulation results, such as the contour of the species mass fractions, fluid density and velocity. The effect of catalyst bed heights on reactor performance was also investigated using these models. These models were adopted to provide guidance for the reactor scaling up in the future.
2.
Models and numerical method
2.1.
Geometric model and meshes
To diminish the error of model, we established geometric models for two-stage reactors completely same with the experimental apparatus (Wang et al., 2011). The particle catalyst and the cordierite monolithic catalyst were placed in a quartz tube (inner diameter 8 mm, length 600 mm) and separated with quartz wool between the two catalyst beds to construct a two-stage reactor as shown in Fig. 1(a). Two sections of 75 mm length were filled with quartz clips above the
391
Fig. 1 – Sketch (a), meshes and geometric model (b and c) of packed bed reactor of monolithic catalyst. particle catalyst and under the cordierite monolithic catalyst (see Fig. 1(a)). For convenience, the particle catalyst bed was denoted as P and the monolithic catalyst bed was denoted as M. When the raw gas was fed from reactor top through the particle catalyst bed firstly and then through the monolithic catalyst bed, the reactor can be denoted as Px My (x, y, the height of particle catalyst bed and monolithic catalyst bed, respectively). In order to study the effect of catalyst heights on the performance of the reactor, a set of reactors, which were denoted as P10 M25 , P10 M50 , P10 M75 , P5 M50 and P15 M50 , was simulated in this work. The geometry models were portioned by meshes was similar with the model mentioned in previous work (Zhang et al., 2015b) and meshes of cross section and catalyst bed were shown in Fig. 1(b) and (c). The numbers of meshes in the reactor models were about 400 000, 650 000, 900 000, 600 000, 700 000 and 750 000, respectively.
2.2.
Governing equations
The simulation of packed bed reactor includes two kind of model: pseudo homogeneous model and heterogeneous ˇ et al., 2004; Stary´ et al., model. The heterogeneous model (Kocí 2006) is an excellent model for monolith bed simulation and can give accuracy results with short time. But heterogeneous model for particle bed simulation is a time consuming work (Wehinger et al., 2015). In this work, the reactor has two catalyst beds and heterogeneous model is not suitable to simulate the particle bed. We chose a pseudo homogeneous phase model (porous medium model) to save the computational time when the error was limited in an acceptable range. According to the operating conditions in literature (Wang et al., 2011), the flow in OCM packed bed reactor is laminar (Re = 5.7), which is suitable to be described by the Navier–Stokes equations (Batchelor, 2000). The porous medium model was used to estimate drag of catalyst bed on reaction gas flow. The temperature gradient from gas to catalyst was estimated by non-equilibrium thermal model in FLUENT software. In the reactor, the presence of molecular diffusion and convection diffusion between species made it necessary to use species transport equation. The set of governing equations was given in previous work (Zhang et al., 2015b).
392
chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 390–399
Table 1 – Simulated performance parameters of the reactors. Reactor
P5 M50 P10 M50 P15 M50 P10 M25 P10 M75
2.3.
CH4 conv. (%)
29.5 32.6 33.8 30.5 34.0
Sel. (%)
Sel. C2 (%)
C2 H4
C2 H6
CO
CO2
45.0 45.4 43.2 43.6 44.9
19.8 21.3 19.9 22.2 21.2
14.9 13.1 13.9 12.9 13.5
20.2 20.2 22.9 21.3 20.4
Kinetic model
The Stansch kinetic model contains 10 chemical reactions and 8 species, where the blocking effect of CO2 is taken into account in 6 reactions (Stansch et al., 1997). In previous works (Zhang et al., 2015a,b), we have modified the Stansch model parameters suitable to simulate the packed bed reactor filled with Na2 WO4 –Mn/SiO2 particle catalyst and Na3 PO4 –Mn/SiO2 /cordierite monolithic catalyst. The modified kinetics model was added to FLUENT software in the form of User Defined Function (UDF). The thermal chemical parameters of each species were mainly provided by the FLUENT database, and the other parameters were obtained from NIST database. Since the reactors are regarded as constant temperature, the viscosity coefficient, thermal conductivity, and diffusion coefficient between each component were assumed to be constant.
2.4. Operating conditions, boundary conditions and numerical methods According to the experiment reported by Wang et al. (2011), the operating condition was same as this literature. The boundary condition and numerical methods of reactor model was similar with the model mention in previous work (Zhang et al., 2015b).
3.
Results and discussion
3.1. Comparison of the simulated and experimental reactor performance Based on the experimental data, we simulated the OCM reaction phenomenon in the two-stage reactors using the mathematic model under the operation conditions introduced above. The CH4 conversion and the selectivity of main products and byproducts obtained by simulation and experiment are all listed in Tables 1 and 2. The relative errors of simulation values to the experimental of these parameters are listed in Table 3. When the monolithic catalyst bed height is fixed at 50 mm and the particle catalyst bed height ranged from 5 mm to 15 mm, the conversion of CH4 increases from 29.5% to 33.8%
64.8 66.7 63.1 65.8 66.1
C2 H4 /C2 H6
C2 yield (%)
2.3 2.1 2.2 2.0 2.1
19.2 21.8 21.3 20.0 22.4
by the simulated results and 31.4% to 32.8% by the experimental, respectively. For the yield and selectivity of C2 H4 , C2 H6 and their sum values, they all achieve their best values when the height of particle catalyst bed is 10 mm. For example, the selectivity of C2 H4 , C2 H6 and their sum values are 45.4%, 21.3% and 66.7% by simulation and 46.3%, 21.2% and 67.5% by experiment. At the same time, the corresponding C2 yield is 21.8% and 22.0%, respectively. Therefore, we fixed the height of particle catalyst bed at 10 mm, and changed the height of the monolithic catalyst bed from 25 mm to 75 mm. The results show that the CH4 conversion increases from 30.5% to 34.0% by calculation and from 31.8% to 32.7% by experiment. The selectivity of C2 H4 and C2 reached their best values when the height of monolithic catalyst bed is 50 mm, but the yield of C2 reaches the optimal value at 75 mm by simulation. The trend is not consistent with the experimental (see Table 2). This may be caused by the positive errors of simulated value of CH4 conversion and C2 selectivity to experimental values in P10 M75 (see Table 3). As can be seen from Table 3, the relative error of simulated values to the experimental is in the range of ±10%. This indicates that the simulated parameters values are agree well with those obtained by experiment, and proves that the OCM reactor model we established in this paper is reliable.
3.2.
Results of P10 M50
The reactor was laid vertically and the raw gas flow into the reactor from the top. The contours of parameters (such as the mass fraction of reactants and products) on a plane which passed through the symmetry axis of the catalyst bed were shown and the performance of two-stage OCM reactor was investigated. This plane was marked as P1 in this article. From Figs. 2 and 3, we can see that the mass fraction of each species in particle catalyst bed of P10 M50 is completely same as the packed bed reactor with 10 mm height particle catalyst (Zhang et al., 2015b). The mass fraction of reactants decrease more slowly along the flow direction in the monolithic catalyst bed than in particle catalyst bed, such as the value of CH4 and O2 change from 0.437 and 0.144 at monolithic bed inlet to 0.404 and 0.101 at the outlet, respectively. Meanwhile, the mass fraction of products increases more slowly in the monolithic catalyst bed in the
Table 2 – Experimental performance parameters of reactors (Wang et al., 2011). Reactor
P5 M50 P10 M50 P15 M50 P10 M25 P10 M75
CH4 conv. (%)
31.4 32.6 32.8 31.8 32.7
Sel. (%)
Sel. C2 (%)
C2 H4
C2 H6
CO
CO2
44.9 46.3 44.5 44.3 43.2
18.7 21.2 20.8 21.4 21.3
16.2 12.1 12.8 14.3 13.0
20.2 20.4 21.9 20.0 22.5
63.6 67.5 65.3 65.7 64.5
C2 H4 /C2 H6
C2 yield (%)
2.4 2.2 2.1 2.1 2.0
20.0 22.0 21.4 20.9 21.1
393
chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 390–399
Table 3 – Relative error between the simulated and the experimental performance parameters. Reactor
P5 M50 P10 M50 P15 M50 P10 M25 P10 M75
CH4 conv. (%)
−5.9 0.1 2.9 −4.2 3.8
Sel. (%) C2 H4
C2 H6
0.3 −1.9 −2.9 −1.5 3.9
5.7 0.5 −4.4 3.6 −0.4
Sel. C2 (%) CO
CO2
−7.7 8.2 9.0 −9.9 4.1
0.2 −0.9 4.8 6.6 −9.3
particle. This phenomenon suggests that the OCM reaction carries out slowly in the monolithic catalyst bed because of the small specific surface area and very short contact time between the reactants and active component (Tang et al., 2009). We can also see from Figs. 2 and 3 that the curvatures of mass fraction contours are smaller in laminar boundary layer of particle catalyst bed than monolithic catalyst bed. It indicates that the effect of boundary layer on the OCM reaction (Zhang et al., 2015a,b) is more apparent in monolithic catalyst bed. According to the mass fractions of reactants and products at the inlet and outlet of the catalyst beds, we can get the overall reaction equation of OCM in particle catalyst bed and monolithic catalyst bed under the simulation conditions as follows:
C2 H4 /C2 H6 (%)
1.9 −1.2 −3.4 0.1 2.4
C2 yield (%)
−5.1 −2.4 1.6 −5.0 4.3
−4.1 0.1 −0.5 −4.0 6.4
In particle catalyst bed, CH4 + 0.782O2 = 0.114C2 H6 + 0.213C2 H4 + 0.235CO2 + 0.111CO + 0.982H2 O + 0.250H2
(1)
and in monolithic catalyst bed, CH4 + 0.666O2 = 0.068C2 H6 + 0.300C2 H4 + 0.033CO2 + 0.231CO + 1.035H2 O + 0.162H2
(2)
Eqs. (1) and (2) tell us that when 1 mol methane converts to products of OCM reaction, there are 0.782 mol and 0.666 mol oxygen consumed and 0.327 mol and 0.368 mol C2 , 0.346 mol and 0.264 mol carbon oxides produced in particle catalyst bed
Fig. 2 – The contours of CH4 (a), O2 (b), C2 H6 (c) and C2 H4 (d) mass fractions on plane P1 in P10 M50 .
Fig. 3 – The contours of CO (a), CO2 (b), H2 (c) and H2 O (d) mass fractions on plane P1 in P10 M50 .
394
chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 390–399
Fig. 4 – The contours of gauge pressure on plane P1 in P10 M50 . with a height of 10 mm and in monolithic catalyst bed with a height of 50 mm, respectively. It suggests that the OCM reaction has a higher C2 selectivity and less O2 consumption in monolithic catalyst bed than in particle catalyst bed with the same CH4 consumption. Less O2 is consumed and less byproducts is produced in the monolithic catalyst bed; in addition, the OCM reactions are more slowly (see Figs. 2 and 3) in this bed. So, the monolithic bed depresses the hot spot more effectively than the particle catalyst bed does. The pressure in P10 M50 gradually decreases along the flow direction due to the friction drag produced by the catalyst beds and the wall surface (see Fig. 4). In this work, the gauge pressure at the outlet of the reactor was set to zero, and then the gauge pressure of the particle catalyst bed and the monolithic catalyst bed would be 9.0–11.0 Pa and 6.7–7.2 Pa. So the pressure drop of each catalyst bed is about 2 Pa and 0.5 Pa. Therefore, the pressure drop in particle catalyst bed is about 20 times as this in monolithic catalyst bed with the same bed height.
3.3.
Effect of particle catalyst bed height
Fig. 5 shows the effect of the height of the particle catalyst bed on the mass fractions of CH4 (see Fig. 5(a)) and O2 (see Fig. 5(b)). When the monolithic catalyst bed height is fixed at 50 mm, the
mass fractions of CH4 and O2 decrease as the height of particle catalyst bed increased from 5 mm to 15 mm and their average values at the outlet of this bed are 0.481, 0.437 and 0.413, and 0.227, 0.144 and 0.093, respectively. In monolithic catalyst bed, the mass fractions of CH4 and O2 also decline with increase of particle catalyst bed height and their average values are 0.423, 0.404 and 0.398, and 0.115, 0.101 and 0.078 at the outlet of the monolithic catalyst bed, respectively. The difference of CH4 and O2 mass fractions between inlet and outlet monolithic catalyst bed decrease with the increase of the particles catalyst bed height. Particularly, the mass fractions of CH4 and O2 almost keep constant in monolithic catalyst bed of P15 M50 reactor. Fig. 6 shows the effect of particle bed height on the mass fractions of C2 H6 (see Fig. 6(a)) and C2 H4 (see Fig. 6(b)). When the height of particle catalyst bed increases from 5 mm to 15 mm, the mass fractions of C2 H6 and C2 H4 increase at the outlet of this bed. The outlet values are 0.0322, 0.0350 and 0.0355 for C2 H6 and 0.0440, 0.0609 and 0.0689 for C2 H4 , respectively. In monolithic catalyst bed of P5 M50 , the mass fraction of C2 H6 increases from 0.0322 at bed inlet to 0.0349 at the middle of the bed firstly and then decreases to 0.0328 at the outlet. The mass fraction of C2 H4 the gradually increases along the direction of gas flow and reaches to 0.0690 at the outlet of the bed. This phenomenon is caused by the conversion from C2 H6 to C2 H4 in the lower half part of monolithic catalyst bed. When the producing rate of C2 H6 is slower than its consumption rate, its selectivity decreases (Zhang et al., 2015a). In monolithic catalyst bed of P10 M50 , it is noted that the mass fraction of C2 H6 increases rapidly from 0.0350 at bed inlet to 0.0391 in the vicinity of the bed inlet with a height about 10 mm and then it nearly keeps constant in the rest portion of the bed with a height about 40 mm. The mass fraction of C2 H4 gradually increases along the direction of gas flow to 0.0778 at the outlet of the bed in P10 M50 . Moreover, the mass fractions of C2 H6 and C2 H4 increase uniformly along the flow direction to 0.0378 and 0.0766 at the outlet of monolithic catalyst bed in P15 M50 . Fig. 7 shows the effect of the height of the particle catalyst bed on the mass fractions of CO (see Fig. 7(a)) and CO2 (see Fig. 7(b)). When the height of particle catalyst bed is changed from 5 mm to 15 mm, the mass fractions of CO and CO2 at the outlet of this bed are 0.0244, 0.0317 and 0.0379, and 0.0590, 0.106 and 0.137, respectively. In monolithic catalyst bed of P5 M50 , P10 M50 and P15 M50 , the mass fraction of CO all gradually increases along the gas flow
Fig. 5 – The contours of CH4 (a) and O2 (b) mass fractions on plane P1 in P5 M50 , P10 M50 and P15 M50 .
chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 390–399
395
Fig. 6 – The contours of C2 H6 (a) and C2 H4 (b) mass fractions on plane P1 in P5 M50 , P10 M50 and P15 M50 .
Fig. 7 – The contours of CO (a) and CO2 (b) mass fractions on plane P1 in P5 M50 , P10 M50 and P15 M50 . direction. But the increasing rate is faster in P5 M50 than in P10 M50 and P15 M50 . At the end of monolithic catalyst bed, the mass fraction of CO in each reactor is 0.0463, 0.0448 and 0.0494, respectively. Moreover, the mass fraction of CO2 increases gradually 0.0590–0.0985 in monolithic catalyst bed of P5 M50 , changes slightly from 0.106 to 0.109 in this bed of P10 M50 , and decreases slowly from 0.137 to 0.128 in this bed of P15 M50 along the gas flow direction. The mass fractions of byproducts CO and CO2 were controlled by the rates of first, third, sixth and tenth reaction steps according to the list of reaction equations (Stansch et al., 1997),
where the third, sixth, and tenth reaction steps are for CO production, the first step for CO2 production and tenth reaction step is for CO2 consumption. Figs. 8 and 9 show the rates of these steps in monolithic catalyst bed of P5 M50 , P10 M50 and P15 M50 , respectively, and illustrate the effect of particle catalyst bed height on the mass fraction contours of each species in monolithic bed, eventually on the performance of two-stage reactor. The mass fraction of O2 in the monolithic catalyst bed of P5 M50 is higher than P10 M50 (see Fig. 5(b)), so the rates of third step (see Fig. 8(b)) and sixth step (see Fig. 9(a)) are faster. The
Fig. 8 – The contours of reaction 1 (a) and 3 (b) rate (kmol m−3 s−1 ) in monolithic bed of P5 M50 , P10 M50 and P15 M50 .
396
chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 390–399
Fig. 9 – The contours of reaction 6 (a) and 10 (b) (kmol m−3 s−1 ) in monolithic bed of P5 M50 , P10 M50 and P15 M50 .
Fig. 10 – The contours of density (a) and velocity magnitude (b) on plane P1 in P5 M50 , P10 M50 and P15 M50 . product of these reactions, CO (see Fig. 7(a)), has a higher selectivity in P5 M50 than in P10 M50 . In the monolithic catalyst bed of P15 M50 , the mass fraction of O2 is lowest (see Fig. 5(b)), so the rate of first step (see Fig. 8(a)) is small, and the CO2 selectivity is little. Meanwhile, the mass fraction of CO2 in this region of P15 M50 is high. These dual reasons promote the CO2 consumption reaction (the tenth step, see Fig. 9(b)). Therefore, the CO mass fraction increases and CO2 mass fraction decreases along the gas flow direction. In brief, the mass fractions of CO and CO2 are dominated by the first, third, sixth and tenth reaction steps with rate
variation when the height of particle catalyst bed changes in P5 M50 , P10 M50 and P15 M50 . This results in that the mass fraction of CO in monolithic catalyst bed of P10 M50 is lower than P5 M50 and P15 M50 , and the mass fraction of CO2 in this bed of P15 M50 is the highest. The longer residence time lead to higher C2 yield in P10 M50 than P5 M50 , but the blocking effect of CO2 on the catalyst in P15 M50 reduces C2 yield. Consequently, the height of particle catalyst bed at 10 mm is the best value for C2 yield. Because OCM reaction belongs to volumetric increase reaction according to the previous work (Zhang et al., 2015a,b),
Fig. 11 – The contours of CH4 (a) and O2 (b) mass fractions on plane P1 in P10 M25 , P10 M50 and P10 M75 .
chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 390–399
397
Fig. 12 – The contours of C2 H6 (a) and C2 H4 (b) mass fractions on plane P1 in P10 M25 , P10 M50 and P10 M75 .
Fig. 13 – The contours of CO (a) and CO2 (b) mass fractions on plane P1 in P10 M25 , P10 M50 and P10 M75 . the density of reaction gas decreases and the velocity magnitude increases along the flow direction. As can be seen from Fig. 10(a), the density decreases from 0.227 to 0.225, 0.222 and 0.220 at the outlet of particle catalyst bed and to 0.221, 0.220 and 0.219 at the outlet of monolithic catalyst bed in P5 M50 , P10 M50 and P15 M50 , respectively. As shown in Fig. 10(b), the velocity magnitude increases at the outlet of particle catalyst bed with its height increasing. In the monolithic catalyst bed, a high velocity region exists in the outer edge of the boundary layer. This is caused by the combined effect of boundary layer, channels in monolithic catalyst and OCM reaction (Zhang
et al., 2015a). However, this high velocity region does not exist in particle catalyst bed, and its area becomes lager with the increase of particle catalyst bed height because of the smaller density.
3.4.
Effect of monolithic catalyst bed height
Figs. 11–13 show the effect of monolithic catalyst bed height (from 25 mm to 75 mm) on the mass fractions of CH4 and O2 , C2 H6 and C2 H4 , and CO and CO2 , when the height of the particle catalyst bed is fixed at 10 mm. The mass fraction contour
Fig. 14 – The contours of density (a) and velocity magnitude (b) on plane P1 in P5 M50 , P10 M50 and P15 M50 .
398
chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 390–399
of each species in particle catalyst bed of P10 M25 , P10 M50 and P10 M75 is completely same as the packed bed reactor with 10 mm height particle catalyst (Zhang et al., 2015b). The reactant mass fractions decrease and the product mass fractions increase along the flow direction. In the monolithic catalyst bed, the reactants are less and less and the products are more and more at the bed outlet when the height of monolithic bed changes from 25 mm to 75 mm. Moreover, the mass fraction contours in P10 M25 is a part of P10 M50 , just as P10 M50 is a part of P10 M75 . We can also see from Figs. 11–13 that the mass fraction contours of species in particle catalyst bed are denser than in monolithic catalyst bed. This phenomenon tells us that the OCM reaction in particle catalyst bed is fiercer than in monolithic catalyst bed again. Especially at the end of monolithic catalyst bed about 25 mm height in P10 M75 , the OCM reaction closes to stagnation. Fig. 14 shows the effect of the height of the monolithic catalyst bed from 25 mm to 75 mm on the gas density and velocity magnitude when the particle catalyst bed fixed at 50 mm. The density and velocity magnitude contour in particle catalyst bed of each reactor is same as the packed bed reactor with 10 mm height particle catalyst reported in previous work. In the monolithic catalyst bed, the density becomes smaller and the velocity magnitude becomes larger when this bed height becomes higher and residence time becomes longer.
4.
Conclusions
A three-dimensional numerical model was established to simulate the OCM tubular two-stage reactors with FLUENT software in the condition of that the volume rate was 80 ml/min under standard state with a CH4 /O2 ratio of 3 and the operating temperature and pressure were 800 ◦ C and 1 atm, respectively. The effect of catalyst bed heights on reactor performance was investigated using these models. The results show that the simulated values of the CH4 conversion, products selectivity and other parameters in the outlet of reactors are consistent with the experimental values in acceptable errors. It showed that two-stage fixed bed reactor had combined advantages of particle catalyst bed and monolithic catalyst bed, such as high conversion of CH4 , high selectivity of C2 , small O2 consumption and low pressure drop, so the reactor has potential of depressing hot spot. We also can find that the OCM reaction was fiercer and the effect of boundary layer on OCM reaction was weaker in particle catalyst bed than in monolithic catalyst bed of two-stage reactors. The effects of residence time caused by bed height and the blocking caused by CO2 in the oxidation reaction induce the P10 M50 reactor having the best performance for C2 yield when the height of particle catalyst bed changed from 5 mm to 15 mm. When the height particle catalyst bed was fixed at 10 mm and the height monolithic catalyst bed increased from 25 mm to 75 mm, the mass fraction contours in P10 M25 was a part of P10 M50 , just as P10 M50 was a part of P10 M75 .
References Batchelor, G.K., 2000. An Introduction to Fluid Dynamics, second ed. Cambridge University Press, Cambridge. Fang, X., Li, S., Lin, J., Chu, Y., 1992. Oxidative coupling of methane on W–Mn catalysts. J. Mol. Catal. China 6, 427–433. Gesser, H.D., Hunter, N.R., 1998. A review of C-1 conversion chemistry. Catal. Today 42, 183–189.
Groppi, G., Ibashi, W., Tronconi, E., Forzatti, P., 2001. Structured reactors for kinetic measurements in catalytic combustion. Chem. Eng. J. (Lausanne) 82, 57–71. Ji, S., Wang, W., 2012. Particle/cordierite monolithic catalysts dual-bed reactor with beds-interspace supplementary oxygen and performance for oxidative coupling of methane. Int. J. Chem. React. Eng. 10, 1–13. Ji, S., Xiao, T., Li, S., Chou, L., Zhang, B., Xu, C., Hou, R., York, A.P.E., Green, M.L.H., 2003. Surface WO4 tetrahedron: the essence of the oxidative coupling of methane over M–W–Mn/SiO2 catalysts. J. Catal. 220, 47–56. Jiang, Z., Fang, X., Yu, C., Li, S., 1992. Mn–Na2 WO4 /SiO2 catalyst for oxidative coupling of methane-I, surface dispersion state and oxide support interaction. J. Mol. Catal. China 6, 477–480. Khammona, K., Assabumrungrat, S., Wiyaratn, W., 2012. Reviews on coupling of methane over catalysts for application in C2 hydrocarbon production. J. Eng. Appl. Sci. 7, 447–455. ˇ P., Kubícek, ˇ Kocí, M., Marek, M., 2004. Modeling of three-way-catalyst monolith converters with microkinetics and diffusion in the washcoat. Ind. Eng. Chem. Res. 43, 4503–4510. Liu, H., Wang, X., Yang, D., Gao, R., Wang, Z., Yang, J., 2008a. Scale up and stability test for oxidative coupling of methane over Na2 WO4 –Mn/SiO2 catalyst in a 200 ml fixed-bed reactor. J. Nat. Gas Chem. 17, 59–63. Liu, H., Yang, D., Gao, R., Chen, L., Zhang, S., Wang, X., 2008b. A novel Na2 WO4 –Mn/SiC monolithic foam catalyst with improved thermal properties for the oxidative coupling of methane. Catal. Commun. 9, 1302–1306. Lunsford, J.H., 2000. Catalytic conversion of methane to more useful chemicals and fuels: a challenge for the 21st century. Catal. Today 63, 165–174. Malekzadeh, A., Khodadadi, A., Dalai, A.K., Abedini, M., 2007. Oxidative coupling of methane over lithium doped (Mn + W)/SiO2 catalysts. J. Nat. Gas Chem. 16, 121–129. Nakisa, Y., Reza, G.M.H., 2008. Oxidative coupling of methane in a fixed bed reactor over perovskite catalyst: a simulation study using experimental kinetic model. J. Nat. Gas Chem. 17, 8–16. Nakisa, Y., Reza, G.M.H., 2009. Modeling the oxidative coupling of methane: heterogeneous chemistry coupled with 3D flow field simulation. J. Nat. Gas Chem. 18, 39–44. Pak, S., Lunsford, J.H., 1998. Thermal effects during the oxidative coupling of methane over Mn/Na2 WO4 /SiO2 and Mn/Na2 WO4 /MgO catalysts. Appl. Catal. A 168, 131–137. Pan, D., Ji, S., Wang, W., Li, C., 2010. Oxidative coupling of methane in a dual-bed reactor comprising of particle/cordierite monolithic catalysts. J. Nat. Gas Chem. 19, 600–604. Schweer, D., Meeczko, L., Baerns, M., 1994. OCM in a fixed-bed reactor: limits and perspectives. Catal. Today 21, 357–369. Stansch, Z., Mleczko, L., Baerns, M., 1997. Comprehensive kinetics of oxidative coupling of methane over the La2 O3 /CaO catalyst. Ind. Eng. Chem. Res. 36, 2568–2579. ˇ ´ T., Solcová, Stary, O., Schneider, P., Marek, M., 2006. Effective diffusivities and pore-transport characteristics of washcoated ceramic monolith for automotive catalytic converter. Chem. Eng. Sci. 61, 5934–5943. Talebizadeh, A., Mortazavi, Y., Khodadadi, A.A., 2009. Comparative study of the two-zone fluidized-bed reactor and the fluidized-bed reactor for oxidative coupling of methane over Mn/Na2 WO4 /SiO2 catalyst. Fuel Process. Technol. 90, 1319–1325. Tang, J., Ji, S., Wang, K., Li, C., 2009. Preparation of M–W–Mn/SiO2 /cordierite monolithic catalysts and their performances for oxidative coupling of methane. Nat. Gas Chem. Ind. (China) 34, 19–26. Taniewski, M., Lachowicz, A., Skutil, K., 1997. Effective utilisation of the catalyst bed deactivating in the course of oxidative coupling of methane. Chem. Eng. Sci. 52, 935–939. Wang, W., Ji, S., Pan, D., Li, C., 2011. A novel particle/monolithic two-stage catalyst bed reactor and their catalytic performance for oxidative coupling of methane. Fuel Process. Technol. 92, 541–546.
chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 390–399
Wehinger, G.D., Eppinger, T., Kraume, M., 2015. Detailed numerical simulations of catalytic fixed-bed reactors: heterogeneous dry reforming of methane. Chem. Eng. Sci. 122, 197–209. Wu, J., Zhang, H., Qin, S., Hu, C., 2007. La-promoted Na2 WO4 /Mn/SiO2 catalysts for the oxidative conversion of methane simultaneously to ethylene and carbon monoxide. Appl. Catal. A 323, 126–134.
399
Zhang, Z., Guo, Z., Ji, S., 2015a. Numerical simulation of monolithic catalyst fixed bed reactor for oxidative coupling of methane. Chin. J. Chem. Eng. Zhang, Z., Guo, Z., Ji, S., 2015b. Numerical simulation of packed bed reactor for oxidative coupling of methane. J. Energy Chem. 24, 23–30.
Contents lists available at ScienceDirect
Chemical Engineering Research and Design journal homepage: www.elsevier.com/locate/cherd
Numerical simulation of particle/monolithic two-stage catalyst bed reactor for oxidative coupling of methane Zhang Zhao, Guo Ziqi, Ji Shengfu ∗ State Key Laboratory of Chemical Resource Engineering, Beijing University of Chemical Technology, 100029 Beijing, China
a r t i c l e
i n f o
a b s t r a c t
Article history:
Three-dimensional models were set up for the oxidative coupling of methane (OCM)
Received 20 March 2015
two-stage packed bed reactors loaded with Na2 WO4 –Mn/SiO2 particle catalyst and
Received in revised form 25 August
Na3 PO4 –Mn/SiO2 /cordierite monolithic catalyst using the computational fluid dynamics
2015
simulation. Firstly, the reactor with particle and monolithic catalyst bed heights of 10 mm
Accepted 2 September 2015
and 50 mm was simulated. Secondly, the effects of particle and monolithic catalyst bed
Available online 8 September 2015
heights on reactor performance were investigated. The results showed that the simulation values matched well with the experimental values on the conversion of CH4 and the selec-
Keywords:
tivity of products (C2 H6 , C2 H4 , CO, CO2 ) in the reactor outlet with an error range of ±10%. The
Oxidative coupling of methane
monolithic catalyst bed had a higher C2 (C2 H4 and C2 H6 ) selectivity and less O2 consumed
Monolithic catalyst
than the particle catalyst bed did. When the heights of particle and monolithic catalyst bed
Packed bed reactor
were 10 mm and 50 mm, respectively, the best values of C2 yield were 21.8% obtained due to
Computational fluid dynamics
the effects of residence time and the CO2 blocking in the oxidation reaction.
Numerical simulation
© 2015 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
1.
Introduction
The oxidative coupling of methane (OCM) is a promising way for the conversion of methane to C2 hydrocarbons (ethylene and ethane) and is considered to be the most concise way of natural gas utilization because that OCM is a direct conversion methods and it can avoid the syngas step (Gesser and Hunter, 1998; Lunsford, 2000). A large number of catalysts, mainly metal oxide based catalysts, have been reviewed by Khammona et al. (2012). Over the past thirty years, one of the most effective catalysts is believed to be Na–W–Mn/SiO2 catalyst which was first reported by Fang et al. (1992). Then, its structure, property for OCM reaction and addition of rare earth oxides were reported by substantial researchers in the literatures (Ji et al., 2003; Jiang et al., 1992; Malekzadeh et al., 2007; Wu et al., 2007). However, one major challenge of the commercialization of the OCM process is that the C2 yield is still not high enough.
∗
Another challenge of commercialization is that the OCM reaction is highly exothermic and hot spots are easily formed in the reactor (Pak and Lunsford, 1998; Schweer et al., 1994; Taniewski et al., 1997). Apart from the exploration of highly selective catalysts, the reactor type and alternated the contact mode between reactant and catalysts to inhibit the formation of hot spots and improve the selectivity of reactant was been researched (Liu et al., 2008a,b; Talebizadeh et al., 2009; Taniewski et al., 1997). Monolithic catalyst, which is prepared by coating the active components onto a regular structure support with appropriate channels, has lower pressure drop, smaller diffusion resistance, and more excellent mass and heat transfer than the traditional catalyst mode, i.e. in the particle form (Groppi et al., 2001). Liu et al. (2008b) and Tang et al. (2009) reported that the hot spot effect was effectively depressed and the selectivity of C2 was improved over the monolithic catalyst. However, the monolithic catalyst channels are relatively large, leading to
Corresponding author. Tel.: +86 10 64419619; fax: +86 10 64419619. E-mail address: [email protected] (S. Ji). http://dx.doi.org/10.1016/j.cherd.2015.09.001 0263-8762/© 2015 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 390–399
a low methane conversion because of the very short contact time between the reactants and active component. It is probably favorable to construct a two-stage reactor by combination of particle and monolithic catalysts in view of their disadvantages and advantages for OCM reaction. Pan et al. (2010) study the OCM reaction in a twostage reactor which is filled with Na2 WO4 –Mn/SiO2 particle catalyst and Ce–Na2 WO4 –Mn/SiO2 /cordierite monolithic catalyst. When particle and monolithic catalyst bed heights are 10 mm and 50 mm, respectively, and the raw gas goes through the particle catalyst first, C2 yield reaches its maximum value of 23.6%. Wang et al. (2011) confirm that similar results can be obtained if the monolithic catalyst is replaced by Na3 PO4 –Mn/SiO2 /cordierite. Since the raw gases first go through the particle catalyst bed, the OCM reaction in monolithic catalyst bed is depressed by the low partial pressure of O2 . A flow of supplementary O2 between the two beds is better to activate the monolithic catalyst (Ji and Wang, 2012). Finally, they improve the C2 yield to 24.3% when the flow rate of supplementary O2 is 3 ml/min. Meanwhile, the flow rate of CH4 and O2 in raw gas is 60 ml/min and 20 ml/min. Computational fluid dynamics (CFD) can accurately predict the effect of reactor flow field on heat transfer and chemical reaction. It has been applied to simulate the OCM reactor performance (Nakisa and Reza, 2008, 2009; Zhang et al., 2015a,b). Reaction kinetic model is the main accuracy factor of reactor model for CFD simulation. The blocking effect of CO2 and computational time must be considered if we plan to select a suitable reaction kinetics model. A kinetic model is built by Stansch et al. (1997), includes 10 step reactions and 8 species, is applied to the OCM reaction catalyzed by La2 O3 /CaO particles first and meet with our selected subjects. The model has been improved for simulation of reactors filled with Na2 WO4 –Mn/SiO2 particle catalyst (Zhang et al., 2015b) and Na3 PO4 –Mn/SiO2 /cordierite monolithic catalyst (Zhang et al., 2015a) in our previous work. In this work, the improved Stansch reaction kinetics models were used again and three-dimensional numerical models were established for the OCM tubular packed two-stage reactors filled with Na2 WO4 –Mn/SiO2 particle catalyst and Na3 PO4 –Mn/SiO2 /cordierite monolithic catalyst. The FLUENT commercial code was used to solve the Navier–Stokes equations and species transport equations. The reaction kinetics models were added by the user-defined function (UDF) of FLUENT software. The advantages of two-stage reactor were illustrated using the simulation results, such as the contour of the species mass fractions, fluid density and velocity. The effect of catalyst bed heights on reactor performance was also investigated using these models. These models were adopted to provide guidance for the reactor scaling up in the future.
2.
Models and numerical method
2.1.
Geometric model and meshes
To diminish the error of model, we established geometric models for two-stage reactors completely same with the experimental apparatus (Wang et al., 2011). The particle catalyst and the cordierite monolithic catalyst were placed in a quartz tube (inner diameter 8 mm, length 600 mm) and separated with quartz wool between the two catalyst beds to construct a two-stage reactor as shown in Fig. 1(a). Two sections of 75 mm length were filled with quartz clips above the
391
Fig. 1 – Sketch (a), meshes and geometric model (b and c) of packed bed reactor of monolithic catalyst. particle catalyst and under the cordierite monolithic catalyst (see Fig. 1(a)). For convenience, the particle catalyst bed was denoted as P and the monolithic catalyst bed was denoted as M. When the raw gas was fed from reactor top through the particle catalyst bed firstly and then through the monolithic catalyst bed, the reactor can be denoted as Px My (x, y, the height of particle catalyst bed and monolithic catalyst bed, respectively). In order to study the effect of catalyst heights on the performance of the reactor, a set of reactors, which were denoted as P10 M25 , P10 M50 , P10 M75 , P5 M50 and P15 M50 , was simulated in this work. The geometry models were portioned by meshes was similar with the model mentioned in previous work (Zhang et al., 2015b) and meshes of cross section and catalyst bed were shown in Fig. 1(b) and (c). The numbers of meshes in the reactor models were about 400 000, 650 000, 900 000, 600 000, 700 000 and 750 000, respectively.
2.2.
Governing equations
The simulation of packed bed reactor includes two kind of model: pseudo homogeneous model and heterogeneous ˇ et al., 2004; Stary´ et al., model. The heterogeneous model (Kocí 2006) is an excellent model for monolith bed simulation and can give accuracy results with short time. But heterogeneous model for particle bed simulation is a time consuming work (Wehinger et al., 2015). In this work, the reactor has two catalyst beds and heterogeneous model is not suitable to simulate the particle bed. We chose a pseudo homogeneous phase model (porous medium model) to save the computational time when the error was limited in an acceptable range. According to the operating conditions in literature (Wang et al., 2011), the flow in OCM packed bed reactor is laminar (Re = 5.7), which is suitable to be described by the Navier–Stokes equations (Batchelor, 2000). The porous medium model was used to estimate drag of catalyst bed on reaction gas flow. The temperature gradient from gas to catalyst was estimated by non-equilibrium thermal model in FLUENT software. In the reactor, the presence of molecular diffusion and convection diffusion between species made it necessary to use species transport equation. The set of governing equations was given in previous work (Zhang et al., 2015b).
392
chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 390–399
Table 1 – Simulated performance parameters of the reactors. Reactor
P5 M50 P10 M50 P15 M50 P10 M25 P10 M75
2.3.
CH4 conv. (%)
29.5 32.6 33.8 30.5 34.0
Sel. (%)
Sel. C2 (%)
C2 H4
C2 H6
CO
CO2
45.0 45.4 43.2 43.6 44.9
19.8 21.3 19.9 22.2 21.2
14.9 13.1 13.9 12.9 13.5
20.2 20.2 22.9 21.3 20.4
Kinetic model
The Stansch kinetic model contains 10 chemical reactions and 8 species, where the blocking effect of CO2 is taken into account in 6 reactions (Stansch et al., 1997). In previous works (Zhang et al., 2015a,b), we have modified the Stansch model parameters suitable to simulate the packed bed reactor filled with Na2 WO4 –Mn/SiO2 particle catalyst and Na3 PO4 –Mn/SiO2 /cordierite monolithic catalyst. The modified kinetics model was added to FLUENT software in the form of User Defined Function (UDF). The thermal chemical parameters of each species were mainly provided by the FLUENT database, and the other parameters were obtained from NIST database. Since the reactors are regarded as constant temperature, the viscosity coefficient, thermal conductivity, and diffusion coefficient between each component were assumed to be constant.
2.4. Operating conditions, boundary conditions and numerical methods According to the experiment reported by Wang et al. (2011), the operating condition was same as this literature. The boundary condition and numerical methods of reactor model was similar with the model mention in previous work (Zhang et al., 2015b).
3.
Results and discussion
3.1. Comparison of the simulated and experimental reactor performance Based on the experimental data, we simulated the OCM reaction phenomenon in the two-stage reactors using the mathematic model under the operation conditions introduced above. The CH4 conversion and the selectivity of main products and byproducts obtained by simulation and experiment are all listed in Tables 1 and 2. The relative errors of simulation values to the experimental of these parameters are listed in Table 3. When the monolithic catalyst bed height is fixed at 50 mm and the particle catalyst bed height ranged from 5 mm to 15 mm, the conversion of CH4 increases from 29.5% to 33.8%
64.8 66.7 63.1 65.8 66.1
C2 H4 /C2 H6
C2 yield (%)
2.3 2.1 2.2 2.0 2.1
19.2 21.8 21.3 20.0 22.4
by the simulated results and 31.4% to 32.8% by the experimental, respectively. For the yield and selectivity of C2 H4 , C2 H6 and their sum values, they all achieve their best values when the height of particle catalyst bed is 10 mm. For example, the selectivity of C2 H4 , C2 H6 and their sum values are 45.4%, 21.3% and 66.7% by simulation and 46.3%, 21.2% and 67.5% by experiment. At the same time, the corresponding C2 yield is 21.8% and 22.0%, respectively. Therefore, we fixed the height of particle catalyst bed at 10 mm, and changed the height of the monolithic catalyst bed from 25 mm to 75 mm. The results show that the CH4 conversion increases from 30.5% to 34.0% by calculation and from 31.8% to 32.7% by experiment. The selectivity of C2 H4 and C2 reached their best values when the height of monolithic catalyst bed is 50 mm, but the yield of C2 reaches the optimal value at 75 mm by simulation. The trend is not consistent with the experimental (see Table 2). This may be caused by the positive errors of simulated value of CH4 conversion and C2 selectivity to experimental values in P10 M75 (see Table 3). As can be seen from Table 3, the relative error of simulated values to the experimental is in the range of ±10%. This indicates that the simulated parameters values are agree well with those obtained by experiment, and proves that the OCM reactor model we established in this paper is reliable.
3.2.
Results of P10 M50
The reactor was laid vertically and the raw gas flow into the reactor from the top. The contours of parameters (such as the mass fraction of reactants and products) on a plane which passed through the symmetry axis of the catalyst bed were shown and the performance of two-stage OCM reactor was investigated. This plane was marked as P1 in this article. From Figs. 2 and 3, we can see that the mass fraction of each species in particle catalyst bed of P10 M50 is completely same as the packed bed reactor with 10 mm height particle catalyst (Zhang et al., 2015b). The mass fraction of reactants decrease more slowly along the flow direction in the monolithic catalyst bed than in particle catalyst bed, such as the value of CH4 and O2 change from 0.437 and 0.144 at monolithic bed inlet to 0.404 and 0.101 at the outlet, respectively. Meanwhile, the mass fraction of products increases more slowly in the monolithic catalyst bed in the
Table 2 – Experimental performance parameters of reactors (Wang et al., 2011). Reactor
P5 M50 P10 M50 P15 M50 P10 M25 P10 M75
CH4 conv. (%)
31.4 32.6 32.8 31.8 32.7
Sel. (%)
Sel. C2 (%)
C2 H4
C2 H6
CO
CO2
44.9 46.3 44.5 44.3 43.2
18.7 21.2 20.8 21.4 21.3
16.2 12.1 12.8 14.3 13.0
20.2 20.4 21.9 20.0 22.5
63.6 67.5 65.3 65.7 64.5
C2 H4 /C2 H6
C2 yield (%)
2.4 2.2 2.1 2.1 2.0
20.0 22.0 21.4 20.9 21.1
393
chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 390–399
Table 3 – Relative error between the simulated and the experimental performance parameters. Reactor
P5 M50 P10 M50 P15 M50 P10 M25 P10 M75
CH4 conv. (%)
−5.9 0.1 2.9 −4.2 3.8
Sel. (%) C2 H4
C2 H6
0.3 −1.9 −2.9 −1.5 3.9
5.7 0.5 −4.4 3.6 −0.4
Sel. C2 (%) CO
CO2
−7.7 8.2 9.0 −9.9 4.1
0.2 −0.9 4.8 6.6 −9.3
particle. This phenomenon suggests that the OCM reaction carries out slowly in the monolithic catalyst bed because of the small specific surface area and very short contact time between the reactants and active component (Tang et al., 2009). We can also see from Figs. 2 and 3 that the curvatures of mass fraction contours are smaller in laminar boundary layer of particle catalyst bed than monolithic catalyst bed. It indicates that the effect of boundary layer on the OCM reaction (Zhang et al., 2015a,b) is more apparent in monolithic catalyst bed. According to the mass fractions of reactants and products at the inlet and outlet of the catalyst beds, we can get the overall reaction equation of OCM in particle catalyst bed and monolithic catalyst bed under the simulation conditions as follows:
C2 H4 /C2 H6 (%)
1.9 −1.2 −3.4 0.1 2.4
C2 yield (%)
−5.1 −2.4 1.6 −5.0 4.3
−4.1 0.1 −0.5 −4.0 6.4
In particle catalyst bed, CH4 + 0.782O2 = 0.114C2 H6 + 0.213C2 H4 + 0.235CO2 + 0.111CO + 0.982H2 O + 0.250H2
(1)
and in monolithic catalyst bed, CH4 + 0.666O2 = 0.068C2 H6 + 0.300C2 H4 + 0.033CO2 + 0.231CO + 1.035H2 O + 0.162H2
(2)
Eqs. (1) and (2) tell us that when 1 mol methane converts to products of OCM reaction, there are 0.782 mol and 0.666 mol oxygen consumed and 0.327 mol and 0.368 mol C2 , 0.346 mol and 0.264 mol carbon oxides produced in particle catalyst bed
Fig. 2 – The contours of CH4 (a), O2 (b), C2 H6 (c) and C2 H4 (d) mass fractions on plane P1 in P10 M50 .
Fig. 3 – The contours of CO (a), CO2 (b), H2 (c) and H2 O (d) mass fractions on plane P1 in P10 M50 .
394
chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 390–399
Fig. 4 – The contours of gauge pressure on plane P1 in P10 M50 . with a height of 10 mm and in monolithic catalyst bed with a height of 50 mm, respectively. It suggests that the OCM reaction has a higher C2 selectivity and less O2 consumption in monolithic catalyst bed than in particle catalyst bed with the same CH4 consumption. Less O2 is consumed and less byproducts is produced in the monolithic catalyst bed; in addition, the OCM reactions are more slowly (see Figs. 2 and 3) in this bed. So, the monolithic bed depresses the hot spot more effectively than the particle catalyst bed does. The pressure in P10 M50 gradually decreases along the flow direction due to the friction drag produced by the catalyst beds and the wall surface (see Fig. 4). In this work, the gauge pressure at the outlet of the reactor was set to zero, and then the gauge pressure of the particle catalyst bed and the monolithic catalyst bed would be 9.0–11.0 Pa and 6.7–7.2 Pa. So the pressure drop of each catalyst bed is about 2 Pa and 0.5 Pa. Therefore, the pressure drop in particle catalyst bed is about 20 times as this in monolithic catalyst bed with the same bed height.
3.3.
Effect of particle catalyst bed height
Fig. 5 shows the effect of the height of the particle catalyst bed on the mass fractions of CH4 (see Fig. 5(a)) and O2 (see Fig. 5(b)). When the monolithic catalyst bed height is fixed at 50 mm, the
mass fractions of CH4 and O2 decrease as the height of particle catalyst bed increased from 5 mm to 15 mm and their average values at the outlet of this bed are 0.481, 0.437 and 0.413, and 0.227, 0.144 and 0.093, respectively. In monolithic catalyst bed, the mass fractions of CH4 and O2 also decline with increase of particle catalyst bed height and their average values are 0.423, 0.404 and 0.398, and 0.115, 0.101 and 0.078 at the outlet of the monolithic catalyst bed, respectively. The difference of CH4 and O2 mass fractions between inlet and outlet monolithic catalyst bed decrease with the increase of the particles catalyst bed height. Particularly, the mass fractions of CH4 and O2 almost keep constant in monolithic catalyst bed of P15 M50 reactor. Fig. 6 shows the effect of particle bed height on the mass fractions of C2 H6 (see Fig. 6(a)) and C2 H4 (see Fig. 6(b)). When the height of particle catalyst bed increases from 5 mm to 15 mm, the mass fractions of C2 H6 and C2 H4 increase at the outlet of this bed. The outlet values are 0.0322, 0.0350 and 0.0355 for C2 H6 and 0.0440, 0.0609 and 0.0689 for C2 H4 , respectively. In monolithic catalyst bed of P5 M50 , the mass fraction of C2 H6 increases from 0.0322 at bed inlet to 0.0349 at the middle of the bed firstly and then decreases to 0.0328 at the outlet. The mass fraction of C2 H4 the gradually increases along the direction of gas flow and reaches to 0.0690 at the outlet of the bed. This phenomenon is caused by the conversion from C2 H6 to C2 H4 in the lower half part of monolithic catalyst bed. When the producing rate of C2 H6 is slower than its consumption rate, its selectivity decreases (Zhang et al., 2015a). In monolithic catalyst bed of P10 M50 , it is noted that the mass fraction of C2 H6 increases rapidly from 0.0350 at bed inlet to 0.0391 in the vicinity of the bed inlet with a height about 10 mm and then it nearly keeps constant in the rest portion of the bed with a height about 40 mm. The mass fraction of C2 H4 gradually increases along the direction of gas flow to 0.0778 at the outlet of the bed in P10 M50 . Moreover, the mass fractions of C2 H6 and C2 H4 increase uniformly along the flow direction to 0.0378 and 0.0766 at the outlet of monolithic catalyst bed in P15 M50 . Fig. 7 shows the effect of the height of the particle catalyst bed on the mass fractions of CO (see Fig. 7(a)) and CO2 (see Fig. 7(b)). When the height of particle catalyst bed is changed from 5 mm to 15 mm, the mass fractions of CO and CO2 at the outlet of this bed are 0.0244, 0.0317 and 0.0379, and 0.0590, 0.106 and 0.137, respectively. In monolithic catalyst bed of P5 M50 , P10 M50 and P15 M50 , the mass fraction of CO all gradually increases along the gas flow
Fig. 5 – The contours of CH4 (a) and O2 (b) mass fractions on plane P1 in P5 M50 , P10 M50 and P15 M50 .
chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 390–399
395
Fig. 6 – The contours of C2 H6 (a) and C2 H4 (b) mass fractions on plane P1 in P5 M50 , P10 M50 and P15 M50 .
Fig. 7 – The contours of CO (a) and CO2 (b) mass fractions on plane P1 in P5 M50 , P10 M50 and P15 M50 . direction. But the increasing rate is faster in P5 M50 than in P10 M50 and P15 M50 . At the end of monolithic catalyst bed, the mass fraction of CO in each reactor is 0.0463, 0.0448 and 0.0494, respectively. Moreover, the mass fraction of CO2 increases gradually 0.0590–0.0985 in monolithic catalyst bed of P5 M50 , changes slightly from 0.106 to 0.109 in this bed of P10 M50 , and decreases slowly from 0.137 to 0.128 in this bed of P15 M50 along the gas flow direction. The mass fractions of byproducts CO and CO2 were controlled by the rates of first, third, sixth and tenth reaction steps according to the list of reaction equations (Stansch et al., 1997),
where the third, sixth, and tenth reaction steps are for CO production, the first step for CO2 production and tenth reaction step is for CO2 consumption. Figs. 8 and 9 show the rates of these steps in monolithic catalyst bed of P5 M50 , P10 M50 and P15 M50 , respectively, and illustrate the effect of particle catalyst bed height on the mass fraction contours of each species in monolithic bed, eventually on the performance of two-stage reactor. The mass fraction of O2 in the monolithic catalyst bed of P5 M50 is higher than P10 M50 (see Fig. 5(b)), so the rates of third step (see Fig. 8(b)) and sixth step (see Fig. 9(a)) are faster. The
Fig. 8 – The contours of reaction 1 (a) and 3 (b) rate (kmol m−3 s−1 ) in monolithic bed of P5 M50 , P10 M50 and P15 M50 .
396
chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 390–399
Fig. 9 – The contours of reaction 6 (a) and 10 (b) (kmol m−3 s−1 ) in monolithic bed of P5 M50 , P10 M50 and P15 M50 .
Fig. 10 – The contours of density (a) and velocity magnitude (b) on plane P1 in P5 M50 , P10 M50 and P15 M50 . product of these reactions, CO (see Fig. 7(a)), has a higher selectivity in P5 M50 than in P10 M50 . In the monolithic catalyst bed of P15 M50 , the mass fraction of O2 is lowest (see Fig. 5(b)), so the rate of first step (see Fig. 8(a)) is small, and the CO2 selectivity is little. Meanwhile, the mass fraction of CO2 in this region of P15 M50 is high. These dual reasons promote the CO2 consumption reaction (the tenth step, see Fig. 9(b)). Therefore, the CO mass fraction increases and CO2 mass fraction decreases along the gas flow direction. In brief, the mass fractions of CO and CO2 are dominated by the first, third, sixth and tenth reaction steps with rate
variation when the height of particle catalyst bed changes in P5 M50 , P10 M50 and P15 M50 . This results in that the mass fraction of CO in monolithic catalyst bed of P10 M50 is lower than P5 M50 and P15 M50 , and the mass fraction of CO2 in this bed of P15 M50 is the highest. The longer residence time lead to higher C2 yield in P10 M50 than P5 M50 , but the blocking effect of CO2 on the catalyst in P15 M50 reduces C2 yield. Consequently, the height of particle catalyst bed at 10 mm is the best value for C2 yield. Because OCM reaction belongs to volumetric increase reaction according to the previous work (Zhang et al., 2015a,b),
Fig. 11 – The contours of CH4 (a) and O2 (b) mass fractions on plane P1 in P10 M25 , P10 M50 and P10 M75 .
chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 390–399
397
Fig. 12 – The contours of C2 H6 (a) and C2 H4 (b) mass fractions on plane P1 in P10 M25 , P10 M50 and P10 M75 .
Fig. 13 – The contours of CO (a) and CO2 (b) mass fractions on plane P1 in P10 M25 , P10 M50 and P10 M75 . the density of reaction gas decreases and the velocity magnitude increases along the flow direction. As can be seen from Fig. 10(a), the density decreases from 0.227 to 0.225, 0.222 and 0.220 at the outlet of particle catalyst bed and to 0.221, 0.220 and 0.219 at the outlet of monolithic catalyst bed in P5 M50 , P10 M50 and P15 M50 , respectively. As shown in Fig. 10(b), the velocity magnitude increases at the outlet of particle catalyst bed with its height increasing. In the monolithic catalyst bed, a high velocity region exists in the outer edge of the boundary layer. This is caused by the combined effect of boundary layer, channels in monolithic catalyst and OCM reaction (Zhang
et al., 2015a). However, this high velocity region does not exist in particle catalyst bed, and its area becomes lager with the increase of particle catalyst bed height because of the smaller density.
3.4.
Effect of monolithic catalyst bed height
Figs. 11–13 show the effect of monolithic catalyst bed height (from 25 mm to 75 mm) on the mass fractions of CH4 and O2 , C2 H6 and C2 H4 , and CO and CO2 , when the height of the particle catalyst bed is fixed at 10 mm. The mass fraction contour
Fig. 14 – The contours of density (a) and velocity magnitude (b) on plane P1 in P5 M50 , P10 M50 and P15 M50 .
398
chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 390–399
of each species in particle catalyst bed of P10 M25 , P10 M50 and P10 M75 is completely same as the packed bed reactor with 10 mm height particle catalyst (Zhang et al., 2015b). The reactant mass fractions decrease and the product mass fractions increase along the flow direction. In the monolithic catalyst bed, the reactants are less and less and the products are more and more at the bed outlet when the height of monolithic bed changes from 25 mm to 75 mm. Moreover, the mass fraction contours in P10 M25 is a part of P10 M50 , just as P10 M50 is a part of P10 M75 . We can also see from Figs. 11–13 that the mass fraction contours of species in particle catalyst bed are denser than in monolithic catalyst bed. This phenomenon tells us that the OCM reaction in particle catalyst bed is fiercer than in monolithic catalyst bed again. Especially at the end of monolithic catalyst bed about 25 mm height in P10 M75 , the OCM reaction closes to stagnation. Fig. 14 shows the effect of the height of the monolithic catalyst bed from 25 mm to 75 mm on the gas density and velocity magnitude when the particle catalyst bed fixed at 50 mm. The density and velocity magnitude contour in particle catalyst bed of each reactor is same as the packed bed reactor with 10 mm height particle catalyst reported in previous work. In the monolithic catalyst bed, the density becomes smaller and the velocity magnitude becomes larger when this bed height becomes higher and residence time becomes longer.
4.
Conclusions
A three-dimensional numerical model was established to simulate the OCM tubular two-stage reactors with FLUENT software in the condition of that the volume rate was 80 ml/min under standard state with a CH4 /O2 ratio of 3 and the operating temperature and pressure were 800 ◦ C and 1 atm, respectively. The effect of catalyst bed heights on reactor performance was investigated using these models. The results show that the simulated values of the CH4 conversion, products selectivity and other parameters in the outlet of reactors are consistent with the experimental values in acceptable errors. It showed that two-stage fixed bed reactor had combined advantages of particle catalyst bed and monolithic catalyst bed, such as high conversion of CH4 , high selectivity of C2 , small O2 consumption and low pressure drop, so the reactor has potential of depressing hot spot. We also can find that the OCM reaction was fiercer and the effect of boundary layer on OCM reaction was weaker in particle catalyst bed than in monolithic catalyst bed of two-stage reactors. The effects of residence time caused by bed height and the blocking caused by CO2 in the oxidation reaction induce the P10 M50 reactor having the best performance for C2 yield when the height of particle catalyst bed changed from 5 mm to 15 mm. When the height particle catalyst bed was fixed at 10 mm and the height monolithic catalyst bed increased from 25 mm to 75 mm, the mass fraction contours in P10 M25 was a part of P10 M50 , just as P10 M50 was a part of P10 M75 .
References Batchelor, G.K., 2000. An Introduction to Fluid Dynamics, second ed. Cambridge University Press, Cambridge. Fang, X., Li, S., Lin, J., Chu, Y., 1992. Oxidative coupling of methane on W–Mn catalysts. J. Mol. Catal. China 6, 427–433. Gesser, H.D., Hunter, N.R., 1998. A review of C-1 conversion chemistry. Catal. Today 42, 183–189.
Groppi, G., Ibashi, W., Tronconi, E., Forzatti, P., 2001. Structured reactors for kinetic measurements in catalytic combustion. Chem. Eng. J. (Lausanne) 82, 57–71. Ji, S., Wang, W., 2012. Particle/cordierite monolithic catalysts dual-bed reactor with beds-interspace supplementary oxygen and performance for oxidative coupling of methane. Int. J. Chem. React. Eng. 10, 1–13. Ji, S., Xiao, T., Li, S., Chou, L., Zhang, B., Xu, C., Hou, R., York, A.P.E., Green, M.L.H., 2003. Surface WO4 tetrahedron: the essence of the oxidative coupling of methane over M–W–Mn/SiO2 catalysts. J. Catal. 220, 47–56. Jiang, Z., Fang, X., Yu, C., Li, S., 1992. Mn–Na2 WO4 /SiO2 catalyst for oxidative coupling of methane-I, surface dispersion state and oxide support interaction. J. Mol. Catal. China 6, 477–480. Khammona, K., Assabumrungrat, S., Wiyaratn, W., 2012. Reviews on coupling of methane over catalysts for application in C2 hydrocarbon production. J. Eng. Appl. Sci. 7, 447–455. ˇ P., Kubícek, ˇ Kocí, M., Marek, M., 2004. Modeling of three-way-catalyst monolith converters with microkinetics and diffusion in the washcoat. Ind. Eng. Chem. Res. 43, 4503–4510. Liu, H., Wang, X., Yang, D., Gao, R., Wang, Z., Yang, J., 2008a. Scale up and stability test for oxidative coupling of methane over Na2 WO4 –Mn/SiO2 catalyst in a 200 ml fixed-bed reactor. J. Nat. Gas Chem. 17, 59–63. Liu, H., Yang, D., Gao, R., Chen, L., Zhang, S., Wang, X., 2008b. A novel Na2 WO4 –Mn/SiC monolithic foam catalyst with improved thermal properties for the oxidative coupling of methane. Catal. Commun. 9, 1302–1306. Lunsford, J.H., 2000. Catalytic conversion of methane to more useful chemicals and fuels: a challenge for the 21st century. Catal. Today 63, 165–174. Malekzadeh, A., Khodadadi, A., Dalai, A.K., Abedini, M., 2007. Oxidative coupling of methane over lithium doped (Mn + W)/SiO2 catalysts. J. Nat. Gas Chem. 16, 121–129. Nakisa, Y., Reza, G.M.H., 2008. Oxidative coupling of methane in a fixed bed reactor over perovskite catalyst: a simulation study using experimental kinetic model. J. Nat. Gas Chem. 17, 8–16. Nakisa, Y., Reza, G.M.H., 2009. Modeling the oxidative coupling of methane: heterogeneous chemistry coupled with 3D flow field simulation. J. Nat. Gas Chem. 18, 39–44. Pak, S., Lunsford, J.H., 1998. Thermal effects during the oxidative coupling of methane over Mn/Na2 WO4 /SiO2 and Mn/Na2 WO4 /MgO catalysts. Appl. Catal. A 168, 131–137. Pan, D., Ji, S., Wang, W., Li, C., 2010. Oxidative coupling of methane in a dual-bed reactor comprising of particle/cordierite monolithic catalysts. J. Nat. Gas Chem. 19, 600–604. Schweer, D., Meeczko, L., Baerns, M., 1994. OCM in a fixed-bed reactor: limits and perspectives. Catal. Today 21, 357–369. Stansch, Z., Mleczko, L., Baerns, M., 1997. Comprehensive kinetics of oxidative coupling of methane over the La2 O3 /CaO catalyst. Ind. Eng. Chem. Res. 36, 2568–2579. ˇ ´ T., Solcová, Stary, O., Schneider, P., Marek, M., 2006. Effective diffusivities and pore-transport characteristics of washcoated ceramic monolith for automotive catalytic converter. Chem. Eng. Sci. 61, 5934–5943. Talebizadeh, A., Mortazavi, Y., Khodadadi, A.A., 2009. Comparative study of the two-zone fluidized-bed reactor and the fluidized-bed reactor for oxidative coupling of methane over Mn/Na2 WO4 /SiO2 catalyst. Fuel Process. Technol. 90, 1319–1325. Tang, J., Ji, S., Wang, K., Li, C., 2009. Preparation of M–W–Mn/SiO2 /cordierite monolithic catalysts and their performances for oxidative coupling of methane. Nat. Gas Chem. Ind. (China) 34, 19–26. Taniewski, M., Lachowicz, A., Skutil, K., 1997. Effective utilisation of the catalyst bed deactivating in the course of oxidative coupling of methane. Chem. Eng. Sci. 52, 935–939. Wang, W., Ji, S., Pan, D., Li, C., 2011. A novel particle/monolithic two-stage catalyst bed reactor and their catalytic performance for oxidative coupling of methane. Fuel Process. Technol. 92, 541–546.
chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 390–399
Wehinger, G.D., Eppinger, T., Kraume, M., 2015. Detailed numerical simulations of catalytic fixed-bed reactors: heterogeneous dry reforming of methane. Chem. Eng. Sci. 122, 197–209. Wu, J., Zhang, H., Qin, S., Hu, C., 2007. La-promoted Na2 WO4 /Mn/SiO2 catalysts for the oxidative conversion of methane simultaneously to ethylene and carbon monoxide. Appl. Catal. A 323, 126–134.
399
Zhang, Z., Guo, Z., Ji, S., 2015a. Numerical simulation of monolithic catalyst fixed bed reactor for oxidative coupling of methane. Chin. J. Chem. Eng. Zhang, Z., Guo, Z., Ji, S., 2015b. Numerical simulation of packed bed reactor for oxidative coupling of methane. J. Energy Chem. 24, 23–30.