Progress In Electromagnetics Research, PIER 83, 435–445, 2008

EDDY CURRENTS IN LAMINATED RECTANGULAR

CORES

S. K. Mukerji, M. George, M. B. Ramamurthy

and K. Asaduzzaman
Faculty of Engineering & Technology
Multimedia University
75450, Melaka, Malaysia

Abstract—A simplified expression for the eddy current loss in

laminated rectangular core is obtained using linear electromagnetic
field analysis. The treatment takes cognizance of current interruption
phenomena, by considering capacitive effects of insulation regions.
Analysis presented in this paper assumes identical field distribution
in each lamination and ignores eddy currents in insulation regions.

1. INTRODUCTION

Eddy currents are induced in conductors subjected to transient

electromagnetic fields [1–7]. Eddy currents are also produced due to
periodically time-varying excitations [8–17]. Alternating magnetic flux
in transformer cores induces eddy currents resulting in eddy current
loss. Time-invariant magnetic flux is established in the poles of a
synchronous machine operating under steady state conditions. The
pole-shoes, however, are subjected to pulsating magnetic field due to
slotted armature surface. To reduce eddy current loss, transformer
cores and pole-shoes of synchronous machines are invariably laminated.
Eddy current phenomena in laminated cores have been studied [8–
11] using electromagnetic field analysis. For this purpose the laminated
core is substituted by an equivalent homogeneous core with anisotropic
conductivity. The treatment is simple and results are concise.
However, the values of conductivity in the directions parallel and
perpendicular to laminations are to be selected rather empirically.
Theoretical and experimental investigations of tooth-ripple
phenomena in laminated pole-shoes are reported [12–17] in literature.
In their treatment for eddy currents in laminated pole-shoes,
authors [15, 16] have considered infinite half-space filled with identical
436 Mukerji et al.

laminations of arbitrary thickness. It is assumed that insulation regions

of negligible thickness, restrains eddy currents in one lamination from
flowing into another. Resulting equations for eddy current loss are
quite complicated. Simplified version of these equations has been also
reported [17].
In an earlier paper [18], it has been indicated that if a high-
resistive region is introduced in the middle of a conductive core,
it amounts to insertion of distributed capacitance in eddy current
path. This reduces eddy current and eddy current loss. The
current-interruption phenomena, thus conceived, have been taken
into cognizance in developing the expression for eddy current loss in
rectangular laminated cores.
In a laminated core, the field distribution in a lamination depends
on the position of the lamination in the stack. The variation in field
distribution from lamination to lamination is due to the finite stack-
thickness. In a core with large stack-thickness, this variation is small
for laminations located near the middle of the stack. The general
solution that takes cognizance of the finite stack-thickness effects is
lengthy and the resulting equation for eddy current loss is indeed
complicated. It can, however, be used for computer aided optimization
studies.
A simplified treatment for eddy currents in laminated cores is
presented here. It is assumed that the core consists of a large number of
laminations so that the field distribution in each lamination is identical.
Further, the simplified treatment ignores eddy currents in insulation
regions by setting zero conductivity for these regions. The advantage
of the analytical approach developed here is that it provides a better
understanding of eddy current loss over larger range of parameter
values.

2. FIELD EQUATIONS

Consider a rectangular core consisting of n insulated laminations, each

of width W and overall thickness T . Let the insulation thickness
on each side of a lamination be T1 /2 and its iron thickness be T2 .
Further, let the corners of the core be located at (−W/2, 0), (W/2, 0),
(−W/2, nT ) and (W/2, nT ), as shown in Fig. 1. In this figure,
insulation regions are indicated as Region-0’, 1’, 2’, 3’, . . . , m
, . . . ,
n
. The iron regions are indicated as Region-1, 2, 3,. . . , m, . . . , n.
The exciting coil wound around the long rectangular core, carrying
alternating current
i = Iejωt (1)
Progress In Electromagnetics Research, PIER 83, 2008 437

is simulated by a surface current density with a peak value which is

modulus of the complex quantity:

Jo = I · N (2)

where N is the number of turns per unit length of the coil. The current
carrying coil will produce time varying magnetic filed in the core,
eddy currents in the conducting regions and displacement currents
in the insulation regions of the core. The magnetic field outside
the coil is neglected. For the long rectangular core with uniformly
distributed current sheet, the magnetic field is entirely axial and
independent of z-coordinate, along the axial direction. It is assumed
that the permeability µ for the iron regions, permittivity ε, for the
insulation regions, and conductivity (σ, σ
), for both types of regions,
are constant. Thus from Maxwell’s equations for harmonic fields, in
charge-free regions:

∂ 2 Hz ∂ 2 Hz
+ = γ 2 Hz (3)
∂x2 ∂y 2
for iron regions, where

γ= (−jωµ)(σ + jωε0 ) (3.1)

and
∂ 2 Hz ∂ 2 Hz  2
+ = − γ
Hz (4)
∂x2 ∂y 2
for insulation regions, where

γ
= (−jωµ0 ) (σ
+ jωε) (4.1)

For perfect insulation, σ’ is zero.

Solutions of Eqs. (3) and (4) can be used to determine the
components of electric fields in iron and insulation regions, since for
iron regions:
1 ∂Hz
Ex = (5.1)
σ + jωε0 ∂y
and
1 ∂Hz
Ey = − (5.2)
σ + jωε0 ∂z
438 Mukerji et al.

(-w/2,nT)
Kx = -J 0 (w/2,nT)

T1/2 REGION - n'

T2 REGION - n'

T1 REGION - (n-1)'

REGION - (n-1)

REGION - (n-2)'

REGION - m'
nT
Ky = -J0 REGION - m K y = J0

(n-1)T
REGION - (m-1)'

REGION - 2' mT

REGION - 2

T1 REGION - 1'
2T

T2 Y REGION - 1 T

T1/2 REGION - 0'

(-w/2,0) K x = J0 X (w/2,0)

Figure 1. Cross-sectional view of a laminated rectangular core.

While, for insulation region:

1 ∂Hz
Ex = (6.1)
jωε ∂y
and
1 ∂Hz
Ey = − (6.2)
jωε ∂x
Progress In Electromagnetics Research, PIER 83, 2008 439

3. FIELD DISTRIBUTIONS

Consider Fig. 1. In view of Eq. (3), the distribution of magnetic field

in the conducting region-m can be given as:
 
pπ
cos x ∞  
W cosh αp (y − mT + T /2) pπ
Hzm = J0  + ap cos x (7)
pπ cosh(αp T2 /2) W
cos x q=1
W
where,
p = 2q − 1, (7.1)
indicating odd integer numbers, and
 2
pπ
αp = − γ2 (7.2)
W

while ap indicates a set of arbitrary constants.

The components of electric field in this region are
∞  
1 sinh αp (y − mT + T /2) pπ
Exm = ap αp cos x (8.1)
σ + jωε0 q=1 cosh(αp T2 /2) W

and
J0 γ sin(γx)
Eym =
σ + jωε0 cos(γW/2)
∞
  
pπ/W cosh αp (y − mT + T /2) pπ
+ ap sin x (8.2)
q=1
σ + jωε0 cosh(αp T2 /2) W

The magnetic field distribution in the non-conducting region-m
,

in view of Eq. (4), can be given as:
∞  
cos (γ
x)  cosh αp
(y − mT ) pπ
Hzm
= J0 + b p
cos x (9)


cos (γ W/2) q=1 cosh α
T1 /2 p
W

where,  2
pπ
αp
= − (γ
)2 (9.1)
W
440 Mukerji et al.

while bp indicates a set of arbitrary constants. The components of

electric field in this region are given as:
∞  
 αp
sinh αp
(y − mT ) pπ
Exm
= bp

cos x (10.1)
q=1
jωε cosh(αp T1 /2) W

and
∞  
γ
sin (γ
x)  pπ/W cosh αp
(y − mT ) pπ
Eym
= J0 + bp
sin x
jωε cos (γ
W/2) q=1 jωε

cosh α T1 /2 W
p
(10.2)

Now, since
Hzm = Hzm
at y = mT − T1 /2, over − W/2 ≤ x ≤ W/2 (11.1)
and
Exm = Exm
at y = mT − T1 /2 , over − W/2 ≤ x ≤ W/2 (11.2)

Therefore one gets:

  
4 pπ (γ
W/π) (γ W/π)
ap − bp = J0 sin 2 − (12.1)
π 2 p2 − (γ
W/π) p2 − (γ W/π)2
and
αp αp

tanh(αp T2 /2) ap = tanh(αp
T1 /2) bp (12.2)
σ + jωε0 jωε
Arbitrary constants found by solving these equations are:
Fp αp

ap = tanh(αp
T1 /2) (13.1)
jωε
and
Fp αp
bp = tanh(αp T2 /2) (13.2)
σ + jωε0
where,
  
4 pπ γ
γ
J0 sin
2
−
W 2 (αp ) (αp )2
Fp =

(14)
αp tanh(αp T1 /2) αp tanh(αp T2 /2)
−
jωε σ + jωε0
Progress In Electromagnetics Research, PIER 83, 2008 441

4. EDDY CURRENT LOSS

Using Poynting theorem, the complex power input per unit core-length,
for n-number of laminations is:
 
mT−T1 /2
 ∗ 
Pc = −n  Eym · Hzm dy 
mT −T1 /2−T2 x=W/2
 
−W/2

 ∗ 
+n  Exm · Hzm dx (15)
W/2 y=mT −T1 /2

Therefore, the expression for the complex power Pc , found is:


J0 γ
Pc = −nJ0∗ tan(γ W/2)T2
σ + jωε0

∞  
1 pπ pπ 2 
+ ap tanh(αp T2 /2) sin
σ + jωε0 q=1 W 2 αp
 ∞   
nJ0∗  pπ 2γ ∗
 σ + jωε ap αp tanh(αp T2 /2) sin
 0 q=1 2 (αp∗ )2 

+
 ∞
 (16)

+ n W 
ap a∗p αp tanh(αp T2 /2)
σ + jωε0 q=1 2

The eddy current loss per unit core-length is the real part of this
complex power, i.e.,
Pe = e[Pc ] (17)

5. APPROXIMATIONS

In view of common numerical values for various parameters, one may

find further simplified, although approximate, expressions for Pc and
Pe . Since σ, the conductivity of iron is a large quantity, and γ’ is very
small, in view of Eqs. (13.1) and (14):
 
4 pπ γ
ap ≈ − J0 sin (18)
W 2 αp2
Further, from Eqs. (7.1) and (7.2), it may be seen that for large values
of p (i.e., for q > Q, say):
pπ
αp ≈ (19.1)
W
442 Mukerji et al.

and
tanh(αp T2 /2) ≈ 1 (19.2)

Further, for small values of p, (i.e., for q ≤ Q, say):

αp ≈ jγ (20.1)

The value of Q should be chosen so as to satisfy Eqs. (19.1) and (20.1),

as best as possible, for a given set of parameter values. Now T2 being
small:
tanh(αp T2 /2) ≈ αp T2 /2 ≈ jγT2 /2 (20.2)

Therefore, one gets an approximate expression for Pc as:

 
∞
J0 J0∗  Q
8 (γW/π) 8 
Pc ≈ n +
σ q=1
π p − (γ W/π)
2 2
q=1
γW
 
J0 J0∗ T2   4π Q
−n (γW ) tan(γW/π) + p (21)
σ W q=1
γW

On summing up the two finite series and the infinite series [19] in the
above equation, the expression for Pc found is:
 
J0 J0∗ 4Q
Pc ≈ n 2 tan(γ W/2) +
σ γW
 
J0 J0∗ T2 4πQ2
−n (γ W ) tan(γ W/2) + (22)
σ W γW

Therefore, using Eqs. (3.1) and (17) one gets:

 
J0 J0∗ sin θ 2Q
Pe ≈ n 2 +
σ cos θ + cosh θ θ
 
∗
J0 J0 T2 sin θ − sinh θ Q2
−n θ + 2π (23)
σ W cos θ + cosh θ θ

where, the parameter θ is the ratio of the core width to the classical
depth of penetration: 
ωµσ
θ=W (23.1)
2
Progress In Electromagnetics Research, PIER 83, 2008 443

From Eq. (23), for small values of θ, one gets:

 
J0 J0∗ T2 1
Pe ≈ n 4Q − 2π Q2 (24)
σ W θ
And for large values of θ:
J0 J0∗ T2
Pe ≈ n θ (25)
σ W
Since the insulation thickness T1 is usually very small, i.e., T1 
T2 , Eqs. (24) and (25) can be rewritten, respectively as:
 
J0 J0∗ T 1
Pe ≈ n 4Q − 2π Q2 (24.1)
σ W θ

for small θ, and

J0 J0∗ T
Pe ≈ n θ (25.1)
σ W

for large θ, where the lamination thickness T is given as:

T = T1 + T 2 (26)

6. CONCLUSION

Using linear electromagnetic field theory, simple expressions for eddy

current loss in laminated rectangular cores have been derived. These
expressions can be readily adapted for cores made of left-handed
materials [20–22].
In view of Eqs. (8.1), (8.2) and (13.1), if the insulation thickness
T1 is zero (i.e., in the absence of any insulation), there will be only y-
component of eddy current density. The presence of insulation layers
interrupts the eddy current path. As a result, the y-component of
eddy current density is modified and an x-component of eddy current
density appears.
It may be seen from Eqs. (13.1), (14) and (16), that due to high
conductivity of iron σ, eddy current loss in a lamination is only mildly
sensitive to the value of the thickness of insulation layers, T1 .
Equation (23) shows that the eddy current loss can be
approximately expressed as a function of a core parameter θ, which
is the ratio of the core width to the classical depth of penetration for
iron. For large values of θ, the eddy current loss in a lamination,
vide Eqs. (25) and (25.1), is linearly proportional to the lamination
444 Mukerji et al.

thickness. However, for small values of θ, as shown by Eqs. (24)

and (24.1), there are two components in the expression for the eddy
current loss in a lamination. One component is independent of
lamination thickness, while the other is proportional to the lamination
thickness.
As shown by Eqs. (24) and (24.1), it is possible that the eddy
current loss in a laminated core may increase if the lamination thickness
is reduced. This is because a reduced eddy current damping results
deeper field penetration in the lamination.

REFERENCES

1. Poljak, D. and V. Doric, “Wire antenna model for transient

analysis of simple grounding systems, Part I: The vertical
grounding electrode,” Progress In Electromagnetics Research,
PIER 64, 149–166, 2006.
2. Poljak, D. and V. Doric, “Wire antenna model for transient
analysis of simple grounding systems, Part II: The horizontal
grounding electrode,” Progress In Electromagnetics Research,
PIER 64, 167–189, 2006.
3. Tang, M. and J. F. Mao, “Transient analysis of lossy nonuniform
transmission lines using a time-step integration method,” Progress
In Electromagnetics Research, PIER 69, 257–266, 2007.
4. Fau, Z., L. X. Ran, and J. A. Kong, “Source pulse optimization
for UWB radio Systems,” Journal of Electromagnetics Waves and
Applications, Vol. 20, No. 11, 1535–1550, 2006.
5. Okazaki, T., A. Hirata, and Z. I. Kawasaki, “Time-domain
mathematical model of impulsive EM noises emitted from
discharges,” Journal of Electromagnetics Waves and Applications,
Vol. 20, No. 12, 1681–1694, 2006.
6. Mukerji, S. K., G. K. Singh, S. K. Goel, and S. Manuja,
“A theoretical study of electromagnetic transients in a large
conducting plate due to current impact excitation,” Progress In
Electromagnetics Research, PIER 76, 15–29, 2007.
7. Mukerji, S. K., G. K. Singh, S. K. Goel, and S. Manuja, “A
theoretical study of electromagnetic transients in a large plate
due to voltage impact excitation,” Progress In Electromagnetics
Research, PIER 78, 377–392, 2008.
8. Bewley, L. V. and H. Poritsky, “Intersheet eddy-current loss in
laminated cores,” Trans. A.I.E.E., Vol. 66, 34, 1937.
9. Beiler, A. C. and P. L. Schmidt, “Interlaminar eddy current loss
in laminated cores,” Trans. A.I.E.E., Vol. 66, 872, 1947.
Progress In Electromagnetics Research, PIER 83, 2008 445

10. Bewley, L. V., Two-dimensional Fields in Electrical Engineering,

1st edition, 90–95, Dover Publications, Inc., 1963.
11. Subbarao, V., Eddy Currents in Linear and Non-linear Media,
60–63, Omega Scientific Publishers, New Delhi, 1991.
12. Gibbs, W. J., “Tooth-ripple losses in unwound pole-shoes,”
J.I.E.E., Vol. 94, Part II, 2, 1947.
13. Carter, G. W., “A note on the surface loss in a laminated pole-
face,” Proceedings of I.E.E., Vol. 102, Part C, 217, 1955.
14. Greig, J. and K. C. Mukherji, “An experimental investigation
of Tooth-ripple flux pulsations in smooth laminated pole-shoes,”
Proceedings of I.E.E., Vol. 104, Part C, 332, 1957.
15. Bondi, H. and K. C. Mukherji, “An analysis of tooth-ripple
phenomena in smooth laminated pole shoes,” Proceedings of
I.E.E., Vol. 104, Part C, 349, 1957.
16. Greig, J. and K. Sathirakul, “Pole-face losses in alternators: An
investigation of eddy current losses in laminated pole-shoes,”
Proceedings of I.E.E., Vol. 108, Part C, 130, 1961.
17. Greig, J. and E. M. Freeman, “Simplified presentation of the eddy
current-loss equation for laminated pole-shoes,” Proceedings of
I.E.E., Vol. 110, No. 7, 1255, 1963.
18. Mukerji, S. K., M. George, M. B. Ramamurthy, and K. Asaduz-
zaman, “Eddy currents in solid rectangular cores,” Progress In
Electromagnetics Research B, Vol. 7, 117–131, 2008.
19. Jolley, L. B. W., Summation of Series, 2nd revised edition, 144,
Dover Publication, Inc., 1961.
20. Chen, H., B. I. Wu, and J. A. Kong, “Review of electromagnetic
waves in left-handed materials,” Journal of Electromagnetics
Waves and Applications, Vol. 20, No. 15, 2137–2151, 2006.
21. Grzegorczyk, T. M. and J. A. Kong, “Review of left-handed
materials: Evolution from theoretical and numerical studies to
potential applications,” Journal of Electromagnetics Waves and
Applications, Vol. 20, No. 14, 2053–2064, 2006.
22. Mahmoud, S. F. and A. J. Viitanen, “Surface wave character on a
slab of metamaterial with negative permittivity and permeability,”
Progress In Electromagnetics Research, PIER 51, 127–137, 2005.