Volume III, Issue IV, April 2014 wwww.ijltemas.

in ISSN 2278 - 2540

Magnetic Flux Distributions in Transformer

Core Joints
S. Sundara Mahalingam
Assistant Professor,Department of Electrical & Electronics
Engineering, Mepco Schlenk Engineering College, Sivakasi,
Tamilnadu, India
[email protected]

Abstract— Transformer is an essential and expensive normally used to manufacturing of core i.e. single step lap
component in the power system. Even though transformer and multi step lap which depends on loss requirements and
losses are small percentage in total power transferred manufacturing capabilities. The aim of this paper is not to
(<0.5% in large power transformers), it produces endorse one method to another but to give general
localized heating which can affect normal operation of the information about the core joint parameters and the loss
transformer. Hence, it is very significant to understand distributions. It has been reported that the power losses in full
how the losses arise and to estimate these losses accurately size transformer cores are reduced as much as up to 5% or
as much as possible. So that we can consider necessary even 10% compared to SSL configurations [1]. But MSL has
steps at design stages itself to find the ways for reducing disadvantage of increased costs of core assembling. It can be
the losses. Accurate estimation of core loss of transformer noted that examinations of the optimum core construction and
is a critical issue at transformer design stage. In order to most suitable magnetic characteristics of core material are
optimize the transformer core design, the influences of possible using finite element method. T. Nakata, N.
core parameters on core loss performance need to be Takahashi [2] also investigated the effect of the step lap joint
analysed. The magnetic flux distribution in transformer design on core losses. This paper presents magnetic flux
core joint regions has been analysed by finite element distribution analysis in Z direction with flux density range of
method as a function of various core parameters. In this 1.2T to 1.7T. From this analysis, we can estimate accurate
paper, the effects of core joints, core materials, overlap core loss and optimize core joint configuration.
length, and air gap length on core loss performance are
examined. It is inferred that losses increased as overlap
length between adjacent lamination increased.

Keywords Finite Element Method, Core loss, Core joints, II ELECTROMAGNETIC ANALYSIS AND
overlap length, step lap joint SIMULATIONS

I INTRODUCTION An electromagnetic analysis with FEM was realized to

determine the magnetic flux distribution and to compute
It is important to understand in detail the behaviour of losses in lamination steel. The analysis of transformer core
localized flux distribution in transformer core in order to model in three-dimensional (3D) models with its exact
optimize the core construction and accurate estimation of core number of laminations causes several problems such as very
losses. Because accurate estimation of core losses are more long simulation time and needs great amounts of computer
important in their design and specifications for economic reasons resources (RAM, hard disk). So by considering these
otherwise higher losses than predicated will lead to financial problems, 3D core model can be replaced by 2D simulations to
penalties and lower losses gives less pay pack to designers and analyse various core parameters [5-6]. In this paper, vector
manufactures. The main factors which increase the loss in potential is used for 2D modelling of the core which is
laminated core are non uniform flux, difference in magnetic path justified by Maxwell equations B = rot A.
Hence ( ∇ x (υ∇ x A)) = J. In 2D problems, we assume that
length, rotating flux, interlaminar short circuit, stress, and
transverse flux [3-4]. These factors can be controlled by selecting
vectors A have only one component, perpendicular to the XY
optimized core design with various core parameters such as core
plane. So the possion equation of magnetic vector potential
joints, core materials, lamination thickness, number of steps,
will be
number of layers per step, overlap length, air gap length. It is fact
that there is no analytical relationship with effect of these
parameters on core loss performance. There are two different
joint stacking patterns

is the magnetic reluctivity i.e. inverse of magnetic permeability

(µ )
262

[Type text] [Type text] [Type text]

 Volume III, Issue IV, April 2014 wwww.ijltemas.in ISSN 2278 - 2540

number of layers per step. Due to periodic repetition of

In 2D fem, flux density is given by lamination steel to whole transformer, consider only finite
number of laminations groups. The average induction in the
laminations steel along X axis is enforced by appropriate
Dirichlet boundary conditions on the upper and lower
boundary of the model.

After flux density calculation, total core loss can be A. Flux Distribution analysis in Single step lap joint
obtained from the following

Where,
Hi - Magnetic field intensity in mesh
(i) Bi- Magnetic field density in mesh
(i) (Vol)i - Volume mesh (i)
n - Number of meshes

Here we assume that vector potential A does not vary along X

Fig.2. Magnetic flux lines in single step lap joint at induction 1.7T
axis so component of By = 0 and A can vary only along Y
direction. So magnetic flux lines are parallel to the
equipotential lines of A i.e. along X axis. By symmetry, the Fig 2 shows flux lines in single step lap joint of core
magnetic field on the horizontal segments in upper region and with air gap length g=0.2mm and overlap length s= 2mm,
lower region should be uniform. So A1 vector potential while overall flux density 1.7T. While assembling the
assigned in upper region where as A2 with same magnitude and transformer core, unavoidable air gaps present at joints. The
opposite polarity. The difference between the potentials A1 and magnetic flux travels between the overlapping laminations of
A2 gives magnetic flux per meter (wb/m). yokes and limbs due to presence of air gaps at the joints. The
magnetic flux passes to next layer of laminations when it
III Z PLANE MODELING meets high reluctance air gap. When these laminations are
saturated, flux crosses the air gaps. This flux distortion causes
additional localized losses.
The magnetic characteristics of cores with single
step lap (SSL) and Multi step lap (MSL) have been analysed
using finite element method taking into account of eddy
current and magnetic saturation. In the single step lap, two
steps were used per group for overlap where as multi step lap
used five steps per group for overlap The magnetic
characteristics of cores with single step lap (SSL) and Multi
step lap (MSL) have been analysed using finite element
method taking into account of eddy current and magnetic
saturation. In the single step lap, two steps were used per
group for overlap where as multi step lap used with five steps per
group for overlap as shown in Fig.1.

Fig.3. Magnetic Flux density distribution in single step lap joint

a) B plot b) Along line

At a sufficiently large distance from the gaps, the flux in the

laminations is nearly uniform. Near the gaps there is a
significant vertical flux component. So flux density in the
Fig.1. Investigated types of overlap arrangement a) SSL b) MSL
lamination steel area shunting the air gap rises upto 2.6T
while in the next gap of same layer is 0.7T as shown in fig. 3.
The following major parameters which affect magnetic
characteristics of cores have been investigated grades of steel, 263
operating flux densities, overlap length, air gap length, and
[Type text] [Type text] [Type text]
 Volume III, Issue IV, April 2014 wwww.ijltemas.in ISSN 2278 - 2540

B. Flux Distribution in Multi step lap joint The flux density distribution along line 6 of Fig. 7 is taken in
the interlaminar gap between two groups i.e. along normal
Similarly in multi step lap joint also, flux lines are direction. The flux density gets reduced in air gap between
distributed uniformly in steel and jumps in air gap region lamination and again increased in steel package.
as shown in fig 4. Due to step arrangement, flux density
distribution almost uniform over whole steel package as
shown in fig. 6. In fig. 5, it shows the representation of
lines to identify flux density distribution in MSL. In case 1.9

of SSL, flux density distribution is sinusoidal in nature 1.7

MSL -M4

where as in MSL, flux density reduced only in air gap 1.5 MSL - MOH

SSL - MOH
region. So MSL is more efficient than SSL. But number of 1.3 SSL - M4

layers per step and number of steps in a group also need 1.1

to be considered. 0.9
1.1 1.3 1.5 1.7 1.9
Flux density(Tesla)

Fig.8. Comparison of core losses between MSL & SSL with

two different materials

The no-load losses depend on the magnetic flux density at

which the transformer operates as well as the physical
characteristics of the magnetic material. So these effects were
studied for different material i.e. 0.27 mm thickness of grain
orientated steel M4 and MOH and for various operating flux
densities from 1.2 T to 1.7T. The results of this simulation on
Fig 4. Magnetic flux lines in Multi step lap joint at B=1.7T how much losses of MSL core deviate from those of a SSL
joint core are represented in fig 8. It can be seen that there are
much improvement which can be attained by replacing SSL by
MSL. Due to joint contribution, losses in SSL are about 5%
loss increase from MSL.

C. Effect of overlap length and air gap length on core loss

1.969
1.968
1.967
Fig 5. Sketch for identification of plot lines in MSL 1.966
1.965
1.964 g = 0.3mm
1.963 g=0.2mm
2.5
1.962 g=0.1mm
1.961
2 Line 1 1.96
1.959
1.5 Line 2 1.958

Line 3 0 2 4 6 8 10
1 Line 4 Overlap length(m m )
Line 5
0.5
a)
0
0 10 20 30 40 50 60
Distance(mm)

Fig. 6 Flux density distribution in MSL at B=1.7T 2 1.998

1.996
1.994
1.992
1.99 g = 0.3m m
1.988 g=0.2m m
1.986
g=0.1m m
1.984
1.982
1.98
0 2 4 6 8 10
Overlap length(mm)

b)
Fig. 9 effect of overlap length and air gap length in MSL at B=1.7T
a) RGO b) HI-B 264
Fig. 7 Flux density distribution in normal direction along line 6

[Type text] [Type text] [Type text]

 Volume III, Issue IV, April 2014 wwww.ijltemas.in ISSN 2278 - 2540

It was applied to analyse core losses and excitation current IV CONCLUSIONS

with multi step lap joint. The effect of overlap length and air gap
length are also analysed for two different core materials in 1. While assembling transformer cores, air gap will
MSL. When increasing overlap length, magnetic mean path naturally arise at the lamination sheets. These air gaps in
length of the core varied. In mitred case, uneven magnetic conjunction with shifted core steel packages cause
path length, air triangles at inner corner of core and transverse interlaminar flux components which leads to flux
flux components becomes more predominant for localized joint inhomogeneties and higher loss.
loss. Core losses increases with overlap length and air gap length 2. Loss increases depend upon the magnetic path
as shown in fig. 9 length and local distribution of air gaps. With increase in air gap
length and overlap length, core losses will be increased.
3. In lower flux density range, core performance is
improved as compared to higher range of induction. The core
loss performance varies with grade of material. Hi-B material
7 has better than RGO steel.
6
4. The percentage increase in loss from non step lap
5 HiB material
4 joint to step lap joint is considerable less in higher operating
3 RGO material flux density.
2
1
1.1 1.3 1.5 1.7 1.9
Flux density(Tesla)

a)
REFERENCES

[1] G. F. Mechler and R. S. Girgis, “Calculation of spatial loss distribution in

stacked power and distribution transformer core,” IEEE Transactionsion
Power Delivery, vol. 13, no. 2, pp. 532-537, Apr. 1998.
100

80 [2]T. Nakata, N. Takahashi, and Y. Kawase, “Magnetic performance of step-

lap joints in distribution transformer cores,” IEEE Trans. Mag., vol. 18, no. 6,
HiB material
60 pp. 1055-1057, Nov. 1982.
RGO material
40
[3]Z. Valkovic, “Some aspects of additional losses in step-lap joints of
20 transformer core,” The Int. J. for Computation and Mathematics,vol.11, no. 1,
1.1 1.3 1.5 1.7 1.9 pp. 137-140, 1986.
Flux density (Tesla)
[4]Girgis, R.S., teNijenhuis, Ed G., Gramm K and Wrethag, J.E,
“Experimental investigations on effect of core production attributes on
b) transformer core loss performance,” IEEE Transactions on Power delivery,
Fig.10. SSL versus MSL, a) % increase in loss b) % increase in excitation Vol. 13, No. 2, April 1998, pp 526-531
current
[5]Mechler, G. F and Girgis, R. S. “Magnetic flux distributions in transformer
More precisely, at lower B the MSL configuration showed core joints”, IEEE Transactions on Power delivery, Vol. 15, No. 1 January
2000, pp. 198-203
higher percentage of loss increase than SSL values, whereas at
higher B this effect became inconsistent as shown in fig 10. [6] Ehsan Hajipour, Pooya Rezaei, Mehdi Vakilian, Mohsen
Normally high permeability (HI-B) material is better Ghafouri,”Power Transformer No-Load Loss Prediction with FEM Modeling
performance than regular grain oriented (RGO) material. and Building Factor Optimization,”Journal of Electromagnetic Analysis and
Applications, October 2011, 3, 430-438

265

[Type text] [Type text] [Type text]