ABSTRACT The physically‐motivated dynamic flocculation model (DFM) is capable of modeling the non‐linear material behavior of filled elastomers, for example, as damping elements in centrifuges. For predicting the system behavior of these dynamic systems, it is planned to incorporate DFM into a multi body simulation (MBS) framework. However, the DFM requires the computationally expensive concept of representative directions (CRD) for a three‐dimensional formulation. Hence, three independent strategies to reduce the computational effort caused by the CRD are investigated. Two approaches capable of approximating internal variables are presented and additionally a method reducing the number of directions by random orientation is proposed. This method contrasts favorably with the other approaches, since it effectively minimizes computational costs while maintaining high accuracy in mean reaction forces.
Niemeyer et al. (Sun,) studied this question.