Escolar Documentos
Profissional Documentos
Cultura Documentos
H I G H L I G H T S
art ic l e i nf o
a b s t r a c t
Article history:
Received 23 May 2014
Received in revised form
2 September 2014
Accepted 4 September 2014
Available online 7 October 2014
Highly endothermic (or exothermic) heterogeneous catalytic reactions are performed commonly in
xed-bed reactors with small tube-to-particle-diameter ratios N both in industrial and lab-scale
applications. For these reactor congurations conventional plug ow models and pseudohomogeneous kinetic models fail. An adequate modeling can be carried out with full computational
uid dynamics (CFD) in combination with detailed reaction mechanisms. In this study, a full threedimensional xed-bed reactor for the catalytic dry reforming of methane (DRM) over rhodium was
simulated with a detailed reaction mechanism. The bed consists of 113 spherical solid particles in which
thermal conductivity was considered. Two different Reynolds numbers were investigated, i.e., 35 and
700. The simulated DRM xed-bed reactor demonstrates the strong interaction between chemical
kinetics and transport of momentum, heat and mass. The observed velocity, temperature and species
elds are characterized by their three-dimensional behavior and interactions highlighting their
complexity and discrepancy from lumped model predictions. In addition, the reaction mechanism
determines regions with catalyst deactivation by carbon deposition. This study demonstrates the
advantages of modeling heterogeneous catalytic xed-bed reactors with small N fully in threedimensional in combination with detailed reaction mechanisms. Finally, this modeling approach reduces
dependencies on empiricism for the calculation of multiscale reaction devices.
& 2014 Elsevier Ltd. All rights reserved.
Keywords:
CFD
Catalysis
Dry reforming of methane
Fixed-bed reactors
Modeling
1. Introduction
The atmospheric concentration of greenhouse gases, i.e.,
carbon dioxide (CO2), nitrous oxide (NO), methane (CH4) and
chlorouorocarbons (CFCs), has increased dramatically during
the last decades (Hartmann et al., 2013). These anthropogenic
emissions have risen a global concern over the current technological practices. Hence, the eld of interest involves CH4 and CO2
disposal, utilization and removal, as well as the effect of these
gases in the atmosphere (Mikkelsen et al., 2010; Centi and
Perathoner, 2009; Hunt et al., 2010; Papadopoulou et al., 2012;
n
http://dx.doi.org/10.1016/j.ces.2014.09.007
0009-2509/& 2014 Elsevier Ltd. All rights reserved.
H 260 kJ=mol
198
0
xi
t
g i
t
xi xj
xj
0
t
xj
xj
for
i 1; ; Ng
where DM
i is the effective diffusivity between species i and the
remaining mixture. Xi represents the molar fraction of species i. Mi
is the molecular weight of species i and T is the temperature. The
binary diffusion coefcients Di are obtained through polynomial
ts. The molar fraction Xi can be written as
Xi
1
N
Yi
Y
j g 1 Mjj M i
vj jk
Sh
t
t
xj
xj
xk
xj
h Y i hi T
i1
as a function of temperature hi hi T.
Ideal gas was assumed connecting pressure, temperature and
density to close the governing equations:
p
RT
N
i g 1 X i M i
11
199
Ng Ns
k1
j1
s_ i ik kf k
cj jk
17
i c i i 1
18
i can be written as
Under steady-state conditions the left side of Eq. (19) will be zero.
The reaction rate expression can be modied by the concentration,
or coverage, of some surface species:
i
i k i
E ak N s
i k exp
20
kf k Ak T k exp
RT
RT
i1
Rep
vin dp
12
vi vj v0i v0j ij g i
13
xj
xi xj
t
The Reynolds stresses v0i v0j have to be put in terms of the
averaged ow quantities to close the system of equations. In our
case we used the realizable k model, developed by Shih et al.
(1995), in combination with a two-layer all-y wall treatment
driven by shear (Wolfshtein, 1969), cf. manual of STAR-CCM (CDadapco, 2014).
2.1.3. CFD and heterogeneous chemical reactions
The chemical reactions at the catalytic surface are coupled via
boundary conditions with the species distribution equation (5).
Under steady-state conditions gas-phase molecules of species
i, which are produced/consumed at the catalytic surface by
desorption/adsorption, have to diffuse from/to the catalyst
(Deutschmann, 2008):
! !
n ji Rhet
14
i
!
with the outward-pointing unit vector normal to the surface n
!
and the diffusion mass ux ji as in Eq. (6). The heterogeneous
reaction term Rhet
is given by
i
F cat=geo M i s_ i
Rhet
i
15
16
s_ i i
19
with two extra coverage parameters, ik and ik. The term including ik indicates the modication of the surface rate expression
proportional to any arbitrary power of a surface species concentration. ik represents a modication of the activation energy as a
function of coverage.
The occurrence of adsorption reactions results in a modication
of the conventional rate coefcient by referencing sticking coefcients Si
s
S0
RT
ads
kf k i
21
2 M i
with S0i as the initial (uncovered surface) sticking coefcient and
Nj s 1 0jk is the sum of all the surface reactant's stoichiometric
coefcients, cf. Kee et al. (2003) and Deutschmann (2008).
Additionally, the operator splitting algorithm was implemented. The algorithm decouples the general species transport equation, due to the different time scales of the ow eld and the
chemical reactions. The time integration of the chemical state
(species mass fractions and enthalpy) was performed in two steps.
For the rst step, only the species source terms were taken into
account for the integration of a time interval. For the second, the
ow eld was integrated over a time interval without the chemical
source terms, cf. Ren and Pope (2008).
All simulations were realized with the simulation software
STAR-CCM version 9.02.005 by CD-adapco (2014). The equation
system for the surface species was solved by DARS, an add-in
solver for chemical reactions for STAR-CCM . The computational
time was high due to the mesh size and chemical reaction steps.
The laminar case with a 3.2 million cell mesh (M3) ran for 35,000
iterations which yielded in a total CPU time of 1:7 107 s or
196 days on a Intel Xeon 3.07 GHz CPU. The turbulent case with a
3.6 million cell mesh (M5) ran for 9000 iterations, i.e., 9:3 106 s
or 107 days. However, the simulations were performed on several
parallel CPUs, which reduced the computational time signicantly.
2.2. Detailed reaction mechanism
As already mentioned detailed uid dynamics call for detailed
reaction mechanisms. Still today, in many CFD simulations LangmuirHinshelwoodHougonWatson (LHHW) models are applied.
200
Table 1
Detailed surface mechanism for the dry reforming of methane, from McGuire et al.
(2011).
No.
Reaction
H2
- Hn Hn
O2
- On On
CH4
- CHn4
H2O
- H2On
CO2
- COn2
CO
Hn Hn - H2
On On - O2
- COn
n
A (a)
E (kJ/mol)
1:0 10 2 b
0.0
1:0 10 2 b
0.0
8:0 10 3 b
0.0
1:0 10 3 b
0.0
4:8 10 2 b
0.0
5:0 10 1 b
0.0
3:0 1021
77.8
1:3 1022
355.2
280c
nO
9
H2On - H2O
10
COn - CO
3:0 10
13
3:5 1013
133.4
4:1 1011
18.0
14
25.1
5:0 1022
83.7
20
37.7
3:0 1020
33.5
5:0 1022
106.4
3:0 1021
100.8
3:0 1021
171.8
5:2 1023
97.9
2:5 1021
169.0
5:5 1018
121.6
3:0 1021
171.8
5:0 1019
108.9
3:7 1021
0.0
15c
CO
11
COn2 - CO2
12
CHn4 - CH4
13
Hn On - OHn
14
OHn
15
Hn OHn - H2On
16
H2On
17
18
19
Cn O - COn
20
COn
21
COn On - COn2
22
COn2 n- COn On
23
COn Hn - HCOn
24
HCOn
nCO
45.0
1:9 10
- Hn On
- Hn OHn
- C n On
n
- COn Hn
3:0 10
50c
n
25
HCO
26
CHn On - HCOn
27
CHn4
28
CHn3 Hn - CHn4
29
CHn3
30
CHn2 Hn - CHn3
31
CHn2
32
CHn Hn - CHn2
33
CHn
34
Cn Hn - CHn
35
36
37
38
39
40
41
CHn On - Cn OHn
42
Cn OHn - CHn On
- CH O
n
n
- CHn3 Hn
n
- CHn2 Hn
n
- CHn Hn
n
- Cn Hn
n
3:7 10
24
59.5
3:7 1021
167.5
3:7 1021
61.0
3:7 1021
51.0
3:7 1024
103.0
3:7 1023
44.1
3:7 1024
100.0
3:7 1021
68.0
3:7 1021
21.0
3:7 1021
172.8
1:7 1024
80.34
21
24.27
3:7 10
3:7 1024
120.31
3:7 1021
15.06
3:7 1024
114.5
3:7 1021
36.82
3:7 1021
30.13
3:7 1021
136.0
However, the fundamental mechanism described by LHHW models can be elusive and the physical signicance of parameters can
be seriously questionable, cf. Salciccioli et al. (2011) and Wehinger
et al. (2014). Taking the strong interactions between the different
transport quantities into account an erroneous kinetics will
directly lead to misleading predictions. Therefore, it is recommended only to use reliable kinetics for detailed uid dynamics
simulations.
In this study a detailed reaction mechanism, see Table 1, was
implemented published by McGuire et al. (2011) for the DRM over
Inlet
Upstream
h u = 1/4 H
Flow direction
Spheres
d p = 4.09 mm
Fixed-bed
H = 40 mm
Reactor diameter
D = 16.2 mm
Downstream
h d = 1/2 H
Outlet
Fig. 1. Scheme of the spherical xed-bed reactor.
201
Fig. 2. Section of xed-bed reactor meshing (A) mesh M3 for Rep 35 and (B) mesh
M5 for Rep 700. Gas phase polyhedral mesh in gray and solid sphere polyhedral
mesh in dark. Flattening is visible between the particles.
Table 2
Characteristics for the investigated meshes.
Mesh Prism layer
thickness (mm)
No. of
layers
Base size
(mm)
Total no. of
cells [106]
Laminar Re 35
M1
0.19
M2
0.19
M3
0.19
2
2
2
8
4
3
0.6
2.6
3.2
0.11
0.65
0.67
Turbulent Re 700
M4
0.0427
M5
0.0427
M6
0.0427
2
2
2
4
3
2
2.8
3.5
10.5
0.66
0.67
3.38
No. solid
cells [106]
BL
dp
22
In addition, as recommended by Dixon et al. (2013), the dimensionless cellwall distance y was kept to approx. 0.52.0. For the
investigated cases the velocity boundary layer (BL) thickness was
smaller than the temperature BL and the concentration BL. The
Prandtl number Pr was of the same order of magnitude as the
Schmidt number Sc. The inuence of the mesh size was investigated
by means of mesh renement, see Table 2. The base size is a
characteristic dimension of the mesh model, to which all other mesh
dimensions refer. It can be interpreted as a scaling factor of the mesh
resulting in the total number of cells (CD-adapco, 2014).
2.4. Boundary conditions
The conditions at the inlet were the following: feed gas
composition xin;CO2 =xin;CH4 =xin;N2 0:20=0:10=0:70, inlet velocity
vin;1 0:886 m=s or vin;2 17:72 m=s. The corresponding particle
Reynolds numbers were Rep 35; 700 calculated with a mean
dynamic viscosity gas;973 K 9:504 10 5 m2 =s and the particle
if z 4 0:637
23
0.8
0.6
0.4
0.2
0
0.5
1.5
= (R-r)/dp (-)
Fig. 3. Comparison of porosity as a function of dimensionless wall distance
R r=dp between computer-generated packing and experimental measurements from Mueller (1992) averaged over the reactor height and a general equation
from de Klerk (2003).
202
!
150 1 2
1
H
1:75
v2in
3
Rep 3
d
p
24
10000
Eisfeld Eq.
Ergun Eq.
Simulation w/o reactions
Simulation w reactions
1000
T = 973 K
T = 873 K
100
10
10
100
1000
Particle Reynolds number ReP = vindP/
154 A2w 1 2 Aw 1
Rep
Bw 3
3
!
H
v2in
dp
25
2
3 D=dp 1
dp
Bw 1:15
D
26
#2
2
0:87
27
Fig. 4. Pressure drop over particle Reynolds number. Comparison between simulation (meshes M3 and M5) and Eqs. (24) and (25), respectively.
!
Fig. 5 shows the specic velocity distribution j v j=vin on a plane
cut through the xed bed for Rep 35 M3 and Rep 700 M5.
Table 3
Results of investigated meshes for laminar and turbulent cases.
Mesh No.
Laminar Re 35
M1
M2
M3
pw=o;chem: (Pa)
34.7
37.9
37.9
Turbulent Re 700
M4
4218
M5
4220
M6
4400
pw;chem: (Pa)
33.2
36.5
36.5
4120
4124
4300
y ()
X CH4 (%)
X CO2 (%)
YCO (%)
YH2 (%)
Y H2O (%)
COn ()
C n ()
32.7
35.2
35.4
19.8
21.9
22.1
28.1
29.7
29.9
27.5
28.7
28.8
1.5
2.2
2.2
0.512
0.534
0.535
0.098
0.058
0.056
0.59
0.59
0.65
11.1
10.9
11.6
8.4
8.2
8.9
5.0
5.0
5.0
5.7
5.7
5.7
0.3
0.3
0.3
0.346
0.346
0.343
0.009
0.009
0.008
203
Fig. 6. Backow regions, i.e., cells with negative velocities. (A) for Rep 35 and
(B) for Rep 700.
2.5
!
Fig. 5. Specic velocity distribution j v j=vin on a plane cut through the xed bed.
(A) for Rep 35 M3 and (B) for Rep 700 M5.
2
(vz ())/vin (-)
Rep = 35
Rep = 700
1.5
1
0.5
0
0.5
1.5
= ( R-r)/dp (-)
Fig. 7. Articial axial specic velocity vz =vin as a function of dimensionless wall
distance for different Reynolds numbers.
204
Position
1
Position
2
Position
3
Fig. 8. Velocity magnitude contours for Rep 35 and positions for the mesh
validation.
whereas it shows a steep rise for the turbulent case. In Fig. 9(a) the
calculated velocities from mesh 1 are slightly different than for
mesh M2 and M3. On the contrary, the meshes M2 and M3 show
almost identical results in the laminar case. For higher Reynolds
numbers, only at position 3 the meshes show different velocities
with mesh renement. Here, the minimum velocity decreases in
the center with a ner mesh.
Finally, Fig. 11 shows the frequency of y values for meshes
M4M6. As it can be seen, most of the cells are small enough that
y o1:5. Hence, velocity boundary layers are well resolved.
3.4. Temperature distribution
The temperature distribution in the xed bed, i.e., gas phase and
solid particles, is shown in Fig. 12. The inlet temperature and the wall
temperature are set constant to 973 K. Due to the endothermic
reactions the temperature inside the bed decreases. Again, strong
axial and radial temperature differences up to approx. 80 K occur.
Low Reynolds numbers result in large residence times. Hence, in
Fig. 12(A) the overall temperature is lower than in Fig. 12(B). In (A), a
cold spot appears after approx. half of the reactor length, whereas in
(B) the temperature in the center decreases constantly. This is due to
the shorter residence time which moves the cold spot out of the bed.
In Fig. 12(B) the hot ow reaches deep inside the xed bed, whereas
for Rep 35 the ow cools down immediately. The solid particles can
be detected easily due to their almost constant temperature, which is
caused by the high thermal conductivity. As a result of channeling
in the near wall region, the thermal penetration into the bed is
declined. Again, the transport property temperature shows highly
three-dimensional behavior.
In Figs. 9 and 10(b), (f) and (j), temperature proles are shown
for the different meshes at the three positions. At position 1 the
temperature decreases from the inlet to the sphere's surface due to
endothermic reactions. However, for Rep 35 it is lower than for
the turbulent case. At position 2, i.e., between wall and sphere, the
temperature decreases from the constant wall temperature T 973
K to the specic surface temperature, which is inuenced rstly by
the surrounding ow and secondly by the surface reactions. In the
laminar case the temperature decreases almost linearly from the
wall to the surface. For the higher Reynolds number boundary
layers can be noticed near both surfaces, which are of the order of
magnitude of the velocity boundary layers. Finally, between the
spheres at position 3 (j) the temperature eld is highly inuenced
by the ow eld. In the laminar case the temperature decreases
from the outside to the center. Therefore, the left surface in (j) is
cooler than the right side. Again, an almost linear prole is shown
although the recirculation zone brings cooler gas. For Rep 700
the endothermic reactions cool down the surface, whereas hot gas
passes this position. The recirculation zone is larger than for lower
Reynolds. Hence, the temperature decreases strongly in the center.
3.5. Surface adsorbed species
As mentioned before, catalyst deactivation through carbon
deposition is one of the major draw backs of the DRM. It has to
be noticed that in reality coke formation takes place including
many carbonaceous atomic layers. However, the present reaction
mechanism only accounts for monolayer carbon on the surface.
Consequently, the model determines the regions where coking
takes place rather than the amount of coke. In Fig. 13 surface site
fractions of adsorbed carbon and some streamlines are illustrated
for different Reynolds numbers. As it can be seen, the carbon
deposition is not only dependent on the Reynolds number but is
due to the interactions between velocity, temperature and gas
composition. In Fig. 13(A) several regions of spheres are totally
blocked by carbon mainly in the center of the inlet region of the
bed. Hence, the catalyst is deactivated resulting in declined or
stopped production of syngas. In Fig. 13(B) almost no carbon is
adsorbed. Catalyst deactivation by carbon deposition for DRM
especially in the inlet regions of xed beds was recently observed
experimentally and numerically by Kahle et al. (2013). Fig. 13
highlights the advantage of this type of reaction mechanism for
DRM that can contribute to identify conditions and regions where
deactivation of the catalyst is likely to occur. In addition, Fig. 14
shows radially and circumferentially averaged surface site fractions of the adsorbed species Cn, COn, Hn and RHn. For the laminar
case (A), surface adsorbed carbon monoxide (COn) becomes the
most abundant reaction intermediate (MARI) after approx. 10 mm
in the xed-bed. Adsorbed carbon is only dominant in the
entrance of the reactor, whereas Hn occurs on less than 1% of the
surface sites. For the turbulent case (B), COn is again the MARI. Due
to the lower residence time, its surface fraction is lower, too. Cn
and Hn are found on less 2% on the surface. These two gures
illustrate that the DRM is kinetically limited. However, it has to be
kept in mind that the two cases are not under iso-conversion.
Therefore, a true comparison of location and quantity of surface
adsorbed species cannot be undertaken.
970
0.4
0.2
950
940
920
1
980
970
Temperature [K]
7
6
5
4
3
0.5
1.5
0
0
CO2
940
930
Wall
0.1
CH4
0.05
0.5
1.5
0.5
1.5
0.2
905
0.15
1.5
1.5
0.04
0.1
CH4
0.03
0.02
0.01
0
0.5
0.5
0.05
CO2
0.05
Temperature [K]
900
890
1.5
0.02
910
895
0.03
0.01
0.5
0.04
0.05
0.15
950
900
Wall
0.2
960
0.02
0.01
Wall
0.03
910
1
0
0.1
920
2
Wall
CH4
0
0
0.15
0.05
930
960
0.04
Mole fraction CO [-]
0.6
0.2
0.8
0.05
CO2
0.25
980
Temperature [K]
1.2
205
0
0
0.5
1.5
0.5
1.5
Fig. 9. Results of mesh renement for laminar case Rep 35. Specic velocity for (A) position 1, (E) position 2, and (I) position 3. Temperature for (B) position 1, (F) position 2,
(J) position 3. Mole fractions CO2 and CH4 for (C) position 1, (G) position 2, and (K) position 3. Mole fractions CO for (D) position 1, (H) position 2, and (L) position 3.
980
0.25
0.012
CO2
0.6
0.4
970
965
960
955
950
0.2
0.01
0.2
Mole fraction [-]
Temperature [K]
975
0.8
0.15
CH4
0.1
206
0.05
940
1
0
0
3
2
960
950
0.5
1.5
0.1
CH4
930
0.006
0.004
Wall
0.5
1.5
0.5
1.5
950
0.2
945
0.15
0.5
1.5
0.02
3
2
Temperature [K]
940
935
CO2
5
Specific velocity v/vin [-]
0.008
Wall
Wall
Wall
0.002
1
0
0.01
CO2
0.05
940
0.012
0.15
Mole fraction [-]
Temperature [K]
970
0.2
6
5
980
0.004
0
0
0.006
0.002
945
0
0.008
0.1
CH4
0.05
0.015
0.01
0.005
1
0
0.5
1.5
930
0.5
1.5
0.5
1.5
0.5
1.5
Fig. 10. Results of mesh renement for turbulent case Rep 700. Specic velocity for (A) position 1, (E) position 2, and (I) position 3. Temperature for (B) position 1,
(F) position 2, and (J) position 3. Mole fractions CO2 and CH4 for (C) position 1, (G) position 2, and (K) position 3. Mole fractions CO for (D) position 1, (H) position 2, and
(L) position 3.
4. Conclusion
Highly endothermic (or exothermic) heterogeneous catalytic
reactions are performed mostly in xed-bed reactors with small
tube-to-particle-diameter ratios N. Inhomogeneities in the bed
structure are dominant especially for small N. This results in
signicant wall effects, local back ows and large axial and radial
three-dimensional behavior and interactions highlighting their complexity and discrepancy from lumped model predictions. Additionally, the reaction mechanism can detect regions where coking takes
place with the help of surface adsorbed carbon. We recommend
meshes with most of the near wall cells being small enough that
y o 1:5. This could be achieved by using two prism layers with a
total thickness calculated by Eq. (22). Meshes with approx. 3 million
total cells show grid independent results for laminar ows. However,
turbulent ows need ner meshes.
1
0.03
0.01
CO*
0.6
0.02
0.01
0
Rep = 35
H*
0.8
0.5
1.5
2.5
3.5
4.5
y+ [-]
Fig. 11. Frequency distribution over dimensionless wall distance y for different
meshes.
0.4
0.005
Rh*
0.2
C*
1
Rep = 700
0.8
0.01
H*
0.6
Rh*
0.4
0.04
Frequency [-]
0.05
207
0.005
CO*
0.2
C*
0.01
0.02
0
0.04
0.03
0.25
Rep = 35
960
CO2
0.15
0.1
940
Temp.
CH4
0.25
900
H2
H2O
880
Rep = 700
980
960
CO2
0.15
0.1
920
CO
0.05
0.2
980
Temp.
CH4
Temperature [K]
0.2
940
920
0.05
CO
H2O
0
-0.01 0 0.01
0.03
0.05
900
H2
880
Fig. 15. Mean mole fractions and mean temperature over reactor length. (A) for
Rep 35 mesh M3 and (B) for Rep 700 mesh M5.
Fig. 13. Catalyst deactivation through carbon deposition on the surface. (A) for Rep 35 mesh M3 and (B) for Rep 700 mesh M5.
208
Fig. 16. Hydrogen production and surface adsorbed carbon on a plane cut through the xed bed. (A) for Rep 35 mesh M3 and (B) for Rep 700 mesh M5.
Pr
r
R
Ri
Rep
s_
Sc
Si
Sh
t
T
vin
vi
v0i
xi
Xi
y
Yi
Prandtl number Pr cp = ()
radial coordinate (m)
ideal gas constant (J/K mol)
production rate of species i (kg/m3 s)
particle Reynolds number Rep vin dp = ()
molar net production rate (mol/m3 s)
Schmidt number Sc =D ()
sticking coefcient ()
heat source (W/m3)
time (s)
temperature (K)
supercial velocity (m/s)
mean velocity components
uctuating velocity components
coordinate in i direction (m)
molar fraction of species i()
dimensionless distance from wall ()
mass fraction of species i ()
Greek letters
Nomenclature
Latin letters
a
Ak
c
cp
dp
D
Di
Ea
F cat=geo
h
H
ji
jq
k
k
m
Mi
N
Ng
p
BL
ij
Acknowledgments
This study is part of the Cluster of Excellence Unifying
Concepts in Catalysis (Unicat) (Exc 314), which is coordinated
by the Technische Universitt Berlin. The authors would like to
209
Hartmann, M., Maier, L., Minh, H., Deutschmann, O., 2010. Catalytic partial
oxidation of iso-octane over rhodium catalysts. Combust. Flame: Exp. Model.
Simul. Study 157, 17711782.
Hickman, D.A., Schmidt, L.D., 1993. Steps in CH4 oxidation on Pt and Rh surfaces:
high-temperature reactor simulations. AIChE J. 39, 11641177.
Horn, R., Korup, O., Geske, M., Zavyalova, U., Oprea, I., Schlgl, R., 2010. Reactor for
in situ measurements of spatially resolved kinetic data in heterogeneous
catalysis. Rev. Sci. Instrum. 81, 064102.
Hunt, A.J., Sin, E.H.K., Marriott, R., Clark, J.H., 2010. Generation, capture, and
utilization of industrial carbon dioxide. ChemSusChem 3, 306322.
Kahle, L.C.S., Roussire, T., Maier, L., Herrera Delgado, K., Wasserschaff, G., Schunk, S.
A., Deutschmann, O., 2013. Methane dry reforming at high temperature and
elevated pressure: impact of gas-phase reactions. Ind. Eng. Chem. Res. 52,
1192011930.
Kee, R.J., Colin, M.E., Glarborg, P., 2003. Chemically Reacting Flow, Theory and
Practice. Wiley, Hoboken, NJ.
de Klerk, A., 2003. Voidage variation in packed beds at small column to particle
diameter ratio. AIChE J. 49, 20222029.
Korup, O., Mavlyankariev, S., Geske, M., Goldsmith, C.F., Horn, R., 2011. Measurement and analysis of spatial reactor proles in high temperature catalysis
research. Chem. Eng. Process.: Process Intensif. 50, 9981009.
Kuroki, M., Ookawara, S., Ogawa, K., 2009. A high-delity CFD model of methane
steam reforming in a packed bed reactor. J. Chem. Eng. Jpn. 42, s73s78.
Leva, M., Weintraub, M., Grummer, M., Pollchik, M., Storch, H.H., 1951. Fluid ow
through packed and uidized systems. US Bur. Mines Bull. 504.
Maestri, M., Vlachos, D.G., Beretta, A., Groppi, G., Tronconi, E., 2008. Steam and dry
reforming of methane on Rh: microkinetic analysis and hierarchy of kinetic
models. J. Catal. 259, 211222.
Maestri, M., Vlachos, D.G., Beretta, A., Groppi, G., Tronconi, E., 2009. A C1
microkinetic model for methane conversion to syngas on Rh/Al2O3. AIChE J.
55, 9931008.
McGuire, N.E., Sullivan, N.P., Deutschmann, O., Zhu, H., Kee, R.J., 2011. Dry reforming
of methane in a stagnation-ow reactor using Rh supported on strontiumsubstituted hexaaluminate. Appl. Catal. A: Gen. 394, 257265.
Mhadeshwar, A.B., Vlachos, D.G., 2005. Hierarchical multiscale mechanism development for methane partial oxidation and reforming and for thermal decomposition of oxygenates on Rh. J. Phys. Chem. B 109, 1681916835.
Mikkelsen, M., Jorgensen, M., Krebs, F.C., 2010. The teraton challenge. A review of
xation and transformation of carbon dioxide. Energy Environ. Sci. 3, 4381.
Mueller, G.E., 1992. Radial void fraction distributions in randomly packed xed
beds of uniformly sized spheres in cylindrical containers. Powder Technol. 72,
269275.
Ookawara, S., Kuroki, M., Street, D., Ogawa, K., 2007. High-delity DEM-CFD
modeling of packed bed reactors for process intensication. In: Proceedings
of European Congress of Chemical Engineering (ECCE-6), Copenhagen.
Papadopoulou, C., Matralis, H., Verykios, X., 2012. Utilization of biogas as a
renewable carbon source: dry reforming of methane. In: Catalysis for Alternative Energy Generation. Springer, New York, pp. 57127.
Reichelt, W., 1972. Zur Berechnung des Druckverlustes einphasig durchstrmter
Kugel- und Zylinderschttungen. Chem. Ingen. Tech. 44, 10681071.
Ren, Z., Pope, S.B., 2008. Second-order splitting schemes for a class of reactive
systems. J. Comput. Phys. 227, 81658176.
Rostrup-Nielsen, J., Hansen, J., 1993. CO2-reforming of methane over transition
metals. J. Catal. 144, 3849.
Salciccioli, M., Stamatakis, M., Caratzoulas, S., Vlachos, D., 2011. A review of
multiscale modeling of metal-catalyzed reactions: mechanism development
for complexity and emergent behavior. Chem. Eng. Sci. 66, 43194355.
Schwiedernoch, R., Tischer, S., Correa, C., Deutschmann, O., 2003. Experimental and
numerical study on the transient behavior of partial oxidation of methane in a
catalytic monolith. Chem. Eng. Sci. 58, 633642.
Shih, T.H., Liou, W.W., Shabbir, A., Yang, Z., Zhu, J., 1995. A new k- eddy viscosity
model for high Reynolds number turbulent ows. Comput. Fluids 24, 227238.
Taskin, M.E., Troupel, A., Dixon, A.G., Nijemeisland, M., Stitt, E.H., 2010. Flow,
transport, and reaction interactions for cylindrical particles with strongly
endothermic reactions. Ind. Eng. Chem. Res. 49, 90269037.
Torniainen, P., Chu, X., Schmidt, L., 1994. Comparison of monolith-supported metals
for the direct oxidation of methane to syngas. J. Catal. 146, 110.
Vortmeyer, D., Winter, R., 1984. On the validity limits of packed bed reactor
continuum models with respect to tube to particle diameter ratio. Chem. Eng.
Sci. 39, 14301432.
Wang, S., Lu, G.Q.M., Millar, G.J., 1996. Carbon dioxide reforming of methane to
produce synthesis gas over metal-supported catalysts: state of the art. Energy
Fuels 10, 896904.
Wehinger, G.D., Eppinger, T., Kraume, M., 2014. Fluidic effects on kinetic parameter
estimation in lab-scale catalysis testinga critical evaluation based on computational uid dynamics. Chem. Eng. Sci. 111, 220230.
Wolfshtein, M., 1969. The velocity and temperature distribution in one-dimensional
ow with turbulence augmentation and pressure gradient. Int. J. Heat Mass
Transf. 12, 301318.
Zikowska, I., Zikowski, D., 1988. Fluid ow inside packed beds. Chem. Eng.
Process.: Process Intensif. 23, 137164.
Zobel, N., Eppinger, T., Behrendt, F., Kraume, M., 2012. Inuence of the wall
structure on the void fraction distribution in packed beds. Chem. Eng. Sci. 71,
212219.