Moisture in Transformer Insulation: Karl Fischer Checks Buyers Should Require

The enemy that entered before the tank was sealed

If you asked a group of electrical engineers what kills transformers, most would answer "overload" or "lightning" or "old age." They would be wrong about the ranking. The most destructive agent inside a transformer is not current or voltage. It is water.

Water accelerates the chemical reactions that break down cellulose insulation. It does this continuously, from the day the transformer is sealed until the day it fails. And in most transformers destined for export markets, the water that will determine the unit's service life enters during the final days of manufacturing—before the tank is bolted shut, before the oil is filled, before the first electrical test is performed.

The moisture content of the insulation at the moment of sealing is a number that the buyer never sees and the manufacturer rarely highlights. It should be the first number you ask for after the ratio and impedance.

The chemistry: what water does to cellulose

Transformer insulation paper is approximately 90% cellulose—a polymer of glucose molecules linked by glycosidic bonds. These bonds are the mechanical backbone of the paper. When they break, the paper loses tensile strength. It becomes brittle. It can no longer withstand the electromagnetic forces that act on windings during faults and inrush events.

Water attacks glycosidic bonds through hydrolysis: a water molecule inserts itself into the bond, splitting the polymer chain. The reaction is acid-catalyzed, meaning that once degradation begins, the organic acids produced by cellulose breakdown accelerate further degradation. It is a self-feeding cycle: water breaks cellulose → acids form → acids catalyze more hydrolysis → more cellulose breaks.

The rate of this reaction follows the Arrhenius equation, which governs chemical kinetics: reaction rate increases exponentially with temperature. This is why the combination of moisture and heat is far more destructive than either factor alone. A transformer with elevated moisture content that operates under high load—and therefore at elevated winding temperature—is aging on an accelerated timeline compared to a dry transformer under the same load.

What the numbers mean

The relationship between moisture content in solid insulation and expected life reduction is documented in the technical literature of transformer diagnostics:

Moisture content of 1% in the solid insulation: the transformer is in excellent condition. Life expectancy is governed by normal thermal aging at design temperature.
Moisture content of 2%: life expectancy reduced by approximately 40% compared to a dry unit. The degradation is already underway, though the transformer may operate for years before symptoms appear.
Moisture content of 3%: life expectancy reduced by approximately 50%. At this level, the paper is losing mechanical strength at a measurable rate.
Moisture content of 4%: life expectancy reduced by approximately 70%. The transformer is on a path to premature failure. The paper will become brittle enough to crack under fault forces long before it would have aged out under thermal degradation alone.

These are not precise predictions for any single unit—operating conditions, loading patterns, and maintenance history all influence actual life. But they represent the direction and magnitude of the effect. More moisture equals shorter life. The physics is unambiguous.

How moisture enters the transformer

Moisture does not enter through leaks during operation—or rather, that is the less common path. Most moisture enters during manufacturing, through one of three routes:

Residual moisture in solid insulation. Cellulose paper is hygroscopic. It absorbs water from ambient air during storage and handling. If the core and windings are assembled, processed, and oil-filled without a complete drying cycle, the water absorbed during storage stays inside.

Moisture in the oil at filling. Transformer oil is hygroscopic, though less so than paper. If the oil is not processed through vacuum dehydration and filtration immediately before filling, it carries dissolved water into the tank. The paper will eventually absorb most of this water, reaching equilibrium over weeks to months.

Exposure during assembly. In high-humidity manufacturing environments, the time between removal from the drying oven and completion of oil filling is critical. Cellulose can absorb measurable moisture from ambient air in minutes. The longer the exposure, the more water is sealed inside.

The scenario: same factory, same spec, different drying protocols

Two transformers of identical design are manufactured in the same factory. Unit A undergoes the full drying protocol: the core and windings are placed in a vacuum oven, heated under controlled vacuum until the dew point stabilizes below -50°C, and then immediately oil-filled with vacuum-degassed and dehydrated oil without exposure to ambient air. Unit B is processed on an accelerated schedule to meet a shipping deadline. The drying cycle is shortened. The oil is filled from a storage tank rather than freshly processed.

Both units pass routine electrical tests. Both are shipped. Five years later, Unit A is operating with normal maintenance intervals. Unit B's oil shows elevated moisture, rising acidity, and declining dielectric strength. Its insulation is aging years faster than its calendar age suggests.

The difference between the two units was determined in the factory, not in the field. And it was measurable: a Karl Fischer titration of the oil at the time of FAT would have shown Unit B's moisture content higher than Unit A's—possibly within the IEC 60814 limit of 20 ppm, but higher nonetheless. The moisture wasn't high enough to fail a test. It was high enough to subtract years from the transformer's service life.

Why Karl Fischer matters

IEC 60814 specifies the Karl Fischer coulometric titration method for determining water content in transformer oil. Acceptance limits should follow the project specification, utility requirement and supplier agreement; 20 mg/kg (ppm) is commonly used as a procurement reference for new oil or oil before energization.

But here is the subtlety that matters for procurement: the Karl Fischer measurement tells you the water content of the oil, not the paper. The paper contains far more water than the oil—typically 100 to 1000 times more, by mass. The relationship between the two is governed by moisture equilibrium curves that depend on temperature and paper type.

A Karl Fischer measurement of 30 ppm in the oil might correspond to 3–4% moisture in the paper if the transformer is cold. The same 30 ppm at operating temperature might correspond to 1–2% in the paper. The absolute number matters, but so does the temperature at which the sample was taken. This is why DGA laboratories report moisture not just as ppm, but as percent moisture by dry weight of the solid insulation, calculated using the temperature at the time of sampling.

What a procurement manager should know: ask for the Karl Fischer moisture value at FAT. And if the value is not provided at all—if moisture is simply not measured—that is a stronger signal than an elevated value. It means the manufacturer does not control what is arguably the single most important variable affecting transformer life.

What to specify

In your next RFQ, add one line to the testing requirements:

"Water content in oil shall be measured by Karl Fischer titration in accordance with IEC 60814. The measured value shall be reported in mg/kg (ppm). Additionally, the percent moisture by dry weight of the solid insulation, calculated at the temperature of sampling, shall be reported. Acceptance criteria: ≤20 ppm for oil; ≤2% estimated moisture in solid insulation."

If your supplier cannot provide this data, ask yourself: what else about the manufacturing process are they not measuring?

Next: short-circuit withstand—the forces that bend conductors and the type test that most procurement specifications omit.

→ Discuss moisture control requirements with our technical team

References: IEC 60814 (Determination of water content in insulating liquids by automatic Karl Fischer coulometric titration; 20 mg/kg is commonly used as a procurement reference for new oil, subject to project specification). Cellulose chemistry: β-1,4-glycosidic bonds, hydrolysis mechanism, autocatalytic effect of organic acids. Arrhenius equation applied to cellulose degradation: reaction rate ∝ exp(−Ea/RT), with Ea typically 100–130 kJ/mol for cellulose hydrolysis. Moisture equilibrium between oil and paper: Oommen curves, temperature dependence. Life reduction estimates: 2% moisture ≈ 40% reduction, 3% ≈ 50%, 4% ≈ 70%, as documented in transformer aging literature (degree of polymerization studies).

Related Transformer Failure and Procurement Guides

Procurement Action

If you are comparing transformer suppliers, send voltage, kVA, application, country, utility requirements and any FAT or test requirements. TransformerGrid can help review the procurement risk before the order is placed.

Moisture: Why Water Destroys Transformer Insulation Faster Than Overload