Cooling towers: the 2,184 hours between legionella results
TL;DR. HSG274 Part 1 asks you to sample a cooling tower for legionella at least quarterly. That leaves roughly 2,184 hours between results. The weekly checks in between are dip slides and hand-held readings, so for most of the year the safety case rests on proxies taken once every 168 hours. Continuous instrumentation does not replace the lab sample, but it does close the gap between them.
Key takeaways
- HSG274 Part 1, published by HSE in 2024, sets the control regime for evaporative cooling systems.
- Legionella sampling is required at least quarterly, which is about 2,184 operating hours between results.
- Microbial activity should be no greater than 1×10⁴ cfu/ml, and legionella not detected or no greater than 100 cfu/l.
- Conductivity, pH, oxidising biocide and microbial activity are all weekly checks, or 168 hours apart.
- Legionella multiplies between 20°C and 45°C, which is the normal operating range of a cooling tower.
- Every notifiable device must be notified to the local authority under the 1992 Regulations, with changes reported within one month.
Table of Contents
On this page
Last updated: 24 August 2026
What does the law require for a cooling tower?
Two duties sit on every evaporative cooling system in the UK. The first is registration. Under the Notification of Cooling Towers and Evaporative Condensers Regulations 1992, whoever controls the premises must ensure no notifiable device sits on them unless the local authority has been notified in writing, and any change to that information has to be reported within one month.
The second is control. HSE’s technical guidance, HSG274 Part 1, published in 2024, gives practical advice on complying with the Health and Safety at Work etc Act 1974, COSHH 2002 as amended, and the Management of Health and Safety at Work Regulations 1999. It requires a named responsible person to take day-to-day responsibility, and it sets out exactly what has to be checked on a cooling tower and how often. HSE’s own cooling towers guidance page points dutyholders straight at it.
Neither duty is unusual. What is unusual about a cooling tower, compared with almost any other water system on a site, is the size of the gap between what the guidance asks you to measure and how often anyone actually measures it.
How often should a cooling tower be tested?
HSG274 Part 1 sets a layered schedule. Daily: check the system is operating as described in the operations manual, and make a visual check of the cleanliness of the water. Weekly: check the physical condition of the system and undertake dip slide analysis or equivalent testing. Every one to three months according to risk: calibrate the conductivity sensor and check the blow-down function.
The water chemistry table is where the frequencies bite. For an evaporative cooling system the guidance lists conductivity, pH, oxidising biocide and microbial activity as weekly checks. Legionella analysis is quarterly. The guidance is explicit that the effectiveness of the biocide regime should be monitored weekly, conventionally by microbial dip slides, and that specific sampling for legionella should be done on at least a quarterly basis.
Convert those cooling tower intervals into operating hours and the picture changes. A daily check is a 24-hour gap. A weekly check is 168 hours. A quarterly legionella sample is roughly 2,184 hours. The tower does not pause between them. It runs continuously, in weather that changes, on make-up water whose quality changes, with a load that changes.
What are the action levels for a cooling tower?
HSG274 Part 1 gives two microbiological targets. Microbial activity, estimated by dip slides or total viable counts, should be not greater than 1×10⁴ cfu/ml. Legionella should be not detected, or not greater than 100 cfu/l. Where risk is elevated, the guidance points to more frequent microbial monitoring, for example monthly legionella sampling rather than quarterly.
The temperature context matters as much as the counts. HSE’s guidance is built around the fact that legionella multiplies between 20°C and 45°C. A cooling tower spends its working life inside that band by design, because rejecting heat into ambient air is the whole point of the machine. You are not trying to keep the water out of the growth range. You are relying entirely on treatment to stop growth happening inside it.
Why is weekly monitoring a problem for a cooling tower?
Because the cooling tower parameters checked weekly are the ones that move fastest. Conductivity climbs as water evaporates and dissolved solids concentrate, and it drops the moment blow-down operates. Biocide is consumed continuously by the load it is controlling. pH shifts with make-up water and with the treatment programme. A reading taken on Tuesday morning describes Tuesday morning.
The failure mode this creates is not a dramatic one. It is a biocide dose that ran out on Thursday, a blow-down valve that stuck closed on Friday, and a conductivity that drifted for four days before anyone looked. By the time the next dip slide is read, the excursion is over and the water looks normal again. Nothing on the record shows it happened, so nothing gets fixed, and it happens again the following month.
HSG274 Part 1 anticipates part of this. It asks for conductivity sensor calibration and a blow-down function check every one to three months according to risk, which tells you the guidance already expects a conductivity sensor to be there and doing something. The question is whether anyone is reading it between visits.
What can continuous monitoring actually cover?
Be precise about this, because overclaiming on cooling tower instrumentation is both wrong and dangerous. Continuous instrumentation cannot replace legionella sampling. Detecting legionella requires culture or an equivalent laboratory method on a physical sample, and no online sensor does that. The quarterly sample stays exactly where the guidance puts it.
What continuous instrumentation does cover is every proxy on the weekly list. Conductivity, pH, oxidising biocide residual and temperature are all measurable online, continuously, with alarms on thresholds. That turns a 168-hour gap into a live trace, and it turns the weekly visit from a data-gathering exercise into a verification exercise.
The full AQUAIOT cooling tower capability stack
Multi-parameter water quality on the circuit
The iSPA-T Multi-Parameter Water Quality Online Monitoring System covers several of the weekly-check parameters from one installation, which suits a plant room where the tower, the chiller circuit and the make-up feed all need watching. Where a compact inline instrument fits the space better, the iSPS-X Multi-Parameter Water Quality Sensor does the same job in a smaller footprint.
Oxidising biocide residual
Where the treatment programme is chlorine-based, the SPA-Chlorine Free/Total Chlorine Water Quality Analyzer reads free and total chlorine continuously rather than as a spot check. That is the single most useful continuous number on a cooling tower, because biocide residual is the control measure and everything else is context for it.
Records that survive an inspection
All of it reports over encrypted cellular and LoRaWAN telemetry into the AQUAIOT Cloud, with trendlines, threshold alarms, exports and role-based access, plus RS485 and Modbus on site and open APIs into BMS, SCADA and CAFM. HSG274 puts record-keeping on the responsible person, and a continuous trace is a better record than a column of handwritten weekly figures.
Where should you start on a cooling tower?
- Confirm the device is notified to the local authority and that the record is current.
- Instrument conductivity and biocide residual first, because those two drive the blow-down and dosing decisions.
- Add temperature so you can see how long the circuit sits in the 20°C to 45°C band under real load.
- Set alarm thresholds below the HSG274 action levels rather than at them, so an alert arrives before an exceedance.
- Keep the quarterly legionella sample exactly as it is, and use the continuous trace to explain any result you do not like.
That last point is the commercial argument. When a quarterly sample comes back high, the first question is always what changed. Without a continuous record nobody can answer it, so the response is a full clean and disinfect and a shrug. With one, you can usually point at the week the biocide residual dropped.
Frequently asked questions
How often must a cooling tower be sampled for legionella?
HSG274 Part 1 states that specific sampling for legionella should be done on at least a quarterly basis. Where risk is elevated the guidance points to more frequent microbial monitoring, giving monthly legionella sampling as its example. The action level is not detected, or not greater than 100 cfu/l.
Do I have to register a cooling tower with the council?
Yes. The Notification of Cooling Towers and Evaporative Condensers Regulations 1992 place a duty on whoever controls the premises to ensure no notifiable device is situated there unless the required information has been notified in writing to the local authority. Changes to that information must be notified within one month, and cessation as soon as is reasonably practicable.
Can online sensors replace legionella sampling?
No. Legionella detection needs a laboratory method on a physical sample, and no continuous sensor performs it. What online instruments replace is the weekly manual reading of conductivity, pH, biocide residual and temperature. They sit alongside the statutory sampling regime rather than in place of any part of it.
What is the microbial action level for a cooling tower?
HSG274 Part 1 sets microbial activity, as estimated by dip slides or total viable counts, at not greater than 1×10⁴ cfu/ml. That is the general bacterial measure and it is checked weekly. Legionella is a separate test with its own action level of not detected or not greater than 100 cfu/l, sampled at least quarterly.
Where this leaves a responsible person
A cooling tower is the one water system on most sites where the control regime is already prescriptive, already written down and already someone’s named responsibility. The weakness is not the scheme. It is that the scheme is executed by hand at intervals measured in days and months, against a process that changes by the hour.
Closing that gap does not require rewriting the risk assessment. It requires putting instruments on the parameters the guidance already lists. Speak to an AQUAIOT engineer about continuous water quality monitoring on your evaporative cooling systems, or read how the same stack applies to Legionella monitoring across the wider estate.
By GP, Digital and IoT lead at AQUAIOT.