KIT | KIT-Bibliothek | Impressum | Datenschutz

Physics-informed neural networks for data-free surrogate modelling and engineering optimization – An example from composite manufacturing

Würth, Tobias ORCID iD icon 1; Krauß, Constantin 1; Zimmerling, Clemens ORCID iD icon 1; Kärger, Luise ORCID iD icon 1
1 Institut für Fahrzeugsystemtechnik (FAST), Karlsruher Institut für Technologie (KIT)

Abstract (englisch):

Engineering components require an optimization of design and manufacturing parameters to achieve maximum performance – usually involving numerous physics-based simulations. Optimizing these parameters is a resource-intensive endeavor, though, especially in high-dimensional scenarios or for complex materials like fiber reinforced plastics. Surrogate models are able to reduce the computational effort, however, data generation still proves to be resource-intensive. Additionally, their data-driven nature may lead to physically implausible results in limit cases. As a remedy, physics-informed neural networks (PINNs) include known physics into the training for enhanced surrogate reliability. This allows to cast a physically consistent, data- and mesh-free manufacturing surrogate for variable process conditions and material parameters. The paper demonstrates how PINNs can be embedded in a design-framework to enhance process understanding, to devise engineering-interpretable processing windows and to support time-efficient process optimization at the example of a thermochemical manufacturing process with fiber-reinforced composite materials. ... mehr


Verlagsausgabe §
DOI: 10.5445/IR/1000159290
Veröffentlicht am 09.06.2023
Originalveröffentlichung
DOI: 10.1016/j.matdes.2023.112034
Scopus
Zitationen: 6
Web of Science
Zitationen: 6
Dimensions
Zitationen: 11
Cover der Publikation
Zugehörige Institution(en) am KIT Institut für Fahrzeugsystemtechnik (FAST)
Publikationstyp Zeitschriftenaufsatz
Publikationsmonat/-jahr 07.2023
Sprache Englisch
Identifikator ISSN: 0264-1275
KITopen-ID: 1000159290
Erschienen in Materials & Design
Verlag Elsevier
Band 231
Seiten Art.-Nr.: 112034
Vorab online veröffentlicht am 26.05.2023
Schlagwörter Machine learning, Physics-based modeling, Mesh-free surrogate modeling, Data-free surrogates, Composite process optimization, Physics-Informed Neural Networks, Artificial intelligence
Nachgewiesen in Dimensions
Web of Science
Scopus
Relationen in KITopen
Globale Ziele für nachhaltige Entwicklung Ziel 9 – Industrie, Innovation und Infrastruktur
KIT – Die Forschungsuniversität in der Helmholtz-Gemeinschaft
KITopen Landing Page