Generalized Boolean logic Driven Markov Processes: a powerful modeling framework for Model-Based Safety Analysis of dynamic repairable and reconfigurable systems

Pierre-Yves Piriou (1) , Jean-Marc Faure (1) , Jean-Jacques Lesage (1)

1 LURPA - Laboratoire Universitaire de Recherche en Production Automatisée

This paper presents a modeling framework that permits to describe in an integrated manner the structure of the critical system to analyze, by using an enriched fault tree, the dysfunctional behavior of its components, by means of Markov processes, and the reconfiguration strategies that have been planned to ensure safety and availability, with Moore machines. This framework has been developed from BDMP (Boolean logic Driven Markov Processes), a previous framework for dynamic repairable systems. First, the contribution is motivated by pinpointing the limitations of BDMP to model complex reconfiguration strategies and the failures of the control of these strategies. The syntax and semantics of GBDMP (Generalized Boolean logic Driven Markov Processes) are then formally defined; in particular, an algorithm to analyze the dynamic behavior of a GBDMP model is developed. The modeling capabilities of this framework are illustrated on three representative examples. Last, qualitative and quantitative analysis of GDBMP models highlight the benefits of the approach.

Mots clés: Model Based Safety Analysis; Generalized Boolean logic Driven Markov Processes; Dynamic and repairable system; Reconfiguration strategies; Moore machine

Domaines: Sciences de l'ingénieur [physics]

Pierre-Yves Piriou, Jean-Marc Faure, Jean-Jacques Lesage. Generalized Boolean logic Driven Markov Processes: a powerful modeling framework for Model-Based Safety Analysis of dynamic repairable and reconfigurable systems. 2016. ⟨hal-01393700⟩