What is Dissolved Gas Analysis (DGA)?
Dissolved Gas Analysis is the single most important diagnostic tool for assessing the health of oil-filled power transformers. When a transformer experiences thermal or electrical stress, the mineral oil and cellulose insulation decompose, generating specific gases that dissolve in the oil. By measuring the types and concentrations of these gases, DGA can:
- Detect incipient faults weeks or months before they become critical
- Identify fault type — partial discharge, overheating, arcing, or insulation degradation
- Estimate fault severity — from mild overheating to active arcing
- Track fault progression — rate-of-change trending provides early warning
DGA is mandated by international standards (IEC 60599, IEEE C57.104) and is used by utilities worldwide as the foundation of transformer condition-based maintenance programs.
The 7 Key Fault Gases
Each gas tells a specific story about what is happening inside the transformer. Understanding their generation mechanisms is the first step to accurate diagnosis.
Hydrogen (H₂)
| Property | Value |
|---|---|
| Chemical formula | H₂ |
| Generation temperature | All temperatures (earliest gas produced) |
| Primary fault association | Partial discharge (PD), corona |
| Secondary fault association | Low-temperature thermal faults, stray gassing |
| IEEE C57.104 90% norm | 80 ppm |
Key insight: Hydrogen is the universal early warning gas. It is produced at relatively low temperatures (<150°C) in partial discharge events and continues to be generated across all fault temperature ranges. A rising H₂ trend is often the first indicator of any developing problem — but by itself, it cannot distinguish between fault types. This is why hydrogen-only monitors are excellent for early screening but insufficient for complete diagnosis.
Generation mechanism: Ionic bombardment from corona discharge breaks C-H bonds in hydrocarbon oil molecules, releasing atomic hydrogen that rapidly recombines into H₂.
Methane (CH₄)
| Property | Value |
|---|---|
| Chemical formula | CH₄ |
| Generation temperature | 150–300°C |
| Primary fault association | Low-temperature thermal faults (T1) |
| IEEE C57.104 90% norm | 90 ppm |
Key insight: Methane is the signature gas for low-temperature overheating. In the Duval Triangle method, methane forms one axis of the diagnostic triangle. High CH₄ with low C₂H₄ suggests a hot spot in the 150–300°C range — often from overloaded connections, circulating currents, or early-stage core faults.
Ethane (C₂H₆)
| Property | Value |
|---|---|
| Chemical formula | C₂H₆ |
| Generation temperature | 300–500°C |
| Primary fault association | Mid-range thermal faults |
| IEEE C57.104 90% norm | 90 ppm |
Key insight: Ethane is produced at intermediate temperatures. The C₂H₆/C₂H₄ ratio is a key indicator of thermal fault severity — a higher proportion of ethane relative to ethylene indicates a lower-temperature hot spot. As temperature rises above 500°C, ethylene production accelerates and C₂H₆/C₂H₄ decreases.
Ethylene (C₂H₄)
| Property | Value |
|---|---|
| Chemical formula | C₂H₄ |
| Generation temperature | >500°C |
| Primary fault association | High-temperature overheating (T2, T3) |
| IEEE C57.104 90% norm | 50 ppm |
Key insight: Ethylene is the primary indicator of severe thermal faults. Above 500°C, ethylene production dominates. A rapidly increasing C₂H₄ trend with low acetylene strongly suggests a hot spot exceeding 500°C — often from bad contacts, circulating currents in core, or tank wall heating from high-current bushings.
Acetylene (C₂H₂)
| Property | Value |
|---|---|
| Chemical formula | C₂H₂ |
| Generation temperature | >700°C (typically >1000°C) |
| Primary fault association | High-energy arcing, severe electrical discharges |
| IEEE C57.104 90% norm | 1 ppm (most critical threshold) |
Key insight: Acetylene is the most diagnostically significant gas in DGA. It is produced almost exclusively by electrical arcing at temperatures above 700°C — conditions that cause rapid and catastrophic insulation failure. Even trace amounts of C₂H₂ (>1 ppm) demand immediate investigation. A transformer with detectable acetylene should be considered at risk of active internal arcing.
Critical note: Acetylene can be as low as 2 ppm during a real arcing fault. A hydrogen-only monitor cannot detect this — this is the strongest argument for multi-gas DGA monitoring on critical transformers.
Carbon Monoxide (CO)
| Property | Value |
|---|---|
| Chemical formula | CO |
| Generation temperature | All temperatures (cellulose pyrolysis) |
| Primary fault association | Paper/cellulose insulation degradation |
| IEEE C57.104 90% norm | 900 ppm |
Key insight: Carbon monoxide is the primary indicator of solid insulation (paper) degradation. Unlike the hydrocarbon gases (H₂, CH₄, C₂H₆, C₂H₄, C₂H₂) which come from oil decomposition, CO comes from the cellulose paper that wraps the transformer windings. Rising CO indicates the paper insulation — which cannot be replaced without a complete rewind — is deteriorating.
Carbon Dioxide (CO₂)
| Property | Value |
|---|---|
| Chemical formula | CO₂ |
| Generation temperature | All temperatures (cellulose oxidation) |
| Primary fault association | Paper/cellulose degradation (broader range than CO) |
| IEEE C57.104 90% norm | 9,000 ppm |
Key insight: The CO₂/CO ratio is diagnostically important. A normal CO₂/CO ratio for healthy transformers is typically >7. When the ratio drops below 3, it indicates pyrolysis of cellulose (thermal paper decomposition) rather than normal oxidation — a more serious condition. However, CO₂/CO interpretation must account for CO₂ from atmospheric oxidation in free-breathing transformers.
Gas Generation Summary Table
| Gas | Symbol | Temperature | Primary Fault | 90% Norm (IEEE) | Key Ratio |
|---|---|---|---|---|---|
| Hydrogen | H₂ | All temps | Partial discharge, corona | 80 ppm | CH₄/H₂ |
| Methane | CH₄ | 150–300°C | Low-temp thermal (T1) | 90 ppm | — |
| Ethane | C₂H₆ | 300–500°C | Mid-range thermal | 90 ppm | C₂H₆/C₂H₄ |
| Ethylene | C₂H₄ | >500°C | High-temp thermal (T2, T3) | 50 ppm | C₂H₄/C₂H₆ |
| Acetylene | C₂H₂ | >700°C | High-energy arcing (D1, D2) | 1 ppm | C₂H₂/C₂H₄ |
| Carbon Monoxide | CO | All temps | Paper degradation | 900 ppm | CO₂/CO |
| Carbon Dioxide | CO₂ | All temps | Paper oxidation/degradation | 9,000 ppm | CO₂/CO |
Ostwald Coefficients and Gas Solubility
Gases dissolve in transformer oil according to Ostwald solubility coefficients. These coefficients are temperature-dependent and gas-specific — they determine the equilibrium between gas concentration in the oil and in the headspace (or gas extraction system).
| Gas | Ostwald Coefficient at 20°C | Ostwald Coefficient at 50°C | Solubility |
|---|---|---|---|
| H₂ | 0.050 | 0.055 | Very low (stays in oil) |
| CO | 0.121 | 0.115 | Low |
| CH₄ | 0.427 | 0.392 | Moderate |
| C₂H₆ | 2.17 | 1.66 | High |
| C₂H₄ | 1.45 | 1.14 | High |
| C₂H₂ | 1.15 | 0.94 | Moderate-High |
| CO₂ | 1.08 | 0.97 | Moderate |
Practical implication: Acetylene is highly soluble — if you detect 1 ppm C₂H₂ in the oil, the actual fault may have generated significantly more. Hydrogen, conversely, partitions strongly into the oil phase. Understanding solubility is essential for comparing online DGA readings with laboratory GC results.
Standards for DGA Interpretation
DGA data is interpreted using internationally recognized standards:
| Standard | Title | Key Content |
|---|---|---|
| IEC 60599:2022 | Guidance on interpretation of dissolved and free gases analysis | Gas ratio methods (R1=CH₄/H₂, R2=C₂H₂/C₂H₄, R3=C₂H₄/C₂H₆); fault classification PD/D1/D2/T1/T2/T3 |
| IEEE C57.104-2019 | Guide for Interpretation of Gases Generated in Mineral Oil-Immersed Transformers | 90th/95th percentile statistical norms; DGA Status 1/2/3 classification; Duval Triangles/Pentagons |
| IEC 60567:2024 | Sampling and analysis of free and dissolved gases | Laboratory procedures; 3 gas extraction methods; new Annex F for ester and silicone fluids |
| IEEE C57.143-2024 | Guide for Application of Monitoring Equipment | Monitoring parameters; communication protocols; cybersecurity; cost-benefit analysis |
| CIGRE TB 771 (2019) | Advances in DGA Interpretation | 330,000+ case database; graded severity; faults by location; expanded sub-types |
From DGA Data to Action
DGA results should trigger a structured response:
| Condition | Criteria | Recommended Action |
|---|---|---|
| Normal | All gases below IEEE 90% norms; stable trends | Continue routine monitoring (online: hourly; lab: annual) |
| Caution (Status 1→2) | Any gas exceeds 90% norm OR rising trend detected | Increase online sampling frequency; schedule confirmatory lab DGA; review load history |
| Abnormal (Status 2) | Gases between 90% and 95% norms with rising trend | Perform comprehensive DGA interpretation (Duval, Rogers, IEC ratios); plan inspection |
| Critical (Status 3) | Any gas above 95% norm OR rapid rate-of-change | Immediate investigation; consider outage; continuous online monitoring if not already installed |
Next Steps in Your DGA Journey
- DGA Diagnostic Methods — Master the Duval Triangle, Rogers Ratio, and IEC ratio methods for fault classification
- Transformer Fault Types — Detailed case studies of PD, arcing, thermal faults, and cellulose degradation
- Online vs. Offline DGA — When to use continuous online monitoring vs. periodic laboratory testing
- DGA Technology Comparison — PAS vs. GC vs. NDIR vs. FTIR — choosing the right technology