2 zeroLength springs in series, convergence problem

[vprakhov]

Hello everyone,

I'm running a response history analysis on a 3D bridge model and facing some covergence issues. I believe I was able to narrow the problem down to the abutment where I have 2 zeroLength springs acting in series.
The setup consists of 3 nodes all having the same x-,y-, and z- coordinates and 2 translational springs. Here is a quick rundown:
- Node 1 is restrained in all 6 DOFs
- Spring 1 representing abutment-soil behavior between Node 1 and 2 in x- and y- directions (1 and 2 directions in the Opensees notation)
- Spring 2 representing bearing behavior between Node 2 and 3, again in x- and y-directions
- Node 3 is connected to beam-column element carrying the mass, but for the sake of discussion you can consider it as a free node
- There are equalDOF commands for node pairs 1-2 and 2-3 constraining DOFs 3,4,5,6, so there are no connectivity issues.
- Both springs are Hysteretic materials
- If I constrain nodes 2 and 3 with equalDOF in x- and y-directions, Spring 1 deforms just fine without any errors
- Similarly, if I fix Node 2, Spring 2 will be able to deform and sustain strength degradation and many cycles of loading
- However, if I use them together, I get convergence issues, not immediately, but at some point throught the analysis

I've tried different algorithms and reducing time steps during the analysis, but without much luck. If both springs go inelastic, they fail at some point, often during unloading.

I'm aware that there is an option to create a Series material that will combine the flexibilities, so I can only use 1 spring. The problem is that I need to know deformation contributions of separate elements for my research, so it's not an ideal choice.

Has anyone dealt with this issue before? Thanks in advance for your help.