
Corrosion zoning of cooling towers and recirculating cooling-water systems is the design practice of dividing an open recirculating cooling-water system, according to water level and exposure condition, into four locations whose corrosion stress differs markedly — the atmospheric splash zone, the variable water-level zone, the immersion zone and the deposition zone of the collection basin — and then specifying, zone by zone, the surface-preparation grade, the coating system and the acceptance criteria. Its value is practical: it replaces "one coating brushed over everything" with "grade by zone, film thickness by zone, acceptance items by zone", so that the owner can verify each item inside an outage window instead of tracing back through water-quality reports after a leak appears.
A cross-flow, mechanically induced draft cooling tower was opened for inspection after 30 months in service. On the web of a carbon-steel support roughly 400 mm above the water-level fluctuation band of the collection basin, 3 through-wall pinholes were found, with black slimy deposits clinging to the hole rims and a distinct sulphide odour when they were scraped away. The above-water part of the very same column showed nothing worse than loss of gloss and local chalking of the topcoat, with no perforation. The owner arrived holding a water-treatment monthly report that read "carbon-steel heat-transfer surface, water side, corrosion rate 0.05 mm/a, below the 0.075 mm/a limit in Clause 3.1.5 of GB/T 50050-2017", and asked a fair question: the numbers are compliant, so why did it still leak? The answer is blunt. The reporting basis of that table is the heat-transfer surface of heat-exchange equipment, whereas the failure happened on the tower structure and in the collection basin. Flow regime, oxygen supply, frequency of wet/dry cycling and coverage by deposits are not of the same order of magnitude between those places, and accepting four different zones on one averaged figure writes the risk straight into a blind spot.
TL;DR — Verifiable Conclusions First
- Zone-by-zone design outranks blanket grading: the operating pH band of an open recirculating cooling-water system is 7.0–9.5 (GB/T 31329-2025, Table 3), yet the coating stress in the splash zone and the variable water-level zone is far higher than in the immersion zone, so the service life of one and the same system at different locations commonly differs by as much as an order of magnitude.
- Water-quality limits are not structural acceptance values: Clause 3.1.5 of GB/T 50050-2017 requires the water-side corrosion rate of carbon-steel heat-transfer surfaces to be below 0.075 mm/a and that of copper alloys and stainless steel below 0.005 mm/a; the deposit (adhesion) rate shall not exceed 15 mg/(cm²·month), or 20 mg/(cm²·month) in the refining industry, and the fouling resistance shall not exceed 3.44×10⁻⁴ m²·K/W; tube-side velocity shall be above 1.0 m/s and shell-side velocity above 0.3 m/s, while the wall temperature on the cooling-water side of the heat-transfer surface should preferably not exceed 70 ℃.
- Grade wet/dry-cycling locations more severely: ISO 12944-2:2017 classes first-year thickness loss of carbon steel as C3 above 25–50 μm/a, C4 above 50–80 μm/a and C5 above 80–200 μm/a; the splash zone and the variable water-level zone are normally assessed from C4 upwards, permanently immersed freshwater locations correspond to Im1, and buried sections in soil to Im3.
- Microbiologically influenced corrosion has checkable criteria: total heterotrophic bacteria should preferably stay below 1×10⁵ CFU/mL and biological slime below 3 mL/m³ (Clause 3.1.12 of GB/T 50050-2017; Clauses 5.2.4 and 5.2.5 of GB/T 31329-2025 tighten the slime figure to not more than 2.0 mL/m³), with sulfate-reducing bacteria (SRB) detected by the MPN method of GB/T 14643.5-2009 and iron bacteria by GB/T 14643.6-2009.
- Holiday testing comes in two equipment families: Method A of ASTM D5162 is the low-voltage wet-sponge test (5–90 V DC) and Method B the high-voltage spark test (above 800 V); concrete substrates follow the scope of ASTM D4787, and NACE SP0188:2024 is the current procedure document for holiday detection.
- Judge film thickness by the 90/10 rule: both SSPC-PA2 and ISO 19840 require 90% of measurement points to be at or above the specified thickness, with the remaining points no lower than 90% of that value, and extra points are needed at edges and welds.
- Read adhesion separately by substrate: pull-off adhesion is measured to GB/T 5210 / ISO 4624. For steel-backed systems the criterion commonly applied in the field is not less than 5 MPa with the failure plane away from the interface; on concrete the failure is expected inside the concrete itself, so the number alone will mislead you.
I. The Four Zones: Location, Medium Condition and Damage Morphology
Dividing the tower into zones is not about colouring a drawing; it decides whether work can be carried out with water present, whether abrasive blasting is feasible, and whether chemical cleaning can restore the surface. The boundaries of the four zones below can be measured on site: mark the fluctuation band from the design level and the level swing in the collection basin, mark the splash zone from the throw of the water distributors and the spray off the fill, mark the immersion zone from the permanently submerged height, and mark the basin-floor zone from where deposits actually accumulate.
1. Atmospheric splash zone: a wet film that keeps drying out, with salts concentrating in place
This zone covers the inner face of the tower wall below the fill, the inside of the air-inlet louvres, the distribution laterals and their hangers, the drift eliminator support frame, and any steel member struck directly by falling water. The metal here is not continuously submerged: a water film spreads, dries in the moving air, and spreads again. Evaporation leaves the bicarbonates and chloride ions of the water behind on the surface, producing a locally concentrated electrolyte, while the same surface repeatedly meets air and is therefore well supplied with oxygen. The typical damage is a mixture of coating blistering, flake-scale rust bloom, chalking and gloss loss, and pitting; the rust layer is thin but scattered, and galvanized parts commonly show white rust — the fluffy, pale zinc corrosion product that forms under damp and ventilated conditions. This is also the first zone that maintenance staff notice with the naked eye, although it is usually not the first zone to fail.
2. Variable water-level zone: oxygen concentration cells under alternating deposit cover
The locations are the water-level fluctuation band of the collection basin, the areas around make-up and blowdown outlets, the suction section of the circulation pumps, and trash racks and strainer supports. What makes the fluctuation band special is that the same surface alternates between an above-water or thin-film state with plentiful oxygen and a submerged state starved of it, setting up an oxygen concentration cell in which the oxygen-poor area becomes the anode and dissolves preferentially. At the same time, leaves, dust and the suspended solids that escape when side-stream filtration underperform are alternately deposited and scoured off, creating occluded regions. The characteristic morphology is rust blisters with steep, narrow pits beneath them — visible to the eye but missed by callipers, so only ultrasonic thickness measurement gives a dependable reading.

3. Immersion zone: stable conductivity, limited oxygen, risk shifting to under-deposit attack and MIC
The locations are the lower basin wall, the permanently submerged lengths of piles and supports, the channel boxes and water boxes of heat-exchange equipment, and the bore of circulating water pipes. On both sides of the coating there is a continuous water phase, ionic conductivity is stable, corrosion is limited by oxygen diffusion, and general thinning is often slow. The real problems come from the occluded regions that deposits create and from microbial metabolites. Black iron sulphide deposits, the localised position of pitting and the odour beneath biological slime are three mutually corroborating clues, and the matching water-side indicators are limits such as a deposit rate of 15 mg/(cm²·month) and heterotrophic bacteria of 1×10⁵ CFU/mL. On austenitic stainless steel, watch chloride enrichment in crevices and under deposits: Table 3 of GB/T 31329-2025 controls chloride to not more than 700 mg/L for stainless heat-exchange equipment with water on the shell side, and requires the water-side wall temperature of the heat-transfer surface to stay at or below 70 ℃ and the circulating cooling-water outlet temperature to remain below 45 ℃.
4. Collection basin deposition zone: anaerobic micro-environments, total iron and inspection blind spots
Think of the low spots on the drainage slope of the basin floor, the areas beneath fill support grids, basin-wall corners, and the roots of manholes and embedded nozzles. Velocity is low, desludging is difficult, and silt, colloidal iron, algal and bacterial exudates together with residual treatment chemicals consolidate into a dense deposit layer. Colloidal iron is strongly adhesive, hard to remove, and feeds iron bacteria; where phosphonate-based formulations are used, ferrous phosphate scale can form and bond tenaciously. Most through-wall pinholes in carbon-steel supports appear in this zone, at the water line and at the lower edge of the deposit layer, because four conditions coincide there — differential aeration, salt enrichment, anaerobic metabolism and the impossibility of drying out — which makes it the zone with the highest specification demands and the poorest working conditions of the four.
Look at the structural materials zone by zone as well. The tower casing and the collection basin are commonly glass-reinforced plastic (isophthalic unsaturated polyester resin with UV-stabilising and anti-ageing additives), and the figures typically written into open-tender technical clauses are a resin glass content of not less than 30%, a minimum casing wall thickness of 8 mm, a minimum basin wall thickness of 10 mm and a long-term temperature resistance of not less than 55 ℃; the upper support frame is hot-dip galvanized carbon steel, fasteners are stainless steel, and dissimilar-metal joints are isolated with insulating gaskets to prevent galvanic corrosion. These are equipment procurement figures, not performance commitments from a coatings manufacturer, and when they are carried into a technical agreement their source and the boundary of responsibility should be stated.
II. Differences in Mechanism: Why One Average Hides a Failure
1. Carbon steel and hot-dip galvanizing behave differently in the same water
Bare carbon steel in neutral water containing dissolved oxygen corrodes mainly by oxygen depolarisation, and its products are reddish-brown hydrated iron oxides. On a hot-dip galvanized part the zinc is consumed first; zinc corrosion products such as basic zinc carbonate are relatively compact in a clean water film, but in chloride- and sulphide-bearing water subject to wet/dry cycling they turn into fluffy white rust that is washed away again and again. Zinc is amphoteric, which makes it pH-sensitive: both acidic and strongly alkaline conditions accelerate dissolution — precisely why an acid-dosing regime and the low-pH window of a cleaning campaign must be written into the coating specification. A zinc layer carried under an organic coating is what the industry calls a duplex system: when galvanizing and paint act together the real service life exceeds the simple sum of the two layers acting alone, but the size of that gain depends on the corrosivity class, edge maintenance and the timing of the overcoat, and must not be extrapolated to every location by applying a factor.
2. Under-deposit corrosion: reading deposit rate and corrosion rate together
When the deposit (adhesion) rate and the corrosion rate are both high, under-deposit corrosion caused by settled material is the likely mechanism, and the order of treatment is to solve the deposition first — side-stream filtration, foul-release cleaning, desludging — rather than to double the inhibitor dose. A low deposit rate with a high corrosion rate usually points to the intrinsic aggressiveness of the water itself (low alkalinity, low hardness, low pH, high chloride) or to insufficient dispersing power in the formulation. A high deposit rate with a low corrosion rate normally means dispersion or side-stream filtration cannot keep up, and is a scaling risk that has not yet turned into metal loss. These three reading paths decide whether the outage window should be spent on coating repair or on water-treatment adjustment.
3. Distinguishing MIC from under-deposit attack: three checkable clues
First, location: pitting concentrated at the lower edge of a deposit layer, at the water line, in dead corners and behind gaskets. Second, products: black iron sulphide deposits accompanied by a sulphide odour point to sulfate-reducing bacteria, while hard reddish-brown nodules point to iron bacteria. Third, testing: GB/T 14643.5-2009 uses the MPN method, incubating under anaerobic conditions and reading a positive as a black ferrous sulphide precipitate, while GB/T 14643.1-2009 and HG/T 4207 each give the plate-count procedure for heterotrophic bacteria in slime. Sampling deserves care: the bacterial concentration inside slime is far higher than in the bulk water, so testing only the circulating water understates the risk; aerating the sample suppresses anaerobes, and unquenched residual chlorine kills the very organisms being counted — either error alone is enough to invalidate the result. Legionella belongs to occupational hygiene rather than corrosion; the detection method is given in GB/T 40392-2021, and although it has nothing to do with coating selection it does bear on the ventilation and personal protection required during sludge removal and jetting.
4. Four sources of water-quality excursion, all pushing the coating the same way
Cycles of concentration: Clause 3.1.11 of GB/T 50050-2017 asks for a design value of not less than 5.0 and never below 3.0 for open recirculating systems with indirect cooling. The water-saving benefit comes from reduced make-up and reduced blowdown, and the price is a parallel rise in dissolved solids and chloride, which raises the osmotic pressure at the coating interface and makes the salt enrichment of each wet/dry cycle harsher. Process leakage: an internal leak in a heat exchanger brings process media into the circulating loop, and oils and low-molecular-weight species undermine the coating interface. Make-up water turbidity: Table 1 of GB/T 31329-2025 requires not more than 3.0 NTU for conventional water and suspended solids not more than 10 mg/L for non-conventional sources; when make-up turbidity runs away, the load passes straight through the side-stream filters onto the basin floor. Biocide dosing regime: continuous versus shock dosing changes the peak duration of oxidative exposure. Working together, these four push the system in one direction: the coating loses its barrier function first, and the water-treatment figures deteriorate afterwards.

III. Zone — Class — System: Turning "What Goes Here" into a Verifiable Table
The first step in system design is to translate zones into the language of corrosivity classes, so that the coatings supplier, the applicator and the owner speak with one vocabulary. ISO 12944-2:2017 classes the atmosphere from C1 to CX by the first-year thickness loss of standard specimens and treats immersion and burial separately (Im1 freshwater, Im2 seawater or brackish water without cathodic protection, Im3 soil, Im4 seawater or brackish water with cathodic protection); the identically adopted Chinese version is GB/T 30790.2-2014. ISO 12944-1 divides durability into four bands: low (not more than 7 years), medium (7–15 years), high (15–25 years) and very high (above 25 years). The peculiarity of a circulating water system is that one and the same member spans more than two classes: a column is classed on atmospheric exposure in the splash zone, more severely under a combined atmospheric-plus-immersion rule in the fluctuation band, and its base foot adds a microbiological factor in the basin deposition zone. The adjacent, complete water-treatment side of the same logic is discussed in Corrosion protection of industrial recirculating cooling-water systems: water boxes, circulation piping and cooling basins; this article deals only with zone boundaries, system comparison and acceptance criteria.
| Zone and location | Medium condition and exposure characteristics | Reference classification | Coating direction | Nominal dry film thickness basis |
|---|---|---|---|---|
| Atmospheric splash zone (inner tower wall, distribution laterals, air-inlet louvres, drift eliminator supports) | Wet film alternating with dry film, salts concentrating in place, plentiful oxygen, with UV and wind speed in the picture | Assessed from C4 upwards; C5 near the coast or in heavy pollution (carbon steel first-year classes of 50–80 and 80–200 μm/a) | Zinc-rich epoxy or epoxy zinc-phosphate primer + epoxy micaceous iron oxide intermediate coat + aliphatic polyurethane or fluoropolymer topcoat | 240–320 μm, with 30% added at edges and welds after stripe coating |
| Variable water-level zone (basin level fluctuation band, pump suction, rack and strainer supports) | Cyclic immersion and drying, deposits covering and releasing, oxygen concentration cells | C4 and Im1 combined and graded to the more severe of the two | Solvent-free epoxy thick build coat + water-resistant sealing topcoat, or a glass-flake reinforced intermediate layer | Not less than 320 μm, applied in several coats with thickness measured after each coat |
| Immersion zone (lower basin wall, submerged lengths of piles and supports, channel boxes and water boxes) | Continuous water phase, stable ionic conductivity, oxygen diffusion limited, with under-deposit and microbial involvement | Im1 (long-term freshwater immersion) | Solvent-free epoxy system as a build coat or a lining, avoiding pinholes and solvent entrapment | 350–500 μm, applied in layers within the per-coat thickness limit |
| Collection basin deposition zone (floor low spots, beneath support grids, corners and manhole roots) | Low velocity, dense deposits, anaerobic micro-environment, never fully dry | Im1 plus a microbiological factor, raised one class on the local-concentration principle | Thick vinyl ester or epoxy glass-flake mortar lining, with spark holiday testing of the completed area in full | Linings controlled at the 1.0–1.5 mm class, recorded coat by coat |
| Concrete basin wall and floor (including repaired old concrete) | Carbonation, chloride ingress, capillary water coexisting with alkalinity, porous surface | Handled under concrete-specific rules rather than by copying steelwork classes | Penetrating sealer primer + alkali-resistant topcoat, or a composite with polymer mortar or acid-resistant brick lining | The sealer primer works by penetration and builds no thick surface film; the topcoat follows the system data sheet |
| Hot-dip galvanized parts and dissimilar-metal joints (upper support frame, fasteners) | Atmosphere and splash together, the zinc layer participates, contact potential difference acting together with the crevice effect | Duplex in C3–C4; crevice locations graded more severely in their own right | Zinc phosphate pre-treatment or epoxy tie-coat + topcoat of the same system; insulating isolation gaskets at joints | Overcoat 60–120 μm, avoiding wet-on-wet application under film tension |
Mature experience with build coats and linings on other assets maps directly onto these zones: the alternating immersion and vapour-space areas of atmospheric storage tank interiors are covered in Tank interior coatings: weld joint rotation, bottom plate zones and cathodic protection compatibility, and the high-humidity acid condensation and osmotic pressure problems on the flue-gas side in Acid-resistant linings for desulphurised chimneys: the boundary between glass flake and foam glass. What all three have in common is a discontinuous medium, a high osmotic pressure at the interface, and acceptance that rests on holiday testing rather than on appearance.
IV. Surface Preparation and Application Windows: The First Link to Break
Coating life is more sensitive to substrate preparation than to the brand of paint. Field failures cluster in four places: mill scale and residual old coating, an anchor profile that is too shallow or too deep, soluble salts above the limit, and application while condensation is present. Graded treatment of rusty and previously coated surfaces is covered in Rusty surface preparation and rust conversion: choosing between St3, Sa2.5 and conversion coatings.
| Preparation item | Method and current standard | Field criterion | Action when the result fails |
|---|---|---|---|
| Rust grade and condition of the existing coating | Visual assessment to ISO 8501-1 grades A–D; cross-cut and pull-off sampling on the old coating | The rust grade is entered in the raw substrate record and countersigned by all three parties | Where adhesion has failed or large areas are hollow, strip completely; cover-over patching is not accepted |
| Blast-cleaning grade | Sa2.5 (very thorough) to ISO 8501-1, with power tooling to St3 where access is restricted | Uniform bare metal by eye, with only slight shadows and streaks permitted | Remaining mill scale or weld slag is re-blasted; priming straight over it is prohibited |
| Anchor profile and surface roughness | Comparison strips or stylus method to ISO 8503-1; replica tape method per GB/T 13288.5 | Rz held within the 40–75 μm band and matched to the chosen per-coat thickness | Below 40 μm, blast a second pass with fine grit; above 75 μm, a light sweep blast flattens the peaks and prevents bridging |
| Soluble salt residue | Bresle patch sampling to ISO 8502-6, conductivity measurement to ISO 8502-9 | The limit commonly quoted in public sources and classification rules is not more than 50 mg/m² as sodium chloride equivalent | If above the limit, give a second low-pressure fresh-water rinse, dry and retest; blowing the dust off alone is not a remedy |
| Dust and oil | Pressure-sensitive tape sampling for dust to ISO 8502-3; solvent wipe followed by a water-break check | Dust distribution grade 1, and the water-break test showing a continuous film | Oily areas are emulsion-cleaned separately and rinsed a second time; silicone contamination requires tracing the equipment source |
| Application microclimate | Steel temperature, relative humidity and dew point monitored and recorded together | Substrate at least 3 ℃ above the dew point, relative humidity not above 85%, temperature 5–35 ℃ | Stop blasting when condensation or rain occurs; restore the window with heating and dehumidification, then re-measure |
| Handover condition of concrete substrates | Assessed jointly on age, moisture content and alkalinity; sealer primer selection is covered in Alkali-resistant primers for concrete: sealing, penetration and topcoat compatibility | Age of at least 28 d, moisture content no more than 8%, pH below 10, laitance and friable layers removed | For new concrete or excessive moisture, dewater and ventilate first; coating against this rule is treated as rework |
The three hard constraints of an outage window: work with water present, cure on damp surfaces, and overcoating intervals
Outage windows for cooling towers and collection basins are typically measured in hours, and complete dryness is rarely achievable. The workable answer is to write the constraints into the specification instead of leaving them to the crew's judgement on the day. The first is zone dewatering and drainage: discharge the standing water below the fluctuation band through the blowdown and drain connections, then clear the basin floor by hand and blow it dry. The second is to choose routes with greater tolerance of a damp interface — penetrating sealer primers on concrete, wet-surface-tolerant solvent-free epoxy on steel — and to require the supplier, in the technical agreement, to provide pull-off adhesion data measured on a damp substrate. The third is overcoating-interval management: once the maximum recoat window stated for the system has passed, the surface must be abraded before the next coat, or intercoat adhesion will fail in the "looks intact, delaminates the moment you pull it" mode. The second of the three is the one most often overlooked: areas applied wet or only half dry tend to declare themselves the following summer, when cycles of concentration run high.

V. Compatibility Boundaries Between the Coating and the Water-Treatment Chemicals
Once a coating enters the system it carries two stresses at once: the water and the chemicals. The mismatch we see most often on the owner's side is that the chemical programme is chosen from corrosion rates, the topcoat is chosen from weatherability, and nobody writes down how much oxidative exposure and how low a cleaning pH this particular coating can actually take. Compatibility on closed systems and secondary-side coolants is discussed separately in Internal corrosion in data-centre liquid-cooling secondary loops: coolant compatibility and copper-aluminium galvanic protection; this section deals only with oxidative and non-oxidative chemical exposure in open systems.
| Coating type | Applicable range of chemical exposure | Early risk signal | Verification method |
|---|---|---|---|
| Acrylics (including waterborne one-pack) | Neutral to mildly alkaline operation, and the conventional case of continuous chlorination with free chlorine on the return header held at not more than 0.5 mg/L | Gloss loss and light chalking appear first in the splash zone and inside the louvres, with fading of colour | Gloss and colour-difference comparison against retained samples, combined with cross-cut sampling of intercoat adhesion |
| Epoxy (solvent-based and solvent-free build coats) | The natural operating band of pH 7.0–9.5; for intermittent shock dosing, confirm against the oxidative exposure ceiling stated by the supplier | Surface blushing, amine spots and rain-like chalking, followed by local blistering | Pull-off adhesion plus hardness before and after immersion; FTIR comparison where degradation must be confirmed |
| Vinyl ester and glass-flake mortar | Wide pH range, dual exposure to oxidative biocides and the low-pH cleaning window, long-term immersion with deposit cover | Interlayer delamination, flake loss and staining along repair joints | Holiday testing coat by coat together with thickness records; delaminated samples analysed by loss on ignition and composition |
| Aliphatic polyurethane topcoat (outdoor weathering layer) | External steelwork, platforms and handrails, where atmosphere and splash dominate | Falling gloss and fine checking, accelerating in hot and humid seasons | Cross-check against accelerated-weathering data and field gloss retention; intercoat compatibility test before recoating |
| Chlorinated rubber and alkyd legacy systems | Not recommended where oxidative biocides are dosed continuously or acid is injected; assess as a whole at refurbishment | Solvent re-attack, lifting of the coat beneath, rapid softening on contact with the new chemical batch | Small-area compatibility trial coating before deciding whether to retain it; never overcoat directly with the new system |
Four quantified boundaries are worth writing into the technical agreement in their own right. The first is oxidative exposure: Table 3 of GB/T 31329-2025 sets free chlorine at the circulating return header at 0.05–0.5 mg/L under continuous chlorination, and the determination of residual and free chlorine is given in GB/T 14424-2008 (measuring range 0.03–2.5 mg/L). Shock dosing lifts the short-term peak several-fold, so the coating specification should state the permitted peak concentration and contact time rather than simply saying "chlorine resistant". The second is the acid and low-pH window: under acid dosing, the permitted sum of calcium hardness and total alkalinity is relaxed to not more than 1800 mg/L as calcium carbonate, against not more than 1100 mg/L at natural pH, and acid addition or softening is required once that sum exceeds 1100 mg/L or the stability index RSI falls below 3.3 (Clause 3.1.10 of GB/T 50050-2017). During chemical cleaning the pH should be held at 4.0–6.0, and pre-filming begins once the flush has brought total iron below 2 mg/L and turbidity below 20 NTU (Clause 5.3.2.1 of GB/T 31329-2025; the cleaning and pre-filming sequence follows HG/T 3778). Throughout that period both the coating and the galvanizing work at low pH, so the specification must state a tolerated duration. The third is the inhibitor and antiscalant package: combinations of organophosphonates, polyacrylates, zinc salts and molybdate change the film that forms at the interface; inhibition performance is assessed by the rotating rack coupon method of GB/T 18175 and the suitability of a formulation is confirmed by the dynamic simulation test of HG/T 2160. The fourth is the analytical basis for scale and corrosion-product reports: HG/T 3610-2026, the general method for analysis of fouling deposits and corrosion products in industrial recirculating cooling water, was issued on 2026-06-01 and takes effect on 2026-12-01, replacing HG/T 3610-2000; the report fields and the wording of judgements for calcium, iron, phosphorus and silicon in scale will move with it, so any comparison that crosses versions must state when the method basis was switched.
VI. Holiday Testing, Film Thickness and Adhesion: An Acceptance Sheet You Can Enforce
The benefit of a zoned system exists only if the acceptance items can be verified. The aim of acceptance is not "the coating looks uniform" but four answers: are there through-coating defects, does the thickness land inside the specified band, do intercoat and interface bond meet the criterion, and is every defect located to coordinates that can be re-measured.
1. Test voltage and equipment choice: new coatings and in-service coatings are not the same method
ASTM D5162 divides discontinuity detection into Method A, the low-voltage wet-sponge test (working voltage 5–90 V DC with an open-cell foam electrode and audible and visual indication), and Method B, the high-voltage spark tester (voltage above 800 V), for metallic substrates, with concrete surfaces following the scope of ASTM D4787; the current 2024 edition of NACE SP0188 sets the procedural requirements for both shop and field equipment. Three boundaries are routinely misapplied and belong in the contract. First, converting test voltage from dry film thickness is only an empirical rule of thumb; the actual value must come from the dielectric strength data supplied by the coatings manufacturer, and coatings loaded with platelet fillers such as glass flake need a correction, because the platelets lengthen the breakdown path. Second, the probe should travel at a steady speed; the range commonly quoted in public sources is 0.15–0.30 m/s, and moving too fast misses holidays while moving too slow can generate false indications from local heating. Third, and least noticed, NACE SP0188 states that the procedure applies only to newly applied coatings: high-voltage testing of an old coating that has already seen immersion service can damage areas that were sound and can also report false holidays through moisture uptake or media penetration. For a tower in service, the sensible combination is low-voltage wet-sponge sampling, re-measurement of dry film thickness, visual grading to the ISO 4628 series, and selective laboratory analysis of taken samples — not sweeping the whole tower with a spark box. Porosity itself is assessed under ISO 29601:2011 on the evaluation of porosity in the dry films of protective coating systems.
2. Layering the thickness and adhesion criteria: nominal value, measured band and failure mode
Film thickness is accepted under the 90/10 rule of SSPC-PA2 or ISO 19840, that is 90% of readings at or above the specified thickness and the remainder no lower than 90% of it; thickness is measured per GB/T 13452.2, using the magnetic method for non-magnetic coatings on steel, and on galvanized substrates the metallic coating and the organic overcoat must be read as two separate layers. Adhesion is measured by the pull-off method of GB/T 5210 or ISO 4624, and the criterion must give a number and a failure mode together: for steel-backed systems the pass line commonly applied in the field is not less than 5 MPa with the failure plane not at the coating-to-steel interface; on concrete, 1.5–2.5 MPa is typical but the failure is expected inside the concrete itself, and a fracture running entirely along the coating-to-concrete interface is judged an adhesive bond failure. Thin films at edges, welds and bolts are the most frequent finding in zone-based acceptance; the remedy is stripe coating before the general spray pass, with a higher proportion of gauging points at those locations.
| Acceptance field | What must be recorded | Source of the criterion | How the owner checks it |
|---|---|---|---|
| Zone and drawing reference | Each inspection lot marked with its zone (splash, variable water-level, immersion, deposition) and member number | The zoning and classification basis in Table 1 of this article | Sample by coordinates and check whether one and the same column crossing zones was written against a single set of criteria |
| Surface preparation and salt residue | Cleanliness grade, anchor profile depth, Bresle sampling locations and measured salt content | ISO 8501-1, ISO 8503-1, ISO 8502-6 and ISO 8502-9 | Verify the number of sample points and the retest records after an exceedance; a single certificate of conformity is not enough |
| Application microclimate | Per shift: substrate temperature, dew point, relative humidity and any periods when blasting or spraying stopped | The system data sheet plus the window criteria in Table 2 of this article | Check continuity of the log; missing shifts require a re-measurement or a written rework statement |
| Number of coats and nominal dry film thickness | Batch number of each coat, wet and dry film readings per coat, and the 90/10 statistical result | SSPC-PA2, ISO 19840, GB/T 13452.2 | Ask for the raw gauging data and the statistical process; an average on its own is not acceptable |
| Holiday test record | Test type (low-voltage wet sponge or high-voltage spark), the basis for the chosen voltage, holiday coordinates and retest after repair | ASTM D5162, ASTM D4787, NACE SP0188:2024 | Check whether one voltage was used for both new and in-service coatings, and look at the post-repair retest results |
| Operating-side linked data | Corrosion rate, deposit rate, fouling resistance, heterotrophic bacteria and slime volume, cycles of concentration and free chlorine | Clauses 3.1.5 and 3.1.12 of GB/T 50050-2017; Section 5.2 of GB/T 31329-2025 | Compare against the coating inspection record for the same period and check whether water-quality excursions preceded coating defects |
VII. Key Figures at a Glance, with Units and Sources
- Water-side corrosion rate of carbon-steel heat-transfer surfaces shall be below 0.075 mm/a, and that of copper alloys and stainless steel below 0.005 mm/a (Clause 3.1.5 of GB/T 50050-2017).
- Fouling resistance on the water side of equipment heat-transfer surfaces shall not exceed 3.44×10⁻⁴ m²·K/W, and the deposit (adhesion) rate shall not exceed 15 mg/(cm²·month), or 20 mg/(cm²·month) in the refining industry; for closed systems the fouling resistance shall be below 0.86×10⁻⁴ m²·K/W (the same clause together with Clause 3.1.6).
- Tube-side velocity of circulating cooling water shall be above 1.0 m/s and shell-side velocity above 0.3 m/s, and the wall temperature on the cooling-water side of the heat-transfer surface should preferably not exceed 70 ℃ (Clause 3.1.5 of GB/T 50050-2017).
- Total heterotrophic bacteria should preferably stay below 1×10⁵ CFU/mL and biological slime below 3 mL/m³ (Clause 3.1.12 of GB/T 50050-2017); Clause 5.2.5 of GB/T 31329-2025 tightens biological slime to not more than 2.0 mL/m³, and not more than 3.0 mL/m³ for the refining industry.
- For open recirculating systems with indirect cooling, the design cycles of concentration should preferably be not less than 5.0 and shall not be below 3.0 (Clause 3.1.11 of GB/T 50050-2017); chloride per Table 3 of GB/T 31329-2025 is limited to 1000 mg/L with water in the tubes, and to 700 mg/L for stainless steel with water on the shell side at a wall temperature not exceeding 70 ℃.
- Circulating cooling water: pH 7.0–9.5, total iron not more than 2.0 mg/L, sulphate not more than 2000 mg/L, and free chlorine 0.05–0.5 mg/L under continuous chlorination (Table 3 of GB/T 31329-2025); turbidity is controlled by equipment type at not more than 20 NTU and not more than 10 NTU (same table).
- ISO 12944-2:2017 first-year thickness loss classes for carbon steel: C3 above 25–50 μm/a, C4 above 50–80 μm/a, C5 above 80–200 μm/a; the corresponding zinc classes are 0.7–2.1, 2.1–4.2 and 4.2–8.4 μm/a.
- Holiday test voltage basis: low-voltage wet sponge 5–90 V DC, high-voltage spark above 800 V (ASTM D5162); a steady probe travel of 0.15–0.30 m/s is the range commonly quoted in public sources.
- Field criteria for thickness and adhesion: 90% of measurement points at or above the specified thickness and the remainder no lower than 90% of it (SSPC-PA2 / ISO 19840); pull-off on steel not less than 5 MPa with the failure plane away from the interface (GB/T 5210 / ISO 4624), and 1.5–2.5 MPa on concrete with the fracture inside the concrete.
VIII. A Checklist the Owner Can Run in the Outage Window
- Put the water-quality report and the coating inspection record on one timeline: check whether step changes in cycles of concentration, free chlorine, total iron or turbidity preceded blistering, chalking or pinholes, and use that same-period comparison to decide whether the problem starts on the water side or on the coating side.
- Measure three coordinates, one per zone: inside a louvre and a distribution lateral for the splash zone, a support in the level fluctuation band for the variable water-level zone, and beneath a floor grid for the deposition zone, recording appearance, dry film thickness and whether any through-coating defect can be detected. Readings taken only on dry, above-water steel are not acceptable.
- Ask where the corrosion coupons are mounted: only coupons in a monitored heat exchanger under simulated heat load approach the velocity and wall-temperature conditions the standard assumes; coupons simply sitting in the tower basin see water near the system average temperature at close to zero velocity, and their corrosion rates usually read low.
- Require the applicator to hand over raw gauging data and the basis for the holiday test voltage, including the supplier's dielectric strength data, the correction applied for platelet filler, and whether low-voltage and high-voltage methods were used respectively for new and in-service coatings.
- Archive the application windows and overcoating intervals: record ambient temperature, touch-dry and hard-dry times for each coat and any abrasion carried out when a coat was applied the following day, so that an intercoat problem can be traced to a specific coat.
- Take scale and corrosion-product samples in duplicate: one for compositional analysis, one sealed and held in reserve. From 2026-12-01 issue reports under HG/T 3610-2026, and state the method change date on any cross-version comparison.
IX. Frequently Asked Questions
1. Every water-quality number is compliant, so why did the tower's steel support perforate?
Because the reporting basis and the failing location are not the same thing. The corrosion-rate limits in Clause 3.1.5 of GB/T 50050-2017 apply to the water side of heat-exchange heat-transfer surfaces and assume tube-side velocity above 1.0 m/s, shell-side velocity above 0.3 m/s and wall temperature no higher than 70 ℃; tower supports and the collection basin are structures sitting in a combined environment of low velocity, deposit cover and wet/dry cycling. What the report measures is an average of general thinning; what destroyed the member is localised pitting. Coupons soaked in the tower basin read lower still. Judging this class of failure means looking at morphology — general thinning, local pitting and under-deposit attack call for three entirely different responses — checking whether deposit rate and corrosion rate are rising together, and running compositional analysis on the corrosion products and the scale, rather than asking whether a single average sits below 0.075 mm/a.
2. Can the splash zone and the immersion zone share one system, with just an extra step of film thickness?
We do not recommend it, because the failure drivers differ. The splash zone is governed by salt enrichment from wet/dry cycling, oxygen availability and UV ageing; the immersion zone is governed by water penetration, ion migration and microbiology in occluded regions. Added thickness helps with osmotic pressure and pinholes in the immersion zone but does almost nothing for UV ageing and wet/dry fatigue in the splash zone, while introducing new risks of build-coat internal stress, solvent entrapment and uneven edge thickness. The right approach is a system per zone: a three-coat barrier with a weather-resistant topcoat and proper stripe coating of edges in the splash zone, and a solvent-free build coat or glass-flake lining with coat-by-coat holiday testing in the immersion and deposition zones. The boundary between them — the water-level fluctuation band — is executed to the more severe class, and the position of that boundary line is written into the acceptance record so that nothing is treated as an "average location" during application.
3. Do galvanized internal frames need painting at all, and when is a duplex system worth doing?
A hot-dip zinc layer protects by forming relatively compact zinc corrosion products under clean, ventilated, near-neutral water films. In chloride- and sulphide-bearing water with frequent wet/dry cycling, however, the zinc forms fluffy white rust that is washed off repeatedly and its protection efficiency falls. In that situation an organic overcoat on the galvanized surface — the duplex approach — does extend the maintenance-free interval, provided three things are done together: the zinc is pre-cleaned with a suitable method and then phosphated or primed with a dedicated tie-coat; the system uses a resin that is compatible with zinc and resistant to saponification, since strongly acid and strongly alkaline conditions both accelerate zinc dissolution; and edges, welds and drilled holes are touch-coated to the stricter local requirement. Conversely, where the original design treats galvanizing as the primary protection and the location is predominantly above-water atmosphere, adding an incompatible topcoat only buys early peeling and leaves acceptance unable to say which layer failed first.
4. Which zone suits a thick glass-flake mortar lining, and what typically goes wrong on site?
Its natural range is the immersion zone and the basin deposition zone — locations permanently under water that also see deposits and cleaning exposure — together with any case where the low-pH window of chemical cleaning has to be tolerated. It works by aligning platelet flakes in parallel overlap so that the penetration path through the film lengthens, and the price is a narrow application window and high sensitivity to interface preparation. Three mistakes dominate field failures. Intercoat intervals and cure temperature are not controlled: at low temperature or high humidity, touch-dry and hard-dry times stretch considerably, and a coat laid on an uncured layer shows up later as interlayer delamination. Corners, embedded nozzle roots and manhole edges receive no local reinforcement: thickness and stress concentration are both unfavourable there, which makes them the usual starting points for staining. Coat-by-coat holiday records are missing, so a single high-voltage sweep after completion costs time and creates disputes at every repair patch. The specification should make the per-coat thickness limit, the environmental window, the local reinforcement detail and coat-by-coat holiday testing mandatory clauses.
5. After oxidative biocide dosing the coating whitens and powders — is that the chemical or the coating?
Investigate both, but in the right order. Start with whether exposure crossed the boundaries: under continuous chlorination, free chlorine on the return header should be held at 0.05–0.5 mg/L (Table 3 of GB/T 31329-2025), while shock dosing multiplies that several-fold for hours, and if the short-term peak or its duration exceeds the tolerance stated by the coating supplier, epoxies will blush and show amine spots while acrylics chalk faster. Then look at the contributing factors outside the chemical: oils and light fractions carried in by an internal process leak swell the coating, and when acid dosing pulls the operating pH down, both the speciation and the oxidation potential of the biocide shift with it. The diagnosis is to sample the whitened area for FTIR and compositional comparison while cross-checking water-quality and dosing records for the same period. If only the splash zone and the inside of the louvres are affected while the immersion zone is sound, oxidative exposure acting together with UV ageing is the more likely cause than poor cure. Corrective action should therefore come from both sides at once: the dosing regime (the rhythm of continuous versus shock feed and the injection point) and the choice of topcoat.
6. Can a hydrostatic test replace spark testing for holiday detection?
No, because the two verify different objects. A hydrostatic test proves the sealing strength of the vessel or pipe body and detects macro leakage, at a test pressure set as a multiple of design pressure, but it cannot find pinholes or coating misses that have not yet reached the substrate. A spark or wet-sponge test proves the continuity of the coating as an insulating barrier and locates defects to coordinate level, but it carries no judgement on structural strength. A sensible combination is: after application, test coating continuity by the methods of ASTM D5162 and record every holiday and its repaired retest; during an outage, prove the body by the hydrostatic or pneumatic test the equipment code requires; after several years of service, stop sweeping old coatings at high voltage — NACE SP0188 warns explicitly about coatings already exposed — and switch to low-voltage sampling with thickness re-measurement and visual grading. Keep the three conclusions in separate acceptance boxes so that one pass cannot cover another gap.
7. The concrete basin wall is cracked from rusting reinforcement — can a coating save it, and what comes first?
A coating cannot address rust-jacking itself. The precondition for concrete protection is to establish carbonation depth, chloride content and reinforcement potential, and only then decide between repair, migratory inhibitor and a new protective layer: rust-jacking cracks mean the bar section or the cover concrete is already damaged, and painting straight over it merely seals water and ions into the interface to keep reacting. The sequence is to break back to sound substrate, remove rust and restore the cover to the structural requirement with polymer mortar or a dedicated repair system, and only then apply the concrete sealer and topcoat. The three quantified handover conditions on the concrete side are an age of at least 28 d, a moisture content no higher than 8% and a pH below 10. Young concrete is alkaline, and direct contact with a resin that is not alkali-resistant leads to saponification, blistering and peeling, so an alkali-blocking sealer primer has to intervene. At acceptance, use pull-off adhesion and look at where the fracture occurred; that reflects the real condition far better than the appearance of the film.
8. To extend coating life, should the money go into the paint or into water management?
Do the low-cost interface management first, and only then talk about a higher-grade coating. The entry point is substrate preparation quality and soluble-salt control; in historical cases, inadequate blast grade, salt exceedance and application during condensation together account for far more failures than the intrinsic performance of the paint. Next, write the zoned system down properly so that one specification does not cover four locations, and above all so that an above-water atmospheric system is not used in the immersion zone. Only after that does it make sense to move up a coating grade or add thickness, and at the same time write the chemical exposure boundaries (free chlorine peaks, cleaning pH, pre-filming window) into the specification so that the supplier gives a checkable commitment on tolerated duration. On the water side, keep an eye on velocity and wall temperature, the two operating conditions most often ignored: when they slip, neither extra thickness nor a better formulation can compensate. Spending in that order usually turns the routine from "patch it at every shutdown" into "postpone the patch with evidence".
X. References and Basis of the Figures
- GB/T 50050-2017, the design code for industrial recirculating cooling-water treatment: Clauses 3.1.5, 3.1.6, 3.1.10, 3.1.11 and 3.1.12 on velocity, wall temperature, fouling resistance, deposit rate, corrosion rate, cycles of concentration and microbiological indicators.
- GB/T 31329-2025, the technical specification for water-saving in circulating cooling water systems (issued 2025-10-31, effective 2026-05-01, replacing the 2014 edition): Tables 1 and 2 on make-up water quality, Table 3 on circulating cooling-water control limits, and Sections 5.2 and 5.3 on system and process control.
- ISO 12944-2:2017 (identically adopted in China as GB/T 30790.2-2014) for corrosivity classification and first-year thickness-loss classes; ISO 12944-1 for the durability bands L, M, H and VH and their year ranges.
- ASTM D5162 (Methods A and B for discontinuity detection), ASTM D4787 (concrete surfaces), NACE SP0188:2024 (holiday testing of new protective coatings on conductive substrates) and ISO 29601:2011 (evaluation of porosity in dry films of protective coating systems).
- ISO 8501-1, ISO 8502-3, ISO 8502-6, ISO 8502-9 and ISO 8503-1 for the surface-cleaning, dust, soluble-salt and roughness methods; SSPC-PA2 and ISO 19840 for film-thickness acceptance rules; GB/T 13452.2, GB/T 5210 and ISO 4624 for thickness and pull-off adhesion methods.
- Microbiology and testing: GB/T 14643.1-2009, GB/T 14643.5-2009 and GB/T 14643.6-2009 (slime-forming bacteria, sulfate-reducing bacteria and iron bacteria), HG/T 4207 (plate count of heterotrophic bacteria), GB/T 40392-2021 (Legionella detection) and GB/T 14424-2008 (residual and free chlorine).
- Chemicals and assessment: GB/T 18175 (rotating rack coupon method for inhibition performance), HG/T 2160 (dynamic simulation test), HG/T 3778 (cleaning and pre-filming), GB/T 15453-2018 (chloride determination), GB/T 15893.1-2014 (turbidity by the light-scattering method) and GB/T 39296-2020 (evaluating operating performance with a monitored heat exchanger).
- Product analysis and equipment side: HG/T 3610-2026, the general method for analysis of fouling deposits and corrosion products in industrial recirculating cooling water (issued 2026-06-01, effective 2026-12-01, replacing HG/T 3610-2000); GB/T 7190.1-2018 and GB/T 7190.2-2018 on product requirements for mechanically draft cooling towers; GB/T 50392-2016 on the process design code for mechanically draft cooling towers; DL/T 1027 on test procedures for industrial cooling towers.
- Basis of the figures: items such as holiday test voltage conversion factors, probe travel speed, duplex life gain and concrete repair criteria follow public sources and industry convention, and are flagged in place throughout the text. For any specific project the governing limits are those in the design documents, the supplier's technical data sheets and the contract appendices, and the data in this article are not a performance guarantee for a particular duty. The figures quoted for cooling tower casing resin content, wall thickness, flame resistance and temperature resistance are taken from publicly available equipment procurement clauses and are manufacturing requirements rather than coating performance indices.
Last updated: 2026-11-18
Compiled by the technical team of Kexin New Materials (Guangdong) Co., Ltd. / Kexin New Materials (Guangdong) Co., Ltd., for engineering and procurement staff working on industrial recirculating water and cooling tower assets. This is a B2B technical compilation written in engineering register; it is not a basis for product design and not a performance commitment, and coating and lining selections for a specific duty should be confirmed by laboratory retesting and field verification data. The company's external business is industrial-grade product supply and technical support, on EXW / FOB trade terms.