Anda di halaman 1dari 14

EARTHQUAKE ENGINEERING AND STRUCTURAL DYNAMICS

Earthquake Engng Struct. Dyn. 2003; 32:151–164 (DOI: 10.1002/eqe.224)

Comparative study of the inelastic response of base


isolated buildings

Daniel Ordoñez1 , Dora Foti2; ∗; † and Luis Bozzo3


1 Universidad Politecnica de Catalunya; Barcelona; Spain
2 Politecnico di Bari; Dipartimento di Ingegneria Civile e Ambientale; Via Orabona 4; 70125 Bari; Italy
3 Universidad de Girona; Girona; Spain

SUMMARY
This article presents a numeric comparative study of the inelastic structural response of base isolated
buildings. The comparative study includes the following isolation systems: laminated rubber bearings,
New Zealand one, pure friction and the frictional pendulum ones. The study is based on obtaining non-
linear response spectra for various design parameters using six earthquake records. Usually the base
isolation of a new building seeks to maintain the structure in the linear elastic range. The response of old
weak buildings or the response of new ones subjected to extreme earthquakes may not be, necessarily,
in the aforementioned ideal elastic range. Consequently, it is important to characterize the response of
isolated buildings responding inelastically. A conclusion from this research is that the isolators aect
signicantly the structural response of weak systems. Rubber isolators seem slightly less sensitive to
plastication that may occur in the structure compared to friction isolators. Ductility demands in the
structure are aected signicantly by friction and neoprene protected systems, in particular sliding ones
where larger demands are obtained. Copyright ? 2002 John Wiley & Sons, Ltd.

KEY WORDS: base isolation; passive control; sliding isolation; hysteretic isolation; friction isolation;
neoprene isolation; structural strengthening; earthquake retrotting

1. INTRODUCTION

The currently accepted design criteria for base isolated buildings is to maintain the supported
structure in the linear elastic range. There are many studies focussed in the assumption of
a linear structural response [1]. There are situations, however, such as the retrotting of old
historical buildings or the design of new ones under severe ground motions, where the acting
forces may exceed the design ones. Besides long term response of rubber and friction isolators

∗ Correspondence to: Dora Foti, Politecnico di Bari; Dipartimento di Ingegneria Civile e Ambientale; Via Orabona
4; 70125 Bari; Italy.
† E-mail: d.foti@poliba.it

Contract=grant sponsor: Spanish Government; contract=grant number: PB98-0455.


Received 10 October 2001
Revised 19 April 2002
Copyright ? 2002 John Wiley & Sons, Ltd. Accepted 29 May 2002
152 D. ORDOÑEZ, D. FOTI AND L. BOZZO

are not fully understood and changes in physical characteristics may occur with time under a
variety of environmental conditions. Consequently, it is important to characterize the response
of inelastic structures supported on base isolation systems identifying general trends and also
possible anomalous responses. Moreover, this study may give the possibility to understand the
mechanisms in order to dissipate energy in the structure and not only in the base, resulting
in more economic designs and more robust structures.
The article presents a comparative study of the inelastic structural response of structures
supported on four base isolation systems. The systems are the pure friction and friction pen-
dulum ones—corresponding to sliding base isolation—and the laminated rubber bearings and
New Zealand systems—corresponding to rubber base isolation. Another comparative study but
based on the aforementioned linear structural response assumption has been developed by Su
et al. [2; 3].
In a previous work [4] the inuence in the variations of the friction coecient during
sliding for inelastic structures has been studied giving simple design guidelines. In this work
a comparative study for a variety of isolators is accomplished, also for structures that re-
spond inelastically. The dynamic structural performance is characterized by inelastic spectra
for base displacements, structural drifts, base shears and ductility demands for various struc-
tural strength coecients. The study includes six seismic records corresponding to sti and
medium local soil conditions.

2. DYNAMIC MODELLING

The equation of motion for a multi-degree-of-freedom structure supported on base isolators


responding in the linear range is

M X + C Ẋ + KX = −MJ [db + a] (1)

where M is the mass matrix, X is a vector representing the story displacements relative to
the base (above the isolators), C; K are the structural damping and stiness matrices, J is
the vector that relates the rigid body displacements to the model degrees of freedom and db
is the base displacement with respect to the ground. Figure 1 illustrates the dynamic model

(a) (b)

Figure 1. (a) One-degree-of-freedom structure with base isolation. (b) Dynamic model.

Copyright ? 2002 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2003; 32:151–164
THE INELASTIC RESPONSE OF BASE ISOLATED BUILDINGS 153

for a one-degree-of-freedom structure. The forces due to the structural stiness and damping
applied to the base can be obtained from Equation (1), as
JC Ẋ + JKX = −JM X − JMJ [db + a] (2)
Consequently, the equation of motion for the base can be represented by
mb (db + a) + JM [X + j(db + a)] + f = 0 (3)
where mb is the mass just above the isolator and f is the isolator force applied to the base
mb . The nature and mathematical expressions for the force f depends on the isolator type
and they are briey included in the following sub-sections for the isolators considered in this
work.

2.1. Pure friction (PF) isolator

The pure-friction system is the simplest base isolator and corresponds to a building supported
on sliding connections. A more detailed description of these isolators and all the ones in-
cluded in this article may be found in Reference [5]. The force f at the sliding connection
corresponding to a pure Coulomb constant friction base isolator and assuming that the sliding
surfaces are always in contact, is rewritten as
f =  g mtot sign(d˙b ) (4)
where  is the friction coecient—typical values range from 0.05 and 0.3, g is the acceler-
ation of gravity, mtot is the total mass above the isolator (J T MJ + mb ) and db is the sliding
displacement. Equation (4) gives the force in the sliding phase. If the base mass sticks to
the foundation a non-sliding condition must be veried [2]. The friction coecient  varies
signicantly with the nature of friction surfaces, the relative velocity, and the axial force at
the connection [6].

2.2. Friction pendulum system (FPS)

The force f for the base corresponding to a building supported on ideal friction pendulum
system (FPS) connections and assuming that the sliding surfaces are always in contact is
f =  gmtot sign(d˙b ) + kb db (5)
where kb is the eective sliding lateral stiness [7]. This additional restoring force is provided
by metallic springs or by the curvature of the sliding surface in the FPS-system. The sliding
lateral stiness is useful to dene the period of the connection, corresponding  to the period
of a perfectly rigid structure
 sliding at the base and it is given as T c = 2 mtot =kb . For the
FPS system, Tc = 2 r=g, where r is the radius of curvature of the sliding surface.
The aforementioned Equation (5) describes the response of the system in the sliding phase.
Similarly to the pure friction system, if the base mass sticks to the foundation, a non-sliding
condition must be veried. The friction coecient and period of the connection in this system
vary between 0:05¡¡0:15 and 2¡Tc ¡3s, respectively. For this study the friction coecient
is considered as 0.05 for both the PF and FPS systems.

Copyright ? 2002 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2003; 32:151–164
154 D. ORDOÑEZ, D. FOTI AND L. BOZZO

2.3. Rubber bearings (LRB) isolator

The force f for the base of a building supported on rubber bearings (LRB), or viscodampers
and excited by an horizontal earthquake ground acceleration can be represented by an equiv-
alent damping (cb ) and stiness (kb ) coecients [1]. The equivalent linear system for the
isolator enables a simple numerical solutionof the problem. A common rubber bearing de-
sign period is about Tb = 2 s (where Tb = 2 mtot =kb ). The equivalent damping coecient of
the rubber varies considerably. For low strain, it is as high as about  = 0:3 (where  = cb =2! b )
but for high strain it is as low as about 0.05. A common value used for design is  = 0:1.
Alternatively a more precise representation of this system can be obtained using the Wen’s
hysteretic model [8]. Accordingly, the expression for the restoring force f(t) in a hysteretic
damper is
f(t) =  kb db + (1 − )kb z (6)
where  is a dimensionless parameter that indicates the degree-of-non-linearity of the system
(for example  = 1 indicates a linear system) and z is a dimensionless hysteretic parameter
satisfying the following non-linear rst-order dierential equation:
ż = Ad˙b − |d˙b ||z |n−1 z − d˙b |z |n (7)
The parameters A; ;  and n are the model parameters which can be rearranged in terms of
the parameters dy and fy which are the yield displacement and force of the hysteretic damper.
In this way the parameters have a more structural meaning to be manipulated. The exponent n
is an integer that controls the smoothness of the transition from elastic to plastic response and
 is the post to pre-yielding stiness ratio. The values fy = 16:5 kN, dy = 0:0015 m,  = 0:3,
 = 66:66,  = 600, A = 1:75 and n = 5 were selected for this study.

2.4. New Zealand (NZ) isolator

The New Zealand (NZ) system is similar to the laminated rubber one but it includes a central
lead core in order to reduce base displacements and to provide an additional source of energy
dissipation. The force–displacement relationship for this system can be represented as well
using the Wen’s model. The values fy = 0:185 kN, dy = 0:031 m,  = 0:157,  = 4:29,  = −1:93,
A = 0:11 and n = 1 were selected for this study.

3. STRUCTURAL MODEL AND SEISMIC ACTION

The dynamic model has two degrees of freedom as illustrated in Figure 1. The rst d.o.f.
corresponds to the isolator and the second d.o.f. corresponds to the structure. The mass above
the isolators as well as the mass at their level are considered. The mass at the isolators’ level
is 3400 kg and above them is 34; 300 kg. The damping factor of the structure is assumed as
 = 5%. The inelastic response of the structure is modelled using, as well, the Wen’s model.
The values of fy = 250 kN, dy = 0:00778 m,  = 0:0843,  = 102:95,  = 102:95, A = 1:75 and
n = 1 are considered. These values give an initial stiness equal to 250 kN= m. In the next
section and for the rst three sub-sections, the strength fy is selected as 35% larger than the

Copyright ? 2002 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2003; 32:151–164
THE INELASTIC RESPONSE OF BASE ISOLATED BUILDINGS 155

sliding threshold force for the friction isolation systems. The last sub-section of next section
presents a comparison considering variations in the strength coecient.
The comparative study considers inelastic spectra for a period in between 0:1 s6T 61:5 s.
These periods are obtained changing the stiness and keeping the mass constant. The structural
resistance is modied varying the base strength coecient
 = Vy =(Mtot a) (8)
where Vy is the limit shear in the structure, Mtot is the total mass and a is the base ac-
celeration. Varying the period and stiness of the structure is accomplished, as indicated
previously, varying the Wen’s parameters. For this study six records corresponding to the
Loma Prieta 1989, Northridge 1994 and Kobe 1995 earthquakes have been considered. For
Loma Prieta the components Gilroy 90◦ and 0◦ are considered with a maximum acceleration
of amax = 433:616 cm= s2 and amax = 426:609 cm= s2 . For Northridge earthquake the Santa Mon-
ica and Sylmar components are considered with a maximum acceleration amax = 865:96 cm= s2
and amax = 826:76 cm= s2 . For Kobe earthquake the East–West and North–South components
are considered with a maximum acceleration of amax = 617:28 cm= s2 and amax = 818:0 cm= s2 .
The rst four records are considered representative of sti local soil conditions while the last
two are considered of soft local soil conditions (besides the clear dierence in the source
mechanism that generates the earthquakes). The peak acceleration for the six records was
scaled to 0:5g, where g is the ground acceleration. The limited number of records does not
allow extrapolating design guidelines, although, in the authors’ opinion, it points out gen-
eral trends. It also illustrates the strong inuence of the structural strength in the response
for individual records. A robust earthquake resistant design for essential facilities and his-
torical buildings must be satisfactory not only for average values but also for individual
cases.

4. PARAMETRIC STUDY

4.1. Base displacement

Figures 2(a)–2(c) show the base displacement for buildings with base isolation responding
inelastically. Plots in Figures 2(a)–2(c) correspond to numerical results for the Gilroy 90◦ ,
Sylmar and Kobe E–W records which are considered representative of this study. The ideal
elastic response is included in the plots just as a reference and not for comparison purposes.
The strength was considered xed and 35% larger than the sliding threshold force for the
friction isolation systems yielding a structural strength coecient  = 0:16. Base displacement
is a relevant design parameter because it denes the separation joint between the foundation
and the ground. In ve base isolated buildings subjected to Northridge earthquake apparently
a base impact was observed in two of them [9].
In general, it is noticed that the curves are relatively smooth and decreasing, especially
those for neoprene isolators. For these systems and considering only Loma Prieta records
(Figure 2(a)), it is possible to notice a clear dierence between the base displacements ob-
tained with NZ and LRB systems. Base displacements measured in medium soil for Kobe
and Northridge records are not very dierent for buildings with NZ and LRB systems. This
observation is signicant because it seems to indicate a dependence between the frequency

Copyright ? 2002 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2003; 32:151–164
156 D. ORDOÑEZ, D. FOTI AND L. BOZZO

0.14
0.35
LRB
LRB
0.12 NZ
0.30 NZ
FPS
FPS
0.10 PF

Relative displacement (m)


0.25 PF
Elastic
Elastic
Displacement (m)

0.20 0.08

0.15 0.06

0.10 0.04

0.05 0.02

0.00 0.00
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
(a) T (s) (d) T (s)

0.50 0.5
LRB LRB
0.45
NZ NZ
0.40 FPS 0.4 FPS
PF PF
Relative displacement (m)

0.35
Elastic Elastic
Displacement (m)

0.30 0.3

0.25

0.20 0.2

0.15

0.10 0.1

0.05

0.00 0.0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
(b) T (s) (e) T (s)

0.35
0.35
LRB LRB
NZ 0.30 NZ
0.30
FPS FPS
PF PF
0.25
Relative displacement (m)

0.25 Elastic Elastic


Displacement (m)

0.20 0.20

0.15 0.15

0.10 0.10

0.05 0.05

0.00 0.00
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
(c) T (s) (f ) T (s)

Figure 2. (a) – (c) Base displacement for Gilroy 90◦ , Sylmar and Kobe E–W records, respectively.
(d) – (f) Structural drift for Gilroy 90◦ , Sylmar and Kobe E–W records, respectively.

Copyright ? 2002 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2003; 32:151–164
THE INELASTIC RESPONSE OF BASE ISOLATED BUILDINGS 157

content of the earthquake and the ecacy of the energy dissipation in the lead core of NZ
system to reduce base displacement, such as is illustrated from Figure 2(a) compared to Fig-
ures 2(b) and 2(c). Maximum displacement varies signicantly with the records: for Loma
Prieta records it is 16 cm while for Sylmar record it is 35 cm.
It is clearly shown in the gures that, except for Kobe record, utilizing friction systems
base displacement is small compared to neoprene systems. Among friction isolators, if FPS
are utilized base displacement is lower than using PF ones. With FPS system it is possible
to notice a tendency in all records to obtain linear displacements that decrease as the struc-
tural period increases. In general, for periods higher than 0:8 s base displacement with friction
isolators is lower than the displacement for conventional structures with linear elastic re-
sponse. These results are consistent with similar observations for linear superstructure response
[10].
Figure 2(c) for Kobe record shows general characteristics dierent from the ones obtained
with Loma Prieta and Northridge records. In this case, base displacements with PF system
are larger than with other systems, except for large periods where isolation is usually not
employed. This response characteristic was also found using the Kobe N–S record, indicating
the convenience of having some additional restoring force in order to prevent large lateral
displacements in friction base isolated buildings. Certainly these results cannot be directly
used for design purposes but they point out the strong inuence of the earthquake frequency
content in the response for weak structural systems.

4.2. Interstorey drift

Figures 2(d) and 2(f) show the interstorey drifts of the structure referred to the three records
dened in the previous sub-section. Again, the structural response is in the inelastic range and,
only as a reference, the displacement of conventional structures with linear elastic response is
included. It is known that the interstorey drift is an important design parameter correlated to
the damage of non-structural elements. Therefore the ecacy of an isolation system to protect
a structure can be measured with this parameter.
In general as it is the case for linear base isolation superstructure [11] the curves are
smoothly linear and increasing, for all the records and for all the considered systems. In in-
creasing order, the relative displacement corresponds to base isolated structures using the NZ,
LRB, PF or the FPS system. This observation is maintained for all the six records included
in this study, except for the Sylmar one (Figure 2(e)) where the drift using the LRB system
increases as about the response using the FPS one. This apparently anomalous response char-
acteristic could be explained by a resonance eect between the isolator period and the ground
motion. The frequency content in a Fourier spectrum of Sylmar record has the maximum
energy content for T = 3 s, having designed the connection of the structure for this period,
this causes the apparently anomalous behaviour. Consequently, this response characteristic in-
dicates again the convenience of having an additional source of energy dissipation as for the
NZ system compared to the LRB one.

4.3. Total displacement of the structure

Figures 3(a)–3(c) show the total structural displacement for the records indicated in the previ-
ous two sub-sections. In general, utilizing LRB and NZ systems the displacement is relatively

Copyright ? 2002 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2003; 32:151–164
158 D. ORDOÑEZ, D. FOTI AND L. BOZZO

0.16 0.13

0.12
0.14 LRB
0.11 NZ
0.12
Total displacement (m)

Seismic coefficient Cs
0.10 FPS

0.10 PF
0.09
UBC
0.08 0.08

0.07
0.06 LRB
NZ 0.06
0.04
FPS 0.05
0.02 PF
0.04
Elastic
0.00 0.03
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
(a) T (s) (d) T (s)

0.50 0.18

0.45
0.16 LRB
0.40
0.14 NZ
Total displacement (m)

Seismic coefficient Cs
0.35 FPS
0.12 PF
0.30
UBC
0.25 0.10

0.20
LRB 0.08
0.15 NZ
0.06
0.10 FPS
PF 0.04
0.05
Elastic
0.00 0.02
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
(b) T (s) (e) T (s)

0.35 0.13

0.12 LRB
0.30
NZ
0.11
Total displacement (m)

Seismic coefficient Cs

0.25 FPS
0.10
PF
0.20 0.09 UBC

0.15 0.08
LRB
0.07
0.10 NZ
FPS 0.06
0.05 PF 0.05
Elastic
0.00 0.04
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
(c) T (s) (f ) T (s)

Figure 3. (a) – (c) Total structural displacement for Gilroy 90◦ , Sylmar and Kobe E–W records,
respectively. (d) – (f) Seismic coecient for Gilroy 0◦ , Sylmar and Kobe E–W records, respectively.

Copyright ? 2002 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2003; 32:151–164
THE INELASTIC RESPONSE OF BASE ISOLATED BUILDINGS 159

constant and independent from the structural period. Among all systems, PF shows the biggest
oscillations depending on the period of the structure, as illustrated in Figure 3(c). The re-
sponse between structures using LRB and NZ systems is very similar, except for the Gilroy
90◦ record where the total displacement using the LRB system is about 30% larger than using
the NZ one. As previously indicated, the eciency in the dissipation in the lead core seems
very much aected by the input earthquake.
In general, the total displacement for PF and FPS friction systems is similar in shape, but
larger using the PF system. The curves in these systems are slightly irregular with very little
signicant variations in their total values. The dierence is more qualitative as the curves
do not clearly tend to increase or decrease. In general, the total displacement in structures
protected with friction systems is lower than the displacement of structures with neoprene
systems as it is the case for linear base isolated superstructure.
For example in Figure 3(b) for T = 1:2 s, the total displacement utilizing LRB and NZ
systems is approximately 30 cm, while this value is reduced to 25 and 22 cm for PF system
and FPS system, respectively.
Figure 3(c) for Kobe E–W record, shows tendencies dierent from those previously de-
scribed. First of all larger displacements appear with PF system in almost all the period range.
Regarding the N–S component not included here, the displacements for all the systems are
similar, but they highly dier for the E–W component. In the Fourier spectrum of Kobe
E–W signal the energy is concentrated at T = 0:2 s and 0:8 s, although for larger periods
the Fourier amplitude has signicant values, especially if compared to the Fourier spectrum
of the N–S component. This seems to explain the dierence in the response between both
components. For FPS friction system the curve is almost linear, soft and little dependant
on the period. Consequently, as previously observed, the results for the Kobe record in-
dicate the convenience of having an additional source of lateral stiness for pure friction
systems.

4.4. Seismic coecient

Figures 3(d)–3(f) show the seismic coecients for the Gilroy 0◦ , Sylmar and Kobe E–W
records, respectively. The seismic coecient dened in Equation (8) is an important design
parameter because it determines the demanding base shear for the structure. As previously
indicated, up to this sub-section the strength was considered xed as  = 0:16. The seismic
coecient of these systems is compared with the one proposed by the UBC-1997:

V Cv I
CS = = (9)
W RT
where V is the base shear, Cv is a UBC dened seismic coecient depending on the local
soil prole type and the seismic zone, I the building importance, R a design force reduc-
tion factor depending on the overall ductility and redundancy of the structural system, T is
the fundamental linear elastic period and W the building weight. The values adopted here
to determine the curve of the normative were the following: Z = 0:4, I = 1:25, Cv = 0:32Nv ,
R = 8:5, corresponding to an important facility built in a zone of maximum intensity, in sti
ground and using a ductile moment resisting frame. For this study the near source factor was
not considered (Nv = 1).

Copyright ? 2002 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2003; 32:151–164
160 D. ORDOÑEZ, D. FOTI AND L. BOZZO

In general, as for linear superstructure response, the seismic coecient is a decreasing


smooth function of the fundamental period. For T ¿0:6 s the isolation systems start to produce
values larger than those recommended by the UBC-1997 for a ductile frame. However, for
very sti buildings the seismic coecient is smaller than the corresponding UBC loads, always
taking into account that structural strength coecient have been maintained constant. In the
neoprene systems the NZ curve is always lower than the LRB curve and it is below the
design forces according to the code, except for the Sylmar record. For frictional systems FPS
is always above PF as the lateral stiness increases the force at the sliding level. For example
in Figure 4(b) for T = 0:2 s, Cs = 0:0078 for PF system and Cs = 0:08 for FPS, that is the shear
forces are 2.5% larger between them. For the Sylmar record, however, the dierence increases
considerably. Again considering a structural period T = 0:2 s, for PF system Cs = 0:08 and for
FPS Cs = 0:16, that is 100% bigger.

4.5. Structural strength

This sub-section presents comparisons for dierent structural parameters changing the struc-
tural strength. The objective is to observe the changes in response as a function of the
structural strength for weak systems. Consequently, this sub-section is important for the ob-
jectives of the present study. Figures 4(a)–4(c) show the base displacement, structural drift
and total displacement, respectively, for the Santa Monica record and considering a structural
strength coecient  = 0:1. Figures 4(d)–4(f) show similar parameters subjected to the same
record but with a strength coecient  = 0:2. Figures 4(a) and 4(c) indicate that for this
record the base displacement and the total structural displacement are not signicantly varied
if the structural strength changes. However, this is not the case for the structural drift, which
changes signicantly for frictional systems. For example considering a structural period of 1 s,
if the structural strength is increased from 0.1 to 0.2, the drift is reduced from 3.3 to 2:5 cm
(32%). Similar results are obtained for the other ve records, not included in the present
paper.
Figure 5 presents a similar comparison for the structural ductility demand but considering
the Gilroy, Sylmar and Kobe records. Figures 5(a)–5(c) illustrate the ductility demands for
a strength coecient  = 0:1 and Figures 5(d)–5(f) for a strength coecient  = 0:2: In this
gure it is observed the inuence of the strength coecient for all the systems, but particularly
for the FPS one. For example, considering a structural strength  = 0:2, a period of 0:4 s and
the Gilroy record (Figure 5(d)), the demanding ductility coecient for PF and FPS systems
are 1.5 and 1.55, respectively. If the strength coecient is reduced up to 0.1 (Figure 5(a)), the
ductility demands increase up to 6 and 8.5. Consequently, a 50% decrease in structural strength
causes a 300% and a 450% increase in ductility demand for PF and PFS systems, respectively.
For LRB and NZ systems similar trends are observed. Considering again a structural period
of 0:4 s and a strength  = 0:2 the ductility demands for the same record are 0.5 and 0.3,
respectively. If the strength coecient is reduced up to 0.1, the ductility demands increase up
to 2.2 and 1.05. These variations indicate that a 50% reduction in structural strength causes
a 340% and a 250% increase in ductility demands.
Clearly, it is observed that for all the isolator the increase in ductility demand is signicant
and not linearly dependent. The absolute value is larger for the frictional systems, although
in percentage it is not as dierent between all the devices.

Copyright ? 2002 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2003; 32:151–164
THE INELASTIC RESPONSE OF BASE ISOLATED BUILDINGS 161

0.20 0.20

0.15 0.15
LRB
LRB
Displacement (m)

Displacement (m)
NZ
NZ
FPS
FPS
0.10 0.10 PF
PF
Elastic
Elastic

0.05 0.05

0.00 0.00
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
(a) T (s) (d) T (s)

0.20 0.20

LRB LRB
NZ NZ
0.15 0.15
FPS FPS
Relative displacement (m)

Relative displacement (m)


PF PF
Elastic Elastic

0.10 0.10

0.05 0.05

0.00 0.00
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
(b) T (s) (e) T (s)

0.20 0.20

0.15 0.15
LRB LRB
Total displacement (m)

Total displacement (m)

NZ NZ
FPS FPS
0.10 PF 0.10 PF
Elastic Elastic

0.05 0.05

0.00 0.00
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
(c) Period (s) (f ) Period (s)

Figure 4. (a) – (c) Base displacement, structural drift and total displacement, respectively, for
a structural strength  = 0:1. (d) – (f) Base displacement, structural drift and total displacement,
respectively, for a structural strength  = 0:2.

Copyright ? 2002 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2003; 32:151–164
162 D. ORDOÑEZ, D. FOTI AND L. BOZZO

11 3.0

10

9 2.5
LRB
LRB
8 NZ
NZ
FPS
Ductility coefficient u

Ductility coefficient u
2.0 FPS
7
PF
PF
6
1.5
5

4
1.0
3

2 0.5
1

0 0.0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
(a) T (s) (d) T (s)

11 2.5

10

9 LRB LRB
2.0
NZ NZ
8
FPS Ductility coefficient u FPS
Ductility coefficient u

7 PF PF
1.5
6

5
1.0
4

3
0.5
2

0 0.0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
(b) T (s) (e) T (s)
8 3.0

7
2.5
LRB LRB
6
NZ NZ
Ductility coefficient u

Ductility coefficient u

FPS 2.0 FPS


5
PF PF
4 1.5

3
1.0
2

0.5
1

0 0.0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
(c) Period (s) (f ) Period (s)

Figure 5. (a) – (c) Structural ductility demand for a strength  = 0:1, for Gilroy, Sylmar and Kobe
records, respectively. (d) – (f) Structural ductility demands for similar structures with a strength  = 0:2.

Copyright ? 2002 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2003; 32:151–164
THE INELASTIC RESPONSE OF BASE ISOLATED BUILDINGS 163

5. CONCLUSIONS

In this paper the non-linear response of weak buildings supported on four base isolated systems
is studied. The systems have been divided into two groups: neoprene isolation systems (LRB
and NZ) and friction isolation systems (PF and FPS). For rubber isolation the non-linearity of
the systems is represented through the Wen’s model. In this study the following parameters
have been compared: base displacements, interstorey drift, total displacement, base shear and
ductility demand for weak structures of dierent strengths.
As in the case of linear base isolated superstructure response, base displacements are smaller
for frictional systems compared to neoprene ones. This general trend is not kept for Kobe
record where large displacements are obtained also for PF systems. This response characteristic
indicates the convenience of having some additional restoring force such as for FPS systems,
in order to prevent large lateral displacements. A similar observation for rubber isolators
indicates the dependance between the frequency content of an earthquake and the ecacy of
the energy dissipation in the lead core of NZ system to reduce base displacements.
The main conclusion from this research is that the structural strength of weak systems
aects signicantly the performance of base isolated buildings. Rubber isolators seem slightly
less sensitive to plastication that may occur in the structure compared to friction isolators,
although, in percentage the dierence is signicant for all systems. Ductility demands in
the structure increase not linearly as the structural strength is reduced. Consequently, design
parameters for a long-term performance of the isolators should be selected with care, and
dierent values should be selected for the desired limit state. In the authors opinion, a robust
earthquake resistant design for essential facilities and historical buildings must be satisfactory
not only for average response values but also for individual cases. Consequently and due to
the uncertainties in the seismic input the results from this research points out the necessity of
a more extensive study in this subject.

ACKNOWLEDGEMENT

This paper has been partially funded by the Spanish Government under Research Grant # PB98-0455
which is greatly appreciated.

REFERENCES

1. Kelly JM. Aseismic Base isolation: review and bibliography. Soil Dynamics and Earthquake Engineering 1986;
5(3):202–216.
2. Su L, Amhadi G, Tadjbakhsh, IG. A comparative study of performances of various base isolation systems. Part
I: Shear beam structures. Earthquake Engineering and Structural Dynamics 1989; 18:11–32.
3. Su L, Amhadi G, Tadjbakhsh IG. A comparative study of performances of various base isolation systems. Part
II: Sensitivity analysis. Earthquake Engineering and Structural Dynamics 1990; 19:21–33.
4. Bozzo LM, Barbat SA. Nonlinear response of structures with sliding base isolation. Journal of Structural
Control 1995; 2(2):59–77.
5. Naeim F, Kelly JM. Design of Seismic Isolated Structures, From Theory to Practice. Wiley: New York, 1999.
6. Moka AS, Constantinou MC, Reinhorn AM. Experimental study and analytical prediction of earthquake response
of a sliding isolation system with a spherical surface. Report NCEER-90-0020, National Center for Earthquake
Engineering Research, State University of New York, Bualo, 1988.
7. Zayas VA, Low SS, Mahin SA. The FPS Earthquake Resisting Systems: Experimental Report. Report EERC-
87/01, 16, Earthquake Engineering Research Center, University of California, Berkeley, 1988; 839–854.

Copyright ? 2002 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2003; 32:151–164
164 D. ORDOÑEZ, D. FOTI AND L. BOZZO

8. Wen YK. Method for random vibration of hysteretic systems. Journal of the Engineering Mechanics Division
(ASCE) 1976; 102:249–263.
9. EERJ. Northridge Earthquake, January 17, 1994. Preliminary Reconnaissance Report. Earthquake Engineering
Research Institute, California, 1994.
10. Barbat A, Bozzo L. Seismic analysis of base isolated buildings. Archives of Computational Methods in
Engineering. State of the Art Reviews. vol. 4(2). CIMNE: Barcelona, 1997; 153–192.
11. Bozzo LM, Mahin SA. Design of Frictional Base Isolation Systems. Proceedings of the 4th US National
Conference on Earthquake Engineering, Palm Springs, California, 1990.

Copyright ? 2002 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2003; 32:151–164

Anda mungkin juga menyukai