RADIOSS: How can I fix penetration properly

Zeynep Aydin_22555

Hi,

I have a model of a composite passenger cell, where I want to simulate a crash analysis. I already set up a model with the material etc. It is a front crash with a rigid wall and self contact Type 7.In a RADIOSS course I heard that we should set Igap to 2. But now I ave a problem with penetrations. There were around 12k .I wanted to detect and fix them but after about an hour it stopped fixing the penetrations and now it says "collision detected". An now it looks like this..Maybe there is a huge mistake in my model, but I cant find it. Can someone please help me? I could also share my model.

Polyvios Romanidis

Hi Zeynep,

Igap = 2, is a good choice for crash applications, but there are many parameters that affect how the minimum gap is calculated and hence how penetrations detected.

In Igap = 2 formulation the gap is calculated based on the following type:

In this type Gap_min is a value set up by user or otherwise calculated by the minimum of average main shell elements thickness and the half of smallest side length.  I see that you have set a Gap_min, just be sure that this value make sense based on the Unit System you use, because that is a common mistake.  For example, if you use S.I. your Gap_Min is 1 m, a value that make no sense.

Fscale is also a value that you set and seems fine.  But, gs and gm are directly affected by the shell elements thickness of main and secondary sets.  Also, if brick elements are used, different definitions for these values are given.  It is not easy to say if you use brick or shell elements because of the great deformation in the photos you sent. 

Gap_max is an upper limit for the gap.

More detailed information for this calculation, you can find in the following link:

https://help.altair.com/hwsolvers/rad/topics/solvers/rad/inter_type7_starter_r.htm

As you can see, there is a great dependency between gap definition and penetration exist.  So, if you made a wrong set up of the parameters Gap_min, Gap_max and F_scale in accordance with the elements size and thickness, your model might detect many penetrations that have to be fixed and probably ended up in this situation that it shouldn't.

Additionally, you can let Radioss solver fix initial penetrations without fixing them from HyperMesh Environment.  That can be done with the Inacti parameter.  I see that you have set it equal to 5, but I will recommend you the value 6 instead.  Of course in order for that to work properly, you should set up the TYPE7 Interface correct.  But, after this you can proceed directly to the solution without fixing the penetrations in HyperMesh.

Anyway a picture of your model before the deformation caused by the penetration fixing tool will be helpful for more detailed examination of the model.

Polyvios

Zeynep Aydin_22555

Thank you a lot, Polyvios!