Modal analysis solver error help

[grigore.andrei1234]

Hello!

I am trying to run a modal analysis of a pump shaft together with the rotors and magnetic drive coupling. In all, there are 6 components present in the simulation.

After setting up my simulation and running it on the cloud, it runs for about 30 minutes showing progress and then i get the following error:

Mesh Details:
Error: Solver Error
Error: An error occurred while solving the model.
Job has finished with errors. The Cloud Credits for this job will be refunded. Receiving results.
1846967 Tetrahedra : 100.0% of elements (100.0% of volume)
Face Angle min: 3.67, max : 172
Dihedral Angle min: 1.12, max : 178
Worst shape ratio : 96.7 on element 1256465
Worst aspect ratio : 12.6 on element 176700, shortest edge: 1.8e-05, longest : 0.00469
Lowest collapse ratio : 0.0128 on element 1256465
Worst Jacobian ratio : 3.63 on element 175711
Error: 8
Error: Solver Error
Error: An error occurred while solving the model.
Error: Solver Error
Error: An error occurred while solving the model.

Solver output file is attached to this post.

The model is attached in the following link:

https://a360.co/3rvAozi

What is the reason for this error?  I suppose it is somehow related with the Impeller component as running the simulation without that one component runs just fine. I also tried replacing the complex geometry of the impeller with a simple disk of the same mass distribution and material and it still fails. So maybe not the component geometry itself is the problem but the contacts it generates?

Thank you in advance!

[mcmillr]

Andrei, I obtained your file from the link in the other forum, and when I open it I see the study cannot be solved because there is an invalid material on Component4:1. There are also interferences in the model which could potentially cause issues. Finally, Component4:1 won't mesh with default mesh settings. You appear to be applying local mesh refinement to get around this, but you also have adaptive mesh refinement enabled. I wonder if the mesh is becoming too fine during refinement and causing a memory error. The solver output you attached shows a system error which could be memory related. Maybe simplifiying the impeller geometry or avoiding adaptive meshing so you can limit the mesh size to something reasonable could get it to solve. I will experiment with your model a bit more to see what might work best.

[mcmillr · Accepted solution]

Andrei, I suspect there is an issue with the geometry of the impeller part. If I set the material to Stainless Steel 316L so I am able to solve, and I also set the overall mesh size to the default 10%, and remove your local mesh controls, then place a 1mm local mesh control on the impeller, and disable adaptive mesh refinement, I get a successful solve. I see that the impeller is coming apart, and when I inspect the geometry I see the 'fins' aren't actually connected to the rest of the impeller on one side. There is zero distance which is the reason there are self-intersection meshing errors if you don't use a small enough mesh size. I would try to fix the impeller geometry to ensure it matches 'real life' otherwise I don't think you will get a correct result even if the solve completes.

[grigore.andrei1234]

Hi Rob!

Thanks, that was apparently the issue. I did not see that the impeller blades where not properly connected with the shroud. I fixed the geometry and now the impeller meshes without additional refinement and the modal analysis runs just fine.

Thank you!