Before San Jose Water invested in permanent acoustic leak detection technology, the utility operated on an assumption common across the water industry: that leaks typically surface within a few days to a week of starting, since a break in a pressurized pipe should eventually work its way to the surface. Once the sensors actually went in, that assumption turned out to be badly wrong. Within just two months of commissioning the first units, San Jose Water found 36 previously unknown leaks and some had been running undetected for more than five years.
San Jose Water deploys the EchoShore-DX system, developed by Echologics (a Mueller Water Products company), which uses acoustic sensors embedded directly into special fire hydrant caps rather than requiring new excavation or dedicated sensor installations. These sensors listen continuously every night for the specific acoustic signature a pressurized leak generates as it travels through the pipe, capable of detecting leaks on pipes up to 12 inches in diameter. When the system identifies an acoustic anomaly, it automatically correlates and maps the likely leak location and sends that geographic information directly to field staff for investigation, rather than requiring a technician to manually walk a route listening for leaks on a fixed schedule.
The system integrates with San Jose Water's existing hydrant network the utility's service territory includes roughly 19,000 fire hydrants, and by recent counts the utility had deployed sensors on about 2,000 of them, with a stated plan to add approximately 8,000 more sensors to reach a full system deployment of roughly 10,000 total. Rather than treating full coverage as an immediate, uniform rollout, San Jose Water has described its expansion approach as strategic and objective prioritizing sensor placement based on a scientific methodology focused on where monitoring delivers the most value, rather than simply moving sequentially through the service territory regardless of actual risk.
The discovery that some leaks had been running for more than five years without ever becoming visible reveals something important about how water loss actually accumulates in an aging distribution system. A small leak the initial finds ranged from as little as 3 gallons per hour to as much as 30 gallons per minute doesn't necessarily grow quickly or announce itself with visible surface water. It can persist for years, quietly wasting water and, over time, potentially eroding the surrounding soil and pipe bedding in ways that increase the risk of a more serious failure later. San Jose Water has specifically noted that leak location matters as much as leak size when prioritizing response a leak near a creek carries a higher consequence of failure than an equivalent leak elsewhere, since any discharge risks contaminating an environmentally sensitive waterway with chlorinated drinking water, giving the utility a reason to prioritize investigation and repair even for a leak that isn't necessarily the largest by volume.
San Jose Water's acoustic leak detection program has reduced the utility's year-over-year water losses by roughly 55 million gallons, with recent monthly analysis showing the utility finding more than 25 leaks a month through this technology alone. Framed against the state's broader water scarcity concerns, the utility's leadership has been direct about why this matters beyond the immediate cost savings: every million gallons that doesn't leak out of the distribution system is a million gallons San Jose Water doesn't need to source, treat, and pump from an already water-constrained supply making leak prevention a genuine supply-side conservation strategy, not just an operational efficiency measure.
Acoustic hydrant-mounted sensors are the centerpiece of San Jose Water's leak detection program, but they operate alongside several complementary technologies. Satellite-based leak detection uses orbital sensors to scan for the specific spectral signature typical of chlorinated drinking water in soil, providing a broader, faster first-pass survey across the service territory that helps the utility decide where to prioritize deploying its limited acoustic sensor inventory rather than trying to instrument the entire system uniformly from day one. Portable correlating radio loggers give field crews a mobile version of the same acoustic detection capability, useful for investigating a suspected leak area more precisely once the fixed sensor network or satellite survey has flagged it. And once a leak's general location is identified, leak correlators and ground microphones let crews pinpoint the exact spot often within a few feet before digging, avoiding the wasted excavation cost of guessing at a leak's precise location.
Beyond finding active leaks, San Jose Water applies machine learning to a genuinely harder problem: predicting which pipe segments are most likely to fail in the future, before they actually do. The utility's in-house model incorporates pipe attributes including diameter, material, age, length, historical leak data, geospatial information, and hydraulic modeling to rank every segment of the water system by failure risk. Every three years, San Jose Water re-runs this ranking and uses the results to build its water main replacement program a genuinely different, more proactive approach than simply replacing pipe on a fixed age-based schedule or waiting for failures to accumulate in a specific area before prioritizing it for replacement.
Investment in this kind of leak detection infrastructure, like San Jose Water's broader AMI program covered elsewhere on this site, ultimately has to be justified through the CPUC rate-setting process that governs any capital expenditure by a regulated investor-owned utility. The financial case has become more straightforward to make as California's water supply constraints have intensified every gallon of prevented loss reduces the volume the utility needs to source from an increasingly expensive and climate-vulnerable supply, giving conservation technology a more direct and quantifiable financial justification than it might have carried in a less water-constrained era. This dynamic is part of why San Jose Water has continued expanding sensor coverage steadily rather than treating the initial pilot results as a one-time proof of concept.
San Jose Water's approach illustrates a broader principle relevant to any utility managing aging underground infrastructure: continuous, sensor-based monitoring and predictive modeling consistently outperform periodic manual inspection and reactive repair, because so much of the actual water loss and failure risk in a distribution system is genuinely invisible until instrumented. The finding that leaks can run undetected for five years or more under a periodic-inspection model is a strong, concrete illustration of exactly the kind of blind spot continuous IoT monitoring is specifically designed to close and it's a useful data point for any water utility, in San Jose or elsewhere, still weighing whether the upfront investment in this kind of technology is worth it relative to continuing with traditional inspection methods.
Acoustic leak detection doesn't operate as an isolated program it sits alongside San Jose Water's AMI smart metering rollout, covered in more detail elsewhere on this site, as a complementary half of the utility's overall water-loss reduction strategy. Smart meters catch leaks occurring on the customer side of the connection, inside a home or business, while acoustic hydrant sensors and satellite monitoring address losses within the utility's own distribution network before water ever reaches a customer's meter at all. A utility investing seriously in only one half of that equation would still be blind to a meaningful category of water loss, which is part of why San Jose Water has pursued both technologies simultaneously rather than treating them as separate, sequential initiatives.
Frequently asked questions
The utility deploys Echologics' EchoShore-DX system, which uses acoustic sensors embedded in fire hydrant caps to continuously listen for the sound signature of pressurized leaks on pipes up to 12 inches in diameter.
San Jose Water found that some leaks had been running undetected for more than five years before acoustic sensor deployment revealed them, contrary to the utility's prior assumption that most leaks surface within days to a week.
The program has reduced year-over-year water losses by roughly 55 million gallons, with the utility finding more than 25 leaks a month through the technology.
The utility had deployed approximately 2,000 sensors across its roughly 19,000-hydrant network, with plans to add about 8,000 more sensors to reach full system deployment of roughly 10,000 total.
Yes, alongside acoustic sensors. Satellite-based detection scans for the spectral signature of chlorinated drinking water in soil, helping the utility prioritize where to focus more detailed acoustic sensor deployment and field investigation.
The utility uses a machine learning model incorporating pipe age, material, diameter, historical leak data, and hydraulic modeling to rank pipe segments by failure risk, re-running the ranking every three years to guide its main replacement program.
Related articles
Smart Utilities
