Manuela Russo
A combined CFD and Deep-Learning approach for the analysis of bubble dynamics in Newtonian fluids.
Rel. Luca Bergamasco, Andrea Bottega, Davide Picchi. Politecnico di Torino, Corso di laurea magistrale in Ingegneria Energetica E Nucleare, 2026
Abstract
This thesis presents a preliminary investigation into the dynamics of a single bubble rising in a Newtonian fluid, with particular emphasis on how the bubble shape is affected by different system parameters. The analysis is carried out in terms of dimensionless numbers to ensure generality. A range of Reynolds and Eötvös numbers is explored in order to identify the various hydrodynamic regimes and to classify the corresponding bubble morphologies. The simulations are performed with the open-source solver Basilisk, which employs adaptive mesh refinement based on a quadtree data structure and a geometrical Volume-of-Fluid (VOF) method for interface reconstruction and tracking. The results show that variations in the relative magnitude of inertial, viscous, and surface-tension forces lead to substantially different bubble behaviours, whose main features are analyzed in detail.
In the second part of the study, a dataset of simulated bubble images is generated to train a Convolutional Neural Network (CNN) for automated identification and classification of bubble regimes features
Relatori
Anno Accademico
Tipo di pubblicazione
Numero di pagine
Informazioni aggiuntive
Corso di laurea
Classe di laurea
URI
![]() |
Modifica (riservato agli operatori) |
