HAL-MINES ParisTech
Not a member yet
27348 research outputs found
Sort by
Time-varying Irradiance Non-Uniformity on PV modules with Horizontal Single Axis Tracker: Modeling vs Measurements
No full textInternational audienceBifacial Horizontal single axis tracker (HSAT) presents an efficient energy production potential due to the capture of both down- and upwelling irradiance on the front and rear faces of the photovoltaic modules and this ability to follow the path of the Sun. This work presents the validation of an optical model based on ray tracing, compared to in-situ measurements consisting of a set of reference cells. The experimental site chosen in this study is the Bifacial Experimental Single-Axis Tracking Field (Best Field) of the National Renewable Energy Laboratory (NREL). The simulations are performed both on the reference cells and on the PV cells of a row consisting of 20 modules. A matrix-based ray-tracing approach allows us to efficiently obtain, at a spatial scale finer than the cells, time series over the period of one year at one minute resolution for all simulations when the industrial standard is typically one hour. Over this period, the relative bias on the front face is 2%. A segmentation according to the angle of incidence and the clear-sky index reveals variations on the relative bias between -16% and 3%. For the back face with several in-situ measurements, the relative bias is between -6% and 0%
Enjeux géonumériques pour l'évaluation environnementale des projets urbains : premiers résultats du projet EcoCIM
No full textInternational audienceReducing the environmental impact of the building stock (construction, use, rehabilitation, end of life) is essential in order to achieve a sustainable rate of consumption of the planet's resources. Life cycle analysis (LCA) enables stakeholders to assess several categories of impact of an urban project (damage to human health, damage to biodiversity, etc.) in order to make the most sober choices. This assessment requires a large amount of data that is as representative as possible of the concerned territories. In view of the difficulties encountered in taking into account the geographical dimension, an exploratory interdisciplinary project has been launched to identify spatial challenges to overcome when carrying out a LCA of an urban project. This article presents the initial results of the project, beginning with a description of the issues involved in the spatialization of LCA, followed by a description of some of the challenges encountered more specifically in the context of LCA of urban projects. The aim is to help define a research programme to improve the way in which the geographical dimension is taken into account in the LCA of urban projects, and to inform the geospatial scientific community of the problems encountered by environmental assessment researchers working to provide professional communities with robust and reliable LCA methods and tools.La réduction des impacts environnementaux du parc bâti (construction, usage, réhabilitation, fin de vie) est nécessaire pour atteindre un rythme soutenable de consommation des ressources planétaires. L’analyse de cycle de vie (ACV) permet à un maître d’ouvrage d’évaluer plusieurs catégories d’impacts environnementaux d’un projet urbain (atteintes à la santé humaine, atteintes à la biodiversité, etc.) afin de faire les choix les plus sobres possibles. Cette évaluation nécessite des données nombreuses et les plus représentatives possible des territoires concernés. Face aux difficultés rencontrées lors de la prise en compte de la dimension géographique, un projet exploratoire et interdisciplinaire a été lancé afin d’identifier les besoins géonumériques émergeant lors de la réalisation de l’ACV d’un projet urbain. Cet article présente les premiers résultats du projet en décrivant d’abord les enjeux liés à la spatialisation de l’ACV puis une partie des défis rencontrés plus spécifiquement dans le contexte de l’ACV des projets urbains. Il a pour objectif de participer à la définition d’un programme de recherche pour améliorer la prise en compte de la dimension géographique dans les ACV de projets urbains et de porter à la connaissance de la communauté en sciences et techniques de l’information géographique les problématiques rencontrées par les chercheurs en évaluation environnementale qui développent des méthodes et outils d’écoconception robustes et fiables à destination des communautés professionnelles
Data-driven inference of Boolean networks from transcriptomes to predict cellular differentiation and reprogramming
No full textInternational audienceBoolean networks provide robust explainable and predictive models of cellular dynamics, especially for cellular differentiation and fate decision processes. Yet, the construction of such models is extremely challenging, as it requires integrating prior knowledge with experimental observation of transcriptome, potentially relating thousands of genes. We present a general methodology, implemented in the software tool BoNesis, for the qualitative modeling of gene regulation behind the observed state changes from transcriptome data and prior knowledge of the gene regulatory network. BoNesis allows computing ensembles of Boolean networks, where each of them is able to reproduce the modeled differentiation process. We illustrate the scalability and versatility of BoNesis with two applications: the modeling of hematopoiesis from single-cell RNA-Seq data, and modeling the differentiation of bone marrow stromal cells into adipocytes and osteoblasts from bulk RNA-seq time series data. For this later case, we took advantage of ensemble modeling to predict combinations of reprogramming factors for trans-differentiation that are robust to model uncertainties due to variations in experimental replicates and choice of binarization method. Moreover, we performed an in silico assessment of the fidelity and efficiency of the reprogramming, and conducted preliminary experimental validation
Static recrystallization of an advanced Ti2AlNb alloy after compression in single- β and two-phase β +O and β + α 2 domains
No full textInternational audienc
From error to inclusion : the transformative potential of mistakes in organisations
No full textInternational audienceIn a context where the inclusion of young people distanced from employment represents a major societal and organisational challenge (recruitment difficulties, retention issues, retirements), this research explores the tensions at play within experimental schemes aimed at fostering their professional and organisational inclusion. It examines error as a potential lever for learning and organisational transformation.This study is grounded in a cross-disciplinary literature that combines perspectives on inclusive organisations (Shore et al., 2011, 2018; Randel, 2025), the paradoxical tensions inherent to inclusion (Ferdman, 2017), learning organisations (Argyris & Schön, 1978, 1986), and communities of practice (Pesqueux, 2022). Error is approached as both an indicator of structural tensions and a catalyst for invisible learning, paving the way for deep organisational learning. The central research question is: how can the mistakes of newcomers and organisational actors be mobilised as learning resources to support the inclusive transformation of organisations?Conducted as an action-research study within the “100% Inclusion” programme of the Compagnons du Devoir, the methodology combines 150 hours of observation and 161 semi- structured interviews, analysed through thematic coding.The findings identify five types of errors (orientation, learning, sanctioned, opportunity-driven, dysfunction-indicating) and five transformative practices (self-evaluation, delayed analysis, trial and error, formative feedback, resilience).This study advocates for a reappraisal of error as a tool for inclusive governance, proposing an innovative articulation between organisational learning and inclusive organisation
Interactive learning in safety: how serious game can enhance retention and collaboration
No full textInternational audienceTraining employees on critical safety protocols is a challenge in many industries, with traditional methods often resulting in disengagement and poor retention. Gamification offers a promising alternative by using game mechanics to foster active participation, thereby enhancing memory retention and learning outcomes. This communication explores the theoretical foundations of gamification in training, focusing on its ability to create immersive, interactive learning environments that stimulate motivation and engagement. However, applying gamification to safety raises important questions about balancing entertainment with the gravity of the content. Are playful elements compatible with the responsibility of ensuring employee safety? To address this, we propose a concrete experiment: a low-tech orienteering-based training for Second Intervention Teams (ESI) in an industrial setting. The gamified training integrates fire safety challenges into a real-world navigation task, prompting participants to recall safety procedures while working together to achieve goals. After a first full-scale experimentation, observations and debriefing with the trainees shows that this approach improves engagement, teamwork, and safety protocol retention. However, the experiment also highlights the limitations of gamification, particularly the need for careful design to prevent trivializing serious safety content. This study concludes by reflecting on the conditions under which gamification can serve as an effective complement to traditional safety training, without compromising the importance and seriousness of the subject matter
Immersive virtual reality for soft-skills training of front-desk agents: Collective creation and appropriation of a virtual learning environment
No full textInternational audienceImmersive virtual reality has gained a renewed popularity in the last decade, following the commercialization of HMDs and the surge in computing power. Allowing the creation of customized virtual worlds and both cognitive and emotional investment in these worlds for its users, immersive virtual reality has been used in education in training across several areas, mostly in the industrial sector and for procedural and technical knowledge. Research on training soft skills with this technology remains scarcer, and often does not question the adaptation of training processes by organizational members to integrate this technology. In this paper, we seek to fill this gap by presenting an intervention research with several partners aiming to create and test a customized learning environment for front-desk agents' training in a French public service organization. We underline how such a collaboration helps creating a customized virtual environment in which professional front-desk agents feel natural, and a virtual training scenario that adequately simulates real-life work situations, allowing professionals to appropriate the technology for their training. Then, we highlight how this technology enables professionals with different positions in the organization to share a same work experience, enabling them to collectively discuss the co-evolution of the technology and their training process
Neural Variational Data Assimilation with Uncertainty Quantification Using SPDE Priors
No full textInternational audienceAbstract The spatiotemporal interpolation of large geophysical datasets has historically been addressed by optimal interpolation (OI) and more sophisticated equation-based or data-driven data assimilation (DA) techniques. Recent advances in the deep learning community enable to address the interpolation problem through a neural architecture incorporating a variational data assimilation framework. The reconstruction task is seen as a joint learning problem of the prior involved in the variational inner cost, seen as a projection operator of the state, and the gradient-based minimization of the latter. Both prior models and solvers are stated as neural networks with automatic differentiation which can be trained by minimizing a loss function, typically the mean-square error between some ground truth and the reconstruction. Such a strategy turns out to be very efficient to improve the mean-state estimation but still needs complementary developments to quantify its related uncertainty. In this work, we use the theory of stochastic partial differential equations (SPDEs) and Gaussian processes (GPs) to estimate both space- and time-varying covariance of the state. Our neural variational scheme is modified to embed an augmented state formulation with both state and SPDE parameterization to estimate. We demonstrate the potential of the proposed framework on a spatiotemporal GP driven by diffusion-based anisotropies and on realistic sea surface height (SSH) datasets. We show how our solution reaches the OI baseline in the Gaussian case. For nonlinear dynamics, as almost always stated in DA, our solution outperforms OI, while allowing for fast and interpretable online parameter estimation
Augmented Quantization: Mixture Models for Risk-Oriented Sensitivity Analysis
No full textA central question in risk analysis is to identify the factors that drive the system toward a specific hazardous outcome, such as the exceedance of a given threshold. When relying on numerical simulators, we propose to study the distribution of the inputs, transformed into uniform variables via their cumulative distributions, conditionally on the occurrence of the hazardous event. To represent this multivariate conditional distribution for sensitivity analysis, we introduce an original quantization approach based on estimating a mixture of Dirac and local uniform distributions. For each marginal of this mixture, a Dirac component indicates a strong influence of the corresponding variable, whereas a uniform component with wide support reflects weak influence. A notable advantage of this method is its ability to identify the regions of the input space that most strongly influence the occurrence of the risk event, while also capturing the joint effects of multiple variables. However, learning mixture models typically relies on likelihood-based methods, which are not well suited to mixtures involving singular or Dirac components. To address this, we propose an \emph{Augmented Quantization} method, a reformulation of the classical quantization problem based on the -Wasserstein distance, which can be computed in very general distribution spaces. The performance of Augmented Quantization in estimating such mixture models is first demonstrated on analytical toy problems, and then applied to sensitivity analysis of both an analytical function and a practical flooding case study on a section of the Loire River
Global decoupling of functional and phylogenetic diversity in plant communities
No full textAll R scripts used for this study can be found in our GitHub repository at : https://github.com/georghaehn/Haehn-et-al-2024-FD-PD-couplingInternational audiencePlant communities are composed of species that differ both in functional traits and evolutionary histories. As species’ functional traits partly result from their individual evolutionary history, we expect the functional diversity of communities to increase with increasing phylogenetic diversity. This expectation has only been tested at local scales and generally for specific growth forms or specific habitat types, for example, grasslands. Here we compare standardized effect sizes for functional and phylogenetic diversity among 1,781,836 vegetation plots using the global sPlot database. In contrast to expectations, we find functional diversity and phylogenetic diversity to be only weakly and negatively correlated, implying a decoupling between these two facets of diversity. While phylogenetic diversity is higher in forests and reflects recent climatic conditions (1981 to 2010), functional diversity tends to reflect recent and past climatic conditions (21,000 years ago). The independent nature of functional and phylogenetic diversity makes it crucial to consider both aspects of diversity when analysing ecosystem functioning and prioritizing conservation efforts