利用改进粒子群优化(PSO)算法优化BP神经网络的权值和阈值,有效地解决了BP算法易陷入局部极小值的缺点,能更快速的实现收敛,不仅具有广泛的映射能力,还明显提高了运算效率。通过对直接转矩控制(DTC)系统进行MATLAB/SIMULINK仿真研究,结果表明:基于PSO-BP神经网络构造的速度辨识器具有良好的辨识效果。 By using the improved Particle Swarm Optimization (PSO) algorithm to optimize the weights and thresholds of BP neural network, BP algorithm can easily solve the disadvantage of falling into local minima and can converge more quickly. It not only has extensive mapping ability but also obvious Increased computational efficiency. The simulation results of MATLAB / SIMULINK for DTC system show that the speed recognizer based on PSO-BP neural network has a good recognition effect.