Fault Identification And Diagnosis Of Rolling Bearing Based On Information Fusion Technology

The rolling bearing is a mechanical device used for the important parts, prone to damagein the device during operation. Once the bearing failure occurred during the work, willdirectly affect the normal safety of production, and even lead to the occurrence of accidents.Generally as a result of bearing’s work is non-steady state, fault features are complex anddifficult to extract effective fault characteristic information and diagnose the fault using asingle method in bearing fault diagnosis process. According to statistics, According tostatistics, because the occurrence of the bearing fault caused by the accident will cause a lotof economic losses. Therefore, a reasonable and effective checks and extract fault featuresof rolling bearings, to avoid accidents is very important.In this paper, the signal is collected by vibration signal test technology, the research ofroller bearing fault identification and diagnosis based on multi-sensor information fusiontechnology, combined with wavelet packet theory, EMD methods, multi-resolution SVD,multi-resolution SVD package, multi-sensor probabilistic neural networks and supportvector machine. Research contents are as follows:Firstly, the process of information fusion technology and its current development andresearch in fault diagnosis and recognition were discussed, and elaborated informationfusion technology to the impact and significance of mechanical science. The division levelinformation fusion and information fusion algorithm were introduced. Gearbox vibrationand noise generation mechanism and vibration mechanism of rolling bearing were analyzed.Secondly, through the application of multi-sensor information fusion technology, weconstructed extraction for rolling bearing fault feature of the multi-domain feature extractionalgorithm and IMF entropy feature extraction algorithm. Multi-sensor networks fusionalgorithm used for rolling bearing fault identification and diagnosis. American Spectra Questcompany produces the comprehensive mechanical fault simulation test bench is used toidentify the fault type of rolling bearing. Rolling bearing fault recognition of mill gearhousing were done. Finally, combining the characteristics of multi-resolution analysis and SVD feature, thepaper applied multi-resolution SVD and multi-resolution SVD package to extract thecharacteristics of rolling bearing, in combination with support vector machine (SVM). Faultidentification and diagnosis conducted with multi-resolution SVD feature fusion and multi-resolution SVD package feature fusion and support vector machine (SVM). AmericanSpectra Quest company’s power transmission fault diagnosis comprehensive simulation testbench is used to identify the fault type of rolling bearing.The final experiment results show that the multi-domain feature, the IMF entropyfeature, multi-resolution SVD feature, multi-resolution SVD packet feature can effectivelyextract and express the different fault types of rolling bearing. The feature fusion recognitiondiagnosis and support vector machine feature fusion recognition diagnosis, can accuratelyclassify the different rolling bearing fault, so as to prove that these methods can effectivelysolve the problem of rolling bearing fault.