I have a model with some contact surfaces. The problem is that 20% of the times I try to solve it, Calculix does not converge and maybe sometimes it will produce multiple result data sets. Is this something expected? Each time I feed the same .inp file in Calculix so probably it is something randomized in the solvers inside Calculix. Has anybody seen this also?
I use Calculix 2.20 and I am using the default solver which is PastiX if I am correct.
By improving convergence you mean to change the covergence criteria?
It would be best to try with other solvers, this one is still relatively new.
This rarely helps. What I mean is changing the model in such a way that it converges better. For example, using displacement control instead of load control.