Meshing of solid part created with Sweep tool?

[19757748]

Hi all,
I am currently using PrePoMax to model a rubber-and-steel-cord conveyor belt featuring a change in elevation – i.e. the belt starts with a flat section of a certain length then slopes down and then continues “flat” (parallel to first section) again after a certain distance. A screenshot of a prototype model is included below (picture 1).

In previous iterations the model was completely flat, and so I was able to use Extrude Mesh items to mesh the cords and rubber parts correctly (picture 2). However, now that I have modified the model and included the sloped section, using the same mesh parameters now returns an error (picture 3). How should I go about meshing my new model? Any help is appreciated.

TIA

Picture 1: Model with Sloped Section

[19757748]

Picture 2: Extrude Mesh Parameters (copied from previous model designs)

[19757748]

Picture 3a: Error Message Generated

[19757748]

Picture 3b: Console Output

[arnie]

Which version of PPM are you using ?

Have you tried Sweep Mesh option ?

[19757748]

Hi Arnie,

I’m running v2.1.5; I didn’t realise there was a new version available! I’ll give that a try and see if it solves my issue.

Thanks

[19757748]

@arnie I’ve just updated to 2.3.0 and re-ran my tests; I was able to create the mesh without issue, but now my analysis is failing with results. The console is also saying that “an element set containing elements with nonpositive jacobian determinant was created.” Any ideas on how to fix it? TIA

[FEAnalyst]

Sweep meshing is available since v2.3.0 stable. Extrusion may also work if there are no cutouts and you select the side faces. All algorithms have certain requirements/limitations described in the documentation. Please share the file if you want more detailed advice.

[FEAnalyst]

Can you share the .pmx file ? If you used sweep algorithm, it may sometimes lead to invalid meshes due to the lack of geometry constraints.

[19757748]

Hi, please find attached the .pmx file for the model I have been testing.
small_scale_1.pmx (5.7 MB)

I did notice that when I created the contact items between the rubber and cords, that for some of the cords the contact appears to also include the end faces? I used the Search Contact Pairs tool in the Interaction tab to create all contact items.

[Gunnar]

I see you only used one mesh item per part: sweep. You should at least always use “meshing parameters” in addition to be able to change the mesh size (your element size needs to be smaller to generate less distored elements over the crooss-section), then you can use “mesh refinement” for the edges in sweep direction to reduce the number of elements (because mesh refinement can also be use to “coarse” the mesh. But that works only when gmsh is involved as mesher).

example

[19757748]

Hi @Gunnar, I’ve done as you suggested and was able to somewhat recreate your mesh above. However, I’m still getting the same nonpositive Jacobian error as before. Do you have any advice on how to solve it? PMX file attached.

small_scale_2.pmx (4.3 MB)

[FEAnalyst]

This mesh is too coarse. Add meshing parameters for the wires and refine them. This should eliminate negative Jacobians.

[Gunnar]

You need to reduce the element size, i used max. 2mm and these Sweep settings (results in more and nice hexa-elements):

[19757748]

Thanks @Gunnar @FEAnalyst , I have increased the fineness of the mesh and was able to obtain results from the simulation. Thanks for your help!

[synt]

interesting models with extra edge exist but swept and quasi-structured quad algorithm succeed.