Fault Diagnosis Of Six-phase Permanent Magnet Synchronous Motor Controller

The six phase permanent magnet synchronous motor with low voltage and high power,small torque ripple and system reliability is widely used in all electric ships driving system,pure electric and hybrid electric vehicle traction system of high power level and high reliability fields.Six phase permanent magnet synchronous motor control system is generally composed of motor,controller and sensors and other parts,in which the controller is prone to failure,its reliability is very important to the whole system's normal operation.Therefore,it is very important to diagnose the fault of six phase permanent magnet synchronous motor controller.In this paper,a fault diagnosis method for six phase permanent magnet synchronous motor(PMSM)controller based on wavelet packet analysis,manifold learning and support vector machine is proposed.Firstly,a mathematic model of six-phase permanent magnet synchronous motor is established,and its vector control is realized by simulation.The structure and working principle of the six-phase permanent magnet synchronous motor are introduced,and the mathematical model under the natural coordinate system and the two-phase rotating coordinate system is deduced.Simulink is used to simulate the mathematical model of six-phase permanent magnet synchronous motor and its vector control.Secondly,the wavelet packet analysis is used to extract the fault feature vector of six phase stator current.The theory of wavelet analysis and wavelet packet analysis are introduced in detail.The six-phase stator current in different IGBT faults is sampled.The wavelet packet is decomposed by wavelet packet analysis.The energy value of each node is calculated and the energy value is normalized as the fault feature vector.Thirdly,the manifold learning method is used to reduce the dimension of the fault feature vector of six phase permanent magnet synchronous motor controller.This paper introduces the theory of manifold learning and its classical algorithms in detail,and uses the local cutting algorithm in the manifold learning method to reduce the dimension vector extracted by wavelet packet analysis.Finally,the IGBT fault of six phase permanent magnet synchronous motor controller is identified by least square support vector machines.This paper introduces the support vector machine theory and the least squares support vector machine theory which is further developed in support vector machine theory.The reduction eigenvectors of local tangent vector space algorithm are used as the training samples and test samples of the least squares support vector machine respectively.The different IGBT open faults are verified by simulation experiments.The simulation results show that the method of fault diagnosis can realize accurate positioning of different IGBT open faults in theory.