Abstract

Systems used e.g. in nuclear power plants and railroad infrastructures require analysis to know about the risk of failures and the effect of different maintenance strategies on this risk. One way to analyse these systems is by building Dynamic Fault Trees (DFTs) and analysing the corresponding Input/Output Interactive Markov Chains (I/OIMCs). Although the I/O IMCs are used for analysis, there is a major drawback, namely the size of the state space. For larger systems this state space becomes too large to do any further calculations. Therefore, analysis of larger FTs is limited by their size. During the conversion from DFTs to I/O-IMCs irrelevant behaviour arises. Therefore, we present smart semantics for fault trees, by applying context-dependent state space generation. This includes an algorithm which sets a boolean to determine the type of the context. This boolean determines whether application of these smart semantics is feasible. Furthermore, we propose a definition of irrelevant behaviour which arises in the conversion of a DFT into an I/O-IMC.

Additional Resources

1. The paper