This article investigates the use of concept drift detection in identifying leakages in water distribution networks. Leakages pose a risk as they cause water loss and increase contamination risks. Detecting these leakages is complex due to the dynamics of water distribution networks. The study explores the potential of model-loss-based and distribution-based drift detection methods for leakage detection. It also discusses the issue of temporal dependencies in the data and proposes a method to deal with it when applying distribution-based detection. The methods are evaluated for leakages of different sizes and detection times.
Publication date: 4 Jan 2024
Project Page: Not provided
Paper: https://arxiv.org/pdf/2401.01733
