Static linear Job run error


#1

https://www.simscale.com/workbench/?pid=6522356165865990968#tab_1-0
“small cuts linear” simulation
I’ve defined fixed support constraint, and rotating motion to the other edge of the live hinge.

I get an error when trying to model.


#2

Hi @gkotelnytskyy and thanks for reaching out to us!

Can you explain what the project is all about and what you want to accomplish first? Makes it easier for the @PowerUsers_FEA and me to fully grasp the concept of your simulation.

Thanks and all the best!

Jousef


#3

That’s a model of a living hinge

First I need to check how many cuts I need to fold it, at the later stage I’m going to load it vertically, – it’s to be used as a leg in a foldable drum stand.


#4

I’ve saved viewer state for a successfull run of a model with bigger cuts.


#5

Any thoughts?


#6

Hi @gkotelnytskyy and sorry for the late response!

I just copied the project so far and still need to have a closer look at it. Tagging @ahmedhussain18 as well as @rszoeke here who might give you some tips in the meantime.

Best,

Jousef


#7

Hi @gkotelnytskyy and thanks @jousefm for tagging me in.

For an initial try you have a big mesh. I will suggest to start with first order smaller mesh just for a check and then go for a second order one. This will help you to not waste your valuable core hours just in finding the right settings for your simulation.

With second order mesh, I would like to further suggest you to refine only valuable areas. For example all the curvatures using local mesh refinements. And when using second order mesh, change solver to MUMPS and select a reasonable core combination under Simulation Setup. For example, 32/8.

All the best,
Ahmed


#8

Initial results with 1st order moderate mesh, 45 degree rotation. Of course not great but can be better with second order mesh. Took approx. 9 to 10 min. to simulate on MUMPS with 32/8 combination.


#10

Hi @gkotelnytskyy,
as @ahmedhussain18 mentioned, this is a memory problem. Your model has more than 2 million nodes on a 2nd order mesh. For this model you would need a 32-core machine, but you used only 4. Same problem happened in the nonlinear case.

We’ll check why no appropriate error message was displayed and try to fix it.

In the meantime I created a new run for the linear case using MUMPS on a 32 core machine with 2 MPI runs (each on 16 OpenMP cores). You find some more information regarding these two settings in this post: FEA - Oscillating Metal Block with Rubber

You can check the project here: https://www.simscale.com/projects/rszoeke/live_hinge/

Best,
Richard