Why a Single Snapshot Is Not Enough
A single dissolved gas analysis (DGA) reading is a snapshot. It tells you what the gas concentrations are today, but not whether the fault is accelerating, stable or cooling down. That is why trend analysis is one of the core elements of DGA application: a single concentration measurement is subject to chance, and the shape of the trajectory over time is far more informative than any one point.
The question for every operator is not just “how much gas is present?” but “how fast is it being generated?” — and that question leads straight to the concept of gas generation rate.
Absolute vs Relative Gas Generation Rate
Gas generation rate is divided into two complementary measures:
- Absolute gas generation rate (ppm/day) — the absolute increase in gas concentration per unit time. A high absolute rate means gas is being produced fast, regardless of where the baseline started.
- Relative gas generation rate (%/month) — the growth of a gas relative to its own baseline. A small absolute rise can still be a large relative rise on a low baseline, which matters for early faults that start near zero.
Neither number replaces the other. The absolute rate tells you how serious the production is; the relative rate tells you how much the situation has changed for that particular gas. Combined, they are the basis for distinguishing a sudden developing fault from slow background aging.
Combining the Two Rates
The distinction is practical, not academic. A slow thermal fault produces a modest, roughly linear absolute rate and a small relative rate — the signature of gradual aging or a low-temperature hot spot. A developing discharge, by contrast, typically shows a sharp jump in absolute ppm/day for acetylene or hydrogen, which shows up immediately in both measures.
Trend analysis also feeds directly into alarm logic. Rather than alarming on an absolute concentration alone, online systems compare current readings against the recent slope and alert when both the level and the rate of rise pass defined thresholds. This is how you catch a fault in its early window instead of after it has already developed.
IEEE C57.104 Condition Ratings and Re-Test Intervals
IEEE C57.104-2019 formalizes this thinking by classifying equipment condition into four ratings according to TDCG and key gas levels, each with a recommended re-test interval:
| Condition rating | Meaning | Recommended re-test interval |
|---|---|---|
| Condition 1 | Normal; gases at low levels | Annual routine |
| Condition 2 | Caution; gas levels rising but below thresholds | Quarterly |
| Condition 3 | Abnormal; approaching or exceeding alarm thresholds | Monthly |
| Condition 4 | Severe; thresholds exceeded, developing fault present | Daily or continuous monitoring |
Notice the escalation path: a transformer can move from an annual sample to daily or continuous monitoring within one deterioration cycle. The re-test interval is itself a diagnostic decision.
Continuous Data vs One-to-Four-Year Lab Sampling
Periodic off-line sampling — even at the recommended intervals — has an inherent blind spot: the time between samples. If a discharge fault begins two weeks after the quarterly sample, the lab result arrives, at best, weeks later. The fault-development window has already passed. This is the core limitation of periodic testing for capturing rapidly developing faults.
Online continuous monitoring removes that blind spot. Continuous trend data allows the gas generation rate to be computed in real time and alarm thresholds to be triggered the moment the trajectory turns. It also provides a confidence reference for the monitor itself: the same data can be used to evaluate the instrument’s step response (for example, T63) and its sensitivity to sudden faults, so a fast gas rise can be separated from a slow sensor response.
For a deeper look at how Condition-based programs use this data, see condition-based transformer maintenance, and for the interpretation frameworks behind it, IEC 60599 vs IEEE C57.104.
PAS DGA for Continuous Trend Data
Trend analysis is only as good as the data stream behind it. The PAS DGA DGA-900 online monitor measures 9 gases plus moisture with laser photoacoustic spectroscopy (L-PAS), delivering the continuous, high-frequency readings needed to compute absolute and relative gas generation rates in real time — including acetylene down to ≤0.1 ppm for early discharge detection.
With near-real-time measurement cycles and no carrier gas or consumables, the DGA-900 keeps a clean, continuous time series on every characteristic gas, ready for Condition 1–4 tracking and rate-of-rise alarms. Explore the DGA fundamentals hub, or contact PAS DGA to discuss a continuous monitoring plan.