Order reduction with partial ARMA-modeling

In this paper a new method for order reduction of large FE-models is presented. Many order reduction methods known from the literature fail when they are applied to large parametric models. The method proposed here is based on numerical data. Therefore this method can be seen as an identification method as well. The reduced system is represented by a difference equation and the parameters are determined in a two-step procedure. Time delays in the original system may be detected. Theoretical as well as real technical systems are used as examples to demonstrate the application of the new method.