Portail des publications scientifiques IMT Mines Alès
Not a member yet
5198 research outputs found
Sort by
Renewable Energy Mix System Modelling with Power Grid Integration: Methods, Approach and Challenges
No full textInternational audienceRenewable energy mix systems (RES) are crucial for sustainable energy transitions, but their integration with power grids poses technical, economic, and regulatory challenges. This paper addresses the critical challenges and methodologies associated with modeling RES integrated with power grids, essential for the global transition to sustainable energy. It examines the complexities of integrating variable renewable sources like wind and solar, highlighting technical, financial, environmental, and regulatory barriers through a comprehensive literature review. To overcome these obstacles, a decision centric framework is proposed, utilizing a Global Information System (GIS) based approach to optimize site selection, energy mix, and grid stability. This framework employs machine learning for enhanced grid feasibility assessments along with contingency analysis, incorporating predictive maintenance and demand supply balancing, alongside detailed simulations using tools like DigSILENT PowerFactory and AnyLogic. The methodology culminates in a decisional model that balances technical, economic, and environmental factors, ensuring the development of robust and efficient renewable energy systems
Examining placebo effects after a 3-week repeated-sprint training program under hypoxic conditions in recreationally trained subjects: a randomized controlled trial
No full textInternational audienceRepeated-sprint training in hypoxia (RSH) has been suggested to significantly enhance anaerobic performance. However, the widespread belief in the benefits of altitude training raises questions about potential placebo effect. The aim of this study was thus to investigate the physiological and placebo effects of normobaric hypoxia combined with repeated-sprint training on performance. Twenty-nine moderately trained participants were randomly assigned to normoxia (repeated-sprint in normoxia (RSN)), placebo (repeated-sprint in normoxia with placebo setup (RSN-P)), or hypoxia (RSH) groups. Participants in RSN-P group were led to believe they were training at simulated altitude (between 2500 and 3500 m), while participants in the RSN and RSH groups knew they were training at sea level and at altitude, respectively. Repeated-sprint training involved six cycling sessions over 3 weeks, consisting of three sets of 8 × 6 s sprint with 24 s of recovery. There was no difference in the estimation of the altitude level to which participants from the RSN-P and RSH groups thought they had been exposed. There was a main significant effect of training on mean power output during Wingate (+7.9%; p < 0.001; ηp² = 0.47) and repeated-sprint ability tests (+7.7%; p < 0.001; ηp² = 0.55). However, contrary to our hypotheses, the enhancement among the three groups did not differ. The lack of greater improvement in the RSH group compared to the other groups raises questions about the added value of hypoxia in these conditions. In conclusion, neither real nor perceived hypoxia enhanced training adaptations following repeated-sprint sessions
Machine Unlearning for Industrial AI Models
No full textInternational audienceIndustrial AI models are increasingly used to replace simulation tools for predicting key performance indicators (KPIs). They offer faster predictions but may introduce errors that can affect critical decisions. Due to the complexity of industrial data, it is often difficult to identify the input conditions under which a model performs poorly. Clustering is therefore used as a preliminary step to detect groups of demands with high prediction errors. This paper investigates a selective machine unlearning strategy applied to these weak groups. Instead of retraining the model from scratch or discarding large portions of data, the approach focuses on removing or reweighting only harmful patterns and preserving useful information. Experiments on a large industrial dataset show that this strategy reduces errors in weak groups and maintains stable performance across the remaining data. This improvement supports the use of AI models as reliable surrogates for industrial decision-making
Enculturation's Rhythmic Imprint: Cross-Cultural Variations in Sensorimotor Synchronization
No full textInternational audienceGroup dynamics are more than a mere aggregation of individual behaviors; they emerge from the complex interplay interactions, which are deeply context dependent. Notably, a growing body of literature shows that enculturation introduces variations, as the rhythmic exposure since childhood shapes how individual synchronize their actions. Thereby, cultural environment molds group dynamics by influencing the most basic element of interpersonal interactions: individual perception-action dynamics. To examine this influence, we conducted two experiments to identify cross-cultural variations between French and Indian participants engaging in classical synchronization tasks. The results were analyzed through the lens of the dynamical system approach using forced oscillator models. In the first study, we documented the synchronization dynamics of both groups. The results revealed that while French participants exhibited negative mean asynchrony, Indian participants had a positive mean asynchrony. We proposed that this difference was partly due to a bigger over-estimation of the frequency of the rhythm for French participants than for Indian participants. In the second study, we explored the frequency adaptation mechanism supposed to be the cause of the observed over-estimation. Overall, our findings suggests that enculturation imprints down our most elementary temporal neural functions, thereby modifying our rhythmic behaviors and altering group dynamics
A Review on the Valorization of Recycled Glass Fiber-Reinforced Polymer (rGFRP) in Mortar and Concrete: A Sustainable Alternative to Landfilling
No full textInternational audienceThe recycling of glass fiber-reinforced polymer (GFRP) in cementitious materials is an interesting way of managing the end of life of this type of material. As the solutions of landfilling and incinerating have reached their limits, the material recovery by recycling approach appears to be suitable to develop cement-based materials with enhanced properties. Different recycling methods, including mechanical, thermal and chemical recycling, are commonly used for the recovery of fibers and resins. Mechanical recycling is more suitable due to its low cost and ease of implementation. Moreover, mechanical recycling has limited environmental impact and is ideal for use with cementitious materials (fiber and resin). Several studies are being conducted to find the best incorporation method, notably the incorporation of recycled GFRP of different sizes (small, medium, large and coarse) and shapes (fibrous, cubic, random) as a substitute for sand and/or aggregate in mortars and concretes or as reinforcement materials. This article aims to establish a state of the art perspective on the incorporation of rGFRP into cement-based materials. The benefits of this incorporation are highlighted as well as the limitations. The various challenges to be overcome to make this incorporation useful from a practical point of view are reported
Thermally conductive filled polymers with low environmental impact
No full textInternational audienceDans le cadre de l’électrification des usages, le développement de matériaux polymères thermoconducteurs à faible impact environnemental constitue un enjeu stratégique pour dissiper la chaleur aux interfaces dans les applications électroniques (Thermal Interface Materials : TIMs). Ce travail s’inscrit dans une démarche d’éco-conception visant à formuler des composites modèles, en utilisant une matrice distincte de celles employées dans les applications finales. Ces composites sont constitués d’une matrice polymère de polyéthylène basse densité, recyclé (rLDPE) ou vierge (LDPE), et de plaquettes de nitrure de bore hexagonal (h-BN). Ces composites doivent pouvoir atteindre une conductivité thermique supérieure à 4 W·m⁻¹·K⁻¹, tout en minimisant leur empreinte environnementale. L’approche adoptée combine l’évaluation environnementale par Analyse de Cycle de Vie (ACV) de la production du composite et des formulations contenant 50% en masse de charges, pour améliorer les propriétés de conductivité thermique des composites. L’ACV est réalisée selon les normes ISO 14040-14044, du berceau à la porte, les impacts environnementaux sont calculés avec la méthode ‘Product Environmental Footprint’
Determinants and consequences of climate commitment among facilitators of environmental awareness workshops: A Mixed-Methods Exploratory Study
No full textProtocoleClimate activism is an expanding form of civic engagement in response to the ecological crisis. Among these initiatives, the facilitation of participatory environmental workshops (e.g., climate, biodiversity, oceans, food systems, digital sustainability) has emerged as a structured and accessible way to disseminate knowledge, foster collective awareness, and promote pro-environmental behavior. In France alone, more than 150 different workshop formats exist, involving thousands of volunteer and professional facilitators. However, little is known about the psychosocial determinants of such engagement, nor about its behavioral and psychological consequences for facilitators themselves. This study aims to explore 1) the determinants of sustained engagement as a facilitator of environmental awareness workshops, and 2) the associated outcomes in terms of pro-environmental behaviors and climate-related emotions.Methods. This cross-sectional mixed-methods study will include (a) an online survey distributed to facilitators of environmental awareness workshops across France, with the objective of reaching at least 500 respondents, and (b) semi-structured interviews with a subsample of approximately 30 participants. The quantitative component will assess multiple dimensions of facilitation engagement, pro-environmental behaviors (mitigation, adaptation, and collective action), climate-related emotions (including eco-anxiety, hope, and helplessness), as well as psychosocial variables such as ecological identity, selfand collective efficacy, sense of community, and political orientation. The qualitative component will provide in-depth insights into facilitatorsʼ motivations, perceived barriers, and subjective experiences of the personal and collective impact of their involvement.Expected Results. The study will generate a multidimensional profile of facilitator engagement and identify key psychosocial predictors of sustained involvement. It will also document potential behavioral and psychological outcomes of facilitation, thereby informing strategies to better support, recognize, and sustain volunteer-driven environmental engagement.</p
The impact of thermal and auditory unpleasant stimulus on explicit motor imagery in healthy individuals: An experimental study
No full textInternational audienceMotor imagery is the ability to mentally simulate a motor task without actually performing it. Furthermore, pain is an unpleasant sensory experience that involves different dimensions – sensory-discriminative, motivational-affective, and cognitive-evaluative – that are known to interfere with motor imagery. However, it remains unclear which specific pain dimension most significantly impairs motor imagery vividness. This study aims to compare the effects of unpleasant auditory (primarily affective and cognitive) and thermal (primarily sensory) stimuli, which can be assimilated to pain, on discrete and continuous explicit motor imagery sessions. Eighteen healthy participants were exposed to these unpleasant stimuli in addition to a control condition. Participants rated the vividness of their explicit motor imagery after performing full back-and-forth flexion and extension of their wrists in discrete and continuous sessions. Results showed that during discrete explicit motor imagery, only the aversive auditory stimulus significantly reduced motor imagery vividness, whereas thermal pain had no effect. In contrast, motor imagery vividness was preserved during the continuous session. These findings suggest that explicit motor imagery may be more affected by the affective dimension of pain induced by aversive auditory stimuli. The preservation of motor imagery vividness in the continuous session provides insight into the optimization of rehabilitation programs
Towards evaluating set-valued predictions with partial observations
No full textInternational audienceIn different situations, one will have to evaluate set-valued predictions while having set-valued/partial observations. This is the case, for instance, when experts provide weak supervision and when predictions are provided by robust classifiers. However, we still miss some principled approach to evaluate such predictions and observations. This is the goal of this paper, where we propose guidelines about such evaluations functions. We will set our study within the framework of belief functions, that seem quite adapted to the considered problem