Transformer Moisture: Why the Same ppm Means Different Risk in Mineral Oil and Ester

1. Interpret moisture in the actual fluid

A water result in ppm cannot be transferred unchanged from mineral oil to an ester transformer. The dissolved-water capacity depends on fluid chemistry, temperature and ageing. Relative saturation describes how much of that capacity is occupied; it answers a different question from the mass concentration. Vaisala's original study compared several mineral oils, aged oils and a synthetic ester and found that comparing relative saturation gave a more useful view of moisture-related breakdown behaviour than comparing ppm alone. It is a laboratory study, not a universal field alarm curve. Moisture and breakdown voltage.

A maintenance decision therefore needs both the identified fluid and the temperature context. A high ester ppm result need not imply the same relative saturation as a lower mineral-oil result. Conversely, a reassuring warm-oil reading can conceal increasing saturation as the oil cools.

2. Request the missing inputs

Record fluid product and batch if available, natural or synthetic ester type, any mineral-oil mixing or retrofill history, sampling point, liquid temperature at sampling, load history and laboratory method. Ask whether the reported ppm is a laboratory mass measurement or a calculated sensor output. For an online calculation, obtain the solubility coefficients and their validation for the actual liquid.

Water activity is expressed as a fraction of saturation; percent relative saturation is 100 times that fraction. Vaisala distinguishes real-time saturation measurement from ppm output. Do not merge the two channels in a trend without preserving their provenance. Moisture measurement.

3. Separate sampling temperature from tank temperature

A monitor may cool the sample before measuring it. Vaisala's OPT100 guidance specifically warns that the measurement chamber temperature does not represent tank or pipe temperature, and that the measured saturation is consequently not automatically the in-tank value. A conversion needs a temperature and solubility model appropriate to that fluid. Temperature-location correction.

For laboratory sampling, use the approved sampling procedure and containers, document the sampling point, avoid water contamination and preserve the chain of custody. Ask the laboratory to state uncertainty and whether visible free water or handling irregularities were observed. A bottle result does not recreate the transformer's full daily thermal cycle.

4. Illustrative calculation: equal ppm, unequal saturation

Use the simple dissolved-water relationship RS = 100 × c / S(T), where c is water concentration in ppm by mass and S(T) is the fluid's saturation concentration at the stated temperature. The example assumes a valid single-liquid model, dissolved water and no loss or gain of water during the comparison.

Assume 30 ppm and illustrative saturation capacities of 60 ppm for mineral oil and 600 ppm for an ester at the same temperature. Then 100 × 30 / 60 = 50% RS, while 100 × 30 / 600 = 5% RS. These capacities are invented calculation inputs, not product specifications. If the mineral oil later cools to a condition with capacity 40 ppm while concentration stays 30 ppm, its calculated saturation becomes 100 × 30 / 40 = 75% RS.

The conclusion is to obtain validated fluid-specific data and examine the cold condition. None of 50%, 5% or 75% is proposed as a universal alarm. In real transformers water also migrates between liquid and cellulose, so constant ppm during cooling is an explicit simplifying assumption.

5. Use each result for its proper decision

Oil moisture is not a direct paper-moisture percentage. Do not prescribe drying from one bottle alone: repeat a questionable sample, reconcile it with temperature and loading, and consider dielectric-response testing and other insulation evidence when assessing the solid insulation.

6. Specify useful monitoring and procurement evidence

Require raw concentration, saturation and temperature channels where available; sensor location; conversion coefficients; calibration records; fluid compatibility; and a written response plan owned by the asset engineer. During retrofill, establish a new baseline rather than forcing the old mineral-oil trend onto the new liquid. Include laboratory cross-checks under comparable conditions and retain sampling metadata with every result.

The ester selection guide explains fluid choice; the retrofill compatibility review addresses equipment interfaces. Neither replaces a moisture model for the actual filled mixture.

7. Frequently asked questions

Is higher ester ppm automatically worse?

No. Compare concentration with validated solubility at the relevant temperature and review the insulation evidence.

Does low RS prove dry paper?

No. Liquid and solid insulation exchange moisture over time; one liquid measurement does not directly quantify cellulose moisture.

Should the monitor and laboratory always agree?

Compare like quantities, sampling times and locations first. Calculated ppm, measured ppm and saturation at different temperatures are not interchangeable.

8. Primary references

• Vaisala: moisture and breakdown voltage
• Vaisala: temperature-location correction
• Vaisala: moisture measurement