Bernd Cebulski, Stephan Paul, Advanced Powertrain Electrification, IAV GmbH Chemnitz
Electrification in automotive systems has an increasing demand on the efficiency and utilization of electric drives. Modern calculation software based on the finite elements approach is nowadays used to accurately predict the performance figure and secures investments. The increasing power of computers makes it possible to run thousands of calculations in parallel in a short time and optimization of the most common geometrical dimensions becomes possible. With the use of adapted evolutionary strategies it will be shown how a clean traceable process of electric motor design has been established.
Based on massive parallel computer farm and the performance of JMAG the evaluation of pareto-optimal design for HEV and EV traction applications makes it possible to achieve the best compromise between cost and performance.
For the development of new machine topologies with the focus on rotor magnet geometry we have developed an advanced topological optimization strategy. With a parametrical model of the geometric elements it is possible to switch the FEA properties for every mesh element to be iron, air or magnet. By using differential evolution optimization use that approach for a structural optimization and finding of new innovative magnet arrangements for minimizing magnet costs and torque ripple as main goals. It will be shown how JMAG can be efficiently being used for even the most sophisticated ideas.
Application of the proposed methods will be shown for a power hybrid system example with an electric drive limited in available space and high power demands.