International Journal of Applied Electromagnetics and Mechanics 41 (2013) 121–141 121

DOI 10.3233/JAE-2012-1598
IOS Press

Numerical simulation and stability analysis

on MHD free convective heat and mass
transfer unsteady flow through a porous
medium in a rotating system with induced
magnetic field
M.M. Haquea , M.M. Alama , M. Ferdowsb,∗ and Qasem M. Al-Mdallalc
a Mathematics Discipline, Khulna University, Khulna, Bangladesh
b Department of Mathematics, University of Dhaka, Dhaka, Bangladesh
c Department of Mathematical Sciences, College of Science, United Arab Emirates University, Al-Ain,

United Arab Emirates

Abstract. A numerical study of Soret and Dufour effects on MHD free convective heat and mass transfer unsteady high
speed flow of viscous fluid through a porous medium with induced magnetic field is investigated in a rotating system. In
this observation, heat and mass fluxes from the medium are considered to be constant for cooling problem with lighter and
heavier particles. Numerical values of velocity, induced magnetic field, temperature and concentration are computed by the
conditionally stable explicit finite difference method with stability and convergence analysis. The shear stress, current density,
Nusselt number and Sherwood number are also calculated here. Effects of associated parameters on the above mentioned
quantities are shown graphically and physical aspects of the problem are discussed in detail. Finally, both the qualitative and
quantitative comparisons of present results with previous work are presented in tabular form.

Keywords: MHD, free convective, thermal-mass diffusion, rotating system, induced magnetic field

Nomenclature

B Magnetic field vector

C̄ Species concentration
C Dimensionless species concentration
cs Concentration susceptibility
cp Specific heat at the constant pressure
C∞ Species concentration of uniform flow
Dm Coefficient of mass diffusivity
Df Dufour number

∗
Corresponding author: M. Ferdows, Department of Mathematics, University of Dhaka, Dhaka-1000, Bangladesh. E-mail:
[email protected].

1383-5416/13/$27.50 
c 2013 – IOS Press and the authors. All rights reserved
122 M.M. Haque et al. / Numerical simulation and stability analysis on MHD free convective heat

e Charge of electron
E Electric field intensity
E Ekman number
Ec Eckert number
F Force vector
g Local acceleration due to gravity
Gr Grashof number
Gm Modified Grashof number
H Induced magnetic field vector
H Dimensionless induced magnetic field component
H0 Constant induced magnetic field
Hw Induced magnetic field at the wall
Hx , Hy Induced magnetic field components
J Current density vector
JL , JA Local and average current density respectively
K Permeability of the medium
m Constant mass flux per unit area
M Magnetic force number
ne Number density of electrons
NuL , NuA Local and average Nusselt number respectively
P Fluid pressure
Pe Pressure of electron
Pm Magnetic diffusivity number
Pr Prandtl number
q Velocity vector
Q Constant heat flux per unit area
Sc Schmidt number
ShL , ShA Local and average Sherwood number respectively
So Soret number
t Dimensionless time
T̄ Fluid temperature
T Dimensionless fluid temperature
Tm Mean fluid temperature
T∞ Fluid temperature of uniform flow
U0 Constant velocity
u, v, w Velocity components
U, V , W Dimensionless velocity components
x, y Cartesian coordinates
X, Y Dimensionless cartesian coordinates

Greek symbols

∇ Divergence vector
Δt Time-step
ΔX, ΔY Mesh sizes along X and Y directions respectively
β Thermal expansion coefficient
β∗ Concentration expansion coefficient
ϕ Dissipation function involving the viscous stress
κ Thermal conductivity
K Permeability number
κT Thermal diffusion ratio
μ Coefficient of viscosity
μe Magnetic permeability
ρ Density of the fluid
σ Electrical conductivity
 M.M. Haque et al. / Numerical simulation and stability analysis on MHD free convective heat 123

τ Time
τ XL , τ XA Local and average X component shear stress respectively
τZ L , τZ A Local and average Z component shear stress respectively
υ Kinematic viscosity
Ω Angular velocity vector
Ω Angular velocity component

Subscripts

A For average
L For local
W At the wall
∞ In uniform flow
i, j Grid points along X and Y axis respectively

Superscript

n Number of time-steps

At the end of a time-step

1. Introduction

In the last decade, considerable improvement has been made in the general theory of rotating fluids
because of its application in rocket propulsion control, crystal growth technology, astrophysical plasma
fluid dynamics, tribological regulation in moving machine parts and MHD energy generators (Bég [1]).
The Ekman boundary layers(common feature of rotating flows) of an incompressible fluid have been
investigated as basic boundary layers in a rotating environment appearing in the oceanic, atmospheric,
cosmic fluid dynamics and solar physics or geophysical problems. Greenspan [2] was the first author
to recognize the Ekman boundary layer near a flat plate in a rotating fluid and find out the viscous
and Coriolis forces are of same order of magnitude. The Ekman layer flow on a horizontal plate was
studied by Batchelor [3]. An analytical study on MHD heat generating heat and mass transfer flow
through a porous medium has been analyzed by Ram [4] in a rotating fluid. Thermal diffusion effects
on MHD free convective mass transfer flow past a moving infinite vertical plate have been investigated
analytically in a rotating system by Nanousis [5]. Alam and Sattar [6] have computed the numerical
solutions of transient MHD heat and mass transfer flow in a rotating system with thermal diffusion by
finite difference technique. Recently, a similarity solution of Dufour and Soret effects on steady MHD
free convective mass transfer fluid flow through a porous medium in a rotating system has been obtained
by Islam and Alam [7].
All the above works are investigated under the action of transverse magnetic field for low speed flow
(absence of viscous dissipation and Joule heating terms in energy equation) but effect of induced mag-
netic field on high speed flow is not studied in a rotating environment. Because of the flow under the
action of a strong magnetic field that induced another magnetic field plays a decisive role in geophysical
and astrophysical problems, an analytical study of the effect of induced magnetic field on combined
heat and mass transfer steady one dimensional flow over a vertical plate has been performed by Chaud-
hary and Sharma [8]. Alam et al. [9] has calculated a similarity solution of the same problem in case
of two dimensional flows by shooting iteration technique. Recently, a steady-state solution of heat and
124 M.M. Haque et al. / Numerical simulation and stability analysis on MHD free convective heat

mass transfer flow with thermal diffusion and induced magnetic field has been analyzed by Haque and
Alam [10]. Quit recently, the effects of induced magnetic field on mixed convective transient heat and
mass transfer two dimensional high speed flows with thermal diffusion, constant heat and mass fluxes
have been observed numerically by Haque and Alam [11]. The problem becomes more complicated
when this problem is investigated in a rotating system with mass diffusion.
Our main goal of the present work is to study numerically thermal and mass diffusion effects on MHD
free convective heat and mass transfer unsteady high speed flow through a porous medium in a rotating
system under the action of strong magnetic field taking into account the induced magnetic field with
constant heat and mass fluxes.

2. Mathematical model of flow

MHD free convective combined heat and mass transfer time dependent flow has been considered in
a rotating system for an electrically conducting incompressible viscous fluid past an electrically non-
conducting semi-infinite vertical porous medium with thermal and mass diffusion. The flow is assumed
to be in the x-direction, which is chosen along the plate in upward direction and y -axis is normal to it. A
uniform strong magnetic field is acted normal to the flow direction. The governing equations in vector
form of the present problem are as follows,
Continuity equation ∇.q = 0
∂q 1 υ 1
Momentum equation + (q.∇) q = F − ∇P + υ∇2 q − 2Ω ∧ q − q + (J × H)
∂t ρ K ρ
∂H 1
Magnetic induction equation + (q.∇) H = (H.∇) q + ∇2 H
∂t σμe
∂T κ 2 J2 1 Dm κT 2
Energy equation + (q.∇) T = ∇ T+ 2
+ ϕ+ ∇ C
∂t ρcp ρcp σμe ρcp cs cp
∂C Dm κT 2
Concentration equation + (q.∇) C = Dm ∇2 C + ∇ T
∂t Tm
σ σ
Generalized Ohm’s law J = σ (E + q × B) − (J × B) + ∇Pe
ene ene
Maxwell’s equations ∇ × B = J, ∇ × E = 0, ∇.H = 0.
Initially (τ  0), it is considered that the medium and fluid are at rest and at same temperature
T̄ (= T∞ ) with species concentration level C̄ (= C∞ ) at all points. At time τ > 0, the whole sys-
tem starts rotating about the y -axis with a constant angular velocity Ω. The physical configuration and
coordinate system of the present investigation are shown in Fig. 1.
The problem is based on the following additional assumptions:
I. In accordance with the Boussinesq’s approximation, all the physical properties of fluid are con-
sidered to be constant while the influence of density variation with temperature is assumed only
in the body force term.
II. Since the system rotates with a angular velocity Ω about the y -axis, the constant angular velocity
is of the form Ω = (0, −Ω, 0).
III. The viscous dissipation and joule heating terms in the energy equation have been considered for
high speed flows. Also the levels of concentration of foreign mass have been taken very high for
observing the Soret and Dufour effects on the flow.
 M.M. Haque et al. / Numerical simulation and stability analysis on MHD free convective heat 125

IV. The magnetic Reynolds number of the flow is taken to be large enough so that the induced mag-
netic field is of the form H = (Hx , Hy , 0) is applicable. The divergence equation of Maxwell’s
equation ∇.H = 0 for the magnetic field gives Hy = constant = H0 (say).
Within the framework of the above stated assumptions, the equations relevant to the recent unsteady
two dimensional problem in a rotating system is governed by the following system of non-linear coupled
partial differential equations under the Ekman boundary-layer approximations,
Continuity equation,
∂u ∂v
+ =0 (1)
∂x ∂y
Momentum equations,

∂u ∂u ∂u   ∂2u υ μe ∂Hx
+u +v = gβ T̄ − T∞ + gβ ∗ (C̄ − C∞ ) + υ 2 + 2Ωw − u + H0 (2)
∂τ ∂x ∂y ∂y K ρ ∂y

∂w ∂w ∂w ∂2w υ
+u +v = υ 2 − 2uΩ − w (3)
∂τ ∂x ∂y ∂y K
Magnetic induction equation,

∂Hx ∂Hx ∂Hx ∂u ∂u 1 ∂ 2 Hx

+u +v = Hx + H0 + (4)
∂τ ∂x ∂y ∂x ∂y σμe ∂y 2
Energy equation,
 2  2  2 
∂ T̄ ∂ T̄ ∂ T̄ κ ∂ 2 T̄ 1 ∂Hx υ ∂u ∂w Dm κT ∂ 2 C̄
+u +v = 2
+ + + + (5)
∂τ ∂x ∂y ρcp ∂y ρcp σ ∂y cp ∂y ∂y cs cp ∂y 2

And species concentration equation,

∂ C̄ ∂ C̄ ∂ C̄ ∂ 2 C̄ Dm κT ∂ 2 T̄
+u +v = Dm 2 + . (6)
∂τ ∂x ∂y ∂y Tm ∂y 2
Since the heat and mass fluxes from the porous medium to fluid are constant, the appropriate initial
and boundary conditions are,
τ  0, u = 0, v = 0, w = 0, Hx = 0, T̄ → T∞ , C̄ → C∞ everywhere (7)
τ > 0, u = 0, v = 0, w = 0, Hx = 0, T̄ → T∞ , C̄ → C∞ at x = 0
∂ T̄ Q ∂ C̄ m
u = 0, v = 0, w = 0, Hx = Hw , =− , =− at y = 0 (8)
∂y κ ∂y Dm
u = 0, v = 0, w = 0, Hx = 0, T̄ → T∞ , C̄ → C∞ as y → ∞
Introducing the dimensionless quantities
xU0 yU0 u v w τ U02
X= , Y = , U= , V = , W = , t= ,
υ υ U0 U0 U0 υ
     
μe H x κU0 T̄ − T∞ Dm U0 C̄ − C∞ H w μe
H= , T = , C= , h=
ρ U0 Qυ mυ U0 ρ
126 M.M. Haque et al. / Numerical simulation and stability analysis on MHD free convective heat

Fig. 1. Physical model and coordinate system. Fig. 2. Finite difference grid space.

and parameters

gβQυ 2 gβ ∗ mυ 2 2Ωυ U 2K H0 μe
Gr = , Gm = , E= 2 , K = 02 , M= ,
κU04 Dm U04 U0 υ U0 ρ
υρcp κU03 κmκT υ 2 κ
QDm T
Pm = υσμe , Pr = , Ec = , Df = , Sc = , So = ,
κ Qυcp Qυcs cp Dm mκυTm
the system of Eqs (1)–(6) become in non-dimensional form as,
∂U ∂V
+ =0 (9)
∂X ∂Y
∂U ∂U ∂U ∂2U 1 ∂H
+U +V = Gr T + Gm C + 2
+ EW − U + M (10)
∂t ∂X ∂Y ∂Y K ∂Y
∂W ∂W ∂W ∂2W 1
+U +V = − EU − W (11)
∂t ∂X ∂Y ∂Y 2 K
∂H ∂H ∂H ∂U ∂U 1 ∂2H
+U +V =H +M + (12)
∂t ∂X ∂Y ∂X ∂Y Pm ∂Y 2
      
∂T ∂T ∂T 1 ∂2T Ec ∂H 2 ∂U 2 ∂W 2 ∂2C
+U +V = 2
+ + Ec + + Df (13)
∂t ∂X ∂Y Pr ∂Y Pm ∂Y ∂Y ∂Y ∂Y 2
∂C ∂C ∂C 1 ∂2C ∂2T
+U +V = + S o (14)
∂t ∂X ∂Y Sc ∂Y 2 ∂Y 2
as well as the initial and boundary conditions Eqs (7)–(8) become
t  0, U = 0, V = 0, W = 0, H = 0, T = 0, C=0 everywhere (15)
t > 0, U = 0, V = 0, W = 0, H = 0, T = 0, C = 0 at X = 0
∂T ∂C
U = 0, V = 0, W = 0, H = h = 1 (say) , = −1, = −1 at Y = 0 (16)
∂Y ∂Y
U = 0, V = 0, W = 0, H = 0, T = 0, C=0 as Y → ∞.
 M.M. Haque et al. / Numerical simulation and stability analysis on MHD free convective heat 127

3. Shear stress, current density, Nusselt and Sherwood number

Since the quantities of chief physical interest are shear stress, current density, Nusselt number and
Sherwood number. From the velocity field, the effects of various parameters on shear stress atthe plate
∂U
have been calculated. The local and average X component shear stress are defined by τXL = μ ∂Y
 ∂U   ∂U  100  ∂U  Y =0
and τXA = μ ∂Y Y =0 dX which are proportional to ∂Y Y =0 and 0 ∂Y Y =0  dX respectively.
The local and average Z component shear stress are represented by τZL = μ ∂W and τZA =
 ∂W   ∂W  100  ∂W  ∂Y Y =0
μ ∂Y Y =0 dX which are proportional to ∂Y Y =0 and 0 ∂Y Y =0 dX respectively.
The effects of different parameters on current density have been observed from the  induced mag-
netic field. The local and average current density are defined by JL = μ − ∂H ∂Y Y =0 and JA =
 ∂H   ∂H  100  ∂H 
μ − ∂Y Y =0 dX which are proportional to − ∂Y Y =0 and 0 − ∂Y Y =0 dX respectively.
From the temperature field, the effects of various parameters on Nusselt number  ∂T have been inves-
tigated. The local and average Nusselt number are represented by NuL = μ − ∂Y and NuA =
 ∂T   ∂T  100  ∂T  Y =0
μ − ∂Y Y =0 dX which are proportional to − ∂Y Y =0 and 0 − ∂Y Y =0 dX respectively.
And the effects of different parameters on Sherwood number have been analyzed  ∂C from the concen-
tration field. Local and average Sherwood number are defined by ShL = μ − ∂Y and ShA =
 ∂C   ∂C  100  ∂C  Y =0
μ − ∂Y Y =0 dX which are proportional to − ∂Y Y =0 and 0 − ∂Y Y =0 dX respectively.

4. Numerical solutions

In this section, we want to solve the non-dimensional system of governing Eqs (9)–(14) subject to the
initial and boundary conditions Eqs (15)–(16). Callahan and Marner [12] solved a transient free con-
vective mass transfer flow by the conditionally stable explicit finite difference method. On the contrary,
Soundalgekar and Ganesan [13] have studied the same problem by the unconditionally stable implicit
finite difference method. It is observed from their results that both technique produce same results. Be-
cause of the only difference between two methods is that the implicit method being unconditionally
stable is less expansive from the point of view of computer time, Haque and Alam [11] have computed
the numerical solution of a mixed convective transient heat and mass transfer problem by the explicit
finite difference method. Hence for simplicity the explicit finite difference method is used here to solve
the present problem.
To obtain a set of finite difference equations, a rectangular region of the flow field is chosen and the
region is divided into a grid of lines parallel to X and Y axis, where X -axis is taken along the plate
and Y -axis is normal to it. It is considered that the plate of height Xmax (= 100) i.e. X varies from 0
to 100 with Ymax (= 25) as corresponds to Y → ∞ i.e. Y varies from 0 to 25. There are m (= 145)
and n (= 145) grid spacing in X and Y axis respectively as shown in Fig. 2. We also assume that
ΔX = 0.69 (0  X  100) and ΔY = 0.17 (0  Y  25) are constant mesh sizes along X and Y
directions respectively with the smaller time-step, Δt = 0.004.
Let U  , V  , W  , H  , T  and C  denote the values of U , V , W , H , T and C at the end of a time-
step respectively. From the basis of the explicit finite difference approximation, we obtain the following
appropriate set of finite difference equations,
 − U
Ui,j i−1,j Vi,j − Vi,j−1
+ =0 (17)
ΔX ΔY
128 M.M. Haque et al. / Numerical simulation and stability analysis on MHD free convective heat

 −U
Ui,j i,j Ui,j − Ui−1,j Ui,j+1 − Ui,j  
+ Ui,j + Vi,j = Gr Ti,j + Gm Ci,j + RWi,j − KUi,j
Δt ΔX ΔY (18)
Ui,j+1 − 2Ui,j + Ui,j−1 Hi,j+1 − Hi,j
+ + M
(ΔY )2 ΔY
 −W
Wi,j i,j Wi,j − Wi−1,j Wi,j+1 − Wi,j
+ Ui,j + Vi,j
Δt ΔX ΔY (19)
Wi,j+1 − 2Wi,j + Wi,j−1
= − RUi,j − KWi,j
(ΔY )2
 −H
Hi,j i,j Hi,j − Hi−1,j Hi,j+1 − Hi,j
+ Ui,j + Vi,j
Δt ΔX ΔY (20)
Ui,j − Ui−1,j Ui,j+1 − Ui,j 1 Hi,j+1 − 2Hi,j + Hi,j−1
= Hi,j +M +
ΔX ΔY Pm (ΔY )2
 −T
Ti,j i,j Ti,j − Ti−1,j Ti,j+1 − Ti,j
+ Ui,j + Vi,j
Δt ΔX ΔY
 
1 Ti,j+1 − 2Ti,j + Ti,j−1 Ec Hi,j+1 − Hi,j 2
= + (21)
Pr (ΔY )2 Pm ΔY
    
Ui,j+1 − Ui,j 2 Wi,j+1 − Wi,j 2 Ci,j+1 − 2Ci,j + Ci,j−1
+Ec + + Df
ΔY ΔY (ΔY )2
 −C
Ci,j i,j Ci,j − Ci−1,j Ci,j+1 − Ci,j
+ Ui,j + Vi,j
Δt ΔX ΔY (22)
1 Ci,j+1 − 2Ci,j + Ci,j−1 Ti,j+1 − 2Ti,j + Ti,j−1
= 2 + So
Sc (ΔY ) (ΔY )2
with the initial and boundary conditions
0 0 0 0 0 0
Ui,j = 0, Vi,j = 0, Wi,j = 0, Hi,j = 0, Ti,j = 0, Ci,j =0 (23)
n n n n n n
U0,j = 0, V0,j = 0, W0,j = 0, H0,j = 0, T0,j = 0, C0,j =0
n n n n n n n n
Ui,0 = 0, Vi,0 = 0, Wi,0 = 0, Hi,0 = 1, Ti,0 = Ti,1 + ΔY, Ci,0 = Ci,1 + ΔY (24)
n n n n n n
Ui,L = 0, Vi,L = 0, Wi,L = 0, Hi,L = 0, Ti,L = 0, Ci,L =0 where L → ∞.
Here the subscripts i and j designate the grid points along X and Y axis respectively and the super-
script n represents a value of time, t = nΔt where n = 0, 1, 2, 3, 4, . . . . . ..
At the end of a time-step Δt, the new temperature (T  ), concentration (C  ), primary velocity (U  ),
secondary velocity (W  ), induced magnetic field (H  ) and V  at all interior nodal points may be ob-
tained by successive applications of Eqs (21), (22), (18), (19), (20) and (17) respectively. This process
is repeated in time and provided that the time-step is sufficiently small, U , V , W , H , T and C should
eventually converge to values which approximate the steady-state solutions of Eqs (9)–(14). These con-
verged solutions are shown graphically in Figs 3–20. The numerical values of local and average shear
stress, current density, Nusselt number, Sherwood number are evaluated by Five-point differential for-
mula and Simpson’s 13 integration formula respectively. The obtained values are displayed graphically
in Figs 21–36.
 M.M. Haque et al. / Numerical simulation and stability analysis on MHD free convective heat 129

5. Stability and convergence analysis

Since an explicit finite difference technique is used, the study will remain incomplete unless we analy-
sis the stability and convergence of the method. For the constant mesh sizes the stability conditions of the
technique are established in this section. The Eq. (17) will be ignored since Δt does not appear in it. At
a time t = 0√, the general terms of the Fourier expansion for U , W , H , T and C are all eiαX eiβY ,
where i = −1. At a time t, these terms become U : ψ (t) eiαX eiβY , W : ϕ (t) eiαX eiβY , H :
φ (t) eiαX eiβY , T : θ (t) eiαX eiβY and C : ξ (t) eiαX eiβY and after a time-step these terms will
become U : ψ  (t) eiαX eiβY , W : ϕ (t) eiαX eiβY , H : φ (t) eiαX eiβY , T : θ  (t) eiαX eiβY and
C : ξ  (t) eiαX eiβY . Substituting the above relations into Eqs (11)–(22), we obtain the following equa-
tions upon simplification,

ψ  = A3 ψ + +A4 ϕ + A5 φ + A6 θ + A7 ξ (25)

ϕ = −A8 ψ + A1 ϕ (26)

φ = B1 ψ + B2 φ (27)

θ  = B3 ψ + B4 ϕ + B5 φ + +B6 θ + B7 ξ (28)
ξ  = B8 θ + B9 ξ (29)
where
Δt   Δt
iβΔY  2Δt
A1 = 1 − U 1 − e−iαΔX − V e −1 + (cos βΔY − 1) − KΔt,
ΔX ΔY (ΔY )2
Δt
iβΔY 
A2 =M e − 1 , A3 = A1 + B3 Gr Δt, A4 = RΔt + B4 Gr Δt,
ΔY
A5 = A2 + B5 Gr Δt, A6 = B7 Gm Δt + B6 Gr Δt, A7 = B8 Gm Δt, A8 = RΔt,
Δt   Δt
iβΔY 
B1 = H̄x 1 − e−iαΔX + M e −1 ,
ΔX ΔY
Δt   Δt
iβΔY  1 2Δt
B2 = 1−U 1 − e−iαΔX − V e −1 + (cos βΔY − 1) ,
ΔX ΔY Pm (ΔY )2
U Ec Δt
iβΔY 2 Δt
iβΔY 2
B3 = e − 1 , B 4 = W Ec e − 1 ,
(ΔY )2 (ΔY )2
Ec Δt
iβΔY 2 2Δt
B5 = H̄x 2 e − 1 , B7 = Df 2 (cos βΔY − 1) ,
Pm (ΔY ) (ΔY )
Δt   Δt
iβΔY  1 2Δt
B6 = 1−U 1 − e−iαΔX − V e −1 + (cos βΔY − 1) ,
ΔX ΔY Pr (ΔY )2
2So Δt
B8 = (cos βΔY − 1)
(ΔY )2
and
Δt   Δt
iβΔY  1 2Δt
B9 = 1 − U 1 − e−iαΔX − V e −1 + (cos βΔY − 1) .
ΔX ΔY Sc (ΔY )2
130 M.M. Haque et al. / Numerical simulation and stability analysis on MHD free convective heat

Fig. 3. Primary Velocity Profiles for Gm = 1.0, Gr = 2.0, Fig. 4. Secondary Velocity Profiles for Gm = 1.0, Gr = 2.0,
E = 0.2, M = 0.5, Df = 0.1, Pr = 0.71, Sc = 0.3, E = 0.2, M = 0.5, Df = 0.1, Pr = 0.71, Sc = 0.3,
K = 0.5, Pm = 1.0 and Ec = 0.01. K = 0.5, Pm = 1.0 and Ec = 0.01.

Fig. 5. Primary Velocity Profiles for Gm = 1.0, Gr = 2.0, Fig. 6. Secondary Velocity Profiles for Gm = 1.0, Gr = 2.0,
So = 1.0, M = 0.5, Df = 0.1, Pr = 0.71, Sc = 0.3, So = 1.0, M = 0.5, Df = 0.1, Pr = 0.71, Sc = 0.3,
K = 0.5, Pm = 1.0 and Ec = 0.01. K = 0.5, Pm = 1.0 and Ec = 0.01.

Fig. 7. Primary Velocity Profiles for Gm = 1.0, Gr = 2.0, Fig. 8. Secondary Velocity Profiles for Gm = 1.0, Gr = 2.0,
So = 1.0, M = 0.5, Df = 0.1, E = 0.2, Sc = 0.3, So = 1.0, M = 0.5, Df = 0.1, E = 0.2, Sc = 0.3,
K = 0.5, Pm = 1.0 and Ec = 0.01. K = 0.5, Pm = 1.0 and Ec = 0.01.
 M.M. Haque et al. / Numerical simulation and stability analysis on MHD free convective heat 131

Fig. 9. Primary Velocity Profiles for Gm = 1.0, Gr = 2.0, Fig. 10. Secondary Velocity Profiles for Gm = 1.0, Gr = 2.0,
E = 0.2, M = 0.5, Df = 0.1, Pr = 0.71, Sc = 0.3, E = 0.2, M = 0.5, Df = 0.1, Pr = 0.71, Sc = 0.3,
K = 0.5, Pm = 1.0 and So = 1.0. K = 0.5, Pm = 1.0 and So = 1.0.

Fig. 11. Primary Velocity Profiles for Gm = 1.0, Gr = 2.0, Fig. 12. Secondary Velocity Profiles for Gm = 1.0, Gr = 2.0,
E = 0.2, M = 0.5, Df = 0.1, Pr = 0.71, Sc = 1.0, E = 0.2, M = 0.5, Df = 0.1, Pr = 0.71, So = 1.0,
K = 0.5, Pm = 1.0 and So = 0.01. K = 0.5, Pm = 1.0 and Ec = 0.01.

Fig. 13. Primary Velocity Profiles for Gm = 1.0, Gr = 2.0, Fig. 14. Secondary Velocity Profiles for Gm = 1.0, Gr = 2.0,
So = 1.0, M = 0.5, Df = 0.1, Pr = 0.71, Sc = 0.3, So = 1.0, M = 0.5, Df = 0.1, Pr = 0.71, Sc = 0.3,
E = 0.2, Pm = 1.0 and Ec = 0.01. E = 0.2, Pm = 1.0 and Ec = 0.01.
132 M.M. Haque et al. / Numerical simulation and stability analysis on MHD free convective heat

Fig. 15. Induced Magnetic Field for Gm = 1.0, Gr = 2.0, Fig. 16. Induced Magnetic Field for Gm = 1.0, Gr = 2.0,
So = 1.0, K = 0.5, Df = 0.1, Pr = 0.71, Sc = 0.3, So = 1.0, M = 0.5, Df = 0.1, Pr = 0.71, Sc = 0.3,
E = 0.2, Pm = 1.0 and Ec = 0.01. E = 0.2, Pm = 1.0 and Ec = 0.01.

Fig. 17. Induced Magnetic Field for Gm = 1.0, Gr = 2.0, Fig. 18. Induced Magnetic Field for Gm = 1.0, Pm = 1.0,
So = 1.0, M = 0.5, Df = 0.1, E = 0.2, Sc = 0.3, So = 1.0, M = 0.5, Df = 0.1, E = 0.2, Sc = 0.3,
K = 0.5, Pr = 0.71 and Ec = 0.01. K = 0.5, Pr = 0.71 and Ec = 0.01.

Fig. 19. Induced Magnetic Field for Gm = 1.0, Gr = 2.0, Fig. 20. Induced Magnetic Field for Gm = 1.0, Gr = 2.0,
So = 1.0, M = 0.5, Df = 0.1, E = 0.2, Sc = 0.3, So = 1.0, M = 0.5, Df = 0.1, E = 0.2, Sc = 0.3,
K = 0.5, Pm = 1.0 and Ec = 0.01. Pr = 0.71, Pm = 1.0 and Ec = 0.01.
 M.M. Haque et al. / Numerical simulation and stability analysis on MHD free convective heat 133

The Eqs (25)–(29) can be expressed in matrix notation as, η  = Γη

⎡ ⎤ ⎡ ⎤ ⎡ ⎤
ψ A3 A4 A5 A6 A7 ψ
⎢ ϕ ⎥ ⎢ −A8 A1 0 0 0 ⎥ ⎢ϕ⎥
⎢ ⎥ ⎢ ⎥ ⎢ ⎥
where η  = ⎢ ⎥ ⎢
⎢ φ ⎥, Γ = ⎢ B1 0 B2 0 0 ⎥ and η = ⎢ φ ⎥.
⎥ ⎢ ⎥
⎣θ ⎦ ⎣ B3 B4 B5 B6 B7 ⎦ ⎣θ⎦
ξ 0 0 0 B8 B9 ξ
For the stability condition we have to find out eigenvalues of the amplification matrix (Γ) but this
study is very difficult because it is a fifth order square matrix and all elements of Γ are different. Hence
the problem requires that the Dufour number (Df ) and Eckert Number (Ec ) are assumed to be very
small that is tends to zero. Under this consideration we have A3 = A1 , A4 = A8 , A5 = A2 and
B3 = B4 = B5 = B7 = 0. Therefore the amplification matrix becomes,
⎡ ⎤
A1 A8 A2 A6 A7
⎢ −A8 A1 0 0 0 ⎥
⎢ ⎥
Γ=⎢
⎢ B1 0 B2 0 0 ⎥ .
⎥
⎣ 0 0 0 B6 0 ⎦
0 0 0 B8 B9

After simplification of the above matrix, we get two eigenvalues, λ1 = B6 , λ2 = B9 and other
three eigenvalues are presented in the equation (λ − B2 ) [(λ − A1 )2 + A28 ] − A2 B1 (λ − A1 ) = 0. For
stability, each eigenvalue λ1 and λ2 must not exceed unity in modulus. Hence the stability conditions
are |B6 |  1 and |B9 |  1 for all α, β .
Now we assume that U is everywhere non-negative and V is everywhere non-positive. Thus B6 =
1 − a − b − P2cr + ae−iαΔX + beiβΔY + P2cr cos βΔY, where a = UΔX Δt
, b = |VΔY |Δt Δt
andc = (ΔY )2
. We
can demonstrated that the maximum modulus of B6 occurs when αΔX = mπ and βΔY = nπ , where
m, n are integers and
hence B6 is real.The value of |B6 | is greater

when both  m and n are odd integers,
2c 2c 2c
in which case B6 = 1 − a − b − Pr − a − b − Pr = 1 − 2 a + b + Pr .
To satisfy the first condition |B6 |  1,
the most negative
 allowable value is B6 = −1.
2c
Hence the first stability condition is 2 a + b + Pr  2 i.e.

Δt Δt 2 Δt
U + |V | + 1 (30)
ΔX ΔY Pr (ΔY )2
Likewise, the second condition |B9 |  1, implies that
Δt Δt 2 Δt
U + |V | + 1 (31)
ΔX ΔY Sc (ΔY )2
Using ΔY = 0.17, Δt = 0.004 and the initial condition, the Eqs (30)–(31) gives Pr  0.28 and
Sc  0.28 respectively. Therefore the stability conditions of the problem are

Δt Δt 2 Δt Δt Δt 2 Δt
U + |V | + 1 and U + |V | + 1
ΔX ΔY Pr (ΔY )2 ΔX ΔY Sc (ΔY )2

as well as the convergence criteria of the problem are Df < 1, Ec << 1, Pr  0.28 and Sc  0.28.
134 M.M. Haque et al. / Numerical simulation and stability analysis on MHD free convective heat

Fig. 21. Temperature Profiles for Gm = 1.0, Gr = 2.0, Fig. 22. Temperature Profiles for Gm = 1.0, Gr = 2.0,
So = 1.0, M = 0.5, Df = 0.1, K = 0.5, Sc = 0.3, So = 1.0, M = 0.5, Df = 0.1, K = 0.5, Sc = 0.3,
Pr = 0.71, Pm = 1.0 and Ec = 0.01. E = 0.2, Pm = 1.0 and Ec = 0.01.

Fig. 23. Temperature Profiles for Gm = 1.0, Gr = 2.0, Fig. 24. Concentration Profiles for Gm = 1.0, Gr = 2.0,
So = 1.0, M = 0.5, Df = 0.1, K = 0.5, Sc = 0.3, Pr = 0.71, M = 0.5, Df = 0.1, K = 0.5, Sc = 0.3,
E = 0.2, Pm = 1.0 and Pr = 0.71. E = 0.2, Pm = 1.0 and Ec = 0.01.

6. Discussion of the results

To obtain the physical insight into the problem under study, the velocity field, induced magnetic field,
temperature field, concentration field and shear stress, current density, Nusselt number, Sherwood num-
ber at the medium are numerically computed by assigning numerical values to the parameters encoun-
tered into the corresponding dimensionless equations with the help of a computer programming language
Compaq Visual Fortran 6.6a. In order to get the steady-state solutions, the computations have been car-
ried out upto t = 80. It is observed that the numerical values of U , W , T and C however, show little
changes after t = 20 but H , show little changes after t = 40. Hence the solutions of all variables for
t = 80 are essentially steady-state. Here the velocity, temperature and concentration distributions are
drawn for t = 10 and 80 as well as the induced magnetic field are drawn for t = 10, 20 and 80.
In this study the Grashof number (Gr = 1.5, 2.0, 2.5) for heat transfer and the modified Grashof num-
ber (Gm = 1.0) for mass transfer are taken to be positive, the values Gr > 0 with Gm > 0 correspond
to cooling to the medium. The cooling problem is often encountered in engineering applications as the
cooling of electronic components and nuclear reactors. The values of Prandtl number (Pr ) are chosen for
 M.M. Haque et al. / Numerical simulation and stability analysis on MHD free convective heat 135

Fig. 25. Concentration Profiles for Gm = 1.0, Gr = 2.0, Fig. 26. Concentration Profiles for Gm = 1.0, Gr = 2.0,
So = 1.0, M = 0.5, Df = 0.1, K = 0.5, Sc = 0.3, So = 1.0, M = 0.5, Df = 0.1, K = 0.5, Pr = 0.71,
E = 0.2, Pm = 1.0 and Ec = 0.01. E = 0.2, Pm = 1.0 and Ec = 0.01.

Fig. 27. Local X-component Shear Stress. Fig. 28. Average X-component Shear Stres.

atmospheric air (Pr = 0.71) , salt water (Pr = 1.0) and water (Pr = 7.0) at 20◦ C temperature. The val-
ues of Schmidt number (Sc ) are liked for a lighter particle helium (Sc = 0.3) , water-vapor (Sc = 0.6)
and a heavier particle methanol (Sc = 0.97) at 25◦ C temperature and one atmospheric pressure. In the
calculations the values of other associated parameters are chosen arbitrarily.
Now we attempt to show the effects of various parameters on flow variables in the following figures
with the discussion of physical aspects of the cooling problem. The velocities, induced magnetic field,
temperature and concentration versus co-ordinate variable (Y ) are illustrated in Figs 3–26 with the help
of data visualization software TECPLOT 10.0.
The primary velocity (U ) and secondary velocity (W ) for different values of Soret number (So ),
Ekman number (E), Prandtl number (Pr ), Eckert number (Ec ), Schmidt number (Sc ) and permeability
number (K) are displayed in Figs 3–14. From these figures, it is observed that U increases while W
decreases with the increase of So or Ec . A reverse effect on U and W is also found for increasing the
values of Pr , Sc or K . We further see that both velocities are decreasingly affected by Ekman number. It
136 M.M. Haque et al. / Numerical simulation and stability analysis on MHD free convective heat

Table 1
Qualitative comparison of present results with previous results of similarity solution
Increased Previous results of similarity solution Present results of numerical solution
parameter U W T C U W T C
So Increase Decrease No Effect Increase Increase Decrease No Effect Increase
Df Increase Decrease Increase No Effect Increase Decrease Increase No Effect
E Decrease Decrease No Effect No Effect Decrease Decrease Increase Increase
Pr Decrease Increase Decrease Increase Decrease Increase Decrease Increase
Sc Decrease Increase No Effect Decrease Decrease Increase Increase Decrease
K Decrease Increase No Effect No Effect Decrease Increase Increase Increase

Fig. 29. Local X-component Shear Stress. Fig. 30. Average X-component Shear Stress.

Fig. 31. Local Z-component Shear Stress. Fig. 32. Average Z-component Shear Stress.
 M.M. Haque et al. / Numerical simulation and stability analysis on MHD free convective heat 137

Table 2
Quantitative comparison of the steady-state present results with previous results
Numerical results due to
Present results
Co-ordinate Haque and Alam [11]
(145 × 145 grid, Δt = 0.004)
variable (Y) (125 × 125 grid, Δt = 0.005)
U H T C U H T C
0.0 1.0 1.0 6.39210 4.86295 1.0 1.0 6.57086 4.79438
5.0 10.9529 –1.55665 0.93423 1.56388 10.9149 –1.56155 0.94056 1.56196
10.0 1.41383 –0.32396 0.07403 0.23999 1.37507 –0.31601 0.07262 0.23699
15.0 0.12336 –0.03186 0.00401 0.02645 0.11701 –0.03026 0.00382 0.02560
20.0 0.00854 –0.00241 0.00019 0.00227 0.00792 –0.00224 0.00018 0.00215
25.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

Fig. 33. Local Z-component Shear Stress. Fig. 34. Average Z-component Shear Stress.

Fig. 35. Local Current Density. Fig. 36. Average Current Density.
138 M.M. Haque et al. / Numerical simulation and stability analysis on MHD free convective heat

Fig. 37. Local Current Density. Fig. 38. Average Current Density.

Fig. 39. Local Nusselt Number. Fig. 40. Average Nusselt Number.

is particularly declared that U is higher while W is smaller for air and helium than water and methanol
respectively. Specially, we have observed that the velocities increase or decrease gradually near the
medium but slowly far away from that medium. It is concluded that the maximum or minimum velocities
occur in the vicinity of the medium.
The effects of magnetic force number (M ), Grashof number (Gr ), magnetic diffusivity number (Pm ),
E , Pr and K on the induced magnetic field (H) are presented in Figs 15–20. The Fig. 15 depicts that H
increases slowly very near the medium and then decreases gradually some distance from that medium
in case of strong M . It is also evident that H rises slowly very near the medium then decreases rapidly
some distance from the medium and again rises far away from that medium with the increase of Gr or
Pm . Further, a reverse effect on H are found for increasing the values of E , Pr or K . Particularly, it is
noted that H is greater for water than air.
 M.M. Haque et al. / Numerical simulation and stability analysis on MHD free convective heat 139

Fig. 41. Local Nusselt Number. Fig. 42. Average Nusselt Number.

Fig. 43. Local Sherwood Number. Fig. 44. Average Sherwood Number.

The temperature (T ) distributions have been shown in Figs 21–23. We observe from these figures
that T is increasingly affected by Df , Gr or Ec but decreasingly affected by Pr . It is also seen a
interesting effect that T decreases for small time values while rises for large time values in case of
strong E or Sc . The temperature is particularly more for air and methanol than water and helium
respectively.
The concentration (C) profiles are graphically represented by Figs 24–26. From these figures, it is
evident that C increases with the rise of So or Pr and decreases in case of strong Ec or Sc . We are further
found that C is same for small time values but increases for large time values with the increase of E or
K . It is particularly noticed that C is greater for water and helium than air and methanol respectively.
Physically, the increase of Sc means decrease of molecular diffusivity, that results in a decrease of
concentration boundary layer. Hence C is higher for small values of Sc and lower for large values of
Sc .
140 M.M. Haque et al. / Numerical simulation and stability analysis on MHD free convective heat

The computed numerical values of steady-state local and average shear stress (τ ), current density
(J), Nusselt number (Nu ) and Sherwood number (Sh ) versus co-ordinate variable (X) and time (t)
respectively are also plotted in Figs 27–44. The X and Z component shear stress (τX & τZ ) for dif-
ferent values of So , Df , E , Pr , Sc and K are presented in Figs 27–34. It is observed from these
figures that both local and average τX increases while τZ decreases with the rise of So or Df . A re-
verse effect on both local and average τX and τZ is also found for increasing the value of Pr , Sc
or K . We further see that both local and average τX and τZ are decreasingly affected by the Ekman
number.
The effects of M , Pm , E , Pr , Gr and K on the current density are shown in Figs 35–38. From these
figures, we are observed that both local and average current density increases in case of strong E , Pr or
K while decreases with the rise of M , Pm or Gr .
The curves of Nusselt number for different values of Df , E , Pr , Ec , Gr and Sc are drawn in Figs 39–
42. A decreasing effect on both local and average Nusselt number are found in case of strong Df , E , Pr ,
Ec or Gr also the Nusselt number increases with the rise of Sc .
The profiles of Sherwood number are displayed in Figs 43–44 for the different values of So , Sc and
K . These figures show that both local and average Sherwood number are increasingly affected by So or
Sc while decreasingly affected by K .
Lastly two comparisons of the present steady-state results are tabulated with previous works. A qual-
itative comparison of our results is presented in Table 1 with the results of the similarity solutions of
free convective mass transfer problem in a rotating system that calculated by Islam and Alam [7]. The
present problem for considering the medium is not subjected to constant heat and mass fluxes, neglecting
the viscous dissipation, joule heating terms in energy equation and absence of induced magnetic field
reduces to the problem given by Islam and Alam [7].
A quantitative comparison of the recent results is also shown in Table 2 with the finite difference
results of mixed convective heat and mass transfer problem with induced magnetic field that computed
by Haque and Alam [11]. If no mass diffusion and rotation effects on the flow past a moving plate are
considered in our problem, it reduces to the problem considered by Haque and Alam [11]. It is concluded
that the accuracy of the present result may be fair in case of all flow variables.

7. Conclusions

In this work thermal and mass diffusion effects on MHD free convective heat and mass transfer un-
steady high speed flow through a porous medium in a rotating system with induced magnetic field is
investigated. The resulting governing equations are solved by explicit finite difference method. The ob-
tained results are graphically presented for the variations of associated parameters. Some of the important
findings of the results are given below,
1. The primary velocity increases with the rise of So or Ec while it decreases with the increase of E ,
Pr , Sc or K .
2. The secondary velocity is increasingly affected by Pr , Sc or K while it decreasingly affected by
So , E or Ec .
3. The induced magnetic field decreases with the increase of M , Gr or Pm but increases with the
rise of E , Pr or K .
4. The temperature profiles are increasingly affected by E or Ec while it decreasingly affected by
Pr .
 M.M. Haque et al. / Numerical simulation and stability analysis on MHD free convective heat 141

5. The concentration increases with the rise of So or Pr while it decreases with the increase of Sc .
6. The X -component shear stress increases with the rise of So or Df while it decreases with the
increase of E , Pr , Sc or K .
7. The Z -component shear stress is increasingly affected by Pr , Sc or K while it decreasingly af-
fected by So , Df or E .
8. The current density decreases with the rise of M , Pm or Gr but it increases with the rise of E , Pr
or K .
9. The Nusselt number is increasingly affected by Sc while decreasingly affected by Df , E , Pr , Ec
or Gr .
10. The Sherwood number increases with the rise of So or Sc while it decreases with the increase of
K.
It is hoped that the findings of the present work may be useful for the study of astrophysical plasma
fluid dynamics, cosmic fluid dynamics and solar physics. The results of the problem are also of great
interest in rocket propulsion control, crystal growth technology, energy generators, astrophysical and
geophysical problems.

References

[1] O.A. Bég, Applications of MHD in supersonic aerodynamics, Technical Report, MFD-08-2, BAE Systems, Aero-
sciences, Academic Wing, KFAA Riyadh KSA July (2007), 144.
[2] H.P. Greenspan, The theory of rotating fluids, Cambridge University Press Cambridge, 1968.
[3] G.K. Batchelor, An introduction to fluid dynamics, Cambridge University Press Cambridge, 1970.
[4] P.C. Ram, Heat and mass transfer on MHD heat generating flow through a porous medium in a rotating fluid, Astrophysics
and Space Science 172 (1990), 273–277.
[5] N. Nanousis, Thermal-diffusion effects on MHD free-convective and mass-transfer flow past a moving infinite vertical
plate in a rotating fluid, Astrophysics and Space Science 191 (1992), 313–322.
[6] M.M. Alam and M.A. Sattar, Transient MHD heat and mass transfer flow with thermal diffusion in a rotating system,
Journal of Energy, Heat and Mass Transfer 21 (1999), 9–21.
[7] N. Islam and M.M. Alam, Dufour and soret effects on steady MHD free convection and mass transfer fluid flow through
a porous medium in a rotating system, Journal of Naval Architecture and Marine Engineering 4 (2007), 43–55.
[8] R.C. Chaudhary and B.K. Sharma, Combined heat and mass transfer by laminar mixed convection flow from a vertical
surface with induced magnetic field, Journal of Applied Physics 99 (2006), 034901–10.
[9] M.M. Alam, M.R. Islam and F. Rahman, Steady heat and mass transfer by mixed convection flow from a vertical porous
plate with induced magnetic field, constant heat and mass fluxes, Thammasat International Journal of Science and
Technology 13(4) (2008), 1–13.
[10] M.M. Haque and M.M. Alam, Steady-state solutions of heat and mass transfer flow from a vertical porous plate with
thermal diffusion and induced magnetic field, Proceedings of the Eighth International Conference on Mechanical Engi-
neering Bangladesh December (2009), 26–28, ICME09-TH-23.
[11] M.M. Haque and M.M. Alam, Transient heat and mass transfer by mixed convection flow from a vertical porous plate
with induced magnetic field, constant heat and mass fluxes, Paper (09) 525(2B), AMSE Journal France October, 2009.
[12] G.D. Callahan and W.J. Marner, Transient free convection with mass transfer on an isothermal vertical flat plate, Inter-
national Journal of Heat and Mass Transfer 19(2) (1976), 165–174.
[13] V.M. Soundalgekar and P. Ganesan, Finite difference analysis of transient free convection on an isothermal flat plate,
Reg Journal of Energy Heat and Mass Transfer 3 (1981), 219–224.