How to prevent rigid body penetration; Rigid body compression deformable body

[LM290]

This is my first post here but have used it a lot to help with my project.

First let me explain what i am trying to achieve here. I am physically testing half an alloy wheel under a compressive load to validate my Abaqus model. The load is dependant on the displacement of the striker by 20mm, stain gauges will be attached to the back of the vertical spoke to detect the compression stress. This is what will be compared to my Abaqus model.

I have modeled 1/4 of the wheel which will be in symm on one face to replicate 1/2 a wheel, the other face will be encastre to represent the floor. I have create a rigid part to represent the striker applying the compressive force which has a reference point with a boundary condition in the U2 direction of the displacement i require. I have given the whole outside surface of the wheel (abaqus detected) and the bottom surface of the rigid body a surface to surface contact interaction with small sliding. The rigid body being the master and the wheel (deformable) being the slave. After the analysis has completed it shows the wheel penetration the rigid body which is what i do not want. I have attached some photographs and a video to show what i mean, hopefully someone has an answer for me as i have spent a long time trying to figur it out as well as many hour computational time to try my modifications to no avail.

http://files.engineering.com/getfil...4a5&file=CapturFiles-Nov-03-2012_12.33.32.png

http://files.engineering.com/getfil...-ae5b-5710321531d2&file=wheel_compression.avi

 

[IceBreakerSours]

Why bother creating a rigid part? Try an analytical rigid surface.

http://www.eng-tips.com/faqs.cfm?fid=376

http://www.eng-tips.com/faqs.cfm?fid=1083

 

[LM290]

I shall try try too, i originally thought of using an analytical surface but initially it started off as a 3d part then i reduced it to a surface part. I suppose now that it is essentially a surface then using analytical surface would make sense.

Should i be using a self contact interaction too ? i assume that in the physical test that the wheel may crumple in on itself but im still unsure if this will happen. If this is not the case and i use self contact will it void the analysis if there is no self contact or will abaqus just remove the interaction automatically ?

I expect someone to say why dont you do the physical test first but i am having to try do both at the same time as it is a final year project an time is limited so have to juggle both tasks as i wait for time on the required equipment.