Hi,
I have developed numerical model with metal matrix reinforced with ceramic inclusion and where imposed cohesive surface interaction between them to predict the debonding mechanism under mode independent behavior, for which, I have considered maximum nominal stress criterion for damage initiation and adopted energy based liner evolution softening law .I performed uni axial tensile test based on displacement control. In this contact based cohesive surface analysis, I have considered small sliding with node to surface discretion. The problem I am encountering in this simulation is that the solution is not converging after reaching maximum stress criterion value say one. This analysis has been conducted by quasistatic condition using abaqus standard solver. Could you please suggest me some hint to make the solution to be converged after material degradation ?
Cohesive elements are strongly mesh dependent, in particular the size of elements near the interface and of cohesive elements themselves is critical. Thus, have you tried to run a couple of simulations with very different elements' size (a coarse and a very fine one)? If nothing changes between them, the problem lies probably in the parameters of the solver, which might need some tuning.
Thank you for the answer but I am not using cohesive element approach. This problem is arising when I am using cohesive surface based method.
as luca said they are mesh dependent. even the surface based method u are using a mesh dependent. refine the meshes of upper and lower surface or try to make sure that the energy is conserved. also check for energy units. did you incorporate stabilization ? Do it otherwise.
Hello,
Following step should be follow to resolve that issue-
->In step module -> other -> solution control manager ->Step1 -> specify -> time incrementation-> activate discontinuous analysis ans change increase IA value (suggest 20). I hope this will help you.
Hello,
In the most of problems with cohesive interfaces you will face the instability in solution when the traction at a cohesive interface reaches its maximum. Roughly speaking, if the the energy dissipated by the crack interface is less than the elastic energy released from the elastic body due to the unloading, then the solver cannot follow the load path. Several techniques have been proposed to handle this issue, e.g. arc-length methods (Riks) and small viscosity parameter. In ABAQUS it is easier to choose an small value for the viscosity parameter as a stabilizer. However, this new unknown parameter (viscosity) must be chosen carefully. For more info please see Gao and Bower 2004.
- Badges
- Science topic