Modeling of turbulence and acoustics in complex compressible flows using Theory, Large Eddy Simulation and Machine Learning - HDR de l'INSA Lyon
Hdr Année : 2024

Modeling of turbulence and acoustics in complex compressible flows using Theory, Large Eddy Simulation and Machine Learning

Modélisation de la turbulence et de l'acoustique dans les écoulements compressibles complexes à l'aide de la théorie, la simulation aux grandes échelles et l'apprentissage automatique.

Résumé

Chapter 3 : Combustion noise modeling: This chapter describes in more details my contributions regarding combustion noise. The question of the modeling of the Flame Transfer Functions (FTF) for the analysis of thermoacoustic stability maps is first adressed. Results concerning the analytical modeling of indirect combustion noise as well as the optimal design implications are then presented. Finally, the work devoted to the derivation and analysis of disturbance energies in reacting flows is summarized. Chapter 4 : Broadband rotor noise modeling: Fan noise sources are presented as well as the LMFA testing facility hosting the ECL5 Fan/OGV stage configuration. After a quick summary of the numerical setup for LES, the numerous results obtained through numerical analysis are described. Namely, the impact on the acoustics of a recirculation bubble is presented. The tip gap region is analyzed through its peculiar aerodynamic and aeroacoustic behavior. Finally, the 360  results are confronted to the experiments. Chapter 5 : Turbulence in dense gas flows modeling: In this chapter, after a brief introduction to the specifics of dense gas flows, DNS of academic flows involving dense gases are presented. The peculiar features of turbulence in Homogeneous Isotropic Turbulence (HIT) and compressible mixing layer are analyzed. The behavior of SubGrid-Scale (SGS) terms is explored through a-priori filtering and through the analysis of the performance of existing SGS model. New SGS models are developed using machine learning techniques (especially Artificial Neural Networks). Their merits are analyzed a-priori and a-posteriori using experimental results from Cambridge University.
Fichier principal
Vignette du fichier
hdr.pdf (102.11 Mo) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
licence

Dates et versions

tel-04660217 , version 1 (23-07-2024)

Licence

Identifiants

  • HAL Id : tel-04660217 , version 1

Citer

Alexis Giauque. Modeling of turbulence and acoustics in complex compressible flows using Theory, Large Eddy Simulation and Machine Learning. Reactive fluid environment. Université Claude Bernard Lyon I, 2024. ⟨tel-04660217⟩
126 Consultations
5 Téléchargements

Partager

More