Numerical simulation on laminar free convection flow and heat transfer over an isothermal vertical plate

IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308
_______________________________________________________________________________________
Volume: 04 Issue: 04 | Apr-2015, Available @ http://www.ijret.org 488
NUMERICAL SIMULATION ON LAMINAR FREE-CONVECTION
FLOW AND HEAT TRANSFER OVER AN ISOTHERMAL VERTICAL
PLATE
Asish Mitra1
1
Associate Prof. & HOD-BSH Dept, College of Engineering & Management, Kolaghat. East Midnapur, West Bengal.
India
Abstract
In the present numerical study, laminar free-convection flow and heat transfer over an isothermal vertical plate is presented. By
means of similarity transformation, the original nonlinear coupled partial differential equations of flow are transformed to a pair
of simultaneous nonlinear ordinary differential equations. Then, they are reduced to first order system. Finally, NewtonRaphson
method and adaptive Runge-Kutta method are used for their integration. The computer codes are developed for this numerical
analysis in Matlab environment. Velocity and temperature profiles for various Prandtl number are illustrated graphically. Flow
and heat transfer parameters are derived as functions of Prandtl number alone. The results of the present simulation are then
compared with experimental data published in literature and find a good agreement.
Keywords: Free Convection, Heat Transfer, Matlab, Numerical Simulation, Vertical Plate.
--------------------------------------------------------------------***------------------------------------------------------------------
List of Symbols
F function defined in eq (6)
g gravitational acceleration, 9.81 m/s2
G function defined in eq (9)
Grx Grashof number based on x, dimensionless
h heat transfer coefficient, W/m2
.K
k thermal conductivity, W/m.K
Nu Nusselt number, dimensionless
Pr Prandtl number, dimensionless
T temperature, K
Ts surface temperature, K
T free streams temperature, K
u velocity component in x, m/s
v velocity component in y, m/s
x coordinate from the leading edge, m
y coordinate normal to plate, m
U dimensionless velocity component in x
V dimensionless velocity component in y
z1, z2, z3, z4, z5 variables, eq (13)
Greek Symbols
 coefficient of thermal expansion, 1/K
 thermal diffusivity, m2
/s
μ dynamic viscosity, N.s/m2
 kinematic viscosity, m2
/s
η similarity variables, eq (5)
τ shear stress, N/m2
ψ stream function, m2
/s
Subscript
s plate surface
 free stream
1. INTRODUCTION
The free convection problem on a vertical plate has been
studied in various ways: Pohlhausen [1], Schmidt and
Beckmann [2], Ostrach [3], LeFevre [4], Finston [5], Finston
[5], Sparrow and Gregg [6], Foote [7], Kuiken [8], Kao [9],
and Elbashbeshy [10], The problem is also discussed in
several text books [11-15].
In the present numerical investigation, a simple accurate
numerical simulation of laminar free-convection flow and
heat transfer over an isothermal vertical plate is developed.
The paper is organized as follows: Mathematical model of
the problem, its solution procedure, development of code in
Matlab, interpretation of the results, comparison with
experimental data.
2. MATHEMATICAL MODEL
We consider a vertical hot flat plate immersed in a quiescent
fluid body. We assume the natural convection flow to be
steady, laminar, two-dimensional, no dissipation, and the
fluid to be Newtonian with constant properties, including
density, with one exception: the density difference
  
is to be considered since it is this density difference between
the inside and the outside of the boundary layer that gives
rise to buoyancy force and sustains flow. (This is known as
the Boussinesq approximation.) We take the upward
direction along the plate to be x, and the direction normal to
surface to be y, as shown in Figure 1.

_______________________________________________________________________________________
Fig 1. Physical Model and its coordinate system
The equations governing the flow are




u
x
v
y
  0
1
u
u
x
v
u
y
u
y
g T T







    
2
2
( )
2
u
T
x
v
T
y
T
y







 
2
2
3
The boundary conditions on the solution are:
At y=0: u=v=0, T=Ts
For large y: u→0, T→T∞ 4
A similarity solution to this problem has been obtained by
Ostrach [2]. The solution involves transforming variables by
introducing a similarity parameter of the form
 
y
x
Grx
[ ]
4
1
4
5
and representing the velocity components in terms of a
stream function defined as
  ( , ) ( )[ ( ) ]x y F
Grx
 4
4
1
4
6
where Grx is the Grashof number based on x, i.e.,
Gr
g T T x
x
s

 

( ) 3
2
7
With the foregoing definition of the stream function, the
velocity components may be expressed as
u
y y
Gr
F
x
Grx x
  






 4
4
1
4
1
4
1
4
( ) '( ) ( )

2 1
2


x
Gr Fx '( )
v
x x
x
Gr
F F xx
   
  







  ( ) [ ( ) ( ) ]
4
3
1
4
1
4
8
The prime, of course, denotes differentiation with respect to
η.
Now, introducing the dimensionless temperature
T T
T T
G
s




( )
9
the three original partial differential equations (1) to (3) may
then be reduced to two ordinary differential equations of the
form
F FF F G''' '' '
( )   3 2 02
10
G FG'' '
Pr 3 0 11
where
Pr 
c
k
p
, Prandtl number of the fluid.
The continuity equation (1) is automatically satisfied through
introduction of the stream function. Eqs (10) and (11)
constitute a pair of simultaneous nonlinear ordinary
differential equations for the velocity and temperature
functions, F’ and G. They must be solved subject to the
following boundary conditions:
At y = 0: u = 0 i.e., at η = 0: F’ = 0
At y = 0: v = 0 i.e., at η = 0: F = 0
At y = 0: T = Ts i.e., at η = 0: G = 1
For large y: u→ 0 i.e., for large η: F’ = 0
For large y: T→ T∞ i.e., for large η: G = 0 12
The fact that the original partial differentials have been
reduced to a pair ordinary differential equations confirms the
assumptions that similarity solutions do in fact exist.

_______________________________________________________________________________________
3. SOLUTION PROCEDURE
Eqs (10) and (11) are coupled and must be solved
simultaneously, which is always the case in free-convection
problems. No analytic solution is known, so numerical
integration is necessary. There are two unknown initial
values at the wall. One must find the proper values of F’’(0)
and G’(0) which cause the velocity and temperature to vanish
for large η. The Prandtl number is a parameter.
The original approximate solutions given Pohlhausen [1] in
1930 were improved with digital-computer solutions
provided by Ostrach [ 3] and others in 1953.
3.1 Reduction of Equations to First-Order System
This is done easily by defining new variables:
z F1 
z z F2 1  
z z z F3 2 1    
   z z z F z z z z3 2 1 1 3 2
2
43 2      
z G4 
z z G5 4  
   Prz z G z z5 4 1 53   
with the following boundary conditions:
z F1 0 0 0( ) ( ) 
z z F2 10 0 0 0( ) ( ) ( )  
 
z z F2 1 0( ) ( ) ( )     
 
z G4 0 0 1( ) ( ) 
z G4 0( ) ( )    14
Eq(10) is third-order and is replaced by three first-order
equations, whereas eq(11) is second-order and is replaced
with two first-order equations.
Three of the boundary conditions are specified at
  0,
while the remaining two at    . One way to overcome
the lack of starting conditions is to guess the missing
boundary values. Since F’’ and G’ are proportional to the
velocity gradient and temperature gradient, F’’ should be
positive and G’ negative at
  0. The initial guessed
values are altered to obtained to obtained solutions at higher
values of Prandtl number.
3.2 Solution to Initial Value Problems
To solve Eq(13), we denote the two unknown initial values
by u1 and u2, the set of initial conditions is then:
z F1 0 0 0( ) ( ) 
z z F2 10 0 0 0( ) ( ) ( )  
 
z z z F u3 2 1 10 0 0 0( ) ( ) ( ) ( )   
  
z G4 0 0 1( ) ( ) 
z z G u5 4 20 0 0( ) ( ) ( )  
 
15
If Eqs (10) and (11) are solved with adaptive Runge-Kutta
method using the initial conditions in Eq(15), the computed
boundary values at    depend on the choice of u1 and
u2. We express this dependence as
z z F f u u2 1 1 1 2( ) ( ) ( ) ( , )     
 
z G f u u4 2 1 2( ) ( ) ( , )   
The correct choice of u1 and u2 yields the given boundary
conditions at    ; that is, it satisfies the equations
f u u
f u u
1 1 2
2 1 2
0
0
( , )
( , )

 16
These are simultaneous nonlinear equations that can be
solved by the Newton-Raphson method. A value of 10 is fine
for infinity, even if we integrate further nothing will change.
3.3 Program Details
This section describes a set of Matlab routines for the
solution of eqns (10) and (11) along with the boundary
conditions (12). They are listed in Table 1.
Table 1: A set of Matlab routines used sequentially to solve
Equations (10) & (11).
Matlab code Brief Description
deqs.m Defines the differential equations
(10) and (11).
incond.m Describes initial values for
integration, u1 and u2 are guessed
values, eq (15)
runKut5.m Integrates the initial value problem
(13) using adaptive Runge-Kutta
method.
residual.m Provides boundary residuals and
approximate solutions.
newtonraphson.
m
Provides correct values u1 and u2
using approximate solutions from
residual.m
runKut5.m Again integrates the initial value
problem (13) using correct values
of u1 and u2.

_______________________________________________________________________________________
The output of the code runKut5.m gives the tabulated values
of F, F’, F’’, G, G’ as function of η for various values of
Prandtl number.
4. INTERPRETATION OF THE RESULTS
4.1 Dimensionless Velocity and Temperature
Profiles
Physical quantities are related to the dimensionless functions
F and G through Eqs (5), (8) and (9). F and G are now
known. Some typical velocity and temperature profiles
obtained using this procedure are shown in Figs (2) and (3).
Some accurate initial values from this computation are listed
in Table 2. These theoretical profiles are in good agreement
with experimental laminar free-convection data on isothermal
vertical plates.
Table 2. Computed parameters from Eqs (10) & (11)
Pr F’’(0) -G(0)
0.01 0.947 0.118
0.05 0.899 0.179
0.10 0.856 0.234
0.50 0.713 0.442
0.72 0.676 0.505
1.0 0.642 0.567
2.0 0.571 0.717
3.0 0.531 0.816
6.0 0.465 1.008
10.0 0.420 1.170
20.0 0.362 1.423
30.0 0.332 1.591
50.0 0.296 1.826
100.0 0.252 2.194
1000.0 0.145 3.970
0 1 2 3 4 5 6 7 8 9 10
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7

F
Pr= 0.01, 0.1, 0.72, 1, 3, 10, 100, 1000
Fig 2. Dimensionless velocity distributions for various
Prandtl numbers
0 1 2 3 4 5 6 7 8 9 10
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1

G
Pr= 0.01, 0.1, 0.72, 1, 3, 10, 100, 1000
Fig 3. Dimensionless temperature distributions for various
Prandtl numbers
4.2 Flow and Heat Transfer Parameters
Besides the velocity and temperature distributions, it is often
desirable to compute other physically important quantities
(for example, shear stress, drag, heat-transfer-rate) associated
with the free-convection flow.
Shear stress is defines as:
 


 ( )
u
y
y 0
17
From Eqs (5) and (8), we can write




u
y x
Gr
Fx
 
2
42
1
4
( ) ( )
Substituting this expression into (17), we get the flow
parameter

2
4
0
2
1
4
x
Gr
F
x
( )
( ) 
18
The flow parameter is presented in Fig (4). From this figure,
various flow quantities for a given set of conditions can be
computed.

_______________________________________________________________________________________
10
-2
10
-1
10
0
10
1
10
2
10
3
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Pr
/(*/x2
)(4Grx
3
)(1/4)
Fig 4. Dimensionless flow parameter as function of Prandtl
number
Heat flux at the plate surface is:
   k
T
y
h T Ty s( ) ( )

 0
19
Accordingly, local Nusselt number can be written as
Nu
hx
k
x
T T
T
y
x
s
y  
 

( )
( )

 0
20
From eqs (5) and (9), we can write






T
y
T
y

   





[ ( )( )] [ ( ) ]T G T T
y
y
x
Gr
s
x
4
1
4
  ( ) '( ) ( )T T G
x
Gr
s
x

1
4
1
4

 ( )
( ) '( )
T T
x
Gr
Gs x
4
1
4

Substitution of this expression into eq (20) yields the heat-
transfer parameter
Nu
Gr
Gx
x( )
'( )
4
01
4
 
21
It is a function of only Prandtl number, shown in Fig.5. Le
Fevre [4] developed the semi-empirical formula (22) from
numerical solutions to Pohlhausen’s approximation [1] as:
Nu
Gr
x
x

 
( )
. Pr
( . . Pr . Pr)4
075
0609 1221 1238
1
4
1
2
1
2
1
4
22
Figure 5 illustrates the comparison of the present
computation (21) with the semi-empirical formula (22).
10
-2
10
-1
10
0
10
1
10
2
10
3
10
-2
10
-1
10
0
10
1
Pr
Nux
/(Grx
/4)(1/4)
computed, eq (22)
empirical correlation, eq(23)
Fig 5. Comparison of the present computation with semi-
empirical formula.
Local Nusselt number can be calculated from this equation,
for example
For Pr=0.72 and Grx =109
Nu
Gr
Gx
x     ( ) '( ) ( ) . .4 0
10
4
05046 6345
1
4
9
1
4
This indicates that large heat transfer coefficients can be
obtained with free convection flows.
5. COMPARISON OF RESULTS WITH
PREVIOUS WORK
Laminar free-convection flows of air as generated by
gravitational forces about isothermal vertical flat plates were
made by Schmidt and Beckmann [2] in which velocity
measurements at various points along the plate were made by
means of a quartz-filament anemometer and the temperature
measurements were obtained by means of manganese-
constantan thermocouples. In particular, they measured
velocity and temperature for a 12.5-cm-high vertical plate for
the conditions Ts=650
C and T∞=150
C. By dimensionalizing
the solution (Figs. 2 & 3), i.e., eqs (5), (8) and (9), we can
compute the dimensional velocity and temperature as a
function of y at a particular value of x:
y x
Grx
  ( )
4 1
4

_______________________________________________________________________________________
u
x
F Grx 
2 1
4

( )
T T G T Ts   ( ) 23
The velocity and temperature profiles of air for the above
mentioned experimental conditions (Ts=650
C, T∞=150
C)
obtained from these computations are then compared with the
experimental data of Schmidt and Beckmann. Figures 6 and 7
illustrate the comparison of the present numerical work with
experimental data.
0 0.5 1 1.5 2 2.5 3
0
5
10
15
20
25
y (cm)
Velocity(cm/s)
x = 11, 7, 4, 2, 1 (cm)
present computation: dotted lines
experimental data: symbols
Fig 6. Comparison of the computed velocity profiles with
experimental data.
0 0.5 1 1.5 2 2.5 3
15
20
25
30
35
40
45
50
55
60
65
y (cm)
Temperature(0
C)
x=11, 7, 4, 2, 1 (cm)
present computation: dotted lines
experimental data: symbols
Fig 7. Comparison of the computed temperature profiles with
experimental data.
6. CONCLUSION
In the present numerical simulation, laminar free-convection
flow and heat transfer over an isothermal vertical plate is
presented. Details of the solution procedure of the nonlinear
coupled partial differential equations of flow are discussed.
The computer codes are developed for this numerical
analysis in Matlab environment. Velocity and temperature
profiles for Prandtl numbers of 0.01, 0.05, 0.1, 0.5, 0.72, 1, 2,
3, 6, 10, 20, 30, 50, 100 and 1000 are computed using these
codes. The computed and experimental velocity and
temperature distributions are in good agreement. Flow and
heat transfer parameters (giving physically important
quantities such as shear stress, drag, heat-transfer-rate) are
derived as functions of Prandtl number alone. The computed
results of the present simulation are compared with
frequently used semi empirical heat transfer correlation and
find a good agreement. A good agreement between the
present results and the past indicates that the present
numerical simulation may be an efficient and stable
numerical scheme in natural convection.
REFERENCES
[1]. Pohlhausen, E.: Der W‫ن‬rmeaustausch zwischen festen
Kِrpern und Flüssigkeiten mit kleiner Reibung und kleiner
W‫ن‬rmeleitung. Z. Angew. Math. Mech. 1, 115-121 (1921).
[2]. E. Schmidt and W. Beckmann, Tech. Mech. U.
Thermodynamik, 1, 341 and 391 (1930).
[3]. S. Ostrach, "An Analysis of Laminar Free-Convection
Flow and Heat Transfer About a Flat Plate Parallel to the
Direction of the Generating Body Force," National Advisory
Committee for Aeronautics Technical Note, No. 2635, 1952.
[4]. E. J. Le Fevre, “Laminar Free Convection from a
Vertical Plane Surface,” Proc. 9th Int. Congress Applied
Mechanics ,Brussels, Vol. 4, pp. 168-174 (1956).
[5]. Finston, M., “Free Convection past a Vertical Plate”,
Journal of Applied Mathematical Physics (ZAMP) 7, pp. 527
– 529, 1956.
[6]. Sparrow, E.M. and Gregg, J.L., “Similar Solutions for
Laminar Free Convection from a Nonisothermal Vertical
Plate”, Trans. ASME, Journal of Heat Transfer 80, pp. 379-
387, 1958.
[7]. Foote, J.R., “An Asymtotic Method for Free Convection
Past a Vertical Plate”, Journal of Applied Mathematical
Physics (ZAMP) 9, pp. 64 – 67, 1958.
[8]. Kuiken, H.K., “General Series Solution for Free
Convection Past a Non-Isothermal Vertical Flat Plate”,
Applied Scientific Research 20, pp. 205 – 215, 1969
[9]. Kao, T.T., “Locally Nonsimilar Solution for Laminar
Free Convection Adjacent to a Vertical Wall”, Journal of
Heat Transfer 98, pp. 321 – 322, 1976.
[10]. Elbashbeshy, E.M.A., “Free Convection Flow with
Variable Viscosity and Thermal Diffusivity along a Vertical
Plate in the Presence of the Magnetic Field”, International
Journalof Engineering Science 38, pp. 207 – 213, 2000.
[11]. Bejan, A., Heat Transfer, John Wiley, New York, 1993.
[12]. Incropera, F.P. and DeWitt D.P., Introduction to Heat
Transfer, Fourth edition, John Wiley, New York, 2002.
[13]. Çengel, Y.A., Heat Transfer, Second edition, McGraw–
Hill, New York, 2003.

_______________________________________________________________________________________
[14]. Lienhard IV, J.H. and Lienhard V, J.H., A Heat
Transfer Textbook, Phlogiston Press, Cambridge, MA, 2003.
[15]. Nellis, G. and Klein, S., Heat Transfer, Cambridge
University Press, London, UK, 2008.
BIOGRAPHY
Dr. Asish Mitra did M.Sc. in Physics in
Presidency College, and PhD and post
PhD research work on Heat Transfer and
CFD in Jadavpur University. He is
presently Associate Professor & HOD of
the Department of Basic Science &
Humanities of College of Engineering &
Management, Kolaghat, East Midnapur, West Bengal. He is
having more than 12 years post-PhD teaching experience in
Physics at UG level. He is a regular reviewer of the
International Journals “Heat and Mass Transfer” and
“WSEAS” (World Scientific & Engineering Academy &
Society). He is the author of 06 papers in International
Journals, 14 papers in International Conferences and 06
papers in Seminars. He has completed a CSIR-sponsored
project as Co-investigator and organized 04 National
Workshops and 01 International Conference as convener. His
field of research is Computational Heat Transfer and Fluid
Flow.

Numerical simulation on laminar free convection flow and heat transfer over an isothermal vertical plate

More Related Content

What's hot (19)

Viewers also liked (6)

Similar to Numerical simulation on laminar free convection flow and heat transfer over an isothermal vertical plate (20)

More from eSAT Journals (20)

Recently uploaded (20)

Numerical simulation on laminar free convection flow and heat transfer over an isothermal vertical plate