[Hamps]

Hi there,

Long-time lurker and finally a member here.

I'm experiencing some perplexing results when trying to generate an unstructured mesh for an intermediary part of my geometry for my master's thesis in combustion. The geometry is a rectangular combustor featuring an inlet and an outlet pipe. I have created structured hex meshes for the inlet pipe and most of the combustor body. However, the region where the inlet connects features several fillets and therefore I decided to make an unstructured mesh for that part, which I can then merge with the two structured parts. I have read Simon's guide here, the Sharcnet manual and watched several Youtube videos on the way to go about this.

My working procedure has been as follows:
1. Generate octree-mesh
2. Remove volume elements
3. Smooth mesh
4. Generate Delaunay mesh with Prism Layers

However, trying out some different combinations of smoother and prism generator, I seem to receive better quality metrics when not using the Laplacian Smoother, coupled with the Pre-Inflation (Fluent) Prism generation. Reading other forum posts and guides on the internet seems to suggest that Laplacian smoothing should produce better results. Am I missing something?

I have uploaded a picture of different mesh metrics as well as what the mesh looks like (roughly).

Thanks in advance,
H