基于优化BP神经网络PID的永磁同步电机控制研究

Research on Control of Permanent Magnet Synchronous Motor Based onPID of Optimized BP Neural Network

针对传统PID控制在永磁同步电机控制系统中未能实现精准控制的问题,提出了一种基于改进蜣螂优化算法的BP神经网络PID控制器,该控制器由BP神经网络通过自适应方法来调整权重系数,解决了PID无法在线调节参数的缺点。针对BP神经网络在进行反向传播时陷入局部最优的概率较大,引入蜣螂优化算法通过适应度值不断更新BP神经网络核心参数,从而提高BP神经网络的优化速率。对于蜣螂优化算法中存在初始种群质量不高及搜索能力不足等问题,对蜣螂优化算法进行混合策略优化,大大提升了蜣螂优化算法求解效率和精度。实验结果表明该改进蜣螂优化算法可以有效地提高控制系统的响应速度,减小超调量,在转速和负载突变的情况下都有较强的鲁棒性。

In view of the problem that traditional PID control fails to achieve accurate control in the permanent magnet synchronous motor control system,a PID controller of BP neural network based on improved dung beetle optimization algorithm is proposed.The controller uses BP neural network to adjust the weight coefficient by adaptive method,which solves the shortcoming that PID can not adjust parameters online.Then,aiming at the high probability of BP neural network falling into local optimum when performing back propagation,the improved dung beetle optimization algorithm is introduced to continuously update the core parameters of the BP neural network through the adaptive value,so as to improve the optimization rate of the BP neural network.Furthermore,addressing issues like low initial population quality and inadequate search capability in the improved dung beetle optimization algorithm,a hybrid strategy is implemented to optimize the improved dung beetle optimization algorithm,which greatly improves the solving efficiency and accuracy of the improved dung beetle optimization algorithm.Experimental results demonstrate that the improved dung beetle optimization algorithm effectively enhances the response speed of the control system,reduces overshoot,and it has strong robustness to the case of speed and load sudden change.