Investigation of Irreversible Demagnetization Constraints in Magnet Volume Minimization Design of IPMSM for Automotive Applications Using Machine Learning

In recent years, interior permanent magnet synchronous motors (IPMSMs) have been widely used as traction motors in electric vehicles. For the design of IPMSMs, finite element analysis is commonly used, but this method is highly time-intensive. To shorten the design period of IPMSMs, various surrogate model construction methods have been proposed to predict the relevant characteristics, and these models have been used in the optimization of IPMSM geometry. However, to date, there have been no surrogate models that are suitable for evaluating irreversible demagnetization of permanent magnets. Here, we show a method for accurately predicting the flux densities of permanent magnets of an IPMSM using machine learning techniques. To improve the prediction accuracy, we set the apparent permeance coefficient as the prediction target of the machine learning methods. We then use the trained surrogate model and a real-coded genetic algorithm to minimize the permanent magnet volumes with irreversible demagnetization constraints and show that the design time can be significantly reduced compared to the case where only finite element analysis is used.