Why does the Block Lanczos solver either not converge or
give erratic results (e.g., frequencies change with element size)
for this piezoelectric model?
There are two independent volt domains used as piezoelectric
actuators on the top and bottom of a solid185 beam. Their terminals
have been coupled in volt, but no voltage constraints have been applied.
One of the coupled sets in each domain should be converted to a Volt=0
surface [NSEL,,,, D,ALL,VOLT,0]