HAL Portal UTC Université de Technologie de Compiègne
Not a member yet
11652 research outputs found
Sort by
Enhancing Research Skills in Engineering Education through Collaborative Digital Tools and Live Experimental Sessions
No full textInternational audienceThis paper investigates the use of digital collaborative tools and live streaming systems to support student initiation into scientific research. A Computer-Supported Cooperative Work in Design (CSCWD) system was implemented to aid two key phases: collaborative planning in a digital workspace and experimental execution in laboratories. In the first phase, multitouch surfaces enabled students to collaboratively define research problems and protocols, fostering idea generation and consensus building. The second phase utilized video capture and streaming systems for documentation and remote participation in experiments. While these tools showed potential, challenges such as camera resolution, latency, and feedback limitations emerged. Initial findings suggest these technologies support research skill development, but further improvements are needed to enhance usability and engagement for both local and remote participants. Future work will focus on addressing these technical and interaction challenges to better support hybrid learning environments
Application des réseaux de neurones convolutionnels à la déconvolution de spectres de neutrons simulé d'un spectromètre d'activation
No full textInternational audienceNeutron studies are of significant interest in fields such as radiation protection, nuclear reactor physics, and criticality safety, where accurate determination of the neutron field is essential. Determining the neutron field typically involves unfolding the detector signals, such as those obtained from the Activation and Counting Neutron Spectrometer (SNAC) detector. Traditional methods, as Bayesian approaches, have been widely integrated for the neutron spectrum unfolding. However, these methods rely on an initial solution estimation, introducing biases or uncertainties. Recent studies in Artificial Intelligence (AI) have demonstrated its potential to address challenges as hysteresis regression. This work is based on our novel convolutional neural network (CNN) architecture to overcome the hysteresis problem in neutron spectrum unfolding. The CNN model predicts the neutron spectrum directly from detector counts, eliminating the need for prior solution predictions. The proposed architecture was trained on a large simulation dataset and validated through a combination of Serpent simulations of various Californium-252 (Cf) spectra and Monte Carlo N-Particles (MCNP) simulations of the Silene reactor. These two complementary simulation approaches are used to evaluate the CNN's evaluation in realistic neutron environments. The results demonstrate the model's high efficiency and accuracy, as evidenced by key performance metrics and the quality of the predicted spectrum (SQ). This approach represents a significant step forward in optimizing and validating AI-based methods for neutron field, especially in criticality dosimetry and radiation protection applications. Preliminary comparisons with Bayesian unfolding codes already indicate than CNN based predictions can capture fine spectral features. A benchmark against MAXED, GRAVEL and Nubay remains a key perspective, together with validation campaigns on neutron facilities
Ontology-Based Approach for Standardization of Sensor Data in Smart Home Environments for Activities of Daily Living (ADL)
No full textInternational audienceOntology-Based Approach for Standardization of Sensor Data in Smart Home Environments for Activities of Daily Living (ADL
A DNN-Based Surrogate Constitutive Equation for Geometrically Exact Thin-Walled Rod Members
No full textInternational audienceKinematically exact rod models were a major breakthrough to evaluate complex frame structures undergoing large displacements and the associated buckling modes. However, they are limited to the analysis of global effects, since the underlying kinematical assumptions typically take into account only cross-sectional rigid-body motion and ocasionally torsional warping. For thin-walled members, local effects can be notably important in the overall behavior of the rod. In the present work, high-fidelity simulations using elastic 3D-solid finite elements are employed to provide input data to train a Deep Neural Newtork-(DNN) to act as a surrogate model of the rod’s constitutive equation. It is capable of indirectly representing local effects such as web/flange bending and buckling at a stress-resultant level, yet using only usual rod degrees of freedom as inputs, given that it is trained to predict the internal energy as a function of generalized rod strains. A series of theoretical constraints for the surrogate model is elaborated, and a practical case is studied, from data generation to the DNN training. The outcome is a successfully trained model for a particular choice of cross-section and elastic material, that is ready to be employed in a full rod/frame simulation
In situ formation and culture of cell spheroids in a low-binding 3D-printed biochip
No full textInternational audienceOrgan-on-chips and microfluidics systems offer new ways to overcome limitations from traditional in vitro models in preclinical studies. However, the lack of standardization and important non-specific binding of tested drugs to devices commonly made in polydimethylsiloxane (PDMS) still slow down their full integration into industrial research pipelines. The goal of this study is to develop a standardized 3D-printed biochip with low-binding properties using perfluoropolyether (PFPE), allowing long-time dynamic cultures of in situ formed cellular spheroids. We first documented the non-specific binding of molecules relevant for pharmaceutical companies, mechanical and surface properties of PFPE as compared with PDMS. A new microstructured biochip was then designed and 3D printed in PFPE to offer a 400 µL chamber containing 384 microwells. The 3D-printing fabrication protocol has been detailed considering key parameters such as UV exposure time or post-curing Finally, 384 HepG2/C3a spheroids were formed per chip in dynamic conditions and maintained for 11 days. The high viability, functionality and polarization of the spheroids cultured in these printed PFPE biochips showed the relevance of this new microphysiological system as alternative to PDMS devices
Development of experimental and numerical methods for understanding and predicting lignocellulosic biomass flows in its valorization processes: preliminary results
No full textInternational audienceThe increasing pressure on industries to reduce their environmental impact and the progressive depletion of natural resources highlight the importance of developing efficient biomass valorization processes. These processes are central to a sustainable bioeconomy but require a deep understanding of the rheological behavior of biomass during handling, transport, and transformation stages. However, biomass flows remain difficult to manage due to the intrinsic complexity of the material (wide granulometric distributions, irregular particle shapes, moisture variability) and the heterogeneity induced by seasonal and biological factors. This research aims to better understand and predict lignocellulosic biomass flow behavior through a combination of experimental and numerical approaches. The methodology is based on four key tasks: (1) physical and physico-chemical characterization of biomass; (2) development of experimental setups to study its rheological properties under various flow regimes; (3) calibration of simulation parameters using DEM and CFD approaches; and (4) validation at pilot scale by comparing experimental and simulated results. Six lignocellulosic biomasses were selected, and preliminary tests were conducted to characterize their physical and rheological properties. The results revealed low repeatability, and additional experiments are currently being carried out to overcome this limitation
Assessing the Sustainability Impacts of Industry 4.0 on Maintenance Policies : A Systematic Literature Review and Future Research Directions
No full textInternational audienceMaintenance strategies have traditionally been designed with a primary focus on cost reduction and operational efficiency, often overlooking their broader environmental and social impacts. However, in the current context where industries must align with European carbon neutrality 2050 objectives and the United Nations Sustainable Development Goals (SDGs), maintenance is recognized as a key lever for enhancing the three pillars of sustainability in industries: economic, social, and environmental. In addition, recent studies have shown that the ongoing digital transformation of industry through Industry 4.0 technologies such as artificial intelligence, Internet of Things, digital twins, and big data analytics, offers new opportunities to improve maintenance strategies. These developments have given rise to the concept of Maintenance 4.0, which opens new perspectives for aligning maintenance practices with broader sustainability objectives.To better understand the impact of these technologies on maintenance sustainability, as well as the existing assessment initiatives in the current state of research, this paper conducts a systematic literature review (SLR). A total of 31 relevant studies were analyzed and classified into literature reviews, conceptual frameworks, and evaluation models. The review reveals that while economic and environmental benefits are increasingly supported by measurable indicators, the social dimension remains underexplored and lacks standardized metrics. In addition, most studies focus on short-term operational gains and do not address life cycle-wide perspective, including manufacturing and end-of-life stages.Based on these findings, this paper (i) clarifies the current maturity of research and its exploratory nature; (ii) identifies major gaps which is the lack of lifecycle-based assessments and operational social indicators; (iii) highlights the weak operationalization of circular economy principles in maintenance 4.0 strategies; and (iv) proposes future research directions to develop holistic, life cycle-oriented, human-centric, and practically validated frameworks. These contributions aim to support the transition toward more sustainable maintenance practices, in alignment with sustainability goals
Flow-induced noise produced by ducted single and double-diaphragm configurations
No full textAéroacoustique : session générale; GAHA - Aéro et Hydro-AcoustiqueNational audienceThe proximity of obstacles in a ventilation network can result in a significant increase of the noise production. In this work, a tandem diaphragm inserted in a rigid duct with rectangular cross section is investigated and compared to a single diaphragm obstruction. The disturbed flow which reaches the second obstacle causes an amplification of the radiated noise which goes beyond a simple doubling of the power. Experimental results, carried with various spacing between the two diaphragms and bulk velocity, allows identifying other aeroacoustic source mechanisms which does not appear with a single diaphragm. Depending on the configuration, the broadband level can increase by more than 10 dB and feedback phenomena can appear at low frequency. Explanations and illustrations of those phenomena are given thanks to compressible-fluid simulations
Méthode Noise Correlation inspired : une approche unique pour l’élastographie par IRM, échographie et imagerie optique
No full textUltrasons biomédicaux - Élasticité et visco-élasticité; GAPSUS - Acoustique Physique, Sous-Marine et Ultra-SonoreNational audienceLes propriétés mécaniques des tissus biologiques, telle que la rigidité, peuvent être modifiées par certaines pathologies. L’élastographie a été développé afin d’estimer la rigidité des tissus de manière non invasive et quantitative. Après avoir fait ses preuves cliniquement en échographie et en IRM, les méthodes d’élastographie ont été étendues au champ de l’imagerie optique. Spécifiquement, l'élastographie par ondes de cisaillement (SWE) est basée sur la génération d’ondes de cisaillement dont la vitesse de propagation est directement reliée à la rigidité du milieu dans lequel elles se propagent. Nous présentons ici une méthode de reconstruction haute résolution et robuste au bruit : la méthode « Noise Correlation inspired » (NCi). L'élastographie par corrélation de bruit permet d'extraire la rigidité des tissus à partir d'un champ d'ondes diffus dans un milieu isotrope. La corrélation entre un point et ses voisins résulte en une tache de refocalisation à partir de laquelle la longueur d'onde moyenne locale de cisaillement est extraite. De celle-ci est estimée la célérité de l’onde et donc la rigidité du tissu étudié. La méthode NCi est une généralisation de cette méthode aux champs d'ondes mécaniques spatialement cohérentes. Ainsi, elle permet d’évaluer la rigidité d’un milieu quel que soit la source des ondes de cisaillement ou la modalité d’imagerie utilisée, permettant ainsi un large champ d’applications dans le domaine médical. La méthode NCi a été validée sur des simulations aux différences finies et testée expérimentalement sur fantôme puis in-vivo : en imagerie ultrasonore avec un échographe (Aixplorer, Supersonic Imagine), en IRM avec un système 7T et en imagerie optique avec un montage d’holographie plein-champ. A chaque modalité est respectivement associée un type de signal : signaux transitoire, signaux harmoniques et champ de bruit. Une comparaison avec les méthodes classiques de reconstruction est présentée afin d’évaluer la méthode NCi
Information Fusion as a Useful Tool to Estimate Parameters from Imprecise Data?
No full textInternational audienceSeveral common estimation problems where one is interested in estimating models from paired input/output observations can be considered as being non-statistical problems: learning preferences from a single user rather than a population, inverse problems where the physical model can be trusted, . . . . In this case, one can easily question whether imperfection in observations and the ensuing estimation should be treated as a statistical problem. In particular, one could challenge the need to model imperfect observation as probabilistic noise, and the fact of considered expected errors as a measure of estimation quality. With these ideas in mind, this work discusses an alternative view, where imperfection is modelled by uncertainty theories that accommodate imprecision by extending set-valued representation, and where estimation is mainly performed by using information fusion tools building upon standard union and intersections rather than averaging and counting considerations