APPLICATION OF A SEARCH ALGORITHM TO MODEL BASED SYSTEM OPTIMIZATION

In this paper a new approach to modeling and optimization of mechatronic systems resp. components of systems is presented. The described methods are applicable in cases where, due to a high complexity of the model, conventional approaches of identification and optimization are not utilizable. This applies to systems consisting of a large number of identical or similar elements as well as to models requiring the simulation of flow or temperature distributions. A model structure consisting of a number of coupled individual control elements has been developed. Two optimization tasks, the adaptation of the model to the given system as well as the system optimization based on model calculations, are accomplished by the same search algorithm. A significant acceleration of the search process is achieved by applying heuristic knowledge. The approach has been developed first for a glass production machine. The combination of model and search algorithm has been applied to a feeder for the conditioning and portioning of melted glass. Simulation results for a typical disturbance are presented.