Chinese Materials Research Society

Progress in Natural Science: Materials International

Progress in Natural Science: Materials International 2012;22(3):244–249

1002-0071 & 2012 Ch

and hosting by Elsev

Peer review under re

Society.

http://dx.doi.org/10.1

n

Corresponding au

E-mail address:

www.elsevier.com/locate/pnsmiwww.sciencedirect.com

ORIGINAL RESEARCH

Reducing the core loss of amorphous cores for

distribution transformers

Deren Lin, Liang Zhang, Guangmin Li, Zhichao Lu, Shaoxiong Zhou

China Iron & Steel Research Institute Group, Advanced Technology and Materials Co., Ltd., No. 76, Xueyuan Nan Rd., Haidian

District, Beijing 100081, China

Received 7 March 2012; accepted 22 March 2012

KEYWORDS

Amorphous core;

Annealing technique;

Reducing core loss;

Loss deterioration;

Distribution

transformer

inese Materials R

ier Ltd. All rights

sponsibility of Chi

016/j.pnsc.2012.04

thor. Tel.: þ86 10

derenli@atmcn.com

Abstract Although amorphous alloy ribbons have lower core loss, serious loss deterioration

occurs in amorphous cores constructed from amorphous alloy ribbon for distribution transformer.

In the present study, the loss deterioration mechanism and influencing factors of amorphous cores

for distribution transformers were investigated. The influence of distributed gaps on core loss of

transformer core was calculated by simulation software and validated by experiments. The boron

fluctuation of amorphous alloy ribbons affects the annealing temperature of amorphous

transformer cores. A superior core annealing technique with annealing parameters was developed,

and appreciable decrease in core loss was obtained.

& 2012 Chinese Materials Research Society. Production and hosting by Elsevier Ltd. All rights reserved.

1. Introduction

Distribution transformers are a significant source of energy

loss in an electric power networks, despite they are very

efficient electrical machines. The losses of transformer consist

of no-load losses and load losses. No-load losses are constant

esearch Society. Production

reserved.

nese Materials Research

.005

5874 2811.

(D. Li).

and appear so long as a transformer is in service, while load

losses vary with the square of the load current carried by the

transformer and are only significant under higher load condi-

tions. The dominant no-load loss is core loss which is

associated with the time-varying nature of the magnetization

and results from hysteresis loss, eddy current loss and

anomalous loss [1]. Therefore, the need to reduce core loss

has become an increasingly important design and manufacture

consideration in all size of distribution transformer. Amor-

phous alloy exhibits low hysteresis loss resulting from a small

magnetic anisotropy and low eddy current loss resulting from

a large resistivity and thinner thickness [2]. Amorphous

ribbons are widely used in electrical power system such as

distribution transformers and inductive devices [3,4] because

of low core loss. Comparing with Si-steel in the application of

distribution transformer, no-load core loss of amorphous core

distribution transformer decreases by 60–70% respecting to

different capacity of transformers. The use of amorphous alloy

ribbon to distribution transformer cores is gradually gaining

widespread acceptance due to lower core loss.

Reducing the core loss of amorphous cores for distribution transformers 245

The core loss of amorphous transformer cores was effec-

tively reduced by optimizing winding tension, core annealing

technique [5], field annealing [6], laser scribing [7], mechanical

scribing [8] and building factor of amorphous cores [9]. Although

low core loss is expected for amorphous transformer cores, large

loss deterioration occurs in amorphous cores even after field

annealing [9]. In order to further reduce the core loss, the aim of

this paper is to systematically study the influence of core

structure, fluctuation of composition and field annealing techni-

que on core loss of large quantity transformer-size cores.

Fig. 2 The structure of a 30 kV A transformer cores constructed

from 38 groups of strips with 40 strip layers per group.

2. Experimental

The amorphous ribbons for preparing transformer cores were

produced by the planar-flow casting method, the nominal

compositions are Fe94.5�xSi5.5Bx (wt%), where x¼2.35–2.55,

the ribbon width is 14270.5 mm and the ribbon thickness is

2570.5 mm. The magnetization and core loss of annealed

sample were measured by IWATSU SY-8232 BH ANALY-

ZER. Examination by X-ray diffraction of the as-cast ribbons

revealed that all samples were amorphous. The core losses of

transformer cores were measured under a controlled sinusoi-

dal induction waveform at 50 Hz. The crystallization tem-

perature of the as-cast ribbons was measured by NETZSCH

DSC 404C, the heating rate was 10 K/s. The boron content

in the as-cast ribbons was measured by iCAP 6300 ICP

SPECTROMETER.

3. Results and discussion

3.1. Structure-related core loss

Over the past years, significant improvements have been made

in amorphous metal transformer core to reduce the core size,

manufacturing costs and the losses introduced into the

distribution transformer by the transformer cores. The no-

cut core has the advantage of lowest core loss comparing to

cut-cores including the step-lap core, the lapped joint core and

Fig. 1 The lapped joints and distributed gaps of a step-lap core.

A group of strips is composed of 40 layers of amorphous strip, the

overlapped length is 14 mm and the distance of each gap is 4 mm.

the butt joint core, but the disadvantage of no-cut core is the

cost resulting from the necessity of coiling around cores. As to

the cut core, the step-lap cores [10,11] are widely adopted in

the present amorphous metal transformer because of the

superiority to the lapped joint core and butt joint core [12].

In the step-lap core, a group of strips with 10–40 strip layers

overlaps 14 mm between the opposite ends of the group and

the distance of each gap is 4 mm. A plurality of this type of

lapped joints is angularly staggered, repeating in a stair-step

fashion as one proceeds from the window to the outer

periphery of the core. Fig. 1 show the lapped joints and

distributed gaps of a step-lap core, where a group of strips is

composed of 40 layers of amorphous strip cut from contin-

uous spool of amorphous ribbon, the overlapped length is

14 mm and the distance of each gap is 4 mm.

The different numbers of strip layers of a group result in

different sizes of gap. The sizes of the distributed gaps have

large influence on the core loss provided that the total strip

layers of a core are identical. To evaluate the influence of

distributed gaps on the core loss, four 30 kV A transformer

cores were made with strip layers of 10, 20, 30 and 40 per

group. The gap size varies with the number of strip layers of a

group and the gap number consequently varies with gap size

due to the same total strip layers of the four cores. For all of

the cores, the weight of each core is 30 kg, the total strip layers

of each constructed core are 1520, the lamination factor is 0.86

and the size of windows A, B and C are 220 mm, 120 mm and

41.5 mm, respectively. Fig. 2 shows the structure of a 30 kV A

transformer core constructed from 38 groups of strips with 40

strip layers per group. The core losses of the four cores were

calculated by using software INFOLYTICA MAGNET 7.2.

In order to simplify the model for calculation, only the gap

number and gap size are considered as variants and the other

parameters are identical. The material parameters of magneti-

zation and core loss at 50 Hz were used during the calculation

0 10 20 30 40 50 60

1.0

1.1

1.2

1.3

1.4

1.5

B (T

)

H(A/m)

1.0 1.1 1.2 1.3 1.4 1.50.04

0.08

0.12

0.16

0.20

Loss

(W/k

g)

B (T)

Fig. 3 (a) The measured static B–H curves of annealed amorphous

ribbon. (b) The dependence of measured core loss of annealed

amorphous ribbon on flux density at 50 Hz.

Fig. 4 (a) The simulated flux density distribution in the lapped

joints of a 30 kV A transformer core constructed from groups of

strips with 10 strip layers per group. (b) The simulated flux density

distribution in the lapped joints of a 30 kV A core constructed

from groups of strips with 40 strip layers per group.

D.R. Li et al.246

process. Fig. 3a shows the measured static B–H curve of

annealed amorphous ribbon and Fig. 3b shows the dependence

of measured core loss of annealed amorphous ribbon on flux

density at 50 Hz. The calculated total core loss of each 30 kV A

transformer core is obtained by adding components of the

building loss of each group to the nominal loss. Fig. 4a and b

show the simulated flux distribution in the lapped joints of

cores constructed from groups of strips with 10 strip layers per

group and 40 strip layers per group, respectively. The config-

uration parameters and calculated results of core loss are listed

in Table 1. Extra losses were introduced to both of the cores in

joint regions due to the distributed gaps which result in extreme

variation of in-plane flux and deviation of flux from long-

itudinal direction of amorphous ribbon. These mechanisms

combined raise the nominal loss by 18.8% and 29.3% for cores

with 10 strip layers per group and 40 strip layers per group,

respectively. The increase of nominal loss resulting from

deviation of flux was also seen in grain-oriented silicon steel

transformer core [13]. Air gaps in joint region reduce the

effective permeability of a core and cause localized variable

flux distribution which in turn increases the losses according to

the gap size [14]. For amorphous transformer cores, the

calculated results also show that the core loss increases with

the increment of gap size. It can be seen from Fig. 4a that a

large magnetization vector occurs near the joint region and

passes through the joint area smoothly at a small angle in core

with 10 strip layers per group. This indicates that the flux

deviation of the core with 10 strip layers per group is small

compared with that of the core with 40 strip layers per group.

Fig. 5 shows the dependence of core loss on the sizes of

distributed gaps in term of strip layers per group. The results

of the core losses were experimentally measured for the

30 kV A transformer cores constructed from different strip

layers per group provided that the total layers of each core are

identical. It can be seen that the core loss increases with the

increment of gap size in term of strip layers of a group. The

experimentally measured results of core loss dependence on

the sizes of distributed gaps are qualitatively in good agree-

ment with the calculated results of core loss dependence on the

sizes of distributed gaps.

3.2. Field annealing

With the growing acceptance of amorphous alloy ribbons by

transformer manufacturers and users, the annealing of large

quantities transformer-size cores can be further optimized to

shorten annealing time and to reduce core loss. The magnetic

properties of amorphous transformer cores are generally

optimized by annealing in an applied magnetic field. The

purpose of the annealing is to minimize the internal stress

which is caused by the rapidly quenched casting and core

making processes. The combination of temperature and time is

chosen such that it is sufficient to relieve as much stress as

possible without causing crystallization of the amorphous

Table 1 The configuration parameters and calculated results of 30 kV A transformer cores.

Gap number One gap size (mm2) Total gap size (mm2) Materials loss (W/kg)a Calculated loss (W/kg)

10 layers core 125 1.08 135.00 0.133 0.158

40 layers core 31 4.32 133.92 0.133 0.172

aMaterials loss was obtained from Fig. 3b by polynomial interpolation.

10 15 20 25 30 35 400.15

0.16

0.17

0.18

0.19

0.20

0.21

Cor

e lo

ss (W

/kg)

Number of strip layers of a group

Fig. 5 The dependence of core losses on the sizes of distributed

gaps in term of strip layers per group.

0 60 120 180 240 3000

50

100

150

200

250

300

350

400

Oven soak temperature Core center temperature Core skin temperature

Tem

pera

ture

(°C

)

Time (min.)

Heat-up Heat preservation Cool-down

Fig. 6 A typical annealing profile of oven soak temperature and

core temperature for a batch of 630 kV A transformer cores.

350 360 370 380 390 4000.16

0.18

0.20

0.22

0.24

0.26

0.5

1.0

1.5

Core loss

Cor

e lo

ss (W

/kg)

Temperature (°C)

Exc

iting

pow

er (V

A/k

g)

Exciting power

Fig. 7 The dependence of core loss and exciting power on

annealing temperature of heat-preservation.

30 60 90 120 1500.17

0.18

0.19

0.20

0.21

0.22

Core loss

Exc

iting

pow

er (V

A/k

g)

Cor

e lo

ss (W

/kg)

Time (min.)

0.2

0.4

0.6

0.8

Exciting power

Fig. 8 The dependence of core loss and exciting power on the

annealing time of heat-preservation.

Reducing the core loss of amorphous cores for distribution transformers 247

alloy. The magnetic field is used for inducing longitudinal

uniaxial anisotropy to reduce the core loss and exciting power.

An annealing technique was developed for large quantities

transformer-size cores. Fig. 6 shows a typical annealing profile

of oven soak temperature and core temperature for a batch of

630 kV A transformer cores, the total number of cores in the

oven soak was 12 and the weight of each core was 102 kg. The

soak temperature was program controlled by a computer to

ensure that the cores in the oven soak were annealed at a

certain temperature and duration. An annealing process

consists of heat-up, heat-preservation and cool-down, the

total annealing cycle was around 5 h. The magnetic field was

longitudinally applied throughout the annealing process to

induce uniaxial magnetic anisotropy.

The combination of annealing temperature, annealing time

and applied magnetic field is optimized to reduce the core loss

and exciting power. Fig. 7 shows the dependence of core loss and

exciting power on the annealing temperature of heat-preservation

ranging from 355 1C to 395 1C for duration of 90 min under

longitudinally applied magnetic field of 46 Oe, the core loss and

exciting power of a batch of 630 kV A cores were measured at

50 Hz and 1.35 T. With the increase of annealing temperature of

heat-preservation, the core loss decreases drastically and then

increases slightly, exhibiting a minimum value around 375 1C.

The exciting power monotonously decreases with the increase of

the annealing temperature of heat-preservation. Fig. 8 shows the

dependence of core loss and exciting power on the annealing time

of heat-preservation ranging from 30 min to 150 min at 375 1C

20 30 40 50 60 700.16

0.18

0.20

0.22

0.24

0.26

0.2

0.4

0.6

0.8

Core loss

Cor

e lo

ss (W

/kg)

Applied field (Oe)

Exciting power

Exc

iting

pow

er (V

A/k

g)

Fig. 9 The dependence of core loss and exciting power on the

longitudinally applied magnetic field.

2.38 2.40 2.42 2.44 2.46 2.48 2.50 2.52 2.54 2.56

510

515

520

525

530

Cry

stal

lizat

ion

tem

pera

ture

(°C

)

B Content (wt.%)

Fig. 10 The dependence of crystallization temperature of amor-

phous ribbons on the boron content in amorphous ribbons.

2.40 2.42 2.44 2.46 2.48 2.50350

355

360

365

370

375

380

Opt

imiz

ed a

nnea

ling

tem

pera

ture

(°C

)

B Content (wt.%)

Fig. 11 The dependence of optimized annealing temperature of

transformer cores on the boron content in amorphous ribbons.

D.R. Li et al.248

under longitudinally applied magnetic field of 46 Oe, the core

loss and exciting power of a batch of 630 kV A cores were also

measured at 50 Hz and 1.35 T. The core loss exhibits a minimum

value around the annealing time of 60 min and the exciting

power also monotonously decreases with the increase of anneal-

ing time of heat-preservation. Fig. 9 illustrates the dependence of

core loss and exciting power on the longitudinally applied

magnetic field ranging from 25 to 67 Oe for cores annealed at

375 1C for 90 min. With the augment of longitudinally applied

magnetic field, both the core loss and exciting power decrease to

minimum values and then increase, indicating that there exists

an optimal magnitude of longitudinally applied magnetic field.

To obtain low core loss and exciting power, it is necessary to

use an optimized longitudinally applied magnetic field during

the annealing process. For amorphous ribbon with longitudin-

ally induced anisotropy, the magnetization process is domi-

nated by domain nucleation and motion of longitudinal domain

wall with concomitance of large Barkhausen jumps. The total

losses of an amorphous core rise with increasing strength of

longitudinally induced anisotropy [15], the observed large total

losses result from domain nucleation and wall pinning [16].

3.3. Boron content and optimal annealing temperature

Typical compositions of amorphous ribbons for transformer

cores are FeSiB ternary alloy [17,18] whereby the metalloids Si

and B are necessary for glass formation and amorphous

structure stabilization. This alloy system has a good glass

forming ability [19] and is easily accessible by planar-flow

casting in large scale production. Although the range of

compositions which can be prepared in the glassy state by

planar flow casting from the melt is wide, the detailed

composition has large influence on the soft magnetic proper-

ties of amorphous ribbons. For Fe94.5�xSi5.5Bx (wt%) alloys,

the boron content has large influence on the crystallization

temperature and the soft magnetic properties of amorphous

ribbon. Fig. 10 shows the dependence of crystallization

temperature of amorphous ribbons on the boron content. It

can be seen from Fig. 10 that the crystallization temperature

of amorphous ribbons increase with the increment of boron

content. As mentioned above, the core loss and exciting power

can be effectively reduced by properly elevating the annealing

temperature. Fig. 11 shows the dependence of optimized

annealing temperature of transformer cores on the boron

content in amorphous ribbons. It is obvious that the optimal

annealing temperature of transformer cores increases with the

content of boron, which is beneficial for the core losses.

4. Conclusions

The loss deterioration factors including the core structure, the

fluctuation of boron content in amorphous ribbons and field

annealing parameters were optimized. The core losses of

amorphous transformer cores increased with an increase of

the sizes of distributed gaps, and the experimental results are

in qualitatively good agreement with the calculated results.

Increasing the boron content in amorphous ribbon resulted in

an increase of the crystallization temperature and of the

optimal annealing temperature of transformer cores. A sophis-

ticated field annealing technique was developed, which effec-

tively reduced the core losses of transformer cores.

Reducing the core loss of amorphous cores for distribution transformers 249

Acknowledgment

This work was supported by the National High-Tech Research

and Development Programme under Grant No. 2012AA030301.

References

[1] Y. Okazaki, H. Kanno, S. Kousaka, E. Sakuma, Magnetic

properties of surface-treated Fe–Si–B amorphous alloy, IEEE

Transactions on Magnetics 23 (1987) 3515–3517.

[2] M.E. McHenry, M.A. Willard, D.E. Laughlin, Amorphous and

nanocrystalline materials for applications as soft magnets, Pro-

gress in Materials Science 44 (1999) 291–433.

[3] R. Hasegawa, Applications of amorphous magnetic alloys, Material

Science and Engineering A375–377 (2004) 90–97.

[4] R. Hasegawa, Applications of amorphous magnetic alloys in

electronic devices, Journal of Non-Crystalline Solids 287 (2001)

405–412.

[5] D.M. Nathasingh, H.H. Liebermann, Transformer application of

amorphous alloys in power distribution system, IEEE Transac-

tions on Power Delivery PWRD 2 (1987) 843–850.

[6] C. Beatrice, C. Appino, E. Ferrara, F. Fiorillo, Loss in transverse

field annealed amorphous ribbon, Journal of Magnetism and

Magnetic Materials 160 (1996) 302–304.

[7] D. Rayboulf, M. Meola, R. Bye, S.K. Das, Reducing the magnetic

losses of amorphous ribbon, Materials Science and Engineering

A241 (1998) 191–201.

[8] K.S. Tan, Amorphous ribbon grain refinement by scribing, IEEE

Transactions on Magnetics 22 (1986) 188–191.

[9] Y. Okazaki, Loss deterioration in amorphous cores for distribu-

tion transformer, Journal of Magnetism and Magnetic Materials

160 (1996) 217–222.

[10] D.E. Ballard, W. Klappert, Method of Manufacturing an

Amorphous Metal Transformer Core and Coil Assembly, US

Patent no. 4,734,957, 1988.

[11] D.E. Ballard, W. Klappert, Method for Making a Transformer

Core Comprising Amorphous Metal Strips Surrounding the Core

Windows, US Patent no. 5,093,981, 1992.

[12] F. G. Grimes, E. Hammack, Butt-lap-step Core Joint, US Patent

no. 4,761,630, 1988.

[13] A.J. Moses, Comparison of transformer loss prediction from

computed and measured flux density distribution, IEEE Transac-

tions on Magnetics 34 (1998) 1186–1188.

[14] A.J. Moses, Prediction of core losses of three phase transformers

from estimation of the components contributing to the building

factor, Journal of Magnetism and Magnetic Materials 254–255

(2003) 615–617.

[15] V. Ungemach, W. Kunz, R. Hilzinger, Influence of the induced

anisotropy on the magnetic properties of amorphous Co66Fe4(-

Mo,Si,B)30 alloys, Journal of Magnetism and Magnetic Materials

41 (1984) 363–365.

[16] J.L. Porteseil, O. Geoffroy, Wall pinning by nanocrystals in an

amorphous matrix, Journal of Magnetism and Magnetic Materi-

als 140–144 (1995) 1855–1856.

[17] F. Luborsky, J. Becker, J. Walter, D. Martin, The Fe–Si–B

ternary amorphous alloys, IEEE Transactions on Magnetics 16

(1980) 521–525.

[18] K. Hoselitz, Magnetic properties of iron–boron–silicon metallic

glasses, Journal of Magnetism and Magnetic Materials 20 (1980)

201–206.

[19] B.S. Dong, S.X. Zhou, D.R. Li, C.W. Lu, F. Guo, X.J. Ni, Z.C.

Lu, A new criterion for predicting glass forming ability of bulk

metallic glasses and some critical discussions, Progress in Natural

Science: Materials International 21 (2011) 164–172.