Research On Fault Diagnosis Method Of Elevator Permanent Magnet Synchronous Traction

In recent years,China’s centralized buildings have generally developed to the upper levels,so the application of elevators is becoming more and more common.It can be said that elevators have become a complex special equipment that is indispensable in the daily work and life of modern people.Elevator safety and reliability are related to people’s lives and property safety.Therefore,ensuring the safety,reliability and stability of elevator operation has become the key and core of modern elevators.The permanent magnet synchronous traction in the elevator is one of the key equipments of the elevator,and its working state is the guarantee and premise of the safety for the entire elevator system.The research on the fault diagnosis method of permanent magnet synchronous traction machine makes it possible to quickly locate the faults and fault types of the permanent magnet synchronous traction machine of the elevator.This is also the trend of safe development of elevators.This paper mainly studies the fault diagnosis method of stator inter-turn short circuit fault and rotor dynamic eccentricity fault of elevator permanent magnet synchronous traction machine.Firstly,the composition of the permanent magnet synchronous traction motor of the elevator and the common fault types are expounded.The mechanism of the stator short circuit fault and rotor dynamic eccentricity fault are analyzed.Thereby,the corresponding fault characteristic frequency is obtained.The KT-6000 high-power elevator permanent magnet synchronous traction motor used in the vertical GLR-M-300 elevator is taken as the research object.Modeling and simulating the two kinds of fault of the permanent magnet synchronous traction machine of the elevator by using Ansoft Maxwell.An eight layers wavelet packet method for the faulty stator current signal using wavelet basis function is proposed.The band energy of the frequency band to which the fault characteristic frequency belongs is extracted to form the fault feature vector.According to this method,the fault feature vectors of 600 sets of current signals in different operating states of the elevator permanent magnet synchronous traction machine are extracted.These vectors are taken as the training and test samples of the diagnosis model of the LS-SVM,and compared with the diagnosis model of the same support vector machine under the same experimental conditions.Through experiments,it is verified that the fault diagnosis of the elevator permanent magnet synchronous tractor based on the least squares support vector machine has a high accuracy.In order to optimize the hyperparameters of the tractor fault classifier based on the LS-SVM,an improved particle swarm optimization algorithm based on the combination of inertia weight and convergence parameters was introduced.Then the high efficiency and accuracy of the fault classification algorithm are verified by comparative experiments.