snap through / snap back problem with NASTRAN

[stressboy]

Hello,I am modeling tape spring with NASTRAN.
It is a non linear problem due to geometry with snap through behaviour.
Usually the solution is to use arc lengh method (I use Crisfield). But with this one, I have ping-pong effect.
If somebody has an idea to solve it ...please
thanks

 

[fkmeyers]

What version of Nastran are you using?

Normally for snap-through you will need arc-length. If the snap-through effect is mild, sometimes you can get away without it. I would try using RIKS or MRIKS on the NLPCI entry to see if that helps. Also, try increasing the number of increments on the NLPARM. You may need to increase the MAX INCR on the NLPCI as well.

Does the loading appear to reverse? Is there contact involved?

 

[stressboy]

I am using version 2005.
I have already test different arc length methods with load increment. I can´t have the post buckling behaviour.

I test also SOL600 with displacement increment. In the best case, I have a ping-pong effect before a jump. I have the global curve with several load reverse.

In the papers, I found in internet about tape spring, it is always Abaqus which is used with a stabilization fonction (damp is added to stabilize the unstability).