Diagnostic Method of Bearing Failure Based on Generalized Jaccard Coefficient and False Degree Melting

Resumen

The generalized Jaccard coefficient is used to assess similarity in evidence. The defects of the conventional Dempster/Shafer(D-S) method are solved and the diagnosis of bearing defects is improved. Determine the confidence of the evidence, weight the Basic Probability Assignment (BPA) value of the original evidence with the confidence and falsehood weighting coefficients, and combine them with the Dempster compositional rule. Determine the reliability of the evidence, weight the BPA value of the original evidence with the confidence and falseness weights, and combine them with the Dempster composition rule. The supporting vibration signal is broken down by the transformation of wavelet packages, the empirical decomposition and the variational decomposition, and the set of default vectors is extracted by combining the sample entropy. The correlation vector machine is used to diagnose and produce the probability of defect, which is the basic probability of D-S evidence theory, for fusion diagnosis. The test results of the defect diagnosis method, the correlation vector machine and the improvement of the D-S evidence theory have good accuracy and reliability in the application of the support defect diagnosis.