Research Publications

Permanent URI for this collection

Browse

Recent Submissions

Now showing 1 - 5 of 1049
  • Item
    Chat GPT and other AI´s – a new language for the industry
    (2023-11-22) Schmid, Philipp A. E.
    Since the free launch of the ChatGPT chatbot in 2022, Artificial Intelligence (AI) has undeniably become a hot topic. More than 50% of ChatGPT users reveal that they utilize AI in their personal lives to streamline daily tasks. With such widespread consumer interest in AI, it prompts a natural inquiry into the current role of AI in the industrial landscape. Philipp Schmid, Head of Industry 4.0 and Machine Learning at CSEM, an internationally recognized Swiss technology innovation center, provides insights in the following interview. Schmid addresses, among other topics, the general significance of AI in the industry, the untapped potential of predictive maintenance and why the industry would do well to rethink its handling and attitude toward data.
  • Item
    LASER POWDER BED FUSION OF HIGH STIFFNESS TITANIUM BASED METAL MATRIX COMPOSITE FOR SPACE APPLICATIONS
    (2023-04-05) Bernard, Gaëtan; Pejchal, Vaclav; Larsson, Joel; Sereda, Olha; Logé, Roland
    This study presents a production method of high stiffness Ti-TiC Metal Matrix Composite by Laser Powder Fusion. Crack and porosity free samples were printed with an improved Young’s modulus of up to 163 GPa showing a 39% increase compared to commercially pure Titanium produced by the same method. Microstructure shows both un-melted TiC particles and sub-stoichiometric TiC dendrites resulting from partial dissolution of TiC particles. The observed low C/Ti ratio leads to an effective reinforcement content more than twice the nominal one. Effects of post-treatments on microstructure and mechanical properties are presented.
  • Item
    Die Technologiewelle rollt – Hochleistungsproduktion in der Schweiz dank Daten und KI
    (2023-11-23) Schmid, Philipp A. E.; Steinecker, Alexander
    AI setzt sich immer stärker im Alltag durch: Übersetzungsprogramme, lustige Fotofilter auf dem Handy oder Fahrassistenzen im Auto. Nun steht der grosse Durchbruch in der produzierenden Industrie an. Perfekte Qualität und immer höhere Effizienz treiben Schweizer Firmen. Viele Betriebe haben begonnen, Daten von Maschinen und Prüfgeräten zu speichern, häufig lokal in unterschiedlichen Formaten und Tabellen auf verschiedenen Geräten. Nun ermöglichen fallende Preise und die einfache Handhabung von IoT-Gateways die Anbindung an die Cloud. Die Daten verweilen nicht länger in verteilten Silos, sondern lassen sich entlang ganzer Wertschöpfungsketten und auch standortübergreifend aggregieren. Neue Konzepte wie Predictive Quality bieten ganzheitliche Lösungsansätze um aus dem Datenberg einen echten Wettbewerbsvorteil zu ziehen. Doch obwohl künstliche neuronale Netzwerke bereits seit einem Jahrzehnt gross in aller Munde sind, gelang es der Industrie bisher kaum davon zu profitieren. Philipp Schmid analysiert in seinem Referat, wo die Technologie heute steht, benennt die Stolpersteine und zeigt anhand konkreter Beispiele das enorme Potential von Data Science und KI in der Industrie auf. Im bevorstehenden INNOtalk #6 möchte die SwissICT Fachgruppe Innovation relevante Aspekte dieses Themas in einer interaktiven Art und Weise diskutieren und freut sich auf die Teilnahme des Experten Philipp Schmid sowie allen anwesenden Gästen. Eine angenehme Atmosphäre mit viel Gelegenheit für wertvolles Networking rundet diesen Anlass ab.
  • Item
    Method to Embed Behavioral Battery Model in Predictive Energy Management Systems
    (2023-10) Sutter, Axel; Gorecki, Tomasz T.; Bhoir, Shubham S.
    In the context of battery use in an energy system, accurate degradation models are crucial to improve economic efficiency, because these models help to ensure that the cost of degradation is lower than the benefits of using the battery. This paper proposes a novel method to integrate a complex empirical degradation model of a lithium-ion battery into the optimization of an Energy Management System (EMS). The model uses relaxed nonlinear equations of a multi-factor degradation model that are based on empirical test data. This approach allows the degradation model to be included in an optimization framework and to monitor the State-of-Health (SoH) of the battery. The paper presents a case study where the proposed degra dation model is applied in an arbitrage scenario to evaluate the benefits of the model in the optimization process. The results show that even a simple degradation model yields a positive benefit of 6,250 EUR/MWh/year, while a more complex model can achieve up to 10,000 EUR/MWh/year. These findings demonstrate the efficacy of the proposed model in achieving efficient battery use in an energy system.
  • Item
    Miniaturized Fluorescence Biosensing Technology for Multiplexed Allergy Screening
    (2022-10-20) Demuru, Silvia; Chai-Gao, Hui; Shynkarenko, Yevhen; Hermann, Nicola; Boia, Patricia-Daiana; Cristofolini, Peter; Generelli, Silvia; Cattaneo, Stefano; Burr, Loïc
    The incidence of immune-mediated diseases such as asthma and allergies is steadily increasing.[1] However, little is known yet about how genetics, environmental factors, and epigenetics can influence the onset and progression of these diseases. Within the Human Exposomic Determinants of Immune-Mediated Diseases (HEDIMED) project, we are developing a portable multi-array system for immune-signature testing. Such a platform could help to quickly screen various biomarkers, such as antibodies, related to immune-mediated diseases.[2] In the present project, a multiarray for antibody detection has been developed for allergy profiling. Based on microfluidic chips of the size of a standardized microscopy slide, the multiarray is embedded in a microfluidic channel with microstructures functionalized with allergen extract or recombinant proteins. The use of a microfluidic chip enables the multiplexed screening of up to 88 different allergens from the patient (blood serum) with low sample volumes (80-150 uL). An automated sample-on-chip processing system has been developed to ensure the reproducible detection of allergy-specific IgEs using fluorescence-labelled antibodies. In addition, a compact, low-cost, and fast fluorescence reader has also been developed for simple measurements of the fluorescence signals and automated quantification of the allergic response. We present our progress regarding the multiarray assay, the optical reader, and preliminary results on fluorescence signal detection in human serum. By testing the microfluidic system with a focus on pregnant women and newborn samples, this technology could help to find new correlations between multiple environmental factors and the onset and progress of immuno-mediated diseases at early ages, improving the prevention of such diseases in the future. References [1] F. Huang, H. Jia, Y. Zou, Y. Yao, and Z. Deng, “Exosomes: an important messenger in the asthma inflammatory microenvironment,” J. Int. Med. Res., vol. 48, no. 2, pp. 1–11, 2020, doi: 10.1177/0300060520903220. [2] E. Sechi and E. P. Flanagan, “Antibody-Mediated Autoimmune Diseases of the CNS: Challenges and Approaches to Diagnosis and Management,” Front. Neurol., vol. 12, no. 673339, pp. 1–18, 2021, doi: 10.3389/fneur.2021.673339.