How Scale Forms and What It Actually Costs

Why this matters

Scale is the most under-rated failure mechanism in the trades because it never presents as a failure. Nothing leaks, nothing trips, and the equipment runs. What actually happens is that a machine sized correctly on the day it was installed quietly stops being able to do its job, runs longer and longer to deliver the same output, and then dies of something that gets written up as a component defect. Customers replace a heater, a coil, or an exchanger, drop the identical unit into the identical water, and start the clock again.

Scale charges in three separate ways at once, and only one of them is obvious. Knowing the arithmetic of all three is what turns "you should probably descale that" into a recommendation somebody accepts.

Before you descale anything

Never mix an acid descaler with a hypochlorite product. The combination releases chlorine gas, and it happens fast in a small mechanical room or in a drain where one was poured after the other. Flush thoroughly between chemical steps and keep the two stored apart on site.

Acid attacking metal generates hydrogen gas. In a closed vessel or a sealed loop being circulated with acid, that gas accumulates, and it is flammable at low concentrations. Provide a vent path, keep ignition sources out of the space, and never cap a vessel that is actively being descaled.

Add acid to water, never water to acid, and wear the eye and skin protection named in Section 8 of the product's safety data sheet, which 29 CFR 1910.1200 requires be available to the person using it.

Before opening a heat exchanger or a heater, isolate it, relieve the pressure, verify at a gauge or an open bleed, and let it cool. Both the pressure and the stored heat are hazardous energy under 29 CFR 1910.147, and a scaled vessel can hold hot water long after the surrounding pipe feels cool.

How scale gets onto a surface

Calcium carbonate, the most common scale, has inverse solubility: it becomes less soluble as water gets hotter. So it comes out of solution precisely where the water is hottest, which is the metal surface you most need to keep clean. An immersion element, a fire-side tube, a brazed-plate exchanger channel, and a burner-side coil all scale on the hot face first and worst.

Three things drive the rate.

• Surface temperature, not bulk temperature. The film of water touching a heat transfer surface is much hotter than the tank. Raising a setpoint raises surface temperature more than proportionally in the places that matter, which is why the same water scales one system and not another.
• Concentration. In any system that evaporates water and makes up the loss, minerals stay behind and concentrate. The multiplier is the cycles of concentration: at four cycles, a makeup water carrying 320 mg/L of hardness leaves the loop running at about 1,280 mg/L, four times the source. Evaporative coolers, humidifiers, cooling towers and steam systems all work this way, and a system running at high cycles will scale on makeup water that would never scale a closed loop.
• Water chemistry. Hardness supplies the mineral, alkalinity and pH decide whether it stays dissolved. A water with a positive saturation index is laying scale down; a negative one is dissolving it.

Not all deposits are carbonate, and it matters, because the treatment differs. Carbonate scale is chalky, white to tan, and dissolves in acid with visible fizzing. Sulfate scale is hard, dense, and resists acid. Silica is glass-hard and does not respond to ordinary descaling at all. Iron and manganese deposits are red-brown or black and are usually a water-quality or corrosion problem rather than a hardness problem. Biofilm is slimy, soft, and comes back on a schedule if the underlying growth is not addressed. Fizz-testing a scraping in a little acid separates the first group from the rest in under a minute.

Cost channel one: heat transfer

Scale is a poor conductor. Its thermal conductivity typically runs one to two orders of magnitude below the metal it sits on, which is why a layer far too thin to see as a restriction dominates the whole heat path.

Work it as resistances per unit area, thickness divided by conductivity.

• A carbon steel tube wall 0.065 in thick is 0.00542 ft. At a conductivity of about 26 Btu per hour per foot per degree F, its resistance is about 0.00021 hour-square-foot-degree F per Btu.
• A calcium carbonate layer 1/32 in thick is 0.00260 ft. At a conductivity of about 1 Btu per hour per foot per degree F, which is in the middle of the published range for carbonate scale, its resistance is about 0.00260.

That scale layer contributes roughly 12.5 times the resistance of the steel wall it is sitting on, at less than half the wall's thickness. The metal was never the bottleneck.

Now put it in a system. Take an illustrative water-to-water exchanger with a clean overall heat transfer coefficient of 200 Btu per hour per square foot per degree F, which is a plausible mid-range value for that service and should be replaced by the real one where it is known. Clean total resistance is 1 divided by 200, or 0.00500. Add the scale layer's 0.00260 and total resistance becomes 0.00760, so the coefficient falls to about 132.

The coefficient dropped from 200 to 132, a loss of about 34 percent. To move the same heat with the same temperature difference, the unit has to run 200 divided by 132, about 1.52 times as long. A machine that ran 8 hours a day now needs about 12.2 hours for the same output, and on a design day, when there were never 12 hours of margin available, it simply falls short and the space or the water never reaches setpoint. That is the call the customer experiences as "it cannot keep up any more," and it arrives without a single fault code.

Cost channel two: flow and pressure

The same layer takes area away from the passage, and the arithmetic is unkind because both effects go with powers of the diameter.

A 1.000 in inside diameter tube with 1/32 in of scale on the wall loses that thickness from both sides, so the bore drops to 0.9375 in. Flow area scales with the square of the diameter: 0.9375 divided by 1.000, squared, is 0.879, so about 12 percent of the area is gone. Pressure drop is harsher. For turbulent flow at a fixed flow rate, treating the friction factor as roughly constant, pressure drop varies close to the fifth power of the diameter, so the ratio is 1.000 divided by 0.9375, raised to the fifth, which is about 1.38. Pushing the same flow now costs about 38 percent more pressure.

Change one thing and it gets much worse. The same 1/32 in layer in a 0.500 in bore leaves 0.4375 in. Area falls to 0.766 of clean, so about 23 percent gone, and the pressure drop factor is 1 divided by 0.875 raised to the fifth, about 1.95, nearly double. Small passages punish scale roughly twice as hard as large ones for the identical deposit thickness, which is why compact exchangers and small-bore tube foul out of service long before a large shell-and-tube in the same water shows a symptom.

In a pumped system with a fixed-speed pump, that extra pressure drop does not stay a pressure problem. The pump rides back up its curve and the flow falls, which reduces heat transfer further, on top of the 34 percent already lost.

Cost channel three: what the deposit hides and starts

• Overheating of the heat source. An element or a tube wall behind scale has to run hotter to push its heat through, and that higher metal temperature is what actually fails it. The element burnout gets logged as an element failure. It was a scale failure with a two-year fuse.
• Under-deposit corrosion. Anything sitting on a metal surface creates a crevice with stagnant fluid underneath, and that is where localized attack starts. Pinholes under scale are common, and they are why aggressive mechanical descaling sometimes reveals leaks rather than causing them.
• Sensors reading the wrong thing. A scaled sensor well responds slowly and reads closer to whatever is around it. Control quality degrades before anyone suspects the sensor.
• Valves and seats not sealing. Mineral particles on a seat hold it open a fraction, and the resulting seep is read as a failed valve.
• Safety devices masked. A scaled relief valve or a scaled low-water cutoff is a device that may not operate when it is needed, which moves this out of the efficiency conversation entirely. Exercise and verify those on schedule regardless of what the rest of the system looks like.

What changes the answer

• Closed loop versus open loop. A genuinely closed loop scales once, using up the minerals in its original fill, and then stops. If a closed loop keeps scaling, it is not closed: it is losing water and taking on fresh makeup, and the leak is the actual finding.
• Setpoint. Lowering a hot water setpoint slows scaling measurably, but it is bounded below by the temperature required for the application and by any legionella or sanitation requirement in force, so this is not a free adjustment and the applicable code or health guidance sets the floor.
• Softening versus balancing. Softening removes the hardness that forms scale and is the right answer when scaling is the problem. It is the wrong answer, and makes matters worse, when the water is already soft and aggressive; see the companion article on water chemistry for that distinction.
• Deposit type. Everything above assumes carbonate. If the fizz test says otherwise, an acid descale will not fix it and may damage the equipment while achieving nothing.

Measuring scale without opening the system

You rarely get to look. These four trends detect fouling from outside.

• Approach temperature. The gap between the leaving process temperature and the entering source temperature on a heat exchanger widens as fouling grows. Record it when the equipment is clean, and that baseline becomes the fouling gauge for the life of the machine. Without the clean baseline there is nothing to compare against, which is the reason to take it at commissioning rather than at the first complaint.
• Pressure drop across the same component at the same flow. The two conditions in that sentence are the whole test. A pressure drop reading at a different flow rate proves nothing, because drop varies with roughly the square of flow.
• Run time for the same duty on comparable days. The 1.52 multiple above shows up here first, as longer cycles, before it ever shows up as a temperature complaint.
• A scraping from an accessible surface, fizz-tested. Even a sight glass, a strainer basket, or a drain plug gives you a sample, and identifying the deposit type decides whether descaling is even the right procedure.

References

• 29 CFR 1910.1200, Hazard Communication, for the safety data sheet governing descaling chemical handling
• 29 CFR 1910.147, control of hazardous energy, for isolating and verifying pressure and stored thermal energy before opening a heater or exchanger
• Equipment manufacturer documentation for the approved descaling method, chemical, contact time and any materials in the unit that a given descaler must not touch
• Published thermal conductivity ranges for calcium carbonate scale and for carbon steel, which vary with deposit density and porosity; use the specific values where a supplier provides them
• See related: The Water Chemistry That Attacks a System; Pitting, Crevice and Uniform Corrosion