It sounds like the Newton-Raphson residual is increasing to a value that is too large to record, might be an ill conditioned problem? Can you plot the residual force distribution from the previous iterations to try to see where the source of trouble might be in the model?
You are navigating away from the AIS Discovery experience