I am analysing a gold ball (very small geometry, lots of dimples, rounds and edges) andn very early in the analysis, windows gives me an error that the Sim CFD Mesher process has crashed.
Is there a way around this?
Thanks.
I accidentally mentioned the edge merging tool had a slider, it doesnt. In response to your questions:
The Creo file is here: http://a360.co/1oJTMQY.
1. I created the air volume in Creo and subtracted the ball from it.
2. I converted it (the air volume) to Absolute accuracy in Creo before launching into Sim CFD (6.0 e-5).
3. When in SimCFD, 1 additional part was generated (the ball).
4. The edge merging tool mentioned that there are 340 edges to be merged, and I did so.
5. When meshing, I used the default Autosize only.
I can recreate the meshing error message you are reporting during mesh adaptation if, after launching from Creo, I simply attempt to mesh the golf ball itself (no edge merging or external volume added) using a fine enough mesh size (0.1 under manual meshing applied to the model edges only). There's an issue with one of the dimples as is shown in the attached images. Not only do model edges not coincide with the dimple boundaries (image 1) but there are topological problems involving multiple edges (those shown with endpoints highlighted in red) indicating that these edges do not meet within tolerance (image 2). The edge highlighted in red in the second image is where meshing is failing. The message handler is reporting the wrong type of message (thanks for uncovering this!), but the model problem giving rise to this is genuine. At this point I'm not sure whether the issue exists within Creo itself or is created by Granite during launching. One thing you might try is to remove this particular dimple and have another go at it.
I did also try another option involving the activation of a second geoemtric kernel in SimCFD (i.e. the launch is completed using another kernel rather than Granite). This secondary kernel can be enabled by setting the flags 'AsmEnabled' to a value of '1', and 'AsmHealing' to a value of '2' in a SimCFD session prior to launching. The golf ball is import cleanly and a repeat of the attempt described above (albeit with a mesh size of 0.0001) is successful. I expect adaptation will be as well.
Do we have permisison to send this model file to PTC? It's likely that this model exposes an issue with the Granite kernel.
Can't find what you're looking for? Ask the community or share your knowledge.