this topic has been started on the elmer/ice emailing before this forum was created. Here a short summary with the final update how the problem has been solved just in case sb. else encounters similar issues.
I faced convergence problems of the nonlinear part of the FreeSurfaceSolver in 100yr prognostic simulations in the regions with "no ice" (ice-thickness = prescribed min. ice-thickness). This problems were caused by problems with the MeshUpdate Solver which made some surface mesh points going below the bedrock.
Playing with the parameters Mesh Elastic Modulus and Mesh Poisson Ratio improved the problem a little.
Not allowing for sliding in areas in this "no ice" regions or not allowing the FreeSurface to evolve there, improved the problem also a little, but finally didn't solve it. These two points were done by creating an ice-mask and using a Condition when defining slip coefficients and imposing an additional Dirichlet Boundary Condition for the FreeSurface.
Now using the StructuredMeshMapper Solver the problem is completly solved, thanks for pointing me to that Thomas.
Extension of Elmer in computational glaciology
1 post • Page 1 of 1