Snap Fit Analysis Issues

[grantchapman_glassboard]

Hello all, 

I am attempting to analyze a snap fit in Nastran and running into some issues. I've Simplified the problem to a near 2D representation of the system.

For now I am modeling this with the Autodesk Material Nylon with linear material models. 

I have separation contacts between 3 parts, one moving part driven by an enforced motion load. The separation contacts are symmetric and manually set. In the nonlinear static version of the set up I have the stiffness factor as 1, 0 friction, 0 penetration surface offset and max activation distance checked by nothing in the field. No advanced settings set.  The same contacts are used in the Explicit version. 

The enforced motion is in one direction but relatively high (17mm). 

In the nonlinear set up options I have it set for 100 iterations. After much adjustment I have finally gotten in to solve to the point right before the snap through. However, I have yet to have it solve past that point. I keep running into the error ER1004. Which is basically non-convergence. 

I decided to try the Explicit solver as it should be better at this style of motion. I however, have not had any success in the separation contacts working correctly as I always have the moving body penetrate the stationary ones. 

Any advice would be greatly appreciated. 

Files attached. 

Best,

Grant Chapman 

[Anonymous · Accepted solution]

I activated your "Analysis 1" and deleted the contact "[M] Surface Contact 2", because I'm seeing no reason for it. Than I added the contact face of Body1 to the master objects of "[M] Surface Contact 1", so this contact can also be used after the snap through.

I started the analysis and everthing worked fine.

I did not look at the other analysis definitions. I guess the probem is the same.

[grantchapman_glassboard]

Sorry for the late reply! Your fix solved it on my end too for this simplified version, I don't know how I hadn't caught that error. Thank you so much for the quick help. 

To add other lessons learned here when a worked the simulation back up the path to more complexity I ended up tweaking the following settings.

Maximum bisections -> higher (~200) 

Tried to use convergence on distance only initially to push through some problematic solves

updated the Stiffness update setting to Auto

Also changing the separation contacts from 1.0 ->.2 helped with convergence and then could be walked back to a stiff value if it was causing inaccuracies. 

Again thank you for the help,

Grant 

[admaiora]

Hi @Anonymous 

When using a finer mesh (mesh control for example) the contacts seems to be lost.

Any idea how to mantain a finer mesh and contacts and converged results?

Thanks!