2025R2 Mesher Curvature error

[kdjones]

I teach an aero class that uses CFX. Late last week our student PCs got updated from 2025R1 to R2, and now we are getting serious issues with the mesher. We import a simple 2D volume to calculate flow around an airfoil, generated in SolidWorks. The airfoil surface is a spline, and while it looks fine, the mesher appears to ignore the Capture Curvature angle setting, or rather, it might be applying the setting, but to a shape that uses the default 18 degree setting. In the attached screenshot, the green line is the airfoil LE, but the mesh fails to follow it. If I add edge sizing, I can improve it, but the cell sizing around the LE is inconsistent, with significant jumps in cell size from cell to cell.

Has anyone else experienced this? I opened a project that was created in R1, and remeshing inside that project doesn't appear to suffer the same issue, so perhaps it has something to do with the geometry import functionality. I tried importing both sldprt files and parasolid files, and both have the same issue.

 

If I import the geometry into a CFX block in R1, save it, open it in R2, when I try to run the mesher, I get the results above. If I go back to R1, and add the curvature settings and edge sizing around the perimeter, and then once again save it, open it in R2, then the mesh looks exactly as it did in R1. However, if I modify it, say to add sweep or inflation, when I generate a mesh, the mangled surface is back.

It seems like we will have to roll back to R1, but getting this done on the 30+ PCs in the lab is a major effort for our IT staff.

 

[Anilkumar]

Hi,

Thanks for reaching to Ansys Learning Forum support. May I know if the geometry is 2D here or 3D(2.5D)?

How are you opening the geometry to Meshing, are you opening it first in Discovery and then to Meshing? OR just attaching the CAD to Geometry tab and opening it within Meshing?

Can you try below settings and see if that improves the mesh capturing on geometry > Set the curvature normal angle to 12-16 deg, Element size to 0.1m from 0.5m currently and see if that gives better results.

Yes, as you described it seems to be geometry faceting issue, so you may also check for the FINE facetting from the geometry properties tab(Import facet quality) in WB as shown below.

Thanks,

Anilkumar

[kdjones]

It is a CFX problem, so inherently 3D, although we create a 1mm thick slice in the z-direction, use sweep with 1-division, and use symmetry BCs on the z-faces, so essentially 2D.

I import the geometry directly into the geometry cell. I've used both solid-parts and parasolids, and get the same results either way.

I should point out that I've been using the same problem for more than a decade, and this is the first time it has failed. Further, it works great in 2025R1, but fails in 2025R2. Here is a simple example generated in R1, with only the curvature controls set up, no sweep, edge sizing or inflation:

Here is the same project opened, but not modified in R2:

Here is the mesh after making a miniscule change to growth rate and remeshing in R2:

There is something weird going on with curvature controls in R2. Perhaps it's related to the spline edge of the airfoil, but this hasn't created a problem for the past decade, so I don't understand what changed.

 

[kdjones]

I checked the properties, and they are set as in your example. Interestingly, if I set max size to 0.1m and leave curvature angle high, the mesh appears to follow the surface better. It seems good up to around 0.25m. At 0.3m it fails. However, If I again reduce the curvature angle to 1 degree, it fails, even with max size of 0.1m. 

[Anilkumar]

Hi,

Thank you for the update and testing out the curvature normal angle sizings to get improved mesh. Let me investigate this and revert back to you.

Thank you,

Anilkumar