Transformer Oil Gas Analysis: Methods, Fault Interpretation, and Standards
What Is Transformer Oil Gas Analysis?
Transformer oil gas analysis — more commonly known as dissolved gas analysis (DGA) — is a diagnostic technique that identifies developing faults in power transformers by measuring the concentrations of gases dissolved in the insulating oil. When a transformer experiences abnormal thermal or electrical stresses, the mineral oil and cellulose (paper) insulation begin to decompose, releasing characteristic gases that dissolve into the oil. By analyzing which gases are present, their relative concentrations, and how those concentrations change over time, engineers can infer the type and severity of a developing fault — often long before the fault triggers protective relays or causes visible damage.
Related Products
The concept is analogous to blood tests in medicine: certain biomarkers indicate specific conditions, and tracking changes over time reveals whether a condition is stable, worsening, or improving. In transformers, the "biomarkers" are dissolved gases, and their patterns point to specific fault mechanisms.
Why DGA Matters
Power transformers are among the most expensive and critical assets in an electrical grid. Many internal faults — partial discharge, low-energy sparking, localized hot spots — develop gradually. They may exist for months or years before growing severe enough to cause a sudden failure. DGA provides an early warning system that detects these incipient faults, allowing operators to schedule maintenance, derate the transformer, or plan replacement before catastrophic failure occurs. This proactive approach avoids unplanned outages, reduces repair costs, and extends transformer service life.
How Gases Are Generated
The gases dissolved in transformer oil originate from two decomposition pathways:
- Mineral oil decomposition: When oil is subjected to temperatures above its normal operating range or to electrical discharges, hydrocarbon chains break apart, producing hydrogen (H2) and a series of hydrocarbon gases — methane (CH4), ethane (C2H6), ethylene (C2H4), and acetylene (C2H2). The higher the temperature or energy of the fault, the greater the proportion of gases with more double bonds and fewer hydrogen atoms, reflecting deeper molecular fragmentation.
- Cellulose (paper) decomposition: The paper insulation wrapping the windings degrades under thermal stress, producing carbon oxides — carbon monoxide (CO) and carbon dioxide (CO2). The relative proportions of these two gases provide information about the severity of paper degradation.
Key Dissolved Gases and Their Fault Indications
DGA focuses on seven key gases, each associated with specific fault mechanisms and approximate temperature ranges. Understanding what each gas signals is the foundation of all interpretation methods.
| Gas | Formula | Primary Fault Indication | Approximate Temperature Range |
|---|---|---|---|
| Hydrogen | H2 | Partial discharge, corona, low-energy arcing | Broad range |
| Methane | CH4 | Low-temperature thermal fault | Low to moderate |
| Ethane | C2H6 | Low-to-medium thermal fault | Low to moderate |
| Ethylene | C2H4 | High-temperature hot spot, localized overheating | Moderate to high |
| Acetylene | C2H2 | High-energy electrical arc | Very high (>700°C) |
| Carbon monoxide | CO | Cellulose/paper insulation degradation | Moderate to high |
| Carbon dioxide | CO2 | Paper overheating; elevated levels confirm paper involvement | General overheating |
Hydrocarbon Gas Patterns
The hydrocarbon gases follow a clear pattern that reflects the energy of the underlying fault. Low-energy faults (partial discharge, mild overheating) predominantly generate hydrogen and methane. As fault energy increases, ethane and then ethylene appear in growing proportions. Acetylene — the gas requiring the highest energy to form — signals the most severe fault condition: a high-energy electrical arc capable of causing rapid, catastrophic transformer failure.
The Critical Importance of Acetylene
Acetylene (C2H2) deserves special attention because it is formed only at very high temperatures — typically above 700°C — and is the hallmark of an active electrical arc within the transformer. Even trace detection of acetylene warrants urgent investigation, as arcing can rapidly escalate to failure. In newly commissioned transformers, any acetylene detection is particularly alarming, since new oil should contain no arc-generated gases.
Carbon Oxides and Paper Insulation Health
The CO2/CO ratio provides qualitative insight into cellulose insulation condition. An elevated CO2/CO ratio generally indicates ongoing, long-term paper aging, while a depressed ratio may suggest a recent rapid thermal event affecting the cellulose. When CO2/CO ratios suggest paper degradation, supplementary furfural analysis (measuring 2-furfuraldehyde derived from cellulose decomposition) can provide more specific information about the remaining paper insulation life.
DGA Interpretation Methods — Duval Triangle, Rogers Ratios, and Key Gas Approach
Multiple interpretation methods have been developed to translate dissolved gas data into actionable fault diagnoses. Each method uses different analytical logic, and applying several methods in parallel provides the most reliable assessment.
Key Gas Method
The Key Gas method is the oldest and simplest DGA interpretation approach. It identifies the dominant gas — the one present at the highest concentration relative to its typical baseline — and maps it to a corresponding fault type. For example, dominant hydrogen suggests partial discharge, dominant ethylene suggests a high-temperature thermal fault, and dominant acetylene suggests arcing.
While straightforward and useful for quick screening and trend analysis, the Key Gas method has limited fault discrimination. Many fault types share the same dominant gas, and moderate faults may produce a mixture of gases that obscures the dominant signal. It is best used as a first-pass indicator rather than a definitive diagnostic tool.
Duval Triangle Method
The Duval Triangle — developed by Michel Duval and recognized by IEC 60599 — is the most widely used graphical method for DGA interpretation. It plots three gases (CH4, C2H4, C2H2) as normalized percentages on a triangular diagram, where each percentage represents that gas's share of the sum of the three:
- %CH4 = CH4 / (CH4 + C2H4 + C2H2)
- %C2H4 = C2H4 / (CH4 + C2H4 + C2H2)
- %C2H2 = C2H2 / (CH4 + C2H4 + C2H2)
The triangle is divided into distinct fault zones:
| Zone | Fault Type | Description |
|---|---|---|
| PD | Partial discharge | Corona or low-energy discharge in gas gaps |
| D1 | Low-energy discharge | Sparking or tracking in oil |
| D2 | High-energy discharge | High-energy arcing; risk of transformer failure |
| T1 | Thermal fault (low) | Insulation or connection overheating |
| T2 | Thermal fault (medium) | Significant hot spot development |
| T3 | Thermal fault (high) | Severe hot spot; rapid oil/paper decomposition |
| DT | Mixed type | Multiple simultaneous fault mechanisms |
The Duval Triangle provides clear visual identification of fault type and is well suited for trend analysis — plotting successive DGA results on the triangle reveals whether a fault is stable, migrating toward a more severe zone, or resolving. Its primary limitation is ambiguity near zone boundaries, where mixed faults may produce gas ratios that straddle two zones.
A related extension — the Duval Pentagon — incorporates additional gases and provides a more expanded diagnostic space. It is particularly useful when the traditional triangle yields uncertain or conflicting results, helping to clarify fault signatures that do not fit classic patterns.
Rogers Ratios Method
The Rogers Ratios method uses ratios of specific gas pairs to classify fault types:
- C2H2/C2H4 — distinguishes electrical discharge from thermal faults
- CH4/H2 — distinguishes thermal decomposition from partial discharge
- C2H4/C2H6 — identifies the temperature level of thermal faults
By comparing these ratios against established diagnostic codes, the method classifies faults into categories such as normal aging, partial discharge, low-energy discharge, high-energy arcing, and thermal faults at various temperature levels. However, Rogers Ratios become unreliable when gas concentrations are very low — near detection limits — because small measurement uncertainties can produce large ratio swings and misleading classifications.
Recommended Interpretation Workflow
Because different methods emphasize different analytical perspectives, they sometimes produce conflicting results for the same data set. This is not a failure of DGA — it reflects the inherent complexity of fault diagnosis. The recommended practice is to apply multiple methods in parallel and use agreement between methods to strengthen diagnostic confidence. When methods disagree, the discrepancies themselves provide valuable information about the nature of the fault, often pointing to mixed or transitional fault conditions that require closer monitoring.
Standards and Practical Considerations for DGA Sampling
Key Industry Standards
Three standards govern DGA practice:
- IEEE C57.104-2019 — Provides a comprehensive framework for evaluating DGA results, including gas concentration thresholds, severity classifications, and recommended actions for each severity level. It also defines Total Dissolved Combustible Gas (TDCG) as a composite metric that sums the concentrations of all combustible gases (H2, CH4, C2H2, C2H4, C2H6, and CO), offering an overall severity assessment.
- IEC 60599 — Offers interpretation guidance with an emphasis on engineering judgment. IEC 60599 stresses that DGA results must always be evaluated in context — considering transformer design, oil type, load history, and sampling consistency — rather than relying solely on rigid numerical thresholds.
- ASTM D3612 — Defines the test method for extracting and analyzing gases from insulating oil samples. It specifies three extraction procedures: vacuum extraction (Method A), headspace sampling (Method C), and gas stripping. Headspace sampling is widely adopted because it enables automated, higher-throughput analysis suitable for routine monitoring programs.
Sampling Procedures and Frequency
A standard DGA workflow involves four steps:
- Collect an oil sample from the transformer sampling valve, following appropriate electrical safety protocols.
- Extract dissolved gases from the oil using a vacuum gas extraction device or headspace autosampler.
- Analyze gas composition by gas chromatography per ASTM D3612.
- Interpret results using the standards and methods described above.
Sampling frequency depends on the transformer's condition and criticality:
- Routine monitoring: Annually for most transformers in normal operation.
- Abnormal gas detection: Increase frequency to quarterly until the trend stabilizes or the fault is resolved.
- New transformer commissioning: Immediate baseline DGA — this is essential, because without a baseline, subsequent trend analysis has no reference point and is largely meaningless.
- Post-fault or post-maintenance: Immediate sampling to establish a new baseline after oil degassing, filtering, or repair.
Gas Generation Rate and Trend Analysis
Beyond absolute concentrations, the rate of gas generation provides critical diagnostic information. The generation rate is calculated as:
Rate = (C2 − C1) / days between samples
where C1 and C2 are the concentrations of a specific gas at two successive sampling dates. A gas that is doubling its concentration within a short period — regardless of whether the absolute concentration has reached a formal threshold — signals an actively worsening fault and warrants urgent planned inspection, even if the absolute values remain below alarm limits.
Online vs. Offline Monitoring
Modern DGA practice increasingly combines offline laboratory analysis with online continuous monitoring:
- Offline (laboratory) DGA provides the highest analytical accuracy, detecting all seven key gases plus oxygen and nitrogen. It is suitable for annual fleet-wide monitoring and for detailed follow-up investigation of transformers flagged by online monitors or routine screening.
- Online DGA monitors — based on gas chromatography or photoacoustic spectroscopy — provide continuous, real-time gas tracking on individual critical transformers. Online monitoring enables early detection of rapidly developing faults that could escalate between annual sampling intervals. The combination of both approaches ensures comprehensive coverage: annual laboratory DGA for the full fleet, and continuous online monitoring for the most critical assets.
Common Pitfalls in DGA Interpretation
Several factors can confound DGA results if not properly accounted for:
- Oil maintenance history: Degassing and oil replacement alter gas concentrations, and CO2 naturally reaccumulates after degassing. Without knowledge of recent maintenance, post-degassing gas levels may be misinterpreted as fault indicators.
- Sampling consistency: Inconsistent sampling techniques, varying sample handling, or different analytical laboratories can introduce apparent gas variations that do not reflect real changes in transformer condition.
- Transformer design and loading: Different transformer designs produce different baseline gas levels under normal operation. High-load transformers naturally generate more gas than lightly loaded units, and this must be factored into interpretation.
Conclusion
Transformer oil gas analysis remains the single most valuable diagnostic tool for identifying incipient faults in power transformers. By tracking the seven key dissolved gases — hydrogen, methane, ethane, ethylene, acetylene, carbon monoxide, and carbon dioxide — and interpreting their patterns through established methods such as the Duval Triangle and Rogers Ratios, engineers can detect developing faults months or years before they escalate to failure. The key to effective DGA practice lies in consistent sampling, baseline establishment, trend monitoring, and the disciplined application of multiple interpretation methods rather than reliance on a single approach. When combined with online monitoring for critical assets, DGA provides a comprehensive early-warning system that protects transformer investments and maintains grid reliability.
Note: If you don't receive our verification email, do the following:
- Confirm that you entered your email address correctly.
- Check if the email is in your spam or junk folder.
- Or you may contact us at
.
