Torque Performance Optimization of MR Fluid Brake Integrated PMSM Based on Brake Constrained Bagging-GA Method

This article presents an optimization design method of torque performance for magnetorheological fluid brake integrated permanent magnet synchronous machine (MRFB-I-PMSM). Machine learning model Bagging is used to build an accurate surrogate model of MRFB-I-PMSM from finite element analysis (FEA) results to improve the efficiency of the multi-objective optimization in the genetic algorithm (GA). Then, the brake torque is added to GA as a constraint related to the EM torque, the initial complex three-objective optimization is simplified into the brake constrained two-objective problem.