Stability problem in soil modelling

[rafetsisman]

Hi,

I am modelling a soil column to perform 1D site response analysis. I started with the example provided in OpenSEES geotechnical examples and changed the material to elastic isotropic. The problem is, whatever the element size or analysis time step, it cannot converge at some time during the analysis. When I illustrate the truncated results, I realized that the displacements goes to infinity in such a pattern that can be regarded as a stability problem. I am sure the model boundary conditions are the same with the provided example. By the way, the analysis yields reasonable results in case of no dashpot, I mean fixed base(infinite rigidity).

What can be the problem?

A fictitious solution to that problem I found is to multiply dashpot coefficient with "-1". In this case the analysis also yields reasonable results.

Is there a problem with viscous material to model dashpots?