WAVELET CONSTRUCTION BASED ON DATA FITTING AND BEARING FAULT DIAGNOSIS

Abstract

Constructing the optimal wavelet function is the key to realize lifting Wavelet de-noising.After determining the optimal decomposition level and effective threshold, it tries to introduce data fitting methodin the process of interpolation subdivision. By selecting different basis function, sample points and the dimensions of the basis function, the new wavelet function is constructed with different smoothness, shock and disappear moment, it introducesthe similarity coefficient to analyze the constructed wavelet function and select the optimum wavelet function.It combines with redundant lifting wavelettransform to improve the predict operators and update operators. The superiority of the wavelet function in vibration signal de-noising is verified by simulation and engineering data analysis, this method is helpful for bearing fault diagnosis.

Keywords

• De-noising;Wavelet structure;Data fitting;Fault diagnosis