Critical gases in transformer monitoring: what they tell us about asset health

Power transformers are among the most critical and expensive components in electrical networks. Failures can lead to widespread outages, environmental incidents, and significant replacement costs. For grid operators, utilities, and industrial facilities transformer health monitoring has evolved from a maintenance luxury to an operational necessity. With transformers expected to remain in service for decades - often well beyond their original life expectancy - the ability to assess their condition accurately has become crucial for maintaining grid reliability.

Traditionally, maintenance relied on offline testing, with oil samples collected periodically, as per industry guidelines, and sent to laboratories for analysis. While effective in providing snapshots of transformer health, this approach left operational gaps in operation between measurements during which incipient faults initiate and developunchecked. The industry has progressively shifted towards continuous monitoring solutions that offer real-time insights into transformer condition.

Among the various monitoring strategies, dissolved gas analysis (DGA) has emerged as the definitive "blood test" for transformers. While comprehensive laboratory analysis examines seven or more gases, three key indicators can be efficiently monitored in real time - hydrogen, acetylene, and moisture – to provide the most critical information about changes in a transformer’s condition.

Key gases and their significance

Hydrogen is the universal fault indicator in transformer oil, appearing in almost all fault conditions from low-energy corona and partial discharge to severe arcing. Its presence at elevated levels (typically above 100 ppm) signals something abnormal is occurring within the transformer. Hydrogen generation begins at relatively low temperatures, around 150°C, making it a reliable marker of most developing faults at their earliest stages. However, while hydrogen reliably indicates that a fault exists, it cannot by itself determine the fault's severity or nature.

Acetylene serves as a critical marker of high-energy faults and provides the most definitive indicator of potentially dangerous conditions. Unlike hydrogen, acetylene forms only at temperatures exceeding 700 °C, which typically occur only during arcing or severe hotspots. The presence of acetylene above the threshold of 2 ppm (as noted in CIGRE Technical Bulletin 783) signals a condition requiring urgent attention. Modern monitoring systems with laboratory-grade sensitivity can detect acetylene at levels as low as 0.5 ppm, offering valuable early warning of developing high-energy faults.

Moisture, while not a fault gas, acts as the silent deteriorator of transformer solid and liquid insulation. Excessive moisture accelerates the aging of paper insulation and reduces its dielectric strength, leading to premature failure. Each doubling of moisture content in paper insulation reduces its expected life by approximately half. In addition, moisture reduces the dielectric strength of oil, allowing fault conditions to occur at lower temperature and under lower load conditions. Furthermore, moisture affects the formation and distribution of fault gases, which can lead to misleading diagnostic results if not properly considered. Monitoring moisture levels alongside hydrogen and acetylene provides a more comprehensive view of transformer health and helps maintenance teams to interpret gas data more accurately, particularly when differentiating between genuinely concerning conditions and normal variations.

Interpreting gas data effectively

Individual gas concentration thresholds form the basis for transformer condition assessment, with industry standards establishing key levels for hydrogen and acetylene. For hydrogen, levels below 100 ppm generally indicate normal conditions, while levels exceeding 700 ppm suggest active fault development. Acetylene thresholds are considerably lower, with levels above 2 ppm indicating present or past arcing conditions.

The relationship between hydrogen and acetylene provides an excellent gauge for determining the severity of a fault type which neither gas alone can offer. When hydrogen levels rise without detectable acetylene, this typically indicates a low-energy fault, such as partial discharge or localised overheating below 700°C. In contrast, when acetylene appears alongside elevated hydrogen, particularly when both are increasing rapidly, this strongly suggests a developing high-energy fault requiring urgent attention. With both gases being measured independently, one has a formidable alarm detector type online DGA instrument.

The rate of change in gas concentrations often provides the earliest and most reliable indication of developing problems. A slow, consistent rise in hydrogen may signal a stable, low-energy fault that can be monitored over time, while a rapid increase suggests an accelerating condition that requires immediate attention. Modern monitoring systems with continuous sampling capabilities excel at detecting these patterns of change.

Contextualising gas data with operational conditions and moisture levels is essential for accurate interpretation. Loading patterns, ambient temperature changes, and moisture fluctuations all influence gas generation and distribution in transformer oil. The most effective monitoring strategies incorporate moisture measurements alongside gas data, utilising algorithms that take these contextual factors into account.

Evolution of monitoring technology

Transformer monitoring began with manual oil sampling, a process that remains relevant today despite technological advances. While it provides high accuracy for multiple parameters, this method creates significant blind spots between sampling intervals, typically 6-12 months for standard transformers.

Online monitoring approaches first emerged in the 1970s, beginning with moisture detection and later expanding to combustible gas monitoring. Early systems had important limitations - hydrogen-only monitors could detect faults but not their nature or severity, composite gas sensors encountered challenges in distinguishing between fault types, and many systems suffered from cross-sensitivity issues. Perhaps most critically, early fault detectors lacked the sensitivity to detect acetylene at the low levels (below 2 ppm) necessary for the early identification of high-energy faults.

Recent advances in detection technology, particularly laser spectroscopy, have transformed transformer monitoring capabilities. Tuneable diode laser spectroscopy (TDLS) technology enables highly selective gas detection by tuning the laser precisely to the absorption spectrum of a specific gas, effectively eliminating interference from other gases present in the oil. This selectivity enables remarkable sensitivity, with modern systems capable of detecting acetylene at levels as low as 0.5 ppm - well below the critical threshold.

The industry has steadily advanced from basic fault detection to more nuanced fault classification. By continuously tracking both hydrogen and acetylene, modern systems can not only detect faults but also provide critical information about fault type and severity. Simultaneous moisture measurement further enhances this capability by accounting for the moisture influence on gas behaviour and offering additional context regarding insulation health.

Implementing effective monitoring strategies

Selecting the appropriate monitoring approach requires careful consideration of transformer criticality, replacement costs, and operational context. For critical transformers where failure could lead to considerable service disruption, continuous monitoring of hydrogen, acetylene, and moisture offers the optimal combination of early fault detection and fault type classification.

Integrating gas monitoring with maintenance programs enhances both. Successful implementations connect monitoring data to maintenance actions, using specific thresholds for hydrogen and acetylene concentrations and rate-of-change values to trigger protocols. Utilities that incorporate hydrogen-acetylene-moisture monitoring into their maintenance programs can extend routine maintenance intervals while maintaining or improving transformer reliability.

Cost-benefit considerations for monitoring deployment should extend beyond simple equipment prices to include the total lifecycle impact. Targeted monitoring systems that precisely track hydrogen, acetylene, and moisture often provide the optimal balance of protection and affordability. The cost calculation should consider not only the monitoring equipment but also the complexity of installation, ongoing maintenance needs, and anticipated service life.

Practical case studies demonstrate the real-world value of implementing strategic monitoring. One large industrial operation detected acetylene at 1.5 ppm in a critical transformer through routine laboratory testing. Rather than immediately removing the transformer from service, they installed a high-precision acetylene-hydrogen monitor to track the condition between laboratory tests. Another example comes from a utility that replaced several dozen composite gas monitors with systems capable of precise acetylene detection alongside hydrogen monitoring, reporting significant improvements in maintenance efficiency by clearly distinguishing between urgent high-energy faults requiring immediate response and lower priority developing conditions that could be addressed during scheduled maintenance windows.

Conclusion

The future of transformer monitoring lies in focused, high-precision tracking of the most critical parameters - hydrogen, acetylene, and moisture - combined with sophisticated analytics that interpret these measurements in context. This approach provides actionable insights while remaining sufficiently cost-effective for deployment across entire transformer fleets.

For maintenance professionals managing transformer fleets, key takeaways include: the ability to detect acetylene with laboratory-grade precision is crucial for identifying high-energy faults before they escalate; monitoring both hydrogen and acetylene provides significantly greater diagnostic value than either parameter alone; and incorporating moisture measurements alongside gas monitoring offers essential context for accurate interpretation.

Organisations looking to enhance transformer health monitoring should implement monitoring systems that accurately track hydrogen, acetylene, and moisture, especially for critical assets where existing monitoring provides insufficient clarity about fault severity. Establishing clear response protocols ensures that monitoring data translates into effective maintenance actions, improving transformer reliability while optimising maintenance resources.