AI tool uses sound waves to detect and repair leaky pipes

In the fight to conserve water, leaky pipes are a sneaky enemy, accounting for an average of 30% of the world’s drinking water from pipes, and up to 70% in some areas. Leaks can be incredibly difficult to pinpoint, but a new AI tool is making the process faster and more accurate.

The white X marks a leaking hydrant in the Santa Monica metropolitan area, southeast of the Phoenix metropolitan area. The leak is moderate, ranging from 3 to 7 gallons per minute. On a typical day, the leak is equivalent to the average water consumption of 43 households.

In this case, there are no typical signs of a leak: no puddles, no unexpected green patches in the desert landscape, no collapsed streets. The water is seeping through the bottom of the fire hydrant.

“It most likely leaked until the next hydrant maintenance cycle, which could be up to five years away,” said Jacob Rogers, division president of EPCOR, Arizona’s largest private water company.

But the leak was detected much sooner thanks to an AI acoustic tool developed by FIDO Tech. The technology not only locates leaks more accurately, but also ranks them by size, allowing utilities to prioritize where to fix them first. FIDO, based in Oxford, UK, is working with several utilities, including EPCOR in Arizona and the State Water Commission in Querétaro, Mexico, as part of a Microsoft program to act responsibly by replenishing water from watersheds around the world where it operates, including data centers.

Leaks cause more waste than water.

Stopping leaks in municipal water systems prevents waste and increases water availability for the customers and communities that the utility serves. For the community involved, leaks mean not only the water lost, but also the costs of pumping and transporting the water (it is so heavy that it requires huge pipes and a lot of energy to push it), filtering it, and using chemicals to treat it. This lost water is called non-revenue water because it never reaches the customer. Instead, everyone pays for it.

“We took on a utility in Arizona that was losing 30 to 40 percent of its water,” said Shawn Bradford, EPCOR’s senior vice president of regulated water for Arizona and New Mexico. EPCOR has been working with FIDO under Microsoft’s program for about a year. “We were having to pump 40 percent more water than we needed to plug the leaks from our wells or treatment plants to get water to our homes, and that’s a huge cost that all of our customers face.”

EPCOR has reduced its non-revenue water from 27 percent to about 10 percent thanks to FIDO. FIDO provides sensors that are placed on easily accessible pipes, fire hydrants, valve chambers or faucets, says FIDO’s Edwards. EPCOR has installed 4,554 of these sensors on its pipe network in its 160-square-mile San Tan service area. The sensors record even the quietest noises.

“Sound propagates differently through materials, just like in a school orchestra. You know, the longer the trombone, the longer the pipe, and the deeper the sound, the shorter the length and the higher the frequency, compared to a trumpet. The biggest leaks are the quietest. The human ear can’t hear them, especially in plastic pipes,” says Edwards.

Utilities have long used acoustic devices to detect leaks. FIDO’s technology goes one step further by running the data collected by the sensors through a deep-learning AI tool. The tool has learned to accurately determine whether a noise is caused by a leak or something else, such as a passing machine or train. It can also assess the size of the leak and pinpoint its exact location. Another benefit of FIDO’s technology is that utility technicians can easily interact with the AI ​​using natural language.

EPCOR has used acoustic leak detectors elsewhere, but Bradford says, “You have to spend a lot of time figuring out the exact location between the two devices.” “The advantage of FIDO is that it takes all that raw data and runs it through AI to help pinpoint where the leak is. FIDO can do that in plastic pipes, which has always been a challenge for the industry. Leaks don’t resonate the way they do in metal pipes.”

The lack of resonance in plastic pipe means that many acoustic leak detectors have difficulty. However, most new pipe infrastructure is made of plastic or PVC, which is lighter and easier to handle, making it easier to install and repair. In addition, plastic pipe is better suited to the soil chemistry of the American Southwest and is dominant in the San Tan Valley, one of the fastest-growing communities in the United States, which has developed over the past 30 years. The utility also has metal pipe. Some are 60 or 70 years old and still in excellent condition. Pipes can fail for a variety of reasons, including changes in ground conditions due to freeze-thaw cycles. In the rapidly growing San Tan, leaks tend to occur in areas under construction.

Before FIDO, the utility relied on satellite imagery to find clues like green plants growing in unexpected places, Rogers says. But such greenery doesn’t sprout overnight. It takes days or weeks to grow. FIDO can find leaks much faster.

Fast and flexible solutions

Edwards says the AI ​​also analyzes maps of the utility grid to advise where to place sensors to avoid hearing loss points. The results come immediately. “You can leave them overnight or leave them for years,” she says. The sensors can be easily moved to monitor different parts of the pipe infrastructure.

Sensors can be placed on either side of a leak after the repair is complete to check if the repair was successful or to see if any other leaks were created by the breakage during the repair. The sensors can then be moved to start detecting in other parts of the network. Meanwhile, FIDO’s AI can assess the amount of water leaking and prioritize repair work for maximum effectiveness.

The leaking hydrant, marked with an X, was found to have a defective seal and was quickly repaired. Not all leaks are easy to fix. Repairing a buried pipe may require digging a hole. Locating the exact location of the leak is essential to avoiding unnecessary tearing up of streets, sidewalks, or yards. It’s not just about minimizing traffic congestion. Utility companies must resurface roads after repairs, which represents a significant expense that increases with the amount of digging.

Image source: Shutterstock