Using Machine Learning to Reduce Design Time for Permanent Magnet Volume Minimization in IPMSMs for Automotive Applications

Interior permanent magnet synchronous motors (IPMSMs) have been widely used as traction in electric vehicles. Finite element analysis is commonly to design IPMSMs but highly time-intensive. To shorten the period for IPMSMs, various surrogate models constructed predict relevant characteristics, and they optimization of IPMSM geometry. However, date, no able accurately characteristics over wide speed range required automotive applications. Herein, we propose a method predicting speed-torque an by using machine learning techniques. improve prediction accuracy, set motor parameters target methods. We then trained model real-coded genetic algorithm minimize volume showed that time can be significantly reduced compared with case where only finite used.