Short-Circuit Withstand Capability of

Large Power Transformers

R. P. P. Smeets#1, L. H. te Paske#2, T. Fogelberg*3
#
KEMA T&D Testing Services, Utrechtseweg 310, 6812 AR Arnhem, the Netherlands
1
rene.smeets@kema.com
2
henk.tepaske@kema.com
*
ABB Transformers, Sweden
3
thomas.fogelberg@se.abb.com

Abstract - The ability to withstand a short circuit is recognised 2. Seismic requirements

more and more as an essential characteristic of power trans- 3. Transport handling
formers. Hence, IEC and IEEE Standards, as well as other Therefore, these must be the basics for a short circuit safe
national standards specify that power transformers have to be design. The short circuit force gives rise to mechanical forces
short circuit proof and how to check this performance.
in milliseconds to hundred of tons.
Unfortunately, however, there is extensive evidence that the
matter is not as simple as the standards make it sound. The The current peak value and the correspondent forces depend
failures caused by short circuits are still a major cause of trans- on many factors. In HV systems, the most probable type of
former outages, though failure rates vary widely in different short circuit is a single-line-to-earth fault, normally due to
countries and systems, depending on various circumstances, environment conditions such as a stroke of lightning on the
network characteristics and the equipment installed. line, equipment failure at the the station, pollution of
insulation strings, etc. Sometimes, short circuit faults develop
Experience is reported of short-circuit testing of large power into other more extensive faults, such as a single-phase-to-
transformers during the past 11 years by KEMA in the Nether- earth fault developing into a double-phase-to-earth and
lands. In total, 77 transformers > 25 MVA participated in the eventually a three-phase fault. The relative severity of the
survey. KEMA statistics show that at first access to standard different types of fault depends on the characteristics of the
IEC 60076-5 short-circuit tests, 30% failed initially. Failures
were observed throughout the complete range of voltage (20 -
system. On the other hand there are factors as arc resistances
500 kV) and power (25 – 440 MVA) tested. and earth network impedances which will have compensatory
effects.
I. DESIGN CONSIDERATIONS REGARDING SHORT-
CIRCUIT WITHSTAND Attributes of well-designed power transformers

- Mechanical sound design and technology

Nowadays fast developing regions, where the demand for - Based on fundamental mechanics
electric power is rising steeply, are adding more and more - Verified by many short-circuit tests
power generating capacity and interconnections to the sys- - Rigid core clamping structure for short-circuit
tems. strength and transport
In addition to this, the changes in the sector are characterized - Accurate manufacturing guided by strict tolerances
by: and quality systems
1. Expanding cross-border electricity trade brings network - Rigid winding mandrels
operations close to their limits - Verified drying and pressing procedures
2. Development of local generation is considered independ- - Rigid low-voltage winding design and clamping
ently from available network capacities
3. Load flows are changing both in magnitude and direction
4. Network components are ageing and tend to be brought to Electromagnetic forces tend to minimize the magnetic energy
their loading limits density in the volume. It means that inner windings will tend
5. The network operation conditions have changed to reduce its radius and the outer windings the contrary,
That means that old and new transformers will see more increase their radius. In axial directions the windings are
severe short circuit stresses than before, creating the need for trying to reduce their heights.
a careful review of design considerations. Forces and related withstand criteria can be splitted into 2
For the future, the mechanical rigidity of a transformer is the components:
most vital performance factor. There are three reasons for
this: 1. Radial forces.
1. Withstand to short circuit stresses The failure modes for radial forces include:
- buckling of inner windings; compacting during known moisture and temperatures are the
- diameter increase of outer winding; key to get the right spring constant for long time service.
- spiralling of end turns in helical windings. Here well-defined processes in the winding shop and active
part assembly are a must. The final pressure setting after the
vapour phase process shall then bring the windings under
pressure during its life time.
Still the most important criteria are that all windings need
their pressure to avoid any displacement between the wind-
ings.
Designing power transformers is an iterative and interactive
process to find an optimal solution from the point of view of:
- masses and losses
- sound level
- short circuit strength
- winding temperature, hot spots and cooling equipment
- dielectric strength between windings and inside windings

Fig. 1: Buckling: Collapse of the cylindrical ABB's designers are supported by the worlds most advanced
winding shell set of design and verification programs for power trans-
An important feature of ABB’s short circuit management is formers. These interactive applications are used today in 14
that inner windings are subject to “free buckling”. It means power transformer plants.
they are self-supporting. There are no radial supports from
core to windings and from winding to winding. It means that More than 140 ABB power transformers of different designs
the strength is determined by Cu hardness (yield point) and have passed short-circuit tests, including around 30 of them
conductor geometry. Spiralling is avoided by tough criteria since 1996 built according to TrafoStar technology.
or non-helical windings are used. Also the dynamic response
from the winding is considered Which transformer units are worth being
considered for short-circuit testing?
2. Axial forces.
The failure mode for axial forces includes: - Key generator step-up transformers and auxili-
- mechanical collapse of yoke insulation, press rings, press ary units in power plants
plates and core clamps; - Key feeding transformers at power plant sub-
- conductor tilting; stations or huge load centers
- conductor axial bending between spacers; - Strategic intertie transformers, three-winding
- possible initial dielectric failures inside windings, followed system transformers (tertiary), auto-transform-
by mechanical collapse. ers
- Transformers with axially split winding con-
The axial forces are cal- nections
culated with FEM programs - Series of transformers, one taken out for a
where axial displacement "type" test
due to workshop tolerances - Track feeding transformers
are fully considered; even - Transformers in networks known for high fault
so axial displacement due to incidence or high fault currents
winding pitch. Windings are
dimensioned for maximum
compression forces, where II. SHORT-CIRCUIT TESTING VS VERIFICATION BY
dynamic effects are em- CALCULATION TOOLS
bedded.
Ampere turn balancing be- Discussions on the pros and cons of short-circuit testing can be
tween the windings is a pre- summarized to the following four statements:
requisite to avoid axial
forces on the windings. 1. "Calculations Can Prove Short-Circuit Withstand".
Therefore strict manufact- It is clear that calculation methods are an indispensable tool in
uring tolerances for wind- the design phase of equipment. Nevertheless, “the test proce-
ings are needed. As wind- Fig. 2: Spiraling: Tangential dure is considered a better means of ascertaining the real per-
ings are springs with about shift of the end turns in helical-
formance of equipment at short-circuit, since such a test
20 % cellulose, the correct type windings
demonstrates that both construction and design are adequate”
(quote from [1]). Recently, this is recognised in an amendment capability show a drastic bending down of an upgoing trend
to IEC 60076-5 Annex A that states (for transformers > 2.5 after short-circuit testing was introduced in China in the mid-
MVA) that "for the purpose of evaluation the unit under con- nineties [10].
sideration may be simultaneously compared with a limited
number of transformers that have passed the short-circuit test 4. "Short-Circuit Tests Reduce the Transformer's Lifetime".
succesfully and match most - but not all - the characteristics Reputable manufactures agree on the fact that for a properly
considered in Annex A" quote from [2]. designed transformer with enough margin to handle the
electrodynamic stresses, the effect of the short-circuit stress
2. "Few Transformers Fail in Service Due to Short-Circuits". will be that the windings undergo a certain settling [9]. The
A large scale international survey on failures in service was effect of the settling is that the stiffness of the windings
conducted in 1983 by CIGRE WG 12.05 [3]. At that time, a increases and this is visible in a small variation (if any) in the
failure rate of 2% was reported for HV transformers, and reactance values measured between the first tests, but
problems with manufacture were quoted as the largest known becoming smaller or nihil at the last tests. Such a transformer
cause of outage, problems with the design being stated as the is even stronger after the short-circuit-tests than before, and
second largest known cause. can be put safely in operation again. Chinese experience [11]
A recent Brazilian 10 year survey shows very high failure rate shows that 40 transformers (110 - 220 kV) put into service
of power transformers (64%) and reactors (24%), partly in after passing short-circuit tests, function without problem
service, partly in acceptance testing [4]. during the monitoring period up to 5 years.
Detailed material was presented by CIGRE WG12.19 [5] on
the failure rate due to poor short-circuit withstand capability. A III. KEMA'S EXPERIENCE FROM TESTING 1996 - 2007
failure rate of 1.2 failures per 10000 transformer years due to
external short-circuits is reported1. It is striking that from the Having available 8400 MVA of direct generator-fed short-
failed transformers, a third had passed a design review succes- circuit power (world's largest), KEMA can test transformers up
fully, whereas none of the failed transformers had been sub- very high MVA and kV ratings. Thanks to the generators,
mitted to a short-circuit test. there is a good match of supply voltage with test-object, as
A much higher failure rate in service due to short-circuit is well as sufficient time constant, and availability of power
reported by Canada. [6]. supply.
An evaluation is made of short-circuit tests in the 12 year
3. "Short-Circuit Testing is Too Expensive and Too Much Time period 1996-2007. The tests are performed on transformers
Consuming". with power 25 - 440 MVA and voltage 20 - 500 kV. in accord-
It is certainly true that considerable costs are involved in short- ance with IEC standard 60076-5 (2006) or IEEE standard
circuit testing of large transformers. These expenses, however C57.12.90- (1993).
must be weighted against the importance of the function of the The population includes single-phase and three-phase trans-
transformer in the network and the time it takes for repair or formers, auto-transformers, step-up -, railway -, auxiliary - and
replacement. three-winding transformers, 16 2/3, 50 and 60 Hz trans-
These costs must be evaluated against the risk of loosing trans- formers, YD- and DY-transformers.
formers in service. The largest transformers tested are 250 MVA single-phase and
Various large power companies have adopted the policy of 440 MVA three-phase.
requiring a short-circuit withstand test of certain transformers
purchased within the established procedures. During the past 12 years, 97 times a test access for a trans-
EDF (France) explicitely reports [7] a significant and positive former > 25 MVA (88 transformers from which 9 are re-
influence of short-circuit testing on the reduction of the rate of tested) has been counted:
winding faults in the overall French experience of many • 62 transformers showed no problem at the test-site. These
decades. Similar experience is also reported from India [8], transformers initially passed the short-circuit test, but the
[9]. Others (among which NTPC-India, EGAT-Thailand, final test-result is not always known because KEMA was
TNB-Malaysia) incite suppliers to pass a learning path towards not involved in the subsequent routine tests, the inspection
a succesful design through full-power short-circuit testing and the identification.
thereby using it as an essential and successful tool for quality • 26 transformers showed a problem due to short-circuit tests
improvement. that became apparent at the test site. 9 Transformers from
In China, all enterprises producing transformers recognise the this group had been retested at KEMA after modification
value of short-circuit testing in quality assurance. The vast and did not show a problem at the test site at the retest.
majority of users of transformers share this opinion.
Transformer failure rate due to lack of short-circuit withstand From these results, an initial failure rate is defined as the ratio
of test objects that resulted in failure to pass the test at first
1
access (26 transformers) and the total number of transformers
"Countries like China and India, where we today hear alot about (88). Thus, the initial failure rate is 30%.
problems with short-circuit failures have not given any contribution
to the survey" (quote from [1]).
Commonly, the reason of not passing short-circuit tests is IV. CONCLUSIONS
because the winding reactance variation is larger than specified It is concluded that short-circuit testing is a vital part of the
in the standards. This is usually confirmed by visual inspect- design process, and a crucial verification tool at the end of it.
ion, that revealed a huge variety of defects such as: The high failure rate at testing (around 30%) is the best de-
- Axial clamping system: Looseness of force axial clamping, monstration that calculation tools alone are not sufficient to
of axial compression force, of axial supporting spacers and design a transformer capable to withstand the worst case of
of top and bottom insulating blocks; short-circuit current.
- Windings: Axial shift of windings, buckling, spiralling ABB’s reports a failure rate from tests (in various labs) in the
- Cable leads: Mechanical movement, for instance from last 11 years of 3 out of 28 tests [12].
tapchanger to regulating windings; deformed or broken
leads, outward displacement and deformation of exit leads 24
initially not OK
from inner windings; broken exit leads; 20

number of transformers
initially OK
- Insulation: damaged conductor insulation; displacement of 16
vertical oil-duct spacers; dielectric flashover across HV-
12
winding or to the tank; displacement of pressboard
insulation; tank current due to damaged conductor in- 8

sulation; damage to tapchanger; 4

- Bushings: Broken or cracked LV-bushings. 0

20-100 100-200 200-300 300-400 >400
kV (rated)
On the other hand, in the cases (the majority) that the reactance
change is within the tolerances set by the standards, it is Fig 4: Initial failure rate 1996-2007 in voltage class
KEMA's observation that (visual) inspection only rarely
(approx. 5% of the cases) still leads to rejection of a certificate. 32
Nevertheless, visual inspection is necessary, because deform- 28
initially not OK

ations and displacements in supporting structures, clamping initially OK

number of transformers

24
systems, insulation materials, winding exit leads, external 20
connections from the coils to the tapchanger and within the on- 16

load tapchanger can not be detected by the reactance measure- 12

ments only. 8

KEMA's experience with the short-circuit reactance measure- 4

0
ments is that for power transformers a variation of more than 25-50 50-100 100-200 >200
1.0% indicates a large deformation in one or more coils. Also a MVA (rated)
gradually increasing variation during the short-circuit tests,
although in total not more than 0.5% to 1.0%, indicates a Fig. 5: Initial failure rate 1996-2007 in MVA class
progressive movement of winding conductors. Variations of
the reactance values between the short-circuit tests in a strange REFERENCES
way form an indication of large flexibility of the windings. [1] G. Bertagnolli: "Short-circuit Duty of Power Transformers", Book
printed by ABB Trasformatori Legano(Milano), ISBN Italy, 1996.
[2] IEC 60076-5 Annex A, "Theoretical Evaluation of the Ability to
One could wonder on what causes the high failure rate, Withstand the Dynamic Effects of Short Circuit", 2006
observed by short-circuit test laboratories world wide com- [3] CIGRE WG12.05: "An International Survey on Failures in Large Power
pared to the much lower failure in service. The main reason Transformers in Service", Electra 88, pp.21-37, 1983.
must be the severity of the tests, compared to actual service [4] G. Bastos et al., "An Analysis of the Increase on Failure Rate
Phenomenon and Measures taken to improve Transformer Reliability",
conditions. From a recent enquiry of CIGRE WG13.08 it can paper A2.01, CIGRE Conference 2006
be concluded that on a statistical basis, large power trans- [5] G. Bertagnolli: "Results of short-circuit performance of transformers",
formers have to face several full and many small short-circuits CIGRE Transformer Colloquium, Budapest, Short-circuit performance;
during their life, more precisely: the 90 percentile was Tests and Failures, report 2, CIGRE TC 12, 1999.
[6] IEEE Transformer Comm. meeting, Presentation Hydro Quebec, 2008
estimated to be 4 full short-circuits in 25 years. [7] G. Macor et al., " The Short-Circuit Resistance of Transformers: The
Thus, it must be assumed that this actual (full) short-circuit Feedback in France Based on Tests, Service and Calculation
current in service is normally (much) smaller than the rated Approaches". CIGRE Conference, paper 12-102, 2000.
short-circuit current for which the transformer is designed. [8] N.V.C. Sastry, H. Gupta, "Short-Circuit Test on EHV Transformers", 9th
Int. Conf. on Short-Circuit Currents in Power Systems, Cracow 2000
Because of the expected future increase in short-circuit power, [9] H. Gupta, N.N. Misra, "Concerns about Short-Circuit Tests", TrafoTech
especially in developing countries, this situation may change, 2002, pp. 121-25, Mumbai, India, January 2002
and the fact must be faced that during the life of the ageing [10] M. Wang, "1995-1999 Fault Statistics & Analysis All Transformer Type
transformer, its withstand against short-circuits will be brought in China", Electrical Equipment 2(1), 2001
[11] Y. He, M. Wang, "The Transformer Short-Circuit Test and the High-
to the limits. Power Laboratory in China - the Past, Present and Future", IEEE
Electrical Insulation Magazine, Jan/Feb. 2004, vol.20 no.1, pp 15 - 19.
[12] T. Fogelberg, "Surviving a short-circuit", ABB Review 1/2008