You are on page 1of 12

Arab J Sci Eng (2014) 39:3499–3510

DOI 10.1007/s13369-014-1011-0

RESEARCH ARTICLE - CIVIL ENGINEERING

Distribution of Damage in Plan of Dual Steel Frames


Gholamreza Abdollahzadeh · Soheil Niknafs ·
Saeed Rabbanifar

Received: 17 September 2012 / Accepted: 31 March 2013 / Published online: 11 March 2014
© King Fahd University of Petroleum and Minerals 2014

Abstract Earthquake events with magnitudes larger than Keywords Seismic damage index · Hysteretic energy ·
eight Richter along subduction zones have been reported Nonlinear dynamic analysis
worldwide. Due to large number of load reversals and ex-
cessive hysteretic energy, the effect of cumulative damage
on structural components due to deterioration becomes crit-
ical for buildings designed based on current seismic codes.
By specifying the damage index of a structure from its real
inelastic behavior, the required criterion for strengthening
would be given. In this paper, three steel structures with
dual systems consisting of intermediate moment-resisting
frames and concentrically braced ones were selected and de-
signed based on ASD method of UBC-97. Then, for eval-
uating inelastic behavior, these structures were subjected to
three earthquake records and the nonlinear dynamic analyses
were carried out by the PERFORM 3D (VER 4.0.1) software.
Next, hysteretic energy and damage measure were computed
for all members of the structures. It is observed that in spite
of uniformity of strength ratios along height and also among
resisting elements of each story, structural damages among
such members do not confirm this uniformity and most of the
damage of columns and beams is correlated to the external
bracing frames. Of course, some regularity in the damage
distribution has been seen in plans of buildings, so that con-
centration of damage in the center of the plan is less than
that of external frames. Thus, approaching the center of the 1 Introduction
plans, the damage imposed on the members decreases.
In recent years, the general evolution of the structural de-
sign criteria adopted for the modern structures and the ma-
jor importance of the evaluation of the seismic behavior of
G. Abdollahzadeh (B) · S. Rabbanifar existing under-designed buildings have extended the objec-
Faculty of Civil Engineering, Babol University of Technology, tives in the seismic design; safety against the collapse re-
Babol, Iran mains the most important objective, while performance in
e-mail: abdollahzadeh@nit.ac.ir
terms of functionality and economy is assumed to play a
S. Niknafs central role in the design criteria [1,2]. Therefore, several
Department of Civil Engineering, Amol University, Amol, Iran authors have discussed the need of an improvement of the

123
3500 Arab J Sci Eng (2014) 39:3499–3510

current earthquake-resistant design methods in order not only damage indexes. Each damage index uses specific damage
to avoid the collapse for destructive earthquake, but also to parameters and the parameters used to categorize the dam-
limit the seismic damage for moderate earthquakes. Mov- age index. Damage indexes are usually normalized so that
ing from a simple force-strength approach the new design their value is equal to zero when there is no damage and is
philosophy tends toward multi-level probabilistic structural equal to unity when total collapse or failure occurs. On the
performance criteria [3–5]. Also in the seismic design of other hand, a damage parameter is a quantity that is used for
structures, the concepts of damage and vulnerability play a estimating the damage. A damage index can involve a com-
central role. It is accepted that standard design procedures, bination of one or more damage variables in its calculation.
based on the concept of the force reduction factor, even if As a result, to calculate damage indexes, damage parameters
adequate in most practical cases, do not result in structures should also be normalized [10].
possessing uniform and indexes or damage indicators have The earliest damage indexes were mostly based on dis-
become popular [6]. Modern codes for seismic design imply placement or rotational ductility only. For example, Banon
large values of the seismic force reduction factor relying on et al. [11] used the rotational ductility (μθ ) at the end of a
the design and detailing strategy, which allows a substantial structural member as its damage index:
dissipation of hysteretic energy supplied by spreading the θm θm − θy
ductility demands in large parts of the structure [7]. μθ = =1+ (1)
θy θy
For more than a decade now, performance-based seismic
design (PBSD) has been at the forefront of earthquake engi- where θm is the maximum rotation (including both elastic
neering research. One of the prime aspects of PBSD is the and plastic rotations) under an earthquake and θy is the yield
realistic characterization of seismic structural damage and its rotation, considering the member’s anti-symmetric double
direct incorporation in the design or performance evaluation curvature bending with the point of contra flexure in its mid-
methodology. In addition, a major emphasis is also placed span. Such ductility-based damage indexes fail to take into
on the consideration of all the uncertainties in the design account the effects of repeated cycles including the strength
and evaluation of structures. The various modes of charac- and stiffness degradations under low-cycle fatigue. The “flex-
terizing the potential seismic damage lead to various PBSD ural damage ratio” [11] and the “modified flexural damage
approaches [8]. ratio” [12] are two other damage indexes also suffer from
The present study focuses on the evaluation of the damage these shortcomings. One of the earliest cumulative damage
distribution pattern in all parts of dual steel buildings accord- parameters was the “normalized cumulative rotation” [11],
ing to Park–Ang damage index, designed in accordance with which is defined as the ratio of the sum of all plastic ro-
UBC-97 [9]. To obtain parameters of damage indexes, non- tations (except for unloading parts) in inelastic springs to
linear dynamic analyses have been carried out using PER- the yield rotation. Several other researchers [13–15] also de-
FORM 3D Software. The PERFORM 3D is a complete and fined similar displacement-/rotation-based cumulative dam-
powerful program to perform nonlinear dynamic analyses age indexes, while some others [16,17] followed a low-cycle
through which many nonlinear parameters could be achieved. fatigue-based approach to define damage indexes in terms of
Therefore, this program had been used for nonlinear dynamic the number of cycles to failure.
modeling. Among the many damage indicators available, the Park–
Ang damage index appears to be the most promising due to its
simplicity and extensive calibration against experimentally
observed seismic damage in reinforced concrete structures,
2 Damage Indexes for Structures although it is less reliable in the case of steel structures [18,
19]. The relation between Park and Ang damage index and
Some parameters of the building response have been pro- damage state is shown in Table 1 [18]. In this method, the
posed as indexes of structural damage. They are shortly called damage index, DPA,I , consists of a linear combination of the

Table 1 The relation between damage index and damage state [18]
Degree of damage Physical appearance Damage index State of building

Slight Sporadic occurrence of cracking <0.1 No damage


Minor Minor cracks; partial crushing of concrete in columns 0.1–0.25 Minor damage
Moderate Extensive large cracks; spalling of concrete in weaker elements 0.25–0.4 Repairable
Severe Extensive crashing of concrete; disclosure of buckled reinforcement 0.4–1.0 Beyond repair
Collapse Partial or total collapse of building >1.0 Loss of building

123
Arab J Sci Eng (2014) 39:3499–3510 3501

ductility and energy dissipation indexes: be based on moment–curvature or fiber stress–strain relation-
ships. PERFORM 3D software provides two components that
Umax β  can be used for this type of model, namely curvature hinges
DPA,I= + dE, (2)
Uu Q r Uu and fiber segments. Figure 2 shows a finite element model
using curvature hinges. The current version of PERFORM
where Umax = is the maximum response deformation, Uu = imposes a limit of all components in a beam or column com-

the ultimate deformation under monotonic loading, dE = pound component.
dissipated hysteretic energy, Q r = yield strength, and β = Each material and each basic structural member have one
is a nonnegative constant. or more actions or forces and corresponding deformations.
The global damage index is a weighted average of the For example, for a simple material the action is stress and the
local damage indexes and the dissipated energy is chosen as deformation is strain, and for a simple plastic hinge the action
the weighting function. The global damage index is given by is bending moment and the deformation is hinge rotation.
the following relation: The relationship between these two parameters is the F–D
n relationship.
i=1 DIL E i
DIG =  n (3) Most of the inelastic members in PERFORM 3D software
i=1 E i have the same form of the F–D relationship. This is a trilinear
where DIG is the global damage index, DIL is the local dam- relationship with optional strength loss, as shown in Fig. 3.
age index after Park–Ang, E i is the energy dissipated at lo- Also, for many members an elastic–perfectly plastic rela-
cation I and n is the number of locations at which the local tionship, rather than a trilinear relationship, may be adequate.
damage is computed. In this case, the Y and U points are the same.
Values of β about 0.15, derived by fitting test results [20], Some members may continue to strain hardening without
were used in the literature for reinforced concrete structures reaching an ultimate load. PERFORM 3D software allows
[21], while in the case of steel structures a value of β = 0.025, this for some members, by an additional parallel stiffness as
which is used in the present study, can be adopted [22]. A shown in Fig. 4.
comparison of the effectiveness of different damage indexes
can be found in many research publications [20,23–26].

3 PERFORM 3D Performance

The PERFORM 3D software [27] has been used for numer-


ical modeling and analyzing of the structures. In a detailed
finite element model, each beam or column member is di-
vided into a number of elements along its length as shown in
Fig. 1, schematically.
There are many different finite elements that might be
used. In general, they can be of low or high order and can
Fig. 3 PERFORM 3D action–deformation relationship

Fig. 1 Detailed finite element model [27]

Fig. 2 Finite element model with hinges [27] Fig. 4 Additional parallel stiffness

123
3502 Arab J Sci Eng (2014) 39:3499–3510

Fig. 5 Strength loss

In a structural member, “brittle” strength loss can be component loses strength, if possible the lost strength is re-
caused by a number of effects, including tensile fracture, con- distributed to adjacent members and the resulting behavior
crete crushing, concrete shear failure, and buckling. When a can be complex.
Usually it is not permissible to deform an inelastic member
beyond the L point. It means that the deformation capacity is
usually smaller than the L point deformation. For example,
the FEMA 356 criteria generally allow deformation beyond
the L point only for certain secondary members at the collapse
prevention performance level. Figure 5 shows the action–
deformation relationships for FEMA 356 (Q– relationship)
and PERFORM 3D (F–D relationship).
In the FEMA 356 relationship, there is sudden strength
loss at Point C and total strength loss at Point E. In the PER-
FORM relationship, strength loss begins at Point L and can
be sudden or gradual. It is likely that strength loss in an actual
structure will be gradual, and hence sudden strength loss is
not realistic. Also in the FEMA 356 relationship, there is to-
Fig. 6 FEMA356 action–deformation relationship

Table 2 Nonlinear parameters


Members a b c

Beams
A: h
tw ≤ √
3,185
, bf
≤ √420 9θy 11θy 0.6
Fye 2tf Fye
B: h
tw ≥ √
5,365 bf
, ≤ √545 4θy 6θy 0.2
Fye 2tf Fye
Columns
A: h
tw ≤ √
2,500
, bf
≤ √420 9θy 11θy 0.6
Fye 2tf Fye
B: h
tw ≥ √
3,850 bf
, ≤ √545 4θy 6θy 0.2
Fye 2tf Fye

Fye expected yield stress of material (kg/cm2 ), h height of the section (cm), tw web thickness (cm), bf flange width (cm), tf flange thickness (cm)

Fig. 7 Buildings plan and sections geometries

123
Arab J Sci Eng (2014) 39:3499–3510 3503

Table 3 Frames basic design properties


Structure Natural period (s) Base shear (ton) Mass of structure (ton)

3-story 0.526 95.21 103.56


5-story 0.783 160.47 174.54
8-story 1.227 205.56 282.88

Table 4 Strong ground motion parameters


Earthquake Date and time Station Component PGA (g) Magnitude (Ms) Duration (s) Scale factor

Northridge 1/17/1994 12:31 Covina—S. Grand Ave. GRA074 0.066 6.7 35 0.91
GRA344 0.062
Loma Prieta 10/18/1989 12:05 APEEL 3E Hayward CSUH A3E000 0.078 6.9 40 1.22
A3E090 0.084
San Fernando 2/9/1971 14:00 Wrightwood—6074 Park Dr WTW025 0.061 6.6 20 1.14
WTW295 0.044

4 Nonlinear Dynamic Analysis of Structures

The steel structures with dual system consisting of intermedi-


ate moment-resisting frames and concentrically braced ones
have been designed according to the requirements of the ASD
method of the UBC-97. All of the structures are symmetrical.
The cross sections of the beams and columns are general I-
shape sections and the box sections are used for the bracings.
In this paper, sections have been optimized and have stress
ratios between 0.95 and 1. So the impact of the earthquake
has been seen on the optimized sections with almost 1 stress
ratio.
Plans and geometry of the buildings are shown in Fig. 7.
All of the plans in all buildings are similar and the bay width
is 4 m and the story height is 3.2 m, which is common in
residential buildings. Some structural design properties are
shown in Table 3. All buildings have been designed consid-
ering a response reduction factor R of 7, corresponding to
the dual system consisting of intermediate moment-resisting
frame and concentrically braced frames. The buildings have
been classified as “importance class III” structures according
to the requirements of UBC-97 code, with subsoil of type C
Fig. 8 Response spectra of the scaled acceleration records and UBC and regional seismicity of category 1. In addition to the dead
design spectrum. a Relative displacement, b acceleration
weight and the seismic loading, snow loads and live loads
have been taken into account, as well as lateral loads due to
tal strength loss at Point E but in the PERFORM relationship column sway. In this paper, the steel type St-37 was used for
total strength loss at Point X is optional. P- and large dis- all structural members.
placement effects can cause nonlinear behavior in elements After designing and detailing the steel structures, nonlin-
and hence in the whole structure. This is usually referred to ear dynamic analyses were carried out for evaluating struc-
as “geometric” nonlinearity. PERFORM-3D gives options to tural seismic responses. In this way, the hysteretic behavior of
include or ignore geometric nonlinearity. Nonlinear parame- the beams, columns and braces has been specified. This hys-
ters in this model are recommended by FEMA356. They are teretic model incorporates stiffness degradation, strength de-
shown in Fig. 6 and Table 2. terioration and non-symmetric response. The inelastic

123
3504 Arab J Sci Eng (2014) 39:3499–3510

DIC= 0.602 DIC= 0.587 DIC= 0.557 DIC= 0.574


DIC= 0.717 DIC= 0.609 DIC= 0.699 DIC= 0.504
DIB= 0.035 DIBR= 0.622 DIB= 0.054 DIBR= 0.627 DIB= 0.053
DIBR= 0.72 DIB= 0.095 DIBR= 0.72 DIB= 0.088 DIB= 0.042 DIBR= 0.481 DIB= 0.038 DIBR= 0.363 DIB= 0.04

DIC= 0.717 DIC= 0.477 DIC= 0.702 DIC= 0.47

DIC= 0.711 DIC= 0.698 DIC= 0.608 DIC= 0.454


DIC= 0.65 DIC= 0.098 DIC= 0.037 DIC= 0.036 DIC= 0.07
DIC= 0.703 DIC= 0.04 DIC= 0.035
DIB= 0.033

DIBR= 0.699 DIB= 0.036 DIBR= 0.698

DIBR= 0.353

DIBR= 0.243
DIC= 0.674 DIC= 0.237 DIC= 0.57

DIC= 0.495 DIC= 0.219 DIC= 0.094 DIC= 0.628

DIBR= 0.441
DIBR= 0.629
DIC= 0.649 DIC= 0.233 DIC= 0.154 DIC= 0.107 DIC= 0.121 DIC= 0.644
DIB= 0.09

DIC= 0.477 DIC= 0.394 DIC= 0.477 DIC= 0.344 DIC= 0.48 DIC= 0.407 DIC= 0.431 DIC= 0.385

DIBR= 0.392 DIB= 0.042 DIBR= 0.377 DIB= 0.04 DIB= 0.036 DIBR= 0.388 DIB= 0.034 DIBR= 0.246 DIB= 0.036

Damage index of story 1 Damage index of story 2

DIC= 0.112 DIC= 0.338 DIC= 0.355 DIC= 0.239 DIC= 0.269
DIBR= 0.707 DIBR= 0.707 DIB= 0.036

DIC= 0.527 DIC= 0.232 DIC= 0.469 DIC= 0.257


DIBR= 0.72 DIB= 0.036

DIC= 0.044
DIC= 0.341 DIC= 0.06 DIC= 0.034 DIC= 0.034
DIBR= 0.582

DIC= 0.207 DIC= 0.038

DIC= 0.265 DIC= 0.044 DIC= 0.047


DIBR= 0.545
DIBR= 0.72

DIC= 0.171 DIC= 0.181 DIC= 0.075 DIC= 0.044 DIC= 0.045

DIC= 0.089 DIC= 0.057

DIBR= 0.615 DIBR= 0.62

Damage index of story 3

Fig. 9 Damage index of 3-story building’s members

behavior of steel elements is determined by trilinear skeleton tal components of far-fault records from major earthquake
curve. Degrading parameters have been chosen from exper- events. Scaling of the original seismic acceleration records,
imental results of cyclic force–deformation characteristics which modifies the nonlinear and the damage effects, has
of typical components of the studied structures [27]. Thus, been carried out in accordance with UBC-97. For each pair
the nominal parameter for stiffness degradation and strength of horizontal ground motion components, the square root of
deterioration has been chosen. the sum of the squares (SRSS) of the 5 % damped spectrum of
The acceleration records, selected for site class C and their the normalized horizontal components has been constructed.
scaled response spectra, of all seismic excitations given in The motions have been scaled such that the average SRSS
Table 4, were used as input data for the nonlinear dynamic spectrum does not fall below 1.4 times the 5 % damped de-
analyses. Earthquake records include three pairs of horizon- sign spectrum for periods from 0.2 to 1.5 T where T is the

123
Arab J Sci Eng (2014) 39:3499–3510 3505

Fig. 10 Damage index of


5-story building’s members
DIB= 0.034 DIBR= 0.215 DIB= 0.034 DIBR= 0.215 DIBR= 0.123 DIB= 0.034 DIBR= 0.123

DIBR= 0.387 DIB= 0.043 DIBR= 0.218 DIB= 0.042


DIB= 0.035 DIBR= 0.536 DIB= 0.078 DIBR= 0.535 DIB= 0.038

DIBR= 0.222 DIC= 0.108

DIBR= 0.209 DIB= 0.039 DIBR= 0.183 DIB= 0.036


DIC= 0.113 DIC= 0.092 DIC= 0.11 DIC= 0.094
DIC= 0.239

DIC= 0.203 DIC= 0.112 DIC= 0.158


DIC= 0.253 DIC= 0.215

DIC= 0.382 DIC= 0.24

DIBR= 0.223
DIB= 0.034 DIC= 0.308

DIB= 0.036

DIB= 0.034

DIB= 0.034
DIC= 0.072 DIC= 0.207 DIC= 0.176 DIC= 0.246 DIC= 0.065 DIC= 0.052

DIBR= 0.331 DIB= 0.068 DIBR= 0.33 DIBR= 0.227 DIB= 0.07 DIBR= 0.177

Damage index of story 1 Damage index of story 2

DIC= 0.268 DIC= 0.233 DIC= 0.345 DIC= 0.167 DIC= 0.042 DIC= 0.349 DIC= 0.326 DIC= 0.349 DIC= 0.363
DIBR= 0.261 DIB= 0.075 DIBR= 0.259 DIB= 0.042 DIB= 0.034 DIBR= 0.465 DIB= 0.08 DIBR= 0.294 DIB= 0.075

DIBR= 0.286 DIB= 0.056 DIBR= 0.285 DIB= 0.048


DIC= 0.124 DIC= 0.233 DIC= 0.176 DIC= 0.209 DIC= 0.141

DIC= 0.184 DIC= 0.249 DIC= 0.177 DIC= 0.269


DIB= 0.038 DIBR= 0.37 DIB= 0.113 DIBR= 0.368 DIB= 0.066

DIC= 0.121 DIC= 0.092 DIC= 0.084 DIC= 0.119 DIC= 0.233 DIC= 0.404 DIC= 0.062 DIC= 0.054
DIB= 0.034

DIB= 0.035 DIBR= 0.452 DIB= 0.08 DIBR= 0.261

DIBR= 0.258 DIB= 0.035 DIBR= 0.18


DIC= 0.285 DIC= 0.137 DIC= 0.088 DIC= 0.233 DIC= 0.128 DIC= 0.664 DIC= 0.071
DIB= 0.034

DIC= 0.309 DIC= 0.173 DIC= 0.233 DIC= 0.11 DIC= 0.155 DIC= 0.455

DIC= 0.242 DIC= 0.237 DIC= 0.14 DIC= 0.165 DIC= 0.084 DIC= 0.7 DIC= 0.107 DIC= 0.057 DIC= 0.06 DIC= 0.104
DIB= 0.039

DIC= 0.212 DIC= 0.32 DIC= 0.451 DIC= 0.356 DIC= 0.402 DIC= 0.508 DIC= 0.49 DIC= 0.387 DIC= 0.455

DIB= 0.036 DIBR= 0.388 DIB= 0.093 DIBR= 0.366 DIB= 0.088 DIB= 0.084 DIBR= 0.562 DIB= 0.089 DIBR= 0.392 DIB= 0.078

Damage index of story 3 Damage index of story 4

DIBR= 0.509 DIB= 0.071 DIBR= 0.504


DIC= 0.092

DIC= 0.084 DIC= 0.038


DIC= 0.171
DIBR= 0.713

DIBR= 0.636

DIC= 0.166
DIC= 0.274

DIC= 0.292
DIBR= 0.659

DIBR= 0.642

DIC= 0.111

DIC= 0.089

DIB= 0.035 DIBR= 0.658 DIB= 0.115 DIBR= 0.654

Damage index of story 5

123
3506 Arab J Sci Eng (2014) 39:3499–3510

Fig. 11 Damage index of


8-story building’s members
DIBR= 0.095 DIBR= 0.095
DIBR= 0.079 DIBR= 0.079

DIB= 0.101 DIBR= 0.58 DIB= 0.117 DIBR= 0.276 DIB= 0.035
DIB= 0.042

DIB= 0.038

DIB= 0.037

DIB= 0.035
DIC= 0.067 DIC= 0.087

DIBR= 0.164
DIB= 0.045 DIBR= 0.453 DIB= 0.118 DIBR= 0.453

DIB= 0.035 DIBR= 0.276 DIB= 0.039 DIBR= 0.276


DIB= 0.035
DIB= 0.04
DIC= 0.055

DIC= 0.071 DIC= 0.074

DIBR= 0.206
DIB= 0.041

DIB= 0.036
DIC= 0.106 DIC= 0.039
DIB= 0.035

DIB= 0.034
DIB= 0.035

DIB= 0.035

DIB= 0.034
DIC= 0.078 DIC= 0.037 DIC= 0.041 DIC= 0.038

DIBR= 0.176 DIB= 0.033 DIBR= 0.173 DIB= 0.035 DIBR= 0.394 DIB= 0.082 DIBR= 0.189

Damage index of story 1 Damage index of story 2

DIBR= 0.073 DIB= 0.038 DIBR= 0.073


DIBR= 0.079 DIBR= 0.079

DIB= 0.084 DIBR= 0.405 DIB= 0.236 DIBR= 0.228 DIB= 0.041

DIB= 0.076

DIB= 0.037

DIB= 0.036
DIB= 0.04
DIBR= 0.465 DIB= 0.146 DIBR= 0.475 DIB= 0.036

DIC= 0.116
DIC= 0.093

DIB= 0.039 DIBR= 0.133 DIB= 0.091 DIBR= 0.132


DIBR= 0.117 DIB= 0.07 DIBR= 0.117
DIC= 0.596 DIC= 0.178
DIC= 0.072

D IC= 0.69 5 D IC = 0 . 4 3 8

DIC= 0.051 DIC= 0.137

DIB= 0.033 DIB= 0.035 D IB = 0 . 0 3 5 D IB = 0 . 0 3 6

DIB= 0.041

DIB= 0.034
DIB= 0.037

DIC= 0.065 DIC= 0.205 DIC= 0.318 DIC= 0.284 DIC= 0.108

DIB= 0.041 DIBR= 0.294 DIB= 0.101 DIBR= 0.275 DIB= 0.044 DIBR= 0.282 DIB= 0.132 DIBR= 0.129 DIB= 0.035

Damage index of story 3 Damage index of story 4

fundamental period of the structure (UBC 1997). Response the figures. The Park–Ang damage index has been used for
spectra of the scaled acceleration records of Loma Prieta, evaluation of dual steel elements and all the buildings have
Northridge and San Fernando earthquakes and the design been designed in accordance with UBC-97. Dynamic non-
spectrum of UBC-97 are shown in Fig. 8. linear analyses have been carried out by PERFORM 3D for
obtaining the parameters of damage indexes. Because of the
large number of analyses with large stiffness matrixes, e.g.,
5 Evaluation of the Results of Nonlinear Dynamic 388 nodes and 1,024 elements for the 8-story building, an-
Analysis in Plan alyzing and evaluating the damage distribution in all parts
of the buildings are very time consuming and difficult. In
Usually perpetual structural damage is seen in the end of an the Figs. 9, 10 and 11: DIBR, damage index of brace; DIC,
earthquake. Therefore, distribution of damage in this time damage index of column; DIB, damage index of beam.
is an expressive of perpetual damage in stories of the build- According to the Fig. 12, maximum damage is seen in the
ing. In this research, the structures have been subjected to external braced frames. In these frames, columns that are
two scaled horizontal components of acceleration records, correlated with brace show the higher damage, both in the
simultaneously. Average damage of the three earthquakes in form of high absorption of hysteretic energy and that of high
the members of 3-, 5- and 8-story buildings has been com- displacement.
puted and is shown in Figs. 9, 10 and 11, respectively. In this section, damage indexes of the buildings are evalu-
In these figures, only the damaged members have been ated along their heights. It seems that higher stories undergo
defined. It means that damage indexes have been computed more serious damages than lower stories. Damages along
and shown for members that have experienced inelastic de- height of buildings are shown in Fig. 12 which is result of
formations. Also, the braced spans have been highlighted in average of Loma prieta, Northridge and San Fernando earth-

123
Arab J Sci Eng (2014) 39:3499–3510 3507

Fig. 11 continued
DIC= 0.154
DIB= 0.053 DIBR= 0.175 DIB= 0.109 DIBR= 0.136
DIC= 0.119
DIB= 0.037 DIBR= 0.04 DIB= 0.05 DIBR= 0.043

DIB= 0.173 DIBR= 0.586 DIB= 0.223 DIBR= 0.431 DIB= 0.123

DIB= 0.049

DIB= 0.089

DIB= 0.093

DIB= 0.035

DIC= 0.29
DIC= 0.253 DIC= 0.315 DIC= 0.237 DIC= 0.135 DIC= 0.153

DIB= 0.174 DIBR= 0.561 DIB= 0.252 DIBR= 0.543 DIB= 0.129

DIB= 0.048
DIB= 0.058

DIB= 0.055

DIC= 0.408 DIC= 0.394 DIC= 0.49 DIC= 0.343


DIB= 0.52
DIC= 0.26 DIC= 0.199 DIC= 0.188 DIC= 0.095 DIC= 0.162 DIC= 0.63 DIC= 0.306 DIC= 0.38 DIC= 0.29 DIC= 0.153
DIC= 0.68 DIC= 0.17 DIC= 0.237 DIC= 0.243 DIC= 0.25

DIB= 0.087 DIBR= 0.446 DIB= 0.116 DIBR= 0.272


DIC= 0.124 DIC= 0.277 DIC= 0.233
DIB= 0.095 DIBR= 0.321 DIB= 0.117 DIBR= 0.316
DIC= 0.598 DIC= 0.308 DIC= 0.266 DIC= 0.142 DIC= 0.194
DIC= 0.655 DIC= 0.098 DIC= 0.237 DIC= 0.233 DIC= 0.1 DIB= 0.043

DIB= 0.035

DIB= 0.034

DIB= 0.034
DIC= 0.563 DIC= 0.201 DIC= 0.192 DIC= 0.156 DIC= 0.295
DIC= 0.69 DIC= 0.227 DIC= 0.14 DIC= 0.076 DIC= 0.24

DIC= 0.577 DIC= 0.304 DIC= 0.48 DIC= 0.483 DIC= 0.313
DIC= 0.717 DIC= 0.178 DIC= 0.247 DIC= 0.24 DIC= 0.167
DIB= 0.04
DIB= 0.035

DIB= 0.034
DIB= 0.035 DIC= 0.063 DIC= 0.342 DIC= 0.387 DIC= 0.411 DIC= 0.291
DIC= 0.065 DIC= 0.207 DIC= 0.37 DIC= 0.423 DIC= 0.219

DIB= 0.087 DIBR= 0.219 DIB= 0.141 DIBR= 0.245 DIB= 0.085 DIB= 0.083 DIBR= 0.387 DIB= 0.154 DIBR= 0.243 DIB= 0.1
DIB= 0.035
Damage index of story 5 Damage index of story 6

DIC= 0.097 DIC= 0.157


DIB= 0.045 DIBR= 0.393 DIB= 0.07 DIBR= 0.403 DIB= 0.035 DIBR= 0.707 DIB= 0.045 DIBR= 0.471 DIB= 0.036
DIB= 0.14 DIBR= 0.717 DIB= 0.186 DIBR= 0.717 DIB= 0.037

DIB= 0.119

DIB= 0.108

DIB= 0.115

DIC= 0.04
DIC= 0.065 DIC= 0.234 DIC= 0.119 DIC= 0.18
DIC= 0.176 DIC= 0.08 DIC= 0.095 DIC= 0.14 DIC= 0.249 DIC= 0.039 DIC= 0.04 DIC= 0.036 DIC= 0.046
DIC= 0.518 DIC= 0.162 DIC= 0.19 DIC= 0.12 DIC= 0.042 DIC= 0.097 DIC= 0.033

DIB= 0.108 DIBR= 0.727 DIB= 0.144 DIBR= 0.663

DIB= 0.034 DIBR= 0.717 DIB= 0.086 DIBR= 0.628


DIB= 0.077 DIBR= 0.419 DIB= 0.105 DIBR= 0.463

DIC= 0.07 DIC= 0.094 DIC= 0.075


DIC= 0.443 DIC= 0.22 DIC= 0.087 DIC= 0.035 DIC= 0.107 DIC= 0.124
DIB= 0.035

DIC= 0.593 DIC= 0.137 DIC= 0.051 DIC= 0.035 DIC= 0.05 DIC= 0.226

DIC= 0.428 DIC= 0.338 DIC= 0.155 DIC= 0.308 DIC= 0.109 DIC= 0.117 DIC= 0.177 DIC= 0.072
DIB= 0.073 DIB= 0.036

DIC= 0.036 DIC= 0.216 DIC= 0.247 DIC= 0.119 DIC= 0.247 DIC= 0.087 DIC= 0.052

DIB= 0.093 DIBR= 0.525 DIB= 0.125 DIBR= 0.506 DIB= 0.098 DIB= 0.076 DIBR= 0.717 DIB= 0.122 DIBR= 0.594 DIB= 0.077
DIB= 0.043 DIB= 0.036 DIB= 0.038 DIB= 0.038
Damage index of story 7 Damage index of story 8

quakes. This figure shows that in all structures, most damage accelerograms are imposed on the four sides of the buildings.
has occurred in the Northridge earthquake. Because of symmetric plans in all the buildings under study
of this survey, to comparison member’s damage or assess-
ment of one member damage, nonlinear dynamic analyse of
6 Variation of Damage on Braces of Surveyed Structures structures; subjected to the one pair of accelerograms which
imposed to one side of structures is sufficient. In compar-
In this section, damage distribution of the elements in the ing damages, the maximum damage of surveyed member
main x-braces and the beams adjacent to them is discussed and three other symmetric members has been chosen as a
and shown. Because of symmetric and square plans, dam- maximum damage in worst mode. In this section, damage
age distribution on two diagonals of the same span is similar. variation in the braces of surveyed structures has been eval-
In present study, two horizontal components of earthquake uated and shown in Figs. 13 and 14. These figures have been
records are exerted on the structural models simultaneously. concluded from the average of three earthquakes of Loma
It is also possible that earthquakes are imposed on the struc- Prieta, Northridge and San Fernando.
tures from every direction. Therefore, in the dynamic analysis In this section, because of the large number of plans,
we must impose two horizontal accelerograms in all angles only three of them have been considered. In accordance with
into the structures to find the worst mode in which members Figs. 13 and 14, some regularity in damage distribution can
have a maximum damage. But this method is time consum- be seen in plans of buildings. Damage has been reduced while
ing and almost impossible. Therefore, only two horizontal moving on diagonals of structure members. Therefore, dam-

123
3508 Arab J Sci Eng (2014) 39:3499–3510

Fig. 12 Damage distribution in stories of a 3-story building, b 5-story building, c 8-story building

Fig. 13 Variation of columns’ damage on diagonal of a story 1 of 3-story building, b story 3 of 5-story building, and c story 5 of 8-story building

123
Arab J Sci Eng (2014) 39:3499–3510 3509

Fig. 14 Variation of beams damage besides of diagonal of a story 1 of 3-story building, b story 3 of 5-story building, and c story 5 of 8-story
building

age concentration in the center of the plan is less than that of get to the center of plans, the less the damage of members
external frames. we see.
4. In upper and lower stories, braces have the most damage
among all members. Also in all parts of buildings, beams
7 Conclusions have a less damage than the other members.

In this study distribution of damage among structural mem-


bers of dual steel buildings was evaluated. Damage indexes
in all members of dual steel buildings with various number of
stories were evaluated, performing nonlinear dynamic analy- References
sis by PERFORM 3D software. The results of the study can
1. Bertero, V.V.: State of art report on design criteria. In: Proceedings
be summarized as followings: of 11th World Conference on Earthquake Engineering (1996)
2. Krawinkler, H.: Research issues in performance based seismic
engineering. In: Fajfar, P.; Krawinkler, H.: (eds.) Seismic design
1. In spite of uniformity of strength ratios in height and methodologies for the next generation of codes Balkema (1997)
among resisting elements of each story, damage values 3. SEAOC Vision 2000 Committee: Performance based seismic de-
do not have that uniformity and do not have a specific sign engineering. Report prepared by Structural Engineers Asso-
regularity in distribution of height and plan. ciation of California, Sacramento, USA (1995)
4. Bertero, R.D.; Bertero, V.V.; Teran-Gilmore, A.: Performance
2. The most damage of columns and beams are correlated based earthquake resistant design based on comprehensive design
to the external bracing frames and the maximum energy philosophy and energy concept. In: Proceedings of 11th World
that imposed to these frames is absorbed by braces, and Conference on Earthquake Engineering (1996)
in the second stage columns absorb the hysteretic energy. 5. Collins, K.R.; Wen, Y.K.; Foutch, D.: Dual-level seismic design: a
reliability based methodology. Earthq. Eng. Struct. Dyn. 25, 1433–
3. In plans of buildings, quite a regular distribution of dam- 1467 (1996)
age can be seen, so that damage concentration in center 6. Colombo, A.; Negro, P.: A damage index of generalized applica-
of plans is less than that in external frames. The closer we bility. Eng. Struct. 27, 1164–1174 (2005)

123
3510 Arab J Sci Eng (2014) 39:3499–3510

7. Dimova, S.L.; Negro, P.: Seismic assessment of an industrial frame 18. Park, Y.J.; Ang, A.H.-S.: Mechanistic seismic damage model for
structure designed according to Eurocodes. Part 2: Capacity and reinforced concrete. Struct. Eng. 111(4), 722–739 (1985)
vulnerability. Eng. Struct. 27, 724–735 (2005) 19. Park, Y.J.; Ang, A.H.-S.; Wen, Y.K.: Damage-limiting a seismic
8. Ghosh, S.; Datta, D.; Katakdhond, A.A.: Estimation of the Park– design of buildings. Earthq. Spectra. 3(1), 1–26 (1987)
Ang damage index for planar multi-storey frames using equivalent 20. Cosenza, E.; Manfredi, G.; Ramasco, R.: The use of damage func-
single-degree systems. Eng. Struct. 33, 2509–2524 (2011) tional in Earthquake resistant design: a comparison among different
9. Uniform building code. In: International conference of building procedures. Struct. Dyn. Earthq. Eng. 22, 68–855 (1993)
officials (1997) 21. Park, Y.J.: Seismic damage analysis and damage-limiting design
10. Estekanchi, H.; Arjomandi, K.: Comparison of damage indexes in of R/C structures. Ph.D thesis. Department of Civil Engineering,
nonlinear time history analysis of steel moment frames. Asian J. University of Illinois, Ubana (IL) (1984)
Civ. Eng. 8, 629–646 (2007) 22. Cosenza, E.; Manfredi, G.: Classificazione e comportamentosis-
11. Banon, H.; Biggs, J.M.; Irvine, H.M.; ASCE: Seismic damage mico di modelliciclicidegradanti. In: Proceedings of workshop
in reinforced concrete frames. J. Struct. Eng, 107(9), 29–1713 on Danneggiamentociclico e prove pseudo-dinamiche, pp. 59–74
(1981) (1994)
12. Roufaiel, M.S.L.; Meyer, C.; ASCE: Analytical modeling of hys- 23. Carr, A.J.; Tabuchi, M.: The structural ductility and the damage
teretic behavior of R/C frames. J. Struct. Eng. 113(3), 57–429 index for reinforced concrete structure under seismic excitation.
(1987) In: 2nd European conference on structural dynamics, vol. 1, pp.
13. Stephens, J.E.; Yao, J.T.P.; ASCE: Damage assessment using re- 76–169 (1993)
sponse measurement. J. Struct. Eng. 113(4), 787–801 (1987) 24. Kunnath, S.K.; Jenne, C.: Seismic damage assessment of inelas-
14. Wang, M.L.; Shah, S.P.: Reinforced concrete hysteresis model tic RC structures. In: 5th US national conference on earthquake
based in the damage concept. Earthq. Eng. Struct. Dyn. 15(8), 993– engineering, vol. 1, pp. 55–64 (1994)
1003 (1987) 25. Ghobarah, A.; Abou-Elfath, H.; Biddah, A.: Response-based dam-
15. Powell, G.H.; Allahabadi, R.: Seismic damage prediction by de- age assessment of structures. Earthq. Eng. Struct. Dyn. 28(1), 79–
terministic methods: concepts and procedures. Earthq. Eng. Struct. 104 (1999)
Dyn. 16(5), 34–719 (1988) 26. Castiglioni, C.A.; Pucinotti, R.: Failure criteria and cumulative
16. Jeong, G.D.; Iwan, W.D.: Effect of earthquake duration on damage damage models for steel components under cyclic loading. J. Con-
of structures. Earthq. Eng. Struct. Dyn. 16(8), 11–1201 (1988) str. Steel Res. 65, 751–765 (2009)
17. Chung, Y.S.; Meyer, C.; Shinozuka, M.: Modeling of concrete dam- 27. Powell, G.H.: CSI Perform 3D, User Guide, Version 4.0.1.RAM
age. ACI Struct. J. 86(3), 71–259 (1989) International, L.L.C., University of California, Berkeley (2006)

123

You might also like