Why Does Hydrogen Form in Transformer Oil During Service?

After the fault is corrected, hydrogen stops increasing and repeat dissolved-gas results establish a stable baseline. A hydrogen indication alone does not identify the fault: validate the sample, examine the accompanying gases, inspect the transformer, and judge the rate of change before selecting corrective work.

What is the DGA screen telling you?

The dissolved-gas analysis result reports hydrogen released into the transformer oil. Transformer oil contains long-chain hydrocarbon molecules. Local electrical or thermal energy can break those molecules into smaller fragments, including hydrogen and hydrocarbon gases such as methane, ethane, and ethylene.

Much of the generated gas dissolves in the oil. Some can migrate into the transformer gas space. The displayed hydrogen value therefore represents gas captured by a particular sampling and analytical path, not a direct measurement at the fault location.

Displayed condition	What it means	Next check
Hydrogen appears in one result	Gas generation, retained historical gas, or a sampling problem may be involved.	Validate the sample record and obtain the previous result.
Hydrogen rises across comparable tests	Gas is being generated faster than it is removed, dispersed, or otherwise lost from the sampled oil.	Calculate the change over the actual interval and inspect for active electrical symptoms.
Hydrogen is present with other gases	The combined gas pattern provides more diagnostic separation than hydrogen alone.	Compare methane, ethane, ethylene, and every other reported gas.
Oil and gas-space results differ	Gas partitioning and the sampling method affect the measured result.	Trend like-for-like samples rather than mixing methods.

The first commissioning check is simple: confirm whether the screen shows an isolated measurement or a repeatable trend from the same transformer and sampling method.

Can the sample be trusted before diagnosing a fault?

The tag is right; the binding is wrong has an analytical equivalent: the laboratory result may be correct for the bottle while the bottle does not represent the transformer. Review sample identity, date, sampling point, transformer operating condition, and analytical method before treating the value as equipment behavior.

Setting or record	Location	Diagnostic effect
Transformer identity	Sample label and laboratory report	Prevents comparison with another unit or compartment.
Sampling point	Field sheet or maintenance record	Keeps successive results representative of the same oil volume.
Sample date	Report header	Defines the interval used to judge the rate of change.
Operating state	Operator log	Connects gas generation with loading, switching, alarms, or abnormal operation.
Oil treatment or maintenance	Maintenance history	Explains a discontinuity caused by degassing, oil replacement, or internal work.
Analysis basis	Laboratory result	Prevents direct comparison of unlike dissolved-oil and gas-space measurements.

Dissolved-oil sampling and gas-space analysis can both detect fault gases. For trending, use the same approach because gas distribution between oil and the gas space changes the reported quantities. If a result conflicts with operating history or previous data, obtain a confirmation sample before escalating the diagnosis.

The check passes when the transformer, compartment, sampling point, dates, and methods match across the results being compared.

How does hydrogen form inside transformer oil?

Hydrocarbon molecules contain hydrogen and carbon bonded in long chains. Intense local energy breaks chemical bonds and creates smaller molecular fragments. Some fragments remain dissolved gases; others migrate toward a gas space where they may also be sampled.

Arcing can supply high localized electrical energy. Corona-type or partial-discharge activity can also decompose oil in a small region. Thermal stress can crack hydrocarbons as local temperature rises. Pressure can accompany an internal fault, but the gas pattern and operating evidence are more useful for diagnosis than pressure alone.

Hydrogen is therefore a sensitive warning gas but not a unique fault label. Its presence does not, by itself, prove that the bulk oil has lost its insulating capability. A localized discharge can generate gas while a separate bulk-oil test still produces an acceptable result. Conversely, poor oil condition can increase electrical risk without identifying where discharge is occurring.

Verify the mechanism by looking for a repeatable hydrogen increase and independent evidence of electrical or thermal activity, rather than assigning the fault from one gas value.

Which accompanying gases separate the likely causes?

Read the complete gas panel. Methane, ethane, ethylene, and hydrogen are smaller products formed when transformer oil decomposes, but different energy conditions produce different mixtures. The relative pattern, its change over time, and the transformer condition must agree before maintenance targets a cause.

Observed pattern	Working interpretation	Decision
Hydrogen without a confirmed rise	A single result cannot distinguish retained gas, early activity, or sampling variation.	Confirm the sample and establish a second comparable point.
Hydrogen rising with little change elsewhere	Investigate low-energy electrical activity, including corona-type discharge, while checking for sampling effects.	Review electrical symptoms and schedule a repeat analysis.
Hydrogen rising with several hydrocarbon gases	The broader pattern indicates oil decomposition under an active energy source.	Use the whole pattern to choose electrical or thermal inspections.
Hydrogen stable while other gases change	Hydrogen is not the controlling diagnostic variable.	Follow the gases that are changing and the matching operating evidence.
All reported gases shift after oil work	Maintenance may have changed the gas inventory or sample basis.	Create a documented post-maintenance baseline.

Do not diagnose arcing solely because hydrogen is present. Arcing and corona-type activity are candidates; the other gases and field observations decide between them. The check passes when one working diagnosis explains the full gas pattern, its trend, and the transformer symptoms without ignoring contradictory data.

What inspections connect the gas result to the transformer?

Trace the indication from report to sample, then from sample to equipment. The objective is to locate corroborating evidence without disturbing a transformer unnecessarily.

1. Review the operating timeline. Compare the sampling dates with loading changes, abnormal temperature indications, protection activity, switching events, unusual sound, and maintenance.
2. Inspect accessible external conditions. Check for leakage, abnormal heating indications, gas accumulation where applicable, contamination paths, and cooling problems.
3. Review electrical diagnostics. Examine protection records and available discharge-related or insulation measurements for evidence that agrees with the gas trend.
4. Test oil properties separately. If the question is whether the oil still performs as insulation, request the applicable oil-quality and dielectric tests. These tests answer a different question from gas analysis and do not locate an internal fault by themselves.
5. Choose internal investigation only from combined evidence. A rising gas trend plus matching electrical, thermal, or protection evidence provides a stronger basis than a single hydrogen result.

Gas analysis and oil-condition testing are complementary. The first detects decomposition products; the second evaluates properties of the sampled oil. Passing one does not cancel a warning from the other.

The check passes when the selected inspection or test addresses the suspected mechanism and produces a result that can be compared with the DGA trend.

When should hydrogen be retested?

Trend direction and rate carry more diagnostic value than one test result. A practical follow-up from this case is to retest in about one month. Treat that interval as a case-specific monitoring choice, not a universal rule. Transformer condition, the size and rate of the change, operating symptoms, criticality, and protection activity determine whether testing must occur sooner.

Compare only equivalent measurements. Record the elapsed time rather than calling two values simply “previous” and “current.” The basic trend calculation is:

Hydrogen rate of change = (current result - previous result) / elapsed time

Use the concentration units and time basis printed on the laboratory reports. Do not invent an alarm threshold when none is supplied. Obtain action limits from the transformer owner’s approved diagnostic procedure, the laboratory interpretation, and the equipment documentation.

Follow-up result	Meaning	Action
Confirmed upward trend	Ongoing gas generation is likely.	Escalate the matching electrical or thermal investigation and shorten monitoring as directed by the asset procedure.
Stable comparable results	No measurable increase occurred over that interval.	Retain the value as a baseline and continue scheduled trending.
Lower result after oil treatment	The gas inventory changed; the underlying source may or may not be removed.	Trend from a documented post-treatment baseline.
Large unexplained step change	An active fault, sampling problem, or changed test basis must be separated.	Check records and obtain confirmation promptly.

The check passes when the repeat interval is documented, the measurements are comparable, and the calculated direction of change is unambiguous.

How is the correction verified end to end?

Correct the condition identified by the combined diagnosis, not the hydrogen number itself. Degassing or replacing oil can lower measured gas, but it does not prove that an electrical or thermal source has been removed. Verification must follow the complete path from equipment condition to laboratory result.

1. Document the pre-work gas results, sample point, operating condition, and suspected mechanism.
2. Record the corrective work and any oil processing that changes the gas inventory.
3. Take a post-work baseline using the same defined sampling point and analytical basis.
4. Return the transformer to its intended monitored service condition.
5. Repeat the analysis at the approved interval and calculate the hydrogen change over elapsed time.
6. Compare the complete gas panel, not hydrogen alone, with the post-work baseline.
7. Confirm that associated electrical, thermal, protection, or operating symptoms have not returned.

The correction is demonstrated when comparable post-work tests show no continuing upward hydrogen trend and the corroborating transformer symptoms remain absent.

FAQ