HAL-Ecole des Ponts ParisTech
Not a member yet
42743 research outputs found
Sort by
Unleashing the transformative power of deliberation with contextual citizens
No full textInternational audienceIn this paper, we investigate deliberation procedures that invite citizens with contextual opinions to explore alternative thinking frames. Contextuality is captured in a simple quantum cognitive model. We show how disagreeing citizens endowed with contextual opinions can reach consensus in a binary collective decision problem with no improvement in their information. A necessary condition is that they are willing to (mentally) experience their fellow citizens’ way of thinking. The diversity of thinking frames is what makes it possible to overcome initial disagreement. Consensus does not emerge spontaneously from deliberations: it requires facilitation.This article is part of the theme issue ‘Quantum theory and topology in models of decision making (Part 1)’
French market design in practice: Some lessons from the 2022 energy crisis
No full textInternational audienceBetween 2005 and 2021, France has generated more electricity from fossil-free resources (491 TWh/year on average) than its gross domestic consumption (481 TWh/year). Therefore, in terms of total surplus, the French electricity sector should have been barely hit, if at all, by the surge in fossil fuel prices during the 2022 energy crisis. In practice, however, the French government spent billions of euros in subsidies to electricity consumers, the incumbent utility – who operates the whole nuclear fleet – recorded its worst yearly financial result to date, and total electricity imports exceeded exports for the first time in more than 40 years. Although these outcomes can largely be attributed to bad luck, the extent to which they could have been mitigated through better market design and public policies is an open question. This article argues that existing policies, through their implied incentives to share and manage long-term risks, played a critical role in how France fared during the energy crisis. Consistently, reforming long-term risk-sharing mechanisms has emerged as the most pressing issue to address. Looking forward, however, updating short-term wholesale market design will likely prove increasingly important to better support a low-cost and reliable energy transition
Application de l'équation de Lippmann-Schwinger à la propagation des ondes acoustiques de Bloch
No full textOndes en milieu hétérogène; GAPSUS - Acoustique Physique, Sous-Marine et Ultra-SonoreNational audienceLes méthodes par FFT offrent un certain nombre d'avantages pour l'homogénéisation des milieux périodiques et aléatoires, tels qu'un faible coût en mémoire et une rapidité de calcul par rapport à la méthode des éléments finis. Depuis les travaux de Segurado et Lebehnson (2021), les méthodes par FFT sont appliquées au calcul des propriétés acoustiques (diagramme de dispersion) de matériaux périodiques arbitraires. Mais ces travaux restent limités aux structures faiblement contrastées tels que les polycristaux, car les algorithmes utilisés sont très sensibles au conditionnement des opérateurs. Dans cette présentation, nous proposons une méthode par FFT alternative, basée sur le théorème de Bloch et l'équation de Lippmann-Schwinger. Nous utilisons les travaux de Sab et al (2024), qui ont introduit un algorithme inconditionnellement convergent, même en présence de pores et de phases infiniment rigides. Ceci permet d'étendre le calcul des diagrammes de dispersion par FFT à des structures très contrastées, avec un faible temps de calcul. Nous présenterons également les premiers résultats sur des micro-structures simples
Production of photovoltaic modules with irradiation enhanced by flat metal reflectors
No full textNational audienceFixed plane reflectors are the simplest and most rustic way of increasing the photovoltaic output of a given surface area of flat modules. The gains from PV-reflector systems depend on climate, latitude and the geometry chosen. Theoretical and experimental studies [1] have shown that annual production gains of the order of 5 to 25% are easily achievable at mid-latitudes.What's more, the geometry makes it possible to select more or less the two periods in spring and autumn when yield is highest.The recent drop in the price of silicon modules means that by 2025, reflectors will no longer be an economically viable option if the sole aim is to generate electricity. However, they could play a role in reducing demand for rare materials in the future. In addition, installing them on flat roofs offers advantages in terms of building insulation, heat recovery heating in winter and limiting urban warming in summer.A system based on low LID PERC crystalline silicon modules with full or half cells and reflectors made of raw rolled aluminium or polished stainless steel was installed on the roof of the ‘École polytechnique novation centre’. Each module is equipped with an optimiser that allows it to operate at its optimal power independently of the other modules in its string. This study, benefits from the international standard meteorological measurements of the SIRTA - Palaiseau platform, a member of the BSRN which is based on Institut Polytechnique de Paris campus . This allows for a detailed analysis of the properties of this energy system.Measurements taken during the first two years showed that the gain due to the reflec-tors was highest in March and October and could reach 13%. Surface oxidation did notcause any measurable degradation in photo-voltaic production during these two years.Half-cells do not appear to offer any decisive advantage in this situation. Similarly, rawaluminium reflectors are not significantly less effective than polished stainless steel reflec-tors. A more in-depth four-year study should provide quantitative answers to questionsconcerning the efficiency of reflectors and modules and degradation. Metal reflectors have theadvantage of being fully recyclable, but optical engineering could improve their reflective properties by increasing the solar radiation reflected in the active band of photovoltaic cells and reducing the infrared flux reflected towards the module. In this way, it may be possible to increase the photoelectric current and reduce cell overheating, which tends to reduce photovoltaic efficiency.The results obtained over four years with this experimental system, will be presented.[1] M. Pavlov, A. Migan-Dubois, V. Bourdin, M. Pons, M. Haeffelin, et al.. Experimental and Numerical Study of the Influence of String Mismatch on the Yield of PV Modules Augmented By Static Planar Reflectors. IEEE Journal of Photovoltaics, 5 (6), (2015) pp.1686 – 1691
Interactive Occlusion Boundary Estimation through Exploitation of Synthetic Data
No full textInternational audienceOcclusion boundaries (OBs) geometrically localize occlusion events in 2D images and provide critical cues for scene understanding. In this paper, we present the first systematic study of Interactive Occlusion Boundary Estimation (IOBE), introducing MS 3 PEa novel multi-scribble-guided deep-learning framework that advances IOBE through two key innovations: (1) an intuitive multi-scribble interaction mechanism, and (2) a 3encoding-path network enhanced with multi-scale strip convolutions. Our MS 3 PE surpasses adapted baselines from seven state-of-the-art interactive segmentation methods, and demonstrates strong potential for OB benchmark construction through our realuser experiment. Besides, to address the scarcity of well-annotated real-world data, we propose using synthetic data for training IOBE models, and developed Mesh2OB, the first automated tool for generating precise ground-truth OBs from 3D scenes with selfocclusions explicitly handled, enabling creation of the OB-FUTURE synthetic benchmark that facilitates generalizable training without domain adaptation. Finally, we introduce OB-LIGM -a high-quality real-world benchmark comprising 120 meticulously annotated high-resolution images advancing evaluation standards in OB research
Network-Wide Traffic Volume Estimation from Speed Profiles using a Spatio-Temporal Graph Neural Network with Directed Spatial Attention
No full textInternational audienceExisting traffic volume estimation methods typically address either forecasting traffic on sensor-equipped roads or spatially imputing missing volumes using nearby sensors. While forecasting models generally disregard unmonitored roads by design, spatial imputation methods explicitly address network-wide estimation; yet this approach relies on volume data at inference time, limiting its applicability in sensor-scarce cities. Unlike traffic volume data, probe vehicle speeds and static road attributes are more broadly accessible and support full coverage of road segments in most urban networks. In this work, we present the Hybrid Directed-Attention Spatio-Temporal Graph Neural Network (HDA-STGNN), an inductive deep learning framework designed to tackle the network-wide volume estimation problem. Our approach leverages speed profiles, static road attributes, and road network topology to predict daily traffic volume profiles across all road segments in the network. To evaluate the effectiveness of our approach, we perform extensive ablation studies that demonstrate the model's capacity to capture complex spatio-temporal dependencies and highlight the value of topological information for accurate network-wide traffic volume estimation without relying on volume data at inference time
Tunability of ice supersaturation in the upper troposphere and lower stratosphere in the LMDZ atmospheric model
No full textModeling the Coupled and Decoupled states of Polar Boundary-Layer Mixed-Phase Clouds
No full textInternational audienceAbstract. Representing mixed-phase clouds (MPCs) is a long-standing challenge for climate models, with major consequences regarding the simulation of radiative fluxes at high-latitudes and uncertainties in future cryosphere melting estimates. Low-level boundary-layer MPCs that prevail at high-latitudes can be either coupled or decoupled to the surface, which modulates their dynamical and microphysical properties. This study leverages a recent physically-based parameterization of phase partitioning considering an explicit coupling between microphysics and subgrid-scale dynamics and involving direct interactions between the cloud and turbulent diffusion schemes. This parameterization makes it possible to capture the structure of the decoupled state of polar boundary-layer MPCs – with a supercooled liquid dominated cloud-top sitting on top of precipitating ice crystals – in single column simulations with the LMDZ Atmospheric General Circulation Model. The positive feedback loop involving cloud-top radiative cooling induced by supercooled liquid droplets, subsequent buoyancy production of turbulence as well as the supercooled liquid water production associated with turbulence, is captured by the model. However, the liquid and cloud ice water path remain slightly underestimated which may be due to an underestimation of the net upward water flux from low layers. The paper further shows that accounting for the detrainment of shallow convective plume's air when diagnosing the in-cloud supersaturation makes it possible to capture the overall vertical structure of surface-coupled clouds, with realistic liquid and ice water contents. A parameteric sensitivity analysis further shows the importance of properly calibrating the parameter controling the supercooled liquid water production term by subgrid turbulence
A Discrete Analog of Tutte's Barycentric Embeddings on Surfaces
No full textInternational audienceTutte's celebrated barycentric embedding theorem describes a natural way to build straight-line embeddings (crossing-free drawings) of a (3-connected) planar graph: map the vertices of the outer face to the vertices of a convex polygon, and ensure that each remaining vertex is in convex position, namely, a barycenter with positive coefficients of its neighbors. Actually computing an embedding then boils down to solving a system of linear equations. A particularly appealing feature of this method is the flexibility given by the choice of the barycentric weights. Generalizations of Tutte's theorem to surfaces of nonpositive curvature are known, but due to their inherently continuous nature, they do not lead to an algorithm.In this paper, we propose a purely discrete analog of Tutte's theorem for surfaces (with or without boundary) of nonpositive curvature, based on the recently introduced notion of reducing triangulations. We prove a Tutte theorem in this setting: every drawing homotopic to an embedding such that each vertex is harmonious (a discrete analog of being in convex position) is a weak embedding (arbitrarily close to an embedding). We also provide a polynomial-time algorithm to make an input drawing harmonious without increasing the length of any edge, in a similar way as a drawing can be put in convex position without increasing the edge lengths
Modeling the Coupled and Decoupled states of PolarBoundary-Layer Mixed-Phase Clouds
No full textRepresenting mixed-phase clouds (MPCs) is a long-standing challenge for climate models, with major consequences regarding the simulation of radiative fluxes at high-latitudes and uncertainties in future cryosphere melting estimates. Low-level boundary-layer MPCs that prevail at high-latitudes can be either coupled or decoupled to the surface, which modulates their dynamical and microphysical properties. This study leverages a recent physically-based parameterization of phase partitioning considering an explicit coupling between microphysics and subgrid-scale dynamics and involving direct interactions between the cloud and turbulent diffusion schemes. This parameterization makes it possible to capture the structure of the decoupled state of polar boundary-layer MPCs -with a supercooled liquid dominated cloud-top sitting on top of precipitating ice crystals -in single column simulations with the LMDZ Atmospheric General Circulation Model. The positive feedback loop involving cloud-top radiative cooling induced by supercooled liquid droplets, subsequent buoyancy production of turbulence as well as the supercooled liquid water production associated with turbulence, is captured by the model. However, the liquid and cloud ice water path remain slightly underestimated which may be due to an underestimation of the net upward water flux from low layers. The paper further shows that accounting for the detrainment of shallow convective plume's air when diagnosing the incloud supersaturation makes it possible to capture the overall vertical structure of surface-coupled clouds, with realistic liquid and ice water contents. A parameteric sensitivity analysis further shows the importance of properly calibrating the parameter controling the supercooled liquid water production term by subgrid turbulence.</div