Automated Leak Detection System for the Improvement of Water Network Management




The need for an efficient Water Management System (WMS) is strongly felt by water 
utilities, municipalities and by medium to large scale corporates that have to face every day with 
problems dealing with water usage and supply [1]. Leveraging a sensor data network, an automated 
system to implement fault detection in a water network at an early stage can be a valuable tool that 
saves water, energy, time and money. This paper introduces a novel FDD (fault detection and 
diagnosis) approach for water networks developed within the FP7 Waternomics Project by modeling 
a water network in the simulation environment EPANET and applying an anomaly detection 
algorithm named ADWICE (Anomaly Detection With fast Incremental ClustEring) [2] to real time 
data of water flow and pressure to infer performance and operational anomalies. The method is 
currently being implemented at the Linate Airport water network in Milan, and initial results are 
presented in this paper.

The full paper is available at this link.