June 13, 2025

New AI leak detection saves water utility $213,000

New AI leak detection saves water utility $213,000

AI leak detection technology saves money and gallons

A suburban Mid-West water utility used a new AI leak detection solution to locate a large water leak that had evaded traditional detection methods for years—saving $213,000 a year and preventing the loss of 350,000 gallons of non-revenue water every day.

The challenge

A major suburban city in the Mid-West was grappling with a non-revenue water loss challenge. 

Utility managers knew that there was a major leak somewhere in the system, but traditional leak detection methods had failed to locate it. Water was leaking from a storm drain, but there was no indication of a leak nearby; despite extensive efforts, the source of the leak itself evaded detection. 

Having tried three different leak detection techniques and technologies without success, after two years the utility decided that a better approach was needed. 

 

The solution: CivilSense™ AI leak detection

The utility partnered with Oldcastle Infrastructure and opted to use CivilSense™ AI leak detection to pinpoint the location of the leak.  

Having assessed the situation and identified the section of network where the leak was likely to be located, Oldcastle’s expert team deployed acoustic sensors at strategic points throughout the network section. 

These sensors gathered acoustic data that was then fed into a proprietary AI for analysis. The AI, powered by FIDO Tech, compared the data to its curated library of over 2.3 million acoustic signatures and confirmed the presence of a large leak in the network. 

Oldcastle experts then deployed two sensors on the section of pipe confirmed to have the leak, and used a technique known as “correlation” to pinpoint the precise location of the leak. 

 

About Correlation

Two sensors take perfectly synchronized data samples that are uploaded to the cloud, where a mathematical formula is used to calculate the precise location of the leak between the two sensors. 

Correlation is accurate to ±1½ ft, even under challenging conditions, which utilities confirm is a sufficient level of accuracy to locate a leak. 

This is equivalent to the accuracy of some alternative techniques, but has the benefit of being significantly less labor-intensive and faster and easier to achieve.

 

The outcome: $213,000 and 350,000 gallons saved

CivilSense™ AI leak detection successfully identified and helped pinpoint the exact location of the leak, which was a complete circumferential 1/16 inch wide break on a 6-inch metallic water main. 

The break was close to a gravel backfill trench around the homeowner’s storm drain, which meant that the water was able to quickly track away from the site of the break, infiltrate the drain and travel to a downstream connection to the city’s main storm drain, where it became detectable at a manhole some 334 feet away. 

At a water pressure of 60 psi, this break represented a significant water loss of almost 243 gallons every minute, or nearly 350,000 gallons per day. Based on the water utility’s local water production costs, this meant that it was costing the municipality up to $213,043 every year.

This costly and unsustainable water loss situation had baffled utility owners and stumped traditional detection methods for two years. Only CivilSense™ was able to detect and locate the leak quickly, reliably and accurately. 

Download this Case Study

Learn more