Issues with running simulations after recomposing and decomposing
 

I am having an issue where if I run a simulation using pimpleDyMFoam on 96 processors for a few time steps, then recompose the case time steps, and then decompose to a different number of processors (say 32), the simulation immediately blows up (Courant numbers go very high!). Does anyone have a suggestion about how to recompose and decompose without this error? I am attempting to use computational resources on campus as they become available.

Thanks,

Bob

which method do you use for decomposing?