• Login
• Admin

Thomson, Michael James (2018) Statistical modelling of equations of state for carbon capture, transport, and storage. PhD thesis, University of Nottingham.

PDF (Thesis - as examined) - Repository staff only - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader Download (21MB) Request Thesis

Abstract

Capturing CO2 produced by power plants and storing it within depleted oil and natural gas reservoirs in the seabed offers a potential means to reduce the emissions that contribute to climate change. This is known as carbon capture and storage (CCS). CO2 captured from power plants is transported to these reservoirs using pipelines. To determine the parameters of pipeline operation this calls for a need to characterise, via an "equation of state", the physical properties of CO2 during transport. Doing so is further complicated by the fact that CO2 captured from power stations is typically a mixture of CO2 with other fluids. This affects the physical properties of the CO2 to be transported and so needs to be accounted for.

In this thesis we develop statistical models for equations of state that can account for the physical properties of CO2 relevant to carbon capture and storage, and which allow us to quantify uncertainty in the predictions from the equation of state. We propose two statistical models for equations of state. Firstly we develop a statistical model which can be applied to any pressure-explicit parametric equation of state. To do so we have developed a novel method by which to rigorously account for uncertainties due to coexistence which is complicated by the fact that it involves perfectly correlated measurements on two fluid phases. We fit this model to pure CO2 obtaining good agreement with data for most temperatures. We then extended this model and our method for accounting for coexistence to mixtures. We fit this model to real CO2-H2 data. Despite mixing well, the results of this fit do not agree well with the data and equations of state need to be developed further to be able to model mixture data well.

Secondly we develop a non-parametric Gaussian process approach which offers greater flexibility and requires fewer assumptions. This non-parametric model is fit to pure CO2 reference data for individual subcritical temperatures. We demonstrate how applying a transformation to the covariance function can account for non-stationarity in the data resulting in good agreement between predictions from the fit model and the data.

Item Type:	Thesis (University of Nottingham only) (PhD)
Supervisors:	Preston, S.P. Graham, R.S.
Keywords:	carbon capture and storage, Bayesian statistics, Markov chain Monte Carlo, Gaussian processes, statistical modelling
Subjects:	Q Science > QA Mathematics > QA276 Mathematical statistics Q Science > QC Physics > QC811 Geomagnetism. Meteorology. Climatology
Faculties/Schools:	UK Campuses > Faculty of Science > School of Mathematical Sciences
Item ID:	48939
Depositing User:	Thomson, Michael
Date Deposited:	16 Mar 2018 15:17
Last Modified:	18 Mar 2018 07:00
URI:	https://eprints.nottingham.ac.uk/id/eprint/48939

Actions (Archive Staff Only)

Edit View