Description

Energy systems are important to the future of UK industry and society. Members of the Statistics group of the Durham Mathematics department (Michael Goldstein and Hailiang Du), are part of the multi-university CESI (Centre for Energy Systems Integration) consortium whose aim is to reduce the risks associated with securing an integrated energy system for the UK.

This area involves a considerable degree of uncertainty. One of the major sources of information in assessing such uncertainty is by the analysis of computer simulators based on mathematical models for the various areas of application. There are many approximations and uncertainties involved in constructing such simulators and matching the output of the simulators with actual observational data. For example, to evaluate a typical model requires the specification of a collection of input parameters, whose true values are unknown.

The aims of this project are, firstly, to explore the general statistical principles involved in learning about real world physical systems by the use of computer simulators (this general area is often termed uncertainty quantification for computer models, and is widely applicable across many areas of science and technology) and secondly to apply these ideas to important energy related real world problems arising from our work with CESI.

The applications that we will choose between are collected from various energy systems, for example i) the records of historical regional energy demand and weather observations which will be used to provide decision support for local wind farm planning; ii) the records of house or building historical energy consumptions which will be used to forecast future energy consumption as well as to provide decision support for retrofitting. In each case the historical records will be augmented by an ensemble of computer simulator evaluations for the problem.

This project will be concerned with uncertainty analysis for computer models in energy systems and in particular with the practical issues that are raised when analysing an ensemble of model evaluations for a physical system and comparing the output to a collection of observational data. A basic tool that we use is the statistical modelling of the relationship between the inputs and the outputs of the computer simulator, by means of a statistical emulator. Therefore, this project draws on methods described in 2H Statistical Concepts and 3H Statistical Methods.

Prerequisites

Resources

Some details of the kind of analysis that we might carry out, in the context of a different but related energy model, are given in this document

More details about the CESI project are given at

the home-page for the CESI research programme

A good web-site which is related to the general types of analyses which this project gives an introduction to is

The MUCM Web-site

This is the web-site for the Managing Uncertainty in Complex Models (MUCM) project, another consortium in which we were involved. That project is now completed but there are many interesting links to follow at this site.