Bearing fault signal noise reduction method based on minimum mean square empirical mode decomposition

Abstract

The invention discloses a bearing fault signal noise reduction method based on minimum mean square empirical mode decomposition. The method comprises the following steps: 1) decomposing an acquired bearing signal by using a CEEMDAN method, and decomposing the signal into different mode spaces; 2) taking the IMF components in different modes as normalized adaptive filtering source signals, adding white Gaussian noise as an expected signal, and filtering the IMF components in different modes by adopting a normalized adaptive filtering method, wherein the difference between the source signals andoutput signals is the IMF components after adaptive filtering; and 3) reconstructing the IMF components of each mode after filtering. Normalized least mean square complete set empirical mode decomposition is applied to noise reduction of the bearing fault signal, the steps are simple, implementation is easy, the noise of the bearing fault signal can be effectively eliminated, fault feature information is enhanced, and the signal to noise ratio of the signal is increased.

Description

technical field [0001] The invention relates to a bearing fault signal processing method, in particular to a bearing fault signal noise reduction method based on least mean square empirical mode decomposition. Background technique [0002] It is the core goal of signal noise reduction to separate the essential features existing in the fault signal from other interference parts, and to mine the fault feature value deeply to obtain as much effective feature information from the fault signal as possible. Since the formation of bearing faults is a slow and complicated process, the fault signal is weak and relatively stable, and a large amount of interference and noise are easily collected during the monitoring process. The weak fault signal is submerged due to noise interference, and it is difficult to extract the fault characteristic information of the signal. [0003] As an effective method for extracting information, EMD has been continuously improved and matured, and has be...

AI Technical Summary

Problems solved by technology

Since the formation of bearing faults is a slow and complicated process, the fault signal is weak and relatively stable, and a large amount of interference and noise are easily collected during the monitoring process. The weak fault signal is submerged due to noise interference, and it is difficult to extract the fault characteristic information of the signal

Images