A monitoring system does not observe “a leak” as a single, unambiguous object. It records sound, pressure, flow, temperature or another physical effect. Software then searches those records for patterns. That distinction matters: an accurate instrument can produce a perfectly accurate record of something other than leakage. Before choosing hardware, describe the decision it should support. Finding a suspect district, narrowing a pipe section and locating an excavation point are different tasks.
Sound travels through a complicated path
Acoustic loggers listen for vibration associated with escaping water. A persistent sound during quieter periods can justify further investigation, especially when neighbouring instruments show a compatible pattern. But pumps, electrical equipment, road traffic and water use also create noise. A loud recording is not automatically a large leak, and a quiet recording does not demonstrate that the pipe is sound.
Pipe material, diameter, pressure, soil conditions and fittings influence transmission. Plastic pipe often attenuates useful leak noise differently from metallic pipe. A logger mounted on a valve may hear a different mixture of sources from a hydrophone in contact with the water. Installation quality and contact condition therefore belong in the evidence record, rather than being treated as minor administrative details.
Correlators compare recordings from separated locations to estimate where a shared sound originates. Their results depend on credible pipe lengths, wave-speed assumptions, timing and the actual route between sensors. An apparently precise location can be misleading if the asset map is wrong or a branch introduces another path. Repeated measurements and an independent listening check can help test the result.
Pressure describes network behaviour
A pressure logger records conditions at its measurement point, not uniformly across an entire district. Elevation, friction, valve positions and demand shape the relationship between locations. An unexpected drop may accompany a burst, but it may also follow a pump change, hydrant use or planned operations. Comparing nearby pressure records with the operations log often provides more insight than adjusting an alarm threshold immediately.
Sampling interval matters. A logger that records occasional averages may describe daily trends well while missing a short transient. Capturing rapid events requires suitable sensor response, acquisition frequency, storage and time synchronisation. The United States Environmental Protection Agency discusses pressure management and distribution-system integrity as connected issues; neither can be reduced to a single minimum reading.
Flow becomes useful through accounting
Flow meters help teams ask whether water entering an area is consistent with expected consumption and storage changes. In a district metered area, commonly called a DMA, the boundary must be understood. Unrecorded connections, bypasses and open boundary valves can undermine an apparently simple balance. A rising night flow can suggest leakage, but hospitals, industrial sites, irrigation and other legitimate overnight users may also contribute.
The International Water Association water-balance approach separates different components of water use and loss. It helps prevent the common mistake of treating all non-revenue water as physical leakage. Meter error and unauthorised consumption need different investigations from a damaged main. London and Singapore illustrate very different demand and network contexts; a useful baseline should come from the actual system, not an assumed universal profile.
Match the instrument to the question
| Measurement | Useful question | Important limitation |
|---|---|---|
| Acoustic | Is persistent noise worth investigating? | Transmission and competing sound vary. |
| Pressure | Did hydraulic conditions change? | Operations and demand can explain the change. |
| Flow | Is movement inconsistent with expected use? | A balance does not pinpoint a defect. |
| Temperature or water quality | Is there a supporting change in conditions? | A change is rarely specific to leakage. |
Combining measurements can reduce uncertainty, provided their errors are not simply repeated. Two alarms derived from the same faulty meter are not independent evidence. Check the measurement chain: sensor, connection, power supply, clock, communications, storage and processing. Record missing data as missing, rather than quietly filling it with zeros that may resemble a hydraulic event.
Technology selection should also consider long-term ownership. Battery replacement, calibration access and staff interpretation time can outweigh a low purchase price. ISO 24516-1 addresses asset management for drinking-water distribution networks; ISO 55000 provides broader asset-management principles. Neither replaces local engineering judgement, but both support the idea that monitoring must serve a managed asset and a defined decision.
Read beyond the instrument
- International Water Association — water losses and water-balance practice.
- US EPA: drinking-water distribution systems — distribution-system context.
- International Organization for Standardization — consult ISO 24516-1 and ISO 55000 through the official catalogue.