Stray Loss Evaluation in Power Transformers - A Review

S.A. Khaparde, Senior Member, IEEE

S. V. Kulkami, Member, IEEE

Indian Institute of Technology

Transformer Division
Crompton Greaves Ltd., Mumbai, India

Mumbai, India

ees&a@ee.iitb.emet

svkgtone cpl co.in

Abstract Survey of current research papers reveals the continued

interest in application of advanced techniques for accurate
estimation and control of stray loss in transformers. This paper
g.ves an overview of research, development and application of
various computational tools for stray loss analysis, based on over 50
published papers. All landmark papers are systematically classified.
Practicality of application of methods by transformer designers is
discussed The report concludes with critical comments on eficacy
of all approaches and directions for pursuing further research.
I. INTRODUCTION
Last couple of decades have seen advent of research
orientation towards application of advanced computational
methods to estimate and control stray losses. Stray load loss
in large ratings of generator transformers and autotransformers can be appreciably lugh. Challenge for designer
today is higher loss capitallzation, optimum performance,
and low cost, space and weight requirements; for which he
needs advanced analysis tools that provide synergistic effect
leading to optunum designs and product performance
improvements. The paper contains a review of about 50
papers, which have dealt with one or more components of
stray loss from the point of view of estimation and reduction.
Stray losses include eddy and circulating current loss m
windings, losses m flitch plate, core edge loss, loss due to
high current field, and frame and tank losses. The paper has
compared, for each stray loss component, pros and cons of
various methods of estmation, and these methods are
reviewed in todays context, where state-of-the-art analysis
softwares are available. Control measures for each stray loss
component are discussed.

11. STRAY LOSS COMPONENTS

References on basic eddy current theory [1-33, since 1950s
have tned to make eddy current analysis understandable and
practically applicable for solving complex engineemg
problems, and even in 1990s such papers give basic
foundation to eddy current researcher for analysis of various
stray losses components mentioned below.

0-7803-5935-6/00/$10.00
(c) 2000 IEEE

in

A . Winding Eddy Loss

Two-dimensional Finite Element Method (2-D FEM) is

most commonly used to compute eddy loss due to axial and
radial leakage fields [4]. The knowledge of flux density
distribution in the winding helps in choosing proper axial and
radial dimensions of conductors. This is particularly very
important for winding having tapings within the main body
where high radial component of flux density can cause
excessive loss and temperature rise. In large generator
transformers, low voltage winding may be designed with
Continuously Transposed Cable (CTC) with lower axial and
radial dimensions of individual strip, to minimize the eddy
loss. Eddy loss 111 individual W d i s c needs to be exactly
estimated to calculate hottest spot in the transformer, which
is generally at the top of the windings. In case of small
distribution transformers with crossmatic LV winding
(copper flat conductor), each turn has to be modeled in FEM,
as the tliiclmess of flat is comparable to or sometimes more
than depth of penetration, in which case modification of
leakage field due to eddy currents cannot be neglected. In
case of foil windings, radial leakage field will cause higher
eddy losses at the ends of the winding. Mullineux et al. [5]
obtain current density as the solution of integral equation
(Fredholm type) for a two winding transformer. For specific
transformer dimensions given in paper, the coefficient of
additional loss is 1.046. Turowskis method [6] assumes the
foil winding as a vertical section of an infinitely wide and
deep conducting plate, which is assumed symmetrically
penetrated on both sides by plane electromagnetic waves.
Today, 2-DFEM can easily analyze the foil winding eddy
losses without any simplifications. The analysis of winding
eddy loss by 3-DFEM analysis will be more accurate than 2D methods, but the complexity in calculations increases
manifold.

B. Circulating Current Loss in Windings

In transformers, loss due to circulating currents in parallel
strands, due to unequal linkage of leakage field, can be
reduced to a negligibly small value by continuous
transposition (like in CTC) or by employing number of
transpositions at pre-determined intervals along the winding
height. Kaul [7] has given analytical method for calculation
of circulating current loss in stranded winding with standard
transposition schemes. The radial field at Winding ends is
not considered in the calculation, which can be easily taken

2269

Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY BOMBAY. Downloaded on December 4, 2008 at 06:16 from IEEE Xplore. Restrictions apply.

In large transformers, penetration of leakage flux into the

cores, in a direction at right angles to plane of lamination,
causes eddy currents and local hot spots. Methods of slitting
the first step or shielding (laminated flitch plate) are adopted
to reduce this core edge loss. Leakage flux penetration into
laminated core poses an anisotropic and three-dimensional
non-linear field problem. Carpenter [13] has solved the
problem by formulating differential equations in terms of an
electric vector potential T and magnetic scalar potential 0.
Solution is expressed in the form of three different
characterisbc modes in a network model, two associated with
the core surfaces and the third describing the flux penetration
into the interior. However, the method is not so general,
since the network is rigid and has to be modified for any
change in geometry. Formulation in the paper was verified
on two experimental models of core [14,15]. Advance of 3-D
FEM with features of anisotropic modeling (of permeability
and conductivity) may help to overcome the difficulties for
accurate calculations.

mto account by 2-D FEM as given in [8]. This paper has

compared currents and loss calculated by analytical
formulation (based on multi-winding network formulation)
and 2-D FEM. The analytical formulation can be used for
any transposition scheme; reactance and resistance of parallel
strands are calculated separately depending on the positions
they take along the winding height, and these are then placed
in a network scheme. Recently a detailed 2-D FEM analysis
has been reported [9], in which analysis of layer and disc
winding has been done for various transpositions cases for
strip as well as bunch conductors. The circulating current
loss is also being calculated even more accurately by 3-D
FEM [lo], but the efforts involved for simulations may not
be justified for the Inappreciable increase in accuracy.

C.Flitch Plate and Core Edge Loss

Stray flux departing radially through the inner surface of
winding hits the core and fittmgs such as the flitch plate
mounted on the core. Although the losses occurring in flitch
plate may not form a significant part of total losses of a
transformer, the local temperature rise is much higher due to
a higher value of incident flux density and poorer cooling
conhtions. The loss density may attain levels that may lead
to hazardous local temperature rise if the matenal and type of
flitch plate are not selected properly. Literature available on
flitch plate loss analysis is quite scarce. Eddy-current losses
arising in metallic parts of rectangular cross-section are
calculated by an analytical method, which is based on certain
approxnnations [l 11. Field strength at the inner edge of the
LV winding is assumed to vary periodically with a sinusoidal
distribution in the space along the height of the winding, and
the non-sinusoidal nature 1s accounted by multiplying loss by
a factor. The analytical formulation cannot analyze flitch
plate with slots of limited length provided in radial flux zone.
These limitations of the analytical formulation can be
overcome in FEM analysls [12]. The paper describes details
of statistlcal analysls used in conjunction with 2-D FEM, for
quantlfying the effect of vanous factors affecting flitch plate
loss. The paper also presents results of 3-D FEM simulations
carried on slotted and laminated flitch plates. Performances
of mild steel and stainless steel flitch plates are compared by
a detailed examination of eddy current patterns. The slots are
more effective in the stainless steel flitch plate than the mild
steel flitch plate, as the field in case of stainless steel is
predommantly radial due to large penetration depth.
Simulation of 1aminated flitch plate proved that the loss in
lammated case is much lower as compared to stainless steel
plate. Hence for higher rabng transformers, particularly
generator transformers, a laminated flitch plate may be
necessary. The eddy loss distribution in the flitch plate
obtained by 3-D FEM electromagnetic analysis was used in a
3-D FEM thermal analysis to compute the temperature
distnbution. Analysis has been verified by measurement of
temperature of flitch plate on a transformer.

0-7803-5935-6/00/$10.00
(c) 2000 LEEE

D. Stray Loss due to High Currents

Poritsky et al. [16] discuss a method to evaluate eddy
losses in a semi-infinite solid nearby a filament current.
Practical cases, however, involve plates of finite thickness
and hence a modification of the solution is required. Deuring
[17] has presented an empirical formula for eddy loss
calculation in steel plates based on an experiment. Jain et al.
[la] evaluated field pattern and eddy current losses in
a l u "
sheet due to current carrying strip bus bars.
Current distribution is expressed as made of an inf~te
number of sinusoidal distributions with the help of Fourier
integral, and then the field due to any current distribution is
obtained by superimposition of fields due to sinusoidal
components. Various curves for loss are given, which are of
practical use. Experimental analysis of eddy current
phenomenon in structure that surrounds the high cment
bushings of large capacity transformers is presented in
[19,20]. Means to prevent overheating are also presented.
Eddy current pattern has been explained for various
configurations of terminations in [19].The 2-D formulation,
used to estimate the eddy current patterns, is based on some
approximations and experimental data. Krakowski et al. [21]
have presented a method for analyzing the electromagnetic
field in a system that comprises of parallel current carrying
bars placed above a steel wall. The current density within the
cross-section of bars is computed using the integral equation
technique. Thus it can be seen that most of the papers
published from 1970s have concentrated on analytical
methods, which due to some assumptions zpd
approximations, are useful to only simplified 2-Dgeometries
and cannot be applied to complex 3-D structures. With
improvements in 3-D FEM software capabilities, now such
complicated structures can be easily simulated and analyzed.
Analysis of eddy current pattern, in a complex three
dimensional model of h a c e transformer LV lead

2270

Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY BOMBAY. Downloaded on December 4, 2008 at 06:16 from IEEE Xplore. Restrictions apply.

termination, is reported in [22]. Total eddy loss estimated is

found to be in close agreement with that observed during
testing of the transformer.

E Frame Loss
Frames (yoke beams), serving to clamp the yoke and
support the windings. are in a vicinity of stray magnetic field
of windings. Due to large surface area and efficient cooling,
hot-spots seldom develop. Losses in frames have been
calculated by Finite Difference Method (FDM) and
analytically, and loss in frames made up of mild steel,
aluminium and non-magnetic steel are compared in [23].
Loss in frames can also be calculated efficiently by 3-D
Reluctance Network Method (RNM) [24], but may not be as
accurately as that by FEM [25]. The loss in frames can be
reduced by either aluminium shielding or by use of nonmetallic platforms for supporting the windings.

F. Tank Loss
Valkovic has [26] presented an analytical method in which
current sheet, a sum of trigonometric functions in between
core and tank (two half spaces), represents "f of winding.
The method can only be applied for a particular tank shape
and effect of shields on tank wall may not be easily
estimated. Szabados et al. [27] have formulated analytical
approach, which requires the knowledge of incident flux
density on tank wall obtamed by any other method. The
incident flux density is expressed in terms of double Fourier
sixies. Since 1960s. the research reported for calculatlon of
tank loss has been mainly concentrating on vanous analytical
methods mvolving complex mathematics, which are mostly
based on approxlmation of 3-phase transformer geometry.
Turowski's 3-D Reluctance Network Method [283 fulfills
designers' requirements of quick estimatlon of tank loss.
Reluctances for conductive tank parts are calculated
analytically by taking into account the skin effect, eddy
current reactlons with phase shift, non-linear permeability
inside solid metals and effect of eddy current shield, The
designer can define locations of magnetic and I or eddycurrent shield on the tank wall. The RNM3D approach for
tank loss estimabon has been verified [29] on various ratmgs
of power transformers.
Karsai et al. [11] have almost analyzed all the above
mentioned stray loss components. In addition to describlng
2-D and 3-D analytical methods for loss calculation, practical
guidelines are given for reducing losses due to leakage and
high-current fields. Performances of various shieldmg
arrangements are compared. Turowski [24] has also covered
number of topics useful to practicing transformer engmeers,
and useful formulae, thumb-rules. curves are given to take
quick and reasonably accurate decisions in the area of stray
loss esnmation and control.

0-7803-5935-6/00/$10.00(c) 2000 IEEE

111. GENERAL OVERVIEW OF METHODS

Krawczyk et al. [30] have presented an overview of
methods for eddy current analysis. The paper compares
methods based on differential formulation (analytical, FDM,
RNh4), integral formulation (volume integral, boundary
element method) and variational methods (weighted residual,
FEM)on attributes such as accuracy, ease of use, practicality
and flexibility. For estimation of stray losses in transformers,
the methods can be broadly classified into three categories,
as they progressed from 2-Danalytical methods to 3-DFEM,
as given below.

A. Two-dimensionalMethodr
Boyajian [31] has given a method for estimating leakage
field, in which any kind of current density distribution can be
resolved into space harmonics by a double Fourier series,
The method is very useful, mathematically less complex and
can be used for calculation of eddy loss and circulating
current loss in windings. Sat0 et al. [32] have presented
analogy between magnetic field equations for twodimensional Cartesian and axi-symmetric problems, and
usefulness of this analogy for numerical calculation has been
mentioned. Komulainen and Nordman [33] have used 2-D
FEM to get static magnetic field solution, and losses in tank
are calculated by analytical formulae. The paper contains test
results of tank loss with magnetic and eddy-current shielding.
Geometric parameters affecting tank loss are explained
through graphs. Pavlik et al. [2S] have emphasized the need
of analyzing the eddy and stray losses as a complete system
and not on an individual component basis. The authors have
done a number of 2-DFEM simulations to understand the
effect of magnetideddy-current shields, placed to cover tank
wall, on the other stray loss components (winding, flitch
plate, frame and core edge losses). Effect of change in
permeability of magnetic shunts on the tank loss has been
analyzed. It can be concluded that even in this era of 3-D
calculations, two-dimensional methods are preferred for
routine calculations of stray losses.

B. Three-dimensional Formulations

Sironi, et al. [34] have formulated quasi 3-D method in

which mi-symmetric leakage flux of the transformer in the
absence of tank wall is superposed with that of the images
reflected in the magnetic surface. El Nahas et al. [35]
developed a method to calculate 3-D magnetic flux density
on tank wall using a 2-Dsolution for one phase of a 3-phase
transformer. These analytical methods may not be easily
applied to complicated tank shapes and for finding the effect
of magnetideddy-current shielding on tank,which now can
be comfortably done by 3-D FEM.

227 1

Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY BOMBAY. Downloaded on December 4, 2008 at 06:16 from IEEE Xplore. Restrictions apply.

C Tliree-niwie~uionalFEM Analysis
Haack and Girgis [36] have verified 3-D FEM calculations
of flux densities along the height and breadth of the windings
on experimental shell type transformer. Subsequently, Girgis
et al. [37] have assessed the need and estimated the benefits
of using 3-D magnetic field calculations, particularly for
eddy and circulatmg current losses in windings. Dexin et al.
[38] have presented three-dimensional FEM analysis of eddy
current problems using the complex magnetic vector
potential. Eddy current losses in steel materials are computed
by combining numencal method with analytical formulation
because of problem of discrehation due to very thm skin
depth of about 1 mm. A power transformer has dimensions in
few meters, whereas skin depths are in millimeters resulting
into errors due to poor aspect ratio of elements. Holland et al.
[39] have outlined the method of modeling tank wall and
other fittings with surface elements removing the need of
complex layers of brick elements to account for skin effects
and this. in turn,reduces the complexity and size of models.
Many commercial 3-D FEM softwares now have the feature
of this surface impedance element modeling, and thus p e m t
designers to calculate tank losses efficiently and accurately.
The 3-D FEM analysis, which started gaining importance in
1980s. is being constantly upgraded to improve its modeling
capabilitles and accuracy for eddy current analysis.
IV. STRAY LOSS CONTROL

Measures for stray loss control are discussed with reference

to large power transformers in [40-42).Methods of reducing
the loss in vanow structural components due to high current
and Ieakage field are described in bnef in [40].Curves are
given for calculation of tank loss with and without shielding
in [41].Kozlowski and Turowski [43]have given formula
for calculating the hmitmg value of conductor width to avoid
hot spots in wmdings. Now by tools such as FEM, it is easily
possible to optimize the total winding losses (IR and eddy
losses) by, for example, using different conductor width and
thickness at the winding ends. The paper has also given
permissible values of tangential components of magnetic
field strength on various constructional elements to eliminate
local overheating. Principles for selection of the type and
thickness of tank shields are presented, and the shielding
efficiency is discussed. Kazmierski et al. [44] have given
guidelines and useful curves for preventing overheating
hazard due to stray field m various components such as
windings, first step of core, flitch plate, frames and tank.
Bose et al. E451 have given experimental results of a 37 MVA
transformer with various propomons of tank shielding by
magnetic shunts. The effect of positioning inter-phase
connections of high current carrying leads of generator
transformer on the tank loss has been explained with
e vperimental verification.

A . Magiretic Shielding

0-7803-5935-6/00/$10.00
(c) 2000 IEEE

Inui et al. [46] have evaluated effect of tank and tank

shields on stray loss in the windings by detailed
measurements on a 150 M V A transformer. Djurovic and
Carpenter [47] have studied the effects of a horizontal shunt
assuming that the shunt is connected directly to the yoke.
The work is extended in [48-493 to study the effect of a small
gap between shunt and yoke on leakage field distribution. It
has been emphasized that the gap between shunt and yoke
must be kept reasonably small for effective control of
leakage flux. Laminated iron is treated as solid anisotropic
block. The parameters of such a block include the directional
effects of both the material anisotropy and lamination
stacking factor [50].The papers [47-49] give useful practical
guidelines for designing of horizontal magnetic shunts.
Bereza [S 13 has compared effectiveness of flat and edge-wise
magnetic shunts by finding their effective anisotropic
permeability. Merits and de-merits of magnetic and eddycurrent shielding are explained in [52]. Loss measured under
various combinations of shielding (yoke shunts, flux
collectors and flux diverters) are reported, which is of
practical significance.
B. Eddy Current Shielding
Mullinem, et al. [53] have given analytical formulation
wherein the windings are replaced by infinite array of line
currents by using the theory of images. The flux carried by
tank, which is shielded by aluminium shield, is calculated by
assuming infinite permeability. Eddy current loss in nonmagnetic shields of air core reactors is evaluated by image
method using Fourier
Bessel integral in [54]; 2-D
approximations and end effects make the formulation less
accurate. With developments in FEM,it is now easy to assess
the effect of such eddy current shields on tank loss
performance even for 3-D complex structures. Eddy current
shields are generally used in case of odd tank shapes, where
magnetic shunts can not be used. This is because, there is
extra loss in the eddy current shield itself, and the diverted
flux from the shield may cause overheating in the nearby
structural part, if not studied properly.

V. CONCLUSIONS

In todays competitive environment, accurate estimation

and subsequent optimization of stray loss by advanced
techniques such as FEM will give a competitive advantage.
The paper has analyzed all the components that constitute the
stray loss in a transformer from the point of view of methods
of estimation, control and elimination of hot-spots. For
computing and controllmg stray loss components in
windings, viz. eddy and circulating current loss, 2-D
methods, analytical or FEM,have been successfully applied;
efforts required for 3-D analysis may be justified only for
large power transformers where improvement in accuracy
will be appreciable.

2272

Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY BOMBAY. Downloaded on December 4, 2008 at 06:16 from IEEE Xplore. Restrictions apply.

Accurate analysis of loss m vanous types of flitch plates

(mild steel, stamless steel and laminated) can be done by 3-0
FEM; analytical formulation is less accurate due to many
approximaQons. Although core edge loss can be easily
controlled. problem of evaluation of exact stray loss in core
due to leakage field is quite complicated. The advance of 3-D
FEM with features of anisotropic modeling may help to
overcome the difficuloes in calculations.
Estmabon of loss due to field of high currents has
progressed from analytical methods to 2-D FEM for simple
geometnes, and then finally to 3-D FEM for complex
geometry like that of LV terminations of large furnace or
generator transformers.
Frame losses can be calculated with reasonable accuracy by
2-D FEM. Tank loss estimation, which again poses a real 3-D
asymmetric problem, has graduated from 2-D/3-D analytical,
RNM to FEM formulations. Use of user-friendly 3-D FEM
software. 011 high speed and memory computers, with features
such as surface impedance elements, will help overcome
computational difficulhes, although tank loss reduction by
magnetdeddy current shielding has been successfilly
practiced by transformer manufacturers.
Survey of papers reveals that judicious choice of method of
estimation of various stray loss components has to be made by
a transformer designer, for which sufficient indications are
given m thx paper.

[9] D A Koppikar, S. V. Kulkami, G. Ghosh, S M Ainapure. J. S.

Bhavsar, Circulating current loss in transformer windings. Roc. IEE
Science. Measurement and Technology, vol. 145, no. 4, July 1998.
pp. 136-140.

[IO] R. S. Girgu. D. J. Scott, D. A Yannucci, and J. B. Templeton.

Calculahon of winding losses in shell form transformers for
improved accuracy and reliability, IEEE Trans.Power Delivery, vol.
PWRD-2.no.2.April1987,pp. 398410
[1I] K Karsai, D.Kerenyi, and L. Kiss, Large Power Transformers, New

York Elsemer Science Publishing Company, 1987.

I121 D. A. Koppikar. S. V. Kulkami. P. N Snnivas, S. A. Khaparde. and
R Jain, Evaluation of flitch plate losses in power transfonnm,
IEEE Trans. POW^ Delivery. PE-388-PWRD-0-08-1998.

[13] C.J. Carpenter, Theory of flux penetrahon into laminated iron and
associated losses, Proc. IEE,vol. 124, no.7, July 1977, pp. 659-664.
[14] R. F. Hemming and G. D.Wale, Heating in transformer cores due to
radial leakage flux, Part-1: Expenmental models and test results,
Pmc IEE. vol. 124, no. 11, November 1977, pp. 1064-1072.
[15] C J Carpenter, K. 0. Sharples, and M Djurovic. Heating in
transformer cores due to radial leakage flux, Part-II: Computed
results, Proc. IEE, vol. 124,n0.12, December 1977, pp. 1181-1 186.
[la] H. Pontsky and R. P. Jenard, Eddy current losses in a sm-infinite
solid due to a nearby altemating current, AIEE Trans., May 1954,
pp. 97-106.
[I71 W. G. Deuring, Induced losses in steel plates in the presence of an
altematmg current, AIEE Trans.. June 1957. pp. 167-173.

VI. ACKNOWLEDGEMENT

P Jam and L. M. Ray, Field pattem and associated losses in

The authors are thankful to the management of Crompton [lS] M.
aluminium sheet in presence of stnp bus bars, IEEE Trans. Power
Greaves Ltd.. India, for permitting publication of the paper.
Apparatus and Systems, vol PAS-89, no 7 Septembed0cmbe.r 1970.
pp. 1525-1539.

VII. REFERENCES
P D Aganval, Eddy current losses
AlEE Trans. May 1959,pp. 169.181

in

solid and laminated iron,

R L Stoll, The Analysis of Eddy Currents. Oxford: Clarendon Press,

IO74
V Subbano. Eddy Currents in Linear and Non-Linear Media. New
Delhi Omega Scientific Publisher, 1991.
0 W Anderson. Transformer leakage flux program based on finite

elemcnt method, IEEE Trans Power Apparatus and Systems, vol.

PAS- 92,1973, pp. 682489.

N Mullineux. J R. Reed. and 1. J. Whyte, Current distribution in

sheet and foil wound transformers, Proc. IEE. vol 116. no. 1,
January 1969. pp 127-129
J Tumwski. Additional losses in foil and bar wound transformers,
Proceedings of the 1976 IEEE PES Winter Power Meehng. New
York, January 25-30.Paper No A-76-151-1

H J

Kaul. Stray current losses in stranded wndings of

transformeis. AIEE Trans, June 1957. pp 137-149

S kaka, T Tokumasu. and K. Kondo. Finite element analysis of

eddy ciimnts in transformer parallel conductors. IEEE Trans. Power
Appai-atus and Systems, vol PAS-104, no. 10. October 1985, pp.
273 1 - 2737

0-7803-5935-6/00/$10.00
(c) 2000 IEEE

1191 S. Saito, K. Inagakl, T. Sam. Y. Inui, K. Okuyama, and H. Otani,

Eddy currents in structure surrounding large current bushings of a
large capacity transformer. IEEE Trans. Power Apparatus and
System, vol. PAS-lW,no.ll,November 1981,pp. 4502-4509.
[20] Y. I. Furman. V. L. Bereza, V. F. Ivankov, and L. P. Nizhnik.
Losses in tanks of large power transformers caused by magnetic
field, and methods of their reduction, Paper no. 12-07. CIGRE 1988.

[21] M R Krakowski, M Kazmierski. and I. Kersz. Effect of current

distnbuhon in parallel bars on clectromagnetx field at nearby steel
wal1,Proc. IEE, vol. 136,pt. B, no. 4. July 1989, pp. 161-167.
[22] D. A Koppikar, S V. Kulkami, S A Khaparde. and S. K. Jha,
Evaluahon of eddy losses due to high current leads in aansformers,
Proc. IEE Science Measurement and Technology, vol. 144, no. 1,
January 1997. pp 34-38.

1231 D. Kerenyi, S h y load losses in yokebeam of transfomrm,

Electromagnebc Fields in ElectncaI Engineenng, Book edited by A.
Savini and J. Turowski, Plenum Fress, New York, 1988, pp. 113-1 18.
[24] J. Tumwski, Technical Electrodynanncs. (in Polish), Warszawa:
WNT. 1993.
[25] D.Pavlik, D.C.Johnson, and R. S.Girgis, Calculation and reduction
of stray and eddy losses in core form transformers using a highly
accurate finite element modeling technique, IEEE Trans. Power
Delivery. vol. 8, no. 1, January 1993, pp 239-245.

2273

Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY BOMBAY. Downloaded on December 4, 2008 at 06:16 from IEEE Xplore. Restrictions apply.

1261 Z Valkoric. Calculation of the losses in three-phase transformer

tanks, Proc. IEE, vol 127, pt C,no 1. January 1980. pp. 20-25

[42] R. Beaumont, Losses in transformers and reactors, Paper no. 12-10,

CIGRE 1988

[27] B Szabados, I El Nahas, M S El Sobki. R. D Findlay. and M.

POlOujadOff, A new approach to determine eddy current losses in the
tank walls ofa power transformer, IEEE Trans. Power Delivery, vol.
PWRD-2.no 3. July 1987. pp 810-816

[43] M. Kozlowski and J. Turowski, Stray losses and local overheating

hazard in transformers, Paper no. 12-10. CIGRE 1972.

[28] J Turowski. M. Turowski, and M. Kopec, Method of 3dimensional

network solution of leakage field in three-phase transformers, IEEE
Trans Magnetics. 26. no 5, 1990.p~.291 1-2917.
129) D A Koppikar. S V. Kulkami, and J. Turowski, lndusmal
verification of 3D hybrid analytically numerical calculation of stray
field and losses in 3-phase power transformers, Intemational
Symposiuin on Electromagnetic Fields, lSEF95. Thessaloniki,
Greece. September 1995, pp 345-348

[44] M. Kazmierski. M. Kozlowski. J Lasocinski. J Pinkiewcz. and J.

Turowski, Hot spot identification and overheating hazard p e n t i n g
when designing a large transformer, Paper no. 12-12, CIGRE 1984.
[45] A. K. Bose, C. Kroon, and G. Hulsink, Some topics on designing
transformers with IOWload loss, Paper no. 12-05, CIGRE 1988.
1461

Y. h i , S. Saito, K. Okuyama, and K. Hiraishi, Effect of tank and

tank shields on magnetic fields and stray losses in transformer
wndings, Proceedings of the 1973 IEEE PES Summer Meeting,
Vancouver, Paper no. C73-401-7

[30] A Krawczyk and J Turowski, Recent developments in eddy current

analysis. IEEE Trans. Magnetics. vol MAG-23, no. 5, September
1987. pp 3032-3037.

[47] M Djurovic and C. J. Carpenter, Three dimensional computation of

transformer leakage fields and associated losses: IEEE Trans.
Magnehcs, vol. MAG-11, no. 5, September 1975, pp. 1535-1537.

Boyajian. Leakage reactance of imgular distributions of

transformer \rindings by the method of double Founer series, AlEE
Trans .October 1954, pp. 1078-1086

[48] M. Djurovic and 1. E. Monson, Three dimensional computation of

the effect of the horizontal magnetic shunts on transformer leakage
fields, IEEE Trans. Magnetics, vol. MAG-13, no. 5, September 1977.
pp.1137-1139.

[31J A

[12] T Sato. Y. Inui. and S Saito, Solution of magnetic field, eddy and
circulating current problems taking magnetic saturation and effect of
eddy current and circulating current paths into account, Proceedings
of the 1977 IEEE PES Winter Power Meeting. New York, Paper no.
A77-168-8

and J. E. Monson, Stray losses in the step of a magnetlc

[49] M. DJU~~VIC
shunt of a transformer yoke with a horizontal magnetic shunt. IEEE
Trans. Power ApparahJs and Systems, vol. PAS-101, no. 8, August
1982, pp. 2995-3000.

L. Barton, Loss calculation in laminated steel utilinng

anisotropic magnetic permeability, IEEE Trans. Power Apparatus
and Systems, vol. PAS-99, no. 3. May/June 1980, pp. 1280-1287.

133) R. Komulainen and H Nordman. Loss evaluation and use of

magnetic and electromagnetic shields in transformers, Paper no. 1203. C E R E 1988

[50] M

[34] G G Simni and J R. Van Hulse, Eddy current losses in solid iron
pieces under three-dimensional magnetic field: IEEE Trans.
Magnetics, vol. Mag -14. no.5. September 1978, pp. 377 -379.

[51] V L. Bereza, Designing magnetic shields for transformer

[35] I El Nahas. B. Szabados. R. D. Findlay, M. Poloujadoff, S . Lee, P.

Burke, and D Perco. Three dimensional flux calculahon on a 3phase traiisfbmr, IEEE Trans Power Delivery. vol PWRD-I, no 3,
Jtlly 1986. pp 156-159.

tanks,

Elektrotekhnika. vol. 52, no. 12,1981, pp. 4446.

[52] T. H. Hamson and B Richardson, Transformer loss reductions,
Paper no. 12-04, CIGRE 1988.
[53J N. Mullineux and J. R Reed,Eddy current shielding of transformer
tanks, ptoc. IEE,vol. 1 13, no 5, May 1966, pp. 8 15-818.

[-Xi] R W Hawk and R S. Girgis, Experimental verification of threedimensional analysis of leakage magnetic fields in large power
tnnsformen, IEEE Trans. Power Apparatus and Systems, vol. PAS102. no 9. September 1983, pp 32 12-32 I7

[54] 1. Woolley, Eddy current losses in reactor flux shields, hoc. IEE,
vol. 117, no. 1 1 , November 1970. pp. 2142-2150.

[37] R. S. Girgis, D A. Yannucci, and J B. Templeton, Performance

panmeters of power transformers using 3D magnetic field
calculations, IEEE Trans. Power Apparatus and Systems, vol PAS103, no 9,September 198J.p~2708-2713

VIII. BIOGRAPHIES

I381 X Dexrn, T Yunqiu, and X Zihong, FEM analysis of 3D eddy

cuiwnt field in power transformer. IEEE Trans. Magnetics, vol.
MAG -23, no 5. September 1987. pp. 3786-3788
1391 S A Holland, G. P OConnell, and L. Haydock. Calculating stray
losses in power transformers using surface impedance with finite
elements. lEEE Trans. Magnetics. vol 28, no 2. March 1992, pp
1355-1358
[40] T Tagaki, T Ishii, T Okada. K Kunta, R. Tamura, and H. Murata,

Reliability improvement of 500 kV large capacity power

transformer. Paper no 12-02, C E R E 1978
[JIJ H W Kerr and S Palmer, Developments in design of large power
transformers. Proc IEE. vol. 1 11.110.4, April 1964, pp. 823-832.

0-7803-5935-6/00/$10.00 (c) 2000 IEEE

S.V. Kulknrnl (M 1999) was bom at Dombwli of Maharashtra state in

India in 1967. He received his B.E. in I988 and M.Tech in 1990 from
Indian Institute of Technology (UT), Mumbai in Electrical Engineenng. He
IS worlong as Deputy Chief Design Engineer at Transformer Dwision,
Crompton Graves Ltd. He has several publicahons to his crrdit and his
research interests include optimizahon and reliability tmprovemt of EHV
Transformers and Reactors. He is currently pursuing his Ph.D. degree in
Electrical Engtneerrng at IIT Mumbai.
S.A. Khnpnrde (M 1987, SM 1991) was bom at Anuavati of
Maharashtra state in India, received his B.E. in 1971, M.Tech. in 1973 and

Ph.D. in 1980 from Indian Institute of Technology, Kharagpur. He IS

cunently professor of electncal engineenng at IIT Mumbai. He has several
publications to his credit and his research interests include pattem
recognition. power system security. Artificial Intelligence and applicahons,
parallel processing and neural networks. He i s senior member of IEEE and
Intemahonal Neural Network society.

2274

Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY BOMBAY. Downloaded on December 4, 2008 at 06:16 from IEEE Xplore. Restrictions apply.