0107DB1401

White Paper 12/2014

Dissolved Gas Analysis

Retain for future use.

Abstract

Oil-filled transformers are a major component in medium voltage distribution. Due to their critical contribution, it is important to
monitor the transformers over their period of operation, typically comprising of twenty years or more. Dissolved Gas Analysis
(DGA) is a diagnostic method used to ascertain normal operating conditions of oil-filled transformers. This paper will cover the
process of dissolved gas analysis, different gases produced during cellulose and mineral oil degradation, gases generated under
fault conditions, and testing procedures. The paper highlights the issue of faulty diagnosis caused by dissolved gases during
normal operating conditions and the effect of transformer size on dissolved gas threshold. Methods to avoid faulty diagnosis and
improving the testing procedure are discussed.

Introduction

During a fault, oil-filled transformers generate gases which help in identifying abnormal working conditions. The generated gases
can be found dissolved in the insulating oil or in the space above the oil. The analysis of the type of dissolved gases generated
based on the insulating medium helps in determining the fault conditions comprising of one or more thermal and electrical faults.
However, it should be noted that the process of diagnosis of dissolved gases and the interpretation of its results is more of an art
than exact science. The variables in the detection process comprise of the following:

1. Fault type, its temperature and location;

2. Type of oil preservation system;
3. Solubility and degree of saturation of different gases in oil;
4. Rate and type of oil circulation;
5. Type of material in contact with the fault; and
6. Differences in sampling and measurement procedures.

Additionally, gases are produced within the transformer during the manufacturing and assembling process. Some of the typical
gases present under normal operating conditions will be discussed in later sections of this paper.
There are some industry standards guides for DGA, consisting of IEEE Standard C57.104-2008 "Guide To the Interpretation of
Gases in Oil Immersed Transformers" and IEC 60599 "Mineral Oil-Impregnated Electrical Equipment in Service - Guide to the
Interpretation of Dissolved and Free Gases Analysis". The scope of the IEEE guide comprises of:

1. Detection of dissolved combustible gases;

2. Operating condition guidelines;
3. Interpretation of results based on Total Dissolved Combustible Gas (TDCG) parameters provided in the guide; and
4. Providing diagnostic techniques and instruments for detecting dissolved combustible gas.

Mineral transformer oil is made up of hydrocarbons and its decomposition results in creation of combustible gases: hydrogen
(H2), methane (CH4), and ethane (C2H6). Further decomposition of mineral transformer oil results in generation of ethylene
(C2H4) and acetylene (C2H2). For example, presence of acetylene gas suggests possibility of high temperature arcing fault, while
the presence of methane suggests low energy electrical or thermal fault. The IEEE TDCG threshold guidelines are used to
determine occurrence and type of faults in mineral oil-filled transformers. The thermal decomposition of oil-impregnated cellulose
insulation produces carbon monoxide (CO) and carbon dioxide (CO2) due to the cellulose, with trace amounts of hydrogen and
methane due to the oil. Carbon dioxide gas is non-combustible; hence, the CO2/CO ratio is used to determine cellulose insulation
degradation. For normal operation this ratio should be in the range of three to ten. The amount of gas produced in cellulose
decomposition is directly proportional to the volume of insulating material; hence, the transformer size is significant in determining
fault conditions.

© 2014 Schneider Electric All Rights Reserved

Dissolved Gas Analysis 0107DB1401
White Paper 12/2014

Types of Faults

Common faults in the transformers are arcing, overheating, overloading, partial discharge, and low-energy sparking. These can
be broadly categorized into either thermal or electrical faults.

Thermal Faults

Depending on the temperature a mix of hydrocarbons are produced during mineral oil decomposition. During the temperature
range of 150 to 500°C (302 to 932°F) large quantities of lower molecular weight gases like hydrogen and methane are produced
with trace amounts of higher molecular weight gases like ethylene and ethane. At slightly higher fault temperatures the amount of
hydrogen gas production exceeds that of methane, accompanied with significant quantities of ethane and ethylene. Finally at
very high temperatures of up to 700°C (1292°F), increasing amounts of hydrogen and ethylene gases are produced with trace
amounts of acetylene.
In case of cellulose, thermal decomposition produces carbon oxides (CO, CO2) and water vapor along with trace amounts of
combustible gases, hydrogen and methane. Since cellulose decomposes at temperatures lower than oil, trace amounts of CO
and CO2 are present during normal operating conditions.

Electrical Faults—Low Intensity Discharges and High Intensity Arcing

Low intensity discharges consist of partial discharges and low intermittent arcing. They result in the production of mainly
hydrogen, some amount of methane and trace amounts of acetylene. As the intensity of the discharge increases to an arc or
continuous discharge proportions, the percentage of acetylene and ethylene increases in the dissolved gas.

Operating and Testing Procedure

Detection, Evaluation and Action are three main steps to follow from an operational point of view. To identify the presence of any
gases which exceed normal operational levels and to ensure that it is detected in earliest possible time to minimize damage, is
termed as Detection. Note that it can be difficult to determine normal operating gas levels of a unit if its DGA history is not
available. This could lead to faulty diagnosis and additional tests to determine the cause of abnormality. IEEE Std. C57.104-2008
states that in the absence of transformer DGA history a four-level criterion for fault diagnosis as shown in Table 1, can be referred
to as a starting point of analysis. The table was developed based on data collected from several large power transformers. In the
four-level criterion Condition 1 indicates satisfactory transformer operation, Conditions 2–4 indicate abnormal transformer
operating conditions where 2 being least severe and 4 being most severe.

Table 1: Dissolved Gas Concentrations

Dissolved Key Gas Concentration Limits [µL/L (ppm)a]

Status Hydrogen Methane Acetylene Ethylene Ethane Carbon Monoxide Carbon Dioxide
TDCGb
(H2) (CH4) (C2H2) (C2H4) (C2H6) (CO) (CO2)
Condition 1 100 120 1 50 65 350 2500 720
Condition 2 101–700 121–400 2–9 51–100 66–100 351–570 2501–4000 721–1920
Condition 3 701–1800 401–1000 10–35 101–200 101–150 571–1400 4001–10000 1921–4630
Condition 4 >1800 >1000 >35 >200 >150 >1400 >10000 >4630
NOTE 1: Table 1 assumes that no previous tests on the transformer for dissolved gas analysis have been made or that no recent history exists. If a previous analysis
exists, it should be reviewed to determine if the situation is stable or unstable. Refer to 6.5.2 for appropriate action(s) to be taken.
NOTE 2: An ASTM round-robin indicated variability in gas analysis between labs. This should be considered when having gas analysis made by different labs.

a The numbers shown in Table 1 are in parts of gas per million parts of oil [µL/L (ppm)] volumetrically and are based on a large power transformer with several
thousand gallons of oil. With a smaller oil volume, the same volume of gas will have a higher gas concentration. Small distribution transformers and voltage regulators
may contain combustible gases because of internal expulsion fuses or load break switches. The status codes in Table 1 are also not applicable to other apparatus in
which load break switches operate under oil.
b The TDCG value does not include CO2 which is not a combustible gas.

2 © 2014 Schneider Electric All Rights Reserved

0107DB1401 Dissolved Gas Analysis
12/2014 White Paper

After determining the abnormality, its impact on the operation of the transformer is evaluated based on a set of guidelines and
recommendations. This process is termed as Evaluation. Finally, appropriate action is taken to remedy the fault condition, which
could vary from reducing the load on the transformer to removing the transformer from service in extreme cases. For further
description on the operating and testing procedure for TDGA refer to IEEE Standard C57.104-2008 [1].

Effect of Transformer Size on TDCG Threshold

“Note a” in Table 1: Dissolved Gas Concentrations in IEEE C57.104-2008 mentions that the values representing possible fault
conditions are based on large power transformer data and the same criterion cannot be applied directly to small transformers.
Small distribution transformers may contain some combustible gases due to internal expulsion fuses and load break switches.
Apart from the "Note a", which forewarns against directly applying the table for transformer DGA irrespective of size, there is
another criteria which needs attention while doing DGA. Carbon monoxide and carbon dioxide gases are produced during
cellulose degradation and as transformer size increases the ratio of oil to cellulose insulation increases. The oil to paper ratio
impacts the generation of carbon oxides, hence, CO and CO2 gases generated are higher for smaller transformers and could flag
fault conditions. An independent study done by Weidmann Diagnostics [2] states that based on the conservative IEEE C57.104-
2008 carbon oxide levels the 90th percentile level for CO and CO2 gases though abnormal for all three size ranges, is worse for
small transformers as shown in Table 2. Note that the normal range for CO2/CO ratio is three to ten.

Table 2: Percentiles for CO2/CO Ratio

Size Range 90th Percentile 10th Percentile

30–2500 kVA 17 3.7
2500–10000 kVA 17 3.2
>10000 kVA 13 3.0

Gases Present under Normal Operating Condition

Based on the analysis of several new transformer samples, it has been noted that dissolved gases are present in the insulating oil
or in the gas blanket just after transformer installation. Gases trapped in winding during processing and testing come out when
the oil is getting filled. Additional presence of moisture should be checked as the transformers are supposed to reach the site dry,
and then are filled up. It should be noted that only very large transformers are shipped without oil. Some of the possible dissolved
gases based on tests are as follows:
Carbon Oxides–Carbon oxides are produced during thermal decomposition of cellulose and the CO2/CO ratio of less than three
and greater than ten indicates abnormality. However, the transformer manufacturing process of removing moisture by subjecting
the units to 100°C (212°F) creates carbon oxides in the paper insulation. Once oil filled, these gases will dissolve into the oil
insulation for a period of time until equilibrium is attained between the gas in the paper and oil causing Condition 2-4 threshold
levels. Test samples taken from these transformers will show abnormality, hence, in order to determine whether cellulose
degradation has actually taken place we have to perform the Furan Test. This is a non-evasive test used to determine the
remaining life of the transformer insulation.
Carbon Oxides are also produced during the process of cover welding in transformers. The test conducted by Westinghouse
Electric Corporation on four three-phase distribution transformers before and after cover welding showed increased amount of
carbon oxides along with trace amount of hydrocarbons like acetylene and ethylene [3].
Acetylene–The transformers are manufactured in such a manner that cover welding is completed post oil filling and testing. It is
common for oil to reach the upper lip while being transferred from the testing station to the cover welding station. Once arriving at
the cover welding station, the cover is placed on the top lip and welded into place. The temperature of the residual oil on the
upper lip can exceed 700°C (1292°F) during welding of the cover, at which point gases, including acetylene, will be produced.
The manufacturing facilities use a gas purging system inside the head space of the transformer to assist the removal of acetylene
and other gases; however there are cases when all of the gases, including acetylene, will not be removed. The acetylene level
could be as high as 3 parts per million which is equivalent to Condition 2 level defined by IEEE [1]. This could cause faulty
diagnosis of possible arcing in the transformer. However, in case of high energy arcing, other hydrocarbons like hydrogen,
methane, and ethylene are also produced and in the absence of these hydrocarbons it is safe to conclude that acetylene is
introduced by process of cover welding.

© 2014 Schneider Electric All Rights Reserved 3

Dissolved Gas Analysis 0107DB1401
White Paper 12/2014

Moisture–Moisture may condense on any surface cooler than the surrounding air. When a transformer is installed at a location
warmer than itself, proper care should be taken to determine that all external condensation has disappeared. Additionally
transformers should not be opened on a day with humidity above 70%, without precautions to prevent entrance of moisture. The
manufacturing process causes trapped moisture in the transformer winding. If insulating oil is filled without drying the transformer
thoroughly then this trapped moisture along with presence of carbon oxides produced in welding will mimic the thermal
decomposition of cellulose.

Conclusion

The dissolved gas analysis process is based on several variants, hence it is considered more of an art than science. However, by
improving the sample taking process, analysis and prior knowledge of possible gases present during normal operating conditions,
faulty diagnoses can be minimized. Multiple samples–instead of one–should be collected over a period of time in order to
determine the trend of the dissolved gases in transformers. DGA and moisture content testing should be done after the first day of
installation in order to identify presence of gases introduced by the manufacturing and installation process. This data will help
determine the normal operating conditions for the transformer and future test results can be analyzed based on these threshold
values. A follow up DGA test should be conducted at three months, then at six months, and finally based on what the sample
looks like, it could be done on one- or three-year intervals.
Transformer manufacturers and laboratories should analyze the first gas sample after transformer installation. The laboratories
follow IEEE threshold values, whereas the transformer manufacturers have prior knowledge of possible gases generated during
the manufacturing process and over the normal operating lifetime of the transformers. Together they can come up with a more
realistic DGA. The transformer size should be considered during sample analysis as smaller transformers could have a higher
CO2/CO ratio without occurrence of a fault. IEEE is looking at 44,000 transformer data samples for differences between small
versus large transformers, however due to dependence on several variants, it will be some time before we see any conclusion.

References:

[1] IEEE Standard C57.104-2008, "Guide for the interpretation of Gases in Oil Immersed Transformers".
[2] "Dissolved Gas Threshold Levels Based on Transformer Size", Thomas A. Prevost, Weidmann Diagnostic Solutions Inc.
[3] "Generation of Combustible Gas During the Cover Welding of Distribution Transformers", D.J. Struemph, Westinghouse
Electric Corporation, Distribution Transformer Division.

Electrical equipment should be installed, operated, serviced, and maintained only by

Schneider Electric USA, Inc. qualified personnel. No responsibility is assumed by Schneider Electric for any
800 Federal Street consequences arising out of the use of this material.
Andover, MA 01810 USA Schneider Electric and Square D are trademarks owned by Schneider Electric
888-778-2733 Industries SAS or its affiliated companies. All other trademarks are the property of
www.schneider-electric.us their respective owners.

4 © 2014 Schneider Electric All Rights Reserved