Professional Documents
Culture Documents
by
HOUMAN GHALIBAFIAN
A THESIS S U B M I T T E D IN P A R T I A L F U L F I L M E N T OF
T H E R E Q U I R E M E N T S F O R T H E D E G R E E OF
M A S T E R OF APPLIED SCIENCE
in
T H E F A C U L T Y OF G R A D U A T E STUDIES
(Department o f C i v i l Engineering)
T H E UNIVERSITY OF BRITISH C O L U M B I A
October 2001
In p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t o f t h e r e q u i r e m e n t s
f o r an a d v a n c e d d e g r e e a t t h e U n i v e r s i t y o f B r i t i s h C o l u m b i a , I
a g r e e t h a t t h e L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e
and s t u d y . I f u r t h e r a g r e e t h a t p e r m i s s i o n f o r e x t e n s i v e c o p y i n g o f
t h i s t h e s i s f o r s c h o l a r l y p u r p o s e s may be g r a n t e d b y t h e h e a d o f my
department o r by h i s o r h e r r e p r e s e n t a t i v e s . I t i s understood that
copying or p u b l i c a t i o n of t h i s t h e s i s f o r f i n a n c i a l gain s h a l l not
be a l l o w e d w i t h o u t my w r i t t e n p e r m i s s i o n .
The U n i v e r s i t y o f B r i t i s h Columbia
Vancouver, Canada
http://vvWvv.library.ubc.ca/spcoll/thesauth.html 9/28/2001
ABSTRACT
substations located on seismically active regions. They have sufficient slack, which
equipment under seismic loading. Generally, in the design process o f the interconnected
equipment and their connections to the flexible conductors, the seismic behavior o f the
conductors is not considered and the dynamic forces generated by them during an
earthquake event are not taken into account. However, these forces could be significant
and could overload the equipment and cause damage. This thesis presents an
evaluating their interaction with and their effect on the equipment to which they are
connected. This was done by performing various experimental tests on one o f the most
the full-scale flexible conductor and four series o f shake table tests on the large scale
models o f the equipment in their stand-alone or interconnected set-up. The design o f the
equivalent models o f the equipment was based on the stiffness and the natural
frequencies o f the actual equipment obtained from field tests. The input signals for the
shake table were synthetic earthquakes. Three records from earthquakes in California
were modified to fit the I E E E (the Institute o f Electrical and Electronics Engineers) and
B C Hydro prescribed design spectra. The specimens were instrumented throughout with
measure the loads generated at the connections o f the conductor to the equivalent
equipment during the tests. The test results are summarized and presented in various
forms. Various analyses including spectral analysis is performed on the test results and
comparisons are made between the results o f various tests and various analyses. This
investigation showed that the dynamic effect o f the flexible conductors should be
ii
T A B L E O F C O N T E N T S
ABSTRACT ii
LIST O F T A B L E S vii
LIST O F F I G U R E S x
ACKNOWLEDGEMENTS xiv
CHAPTER 1
INTRODUCTION 1
1.1 Background 1
CHAPTER2
LITERATURE REVIEW 7
2.1 Introduction 7
Equipment 9
CHAPTER 3
SIGNAL PROCESSING 16
3.1 Introduction 16
iii
3.4.1 Introduction 21
3.4.3 Sampling 23
CHAPTER 4
QUASI-STATIC TEST 39
4.1 Introduction 39
4.3 Instrumentation 42
CHAPTER 5
SHAKE T A B L E TESTS 47
5.1 Introduction 47
5.6.2 Instrumentation 61
iv
5.6.4.2 Damping Ratios 68
5.7.2 Instrumentation 79
CHAPTER 6
v
6.3 Frequency Content o f the Acceleration Response o f the Conductor 151
CHAPTER 7
C O M P A R I S O N O F R E S U L T S A N D DISCUSSION 165
CHAPTER 8
REFERENCES 176
APPENDIX A
APPENDIX B
DRAWINGS 183
APPENDIX C
vi
L I S T O F T A B L E S
Table 3.7.1 Some window functions and their characteristics (adopted from L a t h i ( 1 1 }
) 30
Table 5.7.3 Effect o f the conductor stiffness on resonance frequencies o f the towers... 87
Table 5.7.12 Peak dynamic forces in the connections o f the conductor 100
Table 5.7.13 Force Amplification Factors ( F A F ) 100
Table 5.9.11 Peak dynamic forces i n the connections o f the conductor 126
Table 6.2.1 PI for forces at the conductor connection to East Tower; set-up 1 142
Table 6.2.2 PI for forces at the conductor connection to West Tower; set-up 1 143
Table 6.2.3 PI for forces at the conductor connection to East Tower; set-up 3 147
Table 6.2.4 PI for forces at the conductor connection to West Tower; set-up 3 148
viii
Table 7.3.4 Displacement Response Ratio for West Tower 167
Table 7.4.1 Comparison o f forces obtained from test set-ups 1 and 3 169
Table 7.5.1 Comparison o f forces obtained from test set-ups 1 and 2 171
ix
L I S T O F F I G U R E S
configuration 2 3
Figure 2.4.1 Frequency bands and amplitudes for sine sweep tests performed by Dastous
and Pierre 12
Figure 2.4.3 Force response spectrum predicted by finite element analysis (after Der
Kiureghian et al) 15
from Test 29 35
Figure 4.2.2 Geometry and connection details o f the flexible conductor configuration 2 41
Figure 4.3.1 Connection o f the conductor to the load cell and to the frame o f the shake
table 42
Figure 5.5.3 Comparison o f the B C Hydro target spectrum for 2 % damping and the
Figure 5.5.4 Comparison of the I E E E 693-1997 target spectrum for 2 % damping and the
Figure 5.5.5 Comparison o f the I E E E 693-1997 target spectrum and the shake table
Figure 5.5.6. Input signal generated from Northridge record; I E E E target spectrum 57
Figure 5.5.7. Input signal generated from E l Centro record; I E E E target spectrum 58
Figure 5.5.8. Input signal generated from Landers record; I E E E target spectrum 59
West Tower 68
Figure 5.6.7 Frequency response o f vertical displacement at the tips o f cantilever beams
Figure 5.7.4 L V D T used for measuring the vertical displacement at the tip o f the
Figure 5.7.5 Cable extension transducers used for measuring towers displacement 81
Figure 5.7.15 Comparison o f Force Amplification Factors from different tests 101
Figure 5.7.18 F z vs. F y for East and West Towers - Test 29 106
Figure 5.9.6 Comparison o f Force Amplification Factors from different tests 127
Figure 6.2.1 Force response power spectra, set-up 1, sinusoidal loading 133
Figure 6.2.2 Force response power spectra, set-up 1, IEEE-Northridge at 90% 134
Figure 6.2.3 Force response power spectra, set-up 1, I E E E - E l Centro at 90% 135
Figure 6.2.6 Force response power spectra, set-up 3, IEEE-Northridge at 90% 138
Figure 6.2.7 Force response power spectra, set-up 3, I E E E - E l Centro at 90% 139
xii
Figure 6.2.9 PI for forces at conductor connection to East Tower; set-up 1 144
Figure 6.2.10 PI for forces at conductor connection to West Tower; set-up 1 145
Figure 6.2.11 PI for forces at conductor connection to East Tower; set up 3 149
Figure 6.2.12 PI for forces at conductor connection to West Tower; set-up 3 150
Figure 6.3.2 East North and West North accelerations o f the conductor - Test 29 154
Figure 6.4.5 M o d a l identification of West Tower; stand alone; sinusoidal loading 161
Figure 6.4.6 M o d a l identification of East Tower; stand alone; sinusoidal loading 162
xiii
ACKNOWLEDGMENTS
I would like to gratefully thank my supervisor Dr. Carlos Ventura for his guidance
throughout this project and sharing his valuable knowledge and experience with me,
which made the success o f this work possible. I would also like to thank my industry
supervisor, Dr. Gouri Bhuyan from Powertech Labs Inc., for his guidance which helped
provide perspective to this project. The funding for this project was provided by
Powertech Labs.
Dr. Hans Rainer coordinated the project between U B C and Powertech and played a major
role in the success o f this work. He provided me with valuable guidance and advice
throughout the project and I would like to sincerely acknowledge his contribution i n this
work. Dr. Endi Zhai also contributed in the first stage o f the project which is
acknowledged.
setting up the test specimen and running the tests. Messrs. Harald Schrempp, M a x Nazar,
Doug Hudniuk and John Wong, all U B C Structures Lab technicians, also contributed in
comments on the first draft o f this thesis and her valuable tips and suggestions. I would
also like to thank all my friends, especially those in earthquake engineering group at
Finally, I would like to deeply thank my parents who always supported me throughout
xiv
Chapter 1 Introduction
CHAPTER 1
INTRODUCTION
1.1 Background
Substations are critical links in power distribution networks. They have several functions,
such as changing the voltage and current level o f the power for transmission and
transformers, circuit breakers, surge arresters and disconnect switches which are
loading. In many past earthquakes, damage to the electrical substation equipment caused
power outages and blackouts. One o f the most common types o f failure in substations is
the failure o f the porcelain type elements o f the equipment such as brittle ceramic
insulators. H i g h voltage equipment e.g. 500 and 230 k V equipment are more susceptible
to earthquake damages.
one o f which is flexible conductor. Flexible conductors have enough slack, which allows
under seismic loading without being excessively stretched or compressed. However, past
suggested that the dynamic interaction between the conductor and the equipment, and
resulting amplified forces, can cause the equipment to overload and fail i n the event o f an
earthquake.
Various research projects, experimental and numerical, have been carried out to
investigate the behaviour o f these flexible conductors and their interaction with the
more comprehensive experimental studies were needed, due to the fact that flexible
conductors are highly nonlinear structures with structural and dynamic properties that are
1
Chapter 1 Introduction
W i t h regard to the safety and reliability o f substations throughout British Columbia and
in order to better understand the behaviour o f flexible conductors and thus improve the
experimental study jointly with Powertech Labs. Inc. on the seismic behaviour o f six
1.1.1 shows the six configurations o f the interconnected equipment and the connecting
flexible conductors that are considered by B C Hydro. Configuration 2 is one o f the most
critical configurations and is the subject o f this thesis. It consists o f a 500 k V circuit
aluminum stranded conductor. T w o spacers are used between the two conductors. Figures
1.1.1 and 1.1.2 show the interconnected equipment and the flexible conductor
configuration 2.
Natural
Conductor Approx.
Configuration Frequencies
Equipment Equipment Type Span
No. respectively
(kcmil) (m)
(Hz)
500 k V 500 k V
1 Shunt Surge 1.60, 10.66 2x2305.5 6.4
Reactor Arrester
500 k V 500 k V
2 Circuit Current 0.94, 3.4 2x2303.5 4.0
Breaker Transformer
500 k V
500 k V
3 Bus 2.86, 12.6 2x2303.5 3.0
Pothead
Support
230 k V
230 k V
4 Surge 11.10,3.16 1x1272 3.0
Transformer
Arrester
230 k V 230 k V
5 Circuit Current 4.28, 1.56 2x1272 3.7
Breaker Transformer
230 k V 230 k V
6 Circuit Current 5.57, 1.55 2x2303.5 3.5
Breaker Transformer
2
Chapter 1 Introduction
Figures 1.1.3 and 1.1.4 show examples o f other configurations o f the interconnected
Table 1.1.1.
5CBI5. SCB2«
DUNSMUIR 5 0 0 k V ( 5 2 6 - E 0 5 - D I 6 9 )
2 - 2 3 0 3 . 5 k c m l I ASC 3 3 0 mm c / c
configuration 2
3
Chapter 1 Introduction
4
Chapter 1 Introduction
5
Chapter 1 Introduction
equipment and the forces that it induces on the equipment in an anticipated earthquake.
• To find out whether the flexible conductor changes the response of the
• To find out the amount of forces generated at the connections of the conductor to
the equipment under seismic loading and to estimate the amplification of forces
studying other configurations. The experimental tests included cyclic quasi-static tests
and various shake-table tests on the full scale flexible conductor and large-scale models
of the equipment. The analytical study for this project included spectral analysis on the
6
Chapter 2 Literature Review
CHAPTER 2
LITERATURE REVIEW
2.1 Introduction
In this chapter, flexible conductors, their geometry and their design concepts are
described. A l s o a review on the experiences from past earthquakes and damage observed
in the equipment is presented. Finally, a summary o f the research carried out on the
presented.
Flexible conductors are electrical conductors used i n substations as paths for high voltage
electricity between different equipment. Figures 1.1.2 to 1.1.4 show some configurations
In general, flexible conductors can be bare cables, braids or expansion joints, and can be
connected to the equipment by bolted or welded connections. Two most common sizes o f
flexible conductors used i n substations in British Columbia are 1272 k c m i l and 2303.5
kcmil aluminum stranded conductors with minimum bending radii o f 300 m m and 600
mm, respectively/ !)
The geometry o f flexible conductors is configured such that they accommodate the
relative displacement o f the interconnected equipment under seismic loading. They also
meet the electrical design requirements. There are three basic geometry configurations
which are most suitable for the commonly used 1272 and 2303.5 k c m i l aluminum
stranded conductors. These geometries provide the conductors with the necessary slack
that enables the equipment to have differential movement without excessive stretch or
7
Chapter 2 Literature Review
compression o f the conductors. The three geometry configurations shown i n Figure 2.2.1
are as f o l l o w s : (1)
• Configuration 1 using two 90° connectors or special brackets, for short length
• Configuration 2 using one 90° connector or special bracket and one straight
Flexible conductors may consist o f one stranded cable (single conductor) or two stranded
cables (twin conductor). Spacers between the two cables are required for twin 1272 kcmil
aluminum stranded conductors with length exceeding 1500 m m and for twin 2303.5
kcmil aluminum stranded conductors with length exceeding 2000 mra. (1)
The required conductor length is determined based on the deflections o f the equipment at
moderate seismic qualification level. The deflections o f individual equipment items are
different from each other depending on the type o f the equipment. The typical values o f
8
Chapter 2 Literature Review
deflections are obtained from the results o f past seismic studies and used for design
purposes. The final conductor length however, is determined on site during installation to
account for the diverse field conditions and assuring that the actual installed conductor
earthquakes even i n the equipment that were designed to withstand the earthquake
caused damage to 500, 200 and 115 K V substations. Typical damage included broken
The study o f the damage caused by past earthquakes to the equipment i n substations led
equipment.
study included both experimental and numerical analysis. Both types o f analysis revealed
that the additional pulling force caused by the stretch o f the conductor on the porcelain
9
Chapter 2 Literature Review
column o f the equipment was the reason for failure o f the equipment. The experimental
investigation consisted o f shake table testing o f a full scale model o f a selected pair o f
equipment interconnected with a flexible conductor. The numerical study included finite
element analysis which was found to be too complicated to be used for studying the
dynamic behaviour o f the flexible conductor. The complexity was due to the equipment-
conductor coupling system and the nonlinearity caused by the conductor and its large
deflections. The final recommendation was to avoid the tension forces in the conductors
by configuring them such that they are not stretched during an earthquake. A design
Additional forces due to the stretch o f the conductor, however, had not always been the
reason for the failure o f the interconnected equipment. Damage had also been observed in
the equipment interconnected by flexible conductors with enough slack, which provided
enough room for the displacements o f the equipment without being excessively stretched.
Further research has been carried out recently in order to explain the reason for failure
through understanding the dynamic behaviour o f the conductor and its interaction with
the equipment.
on dynamic behaviour o f flexible conductors. Their main goal was to find whether typical
flexible conductors have natural frequencies that could be excited during an earthquake
and to provide insight into the dynamic behaviour o f those conductors. To achieve this,
they ran cyclic static tests as well as sine sweep tests with a frequency band from 0.5 to 5
H z at expected ground motion amplitudes. Figure 2.4.1 shows the frequency bands and
the corresponding amplitudes for sine sweep tests. The excitation was either only at one
end o f the conductor with the other end fixed or at both ends oscillating out o f phase with
respect to each other. The test specimens, comprising o f either one cable or two cables
with spacers, were two o f the most commonly used conductors in Hydro-Quebec
substations. Properties o f theses conductors are shown in Table 2.4.1. The forces at the
10
Chapter 2 Literature Review
The results showed that the flexible conductors are nonlinear dynamic systems which
are variable due to the nonlinear behaviour o f the conductor and vary according to its
configuration and the amplitude o f excitation and they are likely in the range o f 0.5 to 5
H z . The magnitude o f forces measured at the ends o f the conductors during their
resonance, indicated that it is not sufficient to design these conductors on a static basis.
Figure 2.4.2 shows three force response spectra obtained from a test i n which resonance
of the conductor did not happen, a test interrupted because o f very large forces generated
at the ends o f the conductor due to resonance, and an uninterrupted test showing the
Dastous and Pierre concluded that the resonance o f the conductors must be avoided
during an earthquake and they mentioned the importance o f designing them such that the
range o f the natural frequencies at which they can be excited would be different than the
natural frequencies o f the equipment interconnected by them. They also concluded that a
large sag/span ratio would improve the dynamic stability o f the conductor and reduces
the risk o f its resonance and suggested that i f a large sag/span ratio violates the required
alternatives. They recommended that the severity o f forces be verified for real
earthquakes since the sinusoidal excitation could not completely simulate the effect o f an
earthquake.
11
Chapter 2 Literature Review
Figure 2.4.1 Frequency bands and amplitudes for sine sweep tests performed by Dastous
and Pierre
12
Chapter 2 Literature Review
2000
I I I I
i i i i i i i i i i i i [ i i i i | i i i i i r t t i• t|-it i i i l i • i i |
0.5O 1.00 1.50 2.00 2.90 3.00 1 9 04.00 4.50 3.00
F r e q u e n c y (Hz)
3000.ee-
connection: 2 conductor* 17^96-KICM
•pan: 5.10 m, sag:0.B3 m. eag/Bpan: 0.16,
Z aeee.ee4 out-of-phase excitation, .amplitude: 0.04 m
£ ieee.ee:
F r e q u e n c y (Hz)
a. be 3.08
- it
t connection: 1 conductor 1796-MCM
•pan: 3.60 m. »ag:0.30 m. sag/span: 0.08
excitation at one end only, amplitude: 0.02 m
1111111111 am pltttd . U ' IS*
i i.ee t.co 2.98 2 3.ee' .69 6.00
1
m ii
F r e q u e n c3.00
y (Hz)
13
Chapter 2 Literature Review
In another experimental research project carried out at the University o f California at San
Diego by Filiatrault et al ( 7 )
in order to evaluate the behaviour o f substation equipment
interconnected by both flexible and rigid conductors, quasi-static cyclic tests were
performed on two different types o f flexible conductors. The test specimens were 2300
M C M single conductor and 1113 M C M twin conductors. The results showed increasing
tensile forces due to the stretch o f the conductor and very l o w compressive forces. These
numerical investigation on the interaction between the flexible conductors and the
electrical equipment under dynamic loading by performing finite element analyses. The
equipment items were modeled as single degree o f freedom systems that represented a
linear system with distributed mass, damping and stiffness properties. The presence o f
response amplification was determined using a "response ratio", defined as the ratio o f
the displacement o f the top o f the interconnected equipment relative to its base, to the
relative displacement o f the stand-alone equipment. The model was able to achieve a
good qualitative agreement with the existing experimental results; however, quantitative
agreement was not attained due to the complexity o f defining the exact properties o f the
conductor and the lack o f complete information about the experimental tests conditions.
The results showed dynamic forces in the conductor ends at magnitudes much higher
respect to each other, and s and L are the length and the span o f the conductor
respectively. It was shown for a pair o f interconnected equipment with 1 and 5 H z natural
finding. Figure 2.4.3 shows the finite element prediction for force response spectrum o f
the conductor subjected to an out o f phase sinusoidal excitation at its supports. These
14
Chapter 2 Literature Review
Frequency (Hz)
Figure 2.4.3 Force response spectrum predicted by finite element analysis (after Der
Kiureghian et al)
15
Chapter 3 An Overview on.
CHAPTER 3
3.1 Introduction
This chapter first presents a general overview on the concepts o f presenting signals and
time domain sequences i n frequency domain, and then explains the fundamentals of
Fourier analysis, linear systems, digital signal processing, spectral analysis and system
identification method used as computational and analytical tools throughout this study.
loading. Signals contain lots o f information about the physical quantity they are
representing. However, portions o f this information are hidden when the signal is
presented i n the time domain, but would be revealed i f the signal was presented i n the
frequency domain. Signals can be presented in the frequency domain i n terms o f the
linear combination o f sine and cosine functions. Time domain and frequency domain
descriptions o f a signal compliment each other and taken together they provide a better
In order to better understand the concept o f frequency content, a function /(t) satisfying
2. f is bounded
16
Chapter 3 An Overview on.
3. A t any particular period, the function may have at most a finite number o f
discontinuities and a finite number o f maxima and minima.
oo
f(t) = Ao + An cos najot+Bn cos ncoot (3.2.1)
where
2TT ,„ »
coo = — (3.2.2)
To
Ao = — f" f(t)dt 2
(3.2.3)
J i-To/2
0
B = — ^^/(t)
n sin ncootdt « = 1,2,3,... (3.2.5)
To
Equation 3.2.1 can also be written in the compact trigonometric form o f Fourier series as
shown i n Equation 3.2.6:
CO
in which
Co = Ao (3.2.7)
17
Chapter 3 An Overview on.
^ t a i T ' t — ) (3.2.9)
An
The compact trigonometric Fourier series in Equation 3.2.6 shows that a periodic signal
f(t) can be expressed as a sum o f sinusoids with frequencies 0 , coo, 2coo, ncoo,
which have amplitudes o f C o , C i , Ci, d , . . . and phases o f 0, 9\, 9i, On, ...
respectively. ^ A plot o f C» vs. co is called the amplitude spectrum and a plot o f 6n vs.
(1
co is called the phase spectrum. These two plots together form the frequency spectra o f
f(f). These spectra show the frequency composition o f fit) and are in fact the
The resulting Fourier series is a periodic function o f period To, therefore for computing
the series coefficients we may perform the integration over any interval o f To and not
A more convenient and more compact form o f the Fourier series is the exponential form
/(0=J>e" (3.2.10)
where
(3.2.11)
It can be observed that coefficients D„ are complex numbers. Their relation with
18
Chapter 3 An Overview on...
D„ = — C„e
2
D - n= — Ci£ (3.2.12)
2
Next section describes how a function can be represented i n the frequency domain i f it is
f(t) is not a periodic function then a new periodic signal fn(t) is constructed by
repeating the signal /(t) at interval o f To seconds. The period To is made long enough
to avoid overlaps between the repeating pulses. The periodic signal fr«(t) can then be
represented by an exponential Fourier series. If To -> oo, the pulses i n the periodic signal
Thus the Fourier series representing fn(t) w i l l also represent /(t) in the limit To -> oo .
C
O (3.2.14)
(3.2.15)
19
Chapter 3 An Overview on..
If F(co) is defined as
then
Dn = — F{ncoo) (3.2.17)
To
The integral on the right hand side o f the Equation 3.2.18 is called the Fourier integral
and represents the non-periodic signal f(t) just as Fourier series represent periodic
signals. The Fourier integral is basically a Fourier series in the limit with fundamental
frequency Aco -> 0 and thus the function F(co) given by Equation 3.2.16 acts as a
frequency spectral function. F(co) given by Equation 3.2.16 is called the direct Fourier
transform o f f{t), and f(t) given by Equation 3.2.18 is called the inverse Fourier
transform o f F(co) and both o f them together is called the Fourier transform p a i r . (11J
Note that these integrals do not always exist. Discussion on the existence o f these
integrals is beyond the scope o f this overview and it is assumed here that they both exist.
20
Chapter 3 An Overview on.
3.4.1 Introduction
A s discussed before, Fourier series is an operation that takes a function /(t) and returns
a sequence o f coefficients £>», and the Fourier transform is a operation that maps a
and must be rather uncomplicated which does not usually happen in the real world.
However, there is a numerical way of doing the above-mentioned operations with digital
an operation that takes the signal f(t) as a sequence o f numbers and maps it to F(co) as
can be processed by D F T .
co
M0= T- D e
(3.2.14)
needs to be selected because it cannot be just an infinite number. Then fr»(t) is sampled
at a sampling interval o f At. Hence there are No = Tol At number of samples in one
period To:
(3.2.19)
k=0
21
Chapter 3 An Overview on.
Substitution o f
No = — , o)o = — (3.2.20)
Ar To
yields
I JVo-l
Dn = l i m — Y f{k&t)e- 2MINa
L\t (3.2.21)
Practically, it is impossible to make A f - > 0 for computing the right hand side o f
Equation 3.2.21 and Ar can be small but not zero. So the limit on At in Equation 3.2.21
cannot be too small as very small Ar w i l l cause the data to increase without limit.
Therefore
1 No-]
Dn = — Yf{ ^)e- " At
k 27ti klNn
(3.2.22)
M fo
Equation 3.2.22 is the discrete form o f Dn which can be computed by a digital computer.
The discrete Fourier transform pair can be found in a similar way. Equations 3.2.22 and
Afo-l
1 No-]
fi = — YF,<> * ' " 2 irk N
(3.2.23)
Nof 0
where
fk = Atf(kAt) (3.2.24)
22
Chapter 3 An Overview on...
Performing the D F T requires a very large amount o f computations. The Fast Fourier
computers. There are however, more recent algorithms that are even faster than F F T .
3.4.3 Sampling
fit) requires sample values o f fit) because digital computers can work only with
discrete data. Moreover, a computer can compute Fico) only at some discrete values o f
co which are the samples o f Fico). It is very important to understand the relation
Figure 3.4.1 shows signal fit) which is limited to r seconds along with its Fourier
(3.2.25)
23
Chapter 3 An Overview on.
m—
f-
The periodic signal fr»(t), which is formed by repeating the signal fit) every To
seconds (TO>T ) is also shown in Figure 3.4.1. D» was defined i n Equation 3.2.17 as
F(ncoo) (3.2.17)
Equation 3.2.17 shows that the coefficients o f the Fourier series for fr«(t) are I/To
times the sample values o f the spectrum F(co) taken at intervals coo which means that
these samples are separated by the fundamental frequency coo. In other words, i f the
theorem). 00
fo = \/To (3.2.26)
When we sample fro(t) and compute its discrete Fourier transform, one o f the most
important consequences o f sampling is that the Fourier spectrum repeats itself after every
24
Chapter 3 An Overview on..
F{co) are spaced at I/To and since No = To/At, it can be concluded that the Fourier
fi = \IAt (3.2.27)
One o f the results o f the periodicity o f F{co) is the spectral overlap due to repeating
cycles as depicted i n Figure 3.4.2. The overlap is at No 12 or fill or coo 12. This effect
This overlap causes error i n the reconstructed signal from the sampled signal because i f
the signal is to be reconstructed from the sampled signal, then its Fourier spectrum should
be recovered from its sampled spectrum. For a signal band-limited to B H z , there would
aliasing o f a signal is called Nyquist rate or Nyquist frequency for that signal. In reality
however, signals are time-limited and they are not band-limited and aliasing would be
signal i n H z and then select suitable values for No, At, To. Once B is picked, the
fi>2B (3.2.28)
25
Chapter 3 An Overview on.
At< — (3.2.29)
IB
F(co), To can be selected knowing that fo = 1 / To, and eventually the number o f
Another way o f decreasing the effect o f aliasing is using the anti-aliasing filter before the
signal is sampled. Anti-aliasing filter eliminates the frequency components o f the signal
For data acquisition i n this study, a sampling rate o f 200 H z was chosen. So the Nyquist
frequency o f the data was 100 H z . Since the maximum expected frequency response o f
the test specimen was no more than 30 H z , a Nyquist frequency o f 100 H z safely
prevented the aliasing error and thus the sampling rate was considered to be adequate.
which has an input function f(t), and output function y{t) and a cause-effect
categories is the linearity o f the system, i.e. the system is linear or nonlinear. A system is
linear i f its output is proportional to its input and i f its response to several causes can be
identified by summing up its response to each cause separately. Most practical systems
Transfer function o f a system is a function that describes how the system operates on the
26
Chapter 3 An Overview on.
time-invariant system (a linear system whose parameters do not change with time) the
output signal
H(a>) = (3.2.30)
input signal
2m on
e{t) = e (3.2.31)
The function e(t) is the characteristic function (or eigenfunction) o f the system because
it is the only function to which the response o f the system has the same exponential form
It can furthermore be observed that i f / ( r ) is the input signal with the Fourier transform
F(co)
Comparing equation (3.2.32) with the Fourier transform formula shows that Y(co), the
27
Chapter 3 An Overview on.
Equation (3.2.33) shows that i f the transfer function o f a system is known, then the
Fourier transform o f the output signal can be easily obtained by using the simple equation
of (3.2.33) and then the output signal itself can be found by operating inverse Fourier
Next section includes an example application o f transfer functions which is used in this
study.
A l l the numerical integrations in this study were performed i n the frequency domain. To
do the integration in frequency domain first the transfer function o f the integral operation
must be determined.
calculated as
XO = \e 2!dw,
dt
IrncH i
X0 = — =— e(t) (3.2.34)
2m co 2m co
Thus to integrate a signal, first the Fourier transform o f the signal is evaluated. Then the
Fourier transform o f the signal is multiplied by H(co) (Equation 3.2.35) to yield the
28
Chapter 3 An Overview on.
Fourier transform o f the integrated signal. Finally, by using the inverse Fourier transform,
the integrated signal in time domain can be obtained from its Fourier transform.
In order to check the accuracy o f the results thus obtained, the displacement o f the shake
table in one o f the shake table tests performed for this study (Chapter 5) was calculated
by integrating its acceleration response and was compared to its displacement directly
measured by displacement transducers. Figure 3.6.1 shows both calculated and measured
displacements. It shows that they are almost identical. Therefore the numerical
Mi Ul hiIAA
1/v
A,
J n
vV
V
1
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
Calculated
Measured
shortening the length o f the signal such as for choosing a finite length o f an infinite
signal, or for breaking the data sequence into segments o f equal length while estimating
designing F I R f i l t e r s . (16)
29
Chapter 3 An Overview on.
Data truncation is carried out by multiplying the signal by a window function o f smaller
width. Table 3.8.1 shows some window functions and their characteristics. To describe
the definitions o f mainlobe and sidelobe o f window functions shown in Table 3.7.1,
mainlobe and sidelobes o f a rectangular window are shown in Figure 3.7.1. Rolloff rate is
the decay rate o f sidelobes which is also shown in Figure 3.7.1. In Figure 3.7.1, wi((t)is
Table 3.7.1 Some window functions and their characteristics (adopted from L a t h i ( 1 0
)
. Rolloff Peak
Mainlobe . Rate . Sidelobe
Window w(t) Width dB/oct LeveHn dB
\\.2r.
,6 Raise): -6 -59.9 (o: = 8,168)
\<a<
-
'<>(«} 10 T
%(G>)
1 w(<)
Rolloff Rate
7 -0.217T «B
(b)
Sidelobes
RoloffRatc -20dB/decade
indB
Figure 3.7.1 Rectangular window and its characteristics (adopted form Lathi
30
Chapter 3 An Overview on...
If /(t) is a signal to be windowed and w(t) is the window function, and i f F(co) and
W(co) are the Fourier transform o f /(t) and w(t) respectively, then the windowed signal
is
The Fourier transform o f fw{t) would be the convolution o f / ( r ) and w(t) [give ref.
F (t) = —
w F(co)*W(co) (3.2.27)
Truncation o f a signal causes spectral spreading and leakage i n the truncated signal.
According to the width property o f convolution, the width o f F*(r) is equal to the sum o f
the widths o f F(a>) and W(co). Thus the truncated signal has larger bandwidth by the
amount o f the bandwidth o f w(t). This is called spectral spreading. Since the signal
bandwidth is inversely proportional to its duration, a wider window would cause less
spectral spreading. In other words, wider window means more samples o f data and better
On the other hand, W{co) is not bandlimited and its amplitude approaches zero as co
increases. So even i f F{co) is bandlimited, Fw(t) is not bandlimited and its amplitude
approaches zero at the same rate as that o f W(co). This is the other effect o f data
31
Chapter 3 An Overview on.
Choosing a suitable window would minimize the side effects o f truncation. A wide
window would reduce the spectral spread and a smooth window would improve the
leakage o f the truncated signal. The two effects cannot always both be improved. For
example, for a given width, the rectangular window has the smallest spectral spread but it
has the worst leakage behavior with high sidelobe magnitudes. Hamming window has the
smallest sidelobe magnitude for a given mainlobe width. Harming is a window widely
used for spectral analysis because it has faster rolloff rate compared to the other windows
(-18 dB/oct compared to -6db/oct for hamming and rectangular). T o achieve the best
results, the application o f the truncated data should be taken into account while selecting
Digital filters are systems that are intended to change the frequency content o f the input
signal by keeping the desired frequencies and suppressing the rest. Filters also change the
phase content o f the input signal. In other words, the primary purpose o f digital filtering
is enhancing the output signal by changing the spectral content o f the input s i g n a l . (16)
There are four types o f filters based on their frequency response characteristics:
• Low-pass filters pass low frequency components o f the input signal to the output
• High-pass filters allow the high frequency components to pass to the output signal
• Band-stop filters reject a range o f frequencies and pass high-frequency and low-
frequency components.
32
Chapter 3 An Overview on.
• Infinite impulse response (IIR) filters, also called recursive filters, are systems o f
which the output is determined iteratively or recursively from its past values in
• Finite impulse response (FIR) filters, also called non-recursive filters, are
systems o f which the output is computed from the present and past values o f the
input signal.
Compared to F I R filters, IIR filters can have sharp frequency cutoff characteristics with
lower order structure which means they are faster and require less memory. However, IIR
filters have nonlinear phase characteristics whereas F I R filters can be designed to have
exactly linear phase response which is desirable i n many applications where a linear
In this study, in order to clean up the test results from noise, a bandpass F I R filter was
designed and employed. To design this filter, a Hanning window was chosen because o f
its faster rolloff rate compared to other alternative windows (Section 3.7). To reduce the
spectral spread o f the filter, a very large width was chosen for the window. This filter was
mainly used as a lowpass filter by setting the highpass frequency equal to zero. The filter
lowpass frequency was set to 25 H z for measured accelerations and displacements, and
20 H z for measured forces (Chapter 5). Figure 3.8.1 shows both amplitude gain and
response obtained from one o f the shake table tests (Chapter 5). The peak values are also
indicated.
33
Chapter 3 An Overview on.
1.5
1.25
rt
o
3 0.75
a.
E
< 0.5
0.25
0 5 10 15 20 25 30 35 40 45 50 0 5 10 15 20 25 30 35 40 45 50
Frequency (Hz) Frequency (Hz)
20
9 -20
\
=3
O "40
o -60
"80 w
'1 1 yy
-100
10 15 20 25 30 35 40 45 50 0 5 10 15 20 25 30 35 40 45 50
Frequency (Hz) Frequency (Hz)
34
Chapter 3 An Overview on.
Not Filtered
2000
"2000
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max = 1213 N min = -1152N
Filtered
2000
1000
-1000
-2000
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max =1136 N min = -1032N
from Test 29
The power spectra i n this study were estimated by using averaged periodograms. (16)
Estimation o f power spectra using averaged periodograms gives better results for long
time series o f random data compared to simply using the Fourier transform o f the whole
data sequence.
the Discrete Fourier Transform ( D F T ) o f the signal divided by the number o f samples,
35
Chapter 3 An Overview on.
No. The periodogram is a measure o f power density o f the signal. Recalling from
Fr=Yif e-
k
2M/N
" (6.2.1)
k=0
in which
fk = Atf(kAt) (6.2.2)
p = -L|Fr|
r
2
r = 0,l,-,No-\ (6.2.3)
No
In the average periodogram method, the signal is broken up into segments with equal
length which could be overlapped. Then a data window is applied to each segment to
smoothen or shape the periodogram o f each segment. Finally the periodogram o f each
windowed segment is calculated and the power spectrum is obtained by averaging the
In this study, to calculate the power spectra, signals were broken up into four segments. A
Hanning window was applied to each segment (Section 3.7) and the overlap o f the
segments was chosen at 25 % . Note that it was observed that there was no significant
difference between the results obtained with 25% overlap and 50% overlap, therefore, to
System identification i n this study involves finding the natural frequencies o f the system
and the mode shapes associated with those natural frequencies. For a system subjected to
and input signal, this task is achieved by using the power spectra o f the measured outputs
36
Chapter 3 An Overview on.
of the system, and the cross spectrum and coherence between the selected output
channels.
The resonance frequencies o f the system are the frequencies at which the power spectrum
of the system response peaks. For lightly damped systems, these frequencies could be
The cross spectrum is used to investigate the similarities o f two selected output signals,
and determine whether they follow each other or not. In fact, cross spectrum o f two
signals represents their cross correlation. The phase o f the cross spectrum represents the
phase angle between the two signals and is helpful in identifying the mode shapes of
structures.
The coherence function is used to determine the linear dependence o f two selected output
signals. The coherence function is defined as the ratio o f the squared magnitude o f the
cross spectrum to the power spectra o f both signals. The coherence function has a value
between 0 and 1. While a value o f 1 for a certain frequency band shows a very good
linear dependence between the two signals and indicates that they both are free o f noise
in that frequency band, a value o f 0 indicates the existence o f significant level o f noise in
the measurements. A l o w value o f coherence could indicate either the existence o f noise
The problem with identifying the natural frequencies o f a system is that any peak in the
power spectrum does not necessarily represent a normal mode o f vibration for that
system. A peak could also represent a coupled mode or a source o f noise. That means
significant level o f noise i n the output measurements or coupling between the normal
frequencies. Cross spectrum phase and coherence function are useful i n distinguishing
normal modes from coupled modes and noise. A s mentioned earlier, the existence o f
noise in a measured signal causes the coherence function between that signal and any
37
Chapter 3 An Overview on..
other measured signal to be less than one. O n the other hand, i f the system is lightly
damped, the phase angle between two measured signals related to a normal mode w i l l be
either 0° or 180°. So i f the mode is a coupled mode, then the phase angle between the two
output signals w i l l be other than 0° or 180°. In general, a peak i n two power spectra with
coherence close to 1 and phase angle close to either 0 or 180°, indicates with confidence
38
Chapter 4 Quasi-Static Test
CHAPTER 4
QUASI-STATIC TEST
4.1 Introduction
A quasi-static test was performed on the full scale flexible conductor configuration 2
(Table 1.1.1). In this test, one end o f the conductor was subjected to cyclic displacements
with increasing amplitudes, and the other end o f the conductor was fixed. The purpose o f
this test was to study the behaviour o f the conductor subjected to differential
displacements at its ends caused by the deflections o f the interconnected equipment. This
test investigated the behaviour o f the conductor statically and did not consider its
dynamic behaviour.
Figure 4.2.1 shows the test set-up for the quasi-static test which was performed in the
Structural Laboratory at the University of British Columbia in January 2001. The test
specimen was the high voltage flexible conductor configuration 2 as described in chapter
1. The lower end o f the conductor, the end which would be connected to the circuit
breaker i n the substation, was connected to a steel pedestal. The other end, the end which
would be connected to the current transformer in the substation, was attached to the
frame o f a linear shake table. The geometry o f the conductor was according to the
drawing shown i n Figure 4.2.2. The conductor was connected to its supports by using
aluminum u-shape plates called "angle adapters" and "terminal connectors" as shown in
Figure 4.2.2. The displacements were applied to the conductor by moving the shake table
39
Chapter 4 Quasi-Static Test
40
Chapter 4 Quasi-Static Test
152
/
240 / 2-6-X %•
12 HOLES 16 D1A- ,54,44
-
4
"'<rT
=0- 6-
" 98 .44., V
1 ' ='—= ^
!?>!
of
DETAIL 1
Figure 4.2.2 Geometry and connection details o f the flexible conductor configuration 2
41
Chapter 4 Quasi-Static Test
4.3 Instrumentation
A 10 kip load cell was used to measure the forces at the moving end o f the conductor
during the test. It was inserted between the terminal connector o f the conductor and the
steel frame o f the shake table (Figure 4.3.1). This load cell measured the horizontal forces
at the connection o f the conductor to the shake table. The applied horizontal
displacements at the same end o f the conductor were obtained directly from the
transducers o f the shake table actuator. Details on the specifications o f the transducers are
included in Appendix A .
Figure 4.3.1 Connection o f the conductor to the load cell and to the frame o f the shake
table
A loading protocol was specifically developed to perform this quasi-static test. The
Hz) 500 K V equipment subjected to 0.3g peak ground acceleration, is 200-600 mm. This
42
Chapter 4 Quasi-Static Test
conductor and a maximum o f 600 m m when pushing it. It was decided not to pull the
conductor for more than 200 m m as this could generate excessive tension in the
conductor and consequently cause damage to it. Damage to the conductor was prevented
since the same conductor was to be used for the shake table tests.
Figure 4.4.1 shows the test protocol which consists o f displacement cycles with
increasing amplitudes. There are three cycles at each level o f displacement amplitude.
700 -i i i i 1 1 1 1 1 1 1 1
-300 -I 1 1 1 1 1 1 1 1 1 1 1
0 3 6 9 12 15 18 21 24 27 30 33
No. of Cycle
Figure 4.5.1 shows the results o f various cycles at different displacement amplitudes and
Figure 4.5.2 shows the results for cycles 31 to 33 which correspond to maximum
displacements. They show that a very small force, less than 180 N (40 lb), is required to
push the conductor, even for a displacement as large as 600 mm. The pulling forces are
higher and increase quickly as the conductor straightens up and goes into tension. U s i n g
a 10 kip load cell for measuring such low forces resulted low precision measurements;
43
Chapter 4 Quasi-Static Test
however, the attained accuracy was acceptable for the purpose o f this test. It was
observed that:
caused very low forces in the conductor and at the conductor attachment points.
The conductor has almost no resistance against pushing forces. The very low
displacement range used for this test is negligible since it was observed that the
load-displacement curves for all the cycles lie reasonably on each other (Figure
equivalent stiffness for the conductor. This equivalent stiffness could be helpful
for numerical studies on the conductor. The result is shown in Figure 4.5.3. The
44
Chapter 4 Quasi-Static Test
Displacement (mm)
Cucles 9 to 12
Cycles 12 to 15
— Cycles 21 to 24
Cycles 30 to 33
-300 700
Displacement (mm)
45
Chapter 4 Quasi-Static Test
46
Chapter 5 Shake Table Tests
CHAPTER 5
SHAKE TABLE TESTS
5.1 Introduction
This chapter describes the shake table tests performed on the flexible conductor
configuration 2. In order to study the interaction o f the conductor and the interconnected
equipment, a full scale model o f the conductor and a large scale equivalent model (East
Tower and West Towers as described i n Section 5.4) o f the equipment were used. The
two equivalent equipment items were mounted on two sets o f cantilever beams which
were specially designed and used for extending the shake table to accommodate the large
span o f the specimen. In order to study the effect o f the conductor on the behaviour o f the
equipment, first the stand-alone towers were tested. Then three different set-ups o f the
towers with the connecting conductor were tested. Uni-axial and bidirectional synthetic
earthquakes were used as input signals o f the shake table. These synthetic records were
obtained by modifying three records from past earthquakes to fit B C Hydro and I E E E
693-1997 design spectra at 2 % damping, with a target peak ground acceleration o f 0.5g.
The result was six different synthetic ground motion records that were used at different
payload capacity o f 156 k N (35000 lbs). The motions are produced by hydraulic
actuators, which have a maximum displacement o f ±7.6 c m (±3 inches). The actuator
used to produce the horizontal longitudinal (South-West) motions can generate up to 156
k N (35000 lbs) o f force. The remaining four actuators are used to produce either vertical
or horizontal motions and can generate each up to 67 k N (15000 lbs) o f force. The
actuators are mounted on an isolated concrete pit foundation. Clearance above the table is
47
Chapter 5 Shake Table Tests
Vibration Control Software. This software performs a closed-loop control o f the shakers
and is capable o f reproducing recorded earthquake motions with high accuracy. Since test
set-up parameters are stored in digital form, accurate replication o f shake table motions
tested on the shake table. A l l channels are conditioned by variable gain buffers and
variable cut-off filters to provide optimal control over signal levels and noise reduction.
Since the dimensions o f the shake table were not large enough to accommodate the
specimen, it was necessary to design two sets o f cantilever beams that project out o f the
shake table and provide bases for mounting the towers. These beams were designed to be
versatile such that they would be useful for testing all the flexible conductor
The projected lengths o f the cantilever beams were chosen based on the span o f the
conductor and the mass o f the towers associated with each set o f beams so that both sets
o f beams had the same fundamental natural frequency. Enough stiffness was provided for
these beams so that each had a first mode natural frequency o f more than 50 H z . This was
to ensure that they would not pick up significant vibration during the tests and thus they
would not introduce significant noise into the results o f the tests.
It was impossible to design rigid beams with no deflection during the tests. However, the
effect o f their deflections was mainly changing the overall stiffness o f the equivalent
equipment as discussed in section 5.6.4.4. The details o f these beams are included in
Appendix B .
48
Chapter 5 Shake Table Tests
Two structures were designed to represent the equipment interconnected by the flexible
conductor. The goal i n designing these two structures was to achieve the stiffness and the
first mode natural frequency o f the equipment. However, due to limitations such as height
limitation above the shake table, it was impossible to achieve all o f the above mentioned
equivalent equipment and then interpret the results based on these structures.
four 70 m m diameter solid round bars and stacks o f steel plates attached to the top o f the
bars with fixed connections. The bars and the stack o f plates provided the stiffness and
required mass o f the equivalent equipment. It was possible to add or remove plates in
order to increase or decrease the mass o f the towers for decreasing or increasing their
natural frequencies. The heights o f these towers were selected to provide the right
The 500 k V Current Transformer (5CT) is the shorter equipment which was represented
by the shorter tower (West Tower). The West Tower was located on the west side o f the
shake table. The 500 k V Circuit Breaker ( 5 C B ) is the taller equipment which was
represented by the taller tower (East Tower). The East Tower was located on the east side
Reasonable agreement was achieved between the stiffness and the first mode natural
frequency o f the current transformer with those o f the West Tower. It was not feasible
however, to achieve a good agreement between the fundamental frequency o f the circuit
breaker and the fundamental frequency o f the East Tower. That was because the
design the East Tower for a fundamental frequency equal to the second mode natural
49
Chapter 5 Shake Table Tests
A s mentioned earlier, the deflection and rotation o f the extension cantilever beams
affected the overall stiffness o f the towers and consequently their natural frequencies.
The actual natural frequencies and equivalent stiffness o f the towers are obtained from
In order to verify the equipment design criteria, it was decided to use earthquakes that
had response spectra matching B C Hydro's and I E E E ' s design spectra. The target spectra
were at moderate seismic performance level with 2 % damping. These spectra are
In order to study the possibility o f different behaviour o f the conductor under different
seismic loadings with different time domain characteristics, three different input signals
with different time domain characteristics were generated for each target spectrum. To
50
Chapter 5 Shake Table Tests
do so, three records from the 1984 Northridge, the 1979 E l Centro and the 1992 Landers
earthquakes i n California were used to generate synthetic ground motions (Figure 5.5.2)
earthquake records. This program uses an initial earthquake record and generates another
record that matches a prescribed target spectrum. In order to match the computed
spectrum with the target spectrum, the program raises and suppresses the computed
was run enough times until very good agreement was achieved between the target and
To prepare the final input signals for the shake table, the acceleration records generated
by S Y N T H program were cut off at appropriate times and conditioned with ramps and
times, ramps and post pulses were chosen such that the acceleration records and their
corresponding velocities and displacements had reasonably smooth ramp-ups and ring-
acceleration records. Integration was done numerically in the frequency domain (Section
3.6) with the help o f a custom-developed Mathcad worksheet. Resulting velocities and
displacements were checked to assure that they were within the limits o f the shake table.
Since the peak displacements o f all the records were beyond the limits o f the shake table,
they were filtered by high-pass filters. The cut off frequencies were chosen for each
In preparing the input records, I E E E 693-1997 requirement for the time history shake
table tests ( 9 )
was also taken into account. According to this requirement, "The input
duration o f the time history tests shall be at least 20 s o f strong motion. R i n g down time
or acceleration ramp-up time shall not be included i n the 20 s o f strong motion. The
duration o f strong motion shall be defined as the time range between when the plot o f the
time history first reaches 25% o f the maximum value to the time when it falls for the last
time to 25% o f the maximum value". The signal generated from Northridge record
51
Chapter 5 Shake Table Tests
conformed to this requirement. The signal generated from E l Centro record had a shorter
duration o f strong motion and was used in order to observe different effects that might be
caused by a different type o f record. The record generated from Landers record had a
very long duration o f strong motion which obviously satisfied the I E E E requirements.
Figures 5.5.3 to 5.5.5 show the resulting input records. Since the records whose response
spectra match B C Hydro spectrum were less severe than the ones whose response spectra
match I E E E spectrum, only the later records were used in this investigation. However,
additional tests were carried out by using the less severe records to assure that the
The performance o f the shake table was verified by running the shake table with the
specimen on it. The synthetic generated records were used as the input signals. The actual
acceleration o f the shake table was measured. The peak accelerations were checked to
have the desired value and the response spectra o f these recorded accelerations were
verified to have a good match with the target spectra. It was found that the best match
would be obtained by using these records at 90% amplitude. However, it was decided to
use the signal generated from the Landers record at 75% o f its amplitude since it was a
very long duration record and a 90% scale simulation could cause irreparable damage to
the specimens. Figures 5.5.6 and 5.5.7 show the target spectra and the spectra o f the
generated records. Figure 5.5.8 shows the response spectra from the actual acceleration
response o f the shake table with the specimen attached to it. It shows a very good match
52
Chapter 5 Shake Table Tests
Northridge
< -0.5 4
E l Centro
10 20 30 40 50 60
Time (sec)
Landers
53
Chapter 5 Shake Table Tests
Frequency (Hz)
Frequency (Hz)
Frequency (Hz)
Figure 5.5.3 Comparison o f the B C Hydro target spectrum for 2 % damping and the
54
Chapter 5 Shake Table Tests
Figure 5.5.4 Comparison o f the I E E E 693-1997 target spectrum for 2 % damping and the
55
Chapter 5 Shake Table Tests
-Target
- Shake Table
Frequency (Hz)
•Target
-Shake Table
Frequency (Hz)
-Target
- Shake Table
Frequency (Hz)
Figure 5.5.5 Comparison o f the I E E E 693-1997 target spectrum and the shake table
56
Chapter 5 Shake Table Tests
Figure 5.5.6. Input signal generated from Northridge record; I E E E target spectrum
57
Chapter 5 Shake Table Tests
•0.5 1 1 !
• 1 1 1 1 1
' ! 1 ! !
' ! ! !
• ! 1
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
t 1 1
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
10
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
Figure 5.5.7. Input signal generated from E l Centro record; I E E E target spectrum
58
Chapter 5 Shake Table Tests
50
25
>
o
-25
All
If f ll
1
i
ri
-50
4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 64 68 72 76 80
Time (sec)
Figure 5.5.8. Input signal generated from Landers record; I E E E target spectrum
59
Chapter 5 Shake Table Tests
identify the characteristics o f the equivalent equipment, first the stand-alone towers were
tested before the conductor was mounted. This section discusses these tests and presents
the test results. To find out the effect o f the conductor on the response o f the towers,
these results are compared with the results from the other test set-ups (Chapter 7). In
addition to the analyses carried out in this chapter, a comprehensive spectral analysis is
performed on the results o f these tests in Chapter 6 with the purpose o f studying the
extension cantilever beams. They were not interconnected by the flexible conductor.
Figure 5.6.1 shows a picture o f the test set-up. East Tower, the taller tower, was located
on the east side o f the shake table and West Tower, the shorter tower, was located on the
west side o f the shake table. Uni-axial earthquake motions in E - W direction were used
60
Chapter 5 Shake Table Tests
5.6.2 Instrumentation
The towers and the extension cantilever beams were instrumented as illustrated in Figure
5.6.2. East Tower was instrumented with two accelerometers attached to the sides o f its
response o f the tower and the other one measured its North-South (N-S) horizontal
acceleration response. West Tower was instrumented with three accelerometers attached
to the sides o f its plates. Two accelerometers measured the E - W horizontal acceleration
response o f the tower and the third one measured it along the N - S direction. The N - S
acceleration was measured for examining the possible noise introduced by vibration o f
the towers i n that direction. The E - W acceleration o f West Tower was measured at two
locations on its plates in order to study its torsional modes. This was done because the
mass o f the West Tower could not be considered as a concentrated mass since it was
61
Chapter 5 Shake Table Tests
provided by large plates. Both Towers were instrumented with cable extension
transducers (Section 5.7.2) which measured their displacements in the E-W direction.
The extension cantilever beams were also instrumented for investigating their effect on
the response of the Towers and the noise introduced by them. One accelerometer on each
set of cantilever beams measured N-S acceleration at the cantilevers tip and one Linear
The shake table itself was instrumented with three accelerometers, one measuring the
acceleration of the table in E-W direction, and the other two measuring the N-S
accelerations of the table at its east and west sides. The displacement of the table is
Table 5.6.1 shows a list of measured responses for this test set-up. In this table, "Col No."
indicates which column of numbers in the data file that contains the measurements, is the
corresponding "Type of Response" shown in the table. "Ch No." is the channel number
of the data acquisition system. Further details about transducers are included in Appendix
A.
62
Chapter 5 Shake Table Tests
Ch7
Ch 19
1 w
IT
Ch 16
Ch9
PLAN VIEW
Ch 16
Ch21 Ch8
— ^
(Ti
Ch9 Ch7
East Tower
West Tower
Ch3 Ch4
Ch5 Ch6
L J L
1 I I J=Z1_C
Ch31
Ch 29 Ch 2 Ch 30 Ch 1
ChO
ELEVATION VIEW
63
Chapter 5 Shake Table Tests
Col Ch
Type o f Response
No. No.
1 0 Table E - W Displacement
2 1 Table N - S Displacement/West Side
3 2 Table N - S Displacement/East Side
4 3 East Side Extension Cantilever Beams Tip Vertical Displacement
5 4 West Side Extension Cantilever Beams T i p Vertical Displacement
6 5 East Side Extension Cantilever Beams Tip Horizontal N - S Acceleration
7 6 West Side Extension Cantilever Beams T i p Horizontal N - S Acceleration
8 7 West Tower N - S Acceleration
9 8 West Tower E - W Acceleration/South Side
10 9 West Tower E - W Acceleration/North Side
11 16 East Tower E - W Displacement
12 19 East Tower N - S Acceleration
13 20 East Tower E - W Acceleration
14 21 West Tower E - W Displacement
15 29 Table N - S Acceleration/East Side
16 30 Table N - S Acceleration/West Side
17 31 Table E - W Acceleration
Table 5.7.1 shows the test sequence for stand-alone towers tests. These tests can be
• Sine sweep tests: The purpose o f these tests was to find the natural frequencies o f
the towers. The input o f the shake table was a sinusoidal wave with frequencies
was performed before all the other tests and one was performed after running all
the tests. This was done to see whether the characteristics o f the specimen were
64
Chapter 5 Shake Table Tests
• Step Force tests: The purpose o f these tests was to put the towers i n the state o f
free vibration by applying a step force to the system through the shake table. The
results o f these tests were used to measure the damping ratios o f the towers by
• Seismic loading tests: These tests involved subjecting the towers to the generated
synthetic earthquakes. A l l six records were used and they were used at various
amplitudes. In Table 5.6.2, the acronym before each input signal name indicates
the target spectrum and the second part o f the name indicates which earthquake
Test
Description o f the Test Input Signal Amplitude
No.
1 Frequencies o f Stand-Alone Towers Sinusoidal 0.2g
2 Damping o f Stand-Alone Towers Step Force N/A
3 Damping o f Stand-Alone Towers Step Force N/A
4 Seismic Loading - Stand-Alone Towers B C H - Northridge 50%
5 Seismic Loading - Stand-Alone Towers B C H - Northridge 100%
6 Seismic Loading - Stand-Alone Towers B C H - E l Centro 50%
7 Seismic Loading - Stand-Alone Towers B C H - E l Centro 100%
8 Seismic Loading - Stand-Alone Towers B C H - Landers 50%
9 Seismic Loading - Stand-Alone Towers B C H - Landers 100%
10 Seismic Loading - Stand-Alone Towers I E E E - Northridge 50%
11 Seismic Loading - Stand-Alone Towers I E E E - Northridge 90%
12 Seismic Loading - Stand-Alone Towers I E E E - E l Centro 50%
13 Seismic Loading - Stand-Alone Towers I E E E - E l Centro 90%
14 Seismic Loading - Stand-Alone Towers I E E E - Landers 50%
15 Seismic Loading - Stand-Alone Towers I E E E - Landers 75%
16 Frequencies o f Stand-Alone Towers Sinusoidal 0.2g
17 Damping o f Stand-Alone Towers Step Force N/A
18 Damping o f Stand-Alone Towers Step Force N/A
65
Chapter 5 Shake Table Tests
Natural frequencies were obtained by analyzing the results o f sine sweep tests with the
read and their Fourier amplitudes were calculated by using the Fast Fourier Transform
(FFT) algorithm. The trends o f the signal, i f existed, were removed. Then, normalized
Fourier amplitudes were calculated by dividing the Fourier amplitudes by their maximum
absolute value. The resulting normalized Fourier amplitudes were plotted versus the
frequency. The natural frequencies are identified as the frequencies at which there is a
J
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 10
Frequency (Hz)
66
Chapter 5 Shake Table Tests
The natural frequencies o f the towers obtained from sine sweep tests are given in Table
5.6.3. Results o f the sine sweep tests also showed no difference in the natural frequencies
of the towers before and after applying seismic loadings. This indicated that there was no
yielding or damage in any part o f the structures and they remained elastic during the tests.
The natural frequencies shown i n Table 5.6.3 are the fundamental frequencies o f the
towers. Towers also have other modes o f vibration with different natural frequencies.
Other modes o f vibration were especially noticeable i n the behaviour o f the West Tower,
which experienced torsional effects because o f its large plates. Evidence o f this is shown
in Figure 5.6.4 which plots the F F T o f the subtraction o f the two acceleration responses at
two opposite sides o f the plates. Three peaks at 5.76, 6.80 and 7.80 H z in Figure 5.6.4
indicate the presence o f other modes. The presence o f these modes could also be caused
by the deflection and rotation o f the extension cantilever beams. The identification o f
these secondary modes and their effect on the response o f the system is discussed in more
detail i n Chapter 6.
67
Chapter 5 Shake Table Tests
0 1 2
1
3 4 5
JJ d
6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
West Tower
Damping ratios o f the towers were obtained by using the logarithmic decrement
method. (3)
Damping ratio is amplitude dependent. That means the damping ratio
calculated by choosing two peaks from the high-amplitude portion o f the free vibration
response would be different than the damping ratio obtained by choosing the peaks from
acceleration and displacement response o f the towers obtained from the first two step
load tests (Test 2 and Test 3) are used to calculate the damping ratios. Thus four values o f
damping ratios are obtained for each tower which are then averaged to give the final
damping ratio for the corresponding tower. The first and second peaks are chosen
Tables 5.6.4 and 5.6.5 show the calculation o f the damping ratios.
68
Chapter 5 Shake Table Tests
East Tower
Ratio £ =
of of the First
from the First (%)
Cycles Chosen Peak
Peak
1
In P
'
j Pi 2m' P\ + j
Pl-H
Displacement (mm) 38 1.504 0.355 0.0060
0.54
Acceleration (g) 43 0.081 0.022 0.0048
Test 3 (Step Force II)
Amplitude of Damping
Number Amplitude
the j Peak
t h
Ratio £" =
of o f the First
from the First
Cycles Chosen Peak
Peak
1
.n P
'
j Pi 2m' P\ + j
Pi i +
69
Chapter 5 Shake Table Tests
West Tower
Ratio ^ =
of o f the First
from the First (%)
Cycles Chosen Peak
Peak
1
In P
>
j Pi 2TU PI + j
P. i +
Ratio £ =
of o f the First
from the First
Cycles Chosen Peak
Peak
1
In *
j Pi 2nj Pi + j
Pi + i
70
Chapter 5 Shake Table Tests
In this section, a summary o f the test results from the seismic tests is presented, including
acceleration and displacement responses at the top o f the towers. A s mentioned earlier,
the results o f the tests with input signals having the I E E E target spectra were used in the
analyses. Figures 5.6.5 and 5.6.6 show sample time histories o f the results. Tables 5.6.6
to 5.6.8 present a summary o f the results and include the amplification factors for the
acceleration responses. Note that the displacements o f the towers are relative to the shake
table.
71
Chapter 5 Shake Table Tests
Input Signal
M a x i m u m Absolute
Response IEEE-Northridge I E E E - E l Centro IEEE-Landers
90% 90% 75%
Input Signal
M a x i m u m Absolute
Response IEEE-Northridge I E E E - E l Centro IEEE-Landers
90% 90% 75%
Input Signal
M a x i m u m Absolute
Displacement (mm) IEEE-Northridge I E E E - E l Centro IEEE-Landers
90% 90% 75%
East Tower 52 48 34
West Tower 66 63 59
Towers Relative 90 84 73
Displacement -87 -90 -73
72
Chapter 5 Shake Table Tests
73
Chapter 5 Shake Table Tests
Shake Table
1001 1 1 1 1 1 i 1 1 i 1 r
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max = 56 mm min = -59 mm
East Tower
1 Mi•
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max = 52 mm min = -51 mm
West Tower
-100
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max = 66 mm min = -61 mm
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max = 90 mm min = -87 mm
74
Chapter 5 Shake Table Tests
Results showed that the deflection response at the tips o f the extension cantilever beams
were at the rates o f the deflection o f the towers (Figure 5.6.7). This is because they
deflected under the loads caused by overturning moments o f the towers. Flexibility o f the
beams affected the overall stiffness o f the towers. A n equivalent stiffness can be
calculated for the equivalent equipment by knowing their masses and natural frequencies
and considering them as single degree o f freedom systems. Table 5.6.9 presents the
Natural
Equivalent Modal Equivalent
Frequency
Equipment Mass (kg) Stiffness (N/m)
(Hz)
75
Chapter 5 Shake Table Tests
—K J
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 10
Frequency (Hz)
"5.
E
<
0)
1.97 2.12 2.27 2.42 2.58 2.73 2.88 3.03 3.18 3.33 3.48 3.64 3.79 3.94 4.09 4.24 4.39 4.55 4.7 4.85 5
Frequency (Hz)
Figure 5.6.7 Frequency response o f vertical displacement at the tips o f cantilever beams
After testing the stand-alone towers, the flexible conductor was mounted and connected
to the towers such that its geometry was as depicted in Figure 4.2.2. This was the first set-
This test set-up is similar to the stand-alone towers set-up with the addition o f the
conductor interconnecting the towers (Figure 5.7.1). The conductor was connected to the
towers by using u-shaped aluminum connectors or "angle adapters" (Figure 4.2.2) and
triangular-shaped mounting brackets. The load cells were inserted between the angle
adapters and mounting brackets. Figures 5.7.2 and 5.7.3 show the connection o f the
76
Chapter 5 Shake Table Tests
Chapter 5 Shake Table Tests
Chapter 5 Shake Table Tests
5.7.2 Instrumentation
Various types o f transducers were employed for the dynamic data acquisition. A
were used i n these tests. Piezoelectric accelerometers are transducers that use
base. (10)
These accelerometers are not sensitive to the angle between the axis o f
the accelerometer and the direction o f motion. They always measure the
acceleration along their sensitive axis. This type o f accelerometer was used for the
sensitive to the angle between the axis o f the accelerometer and the direction o f
motion. This type o f accelerometer was used where the direction o f motion did
not change its angle, i.e. it always remained horizontal or vertical. For
proportional to the position o f the core. The L V D T may be used i n many different
measure the tip deflection o f the extension cantilever beams (Figure 5.7.4). The
second type o f displacement transducer used in these tests was cable extension
transducers or string pots. These transducers use a stainless steel cable wound
79
Chapter 5 Shake Table Tests
These transducers are installed by mounting their base to a fixed surface and
attach their cable to the movable object. Figure 5.7.5 shows a cable extension
Figure 5.7.4 L V D T used for measuring the vertical displacement at the tip o f the
80
Chapter 5 Shake Table Tests
Figure 5.7.5 Cable extension transducers used for measuring towers displacement
the forces at the connection o f the flexible conductor to the towers in both
horizontal and vertical directions. These force/torque load cells have six outputs
corresponding to the three force and the three torque components. These
transducers have a 12.57 c m (4.95 in) diameter top and bottom mounting surfaces
equipped with mounting holes and threaded inserts. The standard capacities o f
these load cells are 10680 N (2500 lb) in the direction normal to mounting
surfaces (z direction) and half o f that rating for the forces on the plane o f the
mounting surface (x and y directions) [ref]. Figure 5.7.2 shows one o f the load
Figure 5.7.6 shows the instrumentation plan and Table 5.7.1 has a list o f measured
responses for this test set-up. The towers and cantilever beams were instrumented with
custom-made aluminum clamps were employed for attaching the accelerometers to the
81
Chapter 5 Shake Table Tests
conductor (Figure 5.7.7). The load cells were used to measure the forces at the
connection points o f the conductor to the towers. They measured the vertical force (Fy)
and horizontal force i n E - W direction (Fz) at both connections. Figures 5.7.2 and 5.7.3
show how the load cells were installed. More details on the specifications o f the
82
Chapter 5 Shake Table Tests
Ch7
•
Ch8
Ch 19
1 Ch 1 Ch 10
Ch 16 Ch20 Ch21
• • Ch 14
— H
J / C h 15
Ch 13
Ch 12
E^jLw Ch9
T PLAN VIEW
Ch 17
Ch 18
Ch 16
Ch22
Ch8
— ^
Ch 15 Ch21 — ^
Ch9 Ch 7
East Tower
West Tower
Ch3 Ch4
Ch 5 Ch6
Ch31
Ch 29 Ch 2 Ch 30 Ch 1
ChO
ELEVATION VIEW
83
Chapter 5 Shake Table Tests
Col Ch
Type o f Response
No No.
1 0 Table E - W Displacement
2 1 Table N - S Displacement/West Side
3 2 Table N - S Displacement/East Side
4 3 East Side Extension Cantilever Beams Tip Vertical Displacement
5 4 West Side Extension Cantilever Beams T i p Vertical Displacement
6 5 East Side Extension Cantilever Beams T i p Horizontal N - S Acceleration
7 6 West Side Extension Cantilever Beams T i p Horizontal N - S Acceleration
8 7 West Tower N - S Acceleration
9 8 West Tower E - W Acceleration/South Side
10 9 West Tower E - W Acceleration/North Side
11 10 Conductor Acceleration Normal to the Conductor/West South Side
12 11 Conductor Acceleration Normal to the Conductor/East South Side
13 12 Conductor Acceleration Normal to the Conductor/East North Side
14 13 Conductor Acceleration Normal to the Conductor/West North Side
15 14 Conductor Acceleration Normal to the Conductor/Centre
16 15 Conductor Acceleration A l o n g the Conductor/Centre
17 16 East Tower E - W Displacement
18 17 Vertical Force F y - Conductor Connection to East Tower
19 18 Horizontal Force F z - Conductor Connection to East Tower
20 19 East Tower N - S Acceleration
21 20 East Tower E - W Acceleration
22 21 West Tower E - W Displacement
23 22 Vertical Force F y - Conductor Connection to West Tower
24 23 Horizontal Force F z - Conductor Connection to West Tower
25 29 Table N - S Acceleration/East Side
26 30 Table N - S Acceleration/West Side
27 31 Table E - W Acceleration
84
Chapter 5 Shake Table Tests
Chapter 5 Shake Table Tests
The test sequence for set-up 1 o f the interconnected towers was similar to that o f the
stand-alone towers and included sine sweep tests, step force tests and seismic loading
tests. Input signals were also the same as the ones used for the stand-alone towers, at the
same amplitudes. The test numbers followed the test numbers o f the stand-alone towers
Test
Test Description Input Signal Amplitude
No.
19 Frequencies o f Interconnected Towers Sinusoidal 0.2g
20 Damping o f Interconnected Towers Step Force N/A
21 Damping o f Interconnected Towers Step Force N/A
22 Seismic Loading B C H - Northridge 0.5
23 Seismic Loading B C H - Northridge 100%
24 Seismic Loading B C H - E l Centro 50%
25 Seismic Loading B C H - E l Centro 100%
26 Seismic Loading B C H - Landers 50%
27 Seismic Loading B C H - Landers 100%
28 Seismic Loading I E E E - Northridge 50%
29 Seismic Loading I E E E - Northridge 90%
30 Seismic Loading I E E E - E l Centro 50%
31 Seismic Loading I E E E - E l Centro 90%
32 Seismic Loading I E E E - Landers 50%
33 Seismic Loading I E E E - Landers 75%
34 Frequencies o f Interconnected Towers Sinusoidal 0.2g
35 Damping o f Interconnected Towers Step Force N/A
36 Damping o f Interconnected Towers Step Force N/A
In this section, tests results are summarized and presented in various forms. M a t h C A D
worksheets were developed for the analysis and presentation o f the results.
86
Chapter 5 Shake Table Tests
Resonance frequencies o f the towers are evaluated and compared to their natural
frequencies (Table 5.7.3). This comparison is to see whether the conductor stiffness has a
Table 5.7.3 Effect o f the conductor stiffness on resonance frequencies o f the towers
The results show that the conductor reduced the resonance frequencies o f the towers by
no more than 2%.
Strong shaking o f the conductor indicated that its natural frequencies are within the range
of the towers natural frequencies. However, no specific values for these frequencies could
be detected due to the nonlinearity o f the conductor and the fact that its natural
In this section, the time histories o f the acceleration response at selected locations o f the
specimen from selected tests are shown. A l s o , a summary o f the peak accelerations and
amplification factors is presented. Sample results are presented in Figures 5.7.8 and 5.7.9.
A l l presented data were digitally filtered as described in Chapter 3. More test results can
be found i n Appendix C .
87
Chapter 5 Shake Table Tests
Shake Table
_ | I I ! I I 1 I I I ] I I I I I I I I I I I
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max = 0.53 g min = -0.50 g
East Tower
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max = 1.77 g min = -1.97 g
88
Chapter 5 Shake Table Tests
1
3.5
"3.5
0
IlillRJJUIIA 1111111
4 INN
-7
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max = 6.92 g min =-4.16 g
V
3.5
fflfffl
"3.5
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max = 5.08 g min = -4.76 g
4 NMil
3 3 5
4
o
1 o WW WW
o
o
<
"3.5
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max = 3.91 g min =-2.66 g
3.5
3.5
0
If II MM MIA/
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max = 3.51 g min = -2.48 g
89
Chapter 5 Shake Table Tests
Tables 5.7.4 and 5.7.5 present peak acceleration responses and amplification factors o f
the towers. Peak accelerations o f the shake table are also presented.
Input Signal
M a x i m u m Absolute
Response IEEE-Northridge I E E E - E l Centro IEEE-Landers
90% 90% 75%
Input Signal
M a x i m u m Absolute
Response IEEE-Northridge I E E E - E l Centro IEEE-Landers
90% 90% 75%
Tables 5.7.6 to 5.7.8 show the peak acceleration responses o f the conductor at its
instrumented locations (Figure 5.7.6). A t each location along the conductor, considerable
difference between the two normal accelerations measured on each o f the cables was
observed. This is ascribed to the initial out-of-plane deformation o f the conductor after it
was mounted (Figure 5.7.10). A s a result o f this irregularity in the conductor geometry,
one cable o f the twin conductor picked up more acceleration compared with the adjacent
one. In order to estimate one value for the peak accelerations, the two values at each
90
Chapter 5 Shake Table Tests
location were averaged. These average numbers represent reasonable values for the peak
accelerations since the two cables o f the twin conductor vibrated reasonably in phase (see
Amplification
10.36 7.38 7.88 6.62
Factor
91
Chapter 5 Shake Table Tests
Amplification
11.13 7.23 8.80 8.57
Factor
Amplification
10.37 7.91 8.34 6.74
Factor
92
Chapter 5 Shake Table Tests
transducers. The conductor, however, did not have displacement transducers and its
Figures 5.7.11 and 5.7.12 show sample displacement time-histories obtained from Test
93
Chapter 5 Shake Table Tests
Shake Table
100
50
lit fll A/
-50
VI n V vV
-100
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max = 56 mm min = -59 mm
East Tower
100
J
5
„
£
E
a.
50
-50
i Ii 1 ii lillM l JIM WWmm
-100
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max = 36 mm min = -40 mm
West Tower
100
1
„ 50
ill
£
upp 1 If IN
£
J 0 WW (WWAMMWWwwwwww
5
"50
-100
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max = 52 mm min = -48 mm
I 50
i
o
.2
o.
0
$ INMM l WW
Q "50
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
94
Chapter 5 Shake Table Tests
E 0
-200
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max =133 mm min = -166 mm
E 0
P, 0
-200
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max = 43 mm min = -52 mm
£ 0
-200
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max = 99 mm min = -96 mm
95
Chapter 5 Shake Table Tests
Table 5.7.9 summarizes the peak values o f the displacement response o f the towers
relative to the shake table. Table 5.7.9 also includes the peak values o f the towers
displacement relative to each other. These values indicate the maximum stretch or
compression that the conductor experienced during the tests. Positive number means that
the conductor was compressed and negative number means the conductor was stretched.
Input Signal
Location and Type o f
Response IEEE-Northridge I E E E - E l Centro IEEE-Landers
90% 90% 75%
East Tower; Relative to
40 35 42
Shake Table
West Tower; Relative to
52 62 55
Shake Table
Towers; Relative to Each 65 72 64
Other -66 -75 -59
Table 5.7.10 summarizes the peak displacement response o f the conductor at the
locations o f the accelerometers (Figure 5.7.6). Similar to the acceleration response o f the
conductor, these values are obtained by averaging the displacements o f the two cables at
each location.
96
Chapter 5 Shake Table Tests
Location o f A l o n g the
Normal to the Conductor
Input Signal the Conductor
Transducer
East Centre West Centre
(Location 1) (Location 2) (Location 3) (Location 2)
North Side 166 52 91 99
IEEE-Northridge South Side 89 N/A 38 N/A
90%
Average 127 52 64 99
To better understand the behaviour o f the conductor, its deflected shapes were obtained
reconstruct the deflected shape. The initial geometry o f the conductor was obtained by
measuring its coordinates at 16 locations along its length (Appendix B ) . Figure 5.7.13
shows the deflected shape o f the conductor at various positions obtained from Test 29.
Figure 5.7.14 shows pictures o f the deflection o f the conductor during one o f the tests.
97
Chapter 5 Shake Table Tests
98
Chapter 5 Shake Table Tests
In this section, the forces measured by the load cells are presented. A s mentioned earlier,
two force components were measured at each conductor attachment point. They are the
Table 5.7.11 shows the static forces in the connections o f the conductor. They were
obtained by measuring forces before applying the seismic loadings. These forces are the
amount o f baseline shift i n the measured force time-histories and can be obtained from
The maximum absolute dynamic forces generated i n the connections o f the flexible
conductor due to seismic loadings are summarized in Table 5.7.12. A force parameter,
denoted as Force Amplification Factor ( F A F ) , is used here to compare dynamic and static
forces. This parameter provides a scale for comparing the severity o f dynamic forces.
F A F is defined as:
Table 5.7.13 presents the F A F o f the measured forces obtained from the seismic loading
tests. For easier comparison, these results are also presented by bar charts (Figure 5.7.15).
99
Chapter 5 Shake Table Tests
1
100
Chapter 5 Shake Table Tests
12
10
'3 6
i—I V
12
10
1 6
^ 4
101
Chapter 5 Shake Table Tests
These results show that the force response o f the east tower to I E E E - E l Centro record at
both magnitude scales is consistently higher than the response to the other input records.
This difference is more obvious between the East Tower F y responses for which the F A F
obtained from I E E E - E l Centro seismic loading is 23% higher than that obtained from
observation since the records used for these tests have all similar response spectra and
hence similar frequency content, and elucidates the sensitivity o f the system to other
characteristics o f the ground motion not clearly identified in the response and frequency
spectra. This indicates that the time domain characteristic o f the seismic loading could
play an important role in the force response o f the conductor. These results also show that
at 90%. This is despite the fact that the amplitude o f the latter record is about 20% higher
than that o f the former one. This again indicates the importance o f the time domain
To find out a relation between the generated forces and the response o f the towers,
dynamic force-displacement hysteresis loops were plotted (Figures 5.7.16 and 5.7.17).
These diagrams show that F z at each tower peaks when the displacement o f the same
tower reaches its maximum value (Figure 5.7.16). However it does not necessarily peak
when the towers reach their maximum displacement relative to each other. This indicates
that the horizontal force response at each tower is highly dependent on the displacement
response o f the same tower, and that this force could reach its maximum value even i f the
conductor is not experiencing its maximum stretch or compression. Results also show no
specific pattern i n the F y force-displacement response o f the towers (Figure 5.7.17). This
indicates that the F y response at each tower is dependent on the response o f both towers.
In other words the vertical force response at each tower is affected by the response o f the
adjacent tower.
5.7.18). The results show that F z and F y could reach their maximum value at the same
time. This indicates that the maximum horizontal and vertical forces could affect the
102
Chapter 5 Shake Table Tests
in Chapter 6.
103
Chapter 5 Shake Table Tests
2000
104
Chapter 5 Shake Table Tests
1100
% 550
£ "550
105
Chapter 5 Shake Table Tests
2000
2000
106
Chapter 5 Shake Table Tests
The purpose o f flexible conductor test set-up 2 was to observe the behaviour o f the
conductor under bidirectional seismic loadings. The input signals used for this set-up
were the generated earthquake records applied at l o w amplitude i n both the E - W and N - S
directions.
This test set-up is a modification o f set-up 1. The modification to the specimen consisted
of removing the bracing o f the west tower.
5.8.2 Instrumentation
The responses measured in this test set-up were similar to those o f set-up 1. Only the
arrangement o f the accelerometers on the conductor was modified in order to measure the
acceleration o f the conductor along the N - S direction. The N - S horizontal forces in the
conductor connections (Fx) were also measured by using the same load cells that
measured other force components. Figure 5.8.1 shows the instrumentation plan for this
107
Chapter 5 Shake Table Tests
Col Ch
Type o f Response
No No.
1 0 Table E - W Displacement
2 1 Table N - S Displacement/West Side
3 2 Table N - S Displacement/East Side
4 3 East Side Extension Cantilever Beams T i p Vertical Displacement
5 4 West Side Extension Cantilever Beams T i p Vertical Displacement
6 5 East Side Extension Cantilever Beams T i p Horizontal N - S Acceleration
7 6 West Side Extension Cantilever Beams T i p Horizontal N - S Acceleration
8 7 West Tower N - S Acceleration
9 8 West Tower E - W Acceleration/South Side
10 9 West Tower E - W Acceleration/North Side
11 10 Conductor Acceleration Normal to Conductor/Centre South Side
12 11 Conductor Acceleration N-S/East North Side
13 12 Conductor Acceleration Normal to Conductor/East North Side
14 13 Conductor Acceleration N-S/Centre North Side
15 14 Conductor Acceleration Normal to Conductor/Centre North Side
16 15 Conductor Acceleration A l o n g the Conductor/Centre North Side
17 16 East Tower E - W Displacement
18 17 Vertical Force F y - Conductor Connection to East Tower
19 18 Horizontal Force F z - Conductor Connection to East Tower
20 19 East Tower N - S Acceleration
21 20 East Tower E - W Acceleration
22 21 West Tower E - W Displacement
23 22 Vertical Force F y - Conductor Connection to West Tower
24 23 Horizontal Force F z - Conductor Connection to West Tower
25 25 Conductor Acceleration N-S/West North Side
26 26 East Tower E - W Acceleration/North Side
27 27 Horizontal Force F x - Conductor Connection to East Tower
28 28 Horizontal Force F x - Conductor Connection to West Tower
29 29 Table N - S Acceleration/East Side
30 30 Table N - S Acceleration/West Side
31 31 Table E - W Acceleration
108
Chapter 5 Shake Table Tests
Ch7
•
Ch 8
-+
Ch 19
1 Ch20
•
Ch 10
Ch 16 Ch21
Ch 13- »
Ch 14
Ch 26 /Ch 15
Ch25
ctT 12 Ch 11
Ch9
PLAN VIEW
N
Ch 17
Ch 16
Chl2 Ch22
Ch28
Ch 25 w
ChlO ^ Ch 23 i
Ch 14
Ch8
— ^
J/] Chl5 Ch21 — ^ W
Ch9 Ch7
Ch 13
C=l L"
East Tower
West Tower
Ch3 Ch4
w Ch5 Ch6
tzzj
Ch31
Ch 29 Ch 2 Ch 30 Ch 1
ChO
ELEVATION VIEW
109
Chapter 5 Shake Table Tests
For test set-up 2 only the input records which were generated from Northridge and E l
Centro records were used. The input signals were used in both E - W and N - S directions
simultaneously. They were applied at low amplitudes since the extension cantilever
beams were not originally designed to take large transverse forces. The maximum
amplitude was chosen at 35% in each direction and there was no phase lag between the
two signals. This is equivalent to the same record being applied at 50% magnitude with
45° degree angle from the E - W direction. Before each bidirectional test, one uni-axial test
for each direction was carried out. Test sequence for the set-up 2 is shown i n Table 5.8.2.
110
Chapter 5 Shake Table Tests
Test Amplitude
Test Description Input Signal
No. E-W N-S
1 Frequencies o f Interconnected Towers E / W Sinusoidal E / W 0.2g 0
2 Damping o f Interconnected Towers E / W Step Force I E / W N/A N/A
3 Damping o f Interconnected Towers E / W Step Force II E / W N/A N/A
4 Frequencies o f Interconnected Towers N / S Sinusoidal N / S 0 0.2g
5 Damping o f Interconnected Towers N / S Step Force I N / S N/A N/A
6 Damping o f Interconnected Towers N / S Step Force II N / S N/A N/A
7 Seismic Loading IEEE-Northridge 18% 0%
8 Seismic Loading IEEE-Northridge 0% 18%
9 Seismic Loading IEEE-Northridge 18% 18%
10 Seismic Loading IEEE-Northridge 35% 0%
11 Seismic Loading IEEE-Northridge 0% 35%
12 Seismic Loading IEEE-Northridge 35% 35%
13 Seismic Loading I E E E - E l Centro 18% 0%
14 Seismic Loading I E E E - E l Centro 0% 18%
15 Seismic Loading I E E E - E l Centro 18% 18%
16 Seismic Loading I E E E - E l Centro 35% 0%
17 Seismic Loading I E E E - E l Centro 0% 35%
18 Seismic Loading I E E E - E l Centro 35% 35%
19 Frequencies o f Interconnected Towers E / W Sinusoidal E / W 0.2g 0
20 Damping o f Interconnected Towers E / W Step Force I E / W N/A N/A
21 Damping o f Interconnected Towers E / W Step Force II E / W N/A N/A
22 Frequencies o f Interconnected Towers N / S Sinusoidal N / S 0 0.2g
23 Damping o f Interconnected Towers N / S Step Force I N / S N/A N/A
24 Damping o f Interconnected Towers N / S Step Force II N / S N/A N/A
Since the main objective o f performing tests on the set-up 2 was to investigate the
magnitude o f forces generated when the actual ground motion is not only i n the direction
of the conductor, only a summary o f the force responses is presented here. The
comparison o f these forces with the forces obtained from set-up 1 is included in Chapter
7. Table 5.8.3 shows the maximum absolute force response at the conductor connections
obtained from this test set-up. Table 5.8.4 presents the resultant o f the horizontal forces
111
Chapter 5 Shake Table Tests
(Fz and Fx) calculated by using the Square Root of Sum o f the Squares (SRSS) rule of
combination.
112
Chapter 5 Shake Table Tests
The first mode natural frequencies o f the interconnected equipment configuration 2 are
1.1 and 3.4 H z . So there is 2.3 H z difference between their first mode natural frequencies.
However, the first mode natural frequencies o f the towers used as equivalent equipment
in set-up 1 were 3.4 and 2.85 H z which were 0.55 H z different. So the towers were
modified so that their natural frequencies would be more different. This was to study the
effect o f frequency difference between the towers on the results. Comparing the results of
these tests with the results o f set-up 1 would provide a better insight into the interaction
To achieve the above mentioned goal, the first mode natural frequency o f the East Tower
was increased. This was achieved by removing 5 o f the 8 plates o f the East Tower which
decreased its mass and therefore increased its natural frequency. Everything else was
similar to test set-up 1 and the bracings o f the West Tower which were removed for test
5.9.2 Instrumentation
Instrumentation for this set-up was similar to that o f set-up 1 with a few differences. A
list o f responses measured in this set-up is presented i n Table 5.9.1. Figure 5.9.1 shows
113
Chapter 5 Shake Table Tests
Ch7
Ch8
Ch 19
Ch 20 Ch 11 Ch 10
Ch 16 Ch 13 Ch21
Ch 14
—£ /Ch 15
Ch26 /
• ' Ch25
Ch 12
E ^ J ^ W Ch9
T PLAN VIEW
Ch 16
Ch8
— ^
Ch21 — •
Ch9 Ch7
Ch 13
E a s t T o w e r
W e s t T o w e r
Ch3 Ch4
• c
Ch 5 Ch6 19
L_l L
J u
Ch31
Ch 29 Ch 2 Ch 30 Ch 1
ChO
ELEVATION VIEW
114
Chapter 5 Shake Table Tests
Col Ch
Type of Response
No No
1 0 Table E - W Displacement
2 1 Table N - S Displacement/West Side
3 2 Table N - S Displacement/East Side
4 3 East Side Extension Cantilever Beams Tip Vertical Displacement
5 4 West Side Extension Cantilever Beams Tip Vertical Displacement
6 5 East Side Extension Cantilever Beams Tip Horizontal N - S Acceleration
7 6 West Side Extension Cantilever Beams Tip Horizontal N - S Acceleration
8 7 West Tower N - S Acceleration
9 8 West Tower E - W Acceleration/South Side
10 9 West Tower E - W Acceleration/North Side
11 10 Conductor Acceleration Normal to the Conductor Axes/West South Side
12 11 Conductor Acceleration Normal to the Conductor Axes/East South Side
13 12 Conductor Acceleration Normal to the Conductor Axes/East North Side
14 13 Conductor Acceleration Normal to the Conductor Axes/N-S Direction/Centre
15 14 Conductor Acceleration Normal to the Conductor Axes/Centre
16 15 Conductor Acceleration A l o n g the Conductor Axes/Centre
17 16 East Tower E - W Displacement
18 17 Vertical Force F y - Conductor Connection to East Tower
19 18 Horizontal Force F z - Conductor Connection to East Tower
20 19 East Tower N - S Acceleration
21 20 East Tower E - W Acceleration/South Side
22 21 West Tower E - W Displacement
23 22 Vertical Force F y - Conductor Connection to West Tower
24 23 Horizontal Force F z - Conductor Connection to West Tower
25 25 Conductor Acceleration Normal to the Conductor Axes/West North Side
26 26 East Tower E - W Acceleration/North Side
27 27 Horizontal Force F x - Conductor Connection to East Tower
28 28 Horizontal Force F x - Conductor Connection to West Tower
29 29 Table N - S Acceleration/East Side
30 30 Table N - S Acceleration/West Side
31 31 Table E - W Acceleration
115
Chapter 5 Shake Table Tests
The same test sequence as for test set-up 1 was used for this set-up too. However the less
severe records with B C Hydro target spectrum were not used. The test sequence is shown
in Table 5.9.2.
Test
Test Description Input Signal Amplitude
No.
1 Frequencies o f Interconnected Towers Sinusoidal 0.2g
2 Damping o f Interconnected Towers Step Force N/A
3 Damping o f Interconnected Towers Step Force N/A
4 Seismic Loading IEEE-Northridge 50%
5 Seismic Loading IEEE-Northridge 90%
6 Seismic Loading I E E E - E l Centro 50%
7 Seismic Loading I E E E - E l Centro 90%
8 Seismic Loading IEEE-Landers 50%
9 Seismic Loading IEEE-Landers 75%
10 Frequencies o f Interconnected Towers Sinusoidal 0.2g
11 Damping o f Interconnected Towers Step Force N/A
12 Damping o f Interconnected Towers Step Force N/A
Test results from test set-up 3 are presented in a similar fashion to that o f set-up 1. A
comparison between the results from these two set-ups is done in chapter 7.
The resonance frequencies o f the interconnected towers are presented i n Table 5.9.3. A s
shown in Section 5.7.4.1, the resonance frequencies o f the interconnected towers can be
considered as their natural frequencies. So the natural frequency o f the new East Tower
was obtained by identifying its resonance frequency from this test set-up. Table 5.9.3
116
Chapter 5 Shake Table Tests
shows that a difference o f 1.74 H z was achieved between the natural frequencies o f the
towers.
Figures 5.9.2 and 5.9.3 present sample acceleration time-histories obtained from Test 5.
Tables 5.9.4 and 5.9.5 present peak acceleration responses and amplification factors of
the towers. Tables 5.9.6 to 5.9.8 show the peak acceleration responses o f the conductor
117
Chapter 5 Shake Table Tests
Shake Table
T InIk [to YY
- 1 1
— 1
— 1
— 1
— 1
— —1 1
— 1
— 1
— 1
— — —1 1 1
— — — — — — — — '
1 1 1 1 1 1 1
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max = 0.54 g min = -0.51 g
West Tower
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max = 1.86 g min = -1.77 g
118
Chapter 5 Shake Table Tests
"3.5
1
3 3.5
c
J
CJ
CJ
0
flu in WW W W * MVUW
I
< "3.5 1
-7
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max = 4.89 g min = -6.56 g
3.5
"3.5
-7
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max = 4.39 g min = -3.86 g
3.5
"3.5
II ll Www
-7
0 2 4 6 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max = 3.83 g min = -3.44 g
119
Chapter 5 Shake Table Tests
Input Signal
M a x i m u m Absolute
Response IEEE-Northridge I E E E - E l Centro IEEE-Landers
90% 90% 75%
Input Signal
M a x i m u m Absolute
Response IEEE-Northridge I E E E - E l Centro IEEE-Landers
90% 90% 75%
Amplification
9.96 8.13 9.49 7.23
Factor
120
Chapter 5 Shake Table Tests
Amplification
9.95 7.59 9.51 6.68
Factor
Amplification
12.27 10.27 11.86 8.37
Factor
121
Chapter 5 Shake Table Tests
results from set-up 1. Figures 5.9.4 and 5.9.5 show sample results. Table 5.9.9
summarizes the peak values o f the displacement response o f the towers. Displacements
are relative to the shake table. Towers displacements relative to each other is also
included. Table 5.9.10 summarizes the peak absolute displacement response o f the
122
Chapter 5 Shake Table Tests
Shake Table
l\l Ju AA/idAA
vil V
V
Jul
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max = 56 mm min = -59 mm
East Tower
100
„ 50
c
a
E
D-
0
If H M m mmmm
-50
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
-100 Time (sec)
max = 31 mm min = -31 mm
West Tower
100
„ 50
c
E
i 1 mmMAM wmIWVWwww mm
OJ
J
5
CL
0
1
"50
-100
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max = 54 mm min = -48 mm
I 50
c
OJ
£
(J
a.
0
Hi to tow WW
5 "50
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
-100 Time (sec)
max = 70 mm min = -60 mm
123
Chapter 5 Shake Table Tests
B 0
"200
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max =139 mm min = -179 mm
B 0
-200
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max =188 mm min = -159 mm
B 0
-200
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max = 42 mm min = -52 mm
£ o
-200
0 2 4 6 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
max =103 mm min = -87 mm
124
Chapter 5 Shake Table Tests
Input Signal
Location and Type o f
Response IEEE-Northridge I E E E - E l Centro IEEE-Landers
90% 90% 75%
East Tower; Relative to
31 27 28
Shake Table
West Tower; Relative to
54 63 55
Shake Table
Towers; Relative to Each 70 69 74
Other -59 -65 -69
125
Chapter 5 Shake Table Tests
Results from the force response o f the system are summarized similar to that of set-up 1.
126
Chapter 5 Shake Table Tests
127
Chapter 5 Shake Table Tests
These results show the tendency o f the system for higher response to IEEE-Landers
record, even though its magnitude is lower than the other records. It can also be observed
that in contrast to set-up 1, the response to I E E E - E l Centro is not more severe than the
response to the other records. Since all these records have similar response spectra, the
difference indicates that the time domain characteristics o f the ground motion could be
important. A detailed comparison between these results and results obtained from set-up
1 is included in chapter 7.
Similar to set-up 1, to find out a relation between the generated forces and the response o f
the towers, dynamic force-displacement hysteresis loops were plotted (Figures 5.9.7 and
5.9.8). These diagrams are similar to those o f set-up 1 (Figures 5.7.16 and 5.6.17) and
thus the observations and comments on them are also consistent to those o f set-up 1
(Section 5.7.5.4).
F z is plotted against F y (Figure 5.9.9). Similar to set-up 1, results show that F z and F y
could reach their maximum value at the same time. This indicates that the maximum
horizontal and vertical forces could affect the equipment simultaneously. Detailed
128
Chapter 5 Shake Table Tests
2000
Z 1000
N
% 0
o
H
i3 -looo
2000
2000 I I I I I I I 1 1 1 1 1 1 1 1
129
Chapter 5 Shake Table Tests
-20 -10 0 10 20 70
East Tower Displacement (mm)
130
Chapter 5 Shake Table Tests
2000
2000
131
Chapter 6 Spectral Analysis
CHAPTER 6
SPECTRAL ANALYSIS
6.1 Introduction
This chapter provides further insight into the behaviour o f the conductor and its
interaction with the towers under seismic loading. The main goal is the identification of
frequencies at which the force responses at the conductor connections are maximum and
relate them to the modal characteristics o f the system. This requires modal identification
of the towers and the conductor and it is accomplished by performing spectral analysis,
which is interpretation of data by using correlation and spectral density functions. The
basic concepts of frequency domain analysis, which are the basis for spectral analysis, are
described in Chapter 3.
The first step towards finding the force response frequencies is to plot the power spectra
of the force output signals. The preliminary results showed that there is no significant
power for frequencies above 20 H z , therefore the signal is conditioned with a lowpass
filter at 20 H z in order to clean it up from high frequency noise. The power spectra were
for calculating and plotting the spectra as well as finding and sorting their peaks.
Figures 6.2.1 to 6.2.4 show the power spectra o f the four force responses to all three input
signals as well as the sinusoidal loading o f the sine sweep test for the test set-up 1. The
force responses are East Tower F z and F y and West Tower F z and F y . Figures 6.2.5 to
132
Chapter 6 Spectral Analysis
East Tower Fy
i io 1
MO 4
1
I io 3
1 k ji
100
J \i h 1 . ii i
10
\J V A if A rVi N
[
1 ' i
Hi
0.1
0.01
S
3
10
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
East Tower Fz
10 II 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
West Tower Fy
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
West Tower Fz
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
133
Chapter 6 Spectral Analysis
East Tower Fy
,1
h
L
A
/I A A
m1 w
1
/
rVU/
VI
%
* yv
A/
' V • r
\
lA
v
i
O 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
East Tower Fz
1-10 I — I — I — I — I — I — i — I — I 1 — 1 — 1 1 — I — T
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
West Tower Fy
—j A,
P K AM fc\7
M
{ fi
r
A /V
V
A .A
i-
-A
i ~—
I
1
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
134
Chapter 6 Spectral Analysis
East Tower Fy
1 10
-i
1 io 4
r \H
1 -io 3
A A
Vt
100
10
• t e
1
V
w »
0.1
s 0.01
o 10 3
i -io" 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
East Tower Fz
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
West Tower Fy
1 10
" MO 4
1 10
100
2 to
s 10
5 1
0.1
0.01
I l l 0
_3
°* 1 -10_
1 10 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
West Tower Fz
1 10
•8 MO 6
1 10
00 4
1 10
I
mm
M0 J
tS 100
10
1
o
°" 0.1
0.01 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
135
Chapter 6 Spectral Analysis
East Tower Fy
1-10°
3 1-10 5
3 4
EE
= 1 -10
3)
3 110
i
K
|
j
100
10
M
L)
§• 1
f °-'
0.01
1 10~ 3
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
East Tower Fz
1 10
-S i -io 6
-f-
2 5
c 1 "10
w>
A
I
/n
4
110
-J
m
g MO 3
to u
J -
TV (Mr
10
1
0.1
0.01
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
West Tower Fy
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
West Tower Fz
MO
110 6
1 -IO 5
1 10 J
J
10
I 1
136
Chapter 6 Spectral Analysis
East Tower Fy
1 10
£
|
MO
MO
4
3
-4- -ft-
I V
i
100
I
i •
10
/A
5 1 *\
S. o.i -A I"
"£ o.oi
I MO 3
fc
1 -io" 4
1 10
0 1 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
East Tower Fz
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
West Tower Fy
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
West Tower Fz
1 10
•S MO 5
H 4
•5 1-10 1
on ,
| 1 -10
| 100 t 1 i
5 10
1
(A V ft* LA
oo
| 0.1
°- 0.01
1 10~ 3
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
137
Chapter 6 Spectral Analysis
East Tower Fy
i -io I I i I I I I I j i I I I I r
0 j 1 1 1 1 1 1 1 1 I I I I I I I ! I I I I I I 1 I I I
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
East Tower Fz
n
A
r
VA A /
A
V
(
J V V\ JV
\ ,
W
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
5 West Tower Fy
i -io I I I i i—I—i—I—i—I—j—I—i—i—r
j i—i—i—i—i—i—i—i—i—i—i—: i i i i i i i i i i i i i i
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
138
Chapter 6 Spectral Analysis
East Tower Fy
\ U/
A
/
Ju A
A \ A./A
,i *
J\
V
V \ Aj AA
v
Vi H A/
10 \ ->—
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
East Tower Fz
jr*
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
5 West Tower Fy
i -io I—i—I—I—I—I—i—i—i—I—I—i—i—i—r
o oi—•—•—'—'——————•—*—————i
1 1 1 1 1 1 1 1 1 1
'• i i i i 1 1 1
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
West Tower Fz
A, .uA . f\ rt A
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
139
Chapter 6 Spectral Analysis
East Tower Fy
A
{
/ rNV A
Y //
V /\ J \
A
i\r(V\iKi
A X 1 K
A /
V V f\r\
V \
0 1 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
East Tower Fz
-f
\
H
i
JO. V
K * l pa
TV Art/
w ' W W^ 1
0 1 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
West Tower Fy
4
XA K ft s X \J* v V
-fr
fj
V V
6 7 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
West Tower Fz
/ VJ\
j A
y A
J
V
vA\ K '\A
A
\
j
r A
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
140
Chapter 6 Spectral Analysis
The peaks o f the spectra i n figures 6.2.1 to 6.2.8 show the frequencies at which the forces
mainly respond to the excitation o f the system. These frequencies are basically the
resonance frequencies at which either the towers or the conductor respond. B y examining
the magnitude o f the spectra at the mentioned frequencies, it is possible to find out what
characteristic o f the system is most influential in the amount o f forces generated at the
To make this examination, the magnitude and the corresponding frequency o f the first
four peaks from the presented spectra, excluding the ones obtained from sine sweep tests,
are summarized and presented in Tables 6.2.1 and 6.2.2. To make more sense out o f the
tabulated numbers and in order to summarize the results o f the power spectra, a
normalized parameter is used. This parameter is introduced as the ratio o f the magnitude
of the spectrum at each frequency to the sum o f the four magnitudes o f the peaks. This
The contribution o f the system response at each frequency in generating forces would
closer to 1.0 for PI indicates higher contribution o f the corresponding system response
in generating the corresponding forces. The values o f PI are also shown i n Tables 6.2.1
and 6.2.2. For easier comparison o f PI for different force responses, the results are
141
Chapter 6 Spectral Analysis
Table 6.2.1 PI for forces at the conductor connection to East Tower; set-up 1
142
Chapter 6 Spectral Analysis
Table 6.2.2 PI for forces at the conductor connection to West Tower; set-up 1
143
Chapter 6 Spectral Analysis
East Tower F y
1.00
0.90
0.80
0.70
0.60
East Tower F z
1.00
0.90
0.80
0.70
0.60
E 0.50
0.40
0.30
0.20
0.10
0.00 -I
2.85 3.45 5.78 6.34
Frequency (Hz)
144
Chapter 6 Spectral Analysis
W e s t Tower F y
1.00
0.90 1
0.80
0.70
1
0.60
E 0.50
0.40 - 1
0.30
0.20
0.10
— I
0.00
2.85 3.45 5.78 6.34
Frequency (Hz)
West Tower F z
111
145
Chapter 6 Spectral Analysis
A s can be seen i n the results, the F z responses, which are the horizontal forces at the
towers. The F z response at West Tower is at the fundamental frequency o f the West
Tower and the responses at other frequencies are negligible. Similarly, the F z response at
East Tower is at the fundamental frequency o f the tower and the responses at other
frequencies are negligible. There is however one exception. That is the F z response at
explained in section 6.3, this frequency is one o f the other resonance frequencies o f the
West Tower. This result indicates that the horizontal forces measured at one tower could
The F y responses however, do not respond only at the fundamental frequencies o f the
6.3, all the first four frequencies at which the force responses peak, are one o f the
resonance frequencies o f the towers related to either a normal mode or a coupled mode o f
vibration. This shows that the behaviour o f the conductor and the interaction o f the
conductor with the towers are such that the vibration o f the towers mainly affect the
forces induced in the connections o f the conductor. It can also be observed that F y at each
tower primarily responds at the fundamental frequency o f the tower except for the F y
vibration o f the West Tower as can be seen i n Figure 6.2.12. This indicates that the time
domain characteristic o f the ground motion could play a role i n the interaction o f the
In order to further investigate the force and conductor acceleration response frequency
contents, a similar analysis was carried out on the test results from test set-up 3, i n which
the natural frequency o f the East Tower is higher than its corresponding value in test set-
up 1. The results from both test set-ups reveal similar characteristics for the interaction
between the conductor and the towers with minor differences. The results from set-up 3
are shown in Tables 6.2.3 and 6.2.4 and in Figures 6.2.12 and 6.2.13.
146
Chapter 6 Spectral Analysis
Table 6.2.3 PI for forces at the conductor connection to East Tower; set-up 3
147
Chapter 6 Spectral Analysis
Table 6.2.4 PI for forces at the conductor connection to West Tower; set-up 3
148
Chapter 6 Spectral Analysis
East Tower F y
1.00
East Tower F z
1.00
0.90
0.80
0.70
0.60
0.50
0.40
0.30
0.20
0.10
0.00
2.85 4.60 5.78 7.45
Frequency (Hz)
149
Chapter 6 Spectral Analysis
W e s t Tower F y
1.00
W e s t Tower F z
1.00
0.90
0.80
0.70 +
0.60
E 0.50
0.40 i
0.30
0.20
0.10
0.00
2.85 4.60 5.78 7.45
Frequency (Hz)
150
Chapter 6 Spectral Analysis
To further understand the relation between the responses o f the towers, the response o f
the conductor and the generated forces at the conductor connections, the acceleration
responses o f the conductor are investigated. Examples o f the power spectra obtained from
acceleration response o f the conductor at three locations on the conductor are shown in
Figure 6.3.1. A s can be seen, these results show that the significant peaks are at the same
resonance frequencies o f the system in detail and deals with finding out why there is a
peak at a particular frequency in the power spectra obtained from force and acceleration
151
Chapter 6 Spectral Analysis
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
0 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
152
Chapter 6 Spectral Analysis
each location o f the conductor was observed and therefore in order to estimate one value
for the peak accelerations along the conductor, these two values at each location were
averaged. In order to justify using the averaged number, this section presents the cross
spectrum and coherence between the normal accelerations at East North and East South
sides o f the conductor, obtained from test 29 o f set-up 1 (Figure 6.3.2). Results show that
within the frequency range o f interest, these two measurements are reasonably in phase
(reasonable l o w phase angle between them). Thus it is reasonable to take their average as
a single value representing the acceleration o f the conductor at that location. Similar
results were obtained for the accelerations at the other location o f the conductor.
153
Chapter 6 Spectral Analysis
^ io it
At A y
A. *
•a
n
M k
.3 1 ft fir
C 7 "|J
% m ft
rT
Oil
I 0 1
to An
g 0.01
1$ VWI
I MO"
-
3
£l-10_
| MO 5
0 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
o —
* Frequency (Hz)
cu 1 -10
-4
t % A-%
f- w u Mi
10
t
A/
1
0.1
V
JV 77T7
0.01
3
1 1 10
"MO
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
a- 150
<
E
§ 50
1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
Figure 6.3.2 East North and West North accelerations o f the conductor - Test 29
154
Chapter 6 Spectral Analysis
The method explained i n Chapter 3, is used to identify the vibration modes o f the test
specimen. West tower was instrumented with two accelerometers on its top, which
measured the horizontal absolute accelerations at the sides o f the top plates. The
responses measured by these two transducers were used to identify the modes associated
to the peaks observed in the power spectra o f the results. East tower was instrumented
with one accelerometer for test set-up 1 and stand alone set-up. Another accelerometer
was added to the tower o f the test set-up 3 to obtain more information on the coupled
Figures 6.4.1 to 6.4.6 show the power spectra, the cross spectrum magnitude, the cross
spectrum phase and the coherence function for the acceleration outputs o f the towers
which were obtained from various test set-ups subjected to sinusoidal loading. The power
spectra o f the two signals are plotted together. For stand alone set-up and set-up 1 only
one power spectrum is estimated for the East Tower since it had only one accelerometer
A summary o f the analysis is given in Tables 6.4.1 to 6.4.3. These tables show the
resonance frequencies which are the frequencies at which both the power spectra and the
cross spectrum magnitude peak. They also show the values o f the phase and coherence
For example in Figure 6.4.3 which shows the spectra o f the two acceleration response of
the West Tower from the set-up 3, a very strong peak at 2.85 H z can be identified with
natural frequency o f the tower in its first transversal mode. The second peak is at 5.70 H z
with 136 degree phase between the two measurements and has a coherence value o f 0.56.
The values o f the phase and coherence indicate that there is no clean normal mode at this
frequency and there may be a coupled mode with coupling between transversal and
torsional modes. The l o w value o f coherence is most likely due to the nonlinearity in the
relation between the measurements and not because o f the existence o f significant noise
155
Chapter 6 Spectral Analysis
in them. This nonlinearity could be explained by the effect o f the extension cantilever
beams and the complex modes that they could introduce in the response o f the towers.
The third peak is at 6.34 H z and has a relative phase angle o f about 10 degrees and a
coherence value o f 0.85. This mode is reasonably acceptable as a clean normal mode.
The fourth resonance frequency is 8.55 H z . The phase angle between the two output
signals is 175 degrees and their coherence is 0.98. This mode is an example o f a torsional
mode for the West Tower. Generally, what can be seen in Figure 6.4.3 is a clean response
of the tower in its normal longitudinal mode to frequencies from 0.5 to 5.5 H z , and not-
the East Tower for the same test set-up, as illustrated in Figure 6.4.4, show a better
behaviour o f the tower i n its normal mode as a single degree o f freedom system. This is
because the tower has less o f a tendency to respond in its torsional mode due to its
concentrated mass. Note that there is a low amplitude peak at 5.7 H z , which is the effect
of the excitation o f the West Tower carried to the East Tower through the conductor.
The results o f the spectral analysis o f the stand alone towers are given in table 6.4.3.
These results identify the vibration modes o f the towers without any influence of
conductor on them and they show the natural frequencies o f the towers. Figure 6.4.5
shows six peaks in the power spectrum o f acceleration response o f the West Tower, four
of which match the resonance frequencies identified for other set-ups. The reason for too
many peaks for the west tower in stand alone set-up is the existence o f coupled modes
with frequencies close to each other. Since these peaks have very l o w power, they might
not show up when the towers are interconnected with the conductor.
Finally, there is a peak at 20 H z in all the spectra. This peak presents nothing but noise
since it is in the spectra o f all the measured signals. The source o f the noise is i n the
156
Chapter 6 Spectral Analysis
v
100
1
10
1 / {
0.1
0.01 4-
3 *
-1 10_
1 10
1 10
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
150
100
50
0 1 2 3 4
Ji
5 6 7
m
8 9 10 11 12 13 14 15
'
16
- "
17 18
rWfrl
19 20
A.
21 22 23 24 25
Frequency (Hz)
Y
iff*?
0.5
0 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
157
Chapter 6 Spectral Analysis
A,, J 1
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
158
Chapter 6 Spectral Analysis
4 $
3
£ 1 10
—1
11 io -
—1
• 1 10 £5 SjBE •til*.
fcAr
4y
1 10
£ 1 10
0 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
100
10
1
s 0.1
E
0.01
£M0_
|M0" 1 I
V 110
o 1
8 9 10 11 12 13
Frequency (Hz)
14 15 16 17 18 19 20 21 22 23 24 25
150
100
50
0 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
159
Chapter 6 Spectral Analysis
i
East Tower Acceleration Outputs
100
10
ll
1
0.1
0.01
2
| '' _10
1 M0_ fill
1 WW
aid r A*
£ M 0 _
| M0_
S. i io
0 1 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
100
10
1
0.1
0.01
°-3 0 1
=H
11 io
6 1 -10
! 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
100
ill
M4
6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
160
Chapter 6 Spectral Analysis
w M0_
| 1 -10
<£ i io 0 1 2 3 4 5 6 7 9 10 11 12 13
Frequency (Hz)
14 15 16 17 18 19 20 21 22 23 24 25
/JV
100
1
10
1
c 1
u
0.1
0.01 IH
3 *****
£M0_
| 110
6 MO"
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
a. 150
100
1
50
0
A 1 2 3
_<*ut
4 5
iM 6 7
AM
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
0 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
161
Chapter 6 Spectral Analysis
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Frequency (Hz)
Set-up 1
Resonance Relative
Equivalent
Frequencies Phase Coherence Description o f the Mode
Equipment
(Hz) (deg)
162
Chapter 6 Spectral Analysis
Set-up 3
Resonance Relative
Equivalent
Frequencies Phase Coherence Description o f the Mode
Equipment
(Hz) (deg)
Stand Alone
Resonance Relative
Equivalent
Frequencies Phase Coherence Description of the Mode
Equipment
(Hz) (deg)
East
3.53 0 1.00 Normal Mode - Longitudinal
Tower
163
Chapter 6 Spectral Analysis
The analysis o f the stand alone towers revealed whether any o f the resonance frequencies
identified for the towers i n set-ups 1 and 3 was caused by the response o f the conductor at
its resonance frequencies. This was especially the case for the west tower which showed
numerous peaks in its power spectrum. Distinguishing the resonance frequencies o f the
towers from those o f the conductor provides a better insight into the behaviour o f the
conductor and its interaction with the towers. The comparison o f the identified resonance
frequencies o f the stand alone towers with those o f the interconnected towers from other
set-ups and with those o f the conductor, reveals that all the identified resonance
frequencies o f the conductor coincide with the natural frequencies o f the towers. Hence
the resonance frequencies identified for the conductor are not independent from the
vibration modes o f the towers. The comparison o f the resonance frequencies o f the
towers and the conductor with the frequency contents o f the force responses shows that
responsible for the horizontal dynamic forces (Fz) generated at its attachment points.
modes has a considerable effect in producing vertical forces (Fy) at its connections.
Although the performed analysis did not identify a specific natural frequency for the
conductor due to the fact that the conductor does not have a well-determined natural
164
Chapter 7 Comparison of Results and Discussion
CHAPTER 7
7.1 Introduction
In this chapter the results from various tests are compared, including comparison o f
forces obtained from quasi-static and dynamic tests, comparison o f the response o f the
A s can be seen i n Section 5.9.4.4, for test set-up 3 the horizontal forces in the connections
of the conductor (Fz) reached a peak value o f 1327 N at East Tower (Table 5.9.11). This
force corresponds to a displacement o f less than 30 m m for East Tower. The horizontal
force at the same connection point o f the conductor for the same level o f displacement
obtained from the quasi-static test (Chapter 4), is no more than 55 N which is 24 times
less than the dynamic force o f 1327 N . Similarly for test set-up 1, the comparison would
corresponding to a displacement o f 40 mm, which shows that the dynamic force obtained
from seismic tests is about 16 times more than the same force obtained from the quasi-
static test. These are substantial differences and show that the magnitude o f forces
generated in the connections o f the conductor due to dynamic behaviour o f the conductor
are much higher than the forces obtained by just considering its static behaviour. This
finding indicates that it is not sufficient to determine the forces in the connections o f the
Rather, the dynamic behaviour o f the conductor must be taken into account in order to
earthquake event.
165
Chapter 7 Comparison of Results and Discussion
To find the effect o f the conductor on the response o f the towers, the acceleration and
displacement response o f the towers obtained from stand-alone towers set-up and set-up 1
are compared. Since the peak acceleration o f the shake table were not exactly identical
for the two set-ups, instead o f the absolute accelerations themselves, the acceleration
amplification factors (Tables 5.6.6, 5.6.7, 5.7.4 and 5.7.5) are used to compare the
acceleration amplification factor o f the towers obtained from set-up 1 to that o f obtained
from the stand-alone set-up. Displacement Response Ratio for each tower is defined as
the peak relative displacement o f the tower from set-up 1 to that o f stand-alone towers
set-up. Tables 7.3.1 to 7.3.4 show the results. These results show that i n most cases, the
conductor reduces the response o f the towers. However, i n one case, the response o f the
East Tower to IEEE-Landers record, the conductor amplified the response o f the tower.
This indicates that there is a possibility that interconnected equipment would have a more
severe response compared to its stand-alone response. It also indicates the effect o f the
Input Signal
Test Set-up
IEEE-Northridge I E E E - E l Centro IEEE-Landers
90% 90% 75%
166
Chapter 7 Comparison of Results and Discussion
Input Signal
Test Set-up
IEEE-Northridge I E E E - E l Centro IEEE-Landers
90% 90% 75%
Input Signal
Test Set-up
IEEE-Northridge I E E E - E l Centro IEEE-Landers
90% 90% 75%
Atand-Alone Towers 52.00 48.00 34.00
Set-up 1 40.00 35.00 42.00
Displacement Response
0.77 0.73 1.24
Ratio
Input Signal
Test Set-up
IEEE-Northridge I E E E - E l Centro IEEE-Landers
90% 90% 75%
Stand-Alone Towers 66.00 63.00 59.00
Set-up 1 52.00 62.00 55.00
Displacement Response
0.79 0.98 0.93
Ratio
167
Chapter 7 Comparison of Results and Discussion
To study the effect o f the frequency difference between the two towers on the magnitude
of forces, tests on set-up 3 were carried out. In this section, a comparison is made
between the results obtained from test set-ups 1 and 3 for different levels o f the same
input records. Table 7.4.1 shows the results, where the Force Amplification Factors
( F A F ) from the two set-ups are compared. The " % Change" indicates increase or
decrease o f F A F obtained from set-up 3 with respect to that obtained from set-up 1. The
Landers. It decreased down to 13% when the input record was I E E E - E l Centro.
• F y increased up to 30% when the input record was IEEE-Landers and decreased
never increased.
• F y increased up to 20% when the input record was IEEE-Landers and decreased
East Tower or F y at any one o f the towers, although the East Tower in set-up 3 had
higher natural frequency and less mass compared to the East Tower i n set-up 1. This
indicates that the frequency domain characteristics o f the input records are not sufficient
to predict the response o f the system and indicates the importance o f the time-domain
characteristics o f the input i n the force response o f the conductor at its connection points.
168
Chapter 7 Comparison of Results and Discussion
Table 7.4.1 does not show a very significant change in the F z at the West Tower. This
finding was expected since it was found that the response o f West Tower has the major
169
Chapter 7 Comparison of Results and Discussion
Set-up 2 was tested to include the transverse vibration o f the conductor in this
investigation and in order to study the magnitude o f forces when the input motion is
bidirectional. T o compare the results o f these tests with those obtained from set-up 1,
resultant horizontal forces were calculated for the cases when input signals were at 35%
amplitude in both E - W and N - S directions (Section 5.8.4). These input signals were
equivalent to the same records being applied at 50% magnitude at an angle o f 45° with
the E - W direction. A comparison the resultant horizontal forces from these tests with the
horizontal forces obtained from test set-up 1 subjected to the same input signals at 50% in
motions. A summary o f this comparison is shown in Table 7.5.1. which shows that both
horizontal and vertical forces obtained from bidirectional tests are significantly lower
than those obtained from set-up 1. This observation indicates that for the given
earthquake records at the given magnitude, the highest magnitude o f forces generated at
the connections o f the conductor to the equipment were obtained when the direction o f
motion was along the direction o f the conductor. Thus it is not necessary to consider
further the bidirectional horizontal ground motions as far as critical loading conditions
are concerned.
170
Chapter 7 Comparison of Results and Discussion
171
Chapter 8 Conclusions and Recommendations
CHAPTER 8
CONCLUSIONS AND RECOMMENDATIONS
8.1 Summary
They have shown to be vulnerable despite the fact that the flexible conductors have
enough slack which allows the equipment to experience displacement without facing any
significant resistance from the conductor. The flexible conductors do not apply
significant load on the equipment i f the behaviour o f the system is analysed statically.
However, it has been observed that the conductor applies much larger forces on the
equipment due to its significant dynamic response under seismic loadings. Since these
forces could be one o f the reasons for the failure o f the equipment, it became necessary to
investigate the dynamic behaviour o f the flexible conductors and their interaction with
the equipment and estimate the magnitude forces induced by the conductor on the
equipment under anticipated earthquakes. This research also investigated the possibility
of the amplification o f the response o f the equipment due to presence o f the conductor
designed for.
In this study, the mentioned investigation was carried out through a series of
configuration o f conductors and four series o f shake table tests on large-scale equivalent
models o f the equipment and the full-scale flexible conductor. The input records used for
the shake table tests were synthetic records generated from three different earthquake
records to fit prescribed response spectra. The quasi-static test provided information on
the static behaviour o f the conductor. The dynamic tests included uni-axial (horizontal)
shake table tests on the stand-alone equivalent equipment as well as uni-axial tests on two
set-ups for the interconnected ones. A series o f bidirectional tests was also included. The
172
Chapter 8 Conclusions and Recommendations
comparison o f the results obtained from stand-alone set-up with those obtained from
into the magnitude o f forces generated at the connections o f the conductor to the
equipment under seismic loading. The results o f dynamic tests provided insight into the
dynamic behaviour o f the conductor and its interaction with the equipment either directly
8.2 Conclusions
• The presence o f the conductor could decrease or amplify the response o f the
record. This means that the interconnected equipment can experience higher
inertia forces compared to its stand-alone state. It should also be realized that
characteristics other than frequency content and peak acceleration o f the ground
statically. Thus the dynamic interaction o f the conductor with the equipment
• The horizontal dynamic force response at the connection o f the conductor to any
the same equipment item, while the vertical force response is dependent on the
173
Chapter 8 Conclusions and Recommendations
seismic tests because horizontal and vertical forces generated i n the connections
o f the conductor under bidirectional seismic loadings did not exceed the forces
obtained when the uni-axial loadings with the same magnitude were used.
• The study on the interaction o f the conductor with the towers showed that the
although did not show any obvious natural frequency for the conductor at which it
interaction between the conductor and the equipment and the dependence o f the
equipment.
• The methodology used i n this study can be used as a guideline for studying the
8.3 Recommendations
• T o study the effect o f vertical seismic loading on the behaviour o f the conductor
comprising o f vertical table motion combined with the horizontal one i n the
system with two inputs and multiple outputs could give more information about
174
Chapter 8 Conclusions and Recommendations
conductor and its interaction with the equipment are recommended. The
experimental data o f this study can be used for calibrating the numerical models.
175
REFERENCES
and Spectral Analysis. Second Edition. Copyright 1993 by John W i l e y and Sons,
Inc.
1996. pp.801-807.
176
9. I E E E Std 693-1997, I E E E Recommended Practice for Seismic Design o f
Engineers, Inc.
10. IES Design, Test, and Evaluation Division Recommended Practice 012.1:
11. Lathi, B . P. Signal Processing and Linear Systems. Copyright 1998 by Berkeley-
Cambridge Press.
13. Okada, T., et al. "Seismic Design o f Connecting Leads in Open-air Type
177
APPENDIX A
INSTRUMENTATION DETAILS
Transducer Type: Accelerometer
Brand Name: F C Sensor M o d e l 3140-005
Serial Number: 3B 12743
Sensitivity: 400 m V / g
Location Used: West Tower; E - W Direction
Transducer Type: L o a d C e l l
Brand Name: A M T I M o d e l M C 5
Serial Number: M3253
Sensitivity: 1.894 m V / V - l b f
Location Used: East Tower F y
Transducer Type: L o a d C e l l
Brand Name: A M T I M o d e l M C 5
Serial Number: M3253
Sensitivity: 0.501 m V / V - l b f
Location Used: East Tower F x
Transducer Type: Load C e l l
Brand Name: A M T I M o d e l M C 5
Serial Number: M 3 2 5 4
Sensitivity: 1.871 m V / V - l b f
Location Used: West Tower F y
Transducer Type: L V D T
Brand Name: T R A N S T E K Model 243-000
Serial Number: 2360
Sensitivity: ±0.5 i n @ 7.5 V
Location Used: East Extension Cantilever Beams
Transducer Type: L V D T
Brand Name: T R A N S T E K M o d e l 243-000
Serial Number: 2359
Sensitivity: ±0.5 i n @ 7.5 V
Location Used: West Extension Cantilever Beams
DRAWINGS
Geometry of the Conductor and Location of the Accelerometers
NAME: 5CB
(East Tower)
4 BOLTS
FOR
FASTENING
ONLY
WELDING
DIRECTLY
TO
BOTTOM
LAYER
118.11" LONG
2.756" DIAMETER
SOLID ROUND BAR
WITH MIN YIELD
STRENGTH >=
ASTM A29
GRADE 1040
NAME: 5CT
(West Tower)
l !•
---tr —
o- |" o-
i 1
e- h -e-
'i
• ©- II o S6f o 1:
o
! II H
-fc T" " i '
"i I
! IJ- I
G- o ft-'-©-:
•o- I 1
6 e- 11
-CM
i II !
. I
1
!
•©-
ii
•©•-• 11 o- -©- j.
I- &
&
187
6&»H S i
| - 3
-o- O !|: O
-6- -O
H ,.. -I H
sea
4- ) o
i-0
PLATE PZ
-e- I -e-
r-
S 8 2
4-
2- F ? g Q t > ~ <5&>\
= \o4-o
i i 1 i 1 I 1
r | i j i- i
1110
2- LA3ISO x 3 7
FIELO B O L T S
SOLID R O C K STEEL
PAINT: ONI COAT OF
CUSTOMER U &.-.C
CHECKED BY
0«T8j^i
DATE
\2./o.) CONTRACT ORG. No.
I3>57
p> q & c e s & ftp
- \oAo
A
t==t= 1 III 11
13 (e OP
' 15.3^-
7. R g o ; p ~ e> 1
90] 44o Wo
lO 2J5o * 4-°\
a z
A: ffe 2o2> * \b 0
FIELD BOLTS
MATERIAL: /^SOW .'.
HOLES: "fej ^ UNLESS NOTED
1
SOLID R O C K STEEL
PAINT: ONI COAT OP
CUSTOMER 0,&,c
HAIIIOPSTt
S t LOCATION
A muopoBAwmo BCAfX'S
A DRAWN By ^ 0ATE^M\^/O>
CONTRACT ORG. No.
A CHECKED BY o«TE
1397 f -
192
•^3
OOP- ~ E f e l C z a p x g,^
QUE ~ £€> 2 -
0«£
2-5o II ID 56
FIELD BOLTS
MATEWAU. .£>GQ. W.
SOLID R O C K STEEL
" X
HOLES: .C?~. UNLESS NOTED
PAINT: O N E C O A T O F
CUSTOMER !J . • C>
X / ^^A£0£j£- ""T^V^l^E-
„2C. NAME 0* STRUCTURE
LOCATION
A E5
APPENDIX C
Shake Table
1
t
4 4
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
East Tower
0 2 4 6 8 10
J 1
12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
West Tower
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
ii * •ID M
-j I I I I I I I I 1 I I I I I I I ! I I ! I
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
195
West North Side - Normal to Conductor
V
rWw
1
YVWVI www WWv
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
VV "flr
vv
p WW
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
-•7 I 1 ! 1 1 1 1 1 1 I I I ! I ! ! ! ! ! I I
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
Shake Table
12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
196
West Tower
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
East Tower
mmmm
ill WWW
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
Set-up 1. Test 33
Shake Table
1 4" 4
0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60
Time (sec)
East Tower
0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60
Time (sec)
1
West Tower
N H m
N VIM 4 II i i
0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60
Time (sec)
0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60
Time (sec)
1
1
N M •ILuiLv H\W\\\
k'l\]k\ tAUlnll
ITPlTF
UHuulAlu AIUHJUJI
1 1
0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60
Time (sec)
Centre North Side - Normal to Conductor
MllldL
• Wl IN Wtf
llUrVtu
(tr'fftP
r^»T^^, TfMi
.7 I 1 1 I I I I 1 I I ! I I I 1 1 1 1 1 1
0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60
Time (sec)
1
_y I I I I I I I 1 I I I ! 1 I I I ! I I ! I
0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60
Time (sec)
Shake Table
flA w
, A /I
v v
A ~ i\f\
0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60
Time (sec)
West Tower
0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60
Time (sec)
199
East Tower
Set-up 3, Test 7
Shake Table
null J\ m
ft Hi
-j I 1 1 1 1 1 1 1 I I I I I I ! ! I I I 1 I
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
200
East Tower
. ^ M W
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
West Tower
Shake Table
0 2 4 6
West Tower
100
I 50
,1,1
c
o
e o
o
,.,111111
MM m • w mAIM
C3
CL
a "50
-100
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
201
East Tower
5 -so
B II • HI mm mm mmt www
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
Nl
1 III illlll
It i in Mi WW
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
202
Centre North Side - Normal to Conductor
\w\fr\
i #1
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time (sec)
41 WW WWWW WW
-7 1 1 1
' ' 1 1 1
' 1
' ' 1 1 1 1 1 1 1 1 1
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Time [sec]
Set-up 3. Test 9
Shake Table
— —
"An
_] 1 ! 1 1 I I I I I I I ! I ! 1 1 ! I I I I
0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60
Time (sec)
203
East Tower
0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60
Time (sec)
West Tower
—f«W
ft H 4 HiIi 41i i •
0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60
Time (sec)
iW ft 4W tfl It m
0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60
Time (sec)
H| I i i •
0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60
Time (sec)
204
Centre North Side - Normal to Conductor
TTPTTT
0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60
Time (sec)
www WW)WW
_y I 1 ! i I I I I ! I I 1 I ! 1 ! ! 1 1 ! I
0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60
Time [sec]
Shake Table
l\T"AIT fAfi
0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60
Time (sec)
West Tower
IIti mm m
0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60
Time (sec)
205
East Tower