A meta model framework for risk analysis, diagnosis and simulation

H Kanj (1) , Jean-Marie Flaus (1)

1 G-SCOP_GCSP - Gestion et Conduite des Systèmes de Production

The goal of risk analysis is to identify events that may have one or several undesirable consequences on a system, and to assess the likelihood and severity of these consequences. A lot of methods may be used to conduct risk analysis such as Preliminary Hazard Analysis (PHA), and Failure Mode Effects Analysis (FMEA). In most of these methods, the obtained information may be used to build a risk model. Very often, the next step after risk analysis is, to study the behavior of the system if the undesirable events occur, in order to evaluate its performance in degraded conditions and its robustness or resilience. An approach allowing integrated risk analysis and simulation would be desirable. Such an approach has been proposed for business process management [Tjoa et al., 2011]. The goal of this paper is to present a meta model, suited to socio-technical systems, that allows describing the system to analyze, the result of the risk analysis and the required aspects of dynamical system behavior in order automatically perform simulation under degraded conditions. The model is an extension of the FIS model presented in [Negrichi et al., 2012]. The meta model may also be used for fault diagnosis as it can be used for generating redundancy relation and performing root cause search [Flaus et al., 2011]. Our meta model consists of three main modules: the structural view, the dysfunctional view and the view of the evolution.: • The structural view (SysFis): defines the architecture of the analyzed system, breaks it down into subsystems, and describes the characteristics of each subsystem and the material entities used. This is the basic view. that describes the structure of the installation or the analyzed object in a relatively simple manner, by showing the various interactions systems, and specifying, if necessary their functions and the material components (human, technical or informational) that

Domaines: Automatique

H Kanj, Jean-Marie Flaus. A meta model framework for risk analysis, diagnosis and simulation. Safety and Reliability Conference ESREL, Sep 2014, Wroclaw, Poland. ⟨10.1201/b17399-280⟩. ⟨hal-01240818⟩