It looks like the number of inflation layers and first cell height may be fine. However, it looks like the remaining mesh structure outside of the inflation layers *may* be too coarse. Have you refined the fluid cells outside of the inflation layers? I'm unsure if you have performed a mesh sensitivity analysis other than considering inflation to resolve the boundary layer. Thus, I can't be sure that you have a grid-independent solution anyway. I would make sure that you have sufficient refinement everywhere else. In addition, what were your mesh metrics/statistics? It appears that you may be using tetrahedral cells. If so, what was the maximum and average mesh skewness? You may have convergence issues if you have local areas of highly skewed cells. Finally, if you have already checked the aforementioned items, have you changed the under-relaxation factors? Under-relaxation can often assist in convergence. I hope this can guide you on the correct path.

Dear David Calamas,

Thank you for your suggestion. Actually, I have tried to do variation on y+ (0.5,0.75 and 1). The relative error is from 4-10%. I want to get relative error around 1%. Hence, I choose y+ = 0.25. Yea, I just concentrate on the inflation layer not out side layer. I will try to fix it. I do not change the Under Relaxation Factor value. Actually, I am trying do similar mesh in the attached file. The mesh statistics is also attached.

Thank you for your helps

Best regards

Generally,

- Try to refine mesh

- change relaxation factor

- are you use enhanced wall treatment ?

I have tried to do mesh refinement in the center of the tube and the problem is the simulation is not converged. Should I change all of relaxation factor? In this case I am using K-Omega SST. So I do not use enhanced wall treatment.