Sei sulla pagina 1di 14

Vol.3, No.

EARTHQUAKE ENGINEERINGAND ENGINEERING VIBRATION

December, 2004

Article ID: 1671-3664(2004)02-0223-14

A simplified methd of evaluating the seismic


performance of buildings
Ashutosh Bagchi*
Department of Building, Civil and Environmental Engineering, Concordia University, Montreal, Quebec, H3G I MS, Canada

A b s t r a c t : This paper presents a simplified method of evaluating the seismic performance of buildings. The proposed

method is based on the transformation of a multiple degree of freedom (MDOF) system to an equivalent single degree of
freedom (SDOF) system using a simple and intuitive process. The proposed method is intended for evaluating the seismic
performance of the buildings at the intermediate stages in design, while a rigorous method would be applied to the final
design. The performance of the method is evaluated using a series of buildings which are assumed to be located in Victoria
in western Canada, and designed based on the upcoming version of the National Building Code of Canada which is due to
be published in 2005. To resist lateral loads, some of these buildings contain reintbrced concrete moment resisting frames,
while others contain reinforced concrete shear walls. Each building model has been subjected to a set of site-specific seismic
spectrum compatible ground motion records, and the response has been determined using the proposed method and the
general method for MDOF systems. The results from the study indicate that the proposed method can serve as a useful tool
for evaluation of seismic perfoNnanceof buildings, and carrying out performance based design.

Keywords: seismic hazard; modal analysis; static pushover analysis; dynamic time history analysis; performance-based
design

1 Introduction
The seismic design of a structure is performed
with the anticipation that a severe earthquake would
cause some damage, but that it would be contained to
an acceptable limit. The extent of damage sustained
by a building would determine the level of seismic
performance it achieved. The goal of performancebased design is to analyze the performance objectives
of a structure to specified levels of hazard, and design
the structure to achieve such performance. Evaluation
of the seismic performance of a building often involves
a number of analytical techniques including modal
analysis, static pushover analysis, and inelastic dynamic
time history analysis. The design of a structure typically
undergoes a number of revisions based on the capacity
and performance requirements, and an evaluation
of tile seismic performance corresponding to each
design revision could be quite cumbersome and time
consuming. To address this concern, many researchers
(Fajfar and Gaspersic, 1996; Mazzolani and Piluso,
1997; Ghobarah et al., 1997; Chopra and Goel, 2001)
Correspondence to: Ashutosh Bagchi, Department of Building,

Civil and Environmental Engineering Concordia University


1455 de Maisonneuve Blvd. West, BE-341, Montreal, Quebec,
H3G [MS, Canada
Tel: (514) 848-3213; Fax: (514) 848-7965
E-mail: abagchi@bcee.concordia.ca
tAssistant Professor
Received date: 2003-I1-04; Accepted date: 2004-10-15

have developed simplified methods for calculating the


dynamic response of buildings.
A simplified method, by nature, has limitations in
terms of accuracy. However, it could be very useful for
a quick evaluation of a building's performance in the
intermediate stages of design. Such a method need not be
very complex and computationally intensive. With that
objective in mind, a simplified method for the evaluation
of the seismic performance of buildings is developed in
this paper. The method is based on an equivalent single
degree of freedom system (SDOF) of a multi degree of
freedom (MDOF) system. The equivalent SDOF system
is derived from the load-deformation curve obtained
from static pushover analysis of the MDOF system. The
peak response of the SDOF system is obtained from
dynamic or response spectrum analysis. The pushover
results of the MDOF system could be used to derive
a relationship between the roof displacement and the
maximum story drift, and this relationship could be used
for interpreting the SDOF response to obtain an estimate
of the response of the actual structure.
Fajfar and Gaspersic, (1996) suggested a method
called the N2 method for seismic performance
evaluation. The method is based on an equivalent
SDOF system, derived using the static displacement
shape of the original structure produced by an assumed
distribution of the lateral forces and a two stage
pushover analysis is necessary to estimate the MDOF
response. The simplified analysis approach suggested
by Mazzolani and Piluso (1997), is based on the trilinear representation of the load-displacement curve and

224

EARTHQUAKEENGINEERINGAND ENGINEERINGVIBRATION

elastic and rigid-plastic analysis of the equivalent SDOF


system. Relationships between roof-displacement and
the damage index have been established in Ghobarah et
al. (1997) for the seismic performance evaluation. Such
relationships could be used for interpreting the response
of the SDOF structure and evaluating the performance
of the building based on the SDOF response. Chopra
and Goel (2001) have combined the strengths of various
simplified methods and developed a method based on
modal pushover analysis.
Although there are a number of simplified methods
are currently available, they are still complicated
for common use. In the method suggested here, an
equivalent SDOF system is derived based on the bilinear
idealization of the pushover curve of the original
structure. The peak response of the SDOF system for
a given hazard level can be obtained either by using
inelastic response spectra or performing inelastic
dynamic time history analysis using appropriate ground
motion records. The peak SDOF response is then
translated to the performance of the MDOF system using
the results of the pushover analysis performed earlier.
The proposed method bears some resemblance to the
N2 method mentioned above. However, the procedures
for deriving the equivalent SDOF system and the method
of estimating the MDOF response from the SDOF
response are quite different. Unlike the N2 method, the
static displacement shape of the original structure is not
used in the derivation of the equivalent SDOF system
and a second stage pushover analysis is not required.
The proposed method is illustrated with a number
of reinforced concrete buildings designed according
to the National Building Code of Canada (NBCC).
The National Building Code of Canada is currently
undergoing major revisions to incorporate the location
based seismic hazard spectra and the latest knowledge of
earthquake engineering. While the proposed method is
quite general, the case studies are based on the upcoming
version of NBCC and they are discussed with reference
to NBCC seismic provisions.

2 Evaluation of the seismic performance


Earthquake resistant design of buildings is based on
the concept of acceptable levels of damage under one
or more events of specified intensity. The acceptable
level of damage is related to the performance objective
of the building. For example, the objective may be
specified in the form of a requirement that the building
is fully operational with little or no damage during an
earthquake that has a 50% probability of exceedance
(PE) in 50 years; but may have moderate damage during
an earthquake with a 10% PE in 50 years.
The Geological Survey of Canada (GSC) has
recently developed site-specific spectra corresponding
to various levels of seismic hazard for a number of
locations in Canada (Adams et al., 1999). These spectra
are called the uniform hazard spectra (UHS). Figure 1 (a)

Vol.3

shows the spectra for Victoria. An earthquake event with


a return period of 475 years (10% PE in 50 years) would
be designated as UHS-500 for easy reference in the
text. Similarly, a 970-year event (10% PE in 100 years),
and a 2500-year event (2% PE in 50 years) would be
designated as UHS-1000 and UHS-2500, respectively.
The Vision 2000 committee (1995) of the Structural
Engineers Association of California (SEAOC) has
suggested performance objectives for buildings of
different types. A fully operational performance would
mean that the building has suffered minor damage, and a
repair may not be necessary; an operational performance
means that the building has sustained some damage,
but it is safe for post-earthquake occupancy, and minor
repair may be necessary; a life-safe performance level
means that the building has sustained extensive damage,
and it is not safe for post-earthquake occupancy without
major repair work; a near collapse performance means
the lateral load carrying capacity of the building has
diminished to a great extent, and the building is not
repairable. Similar definitions are adopted by the
upcoming version ofNBCC (Humar, 2000).

2.1 NBCC seismic provisions


The reinforced concrete buildings considered here
designed according to the seismic provisions of the
upcoming version of the National Building Code of
Canada. Thus a brief discussion of the seismic provisions
and the design philosophy of the new version of NBCC
would be of interest.
According to NBCC 1995, buildings are required
to be life-safe under an earthquake with a return
period of 475 years. However, such buildings have
considerable reserve strength to sustain earthquakes of
higher intensity. Suppose that a building has sufficient
over-strength to sustain twice the design shear without
collapse. From the hazard spectral shapes, it could
be observed that a building in western Canada would
be able to sustain a 2400-year earthquake although it
was designed to sustain a 475-year earthquake, while
a building in eastern Canada will be able to sustain an
earthquake with a return period of about 1600 years
(Humar, 2000). Consequently, the two buildings have
different levels of protection. Thus, to design the two
buildings to have the same levels of protection, the
design base shear should be derived for, say a 2500-year
earthquake and then reduced by a factor to account for
the reserve strength. This approach is adopted in the
new version of NBCC. Calculation of the design baseshear according to the new version on NBCC is briefly
discussed below.
The elastic base shear, Vo for a single degree-offreedom building can be obtained from the spectral
acceleration, S(T) corresponding to the period of the
building, T and weight of the building, IV.

vo= s (:r) w

(1)

No.2

Ashutosh Bagchi: A simplifiedmethd of evaluating the seismic performanceof buildings

The design spectral acceleration, S(T) value is calculated based on the site-specific values of the spectral
acceleration So(T) available from the Geological Survey
of Canada (Adams et al., 1999).

from the quantitative values of the damage parameters


is quite difficult. Vision 2000 (1995) provides some
guidelines for evaluating the performance of a building
based on the drift values (Table 1).
2.3

-FaS~ (0.2)

for T < 0.2s

FvSa(0.5 )

or F.S~(0.5)

whichever is less for T =0.5s

S(T)=

F~S (1.O) for T=l.Os


FvS(2.0 )

for T : Z . 0 s

(2)

F~S (2.0)/2 for T > 4.0s

where Fa and F~ are the foundation factors. The


design base shear is calculated using the following
expression.

v - s(ra)MvZw >_s(2.o)Mj

RoR

RoRa

(3)

where, My accounts for higher mode effect, Ic is the


importance factor, Rd and Ro account for ductility and
over-strength, respectively.
2.2 Damage parameters and performance metrics
Seismic performance of a building is generally
expressed in qualitative terms. However, for engineering
evaluation of performance, a quantitative measure of
damage is calculated. It is often necessary to correlate
the quantitative measures of the damage parameters to
the overall performance of the building in qualitative
terms. The selection of appropriate damage parameters
is very important for performance evaluation. Overall
lateral deflection ductility demand, inter-story drift, etc.,
are commonly used damage parameters. The damage
index developed by Park and Ang (! 985) is regarded as a
good representation for structural damage as it accounts
for the damage caused by cyclic deformations into the
post-yield level.
Obtaining a qualitative measure of performance

225

Analysis procedures

The evaluation of seismic performance of any


structure requires the estimation of its capacity, dynamic
characteristics and the prediction of its response to the
ground motion to which it could be subjected during its
service life. The commonly used methods for evaluating
the seismic performance of a building are, (1) static
pushover analysis, (2) modal analysis, and (3) dynamic
time history analysis.
DRAIN-2DX (Prakash et al., 1993), a general
purpose program for dynamic analysis of building
frames, is used for the analysis in the present study.
Selection of appropriate earthquake records is very
important for dynamic time history analysis. If actual
earthquake records are used, only those records whose
spectra closely match the uniform hazard spectrum
corresponding to the design earthquake level for the
site of interest can be used. As an alternative, the UHScompatible ground motion time histories developed by
Atkinson and Beresnev (1998) could be used.
Atkinson and Beresnev (1998) have produced
physically realistic time history records, which not only
match the hazard spectrum, but also are representative
of motions for specified magnitude distance scenarios in
the regions of interest. The UHS of a given location can
be represented by the composite two ground motions;
one corresponding to a short duration, low magnitude
earthquake at a very short distance and the other to a long
duration, high magnitude earthquake far away from the
site (Atkinson and Beresnev, 1998). The characteristics
of these artificial ground motion records are summarized
in Bagchi (2001). Figures 1 (b) and (c) show a set of
records corresponding to UHS-500. There are sets of
four records for UHS-500 and UHS-1000, whi!e a set of
eight records for UHS-2500 available for this study.
3

Details of the proposed m e t h o d

Static pushover analysis is performed to construct


the base shear - top displacement curve for the frame.
A curve representing the relationship between the roofdisplacement and the maximum inter-story drift (let us

Table 1 Permissible structural damage at various performance levels (Vision 2000, 1995)
Damage

Performance level

parameter
Drift

Fully o p e r .

Operational

Life-safe

Near collapse

Collapse

(a) Transient
(b) Permanent

< 0.2%
Negligible

< 0.5%
Negligible

< 1.5%
< 0.5%

< 2.5%
< 2.5%

> 2.5%
> 2.5%

226

EARTHQUAKEENGINEERINGAND ENGINEERINGVIBRATION

Vol.3

Fig. 1 Uniform hazard spectra and sample ground motion records for Victoria

call it the drift curve) is also constructed. The pushover


curve is then idealized as a bilinear force-displacement
relationship as shown in Fig. 2 (a), where Fy and Dy
are the effective yield force and yield displacement,
respectively; and Fr is the force corresponding to a
reference displacement beyond yield. The idealized
bilinear curve is constructed based on the approximate
equal energy criteria. The equivalent stiffness of the
single degree of freedom system (SDOF) is obtained
from the following expressions
k = Fy/De

(4)

where k is the stiffness of the equivalent SDOF system.


The period of the SDOF system is given by
T = 2rt ( m x / ~

(5)

where m is the mass of the equivalent system. The period


T is taken to be equal to the fundamental period of the
structure as obtained from the modal analysis of the
MDOF model. Equation 5 thus allows the determination
of mass m of the equivalent system.
A bilinear behaviour is assumed for the SDOF
system. IfFr is the base shear and Dr is the inter-story drift

No.2

Ashutosh Bagchi: A simplified mcthd of evaluating the seismic performance of buildings

F~

Push-over curve
(MDOF)

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
i~w~ll~.

~D

227

1. . . . . . . . . . . . . . . . . . . . . . . .

-F/ -/r~~.~.......~..:...=..=...:..=
..............
Bilinea.rapproximation of
the push-overcurve

O
O

,'

Dy

UHs:

Roof displacement

Dr

(a) Bilinear idealization of the pushover curve

Maximum inter-storey drift


(b) A drift curve representing the relation between the
maximum inter-story drift and roof displacement

Fig. 2 Derivation of an equivalent S D O F model

corresponding to a reference level of roof displacement, an


equivalent strain hardening ratio could be calculated from
Fig. 2(a). The bilinear hysteretic behaviour of the equivalent
SDOF system can be derived from the loading curve shown
in Fig. 2(a) with a diagonally symmetric unloading curve.
No strength or stiffness degradation is assumed in modeling
the hysteretic behaviour of the SDOF system.
Seismic response of the equivalent SDOF system
is obtained through a series of inelastic dynamic time
history analyses for the UHS-based ground motion
records corresponding to all three levels of seismic
hazard. The envelope of the SDOF response for a
given level of hazard is used as the peak response.
Alternatively, the peak displacement response could be
obtained using the inelastic response spectra for a given
level of seismic hazard, if available. As opposed to many
other simplified methods, such as the N2 method, the
SDOF model derived here does not utilize any algebraic
transformation based on the modal or assumed shape
vectors. Thus, SDOF response corresponding to a given
level of seismic hazard can be taken to be equal to the
peak roof-displacement of the MDOF building frame,
which is then used for predicting the maximum interstory drift in the MDOF structure using the drift curve.

Description of the building models

Two types of lateral load resisting systems are


considered here: concrete moment resisting frame, and
concrete shear wall frame. The buildings are assumed to
be located in Victoria in western Canada. The buildings
with moment resisting frames are designed to be six
and twelve-stories high, while the shear wall buildings
are designed to be twelve and twenty-stories high. The
geometric details of the buildings considered here are
shown in Figs. 3 and 4. The plan layout is common to
all the buildings. It has several 6 m bays in the N-S
direction and 3 bays in the E-W direction. The E-W bays
consist of two 9 m office bays and a central 6 rn corridor
bay. The story height is 4.85 m tbr the first story and

3.65 m for all the other stories. The yield stress, fy for
reinforcing steel, and the 28-day concrete compressive
stress, .s are assumed to be 400 MPa and 30 MPa,
respectively. The following gravity loads are used in
the design. Dead load is assumed to be 3.5 kN/m 2 on
the roof and 5.0 kN/m 2 on all other floors. Live load is
assumed to be 2.2 kN/m 2 on the roof and 2.4 kN/m 2 on
all other floors. The seismic lateral forces are obtained
using the new UHS based methodology (Eqs. 1 through
(3)). The base shear is distributed across the height of the
frame, using the procedure suggested by NBCC 1995 to
obtain the floor level forces (triangular distribution). All
transverse frames are designed to be ductile lateral load
resisting frames.
For moment resisting frame systems, a bare
frame model, and a masonry infilled model have been
considered. The masonry infill has been considered to
account for the effect of non-structural elements in the
building and to match the fundamental period of the
model with the building code specified value (normally
the period of a bare frame model is much higher than
the code specified value). The masonry infilled panels
are modelled as equivalent struts developed by StaffordSmith (1966). Clay masonry with a thickness of 100
ram, and crushing strength of 8.6 MPa has been used
in the study. The CMRF frame models are shown in
Fig. 5. For the six-story building, every alternate frame
is considered to be infilled such that the period of the
structure is close to the NBCC 1995 values. To construct
the analytical model, an infilled frame and a bare frame
are connected together using rigid links. For the twelvestory building, all frames are considered to be infilled.
For a shear wall frame system, a single wall model, and
a model consisting of the shear wall and the associated
frame have been considered. The reinforcement details
are not shown here to conserve space. The building
models as described in Table 2 are considered here to
demonstrate the effectiveness of the simplified method
of seismic performance evaluation of buildings.
A preliminary analysis of the seismic performance

228

EARTHQUAKE ENGINEERING AND ENGINEERING VIBRATION

A
L
1~

B
9m

C
6m _l_ 9m

ill

Vol.3

D
_l

~i ~

q'5
r162

~I

4.85m
2::::::

9:

---2-

. . . . . .

Elevation-six storey
t
- -

I---

: : 2 - I - : :

: : ; _ - ; :

:-

_- 9 - . -

I -

::---

:::-it

2 . = : : : :

At..

j A
tt3
',D

- C : % : % -% I :

C : C :-~lI

--- -% : % : :

--~r~5

Elevation-twelve storey

Plan
Fig. 3 Details of the CMRF buildings

9m _1~ 6m _L
~1 ~

I_

9m ~., 6m _L

9m

~r ~

9m _[
:.:.:->: : . b : . 5 > : . : . > :

r',,,4

A
.

-1

-I -

r.-,-.'.'.','.'.'.'.'.'.'.'.'.l

r.'.-...-:..'=.w.'.'.'.'.l
:':-5:.:-: : ' : ' > : ' : ' : ' : ' : ' : t

rs..'., =.. =..'.'.'.'.'.:

::::::::::::::::::::::::::::::
, . , =., . . . . . . . .
: : : : Z . : : - I : : : : : :

: : - : : : :

..:.K

r . . . , . . : . : . . . . . , 1....1

4.85m

Elevation

Plan
Fig. 4 Details of the CSWF buildings

of these buildings has been reported in Bagchi (1999),


and Humar and Bagchi (2000). A more detailed study
on the seismic performance of these buildings using
the rigorous analysis of the MDOF models has been

reported in Bagchi (2001). The performance of the


simplified method would be evaluated by comparing
the response parameters obtained using this method with
those reported in Bagchi (2001).

No.2

Ashutosh Bagchi: A simplified methd of evaluating the seismic performance of buildings

Frame A

LI .

Rigid links
~

Frame B

. . .

,'!-',',',

Infill panel
(a) CMRF-V6FB

Frame A: Frame with infill panels


Frame B: Neighbouring bare frame
(b) CMRF-V6FP

'''--"
. m m m ,

(c) CMRF-V 12FB

(d) CMRF-V12FP

Fig. 5 Bare and infill frame models of the CMRF buildings

Table 2

Description of the building models

Model

Description

CMRF-V6FB

Six story CMRF building: bare frame model

CMRF-V6FP

Six story CMRF building: infilled flame model

CMRF-VI 2FB

Twelve story CMRF building: bare frame model

CMRF-V 12FP

Twelve story CMRF building: infilled frame model

CSWF-VI2FW

Twelve story CSWF building: wall model

CSWF-VI2FF

Twelve story CSWF building: wall-frame model

CSWF-V20FW

Twenty story CSWF building: wall model

CSWF-V20FF

Twenty story CSWF building: wall-flame model

229

230

EARTHQUAKE ENGINEERING AND ENGINEERING VIBRATION

Performance of the simplified method

Vol.3

displacement is arbitrarily chosen from the push over


curves to get a good representation o f the bilinear
curve, and in the present work it is chosen to be 2% for
CMRF buildings and 3% for the shear wall buildings.
The equivalent SDOF models can be derived from this
information. The properties of the equivalent SDOF
models are given in Table 4.
The drift curves for the models are obtained from
the static pushover analysis. However, These curves are
not shown here to conserve space. Tables 5 through 7
show the response o f the SDOF models and how they

The idealized bilinear load-deformation curves


for the building models considered here are obtained
from the respective pushover curves as shown in Fig.
6. The bilinear curves are derived from these curves.
The details of the models are given in Table 3, where
H is the height, W is the weight, Fy is the yield base
shear, Dy is the lateral drift at yield, Fr is the base
shear corresponding to a reference roof-displacement,
and T is the fundamental period. The reference roof-

Table 3 Details of the building models

Model

H(m)

W(kN)

Fy/W

Dy(%H)

F/W

T (s)

CMRF-V6FB

22.8

3592

0.160

0.55

O.180

1.47

CMRF-V6FP

22.8

3592

0.310

0.45

0.330

0.77

CM RF-V 12FB

44.7

7372

0.070

0.58

0.075

2.91

CMRF-VI2FP

44.7

7372

0.I 50

0.43

0.155

1.29

CSWF-V 12FW

44.7

7372

0.072

0.40

0.095

2.05

CSWF-V 12FF

44.7

7372

0.11

0.40

0.180

1.80

CSWF-V20FW

73.9

12436

0,040

0.48

0.065

3.57

CSWF-V20FF

73.9

12436

0.062

0.48

0.095

2.70

Table 4 Details of the equivalent SDOF models

Model

H (m)

Stiffness, k (kN/m)

Mass, m (t)

CMRF-V6FB

22.8

4583.1

250.9

CMRF-V6FP

22.8

10853.0

163.0

CMRF-V 12FB

44.7

1990.4

426.9

CM RF-V 12FP
CSWF-V12FW

44.7
44.7

5829.8
2968.6

245.7
316.0

CSWF-V 12FF

44.7

4535.4

372.2

CSWF-V20FW
CSWF-V20FF

73.9
73.9

1402.3
2173.6

452.7
401.4

Table 5 Response under UHS-500 events

Model

SDOF response, D~r(% H)

Estimated drift, D~ (% h)

MDOF drift, Dsm(%h)

Ds~/D~m

CMRF-V6FB

0.40

0.57

0.60

0.95

CMRF-V6FP

0.35

0.48

0.50

0.96

CMRF-VI 2FB

0.50

0.90

1.00

0.90

CMRF-V 12FP

0.30

0.42

0.45

0.93

CSWF-V 12FW

0.70

0.94

1.10

0.86

CSWF-V12FF

0.45

0.65

0.75

0.87

CSWF-V20FW

0.45

0.62

0.60

1.03

CSWF-V20FF

0.35

0.50

0.55

0.91

Ashutosh Bagchi: A simplified methd of evaluating the seismic performance of buildings

No.2

compare to the response of the corresponding MDOF


models reported in Bagchi (2001). In these tables, the
SDOF response, D~ is expressed as a percentage of the
building height, H, the estimated inter-story drift, D~ is
expressed as a percentage of the story height, h, and the
corresponding inter-story drift using the MDOF model
is D~.
In Tables 5 through 7, the comparison is made
between D~ and D~m, representing the dynamic response
calculated using the simplified method, and the MODF
time-history analysis, respectively.
It is observed from the results presented here that
the simplified method gives a good estimate of the

maximum story drift corresponding to UHS-500 and


UHS-1000. However, in some cases with UHS-2500, the
estimated story drift deviates as much as 11% from the
MDOF response. This implies that when displacement
demand is very high and the structure is well in the
yield zone, the response by simplified method tends to
be less accurate. In most of the cases, inter-story drift
calculated using the simplified method is lower than the
actual values.
A summary of the single degree of freedom response,
D~ under various levels of seismic hazard is compiled
from Tables 5 through 7, and the values are shown on the
idealized bilinear load-displacement curves in Fig. 6 as

0.5

0.20
. . . . . CMRF-VI2FB
......... CMRF-VI2FP

. . . . . CMRF-V6FB
......... CMRF-V6FP
0.4

0.16
( 0 . 3 1 ~

(0.15,~,~:$
0.4~."
_m"" ..........

0.08

r (0.07, 0.~0)~,..~.-....-..:...'....'."~". . . . . .

",,2
r

0.2
.

"'-.-~ ".,..

0.12

..........~. . . . . . . . . . . . .

0.3

~E
.

'

"-,,

3
r

231

0.1

0.04

!,
9 ]

0.0
0.0

0.4

0.8

1.2

1.6

0.00

2.0

0.0

0.4

Roof displacement (% H)
(a) CMRF six story

0.8

1.2

1.6

2.0

Roof displacement (% H)
(b) CMRF twelve story

0.18

0.11
.

' cswi-vl2FW

......... CSWF-V 12FF

0.16

........ 7

'

. . . . . CSWF-V20FW
......... CSWF-V20FF

0.10
0.09

0.14

(0. 1 , 0 . 4 0 ) . , f ~ . , /

0.08
(0.062, 0.48) 3 o . , ~ . ~

0.12

"~:~''''w~'~--~'''

0.07

0.10

0.06
0.05

0.08

"2 ( J

0.2

0.04

0.06

9" (0.072, 0.40)

0.03

0.04

(0.04, 0.40)

0.02

0.02

0.01

0,00
0.0

0.5

1.0

1.5

2.0
Roof displacement (% H)
(c) CSWF twelve story

0.00

2.5

3.0

0.0

0.5

1.0

1.5

Fig. 6 Pushover curves of the building models

2.0
Roof displacement (% H)
(d) CSWF twenty story

2.5

3.0

232

EARTHQUAKEENGINEERINGAND ENGINEERINGVIBRATION

Vol.3

Table 6 Response under UHS-1000 events

Model

SDOF response, Dsr(%H) Estimateddrift, Dss(%h)

MDOF drift, D~m(%h)

DJD~m

CMRF-V6FB

0.45

0.65

0.70

0.93

CMRF-V6FP
CMRF-V12FB
CMRF-V12FP
CSWF-VI2FW
CSWF-V12FF
CSWF-V20FW
CSWF-V20FF

0.42
0.70
0.35
0.80
0.55
0.55
0.40

0.62
1.25
0.50
1.10
0.75
0.70
0.55

0.60
1.20
0.55
1.20
0.80
0.70
0.60

1.03
1.04
0.91
0.92
0.94
1.00
0.92

Table 7 Response under UHS-2500 events

Model

SDOF response, Dsr (%H)

Estimateddrift, Ds~(%h)

MDOFdrift, Dsm(%h)

DssrDsm

CMRF-V6FB
CMRF-V6FP

1.10
0.80

1.95
1.52

2.20
1.70

0.89
0.89

CMRF-V12FB

1.32

2.50

2.50

1.00

CMRF-V12FP

0.75

1.65

1.50

1.10

CSWF-V12FW

1.81

2.60

2.80

0.93

CSWF-VI2FF

1.33

1.82

2.00

0.91

CSWF-V20FW

1.55

2.20

2.40

0.92

CSWF-V20FF

1.05

1.33

1.50

0.89

points 1, 2 and 3 corresponding to UHS-500, UHS-1000


and UHS-2500, respectively. It can be noted from Fig.6,
that the values of Dsr for UHS-500 and UHS-1000 are
well below the reference roof-displacement and some of
the values are even below the yield displacement. This is
indicative of negligible or minor damage as determined
from the MDOF analysis. Although the SDOF
responses due to UHS-2500 events are beyond the yield
displacements, they are still lower than the reference
roof displacements. One may argue that a smaller value
for the reference displacement could be taken to obtain
the idealized bilinear curves. As indicated earlier in the
paper that the reference roof displacement is chosen
arbitrarily to match the push-over curve as closely as
possible. Thus it is possible to choose a different value
for the reference roof displacement which may alter
the strain hardening segment of the idealized bilinear
curve, and the response may change to some extent. A
sensitivity analysis could be carried out to quantify the
extent of change in the SDOF response due to different
values of the reference roof displacement. To the
author's opinion, the reference roof displacement should
be high enough to capture the entire pushover curve
and the post-yield behaviour of the structure before
collapse, which will be useful for estimating the reserve
deformation capacity of the structure beyond the UHS2500 level of seismic hazard at which it is expected to
have life-safety performance.

5.1

A c c o u n t i n g for the effects o f higher m o d e s

The distribution of lateral loads used in the pushover


analysis is based on the NBCC 1995 guidelines
(triangular distribution). Pushover analysis using the
NBCC 1995 distribution of lateral loads may fail to
capture the effects of higher modes and the pushover
results may not provide an accurate prediction of
the dynamic behaviour of the structure. Humar and
Rahgozar (2000), and Humar and Mahgoub (2000) have
shown that the effect of higher modes on the base-shear
depends on structural types, modal periods and spectral
shapes. The higher mode effect is found to be more
pronounced for flexural wall systems than for shear
flame structures.
To evaluate the effect of higher modes in the present
analysis, the following shear wall models are considered:
CSWF-V12FW and CSWF-V20FW (refer Table 2). For
a building model, the contribution of each mode to the
base shear is computed using the method suggested
in Humar and Rahgozar (2000). Components of the
base shear corresponding to the individual modes are
calculated based on the modal contribution ratios. Each
component is distributed along the height of the building
according to the corresponding modal shape. The modal
distributions of lateral loads are then combined together
using square root of sum squares (SRSS) technique and
scaled to the design base shear. Instead of using the

No.2

Ashutosh Bagchi: A simplified methd of evaluating the seismic performance of buildings

GSC spectrum, S (UHS-2500)

120
I

233

120

.......... Design spectrum, S(T)

100

100

80

80

60

60

e~

e~
t~

t~~
40

40

l ,.o

0.0

ilo:ol ....... ,:

20 l_ i . . . . . . . . .
,0.45

_ x ........

20

r ......:

: 2.05................................
I

II

2.0
3.0
4.0
5.0
Period (s)
(a) Spectral ordinates for CSWF-V12FWmodel
1.0

12

......

0.0

2.0
3.0
4.0
5.0
Period (s)
(b) Spectral ordinates for CSWF-V20FWmodel
1.0

20

%~%~*%.%.%

10

P/>,

>

o
4

/I~//el!! ;)
I~// /lll
~

15

10
NBCC d i s t
[ iiI itl:l~
1st mode dist .......
ml . . . .

NBCCdist
1st mode dist ........
ml . . . . ~ -:r
m2

m2 ---- ,,,
f7tr

ScaledSRSS - - - . , ' j _ .
m l,m2 are modal
contributions
0

-100

Scaled SRSS ~ - ~L./r


m 1,m2 are modal
contributions

0
-50

0
50
100
Lateral force (kN)
(c) Lateral load distribution in CSWF-V12FWmodel

-100

-50

0
50
100
Lateral force (kN)
(d) Lateral load distribution in CSWF-V20FWmodel

Fig. 7 Accounting for higher modes

triangular distribution, this SRSS modal distribution


of lateral forces could be used in the pushover analysis.
The proposed simplified method is then used to evaluate
the seismic response. Only the first two modes are
considered in the present study.
Figures 7 (a) and (b) show the design spectral
acceleration curve (based on UHS-2500) for Victoria.
The first two modal periods for the CSWF-V12FW
model are 2.05 s and 0.45 s. Corresponding values of
effective modal weight are 0.68 W and 0.18 W, W being
the weight of the structure. The spectral acceleration
corresponding to each mode is obtained from Fig. 7 (a).
The values of spectral acceleration corresponding to the

first mode and the second mode are 0.19g and 0.88g,
respectively (g is the acceleration due to gravity). By
multiplying the spectral acceleration with the modal
weight, the modal base shear is obtained. These values
are tabulated in Table 8. The base shear contribution of
the second mode is also quite considerable in the twentystory shear wall building model, CSWF-V20FW.
Figures 7 (c) and (d) show the distribution of
modal contributions to lateral loads. For the purpose
of comparison, the NBCC 1995 distribution and the
distribution of lateral loads according to the first mode
only are also shown. It is observed that the NBCC
distribution follows the first mode only distribution more

234

EARTHQUAKE ENGINEERING AND ENGINEERING VIBRATION

Vol.3

Fig. 8 Effect of higher modes

Table 8 Higher mode effect on the base-shear

Model

Period (s)

Effective modal
weight

Spectral
acceleration

Modal base-shear
factor

Ratio of modal
base-shear

CSWF-V 12FW

2.05

0.68 W

0.19g

O.129

0.45

CSWF-V2OFW

0.45
3.57

0. l 8 W
0.64 W

0.88g
0. I 0g

0.158
0.134

0.55
0.47

0.58

0.20W

0.75~

0.150

0.53

Table 9 Estimated story-drift by the simplified method


Model

Method

UHS-500

UHS- 1000

UHS-2500

CSWF-VI2FW

NBCC-1995 dist.

0.94

1.10

2.60

SRSS-modal dist.

1.05

1.16

2.76

MDOF response

1.10

1.20

2.80

NBCC-1995 dist.

0.62

0.70

1.33

SRSS-modal dist.

0.63

0.71

1.46

MDOF response

0.60

0.70

1.50

CSWF-V20FW

No.2

Ashutosh Bagchi:A simplifiedmethd of evaluatingthe seismicperformanceof buildings

closely than the SRSS modal distribution. To account for


the higher order modes, the NBCC distribution assigns
higher magnitude of lateral force at the top story.
Pushover curves using the SRSS modal distribution
of lateral forces, and the corresponding drift curves are
shown in Fig. 8. It is observed from Fig. 8 that there
is some difference between the pushover response
of a building using the NBCC distribution and the
SRSS modal distribution when first two modes are
considered.
Table 9 shows the values of the maximum inter-story
drift estimated by the proposed simplified method using
the NBCC and SRSS modal distribution of the lateral
forces. The results presented in Table 9 show some
improvement in estimated response. The contribution
of the higher modes could be significant as observed by
Chopra and Goel (2001). The estimated response based
on the simplified method presented here still provides a
valuable guidance on the seismic performance. It should
be noted that the pushover analysis here has been carried
out using the combined distribution of lateral loads, not
individual modal distribution, and the reference roof
displacement is assumed to be 2% or 3% as in the case
of NBCC distribution.

6 Discussion and conclusions


From the results presented above, it is observed that
the accuracy of the simplified method in predicting the
maximum inter-story drift is reasonably good in some
cases. Given the nature of approximation involved in the
proposed method, its performance is acceptable. From the
examples presented by Fajfar and Gaspersic (1997), the
N2 method seems to achieve a similar level of accuracy.
Based on the estimated maximum values of interstory drift obtained using the proposed simplified
method, the following observations could be made on
the performance of the buildings studied here.
9 The six-story CMRF building achieves a
performance level of operational under UHS-500 and
UHS-1000, and life-safe under UHS-2500 when the
bare frame model is considered. When the infill frame
model is considered, its performance levels are fully
operational under UHS-500, operational under UHS1000 and life-safe under UHS-2500.
9 The twelve-story CMRF building achieves a
performance level of operational under UHS-500 and
UHS-1000, and near collapse under UHS-2500 when
the bare frame model is considered. With infill panels,
its performance levels are fully operational under UHS500, operational under UHS-1000 and life-safe under
UHS-2500.
9 The twelve-story CSWF building achieves a
performance level of operational under both UHS-500
and UHS-1000 and life-safe under UHS-2500 when the
wall model is considered. With the wall-frame model,
the pertbrmance level is operational under both UHS500 and UHS-1000 and life-safe under UHS-2500.

235

9 The twenty-story CSWF building in Victoria


achieves a performance level of operational under both
UHS-500 and UHS-1000 and life-safe under UHS-2500
for both wall and wall-frame models.
In most cases, the qualitative performance levels
mentioned above are consistent with the performance
levels evaluated through a rigorous study reported in
Bagchi (2001 ).
Simplified analysis as described above could form an
important intermediate step in the earthquake resistant
design of buildings. The designer can quickly estimate
the possible performance level of the structure and
modify the design based on the results. Once the design
is finalized, a detailed analysis could be performed to
evaluate the seismic performance of the structure.
The simplified method presented here has some
limitations. For cases where the effect of higher order
modes is significant, this technique may not very
effective in estimating the maximum inter-story drift.
Pushover analysis using the SRSS modal distribution of
the lateral loads improves the solution to some extent.
However, the method could be used for an easy and
approximate estimate of the seismic pertbrmance of a
building for given level of hazard.
The proposed method may be criticized for being
overly simple and approximate. Designers must be aware
of its limitations while using it for practical design. The
simplified method presented here would, however, be of
assistance to designers in carrying out a perfbrmancebased design. The estimated response using the method
described here would allow the designers to get an
approximate measure of the qualitative performance
of the structure and could help them in revising the
design through a number of iterations in order to achieve
the desired performance characteristics, strength and
deformation capacity. A detailed MDOF analysis must
be carried out in the final stage for estimating the seismic
response more accurately.

Acknowledgement
The work presented here forms a part of the author's
Doctoral thesis. The author would like to thank his
thesis supervisor Professor Jag Humar at Carleton
University, Ottawa, Canada, for his guidance, help and
encouragement.

References
Adams J, Weichert D, Halchuk S and Basham P (1999),
"Seismic Hazard Maps of Canada," Geological Survey
of Canada, Open File 3724.
Atkinson GM and Beresnev IA (1998), "Compatible
Ground Motion Time-histories for New National Hazard
Maps," Canadian Journal of Civil Engineering, 25:305318.
Bagchi A (1999), "Seismic Performance of Buildings
Designed to National Building Code of Canada," 8'h

236

EARTHQUAKEENGINEERINGAND ENGINEERINGVIBRATION

Canadian Conference on Earthquake Engineering,


Vancouver, B.C., Canada, pp.487-494.
BagchiA (2001), "Evaluation of the Seismic Performance
of Reinforced Concrete Buildings," Ph.D. Thesis,
Department of Civil and Environmental Engineering,
Carleton University, Ottawa, Ontario, Canada.
Building Seismic Safety Council (BSSC) (1997),
"Recommended Provisions for Seismic Regulations for
New Buildings and Commentary," National Earthquake
Hazards Reduction Program (NEHRP), Federal
Emergency Management Agency, Washington, D.C.
Chopra A and Goel R (2001), "A Modal Pushover
Analysis Procedure to Estimate Seismic Demands for
Buildings: Theory and Preliminary Evaluation", PEER
Report #2001/03, Pacific Earthquake Engineering
Research Centre, Berkeley, USA.
Faj far P and Gaspersic P (1996), "The N2 Method for the
Seismic Damage Analysis of RC Buildings," Earthquake
Engineering and Structural Dynamics, 25: 31-46.
Ghobarah A, Aly NM, and El-Attar M (1997),
"Performance Level Criteria Evaluation," Seismic
Design Methodologiesfor the Next Generation of Codes,
edited by Fajfar and Krawinkler, pp.207-215.
Humar JL (2000), "Future Directions of Seismic Design
Provisions," Short Course on Earthquake Resistant
Design of Buildings, Carleton University, Ottawa,
Canada.
Humar JL and Bagchi A (2000), "Seismic Level of
Protection in Shear Wall Frame Structures Designed
According to National Building Code of Canada,"
28'h Annual Conference of Canadian Society for Civil

Vol.3

Engineering, London, Ontario, Canada, pp. 161-168.


Humar JL and Mahgoub MA (2000), "Accounting
for Higher Modes in UHS-based Design," Discussion
paper, Canadian National Committee on Earthquake
Engineering (CANCEE).
Humar JL and Rahgozar MA (2000), "Application
of Uniform Hazard Spectra in Seismic Design of
Multi-story Buildings," Canadian Journal of Civil
Engineering, 27: 563-580.
Mazzolani FM and Piluso V (1997), "A Simple Approach
for Evaluating Performance Levels of Moment-resisting
Steel Frames," Seismic Design Methodologies Jbr
the Next Generation of Codes, edited by Fajfar and
Krawinkler, pp.241-251.
National Research Council of Canada (NRC) (1995),
National Building Code of Canada (NBCC), Ottawa,
Ontario.
Park YJ and Ang AHS (1985), "Mechanistic Seismic
Damage Model for Reinforced Concrete," Journal of
Structural Engineering, ASCE, 111: 722-739.
Prakash V, Powell GH and Campbell S (1993), "DRAIN2DX Base Program Description and User Guide: Version
1.10," Report No. UCB/SEMM-93/17, Deparment of
Civil Engineering, University of California, Berkeley.
Stafford-Smith B (1966), "Behaviour of Square Infilled
Frames," Journal of the Structural Engineering Division,
ASCE, 91, (ST1): 381-403.
Vision 2000 Committee (1995), Performance Based
Seismic Engineering of Buildings, Structural Engineers
Association of California (SEAOC), Sacramento,
California.

Potrebbero piacerti anche