COMPARISON OF MODAL AND NON-MODAL MODEL REDUCTION APPROACHES FROM A MECHANICAL APPLICATION ENGINEER’S PERSPECTIVE

Model reduction is essential when the dynamic of Finite Element (FE) models is investigated. Successful model reduction guarantees that a complex system with several million degrees of freedom (DOF) can be accurately approximated by a set of equations which is significant smaller. This is of particular importance in the field of multi body simulation where model reduction is applied in order to import flexible FE structures into the multi body dynamic system. Vibration mode based model reduction methods like Component Mode Synthesis are compared with methods coming from control engineering, namely moment matching and balanced truncation. The key issue, on which this paper is focused on, is the characteristic of these approaches in terms of a “black box” behaviour for industrial use. In particular the methods strength and limitations in terms of the reduction of metallic structures, the sensitivity of the reduced model to varying boundary conditions, full system response, accurate statics, stress recovery and the possibility to determine ‘a priori’ the number of remaining DOF are investigated and commented. This paper gives insight into the methods nature and clarifies the strength and limitations of the three methods. Conclusions based on the theory and on a generic numerical example are presented. It turns out, that mode based model reduction methods like Component Mode Synthesis have best overall performance.