Hi everyone,

I am now trying to change my cartesian FVM NS solver to a structured one. It uses fractional step method. I am discretizing and interpolating in physical space.

However after debugging for quite a while, I am still not able to obtain the correct results for my test case, which is a skewed(inclined) lid driven cavity. It either converges to the wrong ans or diverges, especially as grid points increase.

does anyone knows what could be the possible error that might occur? I am wondering if the cause could be the interpolation scheme.

I am using the interpolation suggested by Kim & Choi (JCP 162 p411) which they use for unstructured grids.

thanks!

its hard to tell what could be wrong with your code from here, but if the code converges with cartesian meshes, and does not converge when used with inclined meshes, the most probable cause is descetisation of diffusion terms.

having said that, there is another element that works against you is how you calculate the gradients. try to search about this issue.