Study On Fault Feature Extraction Of Rotating Machine Method Based On Blind Source Separation

With the development of modern industrial production, machinery fault diagnosis hasrapidly developed into an emerging discipline. The key to mechanical fault diagnosis is toexact fault features from the vibration signal. The Signal processing and analysis are theusual method for feature extraction. In recent years, there are more and more attentions tosignal processing in machinery fault diagnosis. The paper focuses on the blind sourceseparation (BSS) and the application in mechanical fault feature extraction.This paper deals with the research on BSS principle and algorithm. Aiming at themulti-sources mutual interference of mechanical vibration, propose the idea of blindsource separation applied to mechanical fault diagnosis.Firstly, analyzing the rotating mechanical failure signal, according to the strongbackground noise of rotating machinery working environment, combining the advantagesof the BSS and AR power spectrum, the empirical mode decomposition (EMD) is alsoused to translate non-stationary signals into stationary signals. In the simulationexperiments, though the mixed-signals contain strong noise, the method is successfullyseparated the unknown noise signal, and extract the characteristic information of the faultsignal. The result verify the effectiveness of the method in the mechanical fault featureextraction.Secondly, the BSS problem that the observed signals fewer than the number of thesource signals is studied. The extremum field mean mode decomposition (EMMD) is usedto increase the dimensions of observed signals. Furthermore, aiming at the unknownsource number BSS problem, discuss a source number estimation method based onEMMD and bayesian estimation. The unknown source number and single channel rotatingmachinery fault signal BSS experiment, verify the effectiveness of the method.