I'm trying to implement a multigrid strategy into my code (unsteady compressible N-S cylindric coord.); it seems to me that the restriction phase and relax on coarser grid works well, because the convergence history on the two sub grids is actually fast to achieve a steady state...however the convergence history on the main grid is slow, a little faster than the case without multigrid. This suggests me that there may be something wrong in the interpolation phase, from the Cosarser to finer grid, but I am using a very simple bilinear interpolation for which I have found and checked the formulas. Is there any expert on the topic? I have several other questions

For the simple bilinear interpolation, did you take account for the radial direction weight?

Many thanks Junseok, but could you please give me any details? I can't see: how does the weight in the radial direction come out? daniel

See my thesis Chapter 4 at my webpage and follow similar approach for the interpolation.

http://math.uci.edu/~jskim/papers.html

Junseok