Singapore’s National Water Agency, PUB, has strengthened its approach to detecting leaks on large transmission mains by deploying a new generation of Transmission Pipe Leak Monitoring (TPLM) technology. The advancement supports PUB’s strategy for maintaining resilience across its 5,500 km potable water network, which supplies more than 1.5 million customers.
The Agency has installed the new system across 200 km of large-diameter pipelines that transport high-pressure flows over long distances. Because leaks on these mains can result in significant water loss, accurate and timely detection is essential to reducing non-revenue water and avoiding wider service disruption.
Advanced acoustic monitoring with physics-informed algorithms
The upgraded TPLM system is based on the EchoShore®-TX technology from Echologics. Each monitoring node combines a hydrophone, electronic module and antenna for cellular and GPS connectivity. The nodes collect high-resolution acoustic data and transmit it to a secure cloud platform for analysis.
Recent improvements incorporate physics-driven algorithms and enhanced signal processing, supporting more reliable alerting, streamlined workflows and greater scalability. According to Echologics data analyst Marcin Kloc, physics-based modelling has become essential to unlocking deeper leak insights:
“Anyone can deploy an AI model and feed it a bunch of data, but it takes real scientific skill to figure out how to use physics and maths to transform the data to mine gold out of chaff.”
EchoShore-TX sensors provide continuous acoustic monitoring across PUB’s large transmission mains.
Faster notifications and improved operational efficiency
One key advancement is a reduction in the time required to notify PUB of possible leaks. The previous system required four consecutive noise recordings, typically resulting in a 2–3 day delay. The upgraded platform now collects more frequent acoustic and spectral data, reducing the notification window to 1–2 days and enabling faster intervention.
Single-channel detection has also expanded capabilities. This function allows the system to identify leaks that do not propagate to a secondary sensor for correlation. In one recent case, a leak of under five litres per minute was detected at an air-valve chamber solely through single-sensor acoustic analysis.
Establishing standardised performance metrics
As technologies evolve, PUB has introduced a standardised framework for evaluating leak monitoring effectiveness. The performance metrics focus on:
- Leak classification performance – the ability to differentiate leak sounds from routine network noise, reducing unnecessary investigations.
- Leak sensitivity – the system’s effectiveness in detecting all actual leaks within a monitoring area.
These metrics assess outcomes such as sounds detected, classified correctly, misclassified, or not detected, enabling utilities to benchmark system capability and support data-driven asset management decisions.
Supporting long-term asset resilience
Detecting leaks early on large-diameter mains helps to minimise losses and avoid premature pipeline replacement. With developments in physics-informed analytics and real-time acoustic monitoring, utilities of any scale can adopt more proactive leak management practices and reduce operational costs associated with emergency repairs.
