Turning Smart Meters Into Water Leak Detectors

Startup PowWow Energy taps data analytics for the agricultural water-energy nexus.

Here’s a question with a lot of value to California’s $43 billion agriculture industry: how can you detect and prevent wasteful, expensive and crop-damaging water leaks without installing a single new sensor?

The answer, according to PowWow Energy, lies in the smart meters that measure the electricity used to power irrigation pumps and water delivery systems across California. The electricity usage of those smart meters represents a set of data that, with the proper analysis, can be turned into an accurate view of when -- and soon, where -- water is being lost to leaks in farm and ranch irrigation systems across the state.

That’s the concept the Sunnyvale, Calif.-based startup has been testing in a handful of pilot projects over the pas six months. In November, it won top honors at the annual Cleantech Open contest and a $200,000 prize to add to the bootstrapped startup’s coffers, and recently raised an undisclosed angel investment, PowWow founder Olivier Jerphagnon told me in a recent interview.

Jerphagnon, a native of France and graduate of UC Santa Barbara, came to the green technology field after fifteen years in the telecommunications industry, including work on data analytics for network diagnostics. In his time working in Australia and California, he determined that the water industry was in need of technology to manage what’s become an increasingly scarce and expensive resource.

But he also found that “there are very [few] standards in the water industry, in terms of sharing data.” He also found that cash-strapped farmers and ranchers aren’t willing to spend lots of money to install sensors to detect water losses, which still make up only a fraction of their operating costs. At the same time, water pumps in aggregate use about 8 percent of California’s total electricity, as the chart below shows, making energy wasted in delivering that water a worthwhile efficiency target.

That led him to ask, “Why don’t we use the energy footprint of water as a fingerprint?” Water pumps use a lot of electricity, from 20 to 200 kilowatts on average, and each tends to have its own meter to measure that power use. Old-fashioned electromechanical meters aren’t useful tools for collecting that data -- but with smart meters, “You don’t need to deploy a new sensor -- you can just grab the data from the utility.”

Putting that data through statistical analysis can allow PowWow to match changes in pumping operations to leaks in the system. In fact, during its pilot tests with three agricultural customers this fall, PowWow’s analytics were able to find not only several large-scale leaks that would have otherwise taken weeks or even months to catch, but also some that were invisible to the naked eye, he said.

That’s the example he provided from one test customer, a horse rancher in the Sacramento region, who at first didn’t believe that the second leak PowWow’s software detected was actually happening. “But a week later, it came to the surface -- it was a micro-leak, and the amount of water was very small, but the energy change was visible,” he said. “It’s like white noise -- even if they did have a water meter, they wouldn’t have been able to detect it.”

Finding leaks in real time isn’t just about reducing water waste, he noted -- it’s also about protecting the crops and livestock that rely on that water. In fact, Jerphagnon found that potential customers who weren’t really interested in water conservation as a business case got interested when “we changed the narrative by saying, water is critical to your operation -- can I help you protect your farm or ranch by giving you an alarm whenever you have a problem?”

Now that PowWow has proven its ability to detect when leaks are happening, it’s turning to the next step -- finding out where they’re happening. That requires quite a bit more data-gathering across the network of smart meters and water pumps serving a customer’s operation, as well as getting the customer to drive around the property and turn certain sections of the delivery system on and off on a schedule, to match those changes to the incoming electric data.

Once that’s done, PowWow applies some more sophisticated methods such as regression analysis and artificial intelligence to “train” the software to match electric data changes to spots on the map. While the location capabilities are still under development, the company is seeing results, he said. 

“You might not be 100 percent sure that it is a specific parcel, but you could be pretty sure that it is one of those north parcels -- it’s in that 20-acre region for sure, and it’s probably in that one-acre parcel,” is how he described it. As with much of the insight delivered by data analytics, the more data PowWow can collect, the more accurate the results become.

The agriculture industry doesn’t lack for water meters, drip irrigation system monitors, soil moisture sensors and other high-tech devices that can deliver location-specific data, of course. But “that costs money upfront -- and your return on investment is going to take time,” he said. Likewise, adding wireless water meters to decades-old pumps is an upfront cost few farmers are willing to take on, whereas smart meters are already there in most cases.

Amidst this world of hardware-enabled systems, “We’re a software solution -- and we’re the only one,” Jerphagnon said. But PowWow has also made its platform as simple as possible for busy farmers to use. Most of its leak alerts are sent via text message, for instance, although it can also provide more data via web interface.

As for water use optimization and planning, PowWow is looking to integrate with established agricultural business analytics software partners, or startups like Fresno, Calif.-based OnFarm that integrate multiple farm equipment and device data sources into business intelligence interfaces, he said. Those partners, in return, could offer PowWow’s smart-meter-based leak detection capabilities as part of their broader services -- at least in California, where smart meters have been deployed, and where PowWow is concentrating its efforts today.

In October, PowWow went live with San Diego Gas & Electric’s Green Button Connect program, which gives the utility’s smart-metered customers free ongoing access to the electricity data that feeds the startup’s software-as-a-service (SaaS) analytics platform. Jerphagnon said PowWow is setting up similar relationships with Southern California Edison and Pacific Gas & Electric, both of which support Green Button data delivery, and hopes to follow the spread of smart meters and utility-provided meter data across the Western United States.

Eventually, “everybody will have smart meters,” he said -- a fact that’s led to an increased interest in technologies that municipal water utilities and water districts can use to push conservation and efficiency to customers. 

PowWow estimates that the first leak it detected in its pilot trials would have wasted about 12,000 gallons of water and 800 kilowatt-hours of energy if left unfixed until it was found by ordinary means. Extrapolating those results to 300 megawatts worth of water pumps, that could add up to 342 million gallons of water saved, or more than 1 million acre-feet.

But it could also reduce wasted electricity consumption by an amount that would save even more water, in terms of what would have been required to cool and operate the power plants that generate it, he said. Consider it another demonstration of the way that energy and water savings go hand in hand.