Quad Remesh Thoughts

Firstly, the quad-remesh seems to work works pretty well. The symmetry axis selection hinting is good. I didn’t try the guide curve options yet.

I’ve done some CFD and FEM analyses. Meshing is usually the nexus of most of the problems and torment in doing fluid/mechanical studies.

The quad remesh looks like such a good start, that it’s tempting to want more in the future…

At some point in the future, can a quad-mesh be made directly from a NURBS object, without the added approximation of the initial triangle/quad approximation?

At some point in the future, if Rhino 3D could make 3D structured meshes from the quads, in a convenient way, it would be an advantage in plug-in/solvers.

There would also be a fun little added bonus: that X-Ray view could be made to really work, by volumetric visualization. This is a large part of why Paraview rocks for visualization.

[I love modeling and designing things in Rhino 3D because it’s hella-wicked™ fast for making subsequent generations and iterations of a device. Though, I cheat on Rhino for solving and pre-processing. On the side, I have used Salome Mecha and Freecad for preprocessing and solving with OpenFoam and Code-Saturne CFD, and Calculix, and Code-Aster FEM. I’ve done some modeling in FreeCad, and it makes me appreciate Rhino. I had so much hope for Butterfly CFD, but it hasn’t been updated for BlueCFD, but to be fair, they removed something in BlueCFD. Still, it is centered on buildings and not vehicles with sliding walls and moving grounds. Looking at the code, I reported a few things in Butterfly that can be changed in the code to allow the newer version, (but things have been stacking up for me in a bad way.) I was interested in Scan-and-Solve, but it doesn’t seem like there’s a V7 version, yet. Sigh.]

you can quad remesh a nurbs part in the current release.

can you clarify what you mean here?

Maybe Brenda is looking for Topology Optimization???

Well, from a bystander’s prospective, accuracy-wise it seems wasteful to make a triangle mesh–and then create a quad mesh from it. In other words, to reinterpret it.

Could not a quad-mesh be created directly from the original NURBS?

I am not sure any of this is helpful, but meshing takes so many forms. There are meshing plugins for Rhino that range from the more modern automated meshing to the older, face based semi-manual gridding.

If CFD is what you are interested in there are meshers included in:

  1. Swift | Food4Rhino
  2. Semerics CFD has a series of Rhino based CFD applications
  3. RhinoCFD

There are also specific applications:

https://www.itascacg.com/software/cad-based-meshing-capabilities

Many plugins include some level of meshing needed:

https://www.intact-solutions.com/rev/scan-solve/download-scansolve-pro-for-rhino-6/

Quad meshing, perhaps structured meshing–without a triangle go-between would be a good foundation for many analytical disciplines.

In addition to the accuracy difference the advantage would be: if you created a quad mesh from a triangle mesh, you don’t have as much information about its generative intent. In other words, if you made a cylinder or sphere or something, it’s easier to sort that out from the original NURBs surfaces—than a confused mass of assorted triangles.

1 Like