Cerfacs Enter the world of high performance ...

The 16 March 2021 at 14h00

PhD defense : Rem-Sophia MOURADI – Non-linear data-driven modelling on multidimensional fields: an application to hydro-morphodynamic coastal flows

Nathalie BROUSSET |  zoom : https://inp-toulouse-fr.zoom.us/j/97261515534 |  

Abstract :

This thesis contributions belong to the general framework of data-based and physically- based data-driven modelling. An efficient approach for Machine Learning (ML), as well as a speed-up technique for Data Assimilation (DA), have been developed. For this purpose, Dimensionality Reduction (DR) and stochastic spectral modelling were used. In particular, a coupling between Proper Orthogonal Decomposition (POD) and Polynomial Chaos Expansion (PCE) is at the center of this thesis contributions.

POD and PCE have widely proved their worth in their respective frameworks, and the idea was to combine them for optimal field measurement based forecasting, and ensemble- based acceleration technique for variational DA. For this purpose, (i) a physically inter- pretable POD-PCE ML for non-linear multidimensional fields was developed in the Neural Networks (NN) paradigm and (ii) a hybrid ensemble-variational DA approach for para- metric calibration was proposed with adapted calculations of POD-PCE metamodelling error covariance matrix.

The proposed techniques were assessed in the context of an industrial application, for the study of sedimentation in a coastal power plant's water intake. Water intakes ensure plant cooling via a pumping system. They can be subject to sediment accumulation, which represents a clogging risk and requires costly dredging operations. For monitoring and safety reasons, the power plant stakeholders asked for a predictive tool that could be run in operational conditions. Data collected during many years of monitoring in the study area were provided. The objective was then to achieve comprehensive analysis of the flow and sediment dynamics, as well as to develop an optimal model in terms of forecasting accuracy, physical meaning, and required computational time. Uncertainty reduction and computational efficiency were therefore starting points for all proposed contributions.

In addition to the previously proposed methods, Uncertainty Quantificiation (UQ) studies were undertaken. Specifically, (i) uncertainties related to tidal hydrodynamic modelling, resulting from common modelling choices (domain size, empirical closures) were investigated. POD patterns resulting from measurements and numerical scenarios were compared; (ii) UQ study of the sediment transport modelling in the intake, in a high- dimensional framework, was achieved. Investigations were based on appropriate DR. In fact, POD patterns of Boundary Conditions (BC) and Initial Conditions (IC), resulting from hydrodynamic simulations outputs and from bathymetry measurements respectively, were used.

A perspective of this work would be to implement a hybrid POD-PCE model, using both measured and numerically emulated data, to better understand and predict complex physical processes. This approach would offer a complete, fast and efficient tool for operational predictions.

Keywords :

data-based prediction, physically-based data-driven modelling, statistical learning, data assimilation, uncertainty quantification, sensitivity analysis, geosciences, hydrodynamics, sediment transport, coastal intake.

Jury :

Olivier Le Maitre, Professeur des Universités, Centre de Mathématiques Appliquées of the École Polytechnique (CMAP) (Rapporteur)

Pierre-Olivier Malaterre, Chercheur, INRAE Montpellier (Rapporteur)

Clémentine Prieur, Professeure des Universités, Université Grenoble Alpes, (Examinatrice)

Christine Keribin, Maître de conférences, Université Paris-Sud, (Examinatrice)

Hélène Roux, Maître de conférences, IMFT, (Examinatrice)

Florent Lyard, Professeur des Universités, LEGOS (Examinateur)

Olivier Thual, Professeur des Universités, Institut National Polytechnique de Toulouse (Directeur de thèse)

Cédric Goeury, Ingénieur-Chercheur, EDF R&D LNHE (encadrant principal)

Fabrice Zaoui, Ingénieur-Chercheur, EDF R&D LNHE (encadrant)

Pablo Tassi, Ingénieur-Chercheur, EDF R&D LNHE (encadrant)


The AVBP code from CERFACS at the heart of three PRACE projects from the 23rd call

CERFACS |  30 September 2021

Cerfacs is involved in three PRACE projects of the 23rd call for which hour allocation runs from 01/10/2021 to 30/09/2022. Researchers from ECL/LMFA UMR5509 (Ecole Centrale de Lyon) and IMFT (UMR 5502) laboratories have earned projects entirely based on the use of the LES solver developed by Cerfacs AVBP and involve the support of experts from the CFD and COOP teams underling the importance and effectiveness of collaborations between French labs and Cerfacs. Alexis Giauque from ECL/LMFA UMR5509 (Ecole Centrale de Lyon) has obtained not only one but two PRACE projects! The first project LESFAN (RA0101, 30 000 000 CPU hours on Irene/Rome TGCC) is based on the use of AVBP in the turbomachinery version to study the generation of noise by a fan of a real airplane engine. The second, PRACE-EDGES (RA0101, 40 000 000 CPU hours on Irene/Rome TGCC) focuses on LES modeling of dens gas in complex geometries. To do so, the LMFA Team has developed advanced thermodynamic closures in AVBP allowing the direct simulation of such flows. Laurent Selle from IMFT (UMR 5502) has received CPU hours for the GASTON project (RA0061, 30 000 000 CPU hours on Marenostrum BSC) which aims to study the structure of hydrogen flames in porous materials. For this, IMFT and Cerfacs will perform coupled simulations considering the reactive flow with AVBP as well as the conduction in the porous medium with AVTP which is known to play an central role in the flame stabilization process.Read more

watch the online contest ”my thesis in 180 sec” organized for Marie Curie actions

CERFACS |  24 September 2021

Next Thursday, Sept 30th the #FallingWallsLab #MSCA research presentation competition will take place, . This is a competition in which 15 researchers compete to deliver the best presentation of their research topic in the format "my thesis in 180 seconds" . Javier Crespo-Anadon, 3rd year PhD student in the CFD team is one of the 15 finalists and he will present this thesis topic: ignition in spinning combustion engines for a reduction in CO2 emissions! You can watch it online on the above website on Sept 30th at 3pm. I encourage you to go to the website and register!  You'll be able to listen to the other participants' pitches and vote for your favorite. Come and see for yourself what kind of research the EU funds!Read more