Performing contact analysis in a transient dynamic procedure is always challenging due to the presence of contact chattering (frequent change in contact status). One of the reasons for contact chattering is that the contact constraints enforce constraints on nodal displacements (penetration) but do not impose any constraints on nodal velocities. Since nodal velocities and nodal accelerations are dependent on nodal displacements (see Section 17.2 Transient Analysis in the Mechanical APDL Theory Reference), any constraints on nodal displacements make the nodal velocities and nodal accelerations inconsistent. This introduces numerical errors in the transient dynamic solution in subsequent time increments and, if left untreated, leads to numerical instability (non-convergence or incorrect solution).
Numerical damping is usually added to the time integration scheme to suppress such numerical errors. However, addition of numerical damping does not work in several cases. For example, in problems with multiple or repeated impacts there is constant growth of numerical error in the solution, and the analysis eventually fails to converge in spite of large numerical damping. In other situations the analysis may converge with addition of numerical damping, but the system response may be underdamped or overdamped depending on the growth of numerical error.
A more appropriate solution to the contact chattering problem is to treat nodal displacement and nodal velocities consistently. One such solution is provided by Energy and Momentum Conserving Contact.