Bill Bulat
Ansys Employee

How are you representing this catastrophic material failure? Is it your expectation that the finite elements surrounding the screw will severely distort and "fail" (as the bone material would physically)? Or are you instead using equivalent friction model in the contact that behaves as failed bone would? If the former, it seems that NLAD (nonlinear adaptive feature) would be necessary. Mind you, NLAD might not be sufficiently robust to accomodate the mesh distortion that I imagine one might see in such a simulation. If the latter, it isn't immediately obvious to me how you might determine an equivalent contact friction. But for what it's worth, contact friction related constants, such as TAUMAX, can, with an APDL command object, be made a function of independent variables (e.g., sliding distance, contact pressure) by using a table. You might consider trying to locate the following Help sections for more details about this:

Kind regards,

Bill