Realization-preserving model reduction of object-oriented models using energy-based metrics

This paper deals with the adaptation of energy-based realization-preserving model-reduction techniques for object-oriented models implemented in Modelica. An approach originally developed for bond graphs, where energy-related heuristics were used to identify the components contributing the most to the model’s salient dynamics, was extended to more heterogeneously formulated models in Modelica. A prototype toolbox was developed inside the modeling environment OpenModelica. Its applicability was demonstrated with two examples. Although the model-reduction tool cannot reduce the model completely automatically, it may still offer many useful insights and hints to the user.