evidoria

← Back to browse

Good practice Imported

PUB's Digital Twin and Smart Sensors — Singapore Cuts Water-Pipe Leaks to 5 per 100 km a Year

Singapore · Singapore · See the Singapore profile · See the Singapore profile

Evidence: Descriptive / self-reported Top 32% 67/100 · Ask Evidence Copilot about this practice

Singapore's water agency PUB paired acoustic leak sensors with digital-twin and machine-learning analytics, scaling from a 2017 pilot on 70 km of mains toward 500 km, and now reports about 5 pipe leaks per 100 km a year, among the lowest rates worldwide.

~5 leaks per 100 km of pipe per year
Current leak rate (as of reporting)
120 EchoShore-TX sensors on ~70 km of DN700 mains
Initial pilot sensor deployment (from 2017)
~500 km of mains
Planned coverage extension (over 5 years)
13mm corrosion hole found within 500mm of predicted location
2018 leak detection accuracy (2018)
PUB's Digital Twin and Smart Sensors — Singapore Cuts Water-Pipe Leaks to 5 per 100 km a Year

Details

Maturity
Established
Promoter
PUB, Singapore's National Water Agency
Period
2017-2026
Keywords
water utilities, infrastructure maintenance, digital twin, IoT sensors

Context

Singapore has no natural aquifers, so PUB, its National Water Agency, has strong incentive to minimise losses across its roughly 6,000 km drinking-water pipe network.

Objectives

Use acoustic sensors, in-pipe sensors and digital-twin/machine-learning analytics to detect developing leaks in water mains before they become bursts.

Activities

From 2017, PUB trialled 120 EchoShore-TX acoustic sensors (Echologics) on about 70 km of large DN700 steel mains, streaming data to a cloud server where algorithms searched for leak noise signatures. Following an early success, PUB extended acoustic and ball-shaped in-pipe sensors to survey roughly 500 km of mains over five years, and used digital-twin software (including Bentley's OpenFlows) to combine machine-learning anomaly detection with hydraulic simulation.

Results

In 2018 the system flagged a leak on a DN700 main that field crews confirmed as a 13mm corrosion hole within 500mm of the predicted location. PUB now reports about 5 leaks per 100 km of pipe a year, cited as among the lowest non-revenue-water rates of any water utility worldwide.

Conclusions

The programme shows a concrete early detection success and a low current leak rate, but the specific contribution of the AI/sensor layer versus decades of broader asset renewal is not isolated in available sources.

Implementation
Indicative cost
Medium (€50k–€500k) — Combines commercial acoustic-sensor hardware (120+ units), cloud data infrastructure and licensed digital-twin software (Bentley OpenFlows) deployed at network scale, consistent with a medium-cost infrastructure programme rather than a low-cost pilot.
Time to results
Long (> 3 years) — Nearly a decade of layered rollout: 2017 pilot on ~70 km, extended to ~500 km within five years, with the ~5 leaks/100km figure reported after years of cumulative deployment.
Staffing & skills
PUB (Singapore's National Water Agency) engineering and field-crew teams, vendor Echologics (sensor hardware), digital-twin software partner (Bentley OpenFlows)

Conditions for success

  • Continuous cloud-based monitoring of acoustic sensor data for leak noise signatures
  • Combining machine-learning anomaly detection with hydraulic digital-twin simulation
  • Sustained multi-year funding to extend sensor coverage across the network

Common failure modes

  • No controlled before/after study isolates the AI/sensor component's specific contribution to the leak rate, which also reflects decades of asset renewal
  • Reporting comes largely from trade press and vendor case studies rather than a PUB-published evaluation

Where it fits

Governance type
national statutory water utility
Scale
citywide network, ~6,000 km of pipes
Income level
high-income

Commonly funded by

National / regional programmes

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful