The Fault Detection System Of Rolling Bearing Based On The Vibration Signal

The vibration performance is one of the important indexes to measure the qualityof bearing, and the vibration detection is a necessary part of the bearing factoryinspection. However, at the present time, in the domestic bearing manufacturingenterprises, the vibration detection of the rolling bearing still relies mainly on themethod of combining "ear" and the simple vibration detection instrument, the accuracyis not high and the efficiency is low, which already can’t satisfy the producingrequirement of automation and in large quantities for the bearing enterprises.Therefore, it has an important practical significance to carry out the research onautomatic fault detection system of rolling bearing based on the vibration signal.This paper studied the basic structure of the rolling bearing, the vibrationmechanism, the fault type and the vibration components, and found that the defects onthe surface of bearing components are the main defects in the process of bearingproduction. So the defect detection on the surface of bearing element is the priority ofthis detection system, and the impact vibration caused by the defects on the surface ofthe bearing element was studied emphatically.The machining defects in the process of bearing production are very mild, belongto the early failure, and the vibration signal of bearing collected is disturbed largely bythe foundation vibration component, based on the resonance demodulation method, theEMD decomposition and the kurtosis criterion, a new method used in the faultdetection of the rolling bearing by eliminating basic foundation and improving EEMDwas proposed. Firstly, the EMD decomposition and the kurtosis criterion method wereused to get adaptively the component of the IMF which had the largest kurtosis, thenfor overcoming the shortage of that the selected IMF component which had the largestkurtosis in the extraction of the early fault feature, the process of eliminating the basicvibration of bearing was carried out, at the same time, the improved application of EEMD decomposition solved not only the interruption and intermittency of signal, andthe "pulse" problem caused by eliminating the basic vibration of bearing, and alsogreatly reduced the amount of calculation of the original EEMD decompositionmethod, to make the detection more efficient. Through a lot of experimental analysis,the results show that the fault detection method of bearing by eliminating basicfoundation and improving EEMD both retained the advantage of adaptive of the EMDdecomposition and the kurtosis criterion, and improved the capacity of extracting theearly fault feature of rolling bearing.On the basis of the fault detection method of bearing by eliminating basicfoundation and improving EEMD proposed in this paper, the fault detection system ofbearing is developed on the LabVIEW platform, and cooperated with the mechanicalstructure and the control system of mechanical motion based on PLC, the fullyautomated operation was realized in the vibration detection of bearing.Finally, in the automatic assembly line of a bearing manufacturer, The faultdetection system of rolling bearing based on the vibration signal developed in thispaper was tested for various types, mass, multiple groups. And the results show that,the fault detection system of rolling bearing based on the vibration signal developed inthis paper, the accuracy of which is more than99.9%, and the detection speed cansatisfy completely the production rhythm of3seconds in bearing manufacturingenterprises, so it has a high practical value and a broad application prospect.