PostDoc Project
Model engineering and architecture
2020/12/01 - 2021/11/30
Industrial context :

The work is part of the "Digital Reactor" project, financed within the framework of the call for "Structuring Projects for Competitiveness" (PSPC) of the Future Investment Program (PIA), and in which the main actors of the French nuclear industry (EDF, Framatome, CEA, CORYS, ESI Group, Aneo, Axone, BOOST, CRAN) participate. The objective of the "Digital Reactor" project is to build a reactor-scale "digital twin", covering all life cycle phases, to simplify the design process and secure safety margins in operation.

Context and scientific objectives :

The ISET department of CRAN (Nancy Automatics Research Center, UMR 7039 CNRS and University of Lorraine), has been developing research activities in the field of Complex Systems Engineering for several years (theses of Evrot D. (2008), Dobre D. (2010), Bouffaron F. (2016), Wu Q. (2020)). These works has focused on the formalization of System Engineering practices (Model-Based Management System (MBSE)) with the objective of founding the processes, models and methods of System Engineering (SE), in order to facilitate their appropriation and operational implementation in complex and evolving industrial contexts.

In SE, the problems encountered by industry are the lack of a global system vision, due to a compartmentalized vision (knowledge silos) of the systems to be developed according to the disciplinary fields concerned, as well as the lack of agility of the processes which does not favour their adoption. These problems lead to time and cost overruns in SE projects as well as the failure to meet customer requirements.
To remedy these problems (lack of global vision, lack of process agility), the work developed by the team focused on formalizing the specification process (D. Dobre 2010) involving the different engineering actors (system and disciplines), then on a collaboratively built system specification and its execution (executable co-specification) (F. Bouffaron 2016). In a second step, in order to act on the adoption of an MBSE approach in companies, the thesis of Q. Wu (2020) focuses on the capitalization of engineering know-how in the form of modeling patterns (library creation, identification, pattern search) and on the evaluation of the maturity of the reuse process.

In this work we wish to extend the work on model-based executable system co-specification using co-simulation (Vaubourg et al. 2015 with EDF R&D, Paris et al. 2019). The defended idea is to have from the beginning and throughout the engineering process an executable model of the specification, from a system vision to the detailed design levels of architectures (trades). This model would allow, in an agile way, at each level of system abstraction, by successive refinements, to have a unique, global and shareable vision of the system to be made and to evaluate the specification under development : Verification/Validation and analysis. Beyond engineering, this model would enable the approach towards Integration to be extended, by verifying and validating the business developments of components, modules and software using general mechanisms of the "X in the loop" type (Model, Hardware, Software).
In the community, a new profession has emerged to meet these architecture and model integration needs: Simulation Architect (F. Retho 2015 and G. Sirin 2015), acting as an interface between the system architect (who guarantees the overall vision) and business experts. His role is to specify an intention model for the business experts, based on requests from the system architect, returning an implementation model to be integrated in order to build a system (co-)simulation.
From the point of view of co-simulation, recent work at LORIA (Paris 2019) makes it possible to incrementally integrate simulators to build the co-simulation of a target system (digital twin). The approach is bottom-up, by composing existing software models. However, in the case of complex systems, this integrative approach reaches its limits and does not guarantee consistency and validation. It is therefore a question of having a method for building a co-simulation model of a system (existing or under development) that guarantees the coherence and validity of the co-simulation model.

Scientific locks :

The locks to be lifted in this objective concern :
- The design of the architecture of the co-simulation model from the system model according to the level of progress of the project (passage from system artifacts to the specification of the co-simulation model elements, definition of the interactions between the system model and the simulation business models, evaluation, etc.).
- The extension of system co-specification mechanisms (Bouffaron 2016) to the construction of a co-simulable system model (interaction mechanisms between system architect, simulation/model architect and business experts);
- The definition of the specification process of an intention model, by the simulation architect;
- The validation of contributions on a case study proposed in the framework of the Digital Reactor project.

The main expected results consist in a method for designing a co-simulation system model from a system co-specification, ensuring consistency of the whole from the system level to the most detailed levels of architecture design (business).
Model-Based Systems Engineering, MBSE, Models architecture, Cosimulation
Profile of the candidate sought: the candidate must have strong skills in model-based systems engineering and simulation.
Eco-Technic systems engineering
Digital Reactor Project