Bagging-GA-Driven Multi-Objective Optimization of MR Fluid Brake-Integrated PMSM Considering Brake Torque Constraint

This article presents a Bagging-GA-driven multi-objective optimization method for magnetorheological (MR) fluid brake-integrated permanent magnet synchronous machine (MRFB-I-PMSM) considering brake torque constraint. In order to effectively alleviate the calculation burden and mitigate the convergence challenges posed by the characteristics of the topology itself, the ensemble learning Bagging method is first aided to construct an accurate surrogate model (T avg,T ᵣᵢₚ, andT b) from finite element analysis (FEA)-based data, aiming to enhance the optimization efficiency of the genetic algorithm (GA). To reduce the convergence risk of multi-objective optimization in GA, this article opts not to directly optimizeT bbut rather to constrain it as a passive variable (R T) within a certain range related toT avg. Finally, the performances of the MRFB-I-PMSM with and without optimization are compared based on the FEA simulation, verifying the validity of the proposed method.