ABSTRACT
Piecewise Deterministic Markov Process (PDMP) is a mathematical framework able to support the rigorous definition of Dynamic reliability problems. Modeling a system as a PDMP offers a more realistic description of its dynamic behavior—allowing the engineer to gain several insights during the system design. However, solving a Dynamic reliability problem is not an easy task. Difficulties arise in: the definition of a model able to capture the deterministic and stochastic nature of the system and their interaction, and that model solution often requires large computing resources. This paper proposes: (i) a modeling formalism which couples stochastic transition-systems, e.g., Stochastic Activity Networks, and flow-sheet models, e.g., Simulink®; and (ii) a Discrete Event Simulation algorithm embedding a sampling procedure based on the Thinning method. A case study of an air-cooling system modeled as a deterministic dynamic system which is subjected to failures is used to show the method.
