Application of fuzzy data fusion theory in fault diagnosis of rotating machinery

In this article, the most common induction motor faults including bearing outer race defect, broken rotor bar, and short-circuit of stator windings are diagnosed with high reliability. The decentralized fuzzy-integral data fusion method is used for information fusion in feature level. In the proposed scheme, the feature vectors are constructed using signatures created by time-domain characteristics obtained from stator three-phase current measurements. Partial matching of each feature is calculated by the fuzzy c-mean classifier algorithm, and features with high diagnosis ability are fused by Choquet fuzzy integral. The technique is validated experimentally on the 4hp induction motor of an electropump, and the results are presented.