[Speeder]

Good morning dear forum!

After reading along quietly for some time now, I wanted to turn to you with my problem. First of all, you have already helped me a lot, thank all of you!

As part of my final thesis, I want to investigate the effect of surface damage on rolling bearings.

I have created the corresponding geometry as an STL. The problem is that the ball radius is 4mm but the surface damage is in the range of 2-3 microns (The damage is scanned from a real bearing, so i want to actually use it instead of creating one of my own via blockMesh alone.) The gap between them even goes to less than 1micron. So on the one hand I need a rather coarse grid to resolve the ball surface and on the other hand a fine grid to resolve the surface damage in the center as well as the fluid in-between them.

The grids I have created so far work in principle, but I am not completely satisfied with them, because in some of them a grid point diverges sooner or later.

Adjusting the grading in blockMesh makes the grid very imprecise at the corners. If I grade in X and Y, for example, I get a cross-shaped fine grid which becomes uneven in the quadrant corners. I have also tried to create the blockMesh from 3 adjacent blocks, but the number of cells at the transitions must be the same, and that doesn't help me either. Increasing the cell number just tends to make the mesh even worse.

I have also worked on castellating and snapcontrols in SHM, but they don't help either. refining the surface more just adds more cells, but doesnt change the shape itself. I have added a picture and, if any of you want to try it, the case files. In the current state, the damage is resolved well, but the ball surface not so much. So sooner or later, the pressure will explode in one cell.

Thanks in advance!

PS: We're using FE4.0 in case you're wondering

[Speeder]

Just realized that i forgot to add the case files, sorry!

I uploaded it to dropbox, since its about 1MB in size. Hope thats ok.

Happy eastern to all of you!

https://www.dropbox.com/s/frglbdafbh...fluid.zip?dl=0