Displacement Self-sensing Control of Permanent Magnet Assisted Bearingless Synchronous Reluctance Motor Based on BP Neural Network Optimized by Improved PSO

Abstract

In order to solve the problems of low reliability, low integration, and high cost brought by mechanical sensors in the control system of permanent magnet-assisted bearingless synchronous reluctance motor (PMa-BSynRM), a displacement self-sensing method of the back propagation (BP) neural network left-inverse system under the optimization of an improved particle swarm algorithm is proposed. Firstly, the working principle of PMa-BSynRM is introduced, and the mathematical model of PMa-BSynRM is derived. Secondly, the suspension force model is established to prove the left reversibility of the PMa-BSynRM displacement subsystem on the basis of the observation principle of the left reversible system. Thirdly, the weights of BP neural network are optimized by using the improved particle swarm algorithm to avoid local optimum, and the final weights are obtained to complete the construction of the displacement self-detection control system. On this basis, velocity change and anti-interference simulations are conducted to prove the tracking performance of the displacement system. Finally, static suspension, velocity change and anti-interference experiments are executed which verify the accuracy and feasibility of the proposed displacement self-detection system.