Accelerated deep learning approaches for vibro-acoustic digital twins VAMOR DC2

at  KU Leuven

Accelerated deep learning approaches for vibro-acoustic digital twins VAMOR DC2
(ref. BAP-2024-208)
Laatst aangepast: 08/04/24
The research is hosted by the Mecha(tro)nic System Dynamics division (LMSD), which currently counts >100 researchers and is part of the department of mechanical engineering of KU Leuven. The research group has a long track record of combining excellent fundamental academic research with industrially relevant applications, leading to dissemination in both highly ranked academic journals as well as on industrial fora. More information on the research group can be found on the website: https://www.mech.kuleuven.be/en/research/mod/about and our linkedIn page: https://www.linkedin.com/showcase/lmsd-kuleuven/. The PhD will be co-supervised by the Conservatoire National des Arts et Metiers Laboratory of mechanics of structures and coupled systems (LMSSC). The laboratory, located in the central area of Paris, currently counts >40 researchers mostly working on the development and validation of robust predictive models of dynamic coupled systems using adaptive treatments. The results of this research are mainly applied to the academic world, research centers and R&D department of high technology industries..The team has various industrial collaborations, mainly with aeronautics and naval industry. More information can be found on the website: https://lmssc.cnam.fr/en/content/structural-mechanics-and-coupled-systems-laboratory
Website van de eenheid
Project
This doctoral project is part of a larger, multidisciplinary and international project VAMOR: “Vibro-Acoustic Model Order Reduction” (GA 101119903) funded under the Marie-Sklodowska-Curie Actions Doctoral Networks within the Horizon Europe Programme of the European Commission.
VAMOR contributes to a more sustainable and quieter future for Europe. Noise pollution has arisen as one of the key factors towards the degradation of the quality of life in European societies. In that context, efficient physics-based sound modelling is a key enabler towards not only optimized and sustainable acoustic profiles through efficient design procedures, but also affordable so-called digital twins that monitor product performance in real time. To this end, the overarching goal of VAMOR is to provide high level scientific and transferable skills training on a new generation of efficient vibro-acoustic modelling techniques, so-called model order reduction (MOR) strategies, to a group of high achieving, competent doctoral candidates to promote a quieter and more sustainable environment. VAMOR brings together a remarkable consortium, which combines research leading academic institutions - KU Leuven, Technische Universitaet Munchen (TUM), Technical University of Denmark (DTU), Kungliga Tekniska Hoegskolan (KTH), Universite du Mans, Conservatoire National des Arts et Metiers (CNAM) - with a constantly innovating, wide variety of industrial partners working on software, material, testing, design and sound enhancement (Siemens Industry Software NV, Müller BBM, Trèves, Phononic Vibes, Saint-Gobain Ecophon, Tyréns, Purifi ApS).
Doctoral Candidate 2 (DC2) within VAMOR will develop novel deep learning approaches to obtain a digital twin that is trained using both physics based metamodels/insights and measurements on the physical asset. To accelerate the training of the machine learning architecture, reduced order models of vibro-acoustic systems are expected to play a key role in data generation, due to their rapid evaluation as compared to their full order model counterpart. Besides the usage of physics based reduced order models, also the addition of physical constraints in the machine learning architecture through physics informed neural networks will be considered to speed up the training. Finally, the addition of measurement data in the reduced physical models will be explored with the aim to obtain a more representative parameter dependent digital twin. The approach will be validated by investigating the sound radiated from a car tyre before and during operation. The DC will explore the combination of conventional MOR strategies with deep learning, physics-informed neural networks and inclusion of data in surrogate models.
Main Supervision at KUL: Prof. Konstantinos Gryllias and Prof. Elke Deckers.
Co-Supervision at CNAM: Prof. Lucie Rouleau.
Profiel

Please refer the Job description for details