部分可观Petri网结构信息在故障诊断中的应用

Application of structural information of partially observed Petri net in fault diagnosis

中文关键词  离散事件系统  故障诊断  Petri网  故障函数

英文关键词  discrete event system  fault diagnosis  Petri net  failure function

中文摘要

      针对离散事件系统的故障诊断问题, 本文提出了一种基于部分可观Petri网结构信息的诊断方法. 它包括两 个部分, 第1部分利用故障变迁的可诊断子网确定故障变迁的可诊断性. 第2部分在故障可诊断的基础上提出一种 在线故障诊断方法: 首先, 利用Petri网的几种基本子网来分析故障变迁的可诊断子网的结构信息; 其次, 根据给定的 可观测变迁序列和可诊断子网的结构特征来描述子网内部托肯的流动形式; 最后, 定义故障函数, 并结合具体实例 来描述故障变迁的发生情况. 该故障诊断的方法基于部分可观Petri网结构信息, 无需遍历系统状态空间, 免去多项 式级的计算复杂性, 能够满足实时性的要求.

英文摘要

      For fault diagnosis problems in discrete event systems, we propose an approach based on the structural information of partially observed Petri nets. This work consists two parts. In the first part, the diagnosable subnet is used to determine the diagnosability of the fault transition. In the second part, a method is proposed to online diagnose the fault if it is diagnosable. First, several basic subnets are used to analyze the structural information of the diagnosable subnet of the fault transition. Then, the flow of tokens in the diagnosable subnet is described by the sequence of observable transitions and the structural characteristics of the diagnosable subnet. Finally, by defining the failure function and giving an example, we described the fault. With the structural information of partially observed Petri nets, it is not necessary to traverse all states of the system. Because this approach is with the computational complexity of polynomials only, it meets with the real time requirements.