Optimization framework for desirable and sustainable complex systems – OSCAR

The OSCAR project work programme is structured in 3 phases (Analysis, Elaboration, Experiments), in 5 batches (plus a project management batch) integrating experimentation on several systems (of different levels of complexity in batches 2 and 5) throughout the project, as follows (Figure 5) :
- Work package 1 aims to review the state of the art of ecodesign problems, focusing on modelling and resolution methods adopted. The expected deliverable consists of 2 parts, defined by 2 sub-batches focusing on ArP for the first one and on ArCV for the second one. The objective of this first batch is to allow the project partners to exchange on their respective practices and to synthesise the modelling elements existing in the literature with the resolution methods used. This common knowledge base will enable the partners to better define the modelling elements to be defined in batch 2 and the resolution strategies and algorithms to be implemented in batch 3.
- The aim of work package 2 is to develop a model for different component typologies (ArP components and ArCV components) in order to efficiently explore the design space of a system (vehicle, automotive system, automotive component). This means in particular to formalise the significant Constraints that influence the Environmental impacts (named EC). A distinction has been made between
o Constraints related to the extraction/manufacturing - logistics and eco-use phases
o Constraints related to the maintenance and end-of-life phases
This package also integrates the assessment of environmental impacts in a non-static, non-uniform and non-isolated way. The deliverables materialise through the identification of significant constraints that influence the environmental assessment and that characterise the consideration of spatiality, evolutions over a long temporality or consequential effects.
- Lot 3 aims at proposing a global strategy for the optimisation of complex systems minimising environmental impacts. The main deliverables are the definition of objective functions (environmental and attractiveness) and optimisation algorithms on which an ecodesign methodology is based. The methodology and algorithms developed will be based on the modelling elements of lot 2.
- Lot 4 aims to equip the methodology of lot 3 so that it can be implemented in a concrete manner. The deliverable is a software demonstrator of the OSCAR methodology.
- The aim of lot 5 is to test all the methodological elements developed on two cases of complex system applications in a simplified version in the first instance and then in an enriched version in the second instance; this double test in two stages will guarantee a certain robustness of the models. It is structured in 2 tasks :
o An application of the demonstrator on a passive component (seat) and on an active component (air conditioning);
o An application of the demonstrator on the Electric Vehicle system.

The results obtained during the OSCAR research program are as follows: 1) Development of a methodology for designing attractive and sustainable complex systems. 2) Creation of algorithms for exploring architectures for products and life cycles. 3) Experimentation on two industrial cases. 4) Creation of a software demonstrator to test the method. The OSCAR method includes a global ecodesign approach, supported by 6 tools and powered by the Ibex optimization software, the Eco-Invent environmental databases and the Simapro software. A new exploration strategy, the «OSCAR strategy«, was developed for a homogeneous exploration of the design space, as opposed to the standard in-depth exploration. A methodology for grouping solutions into families of architectures was developed, allowing the «clustering« of product architectures and life cycle architectures. This approach has been successfully tested in the field of electrified vehicles, more specifically on 2 subsystems: 1) The hybrid power generation system «TurboGenerator + batteries«. 2) The «battery pack« system. To test the OSCAR method, a software demonstrator was developed, linked to an environmental database and including several modules associated with the method's tools.

Two perspectives are possible:
- Continuation with Stellantis within the framework of the EasyMoove Project/ Extension by CIFRE or Ademe thesis (operationalization of the Oscar approach).
- Proposal of a project to the ADEME ECONUM call for proposals/feb24

Richard de Latour, T.; Chenouard, R.; Granvilliers, L. Partial design space exploration strategies applied in preliminary design. Int. J. Interact. Des. Manuf. 2023.

Tchertchian, N.; Millet, D.; Richard de Latour, T.; Chenouard, R. Ecodesign approach for attractive complex systems. Life Cycle Management. 2023.

Richard de Latour, T.; Chenouard, R.; Granvilliers, L. Stratégies de résolution anytime de problèmes de satisfaction de contraintes numériques (Journée Francophone de la Programmation par Contraintes 2021).

Tchertchian, N.; Millet, D.; Richard de Latour, T.; Chenouard, R. Nouveau paradigme d’écoconception de système complexe, Séminaire Réseau EcoSD, 24 juin 2021, Paris.

The objective of this research is to develop a new ecodesign approach for complex systems allowing through a wide exploration in the upstream phases of a design project to identify optimal solutions on all the environmental impact categories taking into account dynamics and spatialization aspects. 2 major scientific problems are targeted: 1) the modelling of new environmental aspects in the form of generic constraints at the same time on the Products architectures and Life cycles architectures and 2) the definition of a new ecodesign method integrating environmental criteria and attractiveness for the customer from the upstream phases.
The developped software OSCAR will demonstrate that we can effectively invert the ecodesign problem, and so make a design of complex system truly compatible with the sustainable development (satisfying environmental objectives a priori defined and simultaneously maximizing desirability for the users).