Verdin, Patrick G.Bourdillon, Arnaud2017-08-162017-08-162016-08http://dspace.lib.cranfield.ac.uk/handle/1826/12319Due to the current oil consumption increase, deposits have decreased drastically. Engineers are constantly pushing the limits in order to drill deeper, convey oil further or build more efficient and sturdy pipelines. In an e.ort to obtain this rare ressource, companies are sometimes forced to install pipelines in extreme conditions and isolated terrains to exploit untouched oil deposits. If the structural designs of these devices have come to an agreement, internal phenomena occuring during the oil transport, are, currently not fully understood. In particular, droplets distribution evolution along with freezing events are the two main mechanisms responsible for efficiency loss of pipelines under extreme conditions. The aim of this work is to improve the current knowledge on these phenomena. For many years, oil industry has focused on expensive experiments to better apprehend complex flow phenomena. A promising alternative, computational fluid dynamic (CFD), has been used in this PhD to fill the gap of knowledge in this field of study. Two new single-fluid solidification solvers, an improved population balance model and a novel multi-fluid solidification model have been developed. These solvers have been implemented in an open-source CFD environment (OpenFOAM) to ensure universal acces and a potential extension to this work. It is shown that both single-fluid solidification solvers provide very good results when compared to experimental data. The maximum local discrepancies are evaluated below 20% for the worst case. The population balance model study performed in this PhD has identified important parameters, often under-looked. These findings have led to an improvement of the previous model close to 30% for the best case when compared to experimental results. The multi-fluid solidification model provides accurate ice formation rates (10% of maximal local discrepancies) when compared to experiments. The work presented in this thesis, describes, within the same CFD environment, solvers able to compute both droplet size and distribution evolution and solidification processes. They can be used separately or conjointly to perform the numerical analysis of the flow behavior under extreme conditions, improving the way such problems are currently tackled. They can also be enhanced further to deal with sligthtly di.erent research areas such as hydrates formations and corrosion events.© Cranfield University, 2016. All rights reserved. No part of this publication may be reproduced without the written permission of the copyright holder.Thesis or dissertation