Hi, I am trying to create a Hex Dominant Parametric Mesh for a CFD analysis of an engine head intake port, and after several iterations carefully adjusting parameters to achieve a proper mesh, the result I am getting has many ripples or ondulations at the surface boundaries of the domain.
Could someone suggest any ideas to reduce or prevent these ripples?
Hi Barry, thanks for your reply.
Yes, there are several mesh iterations there. Sorry.
A good example where the ripples are happening is the last one in the list (“Copy of Refined PolishedChan_HexDom”).
As a second question (not sure if it would be related to the ripples), I also wanted to point out that when I check this mesh iwith slices in the post-processor (saved state: “Copy of Refined polished chan - Mesh Check”), the ripples do not seem to be causing serious problems to the inner portion of the mesh, however, I do note that there is no expansion layer being build, although I am requesting it with a layer refinement step, but instead the “refinements” obtained at the wall boundaries seem to be following the feature refinement at edge step that I also have added.
I would recommend setting the min surface refinements to 2 while setting the feature refinement to 0 m and level 3 to see if we can get better mesh conformation to remove the ripples. The valve refinements are fine however so there is no need to adjust them. You can remove the region refinement as you have not created a geometry primitive so that won’t do anything. You may also want to increase the surface refinement levels of the selection shown below to a slightly higher value, maybe min 3 and max 4 to ensure that it conforms better.
Also do ensure you are not overlapping refinements like you did in the mesh you pointed me to. Refer to the “refine-surfaces” where you have selected your entire geometry while also adding refinements for other surfaces. This may cause issues. You want to ensure that you perform a single refinement for any particular surface so as to not also reduce the level of control you have over the mesh.
I applied your recommendations and the result seems much better!! In mesh “220.127.116.11. refined polishd_HexDom_Param” the ripples have significantly been reduced vs. mesh “18.104.22.168. refined polishd_HexDom_Param - Mesh Check”.
Thanks for pointing out the fact that I had an overlap of two surface refinements. This could have been causing some of the issues as well.
Also, thanks for the link on the topic of Y+
Am I correct to assume that it didn’t seem necessary to relax the quality check “Min Cell Volume” to -1e-30, as with the recommendations above it seemed that Layer was built (I think, based on picture below)?.. or is this not the result of a “layer refinement” but just the result of having a tighter refinement level at the port channel wall and valve surfaces?
If you perform the preliminary check as stated in the document, you will notice the following figure below. This means that no layers are generated and you will need to perform some troubleshooting to generate the layers properly.
That is correct. As mentioned earlier, the layers are not present. Instead it is just the higher refinements at the port channel and valve surfaces.