Research On Fault Diagnosis Method Of Motor Based On Wavelet Transform Of Vibration Signal

As mechanical power equipment,the motor is one of the most common equipment in modern production,widely used in agricultural machinery,petrochemical industry,ship power and other fields,its operation status directly affects the working status of the equipment.Once the motor failure,the light will reduce the production efficiency,the heavy will cause economic losses,and endanger the safety of life and property.However,the structure of the motor is more complex,the working condition is bad,the possibility of failure is greater.Therefore,it is of great significance to improve the operation reliability of motor equipment to monitor the running state of the motor in real time,make sure that the motor has no fault,and obtain the fault type in time when the fault occurs.In the field of motor fault diagnosis and monitoring,it is an effective way to obtain the running state of motor to analyze the vibration signal generated during its operation.This thesis selects the wavelet threshold denoising method for denoising by analyzing and comparing the advantages and disadvantages of three methods,namely wavelet decomposition and reconstruction method,wavelet threshold denoising method and modulus maximum reconstruction filter denoising method.Based on the relationship between the wavelet transform modulus maximum(WTMM)and Lipschitz index,the singularities of the de-noised signals are analyzed,and the appropriate eigenvalues are extracted,which accurately reflect the local fault characteristics and the singularities.In this thesis,the three-phase quadrupole permanent magnet synchronous motor is studied,sets up an experimental platform,simulates the motor stator's fault of different degrees of inter-turn short circuit and foundation looseness on the platform,and designs a set of vibration signal acquisition system,realizes the real-time acquisition of vibration signal under different faults.On the basis of theoretical analysis,this thesis carries out signal acquisition,wavelet noise elimination,singularity analysis and feature extraction for motor vibration signals in six operating states: normal operation of the motor,slight loosening of the base,severe loosening of the base and 10%,20% and 30% interturn short-circuit fault of the motor.The extracted eigenvalues are used to construct the eigenvectors,the eigenvectors are input into the probabilistic neural network to create the network model,and the fault diagnosis test is carried out.The results show that the diagnosis method based on the wavelet transform of vibration signal with probabilistic neural network can effectively judge the fault type and the fault severity of the motor.