Fixed Point Convergence Error:¶
Convergence problem in the fixed point loop for the contact nonlinearity resolution. Switching the contact nonlinearity solution method to Newton might resolve this issue.
The convergence for the contact nonlinearity could not be reached using the fixed point iteration method.
What could be the possible reason?¶
There could be multiple reasons why the convergence could not be reached. Some examples might be an incorrect definition of master and slave entities for the contact zone, too high penalty coefficients or a too coarse surface mesh in the contact region.
What can I do now?¶
- Check the master and slave assignments:
- Check if the general recommendations for the selection of master and slave entities for the contact zone are fulfilled. In case of doubt you can use the swap entities functionality to directly swap master and slave entities for a specific contact zone.
- Check the penalty coefficients:
- The correct selection of penalty coefficient may sometimes be a little tricky. If the values are too low, the contact penetration might be too high, if the value is on the other hand too high, convergence might be impossible. As a rule of thumb, the starting guess for the penalty coefficients should be about 5-50 times as high as the softest material of the contacting parts.
- Switch to Newton method:
- Depending of the nature of the contact problem, using the Newton method for the contact nonlinearity resolution might resolve the convergence problems. Generally, the Newton method is faster than the penalty method.
- Check the mesh fineness on the contact regions:
- Sometimes convergence problems occur if the mesh is too coarse on the contact regions. Especially if the slave entities contain highly curved surfaces, make sure the mesh resolution is high enough to have rather small angles between adjacent surface elements. If not, you can either add a mesh refinement on the contact surfaces or enable the contact smoothing in the physical contact settings.