Difference between "Leaf-level Mesh Part" and "A CAD Part created in 3D-CAD"?
 

Hello everybody,

Can someone please describe the difference between a "Leaf-level Mesh Part" and "A CAD Part created in 3D-CAD"?

Thanks a lot in advance

Renee

A mesh part is a discrete surface - an assembly of triangles that describes some volume. It is not CAD.

A 3D-CAD part is a solid CAD geometry built from 3D-CAD. It contains analytic CAD data.

Do you think there might be a difference for StarCCM+ by using either the leaf-mesh or the CAD part? Do you have an idea where starCCM might have problems?
Thanks a lot for your answer.

Using either datatype shouldn't generate any errors. However they may give different answers.

For the mesh part, if the initial mesh is too coarse, then even if you specify smaller sizes for the remeshed surface, you will not get any more detail resolved. A nice analogy to use is a circle.

To the computer, a circle is an assembly of lines. As more lines are used, the shape more resembles a circle.

A mesh part is a circle with N lines, where N is a constant value. The length of each line is also constant, let's call that X. If you create a remeshed surface that has sizes such that edge lengths smaller than X exist, N does not increase. Therefore although the mesh is finer, the geometry is no more resolved than it was with a smaller mesh.

A CAD part is a circle described by an equation (x^2+y^2=r^2). To make a geometry out of it for computation we create a tessellation - a value N. However N is not a constant. We can specify a greater value for N to resolve the CAD as small as we want. Now if our mesh edges fall below X, we can always choose a greater value of N.

Furthermore, if we use CAD, we enable features like Project to CAD, where our final mesh has vertices that are projected to the known curve (the equation of the circle, in this case). If that curve does not exist (we don't have CAD), then that feature will not work.

Does that make sense?

Thanks a lot for your reply. It helped, yes :)

2 years later, still a useful reply. Thanks.

Firstly, is it a good idea to unite two parts 1) 3D CAD Part 2) Leaf Level CAD Part which as a result gives you a mesh part that is going to be used for mesh generation later on?


Secondly, when importing any of these parts as a surface mesh, would it matter if tesselation density is kept fine or it's good to avoid any issues?

It's almost always better to use CAD part whenever it is available and then fall back to discrete when it fails. Both are usable in meshing anyway. If CAD fails, then there is a bug in CAD translation (either in the CAD software, or in Star's converter). These bugs don't live for long, because they're very problematic. Once the bug is fixed in the next product release cycles, you should go back to CAD again until the next bug is found.

A part comes with a tesselation. When you import it, you are re-tesselating. To know whether or not you need to use a fine tesselation requires you to know whether the part was generated already using a fine tesselation or not. Most parts are generated with a coarse or medium tesselation. You selecting "fine" does not improve the quality of the part beyond the initial tesselation and only works to bloat your memory usage. Pretty much, you should already know if fine is required for your problem. If not, stick with medium or coarse.

I used medium tessellation when creating a CAD part as it was by default provided in CCM+.

To be specific, when I import the the geometry and domain as a surface mesh I have them as a CAD part and leaf level CAD part respectively. In order to create the mesh, I unite them. My only concern was to know if it will result in some issues in the future during mesh generation if I am uniting two different part types