International Journal of Structural Stability and Dynamics

Vol. 16 (2016) 1550017 (14 pages)

#.c World Scienti¯c Publishing Company
DOI: 10.1142/S0219455415500170

Buckling of Standing Tapered Timoshenko Columns with

Varying Flexural Rigidity Under Combined Loadings

D.-L. Sun*, X.-F. Li*,‡ and C. Y. Wang†

*School of Civil Engineering, Central South University
Changsha 410075, P. R. China
Int. J. Str. Stab. Dyn. Downloaded from www.worldscientific.com
by MONASH UNIVERSITY on 10/23/15. For personal use only.

†Department of Mathematics and Department of Mechanical Engineering

Michigan State University, East Lansing, MI 48824, USA
‡
[email protected]

Received 11 March 2015

Accepted 1 April 2015
Published 20 May 2015

The stability of a nonuniform column subjected to a tip force and axially distributed loading
is investigated based on the Timoshenko beam theory. An emphasis is placed on buckling of
a standing column with varying cross-section and variable material properties under self-
weight and tip force. Four kinds of columns with di®erent taper ratios are analyzed. A
new initial value method is suggested to determine critical tip force and axial loading at
buckling. The e®ectiveness of the method is con¯rmed by comparing our results with those
for Euler–Bernoulli columns for the case of su±ciently large shear rigidity. The e®ects of shear
rigidity, taper ratio, and gravity loading on the buckling loads of a heavy standing or hanging
column are examined.

Keywords: Timoshenko beam theory; taper ratio; nonuniform cross-section; buckling; gravity;
initial value method.

1. Introduction
Standing columns under the in°uence of gravity and tip force are a class of important
structures in Civil engineering such as tall buildings, chimneys and towers being
representative structural elements. When a compressive tip load exceeds a critical
value, elastic columns deviate from an original stable equilibrium state and buckling
takes place. So, predicting the critical load is of signi¯cance for structural designs
against buckling. There have been many investigations devoted to this subject (e.g.
Refs. 1–3).
Euler ¯rst solved the buckling problem of standing prismatic columns using the
power series method.4 Buckling of heavy columns with di®erent boundary conditions

‡
Corresponding author.

1550017-1
 D.-L. Sun, X.-F. Li & C. Y. Wang

were also found with various methods. For example, for certain particular cases,
Bessel functions were used.5,6 However, standing columns are usually not uniform
but tapered, wide at the base and narrow at the top. Vaziri and Xie7 proposed a
numerical integral method to determine the critical load. Based on the integral
scheme, several di®erent approaches have been put forward.8–11 The stability of
elastic columns with self weight and/or tip load was also studied.12–15 Duan
and Wang16 derived analytical solutions for the elastic buckling of heavy columns
with various combinations of end conditions. Huang and Li17 gave an analytic
approach for exactly determining critical buckling loads of a class of nonuniform
columns.
There have been a number of papers devoted to the topic of the stability of
standing columns subjected to a tip load and axially distributed load. However, these
Int. J. Str. Stab. Dyn. Downloaded from www.worldscientific.com
by MONASH UNIVERSITY on 10/23/15. For personal use only.

works are based on the Euler–Bernoulli beam theory, and there are no information
on study of the buckling of nonuniform standing columns under a tip load and axial
loading (such as gravity) based on the Timoshenko beam theory, which includes the
e®ect of transverse shear deformation of the cross-section.
This paper aims at presenting an approach for analyzing buckling problems of
standing nonuniform columns subjected to combined loadings within the frame-
work of the Timoshenko beam theory. The columns considered possess varying
bending sti®ness including varying cross-section and/or varying material prop-
erties. An emphasis is placed on buckling analysis of a class of columns with
variable bending sti®ness and axial load in a power function form. The proposed
method is con¯rmed by examples of nonuniform heavy columns with and without
tip force using the Euler–Bernoulli beam theory. Lastly, we give some numerical
results to show the e®ect of shear rigidity and taper ratio on the critical buckling
load.

2. Statement of the Problem

Consider buckling of a standing elastic column with variable bending sti®ness and
variable axial loading. An illustrated sketch is shown in Fig. 1. In the following
analysis, the Timoshenko beam theory is applied to solve the problem stated above,

Fig. 1. Schematic of a standing tapered column subjected to varying axial loading and tip force.

1550017-2
 Buckling of Standing Tapered Timoshenko Columns

where the transverse shear deformation of the cross section is taken into account.
Based on the Timoshenko beam theory,1 the transverse de°ection wðxÞ and rotation
of the cross-section
ðxÞ at buckling of an elastic column satisfy the following gov-
erning equations

  
d d
dw
EIðxÞ ¼ kGAðxÞ
þ ; 0
x
L; ð1Þ
dx dx dx

 
Z 
d dw d dw dw L
GAðxÞ
þ ¼ P þ qðxÞdx ; 0
x
L; ð2Þ
dx dx dx 0 dx dx x
where E and G are Young's and shear moduli of the column,  is the shear correction
factor, AðxÞ and IðxÞ are the cross-section area and its moment of inertia, respec-
tively, L is the length of the elastic column, P0 is a tip compressive (P0 > 0) or tensile
Int. J. Str. Stab. Dyn. Downloaded from www.worldscientific.com
by MONASH UNIVERSITY on 10/23/15. For personal use only.

(P0 < 0) force, qðxÞ is axially distributed force along the whole column
ðqðxÞ > 0 down; qðxÞ < 0 upÞ. Axially distributed loading may arise from its own
weight, and in this case we have qðxÞ ¼ q0 AðxÞ; q0 ¼ gðxÞ, where g is the acceler-
ation due to gravity and ðxÞ is the mass density of the column dependent on the
local position. The present paper only considers the case of homogeneous density, i.e.
ðxÞ ¼ 0 . Here, x stands for the axial coordinate measured from the clamped end.
When the own weight of an elastic column is considered, we have q0 > 0 for a
standing column and q0 < 0 for a hanging column. In Eqs. (1) and (2), the positive
direction of rotation of cross-section
> 0 is opposite to that of the slope angle of
de°ection dw=dx > 0, and if de¯ning the rotation of cross section  to have the same
positive direction as the slope angle, we only replace
with . Note that in the
absence of axially distributed loading, the tip force must be compressive and buckling
of a standing or hanging column occurs for the tip force in excess of the critical load.
Likely, in the absence of the tip force, axial loading must be compressive which may
give rise to occurrence of buckling of a column. At the presence of both tip force and
axial loading, the tip force may be tensile or axial loading may be tensile. However, it
is not allowed that both of the tip force and axial loading are tensile simultaneously.
In the present paper, we shall determine the buckling tip forces for a prescribed
axially distributed force.
The governing Eqs. (1) and (2) are a system of coupled partial di®erential
equations with varying coe±cients. So an exact solution is nearly impossible unless
for some special cases with the aid of special functions. In this paper, we propose a
new approach for determining the buckling loads. To this end, let us represent
bending moment M and horizontal shear force V in terms of w and
as follows:
d

M ¼ EIðxÞ ; ð3Þ
dx
 
Z 
dw dw dw L
V ¼ GAðxÞ
þ  P0 þ qðxÞdx : ð4Þ
dx dx dx x

1550017-3
 D.-L. Sun, X.-F. Li & C. Y. Wang

Consequently, for a standing column with clamped end x ¼ 0, the boundary con-
ditions at the two ends can be stated as
wð0Þ ¼ 0;
ð0Þ ¼ 0; ð5Þ
MðLÞ ¼ 0; V ðLÞ ¼ 0: ð6Þ

For most elastic columns with varying bending sti®ness EIðxÞ, axially distributed
loading qðxÞ, of much signi¯cance in practice is the case where the bending sti®ness
and axially distributed force are power functions with respect to the axial distance
from the base. Therefore, we have
 x m
EIðxÞ ¼ EI0 lðxÞ; lðxÞ ¼ 1   ; ð7Þ
L
Int. J. Str. Stab. Dyn. Downloaded from www.worldscientific.com
by MONASH UNIVERSITY on 10/23/15. For personal use only.

 
x n
AðxÞ ¼ A0 rðxÞ; rðxÞ ¼ 1   ; ð8Þ
L
where EI0 is the reference value of bending sti®ness EIðxÞ at the base, A0 is the
reference value of cross-section area AðxÞ at the base. It is obvious that varying
bending sti®ness may result from varying cross section or/and varying Young's
modulus. Here only the former case is analyzed and it is simple to extend the results
obtained in the present paper to the latter case of varying Young's modulus, i.e.
axially functionally graded columns. In Eqs. (7) and (8),  represents the taper ratio,
m and n are constants. Of much interest to us is a combination of di®erent values m
and n, corresponding to a column of various shapes. For example, m ¼ 4 and n ¼ 2
correspond to a solid cone- or pyramid-shaped column (Fig. 2(a)), m ¼ 2 and n ¼ 0
correspond to tapered composite column of N inclined uniform legs (Fig. 2(b)),
m ¼ n ¼ 1 correspond to a column with constant thickness (or depth) and linearly
varying sides (Fig. 2(c)), m ¼ 3 and n ¼ 1 correspond to a column with constant sides
and linearly varying thickness (Fig. 2(d)). Usually, the taper ratio  is required to

(a) (b) (c) (d)

Fig. 2. (a) Solid cone column (m ¼ 4; n ¼ 2); (b) Uniform legs (m ¼ 2; n ¼ 0); (c) Constant thickness and
tapered sides (m ¼ 1; n ¼ 1); (d) Tapered thickness and constant sides m ¼ 3; n ¼ 1.

1550017-4
 Buckling of Standing Tapered Timoshenko Columns

satis¯es 
1. When  ¼ 1 is chosen, the column becomes mathematically sharp at

the end x ¼ L, and in this case the column is impossible to bear any compressive force.
In addition,  ¼ 0 implies that the column has uniform cross section, while  < 0
implies that the cross section becomes larger for a position far away from the base. In
the case of  < 1, it is assumed that the column top is not sharp, meaning IðLÞ 6¼ 0.
This is to say that the boundary conditions at the top in (6) can be rewritten as
 
d
dw dw
¼ 0; GAðLÞ
þ  P0 ¼ 0; x ¼ L: ð9Þ
dx dx dx
To simplify our analysis, we introduce the following dimensionless variables:
x EI0 P0 L 2 q0 L 3
¼ ; ¼ ; ¼ ; ¼ ; ð10Þ
L GL 2 A0 EI0 EI0
Int. J. Str. Stab. Dyn. Downloaded from www.worldscientific.com
by MONASH UNIVERSITY on 10/23/15. For personal use only.

and Eqs. (1) and (2) can be transformed into dimensionless form

  
d d
dw
 lðÞ ¼ rðÞ
þ ; ð11Þ
d d d

  "Z #
1
d dw d 2w d dw
rðÞ
þ ¼  2 þ  rðÞd ; ð12Þ
d d d d  d

where
rðÞ ¼ ð1  Þ n ; lðÞ ¼ ð1  Þ m ð13Þ
and
8
Z 1 <1  ;  ¼ 0;
rðÞd ¼ ð1  Þ nþ1  ð1  Þ nþ1 ð14Þ
 : ; ¼6 0:
ðn þ 1Þ

In particular, we ¯nd that if  ! 0, which indicates that shear rigidity is su±-

ciently large so that shear deformation hardly takes place, this case reduces to the
classical Euler–Bernoulli columns. On the other hand,  ¼ 0 means that the tip force
disappears, while ¼ 0 means that the axial loading is neglected.
Equations (11) and (12) are a set of coupled ordinary di®erential equations with
variable coe±cients. For general cases, those equations subjected to appropriate
boundary conditions cannot be solved analytically except for certain special cases.
For the problem under consideration with any combination of two constants m and
n, it seems unlikely to derive its analytical solution. Here a new approach for de-
termining the buckling loads of elastic Timoshenko columns is used.

3. Initial Value Method

The determination of buckling loads of elastic columns is transformed to seek a
nontrivial solution for ordinary di®erential equations with varying coe±cients sub-
jected to four boundary conditions. Although many numerical methods for seeking

1550017-5
 D.-L. Sun, X.-F. Li & C. Y. Wang

buckling loads based on the Euler–Bernoulli beam theory are available, most of these
methods cannot directly be applied to determine the buckling loads based on the
Timoshenko beam theory since governing equations are coupled for the latter, while
only one governing equation is involved for the former.
Generally speaking, seeking a solution to a boundary value problem is more dif-
¯cult than that to an initial value problem. For this reason, we present a new
approach to convert the corresponding boundary value problem into an initial value
problem. Then a well-known Runge–Kutta method is used to solve the resulting
initial value problem.
To this end, owing to the superposition principle of linear problems, let a desired
solution be represented as a sum of two independent particular solutions as follows:
S ¼ c1 s1 ðw1 ;
1 Þ þ c2 s2 ðw2 ;
2 Þ; ð15Þ
Int. J. Str. Stab. Dyn. Downloaded from www.worldscientific.com
by MONASH UNIVERSITY on 10/23/15. For personal use only.

where c1 and c2 are unknown coe±cients, s1 and s2 are any two independent func-
tions that satisfy Eqs. (11) and (12) subjected to the clamped initial conditions in
(5). Thus the transverse de°ection w and the rotation angle
can be expressed below
w ¼ c1 w1 þ c2 w2 ; ð16Þ

¼ c1
1 þ c2
2 ; ð17Þ
where w1 , w2 ,
1 and
2 are the corresponding transverse de°ections and rotation
angles of cross-section, which are independent respectively and meet the following
initial conditions
s1 : w1 ð0Þ ¼ 0; w 01 ð0Þ ¼ 0;
1 ð0Þ ¼ 0;
01 ð0Þ ¼ 1; ð18Þ
s2 : w2 ð0Þ ¼ 0; w 02 ð0Þ ¼ 1;
2 ð0Þ ¼ 0;
02 ð0Þ ¼ 0; ð19Þ
where the prime denotes di®erentiation with respect to , i.e. w 0 ¼ dw=d and

0 ¼ d
=d. Furthermore, the moment and shear force are also expressed below
M ¼ c1 M1 ðw1 ;
1 Þ þ c2 M2 ðw2 ;
2 Þ; ð20Þ
V ¼ c1 V1 ðw1 ;
1 Þ þ c2 V2 ðw2 ;
2 Þ; ð21Þ
where Mj and Vj ðj ¼ 1; 2Þ are the moment and shear force corresponding to the
solution wj and
j ðj ¼ 1; 2Þ, respectively.
As a result, Eqs. (11) and (12) subjected to (18) and (19) become an initial value
problem, and it can be solved with the aid of the Runge–Kutta method to get w1 , w2 ,

1 and
2 . In addition, at the top tip of the column, considering (3) and (9), the
boundary conditions Mð1Þ ¼ 0; V ð1Þ ¼ 0 in (6) are transformed to dimensionless
form in terms of the introduced wj and
j ðj ¼ 1; 2Þ as follows

0 ð1Þ ¼ 0; rð1Þ
ð1Þ þ ½rð1Þ  w 0 ð1Þ ¼ 0: ð22Þ
Remembering (16) and (17), the above equations are rewritten as
c1
01 ð1Þ þ c2
02 ð1Þ ¼ 0; ð23Þ
c1 rð1Þ
1 ð1Þ þ c1 ½rð1Þ  w 01 ð1Þ þ c2 rð1Þ
2 ð1Þ þ c2 ½rð1Þ  w 02 ð1Þ ¼ 0: ð24Þ

1550017-6
 Buckling of Standing Tapered Timoshenko Columns

To obtain a nontrivial solution of the resulting system of Eqs. (23) and (24), the
determinant of the coe±cient matrix of the system has to vanish. Accordingly, we
obtain the following characteristic equation
 

01 ð1Þ
02 ð1Þ 
 
 rð1Þ
ð1Þ þ ½rð1Þ  w 0 ð1Þ rð1Þ
ð1Þ þ ½rð1Þ  w 0 ð1Þ  ¼ 0: ð25Þ
1 1 2 2

Equation (25) provides us with an exact formula for the buckling load parameter 
or by seeking the lowest positive root of the equation, which gives critical buckling
load. Precisely, for a given value of , by solving (25) we can calculate buckling
loading ¼ q0 L 3 =EI0 if P0 is prescribed or buckling loading  ¼ P0 L 2 =EI0 if q0 is
prescribed. The buckling load parameter  or can be obtained very accurately by
the Runge–Kutta method and bisection algorithm. Such an initial value method was
Int. J. Str. Stab. Dyn. Downloaded from www.worldscientific.com
by MONASH UNIVERSITY on 10/23/15. For personal use only.

probably ¯rst proposed by Barasch and Chen18 using numerical integration.

4. Numerical Results and Discussion

In this section, illustrative examples are given and numerical computations are
carried out to show the e®ects of tip force, axial loading and shear deformation of the
cross-section on the stability of standing tapered columns.
Prior to the presentation of numerical results of buckling loading of Timoshenko
columns, the e®ectiveness of this method will be validated by comparing the buckling
loads for columns su±ciently large shear rigidity with those for Euler–Bernoulli
columns. When adopting the Euler–Bernoulli beam theory, buckling loads of an
elastic column can be determined by many approaches. According to the present
analysis, the Euler–Bernoulli beam theory is equivalent to a limiting case of the
Timoshenko beam theory as  ! 0. Thus, to con¯rm the e±ciency of this approach,
numerical computations are carried out for a tip force and axially distributed load-
ing. The results obtained are tabulated in Tables 1 and 2, respectively, where in
calculations  ¼ 10 6 is chosen. In addition, the manipulation encounters some
di±culty in the case of  ¼ 1, as mentioned before. To overcome the di±culty,  ¼ 1
is approximately replaced by  ¼ 0:9999 in our numerical treatment. From Tables 1

Table 1. Critical buckling load parameter for tapered standing columns with  ¼ 10 6 .


0 0.1 0.3 0.5 0.7 0.9999
m ¼ 1; n ¼ 1 EB[15] 7.8374 8.3047 9.5069 11.289 14.236 26.024
Present results 7.8373 8.3046 9.5068 11.288 14.236 26.016
m ¼ 2; n ¼ 0 EB[15] 7.8374 7.5035 6.8105 6.0718 5.2606 3.6705
Present results 7.8373 7.5034 6.8104 6.0717 5.2606 3.6712
m ¼ 3; n ¼ 1 EB[15] 7.8374 7.9477 8.2281 8.6391 9.3286 13.187
Present results 7.8373 7.9477 8.2280 8.6390 9.3285 13.183
m ¼ 4; n ¼ 2 EB[15] 7.8374 8.4144 9.8887 12.054 15.627 30.53
Present results 7.8373 8.4143 9.8886 12.053 15.626 30.520
Remark: EB means Euler–Bernoulli beams.

1550017-7
 D.-L. Sun, X.-F. Li & C. Y. Wang

Table 2. Critical buckling load parameter  for tapered standing

columns with  ¼ 10 6 , m ¼ 4, n ¼ 2.

4
2

 1 3 5 10
0 EB[11] 1.7207 0.1446 1.5517 6.3766
Present results 1.7206 0.14440 1.5519 6.3771
0.2 EB[11] 1.4029 0.3741 0.7556 4.0871
Present results 1.4029 0.37411 0.75562 4.0871
0.4 EB[11] 1.0418 0.4561 0.2087 2.2946
Present results 1.0418 0.45612 0.20868 2.2947
0.6 EB[11] 0.6454 0.3925 0.0873 1.0022
Present results 0.64538 0.39248 0.087290 1.0022
0.8 EB[11] 0.2439 0.1952 0.1292 0.2176
Int. J. Str. Stab. Dyn. Downloaded from www.worldscientific.com
by MONASH UNIVERSITY on 10/23/15. For personal use only.

Present results 0.24393 0.19520 0.12922 0.21761

Remark: EB means Euler–Bernoulli beams.

and 2, we ¯nd that our numerical results agree well with those for Euler–Bernoulli
columns obtained by other approaches.11,14,15 Consequently, the suggested approach
is con¯rmed to be e®ective for the presence of both tip force and distributed loading.
Example 1. Buckling of heavy standing tapered Timoshenko columns without tip
force.
We ¯rst calculate the buckling load parameter of a tapered standing Timosh-
enko column under self-weight ð ¼ 0Þ with an emphasis on the analysis of the e®ect
of shear rigidity, which may be described by . The obtained values of the critical
buckling load parameter for Timoshenko columns of four di®erent geometries are
presented for various values of  in Tables 3–6, respectively. Owing to the e®ects of
the taper ratio, it can be seen from Tables 3, 5 and 6 that the critical buckling load
parameter increases with the increase of . However, for Table 4 corresponding to
m ¼ 2, n ¼ 0, the trend of the critical buckling load parameter is completely oppo-
site. Buckling loading decreases with rising . In addition, in Tables 3–6, the
critical buckling load parameter drops when the shear rigidity becomes smaller or 
is increasing. Precisely, when  increases from 0 to 0:1, from Table 3 one ¯nds that
the critical buckling load parameter decreases 28:5% for  ¼ 0, 30:5% for  ¼ 0:3 and

Table 3. Critical buckling load parameter for tapered standing

columns with m ¼ 1; n ¼ 1.


 0 0.1 0.3 0.5 0.7 0.9999

0 7.8373 8.3046 9.5068 11.288 14.236 26.016

0.001 7.8083 8.2730 9.4681 11.238 14.164 25.831
0.01 7.5547 7.9968 9.1307 10.804 13.549 24.266
0.05 6.5705 6.9303 7.8444 9.1705 11.293 18.936
0.1 5.5976 5.8846 6.6072 7.6392 9.2520 14.611

1550017-8
 Buckling of Standing Tapered Timoshenko Columns

Table 4. Critical buckling load parameter for tapered standing

columns with m ¼ 2; n ¼ 0.


 0 0.1 0.3 0.5 0.7 0.9999
0 7.8373 7.5034 6.8104 6.0717 5.2606 3.6712
0.001 7.8083 7.4772 6.7894 6.0556 5.2491 3.6667
0.01 7.5547 7.247 6.6043 5.9132 5.1474 3.6266
0.05 6.5705 6.3468 5.8687 5.3378 4.7287 3.4541
0.1 5.5976 5.4460 5.1131 4.7296 4.2714 3.2511

35% for  ¼ 0:7. Hence, the e®ects of  are stronger for higher values of . This
phenomenon can also be seen in Tables 4–6.
Int. J. Str. Stab. Dyn. Downloaded from www.worldscientific.com
by MONASH UNIVERSITY on 10/23/15. For personal use only.

Example 2. Buckling of heavy standing tapered Timoshenko columns with tip

force.
Now we turn our attention to another case of an elastic column with both axially
distributed loading and tip force, i.e. the buckling of a standing tapered Timoshenko
column under self-weight and tip force. For a heavy standing coned column with
circular or squared cross-section ðm ¼ 4; n ¼ 2Þ subjected to tip force, when
neglecting shear deformation of cross-section, the buckling load parameter  was
evaluated for axially distributed loading is known based on the Euler–Bernoulli beam
theory. In the present analysis, if shear deformation of cross-section is taken into
account, the buckling load parameter  is determined for a special case of ¼ 2 =4

Table 5. Critical buckling load parameter for tapered standing

columns with m ¼ 3; n ¼ 1.


 0 0.1 0.3 0.5 0.7 0.9999
0 7.8373 7.9477 8.2280 8.6390 9.3285 13.183
0.001 7.8083 7.9190 8.2001 8.6118 9.3016 13.147
0.01 7.5547 7.6684 7.9555 8.3724 9.0647 12.834
0.05 6.5705 6.6934 6.9975 7.4277 8.1198 11.559
0.1 5.5976 5.7251 6.0362 6.4670 7.1419 10.197

Table 6. Critical buckling load parameter for tapered standing

columns with m ¼ 4; n ¼ 2.


 0 0.1 0.3 0.5 0.7 0.9999

0 7.8373 8.4143 9.8886 12.053 15.626 30.520

0.001 7.8083 8.3832 9.8518 12.008 15.567 30.377
0.01 7.5547 8.1105 9.5306 11.615 15.051 29.128
0.05 6.5705 7.0549 8.2914 10.102 13.069 24.424
0.1 5.5976 6.0150 7.0778 8.6281 11.142 19.995

1550017-9
 D.-L. Sun, X.-F. Li & C. Y. Wang

Table 7. Critical buckling load parameter  for tapered

standing columns with ¼ 2 =4, m ¼ 4; n ¼ 2.


 0 0.2 0.4 0.6 0.8
0 1.7206 1.4029 1.0418 0.64538 0.24393
0.001 1.7142 1.3980 1.0382 0.64316 0.24297
0.01 1.6578 1.3547 1.0073 0.62375 0.23455
0.05 1.4308 1.1828 0.88589 0.54848 0.20242
0.1 1.1896 1.0041 0.76268 0.47397 0.17157

based on the Timoshenko beam theory and the evaluated results are listed in
Table 7. The critical buckling load parameter for  ¼ 0 is calculated by taking  ¼
10 6 in the Timoshenko beam theory. Obviously, our results for  ¼ 10 6 are in
Int. J. Str. Stab. Dyn. Downloaded from www.worldscientific.com
by MONASH UNIVERSITY on 10/23/15. For personal use only.

excellent agreement with those obtained when adopting the Euler–Bernoulli beam
theory in Ref. 11. This is easily understood since Timoshenko columns reduce to
Euler–Bernoulli columns when  ! 0. Nevertheless, with  increasing, the critical
buckling load reduces signi¯cantly, regardless of the taper ratios. This indicates that
the obtained buckling load based on the Euler–Bernoulli beam theory is indeed
overestimated as compared to that based on the Timoshenko beam theory, as
expected. The e®ect of shear deformation cannot be simply neglected and must be
considered in the analysis. On the other hand, for a given value of , the increase of 
leads to a lower critical buckling load. This is to say that the carrying-load capacity
of a sharper coned column is weaker, as expected.
Of the other three cases of combined m and n, we also calculate the critical
buckling load parameter  and numerical results are listed in Tables 8–10. From
Tables 8–10, the critical buckling load parameter has a similar trend as that ob-
served in Table 7. The critical buckling load parameter  takes di®erent values,
depending on the geometry of the column and the taper ratio, although it takes the
same value for a uniform column with di®erent values of m and n. Moreover, the
critical load parameter  decreases with  increasing for all cross-sections under
consideration. Also, it is again seen that with the increase of the values of , the
critical buckling load parameter  decreases.

Table 8. Critical buckling load parameter  for tapered

standing columns with ¼ 2 =4, m ¼ 1, n ¼ 1.


 0 0.2 0.4 0.6 0.8

0 1.7206 1.7001 1.6654 1.6108 1.5226

0.001 1.7142 1.6937 1.6590 1.6040 1.5146
0.01 1.6578 1.6375 1.6021 1.5443 1.4457
0.05 1.4308 1.4145 1.3804 1.3179 1.1950
0.1 1.1896 1.1834 1.1580 1.1008 0.97104

1550017-10
 Buckling of Standing Tapered Timoshenko Columns

Table 9. Critical buckling load parameter  for tapered

standing columns with ¼ 2 =4, m ¼ 2; n ¼ 0.


 0 0.2 0.4 0.6 0.8
0 1.7206 1.4521 1.1755 0.88710 0.57838
0.001 1.7142 1.4471 1.1717 0.88454 0.57687
0.01 1.6578 1.4028 1.1387 0.86170 0.56337
0.05 1.4308 1.2220 1.0014 0.76519 0.50502
0.1 1.1896 1.0253 0.84805 0.65387 0.43488

Table 10. Critical buckling load parameter  for tapered

standing columns with ¼ 2 =4, m ¼ 3, n ¼ 1.
Int. J. Str. Stab. Dyn. Downloaded from www.worldscientific.com
by MONASH UNIVERSITY on 10/23/15. For personal use only.

 0 0.2 0.4 0.6 0.8

0 1.7206 1.4364 1.1414 0.82892 0.48133
0.001 1.7142 1.4315 1.1379 0.82662 0.48020
0.01 1.6578 1.3879 1.1063 0.80631 0.47021
0.05 1.4308 1.2120 0.97884 0.72407 0.42961
0.1 1.1896 1.0250 0.84310 0.63634 0.38615

Another interesting fact is the dependence of the critical buckling load parameter
 on axially distributed loading . Figures 3–6 show the critical buckling load pa-
rameter  against axially distributed loading for four di®erent geometries with
 ¼ 0:05. Note that axially distributed loading < 0 corresponds to a hanging col-
umn, whereas > 0 corresponds to a standing column. In other words, gravity may

Fig. 3. Buckling load parameter  against with m ¼ n ¼ 1 for di®erent taper ratios where solid lines
correspond to Timoshenko columns with  ¼ 0:05, dashed lines to Euler–Bernoulli columns with  ¼ 0.

1550017-11
 D.-L. Sun, X.-F. Li & C. Y. Wang
Int. J. Str. Stab. Dyn. Downloaded from www.worldscientific.com
by MONASH UNIVERSITY on 10/23/15. For personal use only.

Fig. 4. Buckling load parameter  against with m ¼ 2; n ¼ 0 for di®erent taper ratios where solid lines
correspond to Timoshenko columns with  ¼ 0:05, dashed lines to Euler–Bernoulli columns with  ¼ 0:

be equivalent to axial compressive loading for a standing column, and tensile loading
for a hanging column, respectively. For comparison, we also depict the corresponding
buckling load parameter  against axially distributed loading when the Euler–
Bernoulli beam theory is used, which are shown as dashed lines in Figs. 3–6.
It is observed that an increase of the load parameter causes an obvious decline of
the critical buckling load parameter , irrespective of which case. By inspecting
Fig. 4 for the case of m ¼ 2; n ¼ 0, we ¯nd that arbitrary two buckling load curves do
not intersect. Nonetheless, for the other three geometries, any two buckling load

Fig. 5. Buckling load parameter  against with m ¼ 3; n ¼ 1 for di®erent taper ratios where solid lines
correspond to Timoshenko columns with  ¼ 0:05, dashed lines to Euler–Bernoulli columns with  ¼ 0:

1550017-12
 Buckling of Standing Tapered Timoshenko Columns
Int. J. Str. Stab. Dyn. Downloaded from www.worldscientific.com
by MONASH UNIVERSITY on 10/23/15. For personal use only.

Fig. 6. Buckling load parameter  against with m ¼ 4; n ¼ 2 for di®erent taper ratios where solid lines
correspond to Timoshenko columns with  ¼ 0:05, dashed lines to Euler–Bernoulli columns with  ¼ 0.

curves would intersect at a certain value 0 in the range of 0 to 10 for the taper ratios
 < 0:8. When is less than 0 , the higher the  value, the smaller the critical
buckling load parameter , while when is greater than 0 , the trend is completely
opposite. It is noted that when  ¼ 0, the critical buckling load parameter  takes
the same values for all four di®erent geometries.

5. Conclusion
Presented herein is a new approach for determining the critical load of buckling of a
standing nonuniform column subjected to axially distributed force and tip force
within the framework of the Timoshenko beam theory. With the help of boundary
conditions at the two ends, we transform the boundary value problem associated
with the stability of Timoshenko columns into an initial value problem. Then the
critical buckling load can be determined using the Runge–Kutta method to solve the
resulting initial value problem of the governing di®erential equations. The e®ec-
tiveness of the suggested method was con¯rmed by comparing our numerical results
with the previous ones for tapered Euler–Bernoulli columns subjected to axial
loading and tip force simultaneously. New results of the buckling loads for standing
or hanging tapered Timoshenko columns were further presented for four typical
geometries of columns. The dependence of the buckling load is elucidated for various
cases of interest.

Acknowledgments
This work was supported by the Faculty Research Foundation of Central South
University, P. R. China (Grant No. 2013JSJJ020).

1550017-13
 D.-L. Sun, X.-F. Li & C. Y. Wang

References
1. S. P. Timoshenko and J. M. Gere, Theory of Elastic Stability (Courier Dover Publications,
New York, 2009).
2. I. Elishako®, A selective review of direct, semi-inverse and inverse eigenvalue problems for
structures described by di®erential equations with variable coe±cients, Arch. Comput.
Meth. Eng. 7 (2000) 451–526.
3. C. M. Wang, C. Y. Wang and J. N. Reddy, Exact Solutions for Buckling of Structural
Members (CRC Press, Boca Raton, Florida, 2004).
4. L. Euler, Determinatio onerum, quae columnae gestare valent, Acta Academiae Scien-
tiarum Petropolitanae 1 (1778) 121–145 (in Latin).
5. R. Frisch-Fay, On the stability of a strut under uniformly distributed axial forces, Int.
J. Solids Struct. 2 (1966) 361–369.
6. Q. S. Li, H. Cao and G. Q. Li, Stability analysis of bars with varying cross-section, Int.
J. Solids Struct. 32 (1995) 3217–3228.
Int. J. Str. Stab. Dyn. Downloaded from www.worldscientific.com
by MONASH UNIVERSITY on 10/23/15. For personal use only.

7. H. H. Vaziri and J. Xie, Buckling of columns under variably distributed axial loads,
Comput. Struct. 45 (1992) 505–509.
8. F. Arbabi and F. Li, Buckling of variable cross-section columns: Integral-equation
approach, J. Struct. Eng. 117 (1991) 2426–2441.
9. Z. Elfelsou¯ and L. Azrar, Buckling, °utter and vibration analyses of beams by integral
equation formulations, Comput. Struct. 83 (2005) 2632–2649.
10. Y. Huang and X.-F. Li, Buckling analysis of nonuniform and axially graded columns with
varying °exural rigidity, J. Eng. Mech. 137 (2011) 73–81.
11. X.-F. Li, L.-Y. Xi and Y. Huang, Stability analysis of composite columns and parameter
optimization against buckling, Compos. Part B Eng. 42 (2011) 1337–1345.
12. C. Y. Wang, Stability of a braced heavy standing column with tip load, Mech. Res.
Commun. 37 (2010) 210–213.
13. D. J. Wei, S. X. Yan, Z. P. Zhang and X.-F. Li, Critical load for buckling of non-prismatic
columns under self-weight and tip force, Mech. Res. Commun. 37 (2010) 554–558.
14. X.-F. Li, L.-Y. Xi and Y. Huang, Buckling load of tapered ¯bers subjected to axially
distributed load, Fiber Polym. 11 (2010) 1193–1197.
15. C. Y. Wang, In°uence of gravity and taper on the vibration of a standing column, Adv.
Appl. Math. Mech. 4 (2012) 483–495.
16. W. H. Duan and C. M. Wang, Exact solution for buckling of columns including self-
weight, J. Eng. Mech. 134 (2008) 116–119.
17. Y. Huang and X.-F. Li, An analytic approach for exactly determining critical loads of
buckling of nonuniform columns, Int. J. Struct. Stab. Dynam. 12 (2012) 1250027.
18. S. Barasch and Y. Chen, On the vibration of a rotating disk, J. Appl. Mech. 39 (1972)
1143–1144.

1550017-14