MathsNet: Research Student Directory

Name Thomson, Michael J
Group Statistics and Probability
Room C11, Math Sci
Telephone +44 (0) 115 95 14925 [direct dial]
Email
Thesis Title Statistical modelling for some problems in carbon capture and transport
Thesis Abstract Capturing and storing CO2 produced by power plants is a potential means to reduce emissions that contribute to climate change. CO2 is transported using pipelines and a number of key questions require answers: what should be the standards for the construction and operation; what pressure should be maintained; and how should the pressure be maintained. This calls for a need to characterise, via an "equation of state", the physical properties of CO2 (and of mixtures of CO2 with impurities and other gases) during transport. Traditionally, equations of state have been modelled parametrically; however, we will investigate the possibility of using non-parametric approaches, such as Gaussian processes, which require fewer assumptions, have the potential for greater flexibility, and offer a natural way to characterise uncertainty in predictions. Several technical challenges present when using Gaussian processes for equations of state: how do we incorporate asymptotic behaviour? How do we build a non-stationary covariance function to model the effect of pressure, temperature and mixture? How do we overcome the computational difficulties faced when using large datasets? We will examine these problems and more.
School contact information School of Mathematical Sciences
University of Nottingham
University Park
Nottingham, NG7 2RD
UK
Tel: +44 (0) 115 951 4949
Fax: +44 (0) 115 951 4951