Hello everyone,

I am try to apply multigrid to a 2D Navier-Stokes Finite Volume solver. Although I am looking for a steady state solution, I solve the problem with a time marching approach. I manage to code the problem in a 1d simpler case, but moving to the 2d proved more complex than I thought. Any suggestions on how to do this?

Thanks in advance

P.S. I obtained the adjoint equation for the coarser grid. Once linearised, this looks very similar to the equation solved in the fine grid, but the coefficients are now multiplied by the corrections rather than by the actual variables. Everything works well in 1D, but I have serious troubles in 2D

are you solving linear or non linear equation. and what type of problem are you solving.

I am solving the full Navier-Stokes equations in a tube with a time marching approach. Of course this means nn linear equations ...

Hi,

I recommend the paper by Raw: A coupled AMG fpr the 3D Navier-Stokes equations. It is a rather simple AMG which I also implemented. It is not based on very subtle mathematical approximation but on physical conservation properties. It is one of the most useful papers I ever read! Also the Algebraic Multigrid Tutorial by Briggs, Cormick and Henson is a very good literature about multigrid methods. Further the page www.mgnet.org.

Hope, this helps...

Actually I am trying to use geometric multigrid rather than algebraic ... do you have a copy of the paper? Could you please send it to me?

Thanks