Sie sind auf Seite 1von 9

Computers & Fluids xxx (2015) xxxxxx

Contents lists available at ScienceDirect

Computers & Fluids


j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / c o m p fl u i d

Coupled radiative and conjugate heat transfer in participating media


using lattice Boltzmann methods
Richard McCulloch, Hitesh Bindra
Department of Mechanical and Nuclear Engineering, Kansas State University, Manhattan, KS 66506, United States

a r t i c l e

i n f o

Article history:
Received 22 December 2014
Received in revised form 7 May 2015
Accepted 28 May 2015
Available online xxxx
Keywords:
Radiation
Lattice Boltzmann
Conjugate heat transfer
Porous media

a b s t r a c t
In the past, lattice Boltzmann methods (LBM) have been extensively developed for momentum and
energy transport in single-phase and multi-phase uid systems. Recently, LBM based algorithms have
been developed and applied to fundamental Radiative Transport Equations (RTE), including radiation
material interactions and were found very convenient to model radiative energy exchange between
radiation and material medium. This work advances the development of Lattice Boltzmann Equations
(LBE) for radiative transport by integrating them with existing LBEs for energy and momentum transport
for solving multi-physics problems. The multi-physics example problems of thermal energy transport
where radiation, conduction and convection all are considered as important modes are modeled via this
integrated LBM. These integrated LBM models are used to solve one and two dimensional problems, and
highlight the advantage of this approach for solving multi-physics problems in a single framework. First
example involves modeling radiative and conductive heat transfer in one-dimensional slab using LBM.
The numerical results are compared with existing benchmark P1 solutions. Next example is the simulation of two-dimensional radiative porous burner with hot walls. This problem is simulated with two
numerical models: a homogenous porous media and a heterogenous model of packed obstacles which
have differential scattering and absorption interactions. The homogenous model uses analytical velocity
eld and provides a simpler approach, but has limitations in providing detailed analysis. In heterogenous
model velocity eld, temperature eld and radiation eld are computed with a set of coupled LBEs. Fluid
owing through heterogenous porous media undergoes conjugate heat exchange with obstacles and also
interacts with isotropic incident radiation. These two-dimensional example cases with different material
properties are solved with D2 Q 16 LBE template.
2015 Elsevier Ltd. All rights reserved.

1. Introduction
The advancement in the eld of high temperature materials
over the period of last two decades has potential to completely revolutionize the energy industry with the design of high efciency
systems. With the higher temperature in the energy system applications, radiation heat transfer becomes equally important along
with conduction and convection modes of heat transfer.
Therefore, the radiative mode of heat transfer with conduction
and convection has growing signicant practical importance in
many engineering applications [13]. Some of these industrial
applications include the manufacturing of glass, design of insulating material, weather forecasting, porous burners, solar collectors,
high temperature nuclear reactors etc. In the past, most of the
commercial and academic codes and calculations for the
Corresponding author. Tel.: +1 785 532 3039.
E-mail address: hbindra@ksu.edu (H. Bindra).

multi-mode heat transfer treated radiative transport using surface


to surface boundary conditions. There are very few numerical
codes which solve the radiative transport within the participating
media where conduction and convection calculations are performed. The radiative portion in a multi-physics set-up requires
tremendous computational effort due to the complexity of radiation interaction within a medium which is directionally sensitive
process.
The RTE solutions are computationally expensive due to the fact
that at each point of evaluation all the radiative sources within the
medium have to be accumulated from their initial starting point to
the point of interest. This recursive nature of the RTE is compounded by the fact that not all radiation in the media has the
same energy, thus another dimension must be accounted for during the simulation. Moreover, radiation transport from a point to
the surrounding regions is highly directional dependent therefore
higher degree of angular resolution is required in computations
[411]. This implies there are additional independent variables

http://dx.doi.org/10.1016/j.compuid.2015.05.024
0045-7930/ 2015 Elsevier Ltd. All rights reserved.

Please cite this article in press as: McCulloch R, Bindra H. Coupled radiative and conjugate heat transfer in participating media using lattice Boltzmann
methods. Comput Fluids (2015), http://dx.doi.org/10.1016/j.compuid.2015.05.024

R. McCulloch, H. Bindra / Computers & Fluids xxx (2015) xxxxxx

i.e. energy level and angular direction. Due to larger number of


independent variables for RTEs as compared to standard continuum mechanics equations, Monte Carlo methods are preferred
over deterministic solvers [12]. But Monte Carlo methods are not
so convenient for conventional continuum mechanics problems
such as uid mechanics or convectiondiffusion of scalar quantities. Therefore, a new approach or framework is required for
multi-physics problems with radiation transport as one of the signicant physical processes. Some new multi-physics problems
involving radiationmaterial interactions at mesoscale such as
photo-thermal heating and mixing and optical tweezers have not
been understood quantitatively as they required mutually coupled
modeling tool between radiation transport and particle transport.
LBM has been proposed as a new method for solving multi physics
problems involving RTE. However, most of the effort was focused
on coupling Discrete Ordinates Method (DOM) or other conventional deterministic methods for solving RTEs with the LBM based
uid transport or convectiondiffusion solvers [1316]. Recently,
LBM based algorithms have been developed to solve RTE problems
[1722]. This paper proposes the use of those LBM algorithms for
solving coupled multi physics examples. The problems shown in
this paper extend the application of LBM to non-linear temperature
driven coupled conductionradiation and coupled convection
radiation in different one-dimensional and two-dimensional
geometries.
This paper is divided into three main sections. The next section
discusses theoretical background on RTE, its coupling with material energy transport and adoption of LBM to solve those coupled
interactions. The subsequent section describes problem set-up,
numerical details and results for conductionradiation and convectionradiation examples solved using the LBEs. In this section,
thermal energy transport in a simplied version of porous burner
is modeled with convection, conduction, and radiation simultaneously. In the last section nal conclusions are provided.

StefanBoltzmann coefcient. These equations can also be written


using total cross-section or attenuation coefcient b ja js as a
parameter. The heat transport system where radiative transfer in
participating media is signicant, coupled simulation of Eqs. (1)
and (2) is required. The velocity eld in Eq. (2) is computed from
uid-momentum transport equations. The LBE solvers for uid
momentum and energy transport are now widely used for single
phase and two phase calculations. Thus, LBE solver for Eq. (1) will
help in building a single framework to solve uid momentum,
energy and radiation transport. The LBE solutions for uid momentum and energy (convectiondiffusion) transport has been discussed extensively in the past, so this work will use some of
those widely accepted algorithms for the numerical examples. The
LBE for the general form of RTE (Eq. (1)) will be discussed in the following subsection.
2.2. Lattice Boltzmann method for RTE
The standard Lattice Boltzmann Equations (LBE) model as
derived previously in several studies is the discrete model in space,
time and lattice velocities. This discrete model is used to evaluate
distribution function for different lattice velocities. The LBE equations conserve the continuum mechanics equations for the macroscopic quantities such as mass, momentum and kinetic energy. A
similar analogue for linear form of the general Boltzmann equation,
which describes radiative transport process, has been developed
with discrete angular directions for radiation eld [1719]. The
distribution function or radiation intensity for each angular direction provides information for current and ux required in thermal
energy balance calculations. The LBE for solving the radiation
transport equations were formulated by Ma et al. [19], and were
shown to conserve macroscopic energy ux using Chapman
Enskog expansion. They used a standard RTE with source term
but without the scattering integral

2. Mathematical formulation

@Ir; X; t
X  rIr; X; t ra Ir; X; t sX r; t:
@t

2.1. Radiationmaterial energy interactions

and it was shown to have a corresponding LBE equivalent as

IX r XDt; t Dt  IX r; t sX r; tDt  ra IX r; tDt:


The radiationmaterial interaction can be considered as equilibrium interaction, where radiation and material energies reach an
instantaneous equilibrium, or non-equilibrium interaction, when
there is a lag between radiative energy and material energy. The
non-equilibrium interaction of radiation and material energies is
modeled by coupling the radiation transport equation (RTE) with
the material energy balance equation [23]. Assuming homogenous
materials and isotropic scattering, the set of governing equations
for non-equilibrium radiative transfer can be written as



@Ir; X; t
1
rTr; t4  Ir; X; t
X:$Ir; X; t ja
@t
4p


Z
1
Ir; X; tdX  Ir; X; t
js
4p 4p
Sr; X; t

The LBE for a general three dimensional system was modied


by Bindra et al. [20] to incorporate weighted summations for scattering integrals. Using the taylor series expansion on these discrete
LBEs, it can be shown that these equations are rst order accurate
in space and time. In this current work we will use this formulation
to solve the RTE form stated in the previous section. The equivalent
LBE for this RTE, Eq. (1), is

h
i
IX r XDt; t Dt  IX r; t ja wX Tr; t4  IX r; t Dt
"
#
X
js wX IX0 r; t  IX r; t Dt
X0

sX r; tDt:
1

5
For a two dimensional Cartesian LBE lattice, Eq. (5) reduces to

and

qC p

@Tr; t
qC p v :rT r:krT
@t
Z

Ir; X; tdX  rTr; t4
ja

Ii x v i;x Dt; y v i;y Dt; t Dt Ii x; y; t v Dt


2

4p

where Ir; X; t is the radiation intensity at any spatial location r and


time t in the direction X; Tr; t is the temperature; v is the
velocity eld of the uid (material) stream; ja is the absorption
cross-section; js is the scattering cross-section; Sr; X; t is the
volumetric energy source; C p is the specic heat; and r is the

 bIi x; y; t ja wi Tx; y; t4 js wi

!
X
Ij x; y; t ja wi sx; y; t ;
j

6
where Ii x; y; t is the discrete angular neutron ux in the ith lattice
direction at x; y location at time t. Lattice velocities v i;x and v i;y are
chosen such that the distance traveled by particles in a discrete
time interval Dt i.e., v i;x Dt and v i;y Dt equals to the distance between

Please cite this article in press as: McCulloch R, Bindra H. Coupled radiative and conjugate heat transfer in participating media using lattice Boltzmann
methods. Comput Fluids (2015), http://dx.doi.org/10.1016/j.compuid.2015.05.024

R. McCulloch, H. Bindra / Computers & Fluids xxx (2015) xxxxxx

Fig. 1. Various 2D LBM lattice arrangements showing (a) D2 Q 4 , (b) D2 Q 8 , and (c) D2 Q 16 lattices.

the neighboring lattice points. Furthermore, wi represents


directional weighting factor or weights for each direction of
travel. Fig. 1 illustrates these discrete model directions in a
two-dimensional cartesian space
In two-dimensions, the most commonly used template for LBE
is D2 Q 8 , which has 8 directions 1  8 (Fig. 1), and the standard
corresponding notation for v i can be written as

vi

8 

p
i1p
>
< v cos i1
i14
;
sin
2
2



p  
i5
p
i5
p
>
: 2v cos 2 p4 ; sin 2 p4
i58

vi

The effective speed of the rst four cardinal directions i 1  4


are equal to v and for diagonal directions i 5  8 are equal to
p
2v . The weights wi for directions with the same speeds are equal
because of symmetry. The weights for each cardinal and diagonal
direction lead to isotropic generalized lattice tensors of rank 2
and 4 [24,25]. This is achieved by conservation of lattice angular
moments up to the fourth order over scattering or any other radiation interaction. In case of an isotropic function wl, the odd
moments vanish,
8
X

v ia wi

wll dl 0

1

i1
8
X

v ia v ib v ic wi

wlla lb lc dl 0;

1

i1

Table 1
Weights wi for different two-dimensional lattice.

Lattice

14

58

9  16

D2 Q 4
D2 Q 8
D2 Q 16

0.25
0.20
0.196

0.05
0.028

0.013


8 
i1p
i1p
>
i14
>
> v cos 2 ; sin 2
>
< p  



i5p
i5
p

2v cos 2 p4 ; sin 2 p4
i58
>
>






>
p

>
: 5v cos i9p p ; sin i9p p
i 9  16
4
8
4
8

13

The corresponding weights, wi for different discrete directions


as illustrated in Fig. 1 are tabulated in Table 1.
As stated earlier in the introduction, momentum and energy
transport in continuum media has been studied by various
researchers in the past using lattice Boltzmann methods.
Therefore, the details or background on those can be found in listed
references. Appropriate version of Eq. (2), i.e. material energy
equation, with associated boundary conditions and corresponding
LBE will be stated from the literature for particular example studies in the following section.
3. Numerical examples
3.1. 1D Example: conduction and radiation

and the even moments can be expressed as

Z
8
X
wi

wldl 1

1

i1
8
X

v ia v ib wi

1

i1
8
X

wlla lb dl dab

v ia v ib v ic v ig wi

i1

10

11

1

wlla lb lc lg dl dab dcg dac dbg dag dbc ;


12

where a; b; c, and g are the indices for representative dimensions.


These moment equations are solved for the D2 Q 8 LBE system, and
the weights for cardinal and diagonal directions are obtained as 15
1
and 20
, respectively. D2 Q 8 LBE system has been shown to be adequate for modeling uid mechanics problems in two dimensions,
but in the case of RTEs a higher number of angular discretization
schemes may be required where the non-diffusive transport mechanism of radiation is important. Therefore, a higher order angular
discretization scheme D2 Q 16 is adopted, which has 1  8 directions
as described in D2 Q 8 and additional directions 9  16 with specic
lattice velocities (Eq. (13)) as shown in Fig. 1.

The benchmark problem of coupled radiation and conduction in


a one dimensional participating media has been studied previously
in a slab geometry [23,26]. These studies used other deterministic
methods (such as DOM, Finite Volume, Spherical Harmonics) for
solving the radiation transport part and here LBM is used for simulating both radiation and conduction modes. This particular
example of slab geometry, considered here, has two walls (boundaries) at xed temperatures. One of these walls is at higher temperature as compared to rest of the media. Due to this temperature
difference there is heat ux due to radiation and conduction
effects. The radiative eld contributes to material temperature
which contributes back to the radiative eld. The geometry and
boundary conditions for this problem are shown in Fig. 2. The
one-dimensional representation forms of Eqs. (1) and (2) modeling
radiative intensity and material temperature are obtained from
Ref. [23] as Eqs. (14) and (15),

Z
dIx; l ja
j 1
rT 4 s

Ix; ldl  ja js Ix; l


dx
4p
2 1


Z 1
2
d T
Ix; ldl:
k 2 ja rT 4 
1
dx

14
15

Please cite this article in press as: McCulloch R, Bindra H. Coupled radiative and conjugate heat transfer in participating media using lattice Boltzmann
methods. Comput Fluids (2015), http://dx.doi.org/10.1016/j.compuid.2015.05.024

R. McCulloch, H. Bindra / Computers & Fluids xxx (2015) xxxxxx

wi s v i Dt; t Dt wi s; t Dt
 wi s; t 1  xwi hs; t4 xwi

!
X
wj s; t :

23

Fig. 2. Geometry and boundary conditions with uniform radiationconduction heat


transfer parameters.

where l is the direction cosine of azimuthal direction, T is material


temperature, r is the StefanBoltzmann constant, ja is the absorption coefcient, and js is the scattering coefcient. In Eq. (15), k is
thermal conductivity of the material. The non-dimensional forms
of these equations are given by Eqs. (16) and (17).

Z
dw
h4
x 1
1  x
w
w dl
ds
4p
2 1


Z
1
@2h 1  x 4

h 
w dl
@ s2
N
1

16
17

In Eq. (23) the scattering integral is approximated by numerical


quadrature and wi are the weights associated with ith direction.
The time-step Dt is chosen to be equal to ds or Ds as commonly
used
approach
for
lattice
Boltzmann
methods.
The
non-dimensional temperature h is the coupling variable between
radiative and material energy equations.The LBE for material energy
transport equation (Eq. (17)) is derived in literature [27] as Eq. (24),


1
eq

fi  fi
0:5Dt



Z 1
1x 4
h

w
d
l
Dt
wm
i
N
1

f i s ~
ei Dt; t Dt f i s; t 

e2s

1

where f i are the distribution functions associated with each microei and corresponding weight wm
to compute
scopic direction ~
i
non-dimensional temperature. The microscopic direction vectors
and corresponding weights wm
i are obtained from literature [27].
Lattice speed of material energy transport equation is represented
eq
by es DDst . The equilibrium function f i is given in Eq. (25)
eq

f i wm
i h

where w rTI 4 is the dimensionless radiative ux, the dimensionless

and the non-dimensional temperature can be computed as

quantity s bx is the optical thickness, b ja js is the extinction


coefcient, x jajsjs is the scattering albedo, and h TTR is dimen-

sionless temperature. The temperature T R can be any reference temperature but here it is convenient to choose from one of the
boundaries. The dimensionless quantity N rkjT a3 is the
R

conduction-radiation parameter, also known as the Stark number,


that relates radiative and conductive heat uxes.
The boundaries or walls are held at different temperatures of
hL 1 and hR 0:5 for the left and right walls, respectively.
Boundary conditions for the radiation equation are calculated
based on emissivities of the walls along with the known temperatures at the boundary, as shown in Eq. (18).

ww h4w

1  e

lw dl

18

lw

24

X
f i:

25

26

For more details on discretized form of LBE for solving material


energy transport (or temperature) equation see Ref. [27].
The steady-state results with two direction LBM in one
dimension D1 Q 2 using w1;2 wm
1;2 0:5 are compared with P 1
benchmark solutions. The effect of Stark number N, or the radiationconduction parameter, is shown in Fig. 3. These calculations
are performed for perfectly black body surfaces e 1 and used
1
Dt Ds 160
. Other examples as shown in Fig. 4 depict the impact
of emissivity value for non-black body surfaces. All the results
obtain from different numerical exercises show that LBM results
are in good agreement with the P1 results.
Grid convergence study was performed for N 0:01 and e 1
case, the global error was computed as a function of grid size.

In Eq. (18) ww is the radiative intensity at the wall, e is the emissivity


of the wall, hw is the temperature at the wall and lw denotes integration domain over half-space on each wall. This problem has been
previously solved using P1 approximation (Ref. [23]), and those
solutions will be used as benchmark for comparison. The P 1 approximation for this one dimensional problem can be represented as
Eqs. (19)(22) (for details see Ref. [23]),
2

d G
3h4  G 0
ds
2
d h 1 4
h  G
ds2 N
dG 3 eL

h4  G 0
ds 2 2  eL L
dG 3 eR

h4  G 0
ds 2 2  eR R

19
20

s0

21

s1

22

R1
where G 1 w dl is the non-dimensional total radiative ux, and
eL and eR are the emissivities of the left and right faces respectively.
The LBE for one-dimensional slab RTE (Eq. (16)) can be written as an
analogue to generic form (Eq. (5))

Fig. 3. Temperature prole for varying stark numbers. N is the Stark number which
relates the interaction between radiation and conduction. hL 1; hR 0:5hL and
e1 e2 1.

Please cite this article in press as: McCulloch R, Bindra H. Coupled radiative and conjugate heat transfer in participating media using lattice Boltzmann
methods. Comput Fluids (2015), http://dx.doi.org/10.1016/j.compuid.2015.05.024

R. McCulloch, H. Bindra / Computers & Fluids xxx (2015) xxxxxx

transport. Accurate and high delity modeling of such systems


require fully coupled computer simulations of radiative heat transport with conjugate heat transfer in porous media. Talukdar et al.
[29] modeled similar problem in two dimensions using LBM for
evaluating the convective ux and Finite Volume Method (FVM)
for the radiative ux. In this example a porous media is placed
between two plates and uid is allowed to ow through the media.
The geometry and boundary conditions for this problem are shown
in Fig. 6. Steady State thermal transport with convection, conduction and radiation in two dimensional homogenous porous media
is governed by Eq. (28)

qC p U

Global error is computed with respect to the P 1 benchmark solution using

P
k i wi  Gk2
:
kGk2

@2T d T

@x2 @y2

!




@qr @qr

@x
@y

28

where U and V are velocities parallel and perpendicular to the ow,


respectively. In Eq. (28), the left-hand side terms are convection
terms, and the right-hand side terms represent conduction and
radiation heat transfer in the uid/solid continuum. The radiative
heat ux terms are explicitly modeled as

Fig. 4. Temperature prole for different wall emissivities (N 0:01).

EDs

@T
@T
V
@x
@y

27

The grid convergence results presented in Fig. 5 show that error


reduces linearly with grid size. As mentioned in the previous section the current form of LBE for radiative transport is a rst order
accurate method, which can be easily deduced by taylor series
expansion of the LBE for RTE. Other established MOC and DOM
methods for RTE are also rst order accurate method (Ref. [28]).
3.2. Radiationconvection in two-dimensional rectangular geometries
3.2.1. Homogeneous radiative porous burner
The high temperature practical applications such as a radiative
porous burner involve coupled radiative, conductive and convective heat transfer in a highly complex geometrical conguration.
Porous burners are getting increased attention due to their ability
to burn very low combustion energy materials or to allow higher
degree of thermal mixing. This is achieved by uniform material
mixing and allowing the transmission of radiative heat uxes
absorbed by the outer walls, which conduct and convect heat in
the axial direction away from the combustion chamber via uid





Z
@qr @qr
ja rT 4 

Ix; y; XdX
@x
@y
X4p

29

Neglecting the conductive and the radiative heat uxes in the


axial direction (x-direction), and assuming fully developed velocity
eld, the non-dimensional form (Eq. (30)) of the governing energy
transport equation in the porous media was derived in Ref. [29].

  2


@h
2 @ h
bW @ Wr


@ sx
Pe @ s2y
2NPe @ sy

30

The left-hand side of Eq. (30) is the convection of heat due to the
bulk motion of the uid passing through the channel. It is assumed
that the velocity in the transverse direction (y direction) is negligible, thus temperature is only convected axially, or in other words
convection is considered only for the x direction. The rst term
on the right hand side of Eq. (30) represents temperature diffusion
throughout the media in the transverse direction, where
2U W qC

Pe m k p is the Peclet number which relates the rate of convection to the rate of diffusion. It is assumed that in the axial direction
diffusion is negligible, thus diffusion is only considered in the
transverse direction of thickness W. The dimensionless thicknesses
in this example are dened as sx Wx and sy Wy , and dimensionless axial velocity u UUm . The second term on the right-hand side
of Eq. (30) is the thermal source term due to radiation. The stark
number as dened earlier in previous example is slightly modied
to N 4rbkT 3 , where T R is the reference temperature used to
R
non-dimensionalize the energy equation, and Wr rqTr4 is the
R

non-dimensional radiative heat ux. The dimensionless form of


radiative heat ux can be simplied to



Z
@ Wr
bW1  x h4 
w dX
@ sy
X4p

31

where non-dimensional radiation intensity w can be obtained from


two-dimensional RTE (Eq. (32))

@w
@w
h4
x
l g
bW 1  x
w
4p
4p
@ sx
@ sy

!
w dX

32

X4p

The Eq. (32) is similar to one-dimensional RTE (Eq. (16)), with


additional term g @@w
sy , where g is the directional cosine in second

Fig. 5. Grid convergence study with N 0:01 and

e 1 case.

dimension. The separate physics of radiative heat transfer in two


dimensions was modeled using D2 Q 4 LBM in the previous work
[20] for uniform (constant) temperature everywhere in the domain
and the results compared well against the S2 DOM solutions. The

Please cite this article in press as: McCulloch R, Bindra H. Coupled radiative and conjugate heat transfer in participating media using lattice Boltzmann
methods. Comput Fluids (2015), http://dx.doi.org/10.1016/j.compuid.2015.05.024

R. McCulloch, H. Bindra / Computers & Fluids xxx (2015) xxxxxx

1
Fig. 6. Homogenous porous burner with hot radiating walls. Thickness in y-direction (W 1); Aspectio Ratio = 4 and Grid spacing = 100
. The length scales used in picture are
in number of grid points.

LBE equation for RTE (Eq. (32)) with the constant temperature hc
everywhere in the media can be obtained, by re-writing Eq. (6)
with dimensionless variables, as

wi sx v i;sx Dt; sy v i;sy Dt; t Dt wi sx ; sy ; t


DtbW wi 1  x

h4c

!
X
 wi sx ; sy ; t xwi wj sx ; sy ; t : 33
j

Both LBE D2 Q 4 and DOM S2 solutions were relatively less


accurate due to use of only four angular directions in
two-dimensional planar geometry. As expected in solving RTE
the accuracy of solution increases with number of directions.
Higher angular resolution results, i.e. S4 scheme agree with the
exact solution [10]. The S4 conguration has been shown
[10,9,23] to be sufcient to capture the radiative heat transfer physics in two-dimensional rectangular or square enclosures.
Therefore, Eq. (33) was solved with higher angular directions i.e.
D2 Q 8 and D2 Q 16 lattice congurations with weights shown in the
Table 1. Similar to other existing LBE algorithms, problem is solved
by explicit time marching which requires the condition Dtb < 1 for
stability. Results shown in the Fig. 7, abstained from D2 Q 16 LBE
template with 16 directions match well with the benchmark
DOM S4 results. Therefore the two-dimensional problems studied
in this work use D2 Q 16 stencil for solving RTEs and are coupled
with established LBE models for convectiondiffusion models.
The simplication of the geometry into homogeneous porous
media enables the evaluation of velocity eld analytically. An analytical expression for ow inside homogeneous porous media was
obtained from literature [30], i.e.

usy ; c

1  e2c  1  ec ecsy 1 esy c

34

1  e2c  21  ec 2 c1

The fully-developed velocity eld is a function of the porosity


parameter c and dimensionless sy . For c ! 0, a non-porous media,
the velocity prole asymptotically approaches a laminar Poiseuille
ow velocity prole. The velocity eld scaled with maximum
Poiseuille ow velocity of 0.1 porosity parameter value of 0.1
(results shown in Fig. 8) is then substituted in the energy
equation which can be solved using LBM.
Mishra et al. [31] derived LBE to solve material energy equation
with radiation term as a source. The radiation source term was
computed using collapsed dimension method and then substituted
into the LBE form of energy equation. LBE models for convection
diffusion equations with source term have been developed in the
past [32,33]. Based on these existing LBE models for solving
convectiondiffusion equations with source terms, the LBE form
of Eq. (30) can be written as Eq. (35).


@f i
1
eq

~
fi  fi
ci  rf i 
1
2
@t
2Pecs 0:5Dt


bW
wi 
rWR
2NPe

35

where the equilibrium function for Eq. (35) is given as



~
ci  ~
u
eq
:
f i wm
h
1

i
cs

36

In Eq. (36) ~
ci is the unit vector in the streaming direction and cs DDsts
is the lattice speed. D2 Q 5 LBE template with appropriate weights for
distribution functions wm
i as mentioned in the literature [32] for
convection diffusion equation were used for these computations.
1

Numerical stability is ensured if the factor 2Pec2s 0:5Dt was


always maintained greater than 0.6, which has been reported by
Huang et al. [32] ensures stable solution for solving convectiondiffusion equations with source terms.
The physical example problem as presented in Fig. 6, has two
walls maintained initially at higher temperature than the porous
medium. These walls dissipate thermal energy into the porous
media with fully developed uid ow via radiative, conductive
and convective heat transfer. This computations are performed

P
Fig. 7. Total radiative ux
i wi solutions with Lattice Boltzmann and Discrete
Ordinate methods in a square enclosure. The results are line plots for
sx 0:5; hc 1 and all the walls kept at zero dimensionless temperature.

Fig. 8. Velocity prole for the case of both non-participating (left) and participating
(right) media as obtained from analytical expression Eq. (34). Dimensionless
velocity at the inlet (sx 0) is 1 for all sy .

Please cite this article in press as: McCulloch R, Bindra H. Coupled radiative and conjugate heat transfer in participating media using lattice Boltzmann
methods. Comput Fluids (2015), http://dx.doi.org/10.1016/j.compuid.2015.05.024

R. McCulloch, H. Bindra / Computers & Fluids xxx (2015) xxxxxx

for x 0:35; N 0:001; Pe 1600 and surface wall emissivity


e 1. In the case of a non-participating media, b 0 and radiation
does not affect the temperature of the uid in the channel as there
is no absorption or scattering within the media (Fig. 9). For the case
of participating media b 2 and radiation emitted from hot walls
is absorbed by the homogeneous porous matrix resulting in higher
temperature distribution in the domain. It can be seen in Fig. 10
that the radiation interaction causes the entire domain to heat
up instantaneously after entering the channel.
Due to the assumption of homogeneity, which allowed the use
of a single porosity parameter c, the approach used in this problem
cannot capture differential scattering and absorption in a real porous matrix. Therefore, accurate analysis of thermal transport at
high temperatures can only be considered reliable if heterogenous
medium with anisotropic properties is modeled.

tiveconductiveconvective heat transfer. The geometry and


boundary conditions for this problem are shown in Fig. 11. The size
of each obstacle is 0:1  0:1. The velocity eld around the obstacles
is solved by LBM and then substituted in energy transport equation. Conduction, convection, and radiation from the walls, obstacles and uid media were solved simultaneously. In other words
all the multi-physics processes involved here are simulated via a
fully coupled LBM solver. The uid ow in continuum regime is
modeled using the non-dimensional steady-state form of Navier
Stokes equations for incompressible ow with no body force along
with the continuity equation (Ref. [34])

@u
@u
@p
1 @2u @2u
u
v


@ sx
@ sy
@ sx Re @ s2x @ s2y
u

3.2.2. Heterogeneous radiative porous burner


The detailed analysis of radiationconductionconvection in
heterogeneous porous media with realistic void fraction of
3040% and much smaller characteristic dimension of thermal
transport as compared to the entire geometry will be computationally challenging task. Moreover, the analytical velocity prole used
in the homogeneous case will no longer be valid for this
heterogeneous region. A simple heterogenous model with few
solid obstacles is used for the detailed analyses of coupled radia

@v
@v
@p
1 @2v @2v
v


@ sx Re @ s2x @ s2y
@ sx
@ sy

!
37

@u
@v

0
@ sx @ sy
In these equations u and v are non-dimensional velocities in the sx
and sy directions, respectively, p is non-dimensional pressure,
Re 2Umm W is the Reynolds number and m is the kinematic viscosity.
The LBEs to solve this classical non-dimensional uid mechanics
problem have been developed and presented in listed [24,37] and
several other references. The most commonly used LBE form is

@g k
1
g k  g eq
~
ckf  rg k 
k
2 1
@t
2mc 0:5Dt

38

fs

The corresponding equilibrium function for Eq. (38) using LBM


with the BGK approximation [3537] is shown in Eq. (39).

"
f
g eq
k wk 1

Fig. 9. Temperature distribution for the case of non-participating homogeneous


porous media i.e. without any radiation absorption in the domain. Steady state
temperature distribution shows negligible temperature gain in the media due to nill
radiation absorption and very high Peclet number.

In Eq. (39) cfs pk3 is the lattice speed, ~


ckf DDstx ^i

39
Dsy ^
j is
Dt

the unit vec-

wkf

tor in the streaming direction,


is the weight of the k-th direction
1
and q is the density. The standard D2 Q 9 with weights 36
for diagonal
directions, 19 for cardinal directions and 49 for the stationary microscopic velocity [24] were used for solving uid mechanics problem.
The equation governing the temperature distribution is given
by Eq. (40).

ux

Fig. 10. Temperature distribution inside heterogeneous porous media with strong
radiation absorption. Steady state temperature distribution equilibrates to the
temperature of bounding hot walls throughout the domain.

~
ckf  ~
u 1 ~
ckf  ~
u
1~
u2


2
4
2
2 c2fs
cfs
cfs

@h
@h
2 @2h @2h
uy

@ sx
@ sy Pe @ s2x @ s2y

!




bW @ Wr @ Wr

2NPe @ sx
@ sy

40

The signicance of individual terms in Eq. (40) and its LBE form
are identical to the homogeneous case and the LBE form of this
convectiondiffusion equation is identical to the homogeneous
case. The radiative heat ux term rWr is computed using Eqs.
(31)(33) as described before. The uid mechanics in this problem

Fig. 11. Heterogeneous Porous Burner with large ow obstacles with differential material properties. Size, aspect ratio and grid size are chosen to be exactly same as
homogeneous case.

Please cite this article in press as: McCulloch R, Bindra H. Coupled radiative and conjugate heat transfer in participating media using lattice Boltzmann
methods. Comput Fluids (2015), http://dx.doi.org/10.1016/j.compuid.2015.05.024

R. McCulloch, H. Bindra / Computers & Fluids xxx (2015) xxxxxx

Fig. 12. Velocity prole around the ow obstacles solved with LBM.

Fig. 13. Temperature contour for the entire 2-D domain with radiatively transparent obstacles (b 0).

Fig. 14. Temperature contour for the entire 2-D domain with radiation absorbing
obstacles (b 2 and N 0:001).

is not affected by the thermal calculations so this model only presents one-way coupling with the uid model. The parameters
Re 100; Pe 1600 and inlet maximum velocity usx 0 0:1
are used for these calculations to maintain similarity with the
homogeneous case in simulating uid ow and the role of convection term in energy transport. The inlet velocity prole was chosen
to be Poiseuille ow. The velocity calculations as obtained from the
solutions of LBE (Eq. (39)) are presented in Fig. 12.
After obtaining the velocity eld and initialized radiation eld,
Eq. (40) is solved using its LBE form (Eq. (35)). The temperature
and radiative elds are mutually coupled, so they were solved iteratively. The resulting temperature proles for different material
properties of obstacles are shown in Figs. 13 and 14. The uid
media in both these cases is assumed to absorb negligible radiation. In the rst case, the solid obstacles are assumed to be
non-absorbers. The results show no temperature change in solids
and the uid near boundaries is hot due to convection near the
walls. Due to very high Peclet number Pe 1600, there is no
change in the uid temperature away from the walls. The second
case is modeled with the absorbing and conducting material. As
uid is almost transparent to radiation, the wall radiation instantly
starts heating up the obstacles and raise its temperature. Therefore
the uid eventually gets heated up from the wall and obstacles
both due to convection and conduction.

4. Conclusions
The LBM based RTE models are extended for their application
into coupled multi-physics problems. The results of this work on

simple examples in multi-mode heat transfer show that LBM has


a potential in solving coupled multi-physics problems. A benchmark problem with conductionradiation heat transfer in slab
geometry was solved using LBE model and results are in good
agreement with the benchmark P1 results. The effect of stark number and wall emissivity on thermal behavior is studied. Grid convergence study was performed which showed that LBE for RTEs
is a rst order accurate method. D2 Q 8 and D2 Q 16 LBE templates
were used to model radiative transport in two-dimensional square
enclosures with constant temperature hc 1 throughout the
domain, to check the validity of these LBE templates for solving
RTE problems. The results were found to be in good agreement
with existing S4 solutions in the literature [10]. Higher angular
resolution template D2 Q 16 was selected to model the radiation
transport in porous radiative burners in two dimensional rectangular geometry. Porous radiative burner is modeled using homogenous approximation and analytical velocity eld of the uid
stream. The results obtained show qualitative effect of the radiation transport in homogeneous participating media. The challenges
of this modeling approach are highlighted and LBM based
single-framework approach for solving multi-physics in a
heterogenous model is presented. The uid transport around the
solid obstacles, conductionconvection modeling in the entire
domain (soliduid) and radiation transport are solved using
LBEs. A fully coupled LBE model is developed and simulation
results are presented for a simplied heterogenous porous media.
The effect of material properties in the entire spatial domain is
studied and results can be physically analyzed. This work involves
only one-way coupling with the uid model i.e. no feedback is provided to the uid model based on thermal calculations. However,
with this example, LBE approach for modeling the phenomenon
of radiative heating of obstacles and in-turn conjugate heating of
uid stream is demonstrated. Due to its simplied single framework model, LBM is well-suited and has clear advantage over other
deterministic/stochastic numerical techniques for modeling novel
mesoscale applications of photo-thermal heating and optical
tweezers where uid motion is directly or indirectly impacted by
radiation heat transport. Although there is signicant effort
required in the area of LBM based solutions to RTEs prior to its
implementation for solving practical problems. One of the important future task is to improve the accuracy of the method i.e.
exploring the LBE schemes with order of accuracy higher than rst
order. Second important aspect is to understand rigorously the
numerical stability envelopes of using LBE as a single framework
for solving multiphysics problems. Finally, extending the LBM for
solving multienergetic RTEs is important to explore its advantages
as compared to existing multienergetic algorithms in solving RTEs.

References
[1] Ghosh S, Friedrich R, Ptzner M, Stemmer Chr, Cuenot B, El Ha M. Effects of
radiative heat transfer on the structure of turbulent supersonic channel ow. J.
Fluid Mech. 2011;677.
[2] Jimnez J, Cuervo D, Aragons JM. A domain decomposition methodology for
pin by pin coupled neutronic and thermalhydraulic analyses in COBAYA3.
Nucl. Eng. Des. 2010;240(2):313.
[3] Studera E, Beccantinia A, Gounanda S, Dabbenea F, Magnauda JP, Paillrea H,
et al. CAST3M/ARCTURUS: a coupled heat transfer CFD code for thermal
hydraulic analyzes of gas cooled reactors. Nucl Eng Des 2007;237(15
17):1814.
[4] Chai JC, Patankar SV. Finite-volume method for radiation heat transfer. Adv
Numer Heat Transfer 2000;2.
[5] Mishra SC, Chugh P, Kumar P, Mitra K. Development and comparison of the
DTM, the DOM and the FVM formulations for the short-pulse laser transport
through a participating medium. Int J Heat Mass Transfer 2006;49.
[6] Carlson BG. Transport theory; discrete ordinates quadrature over the unit
sphere. Report LA-4554, Los Alamos Scientic Laboratory; 1970.
[7] Lathrop KD, Carlson BG. Discrete ordinates angular quadrature of the neutron
transport equation. Report LA-3186, Los Alamos Scientic Laboratory; 1965.

Please cite this article in press as: McCulloch R, Bindra H. Coupled radiative and conjugate heat transfer in participating media using lattice Boltzmann
methods. Comput Fluids (2015), http://dx.doi.org/10.1016/j.compuid.2015.05.024

R. McCulloch, H. Bindra / Computers & Fluids xxx (2015) xxxxxx


[8] Razzaque MM, Klein DE, Howell JR. Finite element solution of radiative heat
transfer in a two-dimensional rectangular with gray participating media. J
Heat Transfer 1983;105.
[9] Fiveland WA. Discrete ordinate solutions of the radiative transport equation
for rectangular enclosures. J Heat Transfer 1984;106.
[10] Truelove JS. Discrete ordinate solution of radiation transport equation. J Heat
Transfer 1987;109.
[11] Jinbo H, Liming R, Heping T. Effect of anisotropic scattering on radiative heat
transfer in two-dimensional rectangular media. J Quant Spectrosc Radiant
Transfer 2003;78.
[12] Briesmeister JF. MCNP: a general Monte Carlo N-particle transport code,
LA-12625-M, 1993.
[13] Mondal B, Mishra SC. Lattice Boltzmann method applied to the solution of the
energy equations of the transient conduction and radiation problems on nonuniform lattices. Int J Heat Mass Transfer 2008;51(12).
[14] Parida PR, Raj R, Prasad A, Mishra SC. Solidication of a semi-transparent
planar layer subjected to radiative and convective cooling. J Quant Spectrosc
Radiant Transfer 2007;107(2).
[15] Moufekkir F, Moussaoui MA, Mezrhab A, Bouzidi M, Lemonnier Denis.
Combined double-diffusive convection and radiation in a square enclosure
lled with semitransparent uid. Comput Fluids 2012;69.
[16] Moufekkir F, Moussaoui MA, Mezrhab A, Naji H, Lemonnier D. Numerical
prediction of heat transfer by natural convection and radiation in an enclosure
lled with an isotropic scattering medium. J Quant Spectrosc Radiant Transfer
2012;113.
[17] Asinari P, Mishra SC, Borchiellini R. A lattice Boltzmann formulation to the
analysis of radiative heat transfer problems in a participating medium. Numer
Heat Transfer B 2010;57.
[18] Di-Rienzo AF, Asinari P, Borchiellini R, Mishra SC. Improved angular
discretization and error analysis of the lattice Boltzmann method for solving
radiative heat transfer in a participating medium. Int J Numer Methods Heat
Fluid Flow 2011;21:640.
[19] Ma Y, Dong S, Tan H. Lattice Boltzmann method for one-dimensional radiative
transfer. Phys Rev E 2011;84:016704.
[20] Bindra H, Patil DV. Radiative or neutron transport modeling using a lattice
Boltzmann equation framework. Phys Rev E 2012;86:016706.
[21] Zhang Y, Yi HL, Tan HP. The lattice Boltzmann method for one-dimensional
transient radiative transfer in graded index gray medium. J Quant Spectrosc
Radiant Transfer 2014;137.

[22] Zhang Y, Yi HL, Tan HP. Short-pulsed laser propagation in a participating slab
with Fresnel surfaces by lattice Boltzmann method. Int J Heat Mass Transfer
2015;80.
[23] Modest MF. Radiative heat transfer. 2nd ed. San Diego (CA): Academic Press;
2003.
[24] Wolf-Gladrow DA. Lattice-gas cellular automata and lattice Boltzmann models
an introduction. Springer; 2005.
[25] Kandhai BD. Large scale lattice-Boltzmann simulations, computational
methods and applications. Thesis, Universiteit van Amsterdam; 1999.
[26] Mishra SC, Roy HK. Solving transient conduction and radiation heat transfer
problems using the lattice Boltzmann method and the nite volume method. J
Comput Phys 2007;223.
[27] Mohamad AA. Lattice Boltzmann method fundamentals and engineering
applications with computer codes. Springer; 2011.
[28] Lewis EE, Miller WF. Computational methods of neutron transport. American
Nuclear Society; 1993.
[29] Talukdar P, Mishra SC, Trimis D, Durst F. Combined radiation and convection
heat transfer in a porous channel bounded by isothermal parallel plates. Int J
Heat Mass Transfer 2004;47.
[30] Kaviany M. Laminar ow through a porous channel bounded by isothermal
parallel plates. Int J Heat Mass Transfer 1985;28.
[31] Mishra SC, Lankadasu A, Beronov K. Application of the lattice Boltzmann
method for solving the energy equation of a 2D transient conduction radiation
problem. Int J Heat Mass Transfer 2005;48.
[32] Huang HB, Liu XY, Sukop MC. Numerical study of lattice Boltzmann methods
for a convectiondiffusion equation coupled with NavierStokes equations. J
Phys A: Math Theor 2011;44:055001.
[33] Chopard B, Falcone JL, Latt J. The lattice Boltzmann advectiondiffusion model
revisited. Eur Phys J Spec Top 2009;171.
[34] White FM. Fluid mechanics. 7th ed. McGraw Hill; 2011.
[35] Bhatnagar PL, Gross EP, Krook M. A model for collision processes in gases. I.
Small amplitude processes in charged and neutral one-component systems.
Phys Rev 1954;94(3):511.
[36] Aidun CK, Clausen JR. Lattice-Boltzmann method for complex ows. Annu Rev
Fluid Mech 2010;42:439.
[37] He X, Luo L-S. Theory of the lattice Boltzmann method: from the Boltzmann
equation to the lattice Boltzmann equation. Phys Rev E 1997;56(6):6811.

Please cite this article in press as: McCulloch R, Bindra H. Coupled radiative and conjugate heat transfer in participating media using lattice Boltzmann
methods. Comput Fluids (2015), http://dx.doi.org/10.1016/j.compuid.2015.05.024