Escolar Documentos
Profissional Documentos
Cultura Documentos
Safety First
Back by the Best
On record as an industry
From the field, to our
leader in safe work practices
full-service rotating
apparatus services
Cutting-Edge
Trained from utility
transmission to
low-voltage distribution
Our experience in commercial, industrial, generation and specialized sectors, such as oil
refineries, pulp and paper, steel and wind power, uniquely qualifies us to handle the
largest, most complex assignments. Our people are trained on the latest technologies and
safety practices. And, we support them with full-service rotating apparatus services and a
large fleet of service vehicles, ready to respond 24/7.
Corporate Headquarters | 2425 East Pioneer Drive, Irving, Texas 75061 Why trust your vital power services to anyone less? Count on an industry leader. Published by
p. 972.793.5523, 888.SHERMCO | f. 972.793.5542 | www.shermco.com Visit www.shermco.com and learn more today. InterNational Electrical Testing Association
Insulating Oils
Handbook
Published by
InterNational Electrical Testing Association
THE “GO-TO”
STANDARDS FOR
ELECTRICAL SAFETY
AND RELIABILITY
KNOWLEDGE IS POWER
When you harness the power of knowledge, you unleash the possibilities. ANSI/NETA ATS-2009
This standard should always be referenced in design
specifications or when performing acceptance testing
The industry’s knowledge leader is your trusted partner. on power system installations. ANSI/NETA MTS-2011
For over 90 years, Doble has partnered with our utility clients to improve ANSI/NETA ETT-2010
operations and optimize system performance. Through the dedication This standard ensures that your acceptance
of world-class staff, Doble provides products and services in 110 and maintenance tests are being preformed by
countries around the globe, delivering the solutions you need – from qualified technicians who are certified in accordance
laboratory testing services to hands-on, technical training seminars. It’s a with ANSI/NETA ETT requirements.
relationship we value and a partnership that will allow us take the industry
to the next level. Together. Available in Bound, CD ROM, or PDF
STANDARD FOR
MAINTENANCE
TESTING SPECIFICATIONS
FOR ELECTRICAL POWER
When you work with Doble’s laboratories, you’re Doble Laboratory Services Include: EQUIPMENT AND SYSTEMS
accessing the collective knowledge of our expert
Dissolved gas analysis
chemists and engineers. Whether you need routine
testing or help developing an entire testing program Furanic compounds in oil Order Your copy of the
for condition assessment, you’ll get the answers Water-in-oil analysis ANSI/NETA Standards Today!
and support you need from a team dedicated to
quality and accuracy. Degree of polymerization
Oil quality screens
Find more information at Metals-in-oil
www.doble.com Polychlorinated biphenyls (PCBs)
Become a NETA Affiliate and Save!
+1 617.393.2997 In-depth, laboratory seminars conducted
labservices@doble.com by Doble’s experts To order:
visit www.netaworld.org
DOBLE IS AN ESCO TECHNOLOGIES COMPANY
Or call 888-300-NETA (6382)
Insulating Oils
Handbook
Table of Contents
Continuous Moisture-in-Oil Sensing: Field Applications for the Power Industry ..........1
Lance R. Lewand and Paul J. Griffin
Published by
Use of Gas Concentrations Ratios to Interpret LTC Dissolved Gas Data ......................58
Fredi Jakob, Ph.D., Karl Jakob, P.E., Simon Jones, and Rick Youngblood
Condition Assessment of Transformers — Analysis of Oil Data and Its Quality ............84
Lance R. Lewand
Metals Analysis in Transformers, Load Tap-Changers and Oil Circuit Breakers .............95
Lance R. Lewand
Copyright © 2009 by InterNational Electrical Testing Association, all rights reserved. No part of this publication may be
reproduced in any form or by any means, electronic or mechanical, without permission in writing from the publisher.
Insulating Oils Handbook 1
Presenters
Lance R. Lewand and Paul J. Griffin
Doble Engineering Company
This paper discusses the application of a continuous In addition, the success of oil or transformer drying
moisture-in-oil sensor that operates in dielectric fluids, such (dehydration) techniques is often measured by the amount
as transformer mineral oils. The sensor can be used to help of water remaining dissolved in the oil after processing.
in the assessment of the condition of transformers and for The historical approach to the detection of water in these
oil processing. circumstances has been by the Karl Fischer titration, com-
monly referred to in North America as ASTM Test Method
Introduction D 1533B. This chemical analysis has three major drawbacks:
it cannot be developed into a continuous measurement; it
The principal reason for developing a continuous mois-
can be plagued with a number of introduced errors most of
ture-in-oil sensor is that moisture continues to be a major
which result from sampling and sample containers; and cer-
cause of problems in transformers and a limitation to their
tain chemical interferences produce erroneous results [3].
operation. Excessive moisture is particularly problematic in
transformers as it affects the dielectric breakdown strength
of the insulation, the temperature at which water vapor
bubbles are formed and the aging rate of the insulating Principles of Moisture Detection
materials [1]. In the extreme case, transformers can fail The Doble moisture-in-oil sensor (DOMINO ) mea-
because of excessive water in the insulation. The dielectric sures the capacitance of a thin-film polymer. The capacitance
breakdown strength of the paper insulation decreases sub- changes proportionally to the change in RS of water in the
stantially when its water content rises above 2 to 3 % by oil. Relative saturation, expressed in units of percent, is the
weight. The dielectric breakdown voltage of the oil is af- concentration of water (Wc) in the oil relative to the solu-
fected by the relative saturation (RS) of water in oil which bility (S) or concentration of water the oil can hold at the
is constantly changing in a transformer environment. The measurement temperature, as shown in Equation 1.
maximum loading that is possible while retaining reliable
operation (i.e., preventing the formation of water vapor (Equation 1) RS = Wc /S (100%)
bubbles) is a function of the insulation water content. For
example, dry transformers (<0.5% water in paper) are much Where: Wc is in ppm wt./wt.
less susceptible to bubble evolution. In this case emergency S is in ppm wt./wt.
loading at hot-spot temperatures below 180 C may be pos-
sible with little risk of bubble formation. In contrast, a wetter The sensor converts the measured RS to a concentration,
transformer, with 2.0% moisture in the paper, runs the risk stated in parts per million in mg/kg (ppm wt./wt.). To be
of bubble formation with hot-spot temperatures as low as able to perform this calculation the instrument measures
139 C under the same conditions [1]. A more long-term both the temperature and the capacitance of the thin-film
problem is that excessive moisture accelerates the aging of polymer and converts it to RS. Equation 2 provides the
the paper insulation, with the aging rate being directly pro- conversion for mineral oil:
portional to the water content. That is as the water in paper
content doubles the aging rate of the paper also doubles [2]. (Equation 2) Log So = -1567/K + 7.0895
For these reasons it is important to have an assessment of the
moisture content of the insulation systems and to maintain Where: So is the solubility of water in mineral oil
transformers in a reasonably dry state. K is the temperature in Kelvin ( C + 273)
2 Insulating Oils Handbook
The structure of the moisture-in-oil sensor is shown 14 70
port base.
Temperature, Degrees C
8 40
Upper
electrode 6 30
4 20
0 0
101
105
109
113
117
121
125
129
133
137
13
17
21
25
29
33
37
41
45
49
53
57
61
65
69
73
77
81
85
89
93
97
1
5
9
Hours
Lower
electrodes
Figure 2 — Continuous Moisture Detection in a Transformer
80 180
factors can provide a false sense of reliability. Real time rat- Elapsed Time, Hours
44
80 80
40
70 70
60 60
Temp. 32
Temperature, Degree C
50 50 28
24
40 40
20
30 30
16
20 20 12
8
10 10
4
0 0
0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 0
E la p s e d T im e , H o u r s 0 10 20 30 40 50 60 70 80 90 100
Water Content, % RS @ 22 C
Figure 4 — Temperature Cycling in Drier Insulation System Model Figure 6 — Dielectric Strength Versus Relative Saturation (RS)
To properly maintain and operate transformers, an under- The solubility of water in oil is an exponential function
standing of the effects of moisture on the dielectric break- of the oil temperature [4] as shown in Table 1. It is clearly
down strength of the electrical insulating liquids is necessary. apparent that, as the temperature of the oil increases, the
The moisture content reduces the dielectric breakdown volt- amount of water that can be dissolved in the oil increases
age of insulating liquids. It is common practice to measure tremendously.
moisture content as a concentration (ppm). The correlation
between the water content in new, filtered, mineral oils at TABLE 1
room temperature and the dielectric breakdown voltage
using ASTM method D 1816 (0.04 inch gap) is given in Water in Oil Solubility As a Function of Temperature
Figure 4. Of course, the dielectric breakdown voltage will Oil Temperature Water Content Water Content
also be a function of the number and type of particles and in Oil, ppm in Silicone, ppm
their conductivity, not just the water content. 0°C 22 88
Recall that the relative saturation of water in oil is the
water content at a given temperature divided by the solubil- 10°C 36 125
ity of water in the oil at the same temperature. Taking the 20°C 55 174
same dielectric breakdown voltage data as given in Figure 30°C 83 237
5 and converting it to RS provides a much straighter curve 40°C 121 316
except at the extremes. This is shown in Figure 6. It is evident
50°C 173 414
that there is a better correlation between RS and dielectric
breakdown voltage than with moisture concentration and 60°C 242 534
dielectric breakdown voltage. 70°C 331 678
80°C 446 850
90°C 592 1052
48
44
40
100°C 772 1287
Dielectric Breakdown Voltage, kV
36
32
A simple example can be used to illustrate that the di-
28
electric breakdown voltage of insulating oils is proportional
24
to the relative saturation of water in oil rather than the con-
20
centration in ppm. First, a dielectric breakdown test set and
16
cup is placed in a temperature and humidity-controlled envi-
12
ronment. The humidity is controlled so the concentration of
8
water in the oil is held constant at 30 ppm. The temperature
4
starts at 100 C and the first dielectric breakdown voltage
0
measurement is made. At 100 C the solubility of water in
0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 oil is about 772 ppm. The relative saturation of water in oil is
Water Content, ppm. wt./wt.
therefore about 4% and the dielectric breakdown voltage of
a well-filtered oil would be quite high. Now the temperature
Figure 5 — Dielectric Strength Versus Water Content of the chamber is reduced to room temperature or about
4 Insulating Oils Handbook
22 C. The solubility of water in oil is about 60 ppm and the The Importance of an In-line
relative saturation is 50%. The dielectric breakdown voltage Moisture Measurement in Oil Processing
would be expected to be about one half of what it was when
the relative saturation was very low. If the temperature is Applications
cooled to 0 C and the dielectric breakdown voltage mea- DOMINOTM can be used for continuous monitoring
surement is made, the results should be quite low because of oil processing applications. It was determined that the
the solubility of water in oil at this temperature is about 22 following attributes would be required in a moisture-in-oil
ppm. As the water content in the oil is higher than this, sensor used for these applications:
the water forms an emulsion (the oil will appear cloudy2
with the suspended water) and begins to form condensation Provides a fast, continuous measurement
(Figure 7). During all this time the concentration of water
in oil has not changed (Table 2). Have the ability to withstand high flow rates
Can be mounted directly in the oil processing equip-
ment
Performance is not affected by pressure, vacuum, and
temperature
Water Content, 30 ppm Water Content, 30 ppm
Can be adapted to a variety of systems
Temperature, 50°C, Low RS Temperature, 0°C, High RS The application of a continuous moisture-in-oil sen-
Visual Appearance, Clear Visual Appearance, Cloudy
sor would greatly benefit the user in determining the end
point of any oil processing activity involving the removal of
dissolved water from the oil. However, additional benefits
could also be gained such as:
Determination of filter bleed-through
Figure 7 — Low RS Oil Versus High RS Oil Elimination of taking oil samples for water content by
chemical analysis
Determination of oil processing efficiency
TABLE 2 Determination of the amount of water removed by
Relationship Between Dielectric Strength filtering and/or dehydrating processes
and Water Content Determination of the effectiveness of field-drying a
transformer
Oil Water Content, Relative Dielectric
Temperature ppm Saturation Strength Determining the proper time to change filter cartridges
100°C 30 3.9% High due to filter bleed-through is a concern during transformer
22°C 30 50% About half
oil processing. Changing them is usually based on dif-
of high ferential pressure or a pressure increase, which can result
from excessive moisture, or particulate contamination.
0°C 30 >100% Very Low
This is often late in the process, when high amounts of
dissolved water is already in the processed oil. The purpose
Transformers are more complicated systems than given
of a continuous in-line moisture-in-oil sensor would be to
in this simple example. However, the same basic principles
accurately measure the amount of water in the processed
apply for the dielectric breakdown strength of the liquid
oil, clearly indicating when filters should be changed, and
dielectric. That is, it remains a function of the relative satu-
therefore speeding up the drying process. In most cases, a
ration of water in the oil. During the cool-down cycle of
moisture-in-oil sensor would indicate the need for a filter
a thermal transient in a transformer some of the moisture
change long before a pressure increase criterion would be
returns to the paper and some of the moisture remains in the
reached. Typically the water content in ppm remains fairly
oil. The relative saturation of water remaining in the oil will
constant at some low value while the filter is most effec-
influence its dielectric breakdown voltage. For this reason
tive, and then starts to increase gradually. A reasonable end
continuous measurement can be used to detect conditions
point can be chosen to change the filter before it becomes
in which moisture distribution is a problem.
ineffective. Monitoring of the differential pressure is still a
valuable practice as filters can be plugged from particulate
contamination and free water.
One of the most beneficial results from the use of an
in-line continuous moisture-in-oil sensor is the elimination
of the need to take periodic samples for Karl Fischer titra-
tion measurements (and the associated chemicals and their
Insulating Oils Handbook 5
disposal). The algorithm used in this particular moisture- appropriate actions can be taken to improve it. The moisture-
in-oil sensor converts relative saturation of water in oil and in-oil sensor can be used in a similar fashion to determine
temperature into the concentration of water in ppm, and the best operating conditions for the use of filters.
was determined initially by using the Karl Fischer titration. Calculating the total amount of water removed can be a
Figure 8 demonstrates how closely the two measurements useful measure of dehydrating effectiveness. It is quite simple
are related to each other. for oil in a storage tank. For instance, using Equation 3, it
can be determined that a filtering or dehydrating process
60
which removes 35 ppm of water (difference between inlet
and outlet measurements) on 10,000 gallons of oil would
50 remove 1.2 liters of water.
(Equation 3) Wt = 3.31X 10-6 Voil (Ci –Co)
40
Water Content, ppm
Where:
Wt = The total amount of water removed in liters
30
20
Voil = The volume of oil in gallons
Ci = The inlet concentration of water in oil in ppm wt./wt.
10
from the moisture-in-oil sensor
Co = The outlet concentration of water in oil in ppm
0 wt./wt. from the moisture-in-oil sensor
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Figure 8 — Moisture-In-Oil Sensor Versus Karl Fischer Titration Wt = 3.31X 10-6(10,000gallons)(40–5)=1.2 liters of water
The accuracy of this particular sensor in ppm varies with The effectiveness of field drying of the transformer,
the temperature of the measurement because it is a conver- whether the process is conducted out of service or in-ser-
sion from RS and not a direct ppm measurement. Table 3 vice, is much more difficult to assess than drying the oil
below provides some useful values. For very aged oils the alone. This is because the large mass of cellulosic insulation
relationship between RS and concentration (ppm) may need (paper, pressboard, wood structural members, etc.) contains
to be determined experimentally. almost all of the water in a transformer and the moisture
distribution is uneven and unknown. A reasonable estimate
can be produced before and after processing if the following
TABLE 3 variables are known:
Accuracy of Moisture-In-Oil Sensor
Volume of oil
TEMPERATURE % Accuracy, RS Accuracy, PPM Mass of paper
0°C ± 1% ± 0.25
Concentration of water in ppm at a known temperature
10°C ± 1% ± 0.40 (preferably one that is > 50 C)
20°C ± 1% ± 0.60
At higher constant temperatures it is possible to estimate
40°C ± 1% ± 1.25 the average moisture content of the cellulosic insulation
60°C ± 1% ± 2.50 [6,7]. For example, if a transformer has 10,000 gallons
80°C ± 1% ± 4.50 (37,850 liters) of oil, 9125 pounds (4148 kg) of paper, and
30 ppm of water in oil while maintained at 70 C (RS =
The moisture-in-oil sensor can be used to estimate the 9.0), then the calculated amount of water in the oil and
efficiency of an oil dehydrating process using vacuum or paper is as follows:
filters. In the case of vacuum dehydrating, the inlet and outlet
oil can be monitored to determine the best flow rate, dwell The oil contains about 1.0 liter of water
time, applied vacuum, temperature, and if other mechanical The paper contains about 78.5 liters of water
aspects of the process are functioning adequately. This can The entire system would have an estimated total volume
be done by comparing the difference in the water content of of approximately 80 liters
the oil at the inlet and outlet of the processor. The greatest The amount of water in oil was determined by Equation
difference in water contents will indicate optimum efficiency. 3, where:
Once the most efficient processing conditions have been set, Ci = the water-in-oil content
the moisture-in-oil sensor can then be used to continuously Co = 0
monitor the process. If the efficiency begins to decline, then
6 Insulating Oils Handbook
The amount of water in the paper is estimated by multi- Conclusions
plying the mass of paper times the concentration of water A new moisture-in-oil sensor, DOMINOTM, has been
in the paper in percent. developed to provide continuous in-situ measurements of
If the amount of paper is unknown it can be estimated dissolved water in oil. The measurement of the capacitance
from the following: of a thin-film polymer is used to detect the relative satura-
Shell-form transformers, the paper mass = oil mass/3 tion (RS) of water in oil. The relative saturation of water in
Core-type transformers, the paper mass =oil mass/8 oil along with the measured value for temperature are used
Core type transformers with enameled wire insulation, to calculate the concentration in ppm (wt./wt.).
the paper mass = oil mass/20
Current applications in which this moisture-in-oil sensor
The amount of water in the paper can be calculated by can be used are:
two means, using an equation or a family of curves. The
equation is as follows: Continuous moisture measurement in transformers
On-line/off-line oil processing applications
(Equation 4)
Cp = {2.17x10-5x[(RStox10-2)(Ps)]0.6685}xe4726 (Tto+273) Determination of filter bleed-through
An alternative to taking oil samples for water content
Where: by chemical analysis
Cp = moisture content in the paper in %
Determination of oil processing efficiency
RSto = relative saturation of water in oil at temperature,
Tto Determination of the amount of water removed by
Ps = The saturation vapor pressure at temperature, Tto filtering and/or dehydrating processes
(from published data, see Determination of the effectiveness of field-drying a
Reference 8). The values are given in mm of Hg and need transformer
to be converted to
atmospheres by dividing by 760 mm Hg/atmosphere.
e = natural log
Tto = top oil temperature References
[1] Griffin, P. J. “Water in Transformers – So What!”,
The amount of water was calculated as 1.9% of the mass
National Grid Condition Monitoring Conference,
of cellulosic materials, which gives a mass of water of 173
May 1996.
pounds (78.5 kg). This mass can then be easily converted
to liters of water by multiplying by 0.454 liters/pound for a [2] Lewand, L. R. and Griffin, P. J., “How to Reduce the
volume of 78.5 liters. The other way to calculate the amount Rate of Aging of Transformer Insulation”, NETA
of water in paper is to use equilibrium isotherm curves, World, Spring 1995, pp. 6-11.
taking the concentration of water in oil and the top oil [3] Lewand, L. R. and Griffin, P. J., “Transformer Case
temperature to find the equilibrium value [4]. Studies”, Proceedings of the Sixty-Six Annual Interna-
Once the drying process has been completed the same tional Conference of Doble Clients, 1999, Sec.5-7.
estimation process for moisture in the cellulosic materials
can be repeated. This should be done after the transformer [4] Griffin, P. J. “How to Prevent Rain in Power Trans-
has been at an elevated temperature for at least three days formers”, ASTM Standardization News, November
without processing. The difference in the amount of water 1991, ASTM, Philadelphia, PA, pp. 30-33.
before and after filtering is the quantity removed. [5] Griffin, P. J., Bruce, C. M., and Christie, J. D. “Com-
Another way to determine how much water is removed parison of Water Equilibrium in Silicone and Min-
by processing each day is to use Equation 3. The flow rate eral Oil Transformers”, Minutes of the Fifty-Fifth An-
can be used to determine the volume of oil processed, and nual International Conference of Doble Clients, 1988,
the inlet and outlet water contents (while fairly constant) Sec. 10-9.1.
provides the concentration of water removed. If the inlet
water-in-oil content changes significantly then the calcula- [6] “Estimating the Water Content of Cellulosic In-
tion should be restarted and the volumes of water removed sulation”, MKT-AB-12, Rev A, Doble Engineering
summed for each set of calculations. The efficiency of the Company DOMINOTM Application Bulletin, No-
process can readily be observed by examining the difference vember, 1999, 4 pp.
between the inlet and outlet water content [7]. [7] “Transformer Oil Field Processing Applications”,
In the example above, the paper contained 1.9% water MKT-AB-16, Rev A, Doble Engineering Company
or 78.5 liters at the start of the processing. Ideally the water DOMINOTM Application Bulletin, November, 1999,
content of the paper would be reduced to a final concentra- 6 pp.
tion of 0.5% or lower. This would require removal of at least
58 liters of water.
Insulating Oils Handbook 7
[8] “Vapor Pressure of Water Below 100 C” In CRC
Handbook of Chemistry and Physics, Ed. Robert C.
West, Boca Raton, FL. CRC Press Inc., 1978-1979,
P. D-232.
References
1. ASTM, Annual Book of ASTM Standards, Volume
10.03, D 3612-96.
2. Jalbert, J. and Gilbert, R. “Comparison Between Head-
space and Vacuum Gas Extraction Techniques for the
Gas Chromatographic Determination of Dissolved
Gases from Transformer Insulating Oils”, 1994 Inter-
national Symposium on Electrical Insulation, Pitts-
burgh, PA, 1994.
3. IEEE C57.104-1991, “IEEE Guide for the Interpreta-
tion of Gases Generated in Oil-Immersed Transform-
ers”, 1991.
4. IEEE P1258, “IEEE Trial-Use Guide for the Inter-
pretation of Gases Generated in Silicone-Immersed
Transformers”, 1995.
5. Youngblood, Rick, et. al, “Application of DGA to De-
tection of Hot Spots in Load Tap Changers”, Minutes
of the Sixtieth Annual International Conference of
Doble Clients, 1993, Sec. 6-4.
6. Charles Baker, Private Communication.
12 Insulating Oils Handbook
The choices confronting insulating oil users have never development of nonflammable, chlorinated fluids. Taken
been greater. Where conventional mineral oil has been the together, these nonflammable fluids are known as “askerals.”
staple diet of transformers for over a hundred years, today’s These fluids cooled and insulated well and allowed further
technology and market demands have produced a variety of use of transformers inside or near buildings because of their
new fluid options. This article discusses the new insulating nonflammable character.
fluids on the market and describes how each can be used But there were also problems with these fluids. Environ-
to solve specific problems. mental studies showed that askarel fluids were extremely
persistent in the environment. In addition, studies showed
It used to be easy.... that if the fluids were exposed to an electric arc or ex-
tremely high temperatures, highly toxic chemicals would
The first closed-core transformers were developed at the
be formed.
Abraham Ganz foundry in Budapest in 1884. These origi-
The use of askerals was phased out worldwide in the
nal units were air cooled. The drawbacks of using air as a
1970s and 1980s. Several fire-resistant fluids were developed
dielectric and cooling medium were soon apparent, and in
to take their place. The two main types of fire resistant fluids
1890 the first oil-cooled, oil-insulated transformer was build
used in the US are fire-resistant petroleum and silicone
by Brown, GmbH, in Germany. Since that time, napthenic
fluid. An example of fire resistant petroleum oil is shown
petroleum oil has been the standard cooling and insulating
in Table One.
medium in liquid-filled transformers. Oils with a napthenic
molecular structure were preferred over those with paraffinic
structure because they had better low temperature behavior Table One
and because napthenic oil was found in shallower wells, Typical Characteristics of Fire Resistant Petroleum Oils
which made it easier to extract from the ground. Characteristic & ASTM method Beta fluid Transformer
The specific petroleum fractions chosen for use in trans- Oil
formers were relatively light. This facilitated cooling flow Fire Point, ASTM D92, Deg.C 306 145
inside the transformer and the use of transformers in cold
Viscosity, D88, cSt. @ 100C 11.7 3.0
environments.
Napthenic petroleum continues to be the standard for Density @ 20ºC, g/cc 0.87 0.87
insulating oil used in the United States. In other parts of the Color, ASTM units L0.5 L0.5
world, scarcity of napthenic crude oil sources and improved Appearance Clear Clear
dewaxing processes have shifted the markets to the use of
Dielectric Breakdown, D877 55 55
paraffinic oils, but the characteristics of the final product
are roughly the same. Dissipation Factor, 40ºC, D924, % 0.1 0.1
Acid Value, D664, mg KOH/g 0.01 0.01
But then we wanted fire resistance…
In the early part of the 20th century, one of the disad-
vantages of using light petroleum fractions as insulating
fluids was their flammability. This was overcome with the
Insulating Oils Handbook 13
• Equipment operating in low ambient temperatures Color, ASTM units L0.5 L0.5
Appearance Clear Clear
• Equipment situated in environmentally sensitive areas.
Dielectric Breakdown, D877 59 55
In the past decade, new insulating fluids have been de- Dissipation Factor, 40ºC, D924, % 0.1 0.1
veloped to address each of these problems. These fluids are Acid Value, D664, mg KOH/g 0.01 0.01
grouped together under the heading of “functional fluids,”
as their characteristics influence the performance of the
equipment in which they are used. Functional fluids will Equipment at high temperatures
probably not replace the use of conventional transformer
mineral oil, at lest in the foreseeable future, but they are Specialty transformers are now being designed to operate
proving themselves to be an important tool to help equip- with much higher heat rise than would have been accept-
ment makers achieve specific performance. able a few years ago. Upgraded paper insulation systems can
withstand hot spot temperatures that would have severely
degraded old style Kraft insulation. Synthetic fluids are
Highly biodegradable fluids available that resist oxidation and aging better than con-
Because of rising environmental liability and spill cleanup ventional mineral oil, and they have enhanced heat transfer
costs, power utilities have requested transformer and fluid characteristics. Using these fluids in transformers that were
manufacturers to provide alternative dielectric fluids that built for mineral oil can yield temperature decreases of up
are more environmentally friendly than conventional to five degrees centigrade with no additional cooling sur-
transformer oil. face area. Because the synthetic oils are more stable than
In response to these requests, fluid manufacturers have mineral oil, service life can be extended. In addition, these
developed several types of highly biodegradable fluids. Some fluids have a significantly higher fire point than conventional
are made from synthetic hydrocarbons, some from vegetable transformer oil, which provides a higher safety margin at
oils, and some from synthetic chemicals called “esters.” these elevated operating temperatures. The characteristics
Most of these new fluids biodegrade far more completely of a fluid made for high temperatures are shown in Table
and rapidly than conventional transformer oil. For example, Three.
several of the “highly biodegradable” oils biodegrade more
than 95 percent when tested with standard methods. In Table Three
contrast, conventional transformer oil biodegrades to 30–50
Typical Characteristics of Alpha-2 Fluid
percent when tested with the same methods.
The different types of fluids available vary widely in Characteristic & ASTM method Alpha-2 Transformer
characteristics and price. This field is still developing, and fluid Oil
there have not been standards developed for fluid character- Fire Point, ASTM D92, Deg.C 250 145
istics, acceptable performance levels, or even test methods Viscosity, D88, cSt. @ 40ºC 3.9 3.0
to determine minimum levels of biodegradability. To date, Density @ 20ºC, g/cc 0.82 0.87
none of the fluids has found sufficient acceptance or use to
become a standard for highly biodegradable fluids. Color, ASTM units L0.5 L0.5
One example of a highly biodegradable oil is ECO Fluid, Appearance Clear Clear
which is made from synthetic hydrocarbon oils. Typical Dielectric Breakdown, D877 57 55
characteristics of ECO Fluid are shown in Table Two. Dissipation Factor, 40ºC, D924, % 0.1 0.1
Acid Value, D664, mg KOH/g 0.01 0.01
14 Insulating Oils Handbook
When you harness the power of knowledge, you unleash the possibilities. ANSI/NETA ATS-2009
This standard should always be referenced in design
specifications or when performing acceptance testing
The industry’s knowledge leader is your trusted partner. on power system installations. ANSI/NETA MTS-2011
For over 90 years, Doble has partnered with our utility clients to improve ANSI/NETA ETT-2010
operations and optimize system performance. Through the dedication This standard ensures that your acceptance
of world-class staff, Doble provides products and services in 110 and maintenance tests are being preformed by
countries around the globe, delivering the solutions you need – from qualified technicians who are certified in accordance
laboratory testing services to hands-on, technical training seminars. It’s a with ANSI/NETA ETT requirements.
relationship we value and a partnership that will allow us take the industry
to the next level. Together. Available in Bound, CD ROM, or PDF
STANDARD FOR
MAINTENANCE
TESTING SPECIFICATIONS
FOR ELECTRICAL POWER
When you work with Doble’s laboratories, you’re Doble Laboratory Services Include: EQUIPMENT AND SYSTEMS
accessing the collective knowledge of our expert
Dissolved gas analysis
chemists and engineers. Whether you need routine
testing or help developing an entire testing program Furanic compounds in oil Order Your copy of the
for condition assessment, you’ll get the answers Water-in-oil analysis ANSI/NETA Standards Today!
and support you need from a team dedicated to
quality and accuracy. Degree of polymerization
Oil quality screens
Find more information at Metals-in-oil
www.doble.com Polychlorinated biphenyls (PCBs)
Become a NETA Affiliate and Save!
+1 617.393.2997 In-depth, laboratory seminars conducted
labservices@doble.com by Doble’s experts To order:
visit www.netaworld.org
DOBLE IS AN ESCO TECHNOLOGIES COMPANY
Or call 888-300-NETA (6382)
Insulating Oils Handbook 15
by Lance R. Lewand
Doble Engineering Company
Insulating materials used in power transformers have been an oxygen atom in what is referred to as a glycosidic link-
selected because of their abundance, low cost, and longev- age. The long-chain cellulose molecules interact with each
ity under normal operating conditions. Oils in the U.S. are other due to hydrogen bonding resulting in strands, mats
expected to last 30 or more years before forming excessive and paper sheets.
amounts of acids and sludges and can then be rejuvenated Much of the mechanical strength of paper and pressboard
by treatments with absorbents such as clay. They can also be comes from the long-chain cellulose polymer. As the cel-
easily replaced. Modern oil preservation systems are designed lulose ages, the polymers are cleaved and become shorter,
to minimize exposure of the insulating oil to air thus retard- resulting in reduced mechanical strength. The primary forms
ing its oxidation. The solid insulation (paper and pressboard) of degradation of the cellulose polymer are hydrolytic, oxida-
is the main dielectric in transformers and also serves as tive, and thermal. In the case of each of these mechanisms
mechanical support. Localized severe degradation in those free glucose is generated and the ring structure tends to be
materials must be considered most serious as this can result opened to form chains. Although temperature is likely to
in loss of adequate dielectric strength. In addition, cellulosic be the most important factor, oxygen and water have been
materials cannot be easily replaced; therefore, their longev- clearly shown to have a significant effect on the degradation
ity, which is primarily a function of temperature, becomes of Kraft paper. The degradation of cellulose molecules results
a limiting factor in the operation of transformers. The end in the formation of gases, primarily carbon monoxide and
of life criteria, tensile strength, or degree of polymerization carbon dioxide, furanic compounds, and other byproducts.
(DP) are physical characteristics of the paper insulation. If The carbon oxide gases often provide early warning of
paper insulation is maintained in a dry state, its good elec- excessive damage. However, other materials such as paints
trical properties will be retained even as it becomes quite and gaskets can outgas carbon oxide gases when exposed
brittle. However, mechanically weakened paper can break to excessive temperatures and, therefore, are not always
especially as windings vibrate and move, particularly during attributable to the degradation of the cellulosic insulation.
through faults thus reducing insulating capability. Dielectric Confirmatory and complementary tests have been developed
breakdown is then more likely to occur. which detect oil soluble breakdown products of the cellulose
Fortunately, as cellulosic materials are degraded, byprod- chain (called furanic compounds) with the primary indica-
ucts such as carbon oxide gases (carbon monoxide and carbon tor being 2-furfural.
dioxide) and furanic compounds are formed which can serve
as indicators of the aging process. Cellulosic materials, most Furanic Compounds
often paper samples, can be tested directly for DP, a measure
Furanic compounds are five-membered ring structures
of its average molecular weight that correlates well with
that are formed in a manner in which the open-chain glu-
mechanical properties.
cose molecule goes through a series of dehydration reactions
Cellulose is a long straight chain polymer (polysaccha-
(elimination of water molecules) and then recycles into
ride) of glucose molecules (monomers), and is the major
a five-membered ring structure. The furanic compounds,
constituent of paper and pressboard. Glucose is a sugar
unlike sugars such as glucose, are oil soluble and, therefore,
that has six carbons and is typically in the more stable ring
are detectable.
structure called a pyranose. The glucose rings are linked by
16 Insulating Oils Handbook
High concentration of 2-furfural is a clear indication of • For normal aging the rate of 2-furfural generation
cellulose degradation as this is the only type of material in should be much less than 50 ug/L/year and usually in
transformers which yields this byproduct. Under some con- the vicinity of 10-20 ug/L/year
ditions where carbon oxides may be lost, such as when a leak
• If estimating insulation quality from the 2-furfural con-
occurs in the gas space of a nitrogen blanketed transformer
tent, use these guidelines:
or from the conservator tank for those that are free breath-
ing, the furanic compounds will continue to accumulate • Normal <100 ug/L
and provide a gross indication of the relative aging of the
• Midlife (examine rate) > 100 <1000 ug/L
cellulosic insulation or a thermal incipient-fault condition
involving cellulosic materials. Conversely, when cellulosic • Last third of life? > 1000 ug/L (flag for
materials are exposed to extreme temperatures which result further study)
in charring, furanic compounds can be destroyed and the
carbon oxides may be the only byproducts remaining in
significant quantities. Degree of Polymerization (DP)
Experience is required in evaluating the furanic com- The degree of polymerization test is used to assess in-
pound data since there are factors such as the type of insula- sulation aging and is performed on paper samples taken
tion preservation/oil expansion system, type of conductor- directly from the transformer so it is an intrusive test. The
wrapped insulation, and family of transformer, all of which DP provides an estimate of the average polymer size of the
influence the interpretation. For example, the treatment of cellulose molecules in materials such as paper and press-
the oil or the transformer can result in the removal of sig- board. The DP correlates well with mechanical properties
nificant amounts of furanic compounds. Not knowing this such as tensile strength but has the advantage that it can
information may lead to a misdiagnosis of the actual con- be performed on used materials that have taken a set dur-
dition of the transformer. In addition, furanic compounds ing service life. Generally, paper in new transformers has a
are generated from thermal events, not electrical discharge DP of about 1000. Aged paper with a DP of 150-200 has
activity and therefore can be useful in the assessment of little remaining mechanical strength, therefore making the
failure mode and incipient-fault conditions leading to the windings more susceptible to mechanical damage during
failure. Tests for furanic compounds should be performed physical movement, which can cause the paper to tear or
initially for all power transformers to establish a baseline, for crumble. This may occur when transformers are moved or
important or older transformers, when high carbon oxides during events such as through faults. Since paper insulation
are generated, for highly loaded transformers, and when does not age uniformly due to thermal, water, oxygen and
other tests indicate accelerated aging. byproduct concentration gradients, samples from several
In order to detect the degradation of cellulosic materials, distinct locations provide the best diagnosis. The DP test
sufficient quantities must be degraded to increase the con- provides the most reliable indication of the overall aging of
centration of indicator gases and furanic compounds in the the paper insulation as it is a direct measurement. This test
oil to thresholds considered to be problematic. Experience should be performed:
has shown that significant damage, including charring of
the cellulosic insulation, when limited to isolated hot spots • when there is other evidence of very accelerated aging
due to incipient-fault conditions, will produce carbon oxides of the insulation
and furanic compounds below thresholds used to indicate • when an internal investigation is being performed and
problems involving the cellulosic insulation. the transformer is more than 20 years old
The analysis of data for furanic compounds should be
based on the type of insulating paper used and the preserva- • for condition assessment of older transformers for pos-
tion system employed. For Kraft paper insulation, suitable sible refurbishment
guidelines are as follows: • for consideration of a partial rewind
• For normal aging <50 ug/L/year of 2-furfural should be • for failure assessment
generated.
• for condition assessment of insulation when purchasing
• Generation rates >50 ug/L/year of 2-furfural is consid- a service-aged transformer
ered accelerated aging
• to assess the condition of a transformer after an extreme
• Values of 2-furfural > 1000ug/L should raise a flag for overheating event such as loss of cooling
further study
Conclusions
The combination of analyses of furanic compounds in oil,
DP, along with routine dissolved gas-in-oil analysis is a very
powerful set of tools to assess the condition of the cellulosic
insulation. The more specific information known about a
transformer and its family, the better the diagnosis that can be
provided.
Understanding Water
in Transformer Systems
The Relationship Between Relative Saturation
and Parts per Million (ppm)
by Lance R. Lewand
Doble Engineering Company
Water content in transformer oil in parts per million celerates the aging of the paper insulation, with the aging
(ppm) is a familiar concept to most in our industry, and lim- rate being directly proportional to the water content. For
its of 30 to 35 ppm are generally referenced. However, these example, as the water content in the paper doubles so does
simple concentration limits have limited value in diagnosing the aging rate of the paper. The deterioration of the paper
the condition of transformer systems and, thus, the concept insulation results from the weakening of the hydrogen bonds
of relative saturation (RS) of water in transformer oil has of the molecular chains of the paper fibers. For these reasons
been re-introduced over the past 15 years. The concept of it is important to have a means of assessing the moisture
relative saturation of water in transformer oil is not a new content of transformer systems and to maintain transformers
one and was originally championed by Frank Doble as early in a reasonably dry state.
as the mid 1940s.Thus, this article discusses and details the In order to fully understand water and its dynamics in
relationship between RS and ppm. transformer systems, a short explanation of the different
It is well known that moisture continues to be a major types of water encountered and the concepts of solubility
cause of problems in transformers and a limitation to their and relative saturation are provided.
operation. Particularly problematic is excessive moisture
in transformer systems, as it effects both solid and liquid Types of Water in Oil
insulation with the water in each being interrelated. Water
Water can exist in several different states within the
affects the dielectric breakdown strength of the insulation,
transformer. There are three basic types of water found as-
the temperature at which water vapor bubbles are formed,
sociated with transformer oil:
and the aging rate of the insulating materials. In the extreme
case, transformers can fail because of excessive water in the • Dissolved water is hydrogen bonded to the hydrocarbon
insulation. The dielectric breakdown strength of the paper molecules of which oil is composed.
insulation decreases substantially when its water content • Emulsified water is supersaturated in solution but has
rises above two to three percent by weight. Similarly, the not yet totally separated from the oil. It usually gives oil
dielectric breakdown voltage of the oil is also affected by a milky appearance.
the relative saturation (RS) of water in oil. The maximum
loading that is possible while retaining reliable operation • Free water is also supersaturated in solution but in a
(i.e., preventing the formation of water vapor bubbles) is a high enough concentration to form water droplets and
function of the insulation water content. For example, dry separate from the oil.
transformers (<0.5 percent water in paper) are much less
susceptible to water bubble evolution. In this case, emer- In most cases, when one is analyzing or discussing the
gency loading at hot-spot temperatures below 180°C may amount of water in oil, dissolved water is being referred to
be possible with little risk of bubble formation. In contrast, as emulsified, and free water is visually apparent.
a wetter transformer, with 2.0 percent moisture in the pa-
per, runs the risk of water bubble formation with hot-spot
temperatures as low as 139°C under the same conditions.
A more long-term problem is that excessive moisture ac-
Insulating Oils Handbook 19
What is Water in Oil (ppm), Solubility Where: So is the solubility of water in mineral oil
of Water in Oil, and RS of Water in Oil? K is the temperature in Kelvin (°C + 273)
The detection of water in oil performed in the laboratory Relative Saturation (RS) is the actual amount of water
is most often performed by an analytical technique called measured in the oil in relation to the solubility level at that
Karl Fischer titration described in ASTM Test Method temperature. Relative saturation, expressed in units of per-
D 1533 or IEC Method 60814. Both methods are very cent, is the concentration of water (Wc) in the oil relative to
comparable and involve a coulometric titration technique the solubility (So) or concentration of water the oil can hold
involving the reduction of an iodine-containing reagent. The at the measurement temperature, as shown in Equation 2.
methods are used to determine the amount of water in an
oil sample on a weight-to-weight (mg/kg) basis or what is (Equation 2) RS = Wc /So (100%)
commonly known as ppm (parts per million).
The concepts of solubility and relative saturation can Where: Wc is in ppm wt./wt.
sometimes be difficult to understand, but it is an impor- So is in ppm wt./wt.
tant concept when trying to assess the dryness or wetness
of a transformer system. Solubility is defined as the total For example, a sample of oil was taken for determination
amount of water than can be dissolved in the oil at a spe- of the water content. The temperature of the oil at the time
cific temperature. The solubility of water is not constant of sampling was 62°C. The laboratory performed the analy-
in oil but changes due to temperature. As the temperature sis and determined the water content to be 11 ppm. From
increases, the amount of water that can be dissolved in oil Equation 1, it is calculated that the solubility level at 62°C
also increases. The increase is not linear but exponential in is 259 ppm. As discussed previously, relative saturation is
function. For example, at 10°C only 36 ppm of water can the actual measured value compared to the solubility value.
be dissolved in the oil, whereas when the temperature in- In this case it is 11 ppm divided by 259 ppm resulting in a
creases to 90°C, the amount of water that can be dissolved relative saturation of 4.25 percent.
in the oil increases tremendously to almost 600 ppm. The
table shown lists the calculated solubility limits for oil at
various temperatures. These levels are the greatest amount Effects of Relative Saturation on Dielectric
of water that can be dissolved at the temperatures listed. If Strength
the concentration of water in oil is greater than that shown To properly maintain and operate transformers, an
for that specific temperature then, in all likelihood, the oil understanding of the effects of moisture on the dielectric
is supersaturated with water, and free or emulsified water breakdown strength of the electrical insulating liquids is
could exist. necessary. Increasing moisture content reduces the dielectric
breakdown voltage of insulating liquids. The correlation be-
Table 1 — Water in Oil Solubility as a tween the water content in new, filtered, mineral oils at room
Function of Temperature temperature and the dielectric breakdown voltage using
ASTM method D 1816 (0.04 inch gap) is given in Figure
Oil Temperature Water Content in Oil, ppm 1 (water content, ppm). Of course, the dielectric breakdown
0°C 22 voltage is also a function of the number and type of particles
and their conductivity, not just the water content.
10°C 36
Taking the same dielectric breakdown voltage data
20°C 55 and converting it to RS (Figure 1, %RS graph) provides a
30°C 83 much straighter curve except at the extremes. It is evident
40°C 121 that there is a better correlation between RS and dielectric
breakdown voltage than with moisture concentration and
50°C 173
dielectric breakdown voltage.
60°C 242
70°C 331
80°C 446
90°C 592
100°C 772
Decreasing Relative
40
concentration was
36
constant at 30 ppm.
Saturation, %
32 The temperature was
28 changed to change the
Medium RS relative saturation.
24
20
16
Low RS
12
44
Tests for Oils Received in Drums and Small Water Content and D 1533 YES YES
Relative Saturation
Distribution Transformers
TEST ASTM Perform on Perform on
METHOD Drums Transformers
Color D 1500 Composite YES
Tests to Perform on In-Service Oils
Dielectric Breakdown D 877 YES YES
Since the oil and solid insulation deteriorates in a trans-
Voltage former while it is in service, periodic sampling and testing
Dielectric Breakdown D 1816 NO YES
is necessary to ensure that deterioration is detected before
Voltage it becomes excessive. The rate at which the insulating ma-
Interfacial Tension D 971 Composite YES
terials will degrade depends on several factors, such as the
type of oil preservation system (amount of oxygen present),
Neutralization Number D 974 Composite YES
operating temperature, water content of the insulation, and
Oxidation Inhibitor D 2668 Composite YES the amounts and types of contaminants. The frequency at
Content
which testing is conducted varies with the test, importance
Polychlorinated D 4059 Composite YES of the equipment, whether an incipient-fault condition
Biphenyls
is known to be present or if a problem exists, and when
Power Factor at 25°C D 924 Composite YES a family of transformers has been identified as having a
Power Factor at 100°C D 924 Composite YES history of problems. All large power and large distribution
Relative Density 60/60 D 1298 Composite YES transformers should be tested immediately after installation
Visual Examination D 1524 YES YES and periodically thereafter. Those units experiencing rapid
Water Content D 1533 YES YES deterioration of the oil or paper or those units for which
an incipient-fault condition has been identified should be
subjected to more frequent testing. Table 6 provides some
guidance as to which tests to choose.
Baseline Tests for Power
and Distribution Transformers
In addition to the testing specified in Tables 3 and 4, there
are certain tests that should be performed immediately after
a transformer is placed in service or energized (if there will
probably be a long duration before actually being placed
into service). This is recommended in order to provide a
baseline analysis to enable trending over the service life of
a transformer and to assure that the transformer does not
have any gross contamination present. These additional tests
are indicated in Table 5.
Insulating Oils Handbook 25
TABLE 6
Tests for In-Service Oils in Transformers
YES/NO indicates that it is a useful test to perform but it is not always necessary.
Safety First
Back by the Best
On record as an industry
From the field, to our
leader in safe work practices
full-service rotating
apparatus services
Cutting-Edge
Trained from utility
transmission to
low-voltage distribution
Our experience in commercial, industrial, generation and specialized sectors, such as oil
refineries, pulp and paper, steel and wind power, uniquely qualifies us to handle the
largest, most complex assignments. Our people are trained on the latest technologies and
safety practices. And, we support them with full-service rotating apparatus services and a
large fleet of service vehicles, ready to respond 24/7.
Corporate Headquarters | 2425 East Pioneer Drive, Irving, Texas 75061 Why trust your vital power services to anyone less? Count on an industry leader. Published by
p. 972.793.5523, 888.SHERMCO | f. 972.793.5542 | www.shermco.com Visit www.shermco.com and learn more today. InterNational Electrical Testing Association
26 Insulating Oils Handbook
by Lance R. Lewand
Doble Engineering Company
The last several articles discussed understanding water in • Prevent the loss or gain of gases or water when testing
transformer systems and which tests to choose and when. for these properties.
The results obtained from the analysis of a sample can
provide unparalleled information concerning the condi- The volume of the sample is of the utmost importance, as
tion of the insulating materials within electrical apparatus, various analytical tests require very different sample volumes.
life assessment, and the operating condition of the device. For example, a test for inhibitor content (ASTM D 2668)
However, a sample that is not representative of the bulk requires only a few milliliters of oil whereas the test for
oil insulation can provide erroneous information that can impulse breakdown (ASTM D 3300) may require as much
easily mislead personnel to incorrectly assess the condition as two to four liters. In general, it is good practice to provide
of the oil or the electrical apparatus. The art of sampling is the sample volume required for each test plus ten percent.
very important and will be covered in a future article. One If unsure of the sample volume required for specific test or
aspect of the sampling function that is often overlooked is tests, laboratories such as the Doble Materials Laboratory
the sample container itself. Unfortunately, there are many have reference lists that detail such information.
instances in which the sample container plays a major role
in the quality of the sample taken. Which Sample Container to Use
For general oil quality tests, glass bottles — either amber
Purpose of a Sample Container or clear — function well. Amber bottles provide protection
Sample containers are used to retrieve and store samples. against photodegradation while clear bottles enable visual
It is important for them to prevent contamination and to inspection of the sample. To prevent photodegradation
provide the best sample for a specific test. ASTM Practices in clear bottles, shielding from direct sunlight by storing
D 923 and D 3613, the Doble Reference Book on Insulat- them in cardboard or some other type of container works
ing Liquids and Gases, or IEC Methods and Guides 60475 well. Bottle caps must be constructed from a compatible
and 60567 are to be consulted when choosing a sample material that will not contaminate the sample. For example,
container. Sample containers should have the following bottle caps with liners composed of paper or having glue
characteristics. They must: that is soluble in the oil are not appropriate. Liners made
out of foil, Teflon® or polyethylene are usually safe to use.
• Be large enough to hold the volume of liquid necessary If sealed tightly, the glass bottle is an appropriate container
for analysis. from which to draw a sample for water analysis. However,
• Not impart any contamination (chemicals or particles) problems have been encountered with caps working loose
to the sample from the container itself. over time. If not properly protected, glass bottles are more
apt to break than other containers, and they are not suitable
• Seal the sample from external contamination. containers for dissolved gas-in-oil analysis, as gases such as
• Shield the sample from direct sunlight to prevent pho- hydrogen and carbon monoxide will be lost.
todegradation. This can be done either by having a dark Metal cans have become popular because they are more
container or by having a covering for that container. resistant to breakage than glass bottles. Cans with soldered
seams prepared with some fluxes will contaminate the
sample; therefore, welded seams are preferred. Cans made
Insulating Oils Handbook 27
of tin, aluminum, and stainless steel have been used and are Sample Storage and Transport
especially good containers to hold samples for oil quality Samples should not be stored longer then a few days
tests. Metal bulbs constructed out of stainless steel are also before shipping to the laboratory for analysis. The properties
used and can be invaluable when sampling in areas that of the dielectric liquid tend to decay over time, and gases
have a lot of airborne environmental contamination such for dissolved gas analysis (DGA) can be lost or gained if
as coastal areas where salt spray is common. the sample is held too long, as there is a finite time in which
The use of plastic bottles has grown significantly in the the sample container can maintain the sample integrity.
past decade. Like cans, they resist breakage and shield the Remember that the sample is supposed to represent the bulk
sample from sunlight when dark plastics are used. They are liquid insulation and storing it too long will change those
appropriate containers for samples in which oil quality tests properties. Store samples in the dark to prevent photodeg-
are to be performed. Samples being tested for water should radation, and keep them away from temperature extremes
not be stored in plastic bottles as water ingress or egress can and humid environments.
occur in just a few hours. Not all plastics are compatible with Package the sample containers to protect from leakage
oil so selection of the construction material is important, and breakage, especially if the samples are in glass. The
and containers made of substances such as polypropylene, United States Environmental Protection Agency Resource
polyvinyl chloride (PVC), and especially polystyrene are Conservation and Recovery Act (RCRA) listed in 40 CFR
not appropriate. Tests such as power factor will be affected Part 261.4(d) requires that all samples shipped must be
by incompatible plastic bottles because of dissolved com- packaged so they do not leak, spill, or vaporize from their
ponents transferred to the sample that will cause increases packaging. Other countries may have similar requirements
in dielectric loss. Bottles made of high density polyethylene as well. Use methods to prevent sample container breakage
have been found suitable and are one of the preferred sample or leakage such as the use of cardboard, Styrofoam, and other
containers when electrostatic charging tendency or furanic protective materials. Package the samples in such a way as to
compound analysis is to be performed. In the case of furanic avoid container-to-container contact. Also, some shipments
compounds, high-density polyethylene does not have the become too heavy and unwieldy, and the shipping company
silanol groups found in glass that attracts the semipolar may not be able to handle it safely. In this case, it is advisable
furanic compounds to the glass walls and thus removes to separate the shipment into several packages.
them from the sample. Samples with a low concentration
of furanic compounds are not as affected as samples with
larger amounts of furanic compounds. Lance Lewand received his BS degree at St. Mary’s College of Mary-
The most appropriate container for taking samples for land in 1980. He has been employed by the Doble Engineering Company
dissolved gas-in-oil and water content analyses, and the easi- for the past seven years and is currently Project Manager of Research
est to manipulate, is the ground glass syringe whose barrel in the materials laboratory and Product Manager for the DOMINOTM
product line. Prior to his present position at Doble, he was the Manager
and plunger have extremely tight tolerances. This type of of the Transformer Fluid Test Laboratory and PCB and Oil Services at
syringe has been found to satisfactorily prevent the ingress MET Electrical Testing in Baltimore, MD. Mr. Lewand is a member
of gases and water into the sample, and in turn prevent the of ASTM committee D 27.
egress of water and dissolved gases from the sample over
a period of time. Care should be taken with glass syringes
to ensure the stopcock is tight and in the closed position
once the sample is taken. Also, the samples collected in
syringes must be quickly protected from photodegradation
(degradation by light) by immediately placing them in the
dark or in their shielding containers. Stainless steel cylinders
may also be used but can be more difficult to manipulate or
to determine when all the air has been removed from the
cylinder. Metal cylinders will also add significant shipping
costs because of weight but are definitely more durable in
transit.
Sample identification is an extremely vital aspect of the
sampling process, and the identification is usually attached
to the container itself to avoid confusion. Many laboratories
receive samples that can not be related to a specific device.
Even if the sample was taken correctly, the lack of proper
identification makes the sample useless, and is therefore
a bad sample. When sampling personnel retrieve samples
from electrical apparatus, it should be done in a prescribed
sequence so nothing is forgotten.
28 Insulating Oils Handbook
OCB Diagnostics
PowerTest 2002
(NETA Annual Technical Conference)
Presenter
Fredi Jakob, Ph.D., Weidmann-ACTI
Co-Authors
Karl Jakob, P.E., and Simon Jones, Weidmann-ACTI
Rick Youngblood, Cinergy
Alex Salinas, Southern California Edison
Figure 1
Insulating Oils Handbook 29
indicates that the ratio of heating to arcing gases should be
indicative of problems such as increased contact resistance
or contact erosion.
Arc suppression grids deteriorate every time that they
quench an arc. When the grids are new the arc suppression
time is low. The particles generated from the grid degrada-
tion are numerous but small in size. As the grid opening
enlarges the arc suppression time increases and larger sized
degradation particles are produced. A fault is limited only by
the maximum current amplitude of the source impedance
and the interruption time of the breaker. Maximum fault
power (I2T) determines time to failure. Particle size and
count continues to grow until the distance from the point
source of heat and the maximum grid hole diameter cause
a cooling and blast zone buffer. At this point the produc-
tion of larger particles decrease and smaller particles again
“Good” Grid increase. A plot of large particle size production with time
seems to follow a bell shaped curve. The difference in the
leading edge to the trailing edge of the curve is noted by
the measurable increase in contact metals present in the
oil. Arc suppression grids are constructed of cellulose and
these particles can be distinguished from carbon and shiny
metal particles by chemical microscopy. At temperatures
above 300o C cellulose is destroyed and the resulting carbon
particles are observed in the oil.
Arc tip and arc shaft erosion can be measured by determi-
nation of the characteristic metals in oil and with chemical
microscopy, which can distinguish between shiny metal and
carbon particles. Oil quality assessment tests are also useful
in identifying OCB problems. Dielectric breakdown voltage
measurements, ASTM D-1816, are effected by moisture,
metals, carbon particles and cellulose particles in the oil.
Particles, especially carbon, also effect oil color. Since OCB’s
Close up of a Degraded Grid.
are free breathing devices the moisture level is higher than
that found in sealed components.
Figure 2 Table 1 is a summary of the tests that we have incor-
porated in our OCB diagnostic program and the type of
problem that can be determined with these technologies.
B. Laboratory Tests
Dissolved Gas Analysis, DGA, has been extensively
applied to locate incipient thermal or electrical faults in
transformers. Normal OCB operation will produce the
“key gases” associated with arcing under oil. These key
gases include acetylene and hydrogen, which are produced
at the very high temperatures associated with arcing. The
“hot metal gases”, methane, ethane and ethylene are pro-
duced whenever the oil is overheated from any cause. The
temperature required to produce acetylene is considerably
higher than that required to produce the hot metal gases.
Insulating fluids absorb and distribute fault energy. Thus the
temperature of the oil is very high in the vicinity of an arc
and decreases with distance away from the arc. This varia-
tion of temperature results in the production of acetylene
close to the arc and hot metal gases further removed from
the arc location. Very little heating occurs in a healthy
OCB so the amount of hot metal gases generated should
be small. This analysis of key gas production mechanisms
30 Insulating Oils Handbook
Table 1 Table 2
Diagnostic Tests th
90 Percentile Gas Concentrations in OCB Oil Samples
OCB Problem Test(s) Result(s) Fault Gas 90th Percentile Concentration
Increased Contact DGA Increased Hot Hydrogen 62
Resistance Metal Gases, In-
creased Heating to Carbon Monoxide 136
Arcing Gas Ratios
Methane 28
Contact Tip Erosion DGA, Metals, Hot Metal Gases,
Chemical Microsco- Metals Observed Ethane 14
py, Dielectric-1816 in Oil, Metal
Particles Observed Ethylene 71
by Chemical Mi- Acetylene 173
croscopy, Lowered
Dielectric Break- Total Dissolved Combustible Gas 530
down Voltage
Arc Suppression DGA, Particle Increased Fault Diagnostic software has been developed to evaluate OCB
Grid Degradation Count, Chemical Gas Levels, Large samples according to our established norm values. OCB
Microscopy, Dielec- particles in the Oil,
tric-1816, Color Cellulose Fibers,
sample results are placed into three broad categories of
Decreased Dielec- Normal, Caution and Warning. Southern California Edison
tric Breakdown designates OCB condition numerically using a 3, 2, and 1
Voltage scale. Normal indicates that the sample should be retested
according to the initial utility criteria. Caution indicates
that the sample should be tested more frequently, again at
IV. Diagnostic Protocols a rate determined by utility protocols. Warning indicates
that an internal inspection is appropriate. Figures 3, 4 and
Once the appropriate analytical procedures were selected
5 are examples of test reports for samples that fit in each of
to ascertain the effect of OCB problems it became neces-
these evaluation categories.
sary to determine normal values for each of the measured
parameters. We used a statistical approach to evaluate test
data from problem free units in order to establish the norms.
We then determined the ninetieth percentile values for each
parameter, for several thousand OCB samples. For example,
the 90th percentile individual fault gas concentration and the
total dissolved combustible gas concentration for all samples
evaluated are given in Table 2. Data for selected heating to
arcing fault gas ratios was calculated in a similar manner.
Similar calculations combined with field observations were
used to establish norms for particle size distributions and
metals in oil. Relevant IEEE guides are used to evaluate
results of oil condition assessment tests. These values are
all generic in nature and no attempt has made at this time
to develop unit specific flag points.
Insulating Oils Handbook 31
Figure 3
OCB Condition: Normal (3)
Narrative: No problems found. No action taken. Unit stayed in service, reset maintenance schedule.
Figure 4
OCB Condition: Caution (2)
Narrative: Analysis indicated mild contact erosion, and poor oil quality.
Remedial Action: Unit was put on ½ maintenance schedule and remained in service.
32 Insulating Oils Handbook
Figure 5
OCB Condition: Warning (1)
Narrative: Oil sample obtained approximately 4 hours after a 30000 AMP, 3 cycle fault.
Contacts evaluated at 25 percent degraded. Copper present due to shaft wear on movable contact.
Grid not significantly degraded.
Remedial Action:
Rotated movable contact ¼ turn.
Filtered and processed oil.
General maintenance. Returned to service.
V. Sampling Protocols
The overall objective of our OCB diagnostic program is Maintenance cost savings for the SCE program can be es-
to minimize the number of required internal inspections and timated considering the number of reduced internal inspec-
thus to realize considerable maintenance cost savings. Utili- tions. SCE currently inspects about three hundred breakers
ties must decide when and at what frequency to sample their a year, both oil filled and others. The oil filled breakers range
OCB’s. A sampling strategy has been developed by Mr. Alex in voltage class from 7.2 kV to 220 kV. Their largest number
Salinas at Southern California Edison, SCE, to determine of breakers operates at 69 kV. Inspection costs range from
when to draw the initial sample, when to draw subsequent 2K$ - 20K$ on OCB’s in the voltage range from 7.2 kV to
samples and when to conduct an internal inspection. The 220 kV. Based on statistical evidence less than ten percent
entire sampling process is incorporated in their system soft- of the units tested will require an internal inspection. The
ware. The software, which maintains OCB operations data, savings promise to be considerable.
generates an initial work order when any of three conditions Cinergy Corporation currently uses both condition based
are met: twenty interruptions, three hundred operations or (CB) and fault adjusted operation count (FAO) as the
five years from the last test. This first work order will be for initial triggers for breaker maintenance. The non-invasive
an initial oil sample. The results of the initial test generate condition based triggers used are thermography, ultrason-
a code: 1=Schedule internal inspection, 2=Reset oil sample ics, oil quality and dissolved gas analysis. The main invasive
trigger at half the initial values (10 interruptions or 150 tests are ductor and power factor. The invasive tests are
operations) and 3=Reset sample trigger at 20 interruptions used minimally due to limited outage request acceptance.
or 300 operations. The software maintains all of the test data, Fault adjusted operation count is a method that employs
resets the triggers and prints the work orders. This closed knowledge of the maximum fault amplitude based on source
loop operation requires no operator intervention. A flow and circuit impedance and fault duration. This is used to
chart for the SCE protocol is given in Table 3.
Insulating Oils Handbook 33
Table 3
SCE Flow Chart
determine the estimated number of breaker operations (300) criteria are not met. Can this time interval be extended
before maintenance is needed. Unless the breaker employs without increasing system outages due to breaker failures?
the use of some form of I2T monitors it is still very dif-
ficult to determine true fault duty cycle. Additionally, due
to deregulation and customer commitment, outages and Dr. Fredi Jakob received his PhD at Rutgers, the State University of
internal inspections will be kept to a minimum. Although New Jersey, in 1961. He is professor emeritus of analytical chemistry at
California State University-Sacramento and is the founder and laboratory
each test gives different positive information to determine director of Analytical ChemTech International, Inc. (ACTI), which is a
maintenance interval, Cinergy has found the addition of wholly owned subsidiary of Weidmann Systems International. As a long-
oil particle analysis to determine grid and contact health term member of ASTM and IEEE and author of over fifty published
invaluable to determine overall maintenance interval. articles, Dr. Jakob is a traveling lecturer to private and governmental
agencies. He has been invited to speak at American Public Power meet-
ings, ASTM symposia, conferences held by Doble, NETA, and AVO
VI. Conclusions and Future Work conferences, as well as other industrial organizations.
Normal or threshold values for OCB test parameters are
at this time generic in nature. Because of design variations
unit specific normal values may be more appropriate. Unit
specific test values will have to be empirically determined.
We are maintaining a very detailed database so that we can
develop unit specific normal values in the future.
We are cooperating with utility clients to further evalu-
ate testing frequency protocols. For example SCE tests the
OCB oil every five years if the interruption (20) or operation
34 Insulating Oils Handbook
by Lance R. Lewand
Doble Engineering Company
The results obtained from the analysis of an insulating What is a Good Sample?
oil sample can provide unparalleled information concerning Simply put, a good sample is one that is representative
the condition of the insulating materials within electrical of the content of the bulk oil insulation. Since samples are
apparatus, life assessment, and the operating condition of usually retrieved from a drain valve or the attached sam-
the device. However, a sample that is not representative of pling cock, preparation of that area is important to obtain
the bulk oil insulation can provide erroneous information a good sample. Cleaning the drain valve and the sampling
which can easily mislead maintenance personnel to incor- cock inside and out is the first step in avoiding sample con-
rectly assess the condition of the oil or the electrical ap- tamination. Cleaning the outside of the drain valve is just as
paratus. In one case, samples taken from two transformers important as cleaning the inside. The dirt and debris falling
showed very high concentrations of hydrogen and no other off the outside of the valve into the sample container during
gases, prompting maintenance personnel to give these units the sampling process can contaminate many samples.
priority for diagnostic surveys. It was later found that the Most of the contamination in the apparatus consists of
cause of the high hydrogen was a galvanic reaction occur- water and particles (paper fibers, metal particles, etc.), which
ring in the drain valves in which water was converted to over time will settle out on the bottom of the apparatus near
hydrogen because of the interaction of a galvanized fitting the drain valve. This material needs to be flushed out of the
with a dissimilar metal. system to get to the bulk oil insulation. It is necessary to
remove at least one to two liters of oil from the drain valve,
Why Sample? cap the drain valve, and then flush out the sampling cock
For in-service oil-filled electric apparatus, sampling of before proceeding with sampling. On occasion, two liters
the oil provides a method to determine the condition of will not be sufficient, especially when sampling a nonen-
the solid and oil insulation as well as the operating condi- ergized transformer or certain OCBs and LTCs. Specific
tion of the apparatus without opening or de-energizing sampling techniques and precautions, especially those deal-
the apparatus. This is especially important in the present ing with low volume electric equipment, are detailed in the
utility and industrial climates, as equipment outages for Doble Reference Book on Insulating Oils and Gases and ASTM
out-of-service testing have become very limited. Sampling Practices D 923 and D 3613.
provides a means to check the condition of oil in storage,
whether it be new or used, and to determine if it complies Good Samples Versus Bad Samples
with specifications such as TOPS, ASTM D 3487, IEC It is sometimes very clear to the laboratory performing
60296, IEEE C57.106, or company specifications. Sampling the analysis on the oil that the sample was taken improperly.
can also help to determine: For example, the presence of free water or foreign objects
1) If accidental mixing of different dielectric oils has taken such as insects, pipe sealing tape or putty are strong indica-
place; tors that the drain valve was not adequately flushed out prior
2) If the method of transportation contaminated the oil; to sampling. Once analysis has begun and it is determined
that there is a high or free water content coupled with a low
3) If the handling equipment to transfer the oil contami- dielectric strength, with all the other test results being ac-
nated the product. ceptable, then it strongly indicates that the proper sampling
Insulating Oils Handbook 35
technique was not adhered to. It may even imply that some These materials increase dielectric loss, which increases the
chemical reactions were taking place in the drain valve that power factor. Incompatible inorganic and organic materials
were not representative of the bulk oil insulation. from the drain-valve stem packing or drain-plug sealants
can also have the same effect on the power factor.
Lab Tests Most Easily Affected
As indicated previously, the analytical tests most eas- Costs Associated with a Bad Sample
ily affected by sampling are dielectric strength and water In the case of a single sample, the costs for routine oil
content. This is due to the fact that drain valves are usually quality analysis and DGA testing are just a very small frac-
at very low points in the tanks where debris and water ac- tion of the total costs associated with taking and analyz-
cumulation occurs. Water can also be present as a result ing a sample. Some of the items and costs associated with
of condensation that occurs in the drain valve, which is sampling and analysis are:
due to the position of the drain valve on the tank. In most
cases the drain valve protrudes 15 to 30 centimeters (six ITEM COST ($)
to 12 inches) away from the main tank. From experiments Labor to take sample 275
performed at Doble Engineering, it was found that the oil
Materials to take sample 15
in many of these valves varies in temperature from eight
Packaging and shipping cost 8
degrees Celsius to 15 degrees Celsius cooler than the bulk oil
insulation. When oil or air has an elevated relative saturation Analysis cost 70
or humidity and there is a significant cooling, condensation Engineering evaluation of the data (10-15 min) 35
of water will occur. This is exactly what happens in a drain TOTAL 403
valve. Other analytical tests easily affected by sampling are
dissolved metals, particulate metals, particle counts, dis-
solved gases-in-oil, and power factor. Every situation is different, but in many cases the analysis
The concentration of metals, whether dissolved or in a cost is only about 17 percent of the entire sampling and data
particulate state, are especially impacted by the amount of review process. In a situation where the sample has been
cleaning performed on the drain valve and the amount of determined to be nonrepresentative of the bulk oil insula-
flushing that is performed. Debris that settles to the bottom tion, the following additional costs may also be incurred:
of the apparatus and subsequently into the drain valve can
consist of metal particles. In addition, just the simple fact of ITEM COST($)
removing the drain-valve plug or opening the sampling cock
Labor to take original sample 275
will create particulate metals. This is due to the grinding of
Materials to take original sample 15
the surfaces between the valve body and the drain plug or
sampling cock. In fact, it is becoming more apparent that Packaging and shipping cost for original sample 8
that these types of samples should only be retrieved after a Analysis of original sample 70
minimum of two (and sometimes three to four) liters of oil Engineering evaluation of the data (10-15 min)
have been passed through the drain valve. of original sample 35
The same is true of retrieving a sample for particle count Additional engineering time to confirm
sample was nonrepresentative 35
where valve debris, whether inside or outside, can severely
skew the results. The debris, soot, and grime that exist on Labor to take 2nd sample 275
the outside of the drain valve are of serious consequence, Materials to take 2nd sample 15
especially in industrial locations. This debris can be easily Packaging and shipping cost for 2nd sample 8
transferred to the sample bottle while the sampling process Analysis cost of 2nd sample 70
is taking place. This validates the importance of cleaning the Engineering evaluation of the data
outside of the valve prior to taking the actual sample. (10-15 min) of 2nd sample 50
Dissolved gas-in-oil analysis is another test impacted by TOTAL 856
sampling, drain valve components, and sampling materials.
When galvanic fittings (zinc coated) are used in the drain
valve assembly — such as the drain plug — a galvanic reac- The cost of taking a bad or nonrepresentative sample has
tion with water can cause very high levels of hydrogen to be more than doubled from the original total. This is in part
produced. If this residue is not flushed out adequately then it due to the fact that review of data from the second sam-
will be transferred to the sample and included in the analysis, pling takes longer as there is a more critical and thorough
causing a level of concern that is not warranted. Galvanic review.
plumbing fittings such as nipples can also have the same If the original sample was not recognized as bad, the costs
effect. Brass, bronze, stainless steel or black iron should be associated with that sample can be staggering. For example,
the only materials used. In addition, drain valve assemblies a bad sample could cause a customer to try a remedial effort
should not be composed of dissimilar metals as corrosion can in an attempt to improve the condition of the insulating oil
result, which may end up in the sample. Debris, water and — such as processing the oil through clay or vacuum-pro-
other ionic contaminants also affect the power factor test.
36 Insulating Oils Handbook
cessing a transformer to remove moisture — then associ- “ASTM D 3613: Standard Practice for Sampling Electri-
ated costs may skyrocket to between $10,000 and $30,000. cal Insulating Oils for Gas Analysis and Determination
This is one reason why Doble always recommends taking a of Water Content” in Electrical Insulating Oils and Gases;
second sample to confirm the results of the first before any Electrical Protective Equipment, Annual Book of ASTM
remedial activities begin. Other factors, such as accidental Standards, Vol. 10.03, ASTM, West Conshohocken,
sample switching or misidentification, can also be the source PA, 2001.
of an erroneous assessment.
Part two of this series will cover the sampling practices Griffin, P. J. “Water in Transformers – So What!” National
to follow and the science of sampling. Grid Condition Monitoring Conference, May 1996.
The first part of this three-part series defined how and 6. Taking the sample (cleaning and preparation of valves).
why to take a good sample and explored the costs associ- 7. Cleanup after sample has been retrieved.
ated with taking a bad sample. Part two covers sampling
technique and factors influencing the sampling process.
There are a number of industry-recognized standards that Materials Used
define the correct way to retrieve samples from electrical Whoever takes the samples must be fully prepared for any
apparatus or storage containers. Some of the sources that eventuality that could occur at the sample site. For example,
list these sampling techniques include: items such as sheet plastic, plastic bags, absorbent materi-
als, flush oil containers, and catch pans are all important
• Doble: Reference Book on Insulating Liquids and Gases materials to have to prevent or clean up liquid spillage. It
must be remembered that, in order to take a correct sample,
• ASTM D 923: Standard Practice for Sampling Electrical
some liquid waste will be generated. Of course, bottles and
Insulating Liquids
syringes will be needed as sample containers and must be of
• ASTM D 3613: Standard Practice for Sampling Electri- sufficient size to hold the volume of dielectric liquid neces-
cal Insulating Oils for Gas Analysis and Determination of sary for the desired tests. Labels are required to sufficiently
Water Content and correctly identify those containers. Make-up oil may
• IEC 60475: Method of Sampling Liquid Dielectrics be necessary to add oil to low volume devices. Bottled ni-
trogen may be needed to pressurize a transformer to relieve
• IEC 60567: Guide for the Sampling of Gases and of Oil a negative pressure in order to get the dielectric liquid out
from Oil-filled Electrical Equipment and for the Analysis of the apparatus.
of Free and Dissolved Gases Tygon™ tubing or other compatible tubing is also nec-
essary to direct the dielectric liquid from the drain valve
These techniques have been developed over a number to the flush container, sample bottle, and syringe. Tubing
of years and have incorporated the expertise of many indi- should only be used once and then discarded as the walls of
viduals. However, sampling technique involves much more the tubing have memory (gases, water, and other chemical
than just taking the sample. It involves a more thorough compounds are held in the walls of the tubing) which can
knowledge of the information to be gained from taking a be transferred to the next sample. Incompatible tubing such
proper sample and includes sample site preparedness and as natural rubber or polyvinyl chloride will contaminate a
site cleanup after sample retrieval. Some of the items that sample.
are a part of sample technique are: The appropriate tools and plumbing accessories must be
on-site to manipulate the drain valve in order to retrieve the
1. Materials used to aid in retrieval of a sample. sample. Personal protective equipment such as nitrile gloves
2. Safety precautions to adhere to. is used to protect personnel from the liquid dielectric and/
or polychlorinated biphenyls (PCBs). Personal protective
3. Environmental concerns. equipment and safety practices to protect against electrical
4. Identification of the sample and apparatus information. or physical hazards must also be present and observed.
5. Final checks prior to sampling.
38 Insulating Oils Handbook
• Make sure there is positive pressure on the electrical ap- Environmental Concerns
paratus. Dielectric liquid spillage as a result of sampling is a main
• Take into consideration the remaining volume in low environmental concern, as some of these liquids may still
oil volume apparatus (this may require de-energizing contain PCBs. The United States and many other countries
the equipment). have very strict guidelines for spill cleanup and notification
of PCB materials. In the United States even one drop of
• De-energize instrument transformers before sampling. liquid containing more than 50 parts per million of PCB
• Secure electrical dangers. is considered “improper disposal.” Even if the oil does not
contain any PCBs or the spill is small and accidental there
• Make sure Occupational Safety and Health Adminis- may be regulations dealing with the dielectric liquids’ release.
tration (OSHA) requirements are met. It is, therefore, easier to prevent against spillage then to clean
up after the spill has occurred. This is why many sampling
Making sure there is positive pressure on the electrical personnel lay down plastic and absorbent materials under
apparatus prior to sampling is the single most critical fac- the drain valve prior to sampling and then use a catch pan
tor in assuring that the equipment survives the sampling to trap larger volumes of liquid.
procedure. Sampling of electrical equipment while under
negative pressure will allow atmospheric air to be drawn
into the equipment through the drain valve, which will rise Identification of the Sample and Appara-
through the transformer as bubbles. These bubbles are areas tus Information
of weak dielectric strength and can easily cause failure of Sample identification is an extremely vital aspect of the
the apparatus through flashover. Most transformers have sampling process. Many laboratories receive samples that
pressure gauges that allow determination of the actual pres- cannot be related to a specific device. Even if the sample
sure. If positive or negative pressure cannot be determined, was taken correctly, the lack of appropriate identification
then follow the procedure in ASTM D 923 to determine makes the sample useless. When sampling personnel retrieve
the pressure condition. This involves using a slug of oil in samples from electrical apparatus, it should be done in a
clear tubing attached to the sampling cock. If negative pres- prescribed sequence so nothing is forgotten.
sure does exist then no samples are to be drawn until that The lack of information concerning an apparatus severely
negative pressure is relieved. Sometimes this is as simple as limits the laboratory in its ability to provide an in-depth
adding dry nitrogen to the headspace of a transformer to diagnosis. Apparatus information such as the age, type of
pressurize the unit, or waiting until ambient temperature preservation system, and any previous incipient fault condi-
has increased to a sufficient degree to cause the expanding tions or oil reclamation activities can alter a diagnosis. For
dielectric liquid to pressurize the apparatus. instance, the Doble Materials Laboratory relies heavily on
The remaining volume in some electrical apparatus — the type of preservation system that is part of the trans-
especially oil circuit breakers, bushings, load tap-changers, former to provide a diagnosis based on oil quality, dissolved
and small instrument transformers — is of serious concern. gas analysis, and furanic compound results. In transformers
Electrical components are positioned at critical clearances that have a sealed conservator preservation system, oxygen
and that distance is determined with liquid insulation pres- and nitrogen values are expected to be below certain levels. If
ent. If the liquid insulation level is too low the insulation test values are above those levels, there may be several causes
between the energized components is now gas instead of for this such as a breach in the bladder or diaphragm, another
the insulant, and the dielectric integrity becomes compro- leak elsewhere on the transformer, or poor sampling. All
mised. This is why it is important to check the liquid level these items would concern the operator of the equipment.
not only before sampling but after as well, and this will aid However, if no information is provided or the information
in maintaining a safe operating environment. Instrument is incorrect then no diagnosis or an incorrect one will be
transformers must also be de-energized prior to sampling provided by the laboratory. Laboratories also use the age and
in order to secure the electrical hazards. the type of preservation system of a transformer as exceed-
Electrical hazards are especially prevalent in small ingly pertinent information when providing a diagnosis
distribution pole and pad-mounted transformers. Pole based on furanic compound results.
transformers do not usually have a sample valve so the lid
Insulating Oils Handbook 39
Final Checks The next and final part of this series will deal with taking
Adhering strictly to safety and environmental concerns the actual sample, sample storage, and transport.
will assist personnel in adequately preparing the site around
the electrical apparatus to be sampled. However, specific
attention should be given to some final checks before the References
sampling commences. These include: 1. Reference Book on Insulating Liquids and Gases, edited
by the Doble Client Committee on Liquid Insulation,
• Confirmation of positive tank pressure — again. Doble Engineering Company, Watertown, MA, 1993.
• Using the correct sample containers, both in size and
compatibility. 2. “ASTM D 923: Standard Practice for Sampling Elec-
trical Insulating Liquids” in Electrical Insulating Liquids
• Labeling the sample information completed. and Gases; Electrical Protective Equipment, Annual Book
• Protecting the sample from outside contamination. of ASTM Standards, Vol. 10.03, ASTM, West Con-
shohocken, PA, 2001.
• Relative humidity of ambient air less than 50 percent
(avoid rain or snow conditions). 3. “ASTM D 3613: Standard Practice for Sampling Elec-
• Temperature of dielectric liquid higher than or equal to trical Insulating Oils for Gas Analysis and Determina-
the ambient air. tion of Water Content” in Electrical Insulating Liquids
and Gases; Electrical Protective Equipment, Annual Book
Checking for positive pressure is so critical that it requires of ASTM Standards, Vol. 10.03, ASTM, West Con-
a second check and must not be overlooked. As mentioned shohocken, PA, 2001.
in Part one of this series, it is very difficult to perform the
requested analysis when not enough sample volume exists. 4. IEC 60475: Method of Sampling Liquid Dielectrics,
Therefore, it is very important to make sure the correct size International Electrotechnical Commission, 3 rue de
sample container is used. If unsure of the sample volume Varembe, Geneva, Switzerland, 1974.
requirements, it is better to provide more sample than to
return to that apparatus at a later date to re-sample. Labeling 5. IEC 60567: Guide for the Sampling of Gases and of Oil
samples with complete sample information prior to actually from Oil-filled Electrical Equipment and for the Analysis of
taking of samples is recommended. In this way, sample con- Free and Dissolved Gases, International Electrotechnical
tainers will not go unmarked and confusion can be avoided Commission, 3 rue de Varembe, Geneva, Switzerland,
when multiple pieces of equipment are being sampled. 1992.
Some geographic locations are inherently inhospitable
to providing an environment where a good sample can be
taken. Inhospitable locations include salt spray areas and Lance Lewand received his BS degree at St. Mary’s College of Mary-
high wind gusting areas that naturally kick up sand, dirt, land in 1980. He has been employed by the Doble Engineering Company
for the past seven years and is currently Project Manager of Research
soot, and other debris that can easily deposit foreign matter in the materials laboratory and Product Manager for the DOMINOTM
into an open sample container while it is being filled. In product line. Prior to his present position at Doble, he was the Manager
these instances protecting the sample from outside con- of the Transformer Fluid Test Laboratory and PCB and Oil Services at
tamination may require special precautions. Really adverse MET Electrical Testing in Baltimore, MD. Mr. Lewand is a member
environments may require totally enclosed sampling systems of ASTM committee D 27.
such as stainless steel cylinders to protect against outside
contamination.
The ambient environment in which a sample is taken
can also contaminate the sample with excessive amounts
of moisture. Taking samples during rain, hail, or snow
conditions should be avoided. If the situation is completely
unavoidable then necessary precautions must be enacted to
keep the sample from absorbing any external moisture. It is
recommended that sampling be performed only when the
relative humidity conditions are less than 50 percent so as
to minimize the amount of external moisture the sample
will absorb. It is recognized that this may be unavoidable
in certain geographical locations and during certain peri-
ods of the year, but every effort must be made to minimize
external contaminants and have a sample that reflects the
bulk liquid insulation.
40 Insulating Oils Handbook
by Lance R. Lewand
Doble Engineering Company
The first part of this three-part series defined how and Table 1
why to take a good sample and explored the costs associ- Sampling Points for Various
ated with taking a bad sample. Part two covered sampling Dielectric Liquids in Apparatus
technique and factors influencing the sampling process. Part
3 will discuss the technique of taking an actual sample and
Sample from Bottom Sample from Top
methods for correctly storing and transporting a sample
for analysis. Mineral oil Polychlorinated Biphenyls
Different dielectric liquids require sampling from differ- Silicone Trichlorobenzene
ent locations based on their relative density (specific grav- R-Temp Tetrachlorobenzene
ity). In general, dielectric liquids with a relative density less Midel 7131 Wecosol
than one should be sampled from the bottom drain valve Reoloec 138 Perchloroethylene
whereas dielectric liquids with a relative density greater than
one should be sampled from the top fill valve as long as it Beta Fluid
is below the liquid level. There are exceptions to this, and Shell Diala HFX
the sampling point can change throughout the life of the WEMCO-FR
transformer. For instance, mineral oil transformers that have
MEPSOL
no drain valve are usually accessed and sampled through the
top. Another example involves retrofilling of askarel trans- Opticool
formers. When a transformer is filled with askarel, sampling ALPHA-1 FLUID
should be performed from the top fill valve because the Polyalphaolefins (PAOs)
relative density is greater than one. However, many askarel
transformers have been retrofilled and the askarel fluid BIOTEMP
replaced with silicone which has relative density less than BIOTRANS
one. In this case, the transformer should now be sampled ECO Fluid
from the bottom. Table 1 is a list of sampling points for EDISOL TR
various dielectric liquids in apparatus during routine test-
ing. Sampling of drums, tankers, and other types of storage ENVIROTEMP® FR3
containers are performed in a different manner. Consult the ENVIROTEMP 200
previously referenced guides for specific procedures.
Once the correct valve from which to retrieve the sample
has been determined, that valve should be prepared for
taking of the sample. As mentioned previously, check for
positive pressure on the apparatus before opening the drain
valve. Adequate preparation of the valve for sampling con-
sists of the following:
Insulating Oils Handbook 41
Clean the outside of the valve to remove any loose de-
bris that may fall into the sample.
Make sure the valve and sampling cock are closed be-
fore removing the drain plug.
Prepare the area under the valve with absorbent materi-
als and a catch pan. Valve Seat
Transformer Tank
Slowly remove the drain plug.
Clean the inside of the valve with a lint-free cloth. Drain Plug
Hand Wheel
Valve Stem
Packing Nut
Corrosive Sulfur
in Transformer Systems
NETA World, Fall 2003
by Lance R. Lewand
Doble Engineering Company
Table 4 Table 5
Total Sulfur Content of CrossTrans 106 Elemental Sulfur Content of CrossTrans 106
Before Arcing Experiment 0.0118% Before Arcing Experiment <1 ppm
Table 6
Organo-Sulfur Content of CrossTrans 106
COMPOUND Before After COMPOUND Before After
Arcing Arcing Arcing Arcing
hydrogen sulfide <1 ppm <1 ppm ethyl methyl disulfide <1 ppm <1 ppm
sulfur dioxide <1 ppm <1 ppm 2-ethyl thiophene <1 ppm <1 ppm
carbonyl sulfide <1 ppm <1 ppm 2,5-dimethyl thiophene <1 ppm <1 ppm
methyl mercaptan <1 ppm <1 ppm 3-ethyl thiophene <1 ppm <1 ppm
ethyl mercaptan <1 ppm <1 ppm 2,4 & 2,3-dimethly thiophene <1 ppm <1 ppm
methyl sulfide <1 ppm <1 ppm 3,4-dimethyl thiophene <1 ppm <1 ppm
carbon disulfide <1 ppm <1 ppm methyl ethyl thiophenes <1 ppm <1 ppm
isopropyl mercaptan <1 ppm <1 ppm trimethyl thiophenes <1 ppm <1 ppm
ethylene sulfide <1 ppm <1 ppm tetramethyl thiophenes <1 ppm <1 ppm
tert-butyl mercaptan <1 ppm <1 ppm benzothiophene <1 ppm <1 ppm
n-propyl mercaptan <1 ppm <1 ppm methyl benzothiophenes <1 ppm <1 ppm
ethyl methyl sulfide <1 ppm <1 ppm dimethyl benzothiophenes <1 ppm <1 ppm
thiophene <1 ppm <1 ppm trimethyl benzothiophenes <1 ppm <1 ppm
sec-butyl mercaptan <1 ppm <1 ppm tetramethyl <1 ppm <1 ppm
benzothiophenes
isobutyl mercaptan <1 ppm <1 ppm dibenzothiophene <1 ppm 2 ppm
ethyl sulfide <1 ppm <1 ppm methyl dibenzothiophenes 11 ppm 19 ppm
n-butyl mercaptan <1 ppm <1 ppm dimethyl dibenzothiophenes 46 ppm 47 ppm
methyl disulfide <1 ppm <1 ppm trimethyl dibenzothiophenes 54 ppm 59 ppm
2-methyl thiophene <1 ppm <1 ppm tetramethyl dibenzothiophenes 17 ppm 26 ppm
3-methyl thiophene <1 ppm <1 ppm Unidentified volatile sulfur <1 ppm 11 ppm
tetrahydrothiophene <1 ppm <1 ppm Total 128 ppm 147 ppm
46 Insulating Oils Handbook
The arcing did not reduce the total amount of sulfur left
in the oil nor did it convert any of the sulfur compounds
already present into free sulfur. However, conversion of some
of the thiophene compounds did occur (see Table 6). The
presence of sulfates and sulfites cannot be determined by this
method and it may be that those were the compounds that
were converted. This may explain the differences between the
before and after test results. What is of significant interest is
the amount of unidentified volatile sulfur compounds that
were created through the arcing process. No mercaptans
or sulfides (corrosive or reactive sulfur) were formed. The
lack of these sulfur species may be a result of the energy
applied through arcing that could have been so severe that
any reactive or corrosive species that were produced were
instantaneously degraded and thus none remain.
The next article will detail the sources of sulfur in a
transformer system and how it impacts the insulation of a
transformer.
References
1. Clark, F.M. and Raab, E.L. “The Detection of Corro-
sive Sulfur Compounds in Mineral Transformer Oil”,
ASTM Publication, Presented at the Society Meeting,
June 21-25, 1948, pp. 1201-1210.
2. Lipshtein, R.S. and Shakhnovish, M.I. Transformer Oil,
2nd Edition, Israel Program for Scientific Translations,
Jerusalem, 1970.
3. Wilson, A.C.M. Insulating Liquids: Their Uses, Manu-
facture and Properties, IEEE, New York and UK, 1980
4. Oommen, T.V. “Corrosive and Non-corrosive Sul-
fur in Transformer Oils”, Electrical/Electronics Insu-
lation Conference, Chicago, October 4-7, 1993.
Sources of Sulfur
in Transformer Systems
NETA World, Winter 2003- 2004
by Lance R. Lewand
Doble Engineering Company
A previous article provided an introduction to sulfur ments present in the sample and can provide the percentage
in transformer systems and some of the negative effects concentrations of each of these elements, thus providing an
that corrosive or reactive sulfur can have on a transformer. elemental breakdown of the material or particles. Results
However, transformer mineral oil is not the only material from two gaskets are shown in Tables 1 and 2.
containing sulfur, and this article will explore some of the
other sources. Sulfur compounds are also present in gaskets, TABLE 1
some water-based glues, copper, and paper insulation used Elemental Composition of an O-ring
in the manufacture of transformers. Sulfur can also be
introduced into the transformer through accidental means Radiator Gasket
such as through the use of incompatible hoses. ELEMENT Outside Surface Inside Surface
It is generally accepted that older gaskets used in trans-
Organic Component ≈80% ≈80%
former applications were made from cork, cork/glyptal, and
corkprene. In more recent years, the most oil-compatible Silicon 8.0% 2.1%
gaskets have been nitrile rubbers such as BUNA-N, fluo- Zinc 5.4% 8.3%
roelastomers such as VITON®, or fluorosilicone rubbers. Sulfur 4.3% 8.2%
Properly made nitrile rubbers of the correct grade (butadiene
acrylonitrile) and fluoroelastomers (fluorinated hydrocar- Titanium 1.2% 0.0%
bon) are excellent gaskets for use in transformers. In the Calcium 1.1% 1.2%
manufacture of some of these materials such as the nitriles, Aluminum 0.0% 0.2%
sulfur is used in the curing process when the formulations
are being developed into a hardened material. The curing
process is supposed to eliminate all sulfur from the finished TABLE 2
product. Most gasket manufacturers assume the sulfur is
eliminated after the curing process. In some cases, the con- Elemental Composition
centration of sulfur contained in the final gasket product of a Butterfly Valve Flat Gasket
is not monitored. ELEMENT Outside Surface Inside Surface
Doble Engineering performed scanning electron mi-
Organic Component ≈80% ≈80%
croscopy/energy-dispersive x-ray analysis (SEM/EDX)
on numerous gaskets taken from recently manufactured Silicon 9.5% 9.0%
transformers. Each gasket was prepared for analysis by Aluminum 4.9% 5.7%
cleaning the outside surface with a sulfur-free hydrocar- Zinc 2.4% 2.3%
bon solvent. The gasket was then cut lengthwise to reveal
the inside surface. The outside and inside surfaces of each Sulfur 2.1% 2.1%
gasket were coated with evaporated graphite. The samples Copper 0.4% 0.0%
were then subjected to SEM/EDX analysis in which an Calcium 0.3% 0.3%
electron beam of the scanning electron microscope enters
Titanium 0.3% 0.3%
the bulk of a sample producing an x-ray emittance. The
x-ray peak positions along the energy scale identify the ele- Iron 0.1% 0.3%
48 Insulating Oils Handbook
Both gaskets contained a large amount of sulfur, especially Note: SEM/EDX analysis cannot quantify elements
the O-ring gasket. The SEM/EDX analysis was performed such as fluorine, nitrogen, oxygen, carbon, hydrogen, and
on the inside surface of the gasket to determine if the out- boron but can sometimes give a qualitative indication of
side surface had possibly been contaminated with corrosive the amount in high enough concentrations.
sulfur from the oil. It is clear that sulfur is a component of As shown in the two tables, the fluoroelastomer shows
both original gaskets. The original formulations for a nitrile no sulfur on the inside surface and very little on the outside,
rubber, fluoroelastomer, or a fluoro-silicone rubber (a fluoro- suggesting that material was cured correctly. The opposite is
polydimethylsiloxane) do NOT contain any sulfur. true of the Parker nitrile material, which shows a very high
In discussions with elastomer manufacturers, it was found sulfur content on both surfaces, suggesting that the sulfur
that very few manufacturers (except for E.I. DuPont) were was not removed after the curing process.
performing any chemical testing on the finished product to Water-based glues, used to secure the paper insula-
determine what amount of sulfur, if any, remained. There also tion during manufacture, may sometimes contain sulfur
does not appear to be any standard on what percentage of compounds. There has been at least one known instance in
sulfur should remain in the final product. It then becomes which the glue used in the manufacture of the windings has
obvious that the onus is on the final user of the material to contributed to a corrosive sulfur condition.
specify a sulfur-free or low-sulfur material for use or to test Most coppers used in manufacturing transformer wind-
the material prior to use. ings contain some impurities. Sulfur happens to be one of
In light of this information, additional SEM/EDX analy- those impurities, along with silver, arsenic, phosphorous,
sis was performed on gasket material available in the Doble tellurium, and oxygen. The amount of sulfur allowed in most
Materials Laboratory. One sample was a fluoroelastomer, of the electrical grades of copper is 15 parts per million or
and another was a nitrile rubber produced by Parker. The less. Analyses performed on random copper samples from
results are shown in Tables 3 and 4. windings showed that the sulfur contents were very low at
five parts per million or less. However, care still has to be
taken in the selection of materials used in construction so
TABLE 3 that copper with a high sulfur content is not used.
The pulping process for coated transformer paper, such
Elemental Composition of a Fluoroelastomer Gasket as electrical Kraft paper, converts wood chips to cellulose
ELEMENT Outside Surface Inside Surface by removing the majority of lignin (95-98.5 percent) and
other impurities. There are two basic processes:
Organic, Fluorine >90% >90%
Component
• The sulfite process is considered an acidic process and
Calcium 6.88% 7.40% uses sulfur dioxide, sulfuric acid, and calcium bisulfite.
Magnesium 1.06% 2.47% • The main process used today and the one that is used
Chlorine 0.31% 0.13% to produce electrical grade coated transformer papers is
Silicon 0.40% 0.0% the sulfate process, which is also called “alkaline pulp-
ing.” Sodium hydroxide and sodium sulfide are used
Phosphorous 0.40% 0.0%
in what is termed the “cooking process.” The cooking
Aluminum 0.36% 0.0% process under conditions of heat, pressure, and chemi-
Sulfur 0.33% 0.0% cals (pulping liquors) removes the lignin and impuri-
Potassium 0.26% 0.0% ties from the wood chips in order that only cellulose
remains. The pulping liquor is removed and recycled for
use again, and the remaining cellulose pulp is washed
several times to remove as much of the pulping liquor
as possible from the cellulose pulp.
TABLE 4
Elemental Composition of a Parker Nitrile Gasket The Kraft process is slightly different, in that the same
ELEMENT Outside Surface Inside
chemicals are used, but the pulp is intentionally undercooked
Surface
and results in the darker color of the paper as well as ex-
ceptional mechanical strength. The pulp fibers in the Kraft
Organic Component >95% >95% process do absorb some of the sulfur compounds that can-
Sulfur 2.46 2.73 not be removed via the washing/rinsing process. Tests were
Zinc 1.79 1.85 performed to determine how much total sulfur remains in
the finished paper products. The first analysis performed was
Aluminum 0.40 0.21
SEM/EDX analysis of new Kraft and thermally-upgraded
Silicon 0.35 0.21 (TU) Kraft papers from United States manufacturers. These
results are listed in Table 5 and are for the surface of the
paper only.
Insulating Oils Handbook 49
In addition, several different samples of Kraft paper corrosive or reactive. In these cases, appropriate material
insulation were analyzed for total sulfur and total sulfate compatibility testing should screen out these materials
content. The results are present in Table 6. before they are used in transformer construction.
TABLE 5 Conclusions
Many internal and external sources of sulfur exist within a
Surface Composition of Kraft and TU Kraft
transformer besides the oil. Internal sources include copper,
ELEMENT KRAFT TU-KRAFT paper, gaskets, glues, and possibly other materials. External
Organic Component >95% >95% sources usually include inadvertent contamination from
incompatible materials such as oil-transfer hoses.
Calcium 3.6% 2.5%
Sulfur 0.4% 1.1%
Silicon 1.1% 1.4%
References
Casey, James P., ed. Pulp and Paper: Chemistry and Chemi-
cal Technology, Interscience Publishers, Inc., New York,
TABLE 6 1952.
Sulfur Composition in Various Electrical Papers
PAPER Total Sulfur Total Sulfate Lance Lewand received his BS degree at St. Mary’s College of
Content* Content* Maryland in 1980. He has been employed by the Doble Engineering
Kraft Paper-1 700 ppm 205 ppm Company since 1992 and is currently the Laboratory Manager for the
Doble Materials Laboratory and Product Manager for the DOMINO®
Kraft Paper-2 300 ppm <7.5 ppm product line. Prior to his present position at Doble, he was the Manager
of the Transformer Fluid Test Laboratory and PCB and Oil Services at
TU-Kraft 700 ppm 158 ppm MET Electrical Testing in Baltimore, MD. Mr. Lewand is a member
Kraft Crepe Paper-1 600 ppm 93 ppm of ASTM committee D 27.
Kraft Crepe Paper-2 500 ppm 30 ppm
Before you begin oil testing, it is important to identify D1533 Oil Moisture The Karl Fischer Moisture Test
which transformers need tested, how often, and which tests measures water ppm within the
are necessary for analyzing the oil. It is recommended that transformer oil. Note, it does not
measure the water content in the
all substation class transformers have the following tests paper insulation.
performed at least annually. Listed below is a summary of
recommended oil tests that should be performed as well as D-924 Power Factor The power factor test measures
a description of the test11. This list is by no means a com- the dielectric losses of the oil, or
energy that is dissipated as heat. A
plete list of all the tests that can be performed. Rather, it is low value indicates low losses. It is
a recommended list. The decision is up to each customer to a useful test for measuring changes
determine which tests should be conducted and analyzed. within the oil resulting from con-
It is important that when sampling, the individual verifies tamination or deterioration.
that there is positive pressure on the pressure gauge so that
D-877 Dielectric The dielectric test measures the
no air bubbles enter the transformer. Otherwise, serious D-1816 Breakdown ability of the oil to withstand
injury and failure of equipment may result. Voltage electric stress without failure. The
The intention of this paper is not to provide an in depth higher the value, the lower the
chemical analysis on how each test can measure the chemical presence of contaminants such
changes taking place within the oil and the insulation. For as water, dirt, or other conductive
particles.
the purposes of this paper, the information provided about
each test is to serve as a guide for the decision maker. D 3612 Dissolved Gas The DGA test measures various
Analysis gas ppm levels that are present.
Different gases will dissolve in
the oil that indicate various types
of thermal and electrical stress oc-
curring within the transformer.
Oil Maintenance For on-site energized oil treatment, the Hot Oil Rec-
lamation Processing (HORP) rig needs mobilized and set
Thus far we have developed an understanding of what is up at the customer’s substation. Because it is important to
occurring inside the transformer, along with which tests to test the oil initially, as well as monitor the oil as it is being
perform, as well as how to interpret the results. Next, it is reclaimed, the HORP rig needs to be equipped with an
important to consider the options available to protect the on-board laboratory to analyze the oil as it is processed.
transformer. Listed below are different options depend- The process begins by connecting two sets of high-pres-
ing on the specific data for each transformer that needs sured hoses to the transformer and circulating the hot oil
processed. in a closed loop path. The oil needs to be pulled from the
bottom of the transformer and return to the top.
a. Hot Oil Reclamation
b. Alternative Options There are six key components to the removal of harm-
i. Retrofilling ful contaminants within the oil. The six components are as
ii. Vacuum Filling follows:
iii. PCBs
1. Heat
2. Centrifuge
3. Fuller’s Earth
4. Vacuum Degassing/Dehydration Column
54 Insulating Oils Handbook
5. Inhibitor water absorption between paper and oil. Note the higher
6. Filter stages the temperature, the higher the water ppm is in the oil, and
the lower the moisture content is in the paper. Remember,
Different companies have different oil treatment process- water has an affinity for paper, but when you introduce
ing rigs. Below is a general layout of the EPS HORP rigs. heat, it transfers the moisture into the oil. This is important
because water in the paper will decompose the insulation
integrity77.
Safety
The advantage of the hot oil reclamation process is that, in
most cases, it can be performed on energized transformers.
Thus, the customer never has to take an outage and incur
additional expenses of setting up a mobile transformer. In
addition, the energized process helps to aid in the clean-
ing of the transformer. Typically, 230kV is considered the
ceiling for performing hot oil reclamation on energized
transformers. But not all transformers qualify. Performing
a hot oil reclamation process on energized transformers is
safe as long as:
Vacuum-Filling References
In some cases, the primary voltage is too high to perform 1 1
Dr. Fredi Jakob, “W-ACTI Diagnostic Reference
a retrofill. Instead, the transformer must be vacuum-filled. Guide”, 2002
The vacuum-filling process will remove air and water, the
2 2
IEEE Guide for Reclamation of Insulating Oil and Cri-
biggest contaminants of an insulating system. In addition, it
teria for Its Use (IEEE Std 637-1985)
is also important to remove the air so that no voids are left
within the insulation. These voids can create corona and elec- 3 3
S. Myers, “Transformer Life Can Be Extended”, TMI
trically stress the windings. How long a transformer needs Reference Guide
Insulating Oils Handbook 57
4 4
Dr. Fredi Jakob, “W-ACTI Diagnostic Reference
Guide”, 2002
5 5
IEEE Guide for the Interpretation of Gases Generated in
Oil-Immersed Transformers (C.57.104-1991)
6 6
IEEE Draft: Guide for Acceptance and Maintenance of
Insulating Oil in Equipment, (C57.106/D7)
7 7
C. Manger, “High Voltage Equipment Maintenance
101, Mineral Insulating Oil”, TechCon 2002
8 8
C. Bandt, “Filtering Insulating Oil”, Presentation for
American Society for Testing and Materials, 2002
Presenter
Fredi Jakob, Ph.D., Weidmann-ACTI
Co-Authors
Karl Jakob, P.E., and Simon Jones, Weidmann-ACTI
Rick Youngblood, Cinergy Corporation
I. Introduction Table 1
Dissolved Gas Analysis, DGA is based on the fact that “Key” Fault Gasses
the release of energy in oil filled electrical equipment results
in a partial destruction of insulating fluids and/or solid Gasses Indication
insulation. The number of molecules destroyed is relatively Hydrogen Partial Discharge, Heating Arcing
small but detectable amounts of low molecular weight gases Ethylene, Ethane, Methane “Hot Metal” Gasses
are produced. The quantification of these low molecular Acetylene Arcing
weight “fault gases” is the basis of all DGA.
Carbon Oxides Cellulose Insulation Degradation
In the main tank of a transformer very little, if any fault
gas other than carbon oxides should be produced. Carbon
oxides are continuously produced by unavoidable decompo-
sition of cellulose insulation. In contrast, equipment that in- II. Data Interpretation
volves conductor switching operations produces these same A. Concept
fault gases during normal operation. Initially it was assumed
that the gases produced by the arcing process associated The interpretation of DGA data for transformers, LTC’s
with switching would “mask” gases associated with equip- and OCB’s is empirical in nature. The development of in-
ment problems such as carbon deposition on the contacts, terpretation protocols for OCB’s and LTC’s parallels the
which results in excessive heating. Youngblood1 was one of development of DGA diagnostics for the main tanks of
the first investigators to recognize that the gassing pattern transformers. Key gases associated with heating problems
for a healthy load tap changer, LTC, would be significantly are methane, ethane and ethylene. These gases are listed by
different from that observed in a problem LTC. Specifically, Halstead2 in order of increasing energy required for their
he suggested that acetylene and hydrogen are generated production. Initially Youngblood ignored the levels of arcing
during the normal arcing process and the hydrocarbons, gases, acetylene and hydrogen that developed whenever an
methane, ethane and especially ethylene are generated LTC operated. Subsequent work by Youngblood5 indicated
when overheating occurs in a problem LTC. These three that arcing gases are also diagnostically significant. For
hydrocarbons are often called ‘hot metal gases’ since they example, increased acetylene levels were often followed by
are produced when a heated conductor is in contact with increased heating gases. The increased acetylene is due to
mineral oil based dielectric fluid. Correlation of fault gasses changes in the arc duration and/or characteristics as the
with specific gas producing conditions is known as the “key contacts are eroded or covered with carbon.
gas” method. Key gas correlations are given in Table 1. The next step in the development of diagnostic protocols
for LTC’s was the empirical determination of normal or
threshold values. The gas retention rate in an LTC is very
dependent on breathing configuration, so this is a major
Insulating Oils Handbook 59
factor in determining threshold levels. Free breathing LTC’s
rapidly lose gases to the environment while sealed LTC’s Table 2
retain most of the gas produced. Gas Formation as a Function
Threshold levels have been determined for specific models
and types of LTC’s by Doble3 and Baker4. Generic levels of Operation Count. (Duval5)
have been set by Youngblood5, and are useful for specific
Operations: 500 3600 49000 Gas produced/operation
LTC models where the threshold values have not yet been
determined. Fault gas ratios, which are discussed below, Gas/500 Gas/3600 Gas/49000
should be considered applicable for unit evaluation only
when threshold values are reached. Hydrogen 6870 12125 14320 13.74 3.37 0.29
B. Fault Gas Concentration Ratios Methane 1028 5386 10740 2.05 1.50 0.22
8-17.
D a te C o m m e n ts R a ti o 1 R a ti o 2 R a ti o 3 R a ti o 4 R a ti o 5
3 /1 2 /9 2 A n n u a l D G A T e s t C yc le 0 .1 5 0 .1 2 0 .2 1 - 0 .1 5
2 /1 /9 3 6 M o n th T e s t C y c l e 0 .3 2 0 .1 1 0 .2 0 .2 1 0 .3 3 Dr. Fredi Jakob received his PhD at Rutgers, the State University of
8 /1 2 /9 3 Th e rm a l R u n a w a y 155 66 94 1 .0 3 155 New Jersey, in 1961. He is professor emeritus of analytical chemistry at
California State University-Sacramento and is the founder and laboratory
director of Analytical ChemTech International, Inc. (ACTI), which is a
Ratio 1 C2H4/C2H2 wholly owned subsidiary of Weidmann Systems International. As a long-
term member of ASTM and IEEE and author of over fifty published
Ratio 2 C2H4/(C2H2+H2) articles, Dr. Jakob is a traveling lecturer to private and governmental
Ratio 3 (CH4+C2H4+C2H6)/(C2H2+H2) agencies. He has been invited to speak at American Public Power meet-
ings, ASTM symposia, conferences held by Doble, NETA, and AVO
Ratio 4 C2H6/CH4 conferences, as well as other industrial organizations.
Ratio 5 C2H4/C2H6
12-Mar-92
This unit indicated the early stages of mechanical difficul-
ties. While the Acetylene and Hydrogen levels are elevated,
the level of Ethylene is less than 100 ppm. Indicating a
continuance of annual monitoring
1-Feb-93
At this time the unit was placed on a 6 month monitoring,
due to the elevated, Acetylene, Hydrogen, and Ethylene
levels. At 534 ppm Ethylene immediate removal from
service was not indicated.
12-Aug-96
Too late, by August, the unit was in thermal runaway.
As indicated by the extremely high level of Ethylene, at
253,024 ppm. Repairs included a Tap Shaft board, Slip
Rings, and a New Reversing Switch assembly.
References
1. Youngblood, R., Jacob, F., Haupert, T.J. “Application of
DGA to Detection of Hot Spots in Load Tap-Chang-
ers”, Minutes of the Sixtieth Annual International Con-
ference of Doble Clients, 1993, Sec. 6-4.1.
by Lance R. Lewand
Doble Engineering Company
The individual paper fibers are clearly visible in both the In an effort to evaluate the effect that this type of con-
Kraft and crepe papers in Figure 4 with no signs of foreign tamination has on paper insulation, dielectric breakdown
material present. Even the crimping of the crepe paper is strength testing by ASTM Method D 149 was performed
clearly visible. When copper/sulfur or other metal/sulfur on the three layers of insulation that surrounded the copper
contamination of the paper surface occurs, the results can conductor. The first layer, or the one closet to the conduc-
be profound as shown in Figure 5. tor, was the insulation shown in Figure 5A. The other two
layers of paper insulation were also contaminated but not
to the degree of the first layer. The results of this testing is
shown in Figure 6.
64 Insulating Oils Handbook
insulation or the metal surfaces may have already occurred
and should be taken into consideration. However, removal
and replacement of the contaminated oil may help mitigate
any future damage.
Conclusions
The effects of corrosive/reactive sulfur not only adversely
affect the conductor material and other metal surfaces but
also can have drastic effects on insulating materials such as
paper. The information presented clearly demonstrates that
serious contamination due to sulfur-containing surfaces
Figure 6 can drastically reduce the dielectric strength of the paper
insulation while no or very little change has taken place in
Layer 1, shown in Figure 6, had an extremely low dielec- the mechanical strength of the paper.
tric strength at 80 volts/mil. As the contamination is reduced Identifying the sources of sulfur contamination and be-
with each paper layer further out from the conductor, the ing able to monitor concentrations of corrosive and reac-
dielectric strength increases significantly to almost as new tive sulfur will help detect problem areas earlier. Therefore,
condition (1800 volts/mil oil impregnated). Of significant a multifaceted approach is employed in which different
interest is the fact that the mechanical strength was not tests are used to provide complementary and meaningful
impacted by the contamination as the DP values for all three information. Other than draining and flushing with new
layers ranged from 903 to 938. The silvery appearance of oil, a commercially-viable remedial process has not been
the paper in Figure 3 is due to the alteration of the sulfur- developed which can remove corrosive and reactive sulfur
bearing compound due to exposure to higher temperatures, species from the oil. Hopefully, a method for treating the
as copper sulfide is usually gray-black in appearance. This contaminated oil can be developed, and Doble is continu-
reaction of copper and sulfur created deposits on the first ing this research.
and second layers of paper on the same sample. Because
of the deposition of the copper and sulfur, the first inside
wrap became a poor insulator and, instead, was more con- References
ductive. When the corrosive sulfur contamination is this 1. ASTM D 1275: “Standard Test Method for Corrosive
severe, a failure of the transformer is almost inevitable, as Sulfur in Electrical Insulating Oils” in Electrical Insu-
in this case. lating Liquids and Gases; Electrical Protective Equip-
ment, Annual Book of ASTM Standards, Vol. 10.03,
Processing Oil To Remove Corrosive ASTM, West Conshohocken, PA, 2001.
Sulfur 2. DIN 51353: Testing of Insulating Oils, Detection of
Several methods were examined to see if they could be Corrosive Sulfur, Silver Strip Method, Auslandsnor-
used to reduce the corrosive sulfur content in transformer menvermittlung im DIN, Burggrafenstrabe 6, D-1000
oil. An oil that clearly failed the corrosive sulfur ASTM test Berlin 30, Germany, 1985.
D 1275 was used for the experiments. The oil was treated
with Fuller’s earth (clay) at a ratio of 30 grams of clay per
liter of oil (0.25 pounds of clay: 1 gallon of oil) at two tem- Lance Lewand received his BS degree at St. Mary’s College of
Maryland in 1980. He has been employed by the Doble Engineering
peratures, 45 and 85 C. This type of treatment proved to be Company since 1992 and is currently the Laboratory Manager for the
futile in that the treated oils failed the ASTM D 1275 test. Doble Materials Laboratory and Product Manager for the DOMINO®
Another type of process using a synthetically-manufactured product line. Prior to his present position at Doble, he was the Manager
absorptive also proved to be unsuccessful in removing the of the Transformer Fluid Test Laboratory and PCB and Oil Services at
corrosive sulfur. The last test involved the use of activated MET Electrical Testing in Baltimore, MD. Mr. Lewand is a member
of ASTM committee D 27.
carbon. Three hundred milliliters of contaminated oil were
processed through 128 grams of carbon activated at 120 C
for 24 hours. The sample was then tested according to
ASTM D 1275. The treated sample failed the test, indicating
that it was still corrosive.
Currently, the only successful way to remove corrosive
sulfur from transformer oil is to remove it from the trans-
former and dispose of it. Adequate flushing of the trans-
former should be effected to remove or dilute any remaining
contaminated oil in the system. If the oil has been present in
the transformer system for some time then damage to the
Insulating Oils Handbook 65
Nomograph for
LTC-DGA Data Interpretation
PowerTest 2004
(NETA Annual Technical Conference)
Presenter
Fredi Jakob, Ph.D., Weidmann-ACTI
Co-Authors
Karl Jakob, P.E., and Simon Jones, Weidmann-ACTI
Rick Youngblood, Cinergy
*Note: Some gas is always lost with time. Therefore, the gas concentration per
operation is expected to decrease with operation count. Duval did not provide
breathing configurations for this data.
The ratio is fairly consistent and independent of operation more heating gases are produced, the ratio of heating to
count. When the operation count is very high, 49,000, the arcing gases, ethylene to acetylene, increases. This change
ratio increase is probably related to changes in the contact in gas concentrations and gas concentration ratios indicates
condition. problems. The concentration ratios of ethane to methane
and ethylene to ethane are temperature dependent, so both
2. Heating Problems are expected to increase with increasing temperature in a
problem unit. These two heating gas ratios should also reflect
Initially a resistive film develops on contacts, which
increased contact resistance and heating.
results in an increase in contact resistance, increased heat-
ing and an increase in heating gas concentrations. Since
Rl R2 R3 R5 R5 R6
Ethylene Ethylene Ethylene + Ethane + Methane Ethane Ethylene Methane + Ethylene + Ethane
Acetylene Acetylene + Hydrogen Acetylene + Hydrogen Methane Ethane Acetylene
0.36 0.24 0.38 0.31 7.51 0.56
Note the difference between the model specific ratios listed in Table 4 and the generic values which are listed in Table
3. The differences in critical ratios for the Westinghouse and McGraw Edison units illustrate how model specific the gas-
sing characteristics can be.
III. Indicating Problem Severity same escape rates from the oil the ratios should remain
The severity of a problem, or condition of the unit, may fairly 100 ppm then the ratio is one. If a heating problem
be designated as Normal, Caution or Warning. A proposed is superimposed on the normal arcing process and the gas
nomograph has been developed by taking the 90th percentile levels are 175 ppm for ethylene and 150 ppm for acetylene
values as a baseline, and aligning the concentration scales so there would be an “additional” 25 ppm of ethylene due to
that any connecting line with this ratio will have a slope of the heating problem. One could thus normalize our results
zero. Units having a zero slope and/or a slope that resides using a ratio of:
below these threshold values are diagnosed as Normal. A Ethylene Total Ethylene
slope of the line residing within one standard deviation of (due to heating) (Ethylene produced during normal arcing)
————————— = ——————————————————
the threshold values is diagnosed as a Caution, and at two Acetylene (due to arcing) Acetylene
standard deviations is a Warning. A section of the proposed
nomograph is illustrated in Figure 3.
References:
1. Youngblood, R., Jakob, F., Haupert, T.J. “Application of
DGA to Detection of Hot Spots in Load Tap Chang-
ers,” Minutes of the Sixtieth Annual International
Conference of Doble Clients, 1993, Sec. 6-4.1.
2. Halstead, W. D., “A Thermodynamic Assessment of the
Formation of Gaseous Hydrocarbons in Faulty Trans-
formers,” Journal of the Institute of Petroleum, Vol. 59,
Sept. 1959, pp. 239-241.
3. Doble Client Transformer Committee Subcommittee
Report on Transformer Load Tap Changer Dissolved
Gas Analysis, September 24, 2001.
4. Baker, Charles. Personal correspondence. 2002.
5. Youngblood, R., Baker, C, Jakob, R, Perjanik, N.,
“Application of Dissolved Gas Analysis to Load Tap
Changers.”
6. Duval, Michel., “A Review of Faults Detectable by Gas-
in-Oil Analysis in Transformers,” IEEE Electrical In-
sulation Magazine, May/June 2002, Vol.18 No. 3, pp.
8-17.
7. Church, J.O., Haupert, T.J., and Jakob, R, Electrical
World, Vol. 201, No. 10, October 1987, pp. 40-44.
8. Rose, Don. Personal communication.
Figure 3. Nomograph Diagnosis of LTC Condition. Dr. Fredi Jakob received his PhD at Rutgers, the State University of
New Jersey, in 1961. He is professor emeritus of analytical chemistry at
California State University-Sacramento and is the founder and laboratory
director of Analytical ChemTech International, Inc. (ACTI), which is a
IV. Future Work wholly owned subsidiary of Weidmann Systems International. As a long-
term member of ASTM and IEEE and author of over fifty published
The empirical analysis of DGA data for LTCs is well articles, Dr. Jakob is a traveling lecturer to private and governmental
developed. Gas concentration levels and gas concentration agencies. He has been invited to speak at American Public Power meet-
ratios can differentiate between normal and problem units ings, ASTM symposia, conferences held by Doble, NETA, and AVO
We believe that both the concentration and ratio values will conferences, as well as other industrial organizations.
work best if they are model specific The compilation of these
values requires user feedback on problem units. Trending
of both the gas concentrations and ratios is always the best
method to identify incipient problems.
Another concept under investigation is normalization
of fault gas data. We believe that during normal switching
operations, the ratio of ethylene to acetylene is fixed. This
ratio should remain constant for different numbers of opera-
tions. Furthermore, since these two gases have approximately
the same solubility in mineral oil and approximately the
70 Insulating Oils Handbook
How a Condition Assessment is Determined ing of apparatus as to condition so that a priority hierarchy
In order to provide a condition assessment for OCBs and can be established. This permits maintenance activities to
LTCs, a numeric ranking is determined through the use be more thoughtfully developed and managed, thus saving
of the four separate sections of the analytical data: DGA, time, money, and unplanned outages due to failure.
oil quality, particle count, and metals. The rankings from
the four groups are summed or otherwise manipulated to
provide a numeric ranking. The ranking system is further References
reduced to a “Condition Code” from which specific main- Asche, R. G., “Dissolved Gas Analysis of Load Tap Chang-
tenance functions are recommended. This ranking system ers, Subcommittee Project Report,” Proceedings of the
is shown in Table 1. 2002 Annual International Conference of Doble Clients,
Transformer Test and Maintenance, 2002, Sec. 13E.
by Lance R. Lewand
Doble Engineering Company
The newest versions of natural ester dielectric fluids have oils were quickly supplanted by the introduction of Western
been in use in the electrical apparatus industry since ap- Pennsylvania paraffinic mineral oils which were in turn
proximately 1998 and are becoming more popular. Unlike eventually succeeded by naphthenic-based mineral oils.
transformer mineral oil, which is refined from petroleum Beginning in the early 1990s, there was renewed interest
derived from compressed plant and animal tissue (mostly in developing natural ester dielectric liquids able to over-
microorganisms from 70 to 440 million years ago), these come some of the deficiencies that plagued the vegetable
liquids are produced from renewable resources such as oils used in the 1890s and 1900s. ABB Inc. and Cooper
vegetable oils and seeds. Power both began intensive research into the formulation
Reference is given to these liquids as “the newest ver- of natural esters, and both now have commercially available
sion” because vegetable oils have been experimented with products. In addition, Waverly Light and Power, a small
and used as dielectric liquids since the late 1880s. George city-owned utility in Iowa, has used a soy-based dielectric
Westinghouse and his team of engineers used castor oil and liquid developed by the University of Northern Iowa’s Ag-
linseed oil as dielectric liquids and insulation impregnants. Based Industrial Lubricants Research Program in pole top
This helped to reduce the size of the transformers by reduc- transformers since about 1998. This product is now produced
ing dielectric clearances and allowed the solid insulation to by Cargill. Several patents were issued for these products
last longer by minimizing the effects of oxygen. Until Elihu beginning in 1999. [See Oomen and Claiborne (1999),
Thomson patented mineral oil for use in transformers in McShane et al (2000), and Cannon and Honary (2000).]
1892 and its acceptance in the mid-1890s, vegetable oils The main purpose for development of many of these
(natural esters) were the dielectric liquids used. dielectric liquids was to create an environmentally friendly
In fact, many vegetable oils have been tested over the product that was not only stable when used as an insulating
years, including linseed, cod liver, hemp, walnut, poppy, liquid in electrical apparatus but also readily biodegraded
sunflower, cottonseed, sesame, peanut, rapeseed, olive, and when exposed to the environment. This would allow spills
others. Two oils in particular, linseed and tung, were very and leaks to be handled more easily than transformer min-
important in the early years of transformer manufacturing. eral oils. Since the first production and use of these liquids,
These oils were considered to be “drying oils” and, as such, other positive attributes have been discovered. As listed in
polymerized easily when exposed to oxygen. Because of this Lewand (2001), potential users of these types of dielectric
characteristic they were used extensively in the formulation liquids would expect some of the following qualities to be
of insulating varnishes. [See Clark (1962).] present.
These same characteristics, however, presented problems
and were objectionable when used as a liquid dielectric. High level of biodegradability
Formation of organic peroxides occurred and eventually Nontoxic
resulted in the formation of aldehydes and acids. These
vegetable oils polymerized easily when exposed to oxygen. Material compatibility with electrical apparatus
Such by-products are unwanted in transformer systems as components
they accelerate the degradation process of the liquid and Good dielectric strength and insulation properties
solid insulation. In addition, vegetable oils have much higher
viscosity and pour point characteristics and could not be Similar dielectric constant as the solid insulation
manufactured consistently. Since early transformers were all Long-term oxidative and thermal stability
free-breathing, oxygen played a major role. Thus, vegetable Relatively low pour point
74 Insulating Oils Handbook
Can be consistently produced The O represents oxygen, C represents carbon, R repre-
Long service life sents an alkyl or aryl group, OR’ represents an alkyl or aryl
group attached to an oxygen, a single dash represents a single
Compatibility with presently used liquids (mineral bond, and a double dash represents a double bond. For those
oil, etc.) wishing to know even more chemistry, an alkyl group is one
Reasonable cost that contains carbon and hydrogen in the following pattern
CnH2n + 1. The n can be any number and simply refers to the
Examples of current, commercially-known, natural ester length of the molecule. An alkyl group is derived from its
dielectrics are shown in Table 1. alkane parent. An alkane is a hydrogen-to-carbon saturated
compound in which all the groups attached to the carbon are
Table 1 - Natural Ester Dielectric Liquids single-bonded. An aryl group or aryl radical is an aromatic
hydrocarbon in which there is one hydrogen less than its
NAME TYPE MANUFACTURER
parent form.
All these ester oils have a triglyceride component of
BIOTEMP Comprised mostly of mono-un- ABB Inc. which many of the properties of the oil are based on the
saturated high oleic acid triglyc- fatty acid content of the oil. A formula for a triglyceride is
eride vegetable oils. The oleic
acid group is defined as having shown below.
one carbon double bond, part of
the eighteen carbon atoms in the
hydrocarbon chain of a carbox-
ylic acid. Examples of high oleic
oils are sunflower, safflower, and
rapeseed (canola).
References
Abeysundara, D.C., Weerakoon, C., Lucas, J.R., Guna-
tunga, K.A.I., and Obadage, K.C. “Coconut Oil as an
Alternative to Transformer Oil,” in ERU Symposium,
November, 2001.
Cannon, G.S., and Honary, L.A.T., “Soybean Based Trans-
former Oil and Transmission Line Fluid,” United States
Patent 6,159,913, issued December 12, 2000.
Clark, Frank M. Insulating Materials for Design and En-
gineering Practice, John Wiley and Sons, New York,
1962, 1218 pages.
Lewand, Lance R. “Laboratory Evaluation of Several Syn-
thetic and Agricultural-Based Dielectric Liquids” in
Proceedings of Sixty-Eighth Annual International Confer-
ence of Doble Clients, 2001.
McShane, C. P., Corkran, J.L., Harthun, R.A., Gauger, G.
A., Rapp, K.J. and Howells, E., “Vegetable Oil Based
Dielectric Liquid,” United States Patent 6,037,537, is-
sued March 14, 2000.
Oommen, T.V., and Claiborne, C.C., “Electrical Trans-
formers Containing Electrical Insulation Fluids Com-
prising High Oleic Acid Oil Compositions,” United
States Patent 5,949,017, issued September 7, 1999.
Laboratory Testing
of Natural Ester Dielectric Liquids
NETA World, Winter 2004-2005
by Lance R. Lewand
Doble Engineering Company
A previous article presented information on a new breed then, once they are in use, the extent of degradation due to
of liquid dielectrics, namely natural esters. This article thermal, oxidative, and hydrolytic actions and contamina-
provides information on the tests that can be performed tion. As oils age while in service, acidic and polar compounds
to evaluate the chemical, physical, and electrical properties are formed and ultimately develop into a sludge or polymer-
of natural ester oils. After the last article, I became aware ized material if remedial action is not taken. The oil can be
that Dielectric Systems Incorporated also manufactures a replaced or rejuvenated by clay treatment or possibly other
natural ester dielectric liquid called ECO Fluid which is processes such as activated alumina to remove the acidic and
based on a high-oleic acid content canola oil (see Sundin). polar contaminants before they induce excessive aging of the
Information concerning ECO Fluid is presented in this solid insulation and corrosion of metal structures or result in
article along with the other natural esters. In addition, two sludge formation. The generation of sludge and polymerized
oils have been removed from the market, BITORANS material can hamper cooling in the transformer, resulting
1000 and 5000. in an increased rate of aging of the insulation.
Throughout this article, many comparisons are made to
the oil quality test results of transformer mineral oil, as this Physical Tests
is something we are all familiar with. It must be emphasized,
Below are the results of testing of new natural esters.
however, that the properties of a natural ester cannot be
Where Doble did not perform the testing, the manufac-
correlated directly to that of a mineral oil as the chemistries
turer’s data is provided, if available.
of the two types of liquids are very different. They are so
As shown in Table 1, significant differences exist between
different, in fact, that ASTM has produced a new specifi-
natural esters and typical mineral oil – especially aniline
cation just for natural esters, referenced as ASTM D 6871
point, flash point, interfacial tension, pour point, and vis-
“Specification for Natural (Vegetable Oil) Ester Fluids Used
cosity. Certain tests such as aniline point and interfacial
in Electrical Apparatus.”
For the most part, the same tests
used to evaluate mineral oil are used Table 1 – Physical Tests of Natural Ester Dielectric Liquids
to evaluate natural esters. Results and
the meaning of the tests are different ASTM TEST Mineral Oil BIOTEMP Coconut Oil ECO Fluid Envirotemp
between the two oils. Doble evaluates Typical Values FR3
natural esters using five categories Data Source Doble Producer Producer Doble
and are compared to specification Furanic Compounds, D 5837, µg/L all < 5 all < 5 654 HMF
Gassing Characteristics of
Transformer Oil Under Thermal Stress
NETA World, Fall 2005
In June 2005, ASTM International adopted a new test wide range of problems and is used to characterize incipient
method for testing transformer oil. The aim of the test is to fault conditions. Categories of abnormal conditions such
determine the gassing pattern of an oil subjected to thermal as overheating, partial discharge, and arcing can be further
stress under what is considered to be low temperatures, i.e. identified by the insulating materials involved (paper or
120 C. The method is entitled “ASTM D 7150, Standard oil) or by the energy (high, moderate, or low temperatures).
Test Method for the Determination of Gassing Character- DGA is ideally suited for condition-based maintenance
istics of Insulating Liquids Under Thermal Stress at Low programs as it allows early detection of problems and can
Temperature.” then be used to follow most deterioration processes as they
Interest in the development of this test resulted from evolve. Then, the decision can be made as to when to per-
research that Doble performed in the early 1990s and form more definitive tests to identify the specific problem
more recently the research that the CIGRE Task Force TF or take remedial action. Indeed, DGA is sensitive enough
15/12-01-11 (TF11) performed in trying to discern the that it can be used to detect problems during factory testing
differences in gassing behavior of some transformer oils at of new transformers helping manufacturers avoid shipping
low temperatures. CIGRE has termed the unusual gassing defective units to the field.
behavior as “stray gassing.” It is imperative that DGA results are interpreted accu-
The CIGRE research was mainly looking at the gassing rately to ensure the correct action is taken concerning any
behavior of new oils in which transformers were exhibiting incipient faults within the transformer. It is well known that
increasing hydrogen levels, yet no apparent cause for the ab- oils can age differently, particularly where there is ample
normal gassing could be determined. As a result, the research oxygen such as may be found in older units, or free-breath-
centered on the oil and found that some oils produced more ing transformers. A 1991 study revealed that oils exposed
gas (in this case hydrogen) than others when they were aged to relatively low temperatures could have different gassing
under thermal conditions in both sealed transformers (gas behaviors and oxygen consumption rates under the same
blanketed, sealed or bladder/diaphragm transformers) and controlled conditions. Gassing rates were different for two
free breathing transformers (open conservators). Doble had oils tested if the oxygen content was fairly low at 3000 ppm
also observed a similar pattern and as a result conducted or high at 30,000 ppm and this is one of the reasons why
a massive research study into the phenomenon that has the new ASTM D 7150 test is performed using air and
lasted over three years and still goes on today. What was nitrogen sparged samples. It was also shown that passing
also determined was the fact that not only could the refining the dissolved gas-in-oil test for a factory heat run could
process be partially responsible for the abnormal production depend on the oil chosen.
of dissolved gases, but different types of contamination in
the oil, incompatible materials in the transformer, and the Early Studies on Gassing Behavior
addition of additives such as metal passivators could also
In 1994, the Doble Oil Committee performed a study
cause abnormal gassing to occur.
on the low temperature gassing behavior of an oil manufac-
The reason that this information is so important is that
tured in France that was being used for the first time by a
it can significantly affect the results of the dissolved gas-
French transformer manufacturer. The transformer did not
in-oil (DGA) test. DGA is one of the most widely used
pass the dissolved gas-in-oil limits used by the purchaser.
diagnostic tools for assessing the condition of electrical
The manufacturer performed some tests which showed that
transformers and in more recent years load tap-changers
the oil used (Shell Diala F) gassed much more than the oils
and bulk oil circuit breakers. The test is very sensitive to a
Insulating Oils Handbook 81
they used historically. To evaluate the gassing behavior of Nitrogen Purged, 16 hours at 120 C
the Shell Diala F oil in comparison to other products, air- Nitrogen Purged, 164 hours at 120 C
saturated samples and some samples that had been degassed
(all gases removed except for small amounts of oxygen and The aging was performed for differing times to assess
nitrogen) were aged in ground-glass matched barrel and initial gassing rate and when a plateau or equilibrium rate
plunger syringes at 95ºC for 168 hours. The following oils was reached. The samples were either air or nitrogen purged
were tested in the study: to have a range of oxygen contents representative of service
conditions. The aging times were as follows:
Shell Diala F, uninhibited, manufactured in France
16 hours – provides indication of initial generation
Esso Univolt 52, uninhibited, manufactured in France
of gases and would be similar to the time for a fac-
Exxon Univolt N61, inhibited, manufactured in US tory heat run test.
Shell Diala A, uninhibited, manufactured in US 164 hours – indicates if the gases reach a plateau or
The results are provided in Table 1 (air saturated) and a constant rate of generation
Table 2 after the oil had been vacuum processed. The oils were assigned a sequential number based on the
total amount of combustible gases formed during the experi-
TABLE 1 ment with air saturated oils aged for
Gassing Characteristics of New Oils Aged at 95ºC, ppm vol./vol. 16 hours. For example, number 1 had
Oil Hydrogen Methane Ethane Ethylene Acetylene CO CO2 TCG the least amount of total combustible
Diala F 378 6 0 1 63 2 0 215 908 764 gas (TCG) and number 30 the greatest
Univolt 52 122 1 0 1 58 10 0 113 1414 404 amount of TCG. In some cases there
Diala A 81 4 3 20 1 0 65 608 201 was more than one product from a
Univolt N61 35 0 0 0 0 46 489 81 refiner. Only the results of the 164-
hour testing are presented here as this
TABLE 2 is the time that was decided upon for
Gassing Characteristics of Vacuum-Processed New Oils Aged at 95ºC, the ASTM method.
ppm vol./vol.
Oil Hydrogen Methane Ethane Ethylene Acetylene CO CO2 TCG Air Saturated Samples
Diala F 165 133 99 2 0 116 370 515 Aged for 164 Hours
Univolt 52 0 4 2 0 0 0 42 6
at 120°C
For the samples saturated with air
and aged for 164 hours, the data shown in Figure 1 reveals
The results listed in Tables 1 and 2 show clearly that the that there is a large difference in the total combustible gas
Shell Diala F product produced larger amounts of combus- (TCG) content between some of the oils and that very sig-
tible gases than the other oils tested. Even when oxygen is nificant amounts of combustible gases are formed in some
removed or reduced in the test samples, the overall concen- cases. There were four gassing patterns that developed:
tration of gases is reduced but the same oil still produces
the largest quantity of dissolved gases. Table 2 reveals that Mostly carbon monoxide
degassing the two oils to reduce the oxygen content had a Hydrogen and carbon monoxide
much greater effect on the Univolt 52 oil than the Diala
F product. This type of experiment shows that the gassing Mostly hydrogen
characteristics of new oils under thermal stress can be drasti- Mixture of hydrogen, carbon monoxide, methane
cally different. In the case of a transformer heat run, one oil and ethane
would cause the transformer to fail the test while the other
The predominant gas formed was hydrogen. However,
oil would not be of concern.
other gases such as methane, ethane, ethylene, carbon mon-
oxide, and carbon dioxide were also formed. In general, if
Recent Studies on Gassing Behavior the oxygen content remained high, the first three gassing
Doble, with the inspiration of the CIGRE work and its patterns occurred (as listed above). If the oxygen was de-
own previous work, undertook a study to evaluate the gas- pleted, the methane and ethane were generated in greater
sing characteristics of oils commercially available in 2001 concentrations. Ethane was not generated in significant
and 2002 and in many cases still sold today. The study was amounts without methane. The oxygen consumption rate
to look at a wide range of oils to determine the amount of was quite different for the various products.
variation.
Testing was performed under the following conditions:
Air purged, 16 hours at 120 C
Air purged, 164 hours at 120 C
82 Insulating Oils Handbook
Concentration, ppm
Concentration, ppm
Nitrogen Purged Samples Aged 164 Hours low temperature thermal problems could be confused
at 120°C with low energy partial discharge activity. Stray gassing
may also be mistaken for excessive gassing at modest
Samples purged with nitrogen and then aged 164 hours temperatures.
did not contain as much combustible gas as those saturated
with air before aging for the same amount of time; however, Some of the “breaking in” characteristics that are ob-
there were five oils that still exhibited high concentrations served for the gassing behavior of new transformers or
of hydrogen and total combustible gases. In these cases the newly processed transformers may be attributed to the
hydrogen made up a large percentage of the composition oil and will stabilize over time.
of the TCG. Stray gassing can also be due to the presence of con-
What is interesting in nitrogen-sparged samples is that tamination in the oil, the presence of incompatible ma-
the methane and ethane values are consistently low but make terials, and the presence of additives.
up a greater percentage of the TCG for more samples than
was the case when there was ample oxygen at the start of It is clearly important to know the properties and char-
aging. This is due to the fact that the hydrogen produced in acteristics of the oil in individual transformers, especial-
nitrogen-sparged samples is not as high in concentration. ly when performing DGA. Similarly, it is important to
Similarly, the carbon oxide gases are consistently lower for make sure the gassing characteristics of oils used to ‘top
the nitrogen-purged samples than when saturated with off ’ transformers are also well understood.
air.
Conclusions References
There is a new ASTM Method D 7150 that provides
the protocol for determining the gassing characteristics 1. Griffin, P.J., Lewand, L. R., Heywood, R., and Lap-
of oils under air and nitrogen (sealed) conditions and at worth, J. “Gassing Characteristics of Transformer Oils
low temperature (120 C). at Modest Temperatures, Part 1: Transformer Experi-
ences”, Proceedings of the Seventy-First Annual Con-
It is not meant to be a routine test method in the sense ference of Doble Clients, 2004.
that DGA currently is. However, it can be applied in
those situations where there is a question of whether 2. Griffin, P.J., Lewand, L. R., Heywood, R., and Lap-
the oil is involved in producing the gassing that is being worth, J. “Gassing Characteristics of Transformer Oils
observed. at Modest Temperatures, Part 2: Laboratory Experi-
ments”, Proceedings of the Seventy-First Annual Con-
DGA is a very important test, but care needs to be tak- ference of Doble Clients, 2004.
en in the interpretation of results to understand the low
temperature gassing behavior of the oil when certain 3. Schmidt, J., Eitner, R., and Hartwig, R., “Gassing Be-
types of problems are detected. As hydrogen is often havior of Transformer Oils at Temperatures Above
the predominant gas formed for the oils with the high- 60ºC”, 7th International Symposium on High Voltage En-
est combustible gas generation rates, it is possible that gineering, Aug. 26-30, 1991, pg. 33-36.
Insulating Oils Handbook 83
4. Griffn, P. J., “Gassing Characteristics of New Oils Used
in Factory Heat Runs”, Doble Oil Committee Minutes,
1994, pg. 18-21.
Condition Assessment
of Transformers —
Analysis of Oil Data and Its Quality
NETA World, Winter 2005-2006
by Lance R. Lewand
Doble Engineering Company
Condition assessment or life assessment of a transformer really that high? Was the sample taken correctly? Did the
should use a variety of techniques. These techniques en- past test data indicate the same condition? In addition, the
compass electrical testing, a physical assessment, review of temperature of the transformer from which the sample was
historical design and test data, review of future load and use taken plays an extremely important part in understanding
requirements, oil analysis, and others. No one technique will the data that was generated in the laboratory. For example,
be able to accurately assess the condition, and they all must if a sample was taken from a transformer operating at 80°C
be used in concert to provide a good evaluation. with 1.6 percent water in the paper insulation, the water
Insulating oil analysis is one of those very powerful and content of the sample should be around 33 ppm under
invaluable tools to aid in such an investigation. The most re- equilibrium conditions. When the sample gets transported
cent data as well as historical trends are especially important to the laboratory for analysis there will be a reduction in
to provide relevant information of the transformer in ques- temperature. At room temperature, say 21°C, with 33 ppm
tion. Many analytical tests can be performed which provide of water, the relative saturation of water in the oil will be
information that can then be gathered and formulated into around 57 percent and the dielectric strength may be low
an overall diagnosis. For example, tests such as interfacial because of it. Although, there are no temperature correction
tension, color, and neutralization number provide informa- factors for dielectric strength test measurements, a seasoned
tion on the condition of the oil. Water content along with analyst must be able to recognize the importance of the
temperature data can provide information on the wetness or temperature change and its effects on the laboratory results
dryness of the liquid insulation and by calculation, the water compared to the dielectric breakdown voltage of the oil in
content of the solid insulation. Dissolved gas-in-oil analysis the transformer at operating temperatures.
(DGA) is an extremely powerful test as it is a sensitive test Another aspect of transformer condition assessment is
for detecting a wide range of incipient-fault conditions or reviewing the oxygen and nitrogen data from the DGA.
problems that may exist within the unit and may influence Take, for example, the information provided in Table 1.
operation decisions. Furanic compounds are by-products
of paper degradation and their determination provides Table 1: Oxygen and Nitrogen Concentrations
information on the mechanical strength of the paper and
localized overheating of paper insulation. Other tests such from the DGA, ppm
as corrosive sulfur, particle count, and particulate and dis- Sample Date Oxygen Nitrogen
solved metals provide information on unusual conditions
7/14/1990 1150 85,200
that may exist in a transformer and can help in providing
an understanding of the source of a problem or help clarify 9/15/1991 1640 81,400
those problems identified by other methods. 8/7/1992 1200 87,300
The quality of the data is especially important as poor 8/9/1993 1350 82,500
data can lead to a poor diagnosis or misdiagnosis. Take for 7/25/1994 2000 84,100
example a case in which the dielectric breakdown strength is 9/10/1995 1860 86,700
low (poor). A closer look at the results may indicate that the 8/27/1996 1280 85,300
water content is elevated which directly affects the dielectric 8/10/1997 1410 84,900
breakdown strength measurement. However, a series of 9/1/1998 1570 86,200
questions must then be asked such as: Is the water content
Insulating Oils Handbook 85
From a review of the data in Table 1, several items can content due to sampling error, problems in the laboratory,
be gleaned. For instance, the sample taken on 8/7/1992 in or an actual condition? That is a much harder question to
which a high oxygen content is present indicates that there answer, and more data is needed before a conclusion can
was probably a sampling error that occurred that allowed air be drawn. If samples taken after 9/1/1998 show that the
to enter the syringe prior to analysis. The nitrogen content is oxygen content reverts back to the low concentrations, in
too high for a sealed conservator transformer in good condi- most cases it was some error in the way the sample was
tion, so it is likely from a gas-blanketed unit. If the nitrogen taken, stored, and/or analyzed that caused the problem. If
value were much higher it could indicate that regulator on the oxygen content continues to increase, then there may
the nitrogen tank feeding the transformer may be set too be a problem with the transformer in which the seals of
high causing more nitrogen to be dissolved in the oil. This the unit have become compromised and oxygen is now
makes the oil more susceptible to being super-saturated able to enter the system. The presence of high oxygen in a
with nitrogen. Oils that are supersaturated with nitrogen transformer increases the rate of degradation of the oil and
can release bubbles during quick cool downs which cause solid (cellulose) insulation and is unwanted.
a reduction in the oil dielectric breakdown voltage. Most The lesson learned is that every test and result must be
users of nitrogen blanketed transformers set the nitrogen reviewed critically in order to provide a competent diagnosis.
pressure at 1-3 psi to avoid this condition. Sometimes, In addition, many tests are interrelated, and their results
however, the information from the field does not match must be reviewed as such. The transformer nameplate data
the data. For example if the data sheet coming from the arriving from the field is especially important as transform-
field were marked “conservator” for the type of preservation ers of different voltage classes, preservation systems, and
system, then the data in Table 1 would not match as different other factors effects the diagnosis that is provided. Providing
nitrogen concentrations would be expected unless there was comments on oil test data based on IEEE or other limits
a leak in the system. In this case, the type of preservation is not enough and in most cases provides a very limited
system would have to be confirmed. diagnosis. The types of oil analysis tests performed must be
In many cases, the oxygen and nitrogen content data is selected in such a way as to provide the best overall condi-
not so clean. The data such as that presented in Table 2 is tion assessment. To obtain a good condition assessment
often seen. on transformers, more oil testing is needed than that for
routine testing. Good sampling techniques and appropriate
Table 2: Oxygen and Nitrogen Concentrations sample containers are especially important and can impact
all the results and the final condition assessment and have
from the DGA, ppm
been covered in detail in previous articles. The final lesson is
Sample Date Oxygen Nitrogen that the quality of the data from the laboratory is especially
7/14/1990 1150 85,200 important to obtain good information for evaluating the
9/15/1991 1640 81,400 condition of transformers. Poor analytical data takes time
8/7/1992 18,100 60,200 to sort through, can result in a misdiagnosis, and can use
resources for resampling and retesting in order to obtain a
8/9/1993 1350 82,500
clear understanding of apparatus condition.
7/25/1994 2000 84,100
9/10/1995 4100 79,300
8/27/1996 8,200 75,100 Lance Lewand received his Bachelor of Science degree from St.
8/10/1997 1410 84,900 Mary’s College of Maryland in 1980. He has been employed by the
Doble Engineering Company since 1992 and is currently the Laboratory
9/1/1998 15,300 71,900 Manager for the Doble Materials Laboratory and Product Manager for
the DOMINO®. product line. Prior to his present position at Doble, he
As shown in Table 2, there is a lot of variation in the was Manager of the Transformer Fluid Test Laboratory and PCB and
oxygen and nitrogen data. In the case of the sample taken Oil Services at MET Electrical Testing in Baltimore, MD. Mr. Lewand
on 8/7/1992 there was probably a sampling error in which is a member of ASTM Committee D 27.
air was allowed to enter the syringe prior to the analysis. The
syringe itself could have been compromised or the sample
was taken in another container such as a bottle in which
there is no way to preserve the oxygen/nitrogen content in
its original state. If the sample container is not gas tight,
other gases will also be affected. Examples are hydrogen
and carbon monoxide whose solubilities in oil are low and
thus can easily escape from the sample. The oxygen levels
on 9/10/1995 and 8/27/1996 are also slightly elevated and
may be due to a combination of sampling error, syringe
leakage, or sample preparation in the laboratory. The data
from the last sample point, 9/1/1998, is very interesting as it
raises many questions. Is the reason for the elevated oxygen
86 Insulating Oils Handbook
1. Introduction Table 1
During the past forty years Dissolved Gas Analysis, Gases Produced during Normal and Abnormal
DGA, was developed, promoted and accepted as the single Equipment Operation
most effective diagnostic tool for power transformers. There
are many additional types of electrical equipment that also Equipment Normal Gases Stress Gases
use mineral oil as a dielectric fluid and can be subjected to Transformers CO, CO2 C2H2, C2H4, C2H6,
thermal and electrical stresses that result in the production (Power) H2, CH4, CO and CO2
of fault gases. A logical question is: why has there been a LTCs and OCBs C2H2, H2 and all other Increased amounts of
“delay” in applying DGA to other electrical apparatus? There fault gases “Hot Metal” Gases
are two basic reasons for the delayed application of DGA Bushings Very Little Gas, Small All Possible Fault Gases
to an expanded array of equipment. The first reason is that Amounts of CO2 and CO
switching equipment such as load tap changers and oil cir- Voltage Regulators, Same as Power Same as Power
cuit breakers produce gases during normal operation. It was Current and Potential Transformers Transformers
thus assumed that the presence of these normally produced Transformers
fault gases would obscure the observation of additional
gases produced by thermal or electrical faults. Fortunately,
as shown below, analysis of the DGA results can reveal Further consideration of the gas formation mechanism
problems, even in the presence of “background” gases. in LTCs and OCBs led to the conclusion that the gas con-
A second reason for the delayed application of DGA was centration pattern changes when these units are subjected
the limited quantity of oil in apparatus such as bushings, to thermal or electrical stress. During normal operation
current and potential transformers. Removal of sufficient the primary arcing gases, acetylene and hydrogen, are pro-
quantities of oil for DGA analysis could lower equipment duced in abundance. Due to the cooling effect of the oil,
oil volumes below required levels. There was, and still is, temperatures at which the hot metal gases are produced
concern about possible contamination of the equipment exist further from the arcing site. Thus all of the fault gases
during the sampling process. These issues have been suc- are produced in both LTCs and OCBs. Load tap changers
cessfully addressed and, as a result, DGA has once again do not contain any cellulose-based insulation, but carbon
proven to be a powerful diagnostic tool for additional dioxide and carbon monoxide are produced from the heated
electrical equipment. oil in the presence of dissolved oxygen.
A common problem in both LTCs and OCBs is the
2. Load Tap Changers and Oil heating of contacts due to a buildup of carbon. This pro-
cess is known as coking. The deposit increases the contact
Circuit Breakers resistance which increases contact temperature and results
It was commonly assumed that DGA was not applicable in additional carbon formation from the breakdown of the
to equipment in which switching and associated arcing oc- oil. This self- accelerating process leads eventually to “ther-
curred during normal operation. This concept is illustrated mal runaway.” Abnormal heating from the contact coating
in Table 1. results in additional production of the heating gases, meth-
ane, ethane and ethylene. Temperatures are high, but not
Insulating Oils Handbook 87
high enough to produce additional quantities of acetylene.
Analysis of these observations leads one to conclude that
all of the fault gases are produced during normal operation
and that contact coking leads to the production of additional
hot metal gases.
The total gas concentration in an LTC or OCB will be
related to the number of operations and loss of fault gas
from the oil to the atmosphere. The gas concentration ratios
should be relatively independent of the operation count.
One can thus conclude that changes in gas ratios are very
important. Additional reasons to consider gas concentration
ratios as the most appropriate diagnostic data interpreta-
tion tool are:
1. Gases are lost continuously to the atmosphere. Hydrogen
is most rapidly lost and the rate of gas loss depends on tem-
perature and breathing configuration. The ratio of ethylene
(a key heating gas) / acetylene (the key arcing gas) remains
fairly constant since these gases are both lost much more Figure 1 — LTC Damage
slowly than hydrogen.
2. The ratios of heating to arcing gases have been shown
to be strongly dependent on the extent of contact coking. B. OCB Case Study
3. LTC and OCB Case Histories
John Stead1 has published several important papers that C1 Power Factor .31 .44 1.23 .40
show that DGA is often a more reliable indicator of bush- C2 Power Factor .21 .20 .24
ing condition than the widely used power factor test. The Serial # 19147 Nameplate 1989 1993 Shop
reluctance of U.S. manufacturers to recommend DGA for C1 Power Factor .29 .31 1.14 .77
bushings as a diagnostic tool is based on their concern with C2 Power Factor .25 .27 .27
oil contamination during the sampling process. Stead1 has
developed a sampling protocol to minimize the possibility
of bushing oil contamination. The newest DGA laboratory
method, D3612-C, uses a much smaller oil sample (5-9ml) Observations during Disassembly Serial # 25610
than the original D-3612-A (35-50ml) method. Therefore, Large region of severe carbon tracking observed lower
less oil has to be removed from the bushing in order to end of bushing core.
complete a DGA. Inner surface of lower porcelain weathershed showed
At this time interpretation of the DGA results is similar carbon tracking.
to that used for power transformers (see Table 2).
Small amount of carbon tracking had penetrated to
within 1 turn of potential tap layer.
Table 2
DGA Interpretation Considering the condition, it was extremely fortunate
that this bushing had not failed in service.
Predominant Gas(es) Interpretation Overall test (GST-Ground) at 10 kV was 8.80%
H2 Partial Discharge power factor.
C2H2, H2 Arcing Water content was measured at 6 PPM.
CH4, C2H4, C2H6 Heating
Observations during Disassembly Serial # 25600
Oil contained small flakes of discolored paper.
IEC2 recommends a DGA for bushings at a four-year
interval. Several Canadian utilities are relying more on Water content was measured at 3.5 PPM.
DGA than on power factor measurements to evaluate their Dielectric strength measured by ASTM Method D1816
bushing population. was 27 kV.
No carbon tracking was observed on either the bushing
core or the weathershed.
Insulating Oils Handbook 89
Observations during Disassembly Serial # 19147 nologies are important components of a condition-based
Odor of burnt oil and paper was noticed approximately maintenance program. Application of DGA to every type
half way to the core during the unwrapping process. of oil filled equipment is rapidly becoming as significant, as
the historical application to power transformers.
Paper was of a lighter color as we moved inward to the
core and did not appear to be saturated with oil as in the
outer wrappings.
References
X-wax was observed on the copper draw lead tube. 1. John Stead, Fredi Jakob “USE OF DGA TO CON-
Power factor improved after removal from unit FIRM UNSATISFACTORY DOBLE TEST RE-
SULTS OF 115 KV BUSHINGS” Minutes of Six-
Bushing serial number 15940 was not disassembled ty-Third Annual International Conference of Doble
due to time constraints. Clients, 1996 63PAIC96
2. IEC 60599:1999-03, Mineral oil-impregnated electri-
6. Voltage Regulators cal equipment in service – Guide to the interpretation
Voltage regulators provide utility and industrial com- of dissolved and free gases analysis.
panies the ability to maintain constant line voltage under
varying load conditions. Voltage regulators are usually
coupled with tap changers that provide the ability to control Karl Jakob is the Director of Business Development for Weidmann-
ACTI Inc. He holds a Bachelor of Science in Electrical and Electronic
the voltage during peak and reduced demand to optimize Engineering from California State University Sacramento and is a Reg-
operating conditions. DGA is an important diagnostic istered Electrical Engineer in the State of California. He is a co-founder
tool used to identify problems within voltage regulators of Analytical ChemTech International (ACTI) and has co-authored
through the use of threshold values and ratios of heating several papers on oil diagnostic applications which have been published
to arcing gases. Thresholds values are utilized to help deter- in trade journals and presented at technical conferences such as EPRI,
NETA, Doble, etc. Karl is an active member of the IEEE Transformer
mine whether a diagnostic ratio is applicable to a specific Committee.
regulator unit, because low levels of Ethane, Ethylene and
Acetylene are typical for voltage regulators, and the ratios
are only applicable when a minimal level of each gas used
in the ratio is present in the DGA analysis.
The gas ratios used for the voltage regulator diagnostics
are Ethylene/Acetylene, and Ethane/Methane. These ratios
can give an indication of contact coking, contact erosion,
and overheating.
8. Conclusion
Management and stockholder expectations require
utilities to operate at increased efficiencies and generate
profit for investors. Reducing expenses is one way to meet
these demands. It is especially significant if one can simul-
taneously decrease costs and increase system reliability.
Maintenance has always been an area subject to budget
cuts because the effect of these cuts was not immediately
evident. Simply cutting maintenance costs is a short term
option. System reliability must be maintained despite the
cuts. Fortunately, condition based maintenance can both
save money and simultaneously increase system reliability.
Increased utilization of DGA and other diagnostic tech-
90 Insulating Oils Handbook
Passivators —
What They Are and How They Work
NETA World, Spring 2006
by Lance R. Lewand
Doble Engineering Company
In the past several years there have been some failures of One major concern is that there is a population of trans-
very large power transformers and shunt reactors associated formers and reactors that are in service with oils that have
with the presence of corrosive sulfur in the transformer oil. excessive corrosive sulfur or a propensity to form it. Many
This problem occurs despite the fact that the oils involved of the failures seem to have occurred in the past few years
have passed standard specification test criteria for corrosive with apparatus that has been in service from 1-7 years. Dur-
sulfur, such as ASTM D 1275. These failures occur because ing this window of time, the number of units with suspect
corrosive sulfur in the oil reacts with copper to form copper oil that have been placed in service could represent a large
sulfide, a conductive compound. The copper sulfide can form monetary value even though it is a small percentage of the
at the copper surface or with copper ions in the oil and paper. total units in service. Not only is there the cost of the ap-
Some of the copper sulfide formed at the copper surface can paratus, there is also loss of revenue and possible regulatory
also migrate to the paper insulation. The formation of copper fines if power is not delivered as specified in contracts.
sulfide is depicted in Figure 1. The conductive copper sulfide This concern about corrosive sulfur has led several trans-
causes a reduction in dielectric strength of the paper insula- former manufacturers and oil refiners to recommend the use
tion. Failure results when the dielectric breakdown strength of passivators by adding them to the oil. There are different
of the conductor insulation is exceeded by the voltage stress kinds of passivators that can be used, but one that has been
which can be influenced by transient voltages. The result recommended for the corrosive sulfur problem is added to
is arcing between two or more turns or possibly disks [1]. reach 100 ppm (mg/kg). This would mean that in a 10,000
Information about corrosive sulfur has been presented in gallon transformer, 7.5 pounds of this passivator would
this publication previously [2,3, and 4]. have to be added to the oil to make the final concentration
100 ppm.
So what exactly are passivators? Although passivators
have been around for a long time in the lubrication industry
S
Cu S S S S S
S
S
their use is relatively new in the electrical industry. Pas-
Cu
Cu Cu S S S S sivators, also know as metal deactivators, react with reactive
Cu S S
Cu S S metal surfaces and dissolved metals such as copper and silver
and reduce their rate of reaction with compounds in the oil.
This includes oxidation reactions with organic compounds
and reactions with corrosive sulfur. Passivators are composed
of two basic types, sulfur based and nitrogen based.
Cu
Cu S
Cu
Cu S Cu
The first suggested use of passivators in transformer oil of
Cu S which the author is aware was in 1967 by J.J. Melchiore and
Cu Cu Cu
I.W. Mills of the Sun Oil Company. The article that they
Cu
Cu Cu S CuCu Cu S
Cu S S
Cu Cu S CuCu Cu S
Cu S S
Cu
Cu S
wrote [5] dealt with the oxidation stability of transformer
oils in response to accelerated oxidation tests such as the
DOBLE PFVO (power factor valued oxidation) test. The
Figure 1 — Corrosive Sulfur Attack of Copper Conductor and Ions results of the testing that they performed suggested that
(Cu=Copper; S=Sulfur) the use of a metal deactivator (passivator) coated the cop-
per surface that was used as a catalyst in the experiment.
Insulating Oils Handbook 91
The passivator decreased the catalytic effect of the copper sulfide was eventually observed on copper surfaces. The
and the amount of copper that dissolved in the oil. This question is whether the passivator sufficiently retards the
reduced the oxidation of the oil and formation of deleteri- corrosive sulfur reactions so they are not of concern for the
ous by-products. Passivators have been used in some cases normal life of the apparatus. Once copper sulfide is formed
for this purpose. Passivators have also been used by some on copper surfaces and deposited in the paper insulation, it
transformer manufacturers in Japan for static suppression. is not removed by passivator or other means. The passivator
Passivators that have been used in electrical insulating undergoes a variety of reactions in the apparatus and can be
mineral oils are nitrogen based and have been predominantly degraded or consumed and might need to be added periodi-
benzotriazole (BTA) or its derivatives. BTA is a granular cally to retain sufficient binding on the copper surface to
solid at room temperature and requires heating and mixing block the corrosive sulfur reactions.
to dissolve in the oil, so it is less desirable than its derivatives.
One type of BTA derivative has a hydrocarbon molecule tail Alternative Approaches
attached. The BTA derivative is liquid at room temperature
For new oils the best approach is to use an oil that does
and, therefore, much more readily mixes with transformer
not have significant amounts of corrosive sulfur compounds
mineral oils. One passivator that has recently been suggested
as shown by passing a more rigorous modified version of the
for use for suppression of corrosive sulfur reactions is that
ASTM D 1275 corrosive sulfur test. Doble recommends
produced by CIBA named Irgamet 39. There are several
performing the test at 150ºC instead of 140ºC and for 48
other manufacturers and brand names of products that are
hours rather than 19 hours and using better nitrogen purging
commercially available.
and sealing of the test vessel. In cases where oils have been
The action of the passivator is to have the nitrogen group
used that do not meet this more stringent criteria, passivator
of the BTA molecule bind with the copper, silver and other
could be used to retard further corrosion. An alternative ap-
reactive metal surfaces. This process is a chemical bonding
proach is to use full or partial retrofills of the oil with an oil
that, given the right circumstances, can be undone (reversed).
that does not have significant amounts of corrosive sulfur
The passivator molecule is attracted to the metal surface and
compounds even when exposed to very high temperatures.
is held to the reactive site so that same site cannot be oc-
This can be tested on a case by case basis to determine the
cupied (attacked) by a corrosive sulfur molecule. Passivator
required dilution. All these methods deserve further research
molecules can also bind with free copper and silver ions or
to determine the best long-term approaches to solving the
particles present in the bulk oil (see Figure 2). In effect, the
corrosive sulfur problem.
BTA and corrosive sulfur compete for the reactive metal
sites, along with other compounds.
Conclusions
S BTA
S Passivators are metal deactivators that bind with certain
BTA BTA BTA BTA
BTA BTA
S BTA BTA BTA
BTA S BTA
BTA
BTA reactive metal surfaces to retard or delay attack by corrosive
BTA BTA BTA
BTA
sulfur. Passivators act by blocking the sites for corrosive
Step 1 sulfur compounds to attack the copper and form copper
sulfide. As shown in Figure 2 and as discussed, passivation
Cu Copper
does not completely block all metal reactions with corrosive
BTA Cu BTA Cu BTA BTA Cu
Step 2 S BTA BTA BTA BTA
BTA
BTA
BTA BTA BTA BTA
Cu S BTA Cu S
BTA BTA BTA
Conductor sulfur compounds. Further research is needed to determine
if passivation is a good long-term solution for in-service oils
S S
with excessive amounts of corrosive sulfur.
S S
S S
S S
S
Step 3 Cu Cu Cu BTA
BTA Cu Cu S BTA
BTA Cu BTA
Cu S BTA BTA
Cu S
BTA BTA BTA BTA BTA
BTA
BTACu S
BTA
BTA Cu S BTA
BTA BTA References
[1]. Lewand, Lance, “Investigating Copper Sulfide Con-
tamination in a Failed Large GSU Transformer” in Pro-
Figure 2 — Action of a Passivator (BTA = BTA or its derivatives) ceedings of the Seventy-Second Annual International Con-
ference of Doble Clients, Doble Engineering Company,
Much of the study on the use of passivator for control Watertown, MA USA, 2005
of corrosive sulfur reactions to date has been in the labora-
tory. The long-term effectiveness for retarding the action of [2]. Lewand, Lance R. “Corrosive Sulfur in Transformer
corrosive sulfur in apparatus is not known. Investigations at Systems”, NETA World, Summer, 2003.
the Doble Laboratory have shown that passivators retard [3]. Lewand, Lance R. “Sources of Sulfur in Transformer
reactions of metals with corrosive sulfur. However, even Systems”, NETA World, Fall, 2003.
at moderate temperatures (110 C) with passivator added,
under certain laboratory conditions corrosive sulfur attack [4]. Lewand, Lance R. “The Negative Effects of Corro-
still occurred. The reactions occurred over extended periods sive Sulfur on Transformer Components”, NETA World,
compared to experiments without passivator as copper Winter, 2003.
92 Insulating Oils Handbook
[5]. Melchiore, J.J., and Mills, I.W., “Factors Affecting
Stability of Transformer and Cable Oils”, in Proceed-
ings of the Seventh Electrical Insulation Conference, IEEE,
October 1967.
What is Sludge?
NETA World, Winter 2006-2007
by Lance R. Lewand
Doble Engineering Company
Have you ever been reading a laboratory report listing Pentane-Insoluble Sludge (Doble) This test is a quali-
the results from testing of a transformer oil sample and tative test in which insoluble sludge is found to be either
wondered what the results are trying to tell you? Words present or absent based on a visual inspection. The test is
like sludge or metallic soaps, x-wax or ionic species creep conducted by mixing oil from the sample and a defined
into the vernacular and leave you scratching your head and amount of a hydrocarbon solvent, pentane. The mixture is
wondering what significance that terminology really is. I held for 16-24 hours in a cool, dark place and then exam-
am sure you are not alone in this regard. ined by tilting the vial and visually inspecting for tiny solid
This article will discuss one of these terms, sludge. It particles of gelatinous clumps [2]. This solid material forms
will also briefly discuss sludge formation in transformers, from soluble sludge which precipitates due to the addition
what it is, and its significance. Three basic analytical tests of the hydrocarbon solvent.
are useful detectors of the precursors to sludge formation
and precipitation. These tests are important as they are good indicators
of the formation of degradation by-products of both oil
Interfacial Tension (ASTM D 971, mN/m) The inter- and paper that are oil soluble. The value for the interfacial
facial tension of an oil is the force in dynes per centimeter tension decreases and the neutralization number increases
(milliNewtons per meter) required to rupture the oil film in response to the formation of organic acids and polar
existing at an oil-water interface. When certain contami- compounds (compounds with an oxygen-bonded atom),
nants such as soaps, paints, varnishes, and oxidation by- both of which are detrimental to the quality of the oil.
products are present in the oil, the film strength of the oil These by-products will eventually lead to sludge formation
is weakened, thus requiring less force to rupture. For oils which should be avoided. The pentane-insoluble sludge test
in service, a decreasing value of interfacial tension indicates is a quick test that Doble uses to detect the early stages of
the accumulation of contaminants, oxidation by-products, or sludge formation.
both. This reduction is a precursor of objectionable oxidation The development of sludge in a transformer is deleterious
by-products that may attack the insulation and interfere and presents a number of problems. Once formed in the oil,
with the cooling of transformer windings [1]. it can layer on the winding insulation and in the oil ducts. In
these cases it acts as a thermal insulator. That is, it prevents
Neutralization Number (ASTM D 974, mgKOH/g)
the transformer from being cooled properly, and thus more
The neutralization number of an oil is a measure of the
overheating occurs which contributes to even more sludge
amount of acidic or alkaline materials present. It is a simple
formation so it is a continuously compounding problem. The
acid/base titration and some use it interchangeably with acid
additional heat reduces the life of both the oil and paper
content. As oils age in service, the acidity and therefore the
insulation but mainly the paper insulation. Sludge may be
neutralization number increases. Used oil having a high neu-
slightly conductive as well, depending on its composition,
tralization number indicates that the oil is either oxidized
reducing the dielectric strength of the solid/liquid insula-
or contaminated with materials such as varnish, paint, or
tion. Because it is a polar type material it has a tendency to
other foreign matter. (A basic neutralization number results
attract water which increases its conductivity.
from an alkaline contaminant in the oil) [1].
94 Insulating Oils Handbook
The formation of sludge presents maintenance issues Sludge and Carbon: Carbon fines and sludge fines
as well. Oils with a low interfacial tension and an elevated are often confused with each other. Sludge is formed as a
neutralization number are easy to reclaim by removing result of the process already described. Carbon is formed
the precursors of sludge through the use of activated clay from a totally separate process in which the carbon atoms
(Fuller’s earth) or activated aluminum. However, once are completely separated from the hydrocarbon molecules
sludge precipitates as a solid and bakes on the insulation or found in transformer oil. The process also generates hydro-
mechanical structures of the transformer, much more ef- gen and hydrocarbon gases and requires a large amount of
fort, time, and money are needed to remove it. The primary energy such as that resulting from very high temperature
reason is that it tends to attach to the windings and other overheating or discharge conditions. Carbon is present in
surfaces so it is not readily dissolved in oil and thus not transformers but most often present in load tap-changers
easy to remove. and oil circuit breakers as arcing occurs more frequently
in these devices. Once either sludge or carbon is filtered
How does Sludge Form and What is it? Sludge forms
from the oil it is usually very easy to distinguish between
from the degradation of the oil and paper insulation due
the two. Carbon is usually very black or gray depending
to stressors such as oxygen and temperature and catalysts
on how much carbon is present. Sludge can form various
such as copper. The early precursors of sludge are peroxides,
colors including tan, tannish-green, brown, beige, etc., but
organic acids, alcohols, aldehydes, ketones, lacquers, and
not black. See Figures 1 and 2.
other aromatic compounds, particularly those that have
polar functional groups. As these materials attack other
transformer components such as the paper, iron, and cop-
per, they form intermediate by-products in the oil that then References
polymerize together to form a solid type material which is
[1]. Doble Transformer Oil Purchase Specifications, publica-
called sludge. This is the terminal stage of this degradation
tion of the Doble Oil Committee, Revised January 1,
process [3]. Sludge tends to precipitate out in the coldest
2006, Watertown, MA 02472 USA
and hottest regions of the transformer.
[2]. Oliver, F.S., “Doble Power-Factor Valued Oxidation
Test”, in the Minutes of the Twenty-Seventh Annual
Conference of Doble Clients, Sec. 10-404, Belmont,
MA USA
[3]. Myers, S.D., Kelly, J.J. and Parrish, R.H., A Guide to
Transformer Maintenance, Transformer Maintenance In-
stitute, Division of S.D. Myers, Tallmadge, OH USA
Particulate Metals-In-Oil (Bearing Wear) As shown in the table, there is a significant decrease in
For transformers cooled by pumps, the test for bearing the concentration of most metals once the cleaning and
wear metals should be performed every few years. The most flushing were performed. The importance of the sampling
common metals tested for are copper, lead, iron, zinc, alumi- efforts cannot be overemphasized, as incorrect results can
num, and silver. Some of these metals are produced in the lead to the wrong conclusion.
particulate state when the bearings of the transformer pump,
the pump shaft, and/or the impeller start to degrade in some SEM/EDX Analysis
manner. The small pieces of metals are jettisoned into the Scanning electron microscopy (SEM) and energy disper-
oil where, if small enough, they can remain suspended and sive X-ray (EDX) analysis are two separate tests conducted
thus be sampled and analyzed. This type of analysis incor- at the same time. SEM is an analysis in which a beam of
porates the use of a graphite furnace coupled with an atomic electrons, a few hundred angstroms in diameter, systemati-
absorption spectrometer. In addition, metal particulates are cally sweeps over the specimen in an evacuated chamber.
concentrated by passing a certain volume of oil through a The intensity of secondary electrons generated at the point
0.45 micron filter to capture them. The filter is then treated of impact on the specimen surface is measured, and the
in a manner to leave only the metals in a solution of acid resulting signal fed into a cathode-ray tube display which is
and water. The analysis has the ability to detect elements at scanned in synchronism with the scanning of the specimen
extremely low levels, <1 µg/kg (ppb), which is necessary as to produce a picture. All prepared samples stubs are coated
problems with pumps are usually detected where the con- with evaporated graphite before analysis which improves
centrations of certain metals are around 20 ppb. Although imaging resolution and stability.
not an ASTM test method, Doble performs this analysis EDX is when an electron beam of the scanning electron
on a very routine basis, and it has the capability of being microscope enters the bulk of a sample producing an x-ray
expanded to perform metals analysis on solid materials such emittance. The x-ray peak positions along the energy scale
as paper, pressboard, and other materials. identify the elements present in the sample and can provide
Like ICP and AE, there are advantages and disadvan- the percentage concentrations of each of these elements
tages. For one, it uses a much larger volume of oil for analysis, thus providing an elemental breakdown of the material or
500 mL (0.5 quarts) whereas the ICP only requires 2-10 particles.
mL. There is also a significant amount of sample prepara- Obviously, this is no ordinary test. This powerful tool
tion involved, and the analysis time is lengthy. However, can be used in investigative type analysis such as corro-
there is really no limit on the metal particle size that can sive sulfur issues on paper or copper samples and other
be analyzed. contaminants from a transformer, LTC, OCB, bushing,
etc. This type of analysis is not performed on the oil itself
Sampling but on filtered materials from the oil or from components
The analysis of dissolved or particulate metals is very sen- or contamination inside the transformer that needs to be
sitive to sampling methodology. The sample valve has to be identified. Although this analysis tool has many advantages,
prepared in such a manner as to not introduce contaminant like most analytical techniques there are disadvantages as
metals or elements from that area into the sample. Depend- well. The foremost is cost of the instrumentation itself and
ing on the type of valve, contamination from copper, tin, the significant operating expenses associated with it. Other
and zinc (brass, bronze) as well as iron is not uncommon. drawbacks include the amount of time necessary to prepare
Zinc from such items as galvanized pipe nipples and drain the sample for analysis, the length of time for analysis and
Insulating Oils Handbook 97
the very high level of detection (0.5%, 5000 ppm) when
compared to other analysis techniques.
Conclusion
The three main types of metals analysis that can be
performed are dissolved metals-in-oil, particulate metals-
in-oil, and SEM/EDX analysis. The goal of the article is to
explain the differences between each test and to list certain
advantages and disadvantages to help the user in determin-
ing which test to request.
References:
1. Oommen, T.V. and Petrie, E.M. “Particle Contamina-
tion Levels in Oil-Filled Large Power Transformers”,
IEEE Transactions on Power Apparatus and Systems,
May 1983, Vol. PAS-102, No. 5, pp. 1459-1465.
2. Griffin, P.J. “A Cup Full Of Trouble”, Report on the
Minutes of the Doble Oil Committee Meeting, April
1989, pp. 8-14.
by Lance R. Lewand
Doble Engineering Company
Over the past several years there have been a number lowest voltage class transformer known to have failed
of failures of power transformers and reactors where there due to corrosive sulfur is 35 kV, but most failures have
are indicators of copper corrosion and formation of cop- been above distribution voltages.
per sulfides on conductors and insulating materials. The
conductive copper sulfide reduces the dielectric breakdown In the past several years, in response to this problem, there
voltage of the solid insulation. This can result in a dielectric has been a great deal of research activity in developing new
puncture through the paper insulation. The failures have tests to help identify oils with the potential to have corrosive
occurred after the apparatus have been in service for a year sulfur problems or transformers that have already developed
or more, often several years. At normal operating tempera- copper sulfide contamination. Some of the tests are already
tures the corrosion process appears to take time to form ASTM tests and some are in various stages of development
critical amounts of conductive sulfides. To date the failures and are broken down into two groups.
have occurred without prior evidence of abnormal gassing
behavior; therefore, it appears that this problem is difficult Oil Tests (discussed in this article)
to detect and manage.
This problem is fairly wide known in the large power Transformer Electrical Tests (discussed in a future ar-
utilities but less so with industrial and commercial users ticle)
with oil-filled electrical equipment, but the same issues
can arise. Failures due to corrosive sulfur attack have also Oil Tests
been documented in load tap-changers. One of the most These tests aid in determining if corrosive sulfur com-
frequently asked questions is “Which apparatus is most pounds are present in the oil or identifying sulfur com-
susceptible?” In response to this, Doble has developed the pounds that are presently stable but may become corrosive
following list: under conditions that exist in the transformer (temperature,
time, oxygen, etc.). It must be remembered that some of
1) Units that have the following characteristics these tests are accelerated aging tests and do not represent
a) Manufactured from 1999 to present a transformer environment. The purpose of the accelerated
b) Operate at high temperatures over long periods of aging test is to get a reasonable expectation of what may oc-
time such as generator step-up transformers and re- cur without testing the oil for years instead of a few days. A
actors and continuously fully-loaded transformers listing of some of the tests used with descriptions follows:
c) Are either gas blanketed or have a sealed conserva-
tor system (a few failures have occurred with free- ASTM D 1275B, Corrosive Sulfur in Oil
breathing conservators) This test ages 220 mLs of oil in contact with a copper
d) Have oils that fail the corrosive sulfur tests (ASTM strip in a sealed vessel for 48 hours at 150 C. The primary
D 1275B and CCD tests) purpose of the test is to determine if any corrosive sulfur
compounds in the oil will react with the copper strip to
2) Those unit that do not have completely enameled or turn it gray or black. It is a subjective test in that there is a
varnished conductors comparison of colors of the copper strip with some colored
3) Corrosive sulfur problems have been known to occur in standards and a table of descriptions listing what is corrosive
units older than 1999, but to a much lesser extent. Test and noncorrosive (Figure 1A and 1B). The test was recently
the ones that are most critical to the system first. The modified from just D 1275 in April of 2006 to D 1275A
Insulating Oils Handbook 99
and D 1275B. The B method is more rigorous then the old
method D 1275 or D 1275A and was developed in response
to the problem with corrosive sulfur. The issue was that some
of these oils met the requirement of the oil specifications
using the old D 1275 (D 1275A) test yet became corrosive
while in service and eventually caused failures of transform-
ers, LTCs, and reactors. ASTM D 27 committee recognized
this flaw and modified the test method to include D 1275B.
This is a very routine test.
Conclusion
Corrosive sulfur attack causing failure in oil-filled elec-
trical apparatus is occurring more frequently although the
number of failures are still a low percentage of the popula-
tion. However, the loss of revenues from these units as well
as capital equipment is significant. As a result, numerous
researchers are engaged in activities to determine and
eliminate the source(s) of the corrosive sulfur. Because of
this, many tests are being used in concert with each other
to determine the corrosive sulfur aspects of the oils used in
the electrical apparatus.
These tests will be used to screen out or identify oils that
may not be suitable.
CERTIFICATION
NETA Certified Technicians conduct the tests that ensure that electrical power equipment meets the ANSI/NETA
standards’ stringent specifications.
Certification of competency is particularly important in the electrical testing industry. Inherent in the
determination of the equipment’s serviceability is the prerequisite that individuals performing the tests be
capable of conducting the tests in a safe manner and with complete knowledge of the hazards involved. They
must also evaluate the test data and make an informed judgment on the continued serviceability, deterioration,
or nonserviceability of the specific equipment. NETA, a nationally-recognized certification agency, provides
recognition of four levels of competency within the electrical testing industry in accordance with ANSI/NETA
Standard for Certification of Electrical Testing Technicians, (ANSI/NETA ETT).