Hi,

I am working on the simulation of a 3D turbine blade using gambit and fluent. I wanted to capture the boundary layer on the blade surface. I did this by setting up 2D grid on the hub surface and then project it to 3D volume with Cooper method in Gambit. The problem is that when I smoothed the 2D domain, the smoothing function tend to unify all the cell sizes so that the cell attached to the blade surface is greatly enlarged and cannot capture the boundary layer anymore. The 3D Cooper method also did this smoothing automatically.

I tried to use the boundary layer function to build up the blade surface grid, but it doesn't work well for 3D cases. And a special turbo function in Gambit requires an extra license that I don't have.

Does anyone has the experience on that? Thank you very much.

Hi bearcat my experience is about centrifugal compressor meshing. I don't have G/turbo too. My first attemp was to create a topology by other pre-processor or other cad. So I created a H topology splitting surfaces by lines and then I meshed with quad mapped plane elements. After I meshed all surfaces and I started the volume meshing. Unfortunately, a very complicate surface ruined all my work an gambit wasn't able to close the mesh. At the end, I used a unstructured mesh, but by the help of custom line, it's possible to help gambit to preserve small element where required.

regards Dothan