Metamodel-Based Multi-Objective Optimization for the Design of Induction Motors for Vehicle Drives

The deployment of electric machines as traction machines in the automotive sector is becoming increasingly prevalent, necessitating their optimization with respect to a multitude of target variables across a spectrum of driving scenarios. The design process thus entails the selection of a multitude of parameters with the objective of attaining optimal values for all target variables, while simultaneously considering the constraints that exist within the given context. The design and optimization of electrical machines is a highly time-consuming and computationally intensive process, as it often necessitates the simulation of numerous designs. Furthermore, it is frequently necessary to have an initial machine design that can then be optimized to meet new requirements. In this paper, we address the issue of high computational intensity and the challenge of numerous optimization approaches that necessitate an initial machine as a starting point. This paper presents a metamodel-based design process for induction machines with squirrel-cage rotors and round-wire windings. The approach is capable of generating different induction machine variants for given requirements, which are approximately Pareto optimal in terms of efficiency and installation space. The optimization process is shown to be much more time-efficient when using the presented metamodel.