[Adam_K]

Hello,

When meshing in gmsh I get the error "Impossible to mesh periodic surface 143" and then as a consequence there are "No elements in volume 1". I have reported this as a gmsh issue, and Christophe Geuzaine suggested that I try using Salome's healing routines. My geo and brep files can be found at the above link.

Overall, my goal here is to generate a mesh of the spacing between the particles in a porous powder compact and then simulate the diffusion across the film. The contact points between particles generally leads to very low quality meshes. One of the global solutions is the simply shrink all the particles by a few percent. The meshing + simulation work fine when I do that. I would like to try a local method of using an inflated version of a particle as a boolean tool to cut its neighbors, leaving a small gap between the particles. More details can be found in this paper.

I have narrowed an example of my problem down to a specific pair of spheres, where an inflated version of the small sphere is used as a Boolean tool to remove a portion of the larger sphere. Note that this specific pair is problematic when I apply a 5 mm inflation and cut, for other values the problem surface may be elsewhere.

x = (0.122353, 0.155428)
y = (0.104232, 0.104217)
z = (0.761678, 0.799164)
R= (0.04, 0.01)
outerBox = geompy.MakeBox(0.1,0.1,0.75,0.2,0.15,0.85)

When I zoom in closely to the surface that can't be meshed I can see that there are some problematic elements. Ideally I could simply combine/merge all curves that are shorter than my mesh critical length.



I have had success with Salome's commands for removing extra edges in the isolated/two particle mesh.

poreRepair =  geompy.RemoveExtraEdges(poreSpace,True)
poreRepair =  geompy.RemoveExtraEdges(poreSpace,False)

However, if I apply them to the full case of hundreds of particles, then many of the legitimate edges are removed and the final result is pretty horrendous (the BREP file goes from 4 MB to 500 KB and can't be meshed properly). I feel like I could manually find the problematic pairs, then shift one of the particles by 0.5 mm just to avoid having the pairs of surfaces in the unlucky configuration. However, this would be quite tedious as I would like to automate the process in able to automate geometry generation/meshing/simulation/analysis.

Does anyone have any suggestions about how to clean up my geometry in Salome, either before or after it's exported as a BREP file?