An adaptive resonance demodulation method for rolling bearings based on eemd

Abstract

The invention discloses an EEMD-based adaptive resonance demodulation method for a rolling bearing. The method includes the following steps that: step 1, an acceleration sensor is installed near the rolling bearing so as to collect the vibration signals of the bearing during the operation of the bearing; step 2, EEMD and reconstruction are performed on the collected vibration signals, so that theadaptive filtering of the original signals can be realized; step 3, the parameters of a Chebyshev type I bandpass filter are initialized; and step 4, a grid search algorithm under a kurtosis entropy ratio indicator is adopted to adaptively optimize the center frequency and bandwidth of the filter; and step 5, envelope demodulation is performed on the filtered signals, so that an envelope spectrumis obtained, and a diagnosis result is obtained on the basis of bearing-related fault characteristic frequencies. According to the EEMD-based adaptive resonance demodulation method for the rolling bearing of the invention, secondary filtering is adopted, the bearing fault signals can be enhanced, and the adaptability of resonance demodulation is realized. The method has higher reliability and diagnostic precision than other methods do.

Description

technical field [0001] The invention relates to the technical field of mechanical engineering, in particular to an EEMD-based adaptive resonance demodulation method for rolling bearings. Background technique [0002] As an important component of rotating machinery, bearings are prone to failure due to long-term work under harsh conditions of high speed and alternating loads, resulting in downtime or catastrophic accidents. Therefore, the research on fault diagnosis of rolling bearings has an important reality. significance. In actual engineering, the vibration signal of rolling bearings is non-stationary and strongly doped with background noise and multiple interference sources, which leads to complex and difficult extraction of early fault features of bearings. Therefore, enhancing the signal features of early fault bearings is very important to improve the diagnostic accuracy. [0003] In view of the non-stationarity of rolling bearing fault signals, scholars at home and ...

AI Technical Summary

Problems solved by technology

However, the selection of the center frequency and bandwidth parameters of the bandpass filter in the traditional resonant demodulation technology is not adaptive, which greatly limits the promotion of the resonant demodulation technology.

[0004] Domestic patents involving resonance demodulation methods include "An Adaptive Resonance Demodulation Method for Rolling Bearing Fault Diagnosis" (201210429262.0), based on the envelope spectrum kurtosis values ​​of signals in different frequency bands, the control single variable method is used to determine the filter's The center frequency and bandwidth parameters do not optimize the filter center frequency and bandwidth parameters at the same time, making the determined filter parameters not necessarily optimal

Invention patent "A Method for Early Fault Diagnosis of Rolling Bearings" (201810276729.X), uses wavelet and time-domain synchronous averaging for preprocessing, and then analyzes the preprocessing signal by segmental correlation resonance demodulation method, but for early fault signals, Since the fault impact is not obvious and the background noise is large, simply borrowing the cross-correlation coefficient index may not be able to extract early fault features well

Images