Reliability analysis of a safety system using petri net and comparison with smart component methodology

For the reliability analysis of advanced nuclear reactor safety systems, though event tree-fault approach has been used over the years, they are inadequate from a modeling perspective. First, it involves making various levels of approximations depending on the complexity of the system being modeled and second, the responsibility of deriving the correct reliability model rests with the analyst. To overcome the problems mentioned above, various methods for the inclusion of dynamic aspects are being developed. Though many of the methods can more closely reflect the dynamic reliability aspects of the reliability model, they lack the features required for a user-friendly approach. Recently, a Smart Component Methodology based on the object-oriented representation of system structure and behavior, to perform dynamic reliability analysis has been proposed. The dynamic reliability methods could be divided into two categories based on how close the initial formal representation is to the actual system description. For example, in the case of Petri nets, which is often used to perform dynamic reliability analysis, a dynamic system's structure and behavior have to be manually translated (as of now) to a Petri net to perform reliability analysis. Petri net and similar methods like dynamic event tree would fall into this category. In SCM since it uses object-oriented representation, which is closer to the system's design description/representation; this method would require the least reliability expertise to perform dynamic reliability analysis (would be the other category). Future methods which would automatically translate a system description/representation into a reliability model or automatically generate reliability metrics would fall into the latter category. In this paper, we perform a comparative study of the dynamic reliability modeling of shutdown system of fast breeder reactor with Petri net model as well as the newly proposed SCM to bring out the differences and advantages of these two methods. The running time and ease of modeling aspects are brought out.