Most cooling-tower schemes of control fail not because a parameter is wrong but because the document is a list of good intentions nobody can act on at the tower. The fix is a single page that says, for each control, what you measure, the range you want, how often, who does it, and what happens when it drifts. Here is one, built around a composite site.
A scheme of control is the practical half of your duty. The risk assessment finds the hazards; the scheme says how you keep them in check day to day [1]. For an evaporative system the law expects that written scheme to exist and to be followed, reviewed and recorded [2].
The site
The following is an illustrative composite, not a real named installation. Picture a mid-size office on the edge of a city: roughly 4,000 square metres over six floors, comfort cooling served by a single open evaporative cooling tower on the roof feeding the chillers. The tower runs hard from late spring to early autumn and ticks over in milder months. A water-treatment contractor visits monthly; an in-house facilities technician does the weekly checks and holds the keys to the plant room.
When a new responsible person took over, she found a folder with a generic scheme-of-control template, a contractor’s dosing report, and three years of dip-slide photos that nobody had ever compared to a target. Everything existed. Nothing connected. Her first job was to turn that pile into one page she could read in the plant room with a clipboard.
What she had to pull together
The chemistry and the microbiology are not separate problems. As water evaporates, dissolved solids concentrate, scale and corrosion potential rise, and biocide gets harder to hold, so the controls below all lean on each other. The conductivity and bleed settings sit upstream of how clean the water stays, which sits upstream of the Legionella result. A scheme that lists them as unrelated rows misses the point.
She grouped the controls into chemistry (keeping the water in condition), microbiology (keeping the bug count down and proving it), and physical (keeping the tower and its barriers sound). Then she gave every row an owner and an action, because a target with no named person and no escalation is decoration.
The worked scheme of control
Treat the table as the shape to copy, not the numbers to copy. Every target, band and frequency below is illustrative and must be set by your own risk assessment, your water treatment specialist and the current guidance, the figures here stand in for “a sensible-looking entry”, nothing more.
| Control | What you measure | Target (illustrative, set by your RA) | Frequency | Owner | Action if out of range |
|---|---|---|---|---|---|
| Biocide residual | Oxidising/non-oxidising biocide level | A residual maintained per the treatment programme | Weekly on-site; dosing checked at contractor visit | FM technician / contractor | Re-dose, check dosing kit, log; investigate repeat lows |
| Conductivity & cycles | Conductivity (proxy for dissolved solids) | A set point that holds the chosen cycles of concentration | Continuous controller + weekly manual check | Contractor sets; technician verifies | Check bleed valve and controller; correct cycles |
| Scale/corrosion inhibitor | Inhibitor reserve | Within the programme’s reserve range | Monthly at service visit | Contractor | Adjust dose; review make-up water quality |
| Dip-slide aerobic count (TVC) | General microbial cleanliness | Below the programme’s TVC threshold | Weekly | FM technician | Increase biocide, investigate; trigger review if rising |
| Legionella sample | Legionella cfu/l, accredited lab | Below the lower commonly cited band | Quarterly, or per the RA | Contractor takes; RP reviews | Tiered response by result band (below) |
| Clean & disinfect | Whole tower condition | Done and certificated | Before commissioning; scheduled by fouling/risk assessment; after shutdown or disruptive work | Contractor | Schedule out-of-cycle clean if fouled |
| Drift eliminators & physical state | Condition, fouling, deposits | Sound, in place, clean | At every service visit and clean | Technician / contractor | Repair or replace; record |
| Make-up & general inspection | Float valve, basin, packing, sump | Working, no debris or biofilm | Weekly walk-round | FM technician | Clear, repair, escalate |
The Legionella row needs its own response ladder because the sample is the control that actually triggers escalation. A result below the lower band typically means continue and review; the commonly cited middle band (in the order of 100–1,000 cfu/l) usually means review the scheme and resample; above the upper band (in the order of more than 1,000 cfu/l) means immediate action: resample, shot-dose or increase the biocide, reassess the whole control programme, take corrective action, and verify the result [2]. Confirm those bands against current guidance rather than treating these as fixed legal numbers.
What changed once it was on one page
Three things surfaced the moment the controls sat side by side. The dip-slide counts had been creeping up for two summers, but because nobody held them against a threshold, the trend was invisible until it sat in a column with a target. The scheduled comfort-cooling clean had quietly become “once, when we remembered”. And the Legionella quarterly sample owner was blank, the contractor assumed the client booked it; the client assumed the contractor did. The classic split-responsibility gap.
The transferable lesson is not the numbers. It is that a scheme of control earns its keep only when every row names a measurable target, a frequency, a person and an action. Strip any one of those and you get the folder she inherited: full of activity, empty of control.
Before you copy anyone’s scheme
This worked example is a teaching device, not a template to drop onto your tower. Your parameters, targets, frequencies and action levels must come from a site-specific risk assessment carried out by a competent person, informed by your water-treatment specialist and the current published guidance [1][3]. An evaporative system on a hospital roof, a tower on hard make-up water, or one feeding a process load will each justify different entries. Nothing here is legal advice, and no figure in the table should be adopted without that assessment behind it.
FAQ
What must a cooling tower written scheme of control actually contain?
At minimum: the controls you rely on, the target or condition for each, how and how often you check it, who is responsible, and what you do when a reading is out of range, plus how the whole scheme is reviewed and recorded. For an evaporative system that means biocide, conductivity and cycles, inhibitor, microbial monitoring, Legionella sampling, cleaning and physical condition all sitting in one auditable document [2].
How is the scheme of control different from the risk assessment?
The risk assessment identifies the hazards and decides what control is needed; the scheme of control is the operational document that delivers it, the parameters, frequencies, owners and actions [1]. The assessment answers “what could harm people and how serious is it”; the scheme answers “what we do every week to stop it”. One feeds the other, and both must be kept current [3].
Who owns each row, the responsible person or the water-treatment contractor?
Usually a mix, and the scheme should say so explicitly. A contractor may dose chemicals, set the controller and take samples, while in-house staff do weekly checks and walk-rounds. The legal duty to have and follow the scheme stays with the duty holder regardless. Blank or assumed ownership, especially on sampling and cleaning, is where control quietly lapses [4].
How often should the scheme itself be reviewed?
Review it periodically and whenever something changes the risk: plant alterations, a Legionella result in an action band, a change of contractor, recurring out-of-range readings, or a shutdown and restart. The schedule is a living control document, not a once-a-year signature [2].
What to do this week
Take your own tower’s paperwork and try to fill one row of the table above, say, the dip-slide line, with all five columns complete: target, frequency, owner, action, and where the last three results sit against the target. If you cannot complete a single row from the documents you hold, that is the gap to close first, before the next cooling season loads the tower up.
Sources
- BSI, BS 8580-1:2019, clause 10 (risk review and reassessment). https://www.bsigroup.com/
- HSE, HSG274 Part 1 (2024), “Evaporative cooling systems: types, design and operation”, p.9. https://www.hse.gov.uk/pubns/books/hsg274.htm
- HSE, ACoP L8 (2013), “Preventing or controlling the risk from exposure to legionella bacteria”, p.16. https://www.hse.gov.uk/pubns/books/l8.htm
- Legionella Control Association, “Code of Conduct for Service Providers”. https://www.legionellacontrolassociation.co.uk/