When should I change my coolant? A guide to degradation thresholds

One fluid, three condemning limits

If you open three reputable technical sources and search for the total acid number at which a heat transfer fluid should be changed, you will find three different numbers. One source puts the condemning limit at a TAN of 1.0. A second says 0.5. A third says 3.0. A fourth warns that negative effects — insolubles, sludge, fouling of heat-transfer surfaces and localised overheating — can begin at a TAN as low as 0.4.

All four sources are referring to the same parameter, measured by the same general method, on the same class of fluid. They disagree because they were written for different fluids, in different system types, at different operating severities. A thermal oil in a vented extrusion process at 280 °C degrades by a different mechanism and at a different rate from a glycol coolant in a closed chilled-water loop at −5 °C. The number that condemns one does not condemn the other.

This article is about how to read the measurements, rather than which number to believe. It draws on published guidance for thermal oils and water-based secondary coolants, and on the monitoring instructions printed in the Glacier catalogue for specific product families. It does not prescribe a change interval for any specific system.

What the catalogue tells you to monitor

The Glacier catalogue prints monitoring instructions for several product families, and reading them together reveals a pattern that is more useful than any single number.

For the low-conductivity LM-4D grade, the catalogue directs that conductivity and pH be tested every three months. If the conductivity rises above the specified value or the pH falls below it, the user is directed to contact the supplier for adjustment. The stock solution reference is below 1 μS/cm at 20 °C, and the 50% dilution reference is below 5 μS/cm.

For the LM-4D-YE grade, the instruction is to test the pH value every three months and contact the supplier if it drops below the specified value.

For the LMH brine inhibitor programme, the catalogue states that pH and anti-corrosion performance should be tested every three months during use, with pH adjusted and LMH replenished according to the technical standard — generally about 0.5%.

For the LMT cooling-water treatment, the instruction is proportional: replenishment is done regularly according to the amount of make-up water added, not according to a calendar interval.

The common thread is the structure: a parameter is named, an interval is given, a limit is referenced, and an action follows. The limit is stated generically — "the specified value" — because it depends on the concentration, the system materials and the water chemistry. The principle is that monitoring is a programme, not a number, and the trend across several readings decides whether the action is needed.

Why the numbers do not agree

The TAN spread — 0.4 to 3.0, depending on the source and the fluid — is the most visible example, but it is not the only one. The reasons are instructive.

One reason is that different fluids start at different baselines. A new synthetic thermal oil may have a TAN below 0.05. A polyalkylene glycol fluid may start at 0.1 to 0.5 and still be within specification. A condemning limit of 1.0 applied to a fluid that started at 0.5 has a different meaning from the same limit applied to a fluid that started at 0.02.

A second reason is that different systems tolerate different levels of degradation products. A large, closed, well-filtered loop with conservative operating temperatures can tolerate a higher TAN than a small, open, unfiltered system running near the fluid's thermal limit, because the degradation products are diluted in a larger volume, filtered out more effectively and formed more slowly.

A third reason is that the parameter you do not measure is the one that will surprise you. A TAN that is stable while the viscosity is climbing and the flash point is dropping suggests that the fluid is polymerising or cracking, not just oxidising. Monitoring one parameter in isolation will miss the degradation mechanism that is actually at work.

Read the trend, not the threshold

The following approach applies to thermal oils, glycol-based coolants and inhibited brines alike, with the parameters adjusted for the fluid type.

Take a baseline measurement of the fresh fluid at the time of charging: TAN, viscosity, flash point for thermal oils; pH, conductivity, reserve alkalinity and glycol concentration for water-based fluids. Record the values, the test methods and the laboratory. This is the single most valuable data point you will ever have for that charge, and it is the one most often omitted.

Sample at the agreed interval — the catalogue references three months as a starting point for several product families; published industry guidance for data-centre glycol loops ranges from three to six months. Use the same sampling point, the same sample container and the same laboratory each time, because a change of method can produce a shift in results that looks like a trend but is not.

Compare each result against the previous two. A single reading outside a limit may be a sampling error, a transient contamination or a laboratory artefact. Two consecutive readings moving in the same direction, and accelerating, are a genuine trend.

The shapes that matter:

• A TAN that is rising slowly and linearly is being controlled by the inhibitor package. A TAN that begins to rise faster from one interval to the next has exhausted the inhibitor and is entering the autocatalytic region, where degradation products accelerate further degradation. That is the point at which the fluid is on a path to sludge, fouling and corrosion, regardless of whether it has crossed an absolute threshold.
• A pH that has not moved does not mean the inhibitor is intact. Glycol inhibitors buffer the pH, and the reserve alkalinity drops before the pH does. Measure reserve alkalinity alongside pH.
• A conductivity that is rising in a low-conductivity loop may indicate inhibitor depletion, ionic contamination from make-up water, or corrosion products dissolving into the fluid. It cannot tell you which without additional measurements — metals analysis, inhibitor concentration and a visual inspection of the fluid.

A monitoring schedule you can use tomorrow

For a water-based secondary coolant, the minimum monitoring set at each sampling interval:

For a thermal oil, the equivalent set:

In both cases, record the results against the baseline and against the previous two readings. If two parameters are moving in the wrong direction and accelerating, prepare for a fluid change. If only one parameter has moved and the others are stable, re-sample before acting.

What this article does not prescribe

It does not give you a TAN, pH or conductivity number at which your fluid must be changed, and it does not set a sampling interval for your system. The monitoring intervals printed in the Glacier catalogue are reference values for the products and conditions they describe; the interval your system needs depends on its size, its operating severity, its materials and the consequences of a failure.

It also does not replace a fluid analysis report interpreted by the supplier. The numbers in this article are drawn from published sources that disagree with each other, and the value of an individual reading can only be assessed against the baseline of your specific fluid and the trend of your specific system. The system maintenance solution page sets out how treatment programmes fit into the wider maintenance picture, and the technical document request page covers obtaining current product documentation and monitoring advice for your installation.