Study On Gear Fault Diagnosis Technology Based On Gear Meshing Morphology

Gear is the most critical and widely used power transmission components and is also the failure part which has a high fault rate,Therefore,it is of great significance for fault detection of gear transmission system.The disadvantages of traditional methods of gear fault diagnosis based on the vibration response signal exist that that signal-to-noise ratio is not high and fault feature extraction is difficult.Therefor,a method of gear fault diagnosis based on the gear meshing morphology is proposed in this thesis.The gear fault diagnosis is realize through analysis of the gear meshing error signals of gear system.Study on the theory analysis of the gear meshing,the detection method of gear meshing error signal,and the analysis of meshing error signals is focused on in this thesis.the meshing error mathematical model of the gear system is established by analyzing the meshing state of the gear transmission system and comparing the comprehensive error of the normal gear transmission system and the faulted gear transmission system.According to the measurement method of gear meshing error,the existing gear fault simulation platform is reformed,the data acquisition card based on single chip microcomputer is designed to obtain the instantaneous meshing error of the gear system by measuring the angular displacement signals of the two axes.The EEMD decomposition method is used to decompose the gear meshing error signal into IMF components.The correlation coefficient method is adopted to select the IMF components with fault information.The remaining IMF components are reconstructed into new signals,and the new signals are analyzed by order spectrum and reverse order spectrum to obtain the periodic signals with fault information.Finally,the BP neural network is established to identify the different fault types of the gears,by using the energy values of different orders segments as input feature vectors.The result of experiments proves that the analysis method based on EEMD and order spectrum analysis has an effective effect on the gear meshing error signal analysis and can effectively extract the fault feature,the BP neural network can effectively identify the type of gear failure.