Sensorless Control of PMa-BSynRM Based on ISWATS Adaptive Neural Network Left Inverse Soft-Sensing Technique

In order to solve the problems of large volume, high cost and low reliability through utilizing the mechanical sensors for permanent magnet assisted bearingless synchronous reluctance motor (PMa-BSynRM) and improve the tracking performance of sensorless detection, a sensorless control strategy of soft-sensing technique based on improved switch Adam to SGD (ISWATS) adaptive neural network left inverse system (ANNLIS) is proposed. After the introduction of the structure and principle, the left reversibility was deduced based on the flux linkage equation. Then, the left inverse system of the subsystem was constructed by using the fitting ability of neural network to realize the soft-sensing technique. The ISWAST algorithm is used to adaptively adjust the connection weights of neural network and the effectiveness has been verified. Finally, the sensorless control system of the PMa-BSynRM was constructed and the simulation research was carried out. The results indicate that the average peak error is approximately 1.5%, while the delay is merely 6%. This validates the effectiveness of observing speed and radial displacements of the rotor, thereby demonstrating its accuracy and feasibility.