I have modelled an unbaffled stirred tank using the k-e model. The model converged using first order discretisation first, and then using second order. I then switched to the realizable k-e model, ran the model using the second order parameters and it converged. Is this an acceptable approach, or should I have started from scratch with the realizable model (i.e. gone through first order, then second order, instead of going straight to second order)?

I think it's acceptable.

Hi

ap