HAL Portal UTC Université de Technologie de Compiègne
Not a member yet
11652 research outputs found
Sort by
TRACKSIM: A Multi-Level Simulation Framework for Near-Life Battery Data Generation
No full textThe authorswould like to thank Yusheng Zheng and Yunhong Chefrom the Department of Energy, Aalborg University forthe acquirement of experiment data used in this paper.International audienceData-driven methods in the battery intelligence field are highly dependent on data similar to the targeted application. This paper introduces the multi-level simulation framework TRACKSIM as a tool to generate realistic synthetic battery cell data for training data-driven models. A use-case for State-Of-Health (SOH) estimation is presented
A High-Efficiency 5.5kW Battery Backup Unit for OCP Open Rack V3 Using Partial Power Conversion
No full textInternational audienceBattery Backup Units (BBUs) are key components in modern, AI-centric data centers where uptime, efficiency, and high-power capability are critical. Additionally, BBUs can help to reduce the strain on grid infrastructure, due to their peak shaving capability.This paper presents a highly efficient BBU based on a 5.5 kW DC-DC converter using a partial power concept. The output power can be maintained for a battery voltage from 35 to 62 V according to OCP V3 (50V). A power density of 290 W/inch 3 is achieved with a converter size of 75 x 128 x 36 mm while maintaining an efficiency above 99%. A prototype has been built, and experimental results are presented.</p
SiC Solid-state Power Contactors SSPC`s in comparison to E-Fuses and Solid-state Circuit breakers
No full textInternational audienceIn this work, the novel approach of a DC protection system based on an all-solid-state semiconductor PCB solution was presented. The switching element, called the "solid-state power contactor -SSPC", is implemented as a circuit board. Modern system integration technologies are used here. The focus is on circuit board technologies that distribute the power losses of the semiconductors in a targeted manner on the circuit board and can very quickly reduce high temperature gradients in the semiconductors in critical operating conditions. There are fundamental differences between the three systems E-Fuse, SSCB and SSPC, depending on their intended use. This is why E-Fuse is primarily considered a fuse replacement system, SSCB as an on/off switching device and the SSPC presented here as a protective device for functional safety must enable "safe disconnection and switching". In addition, the naming E-Fuse, SSCB or SSPC are not absolute
Good practices to sort 650 V GaN die inside a PCB
No full textInternational audienceThis work studies the embedding of 650V d-mode GaN within a PCB using double-side microvias. The study begins with a comparison of electrical parameters before and after packaging in a sample of devices. This comparison aims to evaluate the packaging impact. After, all packaged devices are characterized and statistically classified. This classification allows detecting some parts with atypical values. X-ray tomography on these atypical parts highlight backside die delamination or voids on them. However, this failure detection method could be time-consuming. Therefore, this study proposes a method based on S-parameters for early failure detection. S-parameters analysis and test results on the identified backside delaminated devices reveal their tendency to have higher intrinsic capacitance. Finally, lessons learned through this study have led to propose some guidelines to improve the embedded PCB process for better sort practices
Battery Efficiency and Aging Measurements for Multilevel Battery Energy Storage Systems -DC vs. 100Hz
No full textInternational audienceBatteries in single phase DC/AC converters can be subject to second harmonic currents. This work examines the difference in battery round trip efficiency due to these currents, by calculation, simulation and physical measurements with batteries. Furthermore, batteries have been aged 1000 cycles of DC and DC+AC current to show no significant difference in aging
Multiscale mechanical constitutive equation of the passive properties of murine skeletal muscle: application to the influence of the Klf10 gene on the extra cellular matrix
No full textInternational audienc
Microrheometric study of damage and rupture of capsules in simple shear flow
No full textInternational audienc
Proposition d’outils à destination des fabricants pour faciliter l’intégration de l’AI Act avec le RDM
No full textRegulation (EU) 2024/1689 laying down harmonized rules on artificial intelligence (AI), also known as the “AI Act”, was published in the Official Journal of the European Union on July 12, 2024. This regulation has a horizontal scope and concerns all sectors using AI, including medical devices (MDs). Most MDs using AI will be classified as high-risk AI systems (HR-AIS) under the AI Act, which imposes regulatory requirements for the various stakeholders during the systems' lifecycle. Manufacturers of MDs with HR-AIS will have to comply with both Regulation (EU) 2017/745 relative to MDs (RDM) and the AI Act, representing a complex regulatory integration challenge. Our thesis aims to answer the following problem : How can we help manufacturers integrate AI Act and MDR requirements ? To address this issue, we first cross-read the two regulations, carried out a review of existing guides and conducted interviews with various stakeholders in the lifecycle of MDs with SIA-HR. We have created two tools for Quality Assurance and Regulatory Affairs (QA/RA) teams at manufacturers of MDs with SIA-HR, in order to understand the additional requirements of the AI Act compared to the MDR, and to assess the manufacturer's maturity regarding these AI-related requirements (interactive PDF and diagnostic tool).Le Règlement (UE) 2024/1689 établissant des règles harmonisées concernant l'intelligence artificielle (IA), aussi appelé « AI Act », a été publié au Journal officiel de l'Union européenne le 12 juillet 2024. Ce règlement a une portée horizontale et concerne tous les secteurs utilisant l'IA, notamment les dispositifs médicaux (DM). La plupart des DM utilisant de l'IA seront classés comme systèmes d'IA à haut risque (SIA-HR) selon l'AI Act, qui impose des exigences règlementaires pour les différentes parties prenantes lors du cycle de vie des systèmes. Les fabricants de DM avec SIA-HR devront se conformer à la fois au règlement (UE) 2017/745 relatif aux DM (RDM) et à l'AI Act, ce qui représente un défi d'intégration réglementaire. Ce mémoire vise à répondre à la problématique suivante : comment aider les fabricants à intégrer les exigences de l’AI Act avec celles du RDM ? Pour répondre à cette problématique, nous avons d’abord réalisé une lecture croisée des deux règlements, dressé un état des lieux des guides existants, et mené des entretiens avec diverses parties prenantes du cycle de vie des DM avec SIA-HR. Nous avons produit deux outils à destination des équipes Assurance Qualité et Affaires Réglementaires (AQ/AR) chez les fabricants de DM avec SIA-HR, afin de comprendre exigences supplémentaires de l’AI Act par rapport au RDM, et d’évaluer la maturité du fabricant concernant ces exigences relatives à l’IA (PDF interactif et outil diagnostic)
Intérêt de la simulation numérique pour la validation des dispositifs médicaux
No full textNumerical simulation is a technique that employs mathematical models to digitally represent systems or their components. These innovative techniques are highly advanced and widely used in various sectors, such as aeronautics and the automotive industry. For instance, IT in the automotive sector leverages digital simulation to predict mechanical issues in structures. This approach helps reduce time consumption while optimizing and developing the final product design. However, the medical device sector has been slower to adopt such methods in manufacturing. This is primarily due to the limited number of specialists in this field and the complexity of creating accurate models for medical devices. Digital simulation techniques can be used as a valuable tool in the medical device certification process. Thus, our memoir aims to address the question : What role could digital simulation play in the approval process for medical devices ? Modeling and digital simulation techniques are being actively developed in the USA, particularly for regulatory submissions concerning medical devices. This progress is according to the ASME V&V40 standard, which provides instructions on using digital simulation to validate computational models for testing medical device specific functions and material or components characteristics. In contrast, such validation and testing methods are currently absent in the European Union, as no harmonized standards concerning these techniques have been established. Therefore, our dissertation seeks to explore the role that digital simulation could play in the approval process of medical devices, specifically within the European Union .La simulation numérique est une technique qui utilise les modèles mathématiques pour la modélisation d’un système ou d’un organe. Elle est très développée dans plusieurs secteurs à savoir l’aéronautique, le secteur de l’automobile où nous avons TI Automotive qui utilise la simulation numérique pour anticiper les problèmes mécaniques de leur structure. Ceci aide à amoindrir le temps de développement tout en permettant l’optimisation du produit final. Le secteur des dispositifs médicaux est faiblement représenté dû au fait de l’unicité des individus ceci rendant difficile la création des modèles précis. Cette technique pourrait intervenir dans la plupart des étapes du processus d’homologation des dispositifs médicaux. C’est dans cette optique que notre mémoire vise à montrer la place que pourrait prendre la simulation numérique dans le processus d’homologation des dispositifs médicaux. Cette technique de modélisation et de simulation numérique est développée aux USA notamment sur les dispositifs médicaux à travers la norme ASME V&V40 qui régit l’utilisation de la simulation numérique. Le processus de validation selon cette norme sera abordé dans le cadre du projet ainsi que les différents outils Medical Devices Development Tools développés par les fabricants de dispositifs médicaux et la Food and Drug Administration permettant de valider les dispositifs médicaux. Etant donné que l’Union Européenne n’a pas harmonisé de normes concernant cette technique, l’approche de notre mémoire vise à proposer des suggestions pour son adoption
Bilan sur la mutualisation dans un service biomédical territorial
No full textThe territorial mutualization of biomedical services within Hospital Groups (GHT) in France aims to optimize the management of medical equipment by centralizing human, technical, and financial resources. This study explores the current practices, challenges, and recommandations to improve operational efficiency and ensure equitable healthcare access. The research draws on a comprehensive survey conducted among 1,056 biomedical professionals, highlighting key aspects such as maintenance organization, mutualization projects, and shared resources management. The results show significant benefits, including cost reduction, enhanced expertise sharing, and improved healthcare access in underserved areas. However, persistent barriers such as resistance to change, governance complexities, and resource disparities limit broader adoption. The study concludes with strategic recommandations : raising awareness, defining clear responsibilities, mapping and sharing competencies, standardizing processes, and implementing performance monitoring systems. Addressing these challenges can foster a more efficient and collaborative healthcare system, ensuring better service delivery and patient safety.La mutualisation territoriale des services biomédicaux au sein des Groupements Hospitaliers (GHT) en France vise à optimiser la gestion des équipements médicaux en centralisant les ressources humaines, techniques et financières. Cette étude explore les pratiques actuelles, les défis et les recommandations pour améliorer l'efficacité opérationnelle et assurer un accès équitable aux soins. L’étude s’appuie sur une enquête exhaustive menée auprès de 1003 professionnels biomédicaux, mettant en évidence des aspects clés tels que l’organisation de la maintenance, les projets de mutualisation et la gestion des ressources partagées. Les résultats montrent des avantages significatifs, notamment la réduction des coûts, le partage accru de l’expertise et l’amélioration de l’accès aux soins de santé dans les zones mal desservies. Cependant, des obstacles persistants tels que la résistance au changement, les complexités de la gouvernance et les disparités de ressources limitent une adoption plus large. L’étude conclut par des recommandations stratégiques : sensibiliser, définir des responsabilités claires, cartographier et partager les compétences, normaliser les processus et mettre en place des systèmes de suivi des performances. Relever ces défis peut favoriser un système de santé plus efficace et plus collaboratif, garantissant une meilleure prestation de services et une meilleure sécurité des patients