Fix AdaptiveBeamForceFieldAndMass

[bakpaul]

This component was broken wince the begining but it wasn't an issue until I needed to accumulate inertia terms on the RHS of the Integration scheme. The problem here is thazt the addMDx method ignores the dx passed and always uses the same value. This broke completely the dynamics, nullifying the contribution of the mass to the RHS.

It was identified by AI, and the code was made by AI, but it works properly now thank's to this.

[bakpaul]

After some investigation, the regression here is indeed a consequence of the PR. Before my PR, because addMDx was broken, when adding the rayleigh damping to the RHS, only the part coming from the rayleigh stiffness was added. Indeed, when adding the part multiplied by M, the forcefield added M*g instead, which is 0 (because here g=0).

After trying to recreate the same behavior by removing the rayleigh stiffness from the RHS, the results are closer than the expected precision. So I'll regenerate the references to fix this regression.