Modelling and simulating a benchmark on dynamic reliability as a Stochastic Activity Network

Several versions of a benchmark on dynamic reliability taken from the literature are examined: each version deals with particular aspects such as state dependent failure rates, failures on demand, and the repair of components. The benchmark was modelled in the past, using two types of Petri Nets; in this paper, we exploit another Petri Net based modelling formalism called Stochastic Activity Network (SAN). This allows a more compact model of the system by exploiting input and output gates, together with the possibility to represent float variables by means of extended places. The SAN model of the system undergoes simulation in order to compute the system unreliability: the results are coherent with those obtained with other methods, and this confirms that Petri Net based models can be a valid approach to dynamic reliability evaluation.