De Ceuster, Frederik;
(2022)
Simulating 3D Radiation Transport, a modern approach to discretisation and an exploration of probabilistic methods.
Doctoral thesis (Ph.D), UCL (University College London).
Preview |
Text
PhD_thesis_FrederikDeCeuster_revised.pdf - Accepted Version Download (44MB) | Preview |
Abstract
Light, or electromagnetic radiation in general, is a profound and invaluable resource to investigate our physical world. For centuries, it was the only and it still is the main source of information to study the Universe beyond our planet. With high-resolution spectroscopic imaging, we can identify numerous atoms and molecules, and can trace their physical and chemical environments in unprecedented detail. Furthermore, radiation plays an essential role in several physical and chemical processes, ranging from radiative pressure, heating, and cooling, to chemical photo-ionisation and photo-dissociation reactions. As a result, almost all astrophysical simulations require a radiative transfer model. Unfortunately, accurate radiative transfer is very computationally expensive. Therefore, in this thesis, we aim to improve the performance of radiative transfer solvers, with a particular emphasis on line radiative transfer. First, we review the classical work on accelerated lambda iterations and acceleration of convergence, and we propose a simple but effective improvement to the ubiquitously used Ng-acceleration scheme. Next, we present the radiative transfer library, Magritte: a formal solver with a ray-tracer that can handle structured and unstructured meshes as well as smoothed-particle data. To mitigate the computational cost, it is optimised to efficiently utilise multi-node and multi-core parallelism as well as GPU offloading. Furthermore, we demonstrate a heuristic algorithm that can reduce typical input models for radiative transfer by an order of magnitude, without significant loss of accuracy. This strongly suggests the existence of more efficient representations for radiative transfer models. To investigate this, we present a probabilistic numerical method for radiative transfer that naturally allows for uncertainty quantification, providing us with a mathematical framework to study the trade-off between computational speed and accuracy. Although we cannot yet construct optimal representations for radiative transfer problems, we point out several ways in which this method can lead to more rigorous optimisation.
| Type: | Thesis (Doctoral) |
|---|---|
| Qualification: | Ph.D |
| Title: | Simulating 3D Radiation Transport, a modern approach to discretisation and an exploration of probabilistic methods |
| Event: | UCL (University College London) |
| Open access status: | An open access version is available from UCL Discovery |
| Language: | English |
| Additional information: | Copyright © The Author 2022. Original content in this thesis is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) Licence (https://creativecommons.org/licenses/by-nc/4.0/). Any third-party copyright material present remains the property of its respective owner(s) and is licensed under its existing terms. Access may initially be restricted at the author’s request. |
| UCL classification: | UCL UCL > Provost and Vice Provost Offices > UCL BEAMS UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Maths and Physical Sciences UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Maths and Physical Sciences > Dept of Physics and Astronomy |
| URI: | https://discovery.ucl.ac.uk/id/eprint/10141997 |
Loading...
Loading...Loading...Loading...Archive Staff Only
 |
View Item |
