Fault diagnosis based on Petri Nets: the case study of a hydropower plant

:Identification and fault diagnosis of industrial discrete event system has been used Petri Nets models in classical Place/Transition approach or even high level nets. A very important issue in this approach is the system monitoring and the integration of structural and behavioral models observed while the system is working, which is captured in the net system. In 2004 a ISO/IEC Petri Nets standard was launched where classic and high level net could be synthesized by folding/unfolding processes. The new standard approximated Petri Nets formal representation to both requirements and fault diagnosis methods. On the other hand, in what concerns the risk analysis of hazardous operations, some new methods appear that approximate diagnostic to the design process or to functional analysis, all requiring discrete process representation. In this paper a fault diagnosis method is proposed, where a HAZOP model is used as input requirement that is transformed in a Goal Oriented Requirements Engineering(GORE) representation called KAOS, which are further analyzed in Petri Nets. KAOS diagrams are translated to Petri Nets, using the classic P/T (place/transition) approach. Such approach is scalable and suitable to be applied to large and critical systems as hydroelectric plants. A case study is presented just to this to show applicability to this kind of application.