Multi-objective design and optimization of generalized switched reluctance machines with particle swarm intelligence

摘要

This paper proposes a fast and generalized multi-objective design and optimization method for the Switched Reluctance Machines (SRM). An analytical design model for SRMs with any feasible stator and rotor slot combinations is firstly developed, which can accurately evaluate a SRM design much faster than the prevalent finite element analysis (FEA) method. In addition, a novel method for multi-objective optimization of SRM is proposed based on this analytical model, and the number of prime variables to be optimized is reduced to only five. A canonical Particle Swarm Optimization (PSO) algorithm with penalty function is applied to find the optimal solution for a user defined objective function. After several rounds of searching process with the PSO, the optimal regions can be found for the design variables in terms of the performance indices (PIs). Finally, the optimized designs are validated by FEA. This method can generate the optimized SRM designs subject to different design requirements and accelerate the entire optimization process.