Cooling Tower Maintenance Checklist

Most cooling tower checklists circulating in facilities teams are somebody else's schedule, copied from a manual or a slide, with the intervals kept and the reasoning lost. That is a poor way to run any asset, and on a cooling tower it is worse than poor, because a tower is not only a mechanical asset. It moves large volumes of warm water through open air and releases a fine aerosol while doing it, which makes its water quality a public health matter and not merely an efficiency one. So this article deliberately does not hand you a schedule. It hands you the task areas, the observations that matter inside each one, and the reasoning, so that you can hold an informed conversation with the people who are actually responsible for setting your intervals.

The message up front: water safety for evaporative cooling systems is legally regulated in most jurisdictions. A written risk assessment and a documented scheme of control prepared by a competent person is normally required, and the specific duties, tests, thresholds and intervals come from your local law and from that person, not from an article. Treat everything below as an orientation checklist to structure the conversation with your water safety adviser and your maintenance provider. It is not a substitute for either, and following it does not make you compliant with anything.

1. What this checklist is, and who actually owns the regime

There are two overlapping regimes on a cooling tower and confusing them is the single most common structural mistake I see in maintenance planning. The first is the mechanical maintenance regime: the fan, the drive, the motor, the structure, the water distribution, the basin. That one you own, and it looks like any other rotating plant asset in a hostile environment. The second is the water safety regime: the treatment programme, the microbiological control, the cleaning and disinfection of the wetted surfaces, the registration or notification of the system where that applies, and the records that prove all of it. That one is normally led by a named competent person working to a written scheme, and in most jurisdictions it carries legal weight that a PM checklist does not.

The two regimes share hardware, which is why they get merged and then quietly degrade. A drift eliminator is a water efficiency item to a mechanical planner and an aerosol control to a water safety adviser. A basin clean is a fouling task to one and a disinfection event to the other. If your schedule was written by the mechanical side alone, the water safety items tend to be present but thin; written by the treatment provider alone, the mechanical items tend to be generic. A good scheme is written jointly and states, for every task, which regime it belongs to and who signs it off.

If you are new to the equipment itself, read the mechanism first: a checklist only makes sense once you can picture what the water and the air are doing. See what a cooling tower is and how it works, and, for the machine on the other side of the condenser loop, what a chiller is and how chillers work.

2. General condition and structure

What you are checking. The casing, frame and supports, fasteners, internal and external coatings, louvres, access doors and panels, the fan deck or plenum floor, and the surrounding area. On a field erected tower this extends to the structure, the cladding and the internal supports for the fill and the distribution system.

Good looks like panels square and fully fastened, coatings intact with no exposed substrate, louvres complete and clear with no daylight through gaps that should not exist, doors closing on sound seals, no standing water outside the basin, and no vegetation, nesting or external contamination sources discharging into the air intake. Bad looks like rust streaks and blistering at water lines and fixings, missing panels or louvre sections, sagging internal supports, cracked concrete at anchor points, and persistent damp on a structure that should be dry.

Why it matters. Structural condition degrades on two clocks at once, corrosion and biological growth, and they accelerate each other. A failed coating gives biofilm a rough surface to colonise. A missing louvre admits light, which encourages algae, and debris, which feeds the nutrient load the treatment programme has to fight. The mechanical and hygiene consequences are not separable, and that is the first idea to internalise about this asset.

3. Fan, drive, belts, gearbox and vibration

What you are checking. Fan blades and hub, blade pitch and tracking, blade tip clearance to the shroud or fan ring, the drive arrangement (direct, belt or gear drive), belt condition and tension, sheave wear and alignment, drive shaft and couplings, the gearbox where fitted including oil level, oil condition and breather, bearing condition and lubrication, mounting integrity, guards, and vibration behaviour including any vibration cut-out or switch that is fitted.

Good looks like blades clean, uncracked, at consistent pitch, tracking within tolerance and clear of the shroud; belts evenly tensioned with no glazing, cracking or fraying and sheaves free of groove wear or misalignment; gearbox oil at level, clear and of the correct grade with a clean breather and no seal leaks; bearings quiet and at stable temperature; and vibration readings that are stable and repeatable against a baseline taken when the machine was known to be healthy. Bad looks like blade cracks at the root, inconsistent pitch, scale or fouling deposits on blades that unbalance the fan and cost airflow at the same time, belt dust under the drive, milky gearbox oil pointing to water ingress through a seal or breather, a rising vibration trend at constant duty, and a vibration switch that has been strapped, bypassed or set so wide it can never trip.

Why it matters. A tower fan is a large, imbalance-prone rotating assembly working in a warm, wet, chemically treated and often corrosive environment, frequently at height, and often with a gearbox nobody looks at because it sits above the fill. Blade liberation is a serious failure with a real safety consequence, and airflow loss is a thermal loss on every hour the plant runs. The good news is that this is the most conventional part of the asset: standard rotating equipment practice applies, and baseline-referenced vibration trending is the most informative single mechanical observation you can take on a tower. The general framework sits in preventive maintenance for HVAC systems.

4. Motor and electrical

What you are checking. Motor condition and mounting, cooling fins and enclosure cleanliness, terminal box and gland integrity, cable condition and support, local isolator condition and identification, starter or variable speed drive condition and enclosure sealing, motor current at a recorded duty, winding and bearing temperature where instrumented, insulation resistance testing where that is part of your electrical regime, earthing and bonding, and any anti-condensation heater fitted.

Good looks like a dry, clean, sealed terminal box with intact glands, cables supported and not chafing, motor current stable at a given speed and ambient condition, enclosures rated and sealed for a wet outdoor location with the gaskets actually present, and isolators labelled unambiguously to the tower and fan they serve. Bad looks like water tracking inside a terminal box, corroded glands and rusted fixings, temporary cable repairs, a current draw that has crept up at the same duty, and unlabelled or shared isolators.

Why it matters. Motors on towers fail for environmental reasons far more often than electrical ones. Moisture ingress, corrosion and condensation do the damage, and the fixes are unglamorous: sealing, supporting, draining, labelling. Motor current is also one of the cheapest condition indicators available to you, because a fouling fan or a binding drive shows up in the current trend before it shows up as a breakdown. Unclear isolator labelling is a safety issue before it is a reliability one, for reasons that become obvious in section fifteen.

5. Fill and drift eliminators

What you are checking. Fill condition, fouling, scaling, biological growth, collapse, distortion, displacement out of its supports, and gaps or bypass paths where air can travel without meeting water. Drift eliminators for completeness, condition, correct orientation and proper seating, with no missing sections and no gaps at the edges where they meet the casing.

Good looks like fill sitting square, fully supported and uniformly wetted across its face with open passages and no deposits, and drift eliminators complete across the full discharge area, tightly seated and correctly oriented with no daylight gaps at joints or edges. Bad looks like hard scale bridging the fill passages, slime or algal growth, sections collapsed under their own fouled weight, or fill lifted and shifted so part of the face is dry and part flooded; and, on the eliminators, sections missing entirely, installed the wrong way round, crushed by somebody standing on them, or edge gaps never closed after a previous repair.

Why it matters. Fill is where the heat transfer happens, so fouled or bypassed fill is a direct and continuous performance loss, and it is also a large sheltered surface area for biological growth. Drift eliminators are the more serious of the two in safety terms. They are the engineered control that stops water droplets leaving the tower with the air stream. A degraded eliminator does not merely waste water and treatment chemical, it releases aerosol from the circulating water into the surrounding environment, which is precisely the exposure route the whole water safety regime exists to control. Treat eliminator condition as a hygiene control that happens to live in the mechanical checklist, and never let it be quietly deferred as a water saving item.

6. Water distribution and nozzles

What you are checking. The hot water basin or distribution header, spray nozzles or orifices, splash plates, distribution laterals and their supports, the pattern and evenness of water across the whole fill face, and the strainers or screens that protect the nozzles.

Good looks like even, full-face wetting with every nozzle flowing to its designed pattern, a distribution basin at consistent level with no dry corners and no overtopping, laterals level and properly supported, and nozzles present, of the correct type and clear. Bad looks like blocked nozzles producing dry patches, missing nozzles producing a jet that floods one area and channels straight through, a distribution basin out of level so one end floods while the other starves, nozzles replaced over the years with whatever was in the store so one face carries a mixture of patterns, and scale rings around orifices.

Why it matters. Uneven distribution destroys tower capacity disproportionately. A dry patch of fill is not just an unused patch, it is a low resistance path that air preferentially takes, so you lose more than the area suggests. Dry and intermittently wetted surfaces are also where scale and biofilm establish most readily. Distribution is the highest value, lowest cost inspection on the whole asset: it needs a look, not an instrument.

7. Basin, strainer and sump cleanliness

What you are checking. Cold water basin or sump cleanliness, silt and debris accumulation, sediment depth, biofilm and slime on wetted surfaces, corrosion and coating condition below the water line, the suction strainer and any anti-vortex arrangement, the drain and overflow, and the condition of the basin sweeper or filtration system where one is fitted.

Good looks like a basin you can see the bottom of, with no sediment bed, no slime, no scale sheets and intact coating; a clean strainer with the correct mesh in place; a drain that actually drains and a clear overflow at the right level; and, where a side stream filter or basin sweeper is installed, evidence that it is running and being maintained rather than isolated years ago and forgotten. Bad looks like a sediment bed, which is the classic finding, plus a holed strainer, a strainer with the wrong mesh, a strainer left out entirely after a previous clean, algal growth where light reaches the water, and a filtration system whose isolation valves have been shut. Sediment is nutrient, it shelters growth from the treatment chemistry, it holds biofilm against the surface, and it puts abrasive solids into the condenser loop and the pumps.

Why it matters. Basin cleanliness is the item where the mechanical and hygiene regimes are least separable. Sediment defeats water treatment more effectively than almost anything else, because chemistry works on the bulk water and sediment protects what is under it. That is why cleaning and disinfection of wetted surfaces sits in the water safety scheme rather than in the housekeeping schedule, and why the frequency and method belong to the competent person and not to whoever happens to be free.

8. Make-up, float control, bleed and conductivity

What you are checking. Make-up water supply, the make-up valve and its float or level control, the float or probe condition and setting, backflow protection on the make-up line, the bleed or blowdown arrangement including its valve and any solenoid, the conductivity controller and its sensor, and the relationship between make-up volume and bleed volume where those are metered.

Good looks like a level held steady at the intended point with the make-up valve seating fully closed and not weeping, a float mechanism moving freely and neither scaled up nor propped, backflow protection present and intact on the make-up connection, a bleed line actually open and flowing when the controller calls for it, a clean conductivity sensor calibrated within your regime and agreeing with an independent check, and both make-up and bleed metered so the water balance can be reconciled rather than assumed. Bad looks like a bleed valve shut to save water or because a nuisance alarm annoyed somebody, a scaled probe reading low so the controller thinks the water is cleaner than it is and never bleeds, a float stuck or tied up so the tower runs permanently high or low, a continuously overflowing basin sending treatment chemical straight to drain, and a make-up line with no backflow protection, which is a drinking water contamination risk and not a plumbing detail.

Why it matters. Bleed and make-up together set the concentration of dissolved solids in the circulating water, which is the boundary condition for everything the treatment programme is trying to do. Get that wrong and you either scale the system up or you waste water and chemical. Both failure directions are common, and both are usually caused by a well-meaning intervention rather than by a broken component, which is why I check valve positions and probe cleanliness before I check anything else when a tower's water chemistry has drifted.

9. Water treatment and the dosing system

What you are checking. That a written treatment regime exists, defined by the competent person as part of the scheme of control, and that the hardware can deliver it: chemical storage and containment, stock and shelf life, dosing pumps, injection points and non-return arrangements, dosing lines, controller settings against the specified regime, sampling points, and the completeness of the monitoring records the scheme requires.

Good looks like a regime you can read, owned by a named person, with the monitoring it requires actually being done and recorded; dosing equipment operating and calibrated to that regime, with intact containment under chemical storage; accessible and representative sample points; and records showing both the results and the actions taken when results fell outside the specified range. Deviations closed out, not just logged. Bad looks like an empty chemical drum with the pump still cycling, cracked or disconnected dosing lines, a controller in manual with nobody able to say why or when, records with long unexplained gaps or records that are complete but show out of range results with no corresponding action, and chemical stored without containment or incompatible chemicals stored together. The version I see most often: a treatment contract being paid for, and a site that cannot produce the scheme document the contract is supposed to be delivering against.

Why it matters. Water treatment on an evaporative system is a specialist function, not a task a maintenance team improvises. It controls scale, corrosion and biological growth simultaneously, in a system whose chemistry changes with load, weather, water source and cycles of concentration, and the biological part carries a public health consequence rather than an efficiency one. Your job on the maintenance side is not to design the regime: it is to verify that a regime exists, that the hardware can deliver it, that the monitoring is happening, and that deviations get acted on. That is a bounded role, and it is where a maintenance function adds most value here.

10. Drift and plume observation

What you are checking. Visible carryover of water droplets from the discharge, the size and behaviour of the plume, where the discharge actually goes in the prevailing wind, and whether anything sensitive sits downwind or nearby: fresh air intakes, windows, walkways, occupied balconies, pedestrian routes, or another tower's air intake.

Good looks like a discharge with no visible droplet carryover, condensate plume behaviour that is understood and expected for the weather, nothing sensitive in the discharge path, and a documented understanding of where that path runs in different wind conditions. Bad looks like visible droplets leaving the tower, wet patches or staining or scale deposits on surfaces downwind, discharge blowing across a walkway or into an air intake, and two towers arranged so one discharges into the other's intake, or a tower short circuiting into its own intake, which also wrecks its thermal performance.

Why it matters. Drift is the exposure pathway. Everything else in the water safety regime is an attempt to make the circulating water safe, and drift observation is the check on how much of that water is leaving the enclosure and where it goes. It costs nothing, it needs no instrument, and it is routinely skipped because it is not on the form. Walk around the tower at ground level and look at the surfaces downwind. Scale streaks on a nearby wall tell you something the log sheets will not.

11. Instrumentation, sensing and controls

What you are checking. Temperature sensing on the water entering and leaving the tower, conductivity sensing, level sensing, flow indication or metering, any vibration switch, fan speed control and its response, the control sequence including staging of multiple cells and fan speed against demand, alarms and their routing, and the relationship between what the local instruments read and what the building management system displays.

Good looks like local and remote readings agreeing with each other and with an independent instrument, sensors clean and properly immersed in the flow and calibrated within your regime, a control sequence matching the documented design intent with fans modulating and cells staging as intended rather than forced, and alarms reaching a human who is expected to act on them with a record of what was done. Bad looks like a temperature sensor out of the flow for years so the trend everybody trusts is measuring nothing useful, a conductivity probe that has never been cleaned, points in manual or overridden with no note of who did it or why, alarms disabled after a nuisance period, and a set point changed on the floor to stop a complaint with the design intent nowhere in sight.

Why it matters. Almost every conclusion you draw about a tower rests on instrumentation you did not verify. If the leaving water temperature sensor is wrong, your performance assessment is wrong. If the conductivity probe is fouled, your water balance is wrong. Instrument verification is not an add-on to the checklist, it is what makes the rest of the checklist mean anything. For how this data becomes usable analysis on the chiller side of the loop, see chiller plant analytics.

12. The task areas in one table

This is the summary I would put in front of a maintenance provider or an in-house team as the frame for a scope discussion. Note deliberately that there is no frequency column. Frequencies are addressed separately below, and for a reason.

13. Common findings and the likely cause

Fault finding on a tower is mostly pattern recognition, and the patterns repeat across sites and climates. This table is the diagnostic shortcut I use, and it is a starting hypothesis rather than an answer.

14. Where the frequencies actually come from

You will have noticed that neither table gives an interval, and that is deliberate rather than evasive. On the water safety side the intervals are set by your jurisdiction's requirements and by the risk assessment and scheme of control prepared for your specific system. Publishing a number here would invite somebody to adopt it, and an interval taken from an article is not a defence anywhere. On the mechanical side the honest answer is that there is no single correct interval either, because the drivers genuinely vary:

• System design. Induced or forced draught, counterflow or crossflow, packaged or field erected, gear or belt or direct drive, open tower or closed circuit cooler. These change what wears and how fast.
• Make-up water quality. The chemistry you start from sets scaling and corrosion tendency, and therefore how quickly wetted surfaces and probes foul.
• Environment. Dust, sand, industrial deposition, coastal salt, pollen and nearby construction all load the tower, and a site with heavy airborne loading fouls on a completely different clock from a clean one.
• Duty and seasonality. A tower running continuously at high load is a different asset from one that runs seasonally, and systems that sit idle for part of the year bring considerations that belong in the scheme rather than the routine.
• Criticality and jurisdiction. A single tower with no spare cell justifies closer attention than one cell of several, and the regulated elements differ by country and sometimes by municipality.

The practical method I would recommend: start from the manufacturer's recommendations for the mechanical items and from the scheme of control for the water safety items, then let your own findings adjust the routine. If a check finds nothing over several cycles, that is evidence to discuss extending it. If it repeatedly finds a problem, tighten it or fix the underlying cause. That loop requires that you are recording findings rather than ticking boxes, and the measurement discipline behind it is covered in preventive maintenance KPIs and schedule compliance. One caution: the mechanical side of that loop is yours to optimise and the regulated water safety side is not. Changing a scheme-mandated activity is a conversation with the competent person, not a planning decision.

15. The safety of the maintenance work itself

This is the part of cooling tower maintenance that checklists handle worst. The checklist tells you to inspect the fill, and says nothing about the fact that inspecting the fill may mean climbing a corroded ladder onto a platform above a live fan to open a panel into a wet enclosure. Four principles, each with a sibling article that goes into the detail properly.

Isolate and lock off before any fan or basin work. A tower fan can rotate under airflow when it is not powered, and it can be started remotely by a control sequence or by somebody at a panel who has no idea anyone is inside. Before anybody reaches into the plenum, opens a fan section or works in the basin, the energy sources need to be identified, isolated, locked, tagged and verified dead at the point of work, with the fan physically restrained where windmilling is possible. Identified, not assumed: the mislabelled isolator becomes a serious matter at exactly this moment. See the lockout tagout guide.

Treat entry as a confined space question, not an assumption. A tower cell, plenum, large basin or sump can meet the characteristics that define a confined space: restricted entry and exit, an enclosure not designed for occupancy, and the potential for a hazardous atmosphere, engulfment or entrapment. Whether a given section qualifies is answered by assessment of that specific structure, not by habit, and where it does qualify the controls that follow are substantial. The failure mode is treating a large accessible basin as ordinary work because it does not look like a tank. See confined space hazards, requirements and procedures.

Working at height is the normal condition here, not the exception. Fan decks, platforms, ladders, walkways over open basins and roof mounted towers are routine on this asset, all in the same wet and corrosive environment that eats the structure. Access provision is itself something to inspect: platform and grating condition, handrail integrity, ladder fixings, anchor points, hatch covers, and whether internal walking surfaces will carry a person. Standing on fill or eliminator media damages both and is not a safe working surface. See working at height hazards, controls and requirements.

Control exposure when cleaning. Cleaning a tower disturbs exactly the material the water safety regime exists to contain, and any method generating spray or dust raises exposure for whoever is doing it. So the method is a controlled decision: who does it, what method, what respiratory and other protection, who else is excluded from the area, what happens to the fan and the system during the work, and how the waste water is handled. Specify it in advance with your water safety adviser as part of the scheme rather than deciding on the day with whatever is in the van. Chemical handling during dosing work sits in the same category, governed by the safety data for the specific products in use.

16. Records and documentation

On most assets, records are the administrative residue of the work. On a regulated water system they are part of the duty itself. The distinction is real and it changes how you should treat the documentation.

What a complete record set generally covers: the risk assessment and the written scheme of control with its revision history and review evidence, the identity and competence basis of the responsible and competent persons, asset records with make, model, capacity, materials and drawings, the monitoring results the scheme requires, cleaning and disinfection records including method and who carried it out, the mechanical maintenance history, corrective work raised with evidence it was closed, and deviations with the action taken. The item most often missing is that last one. Plenty of sites can show you results; far fewer can show what happened when a result fell outside its range.

This is the one place software genuinely belongs in this article. A CMMS is a reasonable home for the asset record, the mechanical schedule, the findings from each visit, the corrective work raised from those findings, and the retention of all of it somewhere searchable, which matters when somebody asks for two years of history at short notice. Any competent system does this; what makes the difference is whether findings and actions are captured as structured data rather than free text in a closing comment. What software cannot do is decide what your scheme requires. It stores the regime, it does not author it. For the plant room context this tower sits in, see boiler and chiller preventive maintenance.

On jurisdiction, one plain statement. Requirements for evaporative cooling systems are national or local and they differ in substance rather than only wording: whether the system must be registered, who may hold the responsible person role, the required form and review cycle of the risk assessment, the monitoring regime, the action thresholds and the retention period. Regulators publish their own requirements directly, for example the UK Health and Safety Executive for Great Britain and US OSHA for the United States, and neither has legal force outside its own jurisdiction. If you operate in the Gulf, the binding requirements are the federal and emirate or national ones plus the local municipality, and an interval lifted from a British or American document is a benchmark at best.

The idea to walk away with

A cooling tower rewards being understood as four assets sharing one enclosure: a structure that corrodes, a rotating machine that wears, a heat exchanger that fouls, and a regulated water system that carries a health consequence. A checklist that only sees the rotating machine will keep the fan running while the hygiene controls quietly degrade, and that is the most common failure pattern on this asset, because the degradation is invisible and the plant keeps working.

So use this framework the way it is intended: to structure a scope conversation and to make sure no task area is missing. Then get the intervals, the tests and the thresholds from your jurisdiction and from the competent person who wrote your scheme of control. The value a maintenance function adds here is not authoring the water safety regime. It is making sure a regime exists, that the hardware can deliver it, that the monitoring happens, that findings become work, and that deviations get closed. That is a substantial contribution, and it is entirely within your control.

Final thoughts

If I had one recommendation for a team that suspects its tower regime is thin, it would be to go and look at the asset with this framework in hand, before touching the schedule. Walk around it at ground level and look at the surfaces downwind. Open what can safely be opened, with the isolation done properly. Check whether the bleed valve is open, whether the conductivity probe is clean, whether the strainer is present, whether the drift eliminators are complete, and whether the level control is doing its job. Then ask to see the scheme of control and the last set of monitoring records, and ask what happened the last time a result fell out of range.

That inspection takes a morning and tells you more about the true state of the regime than any document review. The findings cluster in the same few places every time: the deferred hygiene items, the instrument nobody verified, the valve somebody closed for a good reason years ago, and the deviation recorded but never actioned. None are difficult to fix. They need somebody to look at this asset as the whole thing it is, rather than as a fan with some water under it.

Related reading: What is a cooling tower and how it works, What is a chiller: types and how chillers work, Boiler and chiller preventive maintenance, Preventive maintenance for HVAC systems, Chiller plant analytics, PM KPIs and schedule compliance, Lockout tagout safety guide, Confined space hazards and procedures, Working at height hazards and controls.