Method of time Petri net analysis for analysis of fault trees with time dependencies

In order to make computer-aided control systems as safe as possible, a number of analysis techniques have been developed. One of these is fault tree analysis. A fault tree (FT) represents causal and generalisation relations between events (e.g. a hazard and its causes). However, it can express neither time relations between events nor detection and protection times. Time Petri nets (TPNs) can model all the above aspects. Thus, TPNs can be used for analysing and verifying time-dependent fault trees (FTs). One of the limitations of classical TPN analysis is the large number of TPN states. Even for a small FT this number can turn out to be vast. The authors introduce a new method for analysing such TPNs that model FTs. We do not consider all states that are reachable from the initial marking in classical TPN analysis but only those that lead to the occurrence of a hazard. Such an approach simplifies the procedure and results in cleaner final conclusions. If the hazard is reachable there is a need for safety measures to be taken. FT analysis and modelling of protection using TPNs will be illustrated using an example.