Penalty vs mixed formulation in FEA

m. malik

I will appreciate knowing the views of the CFD workers who have experience with both penalty and mixed formulations for the finite element analysis of non-Newtonian, incompressible flows.

Just briefly the reason for this query: My interest is in the simulation of polymer flows in processing equipment. I use my own three-dimensional pentaly-Galerkin finite element code for the solution of the three conservation equations with 8-node hexahedral elements. At times I face severe problems of negative pressures (which are obtained in post-processing from p = -(viscosity)(penalty parameter)div (v-vector)). There are other less troublesome problems too. For these reasons, I am now thinking of developing the simulation code with mixed formulation and see how that goes.

Unfortunately, I am unable to dig out any information on the comparison of the two formulations. And so the question with me is: Will the efforts on another code be worthwhile?

Regards.