See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/292360921

General Form of the Stiffness Matrix of a Tapered Beam-column

Article · January 2013

CITATIONS READS
6 648

3 authors:

Yasha Zeinali Milad Jamali

Southern Methodist University Babol Noshirvani University of Technology
15 PUBLICATIONS   105 CITATIONS    3 PUBLICATIONS   11 CITATIONS   

SEE PROFILE SEE PROFILE

Saman Musician
Amirkabir University of Technology
4 PUBLICATIONS   9 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Efficient Computer-aided Bridge Impairment Detection Using Machine Learning Algorithms and Computer Vision View project

Context-aware vision-based system for civil and infrastructure engineering View project

All content following this page was uploaded by Yasha Zeinali on 31 January 2016.

The user has requested enhancement of the downloaded file.

 International Journal of Mining, Metallurgy & Mechanical Engineering (IJMMME) Volume 1, Issue 2 (2013) ISSN 2320–4060 (Online)
General Form of the Stiffness Matrix of a
Tapered Beam-column
Yasha H. Zeinali, S. Milad Jamali, and Saman Musician
Abstract—Shear deformation and distributed or concentrated
axial force influence the structural behavior of all kinds of beams
specially tapered ones. A tapered Timoshenko-Euler beam element
(general form of beams) under a concentrated and constant
distributed axial force with box-shaped cross-section is considered
and governing equilibrium differential equation established. By
utilizing the Chebyshev Polynomial approach technique, the
governing differential equation is fully solved and stiffness matrix
established.
Keywords— Buckling Load, Chebyshev Polynomial, Stiffness
Matrix, Tapered beams, Timoshenko-Euler beam-columns.
I. INTRODUCTION
S TRUCTURAL members may have non-prismatic sections.
Utilizing the non-prismatic members causes the stresses
distribute more smoothly and the consumption of material
reduces [1]. These facts lead designers to use widely tapered
members.
During the past decades, several researchers have examined
the structural behavior of this type of elements. Some of them
focused on flexural or torsional buckling of non-prismatic
members [4], [6], [8], [9], [10], [11], and some other focused
on free vibration analysis of tapered members [5], [7], [12].
One of the analysis methods of tapered members is dividing
it into a number of uniform elements which is called "step
representation" [13]. This technique is lowly efficient [1]. An
efficient method is solving the governing differential equation
by using a numerical method.
According to modeling assumptions, four types of elements
for tapered members exist which are illustrated in Table I.
TABLE I
TYPES OF BEAMS-ACCORDING TO CONSIDERED EFFECTS
Beam name Axial Force Shear Deformation
Bernoulli
Bernoulli-Euler ×
Timoshenko ×
Timoshenko-Euler × ×
According to Table. I, Timoshenko–Euler beam element
(the last one) includes effects of axial force and shear
Yasha H. Zeinali, is M.Sc. Student at Amirkabir University of Technology,
Tehran, Iran (corresponding author to provide phone: +98-912-4942-181;
e-mail: yasha_zeinaly@aut.ac.ir).
Milad Jamali, is M.Sc. Student at Tehran University, Tehran, Iran. (e-mail:
m_jamali90@ut.ac.ir).
Saman Musician, is M.Sc. Student at Amirkabir University of Technology,
Tehran, Iran. (e-mail: Saman.Musician@gmail.com).
149
deformation. The Timoshenko–Euler element is the most
general one.
Shear deformation and distributed or concentrated axial
force have been verified to influence significantly the
structural behavior of all kinds of beams specially tapered
ones [14]. These influences reduce the stiffness of members.
Several researches have been done to establish the stiffness
matrix and governing differential equation of equilibrium of a
tapered beam element. In this studies a number of numerical
methods have been proposed like: VIM (Variational Iteration
Method) [17]-[19], General Polynomial Shape Function [4],
Direct Integral approach [20], [21] and Bessel Function
approach [1]. However, the governing differential equation of
equilibrium in these publications did not simultaneously
include effects of axial force and shear deformations.
Shear Deformations have been included in many researches
but this studies were on free vibration analysis of Timoshenko
tapered beams. The effects of axial force on elemental
stiffness in these studies were excluded.
In 2002, Li and Li [1], established the equilibrium
differential equation of a tapered Timoshenko-Euler beam by
considering simultaneously effects of constant axial force and
shear deformation. Li and Li employed power series method
to obtain the expression of the tapered Timoshenko-Euler
beam stiffness equation. In their equations, they assumed that
the beam is just under a concentrated axial force and has an I-
shaped section. They did not consider distributed axial force.
So, to the authors' knowledge there is no publication at present
obtaining the elemental stiffness matrix for tapered
Timoshenko-Euler beam with box-shaped cross-section
considering effects of concentrated and distributed axial force
and shear deformation simultaneously.
This paper drives the governing differential equation of
equilibrium of the Timoshenko-Euler beam axial force by
considering the effects of concentrated or constant distributed
axial force and shear deformation simultaneously.
II. GOVERNING DIFFERENTIAL EQUATION OF EQUILIBRIUM
In this paper, it assumed that cross section of tapered beam
is box-shape and symmetric. The applied forces as well as the
corresponding deformation of tapered element is modeled as
shown in Fig. 1 assuming the axis of the member remains
straight.
 International Journal of Mining, Metallurgy & Mechanical Engineering (IJMMME) Volume 1, Issue 2 (2013) ISSN 2320–4060 (Online)

area. According to basic relations we have

𝜏(𝑥) 1 𝑑
𝛾(𝑥) = = ∙[ 𝑀(𝑥)] (6)
𝐺 𝐺.𝐴𝑊 (𝑥) 𝑑𝑥

(a) In (6), 𝐺 is the shear Modules. In box-section thin-walled

members it can be seen that the slope of the member induced
by shear deformation is equal to total section shear
strain 𝛾(𝑥). So the slope of the member caused by shear
deformation (𝑦 ′ 𝑉 ) is

1 𝑑
𝑦 ′ 𝑉 (𝑥) = ∙[ 𝑀(𝑥)] (7)
𝐺.𝐴𝑊 (𝑥) 𝑑𝑥

(b)
Substituting (4) into (7), we have

−𝑉1 + 𝑦(𝑥)𝑁(𝑥) + 𝑁1 𝑦′(𝑥)

𝑦 ′ 𝑉 (𝑥) = (8)
𝐺. 𝐴𝑤 (𝑥)

Differentiating (8) by 𝑥 yields

(c)
Fig. 1 (a) Element forces (b) Element deformation (c) Element
𝑦"𝑉 (𝑥) =
free-body diagram
1 𝑑 𝑑
�𝑉1 � 𝐴𝑤 (𝑥)� − 𝑦(𝑥)𝑁(𝑥) 𝐴𝑤 (𝑥) −
𝐺 (𝐴𝑤 (𝑥))2 𝑑𝑥 𝑑𝑥
Fig. 1-a and 1-b shows the positive directions of forces and
𝑑 𝑑 𝑑 (9)
deformations. Element deflection (𝑦) (shown in Fig. 1-b) � 𝐴𝑤 (𝑥)� 𝑁1 � 𝑦(𝑥)� + 𝐴𝑤 (𝑥) � 𝑦(𝑥)� 𝑁(𝑥) +
𝑑𝑥 𝑑𝑥 𝑑𝑥
consists of two portions. One is caused by bending
𝑑 𝑑2
deformation (𝑦𝑀 ) and the other one induced by shear 𝐴𝑤 (𝑥)𝑦(𝑥) � 𝑁(𝑥)� + 𝐴𝑤 (𝑥)𝑁1 � 𝑦(𝑥)��
𝑑𝑥 𝑑𝑥 2
deformation (𝑦𝑉 ). Thus we have

𝑦 = 𝑦𝑀 + 𝑦𝑉 (1) And by substituting (4) into (3) we obtain

𝑥
−𝑀1 +𝑉1 .𝑥−∫0 𝑦(𝑥)𝑁(𝑥)𝑑𝑥−𝑁1 .𝑦(𝑥)
And a similar relation can be written for curvatures 𝑦"𝑀 = (10)
𝐸.𝐼(𝑥)

𝑦" = 𝑦"𝑀 + 𝑦"𝑉 (2)

Substituting (9) and (10) into (2) and simplifying yields
The curvature of the element caused by bending is 𝑑
�𝑑𝑥𝐴𝑤 (𝑥)�𝑁1
𝑑2 𝑁1 𝑑
−𝑀 𝑦(𝑥) � − 1� + 𝑦(𝑥)(− 2 +
𝑦"𝑀 = (3) 𝑑𝑥 2 𝐺 𝐴𝑤 (𝑥) 𝑑𝑥 𝐺 �𝐴𝑤 (𝑥)�
𝐸.𝐼(𝑥)
𝑑
�𝑑𝑥𝐴𝑤 (𝑥)�𝑁(𝑥) 𝑑
𝑁(𝑥) (11)
𝑁(𝑥) 𝑑𝑥 𝑁1
In (3) 𝐼(𝑥) is inertia moment of the cross-section at the (𝑥)
) + 𝑦(𝑥)(− 2 + − )−
𝐺 𝐴𝑤 𝐺 �𝐴𝑤 (𝑥)� 𝐺 𝐴𝑤 (𝑥) 𝐸 𝐼(𝑥)
location of distance 𝑥 from left end of the element. 𝐸 is the 𝑥
∫0 𝑦(𝑥)𝑁(𝑥)𝑑𝑥 𝑉1 𝑑 𝑀1 𝑉1 𝑥
elastic modules and 𝑀 is The cross-section moment. 𝑀 can be =− 2 � 𝐴𝑤 (𝑥)� + −
𝐸 𝐼(𝑥) 𝐺 �𝐴𝑤 (𝑥)� 𝑑𝑥 𝐸 𝐼(𝑥) 𝐸 𝐼(𝑥)
obtained by (4)
𝑥
(11) is the governing differential equation of equilibrium. In
𝑀(𝑥) = 𝑀1 − 𝑉1 . 𝑥 + ∫0 𝑦(𝑥)𝑁(𝑥)𝑑𝑥 + 𝑁1 . 𝑦(𝑥) (4) this paper the Chebyshev Polynomial approach have been
utilized to solve this differential equation. It is better to solve
In (4), 𝑉1 and 𝑀1 are the left end shear force and moments the equation in the [0,1] domain [3]. So we need to convert the
and 𝑁(𝑥) and 𝑁1 are distributed and concentrated axial force differential equation to non-dimensional form. This Procedure
𝑥
respectively. In box-shaped sections most of the shear stresses can be done by letting the 𝜉 = and rewriting (11) by this
𝐿
𝜏(𝑥) are acting on webs. According to this fact, the shear assumption. By defining some new functions, the equilibrium
stress approximately is equal to differential equation changes to
𝑉(𝑥) 𝑉(𝑥) 1 𝑑
𝜏(𝑥) = = = ∙[ 𝑀(𝑥)] (5) 𝑑2 𝑑
𝐴𝑆ℎ𝑒𝑎𝑟 (𝑥) 𝐴𝑊 (𝑥) 𝐴𝑊 (𝑥) 𝑑𝑥 𝛼(𝜉) 𝑦(𝜉) + 𝛽(𝜉) 𝑦(𝜉) + 𝛾(𝜉) 𝑦(𝜉) −
𝑑𝜉 2 𝑑𝜉
𝜉 𝑀1
(12)
In (5), 𝑉(𝑥) is the section shear force and 𝐴𝑊 (𝑥) is web ∫0 𝑦(𝜉)𝑁(𝜉)𝑑𝜉 = 𝜅(𝜉)𝑉1 + 𝐿
− 𝑉1 𝜉

150
 International Journal of Mining, Metallurgy & Mechanical Engineering (IJMMME) Volume 1, Issue 2 (2013) ISSN 2320–4060 (Online)

𝑁0 𝑦𝑛−1
∑ni=0 βi (n + 1 − i)yn+1−i + ∑ni=0 γi yn−i + =
𝑛
In which
V1 κn
𝐸 𝐼(𝜉) 𝑁1
⇒
𝛼(𝜉) = � − 1� (13a) yn+2 =
1
∗ [− ∑ni=0 βi (n + 1 −
𝐿3 𝐺 𝐴𝑤 (𝜉)
α0 (n+2)(n+1)
𝑑
𝐸 𝐼(𝜉)
�𝑑𝜉𝐴𝑤 (𝜉)�𝑁1
𝑁(𝜉) (13b) i)yn+1−i − ∑ni=0 γi yn−i + V1 κXn − ∑ni=1 αi (n +2− (17d)
𝛽(𝜉) = (− 2 + ) 𝑁 𝑦
𝐿2 𝐿 𝐺 �𝐴𝑤 (𝜉)� 𝐺 𝐴𝑤 (𝜉) i)(n + 1 − i)yn+2−i − 0 𝑛−1�
𝑛
𝑑 𝑑
�𝑑𝜉𝐴𝑤 (𝜉)�𝑁(𝜉) 𝑁(𝜉)
𝐸 𝐼(𝜉) 𝑑𝜉 𝑁1 (13c)
𝛾(𝜉) = (− 2 + − ) In (17d) we can see that the term yn+2 is a linear function of
𝐿 𝐿 𝐺 �𝐴𝑤 (𝜉)� 𝐿 𝐺 𝐴𝑤 (𝜉) 𝐸 𝐼(𝜉)
𝐸 𝐼(𝜉) 𝑑 y0 , y1 , y2 , y3 and V1 . Series yn may be found when the values
𝜅(𝜉) = − 2 ( 𝐴𝑤 (𝜉)) (13d) of these terms are obtained.
𝐿2 𝐺 �𝐴𝑤 (𝜉)� 𝑑𝜉

III. DIFFERENTIAL EQUATION SOLVING BY USING THE The boundary conditions is

CHEBYSHEV POLYNOMIALS APPROACH
The functions 𝛼(𝜉), 𝛽(𝜉), 𝛾(𝜉) and 𝜅(𝜉) can be expressed For ξ = 0
as 𝑦0 = 0 (18a)
𝐿 𝑉1
𝑦1 = 𝑁1 (𝜃1 − ) (18b)
1− 𝐺∙𝐴𝑤(0)
𝛼(𝜉) = ∑𝐶𝑃𝐷
𝑖=0 𝛼𝑖 𝜉
𝑖 (14a) 𝐺∙𝐴𝑤(0)

𝐶𝑃𝐷 𝑖
𝛽(𝜉) = ∑𝑖=0 𝛽𝑖 𝜉 (14b)
And for ξ = 1
𝛾(𝜉) = ∑𝐶𝑃𝐷
𝑖=0 𝛾𝑖 𝜉
𝑖 (14c)
𝐶𝑃𝐷
𝑦(1) = ∑𝐶𝑃𝐷
𝑖=0 𝑦𝑖 = 𝛿2 − 𝛿1 (19a)
𝜅(𝜉) = ∑𝑖=0 𝜅𝑖 𝜉 𝑖 (14d) 𝐿 𝑉1 −(𝛿2 −𝛿1 ) 𝑁0
𝑦 ′ (1)
= ∑𝐶𝑃𝐷
𝑖=0 𝑖 𝑦𝑖 = 𝑁1 �𝜃2 − � (19b)
1− 𝐺.𝐴𝑤(1)
𝐺.𝐴𝑤(1)
and the general solution of deflection (𝑦(𝜉)) can be
expressed as From the above we know that the series yn may be found
when the values of terms y0 , y1 , y2 , y3 and V1 are obtained.
𝑦(𝜉) = ∑𝐶𝑃𝐷
𝑖=0 𝑦𝑖 𝜉
𝑖 (15a) So we can rewrite the (19a) and (19b) as

too. 𝑐1 . 𝑦1 + 𝑐2 . 𝑦2 + 𝑐3 . 𝑦3 + 𝑐4 . 𝑉1 = (𝛿2 − 𝛿1 ) (20a)

Differentiating with respect to 𝜉, yields 𝐿 (20b)
𝑐5 . 𝑦1 + 𝑐6 . 𝑦2 + 𝑐7 . 𝑦3 + 𝑐8 . 𝑉1 = 𝑁1 �𝜃2 −
1−
𝐺.𝐴𝑤(1)
𝑦′(𝜉) = ∑𝐶𝑃𝐷
𝑖=1 𝑖 𝑦𝑖 𝜉
𝑖−1 (15b) 𝑉1 −(𝛿2 −𝛿1 ) 𝑁(1)
𝐶𝑃𝐷 �
𝑦"(𝜉) = ∑𝑖=2 𝑖 (𝑖 + 1)𝑦𝑖 𝜉 𝑖−2 (15c) 𝐺.𝐴𝑤(1)

By substituting (14a)-(15c) into (12) we have The deflection function (𝑦(𝜉)) is a general solution and it
should be correct for all situations. We can assume that the
∑𝐶𝑃𝐷 𝑛
𝑛=0[∑𝑖=0 𝛼𝑖 (𝑛 + 2 − 𝑖)(𝑛 + 1 − 𝑖)𝑦𝑛+2−𝑖 ]𝜉
𝑛
+ values of y2 , y3 and V1 are equal to zero and just we have
𝐶𝑃𝐷 𝑛 𝑛
∑𝑛=0[∑𝑖=0 𝛽𝑖 (𝑛 + 1 − 𝑖)𝑦𝑛+1−𝑖 ]𝜉 + y1 = 1. So the (20a) and (20b) will be reduce to
∑𝐶𝑃𝐷 𝑛 𝑛 𝐶𝑃𝐷 𝑁0 𝑦𝑛−1 𝑛 (16)
𝑛=0(∑𝑖=0 𝛾𝑖 𝑦𝑛−𝑖 )𝜉 + ∑𝑛=1 𝜉 =
𝑀1
𝑛 𝑐1 = ∑𝐶𝑃𝐷
𝑖=0 𝑦𝑖 (21a)
𝐶𝑃𝐷 𝑛
𝑉1 ∑𝑛=0 𝜅𝑛 𝜉 + − 𝑉1 𝜉 𝑐5 = ∑𝐶𝑃𝐷
𝐿 𝑖=0 𝑖 𝑦𝑖 (21b)

Since (16) must be satisfied regardless of 𝜉, setting the (21a) and (21b) obtains the value of 𝑐1 and 𝑐5 ,
coefficient of 𝜉 𝑛 to zero for 𝑛 = 0,1,2,3, … we obtain three simultaneously. By the same procedure and assuming a
equations. different values for terms y1 , y2 , y3 and V1 we can obtain the
other coefficients. In Table II the assumed valued are shown.
For 𝑛 = 0
𝑀1 TABLE II
= 2𝛼0 . 𝑦2 + 𝛽0 . 𝑦1 + 𝛾0 . 𝑦0 − 𝑉1 . 𝜅0 (17a) ASSUMED VALUES FOR CALCULATING 𝑑𝑖
𝐿
Assumed values for
For 𝑛 = 1 The terms that should be calculated
𝑉1 𝑦3 𝑦2 𝑦1
6𝛼0 . 𝑦3 + 2𝛼1 . 𝑦2 + 2𝛽0 . 𝑦2 + 𝛽1 . 𝑦1 + 𝛾0 . 𝑦1 +
(17b) 𝑐1 , 𝑐5 0 0 0 1
𝛾1 . 𝑦0 − 𝑁0 . 𝑦0 = 𝑉1 . 𝜅1 − 𝑉1
𝑐2 , 𝑐6 0 0 1 0
And For 𝑛 > 1 𝑐3 , 𝑐7 0 1 0 0
∑ni=0 αi (n + 2 − i)(n + 1 − i)yn+2−i + (17c) 𝑐4 , 𝑐8 1 0 0 0

151
 International Journal of Mining, Metallurgy & Mechanical Engineering (IJMMME) Volume 1, Issue 2 (2013) ISSN 2320–4060 (Online)

The equilibrium conditions are representation of the realistic deflection y(z) and the defined
function by Chebyshev Polynomial with definite terms, which
𝑉1 + 𝑉2 = 0 (22a) affects directly no more than the coefficients 𝑐𝑖 and 𝑑𝑖 .
1
𝑉1 𝐿 − 𝑁1 (𝛿2 − 𝛿1 ) − 𝐿 ∫0 𝑦(𝜉)𝑁(𝜉)𝑑𝜉 − 𝑀1 − (22b)
𝑀2 = 0 IV. VERIFICATION MODELS
3B

A. Buckling Load
5B

1
For simplifying the term ∫0 𝑦(𝜉)𝑁(𝜉)𝑑𝜉 in (22b) it can be
shown that

1
∫0 𝑦(𝜉)𝑁(𝜉)𝑑𝜉 =
1
∑𝑖=0 �𝑦𝑖 �∑𝑗=0 𝑁𝑗 � 𝜉 𝑖+𝑗+1 � = (23)
𝑖+𝑗+1 𝜉=1
1
∑𝑖=0 �𝑦𝑖 �∑𝑗=0 𝑁𝑗 �� Fig. 2 Example-A Elastic buckling Load
𝑖+𝑗+1

And if we assume that the distributed axial force is constant The elastic buckling load of the tapered cantilever column
(𝑁(𝜉) = 𝑁0 ), we will have shown in Fig. 2 is obtained by general structural analysis
software SAP2000-14 and the procedure presented in this
1 1 paper assuming local and out-of-plane buckling are prevented.
∑𝑖=0 �𝑦𝑖 �∑𝑗=0 𝑁𝑗 �� = ∑𝐶𝑃𝐷 𝑦 𝑁 =
𝑖+𝑗+1 𝑛=0 𝑛+1 𝑛 0 (24)
𝑑1 . 𝑦1 + 𝑑2 . 𝑦2 + 𝑑3 . 𝑦3 + 𝑑4 . 𝑉1 200

AXIAL FORCE, N (103 KN)

In Which 150

𝑁0
𝑑𝑗 = ∑𝐶𝑃𝐷 𝑦 (𝑗 = 1, … ,4) (25) 100
𝑖=0 𝑖+1 𝑖 SAP2000-14
THIS PAPER
TABLE III
ASSUMED VALUES FOR CALCULATION OF 𝑑𝑖
50
The term that should be Assumed values
calculated 𝑉1 𝑦3 𝑦2 𝑦1 0
𝑑1 0 0 0 1 0.00 0.50 1.00 1.50 2.00
𝑑2 0 0 1 0 Left end deflection in Fig. 2 (mm)
𝑑3 0 1 0 0 Fig. 3 P-Δ graph for example-A
𝑑4 1 0 0 0
Table IV summarizes these comparison results.
If the boundary deformations of the element
TABLE IV
𝛿1 , 𝜃1 , 𝛿2 𝑎𝑛𝑑 𝜃2 are known, the above equations can be EXAMPLE-A RESULTS COMPARISON
combined for solving 𝑉1 and 𝑀1 . Assume that the matrix form Technique Elastic buckling load (KN)
of stiffness equation of the element is SAP2000-14 169792
THIS PAPER 166900
𝛿1 𝑉1 Difference 1.7 %
𝜃 𝑀
� � 1� = � 1�
𝐾 (26)
𝛿2 𝑉2
𝜃2 𝑀2 B. Lateral deflection verification
6B

The results in Table. V represents the comparison between

Then the stiffness matrix would be lateral mid-span deflection of the beam-column shown in Fig.
4 calculated by utilizing this paper procedure and FEM.
−𝜓15 +𝜓13 +𝜓15 +𝜓14
� = �−𝜓18
𝐾
+𝜓16 +𝜓18 +𝜓17
� (27)
+𝜓15 −𝜓13 −𝜓15 −𝜓14
+𝜓19 +𝜓20 +𝜓21 +𝜓22

The expression of 𝜓𝑖 (𝑖 = 1, … ,22) are given in the

Appendix A. Fig. 4 Applied Forces in example-B
In theory, the approach described above is accurate. The
unique possible error comes computationally from the

152
 International Journal of Mining, Metallurgy & Mechanical Engineering (IJMMME) Volume 1, Issue 2 (2013) ISSN 2320–4060 (Online)

The geometrical and material properties of this beam is look 𝜓11 𝑐3 𝜓6 + 𝜓11 𝑐4 )
like the beam assumed in Fig. 4. The results of this paper have 𝜓19 = −𝐿𝜓15 − 𝐿(−𝑑1 𝜓2 𝜓15 + 𝑑2 (−𝜓4 𝜓15 − 𝜓5 𝜓18 ) +
been calculated by dividing the whole model into two 𝑑3 (−𝜓6 𝜓15 − 𝜓7 𝜓18 ) − 𝑑4 𝜓15 ) + 𝜓18 + 𝑁1
elements. The FEM results have been calculated by general 𝜓20 = 𝐿𝜓13 − 𝜓16 − 𝐿(𝑑1 (𝜓1 + 𝜓2 𝜓13 ) + 𝑑2 (𝜓3 + 𝜓4 𝜓13 +
structural analysis software SAP2000-14 and utilizing the step 𝜓5 𝜓16 ) + 𝑑3 (𝜓6 𝜓13 + 𝜓7 𝜓16 + 𝜓8 ) + 𝑑4 𝜓13 )
representation method. SAP2000-14 considers just the shear 𝜓21 = 𝐿𝜓15 − 𝐿(𝑑1 𝜓2 𝜓15 + 𝑑2 (𝜓4 𝜓15 + 𝜓5 𝜓18 ) + 𝑑3 (𝜓6 𝜓15 +
deformation influences on stiffness matrix. 𝜓7 𝜓18 ) + 𝑑4 𝜓15 ) − 𝑁1 − 𝜓18
𝜓22 = 𝐿𝜓14 − 𝐿(𝑑1 𝜓2 𝜓14 + 𝑑2 (𝜓4 𝜓14 + 𝜓5 𝜓17 ) + 𝑑3 (𝜓6 𝜓14 +
TABLE V 𝜓7 𝜓17 ) + 𝑑4 𝜓14 ) − 𝜓17
L
EXAMPLE-B RESULTS COMPARISON 𝑁𝑡 = −𝑁2 = N1 + ∫0 N(z)dz = N1 + N0 L
Technique Left end deflection ,δ1 (mm)
SAP2000-14 0.9462 REFERENCES
THIS PAPER 0.9042
[1] Li GQ, Li JJ. "A tapered Timoshenko-Euler beam element for analysis
Difference 4.6 % of steel portal frames". Journal of Constructional Steel Research
2002;58(12):1531–44.
V. CONCLUSION [2] Li GQ, Li JJ, Chan SL. "A second-order inelastic model for steel frames
of tapered members with slender web". Journal of Engineering
Axial Force and shear deformation influence the tapered Structures 2003;25(14): 1033–1043.
beams stiffness-equation and significantly reduce their [3] Mason JC, Handscomb DC. Chebyshev polynomials. London: CRC
Press LLC; 2003.
stiffness. [4] Bradford MA, Cuk PE. "Elastic buckling of tapered monosymmetric I-
A tapered Timoshenko-Euler beam element (general form beams". J of Struc Eng 1988;114(5):977–96.
of beams) under a concentrated and distributed axial force [5] Dube GP, Dumir PC. "Tapered thin open section beams on elastic
foundation—buckling analysis". Computers & Structures
with box-shaped cross-section is considered and governing
1996;61(5):845–57.
equilibrium differential equation established. By utilizing the [6] Gupta P, Wang ST, Blandford GE. "Lateral-torsional buckling of
Chebyshev Polynomial approach technique, the governing nonprismatic I-beam". J. of Struc. Engng 1996;122(7):748–55.
differential equation of latticed and arbitrary variational cross- [7] Gupta AK. "Frequent-dependent matrices for tapered beams". J. of
section members is fully solved and stiffness matrix Struc. Engng 1986;112(1):85–103.
[8] Kitipornchai S, Trahair NS. "Elastic stability of tapered I-beams". J. of
established. Struc. Divi 1972;98(3):713–28.
Like [1], The verifications proves that the proposed element [9] Lee GC, Szabo BA. "Tortional response of tapered I-girders". J. of
and procedure could be successfully applied to general 2D Struc. Divi 1967;93(5):233–52.
structures. [10] Li JJ. "Research on theory of nonlinear analysis and integrated reliability
design for steel portal frames with tapered members", PhD dissertation,
Department of Buildings and Structural Engineering, Tongji University,
APPENDIX Shanghai, China (in Chinese), 2001.
𝐿𝐺𝐴𝑤 (0) −𝐿 −𝛽0 𝜓1 [11] Polyzois D, Raftoyiannis IG. "Lateral-torsional stability of steel web-
𝜓1 = 𝜓2 = 𝐺𝐴 𝜓3 = , tapered I-beams". J. of Struc. Engng 1998;124(10):1208–16.
𝐺𝐴𝑤 (0)−𝑁1 𝑤 (0)−𝑁1 2𝛼0
−𝛽0 𝜓2 +𝜅0 1 [12] Wekezer JW. "Vibrational analysis of thin-walled bars with open cross
𝜓4 = 2𝛼0
𝜓5 = 2 𝐿 𝛼 ,
0 sections". J of Struc Engng 1989;115(12):2965–78.
𝜅1 −1−𝛾0 𝜓2 −2𝛼1 𝜓4 −2𝛽0 𝜓4 −𝛽1 𝜓2 −𝜓5 (𝛼1 +𝛽0 ) [13] Wang CK. Stability of rigid frames with nonuniform members. J of
𝜓6 = 𝜓7 = ,
6𝛼0 3𝛼0 Struc Divi 1967;93(1):275–94.
−𝛾0 𝜓1 −2𝛼1 𝜓3 −2𝛽0 𝜓3 −𝛽1 𝜓1
𝜓8 = , [14] Li GQ, Shen ZY. "Theory for analysis and calculation of elastic and
6𝛼0 elasto-plastic behavior of steel frameworks", Shanghai Science and
𝐿𝐺𝐴𝑤 (1) −𝐿
𝜓9 = 𝜓10 = Technology Press (in Chinese), 1998.
𝐺𝐴𝑤 (1)−𝑁1 𝐺𝐴𝑤 (1)−𝑁1
𝐿 𝑁0
[15] Banerjee JR. "Exact Bernoulli-Euler static stiffness matrix for a range of
𝜓11 = 𝐺𝐴 (1)−𝑁 tapered beam-columns". Int J Num Meth Engng 1986;23(9):1615–28.
𝑤 1
[16] Gere JM, Carter WO. "Critical buckling loads for tapered columns". J.
𝜓12 = 𝑐1 𝜓2 𝑐7 𝜓7 + 𝑐1 𝜓2 𝑐6 𝜓5 + 𝑐2 𝜓4 𝑐7 𝜓7 + 𝑐3 𝜓6 𝑐6 𝜓5 + 𝑐4 𝑐7 𝜓7 +
of Struc. Divi 1962;88(1):1–11.
𝑐4 𝑐6 𝜓5 − 𝑐3 𝜓7 𝑐5 𝜓2 + 𝑐3 𝜓7 𝜓10 − 𝑐3 𝜓7 𝑐8 − 𝑐3 𝜓7 𝑐6 𝜓4 − [17] J.H.He, Variational "iteration method – A kind of nonlinear analytical
𝑐2 𝜓5 𝑐5 𝜓2 + 𝑐2 𝜓5 𝜓10 − 𝑐2 𝜓5 𝑐7 𝜓6 − 𝑐2 𝜓5 𝑐8 , technique": some examples, International Journal of Nonlinear
−1
𝜓13 = 𝜓 (𝑐7 𝜓7 𝑐1 𝜓1 + 𝑐7 𝜓7 𝑐2 𝜓3 + 𝑐6 𝜓5 𝑐1 𝜓1 + 𝑐6 𝜓5 𝑐3 𝜓8 − Mechanics 34(4) (2000) , 699-708
12 [18] J.H. He, "A review on some new recently developed nonlinear analytical
𝑐3 𝜓7 𝑐5 𝜓1 − 𝑐3 𝜓7 𝑐6 𝜓3 − 𝑐2 𝜓5 𝑐5 𝜓1 − 𝑐2 𝜓5 𝑐7 𝜓8 ) techniques", International Journal of Nonlinear Science Numerical
(−𝑐3 𝜓7 −𝑐2 𝜓5 )𝜓9 Simulation 1(1) , (2000) , 51-70
𝜓14 = 𝜓12 [19] J.H.He, "Variational itration method – some recent result and new
𝑐7 𝜓7 +𝑐6 𝜓5 −𝜓11 𝑐3 𝜓7 −𝜓11 𝑐2 𝜓5
𝜓15 = interceptions:, journal of Copmutational and Applied Mathematics 207
𝜓12 (1) , (2007) , 3-17
1
𝜓16 = 𝜓 (𝑐5 𝜓2 𝑐2 𝜓3 + 𝑐5 𝜓2 𝑐3 𝜓8 + 𝑐7 𝜓6 𝑐1 𝜓1 + 𝑐7 𝜓6 𝑐2 𝜓3 + [20] Just DJ. "Plane frameworks of tapering box and I-section". J. of Struc.
12 Divi 1977;103(1):71–86.
𝑐6 𝜓4 𝑐1 𝜓1 + 𝑐6 𝜓4 𝑐3 𝜓8 − 𝑐1 𝜓2 𝑐6 𝜓3 − 𝑐1 𝜓2 𝑐7 𝜓8 − 𝑐2 𝜓4 𝑐5 𝜓1 − [21] Karabalis DL. Static, "dynamic and stability analysis of structures
𝑐2 𝜓4 𝑐7 𝜓8 − 𝑐3 𝜓6 𝑐5 𝜓1 − 𝑐3 𝜓6 𝑐6 𝜓3 + 𝑐8 𝑐2 𝜓3 − 𝜓10 𝑐2 𝜓3 − composed of tapered beams". Computers & Structures 1983;16(6):731–
𝑐4 𝑐7 𝜓8 + 𝑐8 𝑐3 𝜓8 − 𝜓10 𝑐1 𝜓1 − 𝑐4 𝑐5 𝜓1 + 𝑐8 𝑐1 𝜓1 − 𝜓10 𝑐3 𝜓8 − 48.
𝑐4 𝑐6 𝜓3 )
(𝑐1 𝜓2 +𝑐2 𝜓4 +𝑐3 𝜓6 +𝑐4 )𝜓9
𝜓17 = 𝜓12
1
𝜓18 = 𝜓 (−𝑐5 𝜓2 + 𝜓10 − 𝑐7 𝜓6 − 𝑐8 − 𝑐6 𝜓4 + 𝜓11 𝑐2 𝜓4 +
12

153

View publication stats