The phenomenon of Unaccounted-for-Gas (UAG) in natural gas transmission networks can be summarised as the failure to account for a percentage of network throughput -- typically around 0.3\% per annum -- resulting in an increased transmission cost. This thesis represents the first holistic approach to UAG management in literature. The chief areas of focus are threefold; identifying the causes of UAG within the transmission system, formulating a baseline for UAG under normal operating conditions, and examining the significance of UAG and its role in detecting errors in flow measurement. Aside from these key topics, a statistical analysis of UAG in the UK is carried out and compared to international counterparts. Our research into sources of uncertainty uncovers errors in linepack estimation as significant contributors to UAG in the UK, and we propose measures to rectify them. Regarding the baseline, we combine uncertainty models with statistical methods to produce a hybrid model, which under certain conditions can be expected to account for all sources of uncertainty in a transmission system. We conclude that the sole reliance on baseline methods to uncover large errors in daily flow measurement is insufficient and propose additional statistical monitoring processes with unique considerations for the different node types. Special attention is paid to energy conservation within power stations, and the calculation of downstream distribution UAG as an additional error detection technique. We examine appropriate statistical process control methodologies, and in particular changepoint analysis in detecting systematic flow metering errors in the transmission grid. The thesis presents a case study implementing the developed statistical methods into National Grid's highly regulated operating environment. We provide recommendations for the reduction of UAG and the improvement of analytic processes, which are applicable to most transmission grids across the world.
Date of Award | 1 Aug 2021 |
---|
Original language | English |
---|
Awarding Institution | - The University of Manchester
|
---|
Supervisor | Paul Johnson (Supervisor) & Geoffrey Evatt (Supervisor) |
---|
- regression
- Unaccounted for Gas
- multivariate
- changepoint analysis
- natural gas
- gas transimission
- process control
- statistics
- UAG
- gas distribution
On the phenomenon of Unaccounted for Gas: baseline formulation and error detection techniques
Botev, L. (Author). 1 Aug 2021
Student thesis: Phd