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

net/publication/283008258

A Computer Method for Advanced Pushover Analysis of Reinforced-Concrete

Frameworks

Conference Paper · September 2015

CITATIONS READS

3 1,338

2 authors:

Szabolcs Varga Cosmin G Chiorean

Universitatea Tehnica Cluj-Napoca Universitatea Tehnica Cluj-Napoca
7 PUBLICATIONS   11 CITATIONS    56 PUBLICATIONS   487 CITATIONS   

SEE PROFILE SEE PROFILE

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

Computational Efficient Models for Non-Linear Inelastic Analysis of Building Frameworks View project

Fit-to-nZEB: Innovative training schemes for retrofitting to nZEB-levels View project

All content following this page was uploaded by Szabolcs Varga on 20 October 2015.

The user has requested enhancement of the downloaded file.

 Civil-Comp Press, 2015
Proceedings of the Fifteenth International Conference on
Paper 162 Civil, Structural and Environmental Engineering Computing,
J. Kruis, Y. Tsompanakis and B.H.V. Topping, (Editors),
Civil-Comp Press, Stirlingshire, Scotland

A Computer Method for Advanced Pushover Analysis

of Reinforced-Concrete Frameworks

S. Varga and C.G. Chiorean

Faculty of Civil Engineering
Technical University of Cluj-Napoca, Romania

Abstract

This paper presents an integrated system for advanced structural analysis and
seismic performance evaluation of reinforced concrete frameworks. The advanced
non-linear inelastic static analysis employed herein uses the accuracy of the fiber
element approach for large deflection inelastic frame analysis and addresses its
efficiency both at element level, through the use of only one element to model each
physical member of the frame, and at a cross-sectional level through the use of the
path integral approach for numerical integration of the cross-sectional nonlinear
characteristics. In this way, the states of strain, stress and yield stress are monitored
explicitly during each step of the analysis, the arbitrary cross-sectional shape,
various stress–strain relationships for concrete and reinforcement steel are accurately
included in the analysis. The capacity curve is determined through advanced
pushover analysis, and then the evaluation of the seismic performance is achieved
with an approach that uses nonlinear time-history analysis of a single degree-of-
freedom oscillator. The procedure is developed in the framework of the provisions
of Eurocode 8. The ductility demand is determined directly from a set of
accelerograms, without graphical or numerical approximations. Several
computational examples - for different height-rise frame structures of 5, 9 and 15
stories - are given to validate the effectiveness of the proposed method, the
reliability and time saving of the code. The proposed approach is presented as an
efficient, reliable tool, ready to be implemented into design practice for advanced
analysis and pushover analysis of reinforced concrete frame structures.

Keywords: advanced analysis, pushover analysis, constant ductility inelastic

spectrum, target displacement, Eurocode 8, reinforced concrete frames.

1
1 Introduction
With the rapid advancement of computer technology, research works are
continuously making progress in the development of advanced nonlinear inelastic
analysis methods and integrate them into new and more rational advanced analysis
as well as performance based design procedures. Reliable nonlinear inelastic
analysis techniques are, for instance, essential in performance-based earthquake
engineering, and advanced analysis methodologies, that involves accurate
predictions of inelastic limit states up or beyond to structural collapse [1-5]. In spite
of the availability of some finite element algorithms and powerful computer
programs, the non-linear inelastic analysis of real large-scale frame structures still
makes huge demand on the most powerful of available computers and still
represents unpractical tasks to most designers. The need for accurate yet
computational efficient tools for the nonlinear inelastic analysis of three-
dimensional reinforced concrete frameworks forms one of the main motivations
behind of this work.
Essentially, the nonlinear inelastic analysis employed herein uses the accuracy of
the fiber element approach for inelastic frame analysis and addresses its efficiency
and modelling shortcomings both at an element level, through the use of only one
element to model each physical member of the frame, and at a cross-sectional level
through the use of the path integral approach for numerical integration of the cross-
sectional nonlinear characteristics. This is an essential requirement to approach real
large spatial frame structures, combining modelling benefits, computational
efficiency and reasonable accuracy fulfilling the performance based design and
advanced analysis requirements. A spread-of-plasticity analysis that includes the
gradual development of plastic zones throughout the volume of the elements, initial
geometric and material imperfections, and any other significant second-order effects,
would certainly be suitable to be classified as advanced analysis in which the
checking of beam column interaction is not required.
On the other hand, since the Kobe and Northridge earthquakes, a static inelastic
analysis (pushover analysis) becomes an accepted and simple method for the
seismic evaluation of high-rise buildings. The purpose of the pushover analysis is to
evaluate the expected performance of a structural system by estimating its strength
and deformation demands in design earthquakes by means of a static inelastic
analysis, and comparing these demands to the available capacities at the
performance levels of interest. In this direction, the new design documents have
developed modelling procedures, acceptance criteria and analysis procedures for
pushover analysis. [6-9]. Seismic demands are computed using nonlinear static
analysis of the structure subjected to monotonically increasing lateral forces with an
invariant height-wise distribution until a predetermined target displacement is
reached. As currently formulated, the pushover analysis has limitations because it is
based on static loading, so it cannot represent dynamic phenomena as well as a
nonlinear dynamic analysis. Both the force distribution and target displacement are
based on the assumptions that the response is controlled by the fundamental mode
and the mode shape remains unchanged after the structure yields. To follow more

2
closely the time variant distribution of inertial forces, modal pushover or the
adaptive force distribution procedures can be applied [2-4].
As it was stated previously, pushover analysis is used to evaluate the seismic
performance of a structure to an expected level of excitation. After plotting the
capacity curve through a static inelastic analysis, the second step of the pushover
analysis procedure is the determination of the inelastic seismic demand and of the
displacement encountered by the structure at a given seismic demand, namely the
target-displacement. In the case of EC8 [6] and FEMA 356 [7] &440 [8], seismic
demand is represented by a smoothed design spectrum, which - through simple Rμ-
μ-T relations - facilitates the determination of the inelastic spectrum. Alternatively,
ATC40 calculates the target displacement through the reduction of elastic spectrum
with the aid of equivalent damping and period. In each case, the target-displacement
is obtained through using the Capacity Spectrum Method (CSM) [9] and is
represented by the intersection of the capacity diagram of the equivalent SDOF
system with the diagram of the predetermined inelastic seismic demand.
It should also be noted that smoothed design spectra are generated from multiple
seismic records and represents the envelope of the expected seismic response.
Furthermore, the equivalence between strength of the structural system Rµ and
displacement ductility demand µ is established by means of empirical relationships
that have also been determined from the processing of several accelerograms.
Consequently, the target-displacement determined from pushover analysis, and
smoothed design spectrum has rather a general character, and is useful especially
during the design of structures for an expected probabilistic level of seismic
intensity. For either assessment or research scenarios, structural engineers do not
have a simple analytical method at hand that would rapidly determine the nonlinear
behaviour of structures, for a set of accelerograms, in the framework of design
codes.
Setting off from the considerations listed above, a new method is proposed herein
in which the target displacements are determined using the NTHA of a SDOF
oscillator. The displacement ductility demand μ is calculated using the strength Rµ
and period T* of the equivalent SDOF [10]. The proposed method follows the steps
of the well-known N2 method [5] when it comes to determine the capacity diagram
of the equivalent SDOF, except that instead of inelastic spectrum obtained from
design spectrum with the aid of empirical formulas, in this approach is used inelastic
spectra directly generated from the accelerograms.

2 Proposed method for advanced nonlinear inelastic

analysis of RC frameworks
In the context of reinforced concrete beam-columns, the following assumptions are
often adopted in the formulation of realistic analytical model: (1) Plane section
remain plane after flexural deformation; (2) Full strain compatibility exists between
concrete and steel reinforcement; (3) Reinforcement steel bars cannot buckle under

3
compression; (4) Mechanical properties of concrete may vary according to
confinement levels. A flexibility-based method is used to formulate the distributed
plasticity model of a 3D frame element (12 DOF) subject to the above assumptions
where elasto-plastic behaviour is modelled accounting for the spread-of plasticity
effects in sections and along the element and employs modelling of structures with
only one line element per member, which reduces the number of degree of freedom
involved and the computational time. The first two assumptions allow the
formulation details to be considered on two distinct levels, namely, the cross-
sectional level and the member longitudinal axis level. Thus the nonlinear response
of a beam-column element can be computed as a weighted sum of the response of a
discrete number of cross-sections. In the present elasto-plastic frame analysis
approach, gradual plastification through the cross-section subjected to combined
action of axial force and bi-axial bending moments is described through basic
equilibrium, compatibility and material nonlinear constitutive equations σ-ε of
concrete and reinforcement steel, in any section by an iterative process. In this way
the arbitrary cross-sectional shape and reinforcement layout the effect of concrete
tensile cracking, the nonlinear compressive response of concrete with different
levels of confinement are accurately included in the analysis.

2.1 Cross-section analysis

Gradual plastification through the cross-section subjected to combined action of
axial force and bi-axial bending moments is described through basic equilibrium,
compatibility and material nonlinear constitutive equations. In this way, the states of
strain, stress and yield stress are monitored explicitly during each step of the
analysis, the arbitrary cross-sectional shape and reinforcement layout, the effect of
concrete tensile cracking, the nonlinear compressive response of concrete and strain
softening of concrete in compression with different levels of confinement are
accurately included in the analysis.

 ε ε2 
Stress fc f c = f c''  2 − 2 
 ε c0 ε c0 
  ε − ε c0 
f c = f c'' 1 − γ   
  ε cu − ε c 0  

f''c

f''c(1-γ)

Strain, ε
εc0 εcu

Figure 1: Inelastic analysis of arbitrary shaped cross-section.

4
Considering the cross-section subjected to the action of the external bending
moments about each global axes and axial force as shown in Figure 1a. Under the
above assumptions the resultant strain distribution corresponding to the curvatures
about global axes Φ = [Φ x Φ y ] and the axial compressive strain u can be
expressed in point r = [x y ] in a linear form as:
ε = u + Φ x y + Φ y z = u + Φr T (1)
The basic equations of equilibrium for the axial load N and the biaxial bending
moments M z ,ext , M y ,ext are given in terms of the stress resultants as:
1
∫ σ (u
A
0 , Φ y , Φ z )⋅  T  dA − S Text = 0
r 
(2)

where the vector S ext = [N M x ,ext M y ,ext ]. The Eqs. (2) are solved numerically
using the Newton-Raphson method, and results in three recurrence relationships to
obtain the unknowns u and Φ and then flexural EI and axial EA rigidity modulus can
be computed [11]. To model the concrete under compression the stress-strain
relationship is represented by a combination of a second-degree parabola (for
ascending part) and a straight line (for descending part), as depicted in Figure 1b,
where γ represents the degree of confinement in the concrete [11]. Based on
Green’s integration formula according to which the domain integrals appearing in
the evaluation of internal resultant efforts and tangent stiffness matrix coefficients of
the section can be evaluated in terms of a boundary integral. This approach is
extremely rapid because stress integrals need only be evaluated at a small number of
points on the section boundary. Tangent flexural rigidities for major and minor axis
bending under conditions of constant axial load are evaluated by applying the
procedure described in [11].

2.2 Elasto-plastic tangent stiffness matrix and equivalent nodal

loads
Flexibility-based method is used to formulate the distributed plasticity model of a
3D frame element (12 DOF) subject to the above assumptions. An element is
represented by several cross sections (i.e. stations) that are located at the numerical
integration scheme points (Fig.2a, b). The spread of inelastic zones within an
element is captured considering the variable section flexural EIy and EIz and axial
EA rigidity along the member length, depending on the bending moments and axial
force level, cross-sectional shape and nonlinear constitutive relationships. Figure 2a
shows the deformed shape of a 3D beam-column element in a local system attached
to the initially straight center line, with the rigid body modes removed. The element
stiffness matrix ke implicitly defined by the Clapeyron relation:
1 1 1
∆W = ⋅ u e k e u e = ⋅ p T u e = ⋅ p T k e−1p
T
(3)
2 2 2
will be obtained by recovering the explicit expression for the element increment
strain energy:

5
 (4)

(a)
(b)

Figure 2: Element formulation: (a) Element in local system with rigid body modes
removed (b) Numerical integration scheme points.

where the nodal forces and displacement vectors are denoted with p and ue
respectively. The resulting element stiffness matrix is a 6x6 matrix. To include rigid
body modes, the stiffness matrix is pre- and post-multiplied by a transformation
matrix to result in the required 12 x12 matrix. The tangent flexural rigidities (EA,
EI) that appear in Eq. 4 for major and minor axis bending under conditions of
constant axial load are automatically following the procedure described in [11]. As
aforementioned, tangent stiffness properties of the cross sections are integrated
along the member length to yield member stiffness coefficients and equivalent nodal
loads [11]. In the present investigation, the loading due to the member lateral loads
and transferred to the nodes, taking into account the element nonlinear behaviour,
are allowed and included automatically in the analysis. This leads to a significant
saving in imputing the member loads, without the need to divide a member into
several elements for simulation of these loads. Explicit expressions for the element
stiffness matrix ke and the equivalent nodal forces and further details about the
element are reported in [11]. The local geometrical nonlinear effects for each
element are taken into account in the present analysis, in a beam column approach,
by the use of the inelastic stability stiffness functions and updating at each load
increment the length, axial force and the flexural rigidity about of each principal
axes of the element. This way minimizes both the modelling and solution time,
generally only one or two elements are needed per member [11].

2.3 Geometry updating and analysis algorithm

In order to trace the equilibrium path, for proportionally and non-proportionally
applied loads, the proposed model has been implemented in a simple incremental

6
and incremental- iterative matrix structural-analysis program. In the simple
incremental method, the simple Euler stepping algorithm is used in conjunction with
constant work-load increments. This analysis is simple, reliable and is not sensitive
to convergence failures that can occur in incremental-iterative schemes, and can also
give the full nonlinear load-deformation response including the ultimate load and
post-critical response [11]. Using an updated Lagrangian formulation (UL) the
global nonlinear geometrical effects are considered updating the element forces and
geometry configurations at each load increment. The natural deformation approach
(NDA) in conjunction with the geometrical “rigid body qualified” stiffness matrix is
adopted for the element force recovery and the web plane vector approach is
effectively used to update the frame element coordinates [11]. Based on the analysis
algorithm just described, a computer program, NEFCAD, has been developed to
study the combined effects of material and geometric nonlinear behaviour on the
load-versus-deflection response for spatial reinforced-concrete framed structures. It
combines the structural analysis routine with a graphic routine to display the final
results. The computational engine was written in Compaq Visual Fortran. The
graphic interface was created using Microsoft Visual Basic 6. Dynamic Link
Libraries (DLL) are used to communicate between the interface and engine.

3 Proposed method for determination of inelastic seismic

displacements using advanced pushover analysis and
constant ductility inelastic spectrum

3.1 The determination of the pushover curve

A pushover analysis is performed by subjecting the structure to a monotonically

increasing pattern of lateral forces, representing the inertial forces which would be
experienced by the structure when subjected to ground shacking. The variation of
the displacement of control node (i.e. the node used to monitor displacement of the
structure) versus the base-shear force forms the capacity (pushover) curve.
According to the pushover procedures, the frame is considered to be subjected to the
non-proportional action of gravity loads and lateral seismic loads. Usually the
gravitational loads, modelled as uniform distributed loads applied on the beams,
represent the first sequence of loading and are kept constant during the applying of
the second sequence of loading formed by seismic lateral loads that are continuously
increasing until the collapse of the structures occurs. The seismic loads are
simulated by point loads at every beam-column joints in conjunction with either an
adaptive or invariant height-wise distribution of the loads throughout the height of
the building. The transformation procdure of the capacity curve of the MDOF
(pushover curve) system into the bilinear capacity diagram of the SDOF system is
implemented via the principles of the well-known N2 method (EC8). For further
details regarding the N2 method, the reader is referred to [5].

7
3.2. The determination of the seismic demand
3.2.1 Inelastic response of SDOF oscillators

The determination of the inelastic spectrum directly from an accelerogram in

conjunction with pushover analysis was also used in the ”Yield Point Spectra”
(YPS) method [10]. In the YPS the bilinear capacity curve, is plotted against the
constant ductility spectra for several ductility values of the considered earthquake.
From conceptual point of view, the proposed method is thus similar to the YPS,
except that it allows the identification of the µ ductility demand from the intrinsic
characteristics of the seismic record and those of the equivalent SDOF system. In
contrast with the YPS method, this feature of the proposed approach could make
possible to use either numerical or graphical approximations, without the necessity
to use empirical relationship like in the case of the traditional design code NSA
methods.

The response in the non-linear domain is modelled by using functions known as

hysteretic models, which establish the response of the SDOF system according to its
period of vibration, displacement and yield strength. By solving Eq. (5) the non-
linear response of the SDOF system is obtained. Due to the fact that the function
has a general form, the differential equation can be solved in its general
form:
(5)

(6)
(7)

(8)

in which is the force associated to the capacity of the SDOF system,

is the idealized variant of in relation with the yield strength
fy (Eq. (6)), ωn is the pulsation of the system (Eq. (7))and ξ is damping. (Eq. (8))
[13]

3.2.2. The determination of the constant strength inelastic spectrum

With the aim of solving the nonlinear equations of motion, there are different
methods for selecting the yield strength of the SDOF oscillator. These result in
different types of inelastic spectrum, which can be constant ductility, or constant
strength inelastic spectrum, among others. [14]
In order to determine the ductility demand characteristic to the SDOF system, Rµ
elastic force reduction factor is used. Rµ is the ratio between the elastic acceleration
Sae(T*) , and the yield capacity Say expressed in acceleration of the bilinearized
equivalent SDOF system. Eq. (9) T* is the period of the SDOF equivalent, and

8
similar with the bi-linearized capacity diagram is computed through rules of N2
method (EC8). For the considered accelerogram, the elastic spectrum Sae(T) can be
computed through the linear response history analysis of SDOF oscillators. Figure 3
presents an example of elastic spectrum (Sae(T), μ=1) plotted against the bi-linear
capacity diagram, where the significance of elastic force reduction factor Rµ is self-
explanatory.
S ae (T * )
Rµ = (9)
S ay
In order to determine the µ ductility demand assigned to the Rµ strength demand, the
constant Rµ strength spectrum is plotted. The latter can be achieved easily for the
considered accelerogram, with the aid of the nonlinear response history analysis of
the oscillator with constant Rµ strength. The constant Rµ strength spectra show the
ductility demand µ(T) for all SDOF systems characterized by the spectrum of T
periods (Figure 3). The ductility demand sought is the same as the displacement
ductility achieved by the oscillator with Rµ strength, at the period T*, of the
equivalent SDOF, as depicted in Figure 3.

Figure 3: The determination of ductility demand µ(T) using the constant Rµ strength
spectrum

3.2.3 The determination of the constant ductility inelastic spectrum and of the
target displacement

Once the ductility demand is determined, the target displacement can be determined
either graphicaly or analyticaly. In order to compute the target displacement with the
use of graphic representation is necessary to plot the constant ductility inelastic
spectrum associated with the previously determined µ ductility value. If the user
choose for analytical calculations, only the ductility µ for the period T* -not for the
all spectrum of periods- has to be determined. Alternatively, after having determined
the ductility demand according to those described in chapter 3.2.2., the calculation
of the target displacement Dt of the MDOF system can be initiated on an analytical
way, with the following expression:
(10)

9
In Eq. (10) the product represents the target displacement D*t of the SDOF
system – computed at the intersection of the ascendant slope of the biliniar capacity
diagram with the constant ductility inelastic spectra in Fig.4 - while is the first
mode mass participation factor.

Figure 4: The determination of the target displacement through the graphical variant
of the proposed method.

4 Numerical results

4.1 Summary of application of the proposed method

The key steps of the proposed methodology described above can be summarised as
follows:
1. Determine the capacity curves for each analysed structure by applying an
advanced pushover analysis.
2. Linearly idealize the capacity curves and calculate the main quantities of the
equivalent SDOF system: the period T*, the yield acceleration Say, the yield
displacement D*y, and the ultimate displacement D*u. This step of the procedure
is similar to the methodology adopted in EC8 (N2 method).
3. Determine the elastic spectrum Sae(T) for each earthquake record, through
SDOF response history analysis.
4. Calculate the strength demand Rµ, for each accelerogram and bi-linear capacity
diagram, as described in section 3.2.2. (Eq. 9).
5. Calculate the ductility demand associated to the equivalent SDOF with the aid
of the constant strength demand Rµ spectrum, as described in section 3.2.2.
Different SDOF hysteretic modelling may yield different ductility demands.
6. Determine the inelastic displacement demand analytically or graphically, as it
was already described in section 3.2.3.
7. Perform a pushover analysis by pushing the structure to the target displacement;
determine the internal efforts experienced by the structures, displacements and
rotations at local level.

10
4.2 Structures, materials, loading

Three reinforced concrete frames with different height regimes (5, 9 and 15 stories)
have been proposed for analysis, referred hereinafter as structure 1, 2 and 3. The
frames have two spans of 6 metres each. Dead loads including monolithic reinforced
concrete slab are considered to be in total of 8 kN/m2. For live loads a value of 4
kN/m2 was considered. The strength class of the concrete is C20/25, with the
characteristic values of the compressive strength of cylinder and cube (Rck=16
N/mm2 and Rck,cube=20 N/mm2 ) and with the modulus of elasticity Ecm= 30 GPa.
The reinforcement has a yield strength of fyk=345 N/mm2.

Figure 5: Geometrical properties of the analysed structures: (a) 5 story, (b) 9 story,
(c) 15 story structures

11
 Figure 6: Reinforced concrete beam and column sections

Figure 7: Capacity curves plotted with the aid of advanced pushover software
NEFCAD, with and without considering the nonlinear geometrical effects.

12
4.3 Plotting the pushover curves
The pushover curves were plotted using NEFCAD [11] advanced analysis software
package, in two ways, with and without considering the nonlinear geometrical
effects (Figure 7). In order to determine the characteristics of the equivalent SDOF
system, namely the period T* and yield displacement D*y, the F* -D* format capacity
curve is bi-linearly idealized using the equal energy rule. The areas situated below
the capacity curve represent the energy dissipated by the equivalent system. This can
be calculated by integrating the product of Sa and D*. By imposing the equality
between the yield and ultimate acceleration (Say=Sau), the yield displacement D*y of
the equivalent system can be computed analytically (Eq.(11)). Once the yield
displacement D*y is determined, the equivalent period T* can be computed: it defines
the angle of the ascendant slope of the bilinear system through Eq. (12) [5].

Figure 8: Elastic (a) and design spectrum matched spectra (b) for the
considered earthquake records

(11)

m* ⋅ D *y
T * = 2 ⋅π ⋅ (12)
Fy*

4.4 Determination of the inelastic seismic displacements

With the aim of validating the proposed method, 8 accelerograms were matched to
the design spectrum with NCR=100 years, PGA=0.24g, and corner period Tc=1.6s,
using the earthquake records taken in Bucharest at the 1977 and 1986 earthquakes.
The design spectrum matched accelerograms were generated using SeismoMatch
software [15] and follow the rules stated in section 3.2.3.1.2 of EC8. The smoothed

13
elastic design spectrum and the obtained elastic spectra from the matched
accelerograms are plotted in Figure 8. A 5% value has been assumed for the viscous
damping and a 5% value for the post-yield hardening. The generation of the constant
ductility inelastic spectra has been carried out with the aid of a Clough type
(stiffness degrading) hysteretic model. In every case, the constant strength demand
Rµ, which is related to the yield/ultimate acceleration Say=Sau of the bilinear capacity
diagram, and the period T* of the SDOF equivalent was determined. Then, the
ductility demand µ was calculated through NTHA of the SDOF oscillator,
characterized by period T* and constant strength demand Rµ, as described
previously in sections 3.2.2 and 3.2.3. The constant strength and constant ductility
spectra were developed using the Bispec [14] software.

5 Discussion of the results compared to other types of

seismic analyses

5.1. Upper story displacements

In Figure 9, specific and average upper story displacements, obtained from the
proposed analysis procedure, are plotted for each of the 8 earthquakes and SDOF
hysteretic models used. As mentioned above, the seismic demands have been
represented by the design spectrum in case of NSA, and the totality of the design
spectrum matched seismic records, for the proposed method and NTHA. For both
the code-based NSA, and the proposed procedure, the capacity curves were plotted
using NEFCAD advanced analysis software (Figure 7). In order to highlight the
differences in terms of structural behaviour, strength capacity and ductility, two
types of nonlinear analysis (i.e. with and without considering the nonlinear local and
global geometrical effects) have been conducted. The effectiveness of the proposed
procedure is proven by comparing the mean upper story displacements yielded by
the totality of the proposed method`s analysis (Figure 9), with the mean upper story
displacement of the totality of NTHA-s, and of the code based NSA. At the same
time, specific displacement demands yielded by the proposed procedure, for each of
the 8 acccelerograms are represented along with the mean values. Thus, the
proposed procedure yields reasonable estimates of the NTHA, while performs better
with respect to the code-based NSA in every case, especially in the field of low and
mid-rise structures.
At the same time, with respect to the code-based NSA, the proposed approach
could not significantly improve the prediction of the inelastic displacements in the
case of the high-rise 15 story structure, where inelastic behaviour and failure
mechanism are not governed only by the fundamental mode of vibration. If
nonlinear geometrical effects are disregarded, target displacements are generally
overestimated, due to the greater load factors and thus, the overestimation of the
ductility demand.

14
Figure 9: The comparison of the upper story displacements obtained the proposed
method, NTHA, and code-based NSA, through the 5 story (a), 9 story (b),
15 story (c) frame structure, for earthquake records with PGA 0.12g; 5
story (d), 9 story (e), 15 story (f) frame structure, for earthquake records
with PGA 0.24g

In the case of including the nonlinear geometrical effects, the smaller displacements
can be due to the spread of plasticity, which leads faster element yielding and in this
case cannot be used the whole ductility capacity of the structure.

15
5.2 Story drift profiles
In present section a comparative analysis of the (maximum) story drift distribution is
presented, where proposed method using advanced pushover analysis (NEFCAD)
with and without nonlinear geometric effects is set against fixed and adaptive load
pattern [3] capacity curve based NSA developed through SeismoStruct, code-based
NSA procedures and nonlinear time-history analysis (NTHA). The story drift
profiles, with the corresponding mean error for the set of 8 earthquakes scaled on
two intensity levels, are presented in Figure 10 and 11.

Figure 10: Drift profiles, and errors Figure 11: Drift profiles and errors
calculated for the 8 calculated for the 8
earthquake records with earthquake records with
PGA 0.12g for the 5 story PGA 0.24g for the 5 story
(a), 9 story (b), 15 story (c) (a), 9 story (b), 15 story (c)
frame structures frame structures

In the SeismoStruct software, a fiber-based structural analysis has been applied, in

order to take into account the behaviour of the beam-column element at the section-
based level. The beam-column element is divided into approximately 100-150

16
fibers, and at each fiber a nonlinear uniaxial stress-strain relationship is assigned.
The stress-strain state of the element is obtained by integrating the nonlinear
uniaxial stress-strain response of the individual fibres along the length of the
element [16].
While in the case of the 5 story structure, and 0.12g earthquakes where inelastic
effects are smaller, differences in the lateral drift predictions are not significant, but
this is no longer true in the cases of the higher-rise 9 and 15 story structures
respectively, and 0.24g earthquakes. It can be observed that at the lower stories,
nonlinear static analysis (Seismostruct) with DAP [3] adaptive load distribution is
the most efficient in predicting the NTHA`s lateral drift profile. Also in the case of
lower stories, the drift profiles yielded by the proposed advanced pushover
procedure (NEFCAD) generally show better estimates than the results obtained with
Seismostruct, considering invariable triangular load distribution. An exception to
this are the upper storey drift predictions of the 9 and 15 story structures, where the
proposed NEFCAD advanced pushover analysis with triangular load distribution –
similarly to the Seismostruct adaptive load pattern analyses, overestimate the story
drifts.

6 Conclusions
Due to the well-known limitations of the code-based NSA methods and the
complexity of the NTHA, structural engineers do not have at present a quick and
conceptually simple method for the determination of inelastic displacement of
structures subjected to recorded accelerograms.
A reliable and robust nonlinear inelastic analysis tool has been presented for
seismic design of RC frame structures. The proposed nonlinear static formulation is
intended to model the geometrically nonlinear inelastic behaviour of RC beam-
columns using only one element per member. This is an essential requirement to
approach real large spatial frame structures, combining modelling benefits,
computational efficiency and reasonable accuracy. The proposed approach is able to
determine local and global displacements of two-dimensional frame structures
extremely fast and with sufficient accuracy for a set of registered earthquake
records, by combining the advanced pushover analysis of the MDOF structural
system with the NTHA of a SDOF oscillator.
For low- and mid-rise structures, the proposed method produced fair estimates of
the of upper story displacements obtained by NTHA. As compared to the code-
based NSA, upper story displacements were improved in each case. On the contrary
proposed method yielded better results in estimating the upper story displacements
in the case of low- and mid- rise structures, while in the case of high-rise structures
the estimates did not differ in significant manner, with respect to the code-based
NSA. Compared to the NTHA, the proposed method has the advantage of speed,
repeatability and graphical representation of seismic capacity and demand through
spectral acceleration-displacement format. Further researches are needed both on

17
ineleastic structural analysis (MDOF) and equivalent SDOF levels in order to
capture the tri-dimensional effects and to attain goals of the research, namely the
quick and accurate seismic assessment of 3D RC frameworks.
Future work is envisaged in order to refine the proposed NSA analysis and the
NEFCAD computer program in order to be able to follow more closely the time
variant distribution of inertial forces by adopting adaptive force distribution or
modal pushover procedures and integrating in a single software package the
proposed method for determination of inelastic seismic displacements. The
determination of the capacity curve through advanced pushover analysis can be
achieved extremely fast, while bilinear idealization and determination of the
inelastic ductility demand can be achieved in fractions of a second. Hence, the
software resulted would be a powerful tool for quick assessment of reinforced
concrete frameworks.

References

[1] Vasilopoulos, A.A., Beskos, D.E., Seismic deisgn of space steel frames using
advanced method of analysis, Soil Dynamics and Earthquake Engineering, 29,
194-208, 2009.
[2] A.K. Chopra, R.K., Goel, A modal pushover analysis procedure for estimating
seismic demands for buildings, Earthquake Engineering Structural Dynamics,
31: 561–582, 2002.
[3] S. Antoniou, R. Pinho, Development and verification of a displacement-based
adaptive pushover procedures, J. Earthquake Engineering, 8(5): 643–661,
2004.
[4] K. Shakeri, M. A. Shayanfar, and T. Kabeyasawa, "A story shear-based
adaptive pushover procedure for estimating seismic demands of buildings,"
Engineering Structures, 32(1), 2010.
[5] P. Fajfar, A Nonlinear Analysis Method for Performance‐Based Seismic
Design, Earthquake Spectra, 16(3), pp.573-592, 2000.
[6] CEN, EN 1998-1: Eurocode 8, “Design of structures for earthquake resistance;
Part 1: General rules, seismic actions and rules for buildings”. Brussels:
Comite Europeen de Normalisation, 2004.
[7] FEMA, Report FEMA 356: “Prestandard and commentary for the seismic
rehabilitation of buildings”: Federal Emergency Management Agency,
[8] FEMA, Report FEMA 440: “Improvement of nonlinear static analysis
procedures”. Washington, DC: Federal Emergency Management Agency,
2005.
[9] ATC, ATC-40: “Seismic Evaluation and Retrofit of Concrete Buildings”.
Redwood City, CA: Applied Tehnology Council, 1996.
[10] S. Varga, "An Alternative Method For The Determination of Inelastic
Displacements Using Pushover Analysis and Directly Generated Inelastic
Spectra," Acta Tehnica Napocensis:Civil Engineering & Architecture, 57(1),
2014.

18
 [11] C.G. Chiorean, A computer method for nonlinear inelastic analysis of 3D
composite steel-concrete frameworks, Engineering Structures 57: 125-152,
2013.
[12] M. Aschheim, E. F. Black, Yield point spectra for seismic design and
rehabilitation, Earthquake Spectra, 16(2), pp. 317-336, 2000.
[13] A.K. Chopra, Dynamics of Structures: Theory and Applications to Earthquake
Engineering, 4th ed.: Prentice Hall/Pearson Education, 2012.
[14] M.M. Hachem, BISPEC. Interactive software for the computation of
unidirectional and bidirectional nonlinear earthquake spectr, version 2.20,
Help manual. http://www.eqsols.com/Pages/Bispec.aspx, 2013.
[15] SeismoSoft,SeismoMatch,version2.0.0,http://www.seismosoft.com/en/downlo
ad_details.aspx?ID_Download=6., 2012.
[16] SeismoSoft. SeismoStruct – A Computer Program for Static and Dynamic
Nonlinear Analysis of Framed Structures (Help file), (online),:
http://www.seismosoft.com, 2014.

19

View publication stats