Leveraging Transfer Learning in LSTM Neural Networks for Data-Efficient Burst Detection in Water Distribution Systems

Researchers and engineers employ machine learning (ML) tools to detect pipe bursts and prevent significant non-revenue water losses in water distribution systems (WDS). Nonetheless, many approaches developed so far consider a fixed number of sensors, which requires the ML model redevelopment and collection of sufficient data with the new...

Development of an LSTM-based methodology for burst detection in water distribution systems

Water utilities face many challenges, including pipe bursts that cause significant non-revenue water losses. Detecting those bursts early is important for the water sector in its path to achieve sustainable water resource management. This study presents a scalable data-driven methodology for burst detection in water distribution systems that is...

Sectorization for Water Distribution Systems with Multiple Sources: A Performance Indices Comparison

Sectorization is an effective technique for reducing the complexities of analyzing and managing of water systems. The resulting sectors, called district metering areas (DMAs), are expected to meet some requirements and performance criteria such as minimum number of intervention, pressure uniformity, similarity of demands, water quality and...

Strategy to Transition from an Intermittent to a Continuous Water Supply Using a District Metering Area Approach

In this thesis, I develop and evaluate a demand-side approach to transition from an intermittent to a continuous water supply using a district metered areas approach in Accra, Ghana, an urban metroplex of 4.0 million persons in western Africa. Intermittent water supply (IWS) is inherently inefficient. These systems foment health hazards and are...

Uniformity and Heuristics-Based DeNSE Method for Sectorization of Water Distribution Networks

Sectorization of a water distribution network (WDN) into district metered areas (DMAs) is a proven solution for proactive leakage control. Traditionally, WDN sectorization is conducted by local experts using a trial-and-error approach, often resulting in the identification of arbitrary solutions. Some recently published methods try to improve...

Automated leak localization performance without detailed demand distribution data

Automatic leak localization has been suggested to reduce the time and personnel efforts needed to localize (small) leaks. Yet, the available methods require a detailed demand distribution model for successful calibration and good leak localization performance. The main aim of this work was to analyze whether such a detailed demand distribution...