Laplace wavelet basis sparse representation dictionary construction method based on waveform impact matching

Abstract

The invention discloses a Laplace wavelet basis sparse representation dictionary construction method based on waveform impact matching. The Laplace wavelet basis sparse representation dictionary construction method comprises the following steps: step 1, setting an initial value of a Laplace wavelet support interval; step 2, re-determining a support interval of the wavelet according to a 3 criterion of Laplace wavelet waveform energy; step 3, determining the Laplace wavelet with the optimal shock waveform matching degree with the bearing fault vibration signal; step 4, constructing a Laplace wavelet basis sparse representation dictionary after waveform impact matching through a point-by-point time shifting method; step 5, calculating a correlation coefficient between each atom in the Laplace wavelet basis sparse representation dictionary and a bearing fault vibration signal. According to the invention, the defect that a traditional Laplace wavelet correlation filtering method can only select appropriate wavelet correlation parameters through a large amount of priori knowledge and human experience is overcome. The method is not only suitable for constructing Laplace wavelet basis dictionaries, but also suitable for constructing Morlet wavelets and other wavelet-based dictionaries, and a new method is developed for constructing the wavelet-based dictionaries in the sparse decomposition theory.

Description

technical field [0001] The invention relates to the technical field of mechanical fault diagnosis and signal processing, in particular to a Lapalce wavelet-based sparse representation dictionary construction method based on waveform impact matching. Background technique [0002] In actual engineering, the fault characteristics of fault vibration signals of rotating mechanical components such as rolling bearings and gears appear as a series of shock response waveforms with a certain time interval in the time domain waveform, but these waveform shock characteristics are often covered by strong background noise in the signal. It is difficult to identify accurately, so it is very difficult to extract the fault characteristics corresponding to the waveform shock under the interference of strong background noise. The sparse decomposition and representation method is used to analyze the signal, and the fault impact features are extracted through the sparse dictionary representation...

AI Technical Summary

Problems solved by technology

Although the traditional Laplace wavelet correlation filtering method can theoretically find the Laplace wavelet correlation parameters corresponding to the fault impact, it is extremely difficult in practical application: ① The noise frequency component is prominent under the interference of strong background noise in the signal, and it is impossible to find The peak value of the main frequency components in the signal spectrum is used to determine the accurate oscillation frequency; ② It is difficult to select the two parameters of Laplace wavelet oscillation frequency and viscous damping ratio. If the range is small, it may not be possible to find accurate parameters; ③At present, the oscillation frequency and viscous damping ratio are mainly selected through human experience, which is greatly affected by subjective factors on the one hand; on the other hand, a large amount of prior knowledge is required , which limits its application in engineering applications

In the prior art, there is no relevant technical record that can well solve the above-mentioned problems

Images