Pressure Sensor Placement Evaluation and Optimization for Singapore’s Water Distribution Networks

Abstract Submission: In collaboration with PUB, Singapore's National Water Agency, a novel cloud-based solution, Anomaly Leak Finder (ALF), has been developed to address the critical challenge of leak detection and localization in Singapore's water distribution networks (WDNs). Monitoring 3 selected supply zones with over 1,000 km of underground pipelines and 95 smart sensors, ALF utilizes near real-time hydraulic data and integrates advanced data-driven prediction (DDP) with physics-based simulation (PBS) models to mitigate non-revenue water (NRW) losses. To optimize the pressure monitoring for anomaly detection and localization, an improved pressure sensor placement software technology, utilizing Darwin optimization techniques, has been implemented and applied to evaluate and optimize the pressure sensor placement in Singapore. Using Monte Carlo sampling, the developed tool simulates single or multiple leaks at randomly selected junction nodes in the hydraulic model to evaluate the detectability of the current pressure sensors. The Darwin optimization process then recommends alternative sensor locations to maximize the overall leak coverage range. Additionally, to handle large hydraulic models and numerous sensors efficiently for the sensor placement analysis, the tool integrates parallel computing via Message Passing Interface (MPI) to improve the computational performance. This paper presents the results of the sensor placement evaluation, showcasing alternative sensor placements and the potential benefits of supplementing them with acoustic sensors to maximize the network’s leak coverage. The proposed approach aims to enhance monitoring coverage for improving the hydraulic model and data analytics application including, but not limited to, detecting and localizing leaks, hence reducing NRW losses in Singapore’s networks.

Learning Objectives/Expected Outcome (Optional) :