I am working towards modeling the effect of hydrogen embrittlement (HE) on the plastic behavior of steel alloys. I wrote a Fortran code that solves the plastic model of HE. Afterward, I coupled the Fortran subroutine with UMAT to study the impact of HE on a tensile specimen. The UMAT accesses the HE model if the yield point is reached.
The problem is when I test the UMAT for a single element it works but if the number of elements increased the UMAT does not converge (incorrect data are supplied to the HE model). The HE model requires from Abaqus Mises stress, strain rate, and other state variables. I update the Jacobian matrix in the UMAT according to the tutorials provided by Abaqus. Any possible solutions for such an issue?
Thanks in advance