1: \begin{abstract}

2: % CHALLENGE

3: %Standard finite element discretization associates mass to every node.

4: %The velocity jumps occurring for contact activation and release then lead to spurious high-frequency oscillations.

5: %As a consequence, numerical time step integration suffers poor convergence and/or energy conservation behavior.

6: %This leads to enormous computation effort and thus impedes long-term simulations, as required for predicting vibrations (vibro-impact processes).

7: %Also, the artificial loss/gain of energy leads to an erroneous prediction of the physical damping behavior.

8: % PROPOSITION

9: We propose to combine the ideas of mass redistribution and component mode synthesis.

10: More specifically, we employ the MacNeal method, which readily leads to a singular mass matrix, and an accordingly modified version of the Craig-Bampton method.

11: Besides obtaining a massless boundary, we achieve a drastic reduction of the mathematical model order in this way compared to the parent finite element model.

12: %We then adapt a symmetric Moreau-like time step integrator for the resulting differential-algebraic equation system subjected to set-valued contact laws.

13: We assess the method's computational performance by a series of benchmarks, including both frictionless and frictional contact.

14: The results indicate that the proposed method achieves excellent energy conservation properties and superior convergence behavior.

15: It reduces the spurious oscillations and decreases the computational effort by about 1-2 orders of magnitude compared to the current state of the art (mass-carrying component mode synthesis method).

16: We believe that the computational performance and favorable energy conservation properties will be valuable for the prediction of vibro-impact processes and physical damping.

17: \end{abstract}

18: