溶解氣體分析 (DGA) 是評估電力變壓器內部狀況的關鍵工具。氫氣 (H2) 是最常見的特徵氣體之一，但其來源多種多樣，從正常的材料化學反應到局部放電和過熱等初期故障。準確識別氫氣來源對於變壓器狀態評估和維護決策至關重要。
This article systematically reviews the main sources, generation mechanisms, and identification methods for hydrogen in transformer oil, and highlights the drawbacks of relying solely on single‑gas hydrogen monitoring.
1.1內部故障
Mechanism: Weak discharge causes oil molecule cracking → H₂ generation.
Gas signature: H₂ dominates (>80%), trace CH₄ and C₂H₆ possible, almost no C₂H₂.
Key indicator: Isolated H₂ rise with very low other hydrocarbons.
Mechanism: Thermal breaking of C‑H bonds releases H₂.
Gas signature: CH₄, C₂H₆, C₂H₄ appear depending on temperature; H₂ level usually lower than in discharge cases.
Key indicator: H₂ rise accompanied by significant hydrocarbon gases (especially C₂H₄).
| Type | Mechanism | Typical Scenario | Gas Signature |
|---|---|---|---|
| Moisture‑involved reaction | Fe + H₂O → FeO + 2H → H₂ | High humidity, poor sealing | H₂↑ + moisture↑ |
| Cyclohexane catalytic dehydrogenation | Cyclohexane → Benzene + H₂ (Ni catalyst) | Stainless steel bellows expanders | Isolated H₂↑ (up to thousands ppm) |
| Metal corrosion | Electrochemical rusting produces H₂ | Rust in tank, core, coolers | Isolated H₂↑, possible moisture↑ |
Residual hydrogen in new oil: Dissolved during refining, transport, or filling – common in newly commissioned transformers, decreases with operation.
Improper oil treatment: Insufficient vacuum, short processing time, or poor temperature control can increase dissolved gases including H₂.
Core principle: Never rely on hydrogen concentration alone – combine multi‑dimensional information.
| H₂ Trend | Accompanying Gases | Moisture | Most Likely Source | Action |
|---|---|---|---|---|
| Isolated rise | No hydrocarbons, CO normal | Normal | Cyclohexane dehydrogenation, residual H₂ | Monitor only |
| Isolated rise | No hydrocarbons, CO normal | Elevated | Moisture‑related corrosion | Check sealing, treat moisture |
| Rise | CH₄, C₂H₄ etc. | Normal or elevated | Overheating | Electrical tests, plan inspection |
| Rise (dominant) | Trace CH₄, C₂H₆, no C₂H₂ | Normal | Partial discharge | PD measurement, consider outage |
| Rise | C₂H₂ present | Normal | Arcing (severe) | Immediate outage |
方法3：追蹤氣體趨勢
方法4：結合電氣測試和運行歷史記錄
Rapid continuous increase → Active fault, need outage
Peak then stable/declining → Commissioning “break‑in” or catalytic equilibrium
Seasonal fluctuation → Moisture‑related reactions
方法5：使用一氧化碳 (CO) – 避免誤報的關鍵
決定邏輯：
H2 + CO 483299086b23667a4484H2
Principle: CO is a specific byproduct of solid insulation (cellulose) thermal decomposition. Material/chemical hydrogen sources do not involve insulation heating, so CO remains normal.
CO
3。案例研究
案例1：不鏽鋼膨脹機脫氫（良性）d433c298c7cb4 83299086b23667a4484案例2：濕氣腐蝕+低溫過熱
4。結論：單一氣體氫氣監測的缺陷
Practical value: Using H₂ and CO as a paired indicator significantly reduces false alarms and avoids costly unnecessary inspections.
Situation: 220kV transformer, 3 months in service, H₂ = 1500 μL/L, other gases <1 μL/L, moisture normal, all electrical tests normal.
Diagnosis: Cyclohexane catalytic dehydrogenation.
Outcome: After 1 year, H₂ stabilized at ~800 μL/L – no action needed.
Situation: 110kV transformer, 10 years in service, H₂ = 380 μL/L, CH₄ = 45 μL/L, C₂H₄ = 28 μL/L, moisture rose from 12 to 25 mg/L.
Diagnosis: Moisture‑induced rust reaction with mild overheating.
Outcome: Internal inspection revealed core rusting; after treatment, H₂ returned to normal.
Relying solely on hydrogen monitoring carries significant false‑alarm risks because:
Hydrogen sources are diverse – benign (material/process) and fault‑related (discharge/overheating) are indistinguishable from H₂ alone.
Single‑gas data cannot identify fault type, severity, or location.
It can lead to two extremes: missing major faults (assuming benign) or frequent false alarms → unnecessary outages and wasted resources.
Recommendations:
Prefer multi‑component DGA (at least H₂ + CH₄ + C₂H₂ + CO + moisture).
Allow elevated H₂ in newly commissioned transformers, but track trends and check CO.
For isolated H₂ rise with normal CO – do not rush to outage; enhance monitoring or perform degassing.
Q: What is the typical alarm threshold for hydrogen in transformer oil?
A: Usually >150 μL/L warrants attention, but thresholds vary by voltage class and asset type – trend analysis is more important.
Q: How long does it take for hydrogen to drop in a new transformer?
A: Typically 1‑3 months of operation or hot oil circulation. If H₂ remains high with hydrocarbon gases, further investigation is needed.
Q: Are single‑gas hydrogen monitors still useful?
A: They can provide basic early warning for distribution‑class transformers, but for critical assets, upgrade to multi‑component DGA is strongly recommended.
HERTZINNO’s online DGA systems (DGA900, DGA500, DGA300) support multi‑component gas plus moisture monitoring, effectively avoiding false alarms caused by single‑gas hydrogen monitoring. Learn more →
