Research On Fault Diagnosis Method Of Typical Gear Fault Based On Kurtosis And Wavelet Analysis

As one of the most important power transmission parts, gear has been widely used in machinery, military, aerospace and other industrial production, gear good operation is very important to the whole equipment.The current detection method of the gear state only in relying on artificial experience or simple threshold judgment, automatic on-line fault diagnosis can not accurately realize.Because the operating state of the gear vibration signal can be good reaction gear, and is sensitive to the fault. Therefore, it is very necessary to study a method to obtain the fault information from the vibration signal.How to get the characteristic frequency of fault accurately is the key to fault diagnosis. To this end, the vibration of the gear is analyzed in depth.The vibration mechanism and vibration signal characteristics of different fault types are studied, and the modulation model of the gear vibration signal is established. In the process of rotating gear meshing, the tooth surface damage will impact its meshing gear periodically, resulting in low frequency vibration, the vibration frequency is the fault characteristic frequency. The gear vibration signal is nonlinear, transient and non-stationary signal, time domain and frequency domain signal processing methods cannot achieve the conventional detection of the signal.To solve the above problem this paper proposes a method of gear fault diagnosis based on kurtosis, wavelet analysis.The simulation signal is established according to the vibration characteristic and the modulation model of gear pitting failure condition. The simulation signals were analyzed by Matlab software, and verified the validity of the method. The experimental platform for gear fault diagnosis is designed, and a test system for signal acquisition is builded. The time domain waveform analysis, amplitude spectrum analysis, envelope spectrum analysis of the pre processed data are performed respectively. The advantages and disadvantages of various methods are compared through simulation. The typical fault signals of gear tooth broken are analyzed by using wavelet decomposition, and the feasibility of the method is verified.Finally, in order to improve the accuracy of gear fault diagnosis, the fault diagnosis based on wavelet packet analysis is also carried out. Wavelet packet analysis can improve the frequency resolution of the high and low frequency, and wavelet packet reconstruction can effectively filter out the interference, so as to realize the accurate extraction of fault frequency. The experimental results show that the typical gear fault frequency can quickly and effectively extract though the fault diagnosis method,which is based on kurtosis and wavelet analysis.This paper not only provides an online gear fault diagnosis method, but also provides a basis for improving the extraction accuracy of fault frequency, which is of great significance to the development of automatic fault diagnosis instrument.