Research Of Method Of Harmonic Wavelet Analysis For Damage Of Gear

With the development of science, technology and the increasing of automatic degree of mechanical system, the technology of fault diagnosis for mechanical system is being valued and perfected increasingly. At the present time, gear drive which is used widely is still one of the most general modes in mechanical system. However, because the probability caused gear fault is proportional to 10.3 percentage of mechanical failure, and it is the key factor which inducts equipment failure, therefore that it is necessary to detect the fault for gear is a reasonable research work.During the mechanical system running, the normalvibration of gear drive always exists, when local fault happened in gear (such as pitting, crack of local dedendum, etc.), in the process of gear mesh, it generates periodic, instant strike, and forms vibration of strike. For the transmission path of vibration is very complicated, the vibrating signal is much weak, and useful information of fault is submerged in other frequency components and noises, which turns original stable signal into non-stationary or time-variant signal, so it is difficult to diagnose gear fault. The conventional methods of gear fault diagnosis, which consist of time domain and frequency domain analysis method, are effective to integral and dispersed defect of gear, but are not sensitive to weak fault signal, don't generally reflect the time-variant character of signal, and affect reliability and accuracy of the technology of fault diagnosis.This paper, according to the detachment and extract of the local fault of gear, on the basis of the wavelet transform,presented the method of fault diagnosis based on harmonic wavelet analysis and its application in fault diagnosis for gear drive. Firstly, it introduced damage forms and features, failure analysis of gear; it studied mechanism and characteristic of gear vibration, and common methods of fault diagnosis for gear, and dynamical analysis and features of vibration for signal. Secondly, it systematically represented the theory, character, algorithm of harmonic wavelet and its application. It is a complex wavelet, being astringe in frequency domain, and having definite function expression. Its algorithm speeds quickly, and has high accuracy and more good value of application. Finally, it presented the method of analysis based on harmonic wavelet, which mainly extracted weak singularity fault information for local fault of gear. The method has been verified by making experiment. Experiment results show this method is superior to conventional method of wavelet analysis in local fault diagnosis for gear.This paper, by means of theory and practical analysis, indicated that harmonic wavelet has many merits, such as excellent characteristic of time-frequency, good effectiveness of filtering and denoise, which easily extracts feeble (or singular) fault signal and supplies deficient information of other wavelets analysis. It effectively solves fault diagnosis of non-stationary signal, and is practical in gear fault diagnosis and of application and more thorough research value.