Research on Torque Observer of Permanent Magnet Synchronous Motor Based on Model Integration

Torque control is an important technology for permanent magnet synchronous motor (PMSM) control performance. However, in many cases, torque sensors are not suitable for scenes with high performance and size requirements such as robots. Therefore, it is necessary to conduct in-depth research on the non-sensor torque estimation. Due to the difficulty of motor parameter identification in the operation of permanent magnet synchronous motor, the motor parameters are greatly affected by harmonic interference, and the electromagnetic torque is difficult to be accurately estimated by mathematical model. A model integration algorithm is proposed to fit the nonlinear relationship between electromagnetic torque, motor current and angle. The algorithm contains the strong coupling ability of BP neural network and the stability of mathematical model. In addition, simulations and experiments were carried out on the proposed model integration algorithm. The results show that the torque estimation accuracy is consistent with the measured value of the sensor. Compared with the simple BP neural network model, the scale of the model structure is reduced, the model stability and the torque estimation accuracy are improved.