|
[Sponsors] |
June 28, 2004, 02:15 |
Little confusion
|
#1 |
Guest
Posts: n/a
|
Cfd beginner I have energy equation E_t=k*dT^2/d^2x Where E_t is a time derivative of total energy w.r.t time, k thermal conductivity and T is temperature. Which explicit finite difference scheme can be used? , what about time step to satisfy CFL condition, as I know k very small then very large time step can be used? when using second order central difference scheme What about the stability when T has discontinuous initial conditions ?
|
|
June 28, 2004, 08:35 |
Re: Little confusion
|
#2 |
Guest
Posts: n/a
|
Stability conditions for explicit scheme
E^n+1_i = E^n_i + dt*k*(T^n_i+1 - 2*T^n_i + T^n_i-1)/dx^2 are: dt < dx^2/(2*k) in 1D dt < dx^2/(4*k) in 2D dt < dx^2/(6*k) in 3D and they don't depend on initial conditions (discontinous or smooth). |
|
June 28, 2004, 19:22 |
Re: Little confusion
|
#3 |
Guest
Posts: n/a
|
Thanks for reply Alexei But I think by this scheme there an oscillation when there is discontinues T data.
|
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Confusion of the Courant number definition | bearcat | Main CFD Forum | 3 | December 9, 2013 12:12 |
Confusion face-addressing for surfaceField | ngj | OpenFOAM Programming & Development | 5 | October 2, 2012 16:40 |
Sub-domain Confusion | Flaky | CFX | 6 | October 7, 2010 07:40 |
Confusion abt non-uniform staggered grid | CF | Main CFD Forum | 2 | August 29, 2007 04:46 |
Avoiding Code Confusion | Jonas Larsson | CFX | 3 | April 18, 2001 03:44 |