[ariorus]

Hi everybody,

I'd like to know if it is possible to set the size of the timestep while using the stationary solver simplefoam.
I mean the numerical time step. I guess that the timestep can be different according to the cell size in stationary problems.

In some of my testcases I noticed that in small cells I have convergence problems and I'd like to control the timestep in these cells.

Have I to use icofoam even if I'm interested only in the stationary results?

Thank you for any suggestion.

[mattijs]

simpleFoam uses underrelaxation which is equivalent to local time stepping.

There is no control apart from changing the underrelaxation factor. Try using different discretisation. Or maybe your problem is transient.

[ariorus]

Just after posting my message I had a look at the code and I realized there was actually no local timestepping, but relaxation. I should have known that.
Sorry for my silly question.

Thank you very much for your answer anyway.

By the way I have another stupid question.

Is that possible that recompiling some solver (simpleFoam in this case) without any changing I get different (best in convergence) results? Or must I have missed something?

Thank you again.

[mattijs]

There might be very small differences if you compile with/without optimization or on a different machine. We compile with -O3 -ffast-math on a p4.