Prediction of failure in loaded thin steel plates

Hello everyone,

I am curretly working on predicting the failure in thin loaded plates. The plates are impacted by a hemispherical rigid cylinders at a low speed (4.53 m/s <--> Free fall from 1.05 meter) and are perforated experimentally at their center.

I tried to simulate failure with ls-dyna on a shell based model ( the sheet is very thin : 1 mm thickness Vs  130 x 130 mm ), but it seems that the failure strain measured on the tensile tests is underestimated ( the curves after the begining if necking are false).

Moreover, when I tried simulation whitout a failure criteria and calculted the stress triaxiality , it seems  the stress triaxiality is changing through the impact ( negative and positive values), it is possible? I read that stress triaxiality of plates under uniaxial loading is usually positive and around 1/3.

Finally, the experimental impact tests showed that indeed 1.05 meter is the critical height for perforation ( half of the specimens were perforated), while in a free fall of 1.00 m, all of the plates were safe. So I have a critical velocity for perforation (4.53 m/s) and a safe velocity with no failure (4.42 m/s).

Attached some pictures of the plate, the projectile and the support.

Can you give me please advices?

Best regards,

M.