Before running my simulations I'm trying to validate my grids by performing a sensitivity analysis to ensure I'm converging to a grid independent solution.

However, although I've generated 6 grids ranging from very coarse to extremely dense, all meeting the boundary layer conditions of my turbulence model, I'm finding that my results are not asymptotically converging. That is, each grid had results that consistently differ by 8-9%.

I have never run into this situation before and would like any ideas as to how I can solve this problem.

Thanks!

Try using grid adaption in FLUENT and see what sort of grid you will get.

Hope that helps.

[tsibel]

Hi everyone
I have a problem. You can see from the picture that I tried different mesh sizes. But tendency of the lines are different. This is steady state solution and SST(automatic wall function).

Does somebody has an idea?

Thanks!