Development of Deep Learning-Based Design Optimization Approach for e-Mobility Traction Machines Considering Multi-Physics Problems and the Evaluation of Uncertainties and Tolerances

This work is about ground-breaking new electric machine design approaches for transportation electrification. (i) Multi-physics aspects, (ii) new machine modeling and optimization approaches, and (iii) improved soft magnetic material modeling will be considered to achieve better designs in terms of efficiency, power density, noise and vibration, etc. Additionally, quantifying the impact of tolerances and additional uncertainties on the machine performance will facilitate obtaining designs with both excellent rated performance and high reliability with regard to inevitable variations. A prototype will be built to verify the obtained simulation results.