Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
S. ANTONIOU
Department of Civil Engineering, Imperial College London
Imperial College Road, London SW7 2BU, UK
J. Earth. Eng. 2004.08:497-522. Downloaded from www.worldscientific.com
R. PINHO
European School for Advanced Studies in Reduction of Seismic Risk (ROSE School)
Collegio Alessandro Volta, Via Ferrata 17, 27100 Pavia, Italy
The recent drive for use of performance-based methodologies in design and assessment
of structures in seismic areas has significantly increased the demand for the development
of reliable nonlinear inelastic static pushover analysis tools. As a result, the recent years
have witnessed the introduction of the so-called adaptive pushover methods, which,
unlike their conventional pushover counterparts, feature the ability to account for the
effect that higher modes of vibration and progressive stiffness degradation might have
on the distribution of seismic storey forces. In this paper, the accuracy of these force-
based adaptive pushover methods in predicting the horizontal capacity of reinforced
concrete buildings is explored, through comparison with results from a large number
of nonlinear time-history dynamic analyses. It is concluded that, despite its apparent
conceptual superiority, current force-based adaptive pushover features a relatively minor
advantage over its traditional non-adaptive equivalent, particularly in what concerns the
estimation of deformation patterns of buildings, which are poorly predicted by both types
of analysis.
1. Introduction
Under the pressure of recent developments, seismic codes have begun to explicitly
require the identification of sources of inelasticity in structural response, together
with the quantification of their energy absorption capacity. Ideally, such perfor-
mance evaluation of structural systems subjected to earthquake loading should be
based on nonlinear time history analysis. However, the intrinsic complexity and the
additional computational effort required by the latter (especially if a fibre-based
distributed inelasticity modelling philosophy is adopted) do not justify its use in
ordinary engineering applications.
As a result of the above, nonlinear static, as opposed to dynamic, pushover
analysis has been gaining significance over recent years as a tool for assessment and
497
May 7, 2004 14:47 WSPC/124-JEE 00151
design verification. Indeed, and despite its relative simplicity and ease of use, this
numerical tool can provide information on many important response characteristics
by UNIVERSIDAD DE CHILE SISTEMA DE SERVICIOS DE INFORMACION Y BIBLIOTECAS (SISIB) on 09/05/17. For personal use only.
16 to line 2
accurately Replace “masses”
estimate structural with “mass”
seismic demand, as demonstrated by a number
16 line 3 Replace “the response acceleration” with “seismic forces”
16 line 3 Replace “the response acceleration” with “seismic forces”
16 line
of researchers 4 [Lawson Replace “heights”
et al.,with 1994; with “height”
Krawinkler and Seneviratna, 1998; Kim and
16 line 4 Replace “heights” “height”
17 line 23 Replace “applications” with “application”
17 D’Amore,
17
line 23 1999; Replace
line 30
Naeim and Lobo,
“applications” 1999].
with For instance, the deformation predictions
“application”
Replace “0.5better” with “0.5% plots described the elastic (or better”
17 line 30 Replace “0.5better” with “0.5% plots described the elastic (or better”
can
18 be highly
line 8 inaccurateReplace if“(0.5”
higher with modes
“±0.5” are important and/or if the structure is
18 line 8 Replace “(0.5” with “±0.5”
18 18 line 12
pushed line 12 into
highly Replace
its “versus.”
nonlinear
Replace “versus.” with “versus” with “versus”
post-yield range. The latter is characterised by
20 20 line 5line 5 ReplaceReplace “at with
“at softening
four” four”“at
with “at
thethe the four”
four”
gradual
20 line 13
degradation
line 13 Replace
and
Replace “reasonable”
of structural
with “reasonably”
system during the analysis pro-
20 “reasonable” with “reasonably”
20 cedure,
20 line 15which
line 15 inReplace
turnReplace
leads to significant
“clear “clear
’global ’global
winner’, the”elongation
winner’, the”
with of the
with“global
“clear “clear periods
“global
winner”, the” (see
winner”, the” Fig. 1(a))
21 21 change
and line 6line of
6 modalReplaceReplace
shape “that” withcorresponding”
“that” characteristics
with “the “the corresponding”
(see Fig. 1(b)), as a result of the tendency
21 21 line 12 line 12 Replace Replace
“load”“load” with “loading”
with “loading”
for deformations
21 line 14
to concentrate
line 14 Replace Replace
at the locations
“configurations. Contrary”
that sustain more
with “configurations
damage.
and, contrary”
J. Earth. Eng. 2004.08:497-522. Downloaded from www.worldscientific.com
6 first mode
second first
mode mode
period (sec)
5 third mode
Mode 1
5
second mode
third mode
Mode 1
4
4
3
3 ___ before loading at peak capacity
2 ___ before loading at peak capacity
2
1
Mode 2
1
0
Mode 2
0.0% 0.5% 1.0% 1.5% 2.0%
0 total drift
0.0% 0.5% 1.0% 1.5% 2.0%
total drift
dation, the period elongation and the influence of all significant modes of vibration
can be explicitly considered. In addition, finding a way to somehow incorporate
by UNIVERSIDAD DE CHILE SISTEMA DE SERVICIOS DE INFORMACION Y BIBLIOTECAS (SISIB) on 09/05/17. For personal use only.
the expected ground motion in the analysis might also allow the attainment of
site-specific results, as required in some applications. This can be achieved with
the utilisation of response spectra representative of such areas, derived from actual
accelerograms or code provisions.
In recent years, a number of research endeavours have been carried out with the
objective of introducing the aforementioned developments. The majority of such
are briefly summarised in the subsections that follow.
J. Earth. Eng. 2004.08:497-522. Downloaded from www.worldscientific.com
sented were superior to the predictions of the pushover analyses with the fixed
force distributions suggested by FEMA [ATC, 1997], which greatly underestimated
by UNIVERSIDAD DE CHILE SISTEMA DE SERVICIOS DE INFORMACION Y BIBLIOTECAS (SISIB) on 09/05/17. For personal use only.
drift demands.
Overall, the above multi-modal procedures constitute a significant improvement
of conventional pushover analysis. The former are theoretically more robust and
conceptually more attractive, since they explicitly consider the response of more
than one vibration mode and the influence of the expected ground motion, thus
yielding results that are closer to rigorous inelastic time-history analysis. However,
these methods do not account for damage accumulation and resulting modification
of the modal parameters, which might considerably affect the response character-
istics of a given structure. Therefore, the fully adaptive methods presented in the
J. Earth. Eng. 2004.08:497-522. Downloaded from www.worldscientific.com
Another method proposed by Requena and Ayala [2000], who discussed two vari-
ations of adaptive pushover (referred to as approaches 2-A and 2-B) and compared
by UNIVERSIDAD DE CHILE SISTEMA DE SERVICIOS DE INFORMACION Y BIBLIOTECAS (SISIB) on 09/05/17. For personal use only.
them with a modal fixed-pattern scheme. Whilst in one of the proposed variants
the storey loads were derived through SRSS combination of modal forces, in the
second option an “equivalent fundamental mode” is first determined (through an
SRSS combination of the vibration modes shapes) and then used to compute the
lateral loads. This procedure is repeated whenever the structural stiffness changes,
as new plastic hinges form and develop, hence the lateral load distribution reflects
the current state of inelasticity. The two alternative adaptive methods are indeed
appealing since they are theoretically rigorous and they explicitly account for higher
modes and spectral contributions. In addition, the procedures suggested in the pa-
J. Earth. Eng. 2004.08:497-522. Downloaded from www.worldscientific.com
per for the determination of the target displacement are also of particular interest,
albeit beyond the scope of the present work. However, and unfortunately, the an-
alytical results presented by Requena and Ayala [2000] were only limited and the
accuracy of the procedures could not be effectively assessed or judged.
A wholly alternative adaptive pushover methodology has been recently proposed
by Albanesi et al. [2002], who suggested an Adaptive Energy-based Pushover Anal-
ysis (AEPOA) whereby the imposed lateral force/displacement profiles at each
step supposedly take into account not only the inertial properties of the struc-
ture but also the kinetic energy that the latter is expected to mobilise when sub-
jected to earthquake loading. The results obtained, however, did not seem partic-
ularly exceptional, since, compared to conventional pushovers, the curves derived
by the AEPOA method were very unstable and tended to stop at very low defor-
mation levels (less than 1% in some cases), whereas the predicted values of the
base shear strength of the building were unreliable. In addition, no clear proce-
dure for applying and updating the lateral loads (forces or displacements) was
described.
Finally, Elnashai [2001] proposed an adaptive pushover scheme that seemed
to encompass, within a single-analysis pushover algorithm, all advanced features
described above. This single-run procedure is fully adaptive and multi-modal and
accounts for system degradation and period elongation by updating the force distri-
bution at every step (or at predefined steps) of the analysis. The dynamic properties
of the structure are determined by means of eigenvalue analyses that consider the
instantaneous structural stiffness state, at each analysis step. Site- or record-specific
spectral shapes can also be explicitly considered in the scaling of forces, so as to
account for the dynamic amplification that expected ground motion might have on
the different vibration modes of the structure. This force-based adaptive pushover
algorithm, deemed as a satisfactory representative of the current state-of-the-art
in the field, was further developed and tested by the authors of the present work,
for improved stability and accuracy, and is thus presented and discussed in greater
detail in the remaining parts of this presentation.
May 7, 2004 14:47 WSPC/124-JEE 00151
mass, which, however, can usually be obtained directly from the vertical gravity
loads, already included in any type of pushover analysis.
by UNIVERSIDAD DE CHILE SISTEMA DE SERVICIOS DE INFORMACION Y BIBLIOTECAS (SISIB) on 09/05/17. For personal use only.
P = λ · P0 . (1)
Load control refers to the case where the load factor is directly incre-
mented/controlled by the pushover algorithm whilst the structural response
(e.g. nodal displacements and rotations) corresponding to such loading level (force
level in this case) is subsequently determined. In other words, and in the case of
FAP, a force-controlled pushover is carried out, with the load factor being used to
scale directly the applied force vector until the point of peak capacity.
In the case of response control, on the other hand, it is the response of the
structure (e.g. a given nodal displacement or rotation) that is first directly incre-
mented/controlled, after which the load factor corresponding to such deformation
level can be computed. The load factor variation, therefore, is not prescribed by
the user, but is instead automatically calculated by the program so that the ap-
plied loading vector (force vector in this case) at a particular increment corresponds
to the attainment of the target response displacement/rotation at the controlled
node. It is noted that this type of load factor incremental scheme differs signifi-
cantly from a displacement-controlled analysis since it is the response deformation
of the structure, as opposed to a loading displacement vector, that is controlled by
the program.
A more detailed description of the workings of such loading/solution schemes
is beyond the scope of the current presentation, interested readers being instead
referred to the manual of the FE program, indicated earlier, where the proposed
adaptive pushover algorithm has been implemented. Here, it suffices to note that for
the purpose of the current parametric study, and in general for all force-based adap-
tive pushover applications, the response control algorithm scheme is the preferred
option since it (i) fully captures irregular response features such as soft-storey fail-
ures, (ii) models the softening post-peak loading/response branch of the structure,
(iii) provides an even distribution of force-displacement curve points throughout
both pre- and post-peak loading/response ranges and (iv) provides direct control
of structural deformations, which, as widely acknowledged, provide a better insight
into the damage incurred by the structure.
May 7, 2004 14:47 WSPC/124-JEE 00151
vector (or load increment vector) at each step, is computed at the start of each load
increment. In order for such scaling vector to reflect the actual stiffness state of the
structure, as computed at the end of the previous load increment, an eigenvalue
analysis is firstly carried out. To this end, the Lanczos algorithm [Hughes, 1987] is
employed to determine the modal shapes and participations factors of any given pre-
defined number of modes, after which the modal storey forces Fij can be determined
as:
where i is the storey number and j is the mode number, Γj is the modal participation
factor for the jth mode, φij is the mass normalised mode shape atvalue for the ith
before
beforeloading
loading atpeak
peakcapacity
capacity ___
___
storey
66
and the jth mode, and M i is the mass of the ith storey.
Alternatively, the modal storey forces can be computed using Eq. (2b), given
period (sec)
period (sec)
first
firstmode
mode
second
secondmode
mode Mode
Mode 11
below,
55
where
third
an additional parameter Sa,j that represents the acceleration response
thirdmode
mode
spectrum
44 ordinate corresponding to the period of vibration of the jth mode, here-
after referred to as spectral amplification, is entered into the computation of F ij . In
33
other words, the modal storey forces are weighted by beforethe S value at
loadinga
beforeloading
thecapacity
instanta-
peakcapacity
atatpeak ___
___
neous period of that mode, so as to take into account the effects that the frequency
22
content
11 of a particular input time-history, or spectrum, have in the response of the
Mode
Mode22
structure being analysed:
00
0.0%
0.0% 0.5%
0.5% 1.0%
1.0% 1.5%
1.5% 2.0%
2.0%
=drift
Fijtotal
total Γj φij Mi Sa,j .
drift (2b)
5000
5000
base shear (kN)
6000
6000
4000
4000
5000
5000
3000
3000 4000
4000
3000
3000
2000
2000
2000
2000
dynamic
dynamic dynamic
dynamic
1000
1000 with
withspectral
spectralamplification
amplification with
withspectral
spectralamplification
amplification
without
withoutspectral
spectralamplification
amplification 1000
1000 without
withoutspectral
spectralamplification
amplification
00 00
0%0% 1%
1% 2%
2% 3%
3% 0%0% 1%
1% 2%
2% 3%
3%
total
totaldrift
drift total
totaldrift
drift
(a) Model
(a)
(a) ModelWM15-AR
Model WM15-AR
WM15-AR (b) Model
(b)
(b) ModelRM15-AR
Model RM15-AR
RM15-AR
Fig.2.2.Effect
Fig. Effectof
ofspectral
spectralamplification
amplificationin
inresults
resultsobtained
obtained
Fig. 2. Effect of spectral amplification in adaptive pushover results.
normalised
normalised nominal
nominal
shapeatatstep
shape stept t loadvector
load vector newforces
new forcesapplied
appliedat
atstep
steptt
normalised = 1
normalised
∑∑ shape = 1
Fig. ( ( representation
Fig.3.3. Graphical
Graphical ))
shape
representation ofloading
of loadingforce
forcevector
vectorcalculation
calculationwith
withtotal
totalupdating
updating
May 7, 2004 14:47 WSPC/124-JEE 00151
results. These observations were confirmed in the present study, where the inclu-
sion of ground motion characteristics in the adaptive pushover analysis (through
by UNIVERSIDAD DE CHILE SISTEMA DE SERVICIOS DE INFORMACION Y BIBLIOTECAS (SISIB) on 09/05/17. For personal use only.
the use of response spectra) provided an equally closer fit to the dynamic results,
as shown in Fig. 2. Although in several cases the predictions with and without
spectral amplification coincided, in none did the pushover with spectral amplifica-
tion (i.e. different Sa for each mode) perform worse. Therefore, the utilisation of
a spectral shape, either code-defined or derived from actual records, to determine
modal storey forces is, in general, recommended.
Ideally, in order to assure full consistency between demand and supply duc-
tilities, multiple response spectra, derived for varying values of equivalent viscous
damping, should be employed so as to reflect the actual energy dissipation charac-
J. Earth. Eng. 2004.08:497-522. Downloaded from www.worldscientific.com
teristics of the structure at each deformation level (i.e. at each analysis step). The
implementation of such refinement, however, was beyond the scope of the current
work, for which reason a single constant response spectrum derived for an equiva-
lent viscous damping value of 5%, which the authors acknowledge may contribute
to an overestimation of the first mode contribution, was used throughout the en-
tirety of each analysis. Future work will target the assessment of the effects that
such apparent limitation has on attained results.
The lateral load profiles of each vibration mode are then combined using ei-
ther the Square Root of the Sum of Squares (SRSS, Eq. 3(a)), if the modes can
be assumed as fully uncoupled, or the Complete Quadratic Combination (CQC,
Eq. 3(b)) method, if cross-coupling of modes and respective viscous damping is to
be considered. It is noted that in Eq. (3b), n stands for the number of modes whilst
ρjk is the cross-model correlation coefficient, which can be approximated by Eq. (4)
if one assumes all modes to feature the same damping coefficient ξ:
v
u n 2
uX
Fi = t Fij , (3a)
j=1
v
u n X
n
uX
Fi = t (Fij · ρjk · Fik ) , (3b)
j=1 k=1
8 · ξ 2 · (1 + r) · r1.5 ωk
ρjk = , r= . (4)
(1 − r2 )2 + 4 · ξ 2 · r · (1 + r)2 ωj
Further, since only the relative values of storey forces (Fi ) are of interest in the
determination of the normalised modal scaling vector F̄ , which defines the shape,
not the magnitude, of the load or load increment vector (see Sec. 3.4), the forces
obtained by Eqs. (3a) or (3b) are normalised with respect to the total value, as
follows:
Fi
F̄i = . (5)
ΣFi
May 7, 2004 14:47 WSPC/124-JEE 00151
It is noted that, as highlighted by Priestley [2003], the use of SRSS or CQC rules
to combine modal forces may not always lead to the attainment of an adequate load
by UNIVERSIDAD DE CHILE SISTEMA DE SERVICIOS DE INFORMACION Y BIBLIOTECAS (SISIB) on 09/05/17. For personal use only.
vector shape that accurately represents the dynamic response characteristics of the
structure at a particular deformation level. In fact, if true reproduction of dynamic
analysis response is the objective of the pushover analysis, then the normalised
scaling vector should ideally be obtained through a weighted vectorial addition
of the contribution of each mode, to avoid the not necessarily realistic storey force
increase when the force vector from one mode is summed, through SRSS or CQC, to
the fundamental one. The employment of such alternative modal force combination
procedure, however, calls for additional in-depth studies that are conspicuously
beyond the scope of the present endeavour. Hence, and taking also account of the
J. Earth. Eng. 2004.08:497-522. Downloaded from www.worldscientific.com
fact that in the present work observed modes of vibration at every deformation
stage were always sufficiently apart, the SRSS modal combination procedure was
consistently employed throughout the analyses.
It is equally noteworthy that when the structural response reaches its post-
peak range, the eigen-solver can no longer output real eigen-solutions, due to the
presence of negative values in the diagonal of the stiffness matrix which in turn lead
to imaginary periods of vibration, corresponding to wholly unfeasible modal shapes.
In such cases, the load vector shape is no longer changed (only its magnitude is
updated), effectively meaning that a conventional non-adaptive pushover analysis
is employed thereafter. This is acceptable since, in the majority of applications,
entrance in the post-peak response range corresponds to the formation of a given
failure mechanism, which then remains qualitatively unchanged until collapse is
reached. Furthermore, and as described in Sec. 4.3, when a given storey starts failing
due to damage accumulation, force-based adaptive algorithms tend to generate
unrealistically high load concentrations at that given storey, leading to equally
unfeasible high drift values. Hence, and for this reason also, the employment of
a constant load vector shape in the post-peak softening structural response range
seems justified.
Finally, in cases where a very large number of analysis steps are employed and/or
when the structural model is very large, it might prove advantageous for the load
vector shape to be updated with a frequency lower than the number of analysis
steps, so as to reduce the computation effort. Such feature is explicitly available
in the algorithm developed by the authors. Within the framework of the current
parametric study, however, for reasons of accuracy and analysis stability, the load
vector shape was updated at every analysis step, with up to ten modes of vibration
being considered in its computation.
normalised nominal
shape at step t load vector new forces applied at step t
by UNIVERSIDAD DE CHILE SISTEMA DE SERVICIOS DE INFORMACION Y BIBLIOTECAS (SISIB) on 09/05/17. For personal use only.
Pt = λt × × =
∑=1
Fig. 3. Graphical representation of loading force vector calculation with total updating.
Fig. 3. Graphical representation of loading force vector calculation with total updating.
a newly derived load vector, computed as the product between the current total
load
at factor λt ,drift
0% total the current
at normalised modal scaling
0.5% total drift at 1.0%vector F̄t and the
total drift nominal
at 1.5% total load
drift
vector P0 , as schematically represented in Fig. 3 and numerically translated into
Fig. 4. Total Updated loading force vector at different deformation levels (model WH30-NR1).
Eq. (6a):
IF-H030
Fig. 4.Fig.
Total / IH30
4. Updated loading
Total-Updated force
loading vector
force at atdifferent
vector
High
different deformation levels
deformation levels (model
(model
4.00 0.565
WH30-NR1).
WH30-NR1).
0.30
IF-M030 / IM30 Irregular Medium 3.00 0.536
2 normalised nominal 8 new increment
IF-M015 / IM15 frame
shape at step t load vector Medium
of forces 3.00 0.613
0.15
IF-L015
∆Pt = ∆λ/t IL15
× × = Low 2.00 0.614
∑=1
FW-H030 / WH30 High 3.50 0.569
new increment 0.30
FW-M030 / WM30 existing forces
Regular Medium
of forces 2.625
new forces applied at step t 0.557
3 8
Pt=Pt-1+∆P t=
frame-wall + Medium
= 2.625 0.601
FW-M015 / WM15
May 7, 2004 14:47 WSPC/124-JEE 00151
proportionally increased so that at any given storey the loading force is at least
equal to that of the previous step. This loading vector corrective technique, how-
by UNIVERSIDAD DE CHILE SISTEMA DE SERVICIOS DE INFORMACION Y BIBLIOTECAS (SISIB) on 09/05/17. For personal use only.
ever, cannot, for obvious reasons, be employed with a response control algorithm,
the preferred choice for force-based adaptive pushover (Sec. 3.2, above). There-
fore, Total Updating has not been adopted in the current parametric study, being
instead replaced by the stable, and comparably accurate, Incremental Updating
scheme, described below.
by adding to the load vector of the previous step Pt−1 (existing balanced loads) a
newly derived load vector increment, computed as the product between the current
load factor increment ∆λt , the current modal scaling vector F̄t and the nominal
load vector P0 , as schematically represented in Fig. 5 and numerically translated
into Eq. (6b).
normalised nominal
shape at step t new forces applied at step t
Pt = Pt−1load vector
+ ∆λ t · F̄t · P0 . (6b)
Pt = λt × that the results
It is noteworthy × obtained with = Incremental Updating were con-
sistently very similar to those ∑=1 obtained with Total Updating, in those cases where
Fig. 3.theGraphical
latter couldrepresentation of loading
indeed be applied. force to
This seems vector
comecalculation
as a result ofwith
thetotal
fact updating.
that
in both schemes the trends of the load distributions are similar, as shown by Anto-
niou [2001, 2002], even though the absolute values of the forces differ significantly
in some cases. In addition, even in those limited cases where differences between the
two load updating modalities were not negligible, it was by no means clear which
variant provided better estimates. Finally, it is also noted that Bracci et al. [1997],
Gupta and Kunnath [2000] and Requena and Ayala [2000], whose work has been
briefly
at 0%reviewed
total driftin Sec.at2 0.5%
of this paper,
total drift have also adopted
at 1.0% total drift the use of incremental
at 1.5% total drift
load vector updating.
Fig. 4. Total Updated loading force vector at different deformation levels (model WH30-NR1).
∆Pt = ∆λt × × =
∑=1
new increment
existing forces of forces new forces applied at step t
Pt=Pt-1+∆Pt= + =
Fig. 5. Graphical
Fig. 5. Graphical representation
representation of loadingforce
of loading force vector
vectorcalculation with incremental
calculation updating. updating.
with incremental
4. Parametric Study
In what follows, sample results of an extensive parametric study that aimed at
by UNIVERSIDAD DE CHILE SISTEMA DE SERVICIOS DE INFORMACION Y BIBLIOTECAS (SISIB) on 09/05/17. For personal use only.
4.1. Introduction
Adaptive pushover is intended to be a method for general use in design and as-
sessment. It is therefore imperative to verify its efficiency for different structural
J. Earth. Eng. 2004.08:497-522. Downloaded from www.worldscientific.com
coupling beam
12 × 3.0 = 36.0 m
structural
8 × 3.0 = 24.0 m
J. Earth. Eng. 2004.08:497-522. Downloaded from www.worldscientific.com
Eurocode 8 [CEN, 1995]. Subsequently, they were modelled by Mwafy [2001] under
the framework of a different project, and were then adapted by Rovithakis [2001]
for the purpose of the current research endeavour. Their general characteristics are
defined in Table 1 and schematically illustrated in Fig. 6.
It is also worth noting that many of the modelled buildings feature a column-
sway, rather than beam-sway, type of failure mechanism, and are thus likely to
experience significant modifications of the applied force vector shape at different
deformation levels. This serves well the purpose of the current study, since it is in
such non-beam-sway cases that conventional non-adaptive pushovers struggle the
most to accurately reproduce the dynamic response characteristics of a building
subjected to a given input motion.
Four input time-histories, consisting of one artificially-generated accelerogram
[Campos-Costa and Pinto, 1999] and three natural records (Loma Prieta earth-
quake, USA, 1989), were employed for the dynamic analyses of the study. The
selection of these four records aimed at guaranteeing that the twelve buildings de-
scribed above would be subjected to a wide-ranging type of earthquake action, in
terms of frequency content, peak ground acceleration, duration and number of high
amplitude cycles. Indeed, the original PGA of the records varies between 0.12g and
0.93g, the spectral shapes are markedly distinct and provide high amplifications at
different periods (see Fig. 7), and the ratio between the significant duration (defined
as the interval between the build up of 5% and 95% of the total Arias Intensity
[Bommer and Martinez-Pereira, 1999]) and the total duration ranges from 22% to
72%. The characteristics of the records are summarised in Table 2, whereas their
elastic response spectra for an equivalent viscous damping of 5% are shown in Fig. 7.
The twelve building configurations subjected to four records resulted in a total
number of 12 × 4 = 48 test cases, deemed adequate to provide a representative sam-
ple for the verification of the proposed algorithm. Two conventional pushover pro-
cedures were carried out for each of the 12 buildings, for which the two distributions
May 7, 2004 14:47 WSPC/124-JEE 00151
Peak Peak
DE CHILE SISTEMA DE SERVICIOS DE INFORMACION Y BIBLIOTECAS (SISIB) on 09/05/17. For personal use only.
(g)(g)
(g) (g)
(g)(g)
(g) (g)
0.6 0.6
acceleration
0.6 0.6
acceleration
acceleration
acceleration
0.6 0.6
acceleration
0.6 0.6
acceleration
acceleration
acceleration
(g)(g)
(g)(g)
acceleration
acceleration
acceleration
acceleration
acceleration
acceleration
proposed in the NEHRP Guidelines [ATC, 1997] were employed; the uniform dis-
UNIVERSIDAD
tribution, whereby lateral forces are proportional to the total mass at each floor
8000 8000
level, and the triangular distribution, in which seismic forces are proportional to
(kN)
(kN)
(kN)
(kN)
8000 8000
12000
(kN)
the product of floor mass and storey height. Furthermore, in excess of ten adaptive
(kN)
(kN)
12000
(kN)
7000 7000
by shear
shear
shear
12000
11000 12000
shear
shear
shear
shear
6000 11000
6000 10000
base
base
base
base
6000 6000
in Sec. 3. The results of the 144 pushover10000 analyses
9000
9000carried out were then compared
10000
base
base
base
5000
base
5000 9000
8000 9000
5000to the envelopes derived with the recently proposed Incremental Dynamic Analysis
8000
5000
8000
7000 8000
4000 7000
4000
4000 4000
(IDA) procedure [Hamburger et al., 2000; 7000Vamvatsikos
6000 7000
6000
and Cornell, 2002; Mwafy
3000
3000and Elnashai, 2000; Papanikolaou, 2000], whereby a structure is subjected to a
6000
5000 6000
5000
3000 3000 5000
4000 5000
2000
2000series of nonlinear time-history
dynamic
adaptive
analysis
dynamic of
3000
4000
increasing
4000
4000
intensity (e.g. peak ground
dynamic
adaptive
dynamic
dynamic adaptive 3000 dynamic adaptive
2000 2000 uniform dynamic uniform dynamic
adaptive uniform 3000
2000 3000 adaptive uniform
1000 triangular adaptive 2000 triangular adaptive
1000 uniform triangular
uniform uniform triangular
1000
2000 2000 uniform
1000 1000 triangular triangular 1000 triangular triangular
0 10000 10000
0% 0 1% 2% 3% 0% 1% 2% 3%
0 0% 1% 2% 3% 0 0% 1% 2% 3%
0 total drift 3% 0 total drift
0% 0% 1% 1% 2% 2% total drift 3% 0% 0% 1% 1% 2% 2% 3% drift 3%
total
total drift total drift total drift total drift
(a) Model
(a)RL15-NR1
Model RL15-NR1 (b) Model
(b)WM30-NR3
Model WM30-NR3
Fig.(a) Model
(a)RL15-NR1
8. Adaptive Model
and conventional
Fig. 8. Adaptive
(b) Model
RL15-NR1 pushover curves vs. dynamic (b)WM30-NR3
Model
analysis
and conventional pushover curves vs. dynamic
WM30-NR3
envelopes
analysis envelopes
Fig. 8. Adaptive and conventional pushover curves vs. dynamic analysis envelopes
Fig. 8. Adaptive and conventional pushover curves vs. dynamic analysis envelopes
May 7, 2004 14:47 WSPC/124-JEE 00151
inelastic dynamic analyses with variable scaling factors were performed per case,
leading to a total of approximately 1000 nonlinear time-history runs.
All the analyses, dynamic or static, were carried out up to the point of 3%
global drift, employing the previously introduced FE package, capable of predicting
the large displacement behaviour of space frames under static or dynamic loading,
taking into account both local (beam-column effect) and global (large displace-
ments/rotations effects) geometric nonlinearities as well as material inelasticity. The
spread of the latter along the member length and across the section area is explicitly
represented through the employment of a fibre modelling approach, implicit in the
J. Earth. Eng. 2004.08:497-522. Downloaded from www.worldscientific.com
numerical stability and results accuracy are considered, is that featuring response
control, spectral amplification and incremental updating. Consequently, in all plots
provided hereafter, only such FAP configuration will be compared with conventional
pushover (uniform and triangular distributions), as well as IDA results.
It is also noted that in the current work the dynamic analysis envelopes consist
of the
1.0
locus of maximum total drift versus 1.0corresponding base shear (i.e. peak
base shear1.0
within a time-window (±0.5 seconds 1.0
of the instant of maximum drift
0.8 0.8
occurrence).
0.8 This reflects a pragmatic, rather than
0.8 judicious, choice, since it has
acceleration (g)
acceleration (g)
not been possible to unequivocally demonstrate that any of the three available
acceleration (g)
0.6
J. Earth. Eng. 2004.08:497-522. Downloaded from www.worldscientific.com
0.6
acceleration (g)
0.6 0.6
choices
0.4 for plotting dynamic analyses envelopes 0.4 (max drift versus max shear, max
0.4 0.4
drift
0.2
versus corresponding shear, corresponding 0.2
drift versus max shear) is more
meaningful0.2
than the other, for all structural types 0.2
and analysis cases. Future work
0.0 0.0
will hopefully
0 0.0
0
1 clarify
1
2 this
2
matter,
3
period
3
with
(sec)
the authors 0.0 not excluding
0
0
1
1
2 the possibility
2
3
3period (sec)
that
the use of only (a) one
Recordof ARthe three possible dynamic analysis
period (sec)
results
(b) Record NR1 interpretation
period (sec)
acceleration (g)
acceleration (g)
0.3 1.8
structural0.3 characteristics, and for different records, 1.8 are shown. It is observed that
the IDA
0.2
0.2 points lay between the uniform and the 1.2
1.2 triangular distributions, at least
within
0.1
0.1
the pre-yield range (some noticeable 0.6 exceptions
0.6
where these two distribu-
tions
0.0
did not provide upper and lower bounds, 0.0
even at pre-yield range, have been
0 0.0 1 2 3 00.0 1 2 3
recorded, 0 as discussed
1 2later inperiod
3this section).
(sec)
period (sec)
The adaptive
0 1 pushover 2 curves are
3period (sec)typi-
period (sec)
cally located between (c) Record NR2
the two conventional curves as well, providing (d) Record NR3 a slightly closer
(c) Record NR2 (d) Record NR3
Fig.Fig.
7. Elastic response spectra of of
thethe
four records (5% equivalent viscous
viscousdamping)
fit to the IDA 7. points.
Elastic response spectra four records (5% equivalent damping)
In addition to the above, however, there have also been several examples where
the pushover methods (adaptive or not) failed to correctly predict the dynamic
8000
8000
base shear (kN)
12000
12000
7000 7000
11000
11000
6000 6000 10000
10000
90009000
5000 5000
80008000
70007000
4000 4000
60006000
3000 3000 50005000
40004000
2000 2000 dynamic
dynamic dynamic
dynamic
adaptive
adaptive 30003000 adaptive
adaptive
uniform
uniform 20002000 uniform
uniform
1000 1000 triangular
triangular triangular
triangular
10001000
0 0 0 0
0% 0% 1% 1% 2% 2% 3% 3% 0% 0% 1% 1% 2%2% 3%3%
total driftdrift
total total drift
total drift
(a)(a)
(a) Model
Model
Model RL15-NR1
RL15-NR1
RL15-NR1 (b)
(b)
(b) Model
Model
Model WM30-NR3
WM30-NR3
WM30-NR3
Fig.Fig. 8. Adaptive
8. Adaptive andand conventional
conventional pushover
pushover curves
curves vs.vs. dynamic
dynamic analysisenvelopes
analysis envelopes
Fig. 8. Adaptive and conventional pushover curves versus dynamic analysis envelopes.
44
May 7, 2004 14:47 WSPC/124-JEE 00151
structural behaviour throughout the entire deformation range, even in the pre-yield
region. The most serious complications have been caused by record NR2, which fea-
by UNIVERSIDAD DE CHILE SISTEMA DE SERVICIOS DE INFORMACION Y BIBLIOTECAS (SISIB) on 09/05/17. For personal use only.
As noted in Fig. 9, in some cases, the dynamic base shear was considerably
higher than the static shear estimates provided by the uniform distribution (or, less
frequently, lower than the triangular). These are induced by the highly irregular
shapes of the inertia forces at the time step of maximum displacement, which cannot
be adequately reproduced, throughout the entire deformation range, by any of the
force-based static procedures employed in this work. Moreover, on many of such
cases, the slope of the IDA envelope was seen to change abruptly at a certain
point of the curve (e.g. Fig. 9(a), drift of 1.7%), due to the fact that, for some of
these structures, when subjected to a given accelerogram with a certain frequency
content, the increase of scaling factors, thus deformation demand and hence periods
of vibration, would at a certain point introduce a sudden change in the dominant
base shear (kN)
9000
response mode of vibration, which would then lead to a markedly diverse response
8000
trend, clearly identified in the dynamic analysis
7000
envelopes. Similar observations
have been made by Vamvatsikos and Cornell 6000 [2002] and Antoniou [2001].
5000
4000
4.3. Interstorey drifts and shear profiles
3000
dynamic
2000 adaptive
Interstorey drifts are crucial parameters in terms of structural response,uniform
especially
triangular
1000
in view of the recent developments in the field
0
of performance-based engineering,
0% 1% 2% 3%
total drift
base shear (kN)
9000 3000
8000
2500
7000
6000 2000
5000
1500
4000
3000 1000
dynamic dynamic
2000 adaptive adaptive
uniform 500 uniform
1000 triangular triangular
0 0
0% 1% 2% 3% 0% 1% 2% 3%
total drift total drift
(a) Model RH30-NR2 (a) Model RH30-NR2 (b) Model IM15-NR2 (b) Model IM15-NR2
base shear (kN)
3000
Fig. 9. Adaptive and conventional pushover curves vs. dynamic analysis envelopes
Fig. 9. Adaptive and conventional pushover curves versus dynamic analysis envelopes.
2500
2000
1500
1000
dynamic
adaptive
500 uniform
triangular
0
0% 1% 2% 3%
total drift
May 7, 2004 14:47 WSPC/124-JEE 00151
drift values as close7as possible to the predictions of more rigorous dynamic analyses.
For this reason,6 the drift profiles obtained in each of the pushover and dynamic
5
analyses were examined 4
at the four levels of total displacement introduced earlier;
0.5%, 1.0%, 1.5% and 3 2.5%. It is noted that the drift profiles from the dynamic anal-
ysis represent in fact 2
1
envelopes of peak response at the specified total displacement
levels, not actual profiles
0.0%
at a given instant
0.5%
of time. In other 1.0%
words, they represent
interstorey drift
the drift maxima at each storey as obtained in the one time-history analysis (20
were carried out for 12 each pair of building/record scenarios) where their sum
12
dynamic
comes
dynamic
12 11 adaptive
dynamic
11 the closest adaptive
to each10of the aforementioned total
11
10
adaptive
displacement
uniform limits. uniform
J. Earth. Eng. 2004.08:497-522. Downloaded from www.worldscientific.com
triangular
uniform triangular
10
In Fig. triangular
10, the sequence
9 of the storey deformations
9
for one representative case
9 8
8 is shown. In general, 8
the predictions were reasonably
7
accurate in the pre-yield and
7
7
reasonably early inelastic
6 range for both the6 adaptive and conventional procedures
6 5
5
(Figs. 10(a) and 10(b)).5
Although there was 4
no clear “global winner”, the force-
4
4 based adaptive pushover 3 predictions were 3 found to be slightly more accurate in
3 2
2
the majority of cases, 2 in consistency with the
1
observations on the capacity curves,
1
1
0.0% 0.5% 1.0%
0% 1% 2%
0.0% 0.5% 1.0% interstorey drift interstorey drift
interstorey drift
12
(c) total drift 1.5% (c) total drift 1.5% (d) total drift 122.5%
dynamic (d) total drift 2.5%
11 adaptive
12 dynamic
11
10 Fig. Interstorey
Fig. 10. Model RM30-NR3.
10. Model
dynamic RM30-NR3. Interstorey
uniform
adaptive drift profiles at triangular
drift profiles
11
different deformation
atlevels.
different deformation levels adaptive
uniform
9 uniform 10 triangular
10 triangular
8 9
9
8
7 8 6
6 7
7
5 6
6
4 5
5
3 4
4
2 3
3
1 2
2
0% 1% 2% 3% 4% 1
1
interstorey drift
0% 1% 2% 3% 4% 5% 6% 7% 8%
0% 1% 2% 3% 4% 5% 6% 7% 8% interstorey drift
interstorey drift
shown above. However, in the highly inelastic region, the FAP method failed to
provide correct response estimates, by exaggerating the inelastic deformations in the
by UNIVERSIDAD DE CHILE SISTEMA DE SERVICIOS DE INFORMACION Y BIBLIOTECAS (SISIB) on 09/05/17. For personal use only.
locations of damage and underestimating them elsewhere (Figs. 10(c) and 10(d)).
Such erroneous behaviour seems to be related to the fact that at onset of damage
on a given storey, the resulting reduction in stiffness and consequent changes in
the modal shapes derived by eigenvalue analysis at the corresponding analysis step
induces a concentration at that storey of the increment of forces for the next step,
thus increasing even more its damage and hence resulting in a sort of “vicious
circle” that leads to exaggeratedly large deformation levels at that such damaged
locations.
It is also noteworthy that, whereas the non-adaptive pushover with uniform
J. Earth. Eng. 2004.08:497-522. Downloaded from www.worldscientific.com
distribution yielded equally flawed results in that range, its triangular counterpart,
despite its fixed loading pattern, managed to provide better drift estimates at high
deformation levels. This can be explained by the fact that, typically, the damage
occurred at the lower floor levels for all structural configurations and, contrary
to the uniform and the adaptive patterns, with the triangular distribution the
forces applied to the failed storeys were smaller than those applied to the upper
storeys, resulting in more reasonable drift predictions. However, and although the
above does indeed constitute an interesting observation, it should be stressed that
such good performance of the fixed triangular patterns comes as a result of the
location of damage in the considered case studies. Had damage occurred at the
upper floors, the triangular distribution would have failed to predict accurately
the drifts profiles. Such behaviour was indeed observed by Antoniou [2002], where
non-adaptive pushovers with triangular distributions failed to provide adequate
predictions for those structural models that featured a soft-storey located at mid-
height.
Similar observations were made also for the storey shear profiles obtained from
each of the pushover and dynamic analyses at the aforementioned four levels of
structural deformation (0.5%, 1.0%, 1.5% and 2.5% total drift). Although differ-
ences were not significant, FAP shear distribution predictions were found to be typ-
ically between the results from the uniform and triangular non-adaptive pushover
analyses, and usually closer to the dynamic envelopes, at least for total deformations
levels of up to 1.5%. Thereafter, it was conventional pushover with triangular force
distribution that presented the best predictions, for the reasons discussed above.
There have also been cases where the interstorey drift and shear distribution
predictions of the adaptive and the fixed distributions were totally inadequate, even
at the early stages of deformation. Typical examples of such flawed behaviour are
depicted in the drift profiles of Fig. 11 and shear distributions of Fig. 12, the latter
clearly confirming that for a full reproduction of dynamic response characteristics it
is required that static analysis methods feature the capability of reversing the sign
of applied storey forces (in Fig. 12(a), for instance, dynamic inertia forces applied
at storeys eight and nine are negative). This is a feature that, as highlighted in
Sec. 3.3, current force-based pushover schemes, adaptive or not, seem to be lacking.
May 7, 2004 14:47 WSPC/124-JEE 00151
1212 12
12 dynamic
dynamic
dynamicdynamic 88 8
8
dynamic
dynamic
dynamic
dynamic
1111 11 adaptive
adaptive
adaptive adaptive
adaptive
adaptive
11 adaptive adaptive
uniform
uniform
uniform uniform
uniform
uniform
uniform 77 7 uniform
by UNIVERSIDAD DE CHILE SISTEMA DE SERVICIOS DE INFORMACION Y BIBLIOTECAS (SISIB) on 09/05/17. For personal use only.
1010 10
10 triangular
triangular
triangular 7 triangular
triangular
triangular
triangular triangular
99 9
9 66 6
6
88 8
8
77 7 55 5
5
7
66 6
6 44 4
4
55 5
5
33 3
44 4
4
3
33 3
3 22 2
2
22 2
2
11 1
11 1
1
1
0.0% 0.0%
0.0% 0.5% 0.5%
0.5% 1.0% 1.0% 0.0%
1.0% 0.0% 0.0% 0.5% 0.5%
0.5% 1.0%
1.0%1.0%
0.0% 0.5% interstorey
interstorey
drift drift 1.0% 0.0% 0.5% interstorey
interstorey
drift 1.0%
interstorey drift
interstorey drift interstorey drift drift
interstorey drift
(a) Model
(a) (a)
Model RL15-NR2
Model
RL15-NR2
RL15-NR2 (b)
(b) (b)Model
Model
(b) Model
IH30-NR2
IH30-NR2
IH30-NR2
J. Earth. Eng. 2004.08:497-522. Downloaded from www.worldscientific.com
(a) Model
(a) RL15-NR2
Model RL15-NR2 (b) Model IH30-NR2
Model IH30-NR2
Fig.Fig.
Fig. 11.Interstorey
11.
Fig. 11.
Interstorey
11. Interstorey
driftdrift
drift
Interstorey profiles
profiles
profiles
drift at 0.5%
at
profiles0.5%
at 0.5%totaltotal
total driftdrift
drift for different
for for
different
different
model
model
model configurations
configurations
configurations
Fig. 11. Interstorey drift profiles atat0.5%
0.5% total
total drift
drift for
fordifferent
differentmodel
modelconfigurations
configurations.
12 12
12 12 dynamic
dynamic 8 8 dynamic
dynamic
dynamicdynamic 8 8 dynamic
dynamic
11 11 adaptive
adaptive adaptive
adaptive
11 11 adaptive
adaptive adaptive
adaptive
uniform
uniform uniform
uniform
10 10 uniformuniform 7 7 uniform
uniform
10 10 triangular
triangular 7 7 triangular
triangular
triangular
triangular triangular
triangular
9 9
9 9 6 6
6 6
8 8
8 8
7 7 5 5
7 7 5 5
6 6
6 6 4 4
4 4
5 5
5 5
4 4 3 3
4 3 3
4
3 3 2 2
3 3 2 2
2 2
2 2 1 1
1 1 1 1
1 1
0 0 1000 1000 2000 2000 3000 3000 4000 4000 0 0 1000 1000 2000 2000 3000 3000
0 0 1000 1000 2000 2000 3000 3000 4000 4000 0 0 1000 1000 2000 2000 3000 3000
storey
storey
shearshear
(kN)(kN) storey
storey
shearshear
(kN)(kN)
storeystorey
shear shear
(kN) (kN) storey shearshear
storey (kN) (kN)
(a) (a)
Model
Model
RM15-NR2
RM15-NR2 (b) (b)
Model
Model
IL15-AR
IL15-AR
(a) (a) Model
(a)Model RM15-NR2
RM15-NR2
Model RL15-NR2 (b) Model
(b) Model
(b) Model IL15-AR
IL15-AR
IL15-AR
Fig.Fig.
12. 12.
Storey
Storey
shear
shear
profiles
profiles
at 0.5%
at
at 0.5%
totaltotal
driftdrift
for for
different
for different
model
model
configurations
configurations
Fig. Fig.
12. 12. Storey
Storey shearshear profiles
profiles at 0.5% 0.5%
totaltotal
driftdrift different
for different model
model configurations
configurations
Fig. 12. Storey shear profiles at 0.5% total drift for different model configurations.
It is interesting to note that the above deficiencies were not always reflected
on the capacity pushover curves, discussed previously, which, on occasions, were
seen to be correct even for those cases where storey drift and/or shear profiles were
incorrect. This highlights the importance of considering local parameters, such as
storey drift and/or storey shear profiles, when assessing the efficiency of pushover
methods, adaptive or not.
5. Conclusions
In this work, the accuracy of adaptive and non-adaptive force-based pushover meth-
ods in predicting the seismic response characteristics of reinforced concrete build-
ings has been explored. To this end, a fully adaptive pushover algorithm, which
explicitly accounts for the effects that progressive stiffness degradation and higher-
mode effects have on structural response, has been developed and employed in the 7 7
7 7
verification study. The latter comprised comparisons between the capacity curves
May 7, 2004 14:47 WSPC/124-JEE 00151
and drift/shear profiles obtained through static and dynamic nonlinear fibre-based
analysis on a number of real structural models of different configurations and duc-
by UNIVERSIDAD DE CHILE SISTEMA DE SERVICIOS DE INFORMACION Y BIBLIOTECAS (SISIB) on 09/05/17. For personal use only.
tility classes, subjected to equally diverse input ground motions; each pushover was
compared with the IDA output obtained for each individual accelerogram, as op-
posed to the average of the latter, to avoid smoothening out of important structural
response peculiarities introduced by individual records.
A summary of the main observations and general conclusions of the present
study is presented below:
capacity curves alone, but rather include an evaluation and comparison of local
response quantities, such as storey drift and storey shear profiles.
• Non-adaptive pushover analyses, with the triangular and uniform distributions,
do not always provide curves that constitute a lower and an upper bound to
the IDA response points. Moreover, they fail to reproduce accurately the local
dynamic response characteristics of buildings, particularly within the post-peak
range.
• Force-based adaptive pushover provides slightly improved capacity curves over its
conventional counterparts, with a closer fit to the IDA points. Although the differ-
ences were not always impressive, they were clear in the vast majority of the cases.
• Accordingly, up to global yield, FAP derives interstorey drift and storey shear
profiles closer to the dynamic envelopes. However, in the highly inelastic range
it tends to introduce excessive forces at the locations of plastic hinges and, sub-
sequently, to overestimate the drift values at such storeys.
• In some of the studied cases, none of the force-based static analysis methods em-
ployed in this study could correctly predict the dynamic response characteristics
of buildings, even at pre-yield deformation stages, due to the difficulties faced by
current force-based pushover algorithms in modelling a reversal of applied storey
force sign.
• Overall, it can be stated that force-based adaptive pushover, in its present state
of development, features a relatively minor advantage over its traditional non-
adaptive counterparts, all of which present limitations in the accurate prediction
of dynamic response characteristics of RC buildings.
The above clearly indicates that further research work is required before force-
based adaptive pushover can be considered as a valid alternative to nonlinear time-
history analysis. As highlighted in the body of the document, such developments
might include (i) the implementation of multiple response spectra to determine
the spectral amplification of different modes at varying deformation stages and/or
(ii) the employment of weighted vectorial addition, as opposed to SRSS and CQC
combination rules, in the computation of normalised scaling vector, as suggested
by Priestley [2003].
May 7, 2004 14:47 WSPC/124-JEE 00151
Acknowledgements
The authors would like to acknowledge the important contribution of Professor A.
by UNIVERSIDAD DE CHILE SISTEMA DE SERVICIOS DE INFORMACION Y BIBLIOTECAS (SISIB) on 09/05/17. For personal use only.
S. Elnashai, with whom the current research endeavours were initiated at Imperial
College of London. The first author is also grateful to the financial support provided
by the European Commission at different stages of the current research project,
through the research networks ICONS (Innovative Concepts for Seismic Design of
New and Existing Structures) and SAFERR (Safety Assessment For Seismic Risk
Reduction), and also by means of a Marie Curie Fellowship at the ROSE School
(Pavia, Italy). Finally, the authors thank Professor Nigel Priestley for his valuable
suggestions during the final stages of this work.
J. Earth. Eng. 2004.08:497-522. Downloaded from www.worldscientific.com
References
Albanesi, T., Biondi, S. and Petrangeli, M. [2002] “Pushover analysis: An energy based
approach,” Proc. of the Twelfth European Conference on Earthquake Engineering,
London, United Kingdom, Paper No. 605.
Antoniou, S. [2001] “Pushover analysis for seismic assessment of RC structures,” Transfer
Report, Department of Civil Engineering, Imperial College London, United Kingdom.
Antoniou, S. [2002] “Advanced inelastic static analysis for seismic assessment of struc-
tures,” PhD Thesis, Imperial College London, University of London, United King-
dom.
ATC [1997] “NEHRP Guidelines for the seismic rehabilitation of buildings,” FEMA Re-
port No. 273, Federal Emergency Management Agency, Applied Technology Council,
Washington, D.C.
Bommer, J. and Martinez-Pereira, A. [1999] “The effective duration of earthquake strong
motion,” Journal of Earthquake Engineering 3(2), 127–172.
Bracci, J. M., Kunnath, S. K. and Reinhorn, A. M. [1997] “Seismic performance and
retrofit evaluation of reinforced concrete structures,” Journal of Structural Engineer-
ing 123(1), 3–10.
Campos-Costa, A. and Pinto, A. V. [1999] “European seismic hazard scenarios — An
approach to the definition of input motion for testing and reliability assessment of
Civil Engineering structures,” JRC Special Publication No. X.99.XX, Joint Research
Centre, Ispra, Italy.
CEN [1995] European Prestandard ENV 1998-1-1: Eurocode 8 — Earthquake resistant
design of structures, Part 1: General rules and rules for buildings. Comite Europeen
de Normalisation, Brussels, Belgium.
Chopra, A. K. and Goel, R. K. [2002] “A modal pushover analysis procedure for estimating
seismic demands for buildings,” Earthquake Engineering and Structural Dynamics 31,
561–582.
Elnashai, A. S. [2001] “Advanced inelastic static (pushover) analysis for earthquake ap-
plications,” Structural Engineering and Mechanics 12(1), 51–69.
Fardis, M. N. [1994] “Analysis and design of reinforced concrete buildings according to
Eurocodes 2 and 8. Configurations 3,5 and 6,” Reports on Prenormative Research in
Support of Eurocode 8.
Freeman, S. A., Nicoletti, J. P. and Tyrell, J. V. [1975] “Evaluations of existing build-
ings for seismic risk — A case study of Puget Sound Naval Shipyard, Bremerton,
Washington,” Proc. of the First U.S. National Conference on Earthquake Engineer-
ing, Oakland, California, pp. 113–122.
May 7, 2004 14:47 WSPC/124-JEE 00151
for evaluation, upgrade and design of moment-resisting steel frames,” Proc. of the
Twelfth World Conference on Earthquake Engineering, Auckland, New Zealand, Paper
No. 2543.
Hughes, T. J. R. [1987] The Finite Element Method, Linear Static and Dynamic Finite
Element Analysis (Prentice-Hall, Inc.).
Kim, S. and D’Amore, E. [1999] “Push-over analysis procedure in earthquake engineering,”
Earthquake Spectra 15(3), 417–434.
Krawinkler, H. and Seneviratna, G. D. P. K. [1998] “Pros and cons of a pushover analysis
of seismic performance evaluation,” Engineering Structures 20(4–6), 452–464.
Lawson, R. S., Vance, V. and Krawinkler, H. [1994] “Nonlinear static push-over analysis
J. Earth. Eng. 2004.08:497-522. Downloaded from www.worldscientific.com
— Why, when, and how?” Proc. of the Fifth US National Conference on Earthquake
Engineering, Earthquake Engineering Research Inst., Oakland, California, 1, 283–292.
Lefort, T. [2000] “Advanced pushover analysis of RC multi-storey buildings,” MSc Disser-
tation, Department of Civil and Environmental Engineering, Imperial College London,
United Kingdom.
Moghadam, A. S. and Tso, W. K. [2002] “A pushover procedure for tall buildings,” Proc. of
the Twelfth European Conference on Earthquake Engineering, London, United King-
dom, Paper No. 395.
Mwafy, A. M. and Elnashai, S. A. [2000] “Static pushover versus dynamic-to-collapse
analysis of RC buildings,” ESEE Research Report No. 00/1, Department of Civil and
Environmental Engineering, Imperial College London, United Kingdom.
Mwafy, A. M. [2001] “Seismic performance of code-designed RC buildings,” PhD Thesis,
Imperial College London, University of London, United Kingdom.
Naeim, F. and Lobo, R. [1999] “Avoiding common pitfalls in pushover analysis,” Proc. of
the Eighth Canadian Conference on Earthquake Engineering, Canadian Association
for Earthquake Engineering, Vancouver, British Columbia, Canada, pp. 269–274.
Papanikolaou, V. [2000] “Development and verification of adaptive pushover analysis pro-
cedures,” MSc Dissertation, Department of Civil and Environmental Engineering,
Imperial College London, United Kingdom.
Paret, T. F., Sasaki, K. K., Eilbeck, D. H. and Freeman, S. A. [1996] “Approximate
inelastic procedures to identify failure mechanisms from higher mode effects,” Proc.
of the Eleventh World Conference on Earthquake Engineering, Disc 2, Paper No. 966.
Priestley, M. J. N. [2003] Myths and Fallacies in Earthquake Engineering, Revisited (IUSS
Press, Pavia, Italy).
Requena, M. and Ayala, G. [2000] “Evaluation of a simplified method for the determi-
nation of the nonlinear seismic response of RC frames,” Proc. of the Twelfth World
Conference on Earthquake Engineering, Auckland, New Zealand, Paper No. 2109.
Rovithakis, A. [2001] “Verification of adaptive pushover analysis procedures,” MSc Disser-
tation, Department of Civil and Environmental Engineering, Imperial College London,
United Kingdom.
Sasaki, K. K., Freeman, S. A. and Paret, T. F. [1998] “Multi-mode pushover procedure
(MMP) — A method to identify the effects of higher modes in a pushover analysis,”
Proc. of the Sixth US National Conference on Earthquake Engineering, Earthquake
Engineering Research Inst., Oakland, California, 12 pp.
SeismoSoft [2004] SeismoStruct — A Computer Program for Static and Dynamic Non-
linear Analysis of Framed Structures [online]. Available from URL: http://www.
seismosoft.com.
May 7, 2004 14:47 WSPC/124-JEE 00151
Valles, R. E., Reinhorn, A. M., Kunnath, S. K., Li, C. and Madan, A. [1996] IDARC2D
Version 4.0: A Computer Program for the Inelastic Analysis of Buildings, Technical
Report NCEER 96-0010, State University of New York at Buffalo, Buffalo, NY.
by UNIVERSIDAD DE CHILE SISTEMA DE SERVICIOS DE INFORMACION Y BIBLIOTECAS (SISIB) on 09/05/17. For personal use only.