Seismic Analysis of Asymmetric Buildings with Flexible

Floor Diaphragms
Dhiman Basu1 and Sudhir K. Jain2

Abstract: Even though a rigid floor diaphragm is a good assumption for seismic analysis of most buildings, several building configu-
rations may exhibit significant flexibility in floor diaphragm. However, the issue of static seismic analysis of such buildings for torsional
provisions of codes has not been addressed in the literature. Besides, the concept of center of rigidity needs to be formulated for buildings
with flexible floor diaphragms. In this paper, the definition of center of rigidity for rigid floor diaphragm buildings has been extended to
unsymmetrical buildings with flexible floors. A superposition-based analysis procedure is proposed to implement code-specified torsional
provisions for buildings with flexible floor diaphragms. The procedure suggested considers amplification of static eccentricity as well as
accidental eccentricity. The proposed approach is applicable to orthogonal as well as nonorthogonal unsymmetrical buildings and accounts
for all possible definitions of center of rigidity.
DOI: 10.1061/ ASCE 0733-9445 2004 130:8 1169
CE Database subject headings: Buildings; Earthquake resistant structures; Seismic analysis; Torsion.

Introduction 1.5, 0.1, and 0.5; and as per MCBC 1995 1.5,
0.1, and 1.0. In Eqs. 1a and 1b , the first term on the
Buildings are seldom, if ever, perfectly symmetric and frequently right hand side accounts for the coupled lateral torsional effect
building vibrations involve coupling of lateral and torsional mo- that arises from lack of symmetry as well as amplification due to
tion torsional coupling . Besides the static eccentricity (e s j ) de- dynamic effects whereas the second term incorporates the acci-
fined as the distance between the center of mass CM and the dental torsional effect. The response obtained from Eqs. 1a and
center of rigidity CR at respective floors , codes often require 1b , whichever is higher, should be used in the design.
the designer to incorporate accidental eccentricity. Accidental ec- Wherever building codes specify 1.0 e.g., in UBC
centricity accounts for factors such as the rotational component of 1997; NZS 4203 1992 , the location of the CR need not be ex-
ground motion about the vertical axis, the difference between plicitly calculated to implement torsional provisions. In such
computed and actual values of the mass, stiffness or yield cases, a rigid floor diaphragm building can be analyzed by apply-
strength, and an unfavorable distribution of live load mass. Codal ing design lateral force at a point away from the CM by b j or
provisions on torsion differ significantly. However, the provisions b j , as the case may be. However, where the codes specify
of most codes for design eccentricity e d j at the j-th floor for static values of and different from 1.0 e.g., in NBCC 1995; MCBC
analysis of buildings can be expressed in the following general 1995 the general impression used to be that one needs to locate
form: the CR. There are some difficulties in explicitly calculating the
edj esj bj (1a) location of the CR in a simple manner. Hejal and Chopra 1987
have proposed a stiffness matrix based formulation to locate the
edj esj bj (1b) CR and the approach is applicable to an arbitrarily shaped dia-
where e s j static eccentricity at j-th floor; b j plan dimension of phragm with arbitrary orientation of the principal planes of lateral
the j-th floor normal to the direction of ground motion; and , load resisting elements. However, the approach is restricted to
and are specified constants. For instance, as per UBC 1997 buildings with rigid floor diaphragms and cannot be implemented
and NEHRP 1997a,b 1.0, 0.05, and 1.0; as per NZS with available standard building analysis software. At about the
4203 1992 1.0, 0.1, and 1.0; as per NBCC 1995 same time, a similar method was documented by Alcocer 1986
and Damy-Rios and Alcocer 1987 . A superposition based ap-
1
Scientist, Structural Engineering Research Center, CSIR Complex, proach was proposed Goel and Chopra 1993 to implement Eq.
Taramani, Chennai 600 113, India; formerly, Graduate Student, Dept. of 1 , which does not require locating the CR explicitly. However,
Civil Engineering, Indian Inst. of Technology Kanpur, Kanpur 208 016, this approach is applicable only for buildings with rigid floor
India. E-mail: dhiman [email protected] diaphragms that have an orthogonal system.
2
Professor, Dept. of Civil Engineering, Indian Inst. of Technology A rigid floor diaphragm is a good assumption in most build-
Kanpur, Kanpur 208 016, India. E-mail: [email protected] ings. However, floor diaphragms in some buildings may have
Note. Associate Editor: Brad Cross. Discussion open until January 1, considerable flexibility in their own plane e.g., buildings that are
2005. Separate discussions must be submitted for individual papers. To
long and narrow or buildings with stiff end walls . In such build-
extend the closing date by one month, a written request must be filed with
the ASCE Managing Editor. The manuscript for this paper was submitted ings, design force for a particular floor cannot be applied at one
for review and possible publication on April 19, 2000; approved on July single point say, the CM or at some eccentricity of that floor. If
29, 2003. This paper is part of the Journal of Structural Engineering, the floor slabs are completely flexible, the lateral load distribution
Vol. 130, No. 8, August 1, 2004. ©ASCE, ISSN 0733-9445/2004/8- is governed by the tributary mass concept and the issue of torsion
1169–1176/$18.00. does not enter the picture. However, when the floor slabs have

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / AUGUST 2004 / 1169

intermediate flexibility, i.e., floor diaphragms that are neither rigid
nor completely flexible, floor diaphragm flexibility must be ex-
plicitly accounted for in the analysis. A considerable amount of
literature is available on the dynamics of buildings with flexible
floor diaphragms e.g., Shepard and Donald 1967; Jain and Jen-
nings 1984; Jain 1984a,b; Lu et al. 1984; Jain and Jennings 1985;
Celbi et al. 1989; Jain and Mandal 1992; Tena-Colunga 1992;
Jain and Mandal 1995; Tena-Colunga and Abrams 1995, 1996;
Tremblay and Stiemer 1996; Medhekar 1997; de la Colina 1999;
Ju and Lin 1999; Tremblay et al. 2000 . However, only a limited
number of published work on dynamic analysis of unsymmetrical
flexible floor diaphragm buildings could be located e.g., Tena-
Colunga 1992; Tena-Colunga and Abrams 1995, 1996; de la
Colina 1999 . Further, static seismic analysis of unsymmetrical
buildings with flexible floor diaphragms has not yet been ad-
dressed in the literature. Moreover, the concept of CR for flexible
floor system needs to be formulated.
This paper proposes a procedure for static seismic anlaysis of
buildings with significant floor diaphragm flexibility but not
completely flexible as well as torsional coupling. The approach
proposed by Goel and Chopra 1993 for rigid floor system has
been modified and suitably extended to buildings with flexible
floor diaphragms. Two definitions of center of rigidity, namely, all
floor and single floor, are proposed for buildings with flexible
floor diaphragms. The proposed method is such that, as the floor
diaphragm rigidity increases, the results are close to those appli-
cable for buildings with rigid floor diaphragms. Fig. 1. No-torsion condition in buildings with flexible floor dia-
The building is assumed to have a single wing only e.g., phragm: a Undeformed floor diaphragm; b deflected shape of floor
rectangular or trapeziod in plan and buildings having multiple slab under in-plane loading with torsion; c deflected shape of floor
wings e.g., L, V, Y, etc. shaped are not considered. However, slab under in-plane loading without torsion
lateral load resisting elements need not be oriented orthogonally.
The structure is considered linear and elastic. Further, it is as-
sumed to be rigidly held to the ground, so soil–structure interac-
tion effects and flexibility of the foundation are ignored. In order to apply Eq. 1 to flexible floor buildings, the CR
needs to be defined for such buildings. In the case of a rigid floor
Center of Rigidity system, ‘‘no-torsional rotation of floor’’ in the definition of CR
can be viewed as no twisting of the line connecting the center
Unlike a single-story building, a multistory building with a rigid node of both ends of the diaphragm. Hence, in the case of a
floor diaphragm does not have a unique or generally accepted flexible floor diaphragm system, identical horizontal displacement
single definition of the CR. The following two definitions are of center nodes at both ends of the diaphragm can be considered
commonly used. the same as a no-torsion response. For the floor slab shown in Fig.
1. Centers of rigidity are the set of points located on each floor, 1 a , AB is the line that connects the center nodes at both ends of
through which application of lateral load profile would cause the undeformed diaphragm. Under lateral loads, the floor slab is
no rotation in any floor Poole 1977; Cheung and Tso 1986 . deformed, translated and rotated Fig. 1 b and line A 1 B 1 forms
According to this definition, the location of the CR is depen- an angle with line AB. Thus, to obtain the no-torsion situation, it
dent on building stiffness properties as well as on the profile is necessary that points A and B must undergo equal horizontal
of the applied lateral load. This definition is referred to as displacement so line A 1 B 1 must remain parallel to AB Fig. 1 c .
‘‘all floor CR’’ in this paper. Therefore, the two definitions of CR can be extended to buildings
2. The center of rigidity of a floor is defined as the point on the with flexible floor diaphragms as follows.
floor where application of lateral load passing through that 1. Apply the design lateral load at all floors such that the lateral
point does not cause any rotation of that particular floor al- load at each floor is distributed along the length of the floor
though the other floors may rotate Humar 1984 . This defi- in proportion to the mass distribution. Next, at each floor,
nition is independent of the magnitude of the applied lateral constrain the center nodes of both ends of the diaphragm
load, and is referred to as ‘‘single floor CR’’ in this paper. such that they undergo equal horizontal displacement at each
It should be mentioned here that application of torsion provi- respective floor level. The resultant shear forces of all lateral
sions based on different definitions of CR could result in different load resisting elements in the stories immediately above and
member forces. Since the all floor definition is more convenient to below a particular floor pass through the CR of that floor all
implement in a design office, it is most frequently used. Some floor CR definition; Fig. 2 a .
codes also define eccentricity with respect to the shear center 2. Apply lateral load distributed along length of one of the
SC ; see e.g., MCBC 1995 . Discussion in this paper is limited floors such that it is proportional to the floor mass distribu-
to torsional provisions based on center of rigidity but the treat- tion. Next, constrain the center nodes of both ends of that
ment presented here can be conveniently extended to provisions floor diaphragm so that they undergo equal horizontal dis-
based on the shear center also. placement. The resultant shear forces of all lateral load re-

1170 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / AUGUST 2004

 Fig. 2. Procedure for locating a all floor center of rigidity and b single floor center of rigidity

sisting elements in the stories immediately above and below Figs. 3 b and e can be replaced by the load cases of Figs. 3 f and
that floor pass through the CR of that floor single floor CR g . Let F cr response when the load profile acts through the CR;
definition; Fig. 2 b . F cm response when the load profile acts through the CM; and
R ac response due to the accidental torsional moment vector
b j F j . Now following Eq. 1a , design response Fig. 3 a can
Review of Goel and Chopra’s Approach to Rigid be expressed as summations of response to Figs. 3 f, g, and d ;
Floor Diaphragm
F d F cr F cm F cr R ac (2a)
Here we review a procedure for analysis for implementation of
In a similar way, following Eq. 1b one can express the design
Eq. 1a for rigid floor diaphragm buildings following the proce-
response as
dure suggested by Goel and Chopra 1993 . Goel and Chopra’s
procedure is applicable for the all floor definition of CR only. F d F cr F cm F cr R ac (2b)
Further, this approach is restricted to orthogonal systems. Goel
and Chopra 1993 have pointed out that lateral force profile F j Here, F cr reflects the pure translational response while F cm con-
applied at design eccentricity Fig. 3 a according to Eq. 1a can siders translational response as well as the effect of torsion due to
be considered superposition of the following three cases: the eccentricity e s j . Hence, (F cm F cr ) is the torsional effect due
1. A force profile F j acting through the CM Fig. 3 b ; to the calculated static eccentricity, and (F cm F cr ) implies the
2. A moment profile ( 1)e s j F j anticlockwise moment torsional effect is due to amplified static eccentricity. It should be
is considered positive Fig. 3 c ; and noted that, in Eq. 2 , the sign of the response obtained from R ac
3. A moment profile b j F j Fig. 3 d . should be considered in that it increases the magnitude of the
Here, the moment in case 2 can be written in terms of a load response obtained from the first two terms. This is to account for
profile ( 1)F j that acts at the CM and load profile ( the location of the CR on either side of the CM. Hence, Eq. 2
1)F j that acts through the CR Fig. 3 e . The load cases of may be rewritten as

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / AUGUST 2004 / 1171

Fig. 3. Alternate form of code specified design eccentricity: a De- Fig. 4. Inertia force distribution for flexible floor diaphragm: a
sign case; b force profile applied through CM; c additional tor- In-plane floor displacement and b actual and assumed inertia force
sional moment due to dynamic amplification; d accidental torsional distribution for uniform mass distribution
moment; e alternate representation of additional torsional moment
due to dynamic amplification; f pure translation case; and g total
amplified torsion with dynamic amplification. profile is calculated. Second, using this calculated displacement
profile the load distribution is modified, and so on.
It is shown in Eq. 2a that the design torsional response can
F d F cr F cm F cr R ac (3a) be expressed as superposition of the following three cases: i
no-torsion case (F cr ), ii times the difference in response of
F d F cr F cm F cr R ac (3b) natural torsion and no-torsion cases (F cm F cr ) , and iii ac-
Eq. 3 can easily be implemented with the help of any standard cidental torsion case (R ac ). This concept of superposition is also
building analysis software. F cm can be calculated by applying the applicable to buildings with flexible floor diaphragms except that
lateral load profile at the CM of the floor slabs of the building. To details for each case are different. The natural torsion case (F cm )
obtain F cr , the lateral load profile is applied at the CMs but the is the response with the load profile applied in proportion to the
floor slabs are constrained to move along the direction of applied mass distribution along the floor length. For the no-torsion case
loading. This can be achieved by placing roller supports at the (F cr ) the same force profile is applied while the floor diaphragms
floors in the direction normal to the direction of applied load. of the building are constrained accordingly. Finally, to obtain a
However, when the load profile is applied through single floor response for the accidental torsional moment, a linearly varying
centers of rigidity the building may translate along both orthogo- force distribution can be applied such that it meets the code-
nal directions. A similar situation also arises if the building is specified accidental torsional moment Fig. 5 .
nonorthogonal. Therefore, the ‘‘roller model’’ restricts the ap- The proposed approach differs from that of Goel and Chopra
proach from being applicable to these two cases. 1993 in defining the no-torsion case. It defines the no-torsion
case in terms of constraining the center nodes of both ends of the
diaphragm so that they undergo equal horizontal displacement;
Analysis for Buildings with Flexible Floor this allows the building to translate along both orthogonal direc-
Diaphragms tions. This also enables the proposed approach to become appli-
cable for all definitions of center of rigidity and to orthogonal as
Here the procedure proposed by Goel and Chopra 1993 is ex- well as nonorthogonal buildings.
tended to buildings with flexible floor diaphragms. The proposed
approach considers all floor as well as single floor definitions of
CR and is applicable to orthogonal and nonorthogonal systems.
Analysis of flexible floor diaphragm buildings is different from
that of rigid floor diaphragms in several respects. Lateral load
cannot be applied at a single point at the CM or at an offset on
the diaphragm of the flexible floor building; it has to be distrib-
uted along the length of the floor. Further, earthquake inertia force
associated with any vibration mode is proportional to the mass
distribution along the floor length and depends on the displace-
ment profile of that particular mode Fig. 4 . Estimation of the
actual inertia force distribution that takes into consideration the
displacement profile may be tedious. However, the static force
procedure is an approximation of the building response obtained
from dynamic analysis; thus in most cases, it is adequate to as-
sume that the inertia force is in proportion to the mass distribution
along the length of the floor. In extreme cases of large aspect ratio
Fig. 5. Estimate of response due to accidental torsional moment: a
where the effect of the displacement profile is important, an itera-
floor slab at j-th floor and b load profile to account for positive
tive approach may be used to obtain the inertia force profile. First,
accidental torsional moment
a distribution of the load profile is assumed and the displacement

1172 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / AUGUST 2004

 Table 1. Shear Force in Shear Walls Using the ‘‘All Floor’’ Defini-
tion of Center of Rigidity
Shear Shear
from from Design
Member F cm a F cr b R ac c Eq. 3a) Eq. 3b shear
Wall Story kN kN kN kN kN kN
SW1 3 142.62 270.10 38.27 117.20 180.89 180.89
2 228.96 430.96 58.00 185.96 286.96 286.96
1 263.18 463.56 64.44 227.43 327.62 327.62

SW2 3 4.57 4.64 0.30 4.84 4.87 4.87

2 8.85 7.40 0.42 10.00 9.27 10.00
1 13.86 11.57 0.59 15.60 14.45 15.60

SW3 3 6.70 5.24 0.42 7.85 7.12 7.85

2 14.19 9.43 1.44 18.01 15.63 18.01
Fig. 6. Example of building a elevation and b its plan all dimen- 1 20.27 12.63 2.33 26.42 22.60 26.42
sions are in millimeters
SW4 3 8.58 4.61 1.44 12.01 10.02 12.01
2 16.14 7.63 2.52 22.92 18.66 22.92
The step-by-step algorithm for the proposed static analysis 1 23.30 11.37 3.80 33.07 27.10 33.07
method is as follows. SW5 3 104.32 24.74 23.88 167.99 128.20 167.99
2 196.22 45.79 42.64 314.08 238.86 314.08
Procedure with All Floor Definition 1 245.71 84.97 52.23 378.31 297.94 378.31
a
1. Apply the design lateral load at all floor levels such that F cm response due to earthquake force profile applied in proportion to
the mass distribution along the floor length.
lateral load at each floor is distributed along the length of the b
floor in proportion to the mass distribution. Analyze the F cr response in no-torsion case.
c
structure to obtain response F cm . R ac response due to accidental torsional moment.
2. At all floor levels constrain the center nodes of both ends of
the diaphragm so that they undergo equal horizontal dis-
placement at each respective floor level. Analyze the struc-
ture for the same force profile as that in step 1 above to SW5 is 4 m in length; the length of each intermediate shear wall
obtain response F cr . SW2–SW4 is 1.5 m. The thickness of all the shear walls is 150
3. Apply the force couple that represents accidental torsional mm. All columns are 400 mm 400 mm, the longitudinal beams
moments Fig. 5 at all floors. Analyze the original model of are 250 mm 550 mm, and the transverse beams are 250 mm
the structure subjected to this force distribution to obtain 400 mm. The floor slabs are taken as 200 mm thick. The aspect
response R ac . ratio of the floor plan 50 m 10 m is chosen as 5, slightly more
4. Combine the above responses following Eqs. 3a and 3b . than that of typical buildings to show the effect of floor flexibility.
The first story height is taken as 4.5 m whereas the height of the
other two stories is 3.2 m. The modulus of elasticity and shear
Procedure with Single Floor Definition modulus of concrete are taken as 2.55 107 and 1.06
1. Apply the lateral load profile at all floor levels to the original 107 kN/m2 , respectively. The gross moment of inertia is used in
model of the building in a way similar to that in Procedure the analysis and the extra rigidity in the rigid end zones of the
with all Floor Definition and calculate the response F cm . members is not modeled. Design eccentricity is taken according
2. Constrain the center nodes of both ends of the diaphragm to Eqs. 1a and 1b with constants , and equal to 1.5, 0.1
only at j-th floor so that they undergo equal horizontal dis- and 1.0, respectively.
placement. Next, apply the distributed lateral load at the j-th The building is analyzed using the SAP2000 program. Beams
floor in proportion to the mass distribution and calculate re- and columns are modeled as line elements. Shear walls are mod-
j eled as shell elements by considering the membrane as well as the
sponse F cr .
3. Repeat step 2 for each floor, j 1, . . . ,n and obtain F cr plate behavior while floor slabs are considered pure membrane
n j elements. The design earthquake force on the building is calcu-
j 1 F cr .
4. Calculate the response due to accidental torsion in a way lated using Indian code provisions for seismic zone V; this gives
similar to in the all floor case and combine the responses the design seismic load profile as 62.0 181.6 363.9 T kN starting
following Eqs. 3a and 3b . from the bottommost floor. Resultant force shear force in the
shear walls is presented in Tables 1 and 2 in which all floor and
single floor definitions of CR, respectively, are considered. It is
Illustration of Methodology observed that for this example the building response obtained
from both the definitions is practically the same. Therefore, the all
A three-story building Fig. 6 with floor diaphragms that are floor definition of CR, being easier to implement, is used in this
neither rigid nor completely flexible is analyzed to study its tor- paper as the basis for further comparisons. It should be noted that
sional response. The building consists of two end shear walls and the sum of the shear forces in the walls will not be equal to the
seven interior frames, three of which include a small shear wall. story shear applied since the columns also share some lateral load.
The left shear wall SW1 is 10 m in length while right shear wall The building is then analyzed for torsion without accounting

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / AUGUST 2004 / 1173

Table 2. Shear Force in Shear Walls Using the Single Floor Definition of Center of Rigidity
Shear Shear
Member from from Design
F cm a F 1cr b 2 b
F cr F 3cr b F cr c R ac d Eq. 3a) Eq. 3b) shear
Wall Story kN kN kN kN kN kN (kN) (kN) kN
SW1 3 142.62 0.75 16.58 250.42 267.75 38.27 118.33 180.89 180.89
2 228.96 5.85 122.16 287.44 415.45 58.00 193.72 286.96 286.96
1 263.18 34.66 137.00 292.95 464.61 64.44 226.91 327.62 327.62

SW2 3 4.57 0.00 7.52 12.31 4.79 0.30 4.76 4.87 4.87
2 8.85 2.85 9.69 0.41 7.25 0.42 10.07 9.27 10.07
1 13.86 4.97 4.45 2.39 11.81 0.59 15.48 14.45 15.48

SW3 3 6.70 0.21 9.35 14.88 5.32 0.42 7.81 7.12 7.81
2 14.19 3.43 11.79 1.24 9.60 1.44 17.93 15.63 17.93
1 20.27 6.25 5.51 1.10 12.86 2.33 26.31 22.60 26.31

SW4 3 8.58 0.00 6.56 11.28 4.72 1.44 11.95 10.02 11.95
2 16.14 2.50 9.09 1.67 8.26 2.52 22.60 18.66 22.60
1 23.30 4.76 4.28 2.38 11.42 3.80 33.04 27.10 33.04

SW5 3 104.32 0.17 13.17 11.98 25.32 23.88 167.70 128.20 167.70
2 196.22 5.04 7.23 46.81 59.08 42.64 307.43 238.86 307.43
1 245.71 5.42 22.19 55.39 83.00 52.23 379.30 297.94 379.30
a
F cm response due to the earthquake force profile applied in proportion to the mass distribution along the floor length.
b 1
F cr ,F 2cr ,F 3cr response from no-torsion case of first, second and third floors, respectively.
c
F cr F 1cr F 2cr f 3cr .
d
R ac response due to the accidental torsional moment.

for accidental torsion, i.e., 0 in Eqs. 1a and 1b ; the result- normal tothe direction of applied force. Therefore, accidental tor-
ant force is shown in column 3 of Table 3. A comparison with sion may contribute significantly to the resultant design force for
results when accidental torsion is also considered column 4, long buildings. It should be mentioned, however, that the value of
Table 3 indicates that accidental torsion contributes up to 26.8% in the codes that specify accidental torsion is based on studies
of the wall shear for this example. The accidental torsional mo- of rigid floor diaphragm buildings; its applicability for flexible
ment is directly proportional to the dimension of the building floor diaphragm buildings needs to be established.

Table 3. Comparison of the Contribution of Torsion and Diaphragm Flexibility to Design Response
Flexible floor
Flexible floor Flexible floor Rigid floor diaphragm with
diaphragm without diaphragm with diaphragm without lateral loads applied
Member accidental torsion accidental torsion accidental torsion proportionally to floor mass
Shear force Shear force Difference Shear force Error Shear force Error
Wall Story kN kN % kN % kN %
SW1 3 142.62 180.89 26.8 150.84 5.8 142.62 0
2 228.96 286.96 25.3 243.24 6.2 228.96 0
1 263.18 327.62 24.5 273.44 3.9 263.18 0

SW2 3 4.57 4.87 6.6 4.39 3.9 4.57 0

2 9.58 10.00 4.4 6.09 36.4 8.85 7.6
1 15.01 15.60 3.9 10.62 29.3 13.86 7.7

SW3 3 7.43 7.85 5.7 8.31 11.8 6.70 9.8

2 16.57 18.01 8.7 11.57 30.2 14.19 14.4
1 24.09 26.42 9.7 17.63 26.8 20.27 15.9

SW4 3 10.57 12.01 13.6 12.23 15.7 8.58 18.8

2 20.40 22.92 12.4 17.05 16.4 16.14 20.9
1 29.27 33.07 13.0 24.63 15.9 23.30 20.4

SW5 3 144.11 167.99 16.6 154.47 7.2 104.32 27.6

2 271.44 314.08 15.7 289.10 6.5 196.22 27.7
1 326.08 378.31 16.0 337.72 3.6 245.71 24.7

1174 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / AUGUST 2004

 The building cited as an example is further analyzed by con- contribution of accidental torsion as well as the torsional ampli-
sidering the floor diaphragms as rigid in their own planes and fication terms can be quite significant. However, the usual codal
with the all floor definition of CR as the basis of torsional provi- specification of accidental eccentricity as a fraction of the build-
sions but excluding the contribution by accidental torsion col- ing dimension may be somewhat conservative for such buildings
umn 6, Table 3 . It is observed that the shear force in the end and this issue needs to be addressed in the future.
shear walls SW1 and SW5 is somewhat overestimated up to Horizontal offset buildings constitute a class of structures that
7.2% and that in intermediate shear walls SW2–SW4 is signifi- are particularly prone to in-plane floor deformation and torsion
cantly underestimated up to 36.4% by assumption of a rigid occurring simultaneously. The present work may be useful for
diaphragm. developing a methodology for the treatment of such buildings.
Consider the case when no special attention is given to issues
of torsion and the flexible floor diaphragm building is analyzed
for the lateral load profile in proportion to the mass distribution Acknowledgments
along the floor. This implies consideration only of static eccen-
tricity without dynamic amplification and accidental torsion The writers gratefully acknowledge Dr. Rakesh K Goel and
1.0, 0 and 1.0 . In this case column 8, Table 3 , the wall anonymous reviewers for a comprehensive review of the manu-
forces are underestimated by up to 27.7% in the end walls and by script and valuable suggestions for the improvement of it. Special
up to 20.9% in the intermediate walls. It should be noted here that thanks are due to Kaustubh Dasgupta for independently verifying
the shear force induced in SW1 and also at the top story of SW2 the numerical results. The research reported in this paper was
is the same as that in column 3 of Table 3. This is because the supported by a grant from the Ministry of Human Resource De-
resultant design force for the stiff side elements are governed by velopment, government of India.
Eq. 3b with 1.0.
The results for the building given as an example above clearly
show that torsional effects may be quite significant in buildings Notation
with a flexible floor diaphragm. In such buildings, neither the
floor diaphragm flexibility nor the torsional response can be ig- The following symbols are used in this paper:
nored. Moreover, ignoring either accidental torsion or torsional b cm j distance of CM of j-th floor from center node of left
amplification may cause significant differences in design forces. end of diaphragm;
However, when the floor diaphragm is completely or significantly bj lateral dimension of j-th floor normal to direction of
flexible Tena-Colunga and Abrams 1996 , each individual frame ground motion;
responds almost independently without any interference from the edj design eccentricity of j-th floor;
others and the torsional contribution may be significantly dimin- esj static eccentricity of j-th floor;
ished. F cm response with load profile applied in proportion to
mass distribution along floor length;
F cr response due to design load profile when center
Summary and Conclusions nodes of both ends of diaphragm at all floor
levels are constrained to undergo equal horizontal
In seismic analysis of buildings, the floor slab is usually assumed displacement;
to be rigid in its own plane. However, for many buildings that are F (crj) response due to design load at constrained j-th floor
long and narrow or have stiff end walls, floor diaphragm flexibil- while other floors are unconstrained;
ity must be accounted for in the distribution of lateral load. Con- Fd design response;
siderable research has been reported in the literature on the dy- Fj lateral load at j-th floor;
namics of flexible floor diaphragm buildings; however, the issue Fj lateral load profile;
of seismic design of such buildings that takes into consideration R ac response due to accidental torsion;
torsional provisions of the codes has not yet been addressed. In w lj intensity of load profile at the left end of j-th floor;
this paper we developed a framework for analysis of such build- w rj intensity of load profile at right end of j-th floor;
ings following usual codal requirements for torsion. The building torsional amplification factor;
is assumed to have a single wing only, i.e., buildings with mul- normalized accidental eccentricity; and
tiple wings e.g., L, V, Y, etc. shaped are not considered. torsional amplification factor.
The definition of center of rigidity for rigid floor diaphragm
buildings is extended to flexible floor buildings. The no-torsion
condition for flexible floor buildings is defined such that center References
nodes at either end of the diaphragm are constrained so that they
undergo equal horizontal displacement. The proposed analysis Alcocer, S. M. 1986 . ‘‘Centro de torsion de edificios obtenido matricial-
procedure considers the final response as the superposition of mente.’’ BS thesis, Facultad de Ingenieria, UNAM, Mexico, DF,
three cases: the no-torsion case, amplification of the static eccen- Mexico.
tricity, and accidental torsion. The proposed procedure ensures Celbi, M., Bongiovanni, G., Safak, E., and Brady, A. G. 1989 . ‘‘Seismic
that the resultant member force is close to that of rigid floor response of a large-span roof diaphragm.’’ Earthquake Spectra, 5 2 ,
337–350.
buildings as the floor diaphragm rigidity increases.
Cheung, V. W. T., and Tso, W. K. 1986 . ‘‘Eccentricity in irregular mul-
Analysis results of a sample building clearly show the signifi- tistory buildings.’’ Can. J. Civ. Eng., 13 1 , 46 –52.
cance of considering the torsion provisions of design codes for Damy-Rios, J. E., and Alcocer, S. M. 1987 . ‘‘Obtencion del centro de
asymmetric flexible diaphragm buildings. It is seen that treating torsion de edificios.’’ Proc., 6th Congreso Nacional de Ingenieria
the diaphragms of such buildings as rigid for torsional analysis Estructural, 60– 67 Mexican Society on Structural Engineering,
may cause considerable error. The example also illustrates that the Mexico, DF, Mexico.

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / AUGUST 2004 / 1175

de la Colina, J. 1999 . ‘‘In-plane floor flexibility effects on torsionally Depto. del Distrito Federal, Mexico, DF, Mexico.
unbalanced systems.’’ Earthquake Eng. Struct. Dyn., 28 12 , 1705– National Building Code of Canada NBCC . 1995 . Associate Committee
1715. on the National Building Code, National Research Council of Canada,
Goel, R. K., and Chopra, A. K. 1993 . ‘‘Seismic code analysis of build- Quèbec.
ings without locating centers of rigidity.’’ J. Struct. Eng., 119 10 , National Earthquake Haxards Reduction Program NEHRP . 1997a .
3039–3055. ‘‘NEHRP Recommended Provisions For Seismic Regulations For
Hejal, R., and Chopra, A. K. 1987 . ‘‘Earthquake response of torsionally New Buildings and Other Structures. Part I: Provisions.’’ Rep. No.
coupled buildings.’’ Technical Rep. No. UCB/EERC-87/20, Earth- FEMA 302, Federal Emergency Management Agency, Washington,
quake Engineering Research Center, Berkeley, Calif. D.C.
Humar, J. L. 1984 . ‘‘Design for seismic torsional forces.’’ Can. J. Civ. National Earthquake Hazards Reduction Program NEHRP . 1997b .
Eng., 12 2 , 150–163. ‘‘NEHRP Recommended Provisions For Seismic Regulations For
Jain, S. K. 1984a . ‘‘Seismic response of building with flexible floors.’’ New Buildings and Other Structures. Part II: Commentary.’’ Rep. No.
J. Eng. Mech., 110 1 , 125–129. FEMA 303, Federal Emergency Management Agency, Washington,
Jain, S. K. 1984b . ‘‘Continuum models for the dynamics of buildings.’’ D.C.
J. Eng. Mech., 110 12 , 1713–1730. New Zealand Standard NZS . 4203:1992. 1992 . ‘‘General Structural
Jain, S. K., and Jennings, P. C. 1984 . ‘‘Continuous model for frame and
Design and Design Loadings for Buildings.’’ Standards Assoc. of New
shear wall buildings with flexible floors.’’ Proc., 8th World Conf. on
Zealand, Wellington New Zealand.
Earthquake Engineering, International Association for Earthquake
Poole, R. A. 1977 . ‘‘Analysis for torsion employing provisions of NZRS
Engineering, Tokyo, 4, 743–750.
4203, 1974.’’ Bull., N. Z. Nat. Soc. Earthquake Eng., 10 4 , 219–225.
Jain, S. K., and Jennings, P. C. 1985 . ‘‘Analytical models for the low
Shepard, R., and Donald, R. A. D. 1967 . ‘‘The influences of in-plane
rise buildings with flexible floor diaphragms.’’ Earthquake Eng.
Struct. Dyn., 13 2 , 225–241. floor flexibility on the normal mode properties of buildings.’’ J. Sound
Jain, S. K., and Mandal, U. K. 1992 . ‘‘Dynamics of buildings with Vib., 5 1 , 29–36.
V-shaped plan.’’ J. Eng. Mech., 118 6 , 1093–1112. Tena-Colunga, A. 1992 . ‘‘Seismic evaluation of unreinforced masonry
Jain, S. K., and Mandal, U. K. 1995 . ‘‘Dynamics of buildings with structures with flexible diaphragms.’’ Earthquake Spectra, 8 2 , 305–
Y-shaped plan and flexible floor diaphragms.’’ J. Struct. Eng., 121 6 , 318.
1004 –1012. Tena-Colunga, A., and Abrams, D. P. 1995 . ‘‘Simplified 3-D dynamic
Ju, S. H., and Lin, M. C. 1999 . ‘‘Comparisons of building analyses analysis of structures with flexible diaphragms.’’ Earthquake Eng.
assuming rigid or flexible floors.’’ J. Struct. Eng., 125 1 , 25–31. Struct. Dyn., 24 2 , 221–232.
Lu, L. W., Masayoshi, N., and Huang, T. 1984 . ‘‘Effect of diaphragm Tena-Colunga, A., and Abrams, D. P. 1996 . ‘‘Seismic behavior of struc-
flexibility on seismic response of building structures.’’ Proc., 8th ture with flexible diaphragms.’’ J. Struct. Eng., 122 4 , 439– 445.
World Conf. on Earthquake Engineering, International Association for Tremblay, R., and Stiemer, S. F. 1996 . ‘‘Seismic behavior of single story
Earthquake Engineering, Tokyo, 4, 735–742. steel structures with flexible diaphragm.’’ Can. J. Civ. Eng., 23, 49–
Medhekar, M. S. 1997 . ‘‘Seismic evaluation of steel buildings with 62.
concentrically braced frames.’’ PhD thesis, Dept. of Civil and Envi- Tremblay, R., Berair, T., and Filiatrault, A. 2000 . ‘‘Experimental behav-
ronmental Engineering, University of Alberta, Edmonton, Alta., ior of low-rise steel buildings with flexible roof diaphragms.’’ Proc.,
Canada. 12th World Conf. on Earthquake Engineering, paper No. 2567.
Mexico City Building Code MCBC 1995 . ‘‘Mexico City building code, Uniform Building Code UBC . 1997 . ‘‘Uniform building code.’’ Int.
complementary technical norms for earthquake resistant design.’’ Conf. of Building Officials, Whittier, Calif.

1176 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / AUGUST 2004