I've been studying physics and especially acoustics for the past five years at the Université du Maine in Le Mans, France. The cursus was mainly about physical acoustics, signal analysis, vibroacoustics, mechanics & fluid dynamics. The first two years were held in common with a bachelor degree in computer sciences (C programming language, web development, etc.).
Modelling of thin and imperfect interfaces: tools and preliminary study
On June 13th, 2018, I defended my licentiate thesis before Pr. Emeline Sadoulet-Reboul (FEMTO ST, Université de France-Comté, Besançon, France):
For quite some time, the strive for more efficient acoustic absorbers keeps increasing, driven by a number of psycho-physiological studies on health related dangers of noise exposure. As the global wealth increases and with it the global expectation of quieter living and working environments, manifested in both politics and research, an important market for sound absorbing and noise control systems develops in all industrialised countries. In the acoustic community, the main endeavours of the two last decades have been oriented towards a better understanding of the dissipation phenomena in absorbers (and especially in poroelastic media) as well as proposing new topologies and structures for these elements. These efforts have resulted in an abundant literature and numerous improvements of the characterisation, modelling and design methodologies for a wide range of media and many different systems.
The chosen research direction for the present thesis slightly deviates from this usual path of modelling absorbing materials as bulk media. Here the aim is to investigate the interfaces between the different components of typical absorbers. Indeed, these interface regions are known to be difficult to characterise and controlling their properties is challenging for a number of reasons. Interfaces in sound packages for instance are inherently by-products of the assembly process and, even if they surely have an important impact on the acoustic performance, they remain mostly overlooked in the established modelling practices. Therefore, the overall objective of the current doctoral project is to identify strategies and methods to simulate the effect(s) of uncertainties on the interface physical or geometrical parameters.
The present licentiate thesis compiles three works which together form a discussion about techniques and tools designed in an attempt to efficiently model thin layers and small details in rather large systems. As part of the work a section of physical model simplifications is discussed which will lay the ground for the next stages of the research. Two publications on the first topic are included, presenting Finite-Element-based hybrid methods that allow for coating elements in meta-poroelastic systems to be taken into account and reduce the computational cost of modelling small geometric features embedded in large domains. The third included contribution is an anticipation, to a certain extent, of the remainder of the doctoral project, discussing the use of physical heuristics to simplify porous thin film models. Here a step towards the modelling of interface zones is taken, departing from numerical simulations and reflecting instead on the physical description and modelling of thin poroelastic layers.
Modelling framework for thin interfaces - MSc 2 final project
The end of the Master degree comprises a report (master thesis) discussing the outcomes of a 4 to 6 months internship.
I realised mine at the Royal Institute of Technology (KTH, Stockholm, Sweden) under the supervision of Pr. Olivier Dazel and Pr. Peter Göransson. Conducted in the Marcus Wallenberg Laboratory, the internship was about modelling thin interfaces for multilayer systems involving porous materials.
We studied different types of porous films, their effect on the system's performance and implemented a hybrid FEM-based method to account for the propogation through films without meshing them.
As a conclusion, we analysed the sensibility of the model to randomness of some parameters (airflow resistivity, etc...) and discussed the usability of the hybrid method to model uncertainties on thin interfaces.
The report can be downloaded here : Modelling framework for thin interfaces (en)
The slides for the oral defense can also be downloaded : Slides (en)
DGM/FEM Coupling - MSc 1 final project
A great part of my first year of MSc was evaluated on a 4-6 months project.
Mine was completed under supervision of Pr. Olivier Dazel and treated the subject of coupling between Finite Elements and Discontinuous Galerkin Methods. The project relied heavily on the work of Dr. Gwenaël Gabard, from the University of Southampton, and Oliver on the usage of a set of plane waves as an expansion basis for DG method. He, Gwenaël, also took part in the project.
We discussed the mathematical formulations of both methods and expressed the boundary operators in a way allowing coupling. Finally, noting a loss of one convergence order, we experimented the use of Hermite polynomials as a interpolation basis for FEM which led to better results.
The report (in french) can be downloaded from Github : Couplage des méthodes FEM et DGM (fr)
The project was interesting enough to be discussed in a talk held by Olivier at the French Congress of Mechanics in Lyon, August 2016. The slides can be found here : Couplage FEM/DGM avec ondes planes (fr)