A smart valve will tell you the water leaving it was 51 degrees C at 09:14 this morning. What it won’t tell you is whether the pipe beyond it holds a metre of blended water sitting at 38 degrees C, or whether the cold supply two floors up has been creeping past 20 degrees C all summer. That gap, between a reading and actual control, is where most of the value in temperature-management devices lives, and so does most of the false comfort.
What follows are the questions worth settling before you fit smart valves, sensor-equipped mixing valves or automatic flushing devices: what each one genuinely fixes, what it quietly doesn’t, and how to turn its readings into evidence an assessor would accept.
What actually counts as a “smart valve”?
The phrase covers a few different things, and they don’t do the same job. A thermostatic mixing valve (TMV) blends hot and cold to deliver a safe outlet temperature and prevent scalding; that is its purpose, not Legionella control. A smart or sensor-fitted valve adds continuous temperature logging, sometimes with remote alerts. Automatic or solenoid flushing valves open a quiet outlet on a timer or sensor trigger to keep water moving. Some products roll all three together and feed a building management system. Before you buy anything, be precise about which of these you are actually getting, because a logging TMV solves a recording problem, not a stagnation one.
Do smart valves lower Legionella risk, or just record it?
Mostly the latter, and that is worth being honest about. Temperature and movement control the risk; a sensor measures whether you are achieving them [1]. A valve that emails you when an outlet drops out of range has not changed the water, it has shortened the time between a fault and your knowing about it. That is genuinely useful. A cold tank warming in a roof void, or a circulating loop losing heat at its far end, gets caught in days rather than at the next quarterly check. But the device that earns its place is the one wired to a response, where an out-of-range alert actually triggers someone to act. An alert nobody reads is just a more expensive way of not noticing.
Don’t TMVs create a Legionella problem of their own?
They can, and it is an under-appreciated trade-off. A TMV deliberately produces water in the warm range that Legionella prefers, then holds it in the short length of pipe between the valve and the outlet. HSE guidance is to fit TMVs as close to the point of use as reasonably practicable, keep that blended length short, and clean, descale and disinfect the valves at a frequency your risk assessment sets [2]. So a corridor of TMVs retrofitted for scald safety can quietly add dozens of small warm dead legs. Smart TMVs help here only if their data is used to keep those sections short, flushed and serviced, not if they simply make the same risk easier to ignore.
Can automatic flushing valves replace manual flushing on quiet outlets?
Sometimes, and on a large estate with many little-used outlets it can be a real saving, but only when it is specified properly. An automatic flush has to move enough water, for long enough, to actually turn over the dead leg it serves, and it needs to log that it ran. The failure modes are specific: a valve that flushes but never reaches temperature, a drain that backs up, or a unit that silently stops and leaves you believing an outlet is managed when it is not. Treat an automatic flushing valve as a control you have to verify, exactly like a manual task. Your risk assessment and written scheme should name it, set what “done” looks like, and say what happens when it fails [1][2]. For how disciplined flushing plays out when it works, a case study where flushing headed off a Legionella problem is a useful companion.
Will logged sensor data satisfy an HSE inspector?
Good data helps, but the records have to show more than numbers. What an assessor looks for is the management chain: a current risk assessment, a written scheme, a named responsible person, monitoring, and evidence that exceptions were acted on [4][1]. Continuous logs are strong monitoring evidence, provided the sensors are calibrated and someone reviews and signs off the out-of-range events. A dashboard full of green ticks that nobody has interrogated proves you bought a system, not that you control the water. Keep the calibration records too, because an uncalibrated sensor reading is an assertion, not a measurement.
Which points should the devices actually monitor?
Pick the points that tell you whether the whole system holds, not just the convenient ones. The usual priorities are the sentinel outlets (the first and last on each loop), hot-water storage and return temperatures, the cold-water storage tank, and any outlet that is hard to reach or rarely used. Common UK benchmark figures are cold water kept below 20 degrees C where reasonably practicable, hot water stored at around 60 degrees C, and hot water hot enough to reach roughly 50 degrees C at the outlet within about a minute, with 55 degrees C applied in healthcare premises, though the exact targets and how they are checked come from your risk assessment rather than a device’s factory default [3][2]. Instrument the awkward outlets first; the busy ones flush themselves through constant use.
Do smart devices remove the need for a responsible person or risk assessment?
No, and any supplier implying otherwise is overselling. The duty sits with the duty holder and the responsible person; software and hardware are tools they use, not replacements for them [4]. The risk assessment still defines what to control, what counts as acceptable, and what to do when it is not, and the smart valve simply feeds that decision faster. Delegating to a clever device is the same trap as delegating to a contractor: you can hand over the task, never the accountability.
Before you trust the numbers
A reading is only as honest as the sensor behind it and the person reading it. Devices drift, batteries die, and a probe in the wrong spot can report comfort while a nearby dead leg sits warm and forgotten. Treat every figure as evidence to be checked against a competent, site-specific assessment, not as proof on its own. Sampling for Legionella can support verification or investigation, but it follows the system and the risk assessment and does not certify a clean bill of health from one good reading [5]. The targets, the monitoring frequency and the remedial actions are yours to set for your building, not the manufacturer’s to set for you.
If you change one thing this week
Walk one wing or one plant area and list every device that touches temperature: every TMV, sensor valve and automatic flusher. Against each, write four things down: what it measures, who reads it, the value that triggers action, and the date it was last calibrated or serviced. Any device where you cannot answer all four is decoration, not control, and that is where your attention belongs first. If you are moving toward automating this across the estate, Using BMS for automated temperature control covers tying these devices into a building management system.
Sources
[1] HSE, “Legionnaires’ disease. The control of legionella bacteria in water systems - Approved Code of Practice and guidance (L8)”. https://www.hse.gov.uk/pubns/books/l8.htm [2] HSE, “Legionnaires’ disease: Technical guidance (HSG274)”. https://www.hse.gov.uk/pubns/books/hsg274.htm [3] HSE, “Hot and cold water systems”. https://www.hse.gov.uk/legionnaires/hot-and-cold.htm [4] HSE, “Legionnaires’ disease - what you must do”. https://www.hse.gov.uk/legionnaires/what-you-must-do/index.htm [5] HSE, “Testing and monitoring your water system for legionella”. https://www.hse.gov.uk/legionnaires/testing-monitoring-water-system.htm