Modelling Dynamic Reliability via Fluid Petri Nets

Combinatorial models for reliability analysis (like fault-trees or block diagram) are static models that cannot include any type of component dependence. In the CTMC (Continuous Time Markov Chain) framework, the transition rates can depend on the state of the system thus allowing the analyst to include some dependencies among components. However, in more general terms, the system reliability may depend on parameters or quantities that vary continuously in time (like temperature, pressure, distance, etc.). Systems whose behavior in time can be described by discrete as well as continuous variables, are called hybrid systems. In the dependability literature, the case in which the reliability characteristics vary continuously versus a process parameter, is sometimes referred to as dynamic reliability [1]. The modelling and analysis of hybrid dynamic systems is an open research area. The present paper discusses the evaluation of a benchmark on dynamic reliability proposed in [1] via a modelling framework called Fluid Stochastic Petri Net (FSPN).