Effects of Water on Dissolved Gas Analysis

The chemistry of liquid-filled power transformers that use mineral oil, synthetic hydrocarbon insulating liquids, natural ester fluids, or synthetic ester fluids with Kraft paper and pressboard solid insulation can be very complex. This paper describes how water is involved in the chemistry leading to the formation of dissolved gassess.

Transformer Moisture

Water affects the chemical composition of the transformer components which are detected by dissolved gas analysis (DGA). Water concentration is measured using different methods. Measuring very small concentrations of water is difficult because of the constancy of atmospheric moisture as well as the attraction of water to surfaces.

ASTM D1533 covers coulometric Karl Fischer determination of water in insulating liquids.[1] IEC 60814 covers water determination in insulating liquids and oil-impregnated paper/pressboard by coulometric Karl Fischer titration.[2]

The four meanings of transformer moisture

Moisture information is separated into four related but different quantities.

Quantity	What it means	Diagnostic limitation
Water in oil, ppm or mg/kg	Absolute mass of water in the oil, measured by Karl Fischer titration.	Strongly dependent on fluid type, temperature, aging, and total water in the systems.
Relative saturation, %RS	The fraction of the water concentration related to the total water-holding capacity.	Requires temperature and known water-solubility for the specific fluid.
Paper moisture	Water held in Kraft paper and pressboard.	Not measured directly by a routine oil Karl Fischer result. Estimation requires equilibrium assumptions or dielectric response methods.
Dielectric breakdown strenght	A stress-withsand test for the oil samples.	A low result my be due to water, particles, fibers, bubbles, or other contamination. A high result does not prove that the paper is dry.

Water concentration in ppm tells how much water is in the oil; relative saturation tells how close the oil is to its water-holding limit. Relative saturation is often more useful than ppm for dielectric risk, moisture mobility, and comparison between temperatures or fluid families. Moisture equilibrium studies describe relative humidity or relative saturation in oil as the dissolved water content relative to the maximum moisture the oil can hold at that condition.[3]

Once the moisture reaches the 100% relative saturation any more water introduced will be visible as water droplets or increased airspace humidity.

Dielectric strenght does not measure moisture

Dielectric breakdown voltage measures the ability of the liquid sample to withstand electrical stress. ASTM D877 states that breakdown voltage is reduced by contaminants such as cellulosic fibers, conducting particles, dirt, and water.[4] IEC 60156 specifies a power-frequency breakdown-voltage method for insulating liquids.[5]

Use dielectric strength as a “reality check” when analyzing transformer oils.

Observation	Interpretation
High water ppm but acceptable dielectric strength	The oil may still be below critical relative saturation, or contaminants may be well controlled. Check temperature and %RS.
Low dielectric strength with high %RS	Water is a plausible contributor, especially if particles or fibers are also present.
Low dielectric strength with low water ppm	Suspect particles, fibers, bubbles, oxidation products, or sample contamination.
DGA is typical and dielectric strength is low	This may be a liquid-condition problem rather than an active gas-forming event.

Relative saturation

Relative saturation (RS) of water in any liquid at a given temperature is calculated from,

where W_oil is the Karl Fischer water result in ppm or mg/kg, and Wsat is the saturation water content of that fluid at the sampling temperature using the same units.

Because W_sat increases strongly with temperature, the same water-in-oil ppm value can mean a low relative aturation when the oil is hot and a much higher relative saturation when the oil is cool. This is why moisture results need a measured temperature. Field moisture guidance from Doble illustrates that the same ppm value can correspond to very different relative saturation as oil temperature changes.[6]

Practical use:

• Use ppm to track the absolute water content of the sampled liquid.
• Use relative saturation to assess how close the system is to yielding free-water.
• Always record oil temperature, sampling location, fluid type, and recent load/cooling conditions.
• Do not compare mineral oil ppm directly with ester-fluid ppm. The fluids do not have the same water solubility.

Water reactions in Kraft paper

Kraft paper is mostly cellulose and its strength depends strongly on the degree of polymerization (DP). Water breaks cellulose chains through hydrolysis, which is reaction with water, especially in the presence of acids. Lundgaard and coworkers showed that moisture, oxygen, and acidity strongly affect oil-impregnated paper aging, and that moisture promotes acid hydrolysis of Kraft paper.[7]

Acid hydrolysis lowers DP; it breaks the chain. There are usually no immediate DGA results. Instead, the paper becomes mechanically weaker and produces more oxygenated degradation fragments that subsequently form CO, CO₂, furans, methanol, ethanol, acids, and water.

Acid-catalyzed cellulose hydrolysis

Oxidation of oils and paper produce carboxylic acids. These are weak organic acids having a -COOH atomic group:

Here “R” is a gerneric molecular chain, and [O] is an undifferentiated oxygen source. The acids dissociate readily in water to release a hydrogen ion (a proton):

The hydrogen ion attacks the cellulose oxygen making the cellulose linkage weaker. Then water cleaves the bond and the hydrogen ion is regenerated. Lundgaard’s acid-hydrolysis work found a clear synergy between low molecular-weight acids and water content in accelerating Kraft paper aging.[8]

How water influences CO and CO₂ formation

Water accelerates cellulose chain scission and increases the pool of oxygenated fragments. Each reaction shortens the cellulose chains. The ends of the shortened chains are aldehyde (-CHO) or acid (-COOH) groups. These groups react to form CO and CO2. In general, the paper decomposition produces a large variety of products:

DGA interpretation: rising CO and CO₂ in a wet transformer should increase suspicion of cellulose involvement. The observance of acidity, furans, methanol, or ethanol, are the strong indicators of paper involvement.

Water, partial discharge, and hydrogen

Under electrical stress water participates in discharge and radical chemistry. Radicals are represented by the • notation.

Results in hydrogen formation:

The hydroxyl (OH•) radical attack the oil:

Further reactions of the oil give the usual DGA gases:

Moisture lowers dielectric strength and can enable partial discharge, tracking, bubble formation, or breakdown. Therefore, an H2-rich trend in a wet transformer should raise the question: is moisture enabling electrical stress?

Water and ester fluids

In the presence of water ester liquids are chemically different from mineral and synthetic hydrocarbon oils. IEEE C57.155 is the primary guide for DGA interpretation in natural ester and synthetic-ester immersed transformers.[9] Natural esters and synthetic esters differ slightly as well. The shared ester (R−COO−R′) hydrolysis reaction is:

where R and R’ are separate parts of
the ester molecule.

Natural ester oils
Natural ester transformer fluids are commonly triglyceride-based vegetable oils. The hydrolysis reaction is:

triglyceride + water⟶ glycerol + fatty acids

Natural esters generally hold much more water than mineral oils, and this changes moisture distribution in the liquid-paper system. [10]

Synthetic ester oils
Synthetic ester transformer oils are engineered ester molecules. A generic synthetic polyol ester hydrolysis reaction can be represented as:

synthetic ester + water ⟶ polyol alcohol + organic acids

Consequences of water in esters:

• A natural ester may show higher water ppm than mineral oil.
• The ester liquid may draw water from the paper changing the equilibrium moisture content.
• Natural esters produce glycerol
• Synthetic esters are designed for improved oxidation and thermal properties; but water, heat, and acid remain a degradation pathway.
• Hydrolysis produces fatty acids; acid interpretation differs from mineral oil because the acids remain soluble in the ester and are usually larger molecules.
• CO and CO₂ can reflect paper degradation, ester chemistry, oxidation, or mixed mechanisms; interpret with ester-specific trends and IEEE C57.155.
• CO/CO₂, hydrogen, and hydrocarbon gas trends should be interpreted using ester-specific guidance rather than mineral-oil only ratio intuition.

Studies comparing insulation aging in mineral oil, natural esters, and synthetic esters emphasize that ester-based systems create new interpretation questions because the liquid-paper interactions differ from mineral oil systems.[11],[12]

Moisture-aware DGA interpretation

Moisture should not be interpreted as just another DGA gas. It is a variable that changes how the gases are formed and how patterns are diagnosed.

Observation	Moisture-related interpretation	Recommended diagnostic question
High Karl Fischer ppm, low relative saturation	Liquid contains water, but not close to saturation at sample temperature.	Was the oil hot? What is the fluid type and saturation curve?
Moderate ppm, high relative saturation	Dielectric risk may be higher than ppm suggests.	Was the oil cool? Is free water plausible?
Rising H2 with high %RS	Moisture-assisted partial discharge is possible.	Do dielectric strength, PD tests.
Rising CO and CO2 with rising water	Hydrolytic/oxidative paper aging becomes more plausible.	Are acidity, furans, methanol, ethanol, or load history changing?
C2H2 appears in a wet unit	Sparking/arcing must be considered, but moisture may be enabling breakdown rather than directly forming acetylene.	Is this main-tank oil, OLTC contamination, or a new event?
Low dielectric strength, DGA typical	Water/particles/fibers may be a liquid-condition issue before a gas-forming event.	What is particle count, filtration history, and sample quality?
After drying or oil processing, moisture and gases fall	The diagnostic baseline may have been reset.	Compare future samples to the new post-processing baseline.

Suggested measurement and interpretation workflow

Identify the liquid: mineral oil, synthetic hydrocarbon, natural ester, or synthetic ester.
Record sample temperature and sampling location with every Karl Fischer result.
Report water both as ppm or mg/kg and as relative saturation when possible.
Check absolute gases. What is present now?
Review DGA trends, especially H2, C2H2, CO, and CO2, against moisture trends.
Apply interpretation methods. Rogers, IEC, Duval, and ester specific guidance when applicable.
Look for agreement. Do methods, chemistry, and transformer history tell the same story?
Assign likely fault family. PD, D1, D2, T1, T2, T3, paper aging, stray gassing, contamination, or mixed fault.
Decide next evidence. Repeat DGA, online monitoring, electrical testing, infrared spectroscopy, furan or alcohol markers, moisture, oil quality, load or cooling review, or inspection.

References:

[1] ASTM International, ASTM D1533: Standard Test Method for Water in Insulating Liquids by Coulometric Karl Fischer Titration. West Conshohocken, PA: ASTM International.
[2] International Electrotechnical Commission, IEC 60814:1997, Insulating liquids Oil-impregnated paper and pressboard – Determination of water by automatic coulometric Karl Fischer titration. Geneva: IEC.
[3] Y. Du, M. Zahn, B. C. Lesieutre, A. V. Mamishev, and S. R. Lindgren, “Moisture equilibrium in transformer paper-oil systems,” IEEE Electrical Insulation Magazine, vol. 15, no. 1, pp. 11–20, 1999.
[4] ASTM International, ASTM D877/D877M: Standard Test Method for Dielectric Breakdown Volt-age of Insulating Liquids Using Disk Electrodes. West Conshohocken, PA: ASTM International.
[5] International Electrotechnical Commission, IEC 60156: Insulating liquids – Determination of the breakdown voltage at power frequency – Test method. Geneva: IEC. [6] L. Lewand, “Understanding Water in Transformer Systems,” Doble Engineering Company technical paper, 2008.
[7] L. E. Lundgaard, W. Hansen, D. Linhjell, and T. J. Painter, “Aging of oil-impregnated paper in power transformers,” IEEE Transactions on Power Delivery, vol. 19, no. 1, pp. 230–239, Jan. 2004. doi:10.1109/TPWRD.2003.820175.
[8] L. E. Lundgaard, W. Hansen, D. Linhjell, and T. J. Painter, “Aging of Kraft paper by acid cat-alyzed hydrolysis,” Proceedings of the IEEE International Conference on Dielectric Liquids, Coimbra, Portugal, 2005, pp. 94–97.
[9] IEEE Power and Energy Society, IEEE Std C57.155-2014, IEEE Guide for Interpretation of Gases Generated in Natural Ester and Synthetic Ester-Immersed Transformers. Piscataway, NJ: IEEE, 2014.
[10] R. D. Villarroel Rodríguez, “Moisture dynamics in natural-ester filled transformers,” doctoral the-sis, Universidad Carlos III de Madrid, 2021.
[11] K. Bandara, C. Ekanayake, T. K. Saha, and H. Ma, “Performance of natural ester as a transformer oil in moisture-rich environments,” Energies, vol. 9, no. 4, article 258, 2016. doi:10.3390/en9040258.
[12] T. Münster, M. Beltle, S. Tenbohlen, and M. Koch, “Thermally accelerated aging of insulation pa-per for transformers with different insulating liquids,” Energies, vol. 14, no. 11, article 3036, 2021. doi:10.3390/en14113036.