[A.Devesa]

Hello everybody,

i observe some strange behavior in the convergence process of my simulations. Since i could reproduce the bug on different machines, i am now very sure, it's a problem with OF and not a hardware problem (like a memory bug)...

Did somebody observe these jumps as well? Does anybody have a tip to solve that?

The main problem is that the values (here on the graph, the drag for example) are landing on different levels after those "jumps"... Which makes the convergence last much longer, and may be false!

My setup:
- 46 Mcells mesh, converted from fluent format
- simulating with simpleFoam 1.7.x
- k-omega-SST-Model
- Pressure solving: GAMG with tolerance 1e-7
- Velocity and k/omega solving: PBiCG with tolerance 1e-8
- relaxationFactors: p 0.3; 0.7 for the rest

Thank you!
Antoine