PEOPLE@HES-SO Annuaire et Répertoire des compétences

Smart Process Lab CONSTELLIUM - Contrat cadre de collaboration

Rôle: Collaborateur/trice

Requérant(e)s: VS - Institut Systèmes industriels

Financement: Innosuisse (delta charges sociales); Innosuisse; Constellium Valais SA; LYSR Sàrl; Eversys S.A

Description du projet : We propose a novel Interactive ML framework for advanced manufacturing. We combine incremental learning (few-shot learning and active learning) and human know-how (human-in-the-loop) to face data scarcity in quality control applications, from predictive manufacturability to predictive maintenance. Management Summary Please note, the executive summary will not be published in Aramis (The pain) Applications of Machine Learning (ML) for quality control and predictive maintenance applications are still facing great difficulties due to data scarcity; however, in most cases, practitioners or operators own relevant knowledge acquired over years of their daily work. (Unique Value Proposition) In this project, our innovation is to use interactive ML enforced by practitioner's interactions to train ML models on-the-fly. More specifically, at each of their interactions, the learned models are retrained to take advantage of new data points. Our approach is generic-purpose (process-agnostic) and will be validated on two different use cases, one from Constellium and the second one from Eversys. (Needed research) The research avenue is based on an interactive ML paradigm that combines: ' Incremental learning: merging few-shot learning (i.e., prototypical networks) with effective labeling (i.e., active learning via deep reinforcement learning). ' Human-in-the-loop: enabling operators to transfer their knowledge by interactive training of ML models. (Business impact) The interactive ML approach will be deployed by LYSR within the two use cases with the business goals of: - Reducing Constellium's scrap rate to 6% which represents an additional revenue of 3.4 MCHF in 2028. - Eversys' module of prediction failures reduction will generate an additional revenue of 4 MCHF over three years. Finally, by having demonstrated the generality of applying interactive ML, LYSR is poised to deploy this ML technology across current and future applications, generating 531 kCHF of additional revenue by 2027.

Equipe de recherche au sein de la HES-SO: Devènes Steve , Zahno Silvan , Baer Edouard , Borgeat Rémy , Carrino Francesco , Delgado Pamela , Petrovic Darko

Partenaires académiques: VS - Institut Systèmes industriels; IDIAP

Durée du projet: 01.03.2024 - 31.08.2026

Montant global du projet: 594'061 CHF

Statut: En cours

AT-Com

Rôle: Co-requérant(s)

Requérant(e)s: Remo Schnyder, Syrto AG, Steg

Financement: Innosuisse

Description du projet :

The availability of cost-effective electronic devices for autonomous driving and access to the ultra-fast 5G network bring new perspectives of digital integration of construction machines.
Today, operators sit in a construction machine and operate it directly. They are tied down both spatially and temporally. During unavoidable down-times, they cannot engage in any other activities. Therefore, most operators and machines are not used at full advantage. In addition, working in quarries or in ontaminated sites is inherently hazardous. Teleoperated working could allow one operator to serve many construction sites at once while increasing efficiency and safety.
Essential parts for effective teleoperated work are autonomous loading and unloading. Due to their complexity, these two tasks are still open research topics. This project aims to make autonomous loading/unloading possible for typical regional materials. The teleoperator should be able to monitor the wheel loader operations remotely and to control it by means of 'work
packages'. A work package is a sequence of elementary tasks then executed autonomously by the machine.
A 20-ton Komatsu W380 wheel loader will be converted and equipped to enable teleoperation and autonomous behaviors. The working area (e.g., structures, heaps) will be digitally recorded using laser scanners and cameras. The created 3D models will feed novel control algorithms. An advanced human-machine interface, which will allow the operators to easily define the work packages mentioned above, will be co-designed with experienced operators.

Equipe de recherche au sein de la HES-SO: Zahno Silvan , Maître Gilbert , Carrino Francesco , Pignat Marc , Petrovic Darko

Partenaires professionnels: Remo Schnyder, Syrto AG, Steg

Durée du projet: 01.05.2022 - 01.05.2024

Statut: En cours

Axe Transformation numérique 2022 - HES-SO Valais-Wallis

Rôle: Requérant(e) principal(e)

Financement: VS - Institut Systèmes industriels; VS - HEI; Constellium Valais SA; VS - Institut Systèmes industriels; VS - Institut Systèmes industriels

Description du projet : CONSTELLIUM VALAIS et HES-SO VALAIS souhaitent développer un partenariat fort pour faire croître en commun des compétences industrielles dans le domaine de la digitalisation et de la maitrise des processus industriels dans un contexte de développement de l'lndustrie 4.0.

Equipe de recherche au sein de la HES-SO: Clausen Michael , Devènes Steve , Bianchi Christophe , Maître Gilbert , Pompili Pierre , Zahno Silvan , Fracheboud Loïc , Mottiez Gilles , Amand Axel , Gaspoz Jonathan , Borgeat Rémy , Reynard Landry , Corre Jérôme , Carrino Francesco , Lenoir Cédric , Demion Arnaud , Petrovic Darko

Durée du projet: 01.09.2018 - 31.07.2027

Montant global du projet: 1'185'262 CHF

Statut: En cours

Terminés

P3 - Innosuisse 2021 projet "Press Pressure Prediction" - Prime soutien

Rôle: Co-requérant(s)

Financement: VS - Direction / Ra&D

Description du projet : Projet inter-disciplinaire pour l'axe Transformation numérique de la HES-SO Valais-Wallis Principalement pour la saisie d'heures pour la coordination et la participation aux séances de l'axe Transformation numérique. Pas de remboursement de frais de déplacement inter-site. Ventilation pour les BSM : W00 - centre de coût 90014 - n°projet (ne pas mettre de pilier 6 dans la ventilation car c'est sur un centre de coûts) Répartition du budget sur les projets socle des instituts au 31.12.2021

, , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , Equipe de recherche au sein de la HES-SO: Rey-Gillioz Anne-Christine Antic Dragana Gaudin Déborah Hannart Stéphanie Ehrenzeller Luisa Loup Jimmy Martinovic Jelena Bocchi Yann Verloo Henk Mudry Pierre-André Corthay François Gabioud Dominique Zeder David Zahno Silvan Carrard Sophie Schegg Roland Bürki Florian Doctor Marut Mantzouranis-Baudat Sophie Widmer Antoine Hilfiker Roger Manzo Gaetano Calbimonte Jean-Paul Darbellay Anne Pocklington Benjamin Hanik Nils Amand Axel Balet Sarah Dhrangadhariya Anjani Gay Cathy Allemand Anouk Imboden Alain Kuhn Alexandre Arulraj Bernard Maxvell Fellay Christophe Kullmann Axel Delaloye Matthieu Imboden Serge

Durée du projet: 01.01.2022 - 31.12.2022

Montant global du projet: 200'000 CHF

Statut: Terminé

Solution viable de métrologie à grande échelle, en production - P2

Rôle: Requérant(e) principal(e)

Financement: HES-SO Rectorat

Description du projet : P3 - Innosuisse 2021 projet "Press Pressure Prediction" - Prime soutien

Equipe de recherche au sein de la HES-SO: Devènes Steve , Zahno Silvan , Matter Fabien , Amand Axel , Gaspoz Jonathan

Partenaires académiques: VS - Institut Systèmes industriels

Durée du projet: 01.06.2021 - 30.06.2022

Montant global du projet: 15'000 CHF

Statut: Terminé

Impulse - USP - Unique Stability Plate

Rôle: Collaborateur/trice

Requérant(e)s: VS - Institut Systèmes industriels

Financement: HES-SO Rectorat

Description du projet : Metal extrusion is a versatile way to produce parts with a complex shape, in large quantities at a relatively low cost. The ability to produce custom shapes specific to customer's applications is also a major advantage. As a result, extruded aluminium profiles are widely used in construction, transportation, aerospace and other industries. However, their complexity, large size and relative low cost is also source of engineering problems: As applications become more and more advanced, customer require tighter and tighter dimensional tolerances, both in the cross section profile and along the length of the extruded part. Metrology on those extruded part is complicated. Firstly because of their size: A typical extrusion is delivered in length of 20 to 30 meter. Secondly, because of their complexity: Each profile is unique and most have internal geometry. Finally, because they are produced at relatively low cost, large investments for new technologies must be economically viable. As a result, no commercial solution is available to offer non-contact in production metrology for this type of manufacturing. Extruder carry on using age-old techniques for in production quality control: Samples are cut at the head and tail of each extruded length. The samples are then measured manually using callipers. Each measurement made is compared with the tolerance on a drawing. Each measurement is then entered in a database. This quality control method happens at the end of the production line, after the extrusion, quenching, cooling, stretching and cutting to size stages. When defect are detected at the end of the line, several tones of product have already gone through the extrusion press A Viable solution to precisely measure extruded profile in production would be invaluable to manufacturers. Firstly, it would help automate their quality control process. In the current Industry 4.0 paradigm, it would help collect more and more data. This data could then be used to improve the process and learn how to reach better and better tolerances. This would reduce costly rework and scraping of manufactured parts. It would then allow the manufacturer to offer high precision tolerance on specific geometry for a premium. Manufacturer could also start to offer digitally augmented products, delivered with their full digital quality control 3D model. This would allow customer to use that data for better planning of their process, for instance, by paring specific sections of profile that are due to be soldered together for optimum fit. And so on. As part of this project, we are planning to bring together existing metrology technologies and build a fully functional 'Proof of Concept' system for accurately measuring aluminium extrusion up to 30cm thick, 75cm wide and 30m long. This Viable Large Scale in Production Metrology POC will be tested as part of this project for a 1-month period in production at the Constellium Valais extrusion press. The outcome of this project will act as a springboard to gain further external funding, both in the form of 'mandate' from Constellium and in the form of 'innovation' projects funded by either of Innosuisse or European funds.

Equipe de recherche au sein de la HES-SO: Zahno Silvan , Matter Fabien , Amand Axel , Corre Jérôme

Partenaires académiques: VS - Institut Systèmes industriels

Durée du projet: 01.10.2020 - 01.03.2022

Montant global du projet: 110'750 CHF

Statut: Terminé

Holz 4.0

Rôle: Requérant(e) principal(e)

Financement: HESSO Regionalentwicklung

Description du projet :

Die Digitalisierung eröffnet Unternehmen neue Chancen und Möglichkeiten. Über digitale Vertriebskanäle können schneller neue Kunden akquiriert werden, Prozesse werden automatisiert und es entstehen neue Geschäftsmodelle, wie beispielsweise der ortsunabhängige Kundendienst und eine digitale vertikale Vernetzung.
Aktuell haben die Unternehmen folgende Hauptprobleme. Fehlende Kompetenzen und Ressourcen im Bereich Digitalisierung. Die Daten welche von den Unternehmen aktuell erhoben werden sind unbenutzt und nicht zur Verbesserung der Prozesse benutzt.
Durch die Zusammenarbeit mit den Brachenverband können Firmenübergreifende Probleme angegangen und gelöst werden. Die Innovation liegt in der erarbeitenden Methodik welche auf die vielen Unternhmen in der Wertschöpfungskette multiplizert werden kann. Durch die Vernetzung der Wertschöpfungskette kann ein enormes Potenzial erschlossen werden.
Das Ziel ist es das die Oberwalliser Unternehmen in der Holzindustrie Ihre Daten systematisch nutzten, um Prozesse zu verbessern und effizient abzuwickeln.

Equipe de recherche au sein de la HES-SO: Zahno Silvan , Devènes Steve , Maître Gilbert

Partenaires professionnels: Zanella Bernhard, Zanella AG, Turtmann

Durée du projet: 01.08.2021 - 01.12.2021

Statut: Terminé

DigitAlu

Rôle: Co-requérant(s)

Requérant(e)s: Lionel Thomas, Constellium Valais SA

Financement: Innosuisse

Description du projet :

Industry 4.0, with its digitalisation of manufacturing processes, transforms the shop floor environment into a workspace of human-machine cooperation. However, it must not stop there, everyone along the business chain must participate and should benefit. Current industry 4.0 efforts focus on digitalisation within a company:
We must start thinking beyond that.
The main link between all the actors along a business chain is the physical product itself. Thus, to ensure that all those actors embrace digitalisation, it seems logical to embed it in the products. The DigitALU innovation project targets the digital augmentation of aluminium plates.
Today Constellium Valais SA, as a Swiss pro

Equipe de recherche au sein de la HES-SO: Zahno Silvan , Corre Jérôme , Bocchi Yann , Leite Michael , Amand Axel , Mottiez Gilles , Petrovic Darko , Lenoir Cédric

Partenaires académiques: Benoît Golay, Institute Icare

Partenaires professionnels: Lionel Thomas, Constellium Valais SA

Durée du projet: 01.05.2020 - 28.02.2021

Statut: Terminé

Rôle: Co-requérant(s)

Requérant(e)s: Lionel Thomas, Constellium Valais SA

Financement: Innosuisse

Description du projet :

In order to achieve unique dimensional stability of aluminium plates, we intend to run machine learning algorithms on available large quantity of quality and process data. This project paves the way to invent a new product family where
traditional metallurgy and statistical analyses are limited.

Equipe de recherche au sein de la HES-SO: Zahno Silvan , Corre Jérôme , Devènes Steve , Petrovic Darko

Partenaires académiques: Fleuret François, IDIAP Research Institute

Partenaires professionnels: Lionel Thomas, Constellium Valais SA

Durée du projet: 01.05.2019 - 28.02.2021

Statut: Terminé

S19_SmartProcessLab

Automated surveying of surface water quality by a physical, chemical and biological sensor equipped anguilliform robot. - Nano-Tera.ch RTD

Rôle: Collaborateur/trice

Financement: VS - Institut Systèmes industriels

Description du projet : Projet destiné à l'acquisition de nouveaux projets et à la promotion des activités (publication, représentation,...) dans le domaine concerné.

Equipe de recherche au sein de la HES-SO: Bianchi Christophe , Zahno Silvan , Corre Jérôme

Durée du projet: 01.01.2019 - 31.12.2019

Montant global du projet: 30'000 CHF

Statut: Terminé

SCCER 4 WP 3.1 - Supply of Electricity - Contrat CTI 2013.0288

Rôle: Collaborateur/trice

Financement: VS - Institut Systèmes industriels; OFFT; Nano-Tera.CH

Description du projet : The major aim to develop and construct a demonstrator robot platform (the Envirobot) that samples and measures a set of relevant water quality parameters in surface water bodies by the incorporation of optical, physical, chemical and biological sensors. The Envirobot is invented to perform autonomous surveying or autonavigation. The second major aim of the project consists in particular in the development and incorporation of a range of biological sensors useful for both functioning modes of the Envirobot.

Equipe de recherche au sein de la HES-SO: Clausen Michael , Rudaz Patrice , Walpen Olivier , Vaccari Aldo , Amoos Serge , Gallay Steve , Ferrez Pierre , Mudry Pierre-André , Delacrétaz Romain , Geiser Martial , Maître Gilbert , Berthouzoz David , Zahno Silvan , Truffer Frédéric , Métrailler Christopher , Roduit Pierre , Petrovic Darko , Mabillard Eric , Pignat Marc

Durée du projet: 01.05.2013 - 31.10.2017

Montant global du projet: 745'147 CHF

Statut: Terminé

Centre pédagogique pour la prévention de séismes, réalisation à la route du Rawyl 47

Rôle: Collaborateur/trice

Financement: HES-SO Rectorat; VS - Institut Systèmes industriels; CTI; Loterie Romande; VS - Institut Systèmes industriels

Description du projet : SCCER 4 WP 3.1 - Supply of Electricity - Contrat CTI 2013.0288

Equipe de recherche au sein de la HES-SO: Clausen Michael , Praplan Charles , Rudaz Patrice , Steiner Amandus , Walpen Olivier , Vaccari Aldo , Gallay Steve , Cachelin Christian Pierre , Moerschell Joseph , Nicollier Grégoire , Berthouzoz David , Zahno Silvan , Arcudi Carmine , Mabillard Eric

Partenaires professionnels: Fonds tiers 2015

Durée du projet: 01.11.2013 - 31.12.2016

Montant global du projet: 777'250 CHF

Statut: Terminé

Centre pédagogique pour la prévention de séismes

Rôle: Collaborateur/trice

Requérant(e)s: VS - Institut Systèmes industriels, Moerschell Joseph, VS - Institut Systèmes industriels

Financement: Loterie Romande; VS - Direction / Ra&D; Etat du Valais, DECS; Etat du Valais - DFS

Description du projet : Ce projet englobe la contribution HES-SO à la mise sur pied d'un Centre pédagogique pour la prévention de séismes au Valais, avec une première réalisation à la route du Rawyl 47. Il s'agit en particulier de réaliser un simulateur de séismes et différentes maquettes qui expliquent la genèse et les effets de séismes.

Equipe de recherche au sein de la HES-SO: Steiner Amandus , Truffer Christophe , Walpen Olivier , Vaccari Aldo , Lorenzetti Franco , Roduit Alain , Amoos Serge , Cachelin Christian Pierre , Sauron Anne , Maître Gilbert , Marcuard Jean-Daniel , Moerschell Joseph , Nicollier Grégoire , Berthouzoz David , Zahno Silvan , Rapillard Laurent , Grand Pascal , Carroz Emmanuel , Knupfer Pierre-André , Mabillard Eric

Partenaires académiques: VS - Institut Systèmes industriels; Moerschell Joseph, VS - Institut Systèmes industriels

Durée du projet: 15.02.2014 - 31.03.2016

Montant global du projet: 776'348 CHF

Statut: Terminé

Plateforme générique pour systèmes nomades

Rôle: Collaborateur/trice

Requérant(e)s: VS - Institut Systèmes industriels, Moerschell Joseph, VS - Institut Systèmes industriels

Financement: Etat du Valais, DECS

Description du projet : Ce projet englobe la contribution HES-SO à la mise sur pied d'un Centre pédagogique pour la prévention de séismes au Valais. Il s'agit en particulier de réaliser un simulateur de séismes et différentes maquettes qui expliquent la genèse et les effets de séismes.

Equipe de recherche au sein de la HES-SO: Steiner Amandus , Truffer Christophe , Walpen Olivier , Vaccari Aldo , Sauron Anne , Marcuard Jean-Daniel , Moerschell Joseph , Berthouzoz David , Zahno Silvan , Rapillard Laurent , Grand Pascal , Mabillard Eric

Partenaires académiques: VS - Institut Systèmes industriels; Moerschell Joseph, VS - Institut Systèmes industriels

Durée du projet: 14.02.2013 - 14.02.2014

Montant global du projet: 200'000 CHF

Statut: Terminé

Fluxmètre pour la tête du nerf optique avec acquisition

Rôle: Collaborateur/trice

Financement: HES-SO Rectorat; EIG - INIT; VS - Institut Systèmes industriels; Systèmes embarqués; SIM-TIC; SIM; MNT; FR - EIA - Général école Ra&D; SIM

Description du projet : Ce projet a pour objectif le développement d'une plateforme générique de système nomade, réunissant les différentes technologies dans lesquelles sont actifs les membres du groupe thématique « systèmes nomades » du RCSO-ISYS. La plateforme proposée permettra la réalisation de réseaux de senseurs à très basse consommation. Les applications envisagées sont en particulier dans le contrôle environnemental, en relation avec la problématique du réchauffement climatique. On peut citer : - détection de parasites du bois - prévention d'avalanches, de chutes de pierres ou même de glissement de terrains Le système étudie les problématiques de la transformation et de l'accumulation d'énergie, de la gestion de la consommation des circuits via leur mise en mode oisif (d'où l'acronyme trypano), du traitement de données avec des circuits programmables à faible consommation, du microstockage de l'information, de la transmission sans fil des données ainsi que de l'intégration du système complet. Un aspect important est la possibilité d'explorer les compromis existant entre la consommation minimale des n'uds senseurs et leur fiabilité, et la fiabilité de l'information globale au niveau du système complet.

Equipe de recherche au sein de la HES-SO: Pirrami Lorenzo , Sartoretti Pascal , Walpen Olivier , Gallay Steve , Sauser Florian , Menth Gilbert , Hochet Bertrand , Corthay François , Brunner Nicolas , Maître Gilbert , Pompili Pierre , Huguenin Gérald , Zahno Silvan , Fragnière Eric , Demierre Michel , Crausaz Philippe , Beuchat René

Durée du projet: 11.02.2009 - 27.11.2009

Montant global du projet: 137'950 CHF

Statut: Terminé

Rôle: Collaborateur/trice

Requérant(e)s: HES-SO Valais-Wallis, Geiser Martial, HES-SO Valais-Wallis

Financement: CHU Grenoble

Description du projet : Suite au développement effectué dans le projet AuMaFlo, le Service d'Ophtalmologie du Centre Hospitalier Universitaire de Grenoble aimerait faire l'acquisition du système élaboré, ainsi que d'un système plus simple.

Equipe de recherche au sein de la HES-SO: Martin Simon , Cretton Fabian , Clausen Michael , Barmaz Xavier , Rodier Madeleine , Girard Hervé , Glenz-Mounir Chantal , Pignat Christophe , Pott Julien , Praplan Charles , Rudaz Patrice , Rey-Siggen Janine , Rey-Gillioz Anne-Christine , Salamin Anne-Dominique , Sartoretti Pascal , Steiner Amandus , Sterren Martin , Cretton Viviane , Sterren Thomas , Walker Anne , Dayer Aline , Truffer Christophe , Udrisard Thierry , Walpen Olivier , Hugo Alexandra , Steiner Rico , Dussex Catherine , Vaccari Aldo , Lorenzetti Franco , Crettol Blaise , Eyer Kurt , Prim Denis , Kniel-Fux Lucie , Roduit Alain , Latrille Sophie , Broccard Manu , Charbonnet Gian , Frei Isabelle , Meyer Yasmine , Kalman Franka , Kotanadurage Asitha , Chappuis Aline , Genoud Dominique , Evéquoz Florian , Ellert Christoph , Projer Diane , Münch-Alligné Cécile , Nyanguile Origène , Clivaz Jacinthe , Volluz-Gasdia Martina , Dayer Etienne , Berthod Marc-Antoine , Albrecht Natacha , Depeursinge Adrien , Palazzo Clothilde , Valette Emilie , Alber Jean-Luc , Amoos Serge , Fournier Anne , Moulin Fabien , Willen Stefan , Wilk Rolf , Zufferey Yves-Benoît , De Chastonay Jean-Victor , Imboden Monique , Gafundi Maria Luigia , Jacquod Isabelle , Gallay Steve , Micaux Fabrice , Roduit Romain , Pernatozzi Franco , Praz Ariane , Mathieu Claudia , Rapillard Gérard , Reynard Carmen , Truffer Marie-Jeanne , Vicini Marie-Laure , Villettaz Karine , Zufferey Monique , Vuignier Sonia , Vadala Francesco , Girod Lehmann Béatrice , Crettol Olivier , Duan Hui , Maier Niklas , Bagnoud Laurent , Morard Ken , Hennebert Jean , Xia Leyun , Duc Alain , Andlauer Wilfried , Anthamatten Hermann , Barbey Valérie , Bianchi Christophe , Morand Frédéric , Bocchi Yann , Curdy-Locher Nadia , Bréchet Mottier Barbara , Moghaddam Fariba , Carreno-Morelli Efrain , Roux-Elsig Dominique , Mudry Pierre-André , Corthay François , Crelier Simon , Dubas Michel , Epiney Jacques , Blatter-Robyr Véronique , Fischer Fabian , Frey Urban , Gabioud Dominique , Geiser Martial , Glassey Balet Nicole , Guex Pierre-Yves , Le Calvé Anne , Maier Werner , Maître Gilbert , Marcuard Jean-Daniel , Meyer Richard , Moerschell Joseph , Montani Bruno , Morand Gilbert , Nicolay Laurence , Nicollier Grégoire , Perruchoud Antoine , Piantini Umberto , Pompili Pierre , Rey Jean-Pierre , Rieder Medard , Roten Hans-Peter , Ruppen Paul , Sailer Limani Karin , Sarrasin Natalie , Savioz Christian , Schmid Sergio , Sciboz Laurent , Stampfli Michèle , Steiner Thomas , Studer Leo , Schumacher Michael Ignaz , Zumstein Monica , Schnyder Marc , Hofstädter Martine , Zahno Silvan , Beyrer Michael , Scaglione Miriam , Schegg Roland , Gaillard Thierry , Bastian Christèle , Fux Rainer , Niederhauser Anne , Arcudi Carmine , Doctor Marut , Blatter Didier , Bocquel Dimitri , Cid Leal Francisco Jose , Costa Christian , Darioly Madeleine , Debons Christophe , Delaloye Thierry , Délétroz Nicolas , Evéquoz Blaise , Emery Claude-Alain , Rapillard Laurent , Müller Henning , Germanier Alain , Allet Carmen , Dufresne Rémy , Dervey Sébastien , Roggo Dominique , Savioz Lucie , Grand Pascal , Margot-Cattin Pierre , Truffer Frédéric , Peter Mélanie , Pillet Alain , Mathieu Nicolas , Elsig-Heynen Simone , Furrer Rey Chantal , Baumgartner Marie-Cécile , Jacquier-Delaloye Anne , Jossen Renata , Alberti Sabrina , Bender Gabriel , Bigoni-Barbier Catherine , Boulé Christophe , Broccard Pierre-Marie , Carlen Fernando , Coquoz Emmanuelle , Cortat Pierre-Henri , Debons Marie-Jeanne , Eberhart Jean-Jacques , Fumeaux-Evéquoz Nicole , Favre Bourban Eliane , Furrer-Andenmatten Marcelline , Furrer-Treyer Christa , Gay Marcelle , Henrion-Dourcy Anne , Hilfiker Roger , Lambiel Ville Karine , Lauber Christine , Lehmann-Wellig Brigitte , Lorenz Cottagnoud Susanne , Maruzza Maurizio , Mathier Favre Irma , Maye Stéphane , Mittaz Hager Anne-Gabrielle , Nowak Frédérique , Pinho Jorge , Pirlet Marielle , Pitarelli Emilio , Plaschy Martin , Pott Murielle , Sottaz Pralong Jasmina , Rey Jean-Charles , Roduit Jean-Marc , Roh Pierre-Yves , Schoepf Chris , Schwery Hilar , Seppey Bovier Christine , Solioz Emmanuel , Sougner-Remillieux Christine , Antonin-Tattini Véronique , Thönnissen Chase Evelyne , Van Overbergh Patrick , Barmaz Marie-Nicole , Carron Jean , Brantschen Philippe , Adkins Léonard , Blatter Martin , Briguet Alain , Sattelmayer Martin , Waldis Barbara , Nowak Maxime , Moroni Isabelle , Pistoletti Blanchet Gordana , Foley Rose-Anna , Schnyder Bruno , Moulin Matthieu , Wegmüller Sarah , Denis Marika , Salzmann Stefan , Rey Vincent , Mastelic Joëlle , Salamat Aziz , Emery Caryl , Allemand Anouk , Pitteloud Pascal , Amrein Daniel , Crettenand Bernard , Segura Jean-Manuel , Rion Didier , Dini Sarah , Stumm Nathalie , Jurisch Praz Sarah , Rodriguez Arbaizar Mikel , Dorsaz Laurent , Bournissen Chantal , Lenggenhager Fabian , Gsponer Nadège , Gessler François , Ballestraz Sarjan , Luciani Christine , Voll Peter , Christen Ursula , Gaspoz Véronique , Michellod Eric , Pannatier Pascal , Perruchoud-Massy Marie-Françoise , Pfeifer Marc Emil , Russo David , Rey Bernard , Roduit Pierre , Tschopp Alexis , Zuber Benoît , Cotting Alexandre , Carroz Emmanuel , Bonvin Michel , Dolt Stéphane , Kuchler Manuela , Jacquemettaz Pascal , Correia De Oliveira Dominique , Schär Sabine , Bourban Béatrice , Knupfer Pierre-André , Crettenand Alexia , Métrailler Conny , Imboden Serge , Eggel Ivan , Lehner Bruno , Petrovic Darko , Marcon Nadia , Mabillard Eric , Matos-Wasem Rafael , Fux Michael , Mayor Stéphane , Gutiérrez López Alauda , Todeschini Beverley , Baudin Martine , Seppey Pierre-André , Zenklusen Fabrizio , Pignat Marc , Languetin Etienne

Partenaires académiques: HES-SO Valais-Wallis; Geiser Martial, HES-SO Valais-Wallis

Durée du projet: 01.08.2005 - 10.03.2008

Montant global du projet: 50'000 CHF

Statut: Terminé

2025

Frequency-dependent deconvolution for enhanced THz-TDS scans :

Article scientifique ArODES

accounting for beam width variations in time traces

Arnaud Demion, Linus Leo Stöckli, Nicolas Thomas, Silvan Zahno

IEEE Transactions on Terahertz Science and Technology, 2025, To be published., 1-9

Résumé:

Terahertz Time-Domain Spectroscopy (THz-TDS) is a powerful tool for non-destructive testing, but the quality of scans is limited by beam spreading effects, which vary with frequency. We propose a deconvolution method that accounts for the changing beam width as a function of frequency. By incorporating frequency-specific Gaussian beam profiles, determined through the knife-edge technique, the proposed method enhances the spatial resolution of THz-TDS scans while keeping the ability to analyze the phase of the signals. Our approach significantly improves contrast and sharpness in both reflectance and transmission modes. This technique offers a more accurate and robust enhancement of THz-TDS imaging.

2023

Dynamic project planning with digital twin

Article scientifique ArODES

Silvan Zahno, Jérôme Corre, Darko Petrovic, Gilles Mottiez, Loïc Fracheboud, Axel Amand, Steve Devènes, Gilbert Maître, Francesco Carrino

Frontiers in Manufacturing Technology, 3

Résumé:

The digital twin (DT) concept plays a crucial role in Industry 4.0 and the digitalization of manufacturing processes. A DT is a virtual representation of a physical object, system, or process, designed to accurately reflect its real-world counterpart. In manufacturing, existing process data are often incomplete and do not qualify as a DT. However, with the help of specialized communication frameworks and cheaper, easier-to-use sensors, it is possible to integrate the existing manufacturing execution system (MES) and enterprise resource planning (ERP) data with the missing data gathered from the shop floor to create a comprehensive DT. In this paper, we present a digital shop floor decision support system (DSS) for non-linear aluminum manufacturing production. The system is split into five main components: digitization of shop floor orders; merging and sorting of MES, ERP, and shop floor data; custom and genetic optimization algorithms for the aging furnace production step; layout construction mechanism for optimal placement and stacking of orders in the furnace; and a user-friendly graphical user interface (GUI). The system’s performance was evaluated through three tests. The first test measured the efficiency of digitization, the second aimed to quantify time saved in finding packets in the hall, and the last test measured the impact of the optimizer on furnace productivity. The results revealed a 23.5% improvement in furnace capacity, but limitations were identified due to usability and human intervention.

Dynamic project planning with digital twin

Article scientifique

Zahno Silvan, Carrino Francesco, Petrovic Darko, Devènes Steve, Gilles Mottiez, Loïc Fracheboud, Maître Gilbert, Jérôme Corre, Amand Axel

Fontiers in Manufacturing, 2023

Résumé:

2021

Edge Processing Unit

Article scientifique

The Efficient Way to Edge Computing for Industrie 4.0

github, 2021

Résumé:

Edge computing has proven to be very valuable and is often needed in addition to cloud com- puting. Especially in areas such as image and video processing, where large amount of data is generated that cannot be transferred in their entirety to the cloud for processing, or in areas where security and real-time behavior are essential.

In the just mentioned cases, it is not practical to transfer the data to the cloud for processing and return the result to the device. Edge computers can be very powerful, but also high maintenance and power hungry, depending on the algorithm and hardware. In this work, a Proof of concept (POC) is presented to combine low-cost and low-power Central Processing Unit (CPU) with Field Programmable Gate Array (FPGA) to run an algorithm as optimized as possible. The simplicity of the CPU programming and the modular and parallel nature of the FPGA creates very powerful hardware, that can handle many different problems.

A prototype system was presented in an industrial use case, where an automated industrial ma- chine is allowed to run, only when no human is nearby. We present an Edge Processing Unit (EPU) that monitors a selected region using a camera. The images are preprocessed in CPU and an object detection is performed in real time using a tiny You only look once (YOLO) v3 Machine Learning (ML) implemented in FPGA. A Convolutional Neural Network (CNN) hardware accelerator, performs the person detection. A MQTT broker, which also runs on the edge device, sends the results to the cloud and local Programmable Logic Controller (PLC). The entire sys- tem runs on a Xilinx PYNQ-Z1 board with a ZYNQ XC7Z020 System on Chip (SoC) that houses a dual core Arm A9 CPU and programmable Artix-7 logic. It achieves a higher frame rate and consumed less power than a comparable non-hardware accelerated edge device. The amount of data transferred was drastically reduced, compared to a cloud application.

2012

Microwave barrel reactor use in trimethylolpropane oleate synthesis by Candida antarctica lipase in a biphasic non-solvent process

Article scientifique ArODES

Manuel Happe, Pascal Grand, Sébastien Farquet, Sandrine Aeby, Jean-Claude Héritier, François Corthay, Eric Mabillard, Roger Marti, Ennio Vanoli, Alain-François Grogg, Samuel Nussbaum, Alain Roduit, François Tièche, Sam Salem, Carole Constantin, Esther Schmitt, Silvan Zahno, Christoph Ellert, Ahmed Habib, Julien Wyss, Fabian Fischer

Green Chemistry, 2012, no. 8

Résumé:

A novel microwave barrel reactor (MBR) was constructed and used in lipase catalyzed biolubricant synthesis. The MBR is thought as a versatile process tool for biotransformation and green chemistry that overcomes current size limitations in microwave reactors. A lipase mediated biotransformation in the MBR was compared to a state of the art jacketed reactor with external heat exchanger. Oleic acid and trimethylolpropane converted quantitatively (96%) into biolubricants using microwave induction. The heat dissipation in the MBR was analyzed by thermal imaging and inside thermometry. Conversion rates, rate constants and pseudo reaction orders were in line with conventional processing and no microwave effect was detected. The MBR is a versatile new reactor for non solvent, minimal and common solvent processing in the microwave field. While the subject of investigations was biolubricant synthesis in the MBR, the technology described is of wider potential interest in the field of biomass processing and sustainable chemical manufacture.

Microwave barrel reactor use in trimethylolpropane oleate synthesis by Candida antarctica lipase in a biphasic non-solvent processa

Article scientifique

Marti Roger, Aeby Sandrine, Fischer Fabian, Manuel Happe, Pascal Grand, Sébastian Farquet, Corthay François, Jean-Claude Hértier, Mabillard Eric, Alain-Francois Grogg, Samuel Nussbaum, Roduit Alain, Tièche François, Sam Salem, Carole Constantin, Esther Schmitt, Zahno Silvan, Ellert Christoph, Ahmed Habib, Julien Wyss

Green Chem., 2012 , vol. 14, pp. 2337-2345

Résumé:

2024

PPP - Press Pressure Prediction

Conférence

What Pasta has in common with Aluminum Extrusion

9. F&E-Konferenz zu Industrie 4.0, 24.01.2024 - 24.01.2024, ETH Zürich

Lien vers la conférence

2023

Digital Twin

Conférence

Product Planning

8. F&E Konerenz, 24.01.2023 - 24.01.2023, HSLU Luzern

Lien vers la conférence

2022

Edge Processing Unit

Conférence

The Way to efficient Edge Computing for Industrie 4.0