Small leaks in water distribution networks are major economic and environmental problems as they can go undetected for years. We model the signature of small leaks as a unique Directed Acyclic Graph, called the Lean Graph, and use the model to find the best places fork sensors. We use the sensors to develop dictionaries which map each leak signature to its location. Leaks are quantified by matching out-of-normal flows detected by sensors against dictionary records. Finally, we investigate how much our approach can tolerate corrupted data due to sensor failures by introducing a subspace voting based localization method.
|Qualification||Doctor of Philosophy|
|Award date||13 Mar 2019|
|Publication status||Unpublished - 2018|