Cedric GOEURY

Chercheur LHSV

Thèse

Cédric Goeury. Modélisation du transport des nappes d'hydrocarbures en zone continentale et estuarienne. Sciences de la Terre. Université Paris-Est, 2012. Français. 〈NNT : 2012PEST1131〉. 〈pastel-00778087〉

Publications

Bayesian inference of numerical modeling-based morphodynamics: Application to a dam-break over a mobile bed experiment
- Goeury Cédric
- Souillé Fabien
Computational Geosciences, Springer Verlag , 2025, 30 (1), pp.2 . Numerical modeling of morphodynamics presents significant challenges in engineering due to uncertainties arising from inaccurate inputs, model errors, and limited computing resources. Accurate results are essential for optimizing strategies and reducing costs. This paper presents a step-by-step Bayesian methodology to conduct an uncertainty analysis of 2D numerical modeling-based morphodynamics, exemplified by a dam-break over a sand bed experiment. Initially, uncertainties from prior knowledge are propagated through the dynamical model using the Monte Carlo technique. This approach estimates the relative influence of each input parameter on results, identifying the most relevant parameters and observations for Bayesian inference and creating a numerical database for emulator construction. Given the computationally intensive simulations of Markov chain Monte Carlo (MCMC) sampling, a neural network emulator is used to approximate the complex 2D numerical model efficiently. Subsequently, a Bayesian framework is employed to characterize input parameter uncertainty variability and produce probability-based predictions. (10.1007/s10596-025-10400-7)
DOI : 10.1007/s10596-025-10400-7
A tree-based Polynomial Chaos expansion for surrogate modeling and sensitivity analysis of complex numerical models
- Ben Said Faten
- Alfonsi Aurélien
- Dutfoy Anne
- Goeury Cédric
- Jodeau Magali
- Reygner Julien
- Zaoui Fabrice
, 2025 . This paper introduces Tree-based Polynomial Chaos Expansion (Tree-PCE), a novel surrogate modeling technique designed to efficiently approximate complex numerical models exhibiting nonlinearities and discontinuities. Tree-PCE combines the expressive power of Polynomial Chaos Expansion (PCE) with an adaptive partitioning strategy inspired by regression trees. By recursively dividing the input space into hyperrectangular subdomains and fitting localized PCEs, Tree-PCE constructs a piecewise polynomial surrogate that improves both accuracy and computational efficiency. The method is particularly well-suited for global sensitivity analysis, enabling direct computation of Sobol' indices from local expansion coefficients and introducing a new class of sensitivity indices derived from the tree structure itself. Numerical experiments on synthetic and real-world models, including a 2D morphodynamic case, demonstrate that Tree-PCE offers a favorable balance between accuracy and complexity, especially in the presence of discontinuities. While its performance depends on the compromise between the number of subdomains and the degree of local polynomials, this trade-off can be explored using automated hyperparameter optimization frameworks. This opens promising perspectives for systematically identifying optimal configurations and enhancing the robustness of surrogate modeling in complex systems.
Flow patterns in shallow rectangular reservoirs with open channel inlet or pipe flow inlet at various depths: An experimental study
- Chagdali El Mehdi
- El Kadi Abderrezzak Kamal
- Erpicum Sébastien
- Goeury Cédric
- Secher Matthieu
- Dewals Benjamin
International Journal of Sediment Research, Elsevier , 2025, 40 (2), pp.209-221 . This study experimentally assesses the influence of varying the inlet boundary condition on the flow patterns in rectangular shallow reservoirs. Two types of inlet boundary conditions were compared: a free surface inlet channel, and a pressurized circular inlet jet positioned at three different elevations over the flow depth (centroid of the inlet jet situated at 25%, 50%, or 75% of the flow depth). The outlet boundary condition was a free surface channel in all cases. Twenty-two experiments were done with two distinct reservoir lengths (length-to-width ratios of 1.1 and 2.0) and three hydraulic boundary conditions (Froude numbers of 0.14, 0.16, and 0.21). Velocity fields were measured with Large-Scale Particle Image Velocimetry (LSPIV) at the surface, and with an Acoustic Doppler Velocity Profiler(ADVP) at several cross sections. The flow patterns are greatly influenced by the inlet boundary condition and the reservoir geometry, but to a lesser extent by the hydraulic boundary condition. For an inlet circular jet located near the reservoir bottom, an unstable flow type, changing over time in a chaotic manner, was observed regardless of the reservoir length and of the inlet flow rate. The same type of unstable flow pattern was observed for a relatively long reservoir and the lowest tested flow rate, irrespective of the vertical positioning of the inlet jet. In all other tested configurations, a steady reattached jet was found in the reservoir equipped with a pressurized inlet jet. In addition to providing new knowledge on flow patterns in shallow reservoirs with an inlet jet, the experimental data presented here will prove valuable for evaluating flow computational models. (10.1016/j.ijsrc.2025.01.004)
DOI : 10.1016/j.ijsrc.2025.01.004
A new methodology for the assessment of flood hazard in urban areas due to levee breaches
- Bacchi Vito
- Goeury Cédric
- Zaoui Fabrice
- El Kadi Abderrezzak Kamal
- Bacchi Sophie
- Pavan Sara
, 2022, pp.6672-6679 . The objective of this study is to propose a new methodology for the assessment of flood hazard induced by the formation of levee breaches due to overtopping flows. The proposed methodology relies on the definition of a new breach model and on the flood hazard assessment (i.e. water height in area of interest) through a deterministic simulation. The methodology is implemented in the open-source hydrodynamic suite of solvers TELEMAC-MASCARET (www.opentelemac.org) and applied to the Loire river using the two-dimensional (2D) depth-averaged hydrodynamic code TELEMAC-2D. We chose a flood scenario corresponding to a return period of around 500 years and we compared results to those obtained by performing 3,000 Monte-Carlo simulations using uniform distribution of the model parameters. Results suggest that the deterministic simulation should be completed by a set of well-chosen deterministic scenarios in order to cover the large uncertainty showed by the computational costly Monte-Carlo simulations. (10.3850/IAHR-39WC2521711920221210)
DOI : 10.3850/IAHR-39WC2521711920221210
Physically interpretable machine learning algorithm on multidimensional non-linear fields
- Mouradi Rem-Sophia
- Goeury Cédric
- Thual Olivier
- Zaoui Fabrice
- Tassi Pablo
Journal of Computational Physics, Elsevier , 2021, 428, pp.110074 . In an ever-increasing interest for Machine Learning (ML) and a favorable data development context, we here propose an original methodology for data-based prediction of two-dimensional physical fields. Polynomial Chaos Expansion (PCE), widely used in the Uncertainty Quantification community (UQ), has long been employed as a robust representation for probabilistic input-to-output mapping. It has been recently tested in a pure ML context, and shown to be as powerful as classical ML techniques for point-wise prediction. Some advantages are inherent to the method, such as its explicitness and adaptability to small training sets, in addition to the associated probabilistic framework. Simultaneously, Dimensionality Reduction (DR) techniques are increasingly used for pattern recognition and data compression and have gained interest due to improved data quality. In this study, the interest of Proper Orthogonal Decomposition (POD) for the construction of a statistical predictive model is demonstrated. Both POD and PCE have amply proved their worth in their respective frameworks. The goal of the present paper was to combine them for a field-measurement-based forecasting. The described steps are also useful to analyze the data. Some challenging issues encountered when using multidimensional field measurements are addressed, for example when dealing with few data. The POD-PCE coupling methodology is presented, with particular focus on input data characteristics and training-set choice. A simple methodology for evaluating the importance of each physical parameter is proposed for the PCE model and extended to the POD-PCE coupling. (10.1016/j.jcp.2020.110074)
DOI : 10.1016/j.jcp.2020.110074
Polynomial Surrogates for Open-Channel Flows in Random Steady State
- El Moçayd Nabil
- Ricci Sophie
- Goutal Nicole
- Rochoux Mélanie C.
- Boyaval Sébastien
- Goeury Cédric
- Lucor Didier
- Thual Olivier
Environmental Modeling & Assessment, Springer , 2017, 23, pp.309–331 . Assessing epistemic uncertainties is considered as a milestone for improving numerical predictions of a dynamical system. In hydrodynamics, uncertainties in input parameters translate into uncertainties in simulated water levels through the shallow water equations. We investigate the ability of generalized polynomial chaos (gPC) surrogate to evaluate the probabilistic features of water level simulated by a 1-D hydraulic model (MASCARET) with the same accuracy as a classical Monte Carlo method but at a reduced computational cost. This study highlights that the water level probability density function and covariance matrix are better estimated with the polynomial surrogate model than with a Monte Carlo approach on the forward model given a limited budget of MASCARET evaluations. The gPC-surrogate performance is first assessed on an idealized channel with uniform geometry and then applied on the more realistic case of the Garonne River (France) for which a global sensitivity analysis using sparse least-angle regression was performed to reduce the size of the stochastic problem. For both cases, Galerkin projection approximation coupled to Gaussian quadrature that involves a limited number of forward model evaluations is compared with least-square regression for computing the coefficients when the surrogate is parameterized with respect to the local friction coefficient and the upstream discharge. The results showed that a gPC-surrogate with total polynomial degree equal to 6 requiring 49 forward model evaluations is sufficient to represent the water level distribution (in the sense of the ℓ2 norm), the probability density function and the water level covariance matrix for further use in the framework of data assimilation. In locations where the flow dynamics is more complex due to bathymetry, a higher polynomial degree is needed to retrieve the water level distribution. The use of a surrogate is thus a promising strategy for uncertainty quantification studies in open-channel flows and should be extended to unsteady flows. It also paves the way toward cost-effective ensemble-based data assimilation for flood forecasting and water resource management. (10.1007/s10666-017-9582-2)
DOI : 10.1007/s10666-017-9582-2
Différentiation algorithmique appliquée à la calibration optimale d'un modèle hydraulique à surface libre
- Demangeon Félix
- Goeury Cédric
- Zaoui Fabrice
- Goutal Nicole
- Pascual Valérie
- Hascoët Laurent
La Houille Blanche - Revue internationale de l'eau, EDP Sciences , 2016 (4), pp.57-65 . (10.1051/lhb/2016040)
DOI : 10.1051/lhb/2016040
Algorithmic differentiation applied to the optimal calibration of a shallow water model
- Demangeon Félix
- Goeury Cédric
- Zaoui Fabrice
- Goutal Nicole
- Pascual Valérie
- Hascoët Laurent
La Houille Blanche - Revue internationale de l'eau, EDP Sciences , 2015 . The information on sensitivity provided by derivatives is indispensable in many fields of science. In numerical analysis, computing the accurate value of the derivatives of a function can be a challenge. The classical Finite Differences (FD) method is a simple solution to implement when estimating the value of a derivative. However, it remains highly sensitive numerically and costly in calculation time. Conversely, the Algorithmic Differentiation Method (AD) is a powerful tool for calculating the derivatives of a function described by a computer program. Whatever the complexity of the algorithms implemented in the expression of a function, AD calculates its derivative accurately and reduces development efforts. This article presents the contribution of AD in comparison to FD in the problem of calibrating an industrial class 1D shallow water model. Model calibration is performed by a deterministic mathematical optimiser requiring accurate calculation of the sensitivity of the water surface profile in relation to the friction on a river bed. Two comparative real test cases are presented. They permit validating the better performance expected from AD as a tool used to obtain optimal calibration.
Hydrodynamic modeling and diffusion of the pollutant
- Muttin Frédéric
- Ricchiuto Mario
- Mostefaoui Imene Meriem
- Kirane Mokhtar
- Goeury Cédric
- Hervouet J. M.
, 2014 .
A Lagrangian/Eulerian oil spill model for continental waters
- Goeury Cédric
- Hervouet Jean-Michel
- Baudin-Bizien Isabelle
- Thouvenel François
Journal of Hydraulic Research, Taylor & Francis , 2014, 52 (1), pp.36-48 . (10.1080/00221686.2013.841778)
DOI : 10.1080/00221686.2013.841778
Modélisation du transport des nappes d'hydrocarbures en zone continentale et estuarienne
- Goeury Cédric
, 2012 . L'application de la Directive Cadre sur l'Eau et l'obligation de surveillance de la qualité d'eau pour la consommation humaine et les activités récréatives ou industrielles, telles que la production d'eau potable, entraînent une forte demande pour des systèmes d'évaluation et de suivi de la qualité de l'eau. Le projet de recherche MIGR'HYCAR (http://www.migrhycar.com) a donc été mis en place pour répondre à un besoin opérationnel et à un défaut d'outils d'aide à la décision adaptés face aux déversements d'hydrocarbures en eaux continentales (rivières, lacs et estuaires) qui représente plus de 50% des déversements accidentels en France. Au cours du projet de recherche MIGR'HYCAR, un modèle mathématique de dérive de nappe d'hydrocarbures, composé d'un modèle lagrangien couplé à un modèle eulérien, a été développé dans la plate-forme hydro-informatique TELEMAC (http://www.opentelemac.org). Le modèle lagrangien décrit le mouvement de la nappe en surface en considérant celle-ci comme un ensemble de particules. Ainsi le modèle développé est capable de modéliser les principaux phénomènes agissant sur une nappe d'hydrocarbures une fois celle-ci déversée : convection, diffusion, échouage, re-largage, étalement, évaporation, dissolution et volatilisation. Bien que le phénomène de dissolution ne concerne qu'un très faible volume d'hydrocarbures, ce processus peut avoir des conséquences importantes du point de vue de la toxicité. Afin de suivre l'évolution du pétrole dissous, un modèle eulérien de suivi de traceurs a été adopté. La quantité de traceur dépend directement de la masse dissoute des particules lagrangiennes. Cette approche permet le suivi des hydrocarbures dissous dans la colonne d'eau. Des cinétiques effectuées en laboratoire ont pour but la calibration du modèle numérique. En complément de cas tests issus de la littérature et de cas réels, des résultats expérimentaux issus d'expérimentations effectuées en canal d'essai doivent permettre de vérifier et valider la qualité des simulations numériques sur des situations où les conditions ne sont que partiellement contrôlées