Hi all

I am doing simulation of flow inside a very complex geometry using STAR V.3.2. There can be a question such that how my calcultion is close to exact real flow conditions. In other words, how large is the uncertainty of my calcultion.

In order to give an answer to this, CFD engineers need to examine mesh dependency on the results. The engineers will increase cell number by double or triple to conclude that the results are almost free from error due to mesh size. In this regard, I think adaptive cell refinement is a good alternative. How do you think? If you also think so, What criteria (i.e., 10% top gradient) do you suggest?

Jihwan

There is an error estimator in star which you can use.