In this project we are interested in the development of model order reduction methods applied to aerothermal simulations.
The CHORUS project (Common Horizon of Open Research in Uncertainty for Simulation) gathers academic and industrial researchers to tackle the growing need of quantitative un- certainty management during the risk analysis of complex system in early design phases. This kind of problem involves to solve multiple times (hundred to thousand evaluations) parametrized partial differential equations and then the underlying challenge is to provide reliable highly-scalable reduce models. CHORUS aims to promote existing initiatives on the subject in a structured way and to give access to up-to-date methods to the R&T community. The three main directions defined in the project are :
- Methodological treatment of uncertainty management problems for multi disciplinary purposes,
- Development of new mathematical models and algorithms to face scalability prob- lems (goal-oriented uncertainty analysis, specific reduction techniques using sparsity or multi fidelity),
- Accessibility of interoperable advanced algorithms linked to HPC capabilities for a larger community than the CHORUS consortium in a recognized open source environment
The present use-case, proposed by Airbus Group, is the study of air flow in an airplane cabin. The aim is to develop an efficient reduced model of this system and to couple this model with an Environment Control System (ECS) in order to optimise the passagers confort and energy costs.
This project gathers many challenges. The non-linearity of the governing equations (Navier-Stokes) is a difficulty for the resolution and for the certification of the reduced model. The physical hypothesis are complicated as well with turbulent flow, coupling between fluid and energy equation.
Road Map :
- Implementation of a finit element model for areothermal simulations with turbulent flows
- Application of the Certified Reduced Basis method to this model
- Coupling with ECS
We use Feel++ finit element to develop the finit element model. The library also proposed a very convenient framework for the application of the reduced basis method. This framework is currently in development in order to propose rigorous error estimator for non linear problems.
- Airbus (Large group)
- Phimeca (SME)
- AICES (RWTH)
- Christophe Prud’homme (IRMA, Cemosis)
- Michel Fouquembergh (Airbus)
- Yannick Hoarau (ICube)
- Karen Veroy (RWTH)
- Martin Grepl (RWTH)
- Vincent Huber (Cemosis)
- Jean-Baptiste Wahl (IRMA,Cemosis)
- Vincent Chabannes (Cemosis)
- ANR Chorus
- Labex IRMA