Research on PMSM Sensorless Control Based on Improved RBF Neural Network Algorithm

摘要

In this paper, a method of realizing PMSM sensorless control based on the improved radial basis function (RBF) neural network and high-frequency injection method is studied. The scheme uses the high-frequency injection method to realize the sensorless detection of the rotor position and speed, and adjusts the PI controller parameters online through the improved RBF neural network. The high-frequency injection method can realize the rotor position detection including zero speed. The improved RBF neural network can adjust the changes of hidden nodes online to obtain the best network structure, and can optimize the learning rate in real time to adjust the PI controller parameters. Finally, the experimental simulation is carried out. The results show that the scheme proposed in this paper can detect the position and speed of the PMSM rotor accurately, with fast response speed and good dynamic and static performance.