Hello Zhao,
Try to plot all internal, kinetic, hourglass, and sliding energies on the same graph as follows:
In this case, I am using fully integrated elements and so hourglass energy is zero. Also, without friction, sliding energy should be zero, or less than 10% of internal energy. In the case above, the sliding energy is zero.
Have you tried with shell elform=-16 instead of 2 for the blank?
You may want to increase the NIP to have more integration points through the thickness of the shell.
At last, if contact energy is large, this could be the problem. You can post-process the contact force in the RCFORC database and compare the different cases. Is the contact force much larger when the wheel speed is increased? If so, then the contact may be causing issues. Maybe the speed of the wheel is so fast that some nodes on the blank penetrate the wheel. The speed of the wheel may be too high and the penetration is not detected. Then, the contact force applies a force proportional to the penetration distance times contact stiffness (Force = K * Dpene). This could lead to large forces being generated. You can fix this problem by decreasing TSSFAC on *CONTROL_TIMESTEP to reduce the timestep so that the nodes in contact don't penetrate too much. Another option is to reduce the bucket sort frequency number. See the LS-DYNA User manual for more information on bucket sort frequency. The bucket sort setting will change based on SMP or MPP and SOFT=0, 1 or SOFT=2.
Note that changing the contact type and the SOFT option may solve this problem. Which contact are you using? Could you post a screenshot of your contact card with all options? The contact between the wheel and the blank.
Reno.
You are navigating away from the AIS Discovery experience