November 18, 2025

New leak detection technology saves Hailey 59M gallons of water annually

New leak detection technology saves Hailey 59M gallons of water annually

Leak detection technology reveals 22 previously undetected leaks

The City of Hailey, Idaho, deployed a new AI-driven leak detection technology on 62 miles of its water distribution network and found 22 previously undetected leaks that were losing a total of 59 million gallons of clean water every year. 

 

The challenge:

The City of Hailey is a small city situated in Idaho’s Wood River Valley, with a population of a little over 9,900 people. Though both the Big Wood and Little Wood rivers run through the region, Hailey’s water distribution network was in need of attention.

Between 2019 and 2023 the city estimated that some 29% of the water that it was producing was not reaching consumers. This unaccounted-for water, or non-revenue water (NRW), reached 50% during the winter months.

The city identified three possible causes of this NRW:

  • Leaks in distribution system pipes
  • Faulty or inaccurate water meters
  • Unmetered connections

The problem with leaking water pipes, however, is that they are invisible, and therefore difficult to locate—or even detect—with any degree of confidence using traditional detection methods.

To address this, city leaders solicited assistance from specialized leak detection companies.

 

The solution: CivilSense™ real-time leak detection technology

Solution selection

City leaders asked Eric Landsberg, PE of Clear Solutions Engineering to evaluate leak detection technologies and services from three different providers: Echologics, Asterra, and Oldcastle Infrastructure.

Echologics offered leak detection technology that the city’s workers would install and operate. Relying on existing city resources to undertake both deployment and analysis was considered unviable as the city’s workforce was already stretched thin.

Asterra offered analysis of satellite imagery, using data over a range of wavelengths to identify areas where soils were saturated, suggested the possible presence of a leak. However, local geography is characterized by rocky terrain that does not retain water, meaning that this technique would be unsuited to the requirement.

Mr Landsberg instead recommended that the city use Oldcastle Infrastructure’s CivilSense™ solution, an AI-driven risk analysis and real-time leak detection technology that uses acoustic sensors and an AI trained on over 2.3 million acoustic signatures to detect, locate and size leaks in the water distribution network.

Landsberg recommended the CivilSense™ approach because of its “turnkey” nature, in which the CivilSense™ team would deploy sensors and the AI would conduct the analysis, as it avoided adding to the resource burden on the city’s workforce. He also noted the fact that its >93% accuracy meant that it delivered superior leak detection than either competitor.

The project

Oldcastle Infrastructure deployed CivilSense™ on a 62-mile section of the water distribution network, including mains, service lines and related infrastructure.

CivilSense™ field experts deployed acoustic sensors across the targeted section of the network, and created a total of 807 investigation sessions. These investigation sessions gathered acoustic data that was fed into the CivilSense™ AI for analysis.

The AI, powered by FIDO Tech, performed comparative analysis using its curated library of more than 2.3 million acoustic signatures to generate 220 Waypoints that indicate potential leaks in the system.

What is a Waypoint?

A Waypoint is an acoustic signature that the AI has assessed to be consistent with a leak based on comparison against a comprehensive library of previously detected and validated acoustic signatures. Waypoints are validated and pinpointed using sensors and a technique called correlation, which confirms and then precisely determines the location of the leak.

The field teams used the results of this analysis to redeploy sensors upstream and downstream of each of the Waypoints, and—using cloud computing and a technique known as correlation—this additional acoustic data enabled the AI to eliminate Waypoints that were not leaks, and to validate, locate and size the Waypoints that were leaks. The leaks were then marked up in situ to direct the resulting repair activity.

 

The outcome

The CivilSense™ team created 880 investigation sessions, from which the AI generated 220 Waypoints. Of these Waypoints, CivilSense™ analysis identified 22 previously undetected leaks in the targeted section of network, and pinpointed their locations for repair crews.

The leaks were discovered across a range of network infrastructure assets, including main lines, service lines and fire hydrants. Of these, seven were assessed as large leaks, six as medium and four as small. In addition, CivilSense™ was able to pinpoint the location and assess the sizes of five leaks that were known about but which could not be located.

In total, applying AWWA nominal volumetric values, the leaks were assessed to be losing a combined volume of around 59.2M gallons of water every year.

By revealing this loss and enabling the City of Hailey to conduct targeted repairs to address the leaks, CivilSense™ leak detection technology is helping the city address its leaks, improve system efficiency and make budgetary savings that can be reinvested into the community.

 

Learn more