Research On The Gear Fault Diagnosis Technology Based On EMD

Gear transmission is the main form of mechanical transmission. As the basis part of the mechanical equipment, its quality and performance decides the performance of the whole machine, and the healthy status of the equipment operation plays an important role. Once the gear failure, it is likely to cause the failure of the whole machine. Based on this, this paper will study on the gear fault feature extraction technology improvement to realize the study on the gear fault diagnosis.The traditional diagnosis method for non-stationary vibration signal analysis of the effects is not obvious. The EMD is decomposed the complex vibration signal into a finite number of IMF, and the IMF contains the actual physical information and it is smooth. This method can realize through the analysis of time domain, and get the fault characteristic signal.Aiming at the shortcomings of the traditional filtering get rid of the noisy method, this paper adopts the EMD method to deal with the noise of data signals in the preprocessing stage, and this method can eliminate the noise effectively. This paper aims at diagnosing the gear fault to study, using the method which based on EMD to process the vibration signal to realize fault diagnosis. Aiming at the endpoint effect which caused by the method of cubic spline interpolation,this paper will study the method of cycle continuation, extreme value point symmetry continuation and mirror continuation to analysis the effect of inhibition of the endpoint effect.This paper puts forward the method based on EMD and adaptive filtering to realize fault feature extraction. It can use EMD smooth the non-stationary signal, also AFD algorithm is simple, can improve the boundary oscillation. Finally, the experiment using QPZZ— Ⅱrotating machinery vibration simulation experiment platform. Through the replacement of the defective gear fault, the experiment try to realize the fault about gear tooth broken, crack and abrasive wear to verify the method which this paper used. The experimental results show that this method can effectively and accurately extract the fault characteristics to realize the gear fault diagnosis.