See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/263733864

Test Requirements for Seismic Qualification of Substation High-Voltage

Equipment and Recommendations for New Edition of IEEE693 Standard

Conference Paper · July 2014

CITATIONS READS

0 2,132

1 author:

Shakhzod Takhirov
University of California, Berkeley
49 PUBLICATIONS   562 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Structural assessment based on utilization of point clouds collected by laser scanners View project

Seismic Qualification of High Voltage Equipment View project

All content following this page was uploaded by Shakhzod Takhirov on 08 July 2014.

The user has requested enhancement of the downloaded file.

 Proceedings of the 9th International Conference on Structural Dynamics, EURODYN 2014
Porto, Portugal, 30 June - 2 July 2014
A. Cunha, E. Caetano, P. Ribeiro, G. Müller (eds.)
ISSN:
Proceedings of the 9th International Conference on2311-9020; ISBN: 978-972-752-165-4
Structural Dynamics, EURODYN 2014
Porto, Portugal, 30 June - 2 July 2014
A. Cunha, E. Caetano, P. Ribeiro, G. Müller (eds.)
ISSN: 2311-9020; ISBN: 978-972-752-165-4

Test Requirements for Seismic Qualification of Substation High-Voltage Equipment

and Recommendations for New Edition of IEEE693 Standard
Shakhzod Takhirov1, Eric Fujisaki2, and Amir Gilani3, and
1
Pacific Earthquake Engineering Research Center, UC Berkeley, Richmond, CA, USA
2
Pacific Gas and Electric, San Ramon, CA 94583
3
Miyamoto International, Sacramento, CA USA
email: [email protected], [email protected],[email protected]
ABSTRACT: In the past and recent earthquakes, high voltage substation components have suffered damage and this damage has
led to disruption in power distribution. As a result, electrical utilities have suffered financial losses and the lag in restoring
power, has caused delays in restoring essential services and in post-earthquake recovery. The IEEE693 (Recommended Practice
for Seismic Design of Substations) Working Group was formed in part to address the seismic vulnerability of high voltage
equipment and to formulate procedures for seismic qualification of such equipment. The standard requires shake table testing for
large substation equipment of higher voltage classes. As currently written, the standard provides significant latitude to users
(utilities, manufacturers, and researchers) with respect to the conduct of certain aspects of the test program. This choice was
motivated in part to allow testing at as many facilities as possible, given the limitations of such facilities at the time of
development of the standard. Although applicable at the time of its inception, many shake tables have been constructed in recent
years that would now allow for tighter test requirements. For example, many more facilities are now able to test to the High
Performance Level, conduct triaxial testing, and use more sophisticated arrays of instruments for data collection. A review of the
current test methodology was conducted and shortcomings were identified. Multiple test programs conducted at PEER, UC
Berkeley dealing with seismic qualification of disconnect switches were considered as examples for this purpose. The
recommendations follow the main idea of the IEEE693 standard aiming at ensuring acceptable seismic performance of sub-
station equipment by including utilities, manufacturers, researchers and practicing engineers in the development of the standard.
Recommendations for development of a more robust experimental procedure were formulated. Examples of the application of
such revisions to the test standard were then evaluated and found to provide a more comprehensive mean of characterizing the
equipment response.

KEY WORDS: Lifeline systems, electrical substation equipment, seismic qualification, IEEE693, shake table testing.

1 INTRODUCTION in the development process of the standard has been proven to

In past and recent earthquakes, high voltage substation be very beneficial. One of the representative examples of the
components have suffered damage leading to disruption in close relationship with academia is the development of a
power distribution [1]. As a result, electric utilities have prescribed time history for seismic qualification testing of
suffered financial losses and the lag in restoring power substation equipment [3]. Active participation of all interested
delayed the restoration of essential services and post- parties [4] is always supported by the IEEE693 Working
earthquake recovery. Lengthy power outages may potentially Group.
inflict significant economic damage on affected communities As a result the IEE693-2005 standard represents a
or whole regions. The recent earthquakes in New Zealand and consistent, up-to-date, and self-contained document. The
Japan serve as representative examples of such failures as well standard initially gained acceptance among North American
as the challenges of the recovery efforts. The IEEE693 utilities, but due to its continuous development and its quality,
Working Group was formed in part to address the seismic it has become a popular document used worldwide. The main
vulnerability of such high voltage equipment and to formulate objective of this paper is - as a part of this continuous
procedures for seismic qualification of such equipment in the development process - to review of the current state of seismic
standard. qualification testing of substation equipment and identify the
The current version of the IEEE standard covering the areas of the standard that needed to be addressed. The review
seismic qualification of substation equipment was published is based on more than fifteen years of testing and test
in 2006 [2]. The standard requires that equipment and witnessing experience by the authors of this paper, which
components of higher voltage classes must be seismically have conducted and participated in tests at many laboratories
qualified by multi-axial testing on earthquake simulators worldwide.
(shake tables). In general, equipment requiring shake table
2 AMPLIFICATION OF SUPPORT STRUCTURE
testing possesses vulnerable characteristics, such as massive
tall insulators or other components that have historically Substation components or entire equipment assemblies are
performed poorly in earthquakes. often installed on a supporting structure to provide structural
The IEEE693 Working Group is developing the next integrity as well as achieve electrical clearance from the
version of the standard and bases its requirements on the ground. Examples include: a disconnect switch on a support
feedback received from open discussions with academia, structure, bushings on top of a transformer, a circuit breaker
utility companies, consulting engineers, and equipment on a supporting structure and many others. Since all
manufacturers. The involvement of all such interested parties equipment requiring shake table testing must be tested full-

387
Proceedings of the 9th International Conference on Structural Dynamics, EURODYN 2014

Proceedings of the 9th International Conference on Structural Dynamics, EURODYN 2014

scale (preferably in the as-installed configuration), the Spectral accelerations of the table and support structure
feasibility of seismic qualification by testing can be limited by computed for High PL run are presented in Figure 2, which
the capacity of a test laboratory. These limitations may show that the spectral accelerations on top of the support
include the clearance above the shake table, the footprint of structure were amplified in a wide range of frequencies with
the support structure and the equipment itself. The allowed some significant amplification in the vicinity of the first and
payload of a shake table, displacement and velocity limits are second mode frequencies of the system. This plot is typical for
also significant factors to consider. Utilities may often use all three directions of testing and only the X direction of
support structures of their own design, and certain types of shaking is shown in this figure.
equipment may be installed with widely varying support Ratios of spectral accelerations on top of the support
configurations depending on functional needs. structure to that of the table have a similar trend for all three
To address these issues, the standard allows the testing of directions of testing, as shown in Figure 3. Spectral
electrical equipment without a support structure under amplifications in both horizontal directions significantly
amplified strong motion that accounts for the effects of a exceeded the standard’s factor of 2.5, with the most
support structure. The shortcomings of this approach are amplification seen in the X axis (direction of the switch with
discussed in this paper and are based on the extensive research lower stiffness, perpendicular to the switch base). The latter
program conducted on the 550-kV vertical break disconnect amplification is more than two times greater than the 2.5
switch at the University of California, Berkeley as part of the factor with the largest amplification is as high as 5.5
Lifelines Program [5]. The switch was tested using many 12
configurations, including the two major ones: main blade Shake table
Top of support
closed and main blade open. The latter configuration was 10 IEEE693
tested with amplified strong motion obtained from the top of
the support structure.
Spectral acceleration, g
8

2.1 Support Structure is Unknown 6

The current version of the standard (IEEE, 693) allows

seismic qualifications testing when an actual support structure 4

is not yet known, by imposing an amplification factor to

required response spectra (RRS). A number of test results 2

show that the amplification factor of 2.5 set by the standard

0
for seismic qualification tests without a support structure may 10
0 1
10
be un-conservative in some cases and cannot alone account Frequency, Hz

for the complex 6D motion (3 translations and 3 rotations) of Figure 2. Spectral accelerations of table, top of support
the attachment points of the equipment on the support structure and current IEEE693 amplified spectra for unknown
structure. Although the Standard specifies that the 2.5 structure.
amplification factor be applied to translational as well as
rotational responses, the latter are very difficult to achieve in 6
practice. As a result, practitioners often neglect the effects of X direction
Y direction
support rotation. 5 Z direction

A representative example of this can be seen in test results IEEE693

of 550-kV switch tested at High Performance Level (PL,

Spectral amplification

defined by a standard spectral shape anchored to 1.0g in

IEEE693) at the University of California, Berkeley [5]. The 3

photo and the schematic drawing of the switch are presented

in Figure 1. 2

0
0 1
10 10
Frequency, Hz

Figure 3. Spectral ratios of shake table and top of support

structure.
The top-of-support spectra plotted in Figure 2 represents the
translational component of the coupled switch/ support
system. Since the switch has a significant mass compared to
the support, these spectra only apply to the specific assembly
being supported. A further complication is that the rotational
response at the top of support is not captured by applying a
(a) (b)
simple scale factor to the ground input motion.
Figure 1. Photo (a) and schematic drawing (b) of 550-kV Similar large spectral amplifications in a wide frequency of
vertical disconnect switch tested at UCB. ranges were observed in other tests as well. One example of

388
 Proceedings of the 9th International Conference on Structural Dynamics, EURODYN 2014

Proceedings of the 9th International Conference on Structural Dynamics, EURODYN 2014

seismic qualifications that had a similar outcome was the test This case represents a case when the model is perfectly
program conducted on three-phase disconnect switch with a defined: properties of the “switch’s model” and the “support
voltage rating of 245-kV, tested at the University of California structure’s model” perfectly reflect the actual ones. Three
at Berkeley [6]. Even though it is somewhat expected that the levels of tests with amplified excitation at 0.125g, 0.25g and
spectral amplification will be larger for the 550-kV switch 0.5g were performed. The test results were compared to test
mounted on a tall and flexible structure, it was surprising to data recorded for the switch with support structure. The
discover a somewhat similar amplification in the case of a differential displacement between the tips of so called Jaw
relatively rigid frame supporting the phases of 245-kV switch, post insulator (located at the opening side of the main blade)
as shown in Figure 4 (the switch in closed configuration is and Rotating post insulator (located under the hinge fixture
shown). coupling two posts together) was monitored during the tests.

Figure 5. 550-kV vertical disconnect switch attached directly

to shake table platform
Figure 4. Large spectral amplifications were also observed in
In addition, strains were monitored in all insulator posts
245-kV switch tests with relatively stiffer frame
including the Rigid post (stationary post on the hinge end of
Based on the discussion above, the amplification factor of the switch) where the strains had a tendency to peak at much
2.5 for unknown support structures may be un-conservative in higher values than at other posts. The comparison of these two
some cases. For equipment qualified by the generic 2.5 measurements for the switch on the support structure and
amplification factor, IEEE693 requires that the equipment be without it is shown in Figure 6 and Figure 7.
installed on a support structure designed such that the 40
amplification factor delivered is less than 2.25. Currently, Extension: Support
Retraction: Support
35
analytical parametric research and a detailed analysis of the Extension: No support
Retraction: No support
test results accumulated from fifteen years of seismic 30

qualification testing at the University of California at

Displacement, mm

25

Berkeley are currently in progress. These studies are expected

20
to provide data on the extent of coverage provided by the
currently specified amplification factor, and whether changes 15

are needed. 10

2.2 Structure Known and Well-defined 5

The current version of IEEE693 standard specifies the 0

0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55
following procedure for seismic qualification of substation PGA level of IEEE693 testing, g

equipment when the structure is known and well-defined. The

Figure 6. Change in distance between Rotating and Jaw posts
equipment and the support structure shall be accurately
modelled and a non-stationary analysis shall be performed on Based on the plots, the differential displacement between
the model. The amplified spectral accelerations shall be the Jaw and Rotating posts differ by about 300%. Therefore,
delivered from the analysis and these spectral accelerations the no-support configuration tested with an amplified motion
comprise target response spectra to be met or exceeded by the severely underestimates the amplitude of this motion and
shake table. Except for a 1.1 factor on the calculated behaves differently compared to the case with support.
amplification, the current version of the standard does not Frequency content of the displacement records was different
specify any other means to account for uncertainties and also which is not shown here due to limitation of the paper’s
assumptions associated with modelling. To implement this size. The strains in Rigid insulator post - which was the most
approach, the 550-kV switch was tested in closed overstressed post - were underestimated in the no-support
configuration with (configuration shown in Figure 1) and configuration by about 30% in transverse direction.
without support structure [5] as shown in Figure 5.

389
Proceedings of the 9th International Conference on Structural Dynamics, EURODYN 2014

Proceedings of the 9th International Conference on Structural Dynamics, EURODYN 2014

40
or cannot be expected to indicate acceptability of the part, the
Extension: Support shake table test is required to be supplemented by static
35 Retraction: Support
Extension: No support testing in which the parts are subjected to estimated PL
Retraction: No support
30 loading. Functionality of the equipment is then confirmed
after these tests. If the strains are below or at the allowable
Displacement, mm

25
values, the results of supplemental static testing are
20
acceptable, and electrical functionality is confirmed, the
15 equipment is considered to be qualified, and its performance
10 can be projected to the full High PL. This approach was
introduced in earlier versions of the standard to address the
5
following: (1) limited number of shake tables was available
0
0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 that could deliver the required motions, (2) available shake
PGA level of IEEE693 testing, g
table facilities had limited payload capacities, and (3) risk of
financial loss during testing of expensive equipment, (4)
Figure 7. Strains at bottom section of Rigid post insulator (the
safety concerns with testing large, brittle porcelain
most overstressed post)
components, some of them subjected to high internal gas
Based on the discussions provided above the following can pressures at shaking levels close to structural failure, (5)
be concluded. Even in this case of ideal match between uncertainty regarding the seismic capability of equipment that
physical components and finite element models the latter fail could actually pass full PL testing. After introduction of the
to fully represent complexity of actual equipment with a standard, several brand new shake table facilities have come
support structure. For the case discussed in this paper, as well on line and several old tables have been upgraded to achieve
as many cases encountered in practice, the effects of top-of- adequate payload capacities. In recent years, a number of full
support rotation are very difficult to address adequately. In High PL qualification tests of high voltage equipment have
many seismic qualification programs done commercially, been successfully completed.
modelling is done without any component tests taken into A typical example that shows shortcomings of the approach
account and in many cases no calibration of the model is is provided herein. In a vertical break disconnect switch its
performed which raises concerns about the validity of the final main blade must remain inside of the contacts when the
model. Even in the perfect (albeit ignoring the support switch is in closed configuration. In many cases the blade has
rotation effects) match situation the performance of the additional electrical shielding components in the vicinity of
equipment can differ significantly when support structure’s the jaw contact clip which limits the allowable relative motion
effect is substituted by an amplified motion. between the switch blade and the jaw clip as shown in Figure
This approach needs to be reconsidered and re-evaluated to 8.
achieve consistency in testing and seismic qualification. A
new reliable and conservative enough approach needs to be
developed to assess structural performance of the equipment
and its support as a coupled system with complex and
nonlinear properties.

3 PROJECTED PERFORMANCE OF QUALIFIED

EQUIPMENT
Another common approach of seismic qualification testing
permitted by the standard is extrapolation of the results of
half-level of testing to full PL. Equipment qualification by test
is often conducted at half of the PL (e.g., spectrum anchored Figure 8. Schematic drawing of jaw contact clip and blade
to 0.5g pga for the High level). The equipment is projected to shielding component
be capable of withstanding seismic loading at the full PL (e.g.,
spectrum anchored to 1.0g pga for the High level), provided Depending on the level of excitation, travel of the blade can
that the measured stresses are within acceptable limits. This exceed the allowable causing an impact loading which is
concept is based upon the assumption that the equipment difficult to predict. In addition, the switch components behave
behavior will remain linear up to the PL level of testing. In nonlinearly making prediction of its travel based on low level
many cases this is not a reasonable assumption, especially for testing almost impossible. A change in distance between
the electrical equipment composed of bearings, spring cans, Rotating and Jaw posts in the 550-kV switch tests [5] is
stops or bumpers and other mechanical devices that may discussed here. The displacement between Rotating and Jaw
exhibit nonlinear behavior. In addition, although electrical post was monitored directly by a position transducer. Five
function is verified following testing at the half-PL, there is no levels of seismic qualification testing were performed with
demonstration of electrical function at the full PL when using IEEE693 RRS anchored to peak ground accelerations of
this approach. For example, substation equipment can be 0.125g, 0.25g, 0.5g, 0.75g and 1.0g (High PL). As shown in
tested at half-High PL while the strains are monitored at Figure 8 peak displacement changed nonlinearly and it was
critical locations throughout the test article. For critical parts not possible to extrapolate performance at High PL from that
with complex geometry or where strains cannot be measured, at half-High PL.

390
 Proceedings of the 9th International Conference on Structural Dynamics, EURODYN 2014

Proceedings of the 9th International Conference on Structural Dynamics, EURODYN 2014

120
survive more than one large earthquake. Although aftershocks
Extension are a consideration, the demand loading from such events is
Retraction
100 IEEE693 generally substantially lower than the main shock.
The primary objective for the most if not all utilities is to
80 maintain or provide for rapid restoration of electric service
Displacement, mm

following a large earthquake. Beyond this performance goal is

60
the desire to limit damage to substation during the earthquake,
40
but at reasonable cost; these competing goals will continue to
require a balancing act for utilities and manufacturers. The
20 IEEE693 Working Group is currently considering proposals to
permit in the next version of the standard, replacement of
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
parts or even the entire test article (of the same design)
PGA level of IEEE693 testing, g between different PL tests. The authors support this approach
based upon the foregoing discussion.
Figure 9. Change in distance between Rotating and Jaw posts
For instance, from half-High PL seismic qualification 5 EQUIPMENT WITH PASSIVE SEISMIC
testing it would be assumed that travel of the blade’s contacts PROTECTION
will not exceed 60 mm (30 mm times two) in both retraction Passive seismic protection systems have found limited use
and extension, and the contacts on the Jaw end will never in electric substation equipment due to the general lack of
reach the stopper on the blade (in retraction). In reality, the experience possessed by manufacturers and utilities, life
switch’s performance was nonlinear, leading to significant cycle/ maintenance concerns, and development/ installation
increase in the relative displacement. As a result, the blade’s costs compared to conventional approaches. In addition,
contacts were forced to travel more than 105 mm in extension; conductor slack issues associated with increased (electrical)
the travel of the blade’s contacts in retraction was close to 80 terminal displacements in such systems have discouraged their
mm at which displacement it was restrained by the blade use. Large scale disasters highlighting the vulnerability of
shielding component. The latter can lead to impact loading on lifeline systems and the recent developments in improving
the Jaw post and should be avoided if at all possible. To seismic resilience of substation equipment via isolation/
address this issue the limiting component was relocated damping technologies [8] reveal a compelling need for more
further away as part of the Research and Development (R&D) definitive guidance for qualification of HV equipment with
effort by the switch manufacturer. A necessity of this simple seismic protection. High-voltage DC and ultra-high voltage
improvement of the switch’s blade would be completely AC equipment may have tall, massive porcelain insulators and
missed in a half-PL test. other fragile components such that passive seismic protection
The following conclusions can be drawn from this may provide one of the few feasible means for achieving good
discussion. First, seismic performance of substation seismic performance.
components may be nonlinear and can lead to under-testing of Little guidance for qualification of electric substation
the equipment and un-conservative conclusions about its equipment using seismic protection technologies can be found
performance. Second, functionality of the equipment is not in IEEE 693 and its counterparts from other standard-making
possible to predict and extrapolate from half-scale to full-scale bodies. A framework for such qualification may include the
PL testing. It is recommended that the use of half-PL testing following features.
as a basis of qualification either be removed from the Depending on feasibility of a shake table testing of this type
standard, or limitations added regarding its use. of equipment, the seismic qualification may be conducted via
detailed finite element analysis or by means of shake table
4 OPTION OF REPLACING TEST ARTICLE WITH tests of full-scale as-installed configurations. The seismic
BRAND NEW ITEM IN PL TESTS protection of the HV equipment can include a seismic
The current version of the standard does discuss the isolation system supporting the equipment and its support
replacement of a test article or its parts with brand new items structure, damping devices increasing effective damping of
between test runs with different configurations. As an the system, and many others. To address issues of proper
example, a disconnect switch with ground switch would be modelling, component testing is needed. The component
required to be tested in three configurations: main blade testing is also needed to demonstrate that the protection
closed and grounding blade open (C-O), main blade open and system is reliable with repeatable properties. The issue of the
grounding blade open (O-O), main blade open and grounding protection device’s aging/decaying during years of outdoor
blade closed (O-C). When PL testing is conducted, the installation while exposed to all environmental challenges
qualification test program would require that the test article be needs to be addressed. Maintenance issues are also a
subjected to at least three very large earthquakes, in which the significant concern for utilities.
equipment may be stressed at levels approaching failure. The 5.1 Seismic Qualification by Analysis
current standard specifies that at the Performance Level,
ductile parts may undergo minor yielding, and some If the seismically protected system cannot be tested in full-
permanent deformation is allowed without fracture of any scale as-installed configuration it may be qualified by analysis
part, provided that the functionality is maintained. It would as currently allowed by the standard for transformers and
seem unreasonable for the equipment to be expected to other oversized equipment.

391
Proceedings of the 9th International Conference on Structural Dynamics, EURODYN 2014

Proceedings of the 9th International Conference on Structural Dynamics, EURODYN 2014

FE analysis: isolation system or damping device response spectra should envelop the lowest resonant
component testing. To develop a FE model that closely frequency obtained in the resonant search tests.
represents the isolation system, component testing on the Shake table testing: time history testing. Only triaxial
isolation system is needed. The results of these tests would be full-scale testing is permitted. The seismic qualification
used for calibration of the nonlinear force versus displacement method for equipment with seismic protective devices should
relation assumed in the analysis. This testing would also be be by shake-table response histories testing. An option of
used to verify that the isolation system has the required using a site-specific spectrum as specified by the utility may
displacement/force capacity determined from the analysis of be used for evaluation.
the coupled equipment-isolation system (with an appropriate
factor of safety). The tests are recommended to be conducted 6 REQUIRED RESPONSE SPECTRUM
in a cyclic setting with incrementally increasing amplitude up As currently written, IEEE 693 requires most substation
to displacement level expected at High PL. equipment to be qualified by means of numerical analysis or
FE analysis: bounding analysis. To account in variations by full-scale shake table testing. The demand on the
in the damper/isolator properties, a bounding analysis should equipment is specified by required response spectra (RRS) of
be conducted. The lower bound should equal the value the strong motion imposed at the attachment points of the
obtained from the component tests but not greater than 85% of equipment. In case of numerical analysis, for complex
nominal value and the upper bound should equal the value structures with many modes in the seismic range, a detailed
obtained from the component tests but not smaller than 120% finite element model is required. The damping ratio is
of nominal value. The components and connections in the required to be measured directly or a conservatively low value
equipment and its anchorage should be designed from the (e.g., 2%) is permitted to be used. The lower frequencies of
worst case results obtained from the bounding analysis. the mathematical model should, if possible, also be verified by
FE analysis: dynamic analysis of seismically protected simple-bump or other specified test methods. In case of full-
system. To qualify the equipment to certain seismic scale testing by means shake table tests a resonance search
qualification level, a dynamic analysis of the equipment- test is required. The resonant frequency search test is for the
isolation system needs to be performed. The equipment and its determination of resonant frequencies and damping of
support structure could be modelled as linear if analysis shows equipment before and after each time history test. The data
that the components of the equipment remain elastic. The obtained from the test may be an essential part of an
isolation system needs to be modelled as nonlinear component equipment qualification; however, the test does not constitute
based on the results of the component tests. Alternatively, for a seismic test qualification by itself.
velocity- and frequency-independent devices, it would be Per the current version of IEEE693, a maximum damping
permitted to perform linear analysis using equivalent value of 2% may be assumed on all equipment and structures
properties of the isolation/damping system at the target and any claims of any damping beyond 2% must be
displacement. substantiated by testing. Historically, the determination of
FE analysis: seismic qualification level. For either damping of substation equipment has been focused on the
moderate or high qualification levels, the acceleration damping associated with the horizontal response of equipment
spectrum anchored to the performance level (Moderate or and with particular emphasis on the lower modes of vibration.
High PL) should be used. In reality, structural and non-structural systems can have
5.2 Seismic Qualification by Means of Shake Table Testing damping ratios other than 2%, depending on various factors
such as structural types, construction materials, the level of
If the seismically protected system cannot be tested in full- ground motion excitations and many others. For instance, the
scale as-installed configuration it may be qualified by analysis effective damping of the equipment can change during the
as currently allowed by the standard for transformers and time history tests especially during high PL testing due to
other oversized equipment. some plastic deformations expected during this severe motion.
In many cases the HV substation equipment is relatively For seismically isolated equipment the variation in damping
light weight, with limited foot print and height, and as such it can be related to physical properties of the isolation system
can be seismically evaluated by means of shake table testing. itself.
Shake table testing: component tests and bounding In most building codes and seismic design standards [9] the
analysis. Similar to the requirements of the seismic design spectra is provided at 5% damping ratio. If a structure
qualification by analysis, the isolation system and/or damping has a damping ratio different from 5% a special damping
devices should be subjected to component testing prior to the scaling factor (DSF) is used to modify the design spectra to
shake table testing and bounding analysis performed. that particular damping value. A similar approach was used in
Shake table testing: resonance search tests. Sine sweep the IEEE693 standard which specifies required response
testing should be conducted independently in three directions. spectra at 2%, 5%, and other damping and provides DSF for
For seismically isolated equipment, both the isolation system other damping ratios. The 2%-damped spectrum is specified
and equipment frequencies and damping should be identified. as the target for spectral matching when developing input
For isolated equipment and damping devices that are motions for testing.
displacement-dependent, a minimum of four sweeps with In recent research utilizing the updated, 2011 version of the
amplitudes ranging from 0.05g to 0.25g are proposed to be NGA database of ground motions recorded in worldwide
conducted. The sine sweeps may be replaced by equivalent shallow crustal earthquakes in active tectonic regions (i.e., the
white noise excitations at multiple levels to estimate the NGA-West2 database), dependencies of the DSF on variables
frequency dependency on the level of excitation. The required

392
 Proceedings of the 9th International Conference on Structural Dynamics, EURODYN 2014

Proceedings of the 9th International Conference on Structural Dynamics, EURODYN 2014

including damping ratio, spectral period, moment magnitude, IEEE693 RRS at 2% and RotD50 (M=7,D=10)
4
source-to-site distance, duration, and local site conditions are
examined [10]. 3.5
The comparison between the IEEE693 required response
spectra at 5% damping and the estimated RRS for other 3

damping ratios utilizing results of [10] is provided in Figure

2.5
10.
The plots show that newly modified DSF closely matches 2
the required response spectra specified in the current version
of the standard. The updated spectra would have slightly 1.5

higher demand in frequency range from 2Hz to 8Hz for 1%,

1
2%, and 10% damping ratios. Smaller spectral accelerations
would be required for lightly damped systems with resonant 0.5
TestQke4IEEE@2%
IEEE693@2%
frequency less than 1 Hz (see plot for 1% as an example). The RotD50@2%
latter will lead to lesser displacement demand, which is 0
-1 0 1
10 10 10
especially important and crucial for design of seismically
isolated equipment.
Judging from the magnitude of the spectral differences in Figure 11. Spectral accelerations of shake table motion based
Figure 10, the modifications of the required response spectra on TestQke4IEEE would still envelope the updated [10] RRS
are quite minor and most likely are not practical to introduce at 2%.
in the new version of the standard.
CONCLUSIONS
IEEE693 RRS at 5% and other damping ratios (M=7,D=10); 5% is taken as reference
4 More rigorous testing has led to the development and
IEEE693@1% manufacture of improved equipment. The results of the testing
3.5 IEEE693@2%
IEEE693@5% and analysis presented in this paper show that there is a need
3
IEEE693@10% to improve in areas of IEEE693 such as first support
RotD50@1%
RotD50@2%
amplification, qualification without known support structure,
2.5 RotD50@10% projected performance from half-scale tests, correlation
between finite element models and actual
2
equipment/components, and issues related to required
1.5 response spectra.
More research is needed on HV substation equipment with
1 passive protection devices to develop improved guidance for
0.5
the seismic qualification process of this type system.
Development of supplemental damping devices or other
0
-1 0 1
means for dealing with excessive terminal displacements are
10 10 10
needed.

Figure 10. Required response spectra estimated from IEEE693 ACKNOWLEDGMENTS

at 5% damping ratio by utilizing results of [10] Acknowledgements are due to Southern States, LLC for
To achieve consistency in testing and numerical analysis of providing access to test data and allowing publication of the
seismic qualification of substation equipment, IEEE693 test results. Special thanks are due to Pacific Gas and Electric
standard provides two pre-approved three-component time Co and Southern California Edison for constant attention to
histories satisfying certain requirements of the standard and quality of IEEE693 testing and their support of research in
enveloping the required response spectra at 2% damping ratio. this field.
One of them is empirically based input motion that was
REFERENCES
derived from a real strong motion recorded during the Landers
[1] Fujisaki, E. (2009). Seismic Performance of Electric Transmission
1992 earthquake. The strong motion called TestQke4IEEE Systems, Presentation at PEER Annual Meeting, San Francisco,
[3]. The spectrum of the shake table motion based on California.
TestQke4EEE would still envelope the updated spectra as [2] IEEE (2006). IEEE693: IEEE Recommended Practice for Seismic
shown in Figure 11 since test response spectra envelop the Design of Substations.
[3] Takhirov, S., Fenves, G., Fujisaki, E., and Clyde, D. (2005) Ground
required response spectra with some margin that is usually Motions for Earthquake Simulator Qualification of Electrical
much greater than the spectral improvements derived from Equipment. Pacific Earthquake Engineering Research Center,
[10]. University of California at Berkeley, PEER 2004/07.
This result supports the conclusion that the modifications of [4] Takhirov, S. and Gilani, A. (2009). Earthquake performance of high
voltage electric components and new standards for seismic qualification.
the required response spectra per [10] are most likely can be Proceedings of TCLEE-2009, Oakland, California: June 28-July 1.
avoided (at least for 2% damping ratio). [5] Takhirov, S., Fenves, G., and Fujisaki, E. (2005). Seismic Qualification
and Fragility Study of Line Break 550-kV Disconnect Switches, Pacific
Earthquake Engineering Research Center, University of California at
Berkeley, PEER 2004/08.

393
 Proceedings of the 9th International Conference on Structural Dynamics, EURODYN 2014

Proceedings of the 9th International Conference on Structural Dynamics, EURODYN 2014

[6] Takhirov, S. (2008). Seismic Qualification Report on EC-1, P and EV-1

Types of 245-kV Disconnect Switches with EVG-1 Grounding Switch
Installed. Report No. SS-SCE-2008, Earthquake Engineering Research
Center, University of California, Berkeley.
[7] Takhirov, S. (2011). Seismic Qualification Study of 550-kV Vertical
Break Disconnect Switch. Report No. PEER – STI/2011-02, Pacific
Earthquake Engineering Research Center University of California,
Berkeley, September 2011.
[8] Paolacci, F. and Giannini, R.. (2008). Study of the Effectiveness of Steel
Cable Dampers for the Seismic Protection of Electrical Equipment.
Proceedings of 14th World Conference on Earthquake Engineering.
October 12-17, 2008, Beijing, China.
[9] American Society of Civil Engineers (ASCE) (2010), ASCE/SEI 7-10:
Minimum Design Loads for Buildings and Other Structures, ASCE,
Reston, VA.
[10] Sanaz Rezaeian, Yousef Bozorgnia, I.M. Idriss, Kenneth Campbell,
Norman Abrahamson and Walter Silva. (2012). Spectral Damping
Scaling Factors for Shallow Crustal Earthquakes in Active Tectonic
Regions. PEER 2012/01.

394

View publication stats