Use this URL to cite or link to this record in EThOS:
Title: Data-driven approaches for near real-time forecasting of discolouration events in pipe networks
Author: Meyers, G.
ISNI:       0000 0004 7962 9441
Awarding Body: University of Exeter
Current Institution: University of Exeter
Date of Award: 2019
Availability of Full Text:
Access from EThOS:
Access from Institution:
Water discolouration in potable drinking water networks is an increasingly important and expensive issue due to rising customer expectations, tighter regulatory demands and ageing Water Distribution Systems (WDSs) in the UK and abroad. The increase in real-time data acquisition systems installed in WDSs has opened the door to using data-driven methods to achieve what physically based models have to date been unable to. The research presented in this thesis describes the development of three novel data-driven methodologies to aid in the reduction of discolouration risk and further understand discolouration issues in a WDSs. These methodologies are: a) a continuous turbidity forecasting methodology capable of forecasting if and when a downstream turbidity event will occur by only taking current and past flow and turbidity measurements at a number of upstream locations in the network; b) a methodology for estimating the percentage of downstream turbidity observations that can be linked to an upstream pipe in a network and thus identify network areas (i.e. pipes) where discolouration material accumulates in a WDS; c) an on-line turbidity event forecasting methodology that predicts if a hydraulic event occurring upstream will cause a turbidity event downstream by analysing the current and historic hydraulic forces in WDS pipes. The results of applying these methodologies to data from three real trunk main networks in the United Kingdom (UK) over a period of two years and 11 months are also reported in this thesis. The results obtained illustrate that it is possible to reliably forecast the occurrence of discoloration events in real WDSs by using a data-driven (i.e. non-physically based) methodology only. The results obtained additionally show the potential of the methodologies presented here to be used as an early warning system for discolouration events. This would enable a multitude of cost beneficial proactive discolouration risk management strategies to be implemented as an alternative to expensive trunk mains cleaning programs and thus enabling water companies to save money while improving their customer service and reputation.
Supervisor: Kapelan, Z. ; Keedwell, E. Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available