convergence problem in micromechanical model calculation

[yongby]

recently I engage in the calculation of a RVE of filled rubber. I have built a RVE model with repeated boudary conditions, in ABAQUS with the keyword *Equation. My model tends to imitate the models in Bergsterom's paper (Rub Chem Tech, 1999). But when the job execution carried out to 110%-120% stretch (lambda1), ABAQUS always reports error due to divergence.
I want to know in the calculation of heterogeneous material, whether it need to change the condition of convergence for instance using *control in ABAQUS? OR using ABAQUS/explicit to do this kind of calculation?

The convergence problems that you are describing are common in large deformation finite element models of heterogeneous materials. The problem is that the material deformation becomes inhomogeneous and hence the FE mesh also becomes hightly distorted. To overcome this problem you need to remesh the simulation, either manually or by using an automatic mesh refinement approach. Note, however, that ABAQUS does not currently have an automatic mesh refinment approach for implicit simulations. :(