Bentley Fellow Bentley Systems Inc., United States
Abstract Submission: A cutting-edge cloud-based software platform, Anomaly Leak Finder (ALF), has been developed in collaboration with Singapore’s National Water Agency, PUB, to revolutionize leak detection and localization across Singapore's extensive water distribution networks (WDNs). 3 selected supply zones, spanning over 1,000 km of underground pipelines and monitored by around 95 smart sensors, are continuously analyzed using near real-time hydraulic data. ALF integrates advanced data-driven prediction (DDP) methods with physics-based simulation (PBS) models to proactively identify and localize hidden leaks. To continue improving ALF’s anomaly detection capabilities, the usefulness of leveraging advanced metering infrastructure (AMI) data is explored in a pilot study in Singapore. In principle, the system's total flow balance is assessed using AMI data, that is expected to provide more accurate NRW volume estimates within the closed system. Additionally, the AMI data also improves the hydraulic model by assimilating unique demand patterns localized to specific junctions, hence enhancing ALF’s ability to pinpoint anomaly sources. This paper will thus present the full details from the blind hydrant tests, along with ongoing improvements to ALF’s detection capabilities to address notable field challenges. These enhancements aim to refine ALF’s accuracy in detecting and localizing leaks, with the goal of reducing NRW losses in Singapore’s water distribution networks.
.jpg "Zheng Yi Wu photo")
