topic Re: Not converge for prescribed translation in Analysis

Not converge for prescribed translation

fstopelli-2 — Fri, 21 Jul 2023 14:38:04 GMT

I am using Creo Simulate Release 9.0 and Datecode9.0.0.0

Not converge for prescribed translation

Here are the errors that I faced
** Warning: Convergence was not obtained because the maximum
polynomial order of 4 was reached.

Re: Not converge for prescribed translation

fstopelli-2 — Fri, 21 Jul 2023 14:48:32 GMT

I impose the deflection "f" to close the gap, but the analysis not converged

Can you help me?
Thank you

Re: Not converge for prescribed translation

ChrisKaswer — Fri, 21 Jul 2023 15:08:49 GMT

It would seem your multi-pass (MPA) analysis settings were selected to converge on a measure with the maximum polynomial set to 4 and the convergence percentage identified was not met. I would recommend running the SPA (Single-Pass Adaptive) first to see how it solves, and then if you require convergence to be demonstrated, re-run your MPA study but increase the maximum polynomial order to 7 or 8 (even 9 since it looks like a fairly simple, single part analysis).

SPA analysis is actually very robust, and should solve more quickly as well. Unless you require demonstration of convergence, SPA would be the approach to use more often.

I hope this helps!

Re: Not converge for prescribed translation

SweetPeasHub — Fri, 21 Jul 2023 17:08:23 GMT

This looks like it might have a singularity requiring isolate for exclusion. Can you give a few more details on how you setup the enforced displacement? Is it on a point or edge? Is the model 2D? How does this relate to an actual physical test? Could you actually do this in a test lab? If so how would that be done?

Re: Not converge for prescribed translation

fstopelli-2 — Mon, 24 Jul 2023 06:58:53 GMT

The enforced displacement is a section, it a 3d model I have not a phisical test

Re: Not converge for prescribed translation

SweetPeasHub — Tue, 25 Jul 2023 15:09:19 GMT

Suggested technique to enforce the gap closed as you indicate. Use at own risk.

This a technique when enforcing rotation on 3D because 3D elements only have translational DOF. (degrees of freedom)

Keep in mind strain results are logarithmic for large displacement.
https://support.ptc.com/help/creo/creo_pma/r10.0/usascii/index.html#page/simulate/simulate/lg_deform_strainmeas.html

1. Coordinate system at center of rotation

2. Point on coordinate system

3. rigid link from moving section to Point

4. Simple 2 point spring from Point to anywhere. (very low stiffness)

5. Constraint on the point to rotate the section the amount needed to close the gap. (may need to adjust the amount, radians or degrees)

6. Enable large displacements and enough steps to allow convergence (I show 21 steps, or 5% increments)

Re: Not converge for prescribed translation

hadardor — Thu, 03 Aug 2023 06:14:52 GMT

Hello,

Couple of suggestion :

Run analysis in Multi-Pass Adaptive analysis
Increase maximum polynomial order
Try by changing the model accuracy
Increase the mesh density of the model and more elements in high stress gradient regions in this model.
- Adding some mesh controls.
- Change the AutoGEM settings so that the face and edges minimum angles must be larger.

Re: Not converge for prescribed translation

SweetPeasHub — Mon, 07 Aug 2023 17:40:47 GMT

How is this the solution?

Re: Not converge for prescribed translation

SB_11013836 — Tue, 22 Apr 2025 09:56:27 GMT

Hello, i didn't understand the need of the aplication of the simple spring in this case, since you made the connection between the point and the section with the rigid link, what was the need of the spring?

i had another question, about the coordinate system, what would happen if instead of the coordinate system and then the point on top of that, there was only the point ? why is the coordinate system needed?