Moisture in transformer oil calculator: ppm, relative saturation and IEC 60422 classification
Convert the water in an oil between mg/kg and relative saturation at any temperature, find out when free water would appear, and classify the oil according to IEC 60422.
Water in mineral oil: ppm, relative saturation and saturation temperature
Convert water content between mg/kg (ppm) and relative saturation at the oil temperature, check at which temperature that same water would saturate the oil, and classify the oil against the IEC 60422 condition table.
Result
IEC 60422 classification
| Test | Value | Good | Fair | Poor | Condition |
|---|
Limits from the IEC 60422:2013 (edition 4) condition table as reported by CIGRE (Wang et al., 2015): water at operating temperature; interfacial tension for uninhibited oil; dissipation factor and resistivity only distinguish above and below 170 kV. The ends of each range count as fair. Edition 5.0 (2024) revised the action limits and merged category O into A: for a contractual criterion, check the edition in force.
Water solubility in new mineral oil (Oommen curve): log₁₀(Ws) = 7.0895 − 1567/T, with T in kelvin and Ws in mg/kg. It is the curve CIGRE uses to turn measured water into relative saturation. Aged or oxidised oil dissolves more water, so the saturation shown here is conservative: the real figure would be somewhat lower.
The relative saturation thresholds (5, 20 and 30%) are common maintenance practice, not a normative figure. The normative criterion, in mg/kg and by voltage class, is the IEC 60422 classification above.
Indicative result. The calculation runs in your browser and is not sent to any server. This website may contain errors or omissions: before making any decision about an installation, the information must be checked and validated by a qualified engineer. Legal notice and terms of use
What this calculator does
It converts the water content of a mineral oil between milligrams per kilogram (what a Karl Fischer test gives) and relative saturation in per cent (what an online capacitive probe gives), at whatever temperature you specify. It calculates two things that almost never appear in the report: the solubility of the oil at that temperature and the saturation temperature, below which that same water begins to appear as free water. And it classifies the oil as good, fair or poor using the IEC 60422 condition table, for water and for any of the other five routine tests you enter.
Why ppm is misleading
Water dissolved in oil does not matter for its absolute amount, but for how close it is to saturating the oil. And solubility depends very strongly on temperature. For new mineral oil:
log₁₀(Ws) = 7.0895 − 1567 / T
with T in kelvin and Ws in mg/kg. This is the Oommen curve, the one CIGRE uses to convert measured water into relative saturation. It gives about 55 mg/kg at 20 °C, 243 at 60 °C and 449 at 80 °C.
The practical consequence: an analysis of 25 mg/kg taken with the transformer hot looks reassuring —10 % saturation at 60 °C—, but that same water, when the unit cools down to 20 °C, is at 45 % saturation. That is why a sample without the oil temperature recorded is of little use, and why online probes report relative saturation rather than ppm.
The formula is for new mineral oil. An aged or oxidised oil dissolves more water, so the actual saturation would be somewhat lower than the value obtained here: the result errs on the conservative side.
The saturation temperature
This is the most useful figure in this calculator and the one that is rarely calculated. Given the measured water content, it indicates the temperature at which that oil would reach 100 % saturation. Below it, free water may appear, which is what really collapses the dielectric strength and causes failures on cold starts or after a prolonged shutdown in winter.
A transformer with 40 mg/kg saturates at around 12 °C. If it is installed in a room that drops below that temperature, the risk is not theoretical.
Classification according to IEC 60422
IEC 60422 is the guide for the supervision and maintenance of mineral insulating oils in electrical equipment. Its condition table, in the 2013 edition, distributes six routine tests across three states and three voltage classes:
| Test | Voltage | Good | Fair | Poor |
|---|---|---|---|---|
| Water (mg/kg at operating temperature) | > 170 kV | < 15 | 15 – 20 | > 20 |
| 72.5 – 170 kV | < 20 | 20 – 30 | > 30 | |
| ≤ 72.5 kV | < 30 | 30 – 40 | > 40 | |
| Breakdown voltage (kV, 2.5 mm) | > 170 kV | > 60 | 50 – 60 | < 50 |
| 72.5 – 170 kV | > 50 | 40 – 50 | < 40 | |
| ≤ 72.5 kV | > 40 | 30 – 40 | < 30 | |
| Acidity (mg KOH/g) | > 170 kV | < 0.1 | 0.1 – 0.15 | > 0.15 |
| 72.5 – 170 kV | < 0.1 | 0.1 – 0.2 | > 0.2 | |
| ≤ 72.5 kV | < 0.15 | 0.15 – 0.3 | > 0.3 | |
| Dissipation factor (90 °C, 40–60 Hz) | > 170 kV | < 0.1 | 0.1 – 0.2 | > 0.2 |
| ≤ 170 kV | < 0.1 | 0.1 – 0.5 | > 0.5 | |
| Resistivity (GΩ·m, 90 °C) | > 170 kV | > 10 | 3 – 10 | < 3 |
| ≤ 170 kV | > 3 | 0.2 – 3 | < 0.2 | |
| Interfacial tension (mN/m, uninhibited) | all | > 25 | 20 – 25 | < 20 |
Source: IEC 60422:2013, as reproduced by Wang et al., “Ageing Assessment of Transformers through Oil Test Database Analyses and Alternative Diagnostic Techniques”, CIGRE SC A2 Colloquium 2015, table 1. The voltage classes correspond to the equipment categories of the standard: O, A and D above 170 kV; B and E between 72.5 and 170 kV; C up to 72.5 kV.
What each state means: good, sampling and testing at the normal frequency are sufficient; fair, deterioration is already detectable and more frequent testing is advisable; poor, there is abnormal deterioration and corrective action should be considered, such as drying, reconditioning or regeneration of the oil. The calculator classifies only the tests that are filled in and summarises with the worst of them. A value exactly at the boundary of an interval counts as fair.
Two caveats. The water limit is given at operating temperature: a sample taken cold or hot cannot be compared directly with it, and that is where relative saturation helps. And the standard has a 2024 edition 5.0 that revised the action limits for all tests and merged category O into A; for a contractual criterion, check the values in the edition in force.
Worked example
This is the in-service case published by CIGRE (table 3 of the same paper): a 15-year-old transformer, a sample of 21.5 mg/kg taken with the oil at 57 °C and a breakdown voltage of 57 kV. These are the values the calculator loads when opened.
- Solubility at 57 °C: 220 mg/kg → saturation 9.8 %: moderate moisture.
- Solubility at 20 °C: 55.5 mg/kg → saturation 38.8 %: if the unit cools down, very wet oil.
- Saturation temperature: −1.0 °C.
CIGRE gives 38.7 % and 9.7 %; the difference is only due to rounding. Its conclusion is what the tool shows: measured in the laboratory at 20 °C, the oil looks much wetter than it actually is in the unit.
With the IEC 60422 classification: in a transformer above 170 kV, the 21.5 mg/kg is poor (the cut-off is 20) and the 57 kV is fair; between 72.5 and 170 kV, water becomes fair and breakdown voltage good; up to 72.5 kV, both are good. The same oil changes state depending on the unit it is in.
What this calculator does not do
It does not tell you how much water is in the paper, which is where most of it is and what has to be dried. Equilibrium between paper and oil takes days or weeks to establish and depends on temperature and load history, so estimating paper moisture from a single oil sample is unreliable. Nor does it correct breakdown voltage for temperature: the CIGRE paper revises it from 57 to 78 kV at 57 °C based on saturation, but using a relationship between saturation and breakdown voltage that the paper does not detail. It does not cover esters or silicone oils, whose solubility and limits are different, nor the IEC 60422 categories for tap changers and circuit breakers, and it does not replace the laboratory report.
How it is measured
Water content is determined by Karl Fischer titration in the laboratory, and relative saturation with capacitive probes installed in the oil circuit, which makes it possible to see the trend rather than a snapshot. The dielectric oil analyser and the other oil measurement equipment cover the testing side; drying and degassing, the treatment side. If the oil is wet, it is also worth looking at the dissolved gas analysis: paper degradation leaves a trace in CO and CO2.
Frequently asked questions
- Why are two oils with the same ppm of water not equally wet?
- Because what degrades the insulation is not the absolute amount of water but how close the oil is to saturation, and solubility depends very strongly on temperature: about 55 mg/kg at 20 °C versus about 243 at 60 °C. Twenty mg/kg is 36 % saturation in a cold oil and just over 8 % in one at operating temperature.
- What relative saturation is acceptable?
- The most widespread maintenance practice considers an oil dry below 5 %, moderately wet between 5 % and 20 %, wet between 20 % and 30 %, and very wet above 30 %. The standard uses a different criterion: IEC 60422 sets limits in mg/kg at operating temperature according to voltage class, and the calculator applies them in its classification block.
- What water limits does IEC 60422 give?
- In the condition table of the 2013 edition, above 170 kV the oil is good below 15 mg/kg, fair between 15 and 20 and poor above 20; between 72.5 and 170 kV the cut-offs are 20 and 30 mg/kg, and up to 72.5 kV, 30 and 40 mg/kg. These values apply at the equipment’s operating temperature. Edition 5.0 of 2024 revised the action limits, so for a contractual criterion the edition in force must be consulted.
- Where does the water in a transformer come from?
- From three sources: water that enters from the environment through gaskets, breathers or during an intervention; water released by the paper insulation as it ages, because cellulose degradation produces water; and water left inside since manufacture or the last treatment. Most of the water in a transformer is in the paper, not in the oil: the oil only shows the part that is in equilibrium with it.