Gas chromatography is a laboratory technique that separates and quantifies the dissolved gases extracted from an oil sample.
Two Technologies, One Purpose
GC has been the reference standard for 50+ years. PAS is the 4th-generation technology making continuous multi-gas monitoring practical. Understanding their tradeoffs is essential for specifying a DGA strategy.
Key figures at a glance
- Multi-gas monitors measure on a cycle of 1 to 24 hours (default 4 hours) — up to 6 measurements per day with no operator involvement.
- Palladium-alloy hydrogen sensors hold ±15% of reading or ±5 ppm accuracy with no routine calibration.
- The laser photoacoustic (LPAS) detection cell has no consumables and its laser diode is rated for more than 50,000 hours of continuous operation.
How GC Works
Oil sample → vacuum extraction → carrier gas → chromatographic column (separation by retention time) → FID (hydrocarbons) + TCD (H₂/O₂/N₂). Gold standard: the only technology recognized in every major standard (IEC 60567, ASTM D3612, IEEE C57.104, CIGRE TB 771). Lab GC achieves <1 ppm detection with ±5% accuracy.
How PAS Works
Membrane permeation → tunable laser at gas-specific IR wavelength → photoacoustic signal proportional to concentration (S = k × α × P × C). No carrier gas, no columns, no consumables. R² >0.95 vs GC in field validation. Full PAS guide →
Head-to-Head Comparison
| Criterion | GC (Lab) | PAS (Online) |
|---|---|---|
| Reference Method? | Yes | Field-validated (R²>0.95) |
| Gases | 7–9 + O₂/N₂ | 9 + moisture |
| C₂H₂ Detection | <0.5 ppm | 0.1 ppm |
| Data Points/Year | 1 | 365–8,760 |
| Consumables | $1,200/yr carrier gas + columns | Zero |
| Maintenance | 4–12 visits/year | <5 min/year visual |
| Operator Skill | Trained chemist | None (automated) |
| 10-Year TCO | $35K–55K (lab service) | $12.5K–23K |
| Best For | Reference, compliance, disputes | Continuous monitoring, early warning |
10-Year TCO Breakdown
| Cost | Lab GC | Online GC | PAS (DGM-900) |
|---|---|---|---|
| Hardware | N/A | $35K–55K | $12K–22K |
| Installation | N/A | $3K–5K | $500–1K |
| 10yr Consumables | $30K–50K | $25K–40K | $0 |
| 10yr Labor | $5K | $45K–90K | $0 |
| Total 10-Year | $35K–55K | $80K–150K | $12.5K–23K |
How to use online PAS DGA alongside laboratory gas chromatography
- Step 1: Use laboratory GC for the reference — standardized offline analysis is the established benchmark for dissolved gas measurement.
- Step 2: Install an online PAS monitor for continuity — it measures directly in oil on a 1–24 hour cycle with no carrier gas.
- Step 3: Let the monitor run between lab samples — the trend shows gas development week by week instead of at single points in time.
- Step 4: Compare at the overlap — validate the online reading against each new lab result to keep the two methods aligned.
- Step 5: Use each where it is strong — online PAS for alarms and trending; GC for confirmatory, accredited laboratory reporting.
FAQ
Q: What is the difference between online and offline DGA?
A: Offline (laboratory) DGA involves manually collecting an oil sample and sending it to a laboratory for analysis, typically once per year. Online DGA uses a permanently installed monitor that automatically extracts and analyzes gases at programmable intervals (hourly to daily).
Q: Do PAS DGA monitors require calibration?
A: No routine calibration is required. The palladium thin-film hydrogen sensor is inherently stable — the Pd-H reaction is a fundamental physical property, not a consumable chemical reaction.
Why PAS and GC measure the world differently
The two methods answer the same question — “how much of each dissolved gas is in the oil?” — through completely different physics, and that difference explains every trade-off in the tables above.
| Gas chromatography | Photoacoustic spectroscopy | |
|---|---|---|
| Principle | Physical separation — gases are carried by a carrier gas through a column and detected one by one as they elute | Optical absorption → sound — each gas is excited by modulated IR light at its own wavelength and the resulting acoustic signal is measured |
| What is measured | Peak area per retention time (TCD / FID) | Acoustic amplitude at the modulation frequency (microphone / tuning fork) |
| Reference | The reference method named in IEC 60567, ASTM D3612, IEEE C57.104 and CIGRE TB 771 | Field-validated against GC; not itself named in the standards |
| Operating mode | Batch — a discrete sample is separated and quantified | Continuous — the cell reaches equilibrium with the oil and reports on a cycle |
Engineering trade-offs that matter in the substation
| Dimension | PAS | Online / lab GC |
|---|---|---|
| Gas coverage | Typically 3–7 gases; H₂ needs a separate sensor | 7+ gases including H₂/O₂/N₂ in one run |
| Time to result | ~10–30 minutes, compact | ~30–60 minutes, larger footprint |
| Consumables & maintenance | Essentially none — no carrier gas, no column | Carrier gas + columns (replaced every 6–12 months); heavier maintenance |
| Field weaknesses | Sensitive to vibration, noise and temperature drift; multiple gases need multiple light sources | Large, costly, needs gas supplies and a climate-controlled environment |
A caution on accuracy figures. Side-by-side accuracy numbers for PAS and GC are drawn from different studies run under different conditions, and some originate from vendor-leaning articles. Any single “PAS is X% accurate versus GC” figure should therefore be read as indicative, not as a specification — and the underlying validation data should always be requested. The comparison that matters in procurement is the field correlation on your own gases, not a headline percentage.
See Photoacoustic spectroscopy for the measurement principle in detail, and DGA technology comparison for how PAS, GC, NDIR, FTIR and TDLAS compare across the online-DGA market.