基于RBF神经网络的PMLSM无速度传感器控制

Speed Sensorless Control of PMLSM Based on RBF Neural Network

提出了一种基于径向基函数(RBF)神经网络的永磁直线同步电机(PMLSM)无速度传感器控制方法。该方法构造了一种RBF神经网络速度观测器,利用RBF神经网络逼近定子电流和速度之间的非线性关系,从而实现对电机速度的精确观测。通过Lyapunov理论设计了学习率的取值方法,以保证神经网络算法收敛性。仿真和实验结果表明,该方法具有良好的动静态特性、鲁棒性强、适用速度范围广。

A speed sensorless control method of permanent magnet linear synchronous motor(PMLSM) based on radial basis function(RBF) neural network is proposed.This method constructs a RBF neural network speed observer,and uses the RBF neural network to approximate the non-linear relationship between the stator current and speed,so as to achieve accurate observation of the motor speed.Based on Lyapunov’s theory,a learning rate method is designed to ensure the convergence of the neural network algorithm.Simulation and experimental results show that the method has good dynamic and static characteristics,strong robustness,and wide applicable speed range.