Scribd
Upload
103 views

Finite Difference Method: I I I I I I

The finite difference method numerically approximates differential equations by using finite values of x rather than infinitesimal values. It can approximate the slope and curvature of functions using forward, backward, and central differencing techniques. The method also approximates changes in time using similar differencing of temperature values at present and future time steps. Stability of the solution requires that the Fourier cell constant, which is the ratio of the time step to the square of the distance step, be less than 0.25. There are four common finite difference methods including explicit, implicit, Crank-Nicolson, and alternating direction implicit.

Uploaded by

hazem
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
103 views

Finite Difference Method: I I I I I I

The finite difference method numerically approximates differential equations by using finite values of x rather than infinitesimal values. It can approximate the slope and curvature of functions using forward, backward, and central differencing techniques. The method also approximates changes in time using similar differencing of temperature values at present and future time steps. Stability of the solution requires that the Fourier cell constant, which is the ratio of the time step to the square of the distance step, be less than 0.25. There are four common finite difference methods including explicit, implicit, Crank-Nicolson, and alternating direction implicit.

Uploaded by

hazem
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

Finite Difference Method

The finite difference method is a means of solving differential equations numerically by


using a finite value for x rather than x 0.

The slope dT/dx can be approximated at the point xi in several ways. The forward
differencing technique uses
dT Ti+1 − Ti Ti +1 − Ti
≈ =
dx xi+1 − xi ∆x
The backward differencing technique uses
dT Ti − Ti−1
≈
dx ∆x
The central differencing technique uses
dT Ti +1 − Ti−1
≈
dx 2∆x
Approximations of differential equations of order 2 or higher follow the same general
scheme, using the difference between the forward and backward differences to calculate
the curvature at a point:
 dT   Ti+1 − Ti   Ti − Ti−1 
d −
2
d T  dx   ∆x   ∆x  Ti +1 − 2Ti + Ti−1
= ≈ =
dt 2 dx ∆x ∆x 2
So, this works fine for changes in space, but what about changes in time? For central
differencing we can write

dT T t+ − T to
≈
dt ∆t
where Tto indicates the temperature at the present time step and Tt+ indicates the
temperature at the next (future) time step. Thus, the diffusion equation is written

T 2
T T t+ − T to T t o − 2Ti t o + Ti−1
to
= ≈ = i+1
t x2 ∆t ∆x 2
and the solution for T at the next (future) time step is

Ti t + = Ti t o +
∆x
(
∆t t o
T − 2Ti t o + Ti−1
2 i+1
to
)

The term
∆t
∆x 2
in the above equation is the Fourier cell constant and must be smaller than 0.25 for the
equation to remain stable. You must check your finite-element model for accuracy by
comparing it to an analytical solution or reducing the time step and seeing that the result
does not change.
There are four methods commonly used in finite-difference calculations:

The alternating-direction implicit method divides each time step in two: one half timestep
that is explicit in one direction and implicit in the other and a second half timestep that is
the opposite.

You might also like