Abstract: The majority of the existing fault diagnosis methods using Dempster-Shafer (DS) evidence theory (DST) all provide the “static” fused results by combining several pieces of diagnosis evidence, which only reflect the current running status of monitored equipment. This paper presents the dynamic diagnosis strategy by using recursively the improved linear evidence updating rule. Its updated result can synthesize the diagnosis evidence collected at historical, current and future time steps by dynamically adjusting the proposed smoothing linear combination weights. The diagnosis examples of machine rotor show that the proposed method can provide more reliable and accurate results than the diagnosis methods based on the classical updating strategies.

Key words: dynamic fault diagnosis| information fusion| Dempster-Shafer evidence theory (DST)| evidence updating

摘要： The majority of the existing fault diagnosis methods using Dempster-Shafer (DS) evidence theory (DST) all provide the “static” fused results by combining several pieces of diagnosis evidence, which only reflect the current running status of monitored equipment. This paper presents the dynamic diagnosis strategy by using recursively the improved linear evidence updating rule. Its updated result can synthesize the diagnosis evidence collected at historical, current and future time steps by dynamically adjusting the proposed smoothing linear combination weights. The diagnosis examples of machine rotor show that the proposed method can provide more reliable and accurate results than the diagnosis methods based on the classical updating strategies.

关键词: dynamic fault diagnosis| information fusion| Dempster-Shafer evidence theory (DST)| evidence updating