Moving cone tutorial runs with velocity motion solver but fails with displacement
 

Hi all,
I'm working on fluid-structure interaction with pimpleFoam and I've run into a problem with adaptive time stepping; the time step oscillates very rapidly. This oscillation is caused by oscillation in the Courant number, which is itself caused by oscillation in the mesh. I use the displacementSBRStress mesh motion solver together with the rigidBodyDynamics library in OpenFOAM v2006, but the problem also occurs with displacementLaplacian.

I simplified the problem by ditching FSI and testing forced boundary motion on the same mesh. The problem is still present when solving with displacementLaplacian, but is completely resolved when using the velocityLaplacian solver. I'd use velocityLaplacian for my FSI case as well, but the rigidBodyDynamics library only works with displacement mesh motion solvers.

I figured that maybe something about my case's settings or mesh might be causing the issue, so I moved to testing simple tutorial cases. The moving cone tutorial of pimpleFoam uses the velocityComponentLaplacian mesh motion solver and it runs fine. I changed the tutorial to use displacementComponentLaplacian, and at about t = 0.002 the same issue of mesh oscillation pops up! The mesh starts oscillating back and forth and quickly fails. I've attached the files for this case.

Does anyone have ideas as to what might be causing this? The equation being solved is the same Laplace equation, only the field being solved for is different, so why is the displacement mesh solver so unstable? I've tried different Laplacian schemes, time steps and looping over the Laplace equation multiple times, nothing helped.