基于BP神经网络PID自适应控制的无刷直流电动机

Research on the Controlling Methods of BLDCM Based on Improved BP Neural Network PID

提出了利用变步长法和引入动量项来改进BP神经网络学习算法,有效减小了学习过程的振荡趋势,改善了收敛性,避免了学习过程陷入某些局部最小值,并将其用于在线调整无刷直流电动机控制系统的PID参数,实现具有最佳组合的PID控制。

In this paper, a novel BP neural network PID control strategy is proposed. In order to speed up the network convergence speed and avoid falling into local minimum point, a mixed method was adopted, by means of changing network iteration step-length and introducing the momentum terms. Then the algorithm is used to adjust the parameters of BLDCM PID controller so that the whole system arrives optimal.