If it's for the fluid flow, why not trying the unstructured triangular mesh with boundary layer? I understand that having an structured mesh is nicer and sometimes better. But as you correctly said, skewness can affect the flow simulation convergence.

BTW, the skewness does not "seem" to be very bad in the image. Quantitative values are required to judge accurately, though.

Also, in COMSOL, you can control the mesh generation locally, which may help getting a high quality mesh.

I hope it helps.

Good luck.

Hi Hamed, thanks for answering. A bit more information:

The skewnes quality is around 0.5 in the corner/transition elements (wich is the lowest of the domain). I don't know when to asume is a bad value, but the mesh shown in the video seems to have better quality, and avoids the abrut diference in elements shape.

I've tried the boundary layer option but i have 2 walls, the onein question, and a cilinder with it's axis perpendicular to it. The boundary layer is not easy to configure in that corner. So i use a surface mesh and a sweep, and with an element distribution in the edges i manage to reduce the wall element size.

The image shows a detail of the encounter of this 2 walls. The blue faces are walls BC, the gray one is simmetryc BC

Hi Felipe,

The quality of 0.5 should be good enough. Please take a look at :

https://www.comsol.com/blogs/how-to-set-up-a-mesh-in-comsol-multiphysics-for-cfd-analyses/

Also, the necessity of boundary layer strongly depends on the Reynolds number.