Research On Gear Fault Diagnosisbased On MCSA

Gear-drive equipment has been widely applied due to its accuracy, reliabilit y, and wide range of speed and power. Gear fault diagnosis has important significance in reducing casualties and economic losses caused by industrial accidents. Tradit ional method of gear fault diagnosis, such as vibrat ion diagno sis, requires vibrat ion sensor and may be easily affected by e nvironment and no ise. Gear fault diagnosis based on motor current signature analysis(MCSA) regards motor stator current as the pointcut, detects gear fault fro m the perspect ive o f electric, and belongs to non-invasive measurement, which brings no disturbance to the system.The paper builds a two-mass model of gear-drive equipment, and introduces fault diagnosis mechanism of MCSA. MATLAB simulat ion based on the model is carried out to verify the mechanism. Then this paper adopts three signal processing techno logies, which are FFT, wavelet transform and empirical mode decomposit ion, to extract gear fault features from motor stator current collected in experiment and diagnose gear fault eventually.First ly, the paper adopts FFT analysis as the signal processing techno logy. FFT analysis is able to diagnose gear fault as long as the sampling t ime is guaranteed long enough. Then the paper studies about how gear fault diagnosis is affected by motor speed and load level, it co mes to a conclusio n that the capabilit y of FFT to diagnose gear fault decreases with the motor speed increasing, and improves with the load level increasing. In a word, high load level and low motor speed are beneficial to diagnose ge ar fault.Then wavelet analysis is introduced after illustrating the disadvantage of FFT analysis. Wavelet analysis is acco mplished through wavelet transform, the most important feature of which is mult i-reso lusion analysis. In this sect ion, wavelet transform is co mbined with FFT analysis and envelope spectrum analysis seperately to diagnose gear fault. Experiment results show that even though FFT failed, wavelet analysis is st ill able to eliminate the influence o f load level and speed, diagnose gear fault effect ively, and expand the range of gear fault diagnosis. Wavelet transform combined with envelope spectrum analysis has a better diagnost ic capabilit y.Finally, the paper introduces empirical mo de deco mposit ion, the most advantage of which co mpareded with wavelet transform is self-adaption and no need to select basis function. After explaining the theory of empirical mode decomposit ion, empirical mode deco mposit ion is applied to decompose motor current, then intrinsic mode funct ions of empirical mode deco mposit ion are analysed by envelope spectrum analysis, the results show its effect iveness. Considering that wavelet transform has the character o f band-pass filter, motor stator current is first ly analysed by wavelet transform, then the details o f wavelet transform are chosen as the object of empirical mode deco mposition, envelope spectrum analysis is applied at last to extract fault features, the results show that the fault features are more obvious than that of other methods ment ioned in the paper.