Functional safety: matching the complexity of methods with the complexity of systems

Abstract: In line with the IEC 61508 functional safety standard, it is required to assess the safety integrity of a system due to random hardware failures. For a rarely used function (operating in a low demand mode), the measurement used is average probability of a dangerous failure on demand (PFDavg). In this paper, four methods have been applied to different configurations of a case study: failure tree analysis with the software GRIF/Tree, multi-phase Markov graphs with the software GRIF/Markov, stochastic Petri nets with predicates with the software GRIF/Petri, and approximate equations (developed by DNV and different from those given in the IEC 61508 standard) using the software OrbitSIL. It is shown that all these methods can lead to similar results for the estimating of the PFDavg, taking into account the required characteristics of the standard. The choice of method must be made without bias, based on an agreement between the modelling efforts, goals, and the system properties. To assist the analyst in this task, a discussion of the benefits and limitations of each of these methods is presented.