AI Tool Detects Water Pipe Leaks Using Sound Waves Technology

In the quest for water conservation, leaky pipes pose a significant threat, with global average losses of about 30 percent—and in some regions, up to 70 percent—of piped drinking water. Although pinpointing these leaks can be tricky, an innovative AI tool is streamlining the process for enhanced accuracy.

A white X designates a leaking fire hydrant in the San Tan water district, located just southeast of metropolitan Phoenix. The leak is characterized as moderate, discharging three to seven gallons per minute, resulting in daily losses equivalent to the water consumption of 43 households.

In this instance, the typical signs of a leak—puddles, excessive greenery amidst the desert terrain, or street collapse—are absent. The water has been escaping subtly from the base of the hydrant.

“This leak might have gone unnoticed until the next fire hydrant maintenance cycle, which could be as long as five years,” remarks Jacob Rogers, division director at EPCOR, Arizona’s largest private water utility.

However, this leak was detected sooner, thanks to an AI acoustic tool developed by FIDO Tech. This technology not only improves the precision of leak detection, but also categorizes them by severity, aiding utilities in prioritizing repairs. Based in Oxford, U.K., FIDO is collaborating with multiple utilities, including EPCOR in Arizona and the State Water Commission in Querétaro, Mexico, as part of a Microsoft initiative to responsibly replenish water in key watersheds worldwide, including those with its datacenters.

Leaks waste more than just water

Addressing municipal water leaks curtails waste and enhances water availability for the communities served by utilities. For these communities, leaks equate to not just lost water, but also expense related to pumping, transportation – which necessitates substantial piping and energy due to the weight of water – filtration, and the treatment chemicals required. This lost water is termed non-revenue water because it never reaches a paying customer, burdening all customers with the costs.

“We’ve acquired underperforming utilities in Arizona experiencing 30 to 40 percent water loss,” states Shawn Bradford, senior vice president of regulated water for EPCOR in Arizona and New Mexico, which has partnered with FIDO under Microsoft’s program for approximately a year. “We’re compelled to pump 40 percent more water than necessary to compensate for leaks from the aquifer or water treatment facilities to residences, and that imposes significant costs on all customers.”

EPCOR has reduced non-revenue water from 27 percent to around 10 percent, partly through FIDO’s assistance. FIDO supplies sensors installed to contact pipes, whether on easily reachable sections, hydrants, valve chambers, or taps, as stated by FIDO’s Edwards. EPCOR’s pipe network in the 160-square-mile San Tan service area contains 4,554 such sensors, capturing even the faintest noises.

“Sound travels differently through various materials, similar to the dynamics in a school orchestra. For instance, a longer trombone produces a deeper sound compared to the trumpet, which creates higher frequency sounds due to its shorter length. Surprisingly, the largest leaks are often the faintest—inaudible to the human ear—especially in plastic piping,” explains Edwards.

Historically, utilities have depended on acoustic devices to detect leaks. FIDO’s technology advances this by processing the collected sensory data through a deep learning AI, adept at distinguishing leaks from other noises, such as machinery or trains. This AI can also evaluate leak size and precisely locate it. Another advantage of FIDO’s technology is the ease with which utility technicians can communicate with the AI using natural language.

Previously, EPCOR utilized acoustic leak detection elsewhere, but “identifying the precise location of a leak between two devices was often time-consuming,” Bradford notes. “FIDO’s strength is its ability to analyze raw data through AI, enabling precise leak localization. Notably, FIDO can achieve this in plastic pipes, a longstanding challenge for the industry, as leaks don’t produce the same resonance as they do in metal pipes.”

Plastic pipes, though predominant in new infrastructure due to their lightweight and simpler installation and repair, do not resonate well, complicating detection. However, they are well-suited for the soil conditions in the southwestern U.S., especially within the San Tan Valley, which has rapidly developed over the past three decades. Although the utility manages metal pipes as well—some are 60 to 70 years old and still operational—numerous other factors can lead to pipe failure, such as ground shifts from freeze-thaw cycles. In fast-evolving San Tan, leaks commonly emerge in construction zones.

Prior to employing FIDO, the utility relied on satellite imagery to identify indicators like unusual greenery, according to Rogers. Yet, this greenery doesn’t appear overnight—it can take days or weeks to develop. FIDO’s advanced technology enables much quicker leak detection.

A rapid and adaptable solution

Additionally, the AI evaluates the utility’s network mapping to recommend optimal sensor placements that minimize blind spots, Edwards adds. Results are delivered instantaneously. “Sensors can be left in place for any duration—from overnight to several years,” she states. They are easily relocatable to monitor diverse sections of the pipeline.

After a leak is repaired, sensors can also be repositioned on either side to ensure the repair was effective or to detect if any new leaks have arisen due to the repair-related excavation. The sensors can then be shifted to monitor different sections of the network. Moreover, FIDO’s AI can estimate the volume of leaking water to prioritize repair efforts for maximum effectiveness.

The leaking hydrant marked with an X turned out to have a defective seal, and the repair was completed swiftly. However, not all leaks are remedied so easily. Fixing buried pipes often requires excavation. Detecting leaks accurately is crucial to avoid unnecessary damage to streets, sidewalks, or yards. This is not merely about reducing traffic inconveniences; the utility is mandated to repave the streets post-repair, which can lead to considerable expenses that escalate with excavation scale.

Image source: Shutterstock