In this thesis we aim to answer a number of key questions related to unmodelled gravitational-wave (GW) transients, namely: (1) how can we detect an unmodelled GW transient (‘burst’); (2) how well can we reconstruct GW burst parameters; (3) how can we infer the structure of an unmodelled GW source based on the observed signal. Chapter 1 introduces GW astronomy: how gravitational waves are produced, what are the main categories of GW sources, and how GW detectors work. We end the chapter with a summary of the Advanced LIGO (Laser Interferometer Gravitational-wave Observatory) and Virgo observing runs. In Chapter 2 we describe the most promising sources of unmodelled GW transients such as gamma-ray bursts (GRBs), supernovae (SNe), isolated neutron stars and fast radio bursts (FRBs). We focus on the short GRB–compact binary coalescence (CBC) and long GRB–supernova progenitor models. In the following chapter (Ch. 3) we present X-Pipeline, a coherent search pipeline for GW bursts. We define a theoretical framework necessary to perform a coherent analysis with X-Pipeline, and describe how X-Pipeline can be used for searches for GWs associated with GRBs. In the second part of the chapter we report results of such analyses for the LIGO–Virgo Observing runs 2 and 3a. Chapter 4 presents a study that answers the question no. 2, i.e. how well can we reconstruct GW burst parameters, especially the waveform h(t). We perform an injection study with BayesWave, a Bayesian parameter estimation algorithm, using binary black hole (BBH) signals in LIGO–Virgo data. We assess BayesWave performance against the first-principle estimates in three key areas: sky localisation accuracy, signal/noise discrimination, and waveform reconstruction accuracy. Finally, Chapter 5 introduces a novel technique to reconstruct a source mass density perturbation from the GW signal h(t). We start by deriving the algorithm and testing it with multiple sample sources. We describe in more detail why the algorithm is unable to reconstruct the radial evolution of a BBH merger, and provide a Bayesian framework that could solve this issue by including additional constraints. We end the chapter by discussing the method’s limitations, possible solutions and future development work.
|Date of Award||1 Nov 2020|
|Supervisor||Patrick J. Sutton (Supervisor)|