Stokes: Laplace vs. elasticity formulation (P1nc-P0)
 

Hello,

I have developed a finite element Stokes-solver that can handle both, Laplace (a(u,v)=\int \grad u : \grad v dx) and elasticity formulations (a(u,v)=\int \epsilon(u) : \epsilon(v) dx) of the Stokes problem. It is known that nonconforming linear-constant approximations do not fulfill all stability requirements in case of the elasticity formulation. Does anybody know a good test case to show a significant difference in the solution (Laplace vs. elasticity) for various elements, particularly the mixed P1nc-P0 approximation?

Cheers, Chris