Resonance Band Selection Method Based on Singular Value Component Frequency Domain Spectrum

Abstract

The invention discloses a resonance band selection method based on singular value component frequency domain spectrum, and relates to the technical field of bearing testing methods. The method includes the following steps: performing FFT frequency domain transformation on the faulty bearing signal collected by the acceleration sensor; determining the number of regions divided by the frequency spectrum according to the number of main frequencies of the signal spectrum; constructing a Hankel matrix from the collected faulty bearing signal , perform singular value decomposition on the matrix to obtain singular value components; select components for envelope demodulation analysis, and determine the center frequency and bandwidth of components by the maximum value of the envelope spectrum kurtosis. The method can not only self-adaptively select the optimal resonance band, but also improve the signal-to-noise ratio, greatly improve the diagnostic accuracy, and lay a good foundation for the subsequent bearing fault diagnosis.

Description

technical field [0001] The present invention relates to the technical field of bearing testing methods, in particular to a resonance band selection method based on singular value component frequency domain spectrum Background technique [0002] Among many rolling bearing fault diagnosis methods, resonance demodulation technology has been widely used due to its accuracy, but in this method, the setting of band-pass filter parameters requires a wealth of professional knowledge and a large amount of experience accumulation, resulting in resonance demodulation in Promotion in engineering applications is limited. In addition, the noise in the vibration signal of the rolling bearing also seriously affects the accuracy of the resonance demodulation fault diagnosis. In order to suppress the noise and improve the signal-to-noise ratio, it is often necessary to denoise the signal. [0003] Document 1 (MCFADDEN P D, SMITH J D. Vibration monitoring of rolling e1-element bearing by the ...

AI Technical Summary

Problems solved by technology

However, this kind of kurtosis map is not convenient to use in actual engineering, and it is too time-consuming and energy-consuming

Document 3 (Guo Daifei, Gao Zhenming, Zhang Qiangqiang. Using wavelet threshold method for signal denoising [J]. Shandong University Journal (Natural Science Edition), 2001,36(3); 306-311.) proposed wavelet denoising, Although it has the advantages of multi-resolution, the selection of threshold and basis function needs to be determined by human experience. If the deviation is too large, it will directly affect the noise reduction effect.

Document 4 (HUANG N E, SHEN Z, LONG S R, et al. The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis[J]. Proceeding of Royal Society London: A, 1998.454(1971 ):903-995.) The empirical mode decomposition noise reduction method used can decompose the signal into different frequency bands, and then transform it into a stable signal for further analysis. Although it is not necessary to consider the selection of parameter thresholds and basis functions, there are Certain frequency aliasing issues

Images