基于DBN网络与BP神经网络PID控制的永磁同步电机调速策略比较研究

Comparative study on speed regulation strategies of PMSM PID control based on DBN and BPNN

针对BP神经网络自身存在的学习速率固定、记忆不稳定等缺点,设计了一种基于DBN网络PID的永磁同步电机控制器,通过Matlab/Simulink对基于BP神经网络PID控制器的电机调速策略和基于DBN网络PID控制器的电机调速策略进行建模仿真分析,探讨两者对于PMSM调速策略中控制鲁棒性和稳定性的优劣。仿真结果表明,基于DBN网络PID的永磁同步电机调速控制策略训练效果更佳,具有更好的稳定性和鲁棒性。

Aiming at the shortcomings of fixed learning rate and unstable memory of BP neural network(BPNN),the PID permanent magnet synchronous motor controller based on deep belief network(DBN)is designed.The motor speed regulation strategy based on BPNN PID controller and the motor speed regulation strategy based on DBN PID controller are modeled and simulated by Matlab/Simulink,and their advantages and disadvantages for control robustness and stability in PMSM speed regulation strategy are discussed.The simulation results show that the training effect of permanent magnet synchronous motor speed control strategy based on DBN PID is better,and has better stability and robustness.