Data-Driven Approaches for Electromagnetic Analysis of Traction Electrical Motors: A Proposal for a Benchmark Problem

The rising demand for high-performance traction electrical motors in electric vehicles necessitates efficient design solutions. This paper describes a preliminary approach to data-driven modelling of the electromagnetic design problem of traction electrical motors. The initial step consists of generating the dataset for training the data-driven models, which is performed using an FEA-based procedure. Subsequently, different data-driven approaches are compared to surrogate the output of the electromagnetic analysis, identified in the computation of the torque and the torque ripple of a specific electrical motor structure.