Engineering Structures 24 (2002) 613623

www.elsevier.com/locate/engstruct

Efficient seismic analysis of high-rise building structures with the

effects of floor slabs
Dong-Guen Lee
a

a,*

, Hyun-Su Kim b, Min Hah Chun

Department of Architectural Engineering, Sungkyunkwan University, 300 Chun-chun-dong, Jang-an-gu, Suwon, 440-746, South Korea
b
MIDAS Information Technology Co. Ltd., IT Venture Tower East Wing, Garak-dong, Songpa-gu, Seoul, 138-803, South Korea
c
Skilling Ward Magnusson Barkshire, 1301 Fifth Avenue, Seattle, WA 98101, USA
Received 27 August 2001; received in revised form 25 October 2001; accepted 29 October 2001

Abstract
Box system structures, composed of only reinforced concrete walls and slabs, have been recently adopted for many high-rise
apartment buildings. Commercial software such as ETABS, commonly used for the analysis of high-rise apartment buildings,
assumes a rigid diaphragm for floor slabs. The flexural stiffness of slabs is generally ignored in the analysis. This assumption may
be reasonable for the analysis of framed structures. In box system structures like apartment buildings, however, the floor slabs may
have a significant influence on the lateral response of the structures. If the flexural stiffness of slabs in a box system structure is
totally ignored, the lateral stiffness may be significantly underestimated. In reality, the cracked section property of a slab will
determine the amount of its flexural stiffness that will be included in the analysis. In order to include the flexural stiffness of slabs,
the slab needs to be modeled with plate elements. If the slab is subdivided into many plate elements while keeping each shear wall
with a large element (as generally modeled with commercial software), the compatibility condition will not be satisfied at the
interface of the slab and the shear wall. To enforce the compatibility condition at the interface, a fictitious beam is introduced. It
would cost a significant amount of analysis time and computer memory to model the floor slab with many subdivided plate elements
in every floor of a high-rise building. In this study, an efficient method is proposed to analyze high-rise box system structures
considering the effects of floor slabs. The proposed method will reduce computational time and memory in the analysis by using
the substructuring technique and matrix condensation. 2002 Elsevier Science Ltd. All rights reserved.
Keywords: Flexural stiffness of slab; Matrix condensation; Substructuring technique; Fictitious stiff beam

1. Introduction
Recently, many high-rise apartment buildings using the
box system have been constructed. The box system, which
consists of reinforced concrete walls and slabs, is very
popular particularly in the Asian region because it allows
a more flexible plan layout without columns. In addition,
the construction with the tunnel forms is very fast, and
there is no need for an extra mortar finish on the walls.
Analysis software such as ETABS, commonly used
for the analysis of high-rise apartment buildings, is
adopting the rigid diaphragm assumption for the slab of
a whole floor [1]. In this case, the flexural stiffness of

* Corresponding author. Tel.: +82-31-290-7554; fax: +82-31-2907570.

E-mail address: dglee@yurim.skku.ac.kr (D.-G. Lee).

slabs is usually ignored in the analysis. This assumption

may be reasonable in a framed structure, which consists
of beams and columns. In box system structures, however, the floor slabs may have a significant influence on
the lateral response of structures. If the flexural stiffness
of slabs in the box system is totally ignored, the lateral
stiffness of the structures may be significantly underestimated [2,3]. The underestimation of the stiffness will
lead to longer natural periods that may result in the underestimation of the seismic loads. In order to predict
accurate seismic response of box system structures, it
may be prudent to include an appropriate amount of
flexural stiffness of slabs. Under large lateral movements
of a building, cracks may occur in slabs along the interface with shear walls. The cracks cause a large reduction
in the flexural stiffness of slabs. For accurate results,
appropriate cracked section properties should be
included based on the actual behavior of a building.

0141-0296/02/$ - see front matter 2002 Elsevier Science Ltd. All rights reserved.
PII: S 0 1 4 1 - 0 2 9 6 ( 0 1 ) 0 0 1 2 6 - 2

614

D.-G. Lee et al. / Engineering Structures 24 (2002) 613623

Fig. 1.

Example structures with a rigid diaphragm. (a) Plan type A, (b) Plan type B, (c) Plan type C.

In order to account for the flexural stiffness of slabs,

it is necessary to use a refined finite element mesh in
the analysis. It may be more efficient to model a shear
wall using one plate element per story to reduce the
number of degrees of freedom. To enforce the compatibility condition at the interface of floor slabs and shear
walls, this study introduces fictitious beams, which interconnect one shear wall element with many sub-divided
slab elements.
If the slab were subdivided into many elements in
every floor of high-rise buildings, it would require a significant amount of computational time and memory,
which may not be practical. In this study, an efficient
analysis method is introduced to reduce the computational time and computer memory for the consideration
of the flexural stiffness of slabs. The proposed method
employs the matrix condensation procedure [4], the substructuring technique [5] and the super element.
Through the proposed method, it will be more
efficient to investigate the seismic response of box system structures with taking into account the effect of the
flexural stiffness of slabs.

that represents each floor by three degrees of freedom.

To include the effects of the floor slabs in the analysis,
the slabs are to be modeled as illustrated in Fig. 2 by
subdividing the slabs into many plate elements. To show
the floor slab effect more clearly, the gross section was
used for the slab stiffness. The cracked section effect of
the slab stiffness will be investigated in Section 5 of
this paper.
2.2. Deformation of floor slabs
Fig. 3 shows the deformed shapes of example structures of plan types A and B having 10 stories, due to
lateral loads. The deformed shape of the framed structure, plan type A, shows a typical shear mode deformation. The floor slabs deform similarly to the surrounding beams as shown in Fig. 4(a), so the slab deformation
is almost not seen in Fig. 3(a). Fig. 3(a) also shows that
the bending deformation of the beams in lower stories
is larger than that of the beams in the higher stories as
expected in framed structures. However, the floor slab
deformation is noticeable in plan type B. The bending

2. Effect of floor slabs on seismic response of highrise buildings

2.1. Example structures
Three different plans as shown in Figs. 1 and 2 were
used to investigate the influence of the flexural stiffness
of slabs. Plan type A is a typical framed structure. Plan
type B is a box system structure with a simplified
arrangement of shear walls, and plan type C is a plan of
typical apartment buildings.
In conventional analyses with commonly used commercial software such as the ETABS or MIDAS/BDS,
the floor slab is usually modeled with a rigid diaphragm

Fig. 2.

Fig. 3. Deformed shape of building structures by lateral loads. (a)

Plan type A, (b) Plan type B.

Example structures with refined slabs. (a) Plan type A, (b) Plan type B, (c) Plan type C.

D.-G. Lee et al. / Engineering Structures 24 (2002) 613623

Fig. 4.

615

Deformation of typical floor slabs due to lateral loads. (a) Plan type A, (b) Plan type B, (c) Plan type C.

deformation of the slab is getting larger in upper stories

of the building, because the lateral displacements in
higher levels are increasing due to the bending mode
deformation of the box system.
The slabs of plan type A are bent mainly in one direction while they remain almost straight in the other direction as shown in Fig. 4(a). The strain energy is mainly
stored in the beams although bending deformation
occurs, which means that the effect of floor slabs on
the lateral response of the framed structure may not be
significant [6]. In the box system structures, shear walls
are connected to the slabs without beams and the distance between shear walls is usually shorter than a typical support span in framed structures. Therefore, more
strain energy is stored in the slab. The slabs of plan type
B are bent mainly in one direction due to the regular
arrangement of shear walls. The slabs of plan type C,
which are bent in both directions as shown in Fig. 4(c),
will store much more strain energy. This phenomenon
will result in a large effect of flexural stiffness of slabs
on the seismic response in box system structures.

3. Seismic response of different building systems

The equivalent static analysis, eigenvalue analysis and
response spectrum analysis were performed with the
framed structures and the box system structures to investigate the effect of floor slabs on seismic response. In
these analyses, two models were used for each plan type.
Model D uses rigid diaphragms (conventional
procedure) not including the flexural stiffness of slabs
as shown in Fig. 1, while model S is using plate elements
to introduce the flexural stiffness of slabs as shown in
Fig. 2. Three plan types were analyzed with 10-story and
20-story structures. In order to emphasize the differences, the gross section of the slab was used for slab
stiffness in this comparison.
3.1. Lateral displacements
Lateral displacements from the equivalent static
analysis in the transverse direction are plotted in Figs.
5 and 6 for 10-story and 20-story structures respectively.
In all cases, the lateral displacements are reduced when
the flexural stiffness of slabs is included in the analysis.
In the framed structures, the effects of the floor slabs are

similar for 10-story and 20-story structures as illustrated

in Figs. 5(a) and 6(a). The effects are more significant,
however, in 20-story box system structures. The displacements of plan type C show the most significant
influence of the floor slabs as expected from the assumption of the strain energy stored in the slabs. The roof
displacement of the 20-story framed structure with plan
type A was reduced by only 14% when the flexural stiffness of slabs is considered. The roof displacements of
boxed system structures were reduced, however, by 85%
and 193% with plan types B and C respectively.
3.2. Natural periods of vibration
Natural periods of vibration for the example structures
are shown in Figs. 7 and 8 for 10- and 20-story structures
respectively to demonstrate the accuracy of the analysis
results by the proposed method. They show that in all
cases the natural period is shorter when the flexural stiffness of the slab is included. The floor slab effects are
more noticeable in taller box system structures. The differences in natural periods are more significant in the
first mode which is the most important mode for the seismic response of a structure. In the comparison between
different plan types, similar observations are made as to
those in the lateral displacements.
3.3. Response spectrum analysis results
The different natural periods result in different seismic
responses of the structures. The design response spectrum from Uniform Building Code 97 (UBC97) was
used in this study. In the design of example structures the
soil type, seismic zone, importance factor and response
modification factor were assumed to be Sa, 2B, 1.0 and
4.5 respectively.
Model D has longer natural periods and thus less spectral accelerations than those of model S, and may be used
as shown in Fig. 9. Therefore, if the flexural stiffness of
the floor slab is ignored, the seismic loads scaled to the
code base shear may be underestimated. Even though
the difference in the periods is small, the difference in
the spectral acceleration becomes large in the shorter
period region, because the slope of the response spectrum is steep in that region. As listed in Table 1, the
base shears of model D are less than those of model S.
Therefore, in order to obtain more accurate results, it is

616

D.-G. Lee et al. / Engineering Structures 24 (2002) 613623

Fig. 7.

Fig. 5.

Displacements of 10-story structures. (a) Plan type A, (b) Plan type B, (c) Plan type C.

Fig. 6.

Displacements of 20-story structures. (a) Plan type A, (b) Plan type B, (c) Plan type C.

Natural periods of vibration for 10-story structures. (a) Plan type A, (b) Plan type B, (c) Plan type C.

important to include the flexural stiffness of slabs

adequately based on the actual behavior of a building.

4. Efficient seismic analysis considering floor slabs

As stated above, it was found that the flexural stiffness
of the slabs may have a significant role in the lateral

response of a box system structure. For the flexural stiffness of the slabs, it is necessary to subdivide a slab into
a fine mesh of plate elements. However, the analytical
model with a fine mesh for each floor of a whole building requires a considerable amount of computational
time and memory. An efficient analysis method is now
proposed with the substructuring and matrix condensation techniques to save time and computer memory
without any deterioration in the accuracy of the results.

D.-G. Lee et al. / Engineering Structures 24 (2002) 613623

Fig. 8.

617

Natural periods of vibration for 20-story structures. (a) Plan type A, (b) Plan type B, (c) Plan type C.

Fig. 9. Difference of acceleration caused by modeling technique in

plan type C.
Table 1
Base shear calculated from response spectrum (unit: tonf)
Plan Type

Model

10-Story
20-Story

19.0
25.8

19.8
26.8

36.8
42.0

38.4
47.8

66.0
78.8

71.6
90.0

4.1. Rigid diaphragm assumption and matrix

condensation
It is desirable to subdivide the shear wall and floor
slab using plate elements as shown in Fig. 10, in order
to get more accurate results in the analysis of the box
system structure. However, a shear wall is generally
modeled in a story with only one element to reduce the
number of degrees of freedom (DOFs). It would cost a
significant amount of computational time and memory
if the slab were subdivided into fine meshes in the analysis of high-rise buildings due to the increase of the number of DOFs. If the rigid diaphragm assumption is
applied, the number of in-plane DOFs in a floor can be
reduced to three, and out-of-plane DOFs can be elimin-

Fig. 10.

Refined model of plan type B.

ated by the matrix condensation procedure. Therefore,

the box system, for which the slab is subdivided into
plate elements, was modeled as a stick having 3 DOFs
per story. The static condensation procedure for the rigid
diaphragm assumption is illustrated in Fig. 11. Before
static condensation the model has 840 DOFs for two
slabs as shown in Fig. 11(a). The DOFs can be reduced
to 426 by the rigid diaphragm assumption, and matrix
condensation will eliminate most of the DOFs, leaving
only 3 DOFs per floor as shown in Fig. 11(c). This
matrix condensation procedure may also need some
computations. However, it is compensated by a significant reduction in computational time and memory
required for the whole structure. Even further reduction
for a large-sized structure may be possible through the
substructuring technique and the super element that will
be introduced below.
4.2. Substructuring technique
Most of the high-rise apartment buildings recently
constructed have identical residential unit plans that are
repeated in most of the floors. Thus, it is very efficient
to apply the substructuring technique in the preparation
of the analytical model. A simplified plan of an apart-

618

D.-G. Lee et al. / Engineering Structures 24 (2002) 613623

Fig. 11. Reduction of DOFs by rigid diaphragm assumption and matrix condensation. (a) Model with unmeshed shear wall, (b) Model applied
rigid diaphragm assumption, (c) Condensed model.

ment floor is shown in Fig. 12(a). Fig. 12(b) shows that

the apartment floor is separated into two residential units
and a stair hall. The separated residential units and stair
hall are replaced by three super elements shown in Fig.
12(c). The super element SE-A can be easily derived
using the mirror image of the super element SE-A. Once
the stiffness and mass matrices for the super element SEA are assembled, those of the super element SE-A can
be easily obtained by rearranging the DOFs and changing the sign of some elements in the stiffness and mass
matrices. Since the residential unit types of a high-rise
apartment building are usually limited to one or two
types in a building, a small number of super elements
can be used repeatedly for the entire building. A substructure for a floor slab in a story can be formed by

assembling the super elements as illustrated in Fig.

12(d). A process of composing the substructure from
every story for the entire building is shown in Fig. 12(e).
The effort for the preparation of the entire building
model can be drastically reduced using the super
elements and substructures. A floor slab of an apartment
floor is subdivided into many plate elements as shown
in Fig. 12(a). These plate elements are then combined
to constitute the equilibrium equations. The equilibrium
equations can be rearranged as shown in Eq. (1) separated the DOFs at the inner area of a super element and
at the boundaries at which the super element is connected to the shear walls:

Sii Sib

Di

Sbi Sbb Db

Ai

Ab

(1)

where subscripts i and b represent the inner area and the

boundary of the element respectively. Eliminating the
DOFs for the inner area by matrix condensation, the equation now represents only the boundary DOFs as follows;
[S bb]DbAb

(2)

where
[S bb][Sbb][Sbi][Sii]1[Sib]
and Ab=Ab[Sbi][Sii]1Ai
The matrix [S bb] is the stiffness matrix for the super
element having nodes that are shared with the shear
walls. Eq. (2) represents the equilibrium equation for the
super element that has DOFs only at the nodes connected
to the shear walls.
Fig. 12. Application of super element and substructure to floor slabs
for plan type C. (a) Plan of a floor slab, (b) Division of a floor slab,
(c) Types of super elements, (d) Substructure for a floor slab, (e)
Assemblage of a substructure.

4.3. Finite element for shear walls

A twelve DOFs plane stress element was proposed by
Lee [7] which was modified from the 10 DOFs element

D.-G. Lee et al. / Engineering Structures 24 (2002) 613623

Fig. 13.

Fig. 14.

619

Plane stress element for shear walls. (a) 8 DOFs plane stress, (b) 12 DOFs plane stress (Lee element).

Typical beam element.

developed by Weaver et al. [8]. The Lee element was

used in this study to get a more accurate response of the
box system structure in lieu of a more general plane
stress element with 8 DOFs, because it has the drilling
degrees of freedom. Fig. 13 shows these two elements.
It is easy to connect a typical beam element (Fig. 14) to
a Lee element because of the presence of the rotational
degrees of freedom. Also, the Lee element can quite
accurately represent the shear stress distribution within
the element. A shape function of an edge of the Lee
element is identical to that of a typical beam element as
shown in Fig. 15. This is the theoretical basis for using a
fictitious beam that is explained in the following section.
4.4. Use of the fictitious stiff beam
In order to include the flexural stiffness of slabs in
the analysis of a box system structure, the slab should be

Fig. 15. Displacement shape function along the boundary of the

Lee element.

subdivided into many plate elements. For a compatible

connection between the slabs and shear walls, the shear
walls should also be subdivided into many plate
elements as illustrated by the refined model shown in
Fig. 16(a). It would, however, be more efficient to model
each shear wall in a story with one element to minimize
the number of nodal points used as in the single element
model, which is shown in Fig. 17(a). In this case, however, the compatibility condition will not be satisfied at
the interface of the slabs and the shear walls, because
most of the nodes at the boundary of the slabs are not
shared with those in the shear walls. The lateral stiffness
of the single element model becomes smaller than that
of the refined model. The stress distribution in the floor
slab for these two models is significantly different from
each other as shown in Figs. 16(b) and 17(b).
Therefore, a fictitious beam is introduced to enforce
the compatibility condition at the interface of the slabs
and the shear walls as illustrated in Fig. 18. A virtual
beam may be added to the boundary of the floor slab
that is connected to the shear wall as shown in Fig.
18(b). Then DOFs in the boundary that are not shared
with the shear wall are eliminated by matrix condensation. Because the beam element deforms in a cubic
curve like the edge of the Lee element, the deformation
of the slab including fictitious beams will be nearly consistent with that of the edge of the shear wall. The surplus stiffness introduced by the fictitious beam can be
eliminated by subtracting the stiffness of the fictitious
beam with nodes only at both ends from the stiffness of
the super element as illustrated in Fig. 18(d).
Fig. 19 illustrates the efficiency of the model with fictitious beams. The number of elements used in the model
is identical to the model in Fig. 17(a), but much less
than that of the model in Fig. 16. The deformed shape
and stress distribution of the model with fictitious beams
(Fig. 19) are, however, similar to those of the model in
Fig. 16, which is considered to be the most accurate.

620

D.-G. Lee et al. / Engineering Structures 24 (2002) 613623

Fig. 16.

Fig. 17.

Refined model for a structure in plan type B. (a) Deformation of a floor slab, (b) Von Mises stress distribution.

Single element model for a structure in plan type B. (a) Deformation of a floor slab, (b) Von Mises stress distribution.

Fig. 18. Use of the fictitious beam at the boundary of the slab and the wall. (a) Boundary area, (b) Add fictitious beam, (c) Condense, (d)
Eliminate fictitious beam.

Fig. 19. Model using fictitious beam of plane type B. (a) Deformation of a floor slab, (b) Von Mises stress distribution.

5. Analysis of example structures

Plan type C was used to demonstrate the accuracy and
efficiency of the proposed method. A computer code
MIDAS [9] was used for the analysis of example structures. With the same plan, three different structures were
analyzed: 10-story regular structure (example structure

1), 20-story regular structure (example structure 2) and

20-story irregular structure (example structure 3), as
shown in Fig. 20. For each structure, the slabs were
modeled using three different methods: the rigid diaphragm method (model D), the refined mesh method
(model S), which is considered to be the most accurate
since the slabs and shear walls were modeled with

D.-G. Lee et al. / Engineering Structures 24 (2002) 613623

Fig. 20.

Various structures of plan type C. (a) Example structure 1, (b) Example structure 2, (c) Example structure 3.

refined plate elements, and the proposed analysis method

(model P), which employs the substructuring techniques
and the fictitious beam. In model S and model P, the
analysis was done with 100% of the flexural stiffness
from the slab gross section (1.0EIg) and with 30% of
that (0.3EIg) to investigate the cracked section effect.
Lateral displacements and natural periods of three
example structures are plotted in Figs. 21 and 22. The
lateral displacements of model D were significantly
overestimated with the longer natural period compared
to model S. The results of model P were very close to
those of model S.
The computation time and the number of DOFs used
for the analyses are listed in Table 2. For inclusion of
the flexural stiffness of floor slabs, the analysis for model
S required more than 30 times the computational time
for model D. Model P, however, provided very reasonable accuracy compared to model S with significantly
less time. The figures also clearly show the cracked section effects of three structures. The results for model P
compared to those for model S demonstrated that there
is no significant difference in the accuracy even with the
cracked section effects. Model P can be effectively used

Fig. 21.

621

for the investigation of floor slab effects with various

levels of cracked section properties, depending on the
actual behavior of a building.
The deformations for zone A in the roof slab shown
in Fig. 1(c), in the case of the example structure 2, are
illustrated in Fig. 23. Table 3 demonstrates that the slab
deformations are very different according to the flexural
stiffness of the floor slab. The values of /L in model
S with 0.3EIg and 1.0EIg are only 22% and 8% respectively of that in model D, which indicates that the amount
of slab flexural stiffness is very important in the analysis.
If the flexural stiffness of the floor slabs was totally
ignored in the analysis (as in model D) and the reinforcement was provided accordingly, excessive slab cracks
might cause a large deformation of the slab, which may
warrant a further detailed investigation on the serviceability of the floor slabs under maximum seismic or
wind load.
6. Summaries and conclusions
The effects of floor slabs on seismic response of highrise apartment building structures were investigated in

Comparison of lateral displacements. (a) Example structure 1, (b) Example structure 2, (c) Example structure 3.

622

D.-G. Lee et al. / Engineering Structures 24 (2002) 613623

Fig. 22.

Comparison of natural periods. (a) Example structure 1, (b) Example structure 2, (c) Example structure 3.

Table 2
Comparison of DOFs and computational time
Example structure

Model

DOFs

Computational time (s)

Assemble M & K

Example structure 1

Example structure 2

Example structure 3

Model
Model
Model
Model
Model
Model
Model
Model
Model

D
S
P
D
S
P
D
S
P

30
15840
30
60
31680
60
60
25080
60

3
13
87
5
28
158
4
25
103

Equation solving
4
377
4
9
871
8
5
510
5

Eigenvalue analysis
62
1847
63
131
3828
126
86
2534
87

Total
69
2237
154
145
4727
292
95
3069
195

this study. In order to include the flexural stiffness of

slabs, an efficient analysis method was proposed with
the substructuring technique, matrix condensation, and
fictitious beams. The major observations and findings are
summarised as follows:

Fig. 23.

Deformation of roof slab in Fig. 1(c) zone A.

Table 3
Slab deformation according to flexural stiffness of slab

L (cm)
(cm)

Model D
(0EIg)

Model S
(0.3EIg)

Model S
(1.0EIg)

100
0.5869

100
0.1292

100
0.0472

1. In a box system structure, the effect of the flexural

stiffness of slabs on the lateral response of the structure is relatively significant, especially in taller buildings and in buildings with irregular arrangements of
shear walls. If the flexural stiffness of the slabs is
totally ignored, the lateral displacements may be overestimated and the seismic loads per the building code
base shear may be significantly underestimated. It is
recommended to that the flexural stiffness of slabs is
adequately included in the analysis of box system
structures.
2. It may be important to determine what amount of
flexural stiffness of slabs should be included in the
analysis of a box system structure, since the amount
depends on the lateral response of a building. A
further study may focus on finding the nature of the

D.-G. Lee et al. / Engineering Structures 24 (2002) 613623

deformed slabs in a boxed system structure under lateral loads. In conjunction with the flexural stiffness
of slabs, it may be necessary to consider the out-ofplane flexural stiffness of the shear wall, which might
cause a considerable bending moment requiring
additional reinforcement in the wall. This additional
moment in the shear wall may also be investigated in
a further study.
3. The slab should be subdivided into many plate
elements to include the flexural stiffness of slabs,
while a shear wall may be more efficiently modeled
with only one element in a story. In this modeling
method, the compatibility condition is not satisfied at
the interface of the slabs and the shear walls. To
enforce the compatibility, a fictitious beam was introduced in this study. The analysis result of the proposed model with the fictitious beam was similar to
that of the refined mesh model, which is considered
to be the most accurate solution.
4. The analysis method using matrix condensation and
the substructuring technique was also proposed to
reduce the computational time and computer memory.
The proposed analysis method could provide accurate
results with only 6% of the computational time compared to the refined mesh model.
Acknowledgements
The Brain Korea 21 Project supported this work. This
work was also partially supported by the Korea Science

623

and Engineering Foundation (KOSEF) through the

Korea Earthquake Engineering Research Center
(KEERC) at Seoul National University (SNU).

References
[1] Wilson EL, Habibullah A. ETABS three dimensional analysis
of building systems users manual. Berkeley (CA): Computers &
Structures Inc, 1995.
[2] Lee DG, Kim HS. The effect of the floor slabs on the seismic
response of multi-story building structures. Proceedings of
APSEC2000, 2000 Sep; Malaysia.
[3] Lee DG, Kim HS. An efficient model for seismic analysis of
high-rise building structures with the effects of floor slabs. Proceedings of SEEBUS2000, 2000 Oct; Japan.
[4] Weaver W Jr, Johnson PR. Structural dynamics by finite
elements. Prentice Hall, 1987.
[5] Petersson H, Popov EP. Substructuring and equation system solutions in finite element analysis. Computers and Structures
1977;7:197206.
[6] Ahn SK, Lee DG. Efficient seismic analysis of building structures
with eccentric beams. Proceedings of 6th APCS, 2000 Oct;
Seoul, Korea.
[7] Lee DG. An efficient element for analysis of frames with shear
walls. ICES88, 1988 Apr; Atlanta.
[8] Weaver W Jr., Lee DG, Derbalian G. Finite element for shear
walls in multistory frames. Journal of the Structural Division
ASCE 1981;107:13659.
[9] Lee HW, Park YG. MIDAS/GENw the most intelligent
design & analysis system. MIDAS Information Technology Co.,
Ltd, 2000.