CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   Multigrid method (http://www.cfd-online.com/Forums/main/11270-multigrid-method.html)

Suresh kumar April 20, 2006 09:30

Multigrid method
 
Hello all,

i am solving steady 2d lid driven caviy problem in finite volume method using multigrid method......Can anyone suggest me which multigrid procedure is used for solving governing equations.........earlier i tried FAS(full approximation scheme) for NS equations but i didn't get any result......can anyone tell me how to solve NS equations by finite volume method using multigrid in a detailed manner?....now i am trying to apply linear multigrid method(coarse grid correction scheme) to pressure correction equation of finite volume method only apart from momentum equations......is it correct to solve only pressure correction equation by multigrid method? if so what's the procedure i have to follow and in which part i have to focus mainly?.........please help me in this regard.................thank u in advance

hurricane April 24, 2006 20:07

Re: Multigrid method
 
Hi,

I'm not a guru of Mutligrid methods or Finite Volume Techniques but I have some ideas that I hope will help you :

- First : Are you sure that your problem is well defined, caus' when working on incompressible flows using primitive variables there's some troubles with the pressure equation and pressure boundary conditions which can be the source of your errors.

- Second : I don't think there's a prescribed multigrid method in this case, so you have two solutions : search for papers of previous works on this problem, select the best result's one and try to fellow the same procedure or keep trying methods until u found a good one, and maybe you'll publish :) Of course all multigrid methods are good, but some are more accuracy, or faster, than others. That's why I'm suprised that you don't get results.

- Third : It's correct to use multigrid method to resolve only "elliptic" pressure equation and use a FV scheme to resolve momentum equations.

- Finally : check your program and your algorithm more than one time, maybe you've just done a typing error or you forgot something.

In my point of you, in order to eliminate errors sources, try to resolve the pressure equation by using a classical iterative or direct method, if you have logical but maybe not very good results, then you can move to multigrid methods.

Good luck.


All times are GMT -4. The time now is 05:56.