[psigcochair]

My FMG completes level 4, but when it goes to level 3, it says:
Normalized Residual = 1.#QNAN

So what happened between level 4 and 3?

at level 4, it says:
FMG: FAS reached maximum iterations.
Normalized residual = 729.568

FMG: Interpolate solution on next level...end

then it begins the messed up level 3.
The 1.#QNANs continue through levels 2 and 1.

Finally, it ends with:
Primitive Error at Node 7: Floating point exception
->1.->2.->3.->4.->5.<<<<<

What does all this mean? And how do I fix it?

Thanks!
Michelle

[Chris D]

Originally Posted by psigcochair

My FMG completes level 4, but when it goes to level 3, it says:
Normalized Residual = 1.#QNAN

So what happened between level 4 and 3?

at level 4, it says:
FMG: FAS reached maximum iterations.
Normalized residual = 729.568

FMG: Interpolate solution on next level...end

then it begins the messed up level 3.
The 1.#QNANs continue through levels 2 and 1.

Finally, it ends with:
Primitive Error at Node 7: Floating point exception
->1.->2.->3.->4.->5.<<<<<

What does all this mean? And how do I fix it?

Thanks!
Michelle

The residual are increasing, meaning that the simulation is diverging. Assuming that you have a good grid, you could try lowering the CFL number used in the fmg initialization. To do this, in the text user interface type "solve/initialize/set-fmg" and work through the text menu until you you get to the CFL number. The default is 0.75.

If that doesn't work, you could try increasing the number of multigrid levels, increasing the number of iterations performed on course multigrid levels, or lowering the residual threshold on the course multigrid levels.

You should also make sure that your boundary conditions, grid, and case setup are all ok first. Does your simulation bomb if you run it without fmg?