A Meta-model to Support the Integration of Dependability Concerns into Systems Engineering Processes: an Example from Power Production

Pierre-Yves Piriou (1) , Jean-Marc Faure (1) , Gilles Deleuze (2)

1 LURPA - Laboratoire Universitaire de Recherche en Production Automatisée
2 EDF R&D - EDF R&D

Systems engineering (SE) is a very promising approach to facilitate the development of complex systems. This explains why several SE processes have been already proposed. However, these proposals focus mainly on systems with faultless components. Integration of dependability concerns into SE processes must be supported by a suitable organization of the data which are dealt with during the system life-cycle. A meta-model which defines the concepts used during this cycle as well as the relations between these concepts is a way to rigorously describe this organization. This article proposes such a meta-model developed for power production systems. These systems are phased mission systems composed of repairable and multi-state components; moreover, several redundancy policies shall be defined for each phase. This proposal is illustrated on a small example from a power plant. Last, the merit of this contribution to support the integration of dependability concerns is shown by proposing a method to build systematically, from the instance diagrams derived from the proposed meta-model, the Markov Chains which represent the dysfunctional dynamic behavior of a system.

Mots clés: System Engineering; Dependability; Redundancy Policy; Phased Mission System; UML Class Diagram; Markov Chain

Domaines: Automatique / Robotique

Pierre-Yves Piriou, Jean-Marc Faure, Gilles Deleuze. A Meta-model to Support the Integration of Dependability Concerns into Systems Engineering Processes: an Example from Power Production. IEEE Systems Journal, 2014, PP (99), pp.SCH-ISJ-RE-13-02551.R3. ⟨10.1109/JSYST.2014.2328663⟩. ⟨hal-01059913⟩