Transformer sound level caused by core

magnetostriction and winding stress

displacement variation
Cite as: AIP Advances 7, 056681 (2017); https://doi.org/10.1063/1.4978759
Submitted: 24 September 2016 • Accepted: 19 December 2016 • Published Online: 15 March 2017

Chang-Hung Hsu, Yi-Mei Huang, Min-Fu Hsieh, et al.

ARTICLES YOU MAY BE INTERESTED IN

Correlation of magnetostriction variation on magnetic loss and noise for power transformer
Journal of Applied Physics 117, 17E716 (2015); https://doi.org/10.1063/1.4919122

A study of kinetic friction: The Timoshenko oscillator

American Journal of Physics 86, 174 (2018); https://doi.org/10.1119/1.5008862

Effects of vertical vibration on hopper flows of granular material

Physics of Fluids 14, 3439 (2002); https://doi.org/10.1063/1.1503354

AIP Advances 7, 056681 (2017); https://doi.org/10.1063/1.4978759 7, 056681

© 2017 Author(s).
 AIP ADVANCES 7, 056681 (2017)

Transformer sound level caused by core

magnetostriction and winding stress
displacement variation
Chang-Hung Hsu,1,2 Yi-Mei Huang,2 Min-Fu Hsieh,3 Chao-Ming Fu,4
Shiva Adireddy,5 and Douglas B. Chrisey5
1 Researchand Development Center, Fortune Electric Company, Taoyuan 320, Taiwan
2 Department of Mechanical Engineering, National Central University, Taoyuan 320, Taiwan
3 Department of Systems and Naval Mechatronic Engineering, National Cheng Kung University,

Tainan 701, Taiwan

4 Department of Physics, National Taiwan University, Taipei 10617, Taiwan
5 Department of Physics and Engineering Physics, Tulane University, New Orleans,

Louisiana 70118, USA

(Presented 2 November 2016; received 24 September 2016; accepted 19 December 2016;
published online 15 March 2017)

Magnetostriction caused by the exciting variation of the magnetic core and the current
conducted by the winding wired to the core has a significant result impact on a power
transformer. This paper presents the sound of a factory transformer before on-site
delivery for no-load tests. This paper also discusses the winding characteristics from
the transformer full-load tests. The simulation and the measurement for several trans-
formers with capacities ranging from 15 to 60 MVA and high voltage 132kV to low
voltage 33 kV are performed. This study compares the sound levels for transformers
by no-load test (core/magnetostriction) and full-load test (winding/displacement ε).
The difference between the simulated and the measured sound levels is about 3dB. The
results show that the sound level depends on several parameters, including winding
displacement, capacity, mass of the core and windings. Comparative results of mag-
netic induction of cores and the electromagnetic force of windings for no-load and
full-load conditions are examined. © 2017 Author(s). All article content, except where
otherwise noted, is licensed under a Creative Commons Attribution (CC BY) license
(http://creativecommons.org/licenses/by/4.0/). [http://dx.doi.org/10.1063/1.4978759]

I. INTRODUCTION
Recently, there are many functional requirements related to the environment protection for
power transformer, which discuss reducing energy consumption problem, but also focus to con-
trol disturbance noise from electromagnetic device. It is clear that to make a right strategy to
reduce noise pollution, whether in advance power substation or nearby resident requirement is
necessary.1 Power transformer core noise is generated by energizing the core of transformers.
In order to realize more transformer core electromagnetic and mechanical structure variation,
the simulated core model structure for power transformers is analyzed by finite element analy-
sis (FEA).2 Ilo et. al. presents an accurately-calculated model for power transformer sound level
and electromagnetic response compared with experimental results.3 These studies have covered
electromagnetic structure, mechanical structure and theoretical method to compare experimental
results.
However, aforementioned discussion for reduction of core magnetostriction and electromagnetic
displacement variation of winding on power transformer sound level characteristic, has not yet fully
explored. Section II is introduced the soft material magnetic properties. Theoretical method is intro-
duced in Section III. In Section IV, both core magnetostriction and winding displacement for power
transformer electromagnetic operation is constructed. Finally, experimental results and discussion is
compared in V.

2158-3226/2017/7(5)/056681/7 7, 056681-1 © Author(s) 2017

056681-2 Hsu et al. AIP Advances 7, 056681 (2017)

II. MAGNETOSTRICTION AND MAGNETIC PROPERTIES

The variation in the magnetostriction as a function of the core properties is measured as shown
in Figure 1. 30Z140, 27ZH100, 27PH100, 30ZH105 and 30P120 are presented in this paper.
The magnetostriction of grain-oriented Si steel was found to be extremely sensitive to com-
pressive stress applied along the rolling direction (RD) and to tensile stress along the transverse
direction (TD). The compressive stress increases the magnitude of magnetostriction, which can also
be affected by the sound level in the core. The magnetostriction results from the shorter tetragonal
axis to coinciding with the direction of the applied magnetic field. So, magnetism parameter in B-H
curve, magnetic induction, permeability and remanence for different cases of silicon steel has defined

FIG. 1. Sampling soft material of silicon steel, magnetostriction measurement results.

FIG. 2. FEA Simulation, winding performance: (a) total deformation, (b) three-phase winding modal shape, (c) HV winding
modal shape, (d) HV and tap modal shape.
056681-3 Hsu et al. AIP Advances 7, 056681 (2017)

by FEA, as shown in Figure 2. For magnetostriction theory, the rotation of moments to align with the
applied field can be modeled by the quadratic relation4
!2
3 M
λ = λs (1)
2 Ms
where λ s and M s , respectively, denote the saturation magnetostriction and magnetization. This creates
the need to develop a reliable technique with the multi step-lap joint and V-shape joint to reduce the
magnetostriction in grain-oriented Si steels under various stress conditions.

III. THEORETICAL METHOD AND MODEL STRUCTURE

A. Electromagnetic field theory
Maxwell’s theory can be used to calculate the electromagnetic field where the eddy current effect
of the iron core is ignored in transient magnetic field analysis. Then, the effect of magnetic induction
on magnetostriction stress in the silicon steel of the power transformer is built by using the partial
differential finite element equation as below:
∂ σ ∂A ∂ ∂A
! !
υ + = −J (2)
∂x y ∂x ∂y ∂y
where A is the vector magnetic potential, υyσ and υxσ are the magnetic resistance rates along the
rolling direction and vertical rolling direction, respectively, and J is the source current density. Also,
both the magnetic energy storage W of the core and the electromagnetic force on the winding to
observe displacement of the vibration is considered. Therefore, for transformer structure in a vibrating
situation, both the electromechanical force of the winding in displacement and the electromagnetic
force of the core in magnetostriction, can be simulated by FEA. The motion equation of harmonic
response analysis can be expressed as
 
−ω2 [M] + iω [C] + [K] ({u1 } + i {u2 }) = ({F1 } + i {F2 }) (3)
where M is the mass matrix, C is the damping matrix, K is the stiffness matrix, u is the vibration
displacement of the windings and core including the constraint loads and magnetostrictive strain
applied to the core and u1 is the real part and u2 is the imaginary part and F is the electromagnetic
force applied to the windings and core and F 1 is the real component and F 2 is the imaginary part.
The harmonic constraint load of magnetostrictive strain, and harmonic forces load of Lorentz and
reluctance are applied to the model separately. Bedsides, the analysis of electrodynamics forces in
windings are carried out by applying modern methods of Ansys software, which are correlated with
the voltage level and current flow. Modal analysis provides the definition of modal parameters of
windings as a mechanical structure and those are the resonant frequencies, suppression and modal
shapes.
B. Sound level caused from core and winding
The most popular calculation method for power transformer sound level is defined by the IEC
60076-10 (2001) standard and is determined by the below parameters.
Actually, transformer sound level is caused from rated operation. According to the reference of
S.-J. Cheng, et. al. research work,4 the correlation between the sound level and the magnetic properties
of the power transformer core is used to analyze the sound level of the transformer. The sound level
dBi (dB) of the power transformer is summarized as dBi (dB) = L WAC + L WAW . Both no-load (L WAC )
and full-load (L WAW ) for the sound level calculation are defined below. The transformer core sound
level L WAC calculation formula is considered as the following expression4,5
LWAC = C + 19 log GFe − 20 log Φ + K1 (Bm − K2 ) + K3 (4)
where C denotes the soft material so that is empirical coefficient. The parameters K 1 and K 2 are
constant values, and K 3 is a compensation factor for the operating frequency of 60 Hz. The parameter
GFe stands for the weight (ton) of the core, and the Φ denotes the cross-section length of core. The
056681-4 Hsu et al. AIP Advances 7, 056681 (2017)

notation Bm is the magnetic flux density (Tesla). Electromagnetic forces inside winding conductors
produce axial and radial winding vibrations with pitch frequency twice the frequency of electrical
distribution network (i.e. 120 Hz for network of 60 Hz). Higher order harmonics are insignificant.
That noise inside the windings is called load noise, in contrast to the magnetic core noise called
the no-load noise. Additional noise is produced by vibrations of terminals from the windings. The
transformer winding sound level for a certain load can be calculated using empirical expression:6
LWAW = K1 + K2 log Sr + 40 log α (5)
where S r is rated power of transformer, K 1 and K 2 are coefficients, and α represents relative load (load
current/rated current). In this study, all of transformer performed sound level testing, measurement
points should be taken at least a distance of 2,000 [mm], from the main tank and radiation, as
follow IEEE Std. Also, the number of measurement points is around 32 and should be no less than 6
measuring points.

IV. EXPERIMENTAL AND SIMULATION RESULTS

The sound level radiated by transformers is generated by a combination of magnetostrictive
deformation of the core and electromagnetic forces in the windings, tank walls, and magnetic shields.
Owing to the magnetic flux density in the laminations and the magnetic properties of the core steel,
the amplitude of core vibrations causes sound radiation. Based on this, the magnetic properties of

FIG. 3. FEA Simulation, core performance: (a) transformer core no-load Bm , (b) transformer core full-load Bm , (c) total
deformation at 116Hz, (d) total deformation at 179 Hz.
056681-5 Hsu et al. AIP Advances 7, 056681 (2017)

the magnetostriction of the core are measured. From Eqs. (2) and (3), it is evident that the winding
displacement is independent of the load current.7 Figure 2 shows the simulated results of a transformer
energized at a fundamental frequency of twice of 60 Hz, winding deformation, and modal shape,
occurred at 274Hz, 378Hz, 555Hz, etc. Figure 3(a) and 3(b) shows that the magnetic induction of
the core operated at no-load and full-load condition are different. Moreover, this indicates that the
sound level depends on the magnetic induction of the core. This study also simulated the core modal
shape results, at 116Hz, 179Hz etc., as shown in Figure 3(c) and 3(d). The core design, as well as the
use of lower induction levels, has reduced the amount of sound generated in the core such that the
sound caused by the electromagnetic forces has become significant. Comparative values of designed
transformers, simulation results, and experimental results are summarized in Table I. The force of
amplitude of the vibrations is proportional to the square of the current, and the radiated sound power
is proportional to the square of the vibrational amplitude. Consequently, the radiated sound power
is strongly independence on the load current. Table I shows that higher magnetic induction of the
core assembly with a lower current density in the winding dependent on sound level (case 1) has
shown sound power of full-load, larger than above of 5 dBA. Both case 2 and 3 have significantly
different winding displacement. Basically, as design values in magnetic induction and current get
higher, both simulation and design values for no-load and full-load testing in sound level is almost
the same.
It is interesting that in cases 4 to 6 transformer design condition, especially with the full-load
parameter, the sound level is not only decided by the magnetic core, but the winding current density
and displacement effect the sound level is significant. Comparative results of capacity, core induction,
and mass are almost decided the sound under no-load test. Table I indicates that opposite-results as
current density and displacement of winding, both no-load and full-load sound results are significant.
Furthermore, Eqs. (4) and (5) represent the sound level calculations for the core operating at no-load
and the winding operating at full-load, respectively. The transformer sound level FEA simulation
results are shown in Figure 4. For full-load condition, the winding current around 1.667 times than that
of no-load condition due to energy transformation, are designed and simulated, as shown in Table I.
Figure 5 illustrates the experimental environment and the sound level testing for the transformer
with IEEE standard. To obtain sound level testing results for no-load and full-load condition, the
transformer design, simulation, and measurement results are compared. It is well-known that the
transformer sound levels are almost proportional to the mass of the core and winding, as shown in
Figure 6. It indicates that the winding displacement parameter and the current density for windings are
completely different. The vibration displacement contours of the windings and the core are obtained
as shown in Figure 7. It indicates that the vibration displacement contour is asymmetric compared

TABLE I. Transformer parameter in design, simulation and experimental results.

TR design parameter Sound level (dBA)

Core Winding Calculated FEA Experiment

Bm , GFe , I, GW , No Full HV ε, No Full No Full
Category Tesla Ton A/mm2 Ton load load µm load load load load

Case 1,
1.33 10.4 0.88 0.97 48.3 41.5 8.7 41.5 50.3 40.3 48
15-MVA
Case 2,
1.49 15.6 0.84 9.43 55.4 60.7 12.8 48.6 52.4 46.9 50.2
25-MVA
Case 3,
1.47 14.5 0.83 10.78 54.6 66 13.0 51.9 53.1 50.3 50.9
25-MVA
Case 4,
1.47 25.7 0.33 15.55 60.3 67.1 13.4 58.3 56.6 56.7 54.1
36-MVA
Case 5,
1.46 25.7 0.66 14.46 59.8 66.4 15.8 51.6 55.3 49.9 52.9
40-MVA
Case 6,
1.43 23.2 0.62 15.06 55.2 58.5 14.7 65.1 68.7 64.7 66.1
60-MVA
056681-6 Hsu et al. AIP Advances 7, 056681 (2017)

FIG. 4. Transformer sound level simulation results at full-load test.

FIG. 5. Case study power transformer sound level testing: (a) testing environment, (b) measurement sampling point with
distance 2 meter.

with other parameter such as the current density. A possible reason for the summarized winding
displacement could be that, in reality, the deformation of the high-voltage winding is larger for the
designed core noise than for the full-load noise, and as the winding displacement increases depending
on decrease in current density, the sound level between no-load and full-load is different because the
056681-7 Hsu et al. AIP Advances 7, 056681 (2017)

FIG. 6. Effect of Winding displacement and current density on sound level (dBA) of power transformer, with six case of 15
to 60MVA.

FIG. 7. Transformer simulated and experimental in sound level results (dBA).

displacement variation of winding is changed. Conversely, the designed sound values of full-load
tests might not be proportional to the measured results of the sound level because of the variations in
the winding displacement and the influence of current flow-induced electromechanical vibration on
the transformer winding is different, and between FEA and experimental result is less than 3 dBA,
as shown in Figure 7.

V. CONCLUSION
In this study, the simulation and the measurement for several transformers with capacities ranging
from 15 to 60 MVA and 33 to 132 kV are performed. The paper presents the difference between the
transformer sound levels for no-load (core/magnetostriction) and full-load (winding/displacement)
tests. The sound level is proportional to the designed values, which including magnetic flux density and
core volume. However, the sound level will also depend on winding parameter because of the winding
displacement variation and the current flow-induced electromechanical vibration. The sound level
from an FEA model of electromechanical structure made of an entirely soft material are compared
with the measurement data. The difference between two types of results is less than 3dB.
1 G. Baguley et al., “The impact of vibration due to magnetostriction on the core losses of ferrite toroidals under DC bias,”
IEEE Trans. Magn. 47(8), 2022–2028 (2011).
2 G. F. Mechler, “Magnetic flux distributions in transformer core joints,” IEEE Trans. Power Deli. 15(1), 198–203 (2000).
3 A. Ilo et al., “Influence of geometric parameters on the magnetic properties of model transformer cores,” J. Magn. Magn.

Mater. 160(1), 38–40 (1996).

4 S.-J. Cheng et al., “Correlation of magnetostriction variation on magnetic loss and noise for power transformer,” J. Appl.

Phys. 117, 17E716-1–17E716-4 (2015).

5 C.-H. Hsu et al., “Suppressing magneto-mechanical vibrations and noise in magnetostriction variation for three-phase

power transformers,” J. Appl. Phys. 117(17), 17D524-1–17D524-4 (2015).

6 L. Lukic et al., “Aspects of design of power transformer for noise reduction,” 4th international conference noise and

vibration, 255–262, 17-19, Oct 2012.

7 Y.-H. Chang et al., “Reducing audible noise for distribution transformer with HB1 amorphous core,” J. Appl. Phys. 109,

07A318-1–07A318-3 (2011).