Multi-Objective Optimization of Torque and Torque Ripple in Double V-Type IPMSM Using a Random Forest-Based Modified PSO-GA Hybrid Method

Interior Permanent Magnet Synchronous Machines (IPMSMs) are widely used not only as drive motors for Electric Vehicles (EVs) but also in various industrial fields due to their high efficiency and high power output characteristics. However, because of the embedded magnets, IPMSMs exhibit significant torque ripple during operation, necessitating the use of optimization algorithms to address this issue. For IPMSMs, which have a large number of design variables, the feasible design space is defined by multiple constraints, increasing the complexity of the design optimization process. Therefore, this paper proposes a Random Forest-based Modified PSO-GA hybrid method to perform optimization in the presence of multiple constraints and applies it to the torque and torque ripple improvement design of a Double V-Type IPMSM.