Study And Application Of Wavelet Analysis On Fault Diagnosis For Gearbox

Wavelet Analysis is a theory developed in the past decade. It has been successfully used in image processing, telecommunication and Geophysics. Wavelet Packet Transform is a generation of Wavelet Transform, and both of them are the main block of Wavelet Analysis. In this thesis, some research and application work with wavelet and wavelet packet in the field of fault diagnosis of GearBox has been conducted. And wavelet analysis, a new time and frequency analysis method, is used in the field of Mechanical Fault Diagnosis. This work comprises of the following parts :Firstly, some research on diagnosis theory and fault signal is carried out and concluded that the diagnosis process needs wavelet analysis; Secondly, theory basis of wavelet analysis is researched. Emphasis to its explanation of engineering is given. According to analyze concrete performance, the appropriate wavelet function for diagnosis is obtained; Thirdly, research on the application of wavelet transform in nonsteady signal is conducted. It is made out that wavelet analysis in these signal has a great advantage over fourier analysis. From wavelet analysis, the fault feature of high speed wire rod processer in the Heavy Section Mill of WISCO is successfully extracted and from this effective method the same kind accident can be avoided and the fault situation can be confirmed for the future. Fourthly, research on the application of wavelet transform in signal noise reduction is conducted. The approach of noise reduction based on wavelet denoise and wavelet packet denoise is proposed. The results of experiments with simulation signals and real signals show that the approach can be applied to noise reduction effectively. Lastly, research on the application of wavelet transform in feature extraction is conducted and demonstrated its merit on fault diagnosis according to the experiments of real signals.From above all, some important feature that can not be obtained by other method like fourier analysis can be got by wavelet analysis in mechanical fault diagnosis. A feasible and effective method is provided for detecting hidden trouble of equipment and analyzing fault.