Application Research On Fault Diagnosis Of Gear Motor Based On Vibration Signal Analysis

Gear motors are the key basic equipment for industrial production and are widely used in various industrial production fields.Realizing the monitoring of gear motor health is the key to ensuring the safe and continuous progress of industrial production processes.It is of great importance to carry out research on gear motor fault diagnosis technology significance.The vibration signals generated by various components during the operation of the geared motor can accurately reflect the health status of the geared motor.Therefore,this paper analyzes the vibration signals generated during the operation of the geared motor to diagnose the rotor eccentricity and bearing faults of the geared motor.The main contents of this paper are as followsFor the rotor eccentricity fault of the geared motor,because the actual vibration signal of the rotor of the geared motor contains a strong noise component,it is difficult to extract the fault feature using its synthetic rotor axis trajectory.This paper first uses harmonic wavelet to purify the rotor vibration signal,and extracts the affine invariant moment feature of the axis trajectory graph synthesized by the purified vibration signal,and uses the BP neural network to simulate the different types of eccentric faults of the geared motor rotor.The characteristics of the invariant moments are classified to identify the rotor eccentric fault diagnosis of the geared motor,and the performance of the rotor eccentric fault diagnosis method of the geared motor is verified by simulation experiments and the design of the rotor eccentric fault diagnosis system.The experimental results show that the fault The diagnosis method can accurately identify several typical eccentric faults of the geared motor rotor.Aiming at the bearing faults of geared motors,this paper has designed a gearbox motor fault diagnosis method based on one-dimensional convolutional neural network.This method can directly extract the fault features from the time-domain vibration signals generated by the bearings during the operation of the geared motors and classify them Identification,the performance of the fault diagnosis method is verified on the CWRU bearing data set.The experimental results show that the fault diagnosis method has a fault diagnosis recognition rate of more than 99.59%,and has strong variable load adaptability.In addition,this paper also designs a two-dimensional convolutional neural network with deep separable convolutional layer gear motor bearing fault diagnosis method,which can convert the time-domain vibration signal generated by the bearing during the operation of the gear motor into a grayscale image Extract the fault features and carry out classification and identification.The performance of the fault diagnosis method is also experimentally verified on the CWRU bearing data set.The experimental results show that the fault diagnosis method has a fault diagnosis recognition rate of 100%.Compared with other reducer motor bearing fault diagnosis methods,the above two reducer motor bearing fault diagnosis methods save the signal analysis step and directly extract fault features from time-domain vibration signals to achieve accurate reducer motor bearing fault diagnosis.