Learning Forum

[12OP12]

Hi, I am currently trying to perform a simple composite tube impact simulation in LS-DYNA (through Workbench) where I used ACP to model the composite tube and a mechanical model to model a rigid wall (shell) and rigid impactor. I gave the impactor an initial velocity and the wall a rigid constraint. In addition I used the keyword manager to define the contacts between wall and tube, and tube and impactor. However, whenever I run the simulation the deformation tracker shows the tube being pushed through the wall. I assumed this to be either a contact definition problem or an error due to initial overlap between the wall and the tube. I tried visualising the wall with the Thickness Shell display option but this does not show overlap. I have added screenshots for clarification.

42 % run:

50% run: 

settings:

 

On the other hand I also feel like the simulation takes very long to run. Can this be related to the previous problem or is there more to it/is this normal? Any help is more than welcome. 

 

[Raman Babu]

Please check whether displacement scaling is set to true scale instead of autoscale.

Please share the energy summary curve.

[12OP12]

Thank you for your response. I have added the energy result and I am looking at the results using the true scale.

However, looking at the results, the deformation looks far from being what you expect for a composite tube (almost more like a metal tube being plastically deformed) and the acceleration reaches almost 270g for an impactor moving at 12 m/s (5.3 kg)...

Also, what is the effect of setting the slave segment set and master segment type? Have used both these settings and part/part, results stay strange.

Finally to fully run the simulation, it takes more than two hours, this seems very long no? 

Thank you for your help and time.

 

[Raman Babu]

If you choose segment type, it will generate a segment of select object and use it for contact detection.

If you want to run your model faster, you can try running the case with a higher number of cores. You can also use mass scaling to reduce the simulation.