CFD Online Discussion Forums

CFD Online Discussion Forums (
-   Main CFD Forum (
-   -   solving a Matrix (

Behafarid September 3, 2006 07:07

solving a Matrix
Hello friends.

I'm working on a very larg 2-D zone. I have to divide it to 3000*500 nodes at least. I need a simple method to solve this problem with an acceptable convergence. My computer CPU is a 3GHZ Intel procesor. I'm eagerly waiting for your suggestions.

Best regards. Goodbye.

SAEED MAHLOO September 3, 2006 13:20

Re: solving a Matrix
HELLO I am SAEED MAHLOO,Aerospace(Propulsion) engineer.If your matrix is a tridiagonal you should use TDMA method(THOMAS ALGORITHM) ,If your matrix is 5 diagonal you should use ADI method.Please explain more or Email me? WITH BEST REGARDS GOOOD LUCK


Nico September 6, 2006 10:05

Re: solving a Matrix
Nothing simpler than Jacoby/Gauss seidel iterative methods.

Behafarid September 9, 2006 23:14

Re: solving a Matrix
Thank you very much. Unfortunately the matrix is so big (3000*500) that by this method solving the matrix needs a lot of time. Besides, it's a subroutine in a program and this matrix has to be solved 50 000 time during the specefic period of time. I'm seeking for a faster method to solve big matrices. Goodbye.

All times are GMT -4. The time now is 13:57.