A power transformer can be the largest single investment made in any electrical system, yet it fails in a predictable and preventable manner. Even before a transformer trips, burns, or fails catastrophically, decomposition of its insulation oil due to thermal and electrical stress has begun, producing specific gases that dissolve in the oil. These gases include hydrogen from partial discharges, ethylene produced due to overheating, and acetylene from arcing, among others. Traditionally, dissolved gas analysis is the study of the gases produced in an insulating fluid similar to performing blood tests. However, with real-time dissolved gas analysis, the process progresses from just an annual lab visit to continuous health monitoring whereby the transformer will show whether or not there is a fault developing, how fast it is growing, and when action is required.
This guide highlights the process of real-time transformer gas monitoring, what gases to look for and their significance, the technologies used in real-time monitoring systems, benefits to asset owners, and how to establish monitoring systems that offer a good return on investment.
Response: The real-time transformer gas monitoring service is able to continuously analyze the concentrations of gases typically present in the transformer oil (such as hydrogen, methane, ethane, ethylene, acetylene, carbon monoxide, and carbon dioxide) using online gas monitoring systems (either in the form of gas chromatography or via photoacoustic/infrared spectroscopy, or through the use of hydrogen specific sensors). Each of these gases is indicative of one of the faults which the transformer is suffering from. For critical transformers, the use of online monitors can make a significant savings as the need to do maintenance is on the basis of necessity rather than simply on the calendar, thus avoiding unscheduled outages altogether.
What Is Real-Time Transformer Gas Monitoring?
Real-Time Transformer Gas Monitoring is a process of continuously measuring the concentration of dissolved gases in the insulating oil of a working transformer without taking it offline. Connection is made through the transformer’s oil system — generally via the drain valve or a dedicated sampling valve — to continuously extract oil or gas at predetermined intervals within a time range of minutes to hours. The oil or gas is analyzed, and readings are displayed and/or transmitted to a SCADA or asset management system.
The online method of DGA is the most applied diagnostic method in the power sector for oil-filled transformers. Utility companies are accustomed to the offline method, which involves taking oil samples at definite intervals (usually every year) and shipping them to a laboratory to get back a gas chromatogram a week after. The online method offers all the features of software based DGA while process data continuously, eliminating a lot of time-consuming steps and providing the rate of change of the dissolved gases being released as one of the earliest warning signs a transformer can provide.
The Gases: What Each One Tells You
Transformer oils and paper release a limited range of key diagnostic gases while breaking down under electrical stress. The gas “signature” is determined by identifying both the gases that are present and their concentrations relative to one another.
| Gas | Typical Cause | Fault Signature |
|---|---|---|
| Hydrogen (H2) | Partial discharge, corona, low-energy faults | Leading early indicator; often the first gas to rise |
| Methane (CH4) | Low-temperature thermal faults | Rises with overheating below ~300°C |
| Ethane (C2H6) | Moderate thermal stress | Pairs with methane in thermal signatures |
| Ethylene (C2H4) | High-temperature thermal faults (300-700°C) | Dominates in severe hot spots |
| Acetylene (C2H2) | Arcing, high-energy discharge | The most dangerous signature; indicates arcing |
| Carbon monoxide (CO) | Cellulose (paper) degradation | Rising CO signals overheated insulation |
| Carbon dioxide (CO2) | Normal aging plus cellulose stress | CO2/CO ratio helps age the fault |
Learning only two rules is more important than the memorization of a chart of sample fault gases. First, understanding that the gas patterns are more important than any single gas is important. For example, having a transformer with low hydrogen and methane levels is not as serious a fault as a transformer experiencing an increase in the presence of acetylene. Second, the trend is more important than the specific number of any gas concentration. A transformer may have gases in acceptable concentrations per IEEE C57.104, but if the concentrations are seeing a doubling every month, it represents a much more urgent situation than one having stable higher gas concentrations over the past two or three years.
How Online DGA Works: Sensing Technologies
The price-performance trade-offs between different proven techniques used in online DGA monitors include:
- Gas chromatography (GC): a laboratory gold standard known for its online adaptability. It makes use of a carrier gas that moves the extracted gas through chromatographic columns where gases are separated by molecular characteristics and detected in sequence. The multi-gas online GCs give you the complete package of key gases (H2, CH4, C2H2, C2H4, C2H6, CO, and CO2) making it the most accurate variety of online monitors.
- Photoacoustic and infrared spectroscopy (PAS/IR): gas molecules absorb infrared frequencies with specific absorption providing gas measurement as well as identification. The PAS technology offers high reliability over longer durations with less maintenance that is beneficial for longer periods of time. Presently, the multi-gas systems offer a full range of key gases.
- Hydrogen-specific sensors: considered one of the cheaper varieties with the potential for hydroen to be released first among several different methods of fault detection (partial discharge, arcing, overheating all produce H2). There is great value with hydrogen monitors producing significant early warning data and with lower costs than multi-gas equipment.
Sampling is as vital as sensing: the device must be capable of extracting a representative gas sample from the oil (headspace or vacuum sampling techniques are most used) and the monitor must be calibrated, installed, and performed in order to capture reliable data — an uncalibrated monitor is worse than no monitor at all as it gives manufacturers false confidence. Our guide to transformer oil testing basics covers the sampling fundamentals behind every monitoring program.
Interpreting the Data: Standards and Methods
The process of determining the cause of raw gas concentrations involves determining necessary parameters using various methods instead of using one alone as established by IEEE C57.104 and the Duval Triangle.
- The IEEE C57 standard is the foundation on which such methods are based namely permissible gas limits by nature of oil-immersed transformers for four conditions (Condition 1 = normal; Condition 4 = urgent action) that have with their assembly of required guidelines for sampling taken for each condition, and what is expected of the operators. The 2019 version of the standard provides that Condition 1 must have a sampling rate set to one year while four months is prescribed for Condition 2, and Condition 3 will have a sampling frequency of one month while weekly sampling must occur for Condition 4.
- The Duval Triangle method makes use of a visual representation method (IEC 60599) which predicts fault types based on the relative percentage of three types of gases – methane, ethylene, and acetylene – which fall under the categories of partial discharge, low/high energy discharge, and thermal error,
- The ratio method utilizes the Doernenburg, Rogers, and IEC 60599 methodologies for fault identification based on the ratio of gas pairs according to the tables.
- The key gas method accounts for possible streams based on the dominant gas; for example, which of the four gases H2, C2H4, and CO are prevalent with the conclusion being reached on the particular defect type.
- The trend analysis method of real-time monitoring measures the rate of generation of gas as a critical determinant for effective decision making.
No method can work in isolation, therefore, experienced inspector use their expertise to combine the approaches and provide a context for better decision-making.For a practical walkthrough of reading DGA reports, our DGA interpretation guide steps through a real gas report end to end.
From Monitoring to Condition-Based Maintenance
Condition-based maintenance does not rely on the calendar for maintenance schedules, rather it focuses on what is taking place inside the individual piece of equipment. In simple terms, one is not told to “inspect every transformer every year and test oil every year no matter what condition it is in”. Instead, one is told to “watch everything continuously, do something when the data indicates it is time to do something, and adjust the frequency of sampling as the condition dictates. More specifically:
- Transformers are not touched and budgets and work hours are diverted into pieces of equipment that require attention
- Equipment generates gas monitoring and special investigations are conducted in which fast sampling is conducted along with thermal imaging, oil quality testing (including moisture via the Karl Fischer method, breakdown voltage, furan analysis of the paper, etc.) and acoustic (partial-discharge checks) diagnostics to pinpoint the trouble areas on transformers.
- Equipment alerts systems of certain actions that need to be taken – load is re-dispatched, spares are on stand-by, outages are planned, and equipment is de-energized (if products start indicating a significant increase in acetylene generation).
Essentially, the maintenance program has morphed into a decision tree based upon the system monitoring the asset which is what asset managers mean by “predictive” maintenance as opposed to “reactive” maintenance. Because it is the gas generation rates that are the most informative parameter, it is the continuous online data that makes the decision tree happen as one annual sample cannot indicate whether gas generation is being accelerated.For guidance on the broader inspection program that wraps around monitoring, see our transformer maintenance checklist.
The Business Case: Benefits and Payback
The advantages of real-time gas monitoring systems are specific and economically crucial where the critical assets are being used.
- Fault detection at an early stage – defects such as partial discharge, hot spots, or loose connections can be detected months prior to failure so that the repairs can be made at almost no cost.
- Avoidance of forced outages – unanticipated failure of transformers can lead to production losses. The cost of forced outage is much higher than the mere price of the monitoring system. According to the research conducted by EPRI (Electric Power Research Institute) it can be said that even a single avoided forced outage could cover the cost of the monitoring program.
- Utilization of resources in a better fashion- maintenance and sampling should be focused only on assets that need help so as to use resources efficiently.
- Longer life of assets- if the overheating of the transformer is detected before it causes damage to the insulation, the transformer will last for several years.
- Safety- detection of the gas signature indicates the advent of an arc fault, allowing operators to take preventive actions.
- Making decisions based on facts- changes in the load or purchase and replacement of transformers should be done based on the historical condition data rather than just guessing.
It is considered safe and best practice to fit all important transformers with real-time monitoring systems and not use this technology on less critical units and rely only on annual tests. Our transformer monitoring solutions page shows how monitoring hardware and services are typically packaged for such fleets.
Solutions and Industry Programs
A complete real-time monitoring system consists of four layers. In the industry, each layer can be implemented in various scales.
- Real-time DGA analyzers — single-gas hydrogen analyzers for budget fleets and multi-gas GC/PAS techniques for critical facilities. Prevalent manufacturers are Qualitrol and its competitors like Vaisala and Serveron/TM8, Morgan Schaffer and Camlin.
- Integrated transformer monitoring systems — accumulation of DGA into one platform along with oil temperature and transformer winding temperature monitoring, etc. (e.g., Cambridge and Siemens).
- The interface with SCADA/asset management systems, where the results are integrated into the plant’s management control systems with respect to alerts and dashboards, where CBM decisions are taken.
- Service that includes DGA interpretation and laboratory tests as a baseline for future implementations.
In solar and wind power stations, the first package includes main transformers and generator step-up transformers, whereas, in industries, the package might use feeding and manufacturing transformers. The bottom line is “monitor what you want absolutely reliable data for and monitor what you can afford to lose.” For help sizing a program for your fleet, contact our team — and start with our about page to see how we support transformer selection and monitoring for owners across industries.
Frequently Asked Questions
What gases does DGA detect and what do they mean?
A full DGAs program measures seven essential gases, namely hydrogen — which indicates a partial discharge, methane and ethane for mild thermal collapse, ethylene for overheating at high temperatures, acetylene for arcing and carbon monoxide / carbon dioxide for degradation of paper dielectric. Depending on the presence and trends of these gases, one can determine the reason and degree of fault, rather than using just the individual concentration number of one of those gases.
How often should transformer oil be sampled?
According to IEEE C57.104-2019, the standard states that transformer’s oils should be sampled annually in case of Condition 1 (normal condition), quarterly for Condition2, monthly for Condition 3, and weekly for Condition 4. Newly commissioned transformers can be called for monthly sampling for the first quarter.
Is online DGA monitoring worth the cost?
For critical transformers, it is worth the investment. Because the lost production, in addition to the cost for replacements, in case of forced outages would easily surpass the cost of investment on the monitor. For auxiliary transformers, lab sampling at least once a year or quarterly will be the best practice.
What is the difference between condition-based and preventive maintenance?
Preventive maintenance works on a pre-established schedule, irrespective of any measured state of the equipment. Thus, MBM uses actual measurements (the current gas trends, loads, and/or test results) in order to take action so that the healthy equipment will be left untouched and developing.Thus, CBM saves money for any work performed. The introduction of real-time gas monitoring is the informative layer that allows you to use CBM.
References
- IEEE Technology Navigator — Dissolved gas analysis overview
- Power Prognosis — Integrated DGA interpretation based on IEEE C57.104
- REIG — Solar plant transformer DGA field guide (sampling cadence, Duval, payback)
- Springer — Online dissolved gas analysis: possibilities, experiences, limitations
- Megger — How to interpret DGA results for transformer health
Conclusion
With real-time transformer gas detection, the precious commodity of the electrical installation changes from being a black box with no information into a patient with an electrocardiogram wiring. The dissolved gases that are in the transformer oil, which include hydrogen, methane, ethane, ethylene, acetylene, and carbon oxides, are early-warning signals that have been proven over the years in DGA usage and standards and use of tools for interpretation in converting the signals to maintenance decisions. Transitioning from the use of periodic laboratory sampling to real-time online monitoring brings one key advantage, which is rate of change, and thus, in this case, one key advantage is what lies in between either achieving detection of an issue at routine cost or one will be realizing it existed through forced outage. It is essential to take note of the business case, which is completed and closed with the first avoided failure for the owner of a critical transformer.