SAFER, SMARTER, GREENER

Contact details:
POSITION PAPER
KEMA Laboratories the Netherlands KEMA Laboratories USA KEMA Laboratories Czech Republic
Utrechtseweg 310 4379 County Line Road Podnikatelská 547
6812 AR Arnhem PA 18914 Chalfont 190 11 PRAGUE 9
The Netherlands USA Czech Republic
Tel: +31 26 356 9111 Tel: +1 215 822 4231 Tel: +420 281 012 371

POWER TRANSFORMERS
NEED REAL TESTS TO
DNV GL
PROVE THEY CAN SURVIVE
A SHORT CIRCUIT
Driven by our purpose of safeguarding life, property and the environment, DNV GL enables organizations to advance
the safety and sustainability of their business. We provide classification and technical assurance along with software and
independent expert advisory services to the maritime, oil and gas, and energy industries. We also provide certification
services to customers across a wide range of industries. Operating in more than 100 countries, our 16,000 professionals
are dedicated to helping our customers make the world safer, smarter and greener.

In the energy industry

DNV GL delivers world-renowned testing and advisory services to the energy value chain including renewables and energy
efficiency. Our expertise spans onshore and offshore wind power, solar, conventional generation, transmission and distribu-
tion, smart grids, and sustainable energy use, as well as energy markets and regulations. Our 2,500 energy experts support
clients around the globe in delivering a safe, reliable, efficient, and sustainable energy supply.

The trademarks DNV GL and the Horizon Graphic are the property of DNV GL AS. All rights reserved. www.dnvgl.com/kemalaboratories SAFER, SMARTER, GREENER
©DNV GL 09-2016
2 ENERGY Power transformers need real power to survive a short circuit
CONTENTS
1. Executive summary ___________________________________________________________________ 05
2. Major power disruptions _______________________________________________________________ 06
2.1 Outages: a growing concern _________________________________________________________06
3. Short-circuit stresses in power transformers _______________________________________________08
4. Power transformer in-service failure______________________________________________________ 10
4.1 In-service failure statistics __________________________________________________________ 10
4.2 Consequences of transformer failures _______________________________________________ 11
5. Short-circuit withstand verification ______________________________________________________ 12
5.1 By design review __________________________________________________________________ 12
5.1.1 Comparison with a short-circuit tested reference transformer ___________________ 12
5.1.2 Checking against manufacturer’s design rules for short-circuit strength ___________ 12
5.2 By full-power testing ________________________________________________________________13
Power transformers need real power to survive a short circuit ENERGY 3
5.3 Limitations of the design review approach ____________________________________________ 14
5.3.1 Simulation tools are a simplification of the reality _______________________________ 14
5.3.2 The list of reviewed sub-components is not complete ___________________________ 15
5.3.3 The design review approach is static whereas the phenomena are dynamic ________ 15
5.3.4 Design review does not cover material and production deficiencies _______________15
5.3.5 Many failures occur in other than the ‘design reviewed’ sub-components _________ 15
5.3.6 There is no quality control on the performance of a design review _______________ 15
5.3.7 Conclusion __________________________________________________________________16
5.4 By testing of scale-models (‘mock-ups’) ________________________________________________16
6. Utility considerations _____________________________________________________________________ 17
7. Features of KEMA Laboratories ___________________________________________________________19
8. Laboratory capacity _______________________________________________________________________ 21
9. Test result statistics of power transformers __________________________________________________ 22
10. Failure modes ____________________________________________________________________________ 23
10.1 Observations from short-circuit tests _______________________________________________ 24
11. Discussion ____________________________________________________________________________ 25
12. Conclusions ______________________________________________________________________________ 25
13. References _______________________________________________________________________________ 26
4 ENERGY Power transformers need real power to survive a short circuit Power transformers need real power to survive a short circuit ENERGY 5

1 EXECUTIVE SUMMARY
ABOUT THE AUTHORS

René Smeets René Bruil

Service Area Leader Head of Section
KEMA Laboratories KEMA Laboratories

René Peter Paul Smeets obtained a PdD degree for research work on René Bruil is currently responsible for the planning of the
switchgear. Until 1995, he was an assistant professor at Eindhoven High Power Laboratory and responsible for the transformer testing
University, the Netherlands. During 1991 he worked for Toshiba Co. business of the KEMA Laboratories. He has more than 20 year of
in Japan in the development of vacuum interrupters. In 1995, he joined experience in the field of short-circuit tests on MV and HV equipment,
KEMA, the Netherlands. Currently, he is Service Area Leader with 8 years as test engineer, 5 years in the certification department and
KEMA Laboratories, dealing with innovation and technology the last years as planner of the High Power Laboratory and transformer
management. In 2001 he was appointed parttime professor at business. He was member of the IEC working group MT45 for switches
Eindhoven University, the Netherlands. In 2013 he became vice and is currently member of the STL Technical Committee.
professor at Xi’an Jiaotong University, China. In 2008 he was elected
Fellow of IEEE. He is convener, secretary and member of several CIGRE
working groups, as well as convener of two IEC maintenance teams on
high-voltage switchgear. Since 2008 he is the chairman of the ‘Current
Zero Club’. In 2014, he published the book ‘Switching in Electrical
Transmission and Distribution Systems’ with John Wiley UK. He got
six international awards, authored more than 200 international papers
on several aspects of power switching and testing technology, and
presented many training courses all over the world.

Shankar Subramanay
General Manager
KEMA Laboratories
Shankar Subramany is currently responsible for the High Power
Laboratory of KEMA Laboratories in Arnhem, Netherlands, the largest
short-circuit test laboratory in the world. He received his Master’s degree
in High Voltage Engineering from the Anna University in Chennai, India.
In 1986 he joined the Central Power Research Institute, Bangalore, India
and worked in various roles including construction of the new high power
laboratory and testing of MV and HV T&D components for short-circuit
and switching performance. From 1998-2006, he worked at ABB, Sweden
as Senior Test Engineer and Technical Manager of the High Power
Laboratory and from 2006 as Manager of the High Power, High Current
and Mechanical test laboratories at ABB. During those years he also
held the position as Chairman of SATS Certification and was a member
of TC17 of Swedish Electrotechnical Commission and STL Technical
and Management Committees. Since 2013 he works at the
KEMA Laboratories of DNV GL. In addition to participation in IEC
standardisation teams he is currently also the Deputy Chairman of the
STL Technical Committee.
Wilco Rorijs
Principal Consultant Transformers
Energy Advisory
Wilco Rorijs is a principal consultant with the Energy Advisory unit at
DNV GL, specializing in transformers. He joined DNV GL (formerly
DNV KEMA) in September 2007. He is involved in many transformer
and reactor related matters internationally. His current remit includes
drafting technical specifications; performing factory audits;
preparing tender evaluations; conducting design reviews; undertaking
QA/QC; and witnessing of FAT and SAT of a wide range of
transformers (distribution, GSU, auto, grid, phase-shift and HVDC)
and reactors (shunt and series). He is also involved in power failure
investigations (i.e. RCA) of transformers and reactors both at site as well
as at the premises of manufacturers. Rorijs received his Bachelor degree
of High Voltage Electro Technical Engineering at the HAN University,
Arnhem, the Netherlands in 1992. After graduating he joined (Royal)
Smit Transformers in the position of electrical design engineer.
During his time at Smit, Rorijs became a senior electrical design
engineer and designed many specialized power transformers which
are installed in several countries worldwide. In 2007, he left
Smit Transformers to join DNV GL as a transformers consultant. He is a
member of the Dutch Cigré group A2 Transformers and a member of
transformer related Cigré working groups.
Power transformers are the most expensive pieces of
equipment in power systems. Interruption of service of
transformers needs to be avoided at all time, given the
enormous consequences. International studies have
indicated that the failure rate of transformers is around 0.6%.
Deeper study reveals that a major portion of these failures
(up to 20%) is directly related to short circuits.
During a short circuit, the large currents involved lead to
severe mechanical forces and stresses in the transformer
active parts, which may become deformed when the
structural design of the transformer is not adequate.
The verification of the ability of power transformers to
survive short circuit is the subject of this paper. Two verifica-
tion methods are practiced today. The first one is “design
review”, in which third-party consultants check calculation
results of forces and stresses and compare these with critical
values based on tests or based on internal manufacturer’s
rules. Design review is based on calculation results of
idealized, homogeneous structures, it does not cover
transient phenomena, it excludes a number of key
subcomponents and it is not embedded in a strict quality
surveillance system.
The second verification method is “short-circuit testing”,
in which the complete transformer is subjected to real
short-circuit current and thus to the same stresses as would
occur in service. Short-circuit testing is the only complete
verification method of short-circuit withstand capability of
power transformers. KEMA laboratories are now ready to
test power transformers with rated voltage up to 800 kV and
power up to 500-600 MVA single phase (1500-1800 MVA
three phase).
In spite of the wide application of advanced calculation
methods, still around 20–30% of the transformers, submitted
to a short-circuit test fail to pass the short-circuit test.
Mostly, the reason to fail is a reactance increase beyond the
limit set by the standard, which indicates an unacceptable
internal deformation. In a number of cases, however,
unexpected events are triggered by short-circuit current
which are outside of the “usual” failure modes, like breaking
of a bushing, oil spill, internal flashover etc.
The highest degree of reliability with respect to short-circuit
withstand verification is through full-scale short-circuit testing
in accordance with the international standards.
6 ENERGY Power transformers need real power to survive a short circuit
2 MAJOR POWER DISRUPTIONS
Electricity is the life blood of modern society. Any interruption to its supply can have a huge
impact on individuals, businesses and communities. The transmission and distribution (T&D)
sector and the energy industry as a whole therefore put a great deal of effort into trying to
avoid disruptions to the power supply. But are we doing enough and in the correct way?
The electricity industry is going through a transformation
right now. The need to create an affordable and sustainable
energy supply that can meet the ever-increasing demand
for electricity is driving growth in (often remotely located)
renewable sources and international electricity trading.
That in turn is leading to the rise of so-called super grids:
extremely high voltage, interconnected networks that
transport electricity over much longer distances between
different regions and countries.
2.1 Outages: a growing concern
Super grids increase both the risk and the potential impact
of outages. Losing power in one region is bad, but imagine
what would happen if the supply to half a continent will be
interrupted. Such a threat is already starting to attract
headlines, with articles in many countries speculating on
when and where ‘the big one’ will hit.
Liability issues are becoming a major concern as the general
public, as well as the legislative, financial and regulating
bodies, are increasingly aware that power disruptions are not
inevitable ‘acts of God’, but are in fact avoidable to a great
extent. Investors are increasingly taking into consideration
the quality of their (future) assets using a careful process of
quality assurance.
Electrical power outages, surges and spikes are estimated
to cost more than $150 billion in annual damages to the
U.S. economy[1]. The downtime costs for data centres in
particular are skyrocketing, with the average cost per minute
in 2015 reaching $8,851: up 58% from the $5,600 per
minute tabulated in 2010[1]. The average cost of a data centre
outage rose from $505,502 in 2010 to $690,204 in 2013 to
$740,357 in 2015, representing a 38% increase in the cost of
downtime. Maximum downtime costs are rising faster than
average, increasing 81% percent since 2010 to a 2015 high
of $2,409,991[1].
Power transformers need real power to survive a short circuit ENERGY 7
European experiences are similar: a German/Swiss/Austrian
survey[2] indicates that 59% of outages are due to faulty
equipment/human error, while the costs of outages can
soar to 22 M€/hour in the major metropolises.
What can network operators do to mitigate the risk of major
power disruptions?
To start, we must identify the root causes of power outages.
Repeated research shows that the three main causes of
unscheduled power outages are: weather events, equipment
failure and human error (figure 1).
Animal
Faulty equiment / human error
Planned
For example, a 2015 US outage survey showed that weather
Unknown
is the prime cause of outages (30%), followed by faulty
Vehicle accident equipment/human error (27%). 1100
Weather / trees
Overdemand Both equipment failure and human error can be reduced 1000
through better quality of network components and
900
Number of outages
better educated people. And while the weather cannot
be controlled, its impact on our electricity supply can be
800
Animal Vehicle accident reduced by ensuring that local problems do not escalate
Faulty equipment/human error Weather/trees across the entire grid. So this again comes down to the
700
Planned Overdemand quality of our components and our people. The need to
Unknown reduce the impact of weather-based outages will be even 600
more urgent as the current models of climate change predict
Figure 1 - Reported outages by cause in 2015[1] that more extreme weather events are going to occur. 500
2008 2009 2010 2011 2012 2013 2014 2015
Furthermore, the number of outages caused by faulty
equipment/human error is still increasing (see figure 2): Year
in 2015, it was up by 47% compared to 2008.
Figure 2 - Trend in outages caused by faulty
equipment/human error[1]
8 ENERGY Power transformers need real power to survive a short circuit
SHORT-CIRCUIT STRESSES IN POWER
3 TRANSFORMERS
The power transformer is the most vital substation component since its unavailability creates
a major problem, given the high costs and the long time involved in repair or replacement.
Power transformers are especially sensitive to short-circuit events, as will be made clear in the
following.
The effects of short-circuit currents in transmission and
distribution networks for electric energy are tremendous,
both for the equipment and for the stability of the networks.
Since short circuits are not rare events (as a rule of thumb
one short circuit per 100 km overhead line per year[3]),
short-circuit withstand capability is regarded as belonging
to the main characteristics of the equipment installed.
The capability to withstand a short circuit is recognized as a
major and an essential requirement of power transformers.
Failure to withstand results in damage to the internal (and
even external) parts, and can lead, on short or longer term,
to loss of service.
Short-circuit current leads to electro-dynamic forces on
the windings that cause mechanical stresses in the radial
as well as axial direction. They result from the interaction
of the current with the leakage magnetic field between the
windings, in radial and axial direction, see figure 3.
Core Inner Outer
winding winding
Leakage
flux
Fr = radial force Fa = axial force
Figure 3 - Radial (left, Fr) and axial forces (right, Fa) resulting from
the leakage flux (B) and current (I)
Power transformers need real power to survive a short circuit ENERGY 9
Apart from winding deformation, a variety of deformation
of internal, but also external parts has been observed in
F testing[4]. Permanent deformation of windings may lead
to immediate damage or long-term issues because of
insulation damage, obstruction of oil flow, material weakness
or loose parts. Damage and event break of bushings due to
the mechanical shock has been observed as well.
Radial and axial forces may have the following effect on
the winding: Transformers, like series reactors, have the ability to limit the
short-circuit currents to values pre-dominantly determined
n The radial forces act in the radial direction with a pulsating by the transformer’s impedance. In this way, the design
radial force and tend to compress the turns in the of a power transformer with respect to the short-circuit
inner (normally the lower voltage) winding and to expand current withstand capability is focused towards the
the turns in the outer winding (higher voltage). When the Normal position limitation of short-circuit current. In addition, the control of
Tilted
mechanical design of the supports is not adequate, radial conductors F the forces and stresses exerted by the same short-circuit
stresses may lead to buckling of the inner winding, which currents inside the transformer must be an integral part of
is observed frequently (see figure 4). the design process and quality verification[5].
Figure 5 - Axial forces may result in tilting (schematically) With an increase of the short-circuit power of power systems
during the years, the most severe short-circuit currents will
appear when the transformer is aged. These short-circuit
Circular winding
currents have to be withstood without impairing the
transformer. Short-circuit withstand capability should also
cover the ability to withstand several full asymmetrical
short-circuit currents in each phase and in each
representative tap position without impairing the
transformer suitability for normal service.
Free buckling
External bulge
Axial supporting strips Compressive force
Figure 4 - Radial forces may result in buckling (schematically -
left and after a short-circuit test - right)
n The axial forces act in the axial direction with a pulsating
compression force. Axial and radial forces have been
observed to result in spiralling and/or tilting of
conductors (see figure 5). Through the deformation of
turns, the oil flow may become obstructed, leading to the
formation of hot spots or may lead to future winding short
circuit.
10 ENERGY Power transformers need real power to survive a short circuit Power transformers need real power to survive a short circuit ENERGY 11

4 POWER TRANSFORMER IN-SERVICE FAILURE

From the CIGRE 2015 study cited above, it is reported that

4.1 In-service failure statistics 6
7.2% of the major failures result in a fire, and 6% in explosion
or burst, in total in 126 incidents. Major failures related to
A number of international studies, all conducted between 5 bushings most often lead to severe consequences. From the
External
external short
short circuitcircuit
1974 - 2005 report a failure rate(1) in the range of 0.4 – 3%. 4 115 bushing failures, 41% resulted in fire or explosion.
Design
design

Failure rate (%)

A noticeable increase of fire probability is observed at higher
Ageing
ageing 3
In 2015 a new extensive survey covering the period rated voltages: from 0.02% per year at 120 kV to 0.29% at
unknown
Unknown
1996 - 2010 of 964 major transformer failures(2) during 2 735 kV, most prominently (in around 50%) by bushing
others
Others
167.459 transformer years from 56 utilities in 21 countries failures[12]. From a recent Brazilian study, 19% results in an
1
was reported by CIGRE WG A2.37[6]. This study reports an explosion and 9% in fire[13].
overall failure rate of 0.57% (“one major failure per 200 0 Apart from the direct risk, also from financial and operational
<150 151-200 201-250 251-300 >300
transformers in a year”). In the table below a breakdown is side, the loss of a transformer is very costly and the lead time
given regarding voltage class: Figure 6 - Failure cause analysis based on 964 major failures [6] MVA rating to move from order to complete installation is anywhere
between 6 and 24 months.
One underwriter[14] quotes: “Power transformers have long
Figure 8 - Failure rate of converter transformers by MVA unit
69 ≤ kV < 100 100 ≤ kV < 200 200 ≤ kV < 300 300 ≤ kV < 500 500 ≤ kV < 700 kV ≥ 700 all
rating (LCC projects)[10]
been a major underwriting concern. Failure of a single unit
can result in widespread loss of service with considerable lost
0.93% 0.44% 0.55% 0.75% 0.54% 0.37% 0.57%
revenue as well as replacement and other collateral costs.
Table 1 - Failure rates of transformers in the CIGRE 2015 study From a compilation of failure data of various key As an object class, transformers have for decades been
components of DC installations shows that the converter ranked in the top five by equipment insurers in terms of
transformer is responsible for 52% of the downtime of claims. One insurance company recently reported 25
More detailed analyses on failed subcomponent, failure that their failure rate is significantly higher than that of DC stations, followed by the DC transmission line (28%), transformer claims during one fiscal year. The largest
mode and failure cause reveals the following: conventional station transformers, between 1.6 – 5.4%, see figure 9. transformer loss on record occurred at a power plant, leading
depending on their size, see figure 8, compared to the 0.6 % to business interruption costs of more than $86 million.
n Ageing and external short circuit are the largest known of AC transformers. The failure rate increases as the unit Had this event, which took place in 2000, occurred today,
failure causes (12.3%, 11.6% resp.), see figure 6. rating increases. Higher rating presents a challenge of costs would have been much higher”.
n Windings are the most common failure location (40%), keeping the dimensions as small as possibe for Transmission
Transmission line AC
AC trafo
trafo
line 28%
28%
followed by tapchanger (27%) and HV bushing (14%), transportation reasons and at the same time meet all the AC
AC excexc
Valve
Valve
tr tr
In a five year international survey, covering 94 cases with
see figure 7. electrical clearances and provide sufficient cooling of all Control & protection
Control&Protection an average loss of 183 MVA of capacity per incident, a total
DC
DC equipment
Equipment
n Mechanical failures account for over 20% of all failures, the parts. Other
Other claim of 287M US$ is reported[15]. In figure 10 this is
Transmission
Transmissionline line
the second largest after dielectric failures (36%). AC trafo
quantified in detail. By far the biggest claim come from loss
OtherOther
% AC trafo 52%
n The failure rates of GSU units is in all voltage classes 1% 52% of GSUs (71%) followed by industrial transformer losses
DC equipment
DC Equipment
higher than of substation transformers. 3%
3% (20%) and utility application (7%). The loss per MVA was
Control&Protection
Control & capitalized at US$ 9000. Further narrowing this down,
3%
EPRI (USA) maintains the IDB database[7] that began to protection 3% the major causes of failure (capitalized at 150M US$) are
Valve
be populated in 2006. Meanwhile, it contains more than winding
Winding Valve 6% “insulation failures”, that include defective installation, short
6%

20.000 US power transformers. One result that clearly stands tapchanger

Tap changer AC exc tr
AC7%
exc tr 7% circuit and insulation deterioration.
out is that “inadequate short-circuit strength” is by far the bushings
Bushings
largest failure cause, with 20% of the total of 654 clearly 160
other
Other business interruption
identified failures. The high-voltage winding is reported to Figure 9 - Breakdown of components of HVDC stations,
140 damage
be by far the most probably failure location (at 45% of the responsible for downtime
120 propery damage

Million US$
1112 identified failed components). 100
4.2 Consequences of transformer failures 80
Converter transformers supplying DC links are of a special
60
design. In several publications, their reliability (in LCC based Figure 7 - Failure location analysis based on 675 major failures Power transformers are considered as critical equipment
40
DC projects) is addressed[8, 9, 10]. It is striking to observe for U ≥ 100 kV[6] because of the large quantity of oil in contact with
20
high-voltage/high-current elements. Experience has shown
0
an increasing number of transformer explosions and fires in 1997 1998 1999 2000 2001
all types of power plants worldwide[11].
Year
(1)
In such studies, failure rate is usually defined as: (n1+n2+ … +ni) Figure 10 - Claimed losses per year in property damage and
(N1+N2+ ... +Ni ).T
. 100%; business interruption due to transformer damage[15]
with ni = number of failures in the i-th year, Ni = number of transformers operating in the i-th year, T = reference period (normally one year)
(2)
Any situation which required removal from service longer than 7 days, requiring opening of transformer or tapchanger tank, or bushing exchange.
12 ENERGY Power transformers need real power to survive a short circuit Power transformers need real power to survive a short circuit ENERGY 13

5 SHORT-CIRCUIT WITHSTAND VERIFICATION

5.1 By design review 5.2 By full-power testing

One of the methods for purchasers to assess the short-circuit n it is manufactured according to the same QA/QC In short-circuit testing, the transformer is subjected to the Test experience shows that, in spite of the ubiquitous
current withstand capability of transformers is to conduct practices; actual short-circuit current. Tests are performed in every use of simulation programs in the design of transformers,
a design review, based on calculation results only. CIGRE n the margins for short-circuit strength of both designs phase, each one to be subjected to the full asymmetrical still around 25% of the transformers do not pass the
issued guidelines on this method[16] that are implemented in overlap. current, one time at minimum tap position, one at standardized short-circuit test, see section 9.
the international standard IEC 60076-5, annex A[17]. nominal and one at maximum, during 0.25 s of current. The large costs and the possible delay of projects that results
This annex is an informative document giving guidelines For the evaluation, the parameters in 5.1d, e are tabulated In total, nine tests in a three-phase transformer. After every at a failure to pass makes it clear that manufacturers do
only, and it is not a standard by itself. A design review is together with the reference ones. The design under review is test, the short-circuit reactance is measured. Increase of everything they can to ensure a positive test. The fact that
based upon the following data: considered capable to withstand the dynamic effects of the short-circuit reactance beyond a certain value(3) , laid down nevertheless 25% fails to pass illustrates that the result of
a) Electromagnetic design data sheets; short circuit on condition that none of the force and stress in IEC 60076-5, is an indication of unacceptable winding the test cannot be predicted beforehand by calculation.
b) drawings/sketches of the winding and insulation parameters exceed the corresponding ones calculated for deformation and leads to a negative test result.
arrangement with indication of the types of material; the tested reference transformer by 20% (and in case of As part of the certification procedure, a detailed The great step forward that calculation methods have made
c) calculation of the (asymmetrical and symmetrical) three parameter values by 10%). out-of-tank active part inspection must be carried out after in last decades, together with the increased emphasis on
short-circuit current values, covering the possible types having performed the test, followed by a repetition of the cost reduction, can lead to a practice of designing close to
of fault and tap positions; 5.1.2 Checking against manufacturer’s design rules dielectric tests at 100% of the specified voltage level. the margin, whereas in the past – due to a greater uncertainty
d) calculation of the main forces, with a possibility to for short-circuit strength After passing all the requirement, a type test certificate is and lower cost pressure – a larger safety margin was built in.
simplify complicated configurations. Radial and axial In this method, evaluation is by checking the force and stress issued, see figure 11. This suggests that advanced design methods not
forces on the windings are required, as well as end thrust parameters against the manufacturer’s design rules for automatically guarantee a more reliable product.
forces on windings, on exits leads and on clamping parts, short-circuit strength.
in total 5 parameters; These rules shall be based on a “solid experimental basis”, The highest degree of reliability with respect to short-circuit
e) main tensile and compressive stresses on windings, a number of short-circuit withstand tests of actual withstand capability is through full-scale short-circuit
conductors, spacers, paper insulation, stack structures, transformers or the outcome of tests performed on withstand testing in accordance with the international
press rings and tie rods. In total 11 parameters have to representative transformer models combined to any indirect standards.
be quantified (for core-type transformers); supportive evidence based on long duration, trouble free
f) drawings of the support and clamping structures; operation of a number of transformers in the field. Short-circuit tests do not reduce the life-time of a well-
g) instructions for QA/QC concerning materials and In order to substantiate this, the manufacturer has to submit designed transformer. Reputable manufactures agree on
manufacturing activities; the following data: the fact that a properly designed transformer with enough
h) checks regarding external components, eg. bushings. margin to handle the electro-dynamic stresses, the effect of
n a list of transformers built that were subjected to the short-circuit stress will be that the windings undergo a
In the design evaluation of the transformer, two alternative short-circuit tests, including its nameplate data; certain settling[18]. The effect of the settling is that the
methods can be adopted: comparison with a reference, n results of these tests and possible impacts on design stiffness of the windings increases and this is visible in a
short-circuit tested transformer or checking against rules; small variation (if any) in the reactance values measured
manufacturer’s own design rules for short-circuit strength. n the technical standards of short-circuit strength, used in
Figure 11 - Example of a KEMA Type Test certificate of between the first tests, but becoming smaller or nihil at the
regular design as well as production activity; short-circuit performance last tests, see section 10.1.1. Such a transformer is even
5.1.1 Comparison with a short-circuit tested reference n service records and in-field failure rates regarding stronger after the short-circuit tests than before, and can be
transformer short-circuit performance and units produced; Short-circuit testing is considered a better means of put safely in service.
This method describes comparison with a reference ascertaining the real performance of equipment at Large utilities in for example Italy, Canada, Turkey have not
transformer that passed short-circuit tests successfully on For the evaluation, the forces and stress parameter values short-circuit, since such a test demonstrates that both seen any service problems on the units which have been
the condition that: are to be compared with the critical values that the construction and design are adequate[4]. This is recognized short-circuit tested[4]. Chinese experience[19] shows that
manufacturer has adopted in his design practice. These data in an IEC 60076-5 Annex B that states (for transformers 40 transformers (110 - 220 kV) put into service after passing
n The design under review can be considered similar to the are to be collected in a table that allows direct comparison > 2.5 MVA): “for the purpose of evaluation the unit under short-circuit tests, function without problem during the
reference transformer. In IEC 60076-5, Annex B, “similarity” between “actual”, “allowable” and “critical” values. consideration may be simultaneously compared with a monitoring period from a few months to more than 5 years.
is defined by having an absorbed power in a range The transformer is considered able to withstand the dynamic limited number of transformers that have passed the
between 30% and 130%, and exceeding forces and stresses on condition that none of the force and stress short-circuit test successfully and match most - but not all -
stresses by not more than 120%; figures exceed the allowable force of stress and remains the characteristics considered in Annex A”[17].
n it is designed using the same calculation methods and below 80% of the critical stresses.
withstand criteria;
(3)
For transformers of category III (above 100 MVA) this value is 1%; for smaller transformer this number depends on the winding geometry.
14 ENERGY Power transformers need real power to survive a short circuit
5.3 Limitations of the design review
approach
CIGRE studies show that calculation methods are an
indispensable tool in the design phase of electrical
equipment and modern, multi-physics finite-element 3D
simulation tools can predict internal stresses. However,
they cannot be applied for the verification of equipment
performance verification[20]. This also applies to power
transformers.
For a number of reasons, listed below, the value of design
review is limited, and generally insufficient as replacement of
short-circuit withstand verification by laboratory testing.
5.3.1 Simulation tools are a simplification of the reality
By means of simplifications it is possible to calculate
approximately the highest stresses that occur. In order to
evaluate these stresses, it is necessary to compare these
with critical values (see section 5.1.2) of forces and stresses.
This implies that also data of failed short-circuit tests shall
be available, which can only be the case when a database
of solid testing experience is at hand. More generally: It is
very important that the manufacturer can demonstrate
that the calculations and considerations of the design
characteristics and manufacturing practices used are
based upon and validated by short-circuit tests.
The most obvious simplifications are: disregard of the
influence of the other phases on the magnetic fields in a
certain winding, calculation of the forces and stresses for
the current peaks and RMS values only, consideration of
the windings as rigid (i.e. without any flexibility or settling
effects), consideration of the windings as rotational
symmetrical, etc.
The most onerous circumstances can be selected by such
calculations and simulations. An exclusive focus on the
copper yield strength is not sufficient.
Therefore, at the design stage extra, difficult to quantify,
margins need to be implemented to cover such effects.
Another simplification is the assumption of rotational
symmetry of the windings. IEC 60076-5 does not take into
account deviations from the ideal cylindrical shape, not from
discontinuities both from design and manufacturing point of
view[21].
Power transformers need real power to survive a short circuit ENERGY 15
From the design point of view: n Stability of the axial support of windings. The axial 5.3.4 Design review does not cover material and
n All kind of magnetic field deviations from rotational pressure applied on the individual windings of the production deficiencies
symmetry such as yoke, unwound leg, other phase and entire winding block is sensitive for the shrinkage of the Natural variations in properties of material, quality
tank wall; insulation material, because the height of the winding is assurance, workmen’s skills etc., especially in the less
n all kind of mechanical discontinuities in the cylindrical partly made up by copper and partly by insulation experienced companies, cannot be taken into account, nor
shape, such as transpositions inside the winding, material. The shrinkage of copper is fixed and can any design review reveal deficiencies resulting from this.
crossovers in disk windings, pitch of a winding and negligible compared to insulation material. Difference From CIGRE studies[23], it became clear that models can
insulation supports inside the winding; in shrinkage of the insulation material blocks in the overall predict stresses but the prediction of yield to stresses is
n type of winding with parallel conductors and current winding block assembly can lead to different heights of another matter, since inhomogeneity of material determines
division; the individual windings in the commonly pressed where and when a disruptive event (breakdown, rupture,
n friction between turns and impact on insulation; winding block and therefore different pressure on the burst etc.) occurs.
n the interaction between phases, both magnetic and individual windings. During short circuit it can
mechanical; occur that windings will be over-pressed resulting in 5.3.5 Many failures occur in other than the “design
n dynamic aspects, such as mechanical resonance in axial damage of conductors (crossovers or transpositions) or reviewed” sub-components
direction and the influence of the oil. that windings are under-pressed and insulation blocks will Complicated ‘secondary’ physical phenomena like shock
come loose and may come out destroying the mechanical waves in oil, shocks and vibrations (leading to untimely
From the manufacturing point of view: stability of the winding (block) and resulting in an falling-off of buchholz relays or damage to ancillary
n Quality of cured epoxy (bonding strength as a function electrical and/or mechanical collapse of the winding. equipment - bushings, tap-changers etc.) are normally
of operating temperature) and the quality of the epoxy A large scale study showed that aged transformers are not considered at all in calculation methods and design
bonding process; more prone to short-circuit winding deformation because review.
n roundness of winding; of diminishing winding fastening strength due to
n dimensional tolerances; insulation shrinkage and pieces of insulating material There is no quality control on the performance for the
n balancing windings on axial symmetry; disappear or become displaced[22]. design calculations. The theoretical evaluation of the
n paper and insulation material shrinkage after the drying ability to withstand the dynamic effects of short-circuit will
process; 5.3.3 The design review approach is static whereas the in general be performed by non-specialized purchasers
n final pressing and tightening procedure before assembly phenomena are dynamic and/or (independent) skilled and/or unskilled third party
in the tank. The electrodynamic stresses are varying in time and are inspectors or consultants. No defined procedure is available
varying spatially. Complicated, three-dimensional time how to judge the theoretical evaluation and how it can be
5.3.2 The list of reviewed sub-components is not complete varying fields and forces stressing various assessed. Statement letters provided are in general unique
The design review should check all critical force and stress non-homogeneous and non-linear structures and to the inspecting body. The theoretical evaluation procedure
values, but it is impossible to ensure that all necessary materials cannot be covered adequately by simulation. leaves opportunities for “shopping” to get “somewhere” a
aspects are considered. One omission is that calculations The transient mechanical behaviour of the windings (natural positive statement letter.
related to lead supports and connections to bushings are frequencies, damping, non-linear effects) and the production
not requested. tolerances (tolerances in materials, processing, assembling, The definition of a “similar” design (in IEC 60076-5
Some aspects that are not (or insufficiently) considered etc.) make it very complicated to exactly simulate the annex B) is too ambiguous and may lead to confusion.
and/or deviate from (not validated) assumptions during the winding’s behaviour. This is a difficult point since purchasers generally do not
design stage related to the short-circuit withstand capability As the dynamic behaviour of the windings during a have the specialist’s knowledge to decide that an available
are: short circuit is not mentioned in design review no aspects of short-circuit tested transformer can be considered as
n cross overs of turns (inside the winding) and transpositions inter-winding (dynamic) oil flow or tank stress calculations “similar” to the transformer under evaluation.
of parallel conductors (inside the winding); are considered. Moreover, design review does not assess
n exit leads of the windings, fixation to prevent movement the fact that successive short circuits lead to a sequence of On the other hand, a short-circuit tested transformer is
and friction (wear of insulation) of exit lead; events. Typically, after the first application of a short circuit certified by an accredited test laboratory which covers
n support of cleats and leads and connections to in a test, a slight reactance increase is observed (see figure everything. The certificates, see figure 11, provided by
tapchanger; 25). Successive movement of the winding, core, clamping accredited test laboratories are uniform in style and
n support of leads to bushings; arrangements and leads due to accumulated stresses during acknowledged everywhere without any restriction.
n stability of the radial support of windings, for example a series of short-circuit tests, leading to a failure after a
spacers used during the winding of the coil, untreated, certain number shots cannot be simulated/calculated.
dried, dried and oil impregnated;
16 ENERGY Power transformers need real power to survive a short circuit Power transformers need real power to survive a short circuit ENERGY 17

6 UTILITY CONSIDERATIONS

5.3.6 Conclusion
The short-circuit withstand capability of a transformer
cannot be presented in a set of equations and summarized
in a spreadsheet table as in the IEC 60076-5 Annex A.
Verification based on design review is an oversimplification
of the existing knowledge and it neglects the vast design
and manufacturing experiences of the different
manufacturers[18].
Theoretical evaluation seemed to be introduced because of
short-circuit testing limitations in the past, as an alternative
in case short-circuit testing was not possible. In due time,
test laboratories have increased their short-circuit power and
can cover the vast majority of power transformer ratings, see
section 8.
According to IEC 60076-5 it might seem that design review
unrestrictedly can be chosen as an equivalent for short-
circuit testing, but this was not the original purpose of the
theoretical evaluations that are behind a design review.
There is very limited information on direct evidence of
failures in service after short-circuit withstand evaluation.
From a survey conducted by CIGRE SC12 (covering 121460
As such, testing of models plays an important role in linking
theory and practice and they assist in material selection,
choice of geometric arrangements and manufacturing
procedures.
The contribution that model testing can make is as follows[26]:
n Evaluation of the dynamic characteristics of materials;
n definition of critical stress levels which can lead to failure;
n validation or calibration of analytical and numerical
modelling methods;
n investigation of the effects of material properties and performance before the year 2000 led to the adoption of
manufacturing process variables;
n examination of the mechanical effects of repeated short As an example: short-circuit failure rate was 0.4% in
circuits.
Experiments on models are insufficient to warrant short-
circuit withstand capability of a complete transformer.
Calculations and tests on models alone are not sufficient
to renounce the short-circuit capability of a transformer.
In the past, utilities mainly relied on the selection of trusted manufacturers to secure the
short-circuit capability of their transformers[5]. However, in recent times the situation has
changed, with more and more users asking for tests on critical units or prototypes of series
of identical units. This may be because a long term confidence between purchaser and
manufacturer is less and less achievable because of increasing deregulation[4].
The high failure rate in service due to poor short-circuit Other major utilities[19, 29,34, 35] require suppliers to pass a
learning path towards a successful design through full-power
short-circuit testing as a method for quality improvement. short-circuit testing thereby using it as an essential and
successful tool for quality improvement.
France[28], 0.35% in Italy[29], 1.2% in Turkey[29]. In China 84%
of all internal failures was due to short circuit [19] and in The considerations above have led to a rapid increase in
India over 80% of the failures were caused by winding testing of large power transformers in the world’s test
displacement[30]. facilities. In figure 13, the number of short-circuit tests of
The experience achieved through tests, and sometimes transformers ≥ 25 MVA is plotted vs. time, performed at
even if the test result is negative, turns out to be a precious the KEMA Laboratories. In this figure, a distinction is made
source of information and knowledge. Critical spots can between the test results showing no problem at short-circuit
be detected and simple measures be set up, often with tests and tests that lead to a problem, see chapter 7.
negligible impact on costs, which result in being highly
beneficial also with respect to long term reliability.
40
A significant and positive influence of short-circuit testing 35 initially not OK

transformer years in the period 1993-1997) 15 failures on the reduction of the rate of faults in a large overall 30 initially OK

attributed to short-circuit were identified. In 5 cases (33%) system during many decades is reported[18, 31, 32]. 25
20
design reviews had been performed, whereas none of the In the 1960s it was reported that in the US from 50% - 85% of
15
failed units (or similar designs) was short-circuit tested[5]. the transformer failures were due to short-circuit withstand 10
deficiency, whereas in 1979 this percentage was estimated 5

5.4 By testing of scale-models (“mock-ups”) to have dropped to 20% - and remains to this rather high 0

Sometimes testing of downscaled models (or “mock-up”) of
actual transformers is practiced, mainly because the actual
transformer’s capacity is too big for any laboratory.
Examples are large GSU’s for nuclear power plants[24] and
phase-shifting transformers[25]. The results of such tests are
useful to obtain input data for models, to validate certain
calculation methods, or to study relevant phenomena
because such “research oriented” models allow additional
observations, such as strain gauge measurements or visual
observation of moving parts through observation ports[22].
In figure 12 a test on a mock-up is shown.
Figure 12 - Testing of a mock-up of a 570 MVA transformer[24]
One might state a parallel with the dielectric withstand
capability[27]: “Although the design of the insulation structure
is based on the calculation of voltage distributions along
the windings, also supplemented by thousands of tests and
measurements on simple insulating structures or on more
complex models, nobody would accept to abstain from
performing dielectric tests on a transformer”.
level, see section 4.1.
A major utility’s experience indicates that premature O&M
failures are occurring due to accelerated ageing and/or
weakening of short-circuit withstand capability of
transformer due to repeated short circuits in the underlying
system[33]. The number of through faults seen by their
transformers is high and has a cumulative effect on the
mechanical weakening of the winding supports and
insulation, and it increases the probability of premature
failure of the transformer.
To check that manufacturers ensure short-circuit withstand
capability as per network requirement/specification of their
transformers during design/manufacturing, this utility
introduces short-circuit withstand test for at least one
transformer for each manufacturer which will be type tested
and benchmarked for future projects.
'96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 '10 '11 '12 '13 '14 '15
Figure 13 - Number of power transformer (≥ 25 MVA) short-circuit
tests performed 1996-2015 at KEMA Laboratories.
Indicated is the fraction of initially “not OK” and “OK”.
Of course it is impractical and even impossible to verify
short-circuit withstand of all transformers by testing.
In the following cases, purchasers have good reasons to
specify short-circuit withstand tests[4]:
n In case of utilities, whose transformer fleet mainly consists
of standard units. If several identical units are going to be
purchased at one time, or if it deals with units to be
manufactured on the basis of new or revised designs,
it may be good practice to submit the first unit of that
series to a short-circuit test, and then, only after the
successful completion of that test, release the following
units for production: the “learning path” approach;
18 ENERGY Power transformers need real power to survive a short circuit Power transformers need real power to survive a short circuit ENERGY 19

7 FEATURES OF KEMA LABORATORIES

n in case of transformers performing a key function in Short-circuit test laboratories for power transformers draw n Making switches (MS in figure 14) that switch in the
the network, for which the reliability issue is of prime their short-circuit power either directly from the power grid short-circuit current. Also these switches can be much
importance. This applies, for example, to key GSU (grid-supplied lab, see figure 14, upper) or from generators more accurately controlled at medium voltage level.
transformers (already having a relatively high failure rate, (generator supplied lab, see figure 14, lower). When high-voltage breakers are used (HV CB right side
see section 4.1) and unit auxiliary power supply in figure 14 upper), the statistical nature of the pre-strike
transformers installed in nuclear power stations. upon closing makes it difficult to close at the exact
n strategic interconnection (auto)transformers located SW T HV CB L voltage phase that will result in a full asymmetrical current
in strategic stations or huge consumption centers as required by the standard. This may lead to unnecessary

HV Grid
(server parks); trial testing with insufficient asymmetry levels.
n transformers requiring a special design, for example
entailing a primary winding and two equally-sized TO At present, KEMA Laboratories uses six generators as its
secondary windings, each with half power rating Grid supplied lab power source. These operate at a voltage 15 – 17 kV.
HV CB
compared to the primary (axially split winding type[4]);
n transformers to be installed in networks with high fault A number of circuit topologies have been designed for

HV Grid
incidence, expected to face a heavy duty operation M optimum conformity with the service situation:
consisting, for example, of tens of short circuits per year. MB MS L T
Any forecast of a sharp increase of the short-circuit power G n Three-phase tests. Three-phase transformers should
level in the network should also receive due attention with preferably be tested three-phase. In case the voltage
respect to this issue; range is not sufficient or the short-circuit power is not
TO
n track feeding transformers; D enough, the testing authorities may use single-phase
n transformers having a low short-circuit impedance testing instead of three-phase testing. As in three-phase
and/or installed in solidly earthed systems. This is often Generator supplied lab testing, at each test one phase is subjected to the
the case in USA, where transformers often have lower specified (peak) current value, see figure 15 where the
short-circuit impedance than in Europe. asymmetrical peak is applied to the upper phase. During
Figure 14 - Lay-out of grid- and generator supplied test later tests the other phases are subjected to the required
laboratories. T: laboratory transformer, TO: transformer current.
under test; HV CB: high-voltage circuit breaker,
MB: master breaker, MS: making switch, G: generator,
M: motor, L: reactor

Grid supplied labs have the advantage that very high power
Moment
can be drawn from the grid - when the grid operator gives of Full asymm. current in one phase
permission - but regarding security for the transformer maximum
under test, generator supplied labs offer lower risk because stress Current in three-phase test
switching occurs at the output voltage level of the generator,
usually around 15 kV:

n Master breakers (MB in figure 14), that need to interrupt Current in 1.5 phase test
current, can be designed to act very fast at this voltage
Current in three-phase test
level. Very fast interruption of short-circuit current, in
combination with a special protection system prevents
major damage to the transformer under test in case of an
initial failure, like an internal arc, or fire because of oil spill.
Current in 1.5 phase test
In our case, the master breakers can de-energize the
circuit within 6 ms. In grid supplied laboratories, master Figure 15 - 1.5 Phase test method showing that at the moment of
breakers have to operate at high-voltage, which implies maximum stress, the current in all phases is identical
that breaking times of several tens of ms are unavoidable. to a full three-phase condition

20 ENERGY Power transformers need real power to survive a short circuit Power transformers need real power to survive a short circuit ENERGY 21

8 LABORATORY CAPACITY

n 1.5 phase tests. With the single-phase method, known In order to be prepared for testing of very large transformers In this figure, tests of the past years are plotted using the
Supply Make switch Transformer under test as 1.5 phase method, the phase under test is connected up to the 800 kV class, investments are made for increasing rated voltage and rated power as coordinates, distinguished
l in series with the other two phases, which are the short-circuit capacity of the author’s laboratory by 50%. by the test method. In addition, several new possibilities are
connected in parallel, figure 16. The RMS currents in the This is achieved by installing two additional generators and entered as an indication. The actual feasibility for testing,
1,5Uphase two parallel connected phases are 50% of the specified four short-circuit transformers, making available a direct choice of test method and test circuit will ultimately depend
1/2l
three-phase value. Figure 15 illustrates this condition. test-power of approx. 15000 MVA. Figure 17 shows one of on parameters such as short-circuit reactance of transformer
The evolving stresses in the two parallel connected phases the new generators. to be tested, the specific grid short-circuit power etc.
1/2l are in this case lower than in the case of three-phase tests,
but at the most critical moment, at the asymmetrical Key assets of any laboratory are the short-circuit transformers
current peak in the fully stressed phase, all currents are that need to adapt the supply voltage to the required
Figure 16 - Test-circuit for 1.5 phase short-circuit tests momentarily identical to the situation of a three-phase voltage of the transformer under test.
test. In this topology only half the power is required from
the test station compared to three-phase testing. Special For the laboratory expansion project, four new short-circuit
attention has to be given to the neutral of the Y-connected transformers have been designed, tested and installed.
windings, which in many other test stations is at a voltage They will be used in addition to the six existing laboratory
level of 0.5 pu. In our practice this remains at ground transformers (2x14 kV/2x36 kV, 250 kV isolation to earth).
potential. The ten transformers in series realize a direct testing voltage
of 550 kV. This implies that having the laboratory extension
n Single phase tests. In case the test station’s power is not available, single phase transformers in the 500 – 600 MVA
enough for using the 1.5 phase method, a real single- class can now be short-circuit tested (up to 1500-1800 MVA
phase method could be applied, like the method used for three phase). Voltage wise, transformers with rated
single-phase transformers. In that case the terminals of the (single phase) voltages up to 800 kV can be tested as well,
other phases are open. Figure 17 - New generator, 2500 MVA 16.7 - 60 Hz which has indeed be demonstrated, see figure 19, showing
an 800 kV transformer after a short-circuit test.

Technically, the impact of the extension is plotted in

figure 18 showing the technical feasibility for testing of
the various voltage and power ratings of transformers
before and after the extension of the laboratory.

900

800 1 phase
1.5 phase
700
3 phase
600

Voltage (kV)
500

400

300 New
capability
200
Tested Figure 19 - An 800 kV transformer under test
100
2008-2014
0
0 200 400 600 800 1000 1200

Power (MVA)

Figure 18 - Past and newly added test potential for power

transformers vs. rated power, voltage and test method
22 ENERGY Power transformers need real power to survive a short circuit
TEST RESULT STATISTICS OF POWER
9 TRANSFORMERS
Having 15000 MVA of installed short-circuit capacity (world’s
largest), the author’s laboratory has tested transformers up to
very high MVA and kV ratings. An evaluation has been made
of short-circuit tests performed in the 20 year period
1996-2015. The tests were performed in accordance with
IEC standard[5] or IEEE standards on transformers with rated
power up to 440 MVA and primary voltage up to 800 kV.
The population includes single-phase and three-phase
transformers, auto-transformers, step-up-, converter-,
railway-, auxiliary- and three-winding transformers, 16.7,
50 and 60 Hz transformers, YD-, DY-transformers and
YY autotransformers. The largest transformers tested are
334 MVA single-phase and 440 MVA three-phase.
In detail, the test-experience is as follows:
During the past 20 years, in total 297 times a test access for
a transformer larger than 25 MVA (258 transformers from
which 39 are re-tested) has been counted:
In 230 cases, the transformer showed no problem at the
test-site. These transformers initially passed the short-circuit
test. The final test-result is not always known because there
was in a limited number of cases inspector’s involvement in from inner windings; broken exit leads;
the subsequent routine tests and the visual inspection.
n In 67 cases a transformers showed a problem due 20-100
to short-circuit stresses that became immediately kV (rated)
apparent at the test site. Mostly, this problem was an conductor insulation;
unacceptable increase of short-circuit reactance due to Figure 21 - Initial power transformer failure rate for various
the short-circuit stress, but a range of other, immediately ranges of kV (primary) rating
evident problems also occurred;
n 39 transformers from the latter group had been re-tested
after modification in the factory and most did not show a is in the same order as the experience reported by another
problem at the test site at the re-test;
n In 7 cases, transformers after having experienced no
problem at the test site, did not pass the routine tests
and/or visual inspection after the tests.
From these results, an initial failure rate is defined as the ratio
of tests that resulted in failure to pass the test at first access
(67 times) and the total tests (297). Thus, the initial failure
rate is 23%.
In figures 20 and 21 results are shown, differentiated in both
power- and voltage class. The results suggest a tendency
of the highest initial failure rates for the highest ratings:
the failure rate of the largest transformers (> 300 kV
or > 200 MVA), around 100 tested, is in the range of 30%.
(4)
The data in this figure are compiled based on 4961 tested transformers, of which
1140 (23%) failed to pass. 400 are transformers > 40 MVA, of which 83 failed to pass.
This figure has indicative value only, since various time periods, testing practices and
testing purposes are considered.
Power transformers need real power to survive a short circuit ENERGY 23
10 FAILURE MODES
Commonly, the reason of not passing short-circuit tests is because the winding reactance
initially not OK
change (usually an increase) is larger than specified in the standards.
80 initially OK
Number of tests
60
Evidence of damage, as suggested by measured reactance Another item to be mentioned is the behaviour of buchholz-
40
variation is usually confirmed by visual inspection. relays, as quite often buchholz-relays operate unnecessary
20
In addition, other, most clearly recognizable defect are due to the vibrations that occur during short-circuit
evident directly at the test site or upon inspection of the conditions. Therefore, the behaviour of such relays is
0 internal parts. carefully monitored during testing and the observations
25-50 50-100 100-200 >200 are reported.
MVA (rated) A wide variety of defects are revealed such as:
On the other hand, in the cases (the vast majority) that
Figure 20 - Initial power transformer failure rate for various n Axial clamping system: looseness of force in axial the reactance change is within the tolerances set by the
ranges of MVA rating clamping, of axial compression force, of axial supporting standards, it is the author’s observation that (visual)
spacers and of top and bottom insulating blocks; inspection sometimes still leads to rejection of a certificate.
n windings: axial shift of windings, buckling, tilting of Visual inspection is necessary, because deformations and
100 initially not OK
conductors; displacements in supporting structures, clamping systems,
initially OK
80 n cable leads: mechanical movement, for instance from tap insulating materials, winding exit leads, external connections
Number of tests
60 changer to regulating windings; deformed or broken from the coils to the tap changer and within the on-load tap
leads, outward displacement and deformation of exit leads changer cannot be detected by the reactance measurements
40
only.
20 n insulation: crushed and damaged conductor insulation;
displacement of vertical oil-duct spacers; dielectric In addition, defects to the regulation winding, often not
0
100-200 200-300 300-400 >400 flashover across HV-winding or to the tank; displacement detectable by impedance change can only be confirmed by
of pressboard insulation; tank current due to damaged visual inspection[37]. The authors conclude that the reactance
variation is a very good tool to assess short-circuit withstand
n bushings: broken or cracked bushings (see figure 23), capability right after the short-circuit test. It can not be used
leading to oil spraying; as a ‘passed’ indicator, only as a ‘not-passed’.
n enclosure: spraying of oil (see figure 24), exhaust of
The failure rate at testing, observed by KEMA Laboratories hot gases, evaporated oil, measurement of current to Our experience with the short-circuit reactance
enclosure. measurements is that for power transformers a variation of
major test laboratory that reports a failure rate of 20 – 25% more than 1.0% indicates a large deformation in one or more
out of 20 units > 100 MVA[36]. Other sources state an overall windings. Also a gradually increasing variation during the
failure rate of 23% for a total of 3934 tests[3] and 21.4% short-circuit tests, although in total not more than 0.5% to
failure (2.5 – 100 MVA), 41.9% failure (> 100 MVA)[29]. 1.0%, indicates a progressive movement of conductors.
Test failure rates of 15 test-laboratories worldwide are Variations of the reactance values between the short-circuit
summarized in figure 22. tests in an unusual way are an indication of instability of the
windings.
70
<=2.5 MVA
60
Although the standards do not require (S)FRA measurement,
2.5-40 MVA
41-100 MVA
50
FRA measurements are regularly performed before and after
Test failure rate (%)
>100 MVA
40
30
short-circuit application. In some cases, a significant
20
difference can be observed[38]. Although a FRA pattern
10 Figure 23 Figure 24 contains in principle more information than a single
0 reactance measurement at a given frequency, it appeared
Figure 23 - Breaking of a bushing and oil spill difficult to correlate the observed shifts in the FRA patterns
Figure 24 - Oil spill upon short circuit application to visible internal deformations. More research is needed
Figure 22 - Test failure rates of test-laboratories worldwide to establish knowledge rules.
(data used from [4])4
24 ENERGY Power transformers need real power to survive a short circuit Power transformers need real power to survive a short circuit ENERGY 25

11 DISCUSSION

10.1.1 Observations from short-circuit tests Careful observation of the reactance evolution, measured Failures of transformers in service are well-known, but the Thus, it must be assumed that this actual (full) short-circuit
after each test, can reveal short-circuit current withstand reason for failure is often not clear[6, 39, 40]. current in service is normally smaller than the rated short-
In the case that a continuous increase of winding reactance deficiency of the regulation winding. One could wonder on what causes the high failure rate, circuit current for which the transformer is designed.
is observed during all tests, it is reasonable to assume (by According to IEC 60076-5, 9 tests have to be performed in observed by short-circuit test laboratories worldwide
extrapolation) that the transformer undergoes accelerated total (on a three-phase transformer) with 3 tests in minimum compared to the lower failure in service, found to be below Because of the expected future increase in system power,
ageing due to the short-circuit tests. In figure 25, such an tap position, 3 tests in nominal and 3 tests in maximum tap 1%[6, 41], mainly due to defects in windings. The main especially in rapidly developing countries, this situation may
example is given. Herein, the change in reactance after position. A (measured) example of this procedure is shown reason of this discrepancy must be the severity of the tests, change and the fact must be faced that during the life of the
successive short-circuit tests is plotted. The left graph in figure 26. As a result of this test, the observed short-circuit compared to actual service conditions. From an enquiry of ageing transformer, its withstand against short circuits will be
indicates a common failure in all three phases (after visual reactance change is very close to 1% (in phase 3). CIGRE WG 13.08[3] it can be concluded that on a statistical brought to the limit.
inspection this turns out to be deformation of the helical What is striking, however, is that the total reactance change basis, large power transformers have to face several full and
windings), whereas the right graph suggests a failure of the (after completion of all tests) in the minimum tap position many small short-circuits during their life, more precisely:
W phase only (which turned out to be a defect of the axial (0.1%) is much smaller than the value in the maximum the 90 percentile was estimated to be 4 full short-circuits in
clamping system). The other phases (U, V) will not after-test tap position (1%). Apparently, inclusion of the 25 years.
deteriorate due to high current stress. regulation winding brings the total reactance change from
0.1% to 1%. Given the fact that the contribution of the
regulation winding to the total externally measurable
TAP 5
Isc=82%
TAP 3
Isc=100%
TAP 1
Isc=118%
TAP 5
Isc=90%
TAP 3
Isc=100%
TAP 1
Isc=110% short-circuit reactance is only a fraction of the total
reactance, it is reasonable to expect that the reactance
Short-circuit reactance (0.5%/div)

12
phase U
change of the regulation winding alone is far larger than
CONCLUSIONS
0 0

1%. The effect of this was confirmed in several case in recent

phase V years, where visual inspection of the regulation winding
0 0 revealed unacceptable damage to this winding.

0 phase W0

0
new 1 2 3 4 5 6 0 1 2 3 4 5 6
n Short-circuit current leads to extreme mechanical forces in
new
test number 1,0
Phase U Phase V Phase W
transformers which need to be managed in the design by adequate
Phase U Phase V Phase W
under test under test under test under test under test under test
clamping and support of all relevant subcomponents, not only the
0,8
winding.
Figure 25 - Evolution of short-circuit reactance increase during n Short circuits are a major contributor to damage to power
two test series 0,6
transformers in service (up to 20% of the major failures).
Left: transformer with common failure in all phases phase 3
n Short-circuit testing is the only complete verification method of
Right: transformers with a failure in phase W only 0,4
phase 1 short-circuit withstand capability of power transformers.
KEMA Laboratories are now ready to short-circuit test power
0,2 transformers with rated voltage up to 800 kV and power up to
phase 2 test number 1000 MVA (based on three-phase transformer banks). The actual
0,0 limit depends on many variables.
1 2 3 4 5 6 7 8 9 10 tap11
min.
n Design review is based on calculation results of idealized,
minimum tap rated tap maximum tap after tests
homogeneous structures, it does not cover transient phenomena, it
excludes a number of key subcomponents and it is not embedded
Figure 26 - Evolution of reactance change in short-circuit test in a strict quality surveillance system. It is not a good representation
that reveal large impedance change of regulation of the reality and it cannot serve as a complete and therefore as a
winding reliable verification tool.
n Failure of passing a short-circuit test is in the 20 – 30% range as
confirmed by major test laboratories worldwide.
26 ENERGY Power transformers need real power to survive a short circuit
13 REFERENCES
[1] Blackout tracker, United States Annual Report 2013, Eaton, 2013
[2] Blackout tracker, Jahresreport 2014, Deutschland, Österreich, Schweiz, Eaton, 2014
[3] CIGRE WG 13.08: “Life Management of Circuit Breakers”, CIGRE Technical Brochure 165, August 2000
[4] G. Bertagnolli, “Power Transformers & Short Circuits. Evaluation of short-circuit performance of power transformers”,
Book printed on behalf of ABB Ltd. Transformers, Switzerland, 2014
[5] CIGRE WG 12.19: “The Short-Circuit Performance of Power Transformers”, CIGRE Technical Brochure 209, August 2002
[6] CIGRE WGA2.37, “Transformer Reliability Survey”, CIGRE Technical Brochure 642, December 2015
[7] B. Desai and M. Lebow, “Needed: ASAP Approach”, Power and Energy Magazine, IEEE, vol.8, no.6, pp. 53-60,
November 2010
[8 ] CIGRE Joint TaskForce B4.04/A2-1, Technical Brochure 240: “Analysis of HVDC Thyristor Converter Transformer
Performance”, 2004
[9] CIGRE WG A2/B4.28, Technical Brochure 406: “HVDC Converter Transformers. Design review, test procedures,
ageing evaluation and Reliability in Service”, 2010
[10] CIGRE WG B4.04, Technical Brochure 617, “HVDC LCC Converter Transformers Converter Transformer Failure Survey
Results From 2003 To 2012”, 2015
[11] H.-P. Berg, N. Fritze, “Reliability of Main Transformers”, RT&A # 01 (20), (Vol.2), March 2011
[12] Foata, M., “Transformer fire risk and mitigation”, CIGRE A2 Transformers Session, 2010
[13] G. Bastos, M. A. Marin et al., “AVALIAÇÃO DO DESEMPENHO DE TRANSFORMADORES DE POTÊNCIA E REATORES
NO SISTEMA ELÉTRICO BRASILEIRO,” Cigre Brazil, GTA2-02, 2014.
[14] http://www.willis.com/documents/publications/services/property/Transformer_Problem.pdf
[15] W.H. Bartley, “Analysis of Transformer Failures”, International Association of Engineering Insurers, 36th Annual
Conference – Stockholm, 2003
[16] CIGRE WG 12.22: “Guidelines for Consulting Design Reviews for Transformers 100 MVA and 123 kV and above”,
CIGRE Technical Brochure 204, August 2002
[17] IEC standard 60076-5: Power Transformers - Part 5: “Ability to withstand short-circuit”, 2006
[18] H. Gupta, N.N. Misra, “Concerns about Short-Circuit Tests”, TrafoTech 2002, pp. 121-25, Mumbai, India, January 2002
[19] Y. He, M. Wang, “The Transformer Short-Circuit Test and the High-Power Laboratory in China - the Past, Present
and Future”, IEEE Electrical Insulation Magazine, Jan/Feb. 2004, vol.20 no.1, pp 15 - 19
[20] B. Verhoeven, R.P.P. Smeets, “Secure Verification of T&D Equipment needs Laboratory Testing”, position paper
KEMA Laboratories, https://www.dnvgl.com/energy/publications/download/KEMA_Laboratories_Position_Paper.html, 2015
[21] K. Spoorenberg, “The relevance of insulation supports in large power transformers for short-circuit withstand”, on behalf
of Royal SMIT, Transformer User’s Meeting, 2016
[22] T. Kobayashi, Y. Shirasaka, Y. Ebisawa, H. Murakami: “Expected Life and Maintenance Strategy for Transformers”,
CIGRE 6th Southern Africa Regional Conf., paper P102, 2009
[23] CIGRE WG A3.20, ‘Simulation and Calculations as Verification Tools for Design and Performance Assessment of
High-Voltage Equipment’, Electra 234, pp. 13-18, October 2007
[24] M. Cuesto, J. Porrero, M. Muñoz, J. Camara, P. Hurlet, A. Tanguy, M. Ryadi, “Short Circuit Design Conception and
Validation ``of a 570 MVA, single-phase GSU-Transformer by SC-Withstand Tests on a Mock-up Unit”, CIGRE Conference,
paper A2-206, 2016
[25] IEC 62032 (2005), IEEE C57.135 (2001) “Guide for the application, specification and testing of phase-shifting transformers”,
2005, 2001
[26] W.J. McNutt, Y. Tournier, G. Prieniger, J. Wendlin, “Assurance of transformer short-circuit withstand capability – a review of
existing knowledge, experience, opinions”, CIGRE Conference, paper 12-07, 1980
Power transformers need real power to survive a short circuit ENERGY 27
[27] M. Gallay, P. Boussier, J. Delhaye, A. Fabbro, F. Sandoz, “Short-circuit withstand capability of large transformers – influence
on transformer design and construction, resulting from experience acquired during tests”, CIGRE Conference, 12-08, 1980
[28] J.P. Arthaud, E. Chemin, D. Girardot, P. Macor, T. Ngueguege, J.C. Ribaud, G. Robert, “The short-circuit resistance of
transformers: the feedback in France based on tests, service and calculation approaches”, CIGRE Conference, 12-102, 2000
[29] L. Bergonzi, G. Bertagnolli, G. Cannevale, G. Caprio, F. Iliceto, B. Dill, O. Gülyesil, “Power Transformer reliability –
Technical and Economic Issues related to the short-circuit performance of power transformers”, CIGRE Conference,
12-107, 2000
[30] Seminar HV transformer failures: rising trend, causes and remedies, 1998
[31] G. Macor et al., “The Short-Circuit Resistance of Transformers: The Feedback in France Based on Tests, Service and
Calculation Approaches”, CIGRE Conference, paper 12-102, 2000
[32] N.V.C. Sastry, H. Gupta, “Short-Circuit Test on EHV Transformers”, 9th Int. Conf. on Short-Circuit Currents in Power
Systems, Cracow 2000
[33] V.K. Bhaskar, A. Gupta, G. Agrawal, A. Singh, “Power¬grid Experience in Transformer Life Management and
Condition Assessment”, TrafoTech, Session V, paper 6, 2010
[34] P. Riffon, “Short-circuit strength and short-circuit testing of power and distribution transformers”, IEEE/PES Transformer
Committee meeting, 2008
[35] E. Perez-Moreno, J. Gortazar Rubial, “Importance of the Design Review and the Short-Circuit Tests in Power Transformers.
The experience in Gas Natural Fenosa”, Advanced Research Workshop on Transformers, 2010
[36] M. de Nigris, “Checking the Short-Circuit Withstand of Power Transformers: CESI perspective”, KEMA Int. Symp. on
High Power Testing, Nov. 2004
[37] R.P.P. Smeets, L.H. te Paske, “Thirteen years Test Experience with Short-Circuit Withstand Capability of Large Power
Transformers”, TrafoTech, session IV paper 5, Mumbai, 2010
[38] R.P.P. Smeets, R. Bruil, K. van der Linden, “Short-circuit test experience of power- and distribution transformers”,
CIGRE SC A2 Colloquium, Shanghai, paper FP0671, 2015
[39] Mengyun Wang, “1995-1999 Fault Statistics & Analysis All Transformer Type in China”, Electrical Equipment 2(1), 2001
[40] G. Bastos et al., “An Analysis of the Increase on Transformers Failure rate and Measures to Improve Transformers
Reliability”, CIGRE Conference, paper A2-101, 2006
[41] CIGRE WG A2.27: “Transformer Reliability Survey: Interim report”, Electra nr. 26