Bayesian Network-based Permanent Magnet Synchronous Motor Demagnetization Fault Diagnosis

Permanent magnet synchronous motors(PMSMs),as a high-performance drive equipment,occupy a central position in the field of high-precision control and industrial automation.Due to the permanent magnet synchronous motor with permanent magnet architecture,permanent magnet material is subject to high temperature and stress during operation,there is a risk of loss of magnetism,light affects the efficiency of the motor,heavy burn the motor,causing irreparable losses.Therefore,It is of great importance to ensure the safe and reliable operation of permanent magnet synchronous motors.In this dissertation,the current and vibration signals of the motor in the demagnetized fault state are investigated,the relationship between the current and vibration signals and the demagnetized fault is found,and a Bayesian network is established based on this,and the structure and parameters of the Bayesian network are optimized by simulating the model of the motor and using it for diagnosis.The main contents in this dissertation arc as follows.In order to obtain the relationship between the corresponding variables,we established a finite element model(FEM).The inherent relationship between the stator current,electromagnetic torque and different types and degrees of demagnetization of the motor is analyzed,and the connection between different order harmonics and different demagnetization faults is separated by using wavelet transformations.Mathematical analysis of the vibration signal roots of the motor was carried out,the relationship between the electromagnetic force and the electromagnetic torque was studied,and the relationship between the electromagnetic torque and the degree of demagnetization in different demagnetization states was simulated using finite element analysis,and the spectrum of the electromagnetic torque was obtained using wavelet transform.Combined with the previous analysis of current and torque,a Bayesian network(BN)of hybrid nodes is proposed,which can use the priori knowledge of the previous analysis to predetermine the relationship between some of the nodes,thus reducing the difficulty of training the Bayesian network and using it as a logical network for fault diagnosis to achieve classification of faults.To verify the effectiveness of the proposed method,a permanent magnet synchronous motor fault simulation test bench was established.The proposed method was validated by analysis and comparison of fault simulation data...