Modular Model Reduction of Interconnected Systems: A Top-Down Approach
Models of complex systems often consist of multiple interconnected subsystem/component models that are developed by multi-disciplinary teams of engineers or scientists. To ensure that such interconnected models can be applied for the purpose of simulation and/or control, a reduced-order model for the interconnected dynamics is needed. In the scope of this paper, we pursue this goal by subsystem reduction to warrant modularity of the reduction approach. Clearly, by reducing the complexity of the subsystem models, not only the accuracy of the subsystem models is affected, but, consequently, also the accuracy of the interconnected model. It is typically difficult to predict a priori how the interconnected model accuracy is affected precisely by the subsystem reduction. In this work, we address this challenge by introducing a top-down approach which enables the translation of given accuracy requirements of the interconnected system model to accuracy requirements at subsystem model level, by using mathematical tools from robust performance analysis. This allows for the independent reduction of subsystem models while guaranteeing the desired accuracy of the interconnected model. In addition, we show how this top-down approach can be used to significantly reduce the interconnected model in an illustrative structural dynamics case study.