FEATURE

PROGRESS IN
TRANSFORMER
DISSOLVED-GAS
ANALYSIS
BY J AM ES J . DUK ARM , Delta-X Research Inc.

Dissolved-gas analysis started out many years ago as a quick and simple
test: Someone in the substation would briefly open a valve and sniff the
transformer’s head-space gas. The pungent odor of a trace of acetylene
would signify a potentially serious problem.
By the late 1960s, it was possible to use a gas
chromatograph as a much more sensitive nose for The Fundamental Principle
sniffing transformer oil. Pioneers such as R. R. of Transformer DGA
Rogers and E. Doernenburg realized that the fault
gases in oil came from breakdown of the solid A transformer is designed not to damage its
and liquid insulation and intuitively understood internal insulation in the course of normal
the fundamental principle of transformer DGA operation. Therefore, if insulation deterio-
(sidebar). They collected DGA data and came up ration byproducts dissolved in the oil are
with ideas for (a) detecting problems, (b) assessing increasing (beyond what is expected due to
their severity, and (c) identifying the general normal aging), something is wrong.
nature of the problem.

40 • FALL 2019 PROGRESS IN TRANSFORMER DISSOLVED-GAS ANALYSIS

 FEATURE
and others. A good summary of limits-based
DGA interpretation was published by Hamrick
in NETA World, Winter 2009.

Building on five decades of industry experience

and data collection, plus the wide availability
of computers, a re-examination of transformer
DGA from the point of view of physical
chemistry and advanced statistics is breathing
new life into the subject. This article describes
important advances. An example shows how
they can improve fault detection and provide
new risk assessment information.

RECENT I NNOVATI ONS I N

DGA
Since about 2014, several innovations have
increased the usefulness and effectiveness of
DGA interpretation.

• Accounting for gas loss. Gas loss —

whether by design or by accident such as
leakage through a bushing gasket or air
exposure of a DGA sample — can be a
serious problem for DGA interpretation,
especially when based on gas concentration
and rate of change limits. It is helpful
to work with cumulative data to avoid
overlooking serious problems.
• Fault energy indexes for trend detection
To detect problems, gas concentration and and severity assessment. It is known
rate of increase limits were developed based on that some fault gases are more significant
the reasonable assumption that an unusually than others. For example, ethylene and
high fault gas concentration or rate of change acetylene are associated with extremely
should be a sign of trouble. To assess severity, high-temperature faults. The physical
additional limits and considerations of rates basis for the differing significance of fault
of increase were employed to get a grade- gases is their heats of formation from
school report card result of OK, so-so, or solid or liquid insulation. Those heats
bad, expressed in North America as numeric of formation, weighted by the respective
condition code scores from 1 (OK) to 4 gas concentrations in oil, can be used to
(terrible). The limits-based approach to DGA calculate normalized energy intensity
interpretation has been refined over the years, (NEI) for fault energy indexes to use for
and the IEEE C57.104 and IEC 60599 gas trending, fault detection, and severity
assessment.
guides are considered the authoritative sources
on how to apply it. Those guides include the • Focus on gassing events. Concentrating
gas ratio methods for fault type identification on time intervals in which a fault energy
developed by Doernenburg, Rogers, Duval, index is trending upwards (boxed intervals

PROGRESS IN TRANSFORMER DISSOLVED-GAS ANALYSIS NETAWorld • 41

FEATURE
in Figure 4 and Figure 5) is a natural • Reliability statistics relating DGA results
and very useful way to look at fault gas to transformer failure. Instead of assessing
production. severity in terms of limits exceeded, use a
• Gassing status score based on the statistical model of the fault energy index
fundamental principle of DGA. The distribution in transformers about to fail
fundamental principle of DGA provides to estimate prior risk exposure and risk of
a natural basis for ranking transformers near-term failure.
according to their apparent need for extra
attention. The transformer has either been RELI ABI LI TY STATI S TICS IN
gassing recently or not, and where there DGA
is gassing, it is either more or less severe The statistical model referred to in the previous
(gassing status sidebar). paragraph shows the distribution of fault
energy index values in gassing transformers
that are about to fail. The models (one for each
Gassing Status fault energy index) were derived from a large
1. No significant fault gas production DGA database with additional information
ever about transformer failures. The information
2. Some fault gas production, but not provided by the model of the failure-related
recently values of the hydrocarbon gas fault energy
3. Recent moderate fault gas
production index (NEI-HC) is summarized by the failure
4. Recent extreme fault gas rate graph shown in Figure 1. The four vertical
production dotted lines represent (left to right) the 90th,
95th, 98th, and 99th percentiles of cumulative
hydrocarbon gas fault energy index (NEI-HC)
• Gas increments over gassing events for
in a large DGA database. The peak failure rate
fault type identification. The Duval
occurs at about the 82nd percentile, well below
triangle (Figure 6) is a very good method
for fault type identification. When
0.005

trying to identify the cause of fault gas

production during a gassing event, don’t
use gas concentrations, which include
0.004
Failure rate per unit of NEI

possibly irrelevant pre-event gas. Instead,

use gas increments calculated between the
0.003

earliest and latest oil samples in that time

interval.
0.002

• Percent change in the CO/CO₂ ratio for

locating paper degradation. The carbon
0.001

oxide gas ratio can be used to estimate the

approximate location of a fault affecting
0.000

paper insulation. It also sometimes gives

20 40 60 80 100 120
early warning of a developing problem. A
Cumulative NEI-HC
strong increase (by 200% or more) suggests
a hot spot affecting winding insulation. Figure 1: The curve shows transformer
A smaller increase suggests that paper failure rate (fraction of surviving population)
insulation outside the windings, such as as a function of the cumulative hydrocarbon gas
on bushing or LTC leads, may be affected. fault energy index (NEI-HC). Vertical dotted
A decrease may indicate CO₂ production lines represent the 90th, 95th, 98th, and 99th
due to general low-range overheating of percentiles of cumulative NEI-HC in a large
paper insulation. DGA database.

42 • FALL 2019 PROGRESS IN TRANSFORMER DISSOLVED-GAS ANALYSIS

 FEATURE
the 90th percentile, suggesting that waiting for consider the example of a 250 MVA 230 kV
something to exceed the 90th percentile before nitrogen-blanketed transformer manufactured
investigating may not be a good idea! in the early 1980s. The transformer's fault
gas levels were unexceptional, except for a
If a transformer with very little fault gas begins persistently high CO₂ concentration averaging
to produce hydrocarbon fault gas, it should about 4200 ppm. One day in 2011, the
raise immediate concern since the associated transformer experienced turn-to-turn arcing
failure risk is increasing very steeply. As NEI- and was removed from service. A post-mortem
HC increases further, the failure rate decreases, surprisingly revealed very extensive charring
indicating that — contrary to how higher of winding insulation paper and pressboard
DGA limits are often interpreted — continued spacers, suggesting that the windings had been
gassing does not necessarily imply worsening overheating for a long time.
reliability. This means that either:
The problem had gone unnoticed for years
a. Whatever is causing the gassing is not because periodic gas expulsion by the head
very harmful to the transformer and may space nitrogen pressure regulation system
continue indefinitely; or had prevented upward trends in heat gases
(methane, ethane, and ethylene) from
b. The transformer is gassing because it
developing. Except for the consistently high
is damaged or defective, and the next
CO₂ levels, no gas concentration or rate of
through fault may kill it; or
change limits were ever exceeded until the day
c. Something between the extremes of (a) the transformer failed.
and (b) is going on.
Figure 2 shows that the transformer’s hydrogen
and hydrocarbon gas concentrations were
C AS E S TU DY consistently low to moderate with a lot of
The new DGA approach using reliability bumps. Methane and ethane increased in
statistics was evaluated on 7,200 transformers 2008–2009, but no limits were exceeded.
in 2016 by a large US electric utility. It Those upward trends were reversed in 2009–
performed so well, identifying many previously 2010.
undetected serious problems, that the utility
immediately adopted it as a key part of its Figure 3 shows that oxygen and CO₂ levels
transformer condition assessment system. were consistent with some ups and downs.
The average CO₂ concentration was about
To see how the innovations mentioned above 4200 ppm. CO concentration was variable
improve DGA for oil-filled transformers, and always lower than 60 ppm. Nitrogen

H2
101 CH4
µL/L

C2H6
C2H4
C2H2

100

1994 1996 1998 2000 2002 2004 2006 2008 2010

Figure 2: Hydrogen and Hydrocarbon Gas Concentrations

PROGRESS IN TRANSFORMER DISSOLVED-GAS ANALYSIS NETAWorld • 43

FEATURE
105

104

103 N2
O2
µL/L 102 CO2
CO
101

100

1994 1996 1998 2000 2002 2004 2006 2008 2010

Figure 3: Atmospheric and Carbon Oxide Gas Concentrations

was consistently very high, as expected in a Figure 4 and Figure 5 show how long-term
nitrogen-blanketed transformer. fault gas production is revealed by:

As Figure 2 and Figure 3 illustrate, it can be a. Using cumulative data to compensate

difficult to understand what is happening by partially for gas loss
trending and assessing multiple fault gases. On
the other hand, fault energy indexes — one b. Trending fault energy indexes
for the oil and one for the paper insulation —
show when significant fault gas is produced The raw (non-cumulative) values of the fault
and provide a sound basis for assessing severity. energy indexes are shown as gray plus (+) signs,
Compare Figure 4 and Figure 5 with Figure 2 while the cumulative values are plotted as a
and Figure 3. solid line. Boxed intervals on the accumulated

5
0.4
NEI-HC [kj/kL]

0.2

0.0
1994 1996 1998 2000 2002 2004 2006 2008 2010

Figure 4: The heavy black line represents the hydrocarbon gas fault energy index (NEI-HC)
calculated from cumulative gas concentrations. The dotted line with plus symbols represents NEI-HC
calculated from raw gas concentrations.
6
4
NEI-CO [kj/kL]

20

10 1

0
1994 1996 1998 2000 2002 2004 2006 2008 2010

Figure 5: The heavy black line represents the carbon oxide gas fault energy index (NEI-CO)
calculated from cumulative gas concentrations. The dotted line with plus symbols represents NEI-CO
calculated from raw gas concentrations.

44 • FALL 2019 PROGRESS IN TRANSFORMER DISSOLVED-GAS ANALYSIS

 FEATURE
fault energy index graphs represent gassing of CO. The subject transformer's reported
events or time intervals when there appears CO concentration was consistently very low,
to be active fault gas production. Clearly, the however, and the CO/CO₂ ratio based on
upward trends in both cumulative hydrocarbon cumulative gas concentrations remained near
gas NEI and cumulative carbon oxide gas NEI 0.008 with very little variation. That could
would be difficult to notice by looking at the be explained by loss of CO as fast as it was
spaghetti graphs in Figure 2 and Figure 3 or produced; due to the low solubility of CO in
the raw numbers from the lab reports. Because oil, most of the CO in the oil migrated to the
there is evidently moderate ongoing carbon head space, where it was expelled by frequent
oxide gas production as of the latest sample, the pressure regulation gas release. The lack of any
gassing status of this transformer would be 3. warning of winding insulation deterioration by
the CO/CO₂ ratio in this case is an example of
For each of the gassing events indicated in the potentially serious impairment due to gas
Figure 4 and Figure 5, increments of methane, loss of DGA’s sensitivity to faults, even when
ethylene, and acetylene during the event are the gas loss is partially compensated for by the
used to plot a point on the Duval triangle use of cumulative data.
(Figure 6), with the most recent result plotted
as a red plus sign. The apparent fault type is The overall conclusion of this analysis using
consistently T1 — a thermal problem at less recent improvements in DGA is that, in spite
than 300° Celsius. of very significant gas loss due to headspace
pressure regulation, this transformer’s
What about the CO/CO₂ ratio? In a abnormal fault gas production could have been
transformer without gas loss, extensive charring detected several years before failure, raising the
of winding insulation would be expected to
transformer's gassing status to 3 and providing
cause a large increase (by more than 200%)
an opportunity for investigation and possible
of the CO/CO₂ ratio due to rapid production
mitigation of the thermal problem or at least
planning for eventual replacement of the
PD transformer. The application of DGA limits
DT without consideration of gas loss failed to
T1
detect that the transformer had a problem.

T2 The quantitative statistical results as of

%

August 2010 (seven months before failure)

C2
4
CH

show a 1.1% probability of failure with NEI-

H4
%

CO below 24.2, meaning that about 11

transformers out of 1,000 would have failed
at a lower level of cumulative NEI-CO. The
hazard factor or estimated time-based failure
rate as of that time was 0.18% per year,
calculated by multiplying the NEI-CO model’s
D1 D2 T3
failure rate (0.1% per NEI unit) times the
% C2H2 most recent rate of increase of NEI-CO (1.8/
year). These statistics indicate that the observed
Figure 6: Duval Triangle. Each plotted cross fault gas production to date, underestimated
is based on the increments (amounts of increase) by an unknown amount due to gas loss,
of methane, ethylene, and acetylene during a represents a modest amount of risk exposure,
gassing event. The red plus sign represents the with that risk continualy increasing. Whether
most recent gassing event. A persistent T1 thermal those results alone would have enabled the
fault (below 300° Celsius) is indicated. utility to avoid failure of the transformer is

PROGRESS IN TRANSFORMER DISSOLVED-GAS ANALYSIS NETAWorld • 45

FEATURE
questionable; however, with fair warning that International Electrotechnical Commission,
the dice were being rolled, it would have been IEC 60599-2015, Mineral Oil-Filled
possible to prepare for eventual replacement, Electrical Equipment in Service – Guidance on
and the eventual failure would not have been the Interpretation of Dissolved and Free Gases
surprising. Analysis, 3rd Edition, September 2015.
L. Hamrick, “Dissolved Gas Analysis for
C ONC LU SION Transformers,” NETA World, Winter 2009.
This analysis does not and cannot take into F. Jakob and J. J. Dukarm, “Thermodynamic
account the unknown amount of gas that Estimation of Transformer Fault Severity,”
was lost and never measured, so the severity IEEE Transactions on Power Delivery, Vol.30,
and risk level are understated to an unknown No.4, pp.1941-1948, August 2015.
degree. The new approach, however, could have C. Rutledge and R. Cox, “A Comprehensive
led to early discovery that the transformer was Diagnostic Evaluation of Power Transformers
gassing and that — as suggested by very low via Dissolved Gas Analysis,” 2016 IEEE/PES
hydrogen and CO levels — gas loss could be Transmission and Distribution Conference
masking the problem. Perhaps the pressure and Exposition, May 2016.
regulation system could have been locked
down at that time to stop gas expulsion long J. J. Dukarm and M. Duval, “Transformer
enough to obtain a more accurate assessment of Reliability and Dissolved-Gas Analysis,”
fault gas production, CO/CO₂ trend, and the 2016 CIGRE Canada Conference, Paper
associated hazard factor. CIGRE-807, Vancouver BC, October 2016.

R E F E R E NC E S Dr. James J. Dukarm is the founder and

DGA Sniff Test described by S. R. Lindgren, Principal Scientist of Delta-X Research Inc.
personal communication. in Victoria BC, Canada. As a member of
ASTM and CIGRE and an IEEE Life
CIGRE Working Group 15-01, “Application Member, Jim has been actively involved
of Physico-Chemical Methods of Analysis to in the development and revision of IEEE
the Study of Deterioration in the Insulation Transformers Committee DGA guides
C57.104, C57.139, and C57.155. He has co-authored several
of Electrical Apparatus,” Paper 15-07, 1970 papers in IEEE journals and NETA World and has presented
CIGRE Session, Paris France, 1970. numerous times at industry conferences. Jim received an MS in
mathematics from St. Mary’s University, San Antonio, Texas,
IEEE Std. C57.104-2008, IEEE Guide for and a PhD in mathematics from Simon Fraser University,
the Interpretation of Gases Generated in Oil- Burnaby BC, Canada.
Immersed Transformers, February 2009.

46 • FALL 2019 PROGRESS IN TRANSFORMER DISSOLVED-GAS ANALYSIS