Mathematical Modeling of Transport Processes

Mathematical Modeling of
Transport Processes
Editors
Agus P. Sasmito, Jundika C. Kurnia and Sachin V. Jangam

Transport Processes

Transport Processes

Editors: Agus P. Sasmito, Jundika C. Kurnia and
Sachin V. Jangam

2011

Mathematical Modeling of Transport Processes

Copyright 2011

ISBN: 978-981-08-9179-4

All rights reserved. No part of this publication may be reproduced or
distributed in any form or by any means, or stored in a database or
retrieval system, without the prior written permission of the
copyright holder.
FOREWORD

For a long time I have thought of and even planned to offer a new type of assignment leading
to some new pedagogy in teaching advanced level postgraduate course in Heat and Mass
Transfer. Essentially it is based on the well known and well-practiced Problem-based
learning (PBL) approach. Students are expected to self-learn based on carefully chosen
problems the solutions of which require groups of students to carry out appropriate
searches on their own. This experience is very valuable despite some inherent limitations
that should be covered elsewhere in the curriculum.

My idea was to let groups of students (in my case just 2) carry out independent and original
research project in transport phenomena and learn all aspects of carrying out a research
project from start to finish just in 3 months. In the current case the course was entitled Mass
Transport which included coupled heat transfer and reactions as well. Seven projects were
loosely defined so as to be original with no readymade solutions found by goggling. Most
projects required use of CFD software such as Fluent and MATLAB. My graduate students
agreed to be mentors for the 7 groups formed out of 14 students most PhD students in
mechanical engineering which I defined and sometimes refined the project
scope/objectives. Students were allowed to modify them with justification. Students wrote
technical report following a standard journal practice and then prepared suitable
PowerPoint presentations of about 20 minutes with 10 minutes for questions and
discussion. They were also told that selected papers could be reformatted and compiled as a
free e-book so as to provide a model for future generations of instructors and students who
wish to follow the model.

In the classroom, lectures focused on basic concepts of conservation laws, control volumes
vs. systems, initial and boundary conditions, analytical solutions, dimensionless groups,
scaling of equations etc. Mentors trained most in use of CFD software as required. Students
were allowed to discuss problems and solutions with mentors and instructor but were
required to be as independent and original as possible. It is impossible to make all projects
of equal complexity or difficulty, of course.

The outcome of this effort is presented in this book upon reformatting and minor updates
for consistency. Readers will be impressed by both the breadth and depth achieved just in
about 3 months from start when most did not know what diffusion is and how a mass
transfer coefficient is defined as all had mechanical and civil engineering background.

My personal impression is that most students effectively went through the entire process
they need to go through during their PhD research. They even had experience to deliver
their paper in a conference like setting with an external Chairman and Co-Chairman.

I hope that the readers will find this documented experience of value in defining their own
pedagogical assignments. With some mentoring it seems to me that students can be
motivated to produce high quality work and learn several different aspects of research while
carrying out an essential course exercise.

I want to thank my PhD students Agus P. Sasmito, Jundika C. Kurnia and Sachin V. Jangam for
being such helpful and motivating mentors. We hope some of these term papers will be
refined and upgraded and submitted for journal publication as well.

Congratulations to everyone whose name appears in this e-book.

Arun S. Mujumdar
Professor

http//serve.me.nus.edu.sg/arun

PREFACE

We are pleased at the opportunity to co-edit under Professor A.S. Mujumdar's guidance
this concise e-book which essentially records the outcome of a new pedagogical exercise
Prof. Mujumdar designed and implemented for the first time in his module ME6203
Mass Transport here at NUS. We are pleased to be a part to this unique experiment
which also has been very successful. We are all Professor Mujumdar's PhD students and
also mentors for the seven term paper projects that were assigned and completed
during January-March 2011.

Although given rather short time, students were able to jump-start their work as the
theme was specified to ensure there was opportunity to make a contribution to a new
area not covered in the lectures and indeed not available on "Google" searches either.
Students had to define the problem and then develop mathematical descriptions
followed by numerical or analytical solutions (if possible). Some help was provided with
the numerical aspect. Focus was placed on processing and interpretation of results and
on presenting it in the form of a technical journal paper. To simulate oral presentation
experience, the students were required to present their work to a larger audience of
faculty and students as happens in a conference session. Numerous criteria were taken
into account for evaluation.

Professor Mujumdar terms this Research Project Based Learning" which probably
represents a paradigm shift in teaching of PhD students since they go through all key
aspects of what is needed to complete a PhD thesis. Readers can make their own
judgment. We feel several of the term papers (as submitted) included here can be
enhanced further for journal publication. The copyrights of the content rest with
individual authors. This freely downloadable book is being brought out solely as a
service to faculty members all over the world who wish to follow this or a modified
version of the pedagogical model in their teaching.

We enjoyed working on this unique experiment and recommend it warmly. The only
caution is that it is a major effort for the faculty members and mentors who need to be
extremely passionate and motivated about teaching.

Agus P. Sasmito
Jundika C. Kurnia
Sachin V. Jangam

June 2011

Index

Chapter No Title / Authors Page No
01 Numerical investigation of the performance of various
micro-reactor configurations
Hui An and Ang Li
01
02 Enhanced convective mixing for gaseous microreactors
M. Shaker and H. Ghaedamini
25
03 Modeling conjugate heat and mass transfer between a
laminar impinging jet and planar and curved thin slab of
wet solids
Tong Wei and Wang Yue
47
04 Numerical evaluation of pulsation effect on the reaction
per-formance of a t-junction micro-mixer
Balaji Mohan and Jiang Puqing
61
05 Numerical analysis of single-particle combustion of coal
char
H. Osman and A. Ismail
83

Hui An, Ang Li Numerical Investigation of the Performance of Various Micro-reactor configura-
tions


Chapter 1
Numerical investigation of the performance of various micro-
reactor configurations

Hui An
1
and Ang Li
2
Department of Mechanical Engineering, National University of Singapore,
9 Engineering Drive 1, Singapore
E-mail:

1
anhui@nus.edu.sg
2
a0034298@nus.edu.sg
Contents
1.1. INTRODUCTION ................................................................................................................................ 3
1.2. MATHEMATICAL MODEL .............................................................................................................. 5
1.2.1. GOVERNING EQUATIONS ..................................................................................................................... 6
1.2.2. CHEMICAL REATIONS ........................................................................................................................... 7
1.2.3. CONSTITUTIVE RELATIONS ............................................................................................................... 8
1.2.3.1. THERMODYNAMIC PROPERTIES OF THE WORKING FLUID ........................................ 8
1.2.3.2. REACTOR PERFORMANCE .......................................................................................................... 9
1.2.4. BOUNDARY CONDITIONS .................................................................................................................... 9
1.3. NUMERICS ........................................................................................................................................ 10
1.4. RESULTS AND DISCUSSIONS ...................................................................................................... 11
1.4.1. EFFECT OF CHANNEL CONFIGURATION .................................................................................... 11
1.4.2. EFFECT OF REYNOLDS NUMBER ................................................................................................... 17
1.5. CONCLUSION.................................................................................................................................... 21
NOMENCLATURE .................................................................................................................................... 21
REFERENCES ............................................................................................................................................ 22

Hui An and Ang Li. Numerical Investigation of the performance of various micro-reactor, Sasmito, A.P.; Kurnia,
J.C.; Jangam, S.V. and Mujumdar, A.S., 2011, ISBN 978-981-08-9179-4, Published in Singapore, pp. 1-24

tions

1.1. INTRODUCTION
Micro-reactor is a recent developing technology in chemical and pharmaceutical in-
dustries. It refers to the reactors with inner dimensions of the structure in the order of
millimetres and smaller. A number of impressive advantages have been demonstrated
on the micro-structured device over the conventional chemical reactors. It exposes high
area to volume ratio and offers high heat exchange and mass transport rate. Most of the
flows inside the micro-reactors are laminar in nature which allows precise computer
simulation on the chemical reactions. Its small scale also permits much easier control on
process parameters like temperature, pressure and flow rate, etc, and hence, reduces the
danger of highly exothermic or explosive chemical reactions (Jhnisch et al. 2003; Watts
and Haswell 2005).
Numerous research activities have been conducted to investigate the performance
of various micro-reactor designs. One of commonly used design choice is an array of mi-
cro-channels arranged in parallel with a common inlet and outlet channels perpendicu-
lar to them. This type of structure is easy to manufacture but expose poor flow distribu-
tion (Amador et al. 2004). A fractal tree-like network structure frequently observed
from natural organs (Bejan and Errera 1997; Chen and Peng 2002) and a regular bifur-
cation shape (Amador et al. 2004), on contrast, are able to provide more uniform flow
distribution and lower pressure loss in the structure. Coil-shape micro-reactor has been
demonstrated its ability of offering high heat and mass transfer as a result of the pres-
ence of secondary flows (Sasmito et al. 2011). Agrawal and Nigam (Agrawal and Nigam
2001) also found that its performance lies in between the plug and laminar tube flow
reactors under a premix inlet condition. Another type of design is to introduce posts or
baffles (Chung et al. 2011) inside the microreactor to enhance the mixing and reaction.
Posts in catalytic reaction are often coated or directly made of catalysts and exhibits
very high surface area (Ni et al. 2005; Yeom et al. 2009). A proper designed posted struc-
ture may provide narrow residence time distribution of gas and lead to high reaction
rate under carefully controlled condition (Deshmukh et al. 2004; Regatte and Kaisare
2011a). Furthermore, the shapes of cross-section (Ramanathan et al. 2003; Arrighetti et
al. 2007; Carvalho et al. 2009, Kurnia et al 2011a) of the micro-reactors such as triangle,
rectangle, square, circular, sinusoidal and hexagonal have also been investigated to im-
prove the performance of micro-reactor.
Apart from the reactor geometry, flow configuration also plays an important role in
determining the performance of the micro-reactor. Axial- and cross-flow are two typical
flow configurations implemented in the micro-reactor design. Ajmera et al (Ajmera et al.
2002) studied fixed-bed micro-reactors and found out that cross-flow configuration can
offer high throughput in expense of less pressure loss than that of axial-flow. However,
cross-flow impose considerable flow mal-distribution, resulting in low reaction rate
(Regatte and Kaisare 2011b). A uniform flow distribution is essential to the rate of reac-
tant conversion (Saber et al. 2010).
tions

Figure 1.1. Micro-reactor configuration design: (a) parallel; (b) serpentine; (c) oblique fin; (e) coiled with
outer inlet/outlet; (f) coiled with inner inlet/outlet; (g) coiled with serpentine and (h) coiled with double
serpentine.
Micro-reactors generally work well for low Reynolds number. However, low Rey-
nolds number leads to low throughput, while high flow rate impose a drop in chemical
a b
c
d
c
e
f
g h
tions

reaction rate, which bottleneck the application of micro-structure in the industry. In this
paper the performance of a few configurations of micro-reactors are evaluated at both
low and high Reynolds number. Figure 1.1 shows the configurations for parallel, serpen-
tine, wavy and oblique fin channels, two rectangular coiled channels with opposite flow
directions as well as two novel hybrid channels. Among all these designs, parallel, wavy
and oblique fin implements cross-flow configurations, whereas the rest are single-
channel. Moreover, the shape of oblique fin channel may be similar to that of micro-
reactor with parallelogrammic posts. In addition, a square shape cross-section for all the
configurations was used as it has better heat transfer behaviour over other design,
slightly higher total catalytic surface area as well as less requirement on manufacture
(Kurnia et al. 2011a)
The heat transfer capacity of these configurations has been evaluated in previous
work (Kurnia et al. 2011b) and the latter four designs has been demonstrated higher
rate and more uniform distribution over others. The reaction implemented to demon-
strate the concept is methane oxidation reaction with platinum as catalyst coated on the
inner wall of surfaces (Li 2004 ; Deutschmann et al. 2000). Note that other type of reac-
tions can also be implemented within frame work derived here. It is assumed that the
species are well mixed at the inlet and constant well temperature of each configuration.
The numerical simulation is conducted by Fluent 6.3.26. The performance of eight con-
figurations in terms of reaction rate of species, pressure drop and Figure of Merit (de-
fined later) are evaluated under four Reynolds numbers (they are, 100, 500, 1000, 1500)
with an assumption of premix inlet condition.

1.2. MATHEMATICAL MODEL
The physical model which can be seen in Figure 1.1 comprises of eight micro-
reaction channel designs, e.g., parallel, serpentine, wavy channel, oblique fin, coiled with
outer inlet/outlet, coiled with inner inlet/outlet, coiled with serpentine and coiled with
double serpentine. All the channels have the same cross-section (1mmx1mm). In this
model, the methane and air is assumed to be well mixed before entering the channel
from the inlet. The reactions at the channel wall consist of multi-step reactions: adsorp-
tion reaction, surface reaction and desorption reaction. The detailed multi-step reaction
mechanism and its reaction rate constants are listed in Table 1.1. The wall is assumed to
have constant temperature, whereas the fluid is assumed to be steady, laminar, and
Newtonian. Since this work relates only to laminar flow, a precise numerical solution is
adequate to simulate the reality very closely.

Table 1.1. Surface reaction mechanism
No Reaction A
r

r
E
r

tions
(J/kmol)
1 H2 + 2Pt(s) => 2H(s) 4.36e7 0.5 0
2 2H(s) => H2 + 2Pt(s) 3.7e20 0 6.74e7
3 O2 + 2Pt(s) => 2O(s) 1.8e17 -0.5 0
4 O2 + 2PT(s) => 2O(s) 2.01e14 0.5 0
5 2O(s) => O2 + 2Pt(s) 3.7e20 0 2.13e8
6 H2O + Pt(s) => H2O(s) 2.37e8 0.5 0
7 H2O(s) => H2O + Pt(s) 1e13 0 4.03e7
8 OH + Pt(s) => OH(s) 3.25e8 0.5 0
9 OH(s) => OH + Pt(s) 1e13 0 1.93e8
10 H(s) + O(s) => OH(s) + Pt(s) 3.7e20 0 1.15e7
11 H(s) + OH(s) => H2O(s) + Pt(s) 3.7e20 0 1.74e7
12 OH(s) + OH(s) => H2O(s) + O(s) 3.7e20 0 4.82e7
13 CO + Pt(s) => CO(s) 7.85e15 0.5 0
14 CO(s) => CO + Pt(s) 1e13 0 1.25e8
15 CO2(s) => CO2 + Pt(s) 1e13 0 2.05e7
16 CO(s) + O(s) => CO2(s) + Pt(s) 3.7e20 0 1.05e8
17 CH4 + 2Pt(s) => CH3(s) + H(s) 2.3e16 0.5 0
18 CH3(s) + Pt(s) => CH2(s) + H(s) 3.7e20 0 2e7
19 CH2(s) + Pt(s) => CH(s) + H(s) 3.7e20 0 2e7
20 CH(s) + Pt(s) => C(s) + H(s) 3.7e20 0 2e7
21 C(s) + O(s) => CO(s) + Pt(s) 3.7e20 0 6.28e7
22 CO(s) + Pt(s) => C(s) + O(s) 1e17 0 1.84e8
23 OH(s) + Pt(s) => H(s) + O(s) 1.56e18 0 1.15e7
24 H2O(s) + Pt(s) => H(s) + OH(s) 1.88e18 0 1.74e7
25 H2O(s) + O(s) => OH(s) + OH(s) 4.45e20 0 4.82e7

1.2.1. GOVERNING EQUATIONS
In the reaction channel, the conservation equations of mass, momentum, species
and energy are considered and given by

0 \ = u
(1)

tions

( )
( )
( )
2
3
p
1
l

\ = \ \ \ \ \
l

l ( )
T
u u u u u I
(2)

( ) ( )
i i i i
+ R D \ = \ \ u

(3)
( ) ( )
p eff temp
c T k T S \ = \ \ u (4)
In the above equations, is the fluid density, u is the fluid velocity, p is the pressure,
is the dynamic viscosity, T is temperature,
i
is the mass fraction of species i ,
i
D is
the diffusion coefficient of species i,
i
R is the mass consumed or produced by reactions,
p
c is the specific heat of the nano-fluid,
eff
k is the effective thermal conductivity and
temp
S
is
heat release/absorb due to reactions.
1.2.2. CHEMICAL REATIONS
The reaction model considers the chemical reactions to be occurred on the channel
wall only, and the reactions involving surface deposition are defined as distinct surface
reactions and hence treated differently from bulk phase reactions involving the same
species. In the same way, the chemical species deposited on surfaces are treated as dis-
tinct from the same chemical species in the gas. In this model, seven gas species (CH4,
O2, H2, H2O, CO, CO2 and N2), one bulk/solid species (Pt(b)) and eleven surface species
(H(s), Pt(s), O(s), OH(s), H2O(s), CH3(s), CH2(s), CH(s), C(s), CO(s), CO2(s)) are consi-
dered.
The gas phase species and surface species can be produced and consumed by sur-
face reactions and the general expression is given by:

' ' ' '' '' ''
, , , , , ,
1 1 1 1 1 1
g g
b s b s
r
N N
N N N N
K
i r i i r i i r i i r i i r i i r i
i i i i i i
g G b B s S g G b B s S
= = = = = =

_ _ _ _ _ _
(5)
Here,
i
G represents the gas phase species,
i
B represents the solid species, and
i
S
represents the surface adsorbed species.
r
K ,
'
, i r
b ,
'
, i r
s are the stoichiometric coefficients
for the three reactant species respectively, and
''
, i r
g ,
''
, i r
b ,
''
, i r
s are the stoichiometric coef-
ficients for the three product species respectively.
r
K is the overall reaction rate con-
stant.
The rate of reaction is given

' '
, ,
,
wall wall
1
g
i r i r
N
g s
r f r i i
i
k G S
=
l l
+ =
l l
]
(6)
where
, f r
k is the reaction rate constant which is calculated using the Arrhenius equa-
tions given by:

/
,
r r
E RT
f r r
k AT e

= (7)
tions
and
wall
i
G
l
l
is the molar concentration on the wall. So the net molar rate of consumption
or production for each species i is represented by:

( )
( )
( )
rxn
rxn
rxn
'' '
,gas , ,
1
'' '
,bulk , ,
1
'' '
,site , ,
1
1, 2, 3, ...,
1, 2, 3, ...,
1, 2, 3, ...,
N
i i r i r r g
r
N
i i r i r r b
r
N
i i r i r r s
r
R g g i N
R b b i N
R s s i N
=
=
=
= + =
= + =
= + =
_
_
_
(8)
Furthermore, it is assumed that on the wall surface the mass flux of each gas species
is balanced with its rate of production/consumption which is given by:

,wall
wall dep i,wall , i,gas
1, 2, 3, ...,
i
i w i g
D m M R i N
n

0
= =
0
(9)

wall
i,site
1, 2, 3, ...,
i
s
S
R i N
t
l
0
l
= =
0
(10)
where
dep
m is the rate of mass deposition or etching due to surface reaction which is
given by:

dep , ,
1
b
N
w i i bulk
i
m M R
=
=
_
(11)
and
wall
i
S
l
l
is the site species concentration on the wall, defined as:

site
wall
i i
S z
l
=
l
(12)
where
site
is the site density of the catalyst and
i
z is the site coverage of species i.
The gas concentration at the wall is calculated from the species mass fraction which
is expressed by:

wall i,wall
wall
,
i
w i
G
M

l
=
l
(13)
1.2.3. CONSTITUTIVE RELATIONS
1.2.3.1. THERMODYNAMIC PROPERTIES OF THE WORKING FLUID
The working fluid is assumed to be well mixed air and methane. The thermody-
namic properties of the working fluid are calculated as follows:
The fluid density is given by the ideal gas law:
/ pM RT = (14)
where R is the universal gas constant andM is the molecular weight of the mixture flu-
id given by:
tions


4 2 2 2 2 2
4 2 2 2 2 2
1
CH H O H O CO N
CO
CH H O H O CO CO N
M
M M M M M M M

( )
(15)
The fluid mixture viscosity is

4 2 2 2 2 2
,
with , = CH , H , O , H O, CO, CO , N
x
x
=
1
_
_
(16)
where
,
x
are the mole fractions of species and , and

,
1 is given by the expres-
sion below:

2
1
1
1/2
2 (g)
4
,
(g)
1
1 1
8
M
M
M M
l
l
1
1 1
l

1 =
l

l
( ) ( )
( )
l
l
l
(17)
The multi-component gas mixture thermal conductivity is defined as:

eff i i
k k =
_
(18)
The gas mixture specific heat capacity
p
c is calculated by:

p i p,i
i
c c =
_
(19)
1.2.3.2. REACTOR PERFORMANCE
In the later part of this paper, the results are discussed based on the figure of merit
(). To ensure the fidelity of comparison between various micro-channel designs, the
figure of merit is introduced to evaluate the effect of Reynolds number and the effect of
geometry on the pressure drop and reaction rate. is defined as the ratio of mass
consumption rate per unit pumping power required given by:

r,in r,out
pump
FoM
P

= (20)
where
pump
P is the pumping power that is required to drive the fluid flow through the
channel, calculated by

1
pump
pump
P Q p
= ^ (21)
In the above equation,
pump
is the pump efficiency, Q is the volume flow rate of the fluid,
and p ^ is the pressure drop.
1.2.4. BOUNDARY CONDITIONS
The boundary conditions for the flow through the micro-channels are defined as
follows:
tions
Inlet: The mixture fluid of air and methane is assumed to be well mixed before
entering the inlet of the micro-channels. We define the inlet mixture fluid flow
velocity and inlet temperature as:

2 4 2
2 4 2
in in in
O
O CH H
, , , ,
in in CH H
u u T T = = = = = (22)
Outlet: At the outlet, we specified the pressure. The stream wise gradient of the
temperature and the species mass fraction are both set to zero. The velocity is
not known a priori but needs to be iterated from the neighboring computational
cells.

out
, 0
i
p p T = \ = \ = n n (23)

Flow channel walls: at the walls of the channels, the surface reaction is taken into
account and is resolved based on Eq.9. We specify no slip condition and constant
wall temperature.

i
0, 0 T = \ = \ = u (24)
1.3. NUMERICS
The computational domains shown in Figure 1.1 were created using AutoCAD 2010.
The commercial pre-processor software GAMBIT 2.3.16 was used for building the mesh,
labeling boundary conditions and determining the computation domains. There are
three different amount of mesh (2.5 10
5
, 5 10
5
, 1 10
6
) that were built and com-
pared in terms of pressure, temperature and velocity to ensure a mesh independent so-
lution. The results show that a mesh amount of 5 10
5
only differs about 1% in terms
of p, u, and with the mesh amount of 1 10
6
, whereas a mesh amount of 2.5 10
5

gives a deviation of 7% as compared to the finest mesh. Hence, a mesh amount of
5 10
5
was used for the numerical investigation purposes since it reduces the computa-
tional time and at the same time gives a reasonable accuracy on the results.
Based on equations 1-4 and the boundary conditions shown in Table 1.2, the finite
volume solver Fluent 6.3.26 was used to solve the constitutive relations consisting of
eleven dependent variables-, , , ,
2
,
4
,
2
,
2
,
2
,
,
and . The gas prop-
erties and chemical reaction mechanism was generated by using ChemKIN software; and
a user defined function (UDF) file was also created using C language to account for tem-
perature dependence of the thermo-physical properties of the fluid.
Table 1.2. Base case and operating parameters
Parameter Symbol Value Unit
Inlet flow velocity (Re 100)
1 m/s
5 m/s
10 m/s
tions

15 m/s
Inlet temperature
300 K
Outlet pressure
101,325 Pa
Wall temperature
1290 K

1.4. RESULTS AND DISCUSSIONS
The computational fluid dynamic simulations were conducted for typical conditions
for micro-reactors. The boundary conditions and geometric parameters of the eight con-
figurations listed in Table 1.2 and Table 1.3 In the following, the effect of geometry on
the performance of the configurations under high flow rate (Reynolds number 1500) is
first addressed. Then an evaluation of the behaviour of every design affected various
Reynolds number (100, 500, 1000, 1500) is performed. The concept of the Figure of
Merit is utilized to compare the conversion rate of reactants of all configurations at unit
pumping power.
1.4.1. EFFECT OF CHANNEL CONFIGURATION
Geometry is one of the key factors to determine the performance of the micro-
reactor. Flow distribution caused by reactor structure has a direct impact on the reac-
tion rate, and a mal-distribution reduces the fraction of species to be converted. Figure
1.2 shows the simulated velocity contour at the middle of the flow channel (i.e. z = 5 x
10
-4
m) of each reactor configuration at Reynolds number of 1500. It is observed that the
serpentine (Figure 1.2b), the two rectangular coiled (Figure 1.2e-f) and the two hybrid
coiled (Figure 1.2g-h) designs have higher velocity magnitude across their configuration
compared to the parallel (Figure 1.2a), wavy (Figure 1.2c) and oblique fin (Figure 1.2d)
shape which, however, have overall velocity magnitude approximately one order
smaller. Moreover, the latter three structures have also shown that higher velocity ap-
pears at branch channels near the inlet and outlet but the middle branches have lower
velocity, which in agreement with flow behaviour in a posted micro-reactors simulated
by Regatte and Kaisare(Regatte and Kaisare 2011b). Both phenomena attribute to the
structure of the reactor designs where cross-flow configuration leads to an uneven dis-
tribution of mass flow rate inside the parallel, wavy and oblique fin configurations, while
the rest have only one channel path throughout the configuration with now flow distri-
bution involved. To improve the distribution of the species, one may tapper the inlet and
outlet of the parallel, wavy and oblique fin designs (see (Regatte and Kaisare 2011b)). A
regular bifurcation shape (Amador et al. 2004) (as introduced early) can also be imple-
mented to obtain a more even flow distribution.
Table 1.3. Geometric parameters
Parameter Symbol Value Unit
Channel Width
ch
w 1 x 10
-3
m
tions
Channel Height
ch
h 1 x 10
-3
m
Oblique fin angle
ob
26
o

Oblique fin width
ob
w 4.49 x 10
-4
m
Oblique fin pitch
ob
p 3.06 x 10
-3
m
Number of sinusoidal wave
wv
n 10
Amplitude of sinusoidal wave
wv
A 5.10 x 10
-4
m
Total length of parallel channel
pa
L 1.376 m
Total length of serpentine channel
se
L 1.351 m
Total length of wavy channel
wv
L 1.486 m
Total length of oblique fin channel
ob
L 1.376 m
Total length of coiled channel with outer
inlet/outlet
co
L 1.428 m
Total length of coiled channel with inner
inlet/outlet
ci
L 1.428 m
Total length of coiled channel with serpentine
cs
L 1.428 m
Total length of coiled channel with double serpen-
tine
cd
L 1.428 m

tions


a

b

c

d

e

f

g

h

Figure 1.2. Velocity contours at z = 5 x 10-4 m for (a) parallel; (b) serpentine; (c) oblique fin; (e)
coiled with outer inlet/outlet; (f) coiled with inner inlet/outlet; (g) coiled with serpentine and (h)
coiled with double serpentine for Re 1500.
tions


a

b

c

d

e

f

g

h

Figure 1.3. Mass fraction of CH4 at z = 5 x 10-4 m for (a) parallel; (b) serpentine; (c) oblique fin; (e)
coiled with double serpentine for Re 1500.
Reaction rate of species shows an inverse trend to the velocity contour of reactor
configuration. At z = 5 x 10-4 m and Re 1500 for each configuration with same inlet
tions


condition. The mass fraction of CH4 of the serpentine (Figure 1.3b) and four coil based
designs (Figure 1.3e-h) continuously reduce along the channels, as shown in Figure 1.3.
The outlet mass fraction of CH4 for these five reactors presents a very small value, which
implies that the reaction is almost complete inside the channels. The parallel (Figure
1.3a), wavy (Figure 1.3c) and oblique fin (Figure 1.3d) configurations, however, exhibit
uneven mass distribution. The reaction rate of CH4 of these reactors is not as good as the
serpentine and coil based configurations, as the results show that at the outlet the wavy
and oblique fin channel has the mass fraction of CH4 one order of magnitude larger than
five designs and the parallel channel gives two orders of magnitude larger. The reaction
contour of O2 of each reactor is similar to that of CH4. This large deviation can be ex-
plained by the residence time of the species. Thermodynamically speaking, the longer
the residence time is, the better the conversion rate. Infinitely long residence time may
lead to almost full species conversion. From kinetic point of view, however, this is not
feasible as it requires extremely low flow rate. For the three rectilinear designs (the pa-
rallel, wavy and oblique fin), species finish the chemical conversion in a short channel
distance at their middle branches. As observed in Figure 1.2 that velocity is very low at
these areas, the reactants present longer residence time and hence better reaction rate.
The branches near the entrance or exit, on contrast, are not long enough to accommo-
date the species for reaction due to the higher velocity noted as compared to middle
branches. Thus the mal-distribution of velocity abates the overall reaction performance
of these multi-channel configurations. On the other hand, single channel implemented in
serpentine and coil based configuration in general has much longer travelling distance
and therefore longer effective residence time for reactants, despite of high velocity.
Hence the configurations of micro-reactor with long single channel design such as ser-
pentine and coil are more efficient for the chemical reaction.
High velocity can result considerable pressure drop across the micro-reactors. Fig-
ure 1.4 illustrates the pressure distribution across the reactor configurations introduced
in this paper. The serpentine (Figure 1.4b) and the four coil based channels (Figure 1.4e-
h) which have large velocity reveal a significant difference between inlet and outlet. A
large amount of pumping power is needed to push the fluid to flow in these channels. On
contrast, the requirement of pumping power is much less for the parallel (Figure 1.4a),
wavy (Figure 1.4c) and oblique fin (Figure 1.4d). Their inlet pressures are one approxi-
mately 30 times smaller than the coiled based channels. This can be explained by the
fact that micro-reactors with branch channels distribute the mass flow rate so that the
friction loss is small in branch channels. This also explains that oblique fin has the small-
est pressure drop among all the configurations as it has the largest number of branch
channels.

tions


a

b

c

d

e

f

g
h
Figure 1.4. Pressure distribution at z = 5 x 10-4 m for (a) parallel; (b) serpentine; (c) oblique fin; (e)
coiled with double serpentine for Re 1500
tions


1.4.2. EFFECT OF REYNOLDS NUMBER
Another key factor for investigation in this study is the effect of Reynolds number,
as it directly affects the flow behavior and hence the reaction rate and the pumping
power that is required. In this study, four different Reynolds numbers 100, 500, 1000
and 1500 were used for investigation.
a)

b)

Figure 1.5. Pressure drop for each design at various Reynolds numbers (a) rectilinear designs; (b)
coiled- base designs.

tions

a)

b)

Figure 1.6. Reaction rate for each design at various Reynolds numbers (a) rectilinear designs; (b) coiled-
base designs.
For similar micro-channel designs, we always would like to keep the pressure drop to
be the minimum, as it directly affects the operating cost. Hence, a good microchannel
design in our study should be able to maintain the reaction rate, and at the same time,
lower the pressure drop. Figure 1.5 shows the pressure drop for different micro-channel
designs at various Reynolds numbers. As expected, the pressure drop increases as the
Reynolds number increases for all the designs. This is because larger Reynolds number
means larger flow rate, and will require higher pressure difference to drive the fluid
flow. We can also see that coiled-base designs (Figure 1.5b) require higher pressure
tions


drop compared with that for rectilinear designs (Figure 1.5a) at the same Reynolds
number. This can be explained by the more complex flow patterns inside the channel
and the longer flow passages. Furthermore, among the rectilinear designs, the pressure
drop for the serpentine channel is one order of magnitude higher than that of parallel,
oblique-fin and wavy channels. We can note that if the flow channel does not split, the
pressure drop is much higher than that with flow splitting since the fluid is forced to
flow in a much longer passages.
Another point of interest is the reaction rate. A good micro-channel design not only
gives a lower pressure drop but also a higher reaction rate. Figure 1.6 depicts the reac-
tion rate for various micro-channel designs at different Reynolds numbers. The coiled-
based designs, as shown in Figure 1.6b, gives the highest reaction rate ranges from 98.8
to 100%. This can be expected due to their longer flow passages which in turn provide
longer reaction duration. On the other hand, the reaction rate for the serpentine channel
design is slightly lower compared to that of the coiled-based design; however, it is still
much higher compared to that of oblique fin, wavy channel and parallel channels espe-
cially at higher Reynolds numbers. But at lower Reynolds numbers, the reaction rate for
the oblique-fin and wavy channels does not differ much compared to that of serpentine,
whilst the pressure drop reduce significantly. This provides a clear evidence that the
oblique-fin and wavy channels are also very effective when used at lower Reynolds
numbers. As shown in Figure 1.1a, the parallel channel design gives the simplest geome-
try, and the distance from the inlet to the outlet is also the shortest. Hence, the reaction
time in the parallel channel design would be the shortest and the reaction rate would be
the lowest as reflected in Figure 1.6a.
The fluid velocity and hence the mass fluid rate at the inlet for different micro-
channel designs increases linearly with the Reynolds number since the cross sectional
area for different designs are the same. For the coiled-based designs, the reaction rates
are almost equal to one even at the highest Reynolds number; hence we can see the
mass consumption rate increases linearly with the Reynolds number as shown in Figure
1.7b. However, at higher Reynolds numbers, the reaction rate for the rectilinear designs
drops significantly which results in the drop of mass consumption rate compared to that
of the coiled-based designs at the same Reynolds number. This explains the deviation of
the mass consumption verses Reynolds number curves for oblique-fin, parallel and
wavy channels from that for the serpentine curve as shown in Figure 1.7a.
With respect to the reaction rate and the pressure drop for different channel designs,
the Figure of Merit concept is introduced to compare the effectiveness of the reactor
designs per unit pumping power (see Eq. 20). Table 1.4 lists the calculated figure of me-
rit for various reaction channel designs at different Reynolds numbers. It can be seen
that as the Reynolds number increases, the figure of merit decreases. This is due to the
fact that for coil-based designs, the pressure drops increases significantly as the Rey-
nolds number increases, whereas for rectilinear designs, the reaction rate reduces large-
ly with respect to the increases of Reynolds number. It can also be found that the coil-
based channel designs have lower figure of merit. This is because coil-based channels
require the highest pressure drop as shown in Figure 1.5b. On the other hand, the paral-
lel, wavy channel and oblique fin channels give much higher figures of merit, up to
tions

around two order of magnitude compared to those for coiled and serpentine channel
designs. This results from the lower pressure drop.
a)

b)

Figure 1.7. Mass consumption rate for each design at various Reynolds numbers (a) rectilinear designs;
(b) coiled- base designs.

tions


Table 1.4 Figure of merit for various channel designs at different Reynolds numbers
Channel Design Re 100 Re 500 Re 1000 Re 1500
Coiled with double serpentine 172.4 7.2 1.5 0.6
Coiled with serpentine 173.1 7.2 1.5 0.6
Coiled with inner inlet/outlet 172.5 7.2 1.6 0.6
Coiled with outer inlet/outlet 172.4 7.1 1.5 0.6
Parallel with oblique fins 11363.9 425.8 80.2 26.9
Parallel 7915.8 236.9 41.3 12.0
Serpentine 175.4 7.9 1.7 0.7
Parallel wavy 5501.0 176.1 33.7 11.1

When designing reaction channels, careful balance and consideration has to be given to
the reaction rate and the pumping power that is required. When the reaction rate is
more of the interest, one can consider coil-based designs and the serpentine channel
design. However, if pumping power is the major concern, the oblique fin design may be
the most ideal case for consideration.

1.5. CONCLUSION
A reaction modeling has been done to investigate the reaction rate and the pres-
sure drop for various reaction channel designs at different Reynolds numbers (100, 500,
1000, and 1500). Eight channel designs,-parallel, serpentine, wavy channel, oblique fin,
coild with inner inlet/outlet, coiled with outer inlet/outlet, coiled with serpentine, and
coiled with double serpentine were used for investigation. The reaction performance of
micro-channels is discussed in terms of the figure of merit. It was found that coil based
designs give much higher reaction rate at all Reynolds numbers compared to those of
rectilinear designs due to their complex flow patterns and longer flow passages, but they
also impose a significantly higher pressure drop penalty. As s result, the figures of merit
for coil-based designs are very much lower. However, for the application where pump-
ing power is not an issue, the coil-based design can be the ideal choices for applications.
NOMENCLATURE
Ar pre-exponential factor
Bi bulk/solid species, mol
bi
, bi
stoichiometric coefficient for bulk reactant, and product

cp specific heat, Jkg
-1
K
-1

Di diffusivity of species I, ms
-2

Er activation energy for the reaction, Jkgmol
tions

Gi gas species, mol
gi
, gi
stoichiometric coefficient for gas reactant and product

keff effective thermal conductivity, Wm
-1
K
-1

kf,r reaction rate constant using Arrhenius expression
M mean molecular mass
dep net rate of mass deposition, kg
p pressure, pa
Q volume flow rate, m
3
s
-1

R universal gas constant, Jkg
-1
mol
-1
K
-1

Ri reaction rate of species i, kgm
-3

Si surface-adsorbed/site species, mol
si
, si
stoichiometric coefficient for site reactant and product

Stemp heat release/absorb due to reactions, Wm
-3

T temperature, K
U velocity, ms
-1

X mol fraction
Greek letters
br temperature exponent
density, kgm
-3

m dynamic viscosity, Kgm
-1
s
-1

rate of rth reaction

hpump pump efficiency
mass fraction of species i

Subscripts
b bulk
dep deposition
eff effective
g gas
i species i
r rth wall surface reaction
r, in reactant at inlet
r, out reactant at outlet
s solid/site
temp temperature
REFERENCES
Agrawal, S. and K. D. P. Nigam, 2001, Modelling of a coiled tubular chemical reactor,
Chemical Engineering Journal, 84 (3), 437-444.
Ajmera, S. K., C. Delattre, Schmidt, M.A., and Jense, K.F., 2002, Microfabricated cross-flow
chemical reactor for catalyst testing, Sensors Actuators B, 4155, 1-10.
Amador, C., A. Gavriilidis, Angeli, P., 2004, Flow distribution in different microreactor
scale-out geometries and the effect of manufacturing tolerances and channel
blockage, Chemical Engineering Journal, 101(1-3), 379-390.
tions


Arrighetti, C., S. Cordiner, Mulone, V., 2007, Heat and mass transfer evaluation in the
channels of an automotive catalytic converter by detailed fluid-dynamic and
chemical simulation, Journal of Heat Transfer, 129(4), 536-547.
Bejan, A. and M. R. Errera., 1997, Deterministic tree networks for fluid flow: Geometry
for minimal flow resistance between a volume and one point, Fractals, 5(4), 685-
695.
Carvalho, A. T., R. R. Lima, Silva, L. M., Simoes, E. W., Silva, M. L. P., 2009, Three-
dimensional microchannels as a simple microreactor, Sensors and Actuators, B:
Chemical, 137(1), 393-402.
Chen, Y. and P. Cheng., 2002, Heat transfer and pressure drop in fractal tree-like
microchannel nets, International Journal of Heat and Mass Transfer, 45(13),
2643-2648.
Chung, C. K., T. R. Shih, Chang, C. K., Lai, C. W., Wu, B. H., 2011, Design and experiments of
a short-mixing-length baffled microreactor and its application to microfluidic
synthesis of nanoparticles, Chemical Engineering Journal, 168 (2), 790-798.
Deshmukh, S. R., A. B. Mhadeshwar, Vlachos, D. G., 2004, Microreactor modeling for
hydrogen production from ammonia decomposition on ruthenium, Industrial
Engineering Chemistry Research, 43(12), 2986-2999.
Deutschmann, O., L. I. Maier, Riedel, U., Stroemman, A. H., Dibble, R. W., 2000, Hydrogen
assisted catalytic combustion of methane on platinum, Catalysis Today, 59(1-2),
141-150.
Jhnisch, K., V. Hessel, Lowe, H., Baerns, M., 2003, Chemistry in microstructured reactors,
Angewandte Chemie International Edition, 43, 406-446.
Kurnia, J. C., A. P. Sasmito, Mujumdar, A. S., 2011a, Laminar convective heat transfer for
in-plane spiral coils of noncircular cross section ducts: a computational fluid
dynamics study, Thermal Science, DOI: 10.2298/TSCI100627014K.
Kurnia, J. C., A. P. Sasmito, Mujumdar, A. S., 2011b, Numerical investigation of laminar
heat transfer performance of various cooling channel designs, Applied Thermal
Engineering, 31(6-7), 1293-1304.
Li, G., "Surface reactions in catalytic tubes." Fluent documentations.
Ni, Z., E. G. Seebauer, Masel, R. I., 2005, Effects of microreactor geometry on performance:
Differences between posted reactors and channel reactors, Industrial
Engineering Chemistry Research, 44(12), 4267-4271.
Ramanathan, K., Balakotaiah, V., D. H. West, 2003, Geometry effects on ignition in
catalytic monoliths, AIChE Journal, 50 (7), 1493-1509.
Regatte, V. R. and N. S. Kaisare, 2011a, Propane combustion in non-adiabatic
microreactors: 1. Comparison of channel and posted catalytic inserts, Chemical
Engineering Science, 66(6), 1123-1131.
Regatte, V. R. and N. S. Kaisare, 2011, Propane combustion in non-adiabatic
microreactors: 2. Flow configuration in posted microreactors, Chemical
Engineering Science, DOI: 10.1016/j.ces.2011.01.002.
Saber, M., J. M. Commenge, Falk, L., 2010, Microreactor numbering-up in multi-scale
networks for industrial-scale applications: Impact of flow maldistribution on the
reactor performances, Chemical Engineering Science, 65(1), 372-379.
Sasmito, A. P., Kurnia, J. C., Mujumdar, A. S., 2011, Numerical evaluation of transport
phenomena in a T-junction micro-reactor with coils of different
configurations, Industrial Engineering Chemistry Research, DOI:
10.1021/ie200139s.
Watts, P. and S. J. Haswell, 2005, The application of microreactors for small scale organic
synthesis, Chemical Engineering and Technology, 28(3), 290-301.
Yeom, J., D. D. Agonafer, Shannon, M. A., 2009, Low Reynolds number flow across an
array of cylindrical microposts in a microchannel and figure-of-merit analysis of
tions

micropost-filled microreactors, Journal of Micromechanics and Microengineering,
19(6), 5025.

Chapter 2
Enhanced Convective Mixing for Gaseous Microreactors

M. Shaker and H. Ghaedamini
Department of Mechanical Engineering, National University of Singapore
9 Engineering Drive 1, Singapore 117576

Contents
ABSTRACT .......................................................................................................................................................... 27
2.1. INTRODUCTION ........................................................................................................................................ 27
2.2. SIMULATION METHOD .......................................................................................................................... 28
2.2.1. Mathematical Model and Numerical Procedure ..................................................................................... 29
2.2.2. Basic of Chemical Reaction .............................................................................................................................. 31
2.2.3. Boundary Conditions ......................................................................................................................................... 32
2.2.4. Computational Procedure ................................................................................................................................ 33
2.3. RESULTS AND DISCUSION ..................................................................................................................... 34
2.3.1. Stochiometric Ratio ............................................................................................................................................ 34
2.3.2. Pre-Mixing and Uniform Concentration ..................................................................................................... 35
2.3.3. Effect of Redirection and Sudden Expansion ............................................................................................ 35
2.3.4. Effect of Split and Recombination and Impingement ........................................................................... 40
2.3.5. Effect of Reynolds Number on Yield ............................................................................................................. 41
2.4. CONCLUSIONS ........................................................................................................................................... 42

M. Shaker and H. Ghaedamini, Enhanced Convective Mixing for Gaseous Microreactors, in Mathematical
Modelling of Transport Processes, Ed. Sasmito, A.P.; Kurnia, J.C.; Jangam, S.V., 2011, ISBN - 978-981-08-
9179-4, Published in Singapore, pp. 25-46.

Shaker and Ghaedamini Enhanced convective mixing for gaseous microreactors
Mathematical Modeling of Transport Processes 27

ABSTRACT
This study investigates the steady, laminar flow field and reaction rates of several con-
figurations of a semi-T shaped microreactor by numerical simulations, governing con-
servation equations of mass, momentum, energy and spices were solved using Fluent 6-
3. The reaction is a catalyst surface based gaseous one between methane and air. Micro-
channel geometries can be divided into some innovative circular and rectangular con-
figurations which are designed to study and compare several effects such as those due to
recirculation, redirection, splitting of the flow and impingement. It is observed that the
rectangular designs display a better performance considering reactant utilization and
desired yield. Configurations with splitting and impingement zones have better per-
formance at the expense of higher pressure drop. Effect of pre-mixing of the reactant is
also considered by changing the inlet design. Effect of Reynolds number is to reduce the
yield due to shorter residence time in the microreactor..
2.1. INTRODUCTION
High performance microreactors have received special attention these years as
miniaturization has emerged as a need in industrial and medical applications like Bio
Micro-Electro-Mechanical Systems (BioMEMS) and lab-on-a-chip Microsystems [Wang et
al., 2011]. using devices working at sub millimeter scale, improved mass and heat trans-
fer rates can be achieved which enable us to proceed with reactions which are precisely
controlled and give the opportunity to selectively control the reaction process and
products [Nguyen and Wu, 2005; Aoki and Mae, 2006a].
In a microreactors two phenomena occur; mixing and reaction. Mixing in micro-
reactor is an important factor which needs special attention. Good mixing in a microre-
actor will result in a good reaction rate which it is desirable. In order to enhance mixing,
two approaches can be followed, passive mixing and active mixing [Hessel et al., 2005].
In passive mixing, we make use of the energy of the flow by designing the shape of the
channel to stretch and fold the flow [Aoki and Mae, 2006]. Even in some cases this can
generate chaotic advection [Tabeling et al., 2004]. All these will increase the interfacial
area between fluid segments and hence reducing the diffusion length. In an active mi-
cromixer, external forces help to enhance the mixing [Lam et al., 2005] this external
force can be a simple rotation [Dinh and Ogami, 2009], pulsation of the flow [Tesar,
2009], or an electrically exited flow [ Meisel and Ehrhard, 2006]. A recent review of sin-
gle phase micromixers has been by Kumar et al. which includes the recent improve-
ments and achievements in micromixers and microreactors[Kumar et al. 2011].
In the case of mixing at microchannels, due to the laminar regime in microchan-
nels, diffusion is the governing mode of mixing and mass transport[Aubin et al., 2010].
As a result, proposing special shapes which can enhance advection mixing is desirable.
This can be achieved by forcing the flow to go through special channel shapes which
manipulate the flow and by creating vortices in the flow field; it increases the contact
area between the species and reduces the diffusion length. For this purpose, methods
like splitting and rejoining the flow, chaotic mixing etc. can be implemented. It should be
noted that when the flow regime is laminar, the mixing is controlled by the diffusion.
Hence to have better mixing, reducing the diffusion length scale is the target. This can be

28 Mathematical Modelling of Transport Processes

achieved by adding some recirculation areas to the flow field, splitting and recombina-
tion of the fluid flow and etc.
Surely it is not possible to have a perfect design but it should be possible to have
a design which is least imperfect. In a microreactor, mixing and reaction occur at the
same time. The most optimum condition reaction completes exactly at the outlet of the
microreactor not sooner, the means that the residence time is equal to reaction time.
This is an ideal case and in our evaluations we will pay attention to it.
The case we examine via modeling in this investigation is a surface catalyst based gase-
ous reaction in a semi-T shaped microreactor. Due to the fact the diffusivity of gas
phases is relatively high; it is believed that T shaped micromixers are appropriate for
such applications. However, the numerical work on semi-T shaped microreactor shows
that, there is dependence on the mass flow volume and as the result Re. The reason is in
the fact that the reaction time scale and mixing time scale are the limiting factors which
at higher flow rates makes the T shaped microreactor efficiency to lower down. Thus, in
order to enhance flow mixing three geometric effects are to be considered.
Adding recirculation zones to the flow field by sudden expansions
Adding recirculation and redirecting the flow
Splitting and recombining the flow and impingement of streams
Based on the above three designs, eight types of semi-T shaped microchannels are mod-
elled which they can be classified as circular and rectangular designs.
2.2. SIMULATION METHOD
The model used in this investigation considers a semi-T shaped microreactor with eight
different designs to evaluate several phenomena which may enhance the mixing. Figure
2.1 shows these eight designs. As one can see these reactor designs can be grouped into
two categorizes viz, circular and rectangular. With such designs the effect of recircula-
tion, cross section change, impingement and splitting of the flow streams can be com-
pared. Based on the main subject of this investigation, which is the study a gaseous re-
action, the model reaction of air with methane in the presence of a catalyst (platinum)
coated surface is chosen. This reaction is very well studied and the mechanism of the
chemical reactions can be obtained through available literature. The inlet design of the
reactor plays an important role in reaction rate and since we want to study the effect of
geometry on the efficiency. It was decided to reduce the negative effect of poor inlet de-
sign by introducing the semi-T shaped microchannel fitted with two inlets for the air
and one inlet for methane. The reason for changing the inlet configuration is the fact that
by considering equal area for fuel and air inlet, it is necessary to increase the inlet veloc-
ity of air in order to obtain the necessary Stoichiometric ratio, defined with air to fuel
molar ratio of 9.52 for the reaction modelled in this study. Hence we changed the inlet
configuration and increased the inlet area for air as we need to feed much more air than
methane. Based on Fig.1, there are two inlets to the T shaped microchannel for air and
through the smaller inlet channel, the fuel enters the reactor. The reactants mix and re-
action occurs along the microchannel inside surface.

Figure 2.1. Different two dimensional channel configurations
2.2.1. Mathematical Model and Numerical Procedure
Mixing and reaction which occur in a microchannel lead to different transport
phenomena such as diffusion, convective mass and heat transfer, adsorption and de-
sorption on the surface of the channel etc. In order to translate these physical phenom-
ena to a numerical model, three basic conservation equations are invoked:
Conservation of mass, the continuity equation:
() = 0 (1)
Conservation of momentum:
( ) = p + ( +()

2
3
( )) (2)
Conservation of energy:
c
p
T = (k
eff
T) +S
temp
(3)
As the microchannel is a microreactor, chemical reaction and therefore conservation of
spices should be considered as the fourth conservation equation:


(
i
) = (D
i
i
) +R
i
(4)
where, is the fluid density, u is the fluid velocity, p is the pressure, is the dynamic vis-
cosity, T is temperature,
i
is the mass fraction of species i, D
i
is the diffusion coefficient
of species i, R
i
is the mass consumed or produced by reactions, c
p
is the specific heat,
k
eff
is the effective thermal conductivity and S
temp
is heat release/absorption due to
reaction.
Solving above four equations results in computation of the velocity, pressure and
temperature fields along channel length; also they lead to calculation of outlet mass frac-
tions of all spices including reactants and products.
In this study we assumed steady, laminar and Newtonian flow. Also it is assumed
that the miscible species mixture follows ideal gas law; so the thermodynamic proper-
ties of the mixture such as density (), viscosity (), thermal conductivity (k
eff
) and heat
capacity (c
p
) defined as follow:
The fluid density is given by the ideal gas law:
= pM RT (5)
where R is the universal gas constant and M is the molecular weight of the mixture fluid
given by:

M = (
CH
4
M
CH
4
+
H
2
M
H
2
+
O
2
M
O
2
+
H
2O
M
H
2O
+
CO
2
M
CO
2
+
CO
M
CO
+
N
2
M
N
2
)
1

(6)
The fluid mixture viscosity is computed from:

=
x
with , = CH
4
, H
2
, O
2
, H
2
O, CO, CO
2
, N
2

(7)
where x
,
are the mole fractions of species and , and
,
is given by the expression
below:

,
=
1
8
(1 +
M
)
1 2
[1 +(
(g)
(g)
)
1
2
(
M
)
1
4
]
2

(8)
The multi-component gas mixture thermal conductivity is computed using:
k
eff
= k
i
i
(9)
The gas mixture specific heat capacity c
p
is calculated by:
c
p
=
i
c
p,i i
(10)

2.2.2. Basic of Chemical Reaction
The heterogeneous reaction model considers the chemical reactions to occur on
the channel surface only, and the reactions involving surface deposition are defined as
distinct surface reactions and hence treated differently from bulk phase reactions in-
volving the same species. In the same way, the chemical species deposited on surfaces
are treated as distinct from the same chemical species. In this model, seven gas species
(CH4, O2, H2, H2O, CO, CO2 and N2), one bulk/solid species (Pt(b)) and eleven surface
species (H(s), Pt(s), O(s), OH(s), H2O(s), CH3(s), CH2(s), CH(s), C(s), CO(s), CO2(s)) are
considered [Deutschmann, O.; Maier, L. I.; Riedel, U.; Stroemman, A. H.; Dibble, R. W. ,
2000].
The gas phase species and surface species can be produced and consumed by surface
reactions and the general expression is given by:

g
i,r
G
i
N
g
i=1
+ b
i,r
B
i
N
b
i=1
+ s
i,r
S
i
N
s
i=1
K
r

g
i,r
G
i
N
g
i=1
+ b
i,r
B
i
N
b
i=1
+ s
i,r
S
i
N
s
i=1

(11)
where G
i
represents the gas phase species, B
i
represents the solid species, S
i
represents
the surface adsorbed species. g
i,r
, b
i,r
, s
i,r
are the stoichiometric coefficients for the

three reactant species respectively, and g
i,r
, b
i,r
, s
i,r
are the stoichiometric coefficients

for the three product species respectively. K
r
is the overall reaction rate constant.
The rate of reaction is given by:

r
= k
f,r
[G
i
]
wall
g
i,r
N
g
i=1
[S
i
]
wall
S
i,r

(12)
where k
f,r
is the reaction rate constant which is calculated using the Arrhenius equa-
tions given by:
k
f,r
= A
r
T
r
e
E
r
RT
(13)
and [G
i
]
wall
is the molar concentration on the wall. So the net molar rate of consumption
or production for each species i is represented by:

R
i,gas
= (g
i,r
g
i,r
)
N
rxn
r=1

r
i = 1,2,3, . . , N
g

R
i,bulk
= (b
i,r
b
i,r
)
N
rxn
r=1

r
i = 1,2,3, . . , N
b

R
i,site
= (s
i,r
s
i,r
)
N
rxn
r=1

r
i = 1,2,3, . . , N
s

(14)
Furthermore, it is assumed that on the wall surface the mass flux of each gas species is
balanced with its rate of production/consumption which is given by:

wall
D
i
i,wall
n
m
dep
i,wall
= M
,i
R
i,gas

i = 1,2,3, . . , N
g

(15)


[S
i
]
wall
t
= R
i,site
i = 1,2,3, . . , N
s
(16)
where m
dep
is the rate of mass deposition or etching due to surface reaction which is
given by:
m
dep
= M
,i
N
b
i=1
R
i,bulk
(17)
and [S
i
]
wall
is the site species concentration on the wall, defined as:
[S
i
]
wall
=
site
z
i
(18)
where
site
is the site density of the catalyst and z
i
is the site coverage of species i.
The gas concentration at the wall is calculated from the species mass fraction which is
expressed by:

[G
i
]
wall
=
wall
i,wall
M
,i

(19)
2.2.3. Boundary Conditions
Besides specifying basic set of conservation equations and characteristic of flow, we
need to define the boundary conditions to solve the mathematical model equations.
Definition of the boundary conditions have been done for inlet, outlet and channel wall;
temperature, spices mass fraction and velocity are known parameter for the inlets, pres-
sure and gradient of the temperature and gradient of the species mass fraction are
known for outlet and along channel wall, no slip condition and constant platinum sur-
face temperature are assumed as the boundary conditions.
Boundary conditions at inlet
1. At air side :
= 0.01, 0.05, 0.1 , 1, 3, 7
,
= 300 ,
2
= 0.21 &
2
= 1
2

2. At methane side:
= 0.01, 0.05, 0.1 , 1, 3, 7
,
= 300 ,

4
= 0.9 &
2
= 0.1
Boundary conditions at the outlet
= 1 ,
= 0 &
= 0
Boundary conditions along channel wall
1. No-slip condition
2. No species flux
= 1290 ,
2.2.4. Computational Procedure
In order to solve the above coupled equations, a commercial software (FLUENT 6.3)
based on the finite volume scheme was implemented. For the grid production, the Gam-
bit code was used to produce the quad 2D mesh which is then imported to Fluent. The
pressure based solving procedure is then implemented while the first order upwind spa-
tial discretization was used. the SIMPLE scheme used for the pressure-velocity coupling.
The reaction equations mechanisms were imported from CHEMKIN to Fluent. Table 2.1
contains the surface reactions postulated. Iterations were done until the residuals of the
energy, continuity, momentum and species were less than 1e-6. Grid study was carried
out by reducing the mesh size until the difference in key variables between two similar
cases with different mesh sizes was less than 1%. This grid study showed that a mesh
size of 0.002 for the mapped quad element is suitable for all cases examined here.
Table 2.1. Surface reaction mechanism
[Deutschmann et al., 2000 ]
No Reaction Ar Br Er (J/kmol)
1 H2 + 2Pt(s) => 2H(s) 4.36e7 0.5 0
2 2H(s) => H2 + 2Pt(s) 3.7e20 0 6.74e7
3 O2 + 2Pt(s) => 2O(s) 1.8e17 -0.5 0
4 O2 + 2PT(s) => 2O(s) 2.01e14 0.5 0
5 2O(s) => O2 + 2Pt(s) 3.7e20 0 2.13e8
6 H2O + Pt(s) => H2O(s) 2.37e8 0.5 0
7 H2O(s) => H2O + Pt(s) 1e13 0 4.03e7
8 OH + Pt(s) => OH(s) 3.25e8 0.5 0
9 OH(s) => OH + Pt(s) 1e13 0 1.93e8
10 H(s) + O(s) => OH(s) +
Pt(s)
3.7e20 0 1.15e7
11 H(s) + OH(s) => H2O(s) +
Pt(s)
3.7e20 0 1.74e7
12 OH(s) + OH(s) => H2O(s) +
O(s)
3.7e20 0 4.82e7
13 CO + Pt(s) => CO(s) 7.85e15 0.5 0
14 CO(s) => CO + Pt(s) 1e13 0 1.25e8


15 CO2(s) => CO2 + Pt(s) 1e13 0 2.05e7
16 CO(s) + O(s) => CO2(s) +
Pt(s)
3.7e20 0 1.05e8
17 CH4 + 2Pt(s) => CH3(s) +
H(s)
2.3e16 0.5 0
18 CH3(s) + Pt(s) => CH2(s) +
H(s)
3.7e20 0 2e7
19 CH2(s) + Pt(s) => CH(s) +
H(s)
3.7e20 0 2e7
20 CH(s) + Pt(s) => C(s) + H(s) 3.7e20 0 2e7
21 C(s) + O(s) => CO(s) + Pt(s) 3.7e20 0 6.28e7
22 CO(s) + Pt(s) => C(s) + O(s) 1e17 0 1.84e8
23 OH(s) + Pt(s) => H(s) +
O(s)
1.56e18 0 1.15e7
24 H2O(s) + Pt(s) => H(s) +
OH(s)
1.88e18 0 1.74e7
25 H2O(s) + O(s) => OH(s) +
OH(s)
4.45e20 0 4.82e7

2.3. RESULTS AND DISCUSION
For heterogeneous catalytic reactions, the desired reaction is defined in terms of the
maximum utilization of the reactants and hence maximum production of desirable spe-
cies. This naturally depends on the concentration of the fresh reactants which are in
contact with the catalyst to yield the desired species. Catalyst improves the reaction rate
by decreasing required activation energy of reaction. Thus, it is important that, as much
as possible, the reactants are in contact with the catalyst-coated surface simultaneously.
Therefore by increasing the contact area of micro-channel we should be able to enhance
yield of the desirable reaction.
In catalyst based micro reactors the basic mechanism of the reaction depends on
the near surface conditions. This includes adsorption, i.e. adsorption of fuel and oxygen
onto the platinum-coated surface, surface reactions, i.e. chemical reactions of the ad-
sorbed species, as well as the desorption reaction, i.e. desorption of the resulting prod-
ucts, which overall cause to heat release. Besides these phenomena, convective heat and
mass transfer between catalyst coated surface and gas boundary layer are the other
mechanisms which take place there. Therefore in addition to the effect of increased cata-
lyst surface, increasing the diffusion rate, convective mass and heat transfer rates will
contribute to improved yield. Satisfying Stoichiometric ratio, pre-mixing and uniform
concentrations are the other factors which should be considered.
2.3.1. Stochiometric Ratio
For all reactions stoichiometric ratio is one of the key factors, because the exhaust
species will be different for different values of the coefficients. It is important to distin-
guish between rich and lean conditions of combustion. Beside the importance of the type
of exhaust products, lean combustion may lead to low fuel consumption which is one of
the features of the reaction.

To investigate the effect of stoichiometric ratio, we modeled two reactions with O2
inlet mass fractions of 0.21 and 1. Our results showed that changing the mass fraction of
O2 in inlet from 0.21 to 1 improved the CH4 and O2 (reactants) utilization by 3.48 and
1.24 times, respectively. Also, it changed the composition of the exhaust. In the first case
(mass fraction of O2=0.21) there was a higher percentage of carbon monoxide but by
increasing the mass fraction of O2, production of CO2 composition increased, which is
our desired product in combustion of methane.
2.3.2. Pre-Mixing and Uniform Concentration
To reach a uniform concentration of products at the outlet, an indicator of a com-
plete reaction, it is important that a uniform mixture of the reactants to be in contact
with catalyst surface. Non-uniform mixture would cause low air-to-fuel ratio on one side
of the catalytic surface and thus it leads to an increase in the production of undesirable
products at the surface. Thus, it is suggested that we first mix the reactants together and
then pass the mixture of reactants through the reactor zones (channel coated with cata-
lyst) to yield desired products. It would lead to uniform concentration at the outlet.
Without premixing of reactants a distinct concentration difference is observed at the
outlet of the reactor while much better uniformity in concentration is attained, as ex-
pected, with good premixing.

Figure 2.2. Effect of premixing on a T shaped microreactor-CO2 mass fraction fields
2.3.3. Effect of Redirection and Sudden Expansion
Figure 2.3 shows the consumption of CH4 as one of the reactants vs the Reynolds
number. Increment in area coated with catalyst and vortices which appear along chan-
nel walls because of sudden expansion will lead to increment in the rate of reactant
utilization comparing to strait T junction channel.
As can be seen, at small Re the utilization for all cases is near 100% which is rea-
sonable since the mean residence time is much greater than diffusion time and the reac-


tion time. This can be better seen from the CO2 mass fraction contour at Figure 2.4.
Therefore there is no difference in utilization of the reactants with different microchan-
nel geometries. The superiority of adding redirection to the flow field shows itself at
relatively higher Reynolds numbers. From Figure 2.3, it can be seen that for Re greater
than 100, redirection of the flow will result in better utilization of reactants. Due to the
fact that redirection will increase the residence time as the flow has to go through a
longer path, also from the velocity profile (Figure 2.5) at Re=90 it is obvious that the
two dissimilar counter rotating vortices in a type2 configuration will occur at higher
Reynolds numbers which are stronger and thus, a shorter diffusion length plus greater
interfacial area will result in a better mixing.

Figure 2.3. CH4 utilization at different Reynolds number for rectangular designs
0
10
20
30
40
50
60
70
80
90
100
0 200 400 600 800
c
h
4

U
t
i
l
i
z
a
t
i
o
n

%
Re. Number
Type 1
Type 2

Figure 2.4. CO2 mass fraction contour plots for type1 configuration.


Figure 2.5. Velocity contour plots at Re=90 for a type1 and type2 configuration
Figure 2.6 shows a similar trend for the circular types as the Reynolds number in-
creases. The redirection can increase the consumption of CH4 but it can be seen that in
comparison with rectangular design the increment is less. This is due to the fact that the
recirculation area in the circular design has reduced into one vortex instead of two,
which means smaller contact area and poorer mixing. This is obvious from the Figure
2.7 which shows the velocity field.

0
10
20
30
40
50
60
70
80
90
100
0 100 200 300 400 500 600 700
c
h
4
U
t
i
l
i
z
a
t
i
o
n

%
Re. Number
Type 3
Type 4

Figure 2.6. CH4 utilization for circular designs

Figure 2.7. Velocity contours at Re=90 for a type3 and type4 configurations.
Figure 2.8 shows the pressure drop of the first four types. It can be seen that the chan-
nels with more complicated flow structure, type 2 and 4, have a higher pressure drop.
By increasing the Reynolds number, pressure drop will increase but the increment of
circular designs is greater in compare with rectangular designs. A comparison between
rectangular and circular designs reveals that the rectangular shapes are superior as they
have the same order of utilization while the pressure drop is much lower.

Figure 2.8. Pressure drop
0
5000
10000
15000
20000
25000
30000
35000
0 50 100 150 200 250 300
P
r
e
s
s
u
r
e

D
r
o
p

(
P
a
)
Re. Number
Type 1
Type 2
Type 3
Type 4


2.3.4. Effect of Split and Recombination and Impingement
In order to see the effect of splitting the flow, channel configuration of types 5 to 8
are designed to have even and uneven splitting of the flow which it gives us the oppor-
tunity to study the case with more generality. Figure 2.9 shows the CH4 utilization of six
configurations, it is obvious that the splitting types which have more complicated ge-
ometries have better performance. However, the results show that type 5 and 7 designs,
rectangular designs with splitting, have the highest reactant utilization among all the
microchannel geometries examined. The main trend of all configurations is that the
utilization of reactants decreases with increasing Reynolds number. There are several
implications here. First, splitting of the two reactant streams and their recombination in
type 5 to 8 geometries increases the contact area with the catalyst-coated surface and at
the same time increases the diffusion because splitting of fluid streams contributes to
mixing by increasing the interfacial area and by shortening the diffusion length. Increase
in both factors leads to increase in the reaction yield. Second, impingement of streams in
splitting configurations leads to increase of heat and mass transfer coefficients between
the catalyst-coated area and the gas flow, thus increasing the adsorption of reactants
and desorption of products. Third, centric and eccentric configurations have a same
production rate as it is obvious from Figure 2.9 and Figure 2.10. The reason is that for
two cases, resistance time as well as the diffusion length are as the same order and thus
equal utilization is observed however, pressure drop of centric configurations are less
than eccentric configurations, Figure 2.11. Fourth, with increasing Reynolds number,
the effect of vortex generation and impingement enhances the mass transfer and reac-
tion rates. However, increasing the Reynolds number also leads to shorter residence
time of fluid in the channel; fluid molecules may have not enough time to mix and react,
and can cause reduction of reaction rate. When the negative effect of residence time is
greater than the positive effect of impingement jet and vortex generation, the overall
rate of reactant utilization will decrease.

Figure 2.9. CH4 utilization vs. Reynolds number.
0
10
20
30
40
50
60
70
80
90
100
0 50 100 150 200 250 300
c
h
4

U
t
i
l
i
z
a
t
i
o
n

%
Re. Number
Type 1
Type 3
Type 5
Type 6
Type 7
Type 8

Figure 2.10. Mass fraction of centric and eccentric configurations at Re=90.

Figure 2.11. Pressure drop of centric and eccentric designs.
2.3.5. Effect of Reynolds Number on Yield
Figure 2.12 show that the Reynolds number has an effect on the type of the final prod-
ucts. There is higher percentage of carbon monoxide with increasing Re, which is not the
final desired product in combustion of methane. With increasing Re, since the residence
time of flow in microchannel decreases, this leads to incomplete reaction between
0
5000
10000
15000
20000
25000
30000
35000
40000
45000
0 50 100 150 200 250 300
P
r
e
s
s
u
r
e

D
r
o
p

(
P
a
)
Re. Number
Type 5
Type 6
Type 7
Type 8


methane and air. It can be seen from Table 2.1, CO2 will be generated when CO(s) and
O(s) react on the surface of catalyst. When there is not enough time for the presence of
flow in channel, some of the surface reactions of Table 1 will occur incompletely; there-
fore the ratio of co production to CO2 will increase. Interestingly type 1 microchannel
has minimum undesirable products at high flow rates. However, it should be noted that
the overall production of type1 is much less than in other configurations e.g. type 5 and
6. Moreover, the superiority of the rectangular design again is observable here.

Figure 2.12. Percentage of produced CO over CO2
2.4. CONCLUSIONS
This paper simulates the heterogeneous reaction of methane and air in the presence
of platinum as a catalyst in eight different microchannel configurations by using a nu-
merical approach. The reaction takes place near the surface of microchannel in a laminar
flow region. Effects of recirculation, redirection, splitting and recombination as well as
impingement were examined. It was found that channel geometry with the effect of
splitting and impingement jet stream has a better performance, while it has higher pres-
sure drop, as a reactor as a result of maximum reactant utilizations. It was observed that
the manipulation of the flow shows its effect when we are dealing with relatively higher
flow rates, as in small flow rates; the residence time is much greater than diffusion time.
Although, increasing the Reynolds number for all tested configurations caused weaker
performance due to shorted residence time. By considering reactants utilization and
product composition at the outlet, rectangular designs show a distinct by better per-
formance compared to the circular designs, at the expense of higher pumping power.
However, splitting the flow by centric and eccentric configurations, it was shown that
centric designs has a better performance as they had less pressure drop and equal reac-
tants utilization than centric configurations.
For the future investigations, it is recommended to study similar shapes in three dimen-
sional formats. In this way, it is anticipated to observe secondary flows which will en-
hance mixing and hence reaction rates. Moreover, shapes with sudden expansion are
believed to create chaotic advection which in a steady two dimensional format is not
feasible. Thus it may be a wise idea if this effect is also considered in future works. In
current work it is tried to manipulate the inlet areas so that the Stoichiometric ratios can
0
10
20
30
40
50
60
70
80
90
100
0 100 200 300 400 500 600 700
c
o
/
c
o
2

%
Re. Number
Type 1
Type 2
Type 3
Type 4
Type 5
Type 6

be satisfied. Another idea can be a combination of area-velocity manipulation and study-
ing their characteristics. It is also worth to more explore the splitting configuration by
doing a through parameter study on the shapes to find out optimum geometric ratios.
NOMENCLATURES
pre-exponential factor
bulk/solid species, mol
,
,
"
stoichiometric coefficient for bulk reactant, and product
specific heat, Jkg

1
K
1

diffusivity of species i, ms
2

activation energy for the reaction, Jkgmol
gas species, mol
,
,
"
stoichiometric coefficient for gas reactant, and product
effective thermal conductivity, Wm

1
K
1

,
reaction rate constant using Arrhenius expression
mean molecular mass
net rate of mass deposition, kg

p Pressure, Pa
R universal gas constant, Jkg
1
mol
1
K
1

reaction rate of species i, kgm

3

surface-adsorbed/site species, mol
,
,
"
stoichiometric coefficient for site reactant, and product
heat release/absorb due to reactions, Wm

3

temperature, K
velocity, ms
1

mol fraction
Greek letters

temperature exponent


density, kgm
3

dynamic viscosity, Kgm
1
s
1

rate of rth reaction
mass fraction of species i

Subscripts

bulk
deposition
effective
gas
species i
rth wall surface reaction
solid/site
Temperature
REFERENCES
Aoki, N. and K. Mae, 2006a, Effects of channel geometry on mixing performance of mi-
cromixers using collision of fluid segments. Chemical Engineering Journal, 118 (3), pp.
189-197.
Aoki, N. and K. Mae, 2006b, Improvement of product yield and selectivity in mi-
croreactors by combining fluid segments of different concentrations and sizes. Studies in
Surface Science and Catalysis, 159, pp. 641-644.
Aubin, J., M. Ferrando, and V. Jiricny, 2010, Current methods for characterising mixing
and flow in microchannels. Chemical Engineering Science, 65 (6), pp. 2065-2093.
Deutschmann, O.; Maier, L. I.; Riedel, U.; Stroemman, A. H.; Dibble, R. W. , 2000, Hydrogen
assisted catalytic combustion of methane on platinum, Catalysis Today, 59 (1-2), pp.
141-150.
Dinh, T.X. and Y. Ogami, 2009, An Active Micromixer Enhances Mixing by a Rotat-ing
Shuttlecock Rotor. Proceedings of the ASME Fluids Engineering Division Summer Con-
ference, FEDSM 2008 1 (PART A), pp. 569-575.
Hessel, V., H. Lowe, and F. Schonfeld, 2005, Micromixers - a review on passive and active
mixing principles. Chemical Engineering Science, 60 (8-9 SPEC. ISS.), pp. 2479-2501.

Kumar, V., M. Paraschivoiu, and K.D.P. Nigam, 2011, Single-phase fluid flow and mixing
in microchannels. Chemical Engineering Science, 66 (7), pp. 1329-1373.
Lam, R.H.W., et al., 2005, Integrated vortex micropumps and active micromixers in
polymer substrates for automating bio-fluidic manipulation.27th Annual International
Conference of the IEEE Engineering in Medicine and Biology Society, Vols 1-7, pp. 1297-
1300.
Meisel, I. and P. Ehrhard, 2006, Electrically-excited (electroosmotic) flows in mi-
crochannels for mixing applications. European Journal of Mechanics B-Fluids, 25(4), pp.
491-504.
Nguyen, N.T. and Z.G. Wu, 2005, Micromixers - a review. Journal of Micromechanics and
Microengineering , 15 (2), pp. R1-R16.
Tabeling, P., et al., 2004, Chaotic mixing in cross-channel micromixers. Philoso-phical
Transactions of the Royal Society A: Mathematical, Physical and Engineering Sci-ences,
362 (1818), pp. 987-1000.
Tesar, V., 2009, Oscillator micromixer. Chemical Engineering Journal, 155 (3), pp. 789-
799.
Wang, S.T., et al., 2011, Highly Efficient Capture of Circulating Tumor Cells by Using
Nanostructured Silicon Substrates with Integrated Chaotic Micromixers. Angewandte
Chemie - International , 50 (13), pp. 3084-3088.

Tong Wei and Wang Yue. Modeling Conjugate Heat and Mass Transfer Between a Laminar Impinging Jet and
Planar and Curved Thin Slabs of Wet Solids, in Mathematical Modelling of Transport Processes, Sasmito, A.P.;
Kurnia, J.C.; Jangam, S.V., 2011, ISBN - 978-981-08-9179-4, Published in Singapore, pp. 47-60.

Chapter 3
Modeling Conjugate Heat and Mass Transfer between a Laminar
Impinging Jet and Planar and Curved Thin Slab of Wet Solids

Tong Wei and Wang Yue
A0049232@nus.edu.sg
A0066388@nus.edu.sg

Contents
3.1. INTRODUCTION .................................................................................................................................... 49
3.2. MATHEMATICAL MODEL ................................................................................................................... 49
3.2.1. Governing Equations .............................................................................................................................................. 49
3.2.2. Constitutive Relations............................................................................................................................................ 51
3.2.3. Boundary Conditions ............................................................................................................................................. 53
3.3. NUMERICS .............................................................................................................................................. 53
3.4. RESULTS AND DISCUSSION ............................................................................................................... 54
3.5. CONCLUSION REMARKS AND FUTURE WORK ............................................................................ 58
NOMENCLATURE STYLE ............................................................................................................................ 59
REFERENCE .................................................................................................................................................... 60

Tong Wei, Wang Yue- Modeling Conjugate Heat and Mass Transfer Between a Laminar Impinging
Jet and Planar and Curved Thin Slab of Wet Solids

3.1. INTRODUCTION
Drying technique is widely used in many industrial areas and daily life, such as food
preservation, wood processing, paper and washing powder drying. Various kinds of dry-
ing methods have been investigated and implemented for different materials. Impinging
jet is one of these methods and commonly used in a great number of major industrial
areas for various drying, cooling, and heating processes due to their high local heat and
mass transfer coefficients and processing continuously. One of its industrial applications
is thermal drying of continuous sheets of materials, such as paper, textiles, and food
products (Mujumdar, 2007).
Over the past decades, extensive research works have been conducted on drying
process by using impinging jet. These studies mainly focused on heat and mass transfer
characteristics under different drying conditions. The modes of air input have great in-
fluence on the drying performance; Poh et al. (2005) conducted a study of heat transfer
in an impinging jet with a pulsed laminar flow and found that the separation is asso-
ciated with Nusselt number during the oscillation cycle, and the vortices within one os-
cillation cycle is influenced by the pulsed inlet velocity. Xu et al. (2010a) studied the heat
transfer under a pulsed slot turbulent impinging jet and found that large temperature
difference could cause a significant change of local Nusselt number. Xu et al. (2010b)
also performed a research on the heat transfer enhancement in a turbulent impinging jet
with intermittent pulsation and reported that the intermittent pulsation of the jet could
enhance heat transfer significantly under a wide range of conditions. There are also
some studies on the configuration of the jet nozzle. Bula et al. (2000) analyzed the local
heat transfer variation by using a finite element method to simulate the conjugate heat
and mass transfer to a heated disk impinged by a laminar liquid jet, they found that the
disk thickness as well as the heating sector locations strongly influence the local heat
transfer rate at the same time. Beitelmal et al. (2000) carried out an experimental study
on the influence of the inclination of an impinging two-dimensional air jet on the heat
transfer from a heat flat plat and the results were presented that the local Nusselt num-
bers were determined as a function of several parameters of the nozzle.
Most of these studies were conducted to investigate the heat and mass transport
phenomena with a planar shape substrate with a certain thickness. Few have studied the
influence of a non-planar or curved shape of substrate. That is why in this study, a
curved substrate of concave shape and convex shape was used. This study investigates
the conjugate heat and mass transfer performance of a curved substrate during its dry-
ing process within an impinging jet with the aim of observing the influence of shape and
thickness on the drying performance.
3.2.1. Governing Equations
Here, we consider a wet solid curved substrate placed at bottom of the impinging
jet channel and a laminar air flow comes into the channel form the top nozzle. The confi-
guration of the impinging jet for different substrate shapes are shown in Figure 3.1; L is
half length of the channel, Z is the channel height, D is the nozzle width, Ls is the half length of the
50 Mathematical Modeling of Transport Processes

substrate, and Hs is the height of the substrate. The target substrate is chosen as potato slab.
The basic mechanism is that the drying air comes from the inlet normally to the product
surface and passes the heat to the surface of the substrate by convection; then within in
the product, the heat is transferred by conduction and induces the evaporation of mois-
ture to water vapor which comes to the substrate surface by diffusion and then is car-
ried away by the drying air by convection. These processes are taking place simulta-
neously.
(a)
(b)

(c) (d)

Figure 3.1. Schematic of different configuration of substrate, (a) concave shape, (b)
convex shape, (c) outer planar, (d) inner planar.
Some assumptions are addressed to establish the mathematical model. The drying
substrate is compact and homogeneous with uniform initial temperature and moisture
content. The drying air is applied continuously to the product. Within the substrate, the
diffusivity of water vapor is 100 times of liquid water. Also neither shrinkage nor de-
formation happens. The model is reduced from 3D to 2D based on the assumption that
span-wise direction variation is negligible compared to the cross-sectional variation.
Then, the conservation equations for mass, momentum, and energy for both the sub-
strate and air flow (De Bonis and Ruocco, 2008) are given as follow:
For substrate:

( )
l
lb l l
c
D c Kc
t
+ =
(1)

( )
v
vb v v
c
D c Kc
t
+ =

(2)
( )
b pb b
T
c k T q
t
+ =
(3)
For air:

= u 0

(4)

( )
v
va v v
c
D c c
t
+ =
u (5)

2
( )
a
p
t

+ = +
u
u u u

(6)

( )
a pa a a pa
T
c k T c T
t

+ =
u

(7)
where c
l
and c
v
are the molar concentration of liquid water and water vapor within the
substrate, D
lb
and D
vb
are the diffusivity of liquid water and water vapor inside the sub-
strate, D
va
is the diffusivity of water vapor in air, K is the rate of water evaporation,

and
are the density of air and substrate, u is the air velocity, p is the pressure, is the
dynamic viscosity of the fluid, c
pa
is the specific heat of air, ka is the air conductivity, and
T is the temperature. These two sets of equations are coupled by T and c
v.
3.2.2. Constitutive Relations
Thermo-physical properties of air are given as polynomial functions of tempera-
ture by fitting the data provided in (Kays et al., 2005). The air density is defined by

5 2 2
1.076 10 1.039 10 3.326,
air air air
T T

= +
(8)
while the air dynamic viscosity is given by
15 3 11 2 8 7
5.21 10 4.077 10 7.039 10 9.19 10 .
air air air air
T T T

= + +
(9)
The thermal conductivity of air is calculated from

10 3 7 2 4
4.084 10 4.519 10 2.35 10 0.0147.
air air air air
k T T T

= +
(10)
The specific heat of air is defined as

6 3 3 2
,
4.647 10 4.837 10 1.599 1175.
p air air air air
c T T T

= + +
(11)
The heat of evaporation is given by

1000( 2.394( 273.15) 2502.1).
fg
h T = +
(12)
The thermo-physical properties of wet solid substrate are given as follow.
Density (Srikiatden and Roberts, 2007)

,
(1 )
.
1
b ref
b
X
SbX
+
=
+
(13)
Specific heat (De Bonis and Ruocco, 2008)

,
1750 2345( ).
1
p b
X
c
X
= +
+
(14)
Thermal conductivity

3
0.049 47 1 1 0.611
exp[ ( )] .
1 8.3143 10 273.15 335.15 1+
b
s
X
k
X T X
= +
+ +
(15)
Diffusivity of liquid water and water vapor

6
0.0725 2044
1.29 10 exp( ) exp( ),
273.15
lb
s
D
X T

=
+
(16)

100 .
vb lb
D D =
(17)
Heat of wetting (heat to evaporate bound water)

6 4 6 3 5 2 4
8.207 10 4.000 10 6.161 10 2.368 10 1163.
w
H X X X X = + + +
(18)
for
0.01 0.2 X

Total heat for evaporation

.
evap fg w
h h H = +
(19)
The calculation of moisture content within the substrate is given by

mass of water
.
mass of dry product
l
s
X

= =
(20)

mass of water
.
mass of wet product 1
l l
s l b
X
W
X

= = = =
+ +
(21)
Equilibrium moisture content (GAB model)

, 0.0209, 0.976, 4.416.
(1 )( )
m w
e m
w w w
X CKA
X X K C
KA KA CKA
= = = =
+
(22)
Free moisture content

.
free e
X X X =
(23)
Cooling rate due to evaporation (De Bonis and Ruocco, 2008)

,
evap l l
q h M Kc
=
(24)

0
.
Ea
RT
K K e
=
(25)
Diffusivity of water vapor in air

6 8 10 2
2.775 10 4.479 10 1.656 10 .
va
D T T

= + +
(26)
Relation of moisture content to concentration of water inside substrate

,
(1 )
.
1 1
b ref
w b
X
X
W
X SbX

+
= =
+ +
(27)

2
,
2
( )
.
1
b ref
w w
X X
M c
SbX X SbX
+
=
+ + +
(28)

, 2 2
( 1) 1 ( ).
b ref
w w
SbX Sb X X X
M c
+ + + = +
(29)

, , 2
( ) ( 1 ) 1 0.
b ref b ref
w w w w
Sb X Sb X
M c M c

+ + + =
(30)

2
4
.
2
b b ac
X
a

=
(31)

,
( ).
b ref
w w
a Sb
M c
=
(32)

,
( 1 ).
b ref
w w
b Sb
M c
= +
(33)

1. c =
(34)
3.2.3. Boundary Conditions
Boundary conditions for the substrate are summarized as follows: inlet,
,
in
u u =
0, v = ,
in
T T = ,
in
RH RH =
0,a
v v
c c = outlet, ,
out
p p = ( ) 0,
v
D c = n ( ) 0 k T = n wall,
0, u v = = ( ) 0,
v v
D c c + = n u ( ) 0. k T = n n

is the normal vector, u and v are the
inlet velocity, T

is the air temperature, RH

is the air relative humidity, c
v
is the specific
heat of water vapor in air, p is the pressure, D is the diffusivity of vapor, and k is the air
thermal conductivity. The values for these variables are summarized in Table 3.1.
Table 3.1. Base case and operating parameters
Parameter Value
Temperature of the inlet,
in
T ( C ) 45 and 80
Inlet velocity,
in
V (
1
m s
) 0.5, 2, and 5
Reference substrate density,
, b ref
(
3
kg m
) 1420
Molar mass of water,
l
M (
1
kg mol
) 0.018
Initial moisture content of substrate,
0
X 4.6
Air density at 45 C ,
,45 a c

(
3
kg m
) 1.11
Air density at 80 C ,
,80 a c

(
3
kg m
) 1.0
Activation energy for cooling, Ea (
1
kJ mol
)
48.7
Sb 1.4
3.3. NUMERICS
The computational domains were established in Gambit 2.3.16. With which the
meshing is created, the boundary is specified, and thereby the computational domain is

determined. Three different mesh sizes of 110
5
, 210
5
, and 310
5
were conducted and
compared in terms of temperatures, moistures, and velocities to ensure a mesh inde-
pendency. We found that the mesh size of 210
5
gives a better result.
The mathematical model discussed above with its boundary conditions were then
solved by using commercial Finite Volume Solver Fluent 6.3.26 and user-defined func-
tions (UDFs) written in C++ language to specify the thermo-physical properties of the
fluid and all the boundary conditions.
3.4. RESULTS AND DISCUSSION
A number of parameters have been monitored that will directly affect the heat and
mass transfer in the jet vicinity and thus indirectly influence the flow in the periphery of
the substrate. They are inlet air velocity, temperature, and the thickness of the slab is
simulated for two dimensions: 3.2mm and 0.7mm, respectively. The simulation results
for this study are presented as contours of moisture content in Figure 3.2, velocity pro-
files in Figure 3.3 and temperature profiles in Figure 3.4 in their related domains at dif-
ferent drying moments. The drying kinetics is presented for different shapes of slabs in
Figure 3.5 and Figure 3.6. All the simulations under a constant air relative humidity with
a value of 17.5% and the initial temperature and moisture content of the substrate are
kept as a constant value of 4.6 and 298K, respectively.

(1) (2)

(3) (4)

Figure 3.2. Contours of moisture content for the four shapes (a. convex, b. concave,
c, planar outside, d. planar inside) of thin slabs of wet solid at different drying
moment (1). 10mins. (2). 20mins. (3). 60mins. (4). 120mins.

The moisture content distributions within the substrate as shown in Figure 3.2 are
plotted at 4 time points. It can be observed that, initially, the diffusion happens in the
vertical direction more than in the horizontal direction which can be explained by the
gradient of moisture content in the slabs, with the evolution of time, the middle region
where the heating effect is stronger compared with two sides has faster drying speed
until such a moment when the vertical gradient of moisture content becomes negligible
and the lateral diffusion becomes the dominant. Contrary to the concave shape substrate,
in which the moisture content is accumulated more in the middle region, the convex one
has more moisture on the two sides since heat is less efficiently transferred on the two
sides. There is an interesting phenomenon that the outside placed planar case has two
regions of local minimum moisture content, while the inside placed planar one only has
one region of local minimum moisture content, and may need to be studied further.
These can be considered as the influence of geometry on the heat and mass transfer,
more specifically, the curve shape that is in contact with air. The non-uniformity of mois-
ture distribution is the result of non-uniform heat transfer.

Figure 3.3. Contour of steady state velocity in the impinging jet of four shapes (a).
convex shape. (b). concave shape. (c). planar outside shape. (d). planar inside
shape of the thin slabs of wet solid.
Figure 3.3 shows the velocity magnitude contours of several cases. The difference
in velocity profile can be seen as the direct result due to the change of geometry, since
the results are given under fixed inlet air condition. The velocity tends to be uniform as
it is further away from the drying slabs. The most obvious change happens in the region
close to the slabs. It can be observed that for inside placed convex case and outside
placed planar case, the flow is smoother, experiencing less re-circulation than the out-
side placed concave slab and inside placed planar slab. With thickness change, this phe-
nomenon becomes more significant for the concave substrate of 0.7mm thickness.
Figure 3.4 shows the temperature distribution in the chamber at different time
points, i.e. 10mins, 20mins, 60mins, and 120mins. There appears not much difference
across time for the temperature profiles; the reason is that the temperature will quickly
get to steady state as the velocity field. Less gradient of temperature happens in curved-
shaped slabs although their geometry has greater non-uniformity. The average tempera-
ture is also higher in curved-shaped cases. The aforementioned two are the results of
more efficient heat transfer in these curved shapes.

(1) (2)

(3) (4)

Figure 3.4. Contours of temperature of four shapes (a. convex, b. concave, c, planar
outside, d. planar inside) for the thin slabs of wet solid at different drying moment.
(1). 10mins. (2). 20mins. (3). 60mins. (4). 120mins.
The moisture content- time plots shows the trend of the average moisture content
in the substrate changing with time as shown in Figure 3.5. It can be seen that thickness
can affect the drying speed, the substrate with a thickness of 0.7mm is faster in drying
speed than the one of 3.2mm if we compare the concave shape and planar outside one as
a pair; convex shape and planar inner-side one as a pair. The reason is that it will take
less time for the vapor molecules to move from the inner core to the surface. The diffu-
sion-controlled stage modulates the time by increasing or decreasing the molecule tra-
velling length since the diffusion coefficient is not much different in its scale and value.
However, with the same thickness, the difference is more significant of the case with
thickness of 3.2mm than the one with thickness of 0.7mm. Therefore, the geometry of
the substrate will have greater influence on thicker substrate, under the same condition
of inlet air.
Further, based on Figure 3.3 and Figure 3.5, we postulate the reasons for the ob-
served phenomena. It is perceivable that the convex case outperforms the inner planar
case due to three main reasons: (1) the exchange surface with the fluid domain is
slightly increased compared with the planar one, although not that significant, (2) the
heating surface is shorter in distance to the heating source in the convex case, (3) the re-
circulation flow happened in the planar case deteriorates the heat transfer in the stagna-
tion region. This comparison shows more significant difference with the slab thickness
of 3.2mm than that with thickness of 0.7mm. Moreover, for specific shape in this study, it
is observed that there are some differences between the results of slabs with thickness

of 3.2mm and 0.7mm. In the 3.2mm thickness case, the concave one outperforms the
planar slightly and it is the reverse situation in the 0.7mm thickness case. The possible
reason is that, since part of the reversal flow will go to the downstream, while part of the
reversal flow will join in the main flow that is coming from the impinging jet. This flow
re-circulation will bring part of the air containing moisture back to the surface, which
hinders the mass transfer; at the same time, the reversal flow with low temperature
mixed with the main air jet will decrease the overall temperature of the air jet before it
arrives at the surface of the slabs. Therefore, even though the contact area is increased
in concave case, this effect should be compared with that of re-circulation, so that under
some conditions, e.g. thickness of 0.7mm slabs, this balance of effect is broken and is
dominated by the flow circulation whereas for other conditions, e.g. thickness of 3.2mm,
the effect is dominated by the increased effective contact surface area with air.

Figure 3.5. Drying kinetics for different dryer containers for substrates with
thickness (a). 3.2mm, (b). 0.7mm under the same condition
The average Nusselt number and Nusselt number at stagnant point are also calcu-
lated that are given by (Incropera and DeWitt, 2002)
Average Nusselt number
0
( )
( )
R
Nu x dx
Nu avg
R
=
(35)
where Nusselt number is
( )
( )
j
air
h x D
Nu x
k
= (36)
and heat transfer coefficient is given by
2
1
2 1
( , )
( )
T
T
h x t dt
h x
T T
=
(37)
(a) (b)

0 20 40 60 80 100 120 140 160 180 200
0
1
2
3
4
5
t(min),time
X
(
a
v
e
)
,
k
g

w
a
t
e
r
/
k
g

d
r
y

p
r
o
d
u
c
t

concave slab with 3.2mm thickness inlet Re=343.6,Tin=45
convex slab with 3.2mm thickness inlet Re=343.6,Tin=45
planar slab outside with 3.2mm thickness inlet Re=343.6,Tin=45
planar slab inside with 3.2mm thickness inlet Re=343.6,Tin=45
0 20 40 60 80 100 120 140 160 180 200
0
1
2
3
4
5
t(min),time
X
(
a
v
e
)
,
k
g

w
a
t
e
r
/
k
g

d
r
y

p
r
o
d
u
c
t

concave slab with 0.7mm thickness inlet Re=343.6,Tin=45
convex slab with 0.7mm thickness inlet Re=343.6,Tin=45
planar slab outside with 0.7mm thickness inlet Re=343.6,Tin=45
planar slab inside with 0.7mm thickness inlet Re=343.6,Tin=45

the integration is calculated using rectangular summation method with time step 10min
s since the fluctuation with h(x) is very small with time at the location x, and R is the tota
l curve length in contact with air. The Nusselt number at stagnant point is calculated by
(0)
(0)
j
air
h D
Nu
k
= (38)
where h(0)is calculated as the average value over time point with time step 10mins.
(1) (2)

(3) (4)

Figure 3.6. Average Nusselt Number and Nusselt Number at stagnant point under
different drying conditions: (1). Re=343.6, T(inlet) =45; (2). Re=85.9, T(inlet)=45;
(3). Re=859, T(inlet)= 45; (4). Re= 286.2869, T(inlet) = 80.
The calculation results are shown in Figure 3.6, from which it can be observed that
in most drying conditions, the convex slab has the highest Nusselt Number in the middle
point which is the stagnant point, this is because the middle point for this shape is clos-
est to the drying air and with fastest heat transfer happening at that place, it is also the
region with highest temperature below the surface. Compared with the planar slabs, the
curved slabs get higher average Nusselt number, this coincides with the drying kinemics
that the faster heat transfer brings the faster mass transfer when there exists a balance
between them. However, except the results in Figure 3.6 (1), for all the other conditions,
the concave shaped slab gets higher average Nusselt number, this is possibly because
more heat flux will happen in the peripheral of the contact surface, the hot air accumu-
lates at the region, so that heat transfer may be more efficient in this case, but this does
not necessarily mean that the mass transfer also achieves best in this case.
3.5. CONCLUSION REMARKS AND FUTURE WORK
Based on this study of conjugate modeling of heat and mass transfer between a la-
minar impinging jet and curved thin slabs of wet solid, we discover from the results that
under the same volume and thickness, the non-planar cases perform better than their

counterparts with certain thickness, whereas the performance of convex shape slab per-
forms almost the same with its planar counterpart and the concave shape slab has dete-
riorated drying performance with certain thickness. The thickness seems to be less sen-
sitive to the change of the geometry when it is very thin. However, we do not know the
divergence point due to limited sampling point of thickness. The nozzle inlet air condi-
tions influence on all the shapes with similar pattern. We draw the conclusion of this
study that both the shape and the thickness of the slabs can influence the drying kinetics
at the same time, and this influence is affected by changing the inlet air velocity which
directly affects the convective heat and mass transport. Since we have observed the dry-
ing performance of slabs with different shapes and thicknesses under certain thermo-
physical properties of air, we may need to expand this range so as to draw the above
conclusions more properly. For instance, if the inlet velocity is too low or the tempera-
ture is too high, we need to take the buoyancy effect into consideration, since the large
gradient of temperature will induce the change of the fluid density in space, and in that
case, we have to encounter the problem of buoyancy force and model this into the go-
verning equations. Meanwhile, if the inlet velocity is too high, we may need to consider
the turbulent flow in the fluid domain, from which we need to add turbulent terms to the
governing equations.
NOMENCLATURE STYLE
c
l
molar concentration of liquid water, mol m
-3

c
pa
specific heat of air, J kg
-1
K
-1

c
pb
specific heat of substrate, J kg
-1
K
-1

c
v
molar concentration of water vapor, mol m
-3

D
lb
diffusivity of liquid water inside the substrate, m
2
s
-1

D
va
diffusivity of water vapor in air, m
2
s
-1
D
vb
diffusivity of water vapor inside the substrate, m
2
s
-1

E
a
activation energy, kJ mol
-1

H
w
heat of wetting, J kg
-1

h heat transfer coefficient, W m
-2
K
-1

h
evap
heat of evaporation, J kg
-1

h
fg
latent heat of evaporation, J kg
-1

K

rate of water evaporation, s
-1

k
a
thermal conductivity of air, W m
-1
K
-1

k
b
thermal conductivity of substrate, W m
-1
K
-1

M
l
molar mass of water, kg mol
-1

Nu Nusselt number, -
P pressure, Pa
cooling rate due to evaporation, J m
-3
s
-1

R universal gas constant, 8.314 J mol
-1
K
-1
RH

relative humidity, -

T temperature, K
t time, s
u air velocity, m s
-1

V velocity, m s
-1

W wet basis moisture content, kg kg
-1
w.b.
X dry basis moisture content, kg kg
-1
d.b.
X
e
equilibrium moisture content, kg kg
-1
d.b.
X
free
free moisture content, kg kg
-1
d.b.


Greek letters

density of air, kg m
-3

density of substrate, kg m
-3

,
reference substrate density, kg m
-3

dynamic viscosity of the fluid, Pa s
Subscripts
a air
b substrate
l liquid
v vapor
in inlet
REFERENCE
Beitelmal, A. H., Saad, M. A., Patel, C. D., 2000, The effect of inclination on the heat trans-
fer between a flat surface and an impinging two dimensional air jet, International
Journal of Heat and Fluid Flow, 21, 156-163.
Bula, A. J., Rahman, M. M., Leland, J. E., 2000, Axial steady free surface jet impinging over
a flat disk with discrete heat sources, International Journal of Heat and Fluid Flow, 21,
11-21.
De Bonis, M. V., Ruocco, G., 2008, A generalized conjugate model for forced convection
drying based on an evaporation kinetics, Journal of Food Engineering, 89, 232-240.
Incropera, F. P., DeWitt, D. P., 2002, Fundamentals of Heat and Mass Transfer (5
th
Ed.),
Wiley, Hoboken, New Jersey.
Kays, W., Crawford, M., Weigand, B., 2005, Convective Heat and Mass Transport (4
th
Ed.),
McGraw Hill, Singapore.
Mujumdar, A.S., 2007, Handbook of Industrial Drying 3
rd
ed., Taylor & Francis, Boca Raton.
Poh, H.J., Kumar, K., Mujumdar, A. S., 2005, Heat transfer from a pulsed laminar imping-
ing jet, International Communications in Heat and Mass Transfer, 32, 1317-1324.
Srikiatden, J., Roberts, J. S., 2007, Predicting moisture profiles in potato and carrot dur-
ing convective hot air drying using isothermally measured effective diffusivity, Journal
of Food Engineering, 84, 516-525.
Xu, P., Mujumdar, A. S., Poh, H. J., Yu, B., 2010a, Heat transfer under a pulsed slot turbu-
lent impinging jet at large temperature differences, Thermal Science, 14, 271-281.
Xu, P., Yu, B., Qiu, S., Poh, H. J., Mujumdar, A. S., 2010b, Turbulent impinging jet heat
transfer enhancement due to intermittent pulsation, International Journal of Thermal
Sciences, 49, 1247-1252.

Chapter 4
Numerical Evaluation of Pulsation Effect on the Reaction Per-
formance of a T-junction micro-mixer

Balaji Mohan and Jiang Puqing
9 Engineering Drive 1, Singapore 117576
balajimohan@nus.edu.sg
a0068406@nus.edu.sg

Contents
4.1. INTRODUCTION .................................................................................................................................... 63
4.2. MATHEMATICAL MODEL ................................................................................................................... 64
4.2.1. Governing equations .............................................................................................................................................. 64
4.2.2. Boundary conditions .............................................................................................................................................. 64
4.2.3. Chemical reactions ................................................................................................................................................. 65
4.2.4. Data reduction .......................................................................................................................................................... 66
4.3. NUMERICAL APPROACH .................................................................................................................... 68
4.4. RESULTS AND ANALYSIS .................................................................................................................... 69
4.4.1. Validation ................................................................................................................................................................... 69
4.4.2. Effect of frequency .................................................................................................................................................. 70
4.4.3. Effect of pulsation amplitude ............................................................................................................................. 74
4.4.4. Effect of phase difference ..................................................................................................................................... 75
4.4.5. Effect of different forms of pulsations ............................................................................................................ 77
4.5. CONCLUSIONS ....................................................................................................................................... 78
Balaji Mohan and Jiang Puqing Numerical Evaluation of Pulsation Effect on the Reaction Performance of a
T-junction micro-mixer, in Mathematical Modelling of Transport Processes, Ed. Sasmito, A.P.; Kurnia, J.C.;
Jangam, S.V., 2011, ISBN - 978-981-08-9179-4, Published in Singapore, pp. 61-82.

Balaji, Jiang Numerical evaluation of pulsation effect on the reaction performance of a T-
junction micro-mixer
Mathematical Modelling of Transport Processes 63

4.1. INTRODUCTION
Micro-reactors and micro-mixers have been receiving great attention especially in
the chemical industries. The term micro-reactors here refers to systems with character-
istic length scales that are in the micrometer range. The potential advantages of using
micro reactors over conventional reactors include portability, safety and better control
of reaction conditions. According to Brody et al (1996), the tangible effect of this small
dimension is that fluid properties become increasingly controlled by viscous forces
rather than inertia. On the other hand, the reduced dimensions of the micro reactor sys-
tem lead to a large surface-to-volume ratio, which increases heat and mass transfer effi-
ciency.
Micro-mixers can be divided broadly into passive and active micro-mixers. Passive
micro-mixers only rely on diffusion or chaotic advection, while active micro-mixers rely
on external disturbances. T-junction micro-mixer is one of the most commonly used
passive micro-mixer. Extensive experimental and numerical investigations (Sasmito et
al., 2011; Wang et al., 2007; Ngugen et al., 2005; Kumar et al., 2010; Engler et al., 2004)
on mixing processes in T-junction micro-mixers have been conducted and reported.
However, the study of pulsation effect on T-junction micro-mixers including a catalytic
exothermic reaction has not been reported yet.
Karamercan et al (1979) studied the effect of pulsations on heat transfer and found
that the heat transfer coefficient increases with the pulsations, with the highest en-
hancement observed in the transition flow regimes. Honaker et al (1970) studied the
effect of sonic pulsations on mass transfer rate of naphthalene from a plate to air and
found that within the useful sonic pulsation ranges of amplitude and frequency the local
mass transfer rate is affected by the time averaged gradient of the air velocity at the in-
terface.
Zohir et al (2005) studied the effect of heat transfer by introducing pulsation and
found that heat transfer rate was enhanced for large diameter tubes but reduced for
smaller diameter tubes for same Reynolds number. Zaki et al (2010) found that super-
imposed pulsating flow enhanced mass transfer at the inner wall of an annulus for lami-
nar flow. Xu et al (2010) proved a significant enhancement of heat transfer at the target
surface by the intermittent pulsation in a turbulent impinging jet by simulation results.
Available literature shows that the heat and mass transfer rates can be enhanced
by pulsation flow. Here we propose a novel idea of introducing pulsating inflow in a T-
junction micro-reactor. The objective of the study presented here is to investigate the:
(i) effect of pulsation frequency, (ii) effect of pulsation amplitude, (iii) effect of phase
difference between inflow pulsations and (iv) effect of different waveforms.
First a mathematical model was developed with governing conservation equations
of mass, momentum, energy and species and was solved using CFD solver. Then the ef-
fect of inflow pulsation was evaluated in terms of outlet temperature, pressure drop and
conversion rate defined later. Finally conclusions were drawn based on the results ob-
tained.


The physical model used in the simulation is a two dimensional T junction micro-
channel as shown in Figure 4.1 Air enters the T junction micro-channel from one wing
while gaseous fuel from the other. The gaseous fuel is a mixture of methane and hydro-
gen. Both air and gaseous fuel are made to enter the micro-channel with pulsating ve-
locities. The interior walls of the micro-channel, just after the junction point are coated
with platinum which acts as the catalyst for the well-known combustion reaction be-
tween methane and air. The flow is considered to be unsteady and Newtonian flow.
Both air and gaseous fuel are miscible and follows ideal gas law.

Figure 4.1. Schematic representation of the micro-channel T-junction examined
4.2.1. Governing equations
The conservation equation of mass, momentum, species and energy are by
0
t
+ =
u (1)

( )
2
3
( ( ) ) ( ) p
t
u
u u u u u
( + = + + +

(2)

( ) ( )
p
p e temp
C T
C T k T S
t
u
+ = +
(3)
( ) ( )
( )
i
i i i i
D R
t
+ = +
u (4)
Where is the fluid density, u is the fluid velocity, pis the pressure, is the dynamic vis-
cosity, T is the temperature,
p
c is the specific heat capacity of the fluid,
e
k is the effective
thermal conductivity,
temp
S is the heat release or absorb due to the reaction,
i
is the
mass fraction species i,
i
D is the diffusion coefficient of species i,
i
R is the mass pro-
duced or consumed by the reactions.
4.2.2. Boundary conditions
The boundary conditions for the flow through micro channel T-junction are as follows:
Air inlet: The oxidant flow is introduced with pulsating flow with the following
conditions
Gaseous fuel
Air

( ) for sine wave : sin 2
2
air
A
u U ft = + + (5)

( ) ( )
( )
1
sin 2 1 2
4
for square wave :
2 2 1
air
k
k ft
A
u U
k

=
+
= +
(6)

2 2
300 K, 0.21, 0.79
O N
T = = = (7)
Fuel inlet: The gaseous fuel flow is introduced with pulsating flow with the fol-
lowing conditions
( ) sin 2
2
fuel
A
u U ft = + + (8)

4 2
300 K, 0.9, 0.1
CH H
T = = = (9)
Outlet: The outlet is pressure outlet with the following conditions
0 Pa, . 0, . 0
i
P n T n = = = (10)
Non-reacting walls: The non-reacting walls are specified as no slip condition, no
species flux and adiabatic at the channel wall before T-junction
0, 0 , 0
i
u T = = = (11)
Reacting walls: At the interior channel walls after T junction, surface reaction is
taken into account and it is considered no slip condition with constant wall tem-
perature
0, 1290 K
wall
u T T = = = (12)
4.2.3. Chemical reactions
Chemical reaction involved in the current simulation is mainly catalytic oxidation of
methane on the inner wall of channel which is coated with a thin layer of platinum cata-
lyst adopted from the work of Sasmito et al (2011). Note that many other types of reac-
tion are also possible. ChemKIN was used to set up a chemical reaction model to simu-
late the reaction processes. Species involved include seven gas species (CH4, O2, H2,
H2O, CO, CO2 and N2), eleven surface species (H(s), Pt(s), O(s), OH(s), H2O(s), CH3(s),
CH2(s), CH(s), C(s), CO(s), CO2(s)) and one bulk/solid species (Pt (b)).
The gas phase species and surface species can be produced and depleted by surface
reaction; hence, it is written in a general form as

' ' ' '' '' ''
, , , , , ,
1 1 1 1 1 1
g g
b s b s
r
N N
N N N N
i r i i r i i r i i r i i r i i r i
i i i i i i
g G b B s S g G b B s S
= = = = = =
+ + + +

(13)
where Gi, Bi, and Si represents the gas phase species, the solid species, and the surface-
adsorbed species, respectively. g, b, s are the stoichiometric coefficients for each reac-
tant species; g, b, and s are stoichiometric coefficients for each product species; and
Kr is the overall reaction rate constant. Since only the species involved as reactants or


products will have non-zero stoichiometric coefficient, the rate of reaction is calculated
as

| | | |
' '
, ,
,
1
g
i r i r
N
g s
r f r i i
wall wall
i
k G S
=
=

(14)
where [Gi]wall represents molar concentration on the wall. Thus, the net molar rate of
production or consumption of each species i is given by

( )
( )
( )
^
'' '
,
, ,
1
^
'' '
,
, ,
1
^
'' '
,
, ,
1
1, 2, 3,......,
1, 2, 3,......,
1, 2, 3,......,
rxn
rxn
rxn
N
i gas
i r i r r g
r
N
i bulk
i r i r r b
r
N
i site
i r i r r s
r
R g g i N
R b b i N
R s s i N
=
=
=
= =
= =
= =
(15)
while the reaction rate constant is computed using the Arrhenius expression given by

/
,
r r
E RT
f r r
k AT e

= (16)
On the reacting wall surface, it is assumed that the mass flux of each gas species is bal-
anced with its rate of production/consumption, is given by

. ^
,
,
, ,
1, 2, 3,.....,
i wall
i gas dep
wall i i wall i g
D m M R i N
n

= =
(17)

| |
^
, 1, 2, 3,.....,
i
wall
i site
s
S
R i N
t
= =
(18)
The gas concentration at the wall is calculated from the species mass fraction, defined as

| |
,
,
wall i wall
i
wall
i
G
M

= (19)
while
dep
m is the net rate of mass deposition or etching as a result of surface reaction,
given by

. ^
,
,
1
b
N
i bulk dep
i
i
m M R
=
=
(20)
wall
i
S

is the site species concentration at the wall, and defined as
| |
i site i
wall
S z = (21)
where
site
is the site density of the catalyst and
i
z is the site coverage of species i.
4.2.4. Data reduction
The gas density is given by the ideal gas law:

pM
RT
= (22)
where R is the universal gas constant and M denotes the mixture molecular weight given
by

4 2 2 2 2 2
4 2 2 2 2 2
1
CH H O H O CO N
CO
CH H O H O CO CO N
M
M M M M M M M

| |
= + + + + + + |
|
\ .
(23)
The gas mixture viscosity , (Bird et al., 2002) is defined as

4 2 2 2 2 2
,
, , , , , , ,
x
with CH H O H O CO CO N
x

= =
(24)
where
,
x

are the mole fractions (Bird et al., 2002) of species and

2
1 1
1
( ) 2 2
4
, ( )
1
1 1
8
g
g
M
M
M M
(
| | | |
| |
(
= + +
| |
|
(
| |
\ .
\ . \ .
(

(25)
The multicomponent gas mixture thermal conductivity
e
k is defined by

e i i
i
k k =
(26)
while the gas mixture specific heat capacity,
p
C , is evaluated using

, p i p i
i
C C =
(27)
The results are later discussed in terms of conversion rate. The conversion ratio is
used to evaluate the effectiveness of the mixing and reaction rate in micro channel. The
conversion rate of methane is defined as

in out
in

= (28)
where
in
and
out
are mass fractions of methane at the junction and outlet respec-
tively.
Mass fraction of methane can be exported directly from FLUENT. However, the
exported outlet mass fraction is sometimes larger than the inlet one, which does not
make sense. Instead, the mass flow rate of methane
4
CH
m is calculated by

4 4
CH mol CH
m C VAM = (29)
where
mol
C is the molar concentration of methane, which can be exported from FLUENT
directly, V is the bulk velocity, A is the channels cross area, and M
CH4
is the molar mass of


methane. Then the mass fraction is calculated from mass flow rate of methane and the
total mass flow rate
tot
m as

4
4
CH
CH
tot
m
m
=

(30)

4.3. NUMERICAL APPROACH
The model geometry was meshed with GAMBIT, pre-processor software of FLUENT
package, as illustrated in Figure 4.2. The grid was adopted from the work of Sasmito et al
(2011). Three different mesh sizes - 2.510
5
, 510
5
and 110
6
- were used in their study
to compare the local pressure, velocities, species mass fractions and temperatures to
ensure a mesh independent solution. It was found that mesh size of 510
5
gave about
1% deviation compared to a finer mesh size of 110
6
; whereas the results from the
mesh size of 2.510
5
gave about 10% deviation as compared to those from the finest
mesh design. Therefore, a mesh consisting of around 510
5
elements was found to be
sufficient for the numerical experiments for our study, although a fine structured mesh
was used near the wall to resolve the boundary layer and an increasingly coarser mesh
in the middle of the channel in order to reduce the computational cost.

Figure 4.2. Computational mesh
A segregated time-dependent unsteady solver in FLUENT 6.3.26 was used for the
calculation. Gas properties and reaction mechanisms were defined using ChemKIN soft-
ware, user defined functions (UDF) were written in C language to account for the tem-
perature dependant thermo-physical properties of the fluids used in the study. User-
defined functions (UDF) were written in C language to provide the pulsating inlet veloci-
ties. Data of velocities, mass fractions, pressure and temperature pulsating with time
were monitored as the calculation carries on, and were exported into data files for fur-
ther post-process.
Semi Implicit Pressure Linked Equation (SIMPLE) algorithm was used for pressure-
velocity coupling, and second order upwind differencing scheme was used for the dis-
cretization. Transient calculation was conducted. The time step size, maximum number
of iterations per time step, and the relaxation factors were carefully adjusted to ensure
convergence. In the calculation it was found that by decreasing the size of time step, the
residuals can be reduced, and thus precision can be enhanced. However, decreasing time

step size comes with the price of higher demands for computing resources. In this case,
by setting time step size as 10
-5
s and 20 iterations for each time step, the continuity re-
sidual can be kept lower than 10
-4
, which gives satisfactorily converged results.
4.4. RESULTS AND ANALYSIS
Numerical simulations were carried out to study the effect of inlet pulsation fre-
quency, amplitude and phase difference on the micro-reactor performance. For each
group of simulation, either two of the three parameters were set constant to study the
effect of other parameter. For all the cases the average inlet velocity was set as 5 m/s,
corresponding to an inlet Reynolds number of around 300. Simulation results show that
the highest velocity occurs at the outlet, and is around 24 m/s, corresponding to a Rey-
nolds number of 1400. This indicates that the flow in the channel is laminar.
A base case was studied for comparison reason. This base case is also time-
dependent, with all the same set up with other cases, only that the inlet velocities are set
constant as 5 m/s. Results of the base case are compared with all the other pulsating
cases in the following analysis. If not stated otherwise specifically, the horizontal lines in
the following graphs are values of the base case.
4.4.1. Validation
The validation of any mathematical model involving mixing and reactions is very
important for its complexity of coupled physical phenomena and interaction between
species and its chemical kinetics. As the mathematical model used in this simulation was
an extended work of Sasmito et al (2011). The validation conducted by Sasmito et al
(2011) by comparing their mathematical model with the experimental results of Bond et
al (1996) hold good for this research work. The methane conversion rate in monolithic
reactor was approximated as a repeating single channel flow; see Bond et al (1996) for
details of experimental setup. The results were also compared with simulation data from
Bond et al (1996) and Canu (2001).

Figure 4.3. Validation with experimental data at low and high inlet stoichiometry
The validation is initiated with methane catalytic oxidation at low stoichiometry gas
inlet ( = 0.18); after which the methane conversion rate of higher inlet stoichiometry (
0
10
20
30
40
50
60
70
80
90
0 1 2 3 4 5 6 7 8
C
H
4

c
o
n
v
e
r
s
i
o
n

r
a
t
e
,

%
Length, cm
exp Bond et al, stoich = 0.39
exp Bond et al, stoich = 0.18
present sim, stoich = 0.39
present sim, stoich = 0.18
sim Bond et al, stoich = 0.39
sim Bond et al, stoich = 0.18
sim Canu, stoich 0.39
sim Canu, stoich = 0.18


= 0.39) is compared, as depicted in Figure 4.3. It is found that the model predictions
agree well with the methane conversion rate from experimental counterpart at both low
and high inlet stoichiometry. In addition, their model has better agreement as compared
to model prediction by Bond et al (1996) and Canu (2001) especially at high methane
stoichiometry. This implies that the model correctly accounts for the fundamental phys-
ics associated with the reactions.
4.4.2. Effect of frequency
Cases with the same average velocity and peak amplitude but different pulsation
frequencies were examined to study the effect of inlet pulsation frequency on micro-
mixer performance. Figure 4.4 shows results of cases with average velocity of 5 m/s and
peak amplitude of 2 m/s but pulsation frequency varies as 1/5 Hz, Hz and 10 Hz sepa-
rately.
30 31 32 33 34 35
3
4
5
6
7
V
_
i
n

(
m
/
s
)
Time (s)
velocity_inlet
600
700
800
900
T_out
T
_
o
u
t

(
K
)

30 31 32 33 34 35
0.04
0.06
0.08
0.10
0.12
0.14
C
o
n
v
e
r
s
i
o
n

r
a
t
e

o
f

C
H
4
Flow Time (s)
Conversion rate of CH
4

30.0 30.5 31.0 31.5 32.0
2
3
4
5
6
7
8
V_in
V
_
i
n

(
m
/
s
)
Flow Time (s)
600
650
700
750
800
850
900
T_out
T
_
o
u
t

(
K
)

30.0 30.5 31.0 31.5 32.0
0.04
0.06
0.08
0.10
0.12
0.14
C
o
n
v
e
r
s
i
o
n

r
a
t
e

o
f

C
H
4
Flow Time (s)
4

37.70 37.75 37.80 37.85 37.90
2
3
4
5
6
7
8
V
_
i
n

(
m
/
s
)
Time (s)
velocity_inlet
600
650
700
750
800
850
900
T_out
T
_
o
u
t

(
K
)

37.70 37.75 37.80 37.85 37.90
0.04
0.06
0.08
0.10
0.12
0.14
C
o
n
v
e
r
s
i
o
n

r
a
t
e

o
f

C
H
4
Flow Time (s)
4

Figure 4.4. Effects of frequency on outlet bulk temperature and methane conversion
rate at frequency a) 1/5 Hz, b) Hz and c) 10 Hz.
a)
c)
b)

Comparison of these cases indicates that the outlet bulk temperature and methane
conversion rate pulsate at the same frequencies corresponding to the inlet velocity fre-
quency. Figure 4.4 also shows that when the inlet velocities of methane and air increase,
the conversion rate of methane also increases while the outlet temperature decreases;
and vice versa. Note that the molar proportion of methane and oxygen is constant as
long as their inlet velocities are at constant ratio, the increased conversion rate of meth-
ane corresponding to increased inlet velocity can be explained that by increasing inlet
velocity, the products are removed away more quickly and the mixture of fuel and oxy-
gen has more chance to get in touch with the wall and react. Also note that the chemical
reaction of methane oxidation is exothermic; the decreased outlet temperature corre-
sponding to increased inlet velocity indicates that the bulk fluid temperature is mainly
heated by the wall instead of by heat generation from chemical reaction.
Comparison of pulsating cases with the base case shows that pulsating cases have
slightly higher average outlet temperature and lower average methane conversion rate.
However, the amplitude or average values of outlet bulk temperature and conversion
rate of methane are little affected by different frequencies. Considering that the channel
length is only 20 mm and the average velocity of fluid is 5 m/s, it takes around 0.004 s
for the fluid to pass through the channel. For the cases in Figure 4.4 the inlet pulsation
frequency might be so low that during the interval of 0.004 s the inlet velocity changes
so little that it fails to have significant effect on the performance.
32.12 32.13 32.14 32.15
4.0
4.5
5.0
5.5
6.0
32.12 32.13 32.14 32.15
680
700
720
740
760
780
V
_
i
n

(
m
/
s
)
Time (s)
V_in
T_out
T
_
o
u
t

(
K
)
Time (s)

32.12 32.13 32.14 32.15
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
C
o
n
v
e
r
s
i
o
n

r
a
t
e

o
f

C
H
4
Flow Time (s)
4

32.12 32.13 32.14 32.15
100
150
200
250
300
350
400
P
r
e
s
s
u
r
e

D
r
o
p

(
P
a
)
Time (s)
Pressure Drop

Frequency 100Hz


32.160 32.165 32.170
4.0
4.5
5.0
5.5
6.0
32.160 32.165 32.170
680
700
720
740
760
780
V
_
i
n

(
m
/
s
)
Time (s)
V_in
T_out
T
_
o
u
t

(
K
)
Time (s)

32.160 32.165 32.170
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
C
o
n
v
e
r
s
i
o
n

r
a
t
e

o
f

C
H
4
Flow Time (s)
4

32.160 32.165 32.170
100
150
200
250
300
350
400
P
r
e
s
s
u
r
e

D
r
o
p

(
P
a
)
Time (s)
Pressure Drop

Frequency 200Hz
32.114 32.115 32.116 32.117
4.0
4.5
5.0
5.5
6.0
32.114 32.115 32.116 32.117
680
700
720
740
760
780
V_in
V
_
i
n

(
m
/
s
)
Time (s)
T_out
T
_
o
u
t

(
K
)
Time (s)

32.114 32.115 32.116 32.117
0.04
0.08
0.12
0.16
0.20
C
o
n
v
e
r
s
i
o
n

r
a
t
e

o
f

C
H
4
Flow Time (s)
4

32.114 32.115 32.116 32.117
0.04
0.08
0.12
0.16
0.20
C
o
n
v
e
r
s
i
o
n

r
a
t
e

o
f

C
H
4
Flow Time (s)
4

Frequency 1kHz
Figure 4.5. Effects of frequency on outlet bulk temperature, pressure drop and
methane conversion rate
A new group of cases were studied, with the inlet pulsation frequency further in-
creased to 100 Hz, 200 Hz and 1 kHz, while the peak amplitude kept constant as 0.5 m/s,

as shown in Figure 4.4. Note that it might not be practically possible to achieve a high
frequency as 1 kHz in real systems. This case was chosen here for academic purpose
only and is not justified for practical application.
Figure 4.5 shows that the averaged results of pulsating cases with increased fre-
quencies do not significantly vary from those of the base case; however, their pulsating
peak amplitudes significantly vary, especially for the case of 1 kHz. When the frequency
is further increased to 1 kHz, the outlet temperature is more stabilized while the pres-
sure drop and methane conversion rate pulsate at larger peak amplitudes. Comparison
also shows there is a phase shift of outlet temperature and pressure drop relative to the
inlet pulsation rate when the inlet frequency is increased.
It can thus be inferred that the outlet temperature can be stabilized either by de-
creasing the frequency to zero, or by increasing it to an extremely high value. This infer-
ence can be hardly explained by theory though. Instead the inlet and outlet bulk mass
flow rates of the three cases were monitored, and compared in Figure 4.6.
32.145 32.150 32.155 32.160
0.006
0.007
0.008
0.009
T
o
t
a
l

F
l
o
w

R
a
t
e

(
k
g
/
s
)
Flow Time(s)
outlet
inlet

32.155 32.160 32.165 32.170
0.006
0.007
0.008
0.009
T
o
t
a
l

F
l
o
w

R
a
t
e

(
k
g
/
s
)
Flow Time(s)
outlet
inlet

32.1150 32.1155 32.1160 32.1165
0.006
0.007
0.008
0.009
T
o
t
a
l

F
l
o
w

R
a
t
e

(
k
g
/
s
)
Flow Time(s)
MRF_out
MFR_IN

Figure 4.6. Inlet and outlet mass flow rate with different frequencies
(frequency = 100 Hz, 200 Hz and 1 kHz)
Figure 4.6 shows that for the first two cases (100 Hz and 200 Hz), the inlet and out-
let total mass flow rate pulsate at the same amplitude and the same frequency, with a
slight phase lag that the outlet mass flow rate is behind the inlet. This result is reason-
able that by the law of mass conservation the outlet mass flow rate should be the same
with the inlet; and since it takes time for the fluid to flow through the channel the outlet
flow rate should lag behind the inlet. For the case of 1 kHz, however, the outlet mass
flow rate pulsates at smaller amplitude than the inlet, totally different from the previous
cases. It should be noted that all the cases were calculated under the same procedure
and residuals were checked to insure convergence. The inconsistency can be explained


either by some calculation problems that have not yet been discovered, or that it is a
true phenomenon but not yet well understood.
Figure 4.7 shows the effect of frequency on percentage deviation of conversion rate,
outlet temperature and pressure drop. The deviation was found to be within 3% which
is permissible error for any numerical simulations thus the effect of inflow pulsation
frequency does not aid the heat and mass transfer phenomenon in T-junction micro
mixer significantly.
0.1 1 10 100 1000
-5.0
-2.5
0.0
2.5
5.0
Conversion Rate CH4
P
e
r
c
e
n
t
a
g
e
(
%
)
Frequency (Hz)

0.1 1 10 100 1000
-5.0
-2.5
0.0
2.5
5.0
Outlet Bulk Temperature
P
e
r
c
e
n
t
a
g
e
(
%
)
Frequency (Hz)

0.1 1 10 100 1000
-5.0
-2.5
0.0
2.5
5.0
Pressure Drop
P
e
r
c
e
n
t
a
g
e

(
%
)
Frequency (Hz)

Figure 4.7. Effect of frequency on percentage deviation from non-pulsating conditions

4.4.3. Effect of pulsation amplitude
Cases with same pulsation frequency (100 Hz) but different peak amplitudes (2 m/s and 0.5
m/s) were studied to investigate the effect of pulsation amplitude on micro-reactor performance.
Results were compared in Figure 4.8.
32.125 32.130 32.135
2
3
4
5
6
7
8
32.125 32.130 32.135
600
650
700
750
800
850
900
V
_
i
n

(
m
/
s
)
Time (s)
V_in T_out
T
_
o
u
t

(
K
)
Time (s)

32.12 32.13 32.14 32.15
2
3
4
5
6
7
8
32.12 32.13 32.14 32.15
600
650
700
750
800
850
900
V
_
i
n

(
m
/
s
)
Time (s)
V_in T_out
T
_
o
u
t

(
K
)
Time (s)


32.125 32.130 32.135
100
150
200
250
300
350
400
P
r
e
s
s
u
r
e

D
r
o
p

(
P
a
)
Time (s)
Pressure Drop

32.12 32.13 32.14 32.15
100
150
200
250
300
350
400
P
r
e
s
s
u
r
e

D
r
o
p

(
P
a
)
Time (s)
Pressure Drop

32.125 32.130 32.135
0.00
0.05
0.10
0.15
0.20
C
o
n
v
e
r
s
i
o
n

r
a
t
e

o
f

C
H
4
Flow Time (s)
4

32.12 32.13 32.14 32.15
0.00
0.05
0.10
0.15
0.20
C
o
n
v
e
r
s
i
o
n

r
a
t
e

o
f

C
H
4
Flow Time (s)
4

Figure 4.8. Effect of pulsation amplitude on outlet temperature, pressure drop and
methane conversion rate (peak amplitude = 2 m/s, 0.5 m/s)
Comparison shows that the averaged values of outlet temperature, pressure drop and meth-
ane conversion rate are the same, but they pulsate at higher amplitudes when the inlet velocity
pulsates at higher amplitude. Thus it is concluded that the pulsation amplitude of inlet velocity
does not have particular effect in enhancing methane conversion rate or decreasing average
pressure drop; the smaller the pulsation amplitudes of inlet velocity, the more stable the per-
formance.
4.4.4. Effect of phase difference
The effect of phase difference between air and fuel inlet in terms of
3
, , ,
4 2 4

were stud-
ied in comparison with the base case. The peak amplitude (0.5 m/s) and frequency (1 kHz) of the
flow are kept constant in this study. Figure 4.9.a shows the inlet velocity of air and gaseous fuel
for different phase differences with flow time.
Figure 4.9.b shows the conversion rate over flow time pulsates with different amplitudes
from the base conversion rate though the time averaged conversion rate remains same for all the
cases. It was also found that the amplitude of conversion rate decreases as phase difference in-
creases and then for a phase difference of it again increases. As the phase difference is intro-
duced, the air enters the channel prior followed by methane which gives air time to reach the
convenient temperature for the reaction to occur by heat transfer through convection from the
walls which reduces the chemical and ignition lag between them. The temperature of air in-
creases as the phase difference increases as there is more time for temperature rise to occur. For
phase difference of , though air gets maximum time for temperature rise, the high temperature
air leaves the channel before the fuel enters the channel as the half time period selected for the
study is lesser than the time taken for the flow to leave the channel thus the amplitude of conver-
sion rate increases.


From Figure 4.9.c, the pressure drop also pulsates with different amplitudes from the base
case and the amplitude of pulsating pressure drop decreases with increasing phase difference,
though there is no significant effect on time-average pressure drop. This phenomenon can be
explained by the pumping power requirement. The total pumping power
Total
P is the sum of
pumping power required for air
Air
P and gaseous fuel
Fuel
P .

Tot F A
P P P = + (31)
The pumping power is given as

. .

pump
pump
p aV
P
= (32)
As average pressure drop and average velocity are same, the average total pumping power
required is also same for all the cases. When the velocity of both air and fuel are in phase, the
pressure drop pulsates with high amplitude as pumping power gets added in same magnitude for
air and gaseous fuel. The amplitude of pressure drop decreases as the phase difference increases
as pumping power required for pumping air at any instance of time is either higher or lesser than
the gaseous fuel and thus the net amplitude of pressure drop decreases. Similarly the pressure
drop is almost constant for phase difference of because at any point of time if pumping power
required for gaseous fuel is high, for air it is lesser by same magnitude, and vice versa.
Figure 4.10 shows percentage deviation of time averaged conversion rate, outlet tempera-
ture and pressure drop with non-pulsating inflow conditions. It was found that conversion rate
increases with increase in phase difference between air and fuel and reaches maximum of 6% at
phase difference of
3
4
and
and percentage of pressure drop increases as phase difference

increases but has no effect on outlet temperature, which shows that the increase in conversion
rate has not increased the outlet temperature therefore no heat release due to higher conversion
0.037 0.038 0.039 0.040
4.0
4.5
5.0
5.5
6.0
6.5
7.0
0
1/4
1/2
3/4

air_in
V
_
i
n
l
e
t

(
m
/
s
)
Flow Time (s)

0.037 0.038 0.039 0.040
0.04
0.08
0.12
0.16
0.20
0.24
0
1/4
1/2
3/4

C
o
n
v
e
r
s
i
o
n

r
a
t
e

C
H
4
Flow Time (s)

0.037 0.038 0.039 0.040
100
150
200
250
300
350
400
450
0
1/4
1/2
3/4

P
r
e
s
s
u
r
e

D
r
o
p

(
P
a
)
Flow Time (s)

Figure 4.9.a) Inlet velocity b) Conversion rate and c) Pressure drop for different
phase difference
a)
b)
c)

rate which means that the percentage increase may only due to cumulative error of numerical
solution and not actual benefit.
0 pi/4 pi/2 3pi/4 pi
0
2
4
6
8
10
Conversion Rate CH4

P
e
r
c
e
n
t
a
g
e

(
%
)
Phase Difference

0.0
0.1
0.2
0.3
0.4
0.5
Outlet Temperature

P
e
r
c
e
n
t
a
g
e

(
%
)
Phase Difference

0
1
2
3
4
5
Pressure Drop

P
e
r
c
e
n
t
a
g
e

(
%
)
Phase Difference

Figure 4.10. Effect of phase difference on percentage deviation from non-pulsating
inflow conditions
4.4.5. Effect of different forms of pulsations
The Figure 4.11 shows inlet velocity profile of air and fuel, outlet temperature, pressure
drop and conversion rate between square and sine waveforms for air inflow pulsations
keeping fuel inflow pulsations to be of sine waveform in both cases. For both cases the
frequency (1 kHz), peak amplitude (0.5 m/s) and phase difference ( = 0) of different
inflow pulsations are maintained same for comparison purpose. For numeric purpose
the smooth square waveform was generated by expanding the sine function using the
Fourier series up to 1000 odd integer harmonics.
As in case of sine waveforms for both inflow pulsations there is a gradual increase
or decrease in velocity which also results in gradual increase or decrease in the tem-
perature, conversion rate and pressure drop. For the case where fuel inflow pulsates
with sine waveform and air inflow pulsates with square waveform the temperature,
pressure drop and conversion rate has sudden increase or decrease at every half period
and this effect is due to sudden transition in square waveform from high to low velocity
or vice versa at every half period instant of time.


0.037 0.038 0.039 0.040
4.0
4.5
5.0
5.5
6.0
0.037 0.038 0.039 0.040
680
700
720
740
760
780
V_methane
V_air
V
_
i
n

(
m
/
s
)
Time (s)
T_out
T
_
o
u
t

(
K
)
Time (s)

32.114 32.115 32.116 32.117
4.0
4.5
5.0
5.5
6.0
32.114 32.115 32.116 32.117
680
700
720
740
760
780
V_in
V
_
i
n

(
m
/
s
)
Time (s)
T_out
T
_
o
u
t

(
K
)
Time (s)

0.037 0.038 0.039 0.040
100
150
200
250
300
350
400
P
r
e
s
s
u
r
e

D
r
o
p

(
P
a
)
Time (s)
Pressure Drop

32.114 32.115 32.116 32.117
100
150
200
250
300
350
400
P
r
e
s
s
u
r
e

D
r
o
p

(
P
a
)
Time (s)
Pressure Drop

0.037 0.038 0.039 0.040
0.00
0.05
0.10
0.15
0.20
C
o
n
v
e
r
s
i
o
n

r
a
t
e

o
f

C
H
4
Flow Time (s)
4

32.114 32.115 32.116 32.117
0.00
0.05
0.10
0.15
0.20
C
o
n
v
e
r
s
i
o
n

r
a
t
e

o
f

C
H
4
Flow Time (s)
4

Figure 4.11. Comparison of Inlet velocity, outlet temperature, Pressure drop and Con-
version rate of different waveforms

4.5. CONCLUSIONS
A numerical study was performed to investigate the effect of inflow pulsation parame-
ters i.e. frequency, amplitude, phase difference, and different pulsation waveforms on
the heat and mass transport phenomena in a T junction micro reactor. It was found that:
By introducing inlet pulsation the average outlet temperature becomes slightly
higher while the methane conversion rate slightly lower;
The average value of pressure drop or methane conversion rate is not signifi-
cantly affected by the frequency or peak amplitudes of inlet pulsation;

The pulsation amplitudes of outlet bulk temperature and methane conversion
rate are not significantly affected when it is lower than 1 kHz; however, when
the frequency is larger than 1 kHz the heat transfer rate tends to have less fluc-
tuations and thus more stable;
For much smaller amplitudes, a more stable performance can be achieved;
When there is a phase difference between the inflow pulsations, it was found
that phase difference of
3
4
with fuel lagging behind air gave a more stable per-

formance; but considering the pressure drop, phase difference of is most sta-
ble.
On studying the effect of different waveforms it is found that sine wave pulsa-
tions are more stable in terms of performance than square wave pulsations.
Thus it is conclude from the result of this numerical study that the inflow pulsations
have no positive effect on augmenting transport phenomena.
NOMENCLATURES
r
A
Pre-exponential factor
a
Cross section area,
2
m
i
B
Bulk/solid species, mol
' ' '
,
i i
b b
Stoichiometric coefficient for bulk reactant, and product
p
C
Specific heat,
1 1
J kg K

mol
C
Molar concentration,
3
mol m

i
D
Diffusivity of species i,
2
m s

r
E
Activation energy for the reaction,
1 1
J kg mol

i
G
Gas species, mol
' ' '
,
i i
g g
Stoichiometric coefficient for gas reactant, and product
eff
k
Effective thermal conductivity,
1 1
W m K

, f r
k Reaction rate constant using Arrhenius expression
M Mean molecular mass
.
dep m
Net rate of mass deposition, kg
p
Pressure, Pa
P Pumping power, W
Q
Volume flow rate,
3 1
m s

R
Universal gas constant,
1 1 1
J kg mol K

i
R
Reaction rate of species i,
3
kg m

i
S
Surface-adsorbed/ site species, mol
' ' '
,
i i
s s
Stoichiometric coefficient for gas reactant, and product
temp
S
Heat release/ absorb due to reactions,
3
W m



T Temperature, K
u
Velocity,
1
m s

V
Bulk velocity,
1
m s

x Mole fraction
Greek letters
r

Temperature exponent

Density
3
kg m

Dynamic viscosity,
1
Pa s

Rate of r
th
reaction
i

Mass fraction of species i

Conversion efficiency
pump
Pump efficiency

Subscripts and superscripts
A Air
b Bulk
dep Deposition
e Effective
F Fuel
g
Gas
i Species i
r r
th
wall surface reaction
s Solid/site
temp
Temperature
tot Total

REFERENCES
Bird, R.B., Stewart, W.E., Lightfoot, E.N., 2002, Transport Phenomena, 2
nd
ed., Wiley, New
York.
Bond, T.C., Noguchi, R.A., Chou, C.P., Mongia, R.K., Chen, J.Y., Diblle, R.W., 1996, Catalytic
oxidation of natural gas over supported platinum: flow reactor experiments and detailed
numerical modelling, Proceeding of the 26th Symposium (International) on Combustion, 26
(1), 1771-1778.
Brody, J.P., Yager, P., Goldstein, R.E., Austin, R.H., 1996, Biotechnology at low Reynolds
numbers, Biophysical Journal, 71 (6), 3430-3441.
Canu, P., 2001, Simulation and interpretation of catalytic combustion experimental data, Ca-
talysis Today, 64(3-4), 239-252.

Engler, M., Kockmann, N., Kiefer, T., Woias, P., 2004, Numerical and experimental investiga-
tions on liquid mixing in static micro-mixer, Chemical Engineering Journal, 101 (1-3), 315-
322.
Honaker, D.E., Tao, L.C., 1970, Effect of sonic pulsations on mass transfer rate of naphtha-
lene from a plate into air, Industrial and Engineering Chemistry Fundamentals, 9(3), 325-
329.
Karamercan, O. E., Gainer, J.L., 1979, The effect of pulsation on heat transfer, Industrial and
Engineering Chemistry Fundamentals, 18( 1), 11-15.
Kumar, V., Paraschivoiu, M., Nigam, K.D.P., 2010, Single phase fluid flow and mixing in mi-
crochannels, Chemical Engineering Science, 66 ( 7), 1329-1373.
Nguygen, N.T., Wu, Z., 2005, Micromixers A review, Journal of Micromechanics and Micro-
engineering, 15 (2), R1 Doi: 10.1088/0960-1317/15/2/R01.
Sasmito, A.P., Kurnia, J.C., Mujumdar, A.S., 2011, Numerical Evaluation of Transport Phe-
nomena in a T-Junction Micro-reactor with coils of square cross section tubes, Industrial
Engineering Chemistry Research, DOI: 10.1021/ie200139s.
Wang, S.J., Mujumdar, A.S., 2007, Three Dimensional analysis of flow and mixing character-
istics of a novel in-line opposing jet mixer Industrial Engineering Chemical Research, 46 (2),
632642.
Xu, P., Yu, B., Qiu, S., Poh, H. J., Mujumdar, A. S., 2010b, Turbulent impinging jet heat
transfer enhancement due to intermittent pulsation, International Journal of Thermal
Sciences, 49, 1247-1252.
Zohir, A.E., Habib, M.A., Attya, A.M., Eid, A.I., 2005, An Experimental investigation of heat
transfer to pulsating pipe air flow with different amplitudes, Heat and mass transfer, 42,
625-635.
Zaki, M.M., Nirdosh, I., Sedahmed, G.H., 2010, Mass transfer behaviour of annular ducts un-
der conditions with superimposed pulsating flow, Chemical Engineering Technology, 34(3),
475-481.

Jangam, S.V. and Mujumdar, A.S. Basic concepts and definitions, in Drying of
Foods, Vegetables and Fruits - Volume 1, Ed. Jangam, S.V., Law, C.L. and Mujumdar,
A.S. , 2010, ISBN - 978-981-08-6759-1, Published in Singapore, pp. 1-30.

Chapter 5
Numerical Analysis of Single-Particle Combustion of Coal Char

H. Osman and A. Ismail

Contents
5.1. INTRODUCTION ............................................................................................................... 85
5.2. PHYSICAL DESCRIPTION ............................................................................................... 85
5.3. LITERATURE REVIEW .................................................................................................... 87
5.3.1. Simplifying assumptions ..................................................................................................... 87
5.3.2. Species considered in the reactions ............................................................................... 88
5.3.3. Shrinking core mode ............................................................................................................ 88
5.3.4. Porous nature of core ........................................................................................................... 88
5.4. MODEL FORMULATION ................................................................................................. 89
5.4.1. Transport equations ............................................................................................................. 89
5.4.2. Chemical reaction .................................................................................................................. 90
5.4.3. Boundary and initial conditions ...................................................................................... 91
5.5. RESULTS AND DISCUSSION .......................................................................................... 92
5.5.1. CHAR CONVERSION RATE ................................................................................................. 93
5.5.2. TEMPERATURE PROFILES ................................................................................................. 96
5.5.3. Model validation .................................................................................................................. 100
5.5.4. Tabulated 1d model results ............................................................................................ 102
5.5.5. 2D Model................................................................................................................................. 104
5.6. CONCLUSION .................................................................................................................. 105
ACKNOWLEDGEMENT ......................................................................................................... 105
NOMENCLATURE ................................................................................................................... 106
REFERENCES ........................................................................................................................... 106

Osman, H. and Ismail, A. Numerical Analysis of Single-Particle Combustion of Coal Char, in Ma-
thematical Modeling of Transport Processes, Ed. Sasmito, A.P.; Kurnia, J.C.; Jangam, S.V., 2011,
ISBN - 978-981-08-9179-4, Published in Singapore, pp. 83-108.

Osman, Ismail Numerical Analysis of Single-Particle Combustion of Coal Char

5.1. INTRODUCTION
Combustion of coal is an important process that has been subjected to numerous
studies. The significance of coal in the industrial world is largely attributed to the steel-
making and energy industry. In terms of energy production, the share of coal in world
energy mix has been on the decline due to several factors (Hk et al., 2008). However,
there are renewed interests in the subject of coal utilization partially due to the availa-
bility and relative abundance of low-rank coal compared to other energy sources. Two
most common types of reactors utilizing coal for energy production are the pulverized
fuel combustor and the fluidized bed. The major advantages of combusting coal in a flui-
dized bed include the control over environmental pollutants, the increase in heat trans-
fer capability and the ability to design a more compact combustor. Bituminous coals are
commonly used in these fuel combustors, though coals of all ranks are used (Meyers,
1981). Recently, there is also an increasing interest in the combustion of biomass driven
by its CO2 neutrality of both cultivated and waste biomass. Furthermore, the efficiency of
burning wood is higher compared to coal as the amount of ash is less in wood.
In all reactors, the degree of char burnout as an operating characteristic in coal com-
bustion systems is central, since it affects boiler efficiency, electrostatic precipitator op-
eration and the value of the fly ash which is a sellable byproduct to the concrete industry
(Comrie, 2007). Single particle experiments are thus useful in providing information on
dilute coal combustion conditions, while experimental studies involving particle interac-
tions are essential for practical knowledge of coal combustion.
The aim of this paper is to obtain a fundamental understanding of the effects of vari-
ous process conditions on the reactivity and ignition behavior of coal and char particles.
Information about this behavior is important for the efficient design and operation of an
industrial coal fired power plants and utility boilers. The current study focuses on quan-
tifying the effects of the most important of these parameters, namely the bulk tempera-
ture of the inflow of air, particle size and the concentration of oxygen during the char
burn out stage. Further, the effect of the shape of the char particles on the combustion
process will be studied. The co-combustion of coal with biomass particles is also ex-
plored given its industrial relevance.
5.2. PHYSICAL DESCRIPTION
Coal combustion could be described in four well-defined steps: drying, devolatilization,
volatile combustion, and char combustion. When the coal particle enters the combustion de-
vice, the rapid release of moisture (drying) and volatile substances (devolatilization) occurs
almost immediately into the surroundings as a result of high temperature. The ignition of
these volatiles (volatile combustion) follows leaving behind a porous solid product which
subsequently combusts (char combustion). The focus of this paper is the combustion of char,
which is the limiting step of the overall rate of the combustion process, a major contribution
to the volume of the combustion chamber.
During char combustion, oxygen first is transported from the bulk flow to the exterior
surface of the char particle either by diffusion or convection. Oxygen then diffuses into the
char pores and reacts at the interior surfaces. Combustion of carbon takes effect only in re-
gions where oxygen is available, provided the temperature in that region is sufficiently high to
result in ignition. The energy required to initiate the reaction comes from bulk gas of tempera-

tures between 500-1100 K, depending on the application. Heat transfer from bulk to particle
surface is dominated by convection, whereas within the char particle conduction is dominant.
Char combustion occurs both at the external and internal surfaces, resulting in an increase of
the particle temperature and heat transfer from the particle to the surroundings (Oka et al.,
1995). Thus in the initial period, the particle temperature is low and its surface is dark. As
time progresses, bright high temperature points appear at the particle surface as the number of
hot spot increases. The heterogeneous reactions of the solid coal particle and the oxygen gas
will finally extend throughout the coal particle. During this time, the reacted gas products
continuously diffuse out from the particle exterior surface to the bulk flow. Simultaneously,
the inner particle materials are consumed and fragmentation occurs. This will result in a
shrinking char particle that will diminish with time.
The following are the homogenous and heterogeneous reactions that occur simulta-
neously within and around the char particle.

Figure 5.1. Schematic of char combustion.

Heterogeneous reactions:

C +O
2

CO
2
(I)
C +O

2CO (II)
C +
1
2
O
2

2CO (III)

Homogenous reaction (gas phase):

CO +
1
2
O
2

CO
2
(IV)
volatiles + O
2

CO +CO
2
+H
2
O (V)

The coal particle may undergo a coal combustion process that is limited by the pure hete-
rogeneous reaction instead of the pure homogenous reaction described earlier. This depends
on the rate of devolatilization and the heating rate of the coal particle (Essenhigh et al.,
1989), which in turn depends on the particle size. When the heating rate of the particle is
faster than the devolatilization process, the former occurs since the primary ignition takes
place as a result of the direct attack of the oxidizer, which is oxygen on the solid coal particle
itself. When the devolatilization process occurs faster than the heating rate of the particle, the
homogenous ignition takes place due to the volatiles-oxidizer mixture in the thin layer located

in the vicinity of the particle surface as depicted in Figure 5.1. The gas flame enveloping the
particle screens its surface and prevents oxidizer from penetrating through the particle surface.
The particle temperature thus remains relatively low in this case until the combustion of vola-
tiles has ceased. At this point the heterogeneous reaction of the char combustion takes place.
In this paper, the focus is in the homogenous reaction of the char particle due to its relevance
in both cases.
Char combustion is controlled by a complex coupling between transport phenomena
around and within the particle as well as the intrinsic kinetics. It can be described by the
following processes: boundary layer diffusion, heat transfer, inter-particle diffusion, in-
trinsic reaction and homogenous reaction (Smith, 1994) that will be explained in the
next section on the mathematical modeling. Whether the combustion of coal is governed
by the rate of oxygen diffusion or chemical kinetics is determined by reaction conditions,
such as the bulk gas temperature, gas velocity, particle size and coal type (Sahu et al.,
1988; Miller et al., 1982; Essenhigh, 1981).
5.3. LITERATURE REVIEW
There are various models of char particle combustion available in the literature
which encompasses chemical reactions, mass transfer and transport processes. Models
of global combustion (Baum and Street, 1981) consist of equations that are analytically
solved by assuming steady state conditions. These models take into account the external
mass transfer and chemical kinetics related to the external surface of the char particles
based on apparent activation energy (Chakraborty and Howard, 1981; Avedesian and
Davidson, 1973). Thus, the reaction rates are controlled by either kinetics or by oxygen
diffusion rate.
Models of intrinsic reactivity (Smith, 1982), on the other hand are more suitable for
parametric and mechanistic investigations relating to the combustion of the char par-
ticle such as temperature, chemical reactions, heat and mass transfer as well as the ef-
fect of porosity (Bliek et al., 1986; Bhatia and Permutter, 1980; Simons, 1982). In these
models, the reaction rates are thus expressed in terms of intrinsic chemical and pore
diffusion rate, depending on the effectiveness factor (Smith, 1978).
The complexity of a model depends mainly on the type of the heterogeneous reac-
tion and its simplifying assumptions. This is because the heterogeneous reaction rate,
which needs to be solved in both solid and gas phases, depends on both collision fre-
quency (which is a function of temperature) and the number density of the oxidizer
(which is a function of concentration). Furthermore, the relative rates of formation of CO
versus CO2 not only result in a different heat of reaction but the stoichiometric coeffi-
cient of oxygen in each reaction is different. The homogenous reaction on the other hand
take place only in a thin layer around the particle and have been considered extensively
by Sotirchos and Amundson (1984).
5.3.1. Simplifying assumptions
Some of the simplifying assumptions used by the authors are as follows:
a. Particles are spherical.
b. Shrinking core model is chosen. This means that given that the process is con-
trolled by diffusion, the shape and the diameter of the particle do not change.

Physically, this model assumes that fragmentation and attrition do not occur
during combustion. The density of the particle however changes due to char
conversion.
c. Porosity decreases proportionally with the carbon conversion degree.
d. Carbon monoxide to be the singular product of the reaction, being converted to
carbon dioxide beyond the boundary layer surrounding the reacting particle.
e. Surrounding air is stagnant.
f. Char particle is completely devolatilized.
g. Pseudo steady-state for gas components
5.3.2. Species considered in the reactions
The number of reactions occurring simultaneously during coal combustion is expan-
sive. Thus there is a need to simplify the coal combustion problem in terms of reactions
considered by limiting the species considered. The authors generally considered O2, CO2,
CO, and C as the reactions of carbon containing species have the greatest contributions
to the heat balance inside the particle (Manovic et al., 2008). Sulfur, Nitrogen and miner-
al impurities are thus neglected.
5.3.3. Shrinking core mode
Although the use of the shrinking core model is in agreement with the combustion of
large particles in FB of Winter et al., 1997 this adopted model neglects particle attrition
and fragmentation, and thus an approximation. The influence of attrition and fragmenta-
tion on char reactivity and diffusivity, as well as the consideration of impact of ash on
kinetics are therefore needed unless the process is limited by diffusion such as that de-
picted in the parametric analysis of Manovic et al. In this case, the diffusion resistance of
ash is practically negligible and hence when the temperature of the combusting particle
is considered, this approach is very close to the realistic picture where the char particle
loses ash layer because of attrition.
5.3.4. Porous nature of core
Since gas diffusions significantly affect char combustions, pore structures will great-
ly affect the apparent combustion rates. One of the prevalent methods used in the litera-
ture is the use of the effectiveness factor, , which was developed by Mingle and Smith
(1961). This method takes into account the strongly bimodal (micropores and macro-
pores) distribution that exists within the char particles. Alternatively, He et al (2003),
only included the effects of macro-pores in the specific reaction surfaces. This is in ac-
cordance of the conclusions found in the works of He et al. (2003) which found that the
larger the ratio of the macro pore surface area to the total pore () the higher the char
combustion rate. This is because the larger macropores act as major arteries, supplying
oxygen to the interior of the particle. On the other hand, although micro pores contri-
bute to most of the pore surface areas in the char, it has a relatively less contribution to
the char combustion rate (Wang et al., 2007). This is because all the oxygen is consumed
before accessing all of the smaller pores. Furthermore, the higher the fractal dimension

(), the more complex the geometric shape, and thus the larger the resistance to gas dif-
fusion. In the work of He et al, both the effects of and are included in the modeling.
5.4. MODEL FORMULATION
The model is subjected to the simplifications mentioned in Section 5.2 due to limited
resources and time, and to reduce computational burden of existing machines. Further-
more, due to the difficulty of getting a fully converged solution for a 2D spherical particle
at high oxygen concentration, we perform our simulation in two parts. The first set of
simulations involves the combustion of a char particle at 1000 K, in oxygen concentra-
tions of 10%, 20%, and 30%. The velocity at the inlet of the fluid domain is fixed at 0.2
m/s. The geometry of the model is illustrated in Figure 5.2.
Due to the limited amount of data that can be gathered from the 2D simulation, a
second set of simulations is needed. To ensure that fully converged solutions are poss-
ible within a limited time, the 2D problem described above is reduced to a 1D problem.
It is noted that while first set of simulation is axis-symmetrical, the reduced from is not.
Instead, the problem is set such that the reduced form is spherically symmetric, allowing
for a 1D solution. Effectively, we are solving for a line from particle center to particle
surface, without modeling the fluid domain. To account for external sources of species
and heat, correlations for heat and mass transfer coefficients provided by Chen (2003)
and Ho (2003) are applied at the boundary. The geometry of this model is shown in Fig-
ure 5.3.

Figure 5.2. Geometry and boundary settings for 2D simulation of char combustion.

Figure 5.3. Geometry and boundary settings for 1D reduced form of char combus-
tion

5.4.1. Transport equations
Model equations are derived under the assumption of non-isothermal conditions,
and constant particle size during combustion to reduce computational burden. Constant
particle size-a consequence of the shrinking core model-offers computational advantage

over the shrinking particle model for two fundamental reasons: Firstly, the shrinking
particle model require moving mesh capability in order to exhibit the shrinking of par-
ticle domain. Dynamic meshing significantly increase computational load due to the
need to re-mesh the computational domain at every time step. Second, the shrinking
core model allows the Navier-Stokes equation governing the flow of the bulk region to
be solved as a steady flow problem, provided flow velocity is large enough to neglect any
buoyancy effect. Conservation of momentum for the bulk fluid is then

( ) = () +( +()
) (1)

where is the velocity vector (m/s), is pressure force (N/m
2
), and is dynamic vis-
cosity (Pas) of the fluid. The flow field solution from equation (1) is coupled to the mass
transport equation (2) through the convective term, but is independent of the energy
and mass equations. Therefore, in many situations, equation (1) can be removed from
the system and solved separately. However, due to the complex nature of the problem
and long convergence time, we removed the fluid domain, and consequently the convec-
tive term from the mass transport equation for the second solution set.
Transport of species within the char particle is dominated by diffusion through
complex pore structures. Only species that are taking part in the reaction are considered
in the species balance equation. Other species such as N2 are assumed to be inert and
non-diffusing. CO2 production is directly related to O2 or C consumption from the stoi-
chiometry of the reaction in equation (4)thus removed from our numerical procedure.
The species equation is expressed as

= (

(2)

where
is species concentration,
is effective diffusivity of species in the char particle,
is the source or sink term of species , and is the bulk velocity vector from the solu-
tion of equation (1). The convective term in the above equation is applied only to the
fluid domain in the 2D simulation. For the 1D cases, the convective term is zero since we
are not solving for the fluid region outside the particle domain. Char is assumed to be
entirely composed of C (carbon), which is non-diffusing, i.e.
1
= 0. For gaseous species,
must account for carbon conversion such that as

1
0,

,
. Correlation be-
tween
and the structure of the coal char can be found in Yang (2003).
Heat transfer in particle is governed by the energy equation

= ()
+ (3)

in which we neglect contributions from the convection term for the 1D simulations for
the same reason we neglected the convective term in equation (2). (W/m
3
) is the heat
generation term resulting from the exothermic reaction between C and O2. Radiation
heat transfer is assumed to be negligible following Manovic et al. (2008).
5.4.2. Chemical reaction
The char particle is assumed to be 100% C as a simplification. In reality, this is not
the case since the existence of inherent moisture, sulfur, nitrogen, and other non-carbon
compounds will affect combustion characteristics to some extent. For char particles,
reaction contributions from sulfur and nitrogen are neglected since their contributions

are much smaller than carbon containing compound. Char combustion can be described
by the following single-step heterogeneous reaction

C +O
2

CO
2

(4)

The volumetric reaction rate, (mol/m
3
s) is defined as

=
1

2
(5)

where
1
and
2
are concentration of carbon and oxygen respectively, and , is the reac-
tion rate constant given by the Arrhenius type relation involving the pre-exponential
factor, ; activation energy, ; and temperature .

= exp(

)
(6)

Stoichiometry of reaction (4), gives the heat and species source terms as

1
= (7a)

2
= (7b)

3
= (7c)
= (7d)

5.4.3. Boundary and initial conditions
Spherical symmetry imposes concentration and temperature gradient of zero at the
particle center.

r = 0

= 0
= 0
(8a)

Heat and mass transfer between the particle and the surrounding gaseous environment
is accounted for in the boundary expression related to the particle surface.

r = r
0

(
,

,
)
= (
)
(8b)

At the start of the simulation ( = 0), concentration of species and temperature is con-
stant along the radius of the particle and is equal to concentration in the surrounding.
The surrounding temperature is increased from its initial value to the target value,
,
early in the simulation using a Heaviside-smoothed step function.
Ranz (1952) correlation for the convective heat transfer coefficient, , of a single
particle is presented in terms of Nusselt number and particle Reynolds number is

Nu
p
= 2 +0.6Re
p
0.5
Pr
b
0.33
(9)

where Pr
b
= 0.27, and

Nu
p
=

0
(10)

Re
p
=

(11)

The mass transfer coefficient,
, between a single particle and gas stream is corre-

lated by Nauze and Jung, (1986).

Sh
p
= 2 +0.6Re
p
0.5
Sc
b
0.33
(12)

where

Sh
p
=

b
(13)

Sc
b
=

b

(14)

Constants used in the simulations are presented in Table 5.1.
Table 5.1. Model constants used in the simulations.
Parameter Value Parameter Value
1.27 10
5
mol/m
3
s
,
1 kJ/kg K
160 kJ/mol
,
2 kJ/kg K
393 kJ/mol
1.22 kg/m
3

,
3 10
5
m
2
/s
1000 kg/m
3

,
7.5 10
6
m
2
/s Pr 0.27
0.06 W/m K c
1,
4.27 10
4
mol/m
3

0.55 W/m K
5.5. RESULTS AND DISCUSSION
In the numerical simulation, a single char particle is assumed to be free of moisture,
volatiles, and other non-carbon compounds. The particle is suspended in a flowing ga-
seous medium that provides a constant supply of oxygen and heat at bulk temperature
. The initial conditions of the particle are such that it is at a uniform temperature of
500 K, and oxygen concentration within the particle is equal to the oxygen concentration
in the bulk. This is due to the assumption that prior to the start of the simulation, the
single coal particle was already exposed to the bulk environment long enough to lose all
its moisture and volatiles without ignition. This exposure time includes the time for
drying and pyrolysis of the coal particle at a temperature below ignition point. Since no
reaction occurs in this exposure period, temperature and oxygen concentration in the
particle reach equilibrium with its bulk surrounding by the start of the simulation. The
bulk oxygen mole fractions used in the parametric study are 5%, 10%, 40%, and 80%;
temperature values are 600 K, 700 K, 800 K, 900 K, and 1100 K; bulk gas velocities are
0.2 m/s, 0.3 m/s, 0.4 m/s, 0.5 m/s, 0.7 m/s, and 0.8 m/s; and particle radius are 0.5 mm,
1.0 mm, 1.5 mm, and 2.0 mm. For comparison, results are shown together alongside the
reference case simulation. Values of variables used for simulation of reference case and
parametric study are listed in Table 5.2 Simulation results are validated by comparison

with works reported by other investigators (Manovic et al., 2008; He et al., 2003; Beja-
rano and Levendis, 2008).
Table 5.2. Variable values for reference case and parametric study.
Case

(mm)

(%)
(K) (m/s)
Reference 2.5 20 1000 0.6
Parametric 0.5, 1.0, 1.5, 2.0 5, 10, 40, 60, 80
700, 800, 900,
1100
0.2, 0.3, 0.4,
0.5, 0.7, 0.8
5.5.1. CHAR CONVERSION RATE
Simulation results show that in general, time required for complete conversion of
the char particle is dependent on all the parameters studied. Figure 5.4 a to c show that
increasing either the bulk gas temperature, bulk oxygen concentration, or bulk gas ve-
locity, reduces char burnout time. Reduction in burnout time is marginal for the range of
gas velocities studied, whereas bulk oxygen concentrations and gas temperatures have
more significant effect.
It may be seen that conversion at bulk temperature of 700 K proceeds slowly at first
but approaches the conversion rate achieved by the combustion of char at higher tem-
peratures. The main discrepancy in conversion rates for the range of temperature stu-
died occurs only in the initial rate period, in which the conversion rate difference is
greatest between 700 K and 800 K, but marginal for bulk temperatures above 800 K.
This is physically explained by considering the timescale of heat conduction from the
particle surface to particle center. Figure 5.4a shows that at a relatively low tempera-
ture of 700 K, time required to raise the particle center to the bulk temperature is consi-
derably longer compared to time required to raise the particle center at higher bulk
temperatures. The initial reaction rate must then be related to the number of active sites
having sufficient energy to undergo the combustion reaction. Beyond the initial rate pe-
riod, increase in the bulk temperature has very little effect on the conversion rate as ob-
served in the slope of Figure 5.4a. A large fraction of the combustion period therefore
lies in the diffusion-limited regime, whereas chemical reaction kinetics is dominant in
the initial rate period.
Variation in oxygen concentration has greater impact on char conversion rate. As
oxygen concentration increase, full conversion of the char particle takes considerably
less time. At bulk temperature of 1000 K, the temperature at the center of the char par-
ticle is quickly raised to the bulk value, meeting the minimum energy requirement for
combustion as defined in the Arrhenius relation of equation (6). However, combustion
can only proceed when the active sites are exposed to oxygen. Since reaction mostly oc-
curs within a thin shell (Jelemensk et al., 2000) at the surface of the particle, it can be
deduced that the rate of char conversion is related to the thickness of the reacting shell,
which in turn is dependent on chemical kinetics and species diffusion within the particle.
Thus, increasing the bulk oxygen concentration increases the mass transfer driving force,
resulting in thicker reacting shell and higher conversion rate.
Bulk flow velocity has some effect on conversion rate for the range of velocities stu-
died. As shown in Figure 5.4c, burning rates increase with bulk velocity due to the flow-
enhanced transfer of reactants between gas and particle. The model assumes no particle
attrition and no fragmentation of the forming ash layer, which is generally the case for
FB combustion at low fluidizing velocities (<1 m/s). The particle Reynolds number in

the correlation lies in the laminar regime. It is expected that bulk velocities approaching
the turbulent regime will significantly enhance heat and mass transfer between gas and
particle, increasing combustion efficiency while at the same time result in greater loss
due to fragmentation (Zhang et al., 2002). However, this cannot be studied in the
present model due to the simplifications implemented.
Figure 5.4d and Figure 5.4d shows the influence of particle size on char burning
rate and center temperature of the particle. The observation that time required for com-
plete burnout increase with particle size can be explained by simply considering the
amount of combustible material available. In general, a smaller particle takes less time
to completely burn compared to a larger particle of the same composition. This result is
rather intuitive and therefore consistent with our expectation. Figure 6d also shows that
the conversion rates are high for small particles. This is explained by considering the
ratio of reacting shell thickness , and particle radius
0
. For small particles, /
0
is large,
enabling a larger fraction of the particle region react exothermically with oxygen. Fur-
ther, the mass transfer coefficient
, being inversely proportional to the Sherwood

number, is larger for smaller particles as shown in Figure 5.5.

Figure 5.4a. Effect of bulk temperature on char conversion ratio, for temperatures of
700 K, 800 K, 900 K, 1000 K (reference case), and 1100 K.

Figure 5.4b. Effect of oxygen concentration on char conversion ratio, for oxygen
mole fraction of 5%, 10%, 20% (reference case), 40%, 60%, and 80%.

Figure 5.4c. Effect of bulk velocity on char conversion ratio, for bulk velocities of 0.2
m/s, 0.3 m/s, 0.4 m/s, 0.5 m/s, 0.6 m/s (reference case), 0.7 m/s, and 0.8 m/s.

Figure 5.4. d. Effect of particle size on char conversion ratio, for particle radius of
0.5 mm, 1.0 mm, 1.5 mm, 2.0 mm, and 2.5 mm (reference case).
5.5.2. TEMPERATURE PROFILES
Figure 5.7a present the temperature at the particle as a function of time for differ-
ent bulk temperatures. The temperature profiles show that heating of particle by bulk
gas cause an initial increase in particle temperature. The slope for the heat-up period is
gentle for low bulk temperatures, but increases tremendously as bulk temperature rises.
At a bulk temperature of 600 K, the particle center approaches the bulk temperature
without further increase in temperature, leading to the conclusion that the particle do
not ignite at that temperature. Center temperature rise beyond the bulk temperature
rapidly for bulk temperatures above 800 K, implying that ignition take effect within a
relatively short time from the start of the heat-up period. Particle heat-up is more gra-
dual for the bulk temperature of 700 K, delaying the onset of ignition considerably.
Beyond the heat-up period, diffusion-limited combustion render no distinction in heat-
ing rates for temperatures above 700 K. Practically, this means that once ignition has
taken place, increasing the bulk temperature has no effect on heating rate. From the
slope of the graphs, it can be deduced that the ignition temperature is 800 K.
Combustion in diffusion-limited regime has greater sensitivity towards changes in
bulk oxidizer concentration. Figure 5.4b shows that peak center temperature and heat-
ing rate increase with bulk oxygen concentration. High bulk oxygen concentration intro-
duces sharp concentration gradient at the surface of the particle, resulting is thicker
reacting shell thickness. It is apparent that the time to reach peak temperature is inde-
pendent of bulk oxygen, but is dependent on all other parameters studied. It is also
worth noting that for bulk oxygen concentration of 80%, peak temperature is reached at
around half the time required for full conversion.


Figure 5.5. Sherwood number as a function of particle size for different bulk veloci-
ties from the correlation of [29]

Figure 5.6. Nusselt number as a function of particle size for different bulk velocities
from the correlation of [28].

While peak center temperature increase with bulk temperature and concentration,
Figure 5.7c shows that this is not the case for velocity. For the range of velocities stu-

died, the differences in consecutive heating rates are marginal. Heating rate and maxi-
mum particle temperature respond inversely with bulk gas velocity due to greater cool-
ing effect at higher gas velocities. In the range of particle size studied, Figure 6d reveal
that peak particle temperature do not vary much. It is also observed that smaller par-
ticles tend to heat-up more quickly than larger particles due to a larger /
0
ratio for
smaller particles. Further, Nusselt number is low for small particles, resulting in higher
heat transfer coefficient and greater cooling rate as shown in Figure 5.6 and equation
(10).

Figure 5.7a. Effect of bulk temperature on the temperature at the center of char
particle, bulk temperatures of 600 K, 700 K, 800 K, 900 K, 1000 K (reference case), and
1100 K.

Figure 5.7b. Effect of oxygen concentration on temperature at the center of char
particle, for oxygen mole fractions of 5%, 10%, 20%, 40%, 60%, and 80%.

Figure 5.7c. Effect of bulk velocity on temperature at the center of char particle, for
bulk velocities of 0.2 m/s, 0.3 m/s, 0.4 m/s, 0.5 m/s, 0.6 m/s (reference case), 0.7 m/s,
and 0.8 m/s.

Figure 5.7. d. Effect of particle size on temperature at the center of char particle, for
particle radius of 0.5 mm, 1.0 mm, 1.5 mm, 2.0 mm, and 2.5 mm (reference case).

5.5.3. Model validation
The above results are consistent with the work of Cai et al. (2003), which exhibits a
peaking temperature profile at the center of the particle. The difference however, is in
the absence of ignition characteristics and cooling period following the peak. While our
investigations reveal heat-up, combustion, and cooling-off periods in the combustion
process, the work of Cai et al. 2003 suggests that the heating and cooling rates are al-
most identical. Figure 5.8 shows that the peaking of particle temperature occur at ap-
proximately 40% conversion, marking a tipping point in heat generation and particle
cooling. Physically, this means heat generation rate is greater than heat loss rate for
conversion below 40%. Heat generation is balanced by heat removal at around 40%
conversion, marking a peak center temperature. Beyond 40%, cooling rate has greater
effect than heat generation resulting in decreasing temperature. This observation is ex-
plained by considering the relationship between particle size, Nusselt number and con-
vective heat transfer coefficient as explained in the previous section.

Figure 5.8. Temperature at the center of particle as a function of conversion ratio at
bulk oxygen mole fraction of 20%, and bulk temperatures of 770 K [33] and 800 K.

For the present study, results show that temperature reaches its peak of 1649 K at
around 25% conversion. This is twice the temperature achieved by Cai et al. 2003 in
which the maximum center is approximately 1000 K. The results from Cai et al. 2003 is
for the combustion of 0.5 mm particle (35 mesh) in bulk temperature of 770 K, which is
compared against a subset of the present study for 1 mm particle in bulk temperature of
800 K. In our parametric study of the effect of particle size with center temperature, it
was found that smaller particles tend to achieve higher center temperatures (refer to
Table 5.6 and Figure 5.7d). Hence, it is concluded that the present model tend to pro-
duce exceptionally high temperatures as a result of the simplifications applied.
He et al. (2003) considered the effects of pore structure in their numerical study of
single particle combustion of coal chars from different types of coal. It was shown that
the combustion time at bulk flow temperature of 1273 K is highly influenced by the size
of particles in the range of 50-500 m. At bulk temperature of 1700 K, char combustion
occur so rapidly that particle size has virtually no effect on combustion time. Figure 5.9
present the work of He et al. alongside the present study results. The combustion time at
1000 K in the present study compares with combustion at 1700 K from Hes investiga-
tions. In fact, the combustion time for the present study should be closer to Hes results
at 1000 K. This confirms our earlier conclusion that the present model produces com-
bustion rates and particle temperatures that are generally much higher than those re-
ported by other investigators. This could be due the simplifications implemented, and
also the values of the parameters. Table 5.3 to Table 5.6lists the simulation results from
the present study.

Figure 5.9. Combustion time as a function of particle size for combustion of char
particle in bulk oxygen concentration ~20%. Marked lines are adapted from He et al.
(2003).
5.5.4. Tabulated 1d model results
Table 5.3. Simulated char burnout times (
), maximum center temperature (
),
and maximum temperature difference between bulk and particle center (
), for dif-
ferent bulk oxygen concentrations.
Bulk oxygen mole fraction,

(%)
5 10 20 40 60 80
(s) 2646 1333 679 352 243 188
(K) 1283 1490 1814 2340 2796 3218
(K) 283 490 814 1340 1796 2218

),
), for dif-
ferent bulk temperatures.
Bulk temperature,
(K)
700 800 900 1000 1100
(s) 878 723 693 679 671
(K) 1565 1649 1731 1814 1899
(K) 865 849 831 814 799

),
), for dif-
ferent bulk velocities.
Bulk velocity, (m/s)
0.2 0.3 0.4 0.5 0.6 0.7 0.8
(s) 751 723 705 690 679 670 663


(K) 1910 1878 1853 1833 1814 1800 1786
(K) 910 878 853 833 814 800 786

),
), for dif-
ferent particle size.
Particle radius,
(mm)
0.5 1.0 1.5 2.0 2.5
(s) 39 125 262 447 679
(K) 1931 1900 1865 1838 1814
(K) 931 900 865 838 814


5.5.5. 2D Model

Figure 5.10. Temperature evolution across particle diameter for bulk temperature
of 1000 K. Oxygen mole fractions of A: 10%, b: 20%, and C: 30%.


We performed the 2D simulations with the aim of studying the effect of particle
shape on the combustion characteristics of a single particle coal/biomass conjugate.
This was not achieved due to the difficulty in obtaining a fully converged solution. Figure
10 shows that a 2D formulation is more able to account for spatial differences in tem-
perature and species, producing results that reflect a physical phenomenon more closely.
A 1D simulation assumes that a spherical char particle undergoes uniform combustion
along its surface, shrinking as it is consumed without any change in shape. The above
figure shows the asymmetrical nature of the problem, in which the reaction front is
skewed towards the side that is facing the inlet supplying oxygen. The partial results in
the above figure show that the change in temperature across the particle diameter is
more gradual for lower bulk O2. At higher O2 concentrations (> 20%) the rise tempera-
ture is more rapid. This is consistent with our findings in the 1D simulation. Transient
solution is only available for the first 32 s of simulation time for the O2 concentration of
30%. It was found that as O2 is reduced, convergence is faster. However, full conver-
gence is possible for O2 concentrations below 1%. The authors have attempted numer-
ous techniques in the attempt to mitigate this problem. These include studying mesh
dependencies, adjusting of time-steps, even to the extent of over-simplifying an already
highly simplified model.
5.6. CONCLUSION
A one-dimensional simplified char combustion model is used to numerically investi-
gate single particle char combustion. Numerical solutions show that high bulk tempera-
ture result in high conversion rate and high peak center temperature. Conversion rate is
also found to increase with bulk oxygen concentration. For the range of bulk oxygen
concentration studied, there is significant difference in the time taken to achieve full
char conversion for different bulk concentrations. On the contrary, increase of bulk tem-
peratures after the onset of ignition has very little effect on burning rate. This confirms
the diffusional limitation of the combustion. The effect of bulk flow velocity on conver-
sion rate and particle temperature is marginal for the range of velocities studied. Burn-
ing rates increase with gas velocity due to the convection-enhanced mass transfer be-
tween particle surface and gaseous surrounding. Similarly, higher gas velocity enhances
particle cooling, leading to lower peak temperature. Results also show that smaller par-
ticles tend to achieve higher peak temperatures and burn off faster. Comparisons with
the works of other investigators reveal that the present model tends to produce excep-
tionally high conversion rates and temperature. This could be due to the simplifications
implemented and the values of the parameters used. A more rigorous model and numer-
ical procedure will be implemented in future works.
ACKNOWLEDGEMENT
The authors acknowledge expert advice and extensive guidance from Dr. Sachin V.
Jangam, as well as assistance in multi-physics modeling from Jundika C. Kurnia. Not for-
getting the excellent mentorship of Professor Arun S. Mujumdar.


NOMENCLATURE
pre-exponential factor, kmol/m
2
s
species concentration, mol/m
3

specific heat, J/m

3
K
diffusion coefficient of species, m
2
/s
activation energy, kJ/mol
convective heat transfer coefficient between particle and fluidized bed, W/m
2
K
enthalpy of reaction, kJ/mol
mass transfer coefficient between particle and fluidized bed, m/s

pressure force, N/m
2

volumetric heat source, W/m
3

distance from centre of particle, m
0
radius of char particle, m
universal gas constant, 8.314 J/mol K

rate of reaction, mol/m
3
s
specific surface area of char particle, m
2
/m
3

source/sink term
time, s
temperature, K
velocity vector, m/s
carbon/char conversion ratio

Greek letters
rate constant, W/ mK
thermal conductivity, W/ mK
dynamic viscosity, Pas
density, kg/m
3

species mole fraction

Subscripts
bulk gas
100% char burnout
char particle
C ( = 1), O2 ( = 2), CO2 ( = 3)
initial value
Particle
reference/base case
char particle surface

REFERENCES
Avedesian, M., and Davidson, J., "Combustion of carbon particles in a fluidised bed,"
Chemical Engineering Research and Design, vol. 51, no. a, pp. 121-131, 1973.

Baum, M., and Street, P., "Predicting the combustion behaviour of coal particles,"
Combustion science and technology, vol. 3, no. 5, pp. 231-243, 1971.

Bejarano, P., and Levendis, Y., "Single-coal-particle combustion in O2/N2 and
O2/CO2 environments," Combustion and flame, vol. 153, no. 1-2, pp. 270-287, 2008.


Bhatia, S., and Perlmutter, D., "A random pore model for fluid solid reactions: I. Iso-
thermal, kinetic control," AIChE Journal, vol. 26, no. 3, pp. 379-386, 1980.

Bliek, Lont, J., and Van Swaaij, W., "Gasification of coal-derived chars in synthesis gas
mixtures under intraparticle mass-transfer-controlled conditions," Chemical engineering
science, vol. 41, no. 7, pp. 1895-1909, 1986.

Cai, Y., and Zygourakis, K., "A multiscale transient model for combustion of highly
porous chars," Industrial & engineering chemistry research, vol. 42, no. 12, pp. 2746-2755,
2003.

Chakraborty, R., and Howard, J., "Combustion of char in shallow fluidized bed com-
bustors: influence of some design and operating parameters," Journal of the Institute of
Energy, vol. 54, pp. 48-54, 1981.

Chen, J.C. Heat transfer in Handbook of Fluidization and Fluid-Particle Systems, Ed.
yang, W.C. Marcel Dekker, New York, 2003.

Comrie, D., "Production of cementitious ash products with reduced carbon emis-
sions," 2007.

Essenhigh, R., Misra, M., and Shaw, D., "Ignition of coal particles: a review," Combus-
tion and Flame, vol. 77, no. 1, pp. 3-30, 1989.

Essenhigh, R., 1981, Fundamentals of Coal Combustion, in Chemistry of Coal Utiliza-
tion, M. Ellis,. Editor, Wiley Interscience, New York.

He, R., Suda, T., Fujimori, T., and Sato, J., "Effects of particle sizes on transport phe-
nomena in single char combustion," International journal of heat and mass transfer, vol.
46, no. 19, pp. 3619-3627, 2003.

Hk, M., Zittel, W., Schindler, J., and Aleklett K., "A supply-driven forecast for the fu-
ture global coal production," contribution to ASPO, 2008.

Huang, G., and Scaroni, A., "Prediction and measurement of the combustion time of
single coal particles," Fuel, vol. 71, no. 2, pp. 159-164, 1992.

Jelemensk, L., Marko, J., ajdlk, R., and Remiarov, B., "Modelling of Nonlinear Be-
haviour during Combustion of Single Coal Char Particle," Chemical Papers, vol. 54, no. 6b,
pp. 473-481, 2000.

La Nauze, R., and Jung, K., "Mass transfer relationships in fluidized-bed combustors,"
Chemical engineering communications, vol. 43, no. 4-6, pp. 275-286, 1986.

Manovic, V., Komatina, M., and Oka, S., "Modeling the temperature in coal char par-
ticle during fluidized bed combustion," Fuel, vol. 87, no. 6, pp. 905-914, 2008.

Miller, J., Mitchell, R., Smooke, M., and Kee, R., "Nineteenth Symposium (International)
on Combustion," The Combustion Institute, Pittsburgh, pp. 181, 1982.

Mingle, J., and Smith, J., "Effectiveness factors for porous catalysts," AIChE Journal,
vol. 7, no. 2, pp. 243-249, 1961.


Meyers, R., "Coal handbook," 1981.

Oka, S., Ilic, M., Vukasinovic, B., and Komatina, M., "Experimental investigations of
mass transfer between single active particle and bubbling fluidized bed," Proceedings of
the 13th International Conference on Fluidized Bed Combustion, pp. 1419-1425, 1995.

Ranz, W., "Friction and transfer coefficients for single particles and packed beds,"
Chem. Eng. Prog, vol. 48, no. 5, pp. 247-253, 1952.

Sahu, R., Northrop, P., Flagan, R., and Gavalas, G., "Char combustion: measurement
and analysis of particle temperature histories," Combustion science and technology, vol.
60, no. 1-3, pp. 215-230, 1988.

Simons, G., "The pore tree structure of porous char," Symposium (International) on
Combustion, vol. 19, no. 1, pp. 1067-1076, 1982.

Smith, K., The structure and reaction processes of coal, Plenum Publishing Corpora-
tion, 1994.

Smith, I.W., 1982, Symposium (International) on Combustion In Nineteenth Sympo-
sium (International) on Combustion, 19, pp. 1045-1065.

Smith, I.W., , 1978, The intrinsic reactivity of carbons to oxygen, Fuel, 57, pp. 409-
414.

Sotirchos, S., and Amundson, N., "Diffusion and reaction in a char particle and in the
surrounding gas phase. A continuous model," Industrial & engineering chemistry funda-
mentals, vol. 23, no. 2, pp. 191-201, 1984.

Wang, X., Zhao, D., He, L., Jiang, L., He, Q., and Chen, Y., "Modeling of a coal-fired slag-
ging combustor: Development of a slag submodel," Combustion and flame, vol. 149, no. 3,
pp. 249-260, 2007.

Williams, A., Backreedy, R., Habib, R., Jones, J.M., and Pourkashanian, M., 2002, Mod-
elling coal combustion: the current position, Fuel, 81, pp. 605-618.

Winter, F., Prah, M., and Hofbauer, H., "Temperatures in a fuel particle burning in a
fluidized bed: The effect of drying, devolatilization, and char combustion* 1," Combustion
and flame, vol. 108, no. 3, pp. 302-314, 1997.

Yang, W., Handbook of fluidization and fluid-particle systems, CRC Press, 2003.

ajdlik, R., Jelemensk, L., Remiarova, B., and Marko, J., "Experimental and model-
ling investigations of single coal particle combustion," Chemical engineering science, vol.
56, no. 4, pp. 1355-1361, 2001.

Zhang, H., Chen, K., Yan, J., and Ni, M., "The fragmentation of coal particles during the
coal combustion in a fluidized bed 1," Fuel, vol. 81, no. 14, pp. 1835-1840, 2002.

2011
For more information please visit us at
http://serve.me.nus.edu.sg/arun/
http://www.mujumdar.net78.net/
Dr. Mujumdar, the world-renowned Drying Guru, is Professor of Mechanical Engineering at
National University of Singapore. Winner of numerous prestigious awards and honors, Prof.
Mujumdar has published over 400 refereed papers and over 100 book chapters in heat-mass
transfer and drying. Founder of the International Drying Symposium (IDS) series, he is Editor-
in-Chief of the archival journal Drying Technology (Taylor & Francis) since 1988 and is also the
Editor of over 60 books including the widely acclaimed Handbook of Industrial Drying (CRC
Press) now in third edition.
Dr. Agus Pulung Sasmito received his PhD from Division of Energy and Bio-Thermal System,
Mechanical Engineering Department, National University of Singapore in 2011. He received his
Bachelor Degree in 2005 from Gadjah Mada University, Indonesia. His research interest
encompass fuel cell modeling, industrial transport processes, mine ventilation and application
of computational fluid dynamics. He has made several conference presentations as well as
published several journal papers.
Jundika Candra Kurnia is currently pursuing his PhD in Division of Energy and Bio-Thermal
System, Department of Mechanical Engineering, National University of Singapore. He received
his Bachelor Degree in 2006 from Gadjah Mada University, Yogyakarta, Indonesia. Currently he
is working on transport phenomena and deformation behavior of stimuli-sensitive hydrogel. He
has published several journal and conference papers.
Dr. Sachin Vinayak Jangam received his PhD in Chemical Engineering from Institute of
Chemical Technology, Mumbai, India and is a Research Fellow at NUS working on experimental
and modeling of drying processes. He is co-editor of 5 e-books and author of several papers on
drying including screw conveyor drying, coal and biomass drying, fluid bed drying, solar drying
etc. He has worked on drying of various food products as major part of his PhD thesis using
heat pump and freeze drying systems.

Mathematical Modeling of Transport Processes

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Mathematical Modeling of Transport Processes

Uploaded by

Copyright:

Available Formats

Mathematical Modeling of

are the mole fractions of species and , and

stoichiometric coefficient for bulk reactant, and product

stoichiometric coefficient for gas reactant and product

stoichiometric coefficient for site reactant and product

rate of rth reaction

mass fraction of species i

are the stoichiometric coefficients for the

are the stoichiometric coefficients

bulk/solid species, mol

specific heat, Jkg

activation energy for the reaction, Jkgmol

gas species, mol

effective thermal conductivity, Wm

net rate of mass deposition, kg

reaction rate of species i, kgm

surface-adsorbed/site species, mol

heat release/absorb due to reactions, Wm

rate of rth reaction

mass fraction of species i

and percentage of pressure drop increases as phase difference

with fuel lagging behind air gave a more stable per-

is effective diffusivity of species in the char particle,

must account for carbon conversion such that as

, between a single particle and gas stream is corre-

, being inversely proportional to the Sherwood

), maximum center temperature (

(s) 2646 1333 679 352 243 188

(K) 1283 1490 1814 2340 2796 3218

(K) 283 490 814 1340 1796 2218

), maximum center temperature (

(s) 878 723 693 679 671

(K) 1565 1649 1731 1814 1899

(K) 865 849 831 814 799

), maximum center temperature (

(s) 751 723 705 690 679 670 663

(K) 1910 1878 1853 1833 1814 1800 1786

(K) 910 878 853 833 814 800 786

), maximum center temperature (

(s) 39 125 262 447 679

(K) 1931 1900 1865 1838 1814

(K) 931 900 865 838 814

specific heat, J/m

mass transfer coefficient between particle and fluidized bed, m/s

universal gas constant, 8.314 J/mol K

carbon/char conversion ratio

You might also like