IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 19, NO.

1, JANUARY 2004

131

Seismic Response of Transformer-Bushing Systems

Selahattin Ersoy and M. Ala Saadeghvaziri

AbstractSubstations are critical components in a power

system. Performance of substations during past earthquakes
has not been satisfactory and many key components such as
transformers and porcelain bushings have sustained significant
damage. Recently, several research investigations have been conducted to better evaluate, assess, and qualify seismic performance
of critical substation components with emphasis on transformers
and porcelain bushings. These studies include both analytical
and experimental research. These works have revealed that
understanding the seismic interactions among key equipments of a
substation (transformers, porcelain bushings, foundation, and interconnecting elements) is critical to proper assessment of seismic
performance of substations and in qualification of equipment.
Due to weight limitations, shaking table tests of transformers are
not feasible. Finite-element analysis, however, provides a suitable
platform to assess their seismic performance and to evaluate their
effect on the response of other important components such as
bushings. This paper will present development of finite-element
models and will discuss dynamic time history analyses results for
several transformer-bushings systems.
Index TermsBushings, seismic interaction, seismic response,
transformers.

I. INTRODUCTION

HE reliability of a substation exposed to earthquake

loading is dependent upon the seismic response of its
individual components and interaction of these components
with each other. Transformers perform the vital function of
transferring power between circuits operating at different
voltages. Their performance in past earthquakes has not been
satisfactory (Loma Prieta, Northridge). Unrestrained or poorly
anchored transformers have failed in recent earthquakes [e.g.,
U.S. (Northridge, 1994), Turkey (Izmit, 1999)]. Porcelain
transformer bushings have also failed in recent earthquakes
due to their interaction with transformers or other substation
components and their own inertia forces (dynamic response).
Transformers are usually placed on top of a concrete pad
or on rails without anchorage. Although this may be safe for
regions with no or low seismicity, installation of transformers
without anchorage should be avoided because of the significant
loss associated with the transformer damage during previous
earthquakes. It is recommended to anchor the transformer where
there is the possibility of seismic risk. This can be done by ei-

Manuscript received December 9, 2002. This work was supported by Earthquake Engineering Research Centers Program of the National Science Foundation under Award Number EEC-9701471 to the Multidisciplinary Center for
Earthquake Engineering Research.
S. Ersoy is with Greenman-Pedersen, Inc., Babylon, NY 11703 USA (e-mail:
selersoy@yahoo.com).
M. A. Saadeghvaziri is with the Department of Civil and Environmental Engineering, New Jersey Institute of Technology, Newark, NJ 07102-1982 USA.
Digital Object Identifier 10.1109/TPWRD.2003.820215

ther bolting the transformer to its footing slab for small transformers or by welding it to steel embedded in the slab for larger
transformers [2]. There are many different transformer configurations. Each transformer has its own geometric constraints and
maintenance requirements. This custom nature of transformers
may yield inadequate anchorage designs. Base isolation is another alternative for transformers, especially as a rehabilitation
scheme, in order to lessen the earthquake-induced accelerations
[3], [6], [7]. Base-isolation will also reduce the input acceleration into the bushing and will lessen the interaction between the
transformer and the bushing, which has been the cause of many
bushing failures during past earthquakes. Furthermore, by reducing the inertia forces base-isolation will reduce the level of
forces on internal elements. Cases of internal damage and electrical abnormalities during past earthquakes have been reported
[2]. Although inspections of the transformer internal structural
system after an earthquake are rare, it is reasonable to assume
that reliability and longevity of a transformer is related to the
level of shaking of internal elements. An issue with the use of
base-isolation that demands careful consideration is the effect
of large displacements. Large displacements will aggravate the
interaction between the transformer-bushing systems and interconnecting equipment.
Recently, several experimental studies have been conducted
on evaluating the seismic response of porcelain bushings [4],
[8], [9]. In these studies, bushings tests were performed using a
stiff supporting frame. Even though the most vulnerable flange
to porcelain gasket detail has been used in these tests, the performance of 196- and 230-kV bushings were good in terms of
the general response based on the qualification of bushings set
forth in IEEE 693-1997 [8]. However, many bushings of the
same type have failed in past earthquakes. This points to the
need for reassessment of the current IEEE 693-1997 qualification procedures for both transformers and bushings to include
consideration to their interaction. Electrical equipment components are typically designed for electrical requirements more
than structural performance requirements. Further research is
needed to quantify the effect of transformer on bushing dynamic
characteristics and its seismic response. Furthermore, interconnecting substation components can complicate the seismic response of transformer-bushing system. Interconnecting equipments can cause damage through connectors. That is, the critical condition for bushings can be either due to shaking of the
transformer tank or loads at the terminal end of the bushing due
to the out-of-phase shaking of the transformer-bushing with respect to interconnecting equipment. Therefore, shake table tests
of bushings on a rigid frame will not reveal all of the critical situations. The finite-element method provides the ideal platform
to perform additional studies in order to better understand the
response characteristics of transformer-bushing systems.

0885-8977/04$20.00 2004 IEEE

132

IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 19, NO. 1, JANUARY 2004

In the following sections, IEEE seismic qualification procedures, finite-element modeling techniques, and analysis procedures for transformers and bushings are summarized first. Then,
the finite-element results are discussed in light of transformer
and bushing interaction.

II. IEEE SEISMIC QUALIFICATION PROCEDURES FOR

TRANSFORMERS AND BUSHINGS
Seismic qualification tests are used to demonstrate through
experimentation that a piece of equipment is able to perform
its intended function during and after an earthquake. In the
U.S., electrical equipment is seismically qualified using a
standard developed by the IEEE. The IEEE standard entitled
IEEE 693-1997 Recommended Practices for Seismic Design
of Substations details procedures for qualification of electrical
substation equipment for different seismic performance levels
(high, moderate, and low) [5].
For qualification, the transformer is required to satisfy static
analysis using 0.5 g in two horizontal directions and 0.4 g simultaneously in the vertical direction. Its appendages, except bushings, are required to satisfy static analysis using 1.0 g in two
horizontal directions and 0.8 g simultaneously in the vertical
direction. However, IEEE 693-1997 states that bushings rated
at 161 kV and above must be qualified using three-component
earthquake-simulator testing. Because it is impractical to test
bushings mounted on a transformer, IEEE specifies that bushings must be mounted on a rigid stand for earthquake testing and
qualification. IEEE 693-1997 identifies several response spectra
of identical shape but of different amplitudes for the qualification of a transformer bushing on an earthquake simulator.
Performance levels (PL) for substation equipments are represented by a response spectrum that is anchored to a peak ground
acceleration (PGA) of 0.5 g for moderate-level qualification and
1.0 g for high-level qualification. Since it is often impractical
to test components to the PL, IEEE 693-1997 permits equipment to be tested using a reduced level of shaking called the
required response spectrum (RRS). RRS corresponds to a PGA
of 0.25 g and PGA of 0.5 g, for moderate-level and high-level
qualification, respectively. To account for the amplification of
earthquake motion due to the flexibility of the transformer and
the local flexibility, IEEE 693-1997 states that the input motion
measured at the bushing flange shall match a spectrum with ordinates twice that of the RRS, termed as the test response spectra
for mounted equipment (TRSME). The PGA for the TRSME
spectrum is, therefore, 0.5 g for moderate-level qualification and
1.0 g for high level qualification. For this level of shaking, IEEE
recommendations are as follows.
1) The stresses in nonductile components must be less than
one-half the ultimate stress.
2) The factor of safety against oil leakage must exceed 2.
Anchorage is stated as one of the most cost-effective
measurements to improve the performance of inadequately anchored equipment. Anchorage must withstand the shear, uplift,
and compressive forces resulting from the design earthquake.

III. MODELING AND ANALYSIS ISSUES AND RESPONSE

CRITERIA FOR TRANSFORMER-BUSHING SYSTEMS
A typical power transformer is composed of six parts: transformer tank, radiators, reservoir, core and coil, mineral oil, and
bushings. The transformer tank is the main structural component of power transformers. Side walls of the tank are braced
with channels, however, the top plate is not stiffened for the
transformers studied in this paper. It has a core and a coil centrally placed within 2/3 of transformer height and the tank is
completely filled by mineral oil. Radiators and reservoirs are
appendages that are externally attached to the transformer tank.
Three different sizes of power transformers were selected for
time history analysis. First transformer type is 25 MVA-650
HV BIL and it is called transformer type 1 (TT1) in this study.
This transformer weighs about 179 kips. The dimensions of this
in., L
in., H
in. (B,
transformer are B
L, and H represent width, length, and height of the transformer
tank, respectively). Transformer wall thickness is 0.375 in. and
bushing is supported on 1.5-in.-thick transformer top plate. The
second transformer type is 33/44/55-MVA 230/133-HV threephase transformer and it is called transformer type 2 (TT2) in
this paper. It weighs about 300 kips. The dimensions of this
in., L
in., H
in. Transtransformer are B
former wall thickness is 0.375 in. and bushing is supported on
0.75-in.-thick transformer top plate. Third transformer type is
250 MVA 230/119.5 kV and it is called transformer type 3 (TT3)
hereafter in this study. Its weight is about 512 kips. The dimenin., L
in., H
sions of this transformer are B
in. Transformer wall thickness is 0.375 in. and bushing is supported on 0.75-in.-thick transformer top plate.
The finite-element package ANSYS is used for development of the finite-element model [1]. The transformer tank is
modeled by shell elements. Braces around the transformers are
modeled by offset beam elements. The core and coil inside
the transformer are modeled as mass elements. The radiators
and the reservoir are modeled by three-dimensional (3-D) solid
elements. The contained oil inside the transformer was modeled
with fluid elements in the early stage of this study. Later, the
contained oil is modeled as a solid with modulus of elasticity
equal to the bulk modulus of the fluid since the transformer
is filled completely with oil and there is no sloshing effect
under consideration. The results obtained from both methods
were the same, but modeling of the contained oil by solid
elements is more computing time efficient. These three types of
transformer all support three 196-kV bushings that are located
on top of the transformer tank.
Bushings are composed of several elements like an aluminum
support unit, porcelain units, gaskets, aluminum core, and dome.
The aluminum support has a built-in flange used to mount the
bushing on top of the transformer. The aluminum core runs from
the top to the bottom of the bushing and houses the aluminum
conductor. Bushings are prestressed through the aluminum core
and this prestressing force is distributed evenly to the other components through the dome to hold the units together. Prestressing
force in finite-element model of bushing is introduced by specifying an initial strain to the aluminum core. There are gaskets
located in between the units. The finite-element model of one

ERSOY AND SAADEGHVAZIRI: SEISMIC RESPONSE OF TRANSFORMER-BUSHING SYSTEMS

133

BUSHING (TYP.)
RESERVOIR
(SOLID ELEMENTS)
SHEAR SPRING
(TRUSS ELEMENT)
CONNECTIONS
(BEAM ELEMENTS)

RIGID ELEMENTS
(TOP AND BOTTOM)

GASKET DETAIL

AXIAL SPRINGS (TYP.)

(TRUSS ELEMENTS)

TRANSFORMER TANK
(SHELL ELEMENTS)
DOME
(BEAM ELEMENTS)

RADIATOR
(SOLID ELEMENTS)

PORCELAIN UNITS
(BEAM ELEMENTS)

GASKETS
SUPPORT POINT TO
TRANSFORMER TOP

BRACES (TYP.)
(OFSET BEAM
ELEMENTS)

ALUMINUM SUPPORT
(BEAM ELEMENTS)

BUSHING

TRANSFORMER

Displacement (inch)

Fig. 1. Transformer mesh and element types.

1

max. displacement=0.59747 inch

0.5
0
-0.5
-1

10
Time (sec)

12

14

16

18

20

10
Time (sec)

12

14

16

18

20

Acceleration (g)

10
max. acceleration=5.2369 g
5
0
-5

Normalized Power Spectrum

-10

Fig. 2.

1
[BUC12-X]
f* =9.9 hz
0.5

10

15

20
Frequency (Hz)

25

30

35

40

Displacement, acceleration responses, and normalized power spectrum at BUC12 of TT2 for El-Centro record in x-direction.

of the transformers with element types and details is shown in

Fig. 1. Based on this information, the analytical models for the
bushings were created by beam elements with equivalent density and stiffness to represent the porcelain units (average 11
in. in diameter), the dome, and the aluminum core. Gaskets between these elements are modeled using linear axial and shear
springs. The total axial stiffness is introduced as
and the shear stiffness is defined as
. In these equations, is the area of the gasket, is youngs modulus, is the
shear modulus, and the is the thickness of the gasket.
Since anchoring of a transformer at its corners is a common
practice, the transformer models are fixed at each corner of the
transformer. Time history analyses are performed for ground

motion input with PGA of 1 g in orthogonal horizontal directions and PGA of 0.8 g in the vertical direction as per IEEE
recommendation. For each transformer type, 2-soil (El-Centro,
Hollister airport) and 2-rock (Pacoima dam, Lake Huge array
#4) earthquake records are utilized for 3-D time history analysis. One of the response spectrums of the earthquake motions
(soil record) used for orthogonal horizontal components and
vertical component, together with the IEEE high performance
level spectra is shown in Fig. 3 for 2% damping. In this study,
earthquake records are not anchored to the IEEE response
spectra but they are scaled to desired PGA.
Response is very sensitive to the value of damping ratio. Consistent with IEEE 693-1997, 2% damping value was employed

134

IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 19, NO. 1, JANUARY 2004

4.0

Response Spectra, Acceleration in units of g

IEEE - High SPL (2% Damping)

El-centro (S00E) (2% Damping)
El-centro (S90W) (2% Damping)
El-centro (UP) (2% Damping)

3.0

2.0

1.0

0.0
0.1

1.0

10.0

100.0

Frequency (hz)
Fig. 3.

Acceleration response spectra for components of El-Centro record and IEEE high-performance level.

in the finite-element model. Rayleigh damping is used for all of

the time history analyses and the Rayleigh damping coefficients
were obtained by fixing the damping value at 0.02 for frequencies of 8 and 25 Hz. These frequency values were selected based
on the response frequencies of the transformer bushing systems.
The minimum damping value obtained in this frequency range
is 1.71%.
Five fragility criteria for bushings were identified in this
study:
1) gap between the porcelain units and/or aluminum components and the gasket;
2) stress/strain levels in the gasket;
3) stress/strain levels in porcelain units;
4) top displacement of bushing based on slack provided;
5) reduction in prestressing force.
The first two of these criteria are related to the gasket. The
third and fourth items are associated with the strength of the
porcelain units, which may cause failure of bushing such as
cracking of the porcelain due to high pressure or insufficient
slack provided between bushing and the connecting equipment.
There have been examples of failures caused by insufficient
slack during past earthquakes. Item 4 is not part of the finite-element study discussed in this paper.
A. Discussion of Results
FE responses are monitored at several locations throughout
the height of the transformer and the bushing. Displacement,
velocity, and acceleration responses on the transformer tank are
obtained for five levels at each corner and the center point between the corners through the height of the transformer. Displacement, velocity, and acceleration responses of 12 points for
each of the three bushings are also monitored. These 12 points

are top of bushing, bottom of bushing, and another ten points

corresponding to the top and bottom of the five gaskets per
bushing.
First, several modes of the transformers correspond to peripheral equipments (such as radiator or reservoir) and they do not
affect the overall response. Primary mode of response for all
three transformers is translation in the weak direction. The frequencies in this direction are 14.1, 13.8, and 11.7 Hz for TT1,
TT2, and TT3, respectively. In the orthogonal (strong) direction, the frequencies are 20.6, 22.4, and 18.5 Hz, respectively.
The maximum translation at the top of the transformer in the
direction (weak horizontal axis) is 0.12 in. for TT1. Those
of TT2 and TT3 are obtained as 0.14 in. and 0.21 in., respectively. It is observed that the translations in the direction are
always smaller than the ones in the direction due to higher
frequencies of the transformers in the direction. Comparison
of the displacement values at mid-height of the transformer to
the top of the transformer show that mid-height displacement
values are one-half of the top displacements. It can be deduced
that there is almost always a linear relationship for the displacement values throughout the height of the transformer. The maximum dynamic amplification factor (ratio of maximum acceleration response to PGA) for top acceleration is found to be 2
for TT1. Those for TT2 and TT3 are 2.4 and 2.5, respectively.
Therefore, it can be stated that the dynamic amplification due
to the transformer body stated as 2 by IEEE 693-1997, which is
based on spectral values at the base of the bushing to that at the
base of the transformer, is not always conservative [8].
The maximum vertical support reaction is 230 kips and the
corresponding horizontal reaction is 98 kips in the direction
(weak axis) for TT1. The maximum vertical support reaction is
256 kips and the corresponding horizontal reaction is 103 kips in
the direction for TT2. Similarly, the maximum vertical support

ERSOY AND SAADEGHVAZIRI: SEISMIC RESPONSE OF TRANSFORMER-BUSHING SYSTEMS

The vertical tipping reaction

135

can be calculated as
kips

The total vertical reaction

in the most critical corner is

kips

The horizontal reaction H can be calculated as

kips

Fig. 4.

Calculation of static forces for TT2.

reaction is 440 kips and the corresponding horizontal reaction

is 246 kips in the direction for TT3. The maximum reactions
were also computed by the static analysis method specified in
IEEE recommendations. In this finite-element study, time history analyses are performed for ground input with PGA of 1 g
in orthogonal horizontal directions and PGA of 0.8 g in the vertical direction. However, static analysis recommended for transformer tanks utilizes 0.5 g in horizontal directions and 0.4 g in
vertical directions, applied to the center of gravity of the transformer tank. To be able to compare the static analysis (working
stress) values to the time history analyses (ultimate) results,
static calculations are performed for accelerations of 1.0 g in
horizontal directions and 0.8 g in vertical directions, applied to
the center of gravity of the transformer (2/3 h from the bottom
of the transformer). Dimensions of the transformer tanks were
given earlier in this paper. Static calculations are carried out to
get vertical and horizontal reactions. The vertical reaction is 236
kips and the horizontal reaction is 45 kips in the horizontal weak
axis, for TT1. The vertical reaction is 360 kips and the horizontal reaction is 75 kips in the horizontal weak axis, for TT2.
Similarly, the vertical reaction is 519 kips and the horizontal reaction is 128 kips in the horizontal weak axis, for TT3. Sample
calculation for TT2 (see Fig. 4) is given below. The overturning
moment in the direction, including uplift, is
kips
kips

The vertical tipping reaction

kips

can be calculated as

kips
The overturning moment in the
kips

direction

kips

kips

It is seen that the vertical reaction obtained from the IEEE

recommendations is 3% more than the finite-element analysis
result for TT1. Similarly, the vertical reactions are 41% and 18%
more than the finite-element analysis results for TT2 and TT3,
respectively. The static analysis results for vertical reactions are
well within the finite-element analysis result. It is also noted
that the horizontal reaction obtained from the recommendations
is 54% less than the finite-element analysis result for TT1. Similarly, the horizontal reactions are 27% and 48% less than the
finite-element analysis results for TT2 and TT3, respectively.
The finite-element analysis results for horizontal reactions exceed the static analysis results.
Failure of transformer and its components other than bushing
was not encountered in this study. Stresses in the transformer
tank are below the allowable material stresses. However, the
base forces are so high that providing proper anchorage may
be unwieldy. Failure of the bushings occurs at the gap between
the porcelain units.
Fig. 2 shows relative displacement, acceleration response,
and normalized power spectrum of the acceleration response
at the top of the center bushing (BUC12) mounted on TT2.
Bushing frequency is about 10 Hz from normalized power spectrum, which is the same as the value obtained from frequency
analysis for the first translational mode. The fundamental frequency of the bushing, when fixed at the level of the transformer
flange, is 14.4 Hz. Thus, the transformer flexibility has a significant effect on the dynamic characteristics of the bushing. As a
general tendency, the translation mode of response in the weak
direction (i.e., direction) of the transformer effects the input
into the bushing by filtering the motion and causing excitation of
the fundamental mode of the bushing. For the bushing mounted
on TT1, the fundamental frequency is reduced to 11 Hz. As discussed for the bushing mounted on TT2, the fundamental frequency is 10 Hz, and it is 10.5 Hz for the bushing mounted on
TT3.
The maximum displacement response at the top of the bushings in the direction (weak horizontal axis) considering all
four records is 0.55 in. for bushings mounted on TT1. That of
the bushings mounted on TT2 and TT3 is as 0.56 in. and 1.2
in., respectively. The maximum total acceleration response at
the top of the bushings in x-direction (weak horizontal axis)
considering all four records is 6.2 g for bushings mounted on
TT1. Those for bushings mounted on TT2 and TT3 are 5.5 g
and 12.4 g, respectively. To be able to make comparisons consistent with IEEE 693, a PGA of 2.0 g is used for the bushing

136

when fixed at its flange. Maximum displacement is 0.18 in. and

the maximum acceleration is 3.0 g. This could be one of the reasons for the discrepancy between bushings poor performance
during previous earthquakes and their good to excellent performance under laboratory tests when supported on a rigid frame.
Axial responses of bushing gaskets are also analyzed. The
seismic response of the bushings is dominated by the behavior
of the gaskets between the porcelain units. One of the common
failure modes involves movement of the upper porcelain unit
relative to its support flange causing oil leakage. In this paper,
gap is defined as the opening at interface of any porcelain unit
and gasket. The gap near the aluminum (transformer) flange
of the bushing is always more critical than any other location
throughout the bushing height. The gap forms due to the relative vertical displacements of the units right above and below
the gaskets and the rotation of these units. During the time history analyses, no gap is observed in the bushings on transformer
TT1. However, gaps do form in the bushings during time history analyses of transformers TT2 and TT3. This indicated a
high likelihood of failure and some preventive measures would
be needed.
The porcelain units are fragile components; therefore, stress
and strain levels in these units have also been investigated in
this study. Their ultimate axial strain capacity is 4000 E-6 in/in.
It is observed that stress and strain levels in the porcelain units
are well below the allowable stresses and strains. The bushing
has an aluminum core housing the copper cables between equipment items and the coils throughout its length. The bushing is
post-tensioned through its aluminum core and the springs in
the metallic dome ensure the uniform distribution of this prestressing force. The relaxation effect of the axial dynamic force
is also examined. It is found that the effect of axial vibration of
the bushing on the prestressing is insignificant, indicating that
there has not been any prestress loss. The gap opening discussed
before is due to bending of the bushing as a result of its first
mode vibration in the horizontal direction.
As it was stated previously in this paper, the finite-element
models of the transformers were fixed at each corner of transformer bottom plate. The finite-element results show that the
anchoring of the transformer to its base does not prevent the interaction between transformer and bushing and causes bushing
failure for some of the cases. The levels of reaction forces are
also very high, some are even higher than what are predicted
based on IEEE 693-1997. For these reasons, implementation
of anchorage for retrofit of existing transformers can be difficult and costly and at the end, functionality of the transformerbushing system in the event of an earthquake will not be ensured. Furthermore, in many situations, for both new and existing transformers, a well-designed anchorage may only change
the mode of failure to the foundation. Boundary gaps due to back
and forth motion of transformers and rocking of transformers
and their footings due to soil-structure interaction have been observed during past earthquakes [2]. Therefore, in many cases,
the use of base-isolation for transformers could be an option
to alleviate these problems, especially for existing transformers
in high seismic regions [3]. But large displacement associated
with the use of any base-isolation technology is an issue to be
dealt with. This issue will be presented in another paper by de-

IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 19, NO. 1, JANUARY 2004

veloping a simplified model. The simplified model is developed

based on this finite-element study and the previous base-isolation study performed on transformers.
IV. CONCLUSION
Transformers and bushings are key components of substations. Their seismic performance during past earthquakes has
not been satisfactory. Recent work has revealed that understanding the seismic interactions among key equipments of a
substation (transformers, bushings, foundation, and interconnecting elements) is critical to proper assessment of seismic
performance of substations and in qualification of equipment.
In this paper, the finite-element method is employed to perform
an analytical study to better quantify the effect of transformer
flexibility on the dynamic response of bushing. Furthermore,
the results of the finite-element analyses and those of prior
work are used to develop a simplified model to investigate the
appropriateness of base-isolation as a rehabilitation strategy.
The following conclusions can be made based on this analytical
study.
Transformer tank flexibility affects bushing dynamic characteristics. This effect is mainly due to the flexibility of the top
plate. For fixed transformers, the translational mode of response
is the dominant mode. For those transformers modeled under
this study, the frequency of this translational mode in the weak
direction is near the frequency of the 196-kV bushings supported, resulting in amplification of bushing response. The level
of accelerations in the bushings is much higher than that predicted based on IEEE 693-1997. This explains the discrepancy
between laboratory and actual performance of bushings during
previous earthquakes. Further research on other type (different
material or design) of bushings is needed for better evaluation
of their seismic response.
ACKNOWLEDGMENT
The research was conducted within Thrust Area 1 Seismic
Evaluation and Retrofit of Lifeline Networks led by Professor
Masanobu Shinozuka. The results and conclusions are those of
the authors and do not necessarily reflect the views of the sponsors.
REFERENCES
[1] ANSYS, Engineering Analysis System, ANSYS Inc., Canonsburg, PA.
[2] A. J. Schiff, Ed., Guide to Improved Earthquake Performance of Electric Power Systems. Reston, VA: ASCE Manuals and Reports on Engineering Practice no. 96, 1999.
[3] S. Ersoy, M. A. Saadeghvaziri, K. Liu, and S. T. Mau, Analytical and
experimental evaluation of friction pendulum system for seismic isolation of transformers, Earthquake Spectra, vol. 17, no. 4, Nov. 2001.
[4] A. S. Gilani, A. S. Whittaker, and G. L. Fenves, Seismic evaluation and
retrofit of 230-kV porcelain transformer bushings, Earthquake Spectra,
vol. 17, no. 4, Nov. 2001.
[5] Recommended Practices for Seismic Design of Substations, IEEE Standards Dept., IEEE (1998), IEEE Std. 693-1997, Piscataway, NJ.
[6] N. Murota and M. Q. Feng, Hybrid base-isolation of bushing-transformer systems, in Proc. Structures Congr. Expo., P. Chang,
Ed. Washington, DC: ASCE, 2001.
[7] M. A. Saadeghvaziri, S. Ersoy, and G. Y. Liu, Experimental seismic
evaluation of base-isolated transformers, presented at the Proc.
IMAC-XIX Conf. Structural Dynam., Kissimmee, FL, 2001.

ERSOY AND SAADEGHVAZIRI: SEISMIC RESPONSE OF TRANSFORMER-BUSHING SYSTEMS

[8] A. S. Whittaker, G. L. Fenves, and A. S. J. Gilani, Evaluation of seismic

qualification procedures for high-voltage substation equipment, in
Proc. Structures Congr. Expo., P. Chang, Ed.. Washington, DC, 2001.
[9] J. Wilcoski and S. J. Smith, Fragility Testing of a Power Transformer
Bushing, U.S. Army Construction Engineering Research Laboratories
(CERL), Champaign, IL, Rep. TR 97/57, 1997.

Selahattin Ersoy was born in Turkey in 1973. He received the B.S. degree in
civil engineering from Middle East Technical University, Ankara, Turkey, in
1994, and the M.S. degree in civil engineering from the New Jersey Institute of
Technology, Newark, NJ, in 2002.
Currently, he is a Structural Engineer with Greenman-Pedersen, Inc.,
Babylon, NY. He was also a Structural Engineer with Wagh Engineers, P.C.,
from 1998 to 2002. His research interests include seismic isolation systems
and seismic evaluation and rehabilitation strategies for power systems.
Mr. Ersoy is a member of the American Society of Civil Engineers and AISC.

137

M. Ala Saadeghvaziri received the B.S. (hons.), M.S., and Ph.D. degrees from
the University of Illinois at Urbana-Champaign, in 1981, 1983, and 1988, respectively.
Currently, he is a Professor of Civil and Environmental Engineering at the
New Jersey Institute of Technology (NJIT), Newark, NJ, where he has been
since 1988. His area of specialization is structural engineering with emphasis
on nonlinear response of structures, finite-element and computational methods,
earthquake engineering, and structural applications of composite and renewable
materials.
Dr. Saadeghvaziri is a member of the American Society of Civil Engineers,
Earthquake Engineering Research Institute, and American Society for Engineering Education.