[plm]

Hi all,

Can anyone explain what is happening when a solution starts to converge and then gets so far and starts to go the opposite way, eventually exploding into an error.....

I'm a bit unsure about why this should occur and what it means about how much I can trust the results when they are converged the most (before explosion).

Does this indicate a problem, specifically in the fvSchemes used to solve the problem perhaps?

Any thoughts would be welcome and thanks for your time.
plm

[danishdude]

Can you provide more detail about your case? Post the case if it's not too large, or at least post the boundary condition files, fvSchemes, and fvSolutions.

I've seen similar behavior when using a zeroGradient condition on the outlet velocity; however, during the transients, a local vortex develops leading to a reverse flow condition. If that's what's happening, you can fix it by changing from a zeroGradient to an inletOutlet condition.

[plm]

Hi danishdude, thanks for your time.
I'm trying ultimately to model a turbine blade using LES, but as a starting point i'm trying to use a RANS approach with simpleFoam. I'm using an unstructured mesh generated with gmsh.

I am indeed using a zeroGradient bc on the outlet, and though I'm not familiar with the inletOutlet bc I'll try and give it a go.

Here's a copy of some of the case files, I'd be very interested on your thoughts.

Thanks,
plm