Cooling tower chemical dosing: is the inhibitor dose right?
Summary
Expected cooling tower inhibitor use equals makeup volume divided by cycles of concentration, times the target product dose in mg/L. Compare that figure with the product drawn from the drum or day tank over the same period. The check holds when the product is fed continuously, the tower is at steady cycles and the product leaves mainly with blowdown.
Makeup water replaces evaporation, blowdown and drift: Make-Up = Evaporation + Blowdown + Drift (US DOE FEMP, accessed 2026).
Cycles of concentration equal makeup divided by bleed, and bleed means any water removed other than by evaporation, including pump seal leaks, overflow and drift (Chem-Aqua TB 2-025, 2018).
In Chem-Aqua's 500-ton example with evaporation of 21,600 US gal/day (about 81.8 m3/day), raising cycles from 1.5 to 3.0 cuts bleed from 43,200 to 10,800 gal/day and inhibitor cost by 75% (Chem-Aqua TB 2-025, 2018).
Feeding inhibitor in proportion to makeup can lose control when makeup conductivity varies, when there are several water sources or when treatment pumps lose prime (Solenis, 2018).
This check does not fit closed recirculating loops, slug or batch feed, oxidizing biocides that are consumed by demand, or the first days after a new program starts, before the tower water reaches a steady concentration. It also cannot say whether the target dose itself is right; the treatment program sets that.
How do you calculate expected inhibitor usage from cooling tower makeup water?
Expected cooling tower inhibitor usage is the bleed volume times the target product concentration, and the bleed volume is the makeup volume divided by cycles of concentration. Chem-Aqua gives the mass balance as cycles = makeup / bleed and bleed = evaporation / (cycles - 1) (Chem-Aqua TB 2-025, 2018). Because the product stays in the recirculating water until that water leaves as bleed, the treatment program has to replace what each cubic meter of bleed carries away.
The working formula is:
Expected product use (g) = makeup volume (m3) / cycles x target product dose (mg/L)
One mg/L in one cubic meter of water is one gram, so the units close without a factor. Divide by the product density from the safety data sheet (SDS) or technical data sheet to get liters. Take cycles from the ratio of tower to makeup conductivity or chloride (Chem-Aqua TB 2-025, 2018), averaged over the same period as the makeup readings, not from the controller set point.
Use the dose as product, the way the feed rate is set. A target written as active ingredient, such as phosphate or a polymer, has to be converted with the product's active content first. In one patent example, a 100 ppm dose of treating agent contained 7.5 ppm of the active tagged polymer (US Patent 5,171,450, 1992), so mixing the two bases gives an error larger than any metering fault.
What does an expected cooling tower inhibitor usage calculation look like (illustrative)?
An expected cooling tower inhibitor usage calculation takes four readings: makeup volume, cycles, target dose and product density, all for the same period. The table below is an illustrative 30-day example; every value is invented to show the arithmetic and should be replaced with the site's own readings.
Item (illustrative) | Reading | Result |
|---|---|---|
Makeup meter, day 1 to day 30 | 900 m3 | Makeup volume |
Cycles, conductivity ratio averaged | 4.0 | Bleed = 900 / 4.0 = 225 m3 |
Target product dose | 100 mg/L | Expected use = 225 x 100 = 22,500 g |
Product density from the SDS | 1.15 kg/L | Expected use = 22.5 / 1.15 = 19.6 L |
Day tank drawdown plus drums added | 26.0 L | Gap = +6.4 L, or 33% above expected |
In this illustrative cooling tower case, the day tank lost 26.0 L of inhibitor while the makeup water called for 19.6 L, a gap of 33% above expected. A gap that size is not a reading error; it points to a cause in the table of usage gaps further down. A gap of a few percent, by contrast, is within the error of reading a tank level and a meter, so judge the gap against how precisely the site can read both. Repeat the calculation for the next period before changing the feed rate, because one period can include a one-off event such as a basin drain.
How do you measure actual inhibitor usage in a cooling tower?
Actual cooling tower inhibitor usage is the drop in product level in the day tank or drum between two dates, plus any product added in between, read on the same dates as the makeup meter. Record the tank level at the start and end, the number of drums or totes emptied into it, and convert to volume with the tank's level chart. When the product is fed from the shipping container, weigh it or read its level gauge the same way.
Check the feed pump separately with a drawdown. Blue-White recommends a physical drawdown, pumping the chemical into a calibration cylinder over a measured time, and calibrating with the chemical being dosed rather than water, after every service and at high, low and typical flow rates (Blue-White, 2025). The drawdown gives the pump's rate in mL/min; multiplied by its run time, it should match the tank drop.
Wear the protective equipment that section 8 of the product's SDS lists when opening drums or working on the suction line, and follow the site's chemical handling procedure. Read conductivity and flow meters regularly and keep a log of makeup and blowdown quantities, conductivity and cycles, as the US DOE Federal Energy Management Program (FEMP) recommends (US DOE FEMP, accessed 2026). For how to keep that product record per site and machine, see How do you record what product went into each customer site?.
What does it mean when actual cooling tower inhibitor usage is above or below the expected amount?
A gap between actual and expected cooling tower inhibitor usage means one of the inputs is wrong: the makeup meter, the cycles, the pump, the tank reading, or water leaving the tower that nobody counted. The table below sorts the gap by direction and by the tower water test, so the next check is clear.
Gap between actual and expected use | Likely cause | What to check |
|---|---|---|
Above, expected calculated from metered blowdown or set point, conductivity cycles below the makeup to blowdown meter ratio | Unmetered losses: overflow, leaks, pump seals, drift (Chem-Aqua TB 2-025, 2018); expected figure understated | Recalculate with conductivity cycles; then basin float, overflow line, valves |
Above, tower inhibitor or tracer test high | Pump overfeeding | Drawdown at the current setting; lower the feed rate |
Above, tower test on target | Makeup meter reads low, or a second makeup source is not metered | Meter calibration; every makeup source |
Below, tower test low | Feed pump lost prime, empty tank, blocked suction (Solenis, 2018) | Prime, foot valve, suction line, tank level |
Below, tower test on target | Makeup meter reads high, or a tank level misread | Meter; level chart and density used |
Either, makeup conductivity changed | Cycles at a fixed conductivity set point shifted (Solenis, 2018) | Makeup conductivity log for the period |
Work down the rows in order: the conductivity ratio and a pump drawdown can be checked during one site visit, while a meter calibration takes longer to arrange. The first row starts from a check a federal guide names directly: compare the makeup to blowdown flow ratio with the conductivity ratio. FEMP says that when the makeup to blowdown flow ratio and the conductivity ratio are not about the same, the tower should be checked for leaks or other unauthorized draw-off (US DOE FEMP, accessed 2026). In Lubinpla, the makeup volume can be kept as condition measurement history and the chemical dose as product usage records in the same project, so both can be read for the same period.
How do you confirm the inhibitor level in the cooling tower water itself?
The inhibitor level in the cooling tower water is confirmed by testing the tower water for the inhibitor actives or for a tracer added to the product, and reading the result against cycles of concentration. Chem-Aqua lists phosphate, phosphonate, azole and zinc as inhibitor actives, and PTSA and molybdenum as tracers used to monitor product levels (Chem-Aqua TB 2-027, 2024). Its bulletin states that "Cycles of concentration must be considered in relation to inhibitor tests," because inhibitor results may be proportionally high or low when cycles are (Chem-Aqua TB 2-027, 2024).
A tracer test separates feed problems from losses. A fluorescent tracer such as PTSA can be measured with a handheld or in-line fluorometer in about 30 seconds (Solenis, 2018). An inert tracer is not consumed, so its reading follows water gains and losses; an active polymer can read lower because part of it deposits on surfaces. In one patent example, the inert tracer indicated 105 ppm of treating agent while the tagged polymer indicated 80 ppm, and the difference was attributed to polymer taken up at scale and corrosion sites (US Patent 5,171,450, 1992).
Read the usage check and the tower test together. Usage on target with a low tower test points to losses or consumption in the system; usage high with the tower test on target points to the meters or the cycles figure.
Frequently asked questions
How long a period should the cooling tower usage check cover?
Use a period long enough that the drop in the day tank is large compared with the error of reading its level and the makeup meter, for example one month (illustrative). Start and end the period on dates when both readings are taken, and log any basin drain or refill in between.
Why does inhibitor usage not match in the first days of a new program?
After a new program starts, the tower water first has to reach the target concentration, so usage runs above the steady figure. In one patent example, it took 120 to 150 hours after blowdown began for the product dosage to fall from 200 ppm to the 95 to 100 ppm maintenance value (US Patent 5,171,450, 1992).
What changes when a cooling tower has two makeup water sources?
Proportional feed can lose control when there are several water sources (Solenis, 2018). Meter each source, add the volumes for the usage formula, and measure cycles against the conductivity of the blended makeup, because one source's conductivity gives the wrong cycles.