International Journal of Non-Linear Mechanics 77 (2015) 85–95

Contents lists available at ScienceDirect

International Journal of Non-Linear Mechanics

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

Large-deflection and postbuckling of beam-columns with non-linear

semi-rigid connections including shear and axial effects
Oliver Giraldo-Londoño a, J. Sebastian Monsalve-Giraldo b, J. Dario Aristizabal-Ochoa c,n
a
Department of Civil Engineering, University of Illinois at Urbana-Champaign, USA
b
Department of Civil Engineering, Federal University of Rio de Janeiro, Brazil
c
School of Mines, National University of Colombia at Medellín, Calle 9c #15-165, Medellin, Colombia

art ic l e i nf o a b s t r a c t

Article history: The non-linear large deflection-small strain analysis and post-buckling behavior of an out-of-plumb
Received 16 April 2015 Timoshenko beam-column of symmetrical cross section subjected to end loads (forces and moments)
Received in revised form with non-linear bending connections at both ends, and its top end partially restrained against transverse
16 June 2015
and longitudinal translations are developed in a classical manner. A set of non-linear equations based on
Accepted 28 July 2015
the “modified shear equation” that includes the effects of (1) shear deformation and the shear
Available online 6 August 2015
component of the applied axial forces; and (2) shortening of the beam-column due to both axial forces
Keywords: and “bowing” are presented. The proposed method and corresponding equations can be used in the large
Beam-columns deflection-small strain analysis of Timoshenko beam-columns with non-linear bending connections, as
Large deflections
well as lateral and longitudinal non-linear restraints at the top end. This paper is an extension of
Post-buckling
previous work presented by the senior author on the large deflection and post-buckling behavior of
Non-linear analysis
Non-linear connections Timoshenko beam-columns with linear elastic semi-rigid connections and linear elastic lateral bracing.
Stability Three comprehensive examples are included that show the effectiveness of the proposed method and
corresponding equations. Results obtained in the three examples are verified against analytical solutions
available in the technical literature and against results from models using the FEM program ABAQUS.
& 2015 Elsevier Ltd. All rights reserved.

1. Introduction
The non-linear large deflection-small strain analysis and post-
buckling behavior of beam-columns is of great importance in
structural engineering, especially for members made of materials
with high resilience and low shear stiffness as composite materi-
als. For structures made of this type of materials, the combined
influence of lateral deflections along the element chord (i.e., P–δ
effects), the relative drift between the ends of the member (i.e.,
P–Δ effects), axial shortening due to both axial loads and “bow-
ing”, softening of the end connections and lateral bracing system,
produce a significant non-linear behavior that must be considered
in the analysis of these structures.
In the past several years, a significant number of studies have
been conducted to understand the large deflection and post-
buckling behavior of shear deformable columns. Different applica-
tions where these theories are used involve, for instance, the
analysis of nano- and micro-beams [1,2], and laced built-up
columns [3]. Some of these studies and additional applications
are summarized next.
n
Corresponding author.
E-mail address: jdaristi@unal.edu.co (J.D. Aristizabal-Ochoa).
Ansari et al. [1] developed a continuum model for the post-
buckling analysis of nanobeams incorporating the effects of surface
stress. In their formulation, different boundary conditions were
taken into account (e.g., simply-supported, and clamped boundary
conditions). Although their formulation is based on a continuum
model, their solution involves a discretization in the domain of the
beam, which is then solved via a pseudo arch-length method.
Chaterjee and Pohit [2] investigated the static and dynamic
large deflection behavior of electrostatically actuated micro-
cantilever beams, where geometric and inertial non-linearities
were incorporated. The solution of the governing equations was
approximately solved using the Galerkin method.
Gantes and Kalochairetis [3] developed a framework for the
approximate second order analysis of an imperfect Timoshenko
beam-column based on the finite element method. This research
focused on the analysis of laced built-up columns, in which effects of
shear deformations are significantly important in the stability analysis.
Hjiaj et al. [4] developed a non-linear finite element formula-
tion for the large-displacement analysis of shear-deformable two-
layer beams. In this formulation, interlayer slip was allowed and
transverse shear deformation of each layer was incorporated.
According to their results, deflections in short beams were
influenced by shear deformations, obtaining an increase in deflec-
tions of 14% for a length-to-depth ratio of six.

http://dx.doi.org/10.1016/j.ijnonlinmec.2015.07.009
0020-7462/& 2015 Elsevier Ltd. All rights reserved.
86 O. Giraldo-Londoño et al. / International Journal of Non-Linear Mechanics 77 (2015) 85–95

Nomenclature

A and As cross-sectional area and effective shear area of beam-column, respectively

E and G elastic and shear modulus of the material respectively
Fuh ultimate load capacity of the lateral bracing at the column top end A
Fuv ultimate load capacity of the axial restraint at the column top end A
Hz “bowing” bending functions given by Eqs. (5) and (6)
h initial height of the beam-column AB
hr chord length of the beam-column AB after “bowing” caused by bending and shear actions
I moment of inertia of the beam-column cross-section about its principal axis of bending
kh and kv stiffness of the lateral and axial linear restraints at end A, respectively
khi and kvi initial stiffness of the lateral and axial linear restraints at end A, respectively
M external bending moment applied at A
Ma and Mb end moments of beam-column at A and B respectively
Mua and Mub ultimate moment applied at ends A and B, respectively
na and nb shape parameter of the non-linear bending connections at A and B, respectively
nh and nv shape parameter for the lateral and axial non-linear restraints at end A, respectively
P applied vertical load at A (þ compression,  tension)
Pe π2EI/h2 ¼Euler load
Ph transverse force developed in the non-linear bracing due to lateral deflection of end A of the beam-column
Pt total axial load along the cord AB ( þ compression,  tension)
Pv vertical force in the non-linear restraint due to vertical deflection at end A of the beam-column
Q applied transverse load at A
u(x) lateral deflection of the column center line at x
ψ(x) rotation of cross section of the beam-column at x
κ ia and κ ib initial stiffness of the non-linear bending connections at ends A and B, respectively
ρia and ρib initial fixity factors of the non-linear bending connections at ends A and B, respectively
θ0 initial out-of-plumb angle of the cord AB with respect to the vertical axis (Fig. 1a)
θ angle of the cord AB with respect to the vertical axis after loading (Fig. 1a)
ψa and ψb rotations of cross sections at A and B due to bending, respectively
Δh and Δv horizontal and vertical deflections of top end A of beam-column AB, respectively

Nguyen and Gan [5] proposed a finite element model for the
large deflection analysis of tapered functionally graded beams. The
large deflection analysis was performed using a Newton–Raphson
technique combined with the arch-length method. Santos and Gao
[6] presented a canonical dual mixed finite element method for
the post-buckling analysis of elastic beams undergoing large
deformations.
Yu et al. [7] presented the numerical and analytical approx-
imate solutions for the electromechanical post-buckling analysis of
an actuator fixed at the ends and subjected to a symmetric
electrostatic field. The numerical and analytical results showed
good agreement for a large range of beam deflections.
Ansari et al. [8] studied the thermal post-buckling behavior of
functionally graded Timoshenko microbeams. Stability equations
were derived based on the Timoshenko beam theory and the
method of virtual work. The effects of imperfections were also
discussed, and the behavior of a theoretically perfect beam was
compared with the behavior of a beam with initial imperfections.

Fig. 1. Beam-column with connections at both ends and with initial out-of-plumb θo: (a) structural model; (b) end forces, moments and rotations; and (c) deformed
differential element.
 O. Giraldo-Londoño et al. / International Journal of Non-Linear Mechanics 77 (2015) 85–95
Smith-Pardo and Aristizabal-Ochoa [9] proposed a set of
closed-form equations to analyze the pre-buckling and post-
buckling behavior of a prismatic beam-column with an initial
camber and subject to transverse loads. The equations derived by
Smith-Pardo and Aristizabal-Ochoa are only valid for simply
supported beam-columns with an axial spring at one of its ends.
Smith-Pardo and Aristizabal-Ochoa [9] highlighted the influence
of the axial restraints magnitude on the pre-buckling and post-
buckling stability behavior.
Aristizabal-Ochoa [10] studied the non-linear large deflection-
small strain elastic behavior of slender beam-columns of symme-
trical cross section with linear semi-rigid end connections. Results
from this investigation were within 1.5% of the theoretical results
(e.g., results using elliptical equations) for very large transverse
displacements. Aristizabal-Ochoa [11] also developed a formula-
tion capable to predict the large-deflection behavior of slender
beam-columns considering linear connections requiring the solu-
tion of complex equations with elliptical integrals. However, in this
formulation the effects of axial and shear deformations are not
taken into consideration. In addition, Aristizabal-Ochoa [12] devel-
oped a set of equations to study the large deflection-small-strain
behavior of columns including the combined effects of shear, axial
and bending deformations, axial load eccentricities, lateral bracing
and initial out-of-plumbness. He showed that the contribution of
shear deformations is important in structures made of laminated
composite materials or elastomeric bearings that usually are
implemented for seismic isolation or in bridge supports where
the shear stiffness is relatively low. However, his model is limited
to linear elastic connections and lateral bracing.
Vega-Posada et al. [13] derived a set of highly non-linear
equations based on the theory of the “Elastica” capable of predict
the large-deflection and post-buckling behavior of laterally braced
or unbraced slender columns with linear semi-rigid connections at
both ends. This analysis involved the numerical solution of a set of
elliptical equations with multiple solutions, which can be tedious
to solve. Following the same methodology, Vega-Posada et al. [14]
studied the large-deflection and post-buckling behavior of slender
beam-columns of symmetrical cross sections with both supports
partially restrained against rotation and sidesway totally inhibited
or uninhibited subjected to end loads including the effects of out-
of-plumbness and non-linear connections. However, this model
only considered bending deformations neglecting the effects of
shear and axial deformations.
The current paper deals with beam-columns where shear
deformations play an important role. Thus, the shear component
of the applied axial load and its corresponding shear deformation
must be considered in the analysis [15]. To account for shear
effects, we adopt the “modified shear equation”, discussed by
Timoshenko and Gere [16]. In their book, Timoshenko and Gere
discussed two approaches: Engesser's (referred to as standard
approach) [17], and Haringx's (referred to as modified approach)
[18]. Here, Timoshenko and Gere state that Haringx method may
be more accurate for beam-columns where shear effects are large.
However, some authors question the use of Haringx theory for
some practical cases [19].
Blaauwendraad [19] tested the hypothesis by Timoshenko and
Gere for the cases of a Vierendeel beam and a sandwich beam. For
the case of the Vierendeel beam, the buckled shape was controlled
by shear deformations, and the cross sections remained parallel to
the undeformed configuration. Under these circumstances the
solution using the Haringx hypothesis led to a critical load that
made no physical sense and was not in agreement with numerical
results using FEM. The FEM results, however, showed to be in good
agreement with the prediction using the Engesser approach.
In contrast, Attard [20] used an appropriate strain energy
density for a compressive isotropic hyperelastic Hookean material
87
to derive the buckling load of an initially straight prismatic shear
deformable column including axial deformations. They showed
that the obtained formula for the critical load agreed well with the
one derived by Haringx. In the research by, Bažant [21], he stated
that shear-deformable beams of the type proposed by Haringx and
Engesser are, in principle, equivalent in the sense that they have
different shear stiffness, depending linearly on the applied axial
load. This statement is supported by Aristizabal-Ochoa [15], who
shows that the governing differential equation for this problem is
essentially the same for both Engesser and Haringx approaches. As
shown in Section 2.2 in Ref. [15], the only difference between the
equations based on the Engesser and Haringx approaches lies on a
single parameter ϕ defining the deflection of the centroid of the
beam-column. Based on this assumption, the equations derived in
this paper, even if based on the “modified shear equation” by
Haringx, can be adequately rewritten in terms of the standard
shear equation by Engesser by just rewriting the parameter ϕ as
will be shown later in Section 3 of this paper.
The main objective of this paper is to propose a set of equations
capable of predicting accurately and efficiently the non-linear large-
deflection small-strain analysis and post-buckling behavior of a
Timoshenko beam-column of symmetrical cross section considering
the effects of non-linear connections and non-linear transverse and
axial restraints located at the top end. This paper is an extension of
a publication previously presented by the senior author [12] on the
large deflection and postbuckling behavior of Timoshenko beam-
columns with linear elastic semi-rigid connections and linear elastic
lateral bracing. The proposed method also includes the effects of
shear and longitudinal deformations of the beam-column. An
iterative method to solve the equation system of non-linear large
deflection-small strain analysis is proposed in this paper. Three
comprehensive examples are presented and discussed in detail
showing the effectiveness and accuracy of this algorithm.
2. Structural model
2.1. Assumptions
Consider the 2D prismatic beam-column that connects points A
and B shown in Fig. 1(a). The element is made up of the beam-
column itself AB, and the non-linear flexural and translational
connections at ends A and B. The non-linear bending connections
at ends A and B are defined by the three-parameter power model
[22,23] as shown in Fig. 2. These connections are defined by their
initial stiffness κia, and κib, the ultimate moment Mua, and Mua, and
shape factors na, and nb (where the subscripts “a” and “b” refer to
the connection parameters at ends A and B, respectively). The non-
linear translational connections can be defined following any
linear or non-linear force–displacement relationship.
End n=
Moment Mu
n= 4
M*
n= 2
n=1
M*
ψ* =
i κi [1 − ( M * / M u ) n ]1 / n
1
Mu i
End Rotation *
Fig. 2. Three-parameter power model used for the end bending connections.
88 O. Giraldo-Londoño et al. / International Journal of Non-Linear Mechanics 77 (2015) 85–95

For convenience the following two parameters are introduced: deflection curves for the frame are non-linear, which can be seen as a
non-linear lateral brace as proposed in the current work.
1 1 In addition, non-linear lateral bracing is used in offshore
ρia ¼ ; and ρib ¼ ð1a; bÞ
1 þ 3EI=h
κ ia 1 þ 3EI=h
κ ib
engineering, for example in an articulated tower. Some authors
such as Fehrenkamp et al. [27], Wishahy and Arockiasamy [28],
where ρia and ρib will be called the initial fixity factors. The fixity and Brynjolfsson [29] suggest modeling the cables as non-linear
factors can vary from 0 (for perfectly hinged connections) to 1 (for springs. In this way, for instance, a guyed tower (which is a lattice
perfectly clamped connections). For this reason, we believe that element restrained against horizontal displacement at the top by
the use of fixity factors is more convenient (instead of the initial means of a mooring line system) can be represented using the
stiffness) in the elastic analysis of structures with semirigid model proposed in this paper.
connections [9–15].
It is assumed that the member AB: (1) bends about one of the
3. Proposed equations
principal axis of its cross section with a moment of inertia I, cross
sectional area A, effective shear area As and span h; (2) is made of a
The large deflection-small strain analysis of a Timoshenko
homogeneous linear elastic material with Young and shear moduli E
beam-column is developed in the plane of bending about one of
and G, respectively; (3) its centroidal axis is a perfect straight line with
the principal axes. By satisfying the condition of moment equili-
an initial out-of-plumbness θ0 with respect to the vertical axis; and
brium of the beam-column shown in Fig. 1(b) about B, the
(4) is loaded in the plane of bending at the top end A with P (vertical
following expression is obtained:
load), Q (horizontal load) and M¼ PeþM0 (where M0 is an additional
moment applied at A and e is the eccentricity of the applied axial load M a þM b
hr
þ ðP  P v Þ sin θ þ ðQ  P h Þ cos θ ¼0 ð2Þ
P). Each one of these three loads can vary independently and are h h
applied in the plane of bending. The lateral and axial restraints at A are The ratio between the chord length of the beam-column AB
provided by two independent springs (linear or non-linear) that after “bowing” and the initial length of the beam-column accord-
induce vertical and horizontal forces as a function of the vertical and ing to Ekhande [24] is as follows:
horizontal displacements of the top end A, as shown in Fig. 1a. " #
Notice that non-linear translational and rotational connections hr Pt Hz
¼ 1  2 2 ð3Þ
are of great importance in analyzing structures undergoing h EA 4P h
t
moderate-to-large deformations. Non-linear bending connections,
for example, are important in a variety of engineering applications. where: P t ¼ ðP P v Þ cos θ þ ðP h  Q Þ sin θ
One example of the applicability of such connection behavior is By substituting Eq. (3) into Eq. (2), the following non-linear
found in metallic submarine risers, which are commonly used in equation is obtained:
!
offshore oil floating units [25]. This type of structure is connected M a þM b
 Pt Hz
to the vessel by means of a flex joint whose stiffness varies with þ ðP  P v Þ sin θ þ ðQ  P h Þ cos θ 1   2 2 ¼ 0 ð4Þ
h EA 4P h
t
the rotation. Such connections are known for exhibiting high non-
linearity even for small rotations. To perform the structural Eq. (4) represents the condition of moment equilibrium of the
analysis of this type of structures, a stiffness-rotation curve must beam-column shown in Fig. 1b about the bottom end, and Eq. (3)
be used to model the rotational spring that represents this flex the geometric condition of the actual distance between the two
joint. This curve is, in general, experimentally calibrated to extremes of the beam-column that includes the effects of the axial
represent a realistic behavior of the connection. strain and the “bowing” caused by its bending and shear deforma-
Another engineering application of the non-linearity of the tions. Notice that the axial load Pt is taken as the sum of the actual
rotational connections can be found in beam-to-column connections. components of the vertical and horizontal forces at the top end A
Chen et al. [22] discussed the moment–rotation behavior of a variety along its chord. As proposed by Aristizabal-Ochoa [12], the
of commonly used semi-rigid connections (e.g. T-stub, end plate, top magnitude of the “bowing” factor Hz depend on the actual sign
and seat angle, header plate, double web angle, and single web of Pt and the magnitude of the end moments of the beam-column
angle). As mentioned previously, the three-parameter power model Ma and Mb, as follows:
is used in the paper to model the non-linear semirigid connection.
This is a well-known method developed by Richard and Abbot [23] to (a) For P t Z 0 (compression) and β 4 0 or P t o 0 (tension) and
represent the moment–rotation behavior of beam-to-column con- β o 0 (i.e., P t β Z 0):
nections. The shape parameter n can be determined by using the   2  β  ϕ

method of least squares for the differences between the predicted H z ¼ βϕ M 2a þ M 2b cos ϕ þ
sin ϕ sin ϕ
moments and the experimental data. Chen et al. [22], give empirical   
2β ϕ
equations of the shape parameter n for four types of beam-columns þ 2βϕM a M b 1þ  2ðM a þ M b Þ2 ð5Þ
sin ϕ tan ϕ
connections. In their book, Chen et al. state that the three-parameter
power model is an effective and accurate tool for designers to
perform second-order non-linear structural analysis. (b) For P t o 0 (tension) and β 40 (i.e., P t β o 0):
Furthermore, non-linear translational end springs following a   2  β  ϕ

specific non-linearity rule lead to a better representation of the H z ¼ βϕ M 2a þ M 2b cosh ϕ þ
sinh ϕ sinh ϕ
lateral bracing system of a structure. For example, the design on steel   
2β ϕ
frames often includes lateral bracing systems capable to dissipate þ 2βϕM a M b 1þ  2ðM a þ M b Þ2 ð6Þ
sinh ϕ
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi tanh ϕ
energy when the structure is subject to severe lateral loads. As an  ffi

where ϕ ¼ P t = βEI=h and β ¼ 1= 1 þ P t =ðGAs Þ .
2
example, Aristizabal-Ochoa proposed a practical approach (known as
DKB) for the lateral bracing of single and multistory steel frames [26].
This system is composed of two elements: the knee and the diagonal These equations, even if based on the “modified shear equation”
brace elements. The knee is a fuse-like element designed to deform by Haringx, can be rewritten in terms of the Engesser shear equation
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 ffi
inelastically to dissipate energy. Due to this inelasticity, load–lateral 0 0
by substituting ϕ ¼ P t = β EI=h , where β ¼ ð2β  1Þ=β [15].
2
 O. Giraldo-Londoño et al. / International Journal of Non-Linear Mechanics 77 (2015) 85–95 89

M a sin ϕ  βϕ M b sin ϕ  βϕ cos ϕ

As discussed above, the non-linear moment–rotation relationship þ þ θ  θ0
of the end connections at A and B is described by the three-parameter EI=h βϕ2 sin ϕ EI=h βϕ2 sin ϕ
power model [22,23] as follows: M b 1  ρib
þ h i ¼0 ð12bÞ
n 1=nb
Mn 3ρib 1  M b =M ub b EI=h
ψ ¼ h
n
i1=n ð7Þ
κ i 1  Mn =Mu n
(b) For P t o0 (tension) and β 40 (i.e., P t β o0):
where ψ n is the rotation of the non-linear end connection induced by
the moment M n . M a sinh ϕ  βϕ cosh ϕ M b sinh ϕ  βϕ
  þ θ  θ0
Notice that a perfectly linear-elastic end connection and a EI=h βϕ2 sinh ϕ EI=h βϕ2 sinh ϕ
perfectly elasto-plastic end connection can be modeled using ðM a  M Þ 1  ρia
expression (7) simply by making M u ¼ 1 and n ¼ 1, respectively. þ h i ¼0 ð13aÞ
na 1=na
Applying Eq. (7) at ends A and B: 3ρia 1  ðM a  M Þ=M ua EI=h

M na M nb M a sinh ϕ  βϕ M b sinh ϕ  βϕ cosh ϕ

ψ na ¼ h i ; and ψ nb ¼ h i   þ θ  θ0
na 1=na nb 1=nb
κ ia 1  M a =Muan
κ ib 1  Mnb =Mub EI=h βϕ2 sinh ϕ EI=h βϕ2 sinh ϕ
ð8a; bÞ M b 1  ρib
þ h i ¼0 ð13bÞ
nb 1=nb
which in terms of fixity factors ρia and ρib become as follows: 3ρib 1  M b =M ub EI=h

M na 1  ρia M nb 1  ρib
ψ na ¼ h i1=na ; ψ nb ¼ h i
nb 1=nb The set of Eqs. (4), (12a,b) and (13a,b) is used in this paper for
3ρia 1  M na =M ua
na
EI=h 3ρib 1  M nb =M ub EI=h
the large deflection-small strain analysis of Timoshenko beam-
ð9a; bÞ columns with non-linear connections at both ends and the
Notice that bending moments in the end connections are non-linear restraints at the top end A subjected to concen-
M na ¼ M a  M and M nb ¼ M b . trated loads P, Q and M at the top end (Fig. 1a).
On the other hand, the bending rotations of the cross sections at
ends A and B are formulated using the “modified shear equation” Note: The proposed paper deals with loads applied only at the
proposed by Timoshenko and Gere [16, p. 134]. This approach has two ends. However, modifications to the equations can be done to
been utilized by Aristizabal-Ochoa [12] in the large deflection-small incorporate effects of loads applied through the span (e.g., dis-
strain analysis of Timoshenko beam-columns. Thus, using the tributed loads). This can be achieved by reformulating Eqs. (18)–
expressions derived by Aristizabal-Ochoa [12], the bending rota- (27) in Appendix A of the paper by Aristizabal-Ochoa [12] to
tions of cross section of the beam-column at A and B are as follows: incorporate these effects.
It is pertinent to mention that the proposed equations are valid for
(a) For P t Z 0 (compression) and β 4 0 or P t o 0 (tension) and the large deflection-small strain analysis of Timoshenko beam-
β o 0 (i.e., P t β Z 0): columns with non-linear semirigid connections and non-linear lateral
and axial end restraints. Hence, the equations presented herein are
M a sin ϕ  βφ cos ϕ Mb sin ϕ  βϕ Δ
ψa ¼ þ þ ¼  ψ na ð10aÞ valid only if the beam-column itself behaves elastically during the
EI=h βϕ2 sin ϕ EI=h βϕ2 sin ϕ h
entire range of deformation. One might think that the two assump-
tions (i.e., the beam-column behaving in an elastic manner and the
M a sin ϕ  βϕ M b sin ϕ  βϕ cos ϕ Δ
ψb ¼ þ þ ¼  ψ nb connections behaving inelastically) are in some way, contradictory.
EI=h βϕ2 sin ϕ EI=h βϕ2 sin ϕ h
However, the authors believe that these assumptions have physical
ð10bÞ meaning. The reason behind these assumptions arises in the fact that
even if the structure is subject to deformations comparable to the
(b) For P t o 0 (tension) and β 4 0 (i.e., P t β o 0): member length, the strains in the structural element can be still small.
Thus, the element itself can behave elastically while the connections
M a sinh ϕ  βϕ cosh ϕ M b sinh ϕ  βϕ Δ
ψa ¼   þ ¼  ψ na (which are undergoing large deformations) can behave inelastically.
EI=h βϕ2 sinh ϕ EI=h βϕ2 sinh ϕ h
ð11aÞ
4. Proposed iterative method
M sinh ϕ  βϕ M b sinh ϕ  βϕ cosh ϕ Δ
ψb ¼  a  þ ¼  ψ nb
EI=h βϕ2 sinh ϕ EI=h βϕ2 sinh ϕ h The following iterative procedure is suggested for evaluate the large
deflection-small strain analysis of the Timoshenko beam shown in Fig. 1a.
ð11bÞ
where Δ=h ¼ θ  θ0 1. Knowing the column properties (E, G, A, As, I, h), the properties
of end connections at A and B (i.e., the values of ρai, Mau, na,
Now, substituting expressions (9a,b) into Eqs. (10a,b) and (11a,b) and ρbi, Mbu, nb), the value of θ0 , and end forces P, Q and M,
gives: assume an initial trial value for θ.
2. Calculate the value of Pt iteratively until convergence, accord-
(a) For P t Z 0 (compression) and β 4 0 or P t o 0 (tension) and ing to the equation shown below:
β o 0 (i.e., P t β Z0):
P t; k þ 1 ¼ ðP  P v; k Þ cos θ þ ðP h; k  Q Þ sin θ with k ¼ 1; 2; 3:…
M a sin ϕ  βϕ cos ϕ M b sin ϕ  βϕ
þ þ θ  θ0
EI=h βϕ2 sin ϕ EI=h βϕ2 sin ϕ
where P v; k ¼ kv Δv; k for axial restraint at A, or P v; k ¼
ðM a  M Þ 1  ρia kvi Δv;
þ h i ¼0 ð12aÞ h k
nv
i1=nv for non-linear axial restraint at A (e.g., using
kvi Δv; k
na 1=na
3ρia 1  ðM a  M Þ=M ua EI=h 1þ F uv
90 O. Giraldo-Londoño et al. / International Journal of Non-Linear Mechanics 77 (2015) 85–95

h   i
a power model); Δv; k ¼ cos θ0  hhr cos θ h; the column shortening equation. This can be achieved by setting
k Hz ¼ 0 wherever it is used [e.g., in Eqs. (3)–(6)].
P h;k ¼ kh Δh;k for a linear lateral brace at B, or P h; k ¼
khi Δh;
h  k nh i1=nh for non-linear lateral brace at A (e.g., using a 5. Comprehensive examples and verification
khi Δh; k
1þ F uh

power model); 5.1. Example 1: Large-deflection analysis of a cantilever beam with a

    non-linear connection at end B (step-by-step calculations)
hr hr P H
Δh; k ¼ sin θ  sin θ0 h; ¼ 1  t; k  2z; k 2 ;
h k h k EA 4P h
t; k The cantilever beam shown in Fig. 3a is subjected to a vertical
  2  β  
ϕk

load P at end A. Assume that: (1) the member has geometric
H z; k ¼ βk ϕk M 2a; k þ M 2b; k k
cos ϕk þ
sin ϕk sin ϕk properties A, As, I, and h; (2) material properties G and E; (3) a non-
   linear connection at end B, according to the three-parameter
2  βk ϕk
þ 2βk ϕk M a; k M b; k 1þ power model with parameters Mu ¼EI/h, ρi ¼0.25 and n ¼2. Also
sin ϕk tan ϕk
assume that bending occurs about its major axis. Determine the
 2 M a; k þ M b; k when P t; k βk Z0 or
2
horizontal and vertical deflections of this beam and compare the
  2  β  ϕk

calculated results with those reported by Aristizabal-Ochoa [11]
H z; k ¼ βk ϕk M 2a; k þ M 2b; k k
cosh ϕk þ
sinh ϕk sinh ϕk and validate the results with those obtained using ABAQUS.
  
2  βk ϕk Solution: To perform the analysis of the beam-column shown in
þ 2βk ϕk M a; k M b; k 1þ
sinh ϕk tanh ϕk Fig. 3a for a given set of values of E, G, A, As, I, h, θ0, and connections
characteristics at end B, the following steps are suggested.
 2 M a; k þ M b; k when P t; k βk o0 ;
2
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
  Step 1. Choose an initial value for θ.


ϕk ¼ P t; k = βk EI=h ; and β k ¼ 1= 1 þ P t; k =ðGAs Þ
2 Step 2. Calculate the values of P t ¼ P cos θ Q sin θ,
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 ffi


β ¼ 1= 1 þ P t =ðGAs Þ , and φ ¼ P t = βEI=h2 .
The values for M a; k and M b; k for each iteration k are obtained Step 3. Replace the value of Ma by zero in Eq. (13a,b) and use
through a numerical method (e.g., Bisection, Müller, Newton– the Newton–Raphson method to determine the value of Mb. This
Raphson or the Secant method). The Newton–Raphson method is, use the Newton–Raphson method to solve the following
is recommended and applied to Eqs. (12a,b) and (13a,b) and equation:
replacing ϕ and β by ϕk and βk . The value of k is increased until
P t; k þ 1  P t; k o tolerance, or until a maximum number of (a) For P t Z 0 (compression) and β 4 0 or P t o 0 (tension) and
iterations is reached. To accelerate convergence, the authors β o0:
suggest to assume that the beam-column acts as a rigid body M b sin ϕ  βϕ cos ϕ Mb 1  ρib
for k¼0 (i.e., for the first iteration). This is, assume that
θ  θ0 þ þ h i ¼0
EI=h βϕ2 sin ϕ 3ρib 1  M b =M ub
nb 1=nb
EI=h
P t;0 ¼ ðP  P v;0 Þ cos θ þ ðP h;0  Q Þ sin θ
ð14aÞ
where P v; 0 ¼ kv Δv; 0 , for a linear axial restraint at A, or
kvi Δv;
P v; 0 ¼ h 0
nv
i1=nv for a non-linear axial restraint at A; (b) For P t o0 (tension) and β 4 0:
kvi Δv; 0
1þ F uv
M b sinh ϕ  βϕ cosh ϕ M b 1  ρib
θ  θ0  þ h i ¼0
Δv; 0 ¼ cos θ0  cos θ h; EI=h βϕ2 sinh ϕ n 1=nb
3ρib 1  M b =M ub b EI=h
ð14bÞ
P h; 0 ¼ kh Δh; 0 for a linear lateral brace at A, or P h; 0 ¼
khi Δh;
h  0 nh i1=nh for a non-linear lateral brace at A; and Step 4. Check if Eq. (4) if fulfilled; otherwise choose a new
khi Δh; 0
1þ F hv value for θ and return to step 2. Use the Regula–Falsi method
Δh; 0 ¼ sin θ  sin θ0 h to determine the value of θ for the given applied loads and
3. Once a value for Pt is obtained, check if the condition of moment input data.
equilibrium of the beam-column [i.e., Eq. (4)] is fulfilled. Other- Step 5. The vertical and horizontal deflections at the tip of the
wise, choose a new value for θ and return to step 2. The authors beam are approximately as follows [12]:
" #
recommend using the Regula-Falsi method to determine the value
Pt Hz Δ
of θ for the given applied loads and input data. The Regula-Falsi Δh ¼ 1   2 2 h sin θ; and Δv ¼ h  h ð15a; bÞ
EA 4P h tan θ
method is suggested, since it can be adopted to converge to a t

single or multiple roots in a closed interval. This versatility is

convenient, since for a single value of Pt, the governing Eqs. (4),
(12a,b) and (13a,b) can be satisfied by more than one value of θ.
Note 1. For columns with no vertical and lateral restraints at the
top end A, an iterative procedure to obtain Pt is not necessary since
its value can be calculated directly as P t ¼ P cos θ  Q sin θ.
Note 2. The authors believe that it is necessary to include all
parameters in the proposed equations in order to fully capture the
behavior of the column under large deflections. However, a
reasonable simplification (since the method uses large-displace-
ments-small-strains) is to neglect the component of “bowing” in
A Finite Element (FE) model was created using the FEM program
ABAQUS to validate the results of the proposed method. The type of
elements used in the model was B23 (i.e., two-node cubic beam
elements in plane), and mesh was generated using 50 of these
elements. Material properties were assigned as linear elastic, with
properties E and G. Selecting B23 elements allowed deformations due
to bending only. The non-linear connection, on the other hand, was
simulated as a non-linear connector using a cloud of points corre-
sponding to the moment–rotation relationship of the connection.
Fig. 3c shows the variations of Δv/h and 1–Δh/h with P/(EI/h2)
for the cantilever beam shown in Fig. 3a. The results from the
 O. Giraldo-Londoño et al. / International Journal of Non-Linear Mechanics 77 (2015) 85–95 91

y
h
x

v
h

h
h
P
P
EI/h v
h

M
n /n
M i M / Mu
Mu

v h
h h

Fig. 3. Example 1. Large-deflection analysis of a cantilever beam: (a) structural model; (b) moment-rotationmoment–rotation curve of nonlinearnon-linear connection at B;
and (c) calculated load-deflectionload–deflection curves.

proposed model were obtained using the step-by-step procedure

explained above and using GAs ¼ 1, EA¼ 1, Q¼ 0, and θ0 ¼ π/2. P
Δh
Based on the assumption that the member behaves elastically, the
results are compared with those reported by Aristizabal-Ochoa A P
[11] and those obtained using ABAQUS. According to Fig. 3c, the Q Δv
results calculated with the proposed method were in accordance
with those by Aristizabal-Ochoa [11] and ABAQUS for the entire A Q
range of deflections. This indicated that results obtained from the E, G
proposed model are accurate and compare well with those A, A s
reported by Aristizabal-Ochoa [11] who used elliptic integrals in I, h
the large-deflection analysis.
θ0
5.2. Example 2: Large-deflection and post-buckling analyses of an
out-of-plumb cantilever column including the effects of shear
deformations and non-linear connections

Perform the large-deflection and post-buckling analyses of a κib , nb

cantilever column subjected to concentrated loads P and Q at the
top end A as shown in Fig. 4. Consider that the column has:
(1) elastic properties G and E; (2) geometrical properties A, As, I
and h; (3) an initial out-of-plumb θ0 ; and (4) a non-linear bending
connection at the bottom B with moment–rotation relationship
following the three-parameter power model with Mu ¼2EI/h,
ρi ¼ 0:5and n ¼2. Compare the results with those reported by
Aristizabal-Ochoa [12] for linear elastic connections.
Solution: The first part of the analysis consisted of studying the
large-deflection behavior of the beam-column of Fig. 4 subject to a
lateral load Q and assuming P ¼0. This part of the analysis allowed
studying the combined effects of shear deformations and the non-
linear connection at B in the non-linear response of the beam-
column. The large-deflection behavior of the beam-column is
similar as that of Example 1, where the structural element is made
of elastic material and the results obtained are valid as long as it
remains elastic. The results obtained with the proposed method
are compared to those obtained using the equations derived by
Aristizabal-Ochoa [12] for linear connections. Fig. 5a-c show the
effects of shear deformations in the large-deflection response of
M ub
B
B
Fig. 4. Example 2. Cantilever column connected at the base B with linear or non-
linear bending connection at B subject to P and Q at A: (a) applied loads and
properties and (b) deflected shape.
the cantilever column of Fig. 4 with linear and non-linear
connections. The results show that:
1) Both vertical and horizontal deflections became smaller as the
shear stiffness increases.
2) For low values of Q/(EI/h2) the effects of shear deformations are
not significant, but for high values the shear effects become
significant.
3) The effects of the non-linear bending connection at the base B on
the column's response are studied by comparing the results
obtained using the proposed model assuming non-linear con-
nection with the results reported by Aristizabal-Ochoa [12] for
92 O. Giraldo-Londoño et al. / International Journal of Non-Linear Mechanics 77 (2015) 85–95
Fig. 5. Example 2. Large-deflection analysis of a cantilever column laterally loaded.
Effects of shear deformations and nonlinear bending connection at B for: (a) GAs/
(EI/h2) ¼ ¥; (b) GAs/(EI/h2) ¼ 10; and (c) GAs/(EI/h2)¼ 5.
linear connections. The curves show that for lower values of Q/
(EI/h2) the behavior of the column with linear connection and
non-linear connections are very similar. However, as expected,
for higher values of Q/(EI/h2) both the horizontal and vertical
deflections become significantly larger for the column with non-
linear connection. This is a consequence of the stiffness degrada-
tion of the connection as the bending moment increases at the
base of the column.
The second part of this analysis deals with the effects of shear
deformations and the non-linearity of the connection on the post-
buckling behavior of the beam-column of Fig. 4. For this case, the
horizontal load Q is taken as zero while the vertical load P is varied.
The relationships between the load parameter P/(EI/h2) and the
horizontal and vertical displacements normalized with respect to h
(i.e., Δh/h and Δv/h, respectively) are analyzed. Results are shown in
Fig. 6a–d for the cantilever column with both linear and non-linear
connections and for different values of GAs. According to the results,
for the column with non-linear connection the effects of shear
deformations become significant for low values of P/(EI/h2), but for
the column with linear connections the effects of shear deformation
are always significant.
5.3. Example 3: Post-buckling behavior of a beam-column with non-
linear connections at both ends and linear elastic lateral bracing at
the top end A
Study the large deflection-small strain post-buckling behavior of a
slender column subjected to a concentrated vertical load P at the top
end A (Fig. 7). Since the column is slender, has purely elastic behavior
and a value of β ¼1 is assumed in all the proposed equations. Assume
that: (1) the column is initially out-of-plumb with θ0 ¼0.01 rad; (2)
β ¼1 and the values of E, A, I and h are given; (3) the bending
connections at and B are non-linear with Mua ¼Mub ¼Mu ¼ EI/h,
ρia ¼ ρib ¼ ρi ¼ 0:6, and na ¼ nb ¼n¼1; and (4) the lateral bracing at
3 3
A is linear with stiffness kh ¼ 0:5EI=h and 10EI=h . Consider that the
large-deflections solution is valid as long as the member remains
elastic, the applied load P is vertical and follows the top end A. Study
the effect of axial strain P t =EA on the cord length hr by selecting
different values for EA/(EI/h2), e.g., EA/(EI/h2)¼50, 100, and 1. Com-
pare the calculated results with those obtained using ABAQUS.
Solution: The large deflection-small strain post-buckling beha-
vior of the column is studied first by analyzing and comparing the
horizontal and vertical deflections Δh and Δv of the top end of the
column obtained using the proposed iterative procedure and those
using the ABAQUS model. As in Example 1, 50 linear elements
(B23) were used to generate the mesh, and the material properties
were assigned as linear elastic, with properties E and G. The
connection was simulated as a non-linear connector using a cloud
of points corresponding to its moment–rotation relationship.
Fig. 8c–f, h and i show the horizontal and vertical displace-
ments parameter Δh/h and Δv/h vs P/(EI/h2) for the three lateral
3 3
bracing stiffness values kh ¼ 0, 0:5EI=h and 10EI=h , respectively.
Notice that the results obtained with the proposed method are in
good agreement with those obtained using the ABAQUS model. It
can be observed from the results presented in these figures that:
1) P/(EI/h2) increases rapidly for small values of Δh/h and for Δv/
h o0.001 up to its peak value (i.e., until the axial buckling load
is reached).
2) The value of Δh/h corresponding to the buckling load of the column,
i.e., the peak value of P/(EI/h2) is slightly affected by the lateral
bracing stiffness and is located between Δh/h¼0.1 and Δh/h¼ 0.15.
3) The buckling load of the column increases with the lateral
bracing stiffness parameter kh/(EI/h3).
4) After the peak axial load is reached, P decreases gradually up to a
maximum horizontal displacement parameter Δh/h less than 1,
with the vertical displacement parameter Δv/hffi 1 (i.e., with the
top end A at the same vertical level of end B). Notice that the
maximum horizontal displacement Δh/h in Fig. 8b, e and h occur
for a value of with P/(EI/h2)ffi 1.8 regardless of either the stiffness of
the lateral bracing at end A or the axial stiffness parameter EA/(EI/
h2). In the case of EA/(EI/h2)-1, all curves pass through a common
point located at the maximum value of Δh/h (in Fig. 8b, e and h)
with Δv/h¼1 (in Fig. 8-c, f and i) with P/(EI/h2)ffi 1.8 regardless of
the stiffness of the lateral bracing at end A. After the common point
in each figure, the behavior of the column is affected by the
stiffness of the lateral bracing. Note that for kh/(EI/h3)¼0, P/(EI/h2)
increases as Δh/h decreases, whereas for kh/(EI/h3)¼ 5 and 10, P/(EI/
h2) decreases changing sign and turning into a tension force and
reaching a maximum and then changing sign again before increas-
ing into compression as Δh/h approaches zero. It was noticed both
maximum and minimum local values for P/(EI/h2) are strongly
affected by the stiffness the lateral bracing.
 O. Giraldo-Londoño et al. / International Journal of Non-Linear Mechanics 77 (2015) 85–95 93

3.0 3.0
GAs
EI/h 2 =

8
2.5 2.5
GAs
EI/h 2 = 10
2.0 GAs 2.0
EI/h 2 = 5
P P
1.5 1.5
EI/h 2 EI/h 2
GAs
1.0 1.0 EI/h 2 =

8
GAs
EI/h 2 = 10
0.5 0.5
GAs
EI/h 2 = 5
0 0.2 0.4 0.6 0.8 1.0 1.2 0 0.4 0.8 1.2 1.6 2.0
Δh Δv
h h

3.0 3.0

GAs
2.5 EI/h 2 = 2.5
8

GAs
EI/h 2 = 10
2.0 2.0
GAs
EI/h 2 = 5
P P
1.5 1.5
EI/h 2 EI/h 2
GAs
1.0 1.0 EI/h 2 =

8
GAs
EI/h 2 = 10
0.5 0.5
GAs
EI/h 2 = 5
0 0.2 0.4 0.6 0.8 1.0 1.2 0 0.4 0.8 1.2 1.6 2.0
Δh Δv
h h
Fig. 6. Example 2. Post-buckling analysis of a cantilever column: effects of shear deformations and bending connections at B: (a) (Δh/h) vs P/(EI/h2) for linear connection; (b)
(Δv/h) vs P/(EI/h2) for linear connection; (c) (Δh/h) vs P/(EI/h2) for non-linear connection; and (d) (Δv/h) vs P/(EI/h2) for non-linear connection.

The load–displacement behavior of the column was also

5) affected by axial stiffness parameter EA/(EI/h2). Three values
P
of EA/(EI/h2) were considered for the analysis (e.g., EA/(EI/h2)¼
50, 100, and 1). From the finite element models, it was noticed
that values of EA/(EI/h2) below 40 showed convergence pro- Δh P
blems. This, in order to compare the results from the proposed kh
model with those obtained with the ABAQUS models, the Q
A Δv
minimum value for EA/(EI/h2) used in this analysis was 50.
Fig. 8c, e and f indicate that the member behaves linearly κi , n
Mu A Q
before the first peak of maximum load is reached. According to
these results, the slope of this linear portion was approximately
equal to the axial stiffness parameter EA/(EI/h2). This indicates
that the behavior of the column before the first peak is reached E, G
θ0
is mainly controlled by the axial stiffness EA/h [or the axial A, A s
I, h
stiffness parameter EA/(EI/h2)], independently of either the
properties of the connections or ks/(EI/h3). The magnitude of
the peak, however, increased with kh/(EI/h3) and EA/(EI/h2).
κi , n
The deformed shape was obtained for different points Mu B
(e.g., points 1 through 11) in the load–displacement curves as shown in B
Fig. 8. The elastica was obtained from Eqs. (24)–(27) as presented by
Vega-Posada et al. [14]. Specific points are indicated such as the maximum Fig. 7. Example 3. Out-of-plumb column with sidesway partially uninhibited and
horizontal tip displacement of the column (points 2, 5 and 9), first peak with non-linear bending connections at both ends subjected to a vertical load P at
loads (points 1, 4, and 8), maximum vertical displacement (points 3, 7 and A: (a) applied loads and properties; and (b) deflected shape.
94 O. Giraldo-Londoño et al. / International Journal of Non-Linear Mechanics 77 (2015) 85–95

1 1 15 3 15 3 7 11
EA = 50 11 EA = 50
4 7
EA = 50
EI/h 2 EI/h 2 EI/h 2
8 8 8
0.5
2 10 10
4 4
5
0
9
y P
5
P
5
h EI/h 2 EI/h 2 1
-0.5 1 2
6 9
9 5 2 5
10 0 0
-1 6 6
3
-1.5 11 7 -5 10
-5
10
0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0 0 0.5 1.0 1.5 2.0 2.5
x Δh Δv
h h h

1 1 15 15 3 711
EA = 100 11 3 EA = 100 EA = 100
4 EI/h 2 7 EI/h 2 EI/h
2

8 8
0.5 8
2 10 10
4
4
5
0 9
y P P
h 5 1 5
EI/h 2 5 EI/h 2 1
-0.5 9
6
10 9 2 5
0 2 0 6
-1 6
3
11 7 10 10
-1.5 -5 -5
0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0 0 0.5 1.0 1.5 2.0 2.5
x Δh Δv
h h h

1 1 15 3 15 3 711
11
4 EA = 7 EA = EA
2 =
8

8
8

8 EI/h 2 EI/h 2 EI/h

8
0.5
2 10 10
4 4
5
0 9
y P P
h 5 1 5
EI/h 2 5 EI/h 2 1
-0.5 2 9
6 2 5
0 9 0
10
-1 6 6
3
11 7
10 10
-1.5 -5 -5
0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0 0 0.5 1.0 1.5 2.0 2.5
x Δh Δv
h h h
Fig. 8. Example 3. Post-buckling behavior of an out-of-plumb slender column with non-linear bending connections at both ends and sidesway partially uninhibited. (—) kh/
(EI/h3) ¼ 0, ( ) kh/(EI/h3) ¼5, ( ) kh/(EI/h3) ¼ 10, ( ) ABAQUS.

11), and minimum vertical load (i.e., points 6 and 10). The values of P/(EI/h2)
corresponding to each of these points are shown in Table 1.
6. Summary and conclusions
A set of non-linear equations capable of predicting accurately
and efficiently the non-linear large-deflection small-strain and
post-buckling behavior of Timoshenko beam-columns with sym-
metrical cross section considering the effects of non-linear bending
connections and non-linear lateral and vertical restraints at the top
end of the column are presented. The proposed method includes
the combined effects of out-of-plumbness, transverse shear forces
and deformations, and shortening of the beam-column due to both
axial forces and bowing. An iterative procedure to solve the system
of non-linear equations is proposed, and three comprehensive
examples are presented in detail. The calculated results are com-
pared to those available in the technical literature and also to results
obtained using ABAQUS. It is important to mention that FE model
presented convergence problems for large deformation, while the
 O. Giraldo-Londoño et al. / International Journal of Non-Linear Mechanics 77 (2015) 85–95 95

[4] M. Hjiaj, J.M. Battini, Q.H. Nguyen, Large displacement analysis of shear
Table 1 deformable composite beams with interlayer slips, Int. J. Non-Linear Mech.
Example 3. Analyzed points for the development of the elastic shapes. 47 (8) (2012) 895–904.
[5] D.K. Nguyen, B.S. Gan, Large deflections of tapered functionally graded beams
Point EA/(EI/h2) ¼50 EA/(EI/h2) ¼ 100 EA/(EI/h2) ¼ 1 subjected to end forces, Appl. Math. Model. 38 (11–12) (2014) 3054–3066.
P/(EI/h2) P/(EI/h2) P/(EI/h2) [6] H.A.F.A. Santos, D.Y. Gao, Canonical dual finite element method for solving
post-buckling problems of a large deformation elastic beam, Int. J. Non-Linear
1 4.2 3.8 3.8 Mech. 47 (2012) 240–247.
2 1.8 1.8 1.8 [7] Y. Yu, B. Wu, C.W. Lim, Numerical and analytical approximations to large post-
3 15.0 15.0 15.0 buckling deformation of MEMS, Int. J. Mech. Sci. 55 (2012) 95–103.
4 8.2 8.2 8.2 [8] R. Ansari, M. Faghih Shojaei, R. Gholami, V. Mohammadi, M.A. Darabi, Thermal
5 1.8 1.8 1.8 postbuckling behavior of size-dependent functionally graded Timoshenko
6  0.4  0.6  0.8 microbeams, Int. J. Non-Linear Mech. 50 (2013) 127–135.
7 15.0 15.0 15.0 [9] J.P. Smith-Pardo, J.D. Aristizabal-Ochoa, Buckling reversals of axially restrained
8 11.6 12.0 12.6 imperfect beam-column, J. Eng. Mech. 125 (4) (1999) 401–409.
[10] J.D. Aristizabal-Ochoa, Non-linear large-deflection-small strain elastic analysis
9 1.8 1.8 1.8
of a beam-column with semi-rigid connections, J. Struct. Eng. 127 (1) (2001)
10  3.4  4.0  4.8
92–96.
11 15.0 15.0 15
[11] J.D. Aristizabal-Ochoa, Large-deflection stability of slender beam-columns
with semi-rigid connections: the elastica approach, J. Eng. Mech. 130 (3)
(2004) 274–282.
proposed method was always convergent. The comparisons show [12] J.D. Aristizabal-Ochoa, Large deflection and postbuckling behavior of
that the proposed model is efficient and accurate to study the large- Timoshenko beam-columns with semi-rigid connections including shear and
deflection and postbuckling behavior of Timoshenko beam-columns axial effects, J. Eng. Struct. 29 (6) (2007) 991–1003.
[13] C. Vega-Posada, M. Areiza-Hurtado, J.D. Aristizabal-Ochoa, Large-deflection
with non-linear bending connections and non-linear lateral bracing stability of slender beam-columns with both ends partially restrained against
and axial restraint at the top end. rotation, J. Eng. Mech. 133 (12) (2007) 1394–1400.
Analytical studies indicate that longitudinal and transverse deflec- [14] C. Vega-Posada, M. Areiza-Hurtado, J.D. Aristizabal-Ochoa, Large-deflection
and post-buckling behavior of slender beam-columns with non-linear end-
tions of Timoshenko beam-columns are highly affected by shear restraints, Int. J. Non-linear Mech. 46 (1) (2011) 79–95.
deformations when linear elastic semi-rigid connections are used. [15] J.D. Aristizabal-Ochoa, Stability of columns with semi-rigid connections
However, when non-linear bending connections and end restraints including shear effects using Engesser, Haringx and Euler approaches, J. Eng.
Struct. 33 (2011) 868–880.
with degrading stiffness are used, the effects of shear deformations on
[16] S. Timoshenko, J. Gere, Theory of Elastic Stability, 2nd ed., McGraw-Hill, New
the longitudinal and transverse deflections are reduced. However, York, 1961.
non-linear bending connections produce larger deflections, while [17] F. Engesser, Die Knickfestigkeit gerader Stäbe, Zentrale Bauverwalt. 11 (1891)
increasing the stiffness of the lateral bracing and axial restraint 483–486.
[18] J.A. Haringx, On the buckling and the lateral rigidity of helical compression
produce an increase in the column capacity together with smaller springs, I Proc. Roy. Acad. Amst. 45 (1942) 533–539.
deflections for the same level of load. [19] J. Blaauwendraad, Shear in structural stability: on the Engesser–Haringx
discord, J. Appl. Mech. 77 (3) (2010) 031005.
[20] M. Attard, Finite strain – beam theory, Int. J. Solids Struct. 40 (2003)
Acknowledgments 4563–4584.
[21] Z.P. Bažant, Shear buckling of sandwich, fiber composite and lattice columns,
Bear. Helical Springs: Paradox Resolved. ASME J. Appl. Mech 70 (1) (2003)
The authors wish to thank the Department of Civil Engineering 75–83.
[22] W.F. Chen, Y. Goto, J.Y.R. Liew, Stability Design of Semi-Rigid Frames, John
of the School of Mines of the National University of Colombia at
Wiley & Sons, Inc., New York, 1996.
Medellín for its support. [23] R.M. Richard, B.J. Abbott, Versatile elastic-plastic stress-strain formula, J. Eng.
Mech. Division-ASCE 101 (4) (1975) 511–515.
[24] S.G. Ekhande, M. Selvappalam, M.K. Madugula, Stability functions for three-
References
dimensional beam-columns, J. Struct. Eng. 115 (2) (1989) 467–479.
[25] Y. Bai., Q. Bai., Subsea Pipelines and Risers, Elsevier, Oxford, UK, 2005.
[1] R. Ansari, V. Mohammadi, M. Faghih Shojaei, R. Gholami, S. Sahmani, Post- [26] J.D. Aristizabal-Ochoa, Disposable knee bracing: improvement in seismic
buckling analysis of Timoshenko nanobeams including surface stress effect, design of steel frames, J. Struct. Eng. 112 (7) (1986) 1544–1552.
Int. J. Eng. Sci. 75 (2014) 1–10. [27] K. Fehrenkamp, J.R. Morgan, J.K. Nelson, Equivalent nonlinear spring for
[2] S. Chaterjee, G. Pohit, A large deflection model for the pull-in analysis of catenary moorings, J. Struct. Eng. 110 (10) (1984) 2541–2546.
electrostatically actuated microcantilever beams, J. Sound Vib. 322 (2009) [28] M.A. Wishahy, M. Arockiasamy, Response of guyed tower to irregular waves
969–986. for linear and nonlinear behavior of mooring cables, Can. J. Civ. Eng. 12 (1)
[3] C.J. Gantes, K.E. Kalochairetis, Axially and transversely loaded Timoshenko and (1985) 200–212.
laced built-up columns with arbitrary supports, J. Constr. Steel Res. 77 (2012) [29] S. Bryniolfsson, Response of guyed offshore towers to stochastic loads in the
95–106. presence of steady current (Ph.D. dissertation), Oregon State University, 1987.