Applied Mathematical Modelling 36 (2012) 4918–4930

Contents lists available at SciVerse ScienceDirect

Applied Mathematical Modelling

journal homepage: www.elsevier.com/locate/apm

A simple mathematical model for free vibration analysis of combined

system consisting of framed tube, shear core, belt truss and outrigger
system with geometrical discontinuities
Reza Kamgar a,⇑, Mohammad Mehdi Saadatpour b,⇑
a
Department of Civil Engineering, University of Kerman, P.O. Box 76135-1893, Kerman, Iran
b
Department of Civil Engineering, Isfahan University of Technology, P.O. Box 84156-83111, Isfahan, Iran

a r t i c l e i n f o a b s t r a c t

Article history: Dynamic analysis of beam-like structure is significantly important in modeling real cases
Received 29 May 2011 such as tall buildings, aircraft wings, spacecraft antennas and many other applications. This
Received in revised form 10 December 2011 paper tries to determine the first natural frequency of tall buildings including framed tube,
Accepted 15 December 2011
shear core, belt truss and outrigger system with multiple jumped discontinuities in the
Available online 24 December 2011
cross section of framed tube and shear core. In this regard, the entire length of the tall
building is partitioned into uniform segments between each two successive discontinuity
Keywords:
points. The effect of belt truss and outrigger system is modeled as a concentrated rotational
Free vibration
Framed tube
spring applied at the belt truss and outrigger system location. Many cantilevered tall struc-
Shear core tures can be treated as cantilever bars with multiple jumped discontinuities in the cross
Belt truss section for the analysis of their free vibration. In this paper, the continuous approach
Outrigger system was accepted and by using the Hamilton’s variational principle, the general form of govern-
Geometrical discontinuity ing equation for free vibration of tall building can be obtained. By applying the separation
of variable method on time and space, the governing Partial Differential Equation (PDE) of
motion is reduced to an Ordinary Differential Equation (ODE) with one variable coefficient
while the other coefficients are constant based on the assumption that the transverse dis-
placement is a harmonic vibration. To find exact solution of ODE, we must have exact dis-
tribution of EI(x), AG(x), N(x) and m(x) in the height of the structure. Some of these
parameters such as EI(x), AG(x) and m(x), are constant throughout the height of each seg-
ment. These parameters can be expressed exactly by using of multi criteria function, while
N(x) is variable in the height of each segment. Therefore, the ODE by using the method of
variable separation and partitioned method can be expressed for each segment. We must
apply the continuity conditions in conjunction with different segments for obtaining
unique mode shape for mentioned system. Tall building characteristics matrix can be
derived based on the boundary conditions and the continuity conditions applied at the par-
titioned points. This matrix is particularly used to find combined system first natural fre-
quency and mode shape. Three numerical examples with different stepped discontinuities
in their cross sections are studied to demonstrate the reliability of this method. The results
of the proposed mathematical model give a good understanding of the structure’s dynamic
characteristics; it is easy to use, yet reasonably accurate and suitable for quick evaluations
during the preliminary design stages which require less time.
Ó 2011 Elsevier Inc. All rights reserved.

⇑ Corresponding authors. Tel./fax: +98 311 3912700.

E-mail addresses: reza.kamgar.uk1387@gmail.com (R. Kamgar), mmehdi@cc.iut.ac.ir (M.M. Saadatpour).

0307-904X/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.apm.2011.12.029
 R. Kamgar, M.M. Saadatpour / Applied Mathematical Modelling 36 (2012) 4918–4930 4919

1. Introduction

Free vibration analysis plays an important role in structural design of tall buildings, especially in the first mode shape
because it is the dominant shape in response to wind and earthquake-induced vibrations in tall buildings. Therefore, it is
important to investigate the methods that are used to compute natural frequencies and mode shapes for tall buildings. In
this respect, numerous studies can be found in the literature on the transverse vibration of tall buildings with uniform
cross-section along the height of the building. However, in many real applications, the investigation of the first natural fre-
quency of tall buildings with non-uniform cross-section may provide a realistic distribution of mass and stiffness desirable
for accurate structural analysis. Particularly, for structures with abrupt changes in cross section, stiffness and geometrical
discontinuities affect the behavior of structure which cannot be neglected. Many researchers in structural engineering have
developed a method to obtain accurate theoretical results for the free vibration analysis of tall buildings in the past decades.
Koplow et al. [1] presented an analytical solution for the dynamic response of a free–free discontinuous beam with a single
step change and an aligned neutral axis was considered. They considered the case of free–free boundary conditions to obtain
direct frequency response functions due to harmonic force or couple excitation at both ends location. Naguleswaran [2] pre-
sented an analytical method to calculate the frequencies of beams with up to three step changes in cross-section. They con-
sidered the combinations of the classical clamped, pinned, sliding, free, ‘general’ and ‘degenerate’ types of elastic end
supports.
Lee et al. [3] presented the governing equations of wall-frame structures with outriggers through the continuum ap-
proach and the whole structure was idealized as a shear-flexural cantilever with rotational springs. The effect of shear defor-
mation and flexural deformation of the wall-frame and outrigger trusses were considered and incorporated in the
formulation of the governing equations. Stafford Smith and Salim [4] presented formulae that were developed for estimating
the optimum levels of outriggers to minimize the drift in outrigger-braced buildings. They presented the analyses and for-
mulae for outrigger structures in which the core and columns were uniform through their height and the outriggers had the
same flexural stiffness. Rutenberg and Tal [5] presented the results of an investigation on drift reduction in uniform and non-
uniform belted structures with rigid outriggers under several lateral load distributions, which were likely to be encountered
in practice. Design aids in the form of graphical presentations of the somewhat complex solutions were provided. Stewart
Townend and Andrew [6] presented an approximation dynamics method by using Hamilton’s principle about including
applications to non-conservative and conservative systems. They had shown that this method was suitable for both multi
degree of freedom and constrained systems. Kuang and Ng [7] presented free vibration analysis of asymmetric-plan frame
structures. They emphasized on analysis of lateral-torsional vibration of the structures, where lateral shear vibrations in two
orthogonal directions were coupled with St. Venant torsion vibration. The governing equation of coupled vibration of the
problem was derived by them. Also the corresponding eigenvalue equation was derived. A theoretical method of solution
was proposed to solve the eigenvalue problem and a general solution was given to determine the natural frequencies and
associated mode shapes of the structure. A simplified analytical method for outrigger structure has been presented earlier
by Stafford Smith and Coull [8] and Taranath [9]. Hoenderkamp and Bakker [10] presented a graphical method of analysis
for the preliminary design of tall building structures comprising braced frames with outrigger trusses subjected to horizontal
loading. They also presented a simple procedure for obtaining the optimum location of the outrigger up the height of the
structure and a rapid assessment of the impact of the outrigger on the behavior of the high-rise structure. Geourgoussis
[11] presented an approximate analysis for calculating the deflections of individual cantilever bents, under lateral loading,
as a sum of deflections of two complementary subsystems: a flexural and a shear-flexural subsystem. The analysis accounted
for axial deformations in the vertical members of bents. Geourgoussis, [12] presented a simple mathematical model for
assessing periods of vibration and mode shapes of common cantilever bents used in concrete structures, such as shear walls,
coupled walls, rigid frames and wall-frame assemblies. He used Dunkerley’s formula for calculating natural frequencies of
mentioned structures and considered the effect of column axial shortenings in the analysis of structural bents. Teresa
[13] determined the modes and the frequency equation of Euler–Bernoulli beams with discontinuous properties in the trans-
versal section by using a dynamical basis, which was generated by a fundamental solution of a fourth-order differential
equation. They considered beams subjected to general boundary conditions and proposed a simple method to study the
problem of the free vibration of a spring-restrained free-supported beam.
In this paper, a simple mathematical model is presented to calculate the first natural frequency of combined system
including framed tube, shear core; outrigger and belt truss system. Framed tube system consists of closely spaced exterior
columns along the periphery interconnected by deep spandrel beams at each floor. This produces a system of rigidly con-
nected joint orthogonal frame panels forming a rectangular tube, which acts as a cantilevered hollow box Coull and Ahmad
[14], Coull and Bose [15], and Connor and Pouangare [16]. The effect of belt truss and outrigger system is modeled as a con-
centrated rotational spring located at the belt truss and outrigger system location. Here by adopting the Hamilton’s varia-
tional principle; the following items are derived: Partial Differential Equation of the structural vibration, boundary
displacements (kinematic boundary conditions), boundary forces (natural boundary conditions) and eigenvalue solution
form. By selecting partitioned method along the height of the structure and by using the assumption of the harmonic motion,
the PDE is reduced to ODE with one variable coefficient while the other coefficients are constant. This equation can be solved
exactly by selecting power-series solution method. By applying the boundary conditions and continuity conditions, the
eigenvalue problem for finding the first natural frequency of tall building is obtained. In order to illustrate the efficiency
4920 R. Kamgar, M.M. Saadatpour / Applied Mathematical Modelling 36 (2012) 4918–4930

and accuracy of the proposed method, three numerical examples have been carried out by the proposed method and SAP
2000 software. The numerical examples show that the approximate values of the first natural frequency of tall buildings ob-
tained by the proposed method are within the acceptable range of engineering practice rather than the more accurate finite
element method results and therefore the proposed method may be used to estimate the natural frequency at the prelimin-
ary stage of the design.

2. Program formulation and solution

In this section by adopting the following assumptions and using the Hamilton’s variational principle, the vibration PDE of
the combined system including framed tube, shear core, belt truss and outrigger system is derived by the following
assumptions:
The floor slabs of the system are not deformable in their planes and have no motion perpendicular to their planes.
The effect of the belt truss and outrigger system is considered as a rotational spring with constant rotational stiffness,
which acts on the position of the belt truss and outrigger system.
Spacing of columns and beams are constant throughout the building’s height and the dimensions of all beams and col-
umns are the same in each segment.
Shear core and columns are fully fixed at the base.
Member’s connections of the outrigger system are assumed to be rigid and connections of members of the belt truss are
assumed to be pinned.
The material of structure is linearly elastic, homogeneous and obedient to the Hook’s law.
The structure is assumed symmetric in plan of all stories and therefore cannot twist.
The thickness of the shear core and the dimensions of the columns and beams of the framed tube structure change in a
stepwise in the height of the structure.
The dimension of members of the belt truss and outrigger system are constant and do not vary in the height of the
structure.
The mass of the outrigger and belt truss system is considered as a lumped mass (M) in the segment to which these sys-
tems belong.
With above assumptions, the structure can be modeled as a beam with a box variable cross section in height (see Fig. 1)
and by using the Hamilton’s principle, the differential equation of vibration of combined system can be obtained. The Ham-
ilton’s Variational Principle (HVP), considering fundamental law of dynamic, encompasses Newton’s equations of motion,
Lagrange’s equations for structure dynamics, D’Alembert’s principle and the Principle of Virtual Work (PVW). Vibration of
structures can be conveniently formulated in terms of HVP. HVP is energy functional based, the diverse areas of structure
dynamics, numerical solutions of PDE, finite element methods, and functional analysis can all be linked in a single develop-
ment. Structure energy is expressed in terms of the functional Lagrangian energy, and HVP requires this functional energy
have a stationary value. The term functional is used to denote a general expression for a continuous function of the domain V
of the structure in space and time, Piersol and Paez [17].
Consider a combined system of framed tube, shear core, belt truss and outrigger system as a continuous beam with flex-
ural stiffness EI(x) shear stiffness AG(x) mass per unit height m(x) dynamic displacement w(x, t) and total height L (see Fig. 1).
The effect of the belt truss and outrigger system is considered as a rotational spring at the location of the belt truss and out-
rigger system.
Structure is defined over the closed domains 0 6 x 6 L, where x is the spatial position of any material point of the system
and t is the time of vibration of any material point of the system as shown in Fig. 1. In Fig. 1, bold line shows the variation of
thickness along the height of the structure.
In Fig. 1, the parameters E, G and t are modulus of elasticity, shear modulus and total equal thickness of the shear core and
framed tube in both the flange and the web panel.

The kinetic energy of the dynamic system under consideration is:

Z L
1 1
Kðx; tÞ ¼ mðxÞ½w: ðx; tÞ2 dx þ M½w: ða; tÞ2 ð1Þ
2 0 2
and the elastic energy is:
Z L
1 1
Uðx; tÞ ¼ fEIðxÞ½w00 ðx; tÞ2 þ AGðxÞ½w0 ðx; tÞ2 þ NðxÞ½w0 ðx; tÞ2 gdx þ K r ½w0 ða; tÞ2 ; ð2Þ
2 0 2
where Kr is the equivalent stiffness of the rotational spring including the effect of the belt truss and outrigger system on the
framed tube acting at x = a, M is the mass of the belt truss and outrigger system, N(x) is the axial force, primes and dots on w
denote partial derivatives with respect to x and t, respectively.
Since we intend to derive the governing equations of the system by Hamilton’s principle, it is necessary first to earn the
following energy expression.
Bðx; tÞ ¼ Kðx; tÞ  Uðx; tÞ: ð3Þ
 R. Kamgar, M.M. Saadatpour / Applied Mathematical Modelling 36 (2012) 4918–4930 4921

Fig. 1. Approximate models of tall buildings: (a) 3D model of the combined system. (b) Elevation of combined system.

The action or principle function of dynamics A, can be expressed as the time integral of B between two times t1 and t2, Piersol
and Paez [17].
Z t2 Z t2
A¼ Bðx; tÞdt ¼ ½Kðx; tÞ  Uðx; tÞdt: ð4Þ
t1 t1

HVP states that A has a stationary value expressed as dA = 0 where d is termed as the variation operator. The d operator over a
structure domain V implies that complete description of the structure shape requires an infinite number of degree of free-
dom- one degree of freedom for each shape. HVP requires that the action defined in Eq. (4) has a minimum or stationary
value over all possible structure variations, that is
Z t2 Z t2 Z t2
dA ¼ d Bðx; tÞdt ¼ dBðx; tÞdt ¼ d½Kðx; tÞ  Uðx; tÞdt ¼ 0: ð5Þ
t1 t1 t1

This expression of HVP in terms of dA utilizing the d operator properties and integration by parts provides
4922 R. Kamgar, M.M. Saadatpour / Applied Mathematical Modelling 36 (2012) 4918–4930

1. The differential equation(s) of motion termed the Lagrange’s equation(s).

2. The boundary displacements (kinematic boundary conditions).
3. The boundary forces (natural boundary conditions).
4. The eigenvalue solution form.

By substituting Eqs. (1) and (2) into Eq. (5), the following equation is obtained:
Z t2 Z L Z t2
dA ¼ ½mðxÞw: ðx; tÞdw:  NðxÞw0 ðx; tÞdw0  EIðxÞw0 ðx; tÞdw0  AGðxÞw0 ðx; tÞdw0 dxdt  K r w0 ða; tÞdw0 dt
t1 0 t1
Z t2
þ M½w: ða; tÞdw: dt: ð6Þ
t1

Integration by parts from Eq. (6) with respect to both time and space, gives:
Z Z Z t2
t2 L
@2 @ @
dA ¼  ½mðxÞw ðx; tÞ þ
ðEIðxÞw 0
ðx; tÞÞ  ðAGðxÞw0
ðx; tÞÞ  ðNðxÞw 0
ðx; tÞÞdwdxdt  ½K r w0 ða; tÞ
t1 0 @x2 @x @x t1
Z t2   Z t2
@
þ EIðxÞw0 ðx; tÞdw0 jL0 dt þ ½EIðxÞw0 ðx; tÞ  AGðxÞw0 ðx; tÞ  NðxÞw0 ðx; tÞ dwjL0 dt  M½w ða; tÞdwdt: ð7Þ
t1 @x t1

The equation of motion and related boundary conditions can be obtained as follows:

@2 @ @
mðxÞw ðx; tÞ þ ðEIðxÞw0 ðx; tÞÞ  ðAGðxÞw0 ðx; tÞÞ  ðNðxÞw0 ðx; tÞÞ ¼ 0 for 0 6 x 6 L; 06t ð8Þ
@x2 @x @x
and
@
@x
½EIðxÞw0 ðx; tÞ  AGðxÞw0 ðx; tÞ  NðxÞw0 ðx; tÞ  Mwða;
€ tÞ ¼ 0 at x ¼ L
K r w0 ða; tÞ þ EIðxÞw0 ðx; tÞ ¼ 0 at x ¼ L
; ð9Þ
wðx; tÞ ¼ 0 at x ¼ 0
w0 ðx; tÞ ¼ 0 at x ¼ 0
where the first two of Eq. (9) are boundary forces at the top of the structure and the last two of Eq. (9) are boundary displace-
ments at the base of the structure.

3. Solution

Using the method of variables separation, we assume:

wðx; tÞ ¼ /ðxÞsinxt; ð10Þ
where x is the natural frequency and /(x) is the shape function of the system.
By substituting Eq. (10) into Eq. (8) one obtains:
2
" 2
#    
d d /ðxÞ d d/ðxÞ d d/ðxÞ
2
EIðxÞ 2
 AGðxÞ  NðxÞ ¼ mðxÞx2 /ðxÞ: ð11aÞ
dx dx dx dx dx dx

For a tall building with parametric discontinuities (e.g., jump in the flexural stiffness, shear stiffness and mass distribution)
Eq. (11a) cannot be solved using conventional approaches. An alternative method is to partition the tall building into uniform
segments between any two successive stepped points and apply the continuity conditions at these points. Fig. 1 illustrates a
tall building with N jumped discontinuities in its spatial height. By considering Fig. 1, some of these parameters such as EI(x),
AG(x) and m(x), are constant throughout the height of each segment. These parameters can be expressed exactly by using of
multi criteria function as follows, while N(x) is variable in the height of each segment.
8 9 8 9 8 9
> EI1 0 6 x  x1 > > AG1 0 6 x  x1 > > m1 0 6 x  x1 >
>
> >
> >
> >
> >
> >
>
>
> EI2 x1 6 x  x2 >> >
> AG2 x1 6 x  x2 >> >
> m2 x1 6 x  x2 >>
>
> >
> >
> >
> >
> >
>
>
< EI x 6 x  x > = >
< AG x 6 x  x > = >
< m x 6xx > =
3 2 3 3 2 3 3 2 3
EIðxÞ ¼ ; AGðxÞ ¼ ; mðxÞ ¼ ; n ¼ 1; 2; 3; . . . ; N :
>
> : >
> >
> : >
> >
> : >
>
>
> >
> >
> >
> >
> >
>
>
> : >
> >
> : >
> >
> : >
>
>
> >
> >
> >
> >
> >
>
: ; : ; : ;
EIN xN 6 x  xN1 AGN xN 6 x  xN1 mN xN 6 x  xN1
ð11bÞ
Eq. (11a) for segment n can be presented as follows:
 R. Kamgar, M.M. Saadatpour / Applied Mathematical Modelling 36 (2012) 4918–4930 4923

4 2
 
ðEIÞn d dx
/n ðxÞ
4  ðAGÞn d dx
/n ðxÞ
2
d
 dx Nn ðxÞ d/dx
n ðxÞ
¼ mn x2 /n ðxÞ xn1 6 x 6 xn ; n ¼ 1; 2; 3; . . . ; N x0 ¼ 0 ; ð12Þ

where un(x), (EI)n, (AG)n and mn are the mode shape, flexural stiffness, shear stiffness and mass per unit height of the building
for the nth segment respectively. Eq. (12) is an ODE with one variable coefficient which is in the third term of Eq. (12). This
can describe the distribution of axial force in the height of each segment of about mentioned system. Let,
PN
Nr 2
n 2 4
e2n ¼ cðEIÞ
n An 3
2
a2n ¼ ðAGÞ l
ðEIÞn n
b4n ¼ xðEIÞmn ln
n
d2n ¼ r¼nþ1
ðEIÞn
ln l
n n

ln ¼ xn  xn1 f ¼ xxlnn1 xn1 6 x 6 xn 06f61

; ð13Þ
x0 ¼ 0 h2n 2
¼a þ
n d2n 2
þe
n
P
N
Ni ðfÞ ¼ Nr þ ci Ai li ð1  fÞ Nr ¼ cr Ar lr þ M beamr þ Mslabr i ¼ 1; 2; . . . ; N
r¼iþ1

where ci and Ai are weight per unit volume of materials and total equivalent area of tall building in segment i, respectively.
Mbeamr and M slabr are mass of beams and slab in r story, respectively.
The Eq. (12) can be rewritten in a non-dimensional form for each segment in a local coordinate f as follows:
4 2
d /n ðfÞ d /n ðfÞ d/n ðfÞ
4
 ½h2n  fe2n  þ e2n  b4n /n ðfÞ ¼ 0: ð14Þ
df df2 df
The Eq. (14) can be solved exactly by selecting the power-series solution method for each segment as follows:
P
1
/n ðfÞ ¼ ak;n fk
k¼0

P
1
/0n ðfÞ ¼ kak;n fk1
k¼1

P
1
/00n ðfÞ ¼ kðk  1Þak;n fk2
k¼2

P
1
/000
n ðfÞ ¼ kðk  1Þðk  2Þak;n fk3 ; ð15Þ
k¼3

P
1
/0000
n ðfÞ ¼ kðk  1Þðk  2Þðk  3Þak;n fk4
k¼4

P
1
NN ðfÞ ¼ bk;n ð1  fÞk ¼ b0;n þ b1;n ð1  fÞ
k¼0

P
N P
1
Ni ðfÞ ¼ Nr jf¼0 þ bk;i ð1  fÞk ; i ¼ 1; 2; . . . ; N
r¼iþ1 k¼0

where ak,n and bk,n are coefficients, the k index shows the required number for convergence of power series and n index
shows segment number of tall buildings, respectively. The Eq. (14) by shifting the indices can be expressed in power-series
form as follows:
X
1
½kðk  1Þðk  2Þðk  3Þak;n  h2n ðk  2Þðk  3Þak;n2 þ e2n ðk  3Þ2 ak;n3  b4n ak;n4 fk4 þ 24a4;n  2h2n a2;n þ e2n a1;n
k¼5

 b4n a0;n
¼ 0: ð16Þ
By using the property of identity, for k = 0, 1, 2, 3 and 4, the coefficients can be obtained as:
a0;n
a1;n
a2;n ð17Þ
a3;n
2h2n a2;n þb4n a0;n e2n a1;n
a4;n ¼ 24

and the recursion relationships for any k 6 5 can be determined as:

h2n ðk  2Þðk  3Þaðk2;nÞ  e2n ðk  3Þ2 aðk3;nÞ þ b4n aðk4;nÞ

ak;n ¼ : ð18Þ
kðk  1Þðk  2Þðk  3Þ
The power-series solution of mode shape functions can be obtained from solving Eqs. (17) and (18) as follows:
4924 R. Kamgar, M.M. Saadatpour / Applied Mathematical Modelling 36 (2012) 4918–4930

/n ðfÞ ¼ a0;n /n;1 ðf; kÞ þ a1;n /n;2 ðf; kÞ þ a2;n /n;3 ðf; kÞ þ a3;n /n;4 ðf; kÞ
x 2 4 2 2 6 e2n b2n x7
/n;1 ðf; kÞ ¼ 1 þ bn24 þ bn h
720
n x
 1260
þ 
 2 2
e2n x4 n bn b2n x5
e4n e2n h2n x6
/n;2 ðf; kÞ ¼ x  24
þ h5040
þ þ 1260
120
x7 þ   
 720 ð19Þ
2 4 2 5
 4 2 2 2 7
x e x þbn e hn x
/n;3 ðf; kÞ ¼ x2 þ hn12  n30 þ hn360 x6  n 420 þ 
2 5 2 x6
 4 2

h x e h þb
/n;4 ðf; kÞ ¼ x3 þ n20  n40 þ n840n x7 þ   

in which a0,n, a1,n, a2,n and a3,n are the coefficients which can be determined by applying the boundary and continuity con-
ditions. In this formula n shows the segment number of the tall building. The continuity conditions for displacement, slope,
bending moment, axial force and shear force at point of discontinuity are given by:

/n jf¼1 ¼ /nþ1 jf¼0

d/nþ1
1
ln
 d/ n
df f¼1
1
j ¼ lnþ1  df
jf¼0
2 2
ðEIÞnþ1
ðEIÞn
l2n
 ddf/2n jf¼1 ¼ l2nþ1
 d df
/nþ1
2 jf¼0 : ð20Þ
ðEIÞn d3 / n ðEIÞnþ1 d3 /nþ1
l3n
 j
df3 f¼1
¼ l3nþ1
 df3
jf¼0
Nn ðfÞjf¼1 ¼ Nnþ1 ðfÞjf¼0 ) b0n ¼ 0

For Jth and J + 1th segments, which the belt truss and outrigger system have been located in both of them, the following con-
tinuity conditions are valid:
/j jf¼1 ¼ /jþ1 jf¼0
1 d/j 1 d/jþ1
lj
 j
df f¼1
¼ ljþ1  j
df f¼0

ðEIÞj d2 / j ðEIÞjþ1 d2 /jþ1

l2j
 df2
jf¼fj  l2jþ1
 df2
jf¼fðjþ1Þ ¼ M x2 /j jf¼1
t b

ðEIÞj d3 / j ðEIÞjþ1 d3 /jþ1 2 d/j

l3j
 df3
jf¼fj  l3jþ1
 df3
jf¼fðjþ1Þ ¼ J out xlj j
df f¼1 ð21Þ
t b

xj ð2hÞxj1 ð2hÞ
fjt ¼ xj xj1
fðjþ1Þb ¼ xjþ1 xj
x0 ¼ 0
Nj ðfÞjf¼1 ¼ Njþ1 ðfÞjf¼0
n0
P
J out ¼ mi r 2i ;
i¼0

where Jout, M, ri, n0 and mi are the rotary mass moment of inertia, total mass of the belt truss and outrigger system, distance of
the ith element of the belt truss and outrigger system from the x-axis, the total number of elements of the belt truss and
outrigger system and mass of ith element of the belt truss and outrigger system, respectively.
The boundary conditions of Eq. (9) can be rewritten as follows:
/1 jf¼0 ¼ 0
d/1
j
df f¼0
¼0
2 d/ : ð22Þ
ðEIÞN ld2 df
/N
2 jf¼1 þ K r l dfj jf¼1 ¼ 0
N j

d3 /N 2 3
j
dn3 f¼1
 a2N d/ N
j þM
dn f¼1
x l /j
ðEIÞ N j f¼1
¼0
N

The characteristics matrix of the system can be formed by applying Eqs. (20)–(22) into Eq. (19) at each point of discontinuity
as well as at the boundaries. It is remarked that bn’s are functions of tall building natural frequency with an explicit expres-
sion given in Eq. (13). Therefore, the characteristics matrix becomes only function of a single parameter b. The characteristics
equation is then given by:
H5N15N1 C5N11 ¼ 05N11 ; ð23Þ
where H = H (b) is the characteristics matrix and C is the characteristics vector of the system:

C ¼ ½a0;1 a1;1 a2;1 a3;1 b0;1 a0;2 a1;2 a2;2 a3;2 b0;2 . . . a0;N a1;N a2;N a3;N b0;N1 T15N1 : ð24Þ
In order to obtain a non-trivial solution for Eq. (23) to find the first natural frequency and the associated mode shape, the
determinant of matrix H must be zero.
det½HðxÞ ¼ 0: ð25Þ
 R. Kamgar, M.M. Saadatpour / Applied Mathematical Modelling 36 (2012) 4918–4930 4925

Since this matrix is a function of only parameter x e (0, 1), its determinant can be numerically evaluated for its zero values.
The values of x, which satisfy Eq. (25), lead to the calculation of the natural frequencies of the buildings.
Control of top drift and base overturning moment in the core of a tall structure subjected to lateral loads is the main con-
cern in design of tall buildings. There are many structural forms such as rigid frame, braced frame and shear-walled frame,
frame-tube, braced-tube, bundled-tube and outrigger systems that can be used to enhance the lateral resistance in tall build-
ings [18]. Belt truss system restrains the bending of the core by introducing a point of inflection in its deflection curve. This
reversal in curvature reduces the bending movement above the belt truss and outrigger system. The belt truss functions as
horizontal fascia stiffeners and engages the exterior columns, which are not directly connected to the outrigger trusses.
Placement of a rigid truss at the top of the building eliminates differential movement between interior and exterior columns
by providing compressive restraint for exterior columns in expansion and tension restraint when columns are in compres-
sion [9].
In order to determine the stiffness of belt truss and outrigger system, the work that are done by Stafford Smith and Coull
[8] and Lee et al. [3] can be utilized. It has pointed out by Lee et al. [3] that magnitude of the reductions depends on the flex-
ural rigidities of the core, the outriggers, and the columns acting axially around the core’s centroid. The reductions depend
also on the locations of the outriggers up the height of the core [4].
Equivalent stiffness of rotational linear spring Kr can be calculated by Lee et al. [3].

K r ¼ 1=h; ð26Þ
where h denotes the total rotation in the outrigger and belt truss system due to the restraining moment, and can be obtained
by splitting up the rotation as:

h ¼ ha þ hb þ hs : ð27Þ
First, the restraining forces in the exterior columns will cause rotation of the outrigger resulting from the axial lengthening
and shortening of the columns. The outrigger rotation ha due to the resulting restraining moment can then be defined as the
column change in length divided by the length of the outrigger (d):
2
ha ¼ ð2aÞ=ðd AEÞ ð28Þ
in which d and AE are the distance between center to center of exterior columns and the axial stiffness of the exterior col-
umns, respectively.
The flexural deformation of outrigger due to the action of the column force will cause additional drifts between adjacent
floors. The resulting rotation hb is given by:

hb ¼ ðdÞ=ð12EIoe Þ; ð29Þ
where EIoe is the effective flexural stiffness of the outrigger, modeled as though its length extended from the column to the
centroid of the core. EIoe can be obtained from the outrigger’s actual flexural rigidity EIou by converting the flexural rigidity of
a wide-column beam, (Fig. 2(a)), to that of an equivalent full-spam beam, (Fig. 2(b)) as follows: [4].

EIoe ¼ EIou ð1 þ ððbc =2Þ=ððd  bc Þ=2ÞÞÞ3 ; ð30Þ

where bc is the length of the shear core and EIou can be calculated by using theory of parallel axes.
The rotation caused by the shear force in the outrigger and belt truss system, hs, results from strain in diagonals, and can
be expressed as:

hs ¼ ½1=ðhAGou Þ; ð31Þ

where h is the height of the outrigger, and AGou is racking shear stiffness of the outrigger and belt truss system. This racking
shear stiffness can be calculated for specific outrigger truss types. The racking shear stiffness is a property for which the
method of determination is most particular to the type of bent. It depends on the deformation of the web members as
the structure racks under shearing action. It should be noted that the vertical members do not have any influence on the

Fig. 2. Wide-column effect of core and belt truss.

4926 R. Kamgar, M.M. Saadatpour / Applied Mathematical Modelling 36 (2012) 4918–4930

racking shear stiffness of the segment. For various types of bents, the values of AGou have been given by Stafford Smith and
Coull [8].
The value of Kr which corresponds to stiffness of the spring at x = a, can be derived as follows:
2
K r ¼ ½ð2aÞ=ðd AEÞ þ ðdÞ=ð12EIoe Þ þ ð1=hAGou Þ1 ; ð32Þ
where A is the area of exterior columns in position of belt truss and outrigger system that are perpendicular to the direction
of vibration of structure.

4. Accurate of the results

Example 1. Three numerical examples are given to demonstrate the ease in application and accuracy of the proposed
approximate method. A high-rise 40-storey reinforced concrete building consisting of framed tube, shear core, belt truss and
outrigger system, as shown in Figs. 3 and 4 is analyzed. The sizes of all beams, columns that change in height of the structure
have been listed in Table 1. The height of each storey is 3.0 m and the center-to-center spacing of the columns (S) is 2.5 m .
Young’s and shear modulus of the material which have been used in structural elements such as beams, columns, slabs, shear
2 2
core, belt truss and outrigger system are E ¼ 2  109 ðkg m Þ and G ¼ 8  108 ðkg m Þ , respectively.
The location of belt truss and outrigger system is 30(m) from the base of the structure. The sizes of belt truss and outrig-
ger system members are 0.8  0.8(m). Other specifications of belt truss and outrigger system used in numerical example, as
shown in Fig. 4, as follows:
sv ¼ sh ¼ 5 ðmÞ; soh ¼ sov ¼ 2:5 ðmÞ; sb ¼ 2:5 ðmÞ:
Other specifications used in numerical example are as follows:

Lf ¼ 35 ðmÞ; Lw ¼ 30 ðmÞ; s ¼ 2:5 ðmÞ; q ¼ 2400 ðkg m2 Þ;

tslab ¼ 0:3 ðmÞ; L ¼ 120 ðmÞ; y ¼ 3:0 ðmÞ;

where y, tslab and q are height of each storey, thickness of the floor slab and mass per unit volume of materials of the system,
respectively. Dimensions of the core are 5.0  5.0 m, thickness of shear core panels (tSC) that changes in height of the struc-
ture have been listed in Table 1 and Poisson ratio is assumed to be 0.25.
By using the method presented by Kwan [19] and also by considering Fig. 1, the equivalent section area of plan of framed
tube structure (Aeft) that changes in the height of the structure can be calculated as follows:

Fig. 3. Model of numerical example.

 R. Kamgar, M.M. Saadatpour / Applied Mathematical Modelling 36 (2012) 4918–4930 4927

Fig. 4. Plane of outrigger and belt truss system.

Table 1
The properties of combined system consisting of framed tube, shear core, belt truss and outrigger system.

No. Height from the base of the Dimensions of beams and Thickness of AGt (kg) EIt (kg m2) M (kg) Geft (kg m2)
Storey structure (m) columns (cm) shear core (cm)
10 30 80 25 43.803  108 1.0548  1013 11357280 1.4596  108
11 63 80 25 43.803  108 1.0548  1013 13544256 1.4596  108
40 120 60 20 23.129  108 5.9091  1012 19606176 7.4387  107

tequ ¼ Ac =s
in step 1; 2 t equ ¼ 0:256 ðmÞ
in step 3 t equ ¼ 0:1440 ðmÞ
:
Afts ¼ 2  t equ  ðLw þ 2  t equ Þ
in step 1; 2 Afts ¼ 15:6221 ðm2 Þ
in step 3 Afts ¼ 8:7229 ðm2 Þ
In Fig. 1, both sizes of web and flange panel of framed tube structure do not vary in height of the structure. The total equal
thickness of framed tube and shear core changes in height of the structure and can be calculated as follows:

t tte ¼ tequ þ t sc
in step 1; 2 t tte ¼ 0:506 ðmÞ ;
in step 3 ttte ¼ 0:344 ðmÞ
where ttte, tequ and tsc are the total thickness of framed tube and shear core, the equal thickness for both of the web and flange
panels of framed tube structure and the thickness of shear core are as follows, respectively.
The moment of inertia of the shear core (Isc) and framed tube (Ift) about the Y-axis that changes in height of the structure
can be calculated as follows:

ISC ¼ ð1=12Þ  ½ða þ t sc Þ4  ða  tsc Þ4 

Ift ¼ ð1=12Þ  ½Lf  ðLw þ 2  t equ Þ3  ðLf  2  tequ Þ  L3w 
EIt ¼ E  ðISC þ Ift Þ ¼
in step 1; 2 EIt ¼ 1:0548  1013 ðkg m2 Þ
in step 3 EIt ¼ 5:9091  1012 ðkg m2 Þ:
The effective section area of shear core (Asc) and framed tube (Afts ) for computing the shear rigidity that changes in the height
of the structure can be calculated as follows:
4928 R. Kamgar, M.M. Saadatpour / Applied Mathematical Modelling 36 (2012) 4918–4930

ASC ¼ ða þ t sc Þ  2  tsc
in step 1; 2 ASC ¼ 2:6250 ðm2 Þ
in step 3 ASC ¼ 2:0800 ðm2 Þ
Afts ¼ ðLw þ 2  t equ Þ  2  tequ
in step 1; 2 Afts ¼ 15:6221 ðm2 Þ
in step 3 Afts ¼ 8:7229 ðm2 Þ:
The equivalent elastic parameters for the analogous orthotropic membrane tube (Geft), as evaluated by Kwan [19] which
change in every step of the height of the structure, have been listed in Table 1.
In calculation of the effective shear area of beams and columns; we use the Hutchinson’s k [20]. In rectangular cross sec-
tion Hutchinson’s k is defined as follows:

k ¼  mÞ
2ð1þ 
b2
9 C þm 1 1
4
4a5 b1 a2
1 1
  P
1 npa1
16m2 b51 ðnpa1 b1 tanhð ÞÞ
4 3 2 b1
C 4 ¼ 45 a1 b1 12a21  15ma21 þ 5mb1 þ 5
ðnpÞ ð1þmÞ
;
n¼1

where the depth of the column or beam (y-direction) is 2a1 and the width of the column or beam (z-direction) is 2b1 (see
Fig. 5).
The total shear stiffness (AGt) of the structure which changes in every step of the height of the structure can be calculated
as follows:
AGt ¼ ðASC  GÞ þ ðAfts  Geft Þ ¼
in step 1; 2 AGt ¼ 43:803  108 ðkgÞ and k ¼ 0:86623 :
in step 3 AGt ¼ 23:129  108 ðkgÞ and k ¼ 0:86623
By using Eq. (26), the value of Kr is calculated as follows:

K r ¼ 5:0115  109 ðkg mÞ

d ¼ 30ðmÞ
AE ¼ 1:92  1010 ðkgÞ :
11
EIoe ¼ 1:7591  10
AGou ¼ 916467325:4
The structure has been analyzed and the result of the first natural frequency which has been obtained by the proposed
approximate method is compared with the result of the finite element method which has been obtained from SAP 2000 [22].
By inserting these data into Eq. (25) and solving Eq. (25) through an iterative numerical process, we obtain
x1 = 1.855(rad/s).The value of x1 which can be obtained by using the finite element method is 1.8034(rad/s). The proposed
approximate method overestimates the natural frequency by 2.86%.

Example 2. A high-rise 50-storey reinforced concrete building consisting of framed tube, shear core belt truss and outrigger
system is analyzed. The sizes of all beams and columns that change in the height of the structure have been listed in Table 2.
The height of each storey is 3.0 m and the center-to-center spacing of the columns (S) is 2.5 m. Young’s and shear modulus of

Fig. 5. Plane of a column.

 R. Kamgar, M.M. Saadatpour / Applied Mathematical Modelling 36 (2012) 4918–4930 4929

Table 2
The properties of combined system consisting of framed tube, shear core, belt truss and outrigger system.

No. Height from the base of the Dimensions of beams Thickness of AGt (kg) EIt (kg m2) M (kg) Geft (kg m2)
Storey structure (m) and columns (cm) shear core (cm)
10 30 120 30 1.5327  1010 2.4084  1013 14467161.60 3.5620  108
20 60 100 30 8.4894  109 1.6589  1013 13327200 2.4129  108
30 90 80 25 4.3803  109 1.0548  1013 11419392 1.4596  108
40 120 60 20 2.3129  109 5.9091  1012 9836064 7.4387  107
50 150 40 20 1.7745  109 2.6340  1012 8709120 2.8649  107

Table 3
The properties of framed tube.

No. Height from the base of the Dimensions of beams Thickness of AGt (kg) EIt (kg m2) M (kg) Geft (kg m2)
Storey structure (m) and columns (cm) shear core (cm)
40 180 80 0 2.2803  109 1.0506  1013 44599775.94 1.4596  108

Table 4
The total elements number of examples.

Example’s number The total number of line elements The total number of area elements The total number of nodes
Example 1 4356 1319 3488
Example 2 5396 1649 4338
Example 3 4160 39 2172

Table 5
The total segments number of examples.

Example’s The total number of The total number of

number segments nodes
Example 1 3 4
Example 2 5 6
Example 3 1 2

the material which have been used in structural elements such as beams, columns, slabs, shear core, belt truss and outrigger
2 2
system are E ¼ 2  109 ðkg m Þ and G ¼ 8  108 ðkg m Þ, respectively. The location of belt truss and outrigger system is
24(m) from the base of the structure. The sizes of belt truss and outrigger system members are1.2  1.2(m). Other
specifications of belt truss and outrigger system used in numerical example, as shown in Fig. 4, are as follows:

sv ¼ sh ¼ 5 ðmÞ; soh ¼ sov ¼ 2:5 ðmÞ; sb ¼ 2:5 ðmÞ:

By inserting these data into Eq. (25) and solving Eq. (25) through an iterative numerical process, we obtain x1 = 1.551(rad/s).
The value of x1which can be obtained by using the finite element method is 1.6175(rad/s). The proposed approximate meth-
od underestimates the natural frequency by 4.29%.

Example 3. A high-rise 40-storey reinforced concrete building of framed tube is analyzed. The sizes of all beams and
columns are the same and have been listed in Table 3. The height of each storey is 3.0 m and the center-to-center spacing of
the columns (S) is 2.5 m. Young’s and shear modulus of the material which have been used in structural elements such as
2 2
beams, columns, slabs, shear core, belt truss and outrigger system are E ¼ 2  109 ðkg m Þ and G ¼ 8  108 ðkg m Þ,
respectively.
By inserting these data into Eq. (25) and solving Eq. (25) through an iterative numerical process, we obtain
x1 = 1.096(rad/s). The value of x1which can be obtained by using the finite element method is 1.1562(rad/s). The proposed
approximate method underestimates the natural frequency by 5.49%.
In the three examples, the original structures with real properties are modeled and analyzed. The total numbers of nodes
and the total number of line elements such as beams, columns, belt truss and outriggers members have been shown in Table
4. In addition, the total numbers of area elements such as slabs and shear core’s elements have been shown in Table 4.
The total numbers of segments and nodes in proposed method have been shown in Table 5.
By considering Tables 4 and 5, this fact is shown that by decreasing the total number of elements and nodes, amount of
computing and time that it requires for dynamic analysis by using of FEM will be decreased considerably.
4930 R. Kamgar, M.M. Saadatpour / Applied Mathematical Modelling 36 (2012) 4918–4930

The main sources of error between the proposed approximate method and SAP 2000 are as follows:
Modeling the frame panels as equivalent orthotropic membranes (framed tube), so it can be analyzed as a continuous
structure.
The equivalent elastic properties are derived for the frame tube, shear core, and belt truss.
The equivalent stiffness of the rotational spring is used to model the effect of the belt truss and outrigger system on frame
tube.
The approximate values of EIt and AGt are derived for description of distribution of them in height of the tall building.
The effect of shear lag has been neglected in approximate method.
As shown in Rahgozar et al. [21], by increasing height of the structure or omitting the belt truss and outrigger system from
the tall building, the shear lag phenomenon increases. Also as shown in Kwan [19], stiffness of the structure changes by
changing in height of the structure which can be effect on the accuracy of the natural frequency of tall building. This phe-
nomenon can be seen in three examples.

5. Conclusion

In this paper, a simple approximate method has been developed to determine the first natural frequency of combined
system consists of framed tube, shear core, belt truss and outrigger system. By Using the Hamilton’s variational principle,
the governing equation of vibration of mentioned structure obtained and by using partitioned method, the governing equa-
tion was reduced to an ODE with variable coefficient. The accuracy of the proposed method is verified by three numerical
examples. The numerical examples show the accuracy of the approximate values of the natural frequency obtained by
the proposed method decreased by increasing height of the structure or omitting the belt truss and outrigger system which
lead to increase shear lag phenomenon and as a result stiffness of the structure changes. By increasing the height of the
structure, it is better to increase the number of belt truss and outrigger system and therefore the shear lag phenomenon will
be decreased. The numerical examples show that the approximate values of the natural frequency of tall buildings obtained
by the proposed method are in the acceptable range rather than the more accurate results of the finite element method. From
the point of view of a structural engineer, this error is within the acceptable range of engineering practice and therefore the
proposed method may be used to estimate the natural frequency at the preliminary stage of the design.

References

[1] M.A. Koplow, A. Bhattacharyya, B.P. Mann, Closed form solutions for the dynamic response of Euler–Bernoulli beams with step changes in cross section,
J. Sound Vibration, Elsevier, Science Direct 295 (2006) 214–225.
[2] S. Naguleswaran, Vibration of an Euler- Bernoulli beam on elastic end supports and with up to three step changes in cross-section, Int. J. Mech. Sci.,
Pergamon, Science Direct 44 (2002) pp.2541–2555.
[3] J. Lee, M. Bang, J.Y. Kim, An Analytical model for high-rise wall-frame structures with outriggers, Struct. Design Tall Special Buildings 17 (2008) 839–
851.
[4] B. Stafford Smith, I. Salim, Formulae for optimum drift resistance of outrigger braced tall building structures, J. Comput. Struct. 17 (1983) 45–50.
[5] A. Rutenberg, D. Tal, Lateral load response of belted tall building structures, J. Eng. Struct. 9 (1987) 53–67.
[6] M. Stewart Townend, H. Kerr. Andrew, Approximate dynamics using Hamilton’s principle including applications to non-conservative and conservative
systems, J. Mech. Machine Theor. 3 (1982) 213–220.
[7] J.S. Kuang, S.C. Ng, Lateral shear St. Venant torsion coupled vibration of asymmetric-plan frame structures, Struct. Design Tall Special Buildings 18
(2009) 647–656.
[8] B. Stafford Smith, A. Coull, Tall Building Structures, Analysis and Design, Wiley, New York, 1991.
[9] B.S. Taranath, Structural Analysis and Design of Tall Buildings, McGraw-Hill, New York, 1988.
[10] J.C.D. Hoenderkamp, M.C.M. Bakker, Analysis of high-rise braced frames with outriggers, J. Struct. Design Tall Special Buildings 12 (2003) 335–350.
[11] K.G. Geourgoussis, Approximate analysis of symmetrical structures consisting of different types of bents, J. Struct. Design Tall Special Buildings 16
(2007) 231–247.
[12] K.G. Geourgoussis, A simple model for assessing periods of vibration and modal response quantities in symmetrical buildings, J. Struct. Design Tall
Special Buildings 15 (2006) 139–151.
[13] T. Teresa, The use of a dynamical basis for computing the modes of a beam system with a discontinuous cross-section, J. Sound Vibration, Elsevier,
Science Direct 281 (2005) 1175–1185.
[14] A. Coull, K. Ahmed, Deflection of framed-tube structures, J. Struct. Division 104 (1978) 857–862.
[15] A. Coull, B. Bose, Simplified analysis of frame tube structures, J. Struct. Division 101 (1975) 2223–2240.
[16] J.J. Connor, C.C. Pouangare, Simple model for design of framed-tube structures, J. Struct. Eng. ASCE 117 (1991) 3623–3644.
[17] A.G. Piersol, T.L. Paez, Harri’s shock and vibration handbook, sixth ed., McGraw-Hill, New York, 2010.
[18] M. Halis Gunel, H. Emer Ilgin., A proposal for the classification of structural systems of tall buildings, J. Building Environ., Elsevier, Science Direct 42
(2007) 2667–2675.
[19] A.K.H. Kwan, Simple method for approximate analysis of framed tube structures, J. Struct. Eng. ASCE 120 (1994) 1221–1239.
[20] J.R. Hutchinson, Shear coefficients for Timoshenko Beam Theory, J. Appl. Mech. 68 (2001) 87–92.
[21] R. Rahgozar, A. R Ahmadi, Y. Sharifi, A simple mathematical model for approximate analysis of tall buildings, J. Appl. Math. Model. 34 (2010) 2437–
2451.
[22] Sap, Computers and Structures, Berkeley, California, USA, 2000.