模糊离散事件系统时间维度可诊断性分析

Diagnosability analysis of fuzzy discrete event systems in dimensionality of time

针对观测条件受限时的模糊离散事件系统(fuzzy discrete event systems,FDES)故障诊断问题,文章提出一种以时间维度可观程度(observable degree in dimensionality of time,ODDT)为基础的故障诊断器构造方法。根据观测信息中的时间属性建立时间约束图,并以时间证据合并的策略分析离散事件在时间维度上的可观性,从而获得FDES的时间信息全局状态;计算ODDT,并用于分析FDES的系统状态和故障可诊断程度。该方法可在观测条件受限以致故障事件链难以完全捕获的情况下,根据状态信息隐含的时间线索来辨识模糊事件并发现系统隐患。实例分析表明,相比于单纯依赖故障事件链分析的诊断方法,根据ODDT构造的诊断器提高了系统的可诊断程度。

Aiming at the fault diagnosis of fuzzy discrete event systems（FDES） under restricted obser- vation conditions, a fault diagnosis device construction method based on observable degree in dimen- sionality of time（ODDT） is presented. The time constraint graph is established according to time at- tribute in the observation information, and the strategy of time evidence combination is applied to e- valuating the observability of discrete event in dimensionality of time, so as to acquire the time infor- mation global state of FDES. The ODDT is calculated and used to analyze the state and fault diagnos- ability of FDES. According to the time clue implied in state information, the proposed method can trace fuzzy events and investigate insecure factors even if the observation conditions are restricted or the knowledge acquisition is incomplete. It is shown that, compared with traditional diagnosis ap- proach based on fault event chain analysis, the diagnosis device based on ODDT has higher diagnos- ability.