Improved Particle Swarm Optimization for Parameter Identification of Permanent Magnet Synchronous Motor

Shuai Zhou,Dazhi Wang,Yongliang Ni,Keling Song,Yanming Li

DOI: https://doi.org/10.32604/cmc.2024.048859

2024-01-01

Abstract:In the process of identifying parameters for a permanent magnet synchronous motor, the particle swarm optimization method is prone to being stuck in local optima in the later stages of iteration, resulting in low parameter accuracy. This work proposes a fuzzy particle swarm optimization approach based on the transformation function and the filled function. This approach addresses the topic of particle swarm optimization in parameter identification from two perspectives. Firstly, the algorithm uses a transformation function to change the form of the fitness function without changing the position of the extreme point of the fitness function, making the extreme point of the fitness function more prominent and improving the algorithm's search ability while reducing the algorithm's computational burden. Secondly, on the basis of the multi-loop fuzzy control system based on multiple membership functions, it is merged with the filled function to improve the algorithm's capacity to skip out of the local optimal solution. This approach can be used to identify the parameters of permanent magnet synchronous motors by sampling only the stator current, voltage, and speed data. The simulation results show that the method can effectively identify the electrical parameters of a permanent magnet synchronous motor, and it has superior global convergence performance and robustness.