Shorter development cycles and increasing numbers of applications require faster development cycles. Simulation and design of electric drives must therefore output an optimum between costs and efficiency within a short time. In order to get an optimal design of an electric machine considering transmission and power unit and to speed up the time of design, machine learning methods are used. In this talk we will look at:

• Structure of parameterized EM designs
• DoE building using LHS and automated calculation of all EM designs generated in the statistical design of experiments
• MOP formation using machine learning methods with optiSLang
• Holistic Optimization of the e-drive under consideration of constraints to be met and consideration of selected objective functions based on MOP with subsequent validation of optimal designs