Optimization of Rotor Position Observer with BP Neural Network

This paper presents an optimization method of high frequency signal injection rotor position observer based on BP neural network, which improves the sensorless control technology. Firstly, a permanent magnet synchronous motor (PMSM) model is established. Then, introduce how to build BP neural network model. Next, the rotor position observation data used in the training of neural network is optimized. Finally, the parameters of the rotor position observer are optimized by using the proposed BP neural network to realize the optimal control of PMSM. In the simulation, the optimization effect of BP neural network is verified.