Scribd
Upload
719 views

Finite Difference Method For Solving Heat and Mass Transfer Problems

1. The finite difference method uses central difference approximations to discretize the second derivative in space and forward difference to discretize the first derivative in time, resulting in an explicit scheme to solve heat and mass transfer problems. 2. For the explicit method, the algorithm starts with known initial and boundary conditions and marches forward in time by directly solving for temperatures at interior nodes from the previous time step. 3. The implicit method discretizes both spatial and temporal derivatives at the new time level, resulting in a system of equations that must be solved simultaneously at each step to determine temperatures, without restrictions on time step size.

Uploaded by

hiranyakumar
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
719 views

Finite Difference Method For Solving Heat and Mass Transfer Problems

1. The finite difference method uses central difference approximations to discretize the second derivative in space and forward difference to discretize the first derivative in time, resulting in an explicit scheme to solve heat and mass transfer problems. 2. For the explicit method, the algorithm starts with known initial and boundary conditions and marches forward in time by directly solving for temperatures at interior nodes from the previous time step. 3. The implicit method discretizes both spatial and temporal derivatives at the new time level, resulting in a system of equations that must be solved simultaneously at each step to determine temperatures, without restrictions on time step size.

Uploaded by

hiranyakumar
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 2

FINITE DIFFERENCE METHOD FOR SOLVING HEAT AND MASS TRANSFER PROBLEMS

SIMPLE EXPLICIT METHOD


For simple analysis we consider a one dimensional unsteady state parabolic system
∂T ( x , t ) ∂ ²T (x ,t )
=α , 0<x<L, t>0 (1)
∂t ∂ x²
The region 0≤x≤L is divided into M parts of equal mesh size.
L
Δx=
M
The differential equation is discretized by usingsecond order accurate central difference formula for
the second derivative and the first order accurate forward difference formula for the time derivative
to yield,
T ni +1−T in T ni+1 −2T ni +T ni−1
= + O[ Δt , ( Δx )2]
Δt (Δx) ²
where
T ( x ,t )=T ( iΔx , nΔt ) ¿ T ni
The above equation is rearranged as,
T n+1
i ¿ r T ni−1+ (1−2 r ) T ni +r T ni +1 (2)
Δt
Where, r =α
(Δx) ²
n=0,1,2............ and i= 1,2,............,M-1
2
with a truncation error of order O[ Δt , ( Δx ) ]
Equation (2) is called simple explicit form of finite difference approximation of the
n+1
diffusion equation (1) because it involves only one unknown T i for the time level n+1, which can
directly be evaluated from equation (2) when the potentials T ni , T ni+1∧T ni−1 at the previous time
level n are available.
The system forT n+1
i i=1,2,............,M-1 provides M-1 algebraic equations but contains M+1 unknown
node potentials ,(i=0,1,2,.....M). two additional relations are needed to make the number of
equations equal to number of unknowns are obtained from the two boundary conditions at i=1 and
i=M. If the boundary potentials are prescribed , then the number of equations are equal to the
number of unknowns. However for convection or prescribed heat flux boundary conditions ,the
boundary potentials are unknown . In such situations , two additional relations are obtained by
discretizing the boundary conditions .
ALGORITHM
Suppose the potentials prescribed at i=0 and i=M, we have
T n0=T a=Known
T nM =T b =Known
Then the system of equations (2) provide M-1 explicit relations for determination of M-1
unknown internal node potentials T n+1 n n
i , i=1,2...,M-1 since the boundary potentials T 0andT M are
known for all times. The computational algorithm is as follows
n+1
1. Start the calculations with n=0. Compute T i i=1,2,............,M-1 , at the end of the first time
step from equation (2), since the right hand side of the equation is known from the initial
boundary conditions.
2. Set n=1 and calculate T 2i , i=1,2,............,M-1, at the end of the second time stepfrom
equation (2) because the right hand side of the equation is known from the previous step.
3. Repeat the procedure for each subsequent time step and continue calculations unill a
specified time or some specified value of potential is reached.
SIMPLE IMPLICIT METHOD

If the calculations are performed over a large period of time , the number of steps hence the number
of calculations may become prohibitively large. To alleviate this difficulty finite difference schemes
that are not restrictive to the size of the time step Δt have been developed and one such method is
the simple implicit method . We consider a simple one dimensional diffusion equation.

∂T ∂²T
=α
∂t ∂ x²
The space derivative is discretized at the n+1 time level as

n+1 n+1 n+1


∂ ² T T i +1 −2 T i +T i−1
= +O [ ( Δx )2] (3)
∂x ² ( Δx)²
And the time derivative as

n +1 n
∂T T i −T i
= +O[ Δt ] (4)
∂t Δt

Introducing equation (3) into (4) we obtain the simple implicit finite difference approximation for
diffusion equation as

T ni +1−T in T n+1 −2T n+1 +T n+1


= i+1 i i−1
Δt (Δx) ²

Which is accurate to O[ Δt , ( Δx )2 ]. This is an implicit scheme because each time level equation has
to be solved simultaneously in order to determine the nodal temperatures.

You might also like