SEISMIC ASSESSMENT OF BUILDINGS AND BRIDGES

USING NONLINEAR STATIC PROCEDURES

Pinho, R., MSc, PhD

Department of Structural Mechanics, University of Pavia, Pavia, Italy, [email protected]

ABSTRACT
Estimating seismic demands on structures requires explicit consideration of the structural inelastic behaviour: to this
end, the use of Nonlinear Static Procedures, commonly based on pushover analyses, is inevitably going to be favoured
over complex, impractical for widespread professional use, nonlinear time-history methods. However, in order to
warrant the attainment of reliable predictions of the seismic response of buildings and bridges, it is crucial that these
static procedures manage to somehow incorporate the effects of higher modes of vibration, particularly relevant in
irregular or long-period structures, and of the change in dynamic characteristics of structures as they are subjected to
higher levels of deformation. Analytical parametric studies on a suite of continuous multi-span bridges and framed
buildings, regular and irregular, were employed to scrutinise if conventional and recently proposed adaptive pushover
analytical algorithms, found at the heart of Nonlinear Static Procedures, can lead to the attainment of reliable seismic
response predictions, which would match closely results from dynamic nonlinear analysis. It will be shown that in
particular displacement-based adaptive pushover procedures may indeed lead to the attainment of significantly
improved predictions, which match very closely results from dynamic nonlinear analysis.
------------------------------------------------------------------------------------------------------------------------------------------------

1. INTRODUCTION errors are commonly encountered as the design/assessment

process evolves.
It is unquestionable that nonlinear dynamic analysis is the Thirdly, even in those situations where the expertise and
most accurate method for assessing the response of structures resources for running time-history analyses are available, it is
subjected to earthquake action. Indeed, any type of static often the case that preliminary simpler analysis (i.e. modal and
analysis will always be inherently flawed, given the static analyses) are run to enable a first check of the model;
conspicuous absence of time-dependent effects. However, as errors in the definition/assemblage of a finite elements model
noted by Goel and Chopra (2005a), amongst others, such type are difficult to detect from dynamic analysis results, whilst they
of analysis is not without its difficulties or drawbacks, tend to be relatively evident from the output of eigenvalue or
particularly for what concerns application within a design pushover runs. As an example, inspection of the first modes of
office environment. vibration of a given building model may be used to check if
Firstly, in order to employ dynamic analysis for seismic member stiffness has been correctly allocated or if the mass has
design/assessment of structures, an ensemble of site-specific been appropriately distributed, whilst examination of a force-
ground motions compatible with the seismic hazard spectrum displacement monotonic capacity curve may serve to quickly
for the site must be simulated. As described by Bommer and assess if member strength and ductility has been properly
Acevedo (2004), amongst others, this is, however, a far from assigned. Static analyses, even if representing simplified
simple task, since seismic design codes feature insufficient or methods, also provide important structural response information
inadequate guidance on procedures to either (i) generate such as (i) identification of critical regions, where large inelastic
artificial spectrum-compatible records, (ii) produce synthetic deformations may occur, (ii) individuation of strength
accelerograms from seismological models or (iii) select irregularities in plan and elevation that might cause important
appropriate suites of real acceleration time-series, eventually changes in the inelastic dynamic response characteristics, (iii)
modified to better fit a given code response spectrum. It is evaluation of the force demand in potentially brittle elements,
believed that until better guidance on record and (iv) prediction of the sequence of yielding and/or failure of
selection/generation will be made available to earthquake structural members. In addition, the explicit insight that
engineer designers, this first step will remain as a very pushover-derived base shear vs. top displacement capacity
difficult-to-overcome hurdle to the use of dynamic time- curves provide into the stiffness, strength and ductility of a
history analysis in design office applications. given structure, constitutes the type of qualitative data that is
Secondly, notwithstanding the significant increase in always most informative and useful within a design application,
computing power witnessed in recent years, nonlinear time- even when time-history analysis is then employed for the
history analysis remains computationally demanding, definitive verifications.
especially when fibre-based (distributed inelasticity) structural The above constitute, in the opinion of the author, strong
analysis programs, which are simpler to calibrate than their reasons for Nonlinear Static Procedures to continue to be
plastic-hinge (concentrated plasticity) counterparts, are developed and improved, so that these tools can become ever
employed to model the seismic response of large multi-storey more reliable and useful when employed either as a replacement
irregular buildings, requiring 3D models with thousands of to time-history analysis in the seismic design/assessment of
elements. This problem becomes even the more significant if relatively simple non-critical structures, or as a complement to
one considers that the analyses will need to be repeated a dynamic analysis of more complex/critical facilities.
significant amount of times, not only because design codes or
guidance documents request for a relatively large number of 2. PUSHOVER ANALYSIS
earthquake records to be employed in order to warrant
minimum probabilistic validity of the results, but also, and At the heart of Nonlinear Static Procedures, such as the
perhaps mainly, because the process of analysing any given Capacity Spectrum Method (CSM) introduced by Freeman an
structure is invariably an iterative one, given that modelling co-workers (1975, 1998), the N2 method suggested by Fajfar
and collaborators (1988, 1999), the Multimodal Pushover standard readily-available commercial software packages. The
Analysis (MPA) proposed by Chopra and Goel (2002) and the associated drawback, however, is that the methods are
Adaptive Capacity Spectrum Method developed by Casarotti inevitably more complex than running a single pushover
and Pinho (2007), lies a pushover analysis procedure. The analysis, as noted by Maison (2005). Furthermore, some of the
term pushover analysis describes a modern variation of the proposed “multiple-run” procedures, either they have an
classical collapse analysis method, as fittingly described by adaptive or non-adaptive nature, lead to difficulties when
Kunnath (2004). It refers to an analysis procedure whereby an applied within capacity-spectrum type of procedure due to the
incremental-iterative solution of the static equilibrium need to handle capacity curves associated with higher mode
equations is carried out to obtain the response of a structure force patterns that display a reversal of the roof displacement as
subjected to monotonically increasing lateral load pattern. The inelasticity develops in the structure (Hernàndez-Montes et al.
structural resistance is evaluated and the stiffness matrix is 2004; ATC 2005; Goel and Chopra 2005b). For all of the above,
updated at each increment of the forcing function, up to these multiple-pushover based approaches do not constitute the
convergence. The solution proceeds until (i) a predefined scope of the current presentation, where focus is placed instead
performance limit state is reached, (ii) structural collapse is on “single-run” pushover analysis procedures.
incipient or (iii) the program fails to converge. In this manner, In tandem with the present drive for performance-based
each point in the resulting displacement vs. base shear seismic engineering, there is also currently a thrust for the
capacity curve represents an effective and equilibrated stress- development and code implementation of displacement or, more
state of the structure, i.e. a state of deformation that bears a generally, deformation-based design and assessment methods.
direct correspondence to the applied external force vector. Therefore, it would seem that applying displacement loading,
Within the framework of earthquake engineering, pushover rather than force actions, in pushover procedures would be an
analysis is employed with the objective of deriving, with appropriate option for nonlinear static analysis of structures
relative ease, an envelope of the response parameters that would subjected to earthquake action. However, due to the unvarying
otherwise be obtained through a much more complex and time- nature of the applied displacement loading vector, conventional
consuming Incremental Dynamic Analysis (IDA) procedure, as (non-adaptive) displacement-based pushover analysis can
can be construed by Figure 1. IDA is a parametric analysis conceal important structural characteristics, such as strength
method by which a structure is subjected to a series of nonlinear irregularities and soft storeys, should the displacement pattern
time-history analyses of increasing intensity (Vamvatsikos and adopted at the start of the analysis not correspond to the
Cornell, 2002), with the objective of attaining an accurate structure’s post-yield failure mechanism. Consequently, when
indication of the “true” dynamic response of a structure only non-adaptive static nonlinear analysis tools are available,
subjected to earthquake action. as has been the case throughout the past, force-based pushover
does constitute a preferable choice over its displacement-based
120
counterpart.
Max Ba se She ar (kN)

100
On the other hand, however, if one is able to apply
displacements, rather than forces, in an adaptive fashion, that is,
80 with the possibility of updating the displacement loading pattern
according to the structural properties of the model at each step
60 of the analysis, then a conceptually appealing deformation-
based nonlinear static analysis tool is obtained.
40

20 3. RECENT DEVELOPMENTS IN SINGLE-RUN

PUSHOVER ANALYSIS
0
0 0 .5 1 1.5 2 2.5 3
According to recently introduced code provisions, such as
drift (%) FEMA-356 (BSSC, 2000) and Eurocode 8 (CEN, 2002),
Figure 1. Maximum base-shear and top displacement values pushover analysis should consist of subjecting the structure to
obtained with incremental dynamic analysis. an increasing vector of horizontal forces with invariant
pattern. Both the force distribution and target displacement are
Recent years have also witnessed the development and based on the assumptions that the response is controlled by the
introduction of an alternative type of nonlinear static analysis, fundamental mode and the mode shape remains unchanged
which involves running multiple pushover analyses separately, until collapse occurs. Two lateral load patterns, namely the
each of which corresponding to a given modal distribution, and first mode proportional and the uniform, are recommended to
then estimating the structural response by combining the action approximately bound the likely distribution of the inertia
effects derived from each of the modal responses (i.e. each forces in the elastic and inelastic range, respectively.
displacement-force pair derived from such procedures does not However, a number of recent studies, summarised in the
actually correspond to an equilibrated structural stress state). FEMA-440 report (ATC, 2005), raise doubts on the
Paret et al. (1996) first suggested the Multi-Modal Pushover effectiveness of these conventional force-based pushover
procedure, which was then refined by Moghadam and Tso methods in estimating the seismic demand throughout the full
(2002). Chopra and Goel (2002), on the other hand, have deformation range: (i) inaccurate prediction of deformations
developed and proposed a Modal Pushover Analysis (MPA) when higher modes are important and/or the structure is highly
technique, which Hernández-Montes et al. (2004) have then pushed into its nonlinear post-yield range, (ii) inaccurate
adapted into an Energy-based Pushover formulation. A further prediction of local damage concentrations, responsible for
refinement of such multiple-pushover procedures, with the aim changing the modal response, (iii) inability of reproducing
to account for the alteration of local resistance and modal peculiar dynamic effects, neglecting sources of energy
characteristics of the structure induced by the accumulation of dissipation such as kinetic energy, viscous damping, and
damage, consists in the employment of adaptive updating of the duration effects, (iv) difficulty in incorporating three-
loading pattern (Gupta and Kunnath, 2000; Aydinoglu, 2003), dimensional and cyclic earthquake loading effects.
effectively meaning that the methods may now be categorised as In Figure 2 and Figure 3, examples of inadequate prediction
piecewise linear response spectrum analysis. As highlighted by of both the capacity curve as well as the deformation response
their respective authors, the main advantage of this category of characteristics of a 12-storey reinforced concrete frame
static analysis procedures is that they may be applied using subjected to a natural earthquake recording (case-study RM15-
NR2 in Antoniou and Pinho (2004a)) and of a 4-storey irregular depending on the shape of the response spectrum being
frame subjected to an artificial accelerogram (ICONS full-scale considered (or on the frequency content of an input record), may
test specimen, described in Pinho and Elnashai (2000)) are trigger significant changes in the response characteristics of the
given. It is noted that although the 12-storey building is regular buildings (Figure 5). Krawinkler and Seneviratna (1998)
in height, its response is heavily influenced by higher mode summarised the above with a single statement: fixed load
effects, effectively rendering its seismic behaviour highly patterns in pushover analysis are limiting, be they first modal or
irregular in height, as conspicuously shown by Figure 2a. The multimodal derived, because no fixed distribution is able of
standard pushover results have been carried out using both representing the dynamic response throughout the full
triangular and uniform loading distributions, and are compared deformation range.
with the envelope of results obtained with incremental dynamic
analysis. 6

period (sec)
first mode
second mode
7000
base shear (kN)

5 third mode

6000
4

5000
3
4000
2
3000 dynamic
DAP
uniform
2000 1
triangular

1000
0
0.0% 0.5% 1.0% 1.5% 2.0%
0 total drift
0% 1% 2% 3%
total drift Figure 4. Periods of vibration of 4-storey building under
Figure 2. Capacity curves of a 12-storey building, obtained increasing levels of deformation.
with standard pushover.
12
12 DYNAMIC
TRIANGULAR 10
UNIFORM
10
8

8 6 top drift = 0.5%

STOREY

top drift = 1.0%

6 4 top drift = 1.5%
top drift = 2.5%
4 2

0
2
0.0% 0.5% 1.0% 1.5% 2.0% 2.5% 3.0% 3.5% 4.0%

0
Figure 5. Interstorey drift profiles of a 12-storey building
0.00% 2.00% 4.00% 6.00% subjected to increasing levels of deformation.
INTERSTOREY DRIFT (% )
a) As a result of the aforementioned limitations, recent years
4 DYNAMIC have witnessed the development and introduction of so-called
TRIANGULAR Adaptive Pushover methods whereby the loading vector is
UNIFORM
updated at each analysis step, reflecting the progressive stiffness
3 degradation of the structure induced by the penetration in the
inelastic range (Figure 6). The response of the structure is thus
computed in incremental fashion, through piecewise
STOREY

2 linearization (Figure 7), hence rendering it possible to use the

tangent stiffness at the start of each increment, together with the
mass of the system, to compute modal response characteristics
1 of each incremental pseudo-system through elastic eigenvalue
analysis, and use such modal quantities to congruently update
(i.e. increment) the pushover loading vector.
0 Force-based adaptive pushover procedures have been
0.00 0.50 1.00 1.50 2.00
proposed by Reinhorn (1997), Bracci et al. (1997), Satyarno et
INTERSTOREY DRIFT (% )
al. (1998), Requena and Ayala (2000), Elnashai (2001) and
b) Antoniou et al. (2002). With the exception of the work of
Figure 3. Interstorey drift profiles of: a) 12-storey building and Satyarno et al. (1998), where a single mode adaptive pushover
b) 4-storey irregular frame, obtained with standard pushover. pattern was employed, all other adaptive methodologies
considered the effects of the higher modes and of the input
The main reason behind the underperformance of these frequency content. Furthermore, Elnashai (2001) and Antoniou
conventional pushover methods is the fact that they do not et al. (2002) implemented their adaptive algorithm within a fibre
account for the effect that damage accumulation, induced by the analysis framework, allowing for a continuous, rather than
increasing deformation levels imposed on the structure, has on discrete, force distribution update to be carried out.
the response of the latter. Cumulative material straining
introduces a reduction in stiffness, which, in turn, causes an
elongation of the periods of vibration (Figure 4), which then,
 F t+ ∆ F t
5000 Ft
base shear (kN) O 't Kt
4000
F k+ ∆ F k Kk
Fk
3000 O 'k

2000

1000

0
0% 1% 2% 3%
total drift O dk dk+ ∆dk dt d t+ ∆ d t

Figure 6. Adaptive pushover: shape of loading vector Figure 7. The use of tangent stiffness in updating (i.e.
is updated at each analysis step. incrementing) the loading vector.

Despite their apparent conceptual superiority, or at least the jth mode, φi,j is the mass normalised mode shape value for
despite their conspicuously more elaborated formulation, the the ith storey and the jth mode, Mi is the mass of the ith storey
improvement introduced by such Force-based Adaptive and Saj represents the acceleration response spectrum ordinate
Pushover (FAP) procedures was not-necessarily impressive, corresponding to the period of vibration of the jth mode. Finally,
with respect to its traditional non-adaptive counterparts, αj is a weighting factor that aims at accounting for the varying
particularly in what concerns the estimation of deformation relative importance that each mode j has on the maximum
patterns of buildings, which seemed to be poorly predicted by response of the structure.
both types of analysis, as shown in Figure 8. As described by The employment of such alternative modal combination
Kunnath (2004) and Antoniou and Pinho (2004a), the main procedure, may indeed lead to the attainment of improved
reason for such underperformance seems to be the quadratic results, as demonstrated by the interstorey drift profiles given in
modal combination rules (e.g. SRSS, CQC) used in computing Figure 9, obtained through consideration of the first three modes
the adaptive updating of the load vector; such rules will of vibration of the buildings, and using α1 = 1.0, α2 = -1.0 and
inevitably lead to monotonically increasing load vectors, since α3 = 1.0 in Eq. 1. However, the arbitrary nature of these
the possibility of sign change in applied loads at any location is weighting factors αj renders the method unfeasible for practical
precluded, whilst it may be needed to represent the uneven application, as explicitly acknowledged in Kunnath (2004) and
redistribution of forces after an inelastic mechanism is triggered demonstrated in López-Menjivar (2004). Indeed, in the latter
at some location. work it is demonstrated how values of αj that lead to optimum
With the above in mind, Kunnath (2004) and López- results for some building configurations, lead then to poor
Menjivar (2004) have proposed an alternative modal predictions in buildings with diverse characteristics. Therefore,
combination scheme, consisting of a weighted Direct Vectorial and until a general procedure to correctly determine the values
Addition (DVA) of the different modal shapes that can be of the weighting factors is found, the DVA adaptive pushover
mathematically expressed as: modality cannot really be deemed as a valid solution for
practical application.
n
Fi = ∑ α j Γ jφ j ,i M j Sa j (Eq. 1)
j =1 4. DISPLACEMENT-BASED ADAPTIVE
PUSHOVER (DAP)
where i is the storey number, j is the mode number, n is the
highest mode of interest, Γj is the modal participation factor for With a view to overcome all the limitations described
above, Antoniou and Pinho (2004b) have proposed a paradigm

12 DYNAMIC 4 DYNAMIC

FAP-SRSS FAP-SRSS
10
3
8
STOREY

STOREY

6 2

4
1
2

0 0
0.00% 2.00% 4.00% 6.00% 0.00 0.50 1.00 1.50 2.00
INTERSTOREY DRIFT (% ) INTERSTOREY DRIFT (% )

a) b)
Figure 8. Interstorey drift profiles of : a) 12-storey building and b) 4-storey irregular frame, obtained with
Force-based Adaptive Pushover using SRSS modal combination.
 12 DYNAMIC 4 DYNAMIC
FAP-DVA FAP-DVA
10

3
8
STOREY

STOREY
6 2

4
1
2

0 0
0.00% 2.00% 4.00% 6.00% 0.00 0.50 1.00 1.50 2.00
INTERSTOREY DRIFT (% ) INTERSTOREY DRIFT (% )
a) b)
Figure 9. Interstorey drift profiles of: a) 12-storey building and b) 4-storey irregular frame, obtained with
Force-based Adaptive Pushover using DVA modal combination.

shift in pushover analysis, by introducing the innovative an eigenvalue analysis is carried out. To this end, the Lanczos
concept of Displacement-based Adaptive Pushover (DAP). algorithm (Hughes, 1987) is employed to determine the modal
Contrarily to what happens in non-adaptive pushover, where shape and participation factors of any given predefined number
the application of a constant displacement profile would force of modes. Modal loads can be combined by using either the
a predetermined and possibly inappropriate response mode, Square Root of the Sum of the Squares (SRSS) or the Complete
thus concealing important structural characteristics and Quadratic Combination (CQC) methods.
concentrating inelastic mechanisms at a given location, within Since application to the analysis of buildings is the scope of
an adaptive framework, a displacement-based pushover is the present work, use is made of the interstorey drift-based
entirely feasible, since the loading vector is updated at each scaling algorithm, whereby maximum interstorey drift values
step of the analysis according to the current dynamic obtained directly from modal analysis, rather than from the
characteristics of the structure. difference between not-necessarily simultaneous maximum
floor displacement values, are used to compute the scaling
displacement vector. This comes as a reflection of the fact that
4.1. DAP – METHODOLOGY the maximum displacement of a particular floor level, being
essentially the relative displacement between that floor and the
The implementation of DAP can be structured in four ground, provides insufficient insight into the actual level of
main stages; (i) definition of the nominal load vector and damage incurred by buildings subject to earthquake loading. On
inertia mass, (ii) computation of the load factor, (iii) the contrary, interstorey drifts, obtained as the difference
calculation of the normalised scaling vector and (iv) updating between floor displacements at two consecutive levels, feature a
of the loading displacement vector. Whilst the first step is much clearer and direct relationship to horizontal deformation
carried out only once, at the start of the analysis, its three demand on buildings. Readers are referred to Antoniou (2002)
remaining counterparts are repeated at every equilibrium stage for further details on this formulation.
of the nonlinear static analysis procedure, as described in the In such an interstorey drift-based scaling technique, the
following subsections. eigenvalue vectors are thus employed to determine the
The loading vector shape is automatically defined and interstorey drifts for each mode ∆ij, as shown in Eq. 3, while the
updated by the solution algorithm at each analysis step, for displacement pattern Di at the ith storey is obtained through the
which reason the nominal vector of displacements, U0, must summation of the modal-combined inter-storey drifts of the
always feature a uniform (rectangular) distribution shape in storeys below that level, i.e. drifts ∆1 to ∆i:
height, so as not to distort the load vector configuration
determined in correspondence to the dynamic response i
characteristics of the structure at each analysis step. In addition, Di = ∑ ∆ k (Eq. 3)
it is noteworthy that the adaptive pushover requires the inertia k =1
mass M of the structure to be modelled, so that the eigenvalue
analysis, employed in updating the load vector shape, may be
[ ]
n n
∑ Γ j (φi, j − φi −1, j ) 2
with ∆i = 2
carried out. ∑ ∆ij =
The magnitude of the load vector U at any given analysis j =1 j =1
step is given by the product of its nominal counterpart U0,
defined above, and the load factor λ at that step (see Eq. 2). The
Since only the relative values of storey displacements (Di)
latter is automatically increased, by means of a load control
are of interest in the determination of the normalised modal
strategy (Antoniou and Pinho, 2004a), until a predefined
analysis target, or numerical failure, is reached. scaling vector D , which defines the shape, not the magnitude,
of the load or load increment vector, the displacements obtained
U = λ⋅U0 (Eq. 2) by Eq. 3 are normalised so that the maximum displacement
remains proportional to the load factor, as required within a load
control framework:
The normalized modal scaling vector, D , used to determine
the shape of the load vector (or load increment vector) at each Di
step, is computed at the start of each load increment. In order for Di = (Eq. 4)
such scaling vector to reflect the actual stiffness state of the max Di
structure, as obtained at the end of the previous load increment,
 Once the normalised scaling vector and load factor have 2000

been determined, and knowing also the value of the initial Experimental
Analytical 1500
nominal load vector, the loading vector Ut at a given analysis
step t is obtained by adding to the load vector of the previous 1000

step, Ut-1 (existing balanced loads), a newly derived load vector

500
increment, computed as the product between the current load

kN
factor increment ∆λt, the current modal scaling vector D t and -0.4 -0.3 -0.2 -0.1
0
0 0.1 0.2 0.3 0.4
the nominal vector U0, as mathematically translated into Eq. 5 -500
and graphically depicted in Figure 10.
-1000

Ut = Ut-1+∆λt⋅ Dt ⋅U0 (Eq. 5)

-1500

normalised shape
normalised nominal new increment -2000
at step
shape t t
at step load vector of forces
of displacements m

∆Ptt = ∆λt ×
∆U × =
Figure 12. Verification of fibre element analytical models of a
2-storey steel frame (Pietra, 2006).

balanced displacem
existing forces ents
new increment
of displacements
of forces new
newdisplacements
forces applied applied
at step tat step t
4.2. DAP – ILLUSTRATIVE RESULTS

UPt=U t-1+∆U
t=Pt-1 +∆Pt= + = Two clearly distinct building frames, both of which
featuring an irregular type of dynamic response, are
considered herein. The first of these is a 12-storey five-bay
Figure 10. Updating of the loading displacement vector. reinforced concrete structure designed according to Eurocode
8 (2002). It displayed a highly irregular dynamic behaviour
The DAP algorithm has been implemented in the computer (Figure 3) when subjected to an accelerogram (Hollister
code SeismoStruct (SeismoSoft 2005), a fibre element-based station, Loma Prieta earthquake, USA, 1989) that presented a
program for seismic analysis of framed structures, very high amplification in the short-period and thus lead to a
downloadable from the internet. The program incorporates both response very much dominated by the 2nd and 3rd modes of
local (beam-column effects) and global (large vibration. Indeed, and as can be observed in Figure 13, these
displacements/rotations effects) sources of geometric two higher modes (0.15 < T2, T3 < 0.30 secs) feature a
nonlinearity as well as the interaction between axial force and spectral amplification, in acceleration, that is ten times higher
transverse deformation of the element. The spread of material than that corresponding to first mode of vibration (T1 > 1.4
inelasticity along the member length is explicitly represented secs). Further details on this case-study can be found in
through the employment of a fibre modelling approach, implicit Antoniou et al. (2002).
in the formulation of the inelastic beam-column frame elements
adopted in the analyses. Various verification studies have been 0.5

carried out with the aforementioned program on a four-storey

reinforced concrete frame (Figure 11a), a reinforced concrete 0.4
bridge (Figure 11b) and a two-storey steel frame (Figure 12) all
acceleration (g)

of which show the ability of the analytical models to replicate 0.3

the seismic response of full-scale structures.
0.2
80

60
0.1
40
Displacement [mm]

20
0.0
0 0 1 2 3
-20 period (sec)
-40 a)
Analytical 5
-60 Experimental

-80
0 2 4 6 8 10 12 14
T ime [s] 4
displacement (cm)

a)
60 3

45
Experimental Numerical 2
Top Displacement (mm)

30

15 1

0
0
-15 0 1 2 3
period (sec)
-30
b)
-45 Figure 13. a) Acceleration and b) displacement response
-60
spectra of accelerogram employed in the analysis of 12-storey
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 building.
Time (s)
b) The second structure is a 4-storey three-bay building refers
Figure 11. Verification of a fibre-element analytical model of: to a full-scale test specimen, built to represent typical design and
a) 4-storey RC frame (López-Menjivar, 2004) and b) RC construction practice in most South-European countries in the
bridge (Casarotti et al., 2005) 1950's, and tested under pseudo-dynamic conditions (Pinho and
 12 DYNAMIC 4 DYNAMIC

DAP-SRSS DAP-SRSS
10
3
8
STOREY

STOREY
6 2

4
1
2

0 0
0.00% 2.00% 4.00% 6.00% 0.00 0.50 1.00 1.50 2.00
INTERSTOREY DRIFT (% )
INTERSTOREY DRIFT (% )
a) b)
Figure 14. Interstorey drift profiles of: a) 12-storey building and b) 4-storey irregular frame, obtained with
Displacement-based Adaptive Pushover using SRSS combination..

Elnashai, 2000) at the JRC in Ispra (Italy). The frame was equilibrium to the applied displacement pattern, thus allowing
designed for gravity loads only, without any consideration of for the reproduction of reversal of storey shear distributions,
ductility provisions or capacity design principles. Consequently, observed in dynamic analysis, even if a quadratic rule is
it exhibited a soft-storey type of deformation mechanism at the employed to combine the contribution of the different modes.
third storey level (e.g. Figure 3) caused mainly by the drastic In effect, DAP drift profiles, despite carrying a permanently
stiffness/strength variation present at such location, as well as by positive sign, do, in any case, feature changes of their respective
inadequate lap-splicing and defective column shear capacity. gradient (i.e. the trend with which drift values change from one
The input motion consisted of artificial accelerograms aiming at storey to the next), introduced by the contribution of higher
being representative of European seismicity (Campos-Costa and modes. When such gradient variations imply a reduction of the
Pinto, 1999). drift of a given storey with respect to its adjacent floor levels,
In Figure 14, the interstorey drift profiles of these two case- then the corresponding applied storey horizontal force must also
studies, as obtained with the employment DAP analyses, are be reduced, in some cases to the extent of sign inversion, as
given. It is observed that the predictions now match much closer observed in Figure 16.
the dynamic response of these two structures, which effectively
means that the response irregularities caused by the flexibility of
12
the 12-storey structure, and subsequent amplification of higher
modes, as well as the strength irregularity of the 4-storey 11
adaptive
prototype, have been fully and correctly captured by the 10 uniform
proposed static analysis algorithm. 9 triangular
In Figure 15, on the other hand, the capacity curves of the 8
12-storey building, as derived by both DAP and standard 7
pushover curves are compared with the Incremental Dynamic
6
Analysis envelope. The advantages of using an adaptive
5
displacement-based pushover can be inferred also from this type
of results. 4

The reason behind the most-improved predictions obtained 3

with the displacement-based adaptive pushover procedure is the 2
fact that storey forces or shears are no longer applied directly to 1
the structure but rather come as a result of structural
-500 0 500 1000 1500 2000
7000
storey shear (kN)
Figure 16. Storey shear distributions of a 12-storey building
base shear (kN)

6000 obtained with Displacement-based Adaptive Pushover as well

as with standard non-adaptive pushovers.
5000

In other words, given that in DAP, shear distributions are

4000 automatically derived to attain structural equilibrium with the
imposed storey drifts, rather than being a result of the loads
3000 directly applied to the structure, the previously described
limitations evidenced by force-based adaptive schemes that use
2000
dynamic
quadratic modal combination rules can be overcome and,
adaptive consequently, results as whole (i.e. deformation profiles and
uniform
1000
triangular capacity curves) become more accurate.
0
0% 1% 2% 3% 4.3. DAP – EASE-OF-USE, COMPUTATIONAL
total drift EFFORT AND NUMERICAL STABILITY
Figure 15. Capacity curves of a 12-storey building, obtained
with DAP and standard pushovers, and compared against IDA When compared with nonlinear time-history analysis,
envelopes. pushover methods are advantaged by their (i) higher user-
friendliness, (ii) reduced running time and (iii) increased
numerical stability. Therefore, it is important that the proposed
coupling beam
displacement-based algorithm, capable of producing improved
structural response predictions in comparison with existing

12 × 3.0 = 36.0 m
structural
non-adaptive pushover techniques, does also feature these

4.5+7× 3.0 = 25.5 m

wall

8 × 3.0 = 24.0 m
three advantages over dynamic analysis.
From a usability point-of-view, the proposed displacement-
based adaptive pushover algorithm effectively presents no
additional effort and/or requirements with respect to its
conventional non-adaptive counterparts. In effect, the only
element of novelty, in terms of analysis input, is the introduction REGULARSYSTEM IRREGULARSYSTEM DUAL SYSTEM

of the building’s inertia mass, which, however, can readily be Figure 17. Geometric characteristics of the regular, irregular and
obtained directly from the vertical gravity loads, already dual systems.
included in any type of pushover analysis.
With regards to computational effort, in general, the amount Table 1. Considered building systems
of time necessary to complete an adaptive pushover analysis is Structural Storeys Structure Ductility Design Behavior Tuncracked
typically double the time necessary for a conventional System (Height) Reference Level PGA (g) Factor (q) (s)
procedure, approximately. Obviously, the duration of such finite RH30 High
0.30
5.00 0.697
Regular RM30 Medium 3.75 0.719
element runs will vary according to the computing capacity of 12 (36 m)
Frame RM15 Medium 3.75 0.745
the workstation being used, as well as with the characteristics of RL15 Low
0.15
2.50 0.740
the model (mainly the number of elements and level of fibre IH30 High 4.00 0.565
0.30
discretisation of the sections). In any case, adaptive pushover Irregular
8 (25.5 m)
IM30 Medium 3.00 0.536
Frame IM15 Medium 3.00 0.613
proved to be up to ten times quicker than nonlinear dynamic 0.15
IL15 Low 2.00 0.614
analysis of a same model (keeping in mind that fibre-based WH30 High 3.50 0.569
0.30
finite element modelling has been adopted for the current work), Regular WM30 Medium 2.625 0.557
8 (24 m)
hence the time-advantage of static methods versus their dynamic Wall-Frame WM15 Medium
0.15
2.625 0.601
RH30 High 5.00 0.697
counterparts is not lost with the addition of the adaptive
features.
As far as numerical stability is concerned, no particular 1.0
problems have been observed in the studies described above,
and those given in subsequent sections, noting that structures 0.8

were pushed well into their post-peak inelastic response range.

acceleration (g)

0.6

5. VERIFICATION PARAMETRIC STUDIES 0.4

0.2
In this Section, an analytical comparative study involving
different pushover methods, either single or multi mode, 0.0
0 1 2 3
adaptive or conventional, and dynamic nonlinear analysis of period (sec)

buildings and bridges is presented. The “true” dynamic (a) Record AR

1.0
response is deemed to be represented by the results of the
Incremental Dynamic Analysis procedure (IDA) (e.g. 0.8

Vamvatsikos and Cornell, 2002), which is a parametric

acceleration (g)

0.6
analysis method by which a structure is subjected to a series of
nonlinear time-history analyses of increasing intensity. 0.4

0.2
5.1. REINFORCED CONCRETE FRAMES
0.0
0 1 2 3
Characteristics of input motion and structural models. period (sec)

Three different configurations of common RC structures were (b) Record NR1

employed: a 12-storey regular frame, an eight storey irregular 0.5

frame and a dual wall-frame system (Figure 17). The latter are
0.4
based on buildings previously designed for different ductility
acceleration (g)

classes and design ground acceleration, on medium soil type 0.3

‘B’ of EC8 (Fardis, 1994), resulting in a total of 12 models, as

0.2
described in Table 1. The overall plan dimensions of the three
configurations are 15 m by 20 m. The storey height is 3 m 0.1

except the first storey of the irregular set, which is 4.5 m high.
0.0
A detailed description of models and load conditions, as well 0 1 2 3
period (sec)
as of their FE modelling, can be found in López (2004). (c) Record NR2
Four input time-histories, consisting of one-artificially 3.0

generated accelerogram (A975) and three natural records (Loma 2.4

Prieta earthquake, USA, 1989), were employed: the selection of
acceleration (g)

these four records aimed at guaranteeing a wide-ranging type of 1.8

earthquake action, in terms of frequency content, peak ground 1.2

acceleration, duration and number of high amplitude cycles.
Their acceleration response spectra are shown in Figure 18. 0.6

Upper and lower bounds of the main characteristics of the 0.0

records are summarised in Table 2, where the significant 0 1 2 3
period (sec)
duration is defined as the interval between the build up of 5% (d) Record NR3
and 95% of the total Arias Intensity (Bommer and Martinez- Figure 18. Elastic response spectra of the four records (5%
Pereira, 1999). equivalent viscous damping).
 Table 2. Bounding characteristics of the employed set of 12
DAP-SRSS
records for buildings DYNAMIC
TRIANGULAR
5% Arias Significant Total 10
UNIFORM
Peak Ground Peak Response
Intensity Duration Duration teff / ttot
Acceleration Acceleration
threshold teff ttot
8
Min 0.12 g 0.50 g 1.02 s 7.24 s 10.0 s 22.3%
Max 0.93 g 4.25 g 11.23 s 10.43 s 40.0 s 72.4%

STOREY
6

Analyses and results post-processing. The two non-adaptive

pushover schemes, proposed in the NEHRP Guidelines (ATC, 4

1997), were applied to each set of buildings: the uniform

distribution, whereby lateral forces are proportional to the total 2
mass at each floor level, and the triangular distribution, in
which seismic forces are proportional to the product of floor 0
mass and storey height. The adaptive pushover algorithm was 0.00% 1.00% 2.00% 3.00% 4.00% 5.00% 6.00%

used in both its force and displacement-based variants, with INTERSTOREY DRIFT

spectrum scaling, employing SRSS or CQC modal Figure 19. Representative results obtained with model RM15
combination rules. subjected to one of the natural accelerograms employed in this
The inter-storey drift profiles obtained from each pushover study (NR2).
analysis are compared to the drift profiles from the nonlinear
dynamic analysis and the standard error of the pushover NF accelerograms, on the other hand, consist of motions
results, with respect to the dynamic, is calculated as: recorded close to the epicentre and which contain very strong
velocity pulses, originally included in the FEMA-440 work
with the objective of verifying, at least in preliminary fashion,
1 n ⎛ ∆iD − ∆iP ⎞ the validity of employing pushover methods in areas where
Error (%) = 100 ∑⎜ ⎟ (Eq. 6)
n i =1⎜⎝ ∆iD ⎟
⎠ pulse-like near-fault ground motions are likely to occur. For
further details the reader is referred to ATC (2005) and
Somerville et al. (1997).
The interstorey drift profiles are monitored at four different
deformation levels: the pre-yield state (0.5 % total drift), the 1
point of global yielding (1.0 % and 1.5 %), where the stiffness A1
changes significantly and the local distributions are rapidly A2
updated, and the deeply inelastic range (2.5 %). 0.8 A3
The Standard Error of the non-adaptive and adaptive A4
pushover schemes was computed for all the structures and 0.6
A5
earthquake records considered. In order to identify the presence A6
Sd [m]

of possible response peculiarities introduced by individual input A7

0.4 A8
motions but smoothed out through results averaging, the
A9
standard error was computed separately for each time history A10
analysis, as a unique value, averaging the standard error of all 0.2 A11
the storeys, in the building, and deformation levels.
0
Obtained results. The Mean Standard Error of the DAP, 0 0.5 1 1.5 2 2.5
T [s]
FAP, Triangular and Uniform pushovers, considering all
structures and ground motions, are 19.11 %, 30.90 %, 21.11 % a)
and 38.76 %, respectively. These overall results seem to
ERZMV1
indicate only a marginal advantage of DAP with respect to
LUCMV1
non-adaptive triangular distribution. However, a closer RRSMV1
inspection of interstorey drift profiles (Figure 19) for some SCHMV1
particularly difficult cases, renders much more conspicuous
the gains provided by the employment of displacement-based
adaptive pushover in the prediction of the seismic
demand/capacity of framed buildings subjected to seismic
action.

5.2. STEEL FRAMES

Characteristics of input motion and structural models. In

order to enable a direct comparison with the extensive and 0 0.5 1 1.5 2 2.5
T [s]
most thorough parametric study described in FEMA-440
(ATC, 2005), the same structural models and earthquake b)
records that have been used in such work have been adopted Figure 20. Ground Motions: unscaled displacement response
here. Hence, considered ground motions include Near-Field spectrum for: a) Ordinary- and b) Near-Field-type records
type records (NF records) as well as ordinary records (ATC, 2005).
(Ordinary Ground Motions, OGMs). Their main properties are
summarized in Table 3, whilst displacement response spectra The prototype buildings analyzed in the current endeavour
are represented in Figure 20. OGMs (11 records) consist of consist therefore of two steel moment-resisting frames (nine and
site class C accelerograms selected from strong-motion three storeys, Figure 21 and Figure 22) designed as a part of the
records that do not present near-fault or near-field FEMA-funded SAC joint venture project (Gupta and
characteristics (strong velocity pulses, short duration, high Krawinkler, 1999). These frames were considered in both
frequency content, etc.). regular and irregular (weak-storey at the ground floor)
configurations, thus leading to a total of four frames.
 Table 3. Ground Motion characteristics (ATC, 2005)
Identifier EQ Ms Station Component PGA PGV Effective Source
[g] [cm/s] Length[s]
Ordinary Ground Motions
A1 Superstitn 6.6 El Centro Imp Co. Cent(01335) 000 0.358 46.4 23 CDMG
11-24-87
A2 Northridge 6.7 Canyon Country-W Lost Cany 000 0.41 43 11 USC
1-17-94 (90057)
A3 Loma Prieta 7.1 Gilroy Array #2 (47380) 090 0.322 39.1 14 CDMG
10-18-89
A4 Chi-Chi, 7.6 (TCU122) N 0.261 34 35 CWB
Taiwan
8-20-99
A5 Loma Prieta 7.1 Gilroy Array #3 (47381) 090 0.367 44.7 17 CDMG
10-18-89
A6 Northridge 6.7 Canoga Park-Topanga Can (90053) 196 0.42 60.8 14 USC
1-17-94
A7 Chi-Chi, 7.6 (CHY101) W 0.353 70.6 32 CWB
Taiwan
8-20-99
A8 Superstitn 6.6 El Centro Imp Co. Cent(01335) 090 0.258 40.9 27 CDMG
11-24-87
A9 Northridge 6.7 Canoga Park-Topanga Can (90053) 106 0.356 32.1 16 USC
1-17-94
A10 Imperial 6.9 El Centro Array #2 (5115) 140 0.315 31.5 17 USGS
Valley
10-15-79
A11 Imperial 6.9 El Centro Array #11 (5058) 230 0.38 42.1 18 USGS
Valley
10-15-79
Near-Field Ground Motions
ERZ Erzican 6.9 Erzican Station NA 0.442 126 7 EERL
3-13-92 Caltech
LUC Landers 7.3 Lucerne Valley Station 280 0.732 147 14 EERL
6-28-92 Caltech
RRS Northridge 6.7 Rinaldi Receiving Station 213 0.891 186 6.5 EERL
1-17-94 Caltech
SCH Northridge 6.7 Sylmar County Hospital Parking 190 0.865 138 5.5 EERL
1-17-94 Lot Caltech

Figure 22. 3-Storey frame considered in this study (ATC,

2005).

approach has been tested in the case of OGMs, consisting in

the employment of the average response spectrum of all
records to compute the modal spectral amplification that is
considered in the calculation of the incremental adaptive
loading vector (Displacement-based Adaptive Pushover with
Average Spectrum scaling [DAP-AS]). Whilst such procedure
might be adopted in the case of OGMs, where records have
been scaled to meet the same target displacement, it cannot be
Figure 21. 9-Storey frame considered in this study used for NF records, which were employed without scaling and
(ATC, 2005). thus involve different drift responses for the same prototype
Analyses and results post-processing. The pushover schemes building.
considered include invariant static load patterns, such as (i) Verification of the pushover algorithm at the “global level”
uniform [Uniform], (ii) inverted triangular [Triangular], (iii) was carried out through comparisons between the base shear vs.
first mode shape [1st Mode], (iv) a code-specified period- top floor IDA envelopes (assumed as representative of the true
dependent distribution (where lateral forces changes from a behaviour of the frame) and the pushover curves. With the
linear distribution for low period systems to a parabolic shape objective of assessing also the accuracy of DAP in estimating
for more flexible models) [Code], and the Displacement-based local response parameters, the dynamic response at different
Adaptive Pushover [DAP]. In addition, an alternative adaptive ductility demand levels has been considered, where each
 100

E1 - 3S [%]
75 Mean
St.Dev.
50
25
0
100
E1 - 9Sw [%]

75
50
25
0
Code

Code

Code

Code
DAP

DAP

DAP

DAP
Uniform

Uniform

Uniform

Uniform
DAP-AS

DAP-AS

DAP-AS

DAP-AS
1st Mode

1st Mode

1st Mode

1st Mode
Triangular

Triangular

Triangular

Triangular
(a) Floor Displacement (b) Interstorey Drift ( (c) Interstorey Shear (d) Interstorey Moment
Figure 23. Mean values and standard deviation of error measure E1, averaged along the height of buildings
models 3S and 9Sw at the intermediate drift level (top floor drift level of 2%).

Ductility Level (DL) is identified by means of the total drift bearings featuring a linear elastic response. The total number
value (i.e. top floor displacement/building height). Three of bridges is therefore twelve, as shown in Figure 24, where
different drift levels are assumed (0.5, 2 and 4%) and each the label numbers 1, 2, 3 characterise the pier heights of 7 m,
ordinary record has been scaled in order to get the predefined 14 m and 21 m, respectively.
drift for each prototype building. Response parameters of A sufficiently large number of records has been employed
interest (displacement, drift, shear and moment) recorded in so as to bound all possible structural responses. The employed
time-history analyses are then compared, at each level location, set of seismic excitation is defined by an ensemble of 14 large
with those predicted by the pushover procedures at the same magnitude (6 ÷ 7.3) small distance (13 ÷ 30 km) records
roof displacement magnitude. For the 9-Storey weak frame selected from a suite of historical earthquakes scaled to match
system an ultimate drift level of 2.7% has been selected the 10 % probability of exceedance in 50 years uniform hazard
(corresponding to a top floor displacement of 1m), instead of the spectrum for Los Angeles (SAC Joint Venture, 1997). The
4% defined in a preliminary stage, since higher values were bounding characteristics of the records are summarized in Table
leading to the development of a global failure mechanism of the 4. Further details on modelling and input can be found in
structure (corresponding to an ultimate steel strain in frame Casarotti et al. (2005).
elements conservatively fixed in 15% (Ballio and Mazzolani
1987)) under several records, and would thus prevent a SHORT BRIDGES LONG BRIDGES

complete statistical evaluation of the results. REGULAR

Label 222 Label 2222222

The effectiveness of the different static procedures in

SEMI Label 123 Label 3332111
predicting the local dynamic response is quantified and REGULAR
compared by means of error measure E1, which provides a Label 213 Label 2331312
IRREGULAR
direct insight on how inaccurate is the static method (evidently,
the mean of a pushover response estimate is computed only in
the case of record-dependent DAP analysis, for all other Figure 24. Analysed bridge configurations.
pushover schemes the single response value is used):

MeanPUSHOVER − Mean DYNAMIC (Eq. 7) Table 4. Bounding characteristics of the employed set of
E1 = records for bridges
MeanDYNAMIC
5% Arias Significant Total
Peak Ground Peak Response
Intensity Duration Duration teff / ttot
Obtained results. Figure 23 summarises the results obtained Acceleration Acceleration
threshold teff ttot
through this parametric study, putting in evidence the fact that, Min 0.30 g 0.84 g 1.25 s 5.3 s 14.95 s 9%
1.02 g 3.73 g 12.5 s 19.66 s 80.00 s 52%
when compared with other pushover procedures, DAP leads to Max

higher accuracy in the prediction of global and local response

parameters of steel buildings, particularly in those cases where
Analyses and results post-processing. The response of the
the influence of higher modes of vibration is important (e.g.
bridge models is estimated through the employment of
high-rise buildings). It is also shown that the employment of
Incremental Dynamic Analysis (IDA), Force-based
an average response spectral shape leads to satisfactory
Conventional Pushover with uniform load distribution (FCPu),
results, thus rendering the procedure very much applicable
Force-based Conventional Pushover with first mode
within a design application framework, where standard code
proportional load pattern (FCPm), Force-based Adaptive
spectral shapes are prescribed.
Pushover with Spectrum Scaling (FAP) and Displacement-
based Adaptive Pushover with Spectrum Scaling (DAP).
5.3. RC CONTINUOUS-SPAN BRIDGES
Results are presented in terms of the bridge capacity curve, i.e.
a plot of the reference point displacement versus total base
Characteristics of input motion and structural models. The
shear, and of the deck drift profile.
parametric study has considered two bridge lengths (50 m
Each level of inelasticity is represented by the deck centre
spans), with regular, irregular and semi-regular layout of the
drift, selected as independent damage parameter, and per each
pier heights and with two types of abutments; (i) continuous
level of inelasticity the total base shear Vbase and the
deck-abutment connections supported on piles, exhibiting a
displacements ∆i at the other deck locations are monitored.
bilinear behaviour, and (ii) deck extremities supported on pot
Results of pushover analyses are compared to the IDA median
value out of the responses to the 14 records, of each response to the fact that, for the same base shear, central deck forces are
quantity R, be it total base shear or deck drift: generally higher compared to the other load patterns, thus
results in larger displacement at that location. Then, on
Rˆi, IDA = median j =1:14 Ri, j − IDA[ ] (Eq. 8) occasions, a “hardening effect” in the pushover curve occurs,
which is sometimes reproduced only by employing DAP: once
piers saturate their capacity, abutments absorb the additional
Pushover analyses with spectrum scaling (i.e. adaptive seismic demand, proportionally increasing shear response and
pushovers) are statistically treated in an analogous way: hardening the capacity curve.
medians of each response quantity represent that particular
pushover analysis (i.e. FAP or DAP) with spectrum scaling. x 10
4 Capacity Curve - BridgeA213
Finally, the results of each type of pushover are normalized with
respect to the corresponding “exact” quantity obtained from the
IDA medians, as schematically illustrated in Figure 25, and 2
translated in Eq. 9. Representing results in terms of ratios

Total Base Shear (kN)

between the “approximate” and the “exact” procedures,
1.5
immediately indicates the bias in the approximate procedure, as
the ideal target value of the different pushovers is always one.
1
Ri, PUSHOVERtype
R i, PUSHOVERtype = L ⎯⎯⎯⎯→1 (Eq. 9) DAP
Rˆ i,IDA ideally
0.5
FAP
FCPu
FCPm
∆2 ∆3 IDA
∆1 ∆2 = ∆3 = 0
∆1 = ∆ˆ 2 ,IDA ∆ˆ 3,IDA ∆4 0 0.02 0.04 0.06 0.08 0.1 0.12
∆4 = Deck Displ. @ Central Pier Top (m)
∆ˆ 1,IDA ∆ˆ 4 ,IDA
a)
x 10
4 Capacity Curve - BridgeB3332111D

3.5

Figure 25. Normalised transverse deformed pattern.

Total Base Shear (kN)

2.5

Given the fact that a “realistic” capacity curve does not 2

imply reliable estimations of the inelastic displacement pattern
at increasing levels of inelasticity, the control of the deformed 1.5

pattern is of the same relevance of the capacity curve prediction. DAP

1
Having the same unitary target value, all normalized deck FAP
FCPu
displacements become comparable, and a bridge index BI can 0.5 FCPm
measure the precision of the obtained deformed shape. Per each IDA
level of inelasticity, such bridge index is defined as the median 0
0 0.045 0.09 0.135 0.18 0.225 0.27
of results over the m deck locations (Eq. 10), with the standard Deck Displ. @ Central Pier Top (m)
deviation measuring the dispersion with respect to the median b)
(Eq. 11). The latter indicates the stability of the estimate of Figure 26. Capacity curve results.
displacements along the deck: a small scatter means that
predicted normalised displacements along the deck are In Figure 27, the Bridge Index, as computed at each level of
averagely close to their median value BI. deck centre drift, is plotted as black filled marks so as to cater
for an easier comparison with the IDA-normalised deck
BI PUSHOVERtype = mediani =1:m (∆ˆ i, PUSHOVERtype ) (Eq. 10) displacements, represented as empty marks in the background.
In this manner, it results immediately apparent the level with
which each pushover analysis is able to capture the deformed
m 2 pattern of the whole bridge, at increasing deformation levels.
∑ ⎛⎜ ∆ˆ i, PUSHOVERtype − BI PUSHOVERtype ⎞⎟
i =1⎝ ⎠ (Eq. 11) For the sake of succinctness, only two analysis types are
δ PUSHOVERtype =
m −1 considered, FCPm and DAP, which are those leading to the
worst and best predictions, respectively.
Obtained results. Current code recommendations require Table 5 provides global averages of means, maximum and
performing pushover analysis by pushing the entire structure minimum values of BI and respective dispersion as well as of
with distributed load. In case of bridges, the additional option the normalised total base shear, over the whole bridge ensemble.
of pushing only the deck has been investigated, observing that It is noted that: (i) FCPm heavily underestimates predictions,
the superstructure is the physical location where the most of featuring also a very high BI dispersion value, (ii) FCPu
the structural mass, i.e. the source of the inertia forces on the performs very well for regular bridges and underestimating
bridge, is usually concentrated and where it is relatively free to otherwise, (iii) DAP features the best overall behaviour, despite
be excited. A preliminary investigation indicated a significant the slight underprediction of deformed shape values, with the
improvement in terms of stability and velocity of analysis in lowest values of scatter.
case of DAP and a very poor influence on results with the
application of the latter option, which is thus recommended 6. CONCLUDING REMARKS
and employed in the parametric study.
Two main pertinent observations can be withdrawn from a Given that current performance-based design trends
scrutiny of the capacity curves obtained by the different require simple, yet accurate methods for estimating seismic
pushover analyses in Figure 26: first, FCPm tends to demand on structures considering their full inelastic
significantly underestimate the structural stiffness, mainly due behaviour, in the current work the effectiveness of pushover
 Bridge A2331312 Bridge A213 Imperial College London, University of London,
2 United Kingdom.
1.5 6
5
4
1 3
2
Antoniou, S., Rovithakis, A., and Pinho, R., (2002).
a) 0.5 1
0 Development and verification of a fully adaptive
FCPm 0
0 0.02 0.04 0.06 0.08 0.1 0.12 0 0.02 0.04 0.06 0.08 0.1 0.12
pushover procedure, Proc. 12th Europ. Conf. Earthq.
Deck Displ. @ Deck Displ. @ Eng., London, UK, Paper No. 822.
Central Pier Top (m) Central Pier Top (m)
2 1.5
Antoniou, S., and Pinho, R., (2004a). Advantages and
1.5
1 limitations of adaptive and non-adaptive force-based
1
pushover procedures, J. Earthq. Eng., 8(4), 497-522.
b) 0.5
0.5

DAP 0
0 0.02 0.04 0.06 0.08 0.1 0.12
0
0 0.02 0.04 0.06 0.08 0.1 0.12 Antoniou, S., and Pinho, R., (2004b). Development and
Deck Displ. @ Deck Displ. @ verification of a displacement-based adaptive
Central Pier Top (m) Central Pier Top (m) pushover procedure, J. Earthq. Eng., 8(5), 643-661.
Figure 27. Prediction of the deformed pattern: BI and relative
scatter. Applied Technology Council, (2005). Improvement of
Nonlinear Static Seismic Analysis Procedures,
Table 5. Global averages of the summaries of results FEMA-440, ATC, California, USA.
Means Bridge Index Dispersion Normalised Base Shear
mean min max mean min max mean min max Applied Technology Council, (1997). NEHRP Guidelines for
FCPm 0.74 0.57 0.92 0.79 0.58 1.00 0.80 0.69 0.95
FCPu 0.87 0.75 1.03 0.24 0.17 0.34 1.03 0.92 1.18 the seismic rehabilitation of buildings, FEMA
FAP 0.88 0.78 1.01 0.22 0.13 0.34 0.99 0.89 1.10 Report No. 273, Federal Emergency Management
DAP 0.87 0.78 0.99 0.19 0.14 0.27 1.03 0.95 1.13
Agency, Applied Technology Council, Washington
D.C.
analysis, which serves as the main engine to Nonlinear Static
Procedures, applied to buildings and bridges has been Aydinoglu, M. N., (2003). An incremental response spectrum
investigated. In particular, the effectiveness of applying a analysis procedure based on inelastic spectral
displacement-based adaptive pushover to estimate the seismic deformation for multi-mode seismic performance
response of buildings and bridges subjected to earthquake evaluation”, Bull. Earthq. Eng., 1(1), 3-36.
action was investigated.
It is observed that the employment of such an innovative Ballio, G., and Mazzolani, F. M., 1987. Strutture in acciaio,
adaptive pushover technique lead to the attainment improved Ulrico Hoepli editore S.p.A., Milano, Italy. (in
response predictions, throughout the entire deformation range, Italian)
in comparison to those obtained by force-based methods, either
adaptive or conventional. Indeed, prediction of the global Bommer, J. J., and Acevedo, A. B., (2004). The use of real
behaviour (capacity curves), as well as of the deformed shapes earthquake accelerograms as input to dynamic
and shear/moment distributions, proved to be very effective. analysis, J. Earthq. Eng. 8 (Special Issue 1), 43-91.
In other words, within the scope of buildings and bridge
applications, whereas the application of a fixed displacement Bommer, J. J., and Martinez-Pereira, A., (1999). The effective
pattern is a commonly agreed conceptual fallacy, the present duration of earthquake ground motion, J. Earthq.
work witnesses not only the feasibility of applying an adaptive Eng., Vol. 3, Issue 2, pp.127-17
displacement profile, but also its practical advantages, with
respect to other pushover methods. Bracci, J. M., Kunnath, S. K., and Reinhorn, A. M., (1997).
To fully explore the demonstrated potential of the DAP Seismic performance and retrofit evaluation for
algorithm, it is advisable for the equally adaptive ACSM reinforced concrete structures, ASCE J. Struct. Eng.,
Nonlinear Static Procedure, proposed by Casarotti and Pinho 123(1), 3-10.
(2007) to be used. Currently ongoing work is therefore testing
how such equally innovative NSP fairs in comparison to its Building Seismic Safety Council, (2000). Prestandard and
counterparts (CSM, N2, MPA); first results (Pinho et al., 2007) Commentary for the Seismic Rehabilitation of
seem promising and reassuring. Buildings, FEMA-356, Washington D.C., USA

7. ACKNOWLEDGEMENTS Campos-Costa, A., and Pinto, A.V., (1999). European Seismic

Hazard Scenarios – An Approach to the Definition
The author would like to acknowledge the important of Input Motions for Testing and Reliability
contribution of Dr C. Casarotti, Dr M. López-Menjivar, Mr D. Assessment of Civil Engineering Structures, JRC
Pietra and Ms. M. Faravelli for the assistance in the Special Pubblication, no X.99.XX, ELSA, JRC –
preparation of this manuscript. Further, part of the current Ispra, Italy.
work was carried out under the financial auspices of the 2005-
2008 framework programme established between the Italian Casarotti, C., and Pinho, R., (2007). An Adaptive capacity
Department of Civil Protection and the European Centre for spectrum method for assessment of bridges
Training and Research in Earthquake Engineering subjected to earthquake action, Bulletin of
(EUCENTRE), a support that is thus also gratefully Earthquake Engineering, 5(3), 377-390.
acknowledged.
Casarotti, C., Pinho, R., and Calvi, G.M., (2005). Adaptive
8. REFERENCES Pushover-based Methods for Seismic Assessment
and Design of Bridge Structures, ROSE Report
Antoniou, S., (2002). Advanced inelastic static analysis for 2005/06, IUSS Press, Pavia, Italy.
seismic assessment of structures, PhD Thesis,
Chopra A. K., and Goel R. K., (2002). A modal pushover Krawinkler H., and Seneviratna G. D. P. K., (1998). Pros and
analysis procedure for estimating seismic demands cons of a pushover analysis of seismic performance
for buildings, Earthq. Eng. Struct. Dyn., 31, 561- evaluation, Eng. Struct., 20(4-6), 452-64.
582.
Kunnath, S. K., (2004). Identification of modal combination
Comité Europeen de Normalization, (2002). Eurocode 8: for nonlinear static analysis of building structures,
Design of Structures for Earthquake Resistance - Computer-Aided Civil and Infrastruc. Eng., 19, 246-
Part 2: Bridges, PrEN 1998-2: 2003, 2 April 2002, 259.
CEN, Brussels, Belgium.
Kunnath, S. K., and John, A. Jr., (2000). Validity of static
Elnashai, A. S., (2001). Advanced inelastic static (pushover) procedures in performance-based seismic design,
analysis for earthquake applications, Struct. Eng. Proceedings of ASCE Structures Congress,
Mech., 12(1), 51-69. Philadelphia, USA.

Fajfar, P., (1999). Capacity spectrum method based on López-Menjivar, M. A., (2004). Verification of a
inelastic demand spectra, Earthquake Engineering displacement-based Adaptive Pushover method for
and Structural Dynamics, 28, 979-993. assessment of 2-D Reinforced Concrete Buildings,
PhD Thesis, European School for Advances Studies
Fajfar, P., and Fischinger, M., (1988). N2 – A method for non- in Reduction of Seismic Risk (ROSE School),
linear seismic analysis of regular buildings, University of Pavia, Italy.
Proceedings of the Ninth World Conference in
Earthquake Engineering, Tokyo-Kyoto, Japan, 5, pp. Maison, B., (2005). Discussion of “Evaluation of Modal and
111-116. FEMA Pushover Analysis: SAC Buildings”, Earthq.
Spectra 21(1), 275-275.
Fardis, M. N., (1994). Analysis and design of reinforced
concrete buildings according to Eurocode 2 and 8. Matsumori, T., Otani, S., Shiohara, H., and Kabeyasawa, T.,
Configuration 3, 5 and 6, Reports on Prenormative (1999). Earthquake member deformation demands in
Research in Support of Eurocode 8. reinforced concrete frame structures, Proc., US-
Japan Workshop PBEE Methodology for R/C
Freeman, S.A., (1998). Development and use of capacity Building Structures, PEER Center Report, UC
spectrum method, Proceedings of the Sixth U.S. Berkeley - 79-94, Maui, Hawaii.
National Conf. Earthquake Engineering, Seattle,
Oakland, USA. Moghadam, A. S., and Tso, W. K, (2002). A pushover
procedure for tall buildings, Proc., 12th Europ. Conf.
Freeman, S.A., Nicoletti, J.P., and Tyrell, J. V., (1975). Earthq. Eng., London, UK, Paper 395.
Evaluations of Existing Buildings for Seismic Risk –
A Case Study of Puget Sound Naval Shipyard, Reinhorn, A., (1997). Inelastic analysis techniques in seismic
Bremerton, Washington, Proceedings of U.S. evaluations, in Seismic design methodologies for the
National Conference on Earthquake Engineering, next generation of codes, Krawinkler and Fajfar
Berkley, USA., pp. 113-122. (editors), Balkema, 277-287.

Goel, R. K., and Chopra, A. K., (2005a). Response to B. Paret, T. F., Sasaki, K. K., Eilbeck, D. H., and Freeman, S. A.,
Maison’s Discussion of “Evaluation of Modal and (1996). Approximate inelastic procedures to identify
FEMA Pushover Analysis: SAC Buildings”, Earthq. failure mechanisms from higher mode effects, Proc.,
Spectra, 21(1), 277-279. 11th World Conf. Earthq. Eng., Acapulco, Mexico,
Paper 966.
Goel, R. K., and Chopra, A. K., (2005b). Role of Higher-
“Mode” Pushover Analyses in Seismic Analysis of Pietra, D., (2006). Evaluation of pushover procedures for the
Buildings, Earthq. Spectra 21(4), 1027-1041. seismic design of buildings, MSc Dissertation,
European School for Advanced Studies in Reduction
Gupta, B., and Kunnath, S. K., (2000). Adaptive spectra-based of Seismic Risk (ROSE School), University of
pushover procedure for seismic evaluation of Pavia, Pavia, Italy.
structures, Earthq. Spectra, 16(2), 367-392.
Pinho, R., and Elnashai, A. S., (2000). Dynamic collapse
Gupta, A., and Krawinkler, H., (1999). Seismic demands for testing of a full-scale four storey RC frame, ISET J.
performance evaluation of steel moment resisting Earthq. Eng., Special Issue on Experimental
frame structures (SAC Task 5.4.3), Report No. 132, Techniques, 37(4), 143-164.
John A. Blume Earthquake Engineering Center,
Stanford University, Stanford, California. Pinho, R., Casarotti, C. and Monteiro, R., (2007). An adaptive
capacity spectrum method an other nonlinear static
Hernández-Montes E., Kwon, O. S., and Aschheim, M., procedures applied to the seismic assessment of
(2004). An energy-based formulation for first and bridges, Proceedings of the First US-Italy Workshop
multiple-mode nonlinear static (pushover) analyses, on Seismic Design and Assessment of Bridges,
J. Earthq. Eng., 8(1), 69-88. Pavia, Italy.

Hughes, T. J. R., (1987). The Finite Element Method, Linear Requena, M., and Ayala, G., (2000). Evaluation of a
Static and Dynamic Finite Element Analysis, simplified method for the determination of the
Prentice-Hall, Inc. nonlinear seismic response of RC frames, Proc. 12th
 World Conf. Earthq. Eng., Auckland, New Zealand, Seismosoft, (2005). SeismoStruct - A Computer Program for
Paper No. 2109. Static and Dynamic nonlinear analysis of framed
structures [online], available from URL:
SAC Joint Venture, (1997). Develop Suites of Time Histories, http//www.seismosoft.com.
Project Task: 5.4.1, Draft Report, March 21, 1997,
Sacramento, CA, USA. Sommerville P., Smith N., Punyamurthula S. and Sun J.,
(1997). Development of ground motion time
Sasaki, K. K., Freeman S. A., and Paret T. F., (1998). Multi- histories for phase 2 of the FEMA/SAC steel project,
mode pushover procedure (MMP) – a method to SAC Background Document, Report No. SAC/BD/-
identify the effects of higher modes in a pushover 7/04, Applied Technology Council, Redwood City,
analysis, Proc., 6th US Conf. Earthq. Eng., Seattle, CA.
Washington – Earthquake Engineering Research
Institute, Oakland, California. Vamvatsikos D., and Cornell C. A., (2002). Incremental
dynamic analysis, Earthq. Eng. Struct. Dyn., 31(3),
Satyarno, I., Carr, A. J., and Restrepo, J., (1998). Refined 491-514.
pushover analysis for the assessment of older
reinforced concrete buildings, Proc. NZSEE
Technology Conference, Wairakei, New Zealand,
75-82.