View Single Post
Feb21-12, 09:12 PM
Sci Advisor
HW Helper
P: 7,110
Crank-Nicolson instability in 3D

Googling around, the Douglas-Gunn method seems to be the favorite 3D ADI method.

Re your test problems and interpolation etc, some problems where the steady state is a uniform heat flux in different directions, (including heat flux not parallel to the finite difference grid) might flush out some errors.

Most of my heat transfer work has been with FE codes rather than FD, so unfortunately I've forgottem most of what I used to know about FD algorithms