Design Space Exploration

The design of power electronic or mechatronic systems entails the specification of values of a vast number of design variables such that the best possible solution/multi-objective performance limit can only be determined using advanced search algorithms. PES has established a multi-domain modelling/optimization framework, which includes cost data and is used for comprehensive comparative converter analyses concerning efficiency, power density and costs with special focus on design space diversity, i.e. very different sets of design parameters mapped to a narrow region in the performance space. Furthermore, a hybrid method, which combines the accuracy of 3D-FEM and the low computational cost of ANNs for the optimization of magnetic components has been implemented. Moreover, the design considerations have been extended from single converters to systems, e.g. an airborne wind turbine, an ultra-deep sea hydrone, and PV inverter life-cycle cost analyses.