This section documents how online DGA monitoring is applied to real transformers: published field records from commissioned monitors, application notes by asset class, and clearly-labelled representative scenarios.
Case Studies & Field Records — Online DGA Fault Detection
Documented records from online DGA monitoring on transmission, converter and industrial transformers. Each record covers the monitored asset and its duty, the gas trend that first raised the alarm, the diagnostic cross-check (for example a Duval Triangle or ratio interpretation), the finding at the planned outage, and the action taken.
We publish only records we can document. Where a page describes a representative scenario rather than a specific installation, it says so on the page.
Acetylene at 0.15 µL/L — Three AC-Substation Cases in Early Online DGA
Three anonymized field records from AC transmission networks. Laser-photoacoustic online DGA detected trace acetylene between 0.15 and 0.8 µL/L — well below any classic alarm level — on a 1000 kV UHV shunt reactor and two 500 kV transformers. Online readings matched laboratory results, and the trend drove action before failure.
- Gas: C₂H₂ (acetylene)
- Detection limit: 0.05 µL/L
- Basis: online readings matched laboratory DGA
Rate of Rise, Not the Number — Three HVDC Converter-Transformer Cases
Three anonymized records from ±660 kV and ±800 kV converter stations. In one, acetylene was detected at 0.07 µL/L roughly 32 hours before a rapid climb to 0.3 µL/L. In the others the decisive parameter was the slope, not the absolute value — including a unit taken out of service on a weekly-increment criterion.
- Gas: C₂H₂ (acetylene)
- Decisive parameter: rate of rise
- Basis: weekly-increment criterion
Railway Traction Transformers — Vibration, Thermal Cycling and H₂ Trending
A representative deployment scenario for trackside and onboard traction transformers: solid-state hydrogen monitoring under continuous vibration and repeated 120–180% peak-load cycling, and how a post-overload hydrogen trend that fails to return to baseline separates a real insulation fault from ordinary load-driven variation.
- Product: DGA-500
- Duty: sustained vibration, frequent overload
- Signal: post-overload H₂ persistence
Application Notes — Online DGA by Asset Class
Technical notes on applying online DGA to specific assets and duty cycles: what to monitor, how to set alarm thresholds, and how to read the trend for each transformer class.
- Format: technical note
- Scope: by asset class
- Basis: standards and field practice
Have a documented monitoring record?
If online DGA monitoring on your transformers produced a gas trend worth documenting — a fault caught early, an unexplained rise resolved, a maintenance decision changed — we would like to hear about it. Records are published only with the asset owner’s permission and with site identifiers removed.
How an online DGA monitor catches a developing fault
- Step 1: Establish the baseline — the monitor records the transformer’s normal gas profile after commissioning.
- Step 2: Detect the rise — a sustained increase in hydrogen or fault gases appears across consecutive measurement cycles.
- Step 3: Alert — the rate of change crosses a threshold and the operator is notified through the alarm or SCADA.
- Step 4: Confirm — a laboratory sample validates the online reading, and ratio or Duval analysis classifies the fault.
- Step 5: Act — maintenance inspects and corrects the issue before an unplanned failure.
FAQ
Q: How do I interpret a sudden gas increase?
A: IEEE C57.104-2019 provides rate-of-change severity: an increase of >30 ppm TDCG per day is a Level 4 (most urgent) condition regardless of absolute concentration. Any sudden appearance of acetylene (C₂H₂) — even at 2 ppm — demands immediate investigation: acetylene is produced only above 700°C, indicating active arcing.
Q: What is stray gassing and should I worry about it?
A: Stray gassing is the generation of fault gases (primarily H₂ and CH₄) at normal operating temperatures without an actual fault — caused by oxidation of certain oil components. It is a known false-positive source that IEC 60599:2022 specifically addresses.