Elastomer Vane Contact Model in Code_Aster

Goal

Predict the contact pressure and deflection of a single elastomer vane as it sweeps a housing profile with designed interference, so that housing geometry could be traded against vane life without cutting steel.

Setup

• 2D plane-strain contact model in Code_Aster via salome_meca, running on WSL2
• Vane clamped at the hub, wall motion imposed rather than pressure-driven
• Nearly incompressible elastomer: E = 0.9 MPa from G = 0.30 MPa, nu = 0.47

The convergence fight

This is the part nobody writes up, so here it is.

Revision	Behavior
b	Singular matrix at the first step
c	Seated to ~40% of interference, then a geometry-pairing 2-cycle that was independent of time step down to dt = 3e-9
d	Relaxed geometric residual and contact coefficient cleared the 2-cycle; died later on a chaotic frictionless tangential orbit

Two findings worth keeping:

1. A pairing 2-cycle that does not respond to time-step refinement is not a time-step problem. Cutting dt further just burns wall clock.
2. Frictionless contact on a compliant sweeping surface is not the conservative simplification it looks like. Removing friction removed the only thing damping the tangential solution.

Status

Ongoing. The next revision adds a small friction coefficient and revisits the pairing tolerance rather than continuing to refine the step.