I:tl.t.

'tlH't'ti/tL'

'~ltttiltttc'~.

Wtl]

_~(). ~11 7. p p t~ll,N lltl. ll)~t,X

p u*~ l!l,t"+rcl " , , , c n , c I td PIIllLCd Ill (;tt',ll |'~rll;tnl OJ-~+l l)_~g('~.,'tJ,R ~ItJ{H) - 11(I<)

,\11 II~hl~ Icxcr~,ct[

I'1 SI:VI/iR

P l h S0 ! 4 ! - 0 2 9 6 1 9 7 ) 0 0 0 5 5 - 2

Analyses of seismic performance of a code designed reinforced concrete building

J a v e e d A. Munshi and Satyendra K. Ghosh
I~m'llaml Cement A.v~o(ia/ion. 5420 Oh/ Orchard R+*ad. Skokie. I1.. l ,'SA (Received November 1996: rcvi.~ed vc/'sion o('cepled I"e/~ru(/r~" 1997)

This paper studies the inelastic seismic performance of a 12-story reinforced concrete (RC) building. The building utilizes a structural system with moment-resisting frames in the longitudinal direction and a dual structural system consisting of coupled shearwalls and moment-resisting frames in the transverse direction. The frame elements, the shearwalls and the coupling beams are sized and detailed on the basis of the 1994 edition of the Uniform Building Code for seismic Zone 4 (regions of high seismicity). The global and local inelastic behavior of the building in the two orthogonal directions is studied under several earthquake ground motions. Nonlinear concrete behavior, including stiffness degradation and strength loss caused by cracking, crushing of concrete and yielding of steel, is simulated by using the fiber beam-column element of the DRAIN-2D program. Pushover analysis is used to determine the global ductility of the structure. The study indicates that the design strength may be inadequate for some critical earthquakes which tend to induce biased response in the structure. The global response of the building is not much altered when the effect of vertical accelerations is also included. Weak coupling between walls induces large ductility demands in them, which can be directly reduced by increasing the wall strength. Well designed coupled walls are more efficient and economical than isolated walls or weakly coupled walls. Optimum values of beam and wall stiffness and strength can be chosen to minimize ductility demands on the walls of a coupled wall system. The coupling between walls appears to become ineffective, due to the simultaneous yielding of coupling beams and walls when a structure is subjected to an early and large displacement pulse, such as that produced by the Northridge earthquake in the near-field. Although the increasing of wall strength decreases the ductility demand, the real challenge, in this case, is to reduce the peak displacement, which is not particularly sensitive to the strength values. 1998 Elsevier Science Ltd.

Keywords: reinforced concrete building, seismic analysis, seismic design, seismic performance, coupled wall, code design, moment resisting frame, dual shearwall-frame system

1.

Introduction

The Uniform Building Code (UBC) provides analysis, dcsign and detailing requirements for buildings subject to earthquake ground motions. The intent of the document is to minimize the life safety hazard, avoid catastrophic failure

of structures and, in the case of essential facilities, to maintain continued functionality after a maior earthquake. In seeking to avoid catastrophic structural failure, the code recognizes the ability of carefully detailed concrete ~,truc-

rural components to dissipate significant amounts of earthquake energy through hysteresis. The design philostq3hy.

608

Analysis of seismic performance of a code designed RC building: J. A. Munshi and S. K. Ghosh

thus, inherently allows damage such as cracking, crushing of concrete and yielding of steel at preferred locations, to avoid distress at structurally critical locations. Though the intent of the code is simple, the actual behavior of a structure and its members could deviate substantially from the intended behavior because of the complexity of the actual ground motion, the complexity of the structure, the manner and extent of strength deterioration and the stiffness degradation of members and the redistribution of forces in the inelastic structure. These aspects need to be studied by considering the inelastic behavior of concrete components under probable ground motions, to assess the adequacy of code provisions and identify possible future improvements. The performance of coupled walls is critical to building safety. Both weak and very smmg coupling between walls can induce large ductility demands in the walls. Optimum coupling between walls (to minimize wall ductility) depends upon the beam-to-wall stiffness and strength ratios, and the yield strength of the walls. These parameters should be considered in order to design an efficient and economical coupled wall system. An additional crucial aspect of ensuring structural safety is taking into consideration all possible ground motions of different characteristics at the site of the structure. The Northridge earthquake has brought to light new problems posed by near-field ground motions capable of producing a large drift pulse in a structure at early stages of ground shaking. This may have significant consequences affecting the structural safety, as the energy dissipation capability of the structural system is not fully mobilized. Note that the code design philosophy assumes normal vibration of the structure with many cycles of energy dissipation. For moment resisting frames, a large deformation pulse may result in excessive story drifts and partial collapse mechanisms. For coupled wall-frame systems, a large deformation pulse may result in the yielding of all the cot, piing beams and the walls, simultaneously, and render the coupling ineffective after such pulse. This paper investigates some of these issues and highlights the important parameters that should be considered in the design process.

609

ties. Soil type $2 with a site coefficient of 1.2 was assumed for design of the building as per UBC. Analyses are carried out by using both the gross and effective (cracked) section properties of the concrete elements in order to get the upper bound on both the displacements and ductility demand. For the moment resisting frame, effective stiffness of the frame is simulated by assuming one-half the ~ross section properties for the beams and gross section properties for the columns. For the frame-wall system, the effective stiffness is based on onehalf the gross section properties for the shear walls and the beams and gross section properties for the columns. Note that the use of gross section properties could overestimate the ductility demands, but may underestimate the peak deformations of the structure. A comparison of displacements and ductilities of the structure using both gross and effective section properties is also discussed. The potential hinges at beam ends are idealized by using the point fiber hinges of the DRAIN-21)X program. The effects of stiffness degradation, strength loss and pinching of concrete are included in the analyses. The gravity loads due to the tributary dead and live load are input for columns and shearwalls at each floor level in order to simulate the axial load effect on their flexural capacity. The analysis assumes rigid floor diaphragms, with each node having three degrees-of-freedom.

3.

Earthquake ground motions

2.

Example building

Figure I shows the plan and elevation of a 12-story building used for this investigation. The building utilizes a structural system with moment-resisting frames in the longitudinal direction and a dual structural system consisting of shearwalls and moment-resisting frames in the transverse direction. The frame elements, the shearwalls and the coupling beams are sized and detailed on the basis of the 1994 UBC Zone 4 (regions of high seismicity) requirements. All beams are sized 2024 in (508x610 ram) and all columns 24x24 in (610610 ram). The first- and second-story columns are of 5 ksi (34.48 Mpa) concrete and the remaining columns and all beams are of 4 ksi (27.58 Mpa) concrete. The analyses of the building are carried out separately in the longitudinal and the transverse directions using the DRAIN-2DX program -~. The fundamental period of the elastic building, initially determined by code prescribed method, was verified by analyses as 1.8 s in the longitudinal direction and 0.83 s in the transverse direction. The building is assumed to have 5% damping in its first three deformation modes, The strain-hardening stiffness is assumed to be 5% of the elastic stiffness. The design of the building is based on period computations using gross section proper-

The ground motion records used for the analyyses include the El Centro (1940) NS component, an artificial earthquake whose spectrum characteristics are compatible with the SEAOC (1990) design spectrum, the Hachinohe earthquake (1968), the Parkfield earthquake {1966) and the Northridge earthquake (1994) as recorded at Sylmar county and Newhall stations (NS and EW components of each record). The Parkfield earthquake was selected to represent a harmonic type motion critical for long-period structures, and the Northridge earthquake records were selected to study the near-field ground excitation el'fecr'. The eight records selected were scaled up Io a spectral intensity of 1.5 times that of the El Centro (1940) NS component and 1.5 times that of the Hachinohe earthquake. Note that these two earthquakes are also being used for level-2 earthquake design in Japan. The spectral characteristics of the scaled earthquakes for 5% viscous damping are shown in Figure 2.

4.

Performance of moment resisting frame

4. I. Inelastic dynamic respon.~e The moment resisting frame in the longitudinal direction was analyzed for the selected ground motions (Figure 2). The results of analyses of the building, assuming gross section properties, under four earthquakes El Centro, Hachinohe, Northridge at Sylmar county (EW), and Parkfield are shown here. Figure 3a shows the roof deformation histories of the building. The peak roof displacements of the building under these four earthquake components are 10, 15, 16 and 10 in (254, 381,406 and 254 ram), respectively. The building is likely to develop a biased motion, particularly severe under the Parkfield earthquake. Figure 3b shows that the Parkfield and Northridge (Sylmar EW) earthquakes also produce critical and biased drift concentrations. The

610

Analysis o f seismic performance o f a code designed RC building: J. A. Munshi and S. K. Ghosh

l e 2 - J . '3~ =====================

::

;:

::

II II II
; .. II .111 H

IF--q[ Jl I[ I[-----71
1I )1 II II II---ll

Jl II II
)1 Jl II _11 II 11 II

II II }I
)1 II II H }1 II II

o!:::::::::!::::::::.... ::::::i:::::::::
(a) Plan
t 3
* ~

:::::::::::::::::::::

:
.. ,
7

ii
il

I -11
~1

II II 1

I[---ql

[i

i[

i[

![

i[

[[

II ..... II---ll

..........

i....... :: ...... :4: ........

,
:"

U tl II U II
(b) Elevation

,,:
H

::
',t

!~

(c) Plan of New Configuration

Figure 1 Example building
80"
.... 8 0

~0

SCALED NORTHRIDGE
.

:.,

.." "--.

...

o e 60

t:'.

?,.::...",

....

~o
>

I//':'v' ':
~.,:

- - .........

1.5xEl C e m r o ( 1 9 4 0 ) N S O

I .OxA~ificial
15xHachinohe

I
I

~20,

/~'~ "
g . '

....

o,,P=~

q
-

......... 13~owh~lEw
....
0 8,.N~h,, NS

I
I
"b

/"

PERIOD (Sec.)

). . . 0o'5. . . 1 . 0 . . .1.5 . . . 2 . 0. . . 2.. 5. . . . .

PERIOD (Sec.)

Figure2 Spectral characteristics of earthquakes (5% damping ratio)

maximum interstory drifts reach more than 2% under the Parkfield earthquake. Analyses were also carded out using the effective section properties of the building members. The gross section properties for the columns and one-half of the gross section properties for the beams were assumed. The peak displacemerits, at roof level,of the building, under the four selected earthquakes El Centro, Hachinohe, Parkfield and Northridge (Sylmar EW) are 9.5, 11.5, 17.6 and 12 in (241, 292, 447 and 305 mm), respectively. The above results indicate that the Parkfield and the Nonhridge earthquakes have the potential of causing more critical damage than the El Centro and Hachinohe earthquakes commonly used for similar analyses. The analyses also indicated that stiffness degradation, strength loss and pinching of well designed and detailed concrete components do not critically change the maximum global response of the building. 381, 406 and 254 mm), global displacement ductilities of 2.5, 3.75, 4 and 2.5 are obtained under the El Centro. Ilachinohe, Parktield and Northridge (Sylmar EW) earthquake motions, respectively. Considering a yield displacement of about 7 in ( 178 mm) for the frame with effective section properties and the peak displacements indicated earlier tor this frame, the corresponding ductility ratios resulting under the tour earthquakes are 1.4, 1.9, 2.5 and 1.7 , respectively, it may bc noted that the ductility values are smaller when effective stiffness properties rather than gross section properties are used. The maximum ductility demands on the moment resisting frame seem to be well within the limits of ductility values for which the building is designed as per the code.

5.

Performance of dual shear wail-frame system

4.2. Global ductility by pushover analysis The global yield displacement of the moment resisting frame was obtained by pushover analysis. Considering the global yield displacement of 4 in ( 102 mm) (see Figure 3c) for the frame with gross section properties, and the maximum roof displacements of 10, 15, 16 and 10 in (254,

5. I. Inelastic dynamic response The dual shear wall-moment resisting frame system in the transverse direction (Figure 1) was analyzed for the selected ground motion records of Figure 2. The results for El Centro, Hachinohe, Northridge at Sylmar county (EW) and Northridge at Newhall station (NS) ground motions are shown here.

Analysis of seismic performance of a code designed RC building: J. A. Munshi and S. K. Ghosh

El C e n t r o

611

" ~ 20.
-!
t i

.......... H a c h i n o h e ...... Parkfield

....
tZ
.

Northrid8e (Sy|miu" EW)

z
L~

20.
: ~.~',, ': : ~ ",1 ".,,,.k-,:7"..,_," .',,-" "'-.:"-:.
.

o
n

,J

(D "~--10
G. U3

'ff~,', ~ ' . : ',,Tv'~,.'~V' "':, ';",'

. .

""

..,

",

(a) Roof Displacement

$
El C e n t r o .......... H a c h i n o h e ...... Northridge (Sylmar EW)

' ,'

",-'

,.'

-"

~, - 2 0

TIME (See.)

10

15

20

(a) Roof Displacement

" "

-20

,~ . . . . 1'0 . . . . TIME ( S e e . )

1'5

20

....

Nonhndge (Newhall NS)

El C e n t r o .....
Hachinohe

..... ......

.....

El C e n f r o Hachinohe Pmrkfleld Nonhndge (Sylmar EW)

...... ......

Northndg (Sylmar EW) Nonhndge 0Newhan NS)

t

,i f"~ ;

",,
i

,~"
" 0.61 0.62

-oJo2 -o2o~ o.6o o.61

(b) D r i f t R a t i o

0.62

-0:02

"-0201 ' ~ I I 0

(b) Drift Ratio

la.

,7-.
w all 0

S
. . . .

w t; c o g PL~D WAM'-~'~A[~ ~" I l/I

(c) P u s h o v e r A n a l y s i s
DISPLACEMENT (In.~

g
0

(c) Pushover Analysis

DISPLACEMENT Inelastic response of wall-frame system

Figure 3

Inelastic response

of moment

resisting

frame

Figure 4

Figures 4a, b show the displacement and drift responses of the building using gross section properties in the transverse direction under the four selected earthquakes. The peak roof displacements of the building under the four earthquakes are 6, 7. 10 and 12 in (152, 178, 254 and 305 mm), respectively. The Northridge earthquake components induce a severe deformation pulse resulting in biased motion of the building. This biased motion results due to excessive ductility demand on the walls (see the following section). The maximum displacement reaches about 12 in (305 mm), while the maximum drift ratio is about 0.7% under the Northridge (Newhall NS) earthquake component. Figure 4a,b, however, also show that shearwalls reduce the displacement and drift envelopes of the building in the transverse direction. For analyses using the effective stiffness values for concrete components, gross section properties of columns, and one-half gross section properties of shear walls and beams is assumed. The displacements of 6.2, 1 !.5, 14 and 9.5 in ( 157, 292, 356 and 241 mm) result under El Centro, Hachinohe, Northridge (Sylmar EW) and Northridge (Newhall NS) earthquake motions, respectively. 5.2. Global ductility by pushover analysis
The global displacement ductility of the building in the transverse direction is obtained by using the yield displace-

ment determined through pushover analysis. Figure 4c shows that the yield displacement of the wall-frame structure is very similar to that of the wall system alone, due to the large stiffnesses of the walls. The global yield displacement of the building, assuming gross section properties, is about 2in (51 mm) (see Figure 4c). Considering the maximum displacements of 6, 7, 10 and 12 in (152, 178, 254 and 305mm) under the El Centro, Hachinohe, Northridge (Sylmar EW) and Northridge (Newhall NS) motions, respectively, global ductilities of 3. 3.5, 5 and 6 result under the four selected earthquake components. Considering yield displacement of about 3.6 in (91 mm) for the wall-frame system with effective section properties and the peak displacements shown above, the resulting ductility demands due to the tour selected earthquake motions are 1.7, 3.2, 3.9 and 2.6, respectively. The ductility values are lower when effective stiffness properties rather than gross section properties are used for the concrete members. It may be noted that these ductility demands result despite an overstrength of about 2.1 for the walls with 0.8% reinforcement under the design axial load. The overstrength factor is defined as the ratio of actual wall strength to the strength required by the code-prescribed seismic forces. Walls usually have substantial overstrength, since wall location and proportions are more often governed by functional rather than structural considerations.

612

Analysis of seismic performance of a code designed RC building: J. A. Munshi and S. K. Ghosh

(a) Roof Displacement 1 !c

,~
~-~o~

" ::
.......
......... ~. . . . . . . . . .

/
ORG. WALL-STRENGTH 1.25xWALL-STRENGTH
1'2 . . . . . . . /'6 . . . . . . . .

t ]

~ ........

TIldE (See.)

~o

-o;6i :i~:6i ~ b i~,6Y"6.~2

(b) Drift Ratio

Figure 5 Inelastic

response of strengthened wall-frame system

5.3. I)m'tilitv of weakly cowHed walls

The large ductility demands on the walls is attributed to weak coupling between the walls due to relatively long coupling beams. [n such situations, walls behave primarily as isolated cantilever walls. The ductility demands on such walls can be directly reduced by increasing their strcngth. This in shown in Figure 5. which compares the response history and drift envelope of the building with the original walls (0.8% steel) with those of the same building with strengthened walls I I.5c~ steel) under the Northridge (Ncwhall NS) motion which produced the critical response. The strengthened wall (overstrength = 2.85) reduces the maximum displacement from 12 to 8 in (203 ram) and the corresponding maximt, m global displacement ductility from 6 to 3.2. This substantial reduction in the displacement ductility demand of walls helps clirninatc the biased response of the building (l'Vgure 5).

displacement response of the two Mruclural ~,yslellls ~;.is not significantly altered when the effect of vertical grokinct nlotion wa,~ included. The nlaxinlunl defornlalions ~how a slight increase wiih vertical ground inolions, especial ). for the nlonlenl resisting fralllO, a study of the hinge fOllllalion patterns revealed thai vertical accelerations induced a slightly different hinge fornlalion pallerll and hinge rotation magnitudes. The response (if" the franlc-wall s).~,lClll did not show nluch sensitivity Io the vertical acccleraliollS in this case.

7.

Performance of coupled wall system

6.

Effect of vertical accelerations

The effect of vertical accelerations tin the global displacement of the moment resisting frame and the shear wallframe system was also investigated. The moment resisting frame and the frame-wall system were analyzed for a combined horizontal and vertical component of the Northridge earthquake recorded at Syhnar County station (EW component ). Note that this earthquake component produced critical responses in the two systems. The effect of floor vibration was neglected. Figure 6a, h shows that the global
Horz. Accel. Only ----~-- Horz. & Vert. Accei.
Z lad

The wall system in the transverse direction of the e~,ample building is redesigned as a cot,pied wall system by reducin~ the wall section and coupling beam length (Figure /c)aud increasing the coupling beam depth. Coupled walls are more efficient and economical for lateral load re~,istance, as compared to isolated walls. Short coupling beams between walls substantially increase the stiffness of the system. Larger strengths and ductility demands of the coupling beams increase the energy dissipation capacity of the system. Shorler and stronger coupling beams also increase the degree of coupling between the walls. The increase in the degree of coupling between walls, however, also increases the magnitude of axial forces (compression and tension) in the walls. This complicates the post-yield behavior and the ductility demands on the walls. Thus, increase in the degree of coupling beyond a certain limit is not efficient and would only increase the ductility demand on the wall,,. An efficient coupled wall system will have optimal coupling between walls m terms of bean>to-wall stiffne.,,,, and strength ratios, resulting in minimum wall ductility 4 ".

M
....

/%

..~

':
ill

O -5:

W t/ V V
~i . . . . ,b " "

.kd.- k/..
ZO

Design of COulUed wall system A coupled wall system (Figure It). having its stiffness and 7. I.
period (with gross section properties) similar to those of the original wall-frame system, is designed with a beamto-wall stiffness ratio of 0.04. This optimum stiffiiess ratio is obtained by reducing the wall section and increasing the beam to wall stiffness ratio. The resulting coupling arm is 7 ft (2.13 m) long and 32 in (813 ram) deep. The coupling beams at all floor levels are assumed to have identical strengths. The tributary gravity load of the walls is assumed to act as point loads on the walls in order to simulate the effect of axial force on the moment-axial force strength interaction diagram of the walls. The resulting wall section has only 34% of the gross moment of inertia of the original wall section. The design moments from lateral load analysis are substantially lower for the coupled walls (Figure 7a). The shear forces (I.igure 7h) however, remain similar, whereas the maximum total

F,-~

TIME '~Sac) ,5 (a) Moment Resisting Frame

Horz. Accel. Only

"-"10

-~

~. s

...... Horz. & Ver~. Accel.

%-s
12_

TIME: (See)

(b) Wall-Frame System

Figure6 Global displacement response of the building for combined horizontal and vertical ground motions

Analysis of seismic performance of a code designed RC building: J. A. Munshi and S. K. Ghosh

...... .....
10 10,

613

Design-Original Design-Coupled Inducd-Orlgtnal Inducd-Coupled

WaR Wall Wall Wall

104 ", ""

(a)

.i

,1
4"

",,
"',

(c)

""

"'-,.,
o. :gO . . . .
MOMENT ( I n . - k l p )

i
2.m~.+ t~.~
SHEAR FORCE ( k l p , )

"",,,.'"'-,,.
AXIAL FORCE ( k i p . )

i.0~.4+

....

~.mtg+ . . . .

~,2~+~

o.{]~ " 4.~+,i + i / . o ~ + ~ +

~{it+,

Figure 7 Design force comparison of weakly and strongly coupled wall system (2-lumped walls)

axial forces (Figure 7c) increase by about 60%,. Consequently, the wall section chosen results in an overstrength of 3.25 with 0.8c/c reinforcement.

~
u.

'

0,

B4~lm-lo-WIl| Sit.niCk RatiO 2*/e

7.2. Inelastic r,sponse Pushover analysis of the coupled wall system (Figure 4c) showed that the stiffness and the yield displacement were similar to those of the original wall-frame system. Figure 8 shows that the roof displacement history and the drift response envelope of the coupled wall system and those of the original wall-frame system are similar under the Northridge (Newhall NS) ground motion, as anticipated 4. The coupled wall system results in about 25% saving in concrete and 31)% saving in steel while producing responses similar to those of the original wall-frame design. The maximum moments, shears and axial forces induced by the Northridge (Newhall NS) earthquake were also compared with the design ti)rces of Figure 7. Though the moments induced were consistent with the overstrength of the wails, the induced shear forces were larger (not consistent with moments) m the coupled walls, which is attributed to higher mode effects". 7.3. Ductility olcoul~h'd wall system The ratio of coupling beam strength to wall strength may be varied, to result in minimal wall ductility under a given beam-to-wall stiffness ratio. Saatcioglu 4 indicated that wall ductility can be reduced by increasing wall strength and selecting an optimal beam-to-wall strength ratio for thai wall strength. Though the reduction of ductility by increasing wall strength is obvious, the effect of beam-to-wall strength ratio on the wall ductility can be understood by studying lhe pattern of yielding of the coupled walls having different beam-to-wall strength ratios under monotonically increasing displacement magnitudes. The following section illustrates this through pushover analysis of the coupled wails.

Mo~I~:am~ Yield tension Will Yields (7ompre~ion Wall Yields

~,,~ ~,o :. /

/,

I-/.......
!"--

_/
,v

~-"

/I---~--

./~.~%

o.~"/_...... "

....

oo

i l~ DISPLACEMENT

" - - " ~-. . . . . . (;n.)

Figure9 Pushover of coupled walls having different beam-to wall strength ratios

7.3.1. Pattern of yielding Figure 9 shows the results of pushover analyses of coupled wall systems, with a wall strength of 400000 in kips (45 2(X) kN m) and the beam-to-wall strength ratios of 0.2, 0.5, 1.0 and 2.(F~. The strength of the coupled wall system increases with increase in the beam-to-wall strength ratio. Note that on each curve, the IMlowing distinct points are distinguished: El)beams start yielding: (2)most beams yield: (3)tension wall yields: and (4)compression wall yields. The results show that for the case of weak coupling, when the beam-to-wall strength is small, most beams yield at a very early, stage of loading and within a short increment of displacement. In this situation, the stiffness and the strength of the walls will primarily control the response of the system. Note that the tension wall and the compression wall yield at nearly the same maximum displacement in this case. As the beam-to-wall strength ratio is increased, the yielding of the beams starts at a higher level of strength and continues up to a larger displacement. Walls on the other hand start yielding early at a smaller displacement (/"i,eure

(a) Roof Displacement

lc

--I-'- ~ ~"'"-..'" .-02c ' O j ~ .

f':,
.,* ,." .

:" ,., ... ;"

,: ....

,,

<:E ~d-o
b.l

i. i.
......

- Io

TIME (See. )

- - o ~ 0 2 = 6 : 0 , ~ . . c ,~J " ' o ; d 1 o . 6 2

(b) D r i l l R a t i o

Figure8 Response comparison

of wail-frame and coupled wall system

614

Analysis of seismic performance of a code designed RC building: J. A. Munshi and S. K. Ghosh

Z3.2.

9), and the gap between the displacements causing the yielding of the tension wall and the compression wall widens. Thus, beams becomc increasingly active in contributing to the strength and behavior of the coupled wall system. For beam-to-wall strength ratios of 0.2, 0.5 and 1.0~. the beams still yield well before yielding starts in the walls. However, at a beam-to-wall strength ratio of 2%. yielding in the tension wall starts prior to yielding of most of the beams. Since beams at higher foors are subjected to larger rotation demands than beams at lower floors, the use of higher beam-to-wall strength ratios may also cause yielding of walls at higher floor levels which is undesirable. Higher beam-to-wall strength ratio, though advantageous from the strength point of view, can cause yielding of the walls at relatively smaller displacement magnitudes. Since the maximum displacements are known to be somewhat insensitive to the strength values, this would mean larger ductility demand on the walls for the same maximum displacement. However, larger strength of coupling beams could increase the energy dissipation (and damping) of the system and help reduce the maximum displacement and the ductility demand of the coupled wall system. Thus, the increase of the beam-to-wall strength ratio has two simultaneous and opposite effects, one of increasing, and, other of decreasing the ductility demand of the walls. This explains the rationale for using the optimum beam-to-wall strength ratio for minimizing the wall ductility. In general, the beams should yield prior to the walls, but their strength should be high enough so that they can dissipate enough energy through hysteresis. At the same time, the beam strength should be low enough not to cause much early yielding of" the walls due to the excessive axial forces. A careful selection of the beam-to-wall strength ratio is, thus, essential for adequate coupled wall response.

Dm'tili;v demaml

The yield displacement of the coupled wall system i~, assumed to be the displacement level at which bolh walls yield. Figure 9 indicates that this yield displacement appears to drop with increase in the beam-to-wall strength ratio. Thc yield displacements obtained in this manner arc used to determine the ductility demand of the cot,pied wall system. It is assumed that the overall ductility of the coupled wall system better represents the safety and stability of the system and can bc directly correlated with the design ductility limits prescribed in the codes. The effect of inelastic shear behavior of the compression wall. which becomes important after yielding of the tension wall. and the effect of linite plastic hinge length of the walls is not included in the analyses. Coupled wall systems with two wall strength levels of S1--400000in - kips (45200 kN m) with ().8~ steel and $2=550000in kips (62 150kN m) with l.b;ch steel were analyzed for the El Centro (1940) NS, Hachinohe and Northridge at Ncwhall NS, and Northridge at Sylmar EW ground motions. The higher wall strength $2 is u~,cd in order to reduce the wall ductility. Beam-to-wall strength ratios of" 4, 3, 2, I, 0.5 and 0.2% werc considered for each level of wall strength. Figure I0 shows the wall ductility demand curves for different beam-to-wail strength ratios under the four carthquakes. In general, the ductility demands increasc when thc beam-to-wall strength ratio approaches the limits of 4 and 0.2% and is at its minimum somewhere between these limits (Figure I0). However. in the case of the Northridge earthquake (Newhall station NS), this trend is les.,, clear. Under this earthquake component and the wall strength of SI (400000 in kips or 45 200 kN m), the displacement duc-

.....
- - - . :

Wall Strength=St Wall Strength=S2

0

..... - - - -

Wall Strength--S1 Wall Strength=S2

8;
7

(a) 1.5xEI Centro

1 (b) 1 . S x H a c h l n o h e

p-

0 5

'i
IEAM-TO-WALL STRENGTH RATIO

J.i, ..... eEAM-m-w

. . . . . . 6:6 . . . . . . ..... a. STRENGTH

TIO

II

ii
B

o.sx..w.o,, (.s)

BEAM-TO-WALL STRENGTH RATIO

Figure 10 Ductility demand of the walls under different earthquakes

i L ...... 616i...... ~ ...... a'~ ...... BEAM-TO-WALL STRENGTH RATIO

Analysis of seismic performance of a code designed RC building: J. A. Munshi and S. K. Ghosh

BEAM-TO-WALL ......... o.olo STRENGTH RATIO ...... 0.005
....

615

- -

0.020

BEAM-TO-WALL - o.olo STRENGTH RATIO ...... 0.005

.... 0.002

0.020

0.002

7,o~ +
,.',~..

7 ,o]
: ,. WA L STRENGTH

WALL STRENGTH S ' ,

~%:.:>-"",, ::,.',:, , .:.

'~

., ,

, . ,.
1.5xEl

e, t*+" r "
]
E-,o~ + ....

V"""
1.5xEl Centro ,.~,, . . . .

tw="l
l
L .% ~] g-"l + " ,~,

"'l
Centro ,,

i ~'

hue (sec)
7'~
z ='I

(a) El Centro Earthquake

. ..... "~,o .,

tiME (Soc)
,,., ,, :,:, . . ]

~,;,-,

,~,-,,

A
~

ft.,/~"';'':~''""":':'""'"'"'-'" ~ z t~;,~ VI"-".,/",~' ..J',,'~-,.'~.-.~-'~

, '":':"""~''"'"""'"""": / r ~"i:"-'!I': ,; '

WALL S R N T S1 T E GH
TIME
(S~c)

]WALL

S R N T 52 T E GH
TIME (,SEE)

(b) HachinoheEarthquake displacement u n d e r

Figure 71

Effect of c o u p l i n g b e a m s t r e n g t h on g l o b a l

a L l Centro, and b H a c h i n o h e e a r t h q u a k e s

tilities of 4,4, 5.4, 6, 5.5, 5.6 and 5.7 result (Figure lOd) for beam-to-wall strength ratios of 4, 3, 2, 1.0.5 and 0.2C/c, respectively. For the wall strength of $2 (550000 in kips or 62 150 kN m) under the same earthquake, the ductility values decreased to 2.4, 2.7, 3.5, 4.0, 3.8 and 3.5 for beamto-wall strength ratios of 4, 3, 2, 1, 0.5 and 0.2%. respectively (see Figure lOd), In general, the results indicate that for each level of wall strength, an optimum range of beam-to-wall strength ratios exists, that would result in minimum wall ductility. The wall strength has a more pronounced effect on the wall ductility. Similar results were obtained by SaatciogluL
7.3.3. Displacement history Figures 11 and 12 show the roof displacement histories of the coupled wall systems with wall strengths of S 1 and $2 and beam-to-wall strength ratios of 2, 1, 0.5 and 0.2e~ under the four earthquakes. The beam-to-wall strength ratio plays an important role under E1 Centro and Hachinohe earthquake components. After initial yielding, the displacement histories come distinctly apart and new peaks form later in the response history for different beam-to-wall strength ratios. The beam-to-wall strength ratio controls the peak displacements and wall ductility demands under the two earthquake motions (Figure 11). However, the Northridge earthquake components produce a large displacement pulse early during the ground excitation (Figure 12). The beam-to-wall strength ratio
BEAM-TO-WALL STRENGTH RATIO - ..... .....
....

does not seem to appreciably influence the displacement history or ductility demand of the building in this case. This is particularly noticeable for the Sylmar EW component of the Northridge earthquake which produces a large and reverse displacement pulse in the structure as shown in Figure 12a. Such an early and large displacement pulse results in simultaneous yielding of the coupling beams and the walls, thus minimizing the influence of beam-to-wall strength ratio on subsequent post-yield response. Although increase in the wall strength will reduce the ductility demand directly, the real issue in such situations is to reduce the peak displacement of the building. 8. Conclusions

For the selected earthquake intensity, the global ductility of code-designed moment resisting frame of the example building appeared to be reasonable and well within the design ductility limits prescribed in the code. The Parkfield and Northridge earthquakes, however, produced biased local and global deformation responses in the building, indicating that similar earthquakes could be critical for the moment frames of this building. The Northridge earthquake also induced biased response of the wall-frame system in the transverse direction of the example building. Weak coupling between the walls resulted in large ductility demands, which can be directly reduced by increasing the
- 0.020 BEAM-TO-WALL -0.010 STRENGTH RATIO ...... 0.005 .... 0.002 ] I ~ ' ,o| ~ s

0.020 0.010 0.005

0.002

_~

~t1[

Northrldge (SyImor-EW)

+
" ~ "-" io.. c
z ,

....
TIME (S.c)

ii
Z

]
~" , *

or~hrldge . ytmar-EW)

~,-,o~

(a) Sylmar (EW)

~', . . . .

/
Wall Str~agth S1

~ i[ IS la 20 ~

O! 1
~

-" X * W . l l ~ r c n ~ b S2 I 1 1-2 TIME ($ec.~

i~ i-6 la

;"

'8

hue (s..~

10

12

14

F, - i o 6

2o

(b) Newhall (NS)

Figure 12

Effect o f c o u p l i n g b e a m s t r e n g t h on g l o b a l d i s p l a c e m e n t u n d e r N o r t h r i d g e e a r t h q u a k e c o m p o n e n t s

616

Analysis of seismic performance of a code designed RC building: J. A. Munshi and S. K. Ghosh

establish definite criteria for eflicicnt dcsign of coupled walls.

wall strength. In general safety of the wall-frame systems involving coupled walls can be ensured by limiting the ductility demand in the walls while maximizing the energy dissipation of the system by allowing large ductility demands in the coupling beams. This can be achieved by selecting an adequate wall strength and an optimum value of the coupling beam-to-wall stiffness and strength ratio for a coupled wall system. Note that such a procedure would result in an efficient and economical coupled wall lateral load resisting system. The above procedure of optimization, however, may not work in case of an earthquake likely to produce a large displacement pulse in the coupled wall system early during the ground shaking resulting in simultaneous yielding of the walls and the coupling beams. Such large displacement pulse induced in the building by the Northridge earthquake components seems to render the coupling between walls ineffective for subsequent response. Under such earthquakes, controlling of peak displacement pulse is perhaps the key to ensuring adequate behavior and safety of coupled walls. From this limited study and that carried out by Saatcioglu +, it appears that the design of coupled walls can be optimized to yield efficient and economical lateral load resisting systems. More systematic and complete parameteric studies, considering different periods, wall strengths, beam-to-wall stiffness and strength ratios, and earthquake ground motions, however, will be required to

References
['nifiwm Building ('(~le. Int. ~'m!l. Buildin.~ Otft+iah. Wh.tticr, CA. 1994 2 Prakash. V.. Powell. G. II. and ('amphell. S. I)RAIN-21),k, slam and dynamic anal.v.~i,~ t~l phme structure,~. NISEE, Earthquake I'.ngmeering Research Center, t:ni',ersil) (71 Calilkwnia. Berkcle.'.. ( ' \ . I t~t;5 ~; Naeim, I:. 'On design implications of the Northridge earlhqtlakc rcc ords." Earthq. Spectra. 1995. It (I). 191-117) 4 Saatcioglu. M. Im,/a,~ti+ heh, Li~r and de,~ign ~[ c(trthqtt~t]~c /,'~l~hmt coupled uall~, Department t~l"('ivil Engineering. Norlh,.~o.lcrn t lni'.ersfly, E',anston, 11,. 19~1 5 Paulay, T. and Prie.,tly. M. J. N. Sei.~tnic design o/rcinltJr~ cd, ,,m rot<" and masonry huihlm.~,~. Wile). Nev+ York, 199+) 6 Ghosh. S. K. and .Marke+ icm,,, "Design of earthquake rc,,i',t:ml ,.hearwalls to pre,.unt shear failure." in Pro<. 4th ~,'S ,Vat. ('+m/. I'.arthq. Engn~, Puhn Spring,,. (+A, Vol. 2. Earthquake Engineering Ro, carL:h Institute. t'] ("rrito. (',.\. 199(]. pp. 9()5-913 l

S] conversion I in = 25.4 into I fl = 0.3048 m I kip = 4.448 kN I ill kip = 0.113 kN m I ksi = 6.895 Mpa