Hello, I managed to fix these two problems by adjusting the mesh zones. Now the code is running. Thank you very much.Yvan Fournier wrote:Hello,
The least-squares gradient quality might not be so bad if only 1% of cells are "bad", but I recommend visualizing the mesh quality to see where cells are bad: it may be due to the joining, but also to other factors, such as warping.
The number of iterations for pressure seems pretty high, especially if you are using multigrid and this is a "pseudo-equivalent" (number of iterations), but the first iterations might be more costly, and there might be other factors, such as gradient reconstruction choice, and especially time step/CFL aspects.
Here again, postprocessing/visualizing after a few iterations will probably provide more insight.
Regards,
Yvan
However, the velocity and CFL number blow up after several hundred steps even if I set the time step as a very small value. I attached the CFL contours of 100, 200 and 300 time steps, after that the velocity diverged from the near trailing edge zone. I think this is probably a initial flow behavior of a circular trailing edge aerofoil, because I used same settings to a sharp trailing edge aerofoil and it runs fine.
I will appreciate if you can give me some suggestions on this problem.
Regards,
Xiang