Challenges in simulating geomechanical models using Mohr-Coulomb with Tension Cut-Off in Abaqus

[cm2618]

I am modelling an underground cavern for Compressed Air Energy Storage applications using Abaqus. To begin, I attempted to replicate the results presented in the provided paper:
Damasceno, DR, Spross, J & Johansson, F 2023, 'Rock mass response for lined rock caverns subjected to high internal gas pressure', Journal of Rock Mechanics and Geotechnical Engineering, vol. 15, no. 1, pp. 119-129.

In my model, I utilized the Mohr-Coulomb plasticity model for the geological formation, with a specified tensile strength of zero, which I interpreted as a tension cut-off of zero (a value of 0.1 was used) in Abaqus.

However, during the solving process, I encountered an issue with negative eigenvalues, resulting in the final solution displaying no stress or deformation in the output, despite the completion of the job in Abaqus.

My questions are as follows:
1. Is my interpretation, equating a tension cut-off of zero with a tensile strength of zero, accurate?
2. How can I address the issue of negative eigenvalues?
3. I have already attempted mesh refinement and employed a small initial increment size. Are there any adjustments I can make to the element controls that might help?

 

[FEA way]

This approach makes sense but may cause convergence issues. Maybe you should modify the other parameters of the Mohr-Coulomb model (especially cohesion yield stress) or use a different model for this material. It's rarely a good idea to change the solver controls.

Make sure that the issues in your original model are caused by the material definition, not by something else like contact. You could also prepare a single element test with this material definition to see how it behaves under different types and levels of loading.