Sensor placement starts before anyone opens a chamber. A useful plan describes what the team wants to learn, which parts of the system influence that observation and how a result will change the next action. Buying enough devices to cover a map does not answer those questions. Uniform spacing can be easy to specify while leaving important hydraulic boundaries or difficult pipe materials poorly observed.
Start with the decision, then the location
A district-level investigation might need trustworthy inflow measurements and representative pressure points. An acoustic survey may instead need accessible fittings that transmit useful sound along a known pipe route. A critical-main programme might prioritise consequence and response time rather than coverage density. These objectives can overlap, but they should not be confused in the procurement or acceptance criteria.
Write down the anticipated response to a change at each proposed point. If nobody can identify who reviews an alert or which follow-up test is available, the location may produce data without producing a useful decision. Set a realistic baseline period that includes relevant operating conditions. A brief trial during an unusually quiet week may conceal the effects of demand, pumping or seasonal activity.
Understand what connects to the point
For a district metered area, verify the boundary and account for every significant inlet, outlet and storage influence. A valve marked closed in a geographic information system may not be closed in the field. A cross-connection can turn an apparent rise in leakage into a change in supply route. Review operational records and use authorised checks rather than assuming that the digital drawing describes current conditions.
Pressure locations should reflect the question being asked. A high-elevation point can help assess a vulnerable service condition; an inlet point can describe district supply. Neither alone characterises every customer connection. Record sensor elevation and the reference used for pressure comparisons. In hilly cities such as Sheffield or Bristol, comparing raw readings without considering elevation can obscure the hydraulic explanation.
For acoustic devices, examine the pipe material, fittings, branches and expected sound path. Avoid prescribing a universal spacing distance. The useful separation depends on local transmission, instrument sensitivity and background conditions. A trial installation with controlled listening checks can establish whether a proposed arrangement yields interpretable evidence. Keep a record of uncertain lengths and material changes so that later correlation results are not treated as more precise than the underlying map.
Test the installation as it will operate
Below-ground communications can behave very differently from a signal test made at street level. Chamber depth, metal covers, standing water and nearby construction may affect connectivity. LoRaWAN and NB-IoT are communication options, not guarantees of reception. Verify coverage with the intended enclosure, antenna position and reporting schedule. Confirm what the device stores during an outage and how delayed records retain their original timestamps.
Battery estimates also depend on actual behaviour. Frequent transmissions, repeated connection attempts, high-rate sampling and cold conditions may change service life. A design that works only with more visits than the team can provide is not a reliable design. Evaluate access permissions, traffic management, lifting requirements and the possibility of flooding before selecting the point.
Do not treat an ingress-protection label as a complete environmental assessment. IEC 60529 defines IP classifications, but a rating alone does not settle corrosion resistance, cable-entry quality or suitability for prolonged site conditions. Components in contact with drinking water must meet applicable local requirements. In England and Wales, Drinking Water Inspectorate guidance is a relevant starting point; WRAS approval addresses particular product compliance questions and should not be represented as universal system approval.
Maintain a measurement, not just a device
Every installation needs a durable asset identity, location, photograph, mounting description, units, calibration history and responsible owner. Store the distinction between the physical sensor and the measurement channel: replacing a logger should not silently erase the history of the point. If a fitting is altered, record the change because it may alter both hydraulic behaviour and acoustic coupling.
Plan checks around risk and manufacturer guidance rather than choosing a universal interval. A drifted pressure sensor can create a convincing but false trend. A fouled or poorly installed flow meter can distort a district balance. Compare against suitable references, record uncertainty and retain failed checks. ISO 4064 addresses water-meter requirements and testing; its scope and the applicable instrument specification should be checked before using it as an acceptance reference.
A pilot should finish with an operational review, not simply a dashboard demonstration. Ask which alerts were understandable, which were caused by installation problems and whether the field crew could act on the information. Include IT and operational-technology teams in access control, credential management and recovery planning. IEC 62443 provides an industrial cybersecurity framework, but its application needs competent assessment of the actual architecture.
Useful reference points
- Drinking Water Inspectorate — drinking-water requirements and guidance.
- WRAS Approvals — product approval scope and listings.
- International Electrotechnical Commission — IEC 60529 and IEC 62443.
- ISO — ISO 4064 water-meter standards.