Insulating Oil Aging and DGA Diagnosis in Pad-Mounted Transformers

1. What This Article Covers

This article examines the aging mechanisms of insulating oil in liquid-filled pad-mounted distribution transformers and the use of dissolved gas analysis (DGA) to detect and classify incipient faults. It covers:

• How oil degrades: the chemical and physical processes behind oxidation, moisture-related deterioration, and thermal cracking of oil molecules.
• How DGA works: which gases are produced by which fault types, and how the IEC 60599 three-ratio method translates gas concentrations into fault classifications.
• How to use DGA in procurement: what oil test requirements to specify, what baseline values to expect for new oil, and how to interpret factory oil test reports.
• What happens when oil diagnosis is skipped: the hidden cost of deferred sampling and the failure pathways that develop without warning.

2. What Insulating Oil Does: Dielectric and Cooling Functions

Insulating oil in a pad-mounted transformer performs two inseparable functions. Its failure in either function can cascade into failure of the other, and ultimately into failure of the solid insulation that determines transformer life.

2.1 Dielectric Function

The oil fills the space between windings, between windings and the core, and between the active part and the grounded tank. Its dielectric strength—typically above 30 kV for a 2.5 mm gap in new mineral oil for equipment up to 69 kV—must be maintained for the design life of the transformer. The oil's dielectric function depends on its purity: dissolved water, conductive particles, and polar aging byproducts all reduce the voltage at which breakdown occurs.

The physical mechanism of oil breakdown is fundamentally different from solid insulation breakdown. In a pure liquid, breakdown begins with field emission or ionization at microscopic irregularities on electrode surfaces. In service-aged oil, the dominant mechanism is the "small bridge" theory: impurity particles (cellulose fibers, carbonized oil residue, metal wear particles) polarized by the electric field align into chains that bridge the gap between conductors. Once a chain forms, current flows through the particles, heating the surrounding oil, creating a gas bubble, and triggering full breakdown within the bubble.

2.2 Cooling Function

Transformer losses—no-load (core) loss and load (winding) loss—generate heat. The oil carries this heat from the core-and-coil assembly to the tank walls by natural convection. The cooling effectiveness depends on oil viscosity (which increases as oil oxidizes and forms sludge), on the absence of flow obstructions (sludge deposits on windings and in cooling ducts), and on the external heat dissipation path (radiator fin cleanliness, ambient air circulation).

A transformer that is thermally well-designed for new oil can develop hot spots as the oil ages and its viscosity rises. The winding hot-spot temperature is the primary driver of cellulose insulation aging: each 6–8°C increase above the rated hot-spot temperature approximately doubles the thermal aging rate.

3. How Insulating Oil Ages: Three Degradation Pathways

3.1 Oxidation: The Slow, Continuous Degrader

Mineral insulating oil is a mixture of hydrocarbon molecules—paraffinic, naphthenic, and aromatic. Oxygen attacks these molecules at the carbon-hydrogen bonds, particularly at elevated temperatures. The reaction proceeds in stages:

1. Initiation: Heat or ultraviolet radiation breaks a C–H bond, creating a free radical. Dissolved oxygen attacks the radical, forming a peroxide.
2. Propagation: The peroxide radical attacks another hydrocarbon molecule, creating a hydroperoxide and a new free radical. The reaction becomes self-sustaining.
3. Termination: Two free radicals combine to form a stable molecule, or an antioxidant (inhibitor) interrupts the chain.

The products of oxidation include organic acids (measured as the neutralization number or total acid number, TAN), aldehydes, ketones, and eventually sludge—high-molecular-weight polymerized compounds that precipitate out of the oil. Acids attack cellulose insulation, reducing its mechanical strength. Sludge deposits on windings impede heat transfer, raising hot-spot temperatures, which accelerates further oxidation in a feedback loop.

Procurement relevance: Inhibited oil (containing an oxidation inhibitor, typically 2,6-ditertiary-butyl para-cresol, DBPC, at 0.3% by weight for new oil) resists oxidation longer than uninhibited oil. Natural ester fluids are inherently more resistant to oxidation than mineral oil but have different dielectric and thermal characteristics. Specifying the inhibitor type and concentration in new oil, and requesting a factory certificate of analysis, establishes a known baseline.

3.2 Moisture: The Silent Accelerator

Water enters transformer oil through four paths: leakage past gaskets and seals, diffusion through conservator diaphragms (if fitted), desiccant breather saturation, and—critically—as a byproduct of cellulose insulation aging. Cellulose (paper, pressboard) is a glucose polymer. When it thermally degrades, the reaction releases water. That water is absorbed by the surrounding oil, but more importantly, a portion of it remains in the cellulose, where it catalyzes further degradation.

The relationship between water in oil and water in cellulose is temperature-dependent and described by moisture equilibrium curves. At 20°C, roughly 1% moisture in cellulose corresponds to about 5 ppm water in mineral oil. At 80°C, the same 1% moisture in cellulose corresponds to roughly 40 ppm in oil. This means that an oil sample taken from a warm transformer will show higher moisture content than the same unit at ambient temperature—not because water entered the unit, but because water migrated from the cellulose into the oil. Interpreting moisture-in-oil values requires knowing the oil temperature at the time of sampling.

Moisture reduces the dielectric breakdown voltage of oil and lowers the partial-discharge inception voltage of the oil-paper insulation system. It also participates directly in cellulose degradation: hydrolysis breaks the glucose polymer chains, reducing the degree of polymerization (DP) of the paper. New cellulose insulation has a DP of roughly 1,000–1,200. At a DP of roughly 200, the paper has lost approximately half its mechanical strength and can no longer reliably withstand short-circuit forces.

Field observation: Real-world data from a provincial-scale online DGA monitoring program covering 2,164 transformer units found that roughly one-third of the monitored units had oil chromatography results that did not meet the applicable standards (Hua Ye et al., Digital-Intelligent Inspection Strategy and Application, 2023). This is not a marginal problem; it is a widespread condition across an operating fleet.

3.3 Thermal Decomposition: The Fault Detector

When oil is heated above its normal operating temperature range—typically above about 150°C—thermal cracking begins. The hydrocarbon molecules break into smaller fragments: hydrogen (H₂), methane (CH₄), ethane (C₂H₆), ethylene (C₂H₄), and, above roughly 700°C and in the presence of arcing, acetylene (C₂H₂). This is the chemical basis of dissolved gas analysis: each fault type and temperature range produces a characteristic gas pattern.

3.4 Bubble Evolution and the Bubble Breakdown Mechanism

At a local hotspot—a loose connection, a core bolt with circulating current, or a winding strand with reduced cross-section—the oil in immediate contact with the hot surface can reach temperatures well above the bulk oil temperature. When the oil temperature exceeds its boiling point at the local pressure (which depends on the oil head above the point), a gas bubble forms. The dielectric strength of the gas bubble is far lower than that of the surrounding liquid oil, so partial discharge initiates within the bubble. The discharge further heats the oil, generating more gas, enlarging the bubble, and potentially triggering a full breakdown.

This bubble breakdown mechanism is one reason that acetylene (C₂H₂) in DGA results commands immediate attention: it signals that temperatures exceeded roughly 700°C at some point inside the tank, whether from sustained overheating with bubble formation or from a discrete arcing event.

4. Dissolved Gas Analysis (DGA): The Diagnostic Core

4.1 The Characteristic Gases and Their Fault Associations

4.2 The IEC 60599 Three-Ratio Method

The three-ratio method uses three gas ratios, each mapped to a code (0, 1, or 2), and the three-digit code combination identifies the fault type. This method is standardized in IEC 60599 and widely used as a first-pass diagnostic tool.

4.3 Ratio Code Interpretation

4.4 CO and CO₂: The Cellulose Story

Carbon monoxide (CO) and carbon dioxide (CO₂) are generated when cellulose insulation (paper, pressboard) is heated. The CO₂/CO ratio provides information about the severity of cellulose involvement:

• CO₂/CO > 10: Normal cellulose aging at moderate temperature
• CO₂/CO 3–10: Moderate cellulose overheating
• CO₂/CO < 3: Severe cellulose overheating or involvement in a fault, particularly if accompanied by furanic compounds (furfural)

CO and CO₂ are also present in the atmosphere and can be introduced during oil sampling if the sampling procedure does not adequately exclude air. A single high CO value without corresponding electrical test abnormalities should be verified with a repeat sample before concluding cellulose involvement.

5. Key Diagnostic Acceptance Limits

The polarization index and absorption ratio (K₁ = R_60s/R_15s, typically ≥ 1.3) provide additional information about moisture and contamination in the solid insulation. A low PI with acceptable oil moisture suggests that moisture is concentrated in the cellulose—a condition that oil reconditioning alone cannot fully correct.

6. Procurement Strategies to Reduce Oil-Related Failure Risk

6.1 Oil Type Specification

The choice of insulating oil has direct consequences for aging behavior, fire safety, environmental risk, and maintenance requirements:

For a pad-mounted transformer installed outdoors at ground level in a commercial or residential area, the fire-safety advantage of natural ester can be significant. The moisture-scavenging property of natural ester—it absorbs water from cellulose insulation—can extend paper life in units that operate at elevated temperatures or in humid environments. The premium cost of the fluid is a fraction of the total unit cost and can be justified by reduced fire risk and extended insulation life in the right application.

6.2 Sealed-Tank Design

A sealed-tank transformer with a nitrogen blanket or diaphragm-sealed conservator prevents oxygen and atmospheric moisture from entering the oil. An open-breather design—common in older units and some lower-cost current production—exposes the oil headspace to ambient air through a desiccant breather. When the desiccant saturates (which can happen within months in a humid environment), moist air contacts the oil surface directly.

Sealed-tank construction adds cost but removes the largest continuous source of oxygen and moisture ingress. For procurement, this is a binary specification: sealed-tank or open-breather. There is no middle ground.

6.3 Factory Oil Test Report

Request a factory oil test report for each unit produced, not a type-test certificate for the oil batch alone. The report should include:

• Dielectric breakdown voltage
• Moisture content (Karl Fischer titration)
• Dissolved gas analysis (DGA) on the filled unit after factory testing
• Neutralization number (TAN)
• Interfacial tension
• Inhibitor content (for inhibited oil)
• Oil type and manufacturer identification

This factory report serves as the baseline for all future DGA trend analysis. Without a factory baseline, interpreting the first field DGA sample requires assuming the unit was delivered with clean oil—an assumption that is not always true. Requiring a factory DGA can reduce exposure to the hidden cost of deferred baseline establishment: the diagnostic uncertainty that arises when the first field sample shows elevated gases and no one can determine whether they were present at commissioning or developed in service.

6.4 Oil Sampling Valve

Specify a dedicated oil sampling valve accessible without opening the cable compartment or tank manhole. The valve should be a self-sealing type (syringe-compatible) located below the oil level to avoid drawing from the oil surface where gases may have accumulated. A poorly located or inaccessible sampling valve discourages routine sampling, which is the most cost-effective failure-prevention activity available.

7. When Skipping Oil Analysis Seems Cheaper

1. "The transformer is new; oil testing can wait a few years." A factory-fresh unit can have elevated dissolved gases from the manufacturing process (residual gas from oil processing, gases generated during factory heat-run testing). Without a baseline DGA, the first field sample years later cannot distinguish between a manufacturing artifact and a developing fault. The unit may be flagged for investigation unnecessarily, or a genuine fault may be dismissed as "probably there from the factory."
2. "We only sample when the unit trips or alarms." DGA detects incipient thermal faults and partial discharge months to years before they escalate to a Buchholz alarm or a differential trip. By the time a Buchholz relay operates, the internal fault has already released enough gas to displace oil and trigger the float switch—the fault is no longer incipient. The repair cost and outage duration are substantially higher than if the fault had been caught at the gas-trending stage.
3. "Oil reclamation is cheaper than oil replacement, so we will reclaim when the oil fails." Oil reclamation (filtering, degassing, re-inhibiting) can restore dielectric strength and remove dissolved gases, but it does not reverse cellulose aging. If the oil degraded because moisture was being released from aging cellulose, reclamation treats the symptom while the root cause—weakened paper insulation that may not survive the next through-fault—remains unaddressed.
4. "A DGA test costs several hundred dollars; we have dozens of units." A single unplanned outage of a pad-mounted transformer serving a commercial building, apartment complex, or industrial facility can cost tens of thousands of dollars in lost revenue, emergency replacement procurement, and expedited installation—not counting reputational damage. The annual DGA cost per unit is typically a small fraction of the avoided outage cost, even at low failure probabilities.
5. "Our units are sealed-tank, so moisture cannot get in and oil does not need testing." A sealed tank prevents external moisture ingress but does not prevent internal moisture generation from cellulose aging. It also does not prevent gas generation from thermal or electrical faults. A sealed-tank unit with a developing fault will accumulate fault gases in the oil just as an open-breather unit would. The absence of a breather removes one failure pathway; it does not remove all of them.
6. "We rely on the manufacturer's warranty; if the oil fails, they will cover it." Most transformer warranties exclude consequential damages (lost production, emergency labor, crane rental, disposal costs) and require the owner to demonstrate that the unit was operated within its ratings and maintained per the manufacturer's recommendations. A warranty claim for an oil-related failure is substantially weaker if the owner cannot produce oil test records showing that the oil condition was monitored and maintained.

8. FAQ

How is DGA sampling performed on a pad-mounted transformer?

Oil is drawn from the sampling valve into a clean glass syringe or stainless-steel cylinder using a procedure that minimizes air contact. The sample container must be flushed with the transformer oil before the final sample is taken, and the container must be fully filled to leave no headspace (gases partition between oil and headspace, skewing results). The sample is sent to a laboratory for gas extraction and gas chromatography analysis. Transport time should be minimized; ideally, the sample is analyzed within 48 hours of extraction. For units with online DGA monitors, the monitor provides continuous or periodic readings without physical sampling but requires periodic calibration verification.

What is the difference between IEC 60599 and IEEE C57.104 for DGA interpretation?

Both standards provide gas concentration limits and ratio-based diagnostic methods for mineral oil. IEC 60599 uses the three-ratio method described in this article and provides ratio-code-to-fault mapping. IEEE C57.104 uses gas concentration thresholds organized by condition levels (1–4) and also provides ratio methods (Rogers ratio, Doernenburg ratio) and the Duval triangle for fault classification. The two standards are broadly consistent in their diagnostic conclusions; differences arise mainly in threshold values and in the treatment of borderline cases. For international procurement, specifying either standard is acceptable; specifying both provides cross-validation.

Can a pad-mounted transformer with high moisture in oil be dried in the field?

Field drying is possible but has limitations. Hot-oil circulation with vacuum dehydration can remove moisture from the oil, but removing moisture from the cellulose insulation requires sustained heat and vacuum over days to weeks, depending on the unit size and moisture level. Mobile dry-out rigs exist for this purpose, but the cost and outage duration must be weighed against replacement. For a pad-mounted distribution transformer below roughly 500 kVA, field drying may exceed the replacement cost once labor, equipment, and outage costs are accounted for.

How do I know if the oil needs to be replaced rather than reclaimed?

Decision criteria for replacement vs reclamation include: neutralization number above roughly 0.20 mg KOH/g (indicating advanced oxidation that reclamation may not fully reverse), interfacial tension below roughly 18 mN/m (indicating high polar contaminant content), sludge visible in the oil or on internal surfaces, and a history of multiple reclamation treatments with declining effectiveness. Oil that has been in service beyond roughly 25–30 years without reclamation is a candidate for replacement rather than reclamation, as the additive package (inhibitor, passivator) is likely depleted.

What is the role of furan analysis (furfural) in oil diagnostics?

Furan compounds—particularly 2-furfural (2-FAL)—are chemical byproducts of cellulose degradation. Unlike CO and CO₂, which can also be produced by oil oxidation, furans are exclusively produced by cellulose breakdown. Furfural concentration correlates with the degree of polymerization (DP) of the paper insulation and provides a direct indicator of cellulose condition. A rising furfural trend signals that the solid insulation is degrading at a rate that may require intervention, even if the oil itself appears normal on standard tests. Furfural analysis requires a separate test from standard DGA.

Should a pad-mounted transformer be de-energized for oil sampling?

Routine oil sampling can be performed on an energized transformer provided that the sampling valve is accessible without violating approach distances to live parts. The sampling valve must be below the oil level to avoid drawing gas from the headspace. However, the sampling technician must be qualified to work in proximity to energized equipment and must follow the applicable electrical safety procedures. If the sampling valve is inside the cable compartment and the compartment contains exposed energized terminations, the unit must be de-energized and isolated before the compartment door can be opened.

9. References

• IEC 60599 — Mineral oil-filled electrical equipment in service: Guidance on the interpretation of dissolved and free gases analysis (three-ratio method)
• IEEE Std C57.104 — Guide for the Interpretation of Gases Generated in Mineral Oil-Immersed Transformers
• IEC 60422 — Mineral insulating oils in electrical equipment: Supervision and maintenance guidance
• IEC 60296 — Fluids for electrotechnical applications: Unused mineral insulating oils for transformers and switchgear
• IEC 62770 — Fluids for electrotechnical applications: Unused natural esters for transformers and similar electrical equipment
• IEEE Std C57.147 — Guide for Acceptance and Maintenance of Natural Ester Insulating Liquid in Transformers
• IEEE Std C57.91 — Guide for Loading Mineral-Oil-Immersed Transformers (thermal aging model, hot-spot temperature calculation)
• IEC 60076-7 — Power transformers: Loading guide for mineral-oil-immersed power transformers
• Gao Wei, ed., Medium- and High-Voltage and Insulation Technology, China Machine Press, 2022 — small-bridge theory of liquid insulation, Walton law, bubble breakdown, and the DGA three-ratio method
• Zhou Qiukuan et al., Live Detection Technology and Application for Power Equipment, Jinan University Press, 2017 — DGA characteristic gases and fault correlation, infrared thermography diagnosis of heating mechanisms
• Hua Ye et al., Digital-Intelligent Inspection Strategy and Application, Zhejiang University Press, 2023 — online evaluation of transformer oil chromatography (about one-third of 2,164 monitored units did not meet the applicable criteria), cold-start fault diagnosis
• Huang Junhui and Fu Shenwen, eds., Electrical Engineering Technology, 3rd ed., Posts and Telecom Press, 2016 — transformer insulation test standards (absorption ratio and insulation resistance)