Two Measurement Philosophies
Dissolved gas analysis (DGA) is the fundamental condition-diagnosis method for oil-filled power equipment — the only energized testing technique that covers both early-stage and developed faults, in offline and online modes alike. But “DGA” is not one method. Laboratory gas chromatography (GC) and online continuous monitoring answer different questions: one is a high-accuracy snapshot, the other is a continuous trend. Understanding the trade-off is the first step in deciding when to upgrade an asset to online monitoring.
Laboratory GC: The Benchmark and Its Limits
Offline GC remains the laboratory benchmark. Oil is sampled on site, dissolved gases are extracted by headspace or vacuum degassing, and a high-purity carrier gas (helium, purity ≥99.9995%) carries the mixture through a chromatography column for separation and quantification by thermal conductivity or flame ionization detection. It reaches lower detection limits at the ppm level and even into the sub-ppm range, and laboratory analysis total uncertainty is on the order of ±15% per IEC 60567. The limits are equally clear: consumables such as carrier gas and columns must be replaced periodically, and the full cycle — sampling, transport, laboratory analysis, reporting — takes several hours, which makes it difficult to capture a rapidly developing fault.
| Aspect | Laboratory GC (offline) | Online monitor (L-PAS) |
|---|---|---|
| Analysis principle | Chromatographic separation + detector | Laser photoacoustic detection |
| C2H2 lower detection limit | ppm level (sub-ppm achievable) | ≤0.1 ppm (commercial) |
| Total uncertainty | ≈ ±15% | Same order vs lab when cross-checked |
| Measurement cycle | Several hours (sample to report) | Near-real-time to ~1 h |
| Consumables | Carrier gas + chromatography columns | None |
| Data continuity | Snapshot per sample | Continuous; trend + rate of rise |
| Best role | Periodic preventive testing | Continuous early warning |
What Continuous Monitoring Adds
Continuous monitoring changes the diagnostic picture in three ways. First, it turns single-point concentrations into a trend. A single measurement carries sampling and analysis uncertainty; a trend separates a genuine developing fault from noise. Second, it enables the gas generation rate — both absolute (ppm/day) and relative (%/month) — which IEEE C57.104 uses to distinguish a sudden fault from slow aging. Third, it catches the early rise of acetylene (C2H2), the discharge indicator that can appear at ppm level and escalate quickly. An online monitor with an hourly or faster cycle and a C2H2 detection limit around 0.1 ppm provides the lead time that annual or quarterly lab sampling simply cannot: it shifts the question from “has the fault happened?” to “is a fault starting now?”
When to Upgrade
Upgrading an asset from lab-only to online monitoring is justified when the downside of a missed fault is large. Three situations dominate. First, an aging fleet: approximately 70% of transformers in the United States have been in service for more than 25 years, and aging combined with load growth pushes fault risk upward. Second, critical assets: for a large main transformer, the combined cost of scrapping or major repair, outage losses, and social impact can reach the order of tens of millions of dollars — far above the cost of a monitor. Third, unmanned or remote sites: where a fault would go unnoticed until a routine visit, continuous monitoring with remote alarms is often the only practical protection. For a structured view of the economics, see our 10-year TCO comparison of online vs periodic monitoring.
Complementary, Not Replacement
Online monitoring is not a substitute for laboratory DGA; it is a complement. In practice the two work as a tiered system: the online monitor provides continuous trending and early alarm on critical assets, while periodic laboratory GC continues to cover the wider fleet, provides an independent reference for online readings, and is pulled in as confirmation when an alarm triggers. A cross-check between online and offline data — both at the same order of uncertainty — validates both. This hybrid approach is the industry pattern for condition-based maintenance, and our guide to condition-based transformer maintenance describes how to operationalize it. If you are new to the fundamentals, start with our introduction to dissolved gas analysis.
PAS DGA for Online DGA Monitoring
For the assets that justify continuous monitoring, PAS DGA offers a consumable-free L-PAS line. The DGA-900 measures 9 gases plus moisture with a C2H2 detection limit of ≤0.1 ppm and near-real-time to hourly cycles — built for trending, rate-of-rise alarms, and early discharge detection. For a hydrogen-first screen across a larger fleet, the DGA-200 hydrogen monitor covers the most common incipient-fault gas at lower cost. Review your asset list against the upgrade triggers above, and contact PAS DGA to plan the hybrid monitoring architecture for your fleet.