Detection and interpretation of anomalous water use for non-residential customers. (February 2018)

• Record Type: Journal Article
• Authors: Patabendige, Samitha; Cardell-Oliver, Rachel; Wang, Rui; Liu, Wei
• Is Part Of: Environmental modelling & software. Volume 100(2018)
• Page Start: 291
• Page End: 301
• Publication Date: 2018-02
• Subjects: Water demand management -- Anomaly detection -- Decision support -- Smart meter; Environmental monitoring -- Computer programs -- Periodicals; Ecology -- Computer simulation -- Periodicals; Digital computer simulation -- Periodicals; Computer software -- Periodicals; Environmental Monitoring -- Periodicals; Computer Simulation -- Periodicals; Environnement -- Surveillance -- Logiciels -- Périodiques; Écologie -- Simulation, Méthodes de -- Périodiques; Simulation par ordinateur -- Périodiques; Logiciels -- Périodiques; Computer software; Digital computer simulation; Ecology -- Computer simulation; Environmental monitoring -- Computer programs; Periodicals; Electronic journals; 363.70015118
• DOI: 10.1016/j.envsoft.2017.11.028
• Languages: English
• ISSNs: 1364-8152

Abstract

Abstract: Smart water meters can help businesses save water. But achieving this goal requires trusted algorithms for processing the data and intuitive, interactive software systems to support end users in decision making. This paper presents an algorithm and a web-based software system to detect and visualise anomalous water use. The algorithm calculates an anomaly score for each day, together with a rationale describing the symptoms of unusual water use. The score for each day is based on ten features of daily demand and its historical context. The score and its rationale are posted to users to help them track down the underlying physical causes of anomalies. Using data from two aquatic leisure centres, we demonstrate that anomaly scores give better coverage than traditional threshold-based systems, that end users are able to utilise the timely feedback to save water, and that the algorithm is reasonably robust to parameter settings. Highlights: A software system is proposed for detecting anomalous water use from smart water data. The system automatically generates a context-aware anomaly score for each day with a rationale for the score. Lightweight, on-line, user feedback is used to document the physical causes of identified anomalies. The anomaly detection system is integrated with existing management workflows.