改进粒子群优化BP神经网络的SRM转子位置间接检测∗

摘要

提出了基于改进粒子群优化BP神经网络的开关磁阻电机转子位置建模估算方法。针对BP易陷入局部最优、收敛速度慢等情况，在标准粒子群算法的基础上，改进粒子的速度与位置更新策略，优化BP神经网络的阈值和权值，建立起开关磁阻电机磁链、电流和转子位置的非线性映射关系。借助于MATLAB/Simulink完成仿真。结果表明，与标准的BP神经网络和遗传算法优化的BP神经网络相比，基于改进粒子群优化BP神经网络算法结构简单、训练迅速，能够获得更高的精度，实现了开关磁阻电机转子位置的间接检测。%This paper presented a modeling and estimation method of SRM rotor position based on improved particle swarm optimization BP neural network. Since BP neural network has the disadvantages of low convergence rate, easy to fall into local optimization and so on, an improved updating strategy of particle position based on the stand particle swarm method was proposed. Then, with the improved particle swarm algorithm, the threshold and weight of the neural network were optimized to establish the nonlinear mapping relationships between switched reluctance motor flux linkage, current and rotor position. The MATLAB/Simulink results show that the particle swarm optimization based on improved BP neural network has the advantages of simple structure, training quickly, and can get higher detection precision, compared with the standard BP neural network and BP neural network optimized by genetic algorithm. Thus, the proposed algorithm can achieve a better performance in indirect rotor position detection of SRM.