See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/260587657

Literature survey of eddy-current loss analysis in rotating electrical machines

Article in IET Electric Power Applications · November 2012

DOI: 10.1049/iet-epa.2011.0335

CITATIONS READS

20 1,127

3 authors:

Anoop Jassal Henk Polinder

12 PUBLICATIONS 159 CITATIONS

Delft University of Technology
293 PUBLICATIONS 10,442 CITATIONS
SEE PROFILE
SEE PROFILE

J.A. Ferreira
Delft University of Technology
345 PUBLICATIONS 8,196 CITATIONS

SEE PROFILE

All content following this page was uploaded by Anoop Jassal on 10 July 2021.

The user has requested enhancement of the downloaded file.

 www.ietdl.org
Published in IET Electric Power Applications
Received on 2nd November 2011
Revised on 11th June 2012
doi: 10.1049/iet-epa.2011.0335

ISSN 1751-8660

Literature survey of eddy-current loss analysis in

rotating electrical machines
A. Jassal H. Polinder J.A. Ferreira
Electrical Power Processing (EPP) Group, Faculty of Electrical Engineering, Mathematics and Computer Science,
Mekelweg 4, 2628CD, TU Delft, The Netherlands
E-mail: a.k.jassal@tudelft.nl

Abstract: Eddy current is a very important and intriguing phenomenon. A direct result of Lenz’s law, it has been an integral part
of research on electrical machines. The applications and associated research field of eddy currents are very widespread and
virtually impossible to include in a singlestudy. Therefore this study tends to capture research trends in eddy-current loss
analysis pertaining to the field of rotating electrical machines only. The research period taken up is from 1892 to 2010.
The trends in research and application have been brought out. This study aims to serve as a reference guide to all those
interested in the field of eddy-current loss analysis in electrical machines.

1 Introduction lost in iron loss is composed of two components [1]

Electrical machines are one of the prime inventions of E = hB̂1.6 + 1N B̂2 (1)
mankind. Their utility and importance need no introduction.
Like many engineering utilities, electrical machines have
where E is the energy lost; h is the coefficient of hysteresis; e
been evolving over time. In the past 150 odd years, we
is the coefficient of eddy currents; N is the frequency; and B̂ is
have been able to develop a whole branch of electrical
the peak flux density.
engineering dealing specifically with electrical machines.
The first part of this expression is the hysteresis loss
One of the most important subjects falling under the
contribution and the second part presents eddy current loss
category of electrical machine design is iron loss in the
contribution.
electrical machines. The complex geometry, material
Jordan proposed a similar formulation but with different
properties, presence of time varying electromagnetic fields
exponents [2]
amid electrically conducting parts have made the study of
iron loss rather challenging and interesting. There are
many general iron loss formulations which are used to Pfe = Chyst f B̂2 + hexc Cec f 2 B̂2 (2)
estimate and group the iron losses in electrical machines.
According to the most accepted view, the iron loss in where Pfe is the iron loss; f is the electrical frequency; B̂ is the
electrical machines can be broadly classified into three peak flux density; Chyst and Cec are the coefficients of
categories: hysteresis and eddy current loss; and hexc is the excess or
anomalous loss factor.
(a) Eddy current loss: The resistive loss caused because of With further experience, another loss term called ‘excess
the induced electric currents, which are produced owing to losses’ was added to account for discrepancies in
change in flux density. measurements and calculations [3]
(b) Hysteresis loss: The loss within the structure of magnetic
material (at domain level) because of the changes in flux Pfe = Chyst f B̂2 + Cec f 2 B̂2 + Cex f 1.5 B̂1.5 (3)
density.
(c) Excess loss: The excess loss arises because of internal
correlation fields between magnetic domains. These fields, Another formulation for total iron loss is by dissociating loss
together with other effects such as eddy currents, act as a caused by linear magnetisation, rotational magnetisation and
damping field opposing any changes in the external loss caused by higher harmonics [4]. This development has
magnetising field, hence excess loss is produced. its roots in increased use of power electronics and hence
increased harmonic content in power supply or generated
Triggered by experiments and empirical relations, the early voltage
research focused on the first two types of loss, that is, eddy
current and hysteresis loss. Steinmetz classified that energy Pfe = C1 Plin + C2 Prot + C3 Ph (4)

IET Electr. Power Appl., 2012, Vol. 6, Iss. 9, pp. 743– 752 743
doi: 10.1049/iet-epa.2011.0335 & The Institution of Engineering and Technology 2012
www.ietdl.org
where C1 , C2 and C3 are the empirical coefficients and Plin ,
Prot and Ph denote power loss owing to linear
magnetisation, rotation magnetisation and higher harmonics,
respectively.
However, in this particular paper, we are going to focus on
eddy-current loss analysis in electrical machines. Eddy
current or Foucault current is a well-known phenomenon in
electromagnetics. A lot of literature is available to extract
knowledge about these currents and the associated physical
effects. However, there is hardly any historical survey
which gives a broad picture about the methods of analysing
eddy currents. This paper is an attempt to present a brief
historical development of this intriguing and important
phenomenon. The prime objective of the paper is to present
the reader with information about the development of
various important methodologies used in the analysis of
eddy current losses. The exhaustive chronological reference Fig. 1 Eddy currents induced in a conductor
list itself is an important and useful set of information.
There are numerous eminent people who contributed to this caused when a conductor is exposed to a changing
field of research and it is not possible to name them all magnetic field. This can be owing to relative motion of the
here. In order to limit the scope, only rotating machinery magnetic field source and conductor or owing to variations
and journal publications were considered in the survey. This of the field with time.
is because the trends in the field of development of This changing magnetic field causes a circulating flow of
eddy-current loss analysis can be projected easily even with electrons or a current, within the body of the conductor.
this restricted scope. The research on this subject can be The name eddy current comes from the fact that the
divided into three stages based on the most used/popular circulation of currents resembles vortices caused in fluids
method of analysis: by turbulence of flow. These ‘eddies’ of current create their
own magnetic field that opposes the change in the original
(a) Stage1: Early phase 1892 –1950; magnetic field. This phenomenon of eddy currents has both
(b) Stage2: Middle phase 1951– 1990; desirable and undesirable physical effects. The two major
(c) Stage3: Modern phase 1990 onwards. effects of eddy currents are:
This demarcation in stages is porous and has been done to
arrange the evolution of this research systematically. It is (a) Magnetic field opposition: The eddy currents generate
important to note that in many cases more than one method their own field so that they counteract the primary field that
is employed to reach a conclusion. There is a primary produced them. Induction machines basically exploit this
analysis which is a contribution and a secondary analysis property. It is also used in eddy current braking, crack
which is used to validate the model or experiments. The testing and so on.
phases and their segregation according to the methods are (b) Heating: The circulation of eddy currents leads to ohmic
based on the primary method of analysis used by the authors. heating of the conductor they are induced in. This can be a
The paper starts with a general introduction to iron loss in desirable property in heating applications while it is a
electrical machines and then a fundamental description of drawback in electrical machines.
eddy current loss phenomenon has been presented. After
that, the three stages of development have been described. 3 Stage 1: 1892 – 1950 – experiments and
Some formulations used during the development stage formulas
have been summarised. The formulations were chosen to
represent a variety of effects which scientists/engineers Like many physical discoveries, the eddy currents were
were trying to capture, for example, popular formulation ‘observed’ first and then the science able to describe the
with two-dimensional (2D)/three-dimensional (3D) analysis, phenomenon was built around those observations. Therefore
linear/non-linear materials, scalar/vector potential and so on. initial studies were carried out on giving a theoretical basis
have been chosen. It might be pointed out that the problem for the phenomenon. The scientists of those times were
formulations are not limited to those mentioned in the trying to understand the effects produced by these
survey but it is not possible to include all the formulations. intriguing currents. From the perspective of electrical
Based on the literature material studied in this research, machines, the losses owing to eddy currents were first
certain trends have been brought out in the results section. studied seriously in 1892 by Thomson [5] where he
analysed the heating effects of eddy currents in an iron plate.
2 Eddy currents – physical phenomenon Among the rotating machines, DC machines were the most
popular machines until the invention of AC machines by
The physics of eddy currents follows from Lenz’s law, which Tesla and Ferraris almost simultaneously around 1888
means that: [6, 7]. In DC machines, eddy current losses were studied in
conductors and armature iron especially owing to slotting of
‘An induced current always flows in such a direction as armature [8 – 12]. However, eddy current losses were not
to oppose its cause’ given that much importance before the late 1800s. The
invention of AC current, induction machines and
Eddy current is induced in a conductor to oppose the change transformers triggered a lot of interest in the analysis of
in flux that generates it (see Fig. 1). Eddy current flow is eddy current losses. It was quickly understood that these

744 IET Electr. Power Appl., 2012, Vol. 6, Iss. 9, pp. 743 –752
& The Institution of Engineering and Technology 2012 doi: 10.1049/iet-epa.2011.0335
 www.ietdl.org
losses depend on material conductivity and surface area literature of the time. Readers interested in knowing about
where current could flow. Consequently, sheet steel these formulae can refer to [26]. The formulae thus
laminations were being used to restrict eddy current losses developed were used for refining the design of machines
in cores of machines since (atleast) as early as 1900 [13]. and are given in Table 1.
Thereafter, the research on materials and methods to reduce A loss calculation comparison using different formulae has
such losses was carried out. Meanwhile, electromagnetic been presented by Aston and Rao in figure 20 of [26]. This
field modelling was also getting more elaborate and type of loss comparison however is not very scientific. This is
accurate. Following these developments, M.B. Field wrote because of the complexity of the problem where we cannot
one of the first analytical papers on eddy current losses in determine a common reference for all the formulae. Different
solid and laminated conducting materials [14]. scientists at different times have different correlations within
From about 1904 onwards, research in induction machines their findings. Hence, details on this quantitative comparison
picked up but it saturated soon [15 –17]. This could be of losses have been omitted. It may also be highlighted here
because induction machines reached an optimum that the above table does not list all the available formulations
performance level suitable enough for applications of that for calculation of eddy current losses, which is virtually
time. The same trend can be seen in eddy current loss impossible to include in a single paper. However, the purpose
analysis in the field of power transformers. From the study here is to show the empirical nature of the results. It was the
of methodology employed by most authors of this time, it lack of consistency in expressions and measurements that led
can be said that initial research (late 1800s to early 1900s) to further improvement in the eddy current loss formulation in
was carried out to observe the effects of the eddy currents. stage 2. At this stage analytical, semi-empirical and empirical
Therefore most of the research was based on performing analysis’ availability was a trigger for other applications
experiments and deriving empirical formulas to calculate and for eddy currents as well. Heating, braking and crack
eddy-current losses. This practice continued for almost three detection are only a few to name. In short, we can say that this
decades. During this period, some publications on induction period presents observation of phenomenon of eddy currents
machines and transformers can be seen as in [13, 14, 18]. followed by keen pursuit of understanding it according to
Considerable interest was also shown in material the means available at that time. This laid ground for the
characterisation [19 – 23]. Around 1937, engineers and next period, which dealt extensively with the scientific
scientists started analysing losses in cores of synchronous development of methods for analysis of eddy current and
machines analytically [24, 25]. This work was one of the losses thereof especially in rotating electrical machines.
first scientific works on eddy current losses in AC
machines. Till that time linear formulation (in Cartesian 4 Stage 2: 1951 – 1990 – analytical methods
coordinates) of machines was popular [26, 27]. This was
followed by some more experimental work in 1941 by Rao The 2D analytical methods for determination of
and Aston who summarised previously used formulae in the electromagnetic fields had already developed during stage 1

Table 1 Some formulae used for eddy-current losses calculation

First author Formula Methodology

F.W. Carter [28] √ theoretical

Pe = kv 1.5 B̂ 2 l
B̂ = flux - density (peak); l = tooth - pitch
v = velocity
v 1.5
Pe = ke m2 B̂ 2 here, m = thickness (lam.) theoretical
l
M.B. Field [14]   analytical
sinh mh − sin mh
Pe = 0.315rmĤs2
cosh mh − cos mh
r = resistivity; Ĥs = field - strength:surface

−4 fm
C.A. Adams [9] m = 2p10 ; h = thickness; m = permeability experimental
10r
 2.5 
2.65 B v 1.55 1.88 √
Pe = 4 3
q l
10 10 103
slot - opening
q= (cm); v = velocity (cm/s)
airgap
l = tooth - pitch(cm); B = flux - density
s3
K. Aston [26] Pe = ke B x v 1.75 (g + 0.05) experimental
l4.9
 
0.0207 s 3 experimental
Pe = ke B x v 1.7 1 −
g l3.1
v = speed (ft/ min ); s, l, g, ke are constants
 
v 1.6 s 2.2
T. Spooner [17] Pe = k B̂ 2.6 0.3 experimental
l g

IET Electr. Power Appl., 2012, Vol. 6, Iss. 9, pp. 743– 752 745
doi: 10.1049/iet-epa.2011.0335 & The Institution of Engineering and Technology 2012
www.ietdl.org
of the research on eddy current losses. B. Hague wrote a book Concordia wrote a series of four papers on machines with
on electromagnetic field modelling in engineering solid rotors from 1958 – 1960 [40 – 43]. These papers were
applications as early as 1929 [29]. Till the end of stage 1, completely analytical with rigorous mathematical theory.
rectangular coordinates, linear materials and harmonic Effects of curvature, 3D effects and material non-linearities
formulations were popular. In stage 2, as we will see later were explored analytically. This was also one of the first
in this section, material non-linearity, effects of curvature detailed analytical works in polar coordinates. Although
and 3D effects were taken into account in analytical these papers were focusing more on field modelling
modelling. however eddy current losses could be analysed from these
The most popular machines of stage 2 were the results.
conventional synchronous machines. This can be directly From 1961 to 1970 we can see a meteoric rise in analytical
linked to the development of many hydro-power and work done on magnetic field modelling and eddy current
thermal-power plants erected during this period especially losses in electrical machines. In 1962, Hammond and
1965 – 1990. Therefore an increased use of synchronous N. Kesavamurthy wrote on eddy current losses in solid
machines for generation of electricity can be predicted. conductive slabs followed by research on eddy current
In this period, significant research was done on sheet steel losses in thin sheets [44, 36]. This analysis was taken a step
used in electrical machines and eddy current loss analysis further by Subba Rao by including effect of saturation
thereof [30 – 32]. It is already known and acknowledged that analytically in 1964 [30, 31]. Effect of eddy current reaction
an appreciable amount of research was carried out in field in analytical eddy current loss calculation was aptly
Germany and France related to electrical machines, done in the same period by Mukherji [32].
materials and electromagnetic field modelling. The utility of Although technically possible, analytical modelling started
these works is somewhat restricted because of the language becoming very complex at this stage. The results thus
and hence is not considered for formulating trends in this obtained were also not very accurate (compared with the
survey. measured experimental results) owing to a number of
assumptions required to solve such complex problems
4.1 General formulation for eddy current losses analytically. The engineers and scientists were looking for
somewhat better methods to formulate and solve eddy
The general formulation which almost all the authors current loss problems. One of the approaches used at that
followed consisted of a magnetic field solution starting from time was the circuit approach linked to analytical modelling
Maxwell’s equations. This type of analysis existed in stage [45]. R.L. Stoll and P. Hammond generalised the theory of
1 and scientists were building upon that knowledge to loss modelling owing to eddy currents in 1965 – 1966 [37]
include more complex geometries, material properties and followed by the first attempt of using numerical methods
3D effects. Here, in this section, we briefly explain the for analysing eddy-current losses in 1968 [46]. Oberretl
general formulation of eddy current based problems. while working on pole losses in synchronous generators
The initial formulations of eddy current losses in this stage presented an innovative analogue circuit methodology for
were in 2D. The following assumptions were used to set-up inculcating the effect of variable permeability (saturation) in
the problem. loss calculation [47]. This method received some attention
but never became main-stream. From 1967 to 1970,
1. Cartesian coordinates were used in 2D analysis. Lawrenson, Reece and Ralph worked extensively on the
2. Materials were assumed linear wherein the B– H curve was theory of eddy current losses in conductive media including
a straight line. 3D effects [38, 48– 50]. In 1970, Lim and Hammond
3. The end effects were neglected. developed universal loss charts for eddy current losses in
4. The stator was assumed slotless and the winding was thick conductive plates serving as a guide to machine
replaced by a current sheet placed at an effective air gap. designers [51]. They used analytical as well as experimental
Fourier series was commonly used to take into account results to derive those charts. The 1970s also saw the first
space harmonics. major deviation from analytical to numerical methods for
electrical machines. Stoll presented a general solution
Using these assumptions, the analytical models could then method for eddy-current losses numerically in 1970 [52]
be solved by employing boundary conditions. The problem followed by a power series method together with Muhlhaus
was solved for the magnetic potential. Both scalar and [53] in the same year. Soon after that, finite-element (FE)
vector magnetic potential were used however vector method appeared on the scene. In the field of electrical
magnetic potential seemed more popular. The solution of machines, Chari, Reece and Carpenter were the first
magnetic potential leads to calculation of an induced proponents of this method and it gained immediate
current density and then material resistivity is used to popularity [54 –56]. The reason for this deviation was the
calculate eddy current losses. This analysis is very common capability of the FE method to model difficult geometries
and therefore details of the analysis are not presented here. and inclusion of material properties [57]. The popularity of
Interested readers can refer to [33, 34] for details we have FE and numerical methods is evident from a steep decline
presented some important formulations in Table 2. in the usage of analytical modelling (see Fig. 2). There
were also attempts to include circuital analysis along-with
4.2 Developments in eddy-current loss analysis – numerical methods to simulate eddy current losses in a
stage 2 more physical sense [58].
A sharp decline in research on eddy current loss in rotating
Bondi and Mukherji [35] presented a thorough analytical machines was noticed in the 1980s. One reason could be that
model for tooth ripple losses in pole shoes of synchronous most commonly used machines already reached somewhat
generators in 1957. Following this, Agarwal [39] wrote a optimal performance levels by iterating on design and
benchmark paper on eddy-current losses in solid and experience. There was hardly any new research focusing on
laminated iron in 1959. Apart from that, Wood and eddy-current loss analysis on induction machines or

746 IET Electr. Power Appl., 2012, Vol. 6, Iss. 9, pp. 743 –752
& The Institution of Engineering and Technology 2012 doi: 10.1049/iet-epa.2011.0335
 www.ietdl.org
Table 2 Some analytical formulations used for eddy current loss calculation

First author Problem formulation Loss calculation

2p/v 
H. Bondi [35] –2D Cartesian coordinates. rv
Pe = [<(J × H)ds] dt
–scalar magnetic potential 2p 0 s

–linear material Pe = eddy - current losses;

–end effects neglected J = current density;
H = magnetic field intensity
1
Pe = (Êx0r Ĥy0r cos c10r −Êy 0r Ĥx 0r cos c20r )
4
N. Kesavamurthy [36] –2D Cartesian coordinates. Pe = eddy - current losses
–vector magnetic potential Êx0r , Êy0r = electric field peak x, y components
–linear material
–end effects neglected Ĥx 0r , Ĥy 0r = magnetic field intensity x , y components
c10r , c20r = respective phase angles

4 2
Pe = vab mm Hm F2
p2
V. Subbarao [31] –2D Cartesian coordinates. v = electrical frequency
–vector magnetic potential a, b = dimensions in m
–non-linear material mm = normalised permeability
–end effects neglected Hm = normalised field intensity
R.L. Stoll [37] –2D Cartesian coordinates. F2 = normalised flux (complex)
–vector magnetic potential 1 jrqp 2 (K̂z e −qb )2 (z − jh)
–linear material Pe = Re
2 {(mr + z)2 + h2 }
–end effects neglected 
 

{1 + (1 + p 4 )} { − 1 + (1 + p 4 )}
z= ; h=
2 2

m0 mr v
p= ; q = p/g
q2 r
g = pole pitch; other symbols have usual meanings
P.J. Lawrenson [38] –3D polar coordinates.    L 
av 2p/v 2p a
–vector magnetic potential Pe = {[Sz ]z=0 − [Sz ]z=L }dr− [Sr ]r=a dz du dt
2p 0 0 0 0
–linear material
a = radius of cylinder
Sr , Sz = radial and axial component of pointing vector
L = length of cylinder
other symbols have usual meanings

synchronous machines in the 1980s. However, modelling of analysis. Slemon and Liu [65] proposed an approximate
magnetic fields was still going on [59, 60]. The FE method modelling of eddy-current losses based on analytical flux
for eddy current loss calculation did obtain some attention variation calculation in 1990. Liu and Binns [66] presented
[61 – 64]. a rather useful analytical model to predict eddy-current
We can term this period as a period of analysis and losses analytically. This was the outcome of past
formulation of new tools for problems related to eddy developments on analytical methods and ready verification
currents in rotating electrical machines. The analytical available in the form of FE analysis. Some research on
methods matured and the FE methods came into existence materials and geometrical research is reflected in [67, 68].
during this period. By the late 1980s, FE methods were still The analytical methods developed thus far are still being
limited by the computing capabilities and numerical used today with minor improvements. The most important
algorithms available at that time. This shortcoming was developments during this stage came from numerical and
soon overcome assisted by a revolution in computing in the FE methods for eddy current loss calculations.
early 1990s. In 1992– 1993, Mecrow andJack [69] showed improved
possibilities with non-linearity and 3D FE modelling.
5 Stage 3: 1991 onwards – FE, numerical and Furthermore, electrothermal coupling for eddy current loss
analytical methods calculation using analytical and numerical methods was
proposed [70].
The 1990s saw regenerated interest in eddy current losses The later part of the 1990s saw 3D FE modelling
owing to developments in permanent magnet (PM) techniques being developed as importance of end-effects in
machines, power electronics and concentrated windings. PM machines was being realised [71]. The Japanese had
A significant increase in the utility of PM machines and been working steadily on material properties of PM
combination of power electronic converters with induction material, laminations and steel during the 1990s but most of
machines brought a second wind to eddy-current loss the useful results from Japanese research started appearing

IET Electr. Power Appl., 2012, Vol. 6, Iss. 9, pp. 743– 752 747
doi: 10.1049/iet-epa.2011.0335 & The Institution of Engineering and Technology 2012
www.ietdl.org
in the early 2000s probably because of extensive use of 3D derivable quantities from the original variable solved. It
FE analysis which was better equipped to deal with eddy may be mentioned that because of discretisation of the
current problems now. Many PM machines were used with domain in the FE method, complex geometries and material
power electronic converters which added to eddy-current non-linearities can be easily handled. The FE method has
losses as well. Deng, Nehl and Polinder have analysed developed a lot since its inception into electrical
these effects in detail [72 – 76]. engineering. Various new algorithms and mathematical
techniques have been implemented to improve the
5.1 Numerical methods and FE formulation for modelling and post-processing of the FE method.
eddycurrent losses Description of various innovations in numerical methods is
out of the scope of this survey, which is given in Table 3.
There are many numerical methods which have been used to
solve electrical engineering problems. Power series or 5.2 Developments in eddy-current loss analysis –
successive approximation method and finite difference stage 3
method were the first major numerical methods to be used
for eddy current loss problems [34]. The power series Similar to the rapid increase in analytical methods observed in
method however failed to take into account the reaction the 1960s, another phenomenon that virtually exploded
field of eddy current losses properly which led to its decline in research during 2000 – 2010 was PM machines. Almost
[77]. On the other hand, the finite difference method, based 70% of the research of eddy current losses was conducted
on Taylor’s series was very popular before the advent of on PM machines of one type or another. The neodymium
the FE method. The finite difference method was limited magnets were available for cheaper prices and it led to a
because of the large number of simultaneous equations great impetus in the development of PM machines.
required to be solved. Unavailability of computing power However, this time around the tools, that is, methods to
available at that time (stage 1) led to a gradual decline in deal with eddy current loss problems were at hand.
the popularity of this method. Interested readers can refer to Therefore in this period, we see extensive use of analytical
Binns and Lawrenson [34] for details of this method. and FE methods to reach good machine designs [80 – 86].
However, from the literature survey, the primary and most Z.Q. Zhu, D. Ishak and D. Howe wrote many papers related
popular numerical method used in the field of electrical to the modelling of magnetic fields and eddy current losses
machines is the FE method, which we will discuss here in PM machines [87 – 90]. Although the techniques used
briefly. The FE method uses numerical techniques to solve were not new however the application was rather new.
partial differential equations but the problem is divided into It led to tremendous interest in PM machines for large
a ‘finite number of elements’ and requires minimisation power applications especially in the field of renewable
of the energy functional solved locally for each element. energy like Archimedes wave swing, PM direct drive
For eddy current problems the energy functional can take generators for wind turbines and so on. [91].
the form [62] Many new applications were identified with the advent of
PM machines. High speed machines without slots, fault
   2  
|B| 2 tolerant machines and switched reluctance machines and so
F= − J A + ipf sA
   (5) on indicated a new breed of applications where eddy-
V 2m
current loss analysis was deemed necessary [92 – 95]. At the
where B  is the magnetic field; A  is the magnetic vector same time a lot of research on investigating the eddy
potential; m is the permeability; v is the volume; s is the current losses in conductive neodymium magnets
electrical conductivity; J is the applied current density; and themselves was gathering momentum. 3D FE is the
f is the frequency. favourite tool when it comes to material characterisation
The mathematical problem to be solved is then because of difficulty in modelling material anisotropy and
minimisation of energy functional F, that is end effects analytically. PMs are conductive and therefore
modelling eddy current losses in magnets was imperative
∂F [96 – 98]. Yamazaki and Fukushima et al. have done a lot
 =0
∂A
(6) of work on eddy-current loss analysis in PMs [79, 99– 101].
In Europe also, a significant amount of work on material
The theory of FE is well known and since the paper is focused characterisation for steel laminations, solid iron samples and
on the literature survey of eddy-current losses, the solution so on has been done [102– 105].
methods are not explained here. After solution of A,  the With all these advancements, modelling of machines for
induced eddy current density can be found from new applications benefited and we can see a lot of
publications on the analysis of these special machines.
Markovic and Perriard have presented generalised models
Jind (x, y, z) = −i2pf s(x, y, z)A(x,
 y, z) (7)
for slotless PM machines in 2007– 2009, respectively,
[106, 107]. The low-power machines (fractional kW – few
Eddy current losses can be calculated easily by integrating
kW) can be made cheaper using concentrated windings
this current density over an area
with fractional pitch because of winding automation
 2 possibility. However, these types of windings also have
J ind (x, y, z) eddy current losses associated with them [108 – 112]. There
Pe = dV (8)
V s is a good literature survey on analytical machine modelling
mentioned in [113]. A very thorough literature survey on
Pe is the eddy current loss; Jind is the induced current density practical design aspects of concentrated winding machines
in A/m2; V is the volume; and s is the conductivity. can be found in [33].
The overall solution is compiled and can be easily Therefore any development in the sector of fractional pitch
displayed for a chosen quantity as well as some other concentrated winding machines led to the analysis of eddy

748 IET Electr. Power Appl., 2012, Vol. 6, Iss. 9, pp. 743 –752
& The Institution of Engineering and Technology 2012 doi: 10.1049/iet-epa.2011.0335
 www.ietdl.org
Table 3 Some numerical formulations used for eddy-current loss calculation

First author Problem formulation Loss calculation Methodology

R. Stoll [53] – 2D Cartesian power series

py/g −jpx/g
– vector potential A(1) 2 k
kl = (Kk0 + Kk1 y + Kk2 y + · · · + Kkk y ) × e e
Ak = kth order term in power series for A (magnetic vector potential)
Kk = coefficient
B.C. Mecrow – 3D Cartesian g = pole pitch FE
[69] – scalar magnetic * Eddy - current losses are then calculated from solution of A
potential
1 1 d m 
∇ × ∇ × TA = ∇ × ∇ × TB = − (TA + TB ) − ∇ V
–end effects s s dt 2
included TA , TB = potential functions in adjacent bodies
– linear material V = scalar magnetic potential
s = electrical conductivity
m = permeability
TA , TB and V are solved to get eddy - current density and losses

∂A s
∇ × (v ∇ × A) + s − u = ∇ × (vBr )
∂t lM mn
⎡   ⎤
N M  ∂A umn 2
⎢ sl + dV ⎥
Pe = ⎣ m=1 M Vmn ∂t lM ⎦
n=1 2 2 2
+i1n Rkn + (i1n + iMn )Rkn
Pe = eddy − current losses in magnets
W.N. Fu [78] – 2D Cartesian Circuit element
– vector magnetic A = magnetic vector potential axial component coupled with FE
potential v = reluctivity of the material
–linear material M, N = slices and bars cut out of the domain
–end effects umn = voltage between nodes of the nth bar and mth slice
neglected lM = axial length per slice of magnet
Rkn = terminal resistance between adjacent bars
   
1 ∂A 1
∇× ∇ × A = Ja − s + ∇w + ∇×M
m ∂t m0
  
∂A
∇ s + ∇w =0
∂t
dF
va = voa + Ra ia = + Ra ia
dt
b a
2d
Pe = (|Jx (x, y)|2 + |Jy (x, y)|2 ) dx dy
K. Yamazaki – 3D Cartesian s 0 0 FE + analytical
[79] – vector magnetic A, w = magnetic vector and electric scalar potential, respectively
potential m, s = permeability and conductivity, respectively
–linear material Ja = armature current density; M = Magnetisation of PM
va , Ra , ia = armature voltage, resistance and current, respectively
F = flux linkage of armature coil
a, b = dimensions
Jx (x, y), Jy (x, y) = x, y comp. of induced current density calculated analytically

current losses associated with them. The latest additions in 6 Results and trends
such machines are the ultra-high speed machines
(105 rpm or more) and PM direct drive wind turbines. From the literature survey, it is clear that the main
These machines utilise concentrated windings for cost methodologies used in the analysis of eddy current losses are:
benefits and also high power densities. For some
applications where 3D effects play a major role, analytical (a) Experimental: The analysis is carried out by experimental
methods are rather limited whereby it becomes necessary to observations and empirical formulae derived for eddy-current
resort to the 3D FE calculations. It is important to note that losses thereof.
in this period, both the analytical and FE methods have (b) Analytical: Starting from Maxwell’s equations magnetic
been used extensively. The probable reason could be that and electric field quantities are derived and used for
analytical methods give a quick insight into the dependence evaluating eddy current losses.
of design on various parameters which is good for (c) Numerical: This method is basically the same as the
optimisation. FE methods are more accurate however they analytical method with the exception that instead of a
take a longer time to solve. Thus, a combination of these mathematical expression, the derived partial differential
methods is the preferred way at present. equations are solved numerically.

IET Electr. Power Appl., 2012, Vol. 6, Iss. 9, pp. 743– 752 749
doi: 10.1049/iet-epa.2011.0335 & The Institution of Engineering and Technology 2012
www.ietdl.org
(d) FE: This is an advancement of numerical methods where
the whole space is divided into a very fine mesh and the
quantities are derived by solving for each region
numerically. A big advantage is that any geometry, motion
and non-linearity can be incorporated at the same time
whereas a probable disadvantage is the long time taken to
solve and requirement of specialised software which is
rather expensive.
(e) Equivalent circuits: The machine is modelled in terms of
lumped elements arranged in an equivalent circuit depending
on machine geometry. The eddy current losses are then
derived from this equivalent circuit.

Each of these methods has played an important part in the

development of the theory of eddy-current loss analysis.
However, the popularity and utility of the methods have
been changing over time. These trends have been deduced Fig. 3 Eddy currents loss analysis –focus on type of machine
based on this extensive literature survey. The utility
percentage shown in Fig. 2 is based on the number of
papers used for this survey. It is not an exact number (f) PM machines: This is a general term and there can be
however, the trend projected by these numbers is indicative many different types of PM machines. In this survey this
of the actual utility. refers to the general class of PM machines, which use PM
Furthermore, it can be seen from the literature survey that as field excitation.
the research focus on eddy current losses in the field of (g) Special: These machines appeared in the late 1990s and
electrical machines shifted in different time periods. switched reluctance motors, single phase motors, printed
The main trend of research and the types of machines circuit board motors, flux switching machines and so on fall
studied can be grouped as under this category.

It may also be mentioned here that the percentage of focus

(a) General: This means development of methodology and on research is relative and based on the number of papers used
analysis tools related to eddy-current loss analysis. in this survey only. The purpose here is to present trends
(b) Materials: This indicates the research which did not focus rather than numbers. Fig. 3 shown below projects the focus
on any particular type of machine but on the materials which of research during a given time period relative to other
can be used in the machines. For example, research on eddy machine types in question.
current losses in sheet steel laminations and solid iron.
(c) Conventional DC machines: These machines need no
7 Conclusions
introduction but in this survey, these machines refer to
electrically excited DC machines with rotating armatures Eddy-current loss analysis in rotating electrical machines is an
and commutator. old and intriguing field. It has been constantly developing for
(d) Conventional induction machines: These include squirrel more than a century. In this paper, a broad perspective on the
cage and wound rotor type of induction machines with or field of eddy current loss analysis in rotating electrical
without power electronic drive connected to them. machinery has been presented. A variety of methods have
(e) Conventional synchronous machines: This term refers to been utilised over the last century for analysis of these
the synchronous machines with electrically excited field losses. These methods have been identified and various
winding on the rotor and distributed stator windings in the trends in terms of methodology used and the type of
stator. machines studied during different time periods have been
deduced. A chronological reference of important (but not
all) contributions from eminent scientists and engineers has
been documented. This paper should serve as a reference
guide for any person interested in the field of eddy current
losses in rotating electrical machines.

8 References
1 Steinmetz, C.P.: ‘On the law of hysteresis (originally from 1892)’, Proc.
IEEE, 1984, 72, (2), pp. 197– 221
2 Pry, R.H., Bean, C.P.: ‘Calculation of the energy loss in magnetic sheet
materials using a domain model’, J. Appl. Phys., 1958, 29, (3),
pp. 532–533
3 Bertotti, G.: ‘General properties of power losses in soft ferromagnetic
materials’, IEEE Trans. Magn., 1988, 24, (1), pp. 621– 630
4 Jacobs, S., Hectors, D., Henrotte, F., et al.: ‘Magnetic material
optimization for hybrid vehicle PMSM drives’. EVS24 – Int. Battery
Hybrid and Fuel Cell Electric Vehicle Symp., Stavanger, Norway, 2009
5 Thomson, J.J.: ‘On the heat produced by eddy currents in an iron plate
exposed to alternating magnetic field’, The Electrician, 1892, 28
Fig. 2 Eddy currents loss analysis methods in electrical machines 6 Tesla, N.: ‘Electro magnetic motor’. US Patent 0381968, May 1888

750 IET Electr. Power Appl., 2012, Vol. 6, Iss. 9, pp. 743 –752
& The Institution of Engineering and Technology 2012 doi: 10.1049/iet-epa.2011.0335
 www.ietdl.org
7 Bowers, B.: ‘Scanning our past from London: Galileo Ferraris and 40 Wood, A.J.: ‘An analysis of solid rotor machines part I. Operational
alternating current’, Proc. IEEE, 2001, 89, (5), pp. 790–792 impedances and equivalent circuits’, Trans. Am. Inst. Electr. Eng.
8 Field, A.B.: ‘Eddy currents in large slot-wound conductors’, Trans. Am. Part III, Power Appar. Syst, 1959, 78, (4), pp. 1657–1665
Inst. Electr. Eng., 1905, XXIV, pp. 761–788 41 Wood, A.J., Concordia, C.: ‘An analysis of solid rotor machines part II.
9 Adams, A., Lanier, A.C., Pope, C.C., Schooley, C.O.: ‘Pole-face losses’, The effects of curvature’, Trans. Ame. Inst. Electr. Eng. Part III, Power
Trans. Am. Inst. Electr. Eng., 1909, XXVIII, (2), pp. 1133–1156 Appar. Syst, 1959, 78, (4), pp. 1666–1672
10 Press, A.: ‘Incremental armature copper losses at no-load and armature 42 Wood, A.J., Concordia, C.: ‘An analysis of solid rotor machines part III.
teeth eddy-current losses’, J. Ins. Electr. Eng., 1915, 53, (250), Finite length effects’, Trans. Am. Inst. Electr. Eng. Part III, Power
pp. 820–823 Appar. Syst, 1960, 79, (3), pp. 21–26
11 Hughes, E.: ‘Iron losses in d.c. machines’, J. Inst. Electr. Eng., 1924, 63, 43 Wood, A.J., Concordia, C.: ‘An analysis of solid rotor machines part IV.
(336), pp. 35– 50 An approximate nonlinear analysis’, Trans. Am. Inst. Electr. Eng. Part
12 Gilman, R.E.: ‘Eddy current losses in armature conductors’, Trans. Am. III, Power Appar. Syst., 1960, 79, (3), pp. 26– 31
Inst. Electr. Eng., 1924, XLIII, pp. 246– 251 44 Hammond, P.: ‘The calculation of the magnetic field of rotating
13 Townsend, F.: ‘Eddy current losses in transformers’, Trans. Am. Inst. machines. Part 3: eddy currents induced in a solid slab by a circular
Electr. Eng., 1900, XVII, pp. 319–334 current loop’, Proc. IEE, Part C, Monogr., 1962, 109, (16),
14 Field, M.B.: ‘Eddy currents in solid and laminated masses’, J. Inst. pp. 508–515
Electr. Eng., 1904, 33, (169), pp. 1125–1143 45 Jones, D.A., Leung, W.S.: ‘A theoretical and analogue approach to stray
15 Hanssen, I.E.: ‘Calculation of iron losses in dynamo electric machinery’, eddy-current loss in laminated magnetic cores’, Proc. IEE, Part C,
Trans. Am. Inst. Electr. Eng., 1909, XXVIII, (2), pp. 993 –1001 Monogr., 1961, 108, (14), pp. 509–515
16 Taylor, H.W.: ‘Eddy currents in stator windings’, J. Inst. Electr. Eng., 46 Stoll, R.L.: ‘Numerical method of calculating eddy currents in
1920, 58, (290), pp. 279 –298 nonmagnetic conductors’, Proc. Inst. Electr. Eng., 1967, 114, (6),
17 Spooner, T.: ‘Squirrel-cage induction-motor core losses’, Trans. Am. pp. 775–780
Inst. Electr. Eng., 1925, XLIV, pp. 155 –163 47 Oberretl, K.: ‘Magnetic fields, Eddy currents, and losses, taking the
18 Lloyd, M.G., Fisher, J.V.S.: ‘The testing of transformer steel’, Trans. variable permeability into account’, IEEE Trans. Power Appar. Syst.,
Am. Inst. Electr. Eng., 1909, XXVIII, (1), pp. 439–467 1969, PAS-88, (11), pp. 1646– 1657
19 Scott, K.L.: ‘Variation of the inductance of coils due to the magnetic 48 Lawrenson, P.J., Reece, P., Ralph, M.C.: ‘Tooth-ripple losses in solid
shielding effect of eddy currents in the cores’, Proc. Inst. Radio Eng., poles’, Proc. Inst. Electr. Eng., 1966, 113, (4), pp. 657– 662
1930, 18, (10), pp. 1750–1764 49 Jamieson, R.A.: ‘Eddy-current effects in solid unslotted iron rotors’,
20 Peterson, E., Wrathall, L.R.: ‘Eddy currents in composite laminations’, Proc. Inst. Electr. Eng., 1969, 116, (4), pp. 612– 613
Proc. Ins. Radio Eng., 1936, 24, (2), pp. 275–286 50 Carpenter, C.J., Djurovic, M.: ‘Three-dimensional numerical solution of
21 Reed, M.: ‘An experimental investigation of the theory of eddy currents eddy currents in thin plates’, Proc. Inst. Electr. Eng., 1975, 122, (6),
in laminated cores of rectangular section’, Proc. Wirel Sect., Inst. Elect. pp. 681–688
Eng., 1937, 12, (35), pp. 167–178 51 Lim, K.K., Hammond, P.: ‘Universal loss chart for the calculation of
22 Blake, L.R.: ‘The eddy-current anomaly in ferromagnetic laminae at eddy-current losses in thick steel plates’, Proc. Inst. Electr. Eng.,
high rates of change of flux’, Proc. IEE – Part II, Power Eng., 1949, 1970, 117, (4), pp. 857– 864
96, (53), pp. 705–718 52 Stoll, R.L.: ‘Solution of linear steady-state eddy-current problems by
23 Stewart, K.H.: ‘Losses in electrical sheet steel’, Proc. IEE – Part II, complex successive overrelaxation’, Proc. Inst. Electr. Eng., 1970,
Power Eng., 1950, 97, (56), pp. 121– 125 117, (7), pp. 1317– 1323
24 Bewley, L.V., Poritsky, H.: ‘Intersheet eddy-current loss in laminated 53 Stoll, R.L., Muhlhaus, J.: ‘Power-series method of calculating eddy
cores’, Trans. Am. Inst. Electr. Eng., 1937, 56, (3), pp. 344– 346 currents in nonmagnetic conductors’, Proc. Inst. Electr. Eng., 1972,
25 Concordia, C., Poritsky, H.: ‘Synchronous machine with solid 119, (11), pp. 1616–1620
cylindrical rotor’, Trans. Am. Inst. Elect. Eng., 1937, 56, (1), 54 Chari, M.V.K.: ‘Finite-element solution of the eddy-current problem in
pp. 49–179 magnetic structures’, IEEE Trans. Power Appar. Syst., 1974, PAS-93,
26 Aston, K., Kesava Rao, M.V.: ‘Pole-face losses due to open slots’, J. Ins. (1), pp. 62–72
Elect. Eng. – Part I, General, 1941, 88, (8), pp. 308–312 55 Chari, M.V.K., Reece, P.: ‘Magnetic field distribution in solid metallic
27 Gibbs, W.J.: ‘Tooth-ripple losses in unwound pole-shoes’, J. Ins. Elect. structures in the vicinity of current carrying conductors, and
Eng. – Part II, Power Eng., 1947, 94, (37), p. 2 associated eddy-current losses’, IEEE Trans. Power Appar. Syst.,
28 Carter, F.W.: ‘Pole-face losses’, J. Ins. Elect. Eng., 1916, 54, (253), 1974, PAS-93, (1), pp. 45– 56
pp. 168–170 56 Carpenter, C.J.: ‘Finite-element network models and their application to
29 Hague, B.: ‘Electromagnetic problems in electrical engineering’ (Oxford eddy-current problems’, Proc. Inst. Electr. Eng., 1975, 122, (4),
University Press, 1929), Chapter IV pp. 455–462
30 Subba Rao, V.: ‘Losses in ferromagnetic laminations due to saturation’, 57 Stoll, R.L., Hindmarsh, R.: ‘Reduction of local eddy-current losses in
Proc. Ins. Elect. Eng., 1964, 111, (12), pp. 2111– 2117 turbogenerator stator windings using composite subconductors’, Proc.
31 Subba Rao, V.: ‘Eddy-current losses in finite sections of solid iron under Inst. Electr. Eng., 1976, 123, (11), pp. 1217–1222
saturation’, Proc. Ins. Elect. Eng., 1964, 111, (2), pp. 343– 348 58 Carpenter, C.: ‘A network approach to the numerical solution of eddy-
32 Mukherji, K.C.: ‘Reaction of eddy currents induced in a ferromagnetic current problems’, IEEE Trans. Mag., 1975, 11, (5), pp. 1517– 1522
medium on the inducing field’, Proc. Inst. Elect. Eng., 1965, 112, (2), 59 Sironi, G., Van Hulse, J.: ‘Eddy loss calculation in solid magnetic pieces
pp. 444–445 of electrical machines’, IEEE Trans. Mag., 1975, 11, (5),
33 Bianchi, N., Bolognani, S., Fornasiero, E.: ‘An overview of rotor losses pp. 1523– 1525
determination in three-phase fractional-slot PM machines’, IEEE Trans. 60 Yannopoulos-Lascaratos, P.P., Tegopoulos, J.A.: ‘Eddy-current
Ind. Appl., 2010, 46, (6), pp. 2338– 2345 distribution in cylindrical structures caused by rotating magnetic
34 Binns, K.J., Lawrenson, P.J.: ‘Analysis and computation of electric fields’, IEE Proc. B, Electr. Power Appl., 1982, 129, (2), pp. 64–74
and magnetic field problems’ (Pergamon Press, 1963), pp. 7– 36, 61 Demerdash, N.A., Nehl, T.W., Mohammed, O.A., Miller, R.H.,
251– 302 Fouad, F.A.: ‘Solution of eddy current problems using three
35 Bondi, H., Mukherji, K.C.: ‘An analysis of tooth-ripple phenomena in dimensional finite element complex magnetic vector potential’, IEEE
smooth laminated pole-shoes’, Proc. IEE, Part C, Monogr., 1957, Trans. Power Appar. Syst., 1982, PAS-101, (11), pp. 4222 – 4229
104, (6), pp. 349–356 62 Brauer, J.: ‘Finite element calculation of eddy currents and skin effects’,
36 Kesavamurthy, N., Rajagopalan, P.K.: ‘An analytical method taking IEEE Trans. Mag., 1982, 18, (2), pp. 504– 509
account of saturation and hysteresis for evaluating the iron loss in 63 Ferrari, R., Pinchuk, A.: ‘Complementary variational finite-element
solid-iron cores subjected to an alternating field’, Proc. IEE, Part C, solution of eddy current problems using the field variables’, IEEE
Monogr., 1962, 109, (15), pp. 237–243 Trans. Mag., 1985, 21, (6), pp. 2242–2245
37 Stoll, R.L., Hammond, P.: ‘Calculation of the magnetic field of rotating 64 Jack, A.G., Mecrow, B.C.: ‘Methods for magnetically nonlinear
machines. Part 4: approximate determination of the field and the losses problems involving significant hysteresis and eddy currents’, IEEE
associated with eddy currents in conducting surfaces’, Proc. Inst. Electr. Trans. Mag., 1990, 26, (2), pp. 424– 429
Eng., 1965, 112, (11), pp. 2083–2094 65 Slemon, G.R., Liu, X.: ‘Core losses in permanent magnet motors’, IEEE
38 Lawrenson, P.J., Ralph, M.C.: ‘General 3-dimensional solution of Trans. Mag., 1990, 26, (5), pp. 1653–1655
eddy-current and laplacian fields in cylindrical structures’, Proc. Inst. 66 Liu, Z.J., Vourdas, A., Binns, K.J.: ‘Magnetic field and eddy current
Electr. Eng., 1970, 117, (2), pp. 467– 472 losses in linear and rotating permanent magnet machines with a large
39 Agarwal, P.D.: ‘Eddy-current losses in solid and laminated iron’, AIEE number of poles’, IEE Proc. A, Sci. Meas. Technol., 1991, 138, (6),
Trans., 1959, 78, p. 169 pp. 289–294

IET Electr. Power Appl., 2012, Vol. 6, Iss. 9, pp. 743– 752 751
doi: 10.1049/iet-epa.2011.0335 & The Institution of Engineering and Technology 2012
www.ietdl.org
67 Rabinovici, R.: ‘Nonlinear eddy current effects in ferromagnetic 90 Amara, Y., Wang, J., Howe, D.: ‘Analytical prediction of eddy-current
materials in the presence of DC magnetic fields’, IEEE Trans. Mag., loss in modular tubular permanent-magnet machines’, IEEE Trans.
1991, 27, (4), pp. 3704– 3709 Energy Convers., 2005, 20, (4), pp. 761–770
68 Arkadan, A.A., Vyas, R., Vaidya, J.G., Shah, M.J.: ‘Effect of toothless 91 Polinder, H., Mecrow, B.C., Jack, A.G., Dickinson, P.G., Mueller, M.A.:
stator design and core and stator conductors eddy current losses in ‘Conventional and TFPM linear generators for direct-drive wave energy
permanent magnet generators’, IEEE Trans. Energy Convers., 1992, conversion’, IEEE Trans. Energy Convers., 2005, 20, (2), pp. 260–267
7, (1), pp. 231– 237 92 Mecrow, B.C., Jack, A.G., Haylock, J.A., Coles, J.: ‘Fault-tolerant
69 Mecrow, B.C., Jack, A.G.: ‘The modelling of segmented laminations in permanent magnet machine drives’, IEE Proc. Electric Power Appl.,
three dimensional eddy current calculations’, IEEE Trans. Mag., 1992, 1996, 143, (6), pp. 437– 442
28, (2), pp. 1122–1125 93 Hor, P.J., Zhu, Z.Q., Howe, D., Rees-Jones, J.: ‘Eddy current loss in a
70 Liu, Z.J., Binns, K.J., Low, T.S.: ‘Analysis of eddy current and thermal moving-coil linear tubular permanent magnet brushless motor’, IEEE
problems in permanent magnet machines with radial-field topologies’, Trans. Mag., 1999, 35, (5), pp. 3601–3603
IEEE Trans. Mag., 1995, 31, (3), pp. 1912–1915 94 Atkinson, G.J., Mecrow, B.C., Jack, A.G., Atkinson, D.J., Sangha, P.,
71 Bobbio, S., Del Pizzo, A., Marignetti, F., Pagano, E.: ‘Eddy current iron Benarous, M.: ‘The analysis of losses in high-power fault-tolerant
losses in axially laminated brushless motors’, IEE Proc. Electr. Power machines for aerospace applications’, IEEE Trans. Ind. Appl., 2006,
Appl., 1995, 142, (3), pp. 183– 190 42, (5), pp. 1162–1170
72 Deng, F.: ‘Commutation-caused eddy-current losses in permanent- 95 Klauz, M., Dorrell, D.G.: ‘Eddy current effects in a switched reluctance
magnet brushless DC motors’, IEEE Trans. Mag., 1997, 33, (5), motor’, IEEE Trans. Mag., 2006, 42, (10), pp. 3437–3439
pp. 4310– 4318 96 Kawase, Y., Yamaguchi, T., Sano, S., Igata, M., Ida, K., Yamagiwa, A.:
73 Polinder, H.: ‘On the losses in a high-speed permanent-magnet generator ‘3-D eddy current analysis in a silicon steel sheet of an interior permanent
with rectifier with special attention to the effect of damper cylinder’. magnet motor’, IEEE Trans. Mag., 2003, 39, (3), pp. 1448–1451
PhD thesis, TU Delft, The Netherlands, 1998 97 Aoyama, Y., Miyata, K., Ohashi, K.: ‘Simulations and experiments on
74 Deng, F., Nehl, T.W.: ‘Analytical modeling of eddy-current losses eddy current in Nd-Fe-B magnet’, IEEE Trans. Mag., 2005, 41, (10),
caused by pulse-width-modulation switching in permanent-magnet pp. 3790– 3792
brushless direct-current motors’, IEEE Trans. Mag., 1998, 34, (5), 98 Yamazaki, K., Shina, M., Miwa, M., Hagiwara, J.: ‘Investigation of
pp. 3728– 3736 eddy current loss in divided Nd –Fe–B sintered magnets for
75 Liu, R., Mi, C.C., Gao, D.W.: ‘Modeling of eddy-current loss of synchronous motors due to insulation resistance and frequency’, IEEE
electrical machines and transformers operated by pulsewidth- Trans. Mag., 2008, 44, (11), pp. 4269–4272
modulated inverters’, IEEE Trans. Mag., 2008, 44, (8), pp. 2021– 2028 99 Yamazaki, K., Abe, A.: ‘Loss investigation of interior permanent-
76 Okitsu, T., Matsuhashi, D., Muramatsu, K.: ‘Method for evaluating the magnet motors considering carrier harmonics and magnet eddy
eddy current loss of a permanent magnet in a PM motor driven by an currents’, IEEE Trans. Ind. Appl., 2009, 45, (2), pp. 659– 665
inverter power supply using coupled 2-D and 3-D finite element 100 Yamazaki, K., Fukushima, Y., Sato, M.: ‘Loss analysis of permanent-
analyses’, IEEE Trans. Mag., 2009, 45, (10), pp. 4574–4577 magnet motors with concentrated windings – variation of magnet
77 Stoll, R.L.: ‘The analysis of eddy currents’ (Oxford University Press, eddy-current loss due to stator and rotor shapes’, IEEE Trans. Ind.
1974), pp. 73– 91 Appl., 2009, 45, (4), pp. 1334– 1342
78 Fu, W.N., Liu, Z.J.: ‘Estimation of eddy-current loss in permanent 101 Yamazaki, K., Ishigami, H.: ‘Rotor-shape optimization of interior-
magnets of electric motors using network-field coupled multislice permanent-magnet motors to reduce harmonic iron losses’, IEEE
time-stepping finite-element method’, IEEE Trans. Mag., 2002, 38, Trans. Ind. Electron., 2010, 57, (1), pp. 61– 69
(2), pp. 1225– 1228 102 Sergeant, P., Van den Bossche, A.: ‘Segmentation of magnets to reduce
79 Yamazaki, K., Kanou, Y., Fukushima, Yu., et al.: ‘Reduction of magnet losses in permanent-magnet synchronous machines’, IEEE Trans.
eddy-current loss in interior permanent-magnet motors with Mag., 2008, 44, (11), pp. 4409– 4412
concentrated windings’, IEEE Trans. Ind. Appl., 2010, 46, (6), 103 Chebak, A., Viarouge, P., Cros, J.: ‘Analytical computation of the
pp. 2434– 2441 full load magnetic losses in the soft magnetic composite stator of
80 Polinder, H., Hoeijmakers, M.J.: ‘Eddy-current losses in the segmented high-speed slotless permanent magnet machines’, IEEE Trans. Mag.,
surface-mounted magnets of a PM machine’, IEE Proc. Electr. Power 2009, 45, (3), pp. 952– 955
Appl., 146, (3), pp. 261–266 104 Shah, M.R., EL-Refaie, A.M.: ‘Eddy-current loss minimization in
81 Kawase, Y., Ota, T., Fukunaga, H.: ‘3-D eddy current analysis in conducting sleeves of surface PM machine rotors with fractional-slot
permanent magnet of interior permanent magnet motors’, IEEE Trans. concentrated armature windings by optimal axial segmentation and
Mag., 2000, 36, (4), pp. 1863–1866 copper cladding’, IEEE Trans. Ind. Appl., 2009, 45, (2), pp. 720 –728
82 Atallah, K., Howe, D., Mellor, P.H., Stone, D.A.: ‘Rotor loss in 105 Ruoho, S., Santa-Nokki, T., Kolehmainen, J., Arkkio, A.: ‘Modeling
permanent-magnet brushless AC machines’, IEEE Trans. Ind. Appl., magnet length in 2-D finite-element analysis of electric machines’,
2000, 36, (6), pp. 1612– 1618 IEEE Trans. Mag., 2009, 45, (8), pp. 3114–3120
106 Markovic, M., Perriard, Y.: ‘Analytical solution for rotor eddy-current
83 Drubel, O., Stoll, R.L.: ‘Comparison between analytical and numerical
methods of calculating tooth ripple losses in salient pole synchronous losses in a slotless permanent-magnet motor: the case of current sheet
machines’, IEEE Trans. Energy Convers., 2001, 16, (1), pp. 61– 67 excitation’, IEEE Trans. Mag., 2008, 44, (3), pp. 386–393
107 Markovic, M., Perriard, Y.: ‘An analytical determination of
84 Nerg, J., Niemela, M., Pyrhonen, J., Partanen, J.: ‘FEM calculation of
eddy-current losses in a configuration with a rotating permanent
rotor losses in a medium speed direct torque controlled PM
magnet’, IEEE Trans. Mag., 2007, 43, (8), pp. 3380–3386
synchronous motor at different load conditions’, IEEE Trans. Mag.,
108 Nakano, M., Kometani, H., Kawamura, M.: ‘A study on eddy-current
2002, 38, (5), pp. 3255– 3257
losses in rotors of surface permanent-magnet synchronous machines’,
85 Mi, C., Slemon, G.R., Bonert, R.: ‘Modeling of iron losses of
IEEE Trans. Ind. Appl., 2006, 42, (2), pp. 429–435
permanent-magnet synchronous motors’, IEEE Trans. Ind. Appl.,
109 Wang, J., Atallah, K., Chin, R., Arshad, W.M., Lendenmann, H.:
2003, 39, (3), pp. 734–742
‘Rotor eddy-current loss in permanent-magnet brushless AC
86 Bottauscio, O., Pellegrino, G., Guglielmi, P., Chiampi, M., Vagati, A.:
machines’, IEEE Trans. Mag., 2010, 46, (7), pp. 2701–2707
‘Rotor loss estimation in permanent magnet machines with
110 Han, S.-H., Jahns, T.M., Zhu, Z.Q.: ‘Analysis of rotor core
concentrated windings’, IEEE Trans. Mag., 2005, 41, (10),
eddy-current losses in interior permanent-magnet synchronous
pp. 3913– 3915
machines’, IEEE Trans. Ind. Appl., 2010, 46, (1), pp. 196–205
87 Toda, H., Xia, Z., Wang, J., Atallah, K., Howe, D.: ‘Rotor eddy-current 111 Amara, Y., Reghem, P., Barakat, G.: ‘Analytical prediction of
loss in permanent magnet brushless machines’, IEEE Trans. Mag., 2004, eddy-current loss in armature windings of permanent magnet
40, (4), pp. 2104–2106
brushless AC machines’, IEEE Trans. Mag., 2010, 46, (8),
88 Zhu, Z.Q., Ng, K., Schofield, N., Howe, D.: ‘Improved analytical pp. 3481–3484
modelling of rotor eddy current loss in brushless machines equipped 112 Yamazaki, K., Fukushima, Y.: ‘Effect of eddy-current loss reduction
with surface-mounted permanent magnets’, IEE Proc Electr. Power by magnet segmentation in synchronous motors with concentrated
Appl., 2004, 151, (6), pp. 641– 650 windings’, IEEE Trans. Ind. Appl., 2011, 47, (2), pp. 779– 788
89 Ishak, D., Zhu, Z.Q., Howe, D.: ‘Eddy-current loss in the rotor magnets 113 Pfister, P.D., Perriard, Y.: ‘Slotless permanent-magnet machines:
of permanent-magnet brushless machines having a fractional number of general analytical magnetic field calculation’, IEEE Trans. Mag.,
slots per pole’, IEEE Trans. Mag., 2005, 41, (9), pp. 2462–2469 2011, 47, (6), pp. 1739–1752

752 IET Electr. Power Appl., 2012, Vol. 6, Iss. 9, pp. 743 –752
& The Institution of Engineering and Technology 2012 doi: 10.1049/iet-epa.2011.0335

View publication stats