摘要:为进一步提高永磁直线电机调速系统的动静态性能,提出将BP神经网络与传统PID控制器相结合,以实现PID参数的最优化自整定。在研究常规BP神经网络的控制算法的基础上,对常规BP算法的学习速率和动量因子不能动态调整带来的缺陷进行分析,提出基于误差变化率的动态调整算法,实现对PID控制器参数的自寻优,并将算法应用于永磁直线电机矢量控制系统速度调节器中。仿真实验结果表明,改进后的基于BP神经网络PID控制算法与传统PI控制器相比,控制系统的动静态性能更优,稳定性更好,解决了由于参数整定困难而导致PID控制器性能不能达到最优化的问题。
关键词:永磁直线同步电机;PID控制;BP神经网络;动态调整; Abstract: In order to further improve on the static and dynamic performance of the permanent magnet linear synchronous motor speed regulating system, the traditional PID controller was combined with the BP neural network to achieve PID parameters self regulating to optimize the parameters. The defects that the learning rate and the momentum factor about the normal BP neural network cant regulate dynamically leads to were analysed on the basis of studying the normal BP neural network algorithms, and the dynamic regulating algorithm based on error varietyrate was proposed to realize to the selfoptimization of the PID parameters, and it was applied to the permanent magnet linear synchronous motor to control the motor speed. The simulation results prove the PID controller combined with the improved BP neural network has better static and dynamic performances and stability, and resolved the problems which the traditional PID controller cant achieve the most optimization because of the difficulty of the parameter regulating. Key words: permanent magnet linear synchronous;PID controller;BP neural network;dynamic regulating
|