Fault Diagnosis Using Multi-Source Information Fusion

Fault diagnosis plays an important role in ensuring system reliability. Accurate fault diagnosis relies on data, fault features, diagnostic knowledge, reasoning, and decision-making. The following issues such as certainty and sufficiency of collected data, importance of features in diagnosis, and accuracy of diagnostic knowledge affect the accuracy of fault diagnosis greatly. Dempster-Shafer (D-S) evidence theory based multi-information fusion can provide a mechanism for representation of uncertain and imprecise information. However, it cannot directly handle the issues of evidence sufficiency, evidence importance, or conflicting evidences. An improved D-S evidence theory through the introduction of a fuzzy membership function, importance indexes, and conflict factors is reported in this paper. Experiment analysis results indicate that the proposed method is effective for practical fault diagnosis