-2018- A Methodology for Architecting Collaborative Product Service System of Systems

Authors :

Andreas M. Hein( Laboratoire Genie Industriel, CentraleSupélec, Université Paris-Saclay), Yann Chazal (Renault SA France), Marija Jankovic (Laboratoire Genie Industriel, CentraleSupélec, Université Paris-Saclay), Samuel Boutin (Knowledge Inside Versailles, France)

Keywords: PSS, system of systems, system architecture, collaborative services, service engineering, business model


Product service systems (PSS) can be understood as an innovation / business strategy that includes a set of products and services that are realized by an network of actors. More recently, more complex PSS have been increasingly of interest as they comprise of System of systems (SoS) involving multiple actors contributing to the PSS. This paper presents a methodology for architecting PSS that are SoS specifically driven by technological innovation. In the first step, the domains on which a technology’s impact is going to be analyzed are modeled. Subsequently potential impacts of a new technology are identified. Then, resulting pains and gains from the new technology for specific actors are elicited. Adapting the value proposition design and business model canvas approach, PSS value propositions and business model alterantives are created. Finally, alternative PSS architectures are developed for specific business models. We illustrate the methodology by applying it to the case of an automobile PSS.

-2017- CSDM Paris 12 et 13 décembre

Les 12 et 13 décembre 2017 à Paris a eu lieu la 8ème édition de la conférence internationale « Design & Management de Systèmes complexes » CSD&M.
Le thème de cette 8ème édition a notamment porté « Vers des systèmes plus intelligents et plus autonomes ».
Nous avons présenté des travaux de pointe sur la modélisation des analyses de sûreté de fonctionnement dans le nucléaire


 -2014- ERTS : Interoperability between a dynamic reliability modeling and a Systems Engineering process

Authors: Gilles Deleuze, Aurélie Léger, Pierre Yves Piriou (EDF R&D Clamart, France), Sylvain Chabroux (KI)

Keywords: Systems Engineering, hybrid systems modeling, RAMS, dependability assessment, dynamic reliability, interoperability.


Industrial systems are often complex in terms of size, structure of interactions between systems, components, dynamic operational environment, ageing, etc. The dynamic reliability approach, covering a wider range of phenomena, is a convenient framework to model the behavior of such systems operating in a dynamic environment. However, there is a price to pay, e.g. in terms of data, state graphs, complexity of reliability tools, and combination of various engineering activities. In a previous study, focused on availability assessment of a large size hybrid system, two complementary approaches have been used, one using Stochastic Hybrid Automaton, and one using modeling and Monte Carlo Simulation. Both require a combination of various engineering activities, the use of computational power, data and component level models. A sound Systems Engineering process, benefiting from the improvement of most recent tools may be a fruitful approach. Although feasibility demonstrations have been done for usual approaches of dependability, the study of interoperability between dynamic reliability modeling and Systems Engineering has not the same level of achievement. We start with dynamic models based on Stochastic Hybrid Automaton, as they represent potentially the highest challenges. The case study is focused on the availability of a feed-water control system in a power plant steam generator.

On the basis of definitions of a Systems Engineering process, a Metamodel defines a framework for integrating the safety into SE processes. It supports a “hub automaton”, that is the key element for interoperability between the tools and activities needed for dynamic reliability assessment. The casual SE process supported by a tool like arKItect, supports the engineer in producing and managing the casual processes (description of functional and physical architectures, allocation of functions, preliminary assessment of requirements…).

Then, it is possible to describe the functional and failure features of the components. The RAMS engineer can model realistic failure/repairing scenarios and define redundancy policies for updating the dynamical allocation of functions caused by failure events. The “hub automaton” supports the translation of the dynamic dependability model into specific dynamic dependability tools.


-2013- CSDM : Synchronize MBSE projects in a context of high diversity

Authors: Philippe TOUSSAINT (KI), Joe MATTA (KI)

Keywords: Data sharing, Synchronization, Change management, Diversity management


Current competitiveness and economic environment requires industrials to give priority to the optimization of developments. Different development projects are, whenever possible, grouped in a unique generic study that is later specialized into specific ones. For instance, instead of launching three projects to develop three vehicles of a new product range, automotive manufacturer launches one vehicle range development program that will be refined into three specific vehicles.

Yet it is important to consider upgradeability of each study, generic as well as specifics, and to keep all developments synchronized. Elements have to be identified with diversity criteria to control their affiliations to generic or specific studies. In such context, change management is a key issue. Teams working on specific projects shall be able not only to synchronize evolutions coming from the generic study but also to visualize and control modifications.

The main problems at hand are:

  • Synchronize data between studies
  • Manage diversity criteria
  • Manage versions

In this paper, we will investigate the possibility to manage shared data while creating or modifying them. The proposed process to control this exchange contains several topics:

  • Considering specific elements versus generic elements
  • Manage specificity as a diversity criteria
  • Share studies architecture elements using a common repository
  • Include versions management features
  • Visualize, control and propagate modifications

The result is to propose a way to efficiently share data between different studies while considering diversity criteria. This helps to optimize development costs by grouping studies, ensuring a good synchronization between teams and maintaining a good control of change management.

Consultez notre poster: KI Poster CSDM2013

-2012- CSDM : How to turn spreadsheets into object oriented models

Auteurs : Samuel Boutin (KI) et Joe Matta (KI)

Résumé :

It is very common to see engineers using spreadsheets (e.g. excel) as a small database. For instance, a data dictionary will be a spreadsheet with a column for the data identifier and then different columns for data attributes (e.g. data type, step, min and max value). An advanced use would be to add two columns for functions producing and consuming the data resulting in a cross-reference dictionary. Then many other applications exist related to project management, diversity management, requirements management and specialized fields topics. The important fact is that spreadsheets generally capture a specialized field specification and sometime very transverse project information and that together they form often a big part of a system specification.

A big part of engineering legacy data is provided in this form in many industries. This is partly due to the many functions of spreadsheet e.g. for billing and counting. So using spreadsheet allows most of the time simple checks for data integrity.  A positive aspect is that spreadsheets contain already structured data w.r.t like word or rich text descriptions. But this does not help solving a major problem in systems engineering: keep all the different spreadsheet and specification documents aligned during the life of a project.

In this paper we investigate the possibility to upload a wide variety of spreadsheet descriptions into a synchronized system specification. To reach this goal, we proceed with the following steps:

–       Provide a semantics to each spreadsheet

–       Link this semantics to a part of a DSL (Domain Specific Language)

–       Organize the semantics of all spreadsheets together

–       Import the spreadsheet content in a unified model

So as a result, we raise for many organizations the opportunity to leapfrog from a wide variety of semi-structured specifications to a synchronized multi-aspect specification which is at the same time a model based design approach.

This concept extends more generally to the possibility to transform a database into an object oriented model and conversely. We also discuss how this relates to UML profiles, BMPN, and DSL initiatives.

Accéder à l’article complet