Vous êtes sur la page 1sur 175

# LATERAL-TORSIONAL BUCKLING OF STRUCTURES WITH MONOSYMMETRIC

CROSS-SECTIONS

by

Matthew J. Vensko

## of the requirements for the degree of

Master of Science

University of Pittsburgh

2008
UNIVERSITY OF PITTSBURGH

## This thesis was presented

by

Matthew J. Vensko

It was defended on

and approved by

## Kent A. Harries, Assistant Professor,

Department of Civil and Environmental Engineering
Albert To, Assistant Professor,
Department of Civil and Environmental Engineering
Morteza A. M. Torkamani, Associate Professor,
Department of Civil and Environmental Engineering,

ii

LATERAL-TORSIONAL BUCKLING OF STRUCTURES WITH MONOSYMMETRIC
CROSS-SECTIONS

## University of Pittsburgh, 2008

Lateral-torsional buckling is a method of failure that occurs when the in-plane bending capacity

of a member exceeds its resistance to out-of-plane lateral buckling and twisting. The lateral-

## torsional buckling of beam-columns with doubly-symmetric cross-sections is a topic that has

been long discussed and well covered. The buckling of members with monosymmetric cross-

sections is an underdeveloped topic, with its derivations complicated by the fact that the centroid

and the shear center of the cross-section do not coincide. In this paper, the total potential energy

## predict the lateral-torsional buckling load.

The total potential energy equation is the sum of the strain energy and the potential

energy of the external loads. The theorem of minimum total potential energy exerts that setting

the second variation of this equation equal to zero will represent a transition from a stable to an

unstable state. The buckling loads can then be identified when this transition takes place. This

thesis will derive energy equations in both dimensional and non-dimensional forms assuming

iii

that the beam-column is without prebuckling deformations. This dimensional buckling equation

## will then be expanded to include prebuckling deformations.

The ability of these equations to predict the lateral-torsional buckling loads of a structure

is demonstrated for different loading and boundary conditions. The accuracy of these predictions

is dependent on the ability to select a suitable shape function to mimic the buckled shape of the

beam-column. The results provided by the buckling equations derived in this thesis, using a

suitable shape function, are compared to examples in existing literature considering the same

The finite element method is then used, along with the energy equations, to derive

element elastic and geometric stiffness matrices. These element stiffness matrices can be

transformed into global stiffness matrices. Boundary conditions can then be enforced and a

generalized eigenvalue problem can then be used to determine the buckling loads. The element

elastic and geometric stiffness matrices are presented in this thesis so that future research can

## complex systems containing members with monosymmetric cross-sections.

iv

1.0 INTRODUCTION....................................................................................................................1

## 3.1 STRAIN ENERGY...................................................................................................33

3.1.1 Displacements...........................................................................................34

3.1.2 Strains.......................................................................................................41

## 3.3.1 Displacements and Rotations of Load Points............................................47

3.4 ENERGY EQUATION FOR LATERAL TORSIONAL BUCKLING......................49

## 4.0 LATERAL-TORSIONAL BUCKLING OF MONOSYMMETRIC BEAMS

CONSIDERING PREBUCKLING DEFLECTIONS.....................................................54

## 4.1 STRAIN ENERGY CONSIDERING PREBUCKLING DEFLECTIONS................54

4.1.1 Displacements............................................................................................54

## 4.2 STRAIN ENERGY EQUATION CONSIDERING PREBUCKLING

DEFLECTIONS.....................................................................................................58

## 4.3 POTENTIAL ENERGY OF THE EXTERNAL LOADS CONSIDERING

PREBUCKLING DEFLECTIONS........................................................................59

## 4.4 ENERGY EQUATION CONSIDERING PREBUCKLING DEFLECTION............60

5.0 APPLICATIONS....................................................................................................................64

## 5.1 SIMPLY-SUPPORTED MONOSYMMETRIC BEAM SUBJECTED TO EQUAL

END MOMENTS, M.............................................................................................64

## 6.2 GEOMETRIC STIFFNESS MATRIX.....................................................................106

vi

6.3 FINITE ELEMENT METHOD CONSIDERING PREBUCKLING
DEFLECTIONS........................................................................................................109

## 6.4 ELASTIC STIFFNESS MATRIX CONSIDERING PREBUCKLING

DEFLECTIONS........................................................................................................110

## 6.5 GEOMETRIC STIFFNESS MATRIX CONSIDERING PREBUCKLING

DEFLECTIONS........................................................................................................112

7.0 SUMMARY...................................................................................................................115

APPENDIX A..............................................................................................................................118

## A.6 ELEMENT PREBUCKLING GEOMETRIC STIFFNESS MATRIX....................136

APPENDIX B..............................................................................................................................143

## B.2 MATRIX [B] FROM SECTON 5.4.........................................................................145

BIBLIOGRAPHY........................................................................................................................156

WORKS CITED..............................................................................................................156

WORKS CONSULTED..................................................................................................158

vii

LIST OF FIGURES

## Figure 2.1a Beams of Rectangular Cross-Section...........................................................................6

Figure 2.1b Beams of Rectangular Cross-Section with Axial Force and End Moments.................6

Figure 2.2 Linearly Tapered Beam Subjected to Equal and Opposite End Moments.....................9

## Figure 2.3 Simply-Supported Beam with Concentrated Load, P, at Midspan..............................11

Figure 2.4a Monosymmetric Beam Subjected to End Moments and Axial Load.........................14

## Figure 3.1 Coordinate System of Undeformed Monosymmetric Beam........................................29

Figure 3.2 External Loads and Member End Actions of Beam Element......................................30

## Figure 3.5 Translation of Point Po to Point P................................................................................35

viii

Figure 5.1 Monosymmetric Beam with Subjected to Equal End Moments..................................65

## Figure 5.2 Monosymmetric Beam with Subjected to Concentrated Central Load.......................70

Figure 5.3 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section
Subjected to a Concentrated Central Load ( βx = − 0.6) .........................................75

Figure 5.4 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section
Subjected to a Concentrated Central Load ( βx = − 0.3) .........................................76

Figure 5.5 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section
Subjected to a Concentrated Central Load ( βx = − 01
. ) .........................................76

Figure 5.6 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section
Subjected to a Concentrated Central Load ( βx = 0) .............................................77

Figure 5.7 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section
Subjected to a Concentrated Central Load ( βx = 01
. ) ...........................................77

Figure 5.8 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section
Subjected to a Concentrated Central Load ( βx = 0.3) ...........................................78

Figure 5.9 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section
Subjected to a Concentrated Central Load ( βx = 0.6) ...........................................78

## Figure 5.10 Monosymmetric Beam Subjected to a Uniformly Distributed Load..........................79

Figure 5.11 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section
Subjected to a Uniformly Distributed Load ( βx = − 0.6) .......................................84

ix

Figure 5.12 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section
Subjected to a Uniformly Distributed Load ( βx = − 0.3) .......................................84

Figure 5.13 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section
Subjected to a Uniformly Distributed Load ( βx = − 01
. ) .......................................85

Figure 5.14 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section
Subjected to a Uniformly Distributed Load ( βx = 0) ...........................................85

Figure 5.15 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section
Subjected to a Uniformly Distributed Load ( βx = 01
. ) .........................................86

Figure 5.16 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section
Subjected to a Uniformly Distributed Load ( βx = 0.3) .........................................86

Figure 5.17 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section
Subjected to a Uniformly Distributed Load ( βx = 0.6) .........................................87

## Figure 5.18 Monosymmetric Cantilever Beam Subjected to a Concentrated End Load...............88

Figure 5.19 Buckling Load: Cantilever Beam with Monosymmetric Cross-Section Subjected to a
Concentrated End Load ( βx = − 0.6) ......................................................................94

Figure 5.20 Buckling Load: Cantilever Beam with Monosymmetric Cross-Section Subjected to a
Concentrated End Load ( βx = − 0.3) .....................................................................95

Figure 5.21 Buckling Load: Cantilever Beam with Monosymmetric Cross-Section Subjected to a
Concentrated End Load ( βx = − 01
. ) ......................................................................95

Figure 5.22 Buckling Load: Cantilever Beam with Monosymmetric Cross-Section Subjected to a
Concentrated End Load ( βx = 0) ..........................................................................96

Figure 5.23 Buckling Load: Cantilever Beam with Monosymmetric Cross-Section Subjected to a
Concentrated End Load ( βx = 01
. ) ........................................................................96

Figure 5.24 Buckling Load: Cantilever Beam with Monosymmetric Cross-Section Subjected to a
Concentrated End Load ( βx = 0.3) ........................................................................97

Figure 5.25 Buckling Load: Cantilever Beam with Monosymmetric Cross-Section Subjected to a
Concentrated End Load ( βx = 0.6) ........................................................................97

## Figure 6.3 Element Degrees of Freedom with Nodal Displacements φ .....................................100

xi

NOMENCLATURE

Symbol Description

A area of member

## {de} local nodal displacement vector for an element

E modulus of elasticity

## e non-dimensional concentrated load height

G shear modulus

xii

[ge] element local geometric stiffness matrix for initial load set

## Iω warping moment of inertia

J torsional constant

K beam parameter

L member length

## Mcr classical lateral buckling uniform bending moment

Mx bending moment

xiii

U strain energy

## u2, u4 out-of-plane rotation at nodes 1 and 2

u’ out-of-plane rotation

## u non-dimensional out-of-plane lateral displacement

xiv

v in-plane bending displacement

## v2 , v4 in-plane rotation at nodes 1 and 2

v’ in-plane rotation

w axial displacement

## zp non-dimensional distance to concentrated load

βx monosymmetry parameter

## βx non-dimensional monosymmetry parameter

xv

εP longitudinal strain of point Po

## γP shear strain of point Po

λ buckling parameter

## non-dimensional total potential energy

ρ degree of monosymmetry

## τP shear stress of point Po

ω warping function

## θ rotation of the member cross-section

xvi

1.0 INTRODUCTION

The members of a steel structure, commonly known as beam-columns, are usually designed with

## a thin-walled cross-section. Thin-walled cross-sections are used as a compromise between

structural stability and economic efficiency and include angles, channels, box-beams, I-beams,

etc. These members are usually designed so that the loads are applied in the plane of the weak

axis of the cross-section, so that the bending occurs about the strong axis. However, when a

beam, usually slender in nature, has relatively small lateral and torsional stiffnesses compared to

its stiffness in the plane of loading, the beam will deflect laterally and twist out of plane when

the load reaches a critical limit. This limit is known as the elastic lateral-torsional buckling

The lateral buckling and twisting of the beam are interdependent in that when a member

deflects laterally, the resulting induced moment exerts a component torque about the deflected

longitudinal axis which causes the beam to twist (Wang, et al. 2005). The lateral-torsional

buckling loads for a beam-column are influenced by a number of factors, including cross-

sectional shape, the unbraced length and support conditions of the beam, the type and position of

the applied loads along the member axis, and the location of the applied loads with respect to the

## centroidal axis of the cross section.

This paper will focus on the lateral-torsional buckling of steel I-beams with a

## monosymmetric cross-section. In a beam with a monosymmetric cross-section, the shear center

1
and the centroid of the cross-section do not coincide. The significance of this can be explained

by the Wagner effect (Anderson and Trahair, 1972), in which the twisting of the member causes

the axial compressive and tension stresses to exert an additional disturbing torque. This torque

can reduce the torsional stiffness of a member in compression and increase the torsional stiffness

## of a member in tension. In I-beams with doubly-symmetric cross-sections, these compressive

and tensile stresses balance each other exactly and the change in the torsional stiffness is zero. In

I-beams with monosymmetric cross-sections where the smaller flange is further from the shear

center, the Wagner effect results in a change in the torsional stiffness. The stresses in the smaller

flange have a greater lever arm and predominate in the Wagner effect. The torsional stiffness of

the beam will then increase when the smaller flange is in tension and decrease when the smaller

flange is in compression.

When a structure is simple, such as a beam, an energy method approach may be used

directly to calculate the lateral-torsional buckling load of the structure. Assuming a suitable

shape function, the equations derived using the energy method can provide approximate buckling

loads for the structure. However, when a structure is complex, this is not possible. In this case,

the energy method in conjunction with the finite element method may be used to calculate the

## lateral-torsional buckling load of the structure.

The finite element method is a versatile numerical and mathematical approach which can

encompass complicated loads, boundary conditions, and geometry of a structure. First, element

elastic stiffness and geometric stiffness matrices are derived for an element using the energy

equations for lateral-torsional buckling. The structure in question must be divided into several

elements, and a global coordinate system can be selected for that structure. The element elastic

stiffness and geometric stiffness matrices are transformed to the global coordinate system for

2
each element, resulting in global element elastic and geometric stiffness matrices for the

structure. After this assembly process, boundary conditions are enforced to convert the structure

from an unrestrained structure to a restrained structure. The derived equilibrium equations are in

the form of a generalized eigenvalue problem, where the eigenvalues are the load factors that,

when multiplied to a reference load, result in lateral-torsional buckling loads for the structure.

The main objective of this thesis is to formulate equations for lateral-torsional buckling

of monosymmetric beams using the energy method. Suitable shape functions will be applied to

these equations to provide approximate buckling solutions that can be compared to previous data.

The finite element method will be used to derive element elastic and geometric stiffness matrices

that can be used in future works in conjunction with computer software to determine lateral-

## torsional buckling loads of more complex structures.

3
2.0 LITERATURE REVIEW

This section reviews available literature that explores lateral-torsional buckling as the primary

state of failure for beams used in structures. A beam that has relatively small lateral and

torsional stiffnesses compared to its stiffness in the plane of loading tends to deflect laterally and

twist out of plane. This failure mode is known as lateral-torsional buckling. Two methods are

used to derive the critical load values that result in lateral-torsional buckling beam failure: the

method utilizing differential equilibrium equations and the energy method. The differential

equilibrium method of stability analysis assumes the internal and external forces acting on an

object to be equal and opposite. The energy method refers to an approach where the total

potential energy of a conservative system is calculated by summing the internal and external

energies. The buckling loads for the system can then be approximated if a suitable shape

function for the particular structure is used, thus reducing the system from one having infinite

degrees of freedom to one having finite degrees of freedom. This approach is known as the

Rayleigh-Ritz method. This method will provide acceptable results as long as the assumed shape

function is accurate. Both the differential equilibrium method and energy methods are examined

## in this literature review.

4
2.1 EQUILIBRIUM METHOD

The closed form solutions for various loading conditions and cross-sections are demonstrated

below using the equilibrium method. The beams are assumed to be stationary and therefore the

sum of the internal forces of the structure and the external forces is assumed to be zero. The

## equations are rearranged in terms of displacements resulting in a second order differential

equation from which the buckling loads can be solved. The beams are assumed in this section to

be elastic, initially perfectly straight, and in-plane deformations are neglected. Rotation of the

beam, φ , is assumed small, so for the small angle relationships sin φ = φ and cos φ = 1 can be

used.

Consider a simply supported beam with a uniform rectangular cross section as shown in

Figure 2.1a and Figure 2.1b. Note that u, v and w are the displacements in the x-, y-, and z-

directions, respectively. The section rotates out of plane at an angle φ . The differential

equilibrium equations of minor axis bending and torsion of a beam with no axial force (F = 0)

## are derived from statics as (Chen and Lui. 1987)

d2 u
EI y = − M xφ (2.1)
dz 2

dφ du
GJ = Mx + Mz (2.2)
dz dz

Ehb 3 Ghb 3
where EI y = and GJ =
12 3

5
EI y represents the flexural rigidity of the beam with respect to the y-axis and GJ represents the

torsional rigidity of the beam with respect to the z-axis. M x and M z are the internal moments

of the beam acting about the x-axis and the z-axis, respectively. In Eq. (2.1), the component of

M x in the y-direction is represented by − M x sin φ which, by way of the small angle theorem,

## reduces to − M x φ . In Eq. (2.2), the torsional component of M x acting in the z-direction is

du
represented by M x .
dz

x  y

## Figure 2.1a Beams of Rectangular Cross-Section

M M
F    F
z

L

y

Figure 2.1b Beams of Rectangular Cross-Section with Axial Force and End Moments

6
2.1.1 Closed Form Solutions

Case 1: A beam that is subjected to only equal end moments M about x-axis

This loading case is shown in Figure 2.1b, with F = 0. Since there is no torsional component of

the moment, let Mx = M and Mz = 0. The equilibrium equations given in Eq. (2.1) and (2.2)

reduce to

d2 u
EI y = −M φ (2.3)
dz 2

dφ du
GJ = M (2.4)
dz dz

d 2 u − Mφ
= (2.5)
dz 2 EI y

d 2 u GJ d 2 φ
= (2.6)
dz 2 M dz 2

## Eliminating u yields a single differential equation of the form

d 2φ M2
+ φ=0 (2.7)
dz 2 GJEI y

Solving the second order differential equation yields the general solution as

⎛ Mz ⎞⎟ ⎛ Mz ⎞⎟
φ ( z) = A sin⎜⎜ + B cos⎜⎜ (2.8)
⎝ EI y GJ ⎟⎠ ⎝ EI y GJ ⎟⎠

## By applying the boundary condition φ = 0 at z = 0, B is equal to zero.

7
The constant A may then not be equal to zero because it provides a trivial solution. Therefore at

z=L

M cr L
sin =0 (2.9)
EI y GJ

Solving for Mcr to provide the smallest nonzero buckling load yields

π EI y GJ
M cr = (2.10)
L

where Mcr is the critical value of M that will cause the beam to deflect laterally and twist out of

plane.

Case 2: A linearly tapered beam with a rectangular cross section that is subjected to only equal

## end moments M about x-axis

z
For this case, consider a linearly tapered beam with initial depth ho which increases at a rate δ
L

## as shown in Figure 2.2 where

z
h( z ) = (1 + δ )ho (2.11)
L

In Eq. (2.11), h(z) is the linear tapered depth of the beam as a function of z, ho is the depth of

## beam at z = 0, and (1 + δ) is the ratio of height of tapered beam at z = L to z = 0.

8
M

h o
h  (z)
(1 + δ ) h o

L

Figure 2.2 Linear Tapered Beam Subjected to Equal and Opposite End Moments

## The beam properties of the tapered section can be written as

EI y = EI oη (2.12)

GJ = GJ o η (2.13)

ho b 3 ho b 3 z
where I o = , Jo = , and η = 1 + δ . (2.14)
12 3 L

## As in Case 1, there is no torsional component of the moment so that Mx = M and Mz = 0.

Substituting Eqs. (2.12) and (2.13) into the equilibrium equations yield

d 2u
EI oη 2 = − Mφ (2.15)
dz

dφ du
GJ o η = M (2.16)
dz dz

dη δ
= (2.17)
dz L

## which enables the following relationships

9
(2.18)
dφ dφ dη δ dφ
= =
dz dη dz L dη

and

d 2φ ⎛ δ ⎞ d 2φ
2

=⎜ ⎟ (2.19)
dz 2 ⎝ L ⎠ dη 2

Substituting Eqs. (2.17) – (2.19) into the equilibrium Eqs. (2.15) and (2.16) yield

⎛δ⎞ d u
2 2

EI o η⎜ ⎟ = − Mφ (2.20)
⎝ L ⎠ dη 2

⎛ δ ⎞ dφ ⎛ δ ⎞ du
GJ o η⎜ ⎟ = M⎜ ⎟ (2.21)
⎝ L ⎠ dη ⎝ L ⎠ dη

Differentiating Eq. (2.21) and combining it with Eq.(2.20) in order to eliminate u yields

d 2φ dφ
η2
2 + η + k 2φ = 0 (2.22)
dη dη

where

M 2 L2
k =
2
(2.23)
EI o GJ o δ 2

## φ = A sin( k ln η) + B cos( k ln η) (2.24)

Applying the boundary condition φ = 0 at z = 0, B is equal to zero. The constant A may then not

be equal to zero because it provides a trivial solution. Therefore the boundary condition φ = 0 at

η = (1 + δ ) yields

## sin( k ln(1 + δ )) = 0 (2.25)

Solving for Mcr to provide the smallest nonzero buckling load yields

10
πδ EI o GJ o
M cr = (2.26)
ln(1 + δ ) L

where Mcr is the critical value of M that will cause the beam to deflect laterally and twist out of

plane. It is important to recognize that if δ = 0 , meaning that the beam is not tapered, Eq. (2.26)

reduces to

π EI o GJ o
M cr = (2.27)
L

## which is the result obtained in Case 1.

Case 3: A simply supported beam with a concentrated load, P, at midspan (at z = L/2)

For this case, consider a non-tapered beam with a concentrated load, P, at midspan as shown in

Figure 2.3.

L

## Figure 2.3 Simply Supported Beam with Concentrated Load, P, at Midspan

11
The moments Mx and Mz are derived using basic equilibrium concepts as

Pz
Mx = (2.28)
2

P
Mz = ( u∗ − u) (2.29)
2

Where u* represents the lateral deflection at the centroid of the middle cross section and u

## represents the lateral deflection at any cross section.

By substituting the relationships for M x and M z into Eqs. (2.1) and (2.2), the differential

d2 u Pz
EI y 2
=− φ
dz 2 (2.30)

dφ Pz du P
GJ = + (u∗ − u)
dz 2 dz 2 (2.31)

d 2φ P2z2
+ φ=0 (2.32)
dz 2 4 EI y GJ

z
η= (2.33)
L

P 2 L4
ζ= (2.34)
4 EI y GJ

## Eq. (2.32) can then be reduced to

d 2φ
+ ζ 2η 2φ = 0 (2.35)
dη 2

12
The general solution utilizes Bessel functions (Arfken, 2005) of the first kind of orders 1/4 and -

## 1/4 shown below as

⎡ ⎛ ζη 2 ⎞ ⎛ ζη 2 ⎞ ⎤
φ = η ⎢ AJ1/ 4 ⎜ ⎟ + BJ −1/ 4 ⎜ ⎟⎥ (2.36)
⎣ ⎝ 2 ⎠ ⎝ 2 ⎠⎦

dφ 1
Applying the boundary conditions φ = 0 at η = 0 and = 0 at η =
dη 2

⎛ζ⎞ ζ
gives J − 3/ 4 ⎜ ⎟ = 0 then = 10585
. yielding an expression for the buckling load, Pc, as
⎝ 8⎠ 8

16.94
Pc = EI y GJ (2.37)
L2

## where Pc is the lateral-torsional buckling load (Wang, et al. 2005).

Case 4: A simply supported I –beam with a monosymmetric cross-section subjected to equal end

## moments, M, and an axial load, F, acting through the centroid:

In this case, consider a non-tapered monosymmetric I-beam subjected to end moments, M, and

an axial force, F, as shown in Figure 2.4a with the cross-section of the beam shown in Figure

2.4b.

13
M  M

F  z

L

Figure 2.4a Monosymmetric Beam subjected to End Moments and Axial Force

IyT

shear center, s
yo
x  centroid, o  h

IyB  y

## Figure 2.4b Cross-Section of Monosymmetric Beam

The minor axis distributed force equilibrium equation and the distributed torque equilibrium

equation for the member can be expressed as (Kitipornchai and Wang. 1989)

d 4u d 2φ d 2u
EI y 4 = − ( M + Fy o ) 2 − F 2 (2.38)
dz dz dz

dφ d 3φ du
(GJ − F ⋅ ro + Mβx )
2
− EI w 3 = M (2.39)
dz dz dz

## Where the warping rigidity is

14
EI w = EI y ρ(1 − ρ)h 2 (2.40)

## and the degree of monosymmetry can be expressed as

I yT I yT
ρ= = (2.41)
( I yT + I yB ) Iy

where I yT and IyB are the second moments of inertia about the y-axis of the top and bottom

flanges, respectively, as shown in Figure 2.4b. Because the beam has a monosymmetric cross-

section, the centroid of the beam, o, and the shear center, s, do not coincide. This introduces a

term, y o , which represents the vertical distance between the centroid and the shear center. The

polar radius of gyration about the shear center, ro, can be expressed as

Ix + Iy
ro 2 = + yo 2 (2.42)
A

## The monosymmetric parameter of the beam (Trahair and Nethercot, 1984) is

1⎛ 2 ⎞
βx = ⎜ ∫ x ydA + ∫ y 3 dA⎟ − 2 y o (2.43)
Ix ⎝ A A ⎠

where x and y are coordinates with respect to the centroid. βx accounts for the Wagner effect,

which is the change in effective torsional stiffness due to the components of bending

compressive and tensile stresses that produce a torque in the beam as it twists during buckling.

Recognizing that, since the beam is simply supported, the boundary conditions become

d 2φ
φ = 0 and = 0 at z = 0, L.
dz 2

With the elimination of u and the implementation of the above boundary conditions, Eqs. (2.36)

and (2.39) yield a closed form solution for critical values F and M (Trahair and Nethercot, 1984)

as

15
⎛ F ⎞⎛ F Mβx ⎞
( M + Fy 0 ) 2 = ro 2 Fz FE ⎜ 1 − ⎟ ⎜1− + 2 ⎟ (2.44)
⎝ FE ⎠ ⎝ Fz ro Fz ⎠

π 2 EI y
FE = (2.45)
L2

## and Fz is given as (Wang, et al. 2005)

GJ ⎛ π 2 EI w ⎞
Fz = ⎜ 1 + ⎟ (2.46)
ro 2 ⎝ GJL2 ⎠

In order to obtain a non-dimensional elastic buckling moment, use is made of the non-

## dimensional parameters (Kitipornchai and Wang. 1989)

π 2 EI y h 2
K=
4GJL2 (2.47)

4 ⎛ Ix + Iy ⎞
η= ⎜ ⎟ (2.48)
h2 ⎝ A ⎠

2 y0
υ=
h (2.49)

F
Λ = (2.50)
FE

⎡ βx
λ = ⎢− υΛ + (1 − Λ )⎤⎥ K (2.51)
⎣ h ⎦

ML
γ =
EI y GJ
(2.52)

Where h is the distance between the centroids of the top and bottom flange and K is the beam

parameter. The practical range for values of K is between 0.1 and 2.5, with low values

16
corresponding to long beams and/or beams with compact cross-sections, and higher values

## corresponding to short beams and/or beams with slender cross sections.

Using the above nondimensional parameters, Eq. (2.44) may be rewritten as (Wang, et al. 2005)

## γ = π ⎡⎢λ ± λ2 − υK 2 Λ + (1 − Λ ){1 + K 2 [4 ρ(1 − ρ) − ηΛ ]} ⎤⎥ (2.53)

⎣ ⎦

The non-dimensional buckling equation shown above is the general solution of M for

monosymmetric beams. Eq. (2.53) is a versatile equation because it also accurately predicts the

lateral-torsional buckling loads for beams of doubly symmetric cross-sections by simplifying the

terms in the equation so that the monosymmetric parameter, β, is equal to zero and the degree of

1
monosymmetry, ρ , is reduced to .
2

17
2.2 ENERGY METHOD

The second method used for determining lateral-torsional buckling loads in thin-walled

structures is the energy method. The energy method serves as a basis for the modern finite

## element method of computer solution for lateral-torsional buckling problems of complex

structures. The energy method is related to the differential equations of equilibrium method in

that calculus of variation can be used to obtain the differential equations derived by the first

method. The energy method is based on the principle that the strain energy stored in a member

during lateral-torsional buckling is equal to the work done by the applied loads. The critical

buckling loads can then be obtained by substituting approximate buckled shapes back into the

energy equation if the shape function is known. This approach is known as the Rayleigh-Ritz

method.

The strain energy stored in a buckled member can be broken down into two categories,

the energy from St. Venant torsion and from warping torsion. Pure or uniform torsion exists

when a member is free to warp and the applied torque is resisted solely by St. Venant shearing

stresses. When a member is restrained from warping freely, both St. Venant shearing stresses

and warping torsion resist the applied torque. This is known as non-uniform torsion.

18
2.2.1 Uniform Torsion

When a torque is applied to a member that is free to warp, the torque at any section is resisted by

shear stresses whose magnitudes vary based on distance from the centroid of the section. These

shear stresses are produced as adjacent cross-sections attempt to rotate relative to one another.

The St. Venant torsional resistance must directly oppose the applied torque as

Tsv = GJ
dz (2.57)

where φ is the angle of twist of the cross-section, G is the shearing modulus of elasticity, J is the

torsional constant, and z is direction perpendicular to the cross section, as illustrated in Figure

2.5.

T
z

## Figure 2.5 Twisting of a Rectangular Beam that is Free to Warp

19
2.2.2 Non-uniform Torsion

If the longitudinal displacements in the member are allowed to take place freely and the

longitudinal fibers do not change length, no longitudinal stresses are present and warping is

permitted to take place. However, certain loading and support conditions may be present that

prevent a member from warping. This warping restraint creates stresses which produce a torsion

in the member. Non-uniform torsion occurs when both St. Venant and warping torsion act on the

same cross section. The expression for non-uniform torsion can be given as

T = Tsv + Tw (2.58)

## where Tw is the warping torsion, which, for an I section, is

Tw = V f h (2.59)

where Vf is the shear force in each flange and h is equal to the height of the section. Recognizing

that the shear in the flange is the derivative of the moment present in the flange, Eq. (2.59)

becomes

dM f
Tw = − h (2.60)
dz

The bending moment in the upper flange, Mf, can be written in terms of the displacement in the

x-directon, u , as

d 2u
M f = EI f (2.61)
d z2

Recognizing that

φh
u= (2.62)
2

20
and introducing the cross-sectional property known as the warping moment of inertia

I f h2
Iw = (2.63)
2

## the warping torsion can now be expressed as

d 3φ
Tw = − EI w (2.64)
dz 3

The differential equation for non-uniform torsion is obtained by substituting Eq. (2.57) and Eq.

## (2.64) into Eq. (2.58) is

dφ d 3φ
T = GJ − EI w 3 (2.65)
dz dz

The first term refers to the resistance of the member to twist and the second term represents the

resistance of the member to warp. Together, the terms represent the resistance of the section to

an applied torque.

## 2.2.3 Strain Energy

The strain energy stored in a twisted member can be broken into two categories, the

energy due to St. Venant torsion and the energy due to warping torsion. The strain energy due to

## St. Venant torsion (Chajes, 1993) is

Tsv dφ
dU sv = (2.66)
2

where it can be seen that the change in strain energy stored in element dz due to St. Venant

torsion is equal to one half the product of the torque and the change in the angle of twist.

21
Tsv
dφ = dz (2.67)
GJ

## and substituting it into Eq. (2.66) yields

Tsv 2
dU sv = dz (2.68)
2GJ

Substituting Eq. (2.57) into Eq. (2.68) and integrating results in the expression for strain energy

## due to St. Venant torsion.

⎛ dφ ⎞
L 2

1
U sv = GJ ⎜ ⎟ dz (2.69)
2 0
⎝ dz ⎠

The strain energy due to the resistance to warping torsion of an I-beam, for example, is equal to

the bending energy present in the flanges. The bending energy stored in an element dz of one of

the flanges is equal to the product of one half the moment and the rotation as
2
1 ⎛ d 2u ⎞
dU w = EI f ⎜ 2 ⎟ dz (2.70)
2 ⎝ dz ⎠

## Substituting Eqs. (2.62) and (2.63) into Eq. (2.70) yields

2
1 ⎛ d 2φ ⎞
dU w = EI w ⎜ 2 ⎟ dz (2.71)
4 ⎝ dz ⎠

Integrating Eq. (2.71) over the length of the member, L, and multiplying by two to account for

the energy in both flanges results in the expression for the strain energy in a member caused by

resistance to warping.
2
L
⎛ d 2φ ⎞

1
Uw = EI w ⎜ 2 ⎟ dz (2.72)
2 0 ⎝ dz ⎠

The total strain energy in a member is then represented by the addition of Eqs. (2.69) and (2.72).

2 2
⎛ dφ ⎞ ⎛ d 2φ ⎞
L L

∫ ∫
1 1
U= GJ ⎜ ⎟ dz + EI w ⎜ 2 ⎟ dz (2.73)
2 o
⎝ dz ⎠ 2 0 ⎝ dz ⎠

22
2.2.4 Solutions Using Buckling Shapes

Case 5: A simply supported, doubly symmetric I-beam that is subjected to only equal end

The loading in this case is identical to Case 1, but this case consists of a beam with an I cross-

## section instead of a rectangular cross-section. As in Case 1, Mx = M and Mz = 0.

The boundary conditions for the case of uniform bending are given below.

d 2u d 2v
u= v= = =0 at z = 0, L (2.74)
dz 2 dz 2

d 2φ
φ= 2 =0 at z = 0, L (2.75)
dz

In order to find the critical moment by use of the energy method, it is necessary to find

the moment for which the total potential energy has a stationary value. The strain energy stored

in the beam consists of two parts; the energy due to the bending of the member about the y-axis

and the energy due to the member twisting about the z-axis. The total strain energy for the

section is
2 2 2
⎛ d 2u⎞ ⎛ dφ ⎞ ⎛ d 2φ ⎞
L L L

∫ ∫ ∫
1 1 1
U = EI y ⎜ 2 ⎟ dz + GJ ⎜ ⎟ dz + EI w ⎜ 2 ⎟ dz (2.76)
2 0 ⎝ dz ⎠ 2 0
⎝ dz ⎠ 2 0 ⎝ dz ⎠

The strain energy, U, must now be added to the potential energy of the external loads, Ω, to

determine a stationary value for Π = U + Ω . For a member subjected to uniform bending, the

external potential energy is equal to the negative product of the applied moments and the angles

## through which they act upon the beam.

23
Ω = − 2 Mψ (2.77)

where ψ is the angle of rotation about the x-axis of the beam and can be expressed as

du dφ
L

1
ψ= dz (2.78)
2 0 dz dz

Combining Eqs. (2.77) and (2.78) yields an expression for the potential energy of the external

du dφ
L

Ω = −M
∫ 0 dz dz
dz (2.79)

and yields the following expression for the total potential energy of the beam.
2 2 2
⎛ d 2u⎞ ⎛ dφ ⎞ ⎛ d 2φ ⎞ du dφ
L L L L

∫ ∫ ∫ ∫
1 1 1
Π = EI y ⎜ 2 ⎟ dz + GJ ⎜ ⎟ dz + EI w ⎜ 2 ⎟ dz − M dz
2 0 ⎝ dz ⎠ 2 0 ⎝ dz ⎠ 2 0 ⎝ dz ⎠ 0 dz dz

(2.80)

As stated at the beginning of this section, the Rayleigh-Ritz method for determining critical loads

requires the assumption of suitable expressions for buckling modes. The following buckling

## shapes satisfy our boundary conditions:

πz
u = A sin (2.81)
L

πz
φ = B sin (2.82)
L

Substituting the buckled shapes into Eq. (2.80) and identifying that

L
πz L
πz
∫ ∫
L
sin 2 dz = cos2 dz = (2.83)
0 L 0 L 2

## the total potential energy of the beam, Π , becomes

EI y A 2 L ⎛ π ⎞ 4 GJB 2 L ⎛ π ⎞ 2 E I w B 2 L ⎛ π ⎞ 4 MABL ⎛ π ⎞ 2
Π = ⎜ ⎟ + ⎜ ⎟ + ⎜ ⎟ − ⎜ ⎟ (2.84)
4 ⎝ L⎠ 4 ⎝ L⎠ 4 ⎝ L⎠ 2 ⎝ L⎠

24
Setting the derivative of Π with respect to A and B equal to zero, the critical moment can be

obtained.

d Π EI y π
2

= A − MLB = 0 (2.85)
dA L

dΠ ⎛ E I wπ 2 ⎞
= MLA − ⎜ GJL + ⎟B= 0 (2.86)
dB ⎝ L ⎠

If the deformed configuration of the beam is to yield a nontrivial solution, the determinant of the

## coefficients A and B in Eqs. (2.85) and (2.86) must vanish leaving

EI y π 2 ⎛ E I wπ 2 ⎞
⎜ GJL + ⎟ − M 2 L2 = 0 (2.88)
L ⎝ L ⎠

Solving for M in Eq. (2.88) yields the critical moment for a simply supported beam in uniform

bending as

π ⎛ π2⎞
M cr = EI y ⎜ GJ + EI w 2 ⎟ (2.89)
L ⎝ L ⎠

Case 6: A doubly symmetric I-beam with fixed ends that is subjected to only equal end moments

## For this case, as in Case 1, Mx = M and Mz = 0.

Consider an I-beam whose ends are free to rotate about a horizontal axis but restrained against

25
M  M

L

## The boundary conditions are as follows

du
u= =0 at z = 0, L (2.90)
dz

d 2v
v= =0 at z = 0, L (2.91)
dz 2

φ= =0 at z = 0, L (2.92)
dz

## The following buckling shapes satisfy the geometric boundary conditions

⎛ 2π z ⎞
u = A⎜ 1 − cos ⎟ (2.93)
⎝ L ⎠

⎛ 2 π z⎞
φ = B⎜ 1 − cos ⎟ (2.94)
⎝ L ⎠

Substituting the buckled shapes into Eq. (2.80) and using the simplification in Eq. (2.83), the

## total potential energy of the beam becomes

π2 ⎛ A2π 2 π2 ⎞
Π = ⎜ 4 EI y + GJB 2
+ 4 EI B 2 − 2 MAB⎟ (2.95)
l ⎝ L2 w
L ⎠

26
Setting the derivative of the Π equations with respect to A and B equal to zero, yield the

## following two equations.

dΠ π 2 ⎛ π2 ⎞
= ⎜ 8 EI y A 2 − 2 MB⎟ = 0 (2.96)
dA L ⎝ L ⎠

dΠ π 2 ⎛ π2 ⎞
= ⎜ 2GJB + 8 EI w B 2 − 2 MA⎟ = 0 (2.97)
dB L ⎝ L ⎠

## Eqs. (2.95) and (2.96) expressed in matrix form is

⎡ π2 ⎤
⎢ 4 EI y L2 − M ⎥ ⎧ A⎫
⎢ 2 ⎥⎨ ⎬ = 0 (2.98)
π B
⎢ − M GJ + 4 EI w 2 ⎥ ⎩ ⎭
⎢⎣ L ⎥⎦

If the deformed configuration of the beam is to yield a nontrivial solution, the determinant of the

## coefficients A and B in Eq. (2.98) must vanish leaving

π2 ⎛ π2 ⎞
4 EI y ⎜ GJ + 4 EI ⎟ − M2 = 0 (2.99)
L2 ⎝ w
L2 ⎠

## Solving for M as the critical moment yields

2π ⎛ π2 ⎞
M cr = EI y ⎜ GJ + 4 EI w 2 ⎟ (2.100)
L ⎝ L ⎠

It is interesting to note that the critical moment for the restrained beam is proportional to that of

the simply supported beam. If the warping stiffness is negligible compared to that of the St.

Venant stiffness, the critical moment for the fixed beam is twice that of the hinged beam. If the

St. Venant stiffness is negligible compared to that of the warping stiffness, the critical moment of

the fixed beam is four times that of the hinged beam. The reason for this is that lateral bending

strength and warping strength depend of the length of the beam, where St. Venant stiffness does

## not. St. Venant stiffness is therefore unaffected by a change in boundary conditions.

27
3.0 LATERAL-TORSIONAL BUCKLING OF BEAM-COLUMNS

The energy method detailed in Chapter 2, in conjunction with the Raleigh-Ritz method, is useful

in determining closed form or approximate solutions, with a high degree of accuracy, when a

suitable buckling mode can be identified. In more complex structural systems, identification of

the buckling mode is not possible. In this case, a finite element approach is an ideal method that

may be used to calculate the buckling load. In order to formulate element elastic and geometric

stiffness matrices that represent different load cases, one approach is to derive total potential

energy of a beam-column element with a concentrated force, distributed force, end moments, and

an axial force. Therefore, the objective of this chapter is to derive energy equations for a beam-

## Lateral-torsional buckling of a beam-column occurs when the loads on an element

become large enough to render its in-plane state unstable. When the loads on the member reach

these critical values, the section will deflect laterally and twist out of the plane of loading. At

critical loading, the compression flange of the member becomes unstable and bends laterally

while the rest of the member remains stable restraining the lateral flexure of the compression

flange, causing the section to rotate. This is common in slender beam-columns with insufficient

lateral bracing that have a much greater in-plane bending stiffness than their lateral and torsional

stiffnesses. It is important to know the critical load for lateral-torsional buckling because this

28
method of failure is often the primary failure mode for thin-walled structures. The focus of this

chapter is to formulate energy equations that can be used to derive element elastic and geometric

## 1. The member has a monosymmetric cross-section.

2. The beam-column remains elastic. This implies that the member must be long and

slender.

3. The cross-section of the member does not distort in its own plane after buckling and

## 5. The member is a compact section.

The orientation of the member used to derive the energy equations is depicted in Figure

3.1 using the xyz coordinate system with the origin being at o. The x-axis is the major principle

axis and the y-axis is the minor principle axis with the z-axis being oriented along the length of

the member, coinciding with the centroidal axis of the undeformed beam-column.

z‐axis h
o
z

y

## Figure 3.1 Coordinate System of Undeformed Monosymmetric Beam

29
The displacements in the x, y, and z directions are denoted as u, v, and w, respectively. If

loading occurs in the yz plane, the member will have an in-plane displacement, v, in the y-

dv
direction and an in-plane rotation . A member loaded along the z-axis will have a
dz

## displacement, w. The result of lateral-torsional buckling is an out-of-plane displacement, u, in

du
the x-direction, an out-of-plane rotation, , an out-of-plane twisting rotation, φ , and an out-of-
dz

plane torsional curvature, . It is assumed in this chapter that in-plane deformations w, v, and
dz

dv
are very small and therefore can be neglected. In the next chapter, the energy equations
dz

derived in this chapter will be expanded to include these displacements, which are known as

prebuckling deformations.

The applied loads on the beam column include; (1) a distributed load, q, which acts at a

height a above the centroidal axis (2) a concentrated load, P, which acts at a height e above the

centroidal axis (3) a concentric axial load, F (4) end moments, M1 and M2, as shown below.

z P
q

e
F  F a

M 1 M 2
V 1 V2

Figure 3.2 External Loads and Member End Actions of Beam Element

30
The energy equations for the lateral-torsional buckling of a member with a

monosymmetric cross-section differ from those of a member with a double symmetric cross-

section because the centroid and the shear center do not coincide on a monosymmetric beam-

column, as shown in Figure 3.3. This introduces the term, y o , into the derivation representing

## the distance between the shear center, s, and the centroid, o.

IyT

shear center, s
yo
x  centroid, o  h

IyB  y

## Figure 3.3 Cross-Section of Monosymmetric I-beam

The change in effective torsional stiffness of the member due to the components of

bending compressive and tensile stresses that produce a torque in the beam as it twists during

buckling is referred to as the Wagner effect (Anderson and Trahair, 1972). In a beam with a

doubly-symmetric cross-section, these compressive and tensile stresses balance each other and

do not affect the torsional stiffness of the beam. For the case of monosymmetry, these tensile

and compressive stresses do not balance each other and the resulting torque causes a change in

the effective torsional stiffness of the member from GJ to (GJ + Mx βx) (Wang and Kitipornchai,

1986). Because the smaller flange of the beam is farther away from the shear center than the

31
larger flange, it creates a larger moment arm and predominates in the Wagner effect. This means

that when the smaller flange is in tension, the effective torsional stiffness of the beam is

increased while the effective torsional stiffness is reduced when the smaller flange is in

compression. This inconsistency adds to the complexity of the energy equation derivations for

## lateral torsional buckling.

The energy equation for an elastic thin-walled member is derived by considering the strain

energy stored in the member, U, and the potential energy of the external loads, Ω , as

Π =U +Ω (3.1)

## where Π represents the total potential energy of the member.

The strain energy present in the member is the potential energy of the internal stresses

and strains present in the beam-column, while the potential energy of the loads represents the

negative of the work done by external forces. The total potential energy increment may be

written as

1 2 1
ΔΠ = δ Π + δ Π + δ 3 Π +... (3.2)
2! 3!

The theorem of stationary total potential energy states that of all kinematically admissible

deformations, the actual deformations (those which correspond to stresses which satisfy

equilibrium) are the ones for which the total potential energy assumes a stationary value (Pilkey

## As previously discussed, the structure is unstable at buckling. The theorem of minimum

total potential energy states that this stationary value of Π at an equilibrium position is

minimum when the position is stable. Therefore the equilibrium position can be considered

stable when

32
1 2
δ Π>0 (3.3)
2

## and likewise is unstable when

1 2
δ Π<0 (3.4)
2

Therefore, the critical condition for buckling would be when the total potential energy is equal to

zero, thus representing the transition from a stable to unstable state (Pi, et al., 1992).

1 2
δ Π=0 (3.5)
2

## Substituting Eq. (3.1) yields the critical condition for buckling as

1 2
2
(δ U + δ2Ω) = 0 (3.6)

## 3.1 STRAIN ENERGY

The strain energy portion of the total potential energy may be expressed as a function of the

longitudinal and shear strains as well as stresses. Assume an arbitrary point Po in the cross

section of the thin walled member. The strain energy of the member can be expressed as

∫ ∫ (ε σ )
1
U= p p + γ pτ p dA dz (3.7)
2
L A

where

## γ p = shear strain of point Po

33
τ p = shear stress of point Po

## With its second variation being

∫ ∫ (δε δσ )
1 2 1
δ U= p p + δγ pδτ p + δ 2 ε pσ p + δ 2 γ pτ p dA dz (3.8)
2 2 L A

## ε p , σ p , γ p , and τ p and their variations will be expressed in terms of centroidal displacements

in the following section in order to derive the energy equation for lateral-torsional buckling.

3.1.1 Displacements

## In order to properly investigate deformations in a beam-column, two sets of coordinate systems

are defined. In the fixed global coordinate system oxyz, the axis oz is fixed and coincides with

the centroidal axis of the undeformed beam. The axes ox and oy represent the principle axes of

o
z

ô z

x

## Figure 3.4 Deformed Beam

34
The second set of coordinate systems is a moving, right hand, local coordinate system,

    . The origin of this coordinate system is at point o located on the centroidal axis of the
oxyz

beam and moves with the beam during displacement, as shown in Figure 3.4. The axis

  coincides with the tangent at o after the centroidal axis has been deformed. The principle
oz

  and oy
axes of the deformed beam are ox  .

When the beam column element buckles, point Po on the beam moves to point P. This

deformation occurs in two stages. Point Po first translates to point Pt by the displacements u, v,

and w. The point Pt then rotates through an angle θ to the point P about the line on where the

line on passes through the points o and o . After the rotation, the moving local coordinate

    becomes fixed. The transition of point Po to point P can be seen in Figure 3.5. The
system oxyz

  , oy
directional cosines of the moving axes ox   , and oz
  relative to the fixed global axes ox, oy,

and oz can be determined by assuming rigid body rotation of the axes through an angle θ (Pi, et

al., 1992).

z
o zˆ z
Po

xˆ P
x y
Pt
0

y ô

x̂ n

## Figure 3.5 Translation of Point Po to Point P

35
The displacements of point Po can be expressed as (Pi, et al., 1992)

⎧ up ⎫ ⎧ u ⎫ ⎧ x ⎫ ⎧ x ⎫
⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪
⎨ v p ⎬ = ⎨ v ⎬ + TR ⎨ y − y o ⎬ − ⎨ y − y o ⎬ (3.9)
⎪ w ⎪ ⎪ w⎪ ⎪ − ωk ⎪ ⎪ 0 ⎪
⎩ p⎭ ⎩ ⎭ ⎩ z⎭ ⎩ ⎭

Where up, vp, and wp are total displacements of general point Po ( x , y ,0) . Note that u, v, and w

are shear center displacements and ( y − y o ) is the distance between the centroid and the shear

center, as seen in Figure 3.3. ω is the section warping function (Vlasov, 1961) and − ω k z is

## the warping displacement and represents the deformation in the z-direction.

The first term on the right side of the equation contains the shear center displacement as

the point Po translates laterally to point Pt. The remaining terms on the right side of the equation

represent the rotation of point Pt to its final destination at point P. TR is defined as the rotational

## transformation matrix of the angle of rotation, θ. Assuming small angles of rotation

⎡ θ y 2 θz 2 θ xθ y θ xθ z ⎤
⎢1 − − − θz + θy + ⎥
⎢ 2 2 2 2 ⎥
⎢ θ x θ y θx 2 θz 2 θ yθ z ⎥
TR = ⎢ θ + 1− − − θx + (3.10)
2 2 2 2 ⎥
⎢ θ yθ z θ x θ y ⎥⎥
2
⎢ − θ y + θ xθ z
2

θx + 1− −
⎢⎣ 2 2 2 2 ⎥⎦

where θ x , θ y , and θz are components of the rotation θ in the x, y, and z directions, respectively

(Torkamani, 1998).

## Consider an undeformed element Δ z and its deformed counterpart Δ z(1 + ε ) , where ε

represents the strain. Δ u, Δ v , and ( Δ z + Δ w) are components of the deformed element length

36
Δ z(1 + ε ) on the ox, oy, and oz axes, respectively. The relationship between the deformed

## element and its components can be expressed as

G G G G
Δ z(1 + ε ) N z = Δ ui + Δ v j + ( Δ z + Δ w) k (3.11)
G
  direction.
where N z is a unit vector in the oz
G
Δ z(1 + ε ) N z is projected on the x and y axes as
G G
Δ u = Δ z (1 + ε ) N z ⋅ i = Δ z ( 1 + ε ) l z (3.12)
G G
Δ v = Δ z (1 + ε ) N z ⋅ j = Δ z (1 + ε ) mz (3.13)

##   direction with respect to the oxyz fixed

where lz, mz, and nz are directional cosines of the oz

coordinate system.

Dividing the previous equations by Δ z , and taking the limit as Δ z goes to zero gives

du Δu Δ z (1 + ε ) l z
= lim = lim = (1 + ε ) l z (3.14)
dz Δz→ o Δ z Δz→ o Δz

dv Δv Δ z (1 + ε ) mz
= lim = lim = (1 + ε ) mz (3.15)
dz Δz → o Δ z Δz→ o Δz

## Where lz and mz are defined as . (Torkamani, 1998)

θ xθ z
lz = θ y + (3.16)
2

θ yθ z
mz = −θ x + (3.17)
2

Therefore

du ⎛ θθ ⎞
= ⎜ θ y + x z ⎟ (1 + ε ) (3.18)
dz ⎝ 2 ⎠

dv ⎛ θ yθ z ⎞
= ⎜ − θx + ⎟ (1 + ε ) (3.19)
dz ⎝ 2 ⎠

37
By disregarding higher order terms, strain is eliminated and Eqs. (3.18) and (3.19) become

du ⎛ θθ ⎞
≈ ⎜θ y + x z ⎟ (3.20)
dz ⎝ 2 ⎠

dv ⎛ θ yθz ⎞
≈ ⎜ − θx + ⎟ (3.21)
dz ⎝ 2 ⎠

## Solving for θ x and θ y from the previous equations yields

dv 1 du
θx = − + θz (3.22)
dz 2 dz

du 1 dv
θy = + θz (3.23)
dz 2 dz

##   axis onto the oy axis and the oy

Projecting the unit lengths along the ox   axis onto the ox axis

## yield mx and l y respectively. (Torkamani, 1998)

θ xθ y
l y = −θz + (3.24)
2

θ xθ y
mx = θ z + (3.25)
2

θ xθ y θ xθ y
The projections − θ z + and θz +   and ox
of unit lengths along the oy   axes onto the ox
2 2

and oy axes, respectively, can be used to define the the mean twist rotation, φ , of the ox
  and oy


## axes on the oz axis is

1 ⎡⎛ θ xθ y ⎞ ⎛ θ xθ y ⎞ ⎤
φ= ⎢⎜ θz + ⎟ − ⎜ − θz + ⎟⎥ (3.26)
2 ⎣⎝ 2 ⎠ ⎝ 2 ⎠⎦

Which simplifies to

φ = θz (3.27)

38
Substituting the expressions for θ x , θ y , and θz into the rotational transformation matrix, TR ,

yields

⎡ lx ly lz ⎤
⎢ ⎥
TR = ⎢mx my mz ⎥ (3.28)
⎢ nx ny nz ⎥⎦

Where
2
1 ⎛ du ⎞ 1 1 du dv
lx = 1 − ⎜ ⎟ − φ 2 − φ (3.29)
2 ⎝ dz ⎠ 2 2 dz dz

2 2
1 du dv 1 ⎛ du ⎞ 1 ⎛ dv ⎞
l y = −φ − + ⎜ ⎟ φ− ⎜ ⎟ φ (3.30)
2 dz dz 4 ⎝ dz ⎠ 4 ⎝ dz ⎠

du
lz = (3.31)
dz
2 2
1 du dv 1 ⎛ dv ⎞ 1 ⎛ du ⎞
mx = φ − − ⎜ ⎟ φ+ ⎜ ⎟ φ (3.32)
2 dz dz 4 ⎝ dz ⎠ 4 ⎝ dz ⎠

2
1 ⎛ dv ⎞ 1 1 du dv
my = 1 − ⎜ ⎟ − φ 2 + φ (3.33)
2 ⎝ dz ⎠ 2 2 dz dz

dv
mz = (3.34)
dz

du dv 1 du 2
nx = − − φ+ φ (3.35)
dz dz 4 dz

dv du 1 dv 2
ny = − + φ+ φ (3.36)
dz dz 4 dz
2 2
1 ⎛ du ⎞ 1 ⎛ dv ⎞
nz = 1 − ⎜ ⎟ − ⎜ ⎟ (3.37)
2 ⎝ dz ⎠ 2 ⎝ dz ⎠

## The torsional curvature is (Love, 1944)

d lx d mx d nx
kz = ly + my + n (3.38)
dz dz dz y

39
Substituting l x through n y into the previous expression yields a nonlinear expression for

torsional curvature as

dφ 1 ⎛ d 2 u dv d 2 v du ⎞
kz = + ⎜ − ⎟ (3.39)
dz 2 ⎝ dz 2 dz dz 2 dz ⎠

Eliminating second and higher order terms, the expression for torsional curvature may be

simplified to

kz = (3.40)
dz

Substituting Eqs. (3.29) – (3.37) into Eq. (3.9) yield the displacement of an arbitrary point Po in

terms of shear center displacements, rotations, and the section warping. The total displacements

up, vp, and wp can be considered the sum of linear and quadratic components of the form

⎧ u p ⎫ ⎧ u pl ⎫ ⎧ u pn ⎫
⎪ ⎪ ⎪ ⎪ ⎪ ⎪
⎨ v p ⎬ = ⎨ v pl ⎬ + ⎨ v pn ⎬ (3.41)
⎪w ⎪ ⎪w ⎪ ⎪w ⎪
⎩ p ⎭ ⎩ pl ⎭ ⎩ pn ⎭

where

⎧ ⎫
⎧ u pl ⎫ ⎪ u − ( y − y o )φ ⎪
⎪ ⎪ ⎪ ⎪
⎨ v pl ⎬ = ⎨ v + xφ ⎬ (3.42)
⎪w ⎪ ⎪ du dv dφ ⎪
⎩ pl ⎭ ⎪ w − x − y −ω ⎪
⎩ dz dz dz ⎭

⎧ 1 ⎛ ⎛ du ⎞ 2 ⎛ du ⎞ ⎛ dv ⎞ ⎞ 1 ⎛ du dv 1 ⎛ du ⎞ 2 1 ⎛ dv ⎞ ⎞
2
du dφ ⎫
⎪ − x ⎜ ⎜ ⎟ + φ + ⎜ ⎟ ⎜ ⎟ φ ⎟ − ( y − y o )⎜
2
− ⎜ ⎟ φ + ⎜ ⎟ φ⎟ − ω ⎪
⎪ 2 ⎝ ⎝ dz ⎠ ⎝ dz ⎠ ⎝ dz ⎠ ⎠ 2 ⎝ dz dz 2 ⎝ dz ⎠ 2 ⎝ dz ⎠ ⎠ dz dz ⎪
⎧ u pn ⎫
⎪ ⎪⎪ 1 ⎛ du dv 1 ⎛ dv ⎞ 1 ⎛ du ⎞ ⎞ 1 ⎛ ⎛ dv ⎞ 2 du dv ⎞ dv dφ ⎪⎪
2 2

⎨ v pn ⎬ = ⎨ − x ⎜ + ⎜ ⎟ φ − ⎜ ⎟ φ⎟ − ( y − y o )⎜ ⎜ ⎟ + φ 2 − φ⎟ − ω ⎬
⎪w ⎪ ⎪ 2 ⎝ dz dz 2 ⎝ dz ⎠ 2 ⎝ dz ⎠ ⎠ 2 ⎝ ⎝ dz ⎠ dz dz ⎠ dz dz ⎪
⎩ pn ⎭ ⎪ ⎛ dv 1 du 2 ⎞ ⎛ du 1 dv 2 ⎞ 1 dφ ⎛ ⎛ du ⎞
2
⎛ dv ⎞ ⎞
2 ⎪
⎪ − x ⎜ φ − φ ⎟ + ( y − y o )⎜ φ + φ ⎟ + ω ⎜⎜ 2⎟ + ⎜ 2⎟ ⎟ ⎪
⎪⎩ ⎝ dz 4 dz ⎠ ⎝ dz 4 dz ⎠ 2 dz ⎝ ⎝ dz ⎠ ⎝ dz ⎠ ⎠ ⎪⎭

(3.43)

40
The relationship between the displacement of the shear center, ws, and the displacement of the

## centroid, w, in the z-direction can be expressed as (Pi and Trahair, 1992a)

⎛ dv 1 dv 2 du ⎞
ws = w − y o ⎜ − φ − φ⎟ (3.44)
⎝ dz 2 dz dz ⎠

## The derivatives of up, vp, and wp with respect to z are

du p du dφ dφ ⎛ du dv ⎞
= − y + y o + Ox ⎜ , , φ ⎟ (3.45)
dz dz dz dz ⎝ dz dz ⎠

dv p dv dφ ⎛ du dv ⎞
= − x + Oy ⎜ , , φ ⎟ (3.46)
dz dz dz ⎝ dz dz ⎠

dw p dw d 2u d 2v d 2φ dφ dv d 2v dφ du d 2u ⎛ du dv ⎞
= − x 2 − y 2 − ω 2 − x − xφ 2 + y + y φ 2 + Oz ⎜ , , φ⎟
dz dz dz dz dz dz dz dz dz dz dz ⎝ dz dz ⎠

(3.47)

The terms Ox, Oy, and Oz represent terms that are second order or higher and may be

disregarded.

3.1.2 Strains

The longitudinal normal strain εp of point P can be expressed in terms of the rates of change of

## the displacements of point P as

1 ⎛ ⎛ du p ⎞ ⎞
2 2 2
dw p ⎛ dv p ⎞ ⎛ dw p ⎞
εp = + ⎜⎜ ⎜ ⎟ +⎜ ⎟ +⎜ ⎟ ⎟
⎟ (3.48)
dz 2 ⎝ ⎝ dz ⎠ ⎝ dz ⎠ ⎝ dz ⎠ ⎠

2 2 2
⎛ dw p ⎞ ⎛ du p ⎞ ⎛ dv p ⎞
For small strains, ⎜ ⎟ is small compared to ⎜ ⎟ and ⎜ ⎟ and can be disregarded.
⎝ dz ⎠ ⎝ dz ⎠ ⎝ dz ⎠

Therefore

41
1 ⎛ ⎛ du p ⎞ ⎞
2 2
dw p ⎛ dv p ⎞
εp = + ⎜⎜ ⎜ ⎟ +⎜ ⎟ ⎟
⎟ (3.49)
dz 2 ⎝ ⎝ dz ⎠ ⎝ dz ⎠ ⎠

Substituting in the derivatives of the displacements up, vp, and wp of point Po yields

d 2 φ 1 ⎛ ⎛ du ⎞ ⎛ dv ⎞ ⎞ du dφ
2 2
dw d 2u d 2v d 2v d 2u
εp = − x 2 − y 2 −ω 2 +
  ⎜ ⎟
⎜ ⎟ + ⎜ ⎟ − x 2 φ + y 2 φ + y o
 
dz dz dz dz 2 ⎝ ⎝ dz ⎠ ⎝ dz ⎠ ⎠ dz dz dz dz

( )
2 ⎛ dφ ⎞
2
1 2
+ x + ( y − y o ) ⎜ ⎟ (3.50)
2 ⎝ dz ⎠

## The first variation of the longitudinal strain is

dδ w d 2δ u d 2δ v d 2δφ dδ u du dδ v dv d 2δ v d 2v d 2δ u
δε p = −x
 − y
 −ω + + −x
 φ − x 2 δφ + y 2 φ
 
dz dz 2 dz 2 dz 2 dz dz dz dz dz 2 dz dz

+ y
d 2u
dz 2 δφ + y o

dδ u dφ
dz dz
+ y o
du dδφ
dz dz
(
+ x 2 + ( y − y o ) )
2 dδφ dφ

dz
(3.51)

## The second variation of the longitudinal strain is

⎛ dδ u ⎞ ⎛ dδ v ⎞ d 2δ v d 2δ u dδ u dδφ
( )
2 2
2 ⎛ dδφ ⎞
2

δ 2ε p = ⎜ ⎟ + ⎜ ⎟ − 2 
x δφ + 2 
y 2 δφ + 2 y o
 + x 2 + ( y − y o ) ⎜ ⎟
⎝ dz ⎠ ⎝ dz ⎠ dz 2
dz dz dz ⎝ dz ⎠

(3.52)

It is assumed that the second variations of the displacements in Eq. (3.52) vanishes. The

above equations contain a combination of the strains before and after buckling. In this case, the

prebuckling displacements are defined as v and w. During buckling, the displacements are

defined as δ u and δφ . Therefore, the displacements u, φ , δ v , and δ w are equal to zero and

## may be eliminated from the above equations.

The equations for longitudinal strain and its first and second variation thus become

2
dw d 2 v 1 ⎛ dv ⎞
εp = − y 2 + ⎜ ⎟
 (3.53)
dz dz 2 ⎝ dz ⎠

42
d 2δ u d 2δφ d 2v
δε p = − x −ω − x 2 δφ (3.54)
dz 2 dz 2 dz

⎛ dδ u ⎞ d 2δ u dδ u dδφ
( )
2 ⎛ dδφ ⎞
2 2

δ 2ε p = ⎜ ⎟ + 2 y 2 δφ + 2 y o
 + x 2 + ( y − y o ) ⎜ ⎟ (3.55)
⎝ dz ⎠ dz dz dz ⎝ dz ⎠

The shear strains due to bending and warping of the thin walled section are neglected. The shear

## strain at point Po due to uniform torsion is (Pi and Trahair, 1992a)

γ p = −2t p (3.56)
dz

Where tp represents the distance of Po from the midthickness line of the cross-section.

## The first variation of the shear strain is

dδφ
δγ p = − 2t p (3.57)
dz

## And the second variation of the shear strain is assumed to be zero.

δ 2γ p = 0 (3.58)

## The stresses at point Po can be related to the strains by Hooke’s Law as

⎧σ p ⎫ ⎡ E 0⎤⎧εp ⎫
⎨ ⎬= ⎢ ⎨ ⎬ (3.59)
⎩τp ⎭ ⎣ 0 G ⎥⎦ ⎩ γ p ⎭

Mx =
∫ yσ dA
A
p (3.60)

43
F=
∫ σ dA
A
p (3.61)

So that

F M x y
σp = + (3.62)
A Ix

## 3.1.4 Section Properties

∫ xdA
A
∫ =0
 = ydA
A
(3.63)

∫ x ydA
A
 =0 (3.64)

A=
∫ dA (3.65)

Ix =
∫ y dA
A
2
(3.66)

Iy =
∫ x dA
A
2
(3.67)

These section properties are valid for not only monosymmetric beams, but doubly symmetric

ones as well. To solve the energy equation for monosymmetric beams, additional section

## properties must be introduced.

I yC I yC
ρ= = (3.68)
I yT + I yC Iy

44
Where I yT and I yC are the minor axis second moments of area of the tension and compression

flanges, respectively (Kitipornchai and Trahair, 1980). The value of ρ(can range from 0 to 1,

with doubly symmetric I-beams having a value of 0.5. The warping moment of inertia given in

## Eq. (3.69), with other section properties listed below.

I w = ρ(1 − ρ) I y h 2 (3.69)

J= ∑ bt 3
3
(3.70)

Ix + Iy
ro 2 = + y o 2 (3.71)
A

1⎛ ⎞
βx = ⎜
Ix ⎝ ∫
A
x 2 y dA +

A
y 3 dA⎟ − 2 y o

(3.72)

where ro is the polar radius of gyration for the beam about the shear center. βx is known as the

monosymmetric parameter (Trahair and Nethercot, 1984) . The term βx arises from the Wagner

effect discussed previously in this chapter when the compressive stresses do not equally oppose

the tensile stresses in a member. In the case of doubly symmetric beams, the stresses balance

each other and βx is equal to zero. When the smaller flange of a monosymmetric beam is in

## compression, there is a reduction in the effective torsional stiffness and βx is negative.

Conversely, when the smaller flange is in tension, the effective torsional stiffness is greater and

βx is positive.

45
3.2 STRAIN ENERGY EQUATION FOR MONOSYMMETRIC BEAM-COLUMN

With the newly introduced section properties, the second variation of the strain energy of the

## properties into Eq. (3.8) yields

2
⎧ ⎛ d 2 (δ u) ⎞ ⎛ d (δφ ) ⎞ ⎛ d 2 (δφ ) ⎞
2 2

1 2 1 ⎪ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟
δ U= ⎨ EI y ⎜ 2 ⎟ + GJ ⎜ dz ⎟ + EI ω ⎜ dz 2 ⎟
2 2 ⎪⎩ ⎝ dz ⎠ ⎝ ⎠ ⎝ ⎠
L

⎡ ⎛ d (δ u) ⎞ 2 ⎛ d (δ u) ⎞ ⎛ d (δφ ) ⎞ ⎛ d (δφ ) ⎞ ⎤
2

+ F ⎢ ⎜⎜ ⎜ dz ⎟ ⎜ dz ⎟ + (ro + y o )⎜ dz ⎟ ⎥
⎟ + 2 y o ⎜ ⎟⎜ ⎟ 2 2 ⎜ ⎟ ⎥
⎢ ⎝ dz ⎟⎠ ⎝ ⎠⎝ ⎠ ⎝ ⎠ ⎦

⎡ ⎛ d 2 (δ u) ⎞ ⎛ d (δφ ) ⎞ ⎤ ⎫⎪
2

+ M x ( z ) ⎢ 2⎜⎜ ⎟ δφ + βx ⎜ ⎟ ⎥
⎜ dz ⎟ ⎥ ⎬ dz (3.73)
⎢ ⎝ dz 2 ⎟⎠ ⎝ ⎠ ⎦⎪
⎣ ⎭

## 3.3 POTENTIAL ENERGY OF THE EXTERNAL LOADS

The equations for the potential energy possessed by the loads, or the work done by external

forces, are derived by multiplying the loads by their corresponding displacements, and summing

them up.

46
Ω = −

L
(v q q )dz − ∑ (v P − dvdz
P
M
M + wF F ) (3.74)

## With its second variation being

∫ (δ v q)dz − ∑
1 2 ⎛ 2 dδ 2 v M ⎞
δ Ω =− 2
q ⎜ δ vP P − M + δ 2 wF F ⎟ (3.75)
2 ⎝ dz ⎠
L

Where

## v M = vertical displacement which moment M acts

dv M
= rotation caused by moment M
dz

## acts may be found by

vq = v + my a − a (3.76)

47
Where
2
1 ⎛ dv ⎞ 1 1 du dv
my = 1 − ⎜ ⎟ − φ 2 + φ (3.77)
2 ⎝ dz ⎠ 2 2 dz dz

Therefore,

1 ⎡ ⎛ dv ⎞ 2 du dv ⎤
v q = v − (a − y o )⎢ ⎜ ⎟ + φ 2 − φ⎥ (3.78)
2 ⎣ ⎝ dz ⎠ dz dz ⎦

## concentrated load, P, acts is expressed as

1 ⎡ ⎛ dv ⎞ 2 du dv ⎤
v P = v − (e − y o )⎢ ⎜ ⎟ + φ 2 − φ⎥ (3.79)
2 ⎣ ⎝ dz ⎠ dz dz ⎦

## which concentrated moment Mx acts is

dv M dv
= (3.80)
dz dz

Because the effects of prebuckling are negected, the deformation v and its derivatives are

reduced to zero. These effects will be implemented in a later section. The displacements

## corresponding to the external loads then simplify to

1
vq = −
2
( a − y o )φ 2 (3.81)

1
vP = −
2
(e − y o )φ 2 (3.82)

dv M
=0 (3.83)
dz

## The second variations of the displacements are

48
1
δ 2 vq = −
2
( a − y o )(δφ )
2
(3.84)

1
δ 2vP = −
2
(e − y o )(δφ ) 2
(3.85)

dv M
=0 (3.86)
dz

## Substituting Eqs. (3.84) – (3.86) into Eq. (3.75) yields

1 2
2
δ Ω =
1
2 ∫
q(a − y o )(δφ ) dz +
L
2 1
2 ∑ P(e − y )(δφ)
o
2
(3.87)

## 3.4 ENERGY EQUATION FOR LATERAL TORSIONAL BUCKLING OF

MONOSYMMETRIC BEAMS

The second variation of the total potential energy equation for lateral torsional buckling of

monosymmetric beams is the sum of the second variation of the strain energy equation given in

Eq. (3.74) and the second variation of the potential energy of the loads given in Eq. (3.87), as

shown below.

49
2
⎧ ⎛ d 2 (δ u) ⎞ ⎛ d (δφ ) ⎞ ⎛ d 2 (δφ ) ⎞
2 2

1 2 1 ⎪ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟
δ Π = ⎨ EI y ⎜ 2 ⎟ + GJ ⎜ dz ⎟ + EI ω ⎜ dz 2 ⎟
2 2 ⎪⎩ ⎝ dz ⎠ ⎝ ⎠ ⎝ ⎠
L

⎡ ⎛ d (δ u) ⎞ 2 ⎛ d (δ u) ⎞ ⎛ d (δφ ) ⎞ ⎛ d (δφ ) ⎞ ⎤
2

+ F ⎢ ⎜⎜ ⎜ dz ⎟ ⎜ dz ⎟ + (ro + y o )⎜ dz ⎟ ⎥
⎟ + 2 y o ⎜ ⎟⎜ ⎟ 2 2 ⎜ ⎟ ⎥
⎢ ⎝ dz ⎟⎠ ⎝ ⎠ ⎝ ⎠ ⎝ ⎠ ⎦

⎡ ⎛ d 2 (δ u) ⎞ ⎛ d (δφ ) ⎞ ⎤ ⎫⎪
2

+ M x ( z ) ⎢ 2⎜⎜ ⎟ δφ + βx ⎜ ⎟ ⎥
⎜ dz ⎟ ⎥ ⎬ dz
⎢ ⎝ dz 2 ⎟⎠ ⎝ ⎠ ⎦⎪
⎣ ⎭

+
1
2∫q(a − y o )(δφ ) dz +
L
2 1
2 ∑ P(e − y )(δφ)
o
2
=0 (3.88)

Where

z2
M x = M 1 + V1 z − q for 0 < z < z p
2

M x = M 1 + V1 z − q
z2
2
( )
− P z − z p for z p < z < L

## z p = distance along beam in z-direction of applied concentrated load, P

The terms in the energy equation can be separated into three groups. The first group consists of

the terms that contain the buckling rigidities EIy, GJ, and EI ω and represent strain energy stored

during buckling. The second group consists of the terms that contain the stress resultants F and

Mx, which represent the work done by the applied loads at the shear center. The third group

consists of the remaining terms which represent the work done by transverse forces q and P.

50
3.5 NON-DIMENSIONAL ENERGY EQUATION FOR LATERAL-TORSIONAL

BUCKLING

The energy equation presented in the previous section has limitations in predicting a lateral-

torsional buckling parameter obtained from the solution of the eigenvalue problem because it

depends on beam properties such as E, G, L, etc. Elimination of these properties through a non-

dimensional analysis will provide more general results to determine critical loads and moments

## The stiffness parameter, K, of the beam is defined as

π 2 EI ω π 2 EI y h 2
K= ≈ (3.89)
GJL2 4GJL2

PL2
γP = (3.90)
EI y GJ

qL3
γq = (3.91)
EI y GJ

βx EI y
βx = (3.92)
L GJ

ro EI y
ro = (3.93)
L GJ

51
2 y o
yo = (3.94)
h

δ u EI y
δu = (3.95)
L GJ

FL2
F= (3.96)
EI y

M1 L
M1 = (3.97)
EI y GJ

V1 L2
V1 = (3.98)
EI y GJ

z
z= (3.99)
L

zp
zp = (3.100)
L

2a
a= (3.101)
h

2e
e= (3.102)
h

where

## h = the distance between the centroids of the flanges

The application of these parameters to the energy equation derived in the previous section yield a

non-dimensional energy equation that provides a buckling parameter for lateral torsional

52
2Π L
Π = (3.103)
GJ

## The second variation of the non-dimensional total potential energy equation is

⎧ ⎛ d 2 (δ u ) ⎞ 2 ⎛ d (δφ ) ⎞ 2
2

1 2 ⎛
d 2 (δφ ) ⎞

1 2 1 ⎪⎜ ⎟ +⎜ ⎟ K ⎜ ⎟ ⎪
δ Π = ⎨⎜ ⎜ dz ⎟ + π 2 ⎜ dz 2 ⎟ ⎬ dz
2 2 0 ⎪⎝ dz 2 ⎟⎠ ⎝ ⎠ ⎝ ⎠ ⎪
⎩ ⎭

1 ⎡ ⎛ d (δ u ) ⎞ 2 ⎛ d (δ u ) ⎞ ⎛ d (δφ ) ⎞ ⎛ d (δφ ) ⎞ ⎤
2

+F
∫ 0
⎢⎜
⎢ ⎜⎝ dz ⎟⎠

⎜ dz ⎟ ⎜ dz ⎟ + (ro + y o )⎜ dz ⎟ ⎥dz
⎟ + 2 yo ⎜

⎟⎜
⎠ ⎝

2 2 ⎜

⎟ ⎥
⎠ ⎦

1 ⎡ ⎛ d 2 (δ u ) ⎞ ⎛ d (δφ ) ⎞ ⎤
2

+ M
∫ 0
⎢⎜
⎢ ⎜⎝ dz 2 ⎟⎠

⎟ δφ + βx ⎜ ⎟ ⎥
⎜ dz ⎟ ⎥ dz
⎝ ⎠ ⎦

K⎡
∑ 2⎤
1

∫ q (a − y o )(δφ ) dz + Pi (e − y o )(δφi ) ⎥
2
+ ⎢ (3.104)
π⎣ 0 ⎦

where

z2
M x = M 1 + V1 z − q for 0 < z < z p
2

M x = M 1 + V1 z − q
z2
2
( )
− P z − z p for z p < z < 1

53
4.0 LATERAL-TORSIONAL BUCKLING OF MONOSYMMETRIC BEAMS

## CONSIDERING PREBUCKLING DEFLECTIONS

In the previous chapter, the effects of prebuckling deflections were ignored. These effects are

ignored assuming that the thin-walled object is almost perfectly straight and any deformation is

so small that it may be disregarded. This assumption is only valid when the ratios of minor axis

flexural stiffness and torsional stiffness to the major axis flexural stiffness are very small. In the

case where the ratios are not small, the effects of prebuckling deflections may significantly alter

## 4.1 STRAIN ENERGY CONSIDERING PREBUCKLING DEFLECTIONS

4.1.1 Displacements

In the previous chapter, the expression for torsional curvature was simplified to include just the

first term because the other terms in the expression were deemed to be relatively small. To

consider the effects of prebuckling deflections, the torsional curvature, kz, must be represented

by:

54
1 ⎛ d 2 u dv du d 2 v ⎞
kz = φ 2 + ⎜ 2 − ⎟ (4.1)
2 ⎝ dz dz dz dz 2 ⎠

## Which creates a longitudinal displacement of

⎡ du dv dφ ⎤ ⎡ ⎛ dv 1 du 2 ⎞ ⎛ du 1 dv 2 ⎞
w p = ⎢ w − x − y − ω ⎥ + ⎢ − x ⎜ φ − φ ⎟ + y ⎜ φ + φ ⎟
⎣ dz dz dz ⎦ ⎣ ⎝ dz 4 dz ⎠ ⎝ dz 4 dz ⎠

⎛ 1 ⎛ d 2 u dv d 2 v du ⎞ 1 ⎛ dφ 1 ⎛ d 2 u dv d 2 v du ⎞ ⎞ ⎛ ⎛ du ⎞ 2 ⎛ dv ⎞ 2 ⎞ ⎞ ⎤
− ω ⎜⎜ ⎜ 2 − 2 ⎟− ⎜ + ⎜ 2 − ⎟ ⎟ ⎜ ⎜ ⎟ + ⎜ ⎟ ⎟ ⎟⎟ ⎥
⎝ 2 ⎝ dz dz dz dz ⎠ 2 ⎝ dz 2 ⎝ dz dz dz dz ⎠ ⎠ ⎝ ⎝ dz ⎠ ⎝ dz ⎠ ⎠ ⎠ ⎥

(4.2)

## with the first derivative of w p being

dw p dw d 2u d 2v d 2 φ ω ⎡ d 3 u dv d 3 v du ⎤ ⎡ dφ dv d 2 v 1 dφ du
= − x 2 − y 2 − ω 2 − ⎢ 3 − − x ⎢ +φ 2 − φ
dz dz dz dz dz 2 ⎣ dz dz dz 3 dz ⎥⎦ ⎣ dz dz dz 2 dz dz

1 2 d 2u ⎤ ⎡ dφ du d 2 u 1 dφ dv 1 2 d 2 v ⎤ ⎛ du dv ⎞
− φ 2 ⎥ + 
y ⎢ + φ 2 + φ + φ 2 ⎥ + Oz ⎜ , , φ⎟
4 dz ⎦ ⎣ dz dz dz 2 dz dz 4 dz ⎦ ⎝ dz dz ⎠

(4.3)

where Oz represents functions which are 4th order or higher and disregarded for simplicity.

## The longitudinal strain in the previous chapter is give as

1 ⎛ ⎛ du p ⎞ ⎞
2 2
dw p ⎛ dv p ⎞
εp ≈ + ⎜⎜ ⎜ ⎟ +⎜ ⎟ ⎟
⎟ (4.4)
dz 2 ⎝ ⎝ dz ⎠ ⎝ dz ⎠ ⎠

55
du p dv p
Substituting the displacements and derived in the previous chapter and the new
dz dz

dw p
expression for considering prebuckling deflections yields
dz

dw d 2u d 2v d 2 φ ω ⎡ d 3 u dv d 3 v du ⎤ ⎡ dφ dv d 2 v 1 dφ du
εp = − x 2 − y 2 − ω 2 − ⎢ 3 − 3 − 
x + φ − φ
dz dz dz dz 2 ⎣ dz dz dz dz ⎥⎦ ⎢⎣ dz dz dz 2 2 dz dz

⎡ dφ du d 2 u 1 dφ dv 1 2 d 2 v ⎤ 1 ⎛ ⎛ du ⎞ ⎛ dv ⎞ ⎞
2 2
1 2 d 2u ⎤
− φ + 
y + φ + φ + φ + ⎜ ⎜ ⎟ + ⎜ ⎟ ⎟
4 dz 2 ⎥⎦ ⎢ dz dz
⎣ dz 2 2 dz dz 4 dz 2 ⎥⎦ 2 ⎝ ⎝ dz ⎠ ⎝ dz ⎠ ⎠

1 ⎛ dφ ⎞
(( y − y ) ) du dφ
2
2
+ ⎜ ⎟ + x 2 + y o (4.5)
2 ⎝ dz ⎠ o
dz dz

## The first variation of ε p would then become

dδw d 2δ u d 2δ v d 2δ φ ⎡ d 2v d 2δ v 1 dφ du 1 dδ φ du
δ εp = − x 2 − y
 2 − ω 2 − x⎢
 2 δφ + φ 2 − δφ − φ
dz dz dz dz ⎣ dz dz 2 dz dz 2 dz dz

1 dφ dδ u 1 d 2 u 1 d 2δ u ⎤ ⎡ d 2u d 2δ u 1 dφ dv
− φ − φδ φ 2 − φ 2 2 ⎥ + 
y ⎢ 2 δ φ + 2 φ + δφ
2 dz dz 2 dz 4 dz ⎦ ⎣ dz dz 2 dz dz

1 dδ φ dv 1 dφ dδ v 1 d 2 v 1 d 2δv ⎤ dδ u dφ du dδ φ
+ φ + φ + φδ φ 2 + φ 2 2 ⎥ + yo
 + y o
2 dz dz 2 dz dz 2 dz 4 dz ⎦ dz dz dz dz

(
+ x + ( y − y o )
2 2
) dδφ dφ ω ⎡ d 3δ u dv d 3 u dδ v d 3 v dδ u d 3δ v du ⎤
− ⎢ +
dz dz 2 ⎣ dz 3 dz dz 3 dz

dz dz

dz 3 dz ⎦

dδ u du dδ v dv
+ + (4.6)
dz dz dz dz

## The second variation of ε p is

56
⎛ dδ u ⎞ ⎛ dδ v ⎞ d 2δ v dδ φ du dφ dδ u dδφ dδ u
2 2

δ εp = ⎜
2
⎟ +⎜ ⎟ − x ⎢ 2δ φ 2 − δφ − δφ −φ
⎝ dz ⎠ ⎝ dz ⎠ ⎣ dz dz dz dz dz dz dz

1 2
2 d u d 2δ u ⎤ ⎡ d 2δ u dδ φ dv dφ dδ v
− (δφ ) 2 − φδ φ 2 ⎥ + y ⎢ 2δ φ
 2 + δφ + δφ
2 dz dz ⎦ ⎣ dz dz dz dz dz

dδ φ dδ v 1 d 2δ v dδ u dδ φ
( )
2 ⎛ dδ φ ⎞
2 2
2 d v
+ + (δ φ ) 2 + φδ φ 2 + 2 yo
 + x 2 + ( y − y o ) ⎜ ⎟
dz dz 2 dz dz dz dz ⎝ dz ⎠

⎡ d 3δ u dδ v d 3δ v dδ u ⎤
− ω⎢ 3 − ⎥ (4.7)
⎣ dz dz dz 3 dz ⎦

The deformations are assumed to occur in two stages; a prebuckling state {0, v , w,0}

followed by a lateral buckling state {δ u, 0, 0, δ φ} . This allows for the following simplication to

be applied to the expression for longitudinal strain and it variations (Pi et al., 1992)

u, φ , δ v , δ w = 0

## The simplified expressions are as follows

2
dw d 2 v 1 ⎛ dv ⎞
εp = − y + ⎜ ⎟ (4.8)
dz dz 2 ⎝ dz ⎠

⎡ d 2δ u d 2 v ⎤ ⎡ d 2δ φ 1 ⎛ d 3δ u dv d 3 v dδ u ⎞ ⎤
δ ε p = − x ⎢ 2 + 2 δ φ ⎥ − ω ⎢ 2 + ⎜ − ⎟⎥ (4.9)
⎣ dz dz ⎦ ⎣ dz 2 ⎝ dz 3 dz dz 3 dz ⎠ ⎦

⎛ dδ u ⎞ d 2δ u 1 dδ φ dv ⎤ dδ u dδ φ
2
⎡ 2
2 d v
δ εp = ⎜
2
⎝ dz ⎠
⎟ + y ⎢ 2δ φ
dz 2 +
2
(δ φ ) dz 2 + δφ
dz dz ⎦
⎥ + 2 y o
dz dz

( )
⎛ dδ φ ⎞
2

+ x + ( y − y o )
2
2
⎜ ⎟ (4.10)
⎝ dz ⎠

57
4.1.3 Shear Strain

The effects of shear strain due to bending and warping are neglected. The shear strain due to

uniform torsion is

γ p = −2t p k z (4.11)

Substituting the unsimplified expression for torsional curvature, γ p becomes (Vlasov, 1961)

⎛ dφ 1 ⎛ d 2 u dv d 2 v du ⎞ ⎞
γ p = −2t p ⎜ + ⎜ 2 − 2 ⎟⎟ (4.12)
⎝ dz 2 ⎝ dz dz dz dz ⎠ ⎠

## The first variation of γ p

⎛ dδ φ1 ⎛ d 2δ u dv d 2 v dδ u ⎞ ⎞
δ γ p = −2 t p ⎜ + ⎜ − ⎟⎟ (4.13)
⎝ dz 2 ⎝ dz 2 dz dz 2 dz ⎠ ⎠

## The second variation of γ p is assumed to be zero.

δ 2γ p = 0 (4.14)

## 4.2 STRAIN ENERGY EQUATION CONSIDERING PREBUCKLING

DEFLECTIONS

Substituting the revised expressions for longitudinal and shear stress and their variations into the

strain energy equation given in Chapter 3 with the linearized stress resultants

dw
F = EA (4.15)
dz

58
d 2v
M x = − EI 2 (4.16)
dz

yield the strain energy equation for monosymmetric beam-columns considering prebuckling

## deflections as shown below.

⎡ d 2 (δ u) d 2 v ⎤ ⎡ d (δφ ) 1 ⎛ dv d 2 (δ u) d 2 v d (δ u) ⎞ ⎤
2
⎧ 2
1 2 1 ⎪
δ U = ∫ ⎨ EI y ⎢ 2 + 2 δφ ⎥ + GJ ⎢ + ⎜⎜ 2 − 2 ⎟⎥
⎟⎥
2 2L ⎪ ⎢⎣ dz dz ⎥⎦ ⎢ dz 2 ⎝ dz dz dz dz ⎠⎦
⎩ ⎣

⎡ d 2δφ 1 ⎛ dv d 3 (δ u) d 3 v d (δ u) ⎞ ⎤ ⎡ ⎛ d (δ u) ⎞ 2
2

+ EI w ⎢ 2 +
⎜ − 3 ⎟ ⎥ + F ⎢⎜ ⎟
⎢⎣ dz 2 ⎜⎝ dz dz 3 dz dz ⎟⎠ ⎥ ⎢ ⎜⎝ dz ⎟⎠
⎦ ⎣

## ⎛ d (δ u) d (δφ ) dv dδφ ⎞ dδφ ⎞ ⎤

2
2 ⎛

+ 2 yo ⎜
 + δφ⎟ + (ro + y o )⎜
⎟ 2
 ⎟ ⎥
⎝ dz dz dz dz ⎠ ⎝ dz ⎠ ⎥

⎡ ⎛ d 2 (δ u) ⎞ ⎛ d (δφ ) ⎞
2

⎛ d 2 v ⎞ 2 ⎤⎥ ⎪
⎢ ⎜
+ M x 2⎜ ⎟ ⎜
δφ + βx ⎜ ⎟ + ⎜ 2 ⎟ δφ ⎬ dz (4.17)
⎢ ⎝ dz 2 ⎟⎠ ⎝ dz ⎟⎠ ⎝ dz ⎠ ⎥⎪
⎣ ⎦⎭

DEFLECTIONS

## 4.3.1 Displacements and Rotations of Load Points

The displacements and rotations of load points must be re-derived considering prebuckling

deflections. The second variations of the displacements due to the distributed and concentrated

## loads, v q and v P respectively, considering prebuckling deflections are given as

59
⎛ dδ u dv ⎞
δ 2 v P = − (e − y o )⎜ (δφ ) 2
− δφ⎟ (4.18)
⎝ dz dz ⎠

⎛ dδ u dv ⎞
δ 2 v q = − (a − y o )⎜ (δφ ) 2
− δφ ⎟ (4.19)
⎝ dz dz ⎠

Substituting the new expressions for the second variations of the displacements of the loads into

the potential energy equation given in Chapter 3 yields the new potential energy of the loads

equation as

## ∑ P(e − y )⎛⎜⎝ (δφ)

⎛ dδ u dv ⎞ dδ u dv ⎞

1 2 1 1
q(a − y o )⎜ (δφ ) −
2 2
δ Ω = δφ⎟ dz + − δφ ⎟
2 2 ⎝ dz dz ⎠ 2 o
dz dz ⎠
L

(4.20)

## 4.4 ENERGY EQUATION CONSIDERING PREBUCKLING DEFORMATIONS

The energy equation for lateral torsional buckling of monosymmetric beam-columns considering

prebuckling deformations is determined by substituting Eq. (4.17) and Eq. (4.20) into Eq. (3.6)

as

60
⎡ d 2 (δ u) d 2 v ⎤ ⎡ d (δφ ) 1 ⎛ dv d 2 (δ u) d 2 v d (δ u) ⎞ ⎤
2
⎧ 2

1 2 1
δ Π = ⎨ EI y ⎢ 2 + 2 δφ ⎥ + GJ ⎢ + ⎜⎜ 2 − 2 ⎟⎥
⎟⎥
2 2 ⎪⎩ ⎢
⎣ dz dz ⎥
⎦ ⎢ dz 2 ⎝ dz dz dz dz ⎠⎦
L ⎣

2
⎡ d 2δφ 1 ⎛ dv d 3 (δ u) d 3 v d (δ u) ⎞ ⎤
+ EI w ⎢ 2 + ⎜ 3
− 3 ⎟⎥
⎢⎣ dz 2 ⎝ dz dz dz dz ⎠ ⎥⎦

⎡ ⎛ d (δ u) ⎞ 2 ⎛ d (δ u) d (δφ ) dv dδφ ⎞ 2⎤
dδφ ⎞ ⎥
2 ⎛

+F ⎜⎜ ⎟ ⎜
+ 2 yo ⎜
 + δφ ⎟ + (ro + y o )⎜
⎟ 2
 ⎟
⎢ ⎝ dz ⎟⎠ ⎝ dz dz dz dz ⎠ ⎝ dz ⎠ ⎥
⎣ ⎦

⎡ ⎛ d 2 (δ u) ⎞ ⎛ d (δφ ) ⎞
2

⎛ d 2 v ⎞ 2 ⎤⎥ ⎪
⎢ ⎜
+ M x 2⎜ ⎟ ⎜
δφ + βx ⎜ ⎟ + ⎜ 2 ⎟ δφ ⎬ dz
⎢ ⎝ dz 2 ⎟⎠ ⎝ dz ⎟⎠ ⎝ dz ⎠ ⎥⎪
⎣ ⎦⎭

⎛ dδ u dv ⎞

1
q(a − y o )⎜ (δφ ) −
2
+ δφ ⎟ dz
2 ⎝ dz dz ⎠
L

## ∑ P(e − y )⎛⎜⎝ (δφ)

1 2 dδ u dv ⎞
+ − δφ⎟ (4.21)
2 o
dz dz ⎠

The first two lines on the right side of the equation contain terms involving buckling rigidities

EIy, GJ, and EI ω and represent the strain energy stored during prebuckling and buckling. The

third and fourth line of the equation contain terms involving the stress resultants F and Mx, which

represent the work done by the applied loads at the shear center, considering the effects of

prebuckling. The final two lines of the equation represent the work done by the distributed load,

## q, and the concentrated load, P, considering prebuckling deflections.

In order to linearize the new energy equation, the second order in-plane displacements are

neglected in order to avoid a quadratic eigenvalue equation (Roberts, 2004). The new energy

## equation, disregarding these second order displacements, is

61
⎛ d 2 (δ u) ⎞ d 2 (δ u) d 2 v ⎛ d (δφ ) ⎞
2 2
1 2
2
1
δ Π = ∫ EI y ⎜⎜
2L
[
⎝ dz ⎠
2

⎟ + 2 EI y
dz 2 dz 2
δφ + GJ ⎜
⎜ dz ⎟

d (δφ ) ⎛ dv d 2 (δ u) d 2 v d (δ u) ⎞ ⎛ d 2 (δφ ) ⎞
2

+ GJ ⎜ − 2 ⎟ + EI ω ⎜⎜ ⎟
dz ⎜⎝ dz dz 2 dz dz ⎟⎠ ⎝ dz 2 ⎟

d 2 (δφ ) ⎛ dv d 3 (δ u) d 3 v d (δ u) ⎞
+ EI ω ⎜ − 3 ⎟
dz 2 ⎜⎝ dz dz 3 dz dz ⎟⎠

⎡ ⎛ d (δ u) ⎞ 2 ⎛ d (δ u) d (δφ ) ⎞ 2⎤
dδφ ⎞ ⎥
2 ⎛

+F ⎜⎜ ⎟ ⎜
+ 2 y o ⎜ ⎟ + (ro + y o )⎜
2

⎢ ⎝ dz ⎟⎠ ⎝ dz dz ⎟⎠ ⎝ dz ⎠ ⎥
⎣ ⎦

⎡ ⎛ d 2 (δ u) ⎞ ⎛ d (δφ ) ⎞ ⎤ ⎫⎪
2

+ M x ⎢ 2⎜⎜
⎢ ⎝ dz ⎠

2
⎟ δφ + βx ⎜
⎟ ⎜ ⎟
⎝ dz ⎠ ⎥⎦ ⎪

2 ∫
⎟ ⎥ ⎬ dz + 1 q(a − y o )(δφ ) 2 dz

+
1
2 ∑ P(e − y o )(δφ ) = 0
2
(4.22)

Comparing this equation to Eq. (3.88) in the previous chapter, the only terms that differ are

1 ⎡ d 2 (δ u) d 2 v d (δφ ) ⎛ dv d 2 (δ u) d 2 v d (δ u) ⎞
2 ∫L ⎢⎣
⎢ 2 EI y δφ + GJ ⎜ − 2 ⎟
dz 2 dz 2 dz ⎜⎝ dz dz 2 dz dz ⎟⎠

d 2 (δφ ) ⎛ dv d 3 (δ u) d 3 v d (δ u) ⎞ ⎤
+ EI ω ⎜ − 3 ⎟ ⎥ dz (4.23)
dz 2 ⎜⎝ dz dz 3 dz dz ⎟⎠ ⎥

## The in-plane curvature is

d 2v Mx
2 = − (4.24)
dz EI x

Likewise,

dv Mx
dz
= ∫− EI x
dz + C (4.25)

62
d 3v 1 dM x Vy
= − = − (4.26)
dz 3 EI x dz EI x

dv dv (0)
The constant C is a function of and is calculated at z = 0 as C = .
dz dz

## Substituting these equations into Eq. (4.23) yields

Iy d 2 (δ u) ⎛ Mx ⎞ d (δφ ) d 2 (δ u)
∫− 2I ∫
1
Mx δφ + GJ ⎜ − dz + C⎟
2 x dz 2 ⎝ EI x ⎠ dz dz 2
L

M d (δφ ) d (δ u) ⎛ ⎞ d 2 (δφ ) d 3 (δ u)

Mx
+ GJ x + EI ω ⎜ − dz + C⎟
EI x dz dz ⎝ EI x ⎠ dz 2 dz 3

Iω d 2 (δφ ) d (δ u) ⎤
V ⎥ dz (4.27)
I x y dz 2 dz ⎥⎦

63
5.0 APPLICATIONS

This chapter will present examples using the buckling equation derived in Section 3.4 (Eq.

(3.88)) to determine approximate solutions for specific loading and boundary conditions. These

solutions will be obtained by assuming suitable shape functions for the displacement u and

rotation φ during buckling. The solutions will then be compared to examples presented in other

literature that used different methods, including differential equilibrium equation method for the

first example and the method of finite differences for the remaining examples, to obtain buckling

results to demonstrate the validity of the buckling equations presented in this paper.

## 5.1 SIMPLY-SUPPORTED MONOSYMMETRIC BEAM SUBJECTED TO EQUAL

END MOMENTS, M

This example will derive the lateral-torsional buckling moment, Mcr, for a simply-supported

beam whose ends are restrained from twist subjected to equal end moments, as shown in Figure

5.1.

64
M M

L
y

## The lateral torsional buckling equation derived in Chapter 3 is

⎛ d 2 (δ u) ⎞ ⎛ d (δφ ) ⎞ ⎛ d 2 (δφ ) ⎞
2
⎧ 2 2

1 2 1 ⎪
δ Π = ⎨ EI y ⎜⎜ 2
⎟ + GJ ⎜

⎟ ⎜
⎜ dz ⎟ + EI ω ⎜ dz 2 ⎟

2 2 ⎪⎩ ⎝ dz ⎠ ⎝ ⎠ ⎝ ⎠
L

⎡ ⎛ d (δ u) ⎞ 2 ⎛ d (δ u) ⎞ ⎛ d (δφ ) ⎞ ⎛ d (δφ ) ⎞ ⎤
2

+ F ⎢ ⎜⎜ ⎜ dz ⎟ ⎜ dz ⎟ + (ro + y o )⎜ dz ⎟ ⎥
⎟ + 2 y o ⎜ ⎟⎜ ⎟ 2 2 ⎜ ⎟ ⎥
⎢ ⎝ dz ⎟⎠ ⎝ ⎠⎝ ⎠ ⎝ ⎠ ⎦

⎡ ⎛ d 2 (δ u) ⎞ ⎛ d (δφ ) ⎞ ⎤ ⎫⎪
2

+ M x ( z ) ⎢ 2⎜⎜ ⎟ δφ + βx ⎜ ⎟ ⎥
⎜ dz ⎟ ⎥ ⎬ dz
⎢ ⎝ dz 2 ⎟⎠ ⎝ ⎠ ⎦⎪
⎣ ⎭

+
1
2 ∫
q(a − y o )(δφ ) dz +
L
2 1
2 ∑ P(e − y )(δφ)
o
2
= 0     (5.1)

## and P = 0 , Eq. (5.1) reduces to

65
⎛ d 2 (δ u) ⎞ ⎛ d (δφ ) ⎞ ⎛ d 2 (δφ ) ⎞
2
⎧ 2 2

1 2 1 ⎪ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟
δ Π = ⎨ EI y ⎜ 2 ⎟ + GJ ⎜ dz ⎟ + EI ω ⎜ dz 2 ⎟
2 2 ⎪⎩ ⎝ dz ⎠ ⎝ ⎠ ⎝ ⎠
L

⎡ ⎛ d 2 (δ u) ⎞ ⎛ d (δφ ) ⎞ ⎤ ⎫⎪
2

+ M x ( z ) ⎢ 2⎜⎜ ⎟ δφ + βx ⎜ ⎟ ⎥
⎜ dz ⎟ ⎥ ⎬ dz (5.2)
⎢ ⎝ dz 2 ⎟⎠ ⎝ ⎠ ⎦⎪
⎣ ⎭

For a simply-supported beam, the following shape functions can be used to solve this problem.

⎛ π z⎞
u( z ) = A sin⎜ ⎟                (5.3)
⎝ L⎠

⎛ π z⎞
φ ( z) = B sin⎜ ⎟                (5.4)
⎝ L⎠

The first variation of u and φ , as well as their first and second derivatives are needed to solve

## this problem and are shown below as

⎛ π z⎞
δ u = δ A sin⎜ ⎟                (5.5)
⎝ L⎠

⎛ π z⎞
δ φ = δ B sin⎜ ⎟                (5.6)
⎝ L⎠

d (δ u) ⎛ π⎞ ⎛ π z⎞
= ⎜ ⎟ δ A cos⎜ ⎟              (5.7)
dz ⎝ L⎠ ⎝ L⎠

d (δ φ ) ⎛ π⎞ ⎛ π z⎞
= ⎜ ⎟ δ B cos⎜ ⎟              (5.8)
dz ⎝ L⎠ ⎝ L⎠

66
d 2 (δ u) ⎛ π⎞
2
⎛ π z⎞
= − ⎜ ⎟ δ A sin⎜ ⎟            (5.9)
d z2 ⎝ L⎠ ⎝ L⎠

d 2 (δ φ ) ⎛ π⎞
2
⎛ π z⎞
= − ⎜ ⎟ δ B sin⎜ ⎟            (5.10)
d z2 ⎝ L⎠ ⎝ L⎠

## Substituting Eqs. (5.3) – (5.10) into Eq. (5.1) yields

L
⎧⎛ π ⎞ 4 2⎛
π z⎞ ⎛ π ⎞ 2 ⎛ π z⎞

1 2 1
2
δ Π =
2

⎨⎝ ⎠ ⎟ EI y (δ A) 2
sin ⎜ ⎟ + ⎜ ⎟ GJ (δ B) cos2 ⎜
⎝ L ⎠ ⎝ L⎠
2

⎝ L⎠
⎟
0 ⎩ L

⎛ π⎞ ⎛ π z⎞
4

+ ⎜ ⎟ EI ω (δ B) sin 2 ⎜
2
⎟
⎝ L⎠ ⎝ L⎠

⎡ ⎛ π⎞ 2⎛ ⎛ π z⎞ ⎞ ⎛ ⎛ π z⎞ ⎞ ⎛ π⎞ ⎛ π z ⎞ ⎤ ⎫⎪
2

## + M ⎢ − 2⎜ ⎟ ⎜ δ A sin⎜ ⎟ ⎟ ⎜ δ B sin⎜ ⎟ ⎟ + βx ⎜ ⎟ (δ B) cos2 ⎜

2
⎟ ⎬ dz
⎝ L ⎠ ⎥⎦ ⎪⎭

⎣ ⎝ L⎠ ⎝ ⎝ L ⎠⎠⎝ ⎝ L ⎠⎠ ⎝ L⎠

(5.11)
The following integrations will be necessary to evaluate Eq. (5.11)
L
⎛ π z⎞

L
sin 2 ⎜ ⎟ dz =                 (5.12)
0 ⎝ L⎠ 2

L
⎛ π z⎞

L
cos2 ⎜ ⎟ dz =                 (5.13)
0 ⎝ L⎠ 2

## Integrating Eq. (5.11) with respect to z gives

L⎞ ⎛ π ⎞ ⎛ L⎞ ⎛ π ⎞ ⎛ L⎞ ⎛ π ⎞
4 2 4
1 2 1 ⎧⎛
⎟ ⎜ ⎟ EI y (δ A) + ⎜ ⎟ ⎜ ⎟ GJ (δ B) + ⎜ ⎟ ⎜ ⎟ EI ω (δ B)
2 2 2
δ Π = ⎨⎜
2 2 ⎩⎝ ⎠
2 L⎝ ⎠ ⎝ ⎠
2 L⎝ ⎠ ⎝ ⎠
2 L⎝ ⎠

⎡ ⎛ L⎞ ⎛ π ⎞ 2 ⎛ L⎞ ⎛ π ⎞ 2 ⎤⎫
2

+ M ⎢ − 2⎜ ⎟ ⎜ ⎟ (δ A)(δ B) + βx ⎜ ⎟ ⎜ ⎟ (δ B) ⎥ ⎬ dz     (5.14)
⎣ ⎝ 2 ⎠ ⎝ L⎠ ⎝ 2 ⎠ ⎝ L⎠ ⎦ ⎪⎭

67
The critical buckling load occurs when the second variation of total potential energy is equal to

## zero. Eq. (5.14) can then be written as

⎡⎛ π ⎞ 2 ⎤
⎜ ⎟ EI y − M
1 L ⎛ π ⎞ ⎧ δ A⎫ ⎢⎢ ⎝ L ⎠ ⎥ ⎧ δ A⎫
2 T
1 2 ⎥⎨ ⎬ = 0
δ Π = ⎜ ⎟ ⎨ ⎬
2 2 ⎝ L ⎠ ⎩ δ B⎭ ⎢ ⎛ π⎞ ⎥ ⎩δ B⎭
2
2
⎢ − M GJ + ⎜ ⎟ EI ω + Mβx ⎥
⎣ ⎝ L⎠ ⎦

(5.15)
Since δ A and δ B are not equal to zero

⎡⎛ π ⎞ 2 ⎤
⎢ ⎜⎝ ⎟⎠ EI y − M ⎥ ⎧ δ A⎫
⎢ L ⎥⎨ ⎬ = 0        (5.16)
⎛ π⎞ ⎥ ⎩δ B⎭
2

⎢ − M GJ + ⎜ ⎟ EI ω + Mβx ⎥
⎣ ⎝ L⎠ ⎦

If the deformed configuration of the beam is to yield a nontrivial solution, the determinant of the

## coefficients δ A  and δ B in Eq. (5.16) must vanish leaving

⎛ π⎞
2
⎛ π⎞
2
⎛ ⎛ π⎞
2

M − ⎜ ⎟ EI y βx M − ⎜ ⎟ EI y ⎜ GJ + ⎜ ⎟ EI ω ⎟ = 0
2
(5.17)
⎝ L⎠ ⎝ L⎠ ⎝ ⎝ L⎠ ⎠

Solving Eq. (5.17) for M gives the critical moment for lateral torsional buckling of a

## monosymmetric I-beam subjected to equal end moments as

( EI ) β
2
2
1⎛ π⎞ ⎛ π⎞ ⎛ π⎞ ⎛ π⎞
2 4 2 4
y x
M cr = ⎜ ⎟ EI y βx + ⎜ ⎟ + ⎜ ⎟ EI y GJ + ⎜ ⎟ EI y EI ω (5.18)
2 ⎝ L⎠ ⎝ L⎠ 4 ⎝ L⎠ ⎝ L⎠

The beam stiffness parameter and the monosymmetric parameter were given in Section 3.5 under

non-dimensional analysis as

68
π 2 EI ω
K=                  (5.19)
GJL2

βx EI y
βx =                 (5.20)
L GJ

Rewriting Eq. (5.18) in terms of K and βx yields the elastic critical moment as

⎡ ⎛ πβ ⎞
2 ⎤
π ⎢ πβ ⎥
M cr = EI y GJ 1 + K +
2
⎜ ⎟ +         (5.21)
L ⎢ ⎝ 2 ⎠ 2 ⎥
⎣ ⎦

This result for the lateral torsional buckling moment of monosymmetric I-beams under a

uniform moment matches the solution given by (Kitipornchai, et al. 1986) for F = 0 when the

elastic critical moment was derived using the differential equilibrium method illustrated in

## Chapter 2 Eq. (2.53).

69
5.2 SIMPLY-SUPPORTED MONOSYMMETRIC BEAM SUBJECTED TO

This example will present the lateral-torsional buckling load, Pcr, for a simply-supported beam

whose ends are restrained from twist subjected to a concentrated central load that is applied to

z

L
y

## Figure 5.2 Monosymmetric Beam Subjected to Concentrated Central Load

There is no axial force or uniformly distributed load acting on the beam. Therefore,   F = 0 and

## Therefore, Eq. (5.1) reduces to

70
⎛ d 2 (δ u) ⎞ ⎛ d (δφ ) ⎞ ⎛ d 2 (δφ ) ⎞
2
⎧ 2 2

1 2 1 ⎪ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟
δ Π = ⎨ EI y ⎜ 2 ⎟ + GJ ⎜ dz ⎟ + EI ω ⎜ dz 2 ⎟
2 2 ⎪⎩ ⎝ dz ⎠ ⎝ ⎠ ⎝ ⎠
L

⎡ ⎛ d 2 (δ u) ⎞ ⎛ d (δφ ) ⎞ ⎤ ⎫⎪
2

+ M x ( z ) ⎢ 2⎜⎜ ⎟ δφ + βx ⎜ ⎟ ⎥
⎜ dz ⎟ ⎥ ⎬ dz (5.22)
⎢ ⎝ dz 2 ⎟⎠ ⎝ ⎠ ⎦⎪
⎣ ⎭

Since this beam is also simply supported, the same buckling modes shapes will be used as in

Section 5.1. The moment along the beam, M x ( z ) , for a beam with a concentrated central load

can be expressed as

P L
M x ( z) = z    for 0≤ z≤           (5.23)
2 2

P L
M x ( z) = ( L − z)   for ≤ z ≤ L          (5.24)
2 2

Substituting Eqs. (5.3) – (5.10) and Eqs. (5.23) – (5.24) into Eq. (5.22) yields

L
⎧⎛ π ⎞ 4 2⎛
π z⎞ ⎛ π ⎞ 2 ⎛ π z⎞

1 2 1
⎨ ⎜⎝ ⎟⎠ EI y (δ A) sin ⎜ ⎟ + ⎜ ⎟ GJ (δ B) cos2 ⎜
2 2
δ Π = ⎟
2 2 0 ⎩ L ⎝ L ⎠ ⎝ L⎠ ⎝ L⎠

⎛ π⎞ ⎛ π z⎞ ⎫
4

+ ⎜ ⎟ EI ω (δ B) sin 2 ⎜
2
⎟ ⎬ dz
⎝ L⎠ ⎝ L ⎠⎭

L
⎧ − 2P ⎛ π ⎞ 2 ⎛ ⎛ π z⎞ ⎞ ⎛ ⎛ π z⎞ ⎞

1 2
+ ⎨ z⎜ ⎟ ⎜ δ A sin⎜ ⎟ ⎟ ⎜ δ B sin⎜ ⎟⎟
2 0 ⎩ P ⎝ ⎠
L ⎝ ⎝ L ⎠⎠⎝ ⎝ L ⎠⎠

⎛ π⎞ ⎛ π z⎞ ⎫
2
P
+ zβx ⎜ ⎟ (δ B) cos2 ⎜
2
⎟ ⎬ dz
2 ⎝ L⎠ ⎝ L ⎠⎭

L
⎧ − 2P ⎛ π⎞ ⎛
2
⎛ π z⎞ ⎞ ⎛ ⎛ π z⎞ ⎞

1
+ ⎨ ( L − z ) ⎜ ⎟ ⎜ δ A sin ⎜ ⎟ ⎟ ⎜ δ B sin ⎜ ⎟⎟
2 L
⎩ P ⎝ L⎠ ⎝ ⎝ L ⎠⎠⎝ ⎝ L ⎠⎠
2

71
π ⎛ π z⎞ ⎫
2

( L − z) βx ⎛⎜ ⎞⎟ (δ B) cos2 ⎜ ⎟ ⎬ dz
P 2
+       (5.25)
2 ⎝ L⎠ ⎝ L ⎠⎭

The following integrations will be necessary to evaluate Eq. (5.25)

L
⎛ π z⎞ L2 L2

2
z sin 2 ⎜ ⎟ dz = +             (5.26)
0 ⎝ L⎠ 16 4π 2

L
⎛ π z⎞ L2 L2

2 z
cos2 ⎜ ⎟ dz = −             (5.27)
0 2 ⎝ L⎠ 32 8π 2

L
⎛ π z⎞ L2 L2
∫ L
2
( L − z) sin ⎜ ⎟ dz =
⎝ L⎠
2
+
16 4π 2
(5.28)

L
( L − z) ⎛ π z⎞ L2 L2
∫ L
2
2
cos ⎜ ⎟ dz =
2
⎝ L⎠

32 8π 2
(5.29)

## Therefore integrating Eq. (5.25) with respect to z gives

1 ⎧⎛ L⎞ ⎛ π ⎞ ⎛ L⎞ ⎛ π ⎞ ⎛ L⎞ ⎛ π ⎞
2 2 2
1 2
⎟ ⎜ ⎟ EI y (δ A) + ⎜ ⎟ ⎜ ⎟ GJ (δ B) + ⎜ ⎟ ⎜ ⎟ EI ω (δ B)
2 2 2
δ Π = ⎨⎜
2 2 ⎩⎝ 2 ⎠ ⎝ L⎠ ⎝ 2 ⎠ ⎝ L⎠ ⎝ 2 ⎠ ⎝ L⎠

⎡ ⎛ π ⎞ 2 ⎛ L2 L2 ⎞ ⎛ π⎞
2
⎛ L2 L2 ⎞ 2⎤
+ ⎢ − ⎜ ⎟ P⎜ + ⎟ (δ A)(δ B ) + ⎜ ⎟ Pβ ⎜ − 2 ⎟ (δ B ) ⎥
⎣ ⎝ L ⎠ ⎝ 16 4π ⎠ ⎝ L⎠ ⎝ 32 8π ⎠
2 x

⎡ ⎛ π ⎞ 2 ⎛ L2 ⎛ π⎞ 2 ⎤⎫
2
L2 ⎞ ⎛ L2 L2 ⎞ ⎪
+ ⎢ − ⎜ ⎟ P⎜ + ⎟ (δ A)(δ B ) + ⎜ ⎟ Pβ ⎜ − ⎟ (δ B ) ⎥ ⎬
⎣ ⎝ L ⎠ ⎝ 16 4π ⎠ ⎝ L⎠ ⎝ 32 8π ⎠
2 x 2
⎦ ⎪⎭

(5.30)
Which simplifies to

72
1 L ⎛ π ⎞ ⎡⎛ π ⎞ ⎛ π⎞
2 2 2
1 2
y( ) ( ) ⎜ ⎟ EI ω (δ B)
2 2 2
δ Π = ⎜ ⎟ ⎢⎜ ⎟ EI δ A + GJ δ B +
2 2 2 ⎝ L⎠ ⎣⎝ L⎠ ⎝ L⎠

⎛ π 2 + 4⎞ ⎛ π 2 − 4⎞ 2⎤
− PL⎜
⎝ 4π 2 ⎠
⎟ (δ A)(δ B ) + PLβ x ⎜
⎝ 8π 2 ⎠
⎟ (δ B ) ⎥      (5.31)

The critical buckling load occurs when the second variation of total potential energy is equal to

zero, as

⎡ ⎛ π⎞
2
⎛ π 2 + 4⎞ ⎤
T⎢ ⎜ ⎟ EI y − PL⎜ ⎟ ⎥ δA
1 L ⎛ π ⎞ ⎧ δ A⎫ ⎢ ⎝ L⎠ ⎝ 8π 2 ⎠
2
1 2
δ Π = ⎜ ⎟ ⎨ ⎬ ⎥ ⎧⎨ ⎫⎬ = 0
2 2 ⎝ L ⎠ ⎩ δ B⎭ ⎢ ⎛ π 2 + 4⎞ ⎛ π⎞ ⎛ π 2 − 4 ⎞ ⎥ ⎩ δ B⎭
2
2
⎢− PL⎜ ⎟ GJ + ⎜ ⎟ EI ω + PLβx ⎜ ⎟⎥
⎣ ⎝ 8π 2 ⎠ ⎝ L⎠ ⎝ 8π 2 ⎠ ⎦

(5.32)

## Since δ A and δ B are not equal to zero

⎡ ⎛ π⎞
2
⎛ π 2 + 4⎞ ⎤
⎢ ⎜ ⎟ EI y − PL⎜ ⎟ ⎥ δA
⎢ ⎝ L⎠ ⎝ 8π 2 ⎠ ⎥ ⎧⎨ ⎫⎬ = 0     (5.33)
⎢ ⎛ π 2 + 4⎞ ⎛ π⎞
2
⎛ π − 4 ⎞ ⎥ ⎩ δ B⎭
2

⎢− PL⎜ ⎟ GJ + ⎜ ⎟ EI ω + PLβx ⎜ ⎟⎥
⎣ ⎝ 8π 2 ⎠ ⎝ L⎠ ⎝ 8π 2 ⎠ ⎦

If the deformed configuration of the beam is to yield a nontrivial solution, the determinant of the

## coefficients δ A  and δ B  in Eq. (5.33) must vanish leaving

2
⎛ π 2 + 4⎞ 2 2 ⎛ π ⎞ 2 ⎛ π 2 − 4⎞ ⎛ π⎞
2
⎛ ⎛ π⎞
2

⎜ ⎟ L P − ⎜ ⎟ EI y βx ⎜ ⎟ LP − ⎜ ⎟ EI y ⎜ GJ + ⎜ ⎟ EI ω ⎟ = 0
⎝ 8π 2 ⎠ ⎝ L⎠ ⎝ 8π 2 ⎠ ⎝ L⎠ ⎝ ⎝ L⎠ ⎠

(5.34)
Solving Eq. (5.34) for P gives the critical buckling load for lateral torsional buckling of a

## monosymmetric I-beam subjected to a concentrated central load as

73
1
⎧ ⎡
( )β ( ) ( )
2 2 2
Pcr = ⎨ 4 16 EI y
2
π 8 − 8 EI y βx 2 π 10 + EI y βx 2 π 12
(π + 4) L ⎢
2 3 ⎩ ⎣ 2 x

( ) ( ) ( )
+ 4 EI y GJL2 π 10 + 32 EI y GJL2 π 8 + 64 EI y GJL2 π 6 + 4 EI y EI ω π 12   ( )( )
⎫⎪
] ( )
1

( )( )
+ 32 EI y EI ω π 10
( )
+ 64 EI y ( EI ω )π 8 2
( )
+ 4 EI y βx π 6 − 16 EI y βx π 4 ⎬  (5.35)
⎪⎭

## The non-dimensional loading parameter, γ P , was given in section 3.5 as

PL2
γP =                   (5.36)
EI y GJ

Rewriting Eq. (5.35) in terms of the non-dimensional beam stiffness parameter, K , given

in Eq. (5.19) and the monosymmetric parameter,   βx , given in Eq. (5.20), as well as the loading

parameter, γ P , shown in Eq. (5.36) yields the elastic critical load parameter for a simply-

## supported monosymmetric beam with a concentrated point load as

γP =
(π 2
1
+ 4)
2 {4[(π 12
− 8π 10 + 16π 8 )βx 2 + (4π 10 + 32π 8 + 64π 6 )

1

+ (4π 10
+ 32π + 64π ) K
8 6
] + (4π
2 2 6
− 16π 4 )βx ⎬

(5.37)

This result can then be compared to the results given in (Anderson and Trahair, 1972) where the

critical load parameters were obtained using the method of finite differences. Figures 5.3 – 5.9

compare the solution presented in Eq. (5.37) with the results given by Anderson and Trahair at

various values of K and βx . The results obtained by this research using approximate shape

functions are more accurate as the values for βx become closer to zero, representing a doubly-

74
symmetric beam. Thus, the shape functions used in this example more accurately predict the

## buckled shape of simply-supported beams with doubly-symmetric cross-sections.

Figure 5.3 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section

## Subjected to a Concentrated Central Load ( βx = − 0.6)

75
Figure 5.4 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section

## Subjected to a Concentrated Central Load ( βx = − 0.3)

Figure 5.5 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section

## Subjected to a Concentrated Central Load ( βx = − 01

.)

76
Figure 5.6 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section

## Subjected to a Concentrated Central Load ( βx = 0)

Figure 5.7 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section

## Subjected to a Concentrated Central Load ( βx = 01

.)

77
Figure 5.8 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section

## Subjected to a Concentrated Central Load ( βx = 0.3)

Figure 5.9 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section

## Subjected to a Concentrated Central Load ( βx = 0.6)

78
5.3 SIMPLY-SUPPORTED MONOSYMMETRIC BEAM SUBJECTED TO

This example will present the lateral-torsional buckling load, qcr, for a simply-supported beam

whose ends are restrained from twist subjected to a uniformly distributed load that is applied to

x
z

L
y

## Figure 5.10 Monosymmetric Beam Subjected to Uniformly Distributed Load

There is no axial force or concentrated load acting on the beam. Therefore,   F = 0 and P = 0 .

Since the uniformly distributed load, q, is applied at the beam’s shear center,  (a − y 0 ) = 0 .

## Therefore, Eq. (5.1) reduces to

79
⎛ d 2 (δ u) ⎞ ⎛ d (δφ ) ⎞ ⎛ d 2 (δφ ) ⎞
2
⎧ 2 2

1 2 1 ⎪ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟
δ Π = ⎨ y⎜
EI 2 ⎟ + GJ ⎜ dz ⎟ + EI ω⎜ 2 ⎟
2 2 ⎪⎩ ⎝ dz ⎠ ⎝ ⎠ ⎝ dz ⎠
L

⎡ ⎛ d 2 (δ u) ⎞ ⎛ d (δφ ) ⎞ ⎤ ⎫⎪
2

+ M x ( z ) ⎢ 2⎜⎜ ⎟ δφ + βx ⎜ ⎟ ⎥
⎜ dz ⎟ ⎥ ⎬ dz (5.38)
⎢ ⎝ dz 2 ⎟⎠ ⎝ ⎠ ⎦⎪
⎣ ⎭

Since this beam is also simply supported, the same buckling modes shapes will be used as in

section 5.1. The moment along the beam, M x ( z ) , for a beam with a uniformly distributed load

can be expressed as

qz
Mx = ( L − z)   for 0≤ z ≤ L          (5.39)
2

Substituting Eqs. (5.3) – (5.10) and Eq. (5.39) into Eq. (5.38) yields

L
⎧⎛ π ⎞ 4 2⎛
π z⎞ ⎛ π ⎞ 2 ⎛ π z⎞

1 2 1
⎨ ⎜⎝ ⎟⎠ EI y (δ A) sin ⎜ ⎟ + ⎜ ⎟ GJ (δ B) cos2 ⎜
2 2
δ Π = ⎟
2 2 0 ⎩ L ⎝ L ⎠ ⎝ L⎠ ⎝ L⎠

⎛ π⎞ ⎛ π z⎞ ⎫
4

+ ⎜ ⎟ EI ω (δ B) sin 2 ⎜
2
⎟ ⎬ dz
⎝ L⎠ ⎝ L ⎠⎭

L
⎧ − 2qz ⎛ π⎞ ⎛
2
⎛ π z⎞ ⎞ ⎛ ⎛ π z⎞ ⎞

1
+ ⎨ ( L − z ) ⎜ ⎟ ⎜ δ A sin⎜ ⎟ ⎟ ⎜ δ B sin⎜ ⎟⎟
2 0 ⎩ 2 ⎝ L⎠ ⎝ ⎝ L ⎠⎠⎝ ⎝ L ⎠⎠

π ⎛ π z⎞ ⎫
2

( L − z) βx ⎛⎜ ⎞⎟ (δ B) cos2 ⎜ ⎟ ⎬ dz
qz 2
+       (5.40)
2 ⎝ L⎠ ⎝ L ⎠⎭

## The following integrations will be necessary to evaluate Eq. (5.40).

80
L
⎛ π z⎞ L3 L3

0
( zL − z ) sin ⎜⎝ L ⎟⎠ dz = 12 + 4π 2
2 2
(5.41)

L
(zL − z 2 ) ⎛ π z⎞ L3 L3

0 2
cos ⎜ ⎟ dz =
2
⎝ L⎠

24 8π 2
(5.42)

## Therefore, integrating Eq. (5.40) with respect to z yields

1 ⎧⎛ L⎞ ⎛ π ⎞ ⎛ L⎞ ⎛ π ⎞ ⎛ L⎞ ⎛ π ⎞
4 2 4
1 2
⎟ ⎜ ⎟ EI y (δ A) + ⎜ ⎟ ⎜ ⎟ GJ (δ B) + ⎜ ⎟ ⎜ ⎟ EI ω (δ B)
2 2 2
δ Π = ⎨⎜
2 2 ⎩⎝ 2 ⎠ ⎝ L⎠ ⎝ 2 ⎠ ⎝ L⎠ ⎝ 2 ⎠ ⎝ L⎠

⎡ ⎛ π ⎞ 2 ⎛ L3 L3 ⎞ ⎛ π⎞
2
⎛ L3 L3 ⎞ 2 ⎤⎪ ⎫
+ ⎢ − ⎜ ⎟ q⎜ + ⎟ (δ A)(δ B ) + ⎜ ⎟ qβ ⎜ − 2 ⎟ (δ B ) ⎥ ⎬  (5.43)
⎣ ⎝ L ⎠ ⎝ 12 4π ⎠ ⎝ L⎠ ⎝ 24 8π ⎠
2 x
⎦ ⎪⎭

Which simplifies to

1 L ⎛ π ⎞ ⎡⎛ π ⎞ ⎛ π⎞
2 2 2
1 2
⎜ ⎟ ⎢ ⎜ ⎟ EI y (δ A) + GJ (δ B) + ⎜ ⎟ EI ω (δ B)
2 2 2
δ Π =
2 2 2 ⎝ L⎠ ⎣⎝ L⎠ ⎝ L⎠

⎛ π 2 + 3⎞ ⎛ π 2 − 3⎞ 2⎤
− qL2 ⎜
⎝ 6π ⎠ 2 ⎟ (δ A)(δ B ) + qL2
β x⎜
⎝ 12π ⎠2 ⎟ (δ B ) ⎥       (5.44)

The critical buckling load occurs when the second variation of total potential energy is equal to

zero, as

⎡ ⎛ π⎞
2
⎛ π 2 + 3⎞ ⎤
T⎢ ⎜ ⎟ EI y − qL ⎜
2
⎟ ⎥ δA
1 L ⎛ π ⎞ ⎧ δ A⎫ ⎢ ⎝ L⎠ ⎝ 12π 2 ⎠
2
1 2
δ Π = ⎜ ⎟ ⎨ ⎬ ⎥ ⎧⎨ ⎫⎬ = 0
2 2 ⎝ L ⎠ ⎩ δ B⎭ ⎢ ⎛ π 2 + 3⎞ ⎛ π⎞ ⎛ π − 3⎞ ⎥ ⎩ δ B ⎭
2 2
2
⎢− qL2 ⎜ ⎟ GJ + ⎜ ⎟ EI ω + qL2 βx ⎜ ⎟⎥
⎣ ⎝ 12π 2 ⎠ ⎝ L⎠ ⎝ 12π 2 ⎠ ⎦

(5.45)
Since δ A and δ B are not equal to zero

81
⎡ ⎛ π⎞
2
⎛ π 2 + 3⎞ ⎤
⎢ ⎜ ⎟ EI y − qL2 ⎜ ⎟ ⎥ δA
⎢ ⎝ L⎠ ⎝ 12π 2 ⎠ ⎥ ⎧⎨ ⎫⎬ = 0     (5.46)
⎢ ⎛ π 2 + 3⎞ ⎛ π⎞
2
⎛ π − 3⎞ ⎥ ⎩ δ B ⎭
2

⎢− qL ⎜2
⎟ GJ + ⎜ ⎟ EI ω + qL2 βx ⎜ ⎟⎥
⎣ ⎝ 12π 2 ⎠ ⎝ L⎠ ⎝ 12π 2 ⎠ ⎦

If the deformed configuration of the beam is to yield a nontrivial solution, the determinant of the

## coefficients δ A  and δ B in Eq. (5.46) must vanish leaving

2
⎛ π 2 + 3⎞ 4 2 ⎛ π ⎞ 2 ⎛ π 2 − 3⎞ 2 ⎛ π⎞
2
⎛ ⎛ π⎞
2

⎜ ⎟ L q − ⎜ ⎟ EI β ⎜ ⎟ L q − ⎜ ⎟ EI ⎜ GJ + ⎜ ⎟ EI ⎟ =0
⎝ 12π 2 ⎠ ⎝ L⎠ y x
⎝ 12π 2 ⎠ ⎝ L⎠ y
⎝ ⎝ L⎠ ω

(5.47)
Solving Eq. (5.47) for q gives the critical buckling load for lateral torsional buckling of a

## monosymmetric I-beam subjected to a uniformly distributed load as

1
⎧ ⎡
( )β ( ) ( )
2 2 2
q cr = ⎨ 6 9 EI
2
π 8 − 6 EI y βx 2 π 10 + EI y βx 2 π 12
(π + 3) L ⎢ y
2 4 ⎩ ⎣ 2 x

( ) ( ) ( )
+ 4 EI y GJL2 π 10 + 24 EI y GJL2 π 8 + 36 EI y GJL2 π 6 + 4 EI y EI ω π 12   ( )( )
⎫⎪
] ( )
1

( )( )
+ 24 EI y EI ω π 10
( )
+ 36 EI y ( EI ω )π 8 2
( )
+ 4 EI y βx π 6 − 12 EI y βx π 4 ⎬
⎭⎪

(5.48)
The non-dimensional loading parameter, γ q , was given in section 3.5 as

qL3
γq =                   (5.49)
EI y GJ

Rewriting Eq. (5.48) in terms of the non-dimensional beam stiffness parameter, K , given

in Eq. (5.19) and the monosymmetric parameter,   βx , given in Eq. (5.20), as well as the loading

82
parameter, γ q , shown in Eq. (5.49) yields the elastic critical load parameter for a simply-

## supported monosymmetric beam with a uniformly distributed load as

γq =
(π 2
1
+ 3)
2 {4[(π 12
− 6π 10 + 9π 8 )βx 2 + (4π 10 + 24π 8 + 36π 6 )

1

+ (4π 10
+ 24π + 36π ) K
8 6
] + (4π
2 2 6
− 12π 4 )βx ⎬

(5.50)

As in section 5.1, this result can then be compared to the results given in (Anderson and Trahair,

1972) where the critical load parameters were obtained using the method of finite differences.

Figures 5.11 – 5.17 compare the solution presented in Eq. (5.50) with the results given by

Anderson and Trahair at various values of K and βx . As in the example in Section 5.2, the

results obtained by this research using approximate shape functions are more accurate as the

## values for βx become closer to zero, representing a doubly-symmetric beam.

83
Figure 5.11 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section

## Subjected to a Uniformly Distributed Load ( βx = − 0.6)

Figure 5.12 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section

## Subjected to a Uniformly Distributed Load ( βx = − 0.3)

84
Figure 5.13 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section

## Subjected to a Uniformly Distributed Load ( βx = − 01

.)

Figure 5.14 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section

## Subjected to a Uniformly Distributed Load ( βx = 0)

85
Figure 5.15 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section

## Subjected to a Uniformly Distributed Load ( βx = 01

.)

Figure 5.16 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section

## Subjected to a Uniformly Distributed Load ( βx = 0.3)

86
Figure 5.17 Buckling Load: Simply Supported Beam with Monosymmetric Cross-Section

## Subjected to a Uniformly Distributed Load ( βx = 0.6)

87
5.4 CANTILEVER WITH END POINT LOAD, P

This example will present the lateral-torsional buckling load, Pcr, for a cantilever beam subjected

to a concentrated central load that is applied at the end of the beam to its shear center, as shown

in Figure 5.18.

z

L
y

## Figure 5.18 Monosymmetric Cantilever Beam Subjected to Concentrated End Load

There is no axial force or uniformly distributed load acting on the beam. Therefore,   F = 0 and

## Therefore, Eq. (5.1) reduces to

88
⎛ d 2 (δ u) ⎞ ⎛ d (δφ ) ⎞ ⎛ d 2 (δφ ) ⎞
2
⎧ 2 2

1 2 1 ⎪ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟
δ Π = ⎨ EI y ⎜ 2 ⎟ + GJ ⎜ dz ⎟ + EI ω ⎜ dz 2 ⎟
2 2 ⎪⎩ ⎝ dz ⎠ ⎝ ⎠ ⎝ ⎠
L

⎡ ⎛ d 2 (δ u) ⎞ ⎛ d (δφ ) ⎞ ⎤ ⎫⎪
2

+ M x ( z ) ⎢ 2⎜⎜ ⎟ δφ + βx ⎜ ⎟ ⎥
⎜ dz ⎟ ⎥ ⎬ dz (5.51)
⎢ ⎝ dz 2 ⎟⎠ ⎝ ⎠ ⎦⎪
⎣ ⎭

The moment along the beam, M x ( z ) , for a beam with a concentrated central load can be

expressed as

M x ( z ) = − P( L − z )     for   0≤ z ≤ L        (5.52)

⎛ πz⎞ ⎛ πz⎞
Several shape functions including u( z ) = A⎜ 1 − cos ⎟ and φ ( z ) = B⎜ 1 − cos ⎟ were
⎝ 2 L⎠ ⎝ 2L⎠

experimented with to accurately predict the buckled shape of a cantilever beam. It was

determined that in order to obtain acceptable results, a more complicated shape function in the

## form of a trigonometric series must be used (Wang and Kitipornchai, 1986).

( 2r − 1)π z ⎤
n

u( z ) = Δ r ⎢1 − cos ⎥            (5.53)
r =1 ⎣ 2L ⎦

n
⎡ ( 2r − 1)π z ⎤
φ ( z) = ∑ θr ⎢1 − cos 2L ⎥            (5.54)
r= 1 ⎣ ⎦

The first variation of u and φ , as well as their first and second derivatives are needed to solve

## this problem and are shown below as

n
⎡ ( 2r − 1)π z ⎤
δu = ∑ δ Δ r ⎢1 − cos 2L ⎥            (5.55)
r= 1 ⎣ ⎦

89
n
⎡ ( 2r − 1)π z ⎤
δφ = ∑ δθr ⎢1 − cos 2L ⎥            (5.56)
r= 1 ⎣ ⎦

d (δ u) n
⎛1 ⎛ 1 2r − 1 ⎞ 2r − 1 ⎞

dz
= ∑ ⎜⎝ 2 δ Δ r sin⎜⎝ 2 L
π z⎟
⎠ L
π⎟

(5.57)
r= 1

d (δφ ) n
⎛1 ⎛ 1 2r − 1 ⎞ 2r − 1 ⎞

dz
= ∑ ⎜⎝ 2 δθr sin⎜⎝ 2 L
π z⎟
⎠ L
π⎟

(5.58)
r= 1

d 2 (δ u) ⎛1n
⎛ 1 2r − 1 ⎞ ( 2r − 1) 2 2 ⎞
= ∑ ⎜⎜ δ Δ r cos⎜ π z⎟ π ⎟⎟       (5.59)
dz 2 r= 1⎝ 4
⎝2 L ⎠ L2 ⎠

d 2 (δφ ) ⎛1
n
⎛ 1 2r − 1 ⎞ ( 2r − 1)
2 ⎞

dz 2
= ∑ ⎜⎝ 4 r ⎝ 2 L ⎠ L2 π 2 ⎟⎟⎠
⎜ δ θ cos⎜ π z ⎟     (5.60)
r=1

## Substituting Eqs. (5.52) – (5.60) into Eq. (5.51) yields

90
⎧ ⎛ n ⎛1
2
1 2 1 ⎪ ⎛ 1 2r − 1 ⎞ ( 2r − 1) 2 2 ⎞ ⎞
δ Π = ∫ ⎨ EI y ⎜⎜ ∑ ⎜⎜ δ Δ r cos⎜ π z⎟ π ⎟⎟ ⎟⎟ +
2 2L ⎪ ⎝ r=1⎝ 4 ⎝ 2 L ⎠ L 2
⎠⎠

2
⎛ n ⎛1 ⎛ 1 2r − 1 ⎞ 2r − 1 ⎞ ⎞
+ GJ ⎜ ∑ ⎜ δ θr sin⎜ π z⎟ π⎟ ⎟
⎝ r= 1⎝ 2 ⎝2 L ⎠ L ⎠⎠

2
⎛ n ⎛1 ⎛ 1 2r − 1 ⎞ ( 2r − 1)
2 ⎞⎞
EI ω ⎜ ∑ ⎜ δ θr cos⎜
⎜ ⎜ π z⎟ π 2⎟ ⎟
⎟⎟
⎝ r= 1⎝ 4 ⎝2 L ⎠ L2 ⎠⎠

⎡ ⎛ n ⎛1 ⎛ 1 2r − 1 ⎞ ( 2r − 1)
2 ⎞⎞⎛ n ⎡ ( 2r − 1)π z ⎤ ⎞
− P( L − z) ⎢ 2⎜⎜ ∑ ⎜⎜ δ Δ r cos⎜ π z⎟ π 2⎟ ⎟ ⎜
⎟ ⎟ ∑ δ θr ⎢1 − cos ⎥⎟
⎢⎣ ⎝ r = 1 ⎝ 4 ⎝2 L ⎠ L 2
⎠ ⎠ ⎝ r= 1 ⎣ 2L ⎦⎠

2 ⎤⎫
⎛ n ⎛1 ⎛ 1 2r − 1 ⎞ 2r − 1 ⎞ ⎞
+ β x ⎜ ∑ ⎜ δ θr sin⎜ π z⎟ π⎟ ⎟ ⎥ ⎪⎬ dz       (5.61)
⎝ =
⎝2 ⎝2 L ⎠ L ⎠⎠ ⎥⎪
r   1   ⎦⎭
Therefore integrating Eq. (5.51) with respect to z gives

π 4 EI y ⎡ n ⎤ GJπ 2 ⎡ n ⎛ ⎛ ( 2r − 1) 2 π 2 EI ω ⎞ 2 ⎞ ⎤
1 2
2
δ Π =
64 L3
(
⎢ ∑ ( 2r − 1) δ Δ r 2
⎢⎣ r = 1
4
)⎥+
⎥⎦ 16 L
⎢ ∑ ⎜ ( 2r − 1) 4 ⎜⎜ 1 +
⎢⎣ r = 1 ⎜⎝ ⎝ 4GJL2
⎟⎟ δ θr ⎟ ⎥
⎠ ⎟⎥
⎠⎦

Pπ 2 ⎡ n ⎛ ⎛ 1 3 ⎞ ⎞⎤
+ ⎢ ∑ ⎜⎜ ⎜ − ⎟ ( 2 r − 1) 2
δ θ r δ Δ ⎟
r⎟⎥
4 ⎢⎣ r = 1 ⎝ ⎝ 4 ( 2r − 1) 2 π 2 ⎠ ⎠ ⎥⎦

⎡ ⎞⎞⎞⎤
P ⎢ n ⎛⎜ ⎛ n ⎛ 1 − ( − 1) r + s− 1 1 − ( − 1) r − s 4
+ ∑ ( 2 r − 1) 2
δ Δ r⎜ ∑
⎜ δ θ ⎜
s⎜ + − ⎟⎟⎟⎥
2⎟⎟⎟⎥
4 ⎢ r = 1 ⎜⎝ ⎝ s= 1, s≠ r ⎝ 2( r + s − 1)
2
2 ( r − s) 2
(2r − 1) ⎠ ⎠ ⎠
⎣ ⎦

Pβ x π 2 ⎡ n ⎛⎛ 1 1⎞ ⎞⎤
+ ⎢ ∑ ⎜⎜ ⎜ ⎟( )2
2 2 − 4 2r − 1 δ θr
2 ⎟⎟ ⎥       (5.62)
8 ⎢⎣ r = 1 ⎝ ⎝ ( 2r − 1) π ⎠ ⎠ ⎥⎦

Taking the derivative of Eq. (5.62) with respect to δΔr and δθr yields the following set of

## homogeneous linear equations.

91
⎛1 ⎞
d ⎜ δ 2Π ⎟
⎠ π 2 ( 2r − 1) 4 γ c ( 2r − 1) ⎛ 1
2
⎝2 3 ⎞
= δΔ r + ⎜ − 2 2 ⎟ δθ r
dδ Δ r 32 4 ⎝ 4 ( 2r − 1) π ⎠

γ c ( 2r − 1) 2 ⎡ n ⎛ ⎛ 1 − ( − 1) r + s− 1 1 − ( − 1) r − s 4 ⎞ ⎞⎤
+ ⎢ ∑ ⎜ ⎜⎜ + − ⎟ δθ ⎟ ⎥ = 0
4π 2 ⎢⎣ s= 1, s≠ r ⎜⎝ ⎝ 2 ( r + s − 1) 2 2(r − s) 2 ( 2r − 1) 2 ⎟⎠ s ⎟⎠ ⎥⎦

(5.63)

for r = 1, ....., n.

And

⎛1 ⎞
d ⎜ δ 2Π ⎟ ⎡
⎝2 ⎠ ( 2r − 1) 2 ( 2r − 1) 4 K 2 γ c ( 2r − 1) 2 ⎛ 1 1⎞ ⎤
= ⎢ + + β⎜ − ⎟ ⎥ δθ
dδθr ⎣ 8 32 4 ⎝ ( 2r − 1) 2 π 2 4 ⎠ ⎦ r

γ c ( 2r − 1) 2 ⎡ 1 3 ⎤
+ ⎢ − 2 2 ⎥δ Δ r
4 ⎣ 4 ( 2r − 1) π ⎦

γc ⎡ n ⎛ ⎛ 1 − ( − 1) r + s− 1 1 − ( − 1) r − s 4 ⎞ ⎞⎤
+ ⎢ ∑ ( 2 s − 1) ⎜⎜
⎜ 2
+ − ⎟⎟ δ Δ s ⎟ ⎥
4π 2 ⎢⎣ s= 1, s≠ r ⎜⎝ ⎝ 2 ( r + s − 1)
2
2 ( r − s) 2 ( 2r − 1) 2 ⎠ ⎟⎥
⎠⎦

γc ⎡ n ⎛⎛ 1 1 − ( − 1) r + s− 1 1 − ( − 1) r − s ⎞ ⎞⎤
+ ⎢ ∑ ⎜⎜ ⎜⎜ ( 2r − 1)( 2 s − 1) β − ⎟ ⎟
⎟ s⎟ ⎥ = 0
δθ
π2 ⎢⎣ s= 1, s≠ r ⎝ ⎝ 4 2( r + s − 1) 2 2(r − s) 2 ⎠ ⎠ ⎥⎦

(5.64)

for r = 1, ....., n.

## where the non-dimensional critical load is defined as

92
PL2
γP =                 (5.65)
EI y GJ

Eqs. (5.63) and (5.64) are evaluated at n = 7 with the results in the form

⎧Δ1⎫ ⎧Δ1⎫
⎪θ ⎪ ⎪θ ⎪
⎪ 1⎪ ⎪ 1⎪
⎪ # ⎪ ⎪ # ⎪
[ A]⎨ ⎬ − γ P [ B]⎨ ⎬ = 0 (5.66)
⎪ # ⎪ ⎪ # ⎪
⎪Δ 7 ⎪ ⎪Δ 7 ⎪
⎪ ⎪ ⎪ ⎪
⎩ θ7 ⎭ ⎩ θ7 ⎭

## [ D]{α} = λ{α}                 (5.67)

Where

[ D] = [ A] −1 [ B] (5.68)

⎧Δ1⎫
⎪θ ⎪
⎪ 1⎪
⎪ # ⎪
α=⎨ ⎬ (5.69)
⎪ # ⎪
⎪Δ 7 ⎪
⎪ ⎪
⎩ θ7 ⎭

1
λ= (5.70)
γP

Where [A] and [B] are 14 by 14 matrices with the terms of each matrix given in Appendix B.

The eigenvalues, λ, can be derived from Eq. (5.67), which yield the critical load, as shown in Eq.

(5.70).

93
This result can then be compared to the results given in (Anderson and Trahair, 1972) where

the critical load parameters were obtained using the method of finite differences. Figures 5.19 –

5.25 compare the results obtained using Eqs. 5.63 and 5.64 at values of n = 3, 5, and 7 with the

results given by Anderson and Trahair at various values of K and βx . The shape functions

using trigonometric series appear to have the same limitations as the ones used for the simply-

supported beams where the predicted buckling moment is more accurate when the beam is closer

to being doubly-symmetric, with the best predicted result occurring for higher values of n.

## Subjected to a Concentrated End Load ( βx = − 0.6)

94
Figure 5.20 Buckling Load: Cantilever Beam with Monosymmetric Cross-Section

## Subjected to a Concentrated End Load ( βx = − 01

.)

95
Figure 5.22 Buckling Load: Cantilever Beam with Monosymmetric Cross-Section

## Subjected to a Concentrated End Load ( βx = 01

.)

96
Figure 5.24 Buckling Load: Cantilever Beam with Monosymmetric Cross-Section

97

## 6.0 FINITE ELEMENT METHOD

The finite element method is a numerical technique used to solve problems that may be

otherwise difficult to solve analytically. In this chapter, the finite element method will be used in

conjunction with the energy method to establish finite element equations that can be used to

solve for the elastic lateral buckling load. The basic concept behind the finite element method is

to model a continuum with infinite degrees of freedom and as a system of elements having finite

degrees of freedom. These elements are assembled to accurately approximate the behavior of the

entire system.

The first step toward formulating a finite element solution is to divide the system into a

number of discrete elements. These elements are connected by nodes, which are common points

shared by adjacent elements that establish the continuity of the system. The size of the elements

are arbitrary and should be selected to closely model the behavior of the entire system. After the

elements have been defined and nodes selected, a displacement function is established for each

## element. A displacement function is normally a linear combination of shape functions. Shape

functions are usually polynomial functions representing a unit displacement of a particular node

and zero for the other nodes. The number of polynomial functions used to describe each element

## is based on the number of degrees of freedom of that element. A strain-displacement

relationship and a stress-strain relationship are then defined for each element from the shape

functions.

98
The principle of minimum total potential energy will be used to derive the element

stiffness matrix and an element geometric stiffness matrix. Once these matrices are obtained,

they can be converted to the global coordinate system and assembled into a global stiffness

matrix to represent the entire system. The matrix can then be partitioned into free and restrained

degrees of freedom by the application of boundary conditions. The section of the global stiffness

matrix and geometric stiffness matrix containing the free degrees of freedom can then be used to

## obtain the buckling loads for lateral-torsional buckling.

In this project, the structural system that the finite element method is being applied to is

any plane frame. Each frame element has six nodal degrees of freedom, which means twelve

total degrees of freedom for each element. The coordinate system for the beam-column elements

## of the plane frame is shown in Figures 6.1 – 6.3.

Figure 6.1 shows the top view of the element with a displacement u(z) at a distance z

along the element, which is the lateral bending in the x direction. Of the four out-of-plane nodal

coordinates shown, u1 and u3 are the out-of-plane nodal displacements at nodes 1 and 2,

respectively, and u2 and u4 are the out-of-plane nodal rotations of nodes 1 and 2, respectively.

Figure 6.2 shows the elevation view of the element with a displacement v(z) at a distance

z along the element, which is the in-plane bending in the y direction. Of the four in-plane nodal

coordinates shown, v1 and v 3 are the in-plane nodal displacements at nodes 1 and 2,

respectively, and v 2 and v 4 are the in-plane nodal rotations of nodes 1 and 2, respectively.

Figure 6.3 shows the elevation view of the element with a displacement φ ( z ) at a distance

z along the element, which is the torsional rotation in the z-direction. Of the four nodal

coordinates shown, φ1 and φ3 are the torsional rotations at nodes 1 and 2, respectively, and φ2

99
u1 u3

u4

u2

u (z)

v1  v3

v 4

v 2

v (z)

φ1 φ3

φ2 φ4
φ(z)

## Figure 6.3 Element Degrees of Freedom with Nodal Displacements φ

100
The displacement functions for the generalized displacements u(z), v(z), and φ(z) are

## assumed to be cubic polynomials. These displacement functions are expressed below as

(Roberts, 2004)

u( z ) = [ N ]{ u} (6.1)

v ( z ) = [ N ]{ v} (6.2)

φ ( z) = [ N ]{φ} (6.3)

where

⎡ 1 1 3 1 1 3 2 2 ⎤
N = ⎢ 3 (2 z 3 − 3z 2 L + L3 ) (z L − 2z 2 L2 + zL3 ) (− 2z 3 + 3z 2 L) 3 ( z L − z L )⎥
⎣L L3 L3 L ⎦

(6.4)

and

{ u} = {u1 u2 u3 u4 }
T
(6.5)

{ v} = {v1 v 2 v 3 v4 }
T
(6.6)

{φ} = {φ 1 φ2 φ3 φ4 }
T
(6.7)

The matrix [N] is the shape function matrix for each element. Each term in the shape

function matrix represents the shape of the displacement function when the element degree of

freedom corresponding to the shape function has a unit value and all other degrees of freedom

## δu( z) = [ N ]{δu} (6.8)

δv ( z) = [ N ]{δv} (6.9)

δφ ( z) = [ N ]{δφ} (6.10)

101
The element stiffness matrix for the structure is derived using the principle of minimum

total potential energy. In order to apply the finite element method, the structure must be

separated into a finite number of elements. The total potential energy for the system may be

expressed as

1 2 1
δ Π = ∑ (δ 2U e + λδ 2 Ω e ) = 0 (6.11)
2 2

1 2 1
where δ U e and λδ 2 Ω e are the second variation of the strain energy stored in each element
2 2

and the work done on each element, respectively. The term λ represents the buckling load

factor which the initial load set has to be multiplied by to obtain the buckling load.

The strain energy stored and the work done on each individual element may be expressed

as

1 2
2
1
2
T
([ ] [ ]){δ d }
δ Π = {δ d e } k e + λ g e e (6.12)

where

⎧ u1 ⎫
⎪u ⎪
⎪ 2⎪
⎪ φ1 ⎪
⎪ ⎪
φ
{d e } = ⎪⎨ u2 ⎪⎬ = the local nodal displacement vector for each element (6.13)
⎪ 3⎪
⎪ u4 ⎪
⎪ ⎪
⎪ φ3 ⎪
⎪⎩ φ4 ⎪⎭

102
λ = the buckling parameter for each element

e

## [ g ] = the element local geometric stiffness matrix corresponding with the

e

The element local stiffness matrix and geometric stiffness matrix are both 8 by 8 matrices

representing eight local displacements for each element corresponding to the displacements

when buckling occurs. The arrangement of each matrix is shown below with both matrices being

## u1 ⎡ k11 k12 k13 k14 k15 k16 k17 k18 ⎤

u2 ⎢ k 22 k 23 k 24 k 25 k 26 k 27 k 28 ⎥
⎢ ⎥
φ1 ⎢ k 33 k 34 k 35 k 36 k 37 k 38 ⎥
⎢ ⎥
φ2 ⎢ k 44 k 45 k 46 k 47 k 48 ⎥
(6.14)
u3 ⎢ k 55 k 56 k 57 k 58 ⎥
⎢ ⎥
u4 ⎢ k 66 k 67 k 68 ⎥
φ3 ⎢ k 77 k 78 ⎥
⎢ ⎥
φ4 ⎢⎣ k 88 ⎥⎦
u1 u2 φ1 φ2 u3 u4 φ3 φ4

## u1 ⎡ g11 g12 g13 g14 g15 g16 g17 g18 ⎤

u2 ⎢ g 22 g 23 g 24 g 25 g 26 g 27 g 28 ⎥
⎢ ⎥
φ1 ⎢ g 33 g 34 g 35 g 36 g 37 g 38 ⎥
⎢ ⎥
φ2 ⎢ g 44 g 45 g 46 g 47 g 48 ⎥
(6.15)
u3 ⎢ g55 g56 g57 g58 ⎥
⎢ ⎥
u4 ⎢ g 66 g 67 g 68 ⎥
φ3 ⎢ g 77 g 78 ⎥
⎢ ⎥
φ4 ⎢⎣ g88 ⎥⎦
u1 u2 φ1 φ2 u3 u4 φ3 φ4

103
To obtain the stiffness matrices, the lateral torsional buckling equation must be re-

written in terms of the initial load set. The second variation of total potential energy then

becomes

2
⎧ ⎛ d 2 (δ u) ⎞ ⎛ d (δφ ) ⎞ ⎛ d 2 (δφ ) ⎞
2 2

1 2 1 ⎪ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟
δ Π = ⎨ EI y ⎜ 2 ⎟ + GJ ⎜ dz ⎟ + EI ω ⎜ dz 2 ⎟
2 2 ⎪⎩ ⎝ dz ⎠ ⎝ ⎠ ⎝ ⎠
L

⎡ ⎛ d (δ u) ⎞ 2 ⎛ d (δ u) ⎞ ⎛ d (δφ ) ⎞ ⎛ d (δφ ) ⎞ ⎤
2

+ λF ⎢ ⎜⎜ ⎜ dz ⎟ ⎜ dz ⎟ + (ro + y o )⎜ dz ⎟ ⎥
⎟ + 2 y o ⎜ ⎟⎜ ⎟ 2 2 ⎜ ⎟ ⎥
⎢ ⎝ dz ⎟⎠ ⎝ ⎠⎝ ⎠ ⎝ ⎠ ⎦

⎡ ⎛ d 2 (δ u) ⎞ ⎛ d (δφ ) ⎞ ⎤ ⎫⎪
2

+ λ M x ( z) ⎢ 2⎜⎜ ⎟ δφ + βx ⎜ ⎟ ⎥
⎜ dz ⎟ ⎥ ⎬ dz
⎢ ⎝ dz 2 ⎟⎠ ⎝ ⎠ ⎦⎪
⎣ ⎭

+
1
2 ∫
L
2 1
λ q(a − y o )(δφ ) dz + λ
2 ∑ P(e − y )(δφ)
o
2
=0 (6.16)

[ ]
where the first three terms will contribute to the element elastic stiffness matrix, k e , and the

rest of the equation will contribute to the element geometric stiffness matrix, g e . [ ]

## 6.1 ELASTIC STIFFNESS MATRIX

The first three terms of the buckling equation, Eq. (3.88), that will contribute to the

## element stiffness matrix are

104
⎧ ⎛ d 2 (δ u) ⎞
2
⎛ d (δφ ) ⎞
2
⎛ d 2 (δφ ) ⎞
2

1 ⎪ ⎪
⎨ EI y ⎜⎜ 2
⎟ + GJ ⎜

⎟ ⎜
⎜ dz ⎟ + EI ω ⎜ dz 2 ⎟
⎟ ⎬ dz (6.17)
2 ⎪⎩ ⎝ dz ⎠ ⎝ ⎠ ⎝ ⎠ ⎪
L ⎭

## which can be expressed as

1
{δε} T [ D]{δε} dz (6.18)
2 L

## where the generalized strain vector is

⎧⎪ d 2 (δ u) d (δφ ) d 2 (δφ ) ⎫⎪
T

{δε} = ⎨ 2 − ⎬ (6.19)
⎪⎩ dz dz dz 2 ⎪⎭

## and the generalized elasticity matrix is

⎡ EI y 0 0 ⎤
⎢ ⎥
[ D] = ⎢ 0 GJ 0 ⎥ (6.20)
⎢⎣ 0 0 EI w ⎥⎦

Substituting the first variation of the displacement functions into the generalized strain vector

yields

⎡ [ N , zz] [ 0] ⎤
⎧ {δ u}⎫

{δε} = ⎢ [ 0] [ N , z] ⎥⎥ ⎪⎨ ⎪⎬ (6.21)
⎪ {δφ}⎪⎭
⎢⎣ [ 0] − [ N , zz ]⎥⎦ ⎩

## which is substituted into Eq. (6.17) as follows.

⎡ [ N , zz] ⎡ [ N , zz]
T
[ 0] ⎤ [ 0] ⎤
⎧⎪ {δ u}⎫⎪ ⎥ ⎧⎪ {δ u}⎫⎪
T

1 ⎢ ⎥ ⎢
⎨ ⎬ [ 0] [ N , z ] [ D ] [ 0] [ N , z ] ⎥ ⎨⎪ {δφ}⎬⎪dz (6.22)
⎩ {δφ}⎪⎭
2 ⎢ ⎥ ⎢
L⎪
⎢⎣ [ 0] − [ N , zz]⎥⎦ ⎢⎣ [ 0] − [ N , zz]⎥⎦ ⎩ ⎭

## The stiffness matrix can then be expressed as

105
⎡ [ N , zz ] ⎡ [ N , zz]
T
[ 0] ⎤ [ 0] ⎤

⎢ ⎥ ⎢
[k ]
e = ⎢ [ 0] [ N , z] ⎥ [ D] ⎢ [ 0] [ N , z] ⎥⎥dz (6.23)
⎢⎣ [ 0] − [ N , zz]⎥⎦ ⎢⎣ [ 0] − [ N , zz ]⎥⎦
L

## u1 ⎡ k11 k12 k13 k14 k15 k16 k17 k18 ⎤

u2 ⎢ k 22 k 23 k 24 k 25 k 26 k 27 k 28 ⎥
⎢ ⎥
u3 ⎢ k 33 k 34 k 35 k 36 k 37 k 38 ⎥
⎢ ⎥
u4 ⎢ k 44 k 45 k 46 k 47 k 48 ⎥
(6.24)
φ1 ⎢ k 55 k 56 k 57 k 58 ⎥
⎢ ⎥
φ2 ⎢ k 66 k 67 k 68 ⎥
φ3 ⎢ k 77 k 78 ⎥
⎢ ⎥
φ4 ⎢⎣ k 88 ⎥⎦
u1 u2 u3 u4 φ1 φ2 φ3 φ4

The terms of the elastic stiffness matrix are then rearranged to their appropriate locations as

## 6.2 GEOMETRIC STIFFNESS MATRIX

The contribution of the lateral torsional buckling equation in terms of the initial load set to the

## geometric stiffness matrix is

106
⎪ ⎢ ⎜ ( )⎟ ⎛ d (δ u) ⎞ ⎛ d (δφ ) ⎞ ⎛ d (δφ ) ⎞ ⎤
⎧ ⎡⎛ d δ u ⎞ 2 2

1
⎟ + 2 y o ⎜ dz ⎟ ⎜ dz ⎟ + (ro + y o )⎜ dz ⎟ ⎥
λ F⎨ ⎜ ⎜ ⎟⎜ ⎟ 2 2 ⎜ ⎟ ⎥
2 ⎢
⎪⎩ ⎣ ⎝ dz ⎠ ⎝ ⎠⎝ ⎠ ⎝ ⎠ ⎦
L

⎡ ⎛ d 2 (δ u) ⎞ ⎛ d (δφ ) ⎞ ⎤ ⎫⎪
2

+ M x ( z ) ⎢ 2⎜⎜ ⎟ δφ + βx ⎜ ⎟ ⎥
⎜ dz ⎟ ⎥ ⎬ dz
⎢ ⎝ dz 2 ⎟⎠ ⎝ ⎠ ⎦⎪
⎣ ⎭

+
1
2 ∫
L
2 1
λ q(a − y o )(δφ ) dz + λ
2 ∑ P(e − y )(δφ) o
2
=0 (6.25)

## which can be expressed as

1

2 L
1
{δε} T [ D]{δε} dz + λ
2 ∑ P(e − y )(δφ)
o
2
(6.26)

## where the generalized strain vector is

⎧ d (δ u) d 2 (δ u) d (δφ ) ⎫
T

{δε} = ⎨ δφ ⎬ (6.27)
⎩ dz dz 2 dz ⎭

## and the generalized elasticity matrix is

⎡ F 0 0 y o F ⎤
⎢ 0 0 M x ( z) 0 ⎥
[ D] = ⎢ ⎥
M x ( z ) q(a − y o )
(6.28)
⎢ 0 0 ⎥
⎢ ⎥
⎣ y o F 0 0 F (ro 2 + y o 2 ) + ( M x ( z ) βx )⎦

Substituting the first variation of the displacement functions into the generalized strain vector

yields

107
⎡ [ N , z] [0] ⎤

[ N , zz] [0] ⎥⎥ ⎧{δ u}⎫
{δε} = ⎢⎢ 0 (6.29)
[] [ N ] ⎥ ⎨⎩{δφ}⎬⎭
⎢ ⎥
⎢⎣ [0] [ N , z]⎥⎦
which is substituted into Eq. (6.26) as follows.

⎡ [ N , z] ⎡ [ N , z]
T
[ 0] ⎤ [ 0] ⎤
T⎢ ⎥ ⎢ ⎥⎧
1 ⎧⎪ {δ u}⎫⎪ ⎢[ N , zz] [ 0] ⎢[ N , zz ] [ 0] ⎥ ⎪⎨ {δ u}⎪⎬dz

∫ ⎨ ⎬
2 L⎪⎩ {δφ}⎪⎭ ⎢ [ 0]

[ N]
⎥ [ D]

⎢ [ 0]

[ N] ⎥ ⎪ {δφ}⎪
⎥⎩ ⎭
⎢⎣ [ 0] [ N , z]⎥⎦ ⎢⎣ [ 0] [ N , z]⎥⎦

1 ⎪⎧ {δ u}⎪⎫ ⎪⎧ {δ u}⎪⎫
T

+ λ⎨
2 ⎪⎩ {δφ}⎪⎭
[
T
]
⎬ [ 0] [ N ] P(e − y o ) [ 0] [ N ] ⎨ ⎬
⎪⎩ {δφ}⎪⎭
[ ] (6.30)
z= zp

## The geometric stiffness matrix can then be expressed as

⎡ [ N , z] ⎡ [ N , z]
T
[ 0] ⎤ [ 0] ⎤
⎢ ⎥ ⎢ ⎥
[ N , zz] [ 0] ⎢ [ N , zz ] [ 0]
[ ]
ge = ⎢
L

⎢ [ 0] [ N]
⎥ [ D]

⎢ [ 0]

[ N]
⎥dz

⎢⎣ [ 0] [ N , z]⎥⎦ ⎢⎣ [ 0] [ N , z]⎥⎦
1
+
2
[ ]
[ 0] [ N ] P(e − yo ) [ 0] [ N ]
T
[ ] (6.31)
z= zp

## u1 ⎡ g11 g12 g13 g14 g15 g16 g17 g18 ⎤

u2 ⎢ g 22 g 23 g 24 g 25 g 26 g 27 g 28 ⎥
⎢ ⎥
u3 ⎢ g 33 g 34 g 35 g 36 g 37 g 38 ⎥
⎢ ⎥
u4 ⎢ g 44 g 45 g 46 g 47 g 48 ⎥
(6.32)
φ1 ⎢ g55 g56 g57 g58 ⎥
⎢ ⎥
φ2 ⎢ g 66 g 67 g 68 ⎥
φ3 ⎢ g 77 g 78 ⎥
⎢ ⎥
φ4 ⎢⎣ g88 ⎥⎦
u1 u2 u3 u4 φ1 φ2 φ3 φ4

108
The terms of the geometric stiffness matrix are then rearranged to their appropriate locations as

## 6.3 FINITE ELEMENT METHOD CONSIDERING PREBUCKLING

DEFLECTIONS

The terms of the second variation of the total potential energy equation that account for

prebuckling are

⎡ Iy d 2 (δ u) ⎛ z
⎞ d (δφ ) d (δ u)
2

∫ ∫
1 M x ( z)
⎢ − 2 M x ( z) δφ + GJ ⎜ − dz + C⎟
2 ⎣ Ix dz 2 ⎝ 0 EI x ⎠ dz dz 2
L

M x ( z ) d (δφ ) d (δ u) ⎛ z
M x ( z) ⎞ d 2 (δφ ) d 3 (δ u)
+ GJ
EI x dz dz
+ EI ω ⎜
⎝ ∫0

EI x
dz + C⎟
⎠ dz 2 dz 3

Iω d 2 (δφ ) d (δ u) ⎤
+ V ⎥ dz (6.33)
I x y dz 2 dz ⎥⎦

⎛ z
M x ( z) ⎞
When integrating ⎜
⎝ ∫
0

EI x
dz⎟ , M x ( z) will be redefined as

H x ( z) =
∫ M (η)dη
0
x (6.34)

## Further substituting the definitions given for M x ( z) in Chapter 4 yields

1 2 1 3
H x ( z) = M 1 z + V z − qz for 0 ≤ z ≤ zp (6.35)
2 1 6

1 2 1 3 1
H x ( z) = M 1 z + V1 z − qz − P( z 2 − zz p ) for zp ≤ z ≤ L (6.36)
2 6 2

109
Substituting H x ( z) into the prebuckling energy equation and rearranging in terms of initial load

set gives

⎡ d 2 (δφ ) d 3 (δ u) d (δφ ) d 2 (δ u) ⎤ ⎡ Iy d 2 (δ u)
∫ ∫
1 1
⎢ EI ω C + GJC ⎥ dz + λ ⎢ − 2 M ( z ) δφ
2 ⎢⎣ dz 2 dz 3 dz dz 2 ⎥⎦ 2 ⎣ Ix x dz 2
L L

Iω d (δφ ) d (δ u) GJ
2 3
d (δφ ) d (δ u) GJ 2
d (δφ ) d (δ u)
− H x ( z) − H ( z ) + M ( z )
Ix dz 2 dz 3 EI x x dz dz 2 EI x x dz dz

Iω d 2 (δφ ) d (δ u) ⎤
+ V ⎥ dz (6.37)
I x y dz 2 dz ⎥⎦

where the first bracketed term of this equation contributes to the elastic stiffness matrix and the

remaining terms contribute to the geometric stiffness matrix. Eq. (6.12) can then be expressed as

1
2
( (
{δ d e }T [ k e ] + [ k e ] P + λ [ ge ] + [ ge ] P )){δ d }
e (6.38)

where [ke] and the [ge] are the same elastic and geometric stiffness matrices, respectively, that

were derived in the previous two sections and [ke]P and [ge]P represent the prebuckling effects

## 6.4 ELASTIC STIFFNESS MATRIX CONSIDERING PREBUCKLING

DEFLECTIONS

The terms from the prebuckling energy equation (Eq. 4.27) that contribute to the elastic

## stiffness matrix, [ke]P, are

110
⎡ d 2 (δφ ) d 3 (δ u) d (δφ ) d 2 (δ u) ⎤

1
⎢ EI ω C + GJC ⎥ dz (6.39)
2 ⎢⎣ dz 2 dz 3 dz dz 2 ⎥⎦
L

## which can be expressed as

1
{δε} T [ D]{δε} dz (6.40)
2 L

## where the generalized strain vector is

⎧⎪ d 2 (δ u) d 3 (δ u) d (δφ ) d 2 (δφ ) ⎫⎪
T

{δε} = ⎨ 2 ⎬ (6.41)
⎪⎩ dz dz 3 dz dz 2 ⎪⎭

## and the generalized elasticity matrix is

⎡ GJ ⎤
⎢ 0 0
2
0 ⎥
⎢ EI ω ⎥
⎢ 0 0 0 ⎥
[ D] = ⎢ GJ 2 ⎥ (6.42)
⎢ ⎥
⎢ 2 0 0 0 ⎥
⎢ EI ω ⎥
⎢ 0 0 0 ⎥
⎣ 2 ⎦

Substituting the first variation of the displacement functions into the generalized strain vector

yields

⎡ [ N , zz] 0 ⎤
⎢ ⎥
[ N , zzz] 0 ⎥ ⎧{δ u}⎫
{δε} = ⎢⎢ 0 (6.43)
[ N , z] ⎥ ⎨⎩{δφ}⎬⎭
⎢ ⎥
⎢⎣ 0 [ N , zz]⎥⎦
which is substituted into Eq. (6.40) as follows.

111
⎡ [ N , zz] ⎡ [ N , zz]
T
0 ⎤ 0 ⎤
T⎢ ⎥ ⎢ ⎥
⎧⎪ {δ u}⎫⎪ [ N , zzz] 0 ⎥ ⎧⎪ {δ u}⎫⎪
⎢[ N , zzz]

1 ⎢ 0 ⎥
C ⎨ ⎬ [ D]
[ N , z] ⎥ ⎨⎪⎩ {δφ}⎬⎪⎭
dz (6.44)
2 L⎪⎩ {δφ}⎪⎭ ⎢ 0 [ N , z] ⎥ ⎢ 0
⎢ ⎥ ⎢ ⎥
⎢⎣ 0 [ N , zz]⎥⎦ ⎢⎣ 0 [ N , zz]⎥⎦
The stiffness matrix can then be expressed as

⎡ [ N , zz] ⎡ [ N , zz]
T
0 ⎤ 0 ⎤
⎢ ⎥ ⎢ ⎥
[ N , zzz] ⎥ [ D] ⎢ [ N , zzz]

0 0
[k ]e P =C ⎢
⎢ 0 [ N , z] ⎥ ⎢ 0
⎥ dz
[ N , z] ⎥
(6.45)
L
⎢ ⎥ ⎢ ⎥
⎢⎣ 0 [ N , zz]⎥⎦ ⎢⎣ 0 [ N , zz]⎥⎦
The elastic stiffness matrix considering prebuckling deflections, [ k e ] P , is again an 8 by 8

matrix. The terms of [ k e ] P are then rearranged to their appropriate locations as shown in Eq.

## 6.5 GEOMETRIC STIFFNESS MATRIX CONSIDERING PREBUCKLING

DEFLECTIONS

The terms from the prebuckling energy equation Eq. (4.27) in terms of the initial load set

## that contribute to the geometric stiffness matrix, [ge]P, are

112
⎡ Iy d 2 (δ u) Iω d 2 (δφ ) d 3 (δ u) GJ d (δφ ) d 2 (δ u)

1
λ ⎢ − 2 M x ( z) δφ − H ( z) − H ( z)
2 ⎣ Ix dz 2 Ix x dz 2 dz 3 EI x x dz dz 2
L

GJ d (δφ ) d (δ u) I ω d 2 (δφ ) d (δ u) ⎤
+ M x ( z) + Vy ⎥ dz (6.46)
EI x dz dz Ix dz 2 dz ⎥⎦

## which can be expressed as

1
λ {δε} [ D]{δε} dz
T
(6.47)
2 L

## where the generalized strain vector is

⎧ d (δ u) d 2 (δ u) d 3 (δ u) d (δφ ) d 2 (δφ ) ⎫
T

{δε} = ⎨ δφ ⎬ (6.48)
⎩ dz dz 2 dz 3 dz dz 2 ⎭

## and the generalized elasticity matrix is

⎡ GJM x ( z) Iω Vy ⎤
⎢ 0 0 0 0 ⎥
⎢ 2 EI x 2I x ⎥
⎢ I y M x ( z) GJH x ( z) ⎥
⎢ 0 0 0 − − 0 ⎥
Ix 2 EI x
⎢ ⎥
⎢ I ω H x ( z) ⎥
0 0 0 0 0 −
⎢ 2I x ⎥
[ D] = ⎢ I y M x ( z) ⎥
⎢ 0 − 0 0 0 0 ⎥
⎢ Ix ⎥
⎢ GJM x ( z ) GJH x ( z ) ⎥
⎢ − 0 0 0 0 ⎥
⎢ 2 EI x 2 EI x ⎥
⎢ IωVy I ω H x ( z) ⎥
⎢ 2I 0 − 0 0 0 ⎥
⎣ x 2I x ⎦

(6.49)

113
Substituting the first variation of the displacement functions into the generalized strain vector

yields

⎡ [ N , z] 0 ⎤
⎢ ⎥
⎢ [ N , zz] 0 ⎥
⎢[ N , zzz] 0 ⎥ ⎧{δ u}⎫
{δε} = ⎢ 0 ⎥ (6.50)
⎢ [ N ] ⎥ ⎨⎩{δφ}⎬⎭
⎢ 0

[ N , z] ⎥⎥
⎢⎣ 0 [ N , zz]⎥⎦
which is substituted into Eq. (6.47) as follows.

⎡ [ N , z] ⎡ [ N , z]
T
0 ⎤ 0 ⎤
⎢ ⎥ ⎢ ⎥
⎢ [ N , zz] 0 ⎥ ⎢ [ N , zz ] 0 ⎥
⎪⎧ {δ u}⎪⎫ ⎢ [ N , zzz] ⎢ [ N , zzz ] ⎥ ⎪⎧ {δ u}⎪⎫
T

1 0 0
λ ⎨ ⎬ ⎢ ⎥ [ D] ⎢ ⎥⎨ ⎬ dz (6.51)
2 L⎪⎩ {δφ}⎪⎭ ⎢ 0 [ N] ⎥ ⎢ 0 [ N] ⎥ ⎪⎩ {δφ}⎪⎭
⎢ 0

[ N , z] ⎥⎥ ⎢ 0

[ N , z] ⎥⎥
⎢⎣ 0 [ N , zz]⎥⎦ ⎢⎣ 0 [ N , zz]⎥⎦
The geometric stiffness matrix can then be expressed as

⎡ [ N , z] ⎡ [ N , z]
T
0 ⎤ 0 ⎤
⎢ ⎥ ⎢ ⎥
⎢ [ N , zz ] 0 ⎥ ⎢ [ N , zz ] 0 ⎥
⎢ [ N , zzz ] 0 ⎥ ⎢ [ N , zzz ] ⎥

0
[g ]
e P = ⎢
0 [ N] ⎥
⎥ [ D] ⎢
[ N] ⎥
⎥ dz (6.52)
L⎢ ⎢ 0
⎢ 0

[ N , z] ⎥⎥ ⎢ 0

[ N , z] ⎥⎥
⎢⎣ 0 [ N , zz]⎥⎦ ⎢⎣ 0 [ N , zz]⎥⎦
The geometric stiffness matrix considering prebuckling deflections, [ g e ] P , is again an 8 by 8

matrix. The terms of [ g e ] P are then rearranged to their appropriate locations as shown in Eq.

## (6.15) and listed in Appendix A.

114
7.0 SUMMARY

Lateral-torsional buckling occurs when a beam has a relatively small lateral and torsional

stiffness compared to its stiffness in the plane of loading, causing the beam to deflect laterally

and twist out of plane when it reaches a critical load. This load is known as the lateral-torsional

buckling load, or LTB load. A review of existing literature was presented to demonstrated

previous methods used to derive the LTB load, including the differential equilibrium method of

stability and energy methods. The derivations of the LTB load for structures with doubly-

symmetric cross-sections have been long discussed and readily available. Lateral-torsional

## buckling of a structure with a monosymmetric cross-section is an underdeveloped topic, with

derivations complicated by the fact that the centroid and the shear center do not coincide in these

cross sections. The purpose of this study was to derive equations for lateral-torsional buckling of

structures with monosymmetric cross-sections, examine the validity of these equations using

approximate shape functions and comparing these results to other analysis, and use the finite

element method to obtain element elastic stiffness and geometric stiffness matrices that may be

used in future research, in conjunction with computer software, to predict the LTB load for

complex systems.

## The energy equation for lateral-torsional buckling of a beam-column element is based on

the theorem of minimum total potential energy. The total potential energy of the system is the

sum of the strain energy and the potential energy of the external loads. This theorem indicates

115
that the critical condition for buckling occurs when the second variation of the total potential

energy is equal to zero, representing the transition from a stable to an unstable state.

The energy equations in this paper are derived for cases both ignoring and considering

## prebuckling displacements. Prebuckling, or in-plane, displacements are considered so small for

thin-walled structures that their effect on the lateral-torsional buckling load is negligible. This

assumption is only valid when the ratios of minor axis flexural stiffness and torsional stiffness to

the major axis flexural stiffness are very small. When these ratios are not small, the effects of

prebuckling deformations will significantly alter the LTB load and cannot be ignored. A non-

dimensional buckling equation is also presented for cases without prebuckling displacements.

The advantage of this form is that the solution can be transferred to other structural systems with

The validity of these energy equations for the lateral-torsional buckling of beam-column

## elements with monosymmetric cross-sections is examined in the applications section of the

paper. Suitable trigonometric shape functions for beams that are simply supported and cantilever

are used to compare the buckling results obtained from the present research to results obtained in

previous literature using the method of finite differences. The derived energy equations prove to

degree of precision based on the ability of the shape function to predict the buckled shape of the

beam.

The finite element method is used to project the energy equations for lateral-torsional

## buckling of a beam-column element onto a structure with complicated loads, boundary

conditions, and geometry. The expression for the second variation of total potential energy is

116
used to derive element elastic stiffness and geometric stiffness matrices for the structure. These

matrices can then be transformed to a global coordinate system for each element and assembled

so boundary conditions can be used to transform the structure from unrestrained to restrained.

The result is a generalized eigenvalue problem that will produce lateral-torsional buckling loads

for the structure. The objective is that future research can utilize these stiffness matrices, along

with computer software, to develop models of complex systems with monosymmetric beam-

117
APPENDIX A

## A.1 ELEMENT ELASTIC STIFFNESS MATRIX

12E I 2E I
y y
k = k =
11 3 26 L
L

12E I
6E I ω 6G J
y k = +
k = 33 3 5L
12 2 L
L

6E I
−12E I ω GJ
y k = +
k = 34 2 10
15 3 L
L

−12E I
6E I ω 6GJ
y k = −
k = 37 3 5
16 2 L
L

6E I
4E I ω GJ
y k = +
k = 38 2 10
22 L L

−6E I 4E I
ω 2G J L
y
k = k = +
25 2 44 2 15
L   L

118
−6E I 4E I
ω GJ y
k = − k =
47 2 10 66 L
L

12E I
2E I ω 6GJ
ω GJ L k = +
k = − 77 3 5L
48 L 30 L

12E I −6E I
y ω GJ
k = k = −
55 3 78 2 10
L   L

−6E I 4E I
y ω 2GJ L
k = k = +
56 2 88 L 15
L

119
A.2 ELEMENT GEOMETRIC STIFFNESS MATRIX

6F
g11 =
5 L

F
g12 =
10

1 ⎡
( −q) L − 7 L V 1 + 7 L P − 84 M 1 − 84 P z p + 84 y o F⎤⎦
2
g13 = ⎣
70 L

−1 2 1 1 1
g14 = qL − M1 − P zp + yo F
140 10 10 10

−6F
g15 =
5L

F
g16 =
10

g17 =
−1
70 L
( 34 q L − 77 L V 1 + 77 L P − 84 M 1 − 84 P z p + 84 y o F)
2

3 2 1 1 1 1 1
g18 = qL − L V1 + LP− M1 − P zp + yo F
70 10 10 10 10 10

120

2F L
g22 =
15

11 2 1 1 11 11 1
g23 = qL − L V1 + LP− M1 − P zp + yo F
420 5 5 10 10 10

1 3 1 2 1 2 2 2 2
g24 = qL − L V1 + L P− L M1 − L P zp + L yo F
210 30 30 15 15 15

−F
g25 =
10

−F L
g26 =
30

−23 2 1 1 1 1 1
g27 = q L + L V1 − L P + M1 + P zp − yo F
210 5 5 10 10 10

1 3 1 1 1
g28 = qL + L M1 + L P zp − L yo F
210 30 30 30

121
1⎡⎣13 L2 q a − 13 L2 q y o − 6 L2 q β x + 21 L β x V 1 − 21 L β x P + 42 F ( ro) 2
g33 =
35 L

1 ⎡ 2
+ 42 F ( y o) + 42 β x M 1 + 42 β x P z p ⎤⎦ +
3⎤
2
6
⎣ p
2 ( z ) 3
− 3 ( z p ) 2
L + L ⎦ P ( e − y o)
L

11 2 11 2 1 2 1 1 1 2 1
L q yo − L q β x + L β x V1 − L β x P + F ( ro) + F ( yo)
2
g34 = L q a−
210 210 28 10 10 10 10

1 1 1 ⎡
2 ( zp) − 3 ( zp) L + L ⎤⎦ P ( e − yo) ⎡⎣( zp) L − 2 ( zp) L + zp L ⎤⎦
3 2 3 3 2 2 3
+ β x M1 + β x P zp + ⎣
10 10 6
L

g35 =
1
70 L
(
2
q L + 7 L V 1 − 7 L P + 84 M 1 + 84 P z p − 84 y o F )

−17 2 1 1 1 1 1
g36 = qL + L V1 − LP− M1 − P zp + yo F
420 10 10 10 10 10

−3 ⎡
(−3) L q a + 3 L q y o − 4 L q β x + 14 L β x V 1 − 14 L β x P + 28 F ( ro) + 28 F ( y o) +
2 2 2 2 2
g37 = ⎣
70 L

1 ⎡
+ 28 β x M1 + 28 β x P zp⎤⎦ + 2 ( zp ) − 3 ( zp ) L + L ⎤⎦ P ( e − y o) ⎡⎣(−2) ( zp ) + 3 ( zp ) L⎤⎦
3 2 3 3 2
6

L

122
−13 2 13 2 1 2 1 1 1
F ( ro) + F ( y o) +
2 2
g38 = L qa+ L q yo + L q βx+ β x M1
420 420 70 10 10 10

1 1 ⎡
2 ( z p ) − 3 ( z p ) L + L ⎤⎦ P ( e − y o) ⎡⎣( z p ) L − ( z p ) L ⎤⎦
3 2 3 3 2 2
+ β x P zp + ⎣
10 6
L

1 1 1 1 1 2
L F ( ro)
3 3 3 2 2 2
g44 = L qa− L q yo − q βx L + L β x V1 − L βxP+
105 105 105 30 30 15

2 2 2 1 ⎡ 2
L F ( y o) + ⎣( zp ) L − 2 ( zp ) L + zp L ⎤⎦ P ( e − y o)
2 3 2 2 3
+ L β x M1 + L β x P zp +
15 15 15 6
L

1 2 1 1 1
g45 = qL + M1 + P zp − yo F
140 10 10 10

−1 3 1 2 1 2 1 1 1
g46 = qL + L V1 − L P+ L M1 + L P zp − L yo F
84 30 30 30 30 30

13 2 13 2 1 2 1 1 1 2 1 2 1
g47 = L q a− L q yo + L q β x − L β x V1 + L β x P − F ( ro) − F ( yo) − β x M1
420 420 28 10 10 10 10 10

1 1 ⎡ 3
1− β x P zp +
6 ⎣( zp) L − 2 ( zp) 2 L2 + zp L3⎤⎦ P ( e − yo) ⎡⎣(−2) ( zp) 3 + 3 ( zp) 2 L⎤⎦
10 L

123
−1 3 1 3 1 3 1 2 1 2 1 2 1 2 1
g48 = L q a+ L q yo + q β x L − L β x V1 + L β x P − L F ( ro) − L F ( yo) − L β x M1 −
140 140 140 60 60 30 30 30

1 1 ⎡ 3
( zp ) L − 2 ( zp ) L + zp L ⎤⎦ P ( e − yo) ⎡⎣( zp ) L − ( zp ) L ⎤⎦
2 2 3 3 2 2
1− L β x P zp + ⎣
30 6
L

6F
g55 =
5L

−F
g56 =
10

−1 ⎡
( −34) q L + 77 L V 1 − 77 L P + 84 M 1 + 84 P z p − 84 y o F⎤⎦
2
g57 = ⎣
70 L

−3 2 1 1 1 1 1
g58 = qL + L V1 − LP+ M1 + P zp − yo F
70 10 10 10 10 10

2F L
g66 =
15

−79 2 9 9 11 11 1
g67 = qL + L V1 − LP+ M1 + P zp − yo F
210 10 10 10 10 10

4 3 1 2 1 2 2 2 2
g68 = qL − L V1 + L P− L M1 − L P zp + L yo F
105 10 10 15 15 15

124

1 ⎡
13 L q a − 13 L q y o − 6 L q β x + 21 L β x V 1 − 21 L β x P + 42 F ( ro) + 42 F ( y o)
2 2 2 2 2
g77 = ⎣
35 L

1 ⎡ 2
+ 42 β x M 1 + 42 β x P z p ⎤⎦ + ( −2) ( z p ) + 3 ( z p ) L⎤⎦ P ( e − y o)
3 2
6

L

−11 11 1 1 1 1 1
F ( ro) − F ( y o) −
2 2 2 2 2
g78 = L qa+ L q yo − L q βx− βx M1 − βx P zp
210 210 70 10 10 10 10

1 ⎡
( −2) ( z p ) + 3 ( z p ) L⎤⎦ P ( e − y o) ⎡⎣( z p ) L − ( z p ) L ⎤⎦
3 2 3 2 2
+ ⎣
6
L

1 1 3 1 1 2
L F ( ro)
3 3 3 2 2 2
g88 = L qa− L q yo − qβxL + L β x V1 − L βxP+
105 105 70 10 10 15

2 2 2 1 ⎡ 2
2 2⎤
L F ( y o) + ( ) ( ) ⎦ P ( e − y o)
2 3
+ L β x M1 + L β x P zp + ⎣ z p L − z p L
15 15 15 6
L

125
A.3 ELEMENT NON-DIMENSIONAL STIFFNESS MATRIX

k = 12 2
11   1 6K
k = +
34 10 2
π

k =6
12
2
−6 12 K
k = −
37 5 2
π
k = −12
15

2
1 6K
k =6 k
38
= +
16   10 2
π

k =4
22   2
2 4K
k = +
44 15 2
π
k = −6
25

2
−1 6K
k = −
47 10 2
k =2 π
26

2
6 12 K
2 −1 2K
k = + k = +
48 30 2
33 5 2 π
π

126
k = 12 2
55   6 12 K
k = +
77 5 2
π

k = −6 2
56 −1 6K
k = +
78 10 2
π

2
k =4 2 4K
66   k = +
88 15 2
π

127
A.4 ELEMENT NON-DIMENSIONAL GEOMETRIC STIFFNESS MATRIX

6F
g =
11 5

−2 F
g =
12 5

6M V 6F y 6Pz
P 1 1 P o p
g = − − − + −
13 10 5 10 70 5 5

M Pz 2F y
−P 1 p o
g = − − −
14 140 10 10 5

−6 F
g =
15 5

F
g =
16 10

11 V 6M 6Pz 6F y
−17 P 1 11 P 1 p o
g = + − + + −
17 35 10 10 5 5 5

128

V M Pz Fy
3P 1 P 1 p o
g = − + − − +
18 70 10 10 10 10 10

7F
g =
22 15

4V 3M 3Pz 2F y
P 1 4P 1 p o
g = − + − − −
23 14 15 15 5 5 5

V 2M 2Pz 7F y
2P 1 P 1 p o
g = − + − − +
24 105 15 15 15 15 15

2F
g =
25 5

−F
g =
26 5

16V 7M 7Pz 2F y
3P 1 16P 1 p o
g = − + − − +
27 7 15 15 5 5 5

2V M Pz Fy
−P 1 2P 1 p o
g = + − + + −
28 21 15 15 5 5 5

129

## ⎡13P K a − 13P K y − 6 P β π + 21β π V − 21β π P + 84 r 2 π + 84 y 2 π

( o) ( o)
1
g =
33 70π ⎣ o x x 1 x

2
p⎦ 2 ⎣ ( p )

+ 42 β π M + 42 β π P z +
1⎡
2 z
3
− 3 (z ) + 1 P (e − y )
2 ⎤ K

x 1 x p ⎦ o π

−1
g
34
=
840π ⎣ o x x 1 x ( o)
⎡(−22) P K a + 22P K y − 3 P β π + 42β π V − 42β π P + 336 r 2 π

+ 336 y ( o)2 π + 168β x π M1 + 168β x π P zp⎤⎦
( ) ( p) ⎦ ⎣( p) ( p) p⎦ ( o) π
⎡2 z 3 − 3 z 2 + 1⎤ ⎡ z 3 − 2 z 2 + z ⎤ P e − y K
1
+
2⎣ p

6M V 6F y 6Pz
−P 1 1 P o p
g = + + + − +
35 10 5 10 70 5 5

M V Fy Pz
−P 1 1 17 P o p
g = − + − + −
36 10 10 10 420 10 10

−3
g
37
=
140π ⎣ o x x 1 x ( o) ( o)
⎡(−3) P K a + 3 P K y − 4 P β π + 14β π V − 14β π P + 56 r 2 π + 56 y 2 π

+ 28β π M + 28β π P z ⎤ + ( ) ( ) ( ) ( ) ( )
1⎡
2 z − 3 z + 1⎤ ⎡(−2) z + 3 z ⎤ P e − y
3 2 3 2 K
x 1 x p⎦ 2⎣ p p ⎦⎣ p p ⎦ o π

130
( o) ( o)
⎡(−13) P K a + 13P K y + 6 P β π + 84 r 2 π + 84 y 2 π + 42β π M
1
g =
38 840π ⎣ o x x 1

+ 42β π P z ⎤ + ( ) ( ) ( ) ( ) ( )
1⎡
2 z − 3 z + 1⎤ ⎡ z − z ⎤ P e − y
3 2 3 2 K
x p⎦ 2⎣ p p ⎦⎣ p p ⎦ o π

## ⎡2 P K a − 2 P K y + 196 r 2 π − 11P β π + 196 y 2 π + 35β π V − 35β π P

( o) ( o)
1
g =
44 420π ⎣ o x x 1 x

2
+ 98β π M + 98β π P z ⎤ +
1⎡
( ) ( p) + z ⎤ P e −y
( o) π
3 2 K
z −2 z
x 1 x p⎦ 2⎣ p p⎦

M Pz 2F y
P 1 p o
g = + + +
45 140 10 10 5

V M Pz Fy
−P 1 P 1 p o
g = + − + + −
46 84 30 30 30 30 5

( o) ( o)
⎡13P K a − 13P K y − 3P β π + 42β π V − 42β π P + 336 r 2 π + 336 y 2 π
1
g =
47 840π ⎣ o x x 1 x

## π + 168β π M + 168β π P z ⎤ + ( z ) − 2 ( z ) + z ⎤ ⎡(−2) ( z ) + 3 ( z ) ⎤ P ( e − y )

1⎡ 3 2 3 2 K
x 1 x p⎦ 2⎣ p p p⎦⎣ p p ⎦ o π

−1
g
48
=
840π ⎣ o x x 1 x ( o) ( o)
⎡3 P K a − 3 P K y − 19P β π + 49β π V − 49β π P + 168 r 2 π + 168 y 2 π

π + 84β π M + 84β π P z ⎤ + ( z ) − 2 ( z ) + z ⎤ ⎡( z ) − ( z ) ⎤ P ( e − y )
1⎡ 3 2 3 2 K
x 1 x p⎦ 2⎣ p p p⎦⎣ p p ⎦ o π

131

6F
g =
55 5

−F
g =
56 10

11V 6M 6Pz 6F y
17P 1 11P 1 p o
g = − + − − +
57 35 10 10 5 5 5

V M Pz Fy
−3 P 1 P 1 p o
g = + − + + −
58 70 10 10 10 10 10

2F
g =
66 15

9V 11M 11P z Fy
−79 P 1 9P 1 p o
g = + − + + −
67 210 10 10 10 10 10

V 2M 2Pz 2F y
4P 1 P 1 p o
g = − + − − +
68 105 10 10 15 15 15

## ⎡13P K a − 13P K y − 6 P β π + 21β π V − 21β π P + 84 r 2 π + 84 y 2 π

( o) ( o)
1
g =
77 70π ⎣ o x x 1 x

132
2

+ 42 β π M + 42 β π P z +
1⎡
(−2) ( z ) + 3 ( z )
3 2⎤
P (e − y )
K
x 1 x p⎦ 2 ⎣ p p ⎦ o π

−1
g =
78 420π ⎣ o x ( o) ( o)
⎡11PK a − 11PK y + 3Pβ π + 42 r 2 π + 42 y 2 π + 21β π M + 21β π Pz ⎤
x 1 x p⎦

( ) ( ) ( ) ( ) ( )
1⎡
(−2) z + 3 z ⎤ ⎡ z − z ⎤ P e − y
3 2 3 2 K
+
2⎣ p p ⎦⎣ p p ⎦ o π

## ⎡2 P K a − 2 P K y − 9 P β π + 21β π V − 21β π P + 56 r 2 π + 56 y 2 π + 28β π M

( o) ( o)
1
g =
88 420π ⎣ o x x 1 x x 1

2
p⎦ 2 ⎣( p)

+ 28β π P z +
1⎡ 3 2⎤
z − (z ) P (e − y )
K
x p ⎦ o π

133

−6C E I
ω CGJ
k = −
p11 3 2L
L

6C E I
ω CGJ
k = +
p18 3 2L
L

CGJ
k =
p23 2L

−3C E I
ω CGJ
k = −
p24 2 4
L

−C G J
k =
p27 2L

3C E I
ω CGJ
k = +
p28 2 4
L

−C G J
k =
p36 2L

134
6C E I
ω CGJ
k = +
p45 3 2L
L

−3C E I
ω CGJ
k = −
p46 2 4
L

−6C E I
ω CGJ
k = −
p58 3 2L
L

CGJ
k =
p67 2L

3C E I
ω CGJ
k = +
p68 3 4
L

135

## A.6 ELEMENT PREBUCKLING GEOMETRIC STIFFNESS MATRIX

g =
1 ⎛ 5 L4 G J P − 15 G J L3 V + 15 G J L3 P − 180L I V E − 3 L4 I E P
13 ⎝ 1 w 1 y
4
10 L E I ( x)
3 3 2 2 2
+ 6 L I E V − 6 L I E P − 30 L G J M − 30 L G J P z − 360I E M + 30 I E P L
y 1 y 1 p w 1 w

+ 120I E L P − 360I E P z ⎞
w w p⎠

−1 ⎛ 180L I V E − 39 L4 I E P + 84 L3 I E V − 84 L3 I E P + 90 L2 G J M
g =
14 ⎝ w 1 y y 1 y 1
3
30 L E I ( x)
2 4 3
− 30I E P L − 120I E L P + 360I E P z − 14L G J P − 45G J L P + 360I E M
1 w w w p w 1

3 5 4 3 2 3
+ 60I E P L + 28L G J P + 90G J L P − 180L G J P z + 90L G J P z + 45G J L V
w p p 1

2 4 3 2 2 5
+ 240I E L P − 90L G J V − 180L G J M − 360L I V E + 60L I E M + 48L I E P −
w 1 1 w 1 y 1 y

3 3 4 4 2
− 120L I E M − 120L I E P z − 108L V I E + 108L P I E + 60L I E P z − 720I E L P z
y 1 y p 1 y y y p w p

− 720I E L M ⎞
w 1⎠

136

−1 ⎛ 5 L4 G J P − 15G J L3 V + 15G J L3 P − 180L I V E − 3 L4 I E P
g =
17 ⎝ 1 w 1 y
4
10L E I ( x)
3 3 2 2 2
P + 6 L I E V − 6 L I E P − 30L G J M − 30L G J P z − 360I E M + 30I E P L
y 1 y 1 p w 1 w

+ 120I E L P − 360I E P z ⎞
w w p⎠

g =
1 ⎛ 8 L4 G J P − 25G J L3 V + 25G J L3 P − 50L2 G J M − 50L2 G J P z
18 ⎝ 1 1 p
3
20L E I ( x)
4 3 3 2 2
− 180L I V E + 6 L I E P − 14L I E V + 14L I E P − 20L I E M − 20L I E P z
p w 1 y y 1 y y 1 y p

## − 360I E M + 30 I E P L + 120I E L P − 360I E P z

2 ⎞
w 1 w w w p⎠

g =
1 ⎛ 360L I V E − 24L4 I E P + 72L3 I E V − 72L3 I E P − 60L2 G J M
23 ⎝ w 1 y y 1 y 1
3
60L E I ( x)
2 4 3
+ 60I E P L + 240I E L P − 720I E P z − 23L G J P − 30G J L P − 720I E M
w w w p w 1

3 5 3 2 3 2
− 120I E P L + 16L G J P + 120L G J P z − 60L G J P z + 30G J L V − 480I E L P +
w p p 1 w

3 2 5 3 3 4
P + 120L G J M + 120L I E M + 48L I E P − 240L I E M − 240L I E P z − 144L V I E +
1 y 1 y y 1 y p 1 y

4 2
+ 144 L P I E + 120 L I E P z + 1440I E L P z + 1440I E L M ⎞
y y p w p w 1⎠

137
−1 ⎡(−360) L I V E − 210L4 I E P + 528L3 I E V − 528L3 I E P − 60I E P L2
g =
24 ⎣ w 1 y y 1 y w
2
( x)
180L E I

4 3 3
− 240I E L P + 720I E P z + 65L G J P + 120G J L P + 720I E M + 240I E P L
w w p w 1 w

5 4 3 3 2 4
− 170L G J P − 240G J L P − 360L G J P z − 120G J L V + 960I E L P + 240L G J V
p 1 w 1

3 2 2 5 3 3
− 360L G J M + 720L I V E + 720L I E M + 660L I E P − 2160L I E M − 2160L I E P z
1 w 1 y 1 y y 1 y p

4 4 2 6
− 1632L V I E + 1632L P I E + 720L I E P z − 2880I E L P z − 2880I E L M − 480L I E P
1 y y y p w p w 1 y

4 4 5 5 2
+ 1440L I E M + 1440L I E P z + 1152L V I E − 1152L P I E + 2880I E L P z
y 1 y p 1 y y w p

## + 2880I E L M − 240I E P L + 80L G J P + 720L G J P z − 960I E L P + 720L G J M ⎤

2 4 6 4 3 4
w 1 w p w 1⎦

g =
1 ⎡(−360) L I V E + 24 L4 I E P − 72 L3 I E V + 72 L3 I E P + 60 L2 G J M
27 ⎣ w 1 y y 1 y 1
3
60 L (E Ix)
2 4 3
− 60 I E P L − 240I E L P + 720I E P z + 23 L G J P + 30 G J L P + 720I E M
w w w p w 1

3 5 3 2 3 2
+ 120I E P L − 16L G J P − 120L G J P z + 60L G J P z − 30G J L V + 480I E L P
w p p 1 w

3 2 5 3 3 4
− 120L G J M − 120L I E M − 48L I E P + 240L I E M + 240L I E P z + 144L V I E
1 y 1 y y 1 y p 1 y

## − 144L P I E − 120L I E P z − 1440I E L P z − 1440I E L M ⎤

4 2
y y p w p w 1⎦

138

g =
1 ⎛ 360L I V E + 24L4 I E P − 48L3 I E V + 48L3 I E P − 40L2 G J M +
28 ⎝ w 1 y y 1 y 1
2
120L E I ( x)
2 4 3
+ 60I E P L + 240I E L P − 720I E P z − 37L G J P − 60G J L P − 720I E M
w w w p w 1

3 5 3 2 3 2
− 120I E P L + 24L G J P + 200L G J P z − 40L G J P z + 60G J L V − 480I E L P
w p p 1 w

3 2 5 3 3 4
+ 200L GJ M − 40L I E M − 48L I E P + 80L I E M + 80L I E P z + 96L V I E
1 y 1 y y 1 y p 1 y

4 2
− 96 L P I E − 40 L I E P z + 1440I E L P z + 1440I E L M ⎞
y y p w p w 1⎠

g =
1 ⎡(−5) L4 G J P + 15G J L3 V − 15G J L3 P + 180L I V E + 3 L4 I E P
35 ⎣ 1 w 1 y
4
10L E I( x)
3 3 2 2 2
− 6 L I E V + 6 L I E P + 30L G J M + 30L G J P z + 360I E M − 30I E P L
y 1 y 1 p w 1 w

− 120I E L P + 360I E P z ⎤
w w p⎦

g
36
=
−1 ( )
⎡ −L4 G J P + 5 G J L3 V − 5 G J L3 P + 15L2 G J M + 15L2 G J P z
⎣ 1 1 p
3
10L E I( x)
4 3 3 2 2
+ 3 L I E P − 8 L I E V + 8 L I E P − 10L I E M − 10L I E P z + 180I E M
y y 1 y y 1 y p w 1

## + 60L I V E − 15I E P L − 60I E L P + 180I E P z ⎤

2
w 1 w w w p⎦

139
g =
1 ⎛180L I V E − 39L4 I E P + 84L3 I E V − 84L3 I E P + 90L2 GJ M
45 ⎝ w 1 y y 1 y 1
3
30L E I ( x)
2 4 3
− 30I E P L − 120I E L P + 360I E P z − 14L G J P − 45G J L P + 360I E M
w w w p w 1

3 5 4 3 2 3
+ 60I E P L + 28L GJ P + 90G J L P − 180L GJ P z + 90L GJ P z + 45GJ L V
w p p 1

2 4 3 2 2 5
+ 240I E L P − 90L G J V − 180L G J M − 360L I V E + 60L I E M + 48L I E P
1 w 1 1 w 1 y 1 y

3 3 4 4 2
− 120L I E M − 120L I E P z − 108L V I E + 108L P I E + 60 L I E P z − 720I E L P z
y 1 y p 1 y y y p w p

− 720I E L M ⎞
w 1⎠

g =
1 ⎡(−120) L I V E + 61L4 I E P − 144L3 I E V + 144L3 I E P − 90L2 GJ M
46 ⎣ w 1 y y 1 y 1
2
60L E I ( x)
2 4 3
+ 30I E P L + 120I E L P − 360I E P z + 6 L G J P + 30G J L P − 360I E M
w w w p w 1

3 5 4 3 2 3
− 60I E P L − 12L GJ P − 60GJ L P + 180L GJ P z − 90L GJ P z − 30GJ L V
w p p 1

2 4 3 2 2 5
− 240I E L P + 60L GJ V + 180L GJ M + 240L I V E − 160L I E M − 72L I E P
w 1 1 w 1 y 1 y

3 3 4 4 2
+ 200L I E M + 200L I E P z + 168L V I E − 168L P I E − 160L I E P z + 720I E L P z
y 1 y p 1 y y y p w p

+ 720I E L M ⎤
w 1⎦

140

−1 ⎡(−5) L4 G J P + 15G J L3 V − 15G J L3 P + 180L I V E + 3 L4 I E P
g =
57 ⎣ 1 w 1 y
4
10L E I( x)
3 3 2 2 2
− 6 L I E V + 6 L I E P + 30L G J M + 30L G J P z + 360I E M − 30I E P L −
y 1 y 1 p w 1 w

− 120I E L P + 360I E P z ⎤
w w p⎦

g =
1 ⎡(−8) L4 G J P + 25G J L3 V − 25G J L3 P + 50L2 G J M + 50L2 G J P z
58 ⎣ 1 1 p
3
20L E I( x)
4 3 3 2 2
+ 180L I V E − 6 L I E P + 14L I E V − 14L I E P + 20L I E M + 20L I E P z
w 1 y y 1 y y 1 y p

2
+ 360I E M − 30 I E P L − 120I E L P + 360I E P z ⎤
w 1 w w w p⎦

−1 ⎛ L4 G J P − 5 G J L3 V + 5 G J L3 P − 15L2 G J M − 15L2 G J P z
g =
67 ⎝ 1 1 p
3
10L E I( x)
4 3 3 2 2
− 3 L I E P + 8 L I E V − 8 L I E P + 10L I E M + 10L I E P z − 180I E M
y y 1 y y 1 y p w 1

2
− 60L I V E + 15I E P L + 60I E L P − 180I E P z ⎞
w 1 w w w p⎠

141
−1 ⎡(−5) L4 GJ P + 25GJ L3 V − 25GJ L3 P + 75L2 GJ M + 75L2 GJ P z
g =
68 ⎣ 1 1 p
2
( x)
60L E I

4 3 3 2 2
− 12L I E P + 26L I E V − 26L I E P + 30L I E M + 30L I E P z + 540I E M
y y 1 y y 1 y p w 1

2
w 1 w w w p⎦

142
APPENDIX B

## B.1 MATRIX [A] FROM SECTION 5.4

2 2
π K 1
A 11 = A 82 = +
32   32 8

2 2
81 π 81 K 9
A 23 = A 94 = +
32   32 8

2 2
625 π 625 K 25
A 35 = A 10.6 = +
32   32 8

2 2
2401 π 2401 K 49
A 47 = A 11.8 = +
32   32 8

2 2
6561 π 6561 K 81
A 59 = A 12.10 = +
32   32 8

2 2
14641 π 14641 K 121
A 6.11 = A 13.12 = +
32   32 8

2 2
28561 π 28561 K 169
A 7.13 = A 14.14 = +
32   32 8

143
*Only non-zero terms from Matrix [A] are listed.

144
B.1 MATRIX [B] FROM SECTION 5.4

1 3 5
B12 = − B22 =
16 2 2
4π   4π

−3 9 3
B14 = B24 = −
2 16 2
4π   4π

−35 5
B16 = B26 =
2 2
36 π   4π

−35 −91
B18 = B28 =
2 2
36 π   100 π

−99 −3
B1.10 = B2.10 =
2 2
100 π   4π

−99 −187
B1.12 = B2.12 =
2 2
100 π   196 π

−195 −91
B1.14 = B2.14 =
2 2
196 π   100 π

145

−11 13
B32 = B42 =
2 2
36 π   36 π

21 −51
B34 = B44 =
2 2
4π   100 π

25 3 45
B36 = − B46 =
16 2 2
4π   4π

21 49 3
B38 = B48 = −
2 16 2
4π   4π

−171 45
B3.10 = B4.10 =
2 2
196 π   4π

−11 −275
B3.12 = B4.12 =
2 2
36 π   324 π

−299 13
B3.14 = B4.14 =
2 2
324 π   36 π

146

−19 21
B52 = B62 =
2 2
100 π   100 π

5 −75
B54 = B64 =
2 2
4π   196 π

−115 85
B56 = B66 =
2 2
196 π   36 π

77 −203
B58 = B68 =
2 2
4π   324 π

81 3 117
B5.10 = − B6.10 =
16 2 2
4π   4π

77 121 3
B5.12 = B6.12 = −
2 16 2
4π   4π

−403 117
B5.14 = B6.14 =
2 2
484 π   4π

147

−27
B72 =
2
196 π

69
B74 =
2
100 π

−155
B76 =
2
324 π

133
B78 =
2
36 π

−315
B7.10 =
2
484 π

165
B7.12 =
2
4π

169 3
B7.14 = −
16 2
4π

148

B81 =
1

3 −7 β x
16 2 B88 =
4π 2
36 π

−β x βx −8019
B82 = + B89 =
16 2 2
4π   100π

−27 9βx
B83 = B8.10 =
2 2
4π   100 π

−3 β x −11979
B84 = B8.11 =
2 2
4π   100π

−875 −11 β x
B85 = B8.12 =
2 2
36π   100 π

5βx −32955
B86 = B8.13 =
2 2
36 π   196π

−1715 13 β x
B87 = B8.14 =
2 2
36π   196 π

149
5 21 β x
B91 = B98 =
2 2
36π   100 π

−3 β x −27
B92 = B99 =
2 2
4π   4π

B93 =
9

3 −3 β x
16 2 B9.10 =
4π 2
4π

−9 β x βx −22627
B94 = + B9.11 =
16 2 2
4π   1764π

125
B95 = 33 β x
2 B9.12 =
36π   2
196 π

−15 β x −15379
B96 = B9.13 =
2 2
4π   900π

−4459 −39 β x
B97 =
2 B9.14 =
900π   2
100 π

150

−11 −35 β x
B10.1 = B10.8 =
2 2
900π   4π

5βx −13851
B10.2 = B10.9 =
2 2
36 π   4900π

189 45 β x
B10.3 = B10.10 =
2 2
100π   196 π

−15 β x −1331
B10.4 = B10.11 =
2 2
4π   900π

B10.5 =
25

3 −55 β x
16 2 B10.12 =
4π 2
36 π

−25 β x βx −50531
B10.6 = + B10.13 =
16 2 2
4π   8100π

1029 65 β x
B10.7 = B10.14 =
2 2
100π   324 π

151

B11.1 =
13 −49 β x βx
2 B11.8 = +
1764π 16 2
4π

−7 β x 3645
B11.2 = B11.9 =
2 2
36 π   196π

−459 −63 β x
B11.3 = B11.10 =
2 2
4900π   4π

21 β x −33275
B11.4 = B11.11 =
2 2
100 π   15876π

1125 77 β x
B11.5 = B11.12 =
2 2
196π   324 π

−35 β x 2197
B11.6 = B11.13 =
2 2
4π   1764π

B11.7 =
49

3 −91 β x
16 2 B11.14 =
4π 2
36 π

152

−19 −63 β x
B12.1 = B12.8 =
2 2
8100π   4π

9βx 81 3
B12.2 = B12.9 = −
2 16 2
100 π   4π

B12.3 =
5 −81 β x βx
2 B12.10 = +
36π 16 2
4π

−3 β x 9317
B12.4 = B12.11 =
2 2
4π   324π

−2875 −99 β x
B12.5 = B12.12 =
2 2
15876π   4π

45 β x −68107
B12.6 = B12.13 =
2 2
196 π   39204π

3773 117 β x
B12.7 = B12.14 =
2
324π 2
484 π

153

21 77 β x
B13.1 = B13.8 =
2 2
12100π   324 π

−11 β x 9477
B13.2 = B13.9 =
2 2
100 π   484π

−675 −99 β x
B13.3 = B13.10 =
2 2
23716π   4π

33 β x 121 3
B13.4 = B13.11 = −
2 16 2
196 π   4π

B13.5 =
2125 −121β x βx
2 B13.12 = +
4356π 16 2
4π

−55 β x 19773
B13.6 = B13.13 =
2 2
36 π   484π

−9947 −143 β x
B13.7 = B13.14 =
2 2
39204π   4π

154

−27 −91 β x
B14.1 = B14.8 =
2 2
33124π   36 π

13 β x −25515
B14.2 = B14.9 =
2 2
196 π   81796π

621 117 β x
B14.3 = B14.10 =
2 2
16900π   484 π

−39 β x 19965
B14.4 = B14.11 =
2 2
100 π   676π

−3875 −143 β x
B14.5 = B14.12 =
2 2
54756π   4π

65 β x 169 3
B14.6 = B14.13 = −
2 16 2
324 π   4π

B14.7 =
6517 −169β x βx
2 B14.14 = +
6084π 16 2
4π

155

BIBLIOGRAPHY

WORKS CITED

## 1. Anderson, J. M. and Trahair, N.S. (1972). Stability of Monosymmetric Beams and

Cantilevers. Journal of the Structural Division, ASCE, 98(1), 269-285.

2. Arfken, G. and Weber, H. (2005). Mathematical Methods for Physicists, 6th ed.

## 3. Chajes, A. (1993). Principles of Structural Stability Theory. Englewood Cliffs, New

Jersey: Prentice-Hall.

## 4. Chen, W. F. and Lui, E. M. (1987). Structural Stability Theory and Implementation.

Upper Saddle River, New Jersey: Prentice-Hall.

## 5. Kitipornchai, S. and Trahair, N. S. (1980). Buckling Properties of Monosymmetric I-

Beams. Journal of the Structural Division. ASCE, 106(5), 941-957.

## 6. Kitipornchai, S. and Wang, C. M. (1989). New Set of Buckling Parameters for

Monosymmetric Beam-Column/Tie-Beams. Journal of Structural Engineering.
ASCE, 115(6), 1497-1513.

## 7. Lee, L. H. N. (1959). On the Lateral Buckling of a Tapered Narrow Rectangular

Beam. Journal of Applied Mechanics. 26, 457-458.

## 8. Love, A. E. H. (1944). A Treatise on the Mathematical Theory of Elasticity. 4th ed.,

Dover, New York, N. Y.

9. Mohri, F., Brouki, A., and Roth, J. C. (2003). Theoretical and Numerical Stability
Analyses of Unrestrained, Monosymmetric Thin-Walled Beams. Journal of
Constructional Steel Research. Elsevier, 59, 63-90.

10. Pi, Y. L. and Trahair, N. S. (1992a). Prebuckling Deflections and Lateral Buckling. I:
Theory. Journal of Structural Engineering, ASCE, 118(11), 2949-2966.

156
11. Pi, Y. L. and Trahair, N. S. (1992b). Prebuckling Deflections and Lateral Buckling.
II: Theory. Journal of Structural Engineering, ASCE, 118(11), 2967-2986.

12. Pilkey, W.D. and Wunderlich, W. (1994). Mechanics of Structures Variational and
Computational Methods. Boca Patom, Florida: CRC Press Inc.

## 13. Roberts, E. R. (2004). Elastic Flexural-Torsional Buckling Analysis Using Finite

Element Method and Object-Oriented Technology with C/C++. M.S. Thesis. School
of Engineering, Univ. of Pittsburgh.

## 14. Torkamani, M. A. M. (1998). Transformation Matrices for Finite and Small

Rotations. Journal of Engineering Mechanics, ASCE, 124(3), 359-362.

## 15. Trahair, N. S. and Nethercot, D. A. (1984) Bracing Requirements in Thin-Walled

Structures.

16. Vlasov, V. Z. (1961). Thin-walled Elastic Beams (2nd ed.). Jerusalem, Israel: Israel
Program for Scientific Translation.

## 17. Wang, C.M., and Kitipornchai, S. (1986). On Stability of Monosymmetric

Cantilevers. Engineering Structures, ASCE, 8, 169-180.

18. Wang, C.M., Wang, C.Y., and Reddy, J.N. (2005). Exact Solutions for Buckling of
Structural Members. Boca Raton, Florida: CRC Press LLC.

157

WORKS CONSULTED

## 1. Andrade, A. and Camotim, D. (2005). Lateral-Torsional Buckling of Singly

Symmetric Tapered Beams: Theory and Applications. Journal of Engineering
Mechanics, ASCE, 586-597.

## 2. Bradford, M. A. (1988). Elastic Buckling of Tapered Monosymmetric I-Beams.

Journal of Structural Engineering, ASCE, 114(5), 977-992.

## 3. Boyce, W. E. and DiPrima, R. C. (2001). Elementary Differential Equations and

Boundary Value Problems. New York, New York: John Wiley & Sons, Inc.

## 4. Clark, J. W. and Hill, H. N. (1960). Lateral Buckling of Beams. Journal of the

Structural Division, ASCE, 86(7), 175-196.

5. Clark, J. W. and Hill, H. N. (1962). Lateral Buckling of Beams and Girders. Civil
Engineering Transactions, ASCE, 127(7), 180-201.

## 6. Dontree, P. (1994). Elastic Flexural-Torsional Buckling Analysis Using Object-

Oriented Technology and C/C++. M.S. Thesis. School of Engineering, Univ. of
Pittsburgh

## 7. Kitipornchai, S. and Trahair, N. S. (1975). Elastic Behavior of Tapered

Monosymmetric I-Beams Under Moment Gradient. Journal of the Structural
Division, ASCE, 101(8), 1661-1678.

## 8. Kitipornchai, S. and Trahair, N. S. (1972). Elastic Stability of Tapered I-Beams.

Journal of the Structural Division. ASCE, 98(3), 713-728.

## 9. Kitipornchai, S., Wang, C. M. and Trahair, N. S. (1986). Buckling of Monosymmetric

I-Beams Under Moment Gradient. Journal of Structural Engineering. ASCE, 112(4),
781-787.

10. Mohri, F., Brouki, A., and Roth, J. C. (2003). Theoretical and Numerical Stability
Analyses of Unrestrained, Monosymmetric Thin-Walled Beams. Journal of
Constructional Steel Research. Elsevier, 59, 63-90.
158
11. Pi, Y. L., Trahair, N. S. and Rajasekaran, S. (1991). Energy Equation for Beam
Lateral Buckling. Research Report No. R635, School of Civ. And Mining Engrg.,
Univ. of Sydney, Australia.

12. Powell, G. and Klingner, R. (1970). Elastic Lateral Buckling of Steel Beams. Journal
of the Structural Division, ASCE, 96(6), 1919-1932.

## 13. Roberts, T. M. and Burt, C. A. (1984). Instability of Monosymmetric I-Beams and

Cantilevers. , ASCE, , 313-324.

14. Tedesco, J. W., McDougal, W. G., and Ross, C. A. (1999). Structural Dynamics:
Theory and Applications. Boston, MA: Addison-Wesley.

15. Tehe, Lip H. (2005). Spatial Rotation Kinematics and Flexural-Torsional Buckling.
Journal of Engineering Mechanics, ASCE, 131(6), 598-605.

## 16. Trahair, N. S. (1993). Flexural-Torsional Buckling of Structures. Boca Raton,

Florida: CRC Press.

17. Vacharajittiphan, P., and Trahair, N.S. (1974). Effect of In-plane Deformation on
Lateral Buckling. Journal of Structural Mechanics, ASCE, 3(1), 29-60.

## 18. Wang, C.M., and Kitipornchai, S. (1986). On Stability of Monosymmetric

Cantilevers. Engineering Structures, ASCE, 8, 169-180.

19. Weaver, W. Jr., and Gere, J.M. Matrix Analysis of Framed Structures. 3rd edition.
New York: Van Nostrand Reinhold.

20. Zhang, L. and Tong, G.E. (2008). Elastic Flexural-Torsional Buckling of Thin-
Walled Cantilevers. Thin-Walled Structures, Elsevier, 46, 27-37.

159