Dynamic time-history analysis using gap elements

[mlee]

Dear all,

I've come across a strange behaviour when using the gap element in a dynamic time-history analysis.

I want to model the collision of a point mass on a rigid wall.

The lumped mass is attached to node I. To simulate the deformation of the impacting body a spring is also attached. Since there are only contact forces, a gap element is implemented between the spring and the support node.

I'm testing two different link elements. One is linear elastic and works perfectly fine. However, if I use a bilinear link element, there is an abortion inmidst the calculation with the error message "Unable to apply the next load step."
The problem occurs even if the spring stays within the elastic range.

Here are two models, same setup, just two different link elements (linear elastic and bilinear).
https://www.dropbox.com/sh/29g0d93jd9oi ... em73a?dl=0

In the end, the system would consist of multiple masses and link elements (in series and in parallel). Examples are simplified for presentation of the main problem.

Is there anything special to consider when using non-linear link elements with gap/hook?

Thank you in advance for any help!

[seismosoft]

Hi,
Some comments on your models:
1) The plastic model that you sent converges until the end of the analysis without problems
2) you need to fix the II node, if you want to simulate the correctly the pulse
3) The gap element is not activated since it is connected to the II node rather than the I node (where the force is applied), i.e. you connect two nodes that should be fixed.
4) The gap element limits are way larger than the vibration max deformation, 0.10 vs. 0.0055. Decrease them to a lower limit

Regards,
Seismosoft Support