cyclic analysis of cantilever column

[trin]

Hello,
I am trying to simulate a quasi-static test of a RC cantilever column which is subjected to cyclic displacement at its top. The cantilever model has a U-shaped fiber section and I modelled the reinforcement using Steel02 and the concrete with Concrete01 material model. I modelled the cantilever using one or several nonlinearBeamColumn elements.

During the analysis I am encountering problems regarding convergence. In particular I am getting 2 error messages:
- failed to get compatible element forces and displacements
- failed to convergence

I tried different algorithms (Newton, Newton –initial, Broydon, …), different convergence tests (NormUnbalance, NormDispIncr), different settings of the convergence criteria but the only model I got to run was the one where the cantilever was modelled with one element and only 3 integration points. When I reduced the axial load the problem seemed more stable.

Any ideas what might cause these troubles? I was also wondering whether there is a limit to the integration points one can specify. Is it in general better to model a cantilever with just one element but specifying a larger number of integration points or can/should several elements be used?

Thanks a lot for any kind of help and ideas!!
Trin