Structural Mechanics

Structural Mechanics
Column Behaviour

2008/9

Dr. Colin Caprani,

1 Dr. C. Caprani
Structural Mechanics

Contents
1. Introduction ......................................................................................................... 3
1.1 Background...................................................................................................... 3
1.2 Stability of Equilibrium ................................................................................... 4
2. Buckling Solutions............................................................................................... 6
2.1 Introduction...................................................................................................... 6
2.2 Pinned-Pinned Column.................................................................................... 7
2.3 Column with Initial Displacements ............................................................... 18
2.4 The Effective Length of Columns ................................................................. 30
3. Column Design................................................................................................... 32
3.1 Background to BS5950.................................................................................. 32
3.2 Column Design Examples ............................................................................. 38
4. Appendix ............................................................................................................ 47
4.1 Solutions to Differential Equations ............................................................... 47
4.2 Code Extracts................................................................................................. 51
4.3 Past Exam Questions ..................................................................................... 62

2 Dr. C. Caprani
Structural Mechanics

1. Introduction

1.1 Background
In the linear elastic analysis of structures, we have assumed that compression
members are limited in load capacity in the same way that tension members are, by
ensuring the yield stress of the material is not exceeded. However, as can easily be
checked with a ruler, compression members often fail long before the material yields
due to buckling. So our problem is to identify reduced stress limits that should apply
for compression members so that buckling does not occur.

The first person to study this problem was Euler (‘oil-er’) as a means to demonstrate
his ability to solve differential equations. Some of the important results in buckling
retain his name.

Leonhard Euler (1707 – 1783)

3 Dr. C. Caprani
Structural Mechanics

1.2 Stability of Equilibrium

A structure will be in an initial equilibrium position. The stability of its equilibrium
can be assessed by examining the structure’s behaviour in an adjacent position. There
are three states:
• Stable equilibrium: the structure tends to return to its initial position. This is
the best situation to have structures in.
• Neutral (or critical) equilibrium: the structure moves to a displaced
configuration and remains in that position. This does not make for good
structure.
• Unstable equilibrium: any movement from the initial position causes further
movement resulting in a ‘runaway’ failure of the structure.

4 Dr. C. Caprani
Structural Mechanics

In relation to columns, the stability of equilibrium takes the form:

• Stable: deflections to not result in extra bending moments, and hence extra
deflections.
• Critical P = Pcr : the load is at a critical value where the column remains in any
displaced position.
• Unstable P > Pcr : the load is greater than the critical load and so divergent
displacements occur, leading to failure.

These situations look like this:

5 Dr. C. Caprani
Structural Mechanics

2. Buckling Solutions

2.1 Introduction
A perfect column (perfectly straight) is one which is perfectly straight and so carries
axial load up to the yield stress of the material. Since in reality columns are not
perfectly straight, buckling occurs:

In our solutions for buckling, we will find that both the perfectly straight and buckled
profiles are both possible theoretically. However, since it is the real behaviour that is
of interest, we will focus on the buckled solutions.

6 Dr. C. Caprani
Structural Mechanics

2.2 Pinned-Pinned Column

Formulation

Firstly, consider the buckled configuration of a pin-ended column and draw a free
body diagram of part of the column:

From this we see:

M = − Py

7 Dr. C. Caprani
Structural Mechanics

So for equilibrium:

M + Py = 0

We know from Euler-Bernoulli bending theory that:

d2y
M = EI 2
dx

And so we have:

d2y
EI + Py = 0 (1)
dx 2

Dividing across by EI gives:

d2y P
+ y=0
dx 2 EI

If we make the substitution:

P
k2 = (2)
EI

We then have:

d2y
2
+ k2y = 0 (3)
dx

8 Dr. C. Caprani
Structural Mechanics

This is a second-order linear homogenous differential equation in y. We seek a

solution for y which will be some function of x. The Appendix shows that the general
solution to this equation is:

y = A cos kx + B sin kx (4)

where A and B are constants to be evaluated from the boundary conditions of the
problem.

9 Dr. C. Caprani
Structural Mechanics

Relevant Solution

To get the particular solution to our problem, we know that we have no deflection at
the pinned end, that is:

At x = 0, y = 0

Substituting this into equation (4):

0 = A cos k 0 + B sin k 0

Since cos ( 0 ) = 1 and sin ( 0 ) = 0 , we have:

0 = A (1) + B ( 0 )
A=0

Thus equation (4) becomes:

y = B sin kx (5)

Using the second boundary condition, at x = L, y = 0 , we have:

0 = B sin kL (6)

There are two possibilities now. The first is B = 0 which makes y = 0 by equation (5)
. This means that a possible solution is for no buckling to occur, in other words, the
perfect column. Since we know that this is highly unlikely, and that buckling doesn’t
occur, we must consider the other possibility from equation (6):

10 Dr. C. Caprani
Structural Mechanics

sin kL = 0 (7)

We know that this only happens at values of:

kL = 0, π , 2π ,3π ,...
= nπ

where n = 0,1,2,3,... . Therefore we have:

nπ
k= (8)
L

So from equation (2) we have:

P n 2π 2
k =
2
= 2
EI L

And so the critical loads at which the column buckles are:

n 2π 2 EI
Pcr = (9)
L2

Further, by using equation (8) in equation (5) the buckled shape is got as:

nπ
y = B sin x (10)
L

11 Dr. C. Caprani
Structural Mechanics

Euler Buckling Load

Since we are interested in the lowest load that the column will buckle at, we use the
value n = 1 to find the Euler Buckling Load, PE , as:

π 2 EI
PE = (11)
L2

And we also find the displaced shape from equation (10) as:

π
y = B sin x (12)
L

This defines a half sine-wave curve as being the buckled shape of the column. Notice
that we have no information about B, the amplitude of the displacement. This is
because the column is in neutral equilibrium at PE and will be in equilibrium at any
displacement amount.

12 Dr. C. Caprani
Structural Mechanics

Other Buckling Modes

In general we can see that the column can buckle in the shapes:

n 0 1 2 3 …
Critical
π 2 EI 4π 2 EI 9π 2 EI n 2π 2 EI
Load Infinite PE =
L2 L2 L2 L2
Pcr =
Mode
π 2π 3π nπ
Shape 0 y = B sin x B sin x B sin x B sin x
L L L L
y=
Plot

13 Dr. C. Caprani
Structural Mechanics

However, to achieve these other buckling loads, the lower modes must be prevented
from occurring by lateral restraints:

14 Dr. C. Caprani
Structural Mechanics

Critical Stress

For design, we are interested in the stress that the material undergoes at the time of
buckling – the critical stress, σ cr :

PE π 2 EI
σ cr = = 2 (13)
A LA

Looking at the factor I A , we see that it is a property of the shape of the cross

section, and is in units of [ length ] [length ] = [length ] . Therefore, we define a new

4 2 2

geometric property, r, called the Radius of Gyration as:

I
r= (14)
A

And so r has units of length. The radius of gyration can be thought of as a distance
from the centroid at which the area of the cross section is concentrated for calculating
the second moment of area, I, since by (14),

I = Ar 2 (15)

The critical stress can now be expressed as:

π 2 Er 2
σ cr =
L2

And we can see that the dimensional properties of the column are summed up by the
factor r 2 L2 , which represents a ratio of r to L. Thus we define the Slenderness
Ratio, λ , as:

15 Dr. C. Caprani
Structural Mechanics

L
λ= (16)
r

Finally then, we have the equation for critical stress as:

π 2E
σ cr = 2 (17)
λ

A plot of the critical stress against slenderness is called a strut curve and looks like:

16 Dr. C. Caprani
Structural Mechanics

As can be seen, at low slendernesses (that is short stocky columns), the critical stress
(to cause bucking) reaches very high values. Since the maximum stress in the
material is the yield stress, me must cap the curve at σ y .

Finally, notice that typical experimental results fall below the Euler strut curve. This
is because the theory examined so far is for perfectly straight columns that have
somehow begun to buckle. In real columns there will be some initial imperfections
which have the effect of reducing the strength of the column. These initial
imperfections can be represented by an initial displacement curve.

17 Dr. C. Caprani
Structural Mechanics

2.3 Column with Initial Displacements

Problem Formulation

The imperfections in the manufacture of real columns mean that an initial

displacement curve exists in the column, prior to loading. Since any curve can be
represented by a Fourier series expansion, we will approximate the initial displaced
shape by the first term of a Fourier series – a sine curve, the equation of which is:

πx
y0 ( x ) = a sin (18)
L

where, at the midpoint of the column, the initial displacement is a.

18 Dr. C. Caprani
Structural Mechanics

Considering a free-body diagram as before:

gives the equilibrium equation as:

d2y
EI 2 + P ( y + y0 ) = 0
dx
Thus we have:

d2y P P
2
+ y+ y0 = 0
dx EI EI

Using equation (2) gives:

19 Dr. C. Caprani
Structural Mechanics

d2y
2
+ k 2 y + k 2 y0 = 0
dx

And so using equation (18) we have

d2y πx
+ k 2
y = − k 2
a sin (19)
dx 2 L

This is a non-homogenous second order differential equation.

20 Dr. C. Caprani
Structural Mechanics

Solution

The solution to non-homogenous differential equations is made up of two parts:

• The complimentary solution (denoted yC ): this is the solution to the
corresponding homogenous equation. That is, the solution when the right hand
side is zero. We have this from before as equation (4):

yC = A cos kx + B sin kx (20)

• The particular solution (denoted yP ): for the function on the right hand side,
the solution is verified in the Appendix as:

−k 2a πx
yP = sin
⎛π ⎞
2
L
−⎜ ⎟ + k 2
⎝L⎠ (21)
k 2a πx
= 2 sin
π L
2
− k2
L

Thus the total solution is:

y = yC + yP
k 2a πx
= A cos kx + B sin kx + sin (22)
π 2
L
2
− k2
L

To find the constants, we know that for the pinned-pinned column, y = 0 at x = 0 :

21 Dr. C. Caprani
Structural Mechanics

k 2a π
y ( 0 ) = A cos k ( 0 ) + B sin k ( 0 ) + sin ( 0)
π 2
L
−k 2

L2
0= A

Also, y = 0 at x = L , giving:

k 2a π
y ( L ) = B sin k ( L ) + sin ( L)
π 2
L
−k 2

L2
0 = B sin kL

Although this is the same equation as found for the perfectly straight column, we
must consider the implications. If B ≠ 0 then sin kL = 0 and so kL = π as before. This
yields k = π L , or k 2 = π 2 L . Substituting this into equation (22) means that the
2

third term is infinite and so the deflection is infinite. Since this is impossible for a
stable column with P < Pcr , we conclude that B = 0 and we are left with:

k 2a πx
y= sin (23)
π 2
L
−k 2

L2

This equation represents the deflections of the column caused by the loading. The
total deflection will be that caused by the loading, in addition to the initial
imperfection deflection curve:

ytot = y + y0

And so:

22 Dr. C. Caprani
Structural Mechanics

k2 πx πx
ytot = a sin + a sin
π 2
L L
2
− k2
L

Solving out, and dropping the tot subscript on y gives::

⎛ k2 ⎞ πx
y =⎜ 2 2 + 1 ⎟ a sin
⎝π L −k
2
⎠ L

And so:

⎛ π 2 L2 ⎞ πx
y =⎜ 2 2 2 ⎟
a sin (24)
⎝π L −k ⎠ L

Now, using the expression for PE (equation (11)), we have:

π2 PE
=
L2 EI

And with the expression for k 2 (equation (2)), equation (24) becomes:

PE EI πx
y= a sin
PE EI − P EI L

And so we have:

⎡ P ⎤ πx
y = ⎢ E ⎥ a sin (25)
⎣ PE − P ⎦ L

23 Dr. C. Caprani
Structural Mechanics

The term in brackets thus amplifies the initial deflection, depending on how close we
are to the critical buckling load. A plot of load against deflection shows:

24 Dr. C. Caprani
Structural Mechanics

Maximum Stress Consideration

At the mid-height of the column, the deflection will be largest, and thus so will the
bending moment. The deflection at the mid-height is got from equation (25), with
x = L 2:

⎛L⎞ ⎡ P ⎤ π ⎛L⎞
y ⎜ ⎟ = ⎢ E ⎥ a sin ⎜ ⎟
⎝ 2 ⎠ ⎣ PE − P ⎦ L⎝ 2⎠
⎡ P ⎤
= ⎢ E ⎥a
⎣ PE − P ⎦

We can equally interpret this equation in terms of stresses by dividing each of the Ps
by A:

⎛ L ⎞ ⎡ σE ⎤
y⎜ ⎟ = ⎢ ⎥a (26)
⎝ ⎠ ⎣ E
2 σ − σ ⎦

where σ E is the stress associated with the critical Euler load (equation (13)).

Consider again the free-body diagram of the column from mid-height to pin. There
are two sources of stress:

1. The stresses due to the moment alone are:

Mz
σ Moment =
I

where z is the distance from the neutral axis of the fibre under consideration.

25 Dr. C. Caprani
Structural Mechanics

2. The stresses due to the axial force are:

P
σ Axial =
A

Thus the stresses at any point are given by:

P Mz
σ= + (27)
A I

Superposition of the stress diagrams shows this:

The maximum stress is on the outside face and is thus:

P Mc
σ max = + (28)
A I

where c is the distance from the neutral axis to the inside face of the column.

26 Dr. C. Caprani
Structural Mechanics

Next, into equation (28), we introduce the relevant properties of equation (15) and the
fact that M = Py ( L 2 ) yields:

Pcy ( L 2 )
σ max = σ +
Ar 2

But, from equation (26) we know the displacement at mid height, y ( L 2 ) :

Pc ⎡ σ E ⎤
σ max = σ + ⎢ ⎥a
Ar 2 ⎣ σ E − σ ⎦

At failure the maximum stress is the yield stress, σ y . The stress associated with the

load P when this occurs is σ cr . Hence the governing equation becomes:

c ⎡ σE ⎤
σ y = σ cr + σ cr ⎢ ⎥a
r 2 ⎣ σ E − σ cr ⎦

Giving:

⎡ ac ⎛ σ E ⎞⎤
σ y = σ cr ⎢1 + ⎜ ⎟⎥ (29)
⎣ r 2 ⎝ σ E − σ cr ⎠⎦

We are looking to find the value of σ cr that solves this equation. At the load
corresponding to σ cr failure occurs. As can be seen, this failure stress is a function of
the section (through r and c) and the initial imperfection, a, as well as the usual Euler
buckling load for the column (through σ E ).

27 Dr. C. Caprani
Structural Mechanics

Critical Stress for Buckling

To solve equation (29) for σ cr we proceed as follows:

⎡ ac ⎛ σ E ⎞ ⎤
σ y = σ cr ⎢1 + 2 ⎜ ⎟⎥
⎣ r ⎝ σ E − σ cr ⎠ ⎦
σ y r 2 (σ E − σ cr ) = σ cr r 2 (σ E − σ cr ) + acσ crσ E
σ yσ E r 2 − σ yσ cr r 2 = σ crσ E r 2 − σ crσ cr r 2 + acσ crσ E
σ yσ E r 2 − σ yσ cr r 2 − σ crσ E r 2 + σ crσ cr r 2 − acσ crσ E = 0
σ cr2 ( r 2 ) + σ cr ( −σ y r 2 − σ E r 2 − acσ E ) + σ yσ E r 2 = 0
⎛ ac ⎞
σ cr2 + σ cr ⎜ −σ y − σ E − σ E ⎟ + σ yσ E = 0
⎝ r2 ⎠

We call the parameter that accounts for the initial imperfections called the Perry
Factor:

ac
η= (30)
r2

And this gives:

σ cr2 + σ cr ⎡⎣ −σ y − σ E (1 + η ) ⎤⎦ + σ yσ E = 0 (31)

This is a quadratic equation in σ cr and so is solved by the usual:

−b ± b 2 − 4ac
x=
2a

where:

28 Dr. C. Caprani
Structural Mechanics

a =1
b = ⎡⎣ −σ y − σ E (1 + η ) ⎤⎦
c = σ yσ E

And this gives:

0.5
σ y + σ E (1 + η ) ⎧⎪ ⎡ σ y + σ E (1 + η ) ⎤ ⎫⎪
2

σ cr = − ⎨⎢ ⎥ − σ σ
y E⎬
(32)
⎩⎪ ⎣ ⎦
2 2
⎭⎪

Notice that we have chosen the lower root of the two possible solutions.

Equation (32) is called the Perry-Robertson formula and it gives the buckling stress
in terms of the yield stress and initial imperfections of the column, as well as its Euler
buckling load.

It is useful to introduce the following:

σ y + σ E (1 + η )
φ= (33)
2

And so the Perry-Robertson formula (equation (32)) becomes:

σ cr = φ − φ 2 − σ yσ E (34)

29 Dr. C. Caprani
Structural Mechanics

2.4 The Effective Length of Columns

So far we have only considered pinned-end columns but clearly other end conditions
are possible. Analysis of the buckling loads for such columns can be carried out along
the same lines as for pinned-end columns. However, a significant advantage can be
got by remembering that a point of contraflexure (zero moment) behaves as a pin
(zero moment). The distance between points of contraflexure can be considered as a
pin-pin column, but with a smaller length than the overall column. We call this length
the effective length, LE . Analysis then proceeds as for pinned-end columns, but
using the effective, rather than actual, length.

The effective lengths of some typical columns are:

Non-Sway Modes

Pin-Pin Fix-Pin Fix-Fix

30 Dr. C. Caprani
Structural Mechanics

Sway Modes

Fix-fix Fix-roller Cantilever (fix-free)

Notice from the above that the locations of the points of contraflexure do not have to
be and can be located outside the column. That is, the column is buckling over a
notional length of LE .

The effective length only affects the slenderness and so the general case for
slenderness is:

LE
λ= (35)
r

31 Dr. C. Caprani
Structural Mechanics

3. Column Design

3.1 Background to BS5950

Initial Imperfections

Robertson performed many tests on struts to arrive at a suitable value for the initial
imperfections in the approach outlined in the previous section. He suggested:

η = 0.003λ (36)

where λ is the slenderness of the column, given by equation (16). More recently, the
initial imperfection has been taken as:

⎛ λ ⎞
2

η = 0.3 ⎜ ⎟ (37)
⎝ 100 ⎠

The idea of linking the initial imperfections to the slenderness is intuitively appealing
– the slimmer a column is, the more likely it is to have imperfections.

The steel design code BS5950 is based on the following:

a ( λ − λ0 )
η= (38)
1000

In which:
• a is the Robertson constant (and is not the same as the a we had for the
deflection of the column previously);
• λ0 is called the limiting slenderness as is given by:

32 Dr. C. Caprani
Structural Mechanics

π 2E
λ0 = 0.2 (39)
σy

By rearranging equation (17) it can be seen that this is:

λ0 = 0.2λcr

where λcr is the slenderness at which the Euler stress reaches the yield stress of the
material.

As can be seen, the higher the value of a, the more initial imperfection is accounted
for and the compressive strength reduces as a result.

33 Dr. C. Caprani
Structural Mechanics

Code Expressions

In BS5950, the Perry-Robertson formula is given in a slightly different form to that

presented in equation (34). To get the code expression, we multiply top and bottom of

equation (34) by φ + φ 2 − σ yσ E to get:

⎡φ − φ 2 − σ yσ E ⎤ ⎡φ + φ 2 − σ yσ E ⎤
σ cr = ⎣ ⎦⎣ ⎦
φ + φ − σ yσ E
2

And multiplying out gives:

φ 2 − φ 2 + σ yσ E
σ cr =
φ + φ 2 − σ yσ E
σ yσ E
=
φ + φ 2 − σ yσ E

Lastly, to get the code expression, we must use the code notation which is:

pC ≡ σ cr
py ≡ σ y
pE ≡ σ E

So finally we have the expression in Appendix C of BS5950:

pE p y
pc = (40)
φ + φ 2 − pE p y

34 Dr. C. Caprani
Structural Mechanics

Strut Curves

BS5950 provides four values of the Robertson constant that may be used in design. It
also specifies what value of a to use for the various types of steel section and the axis
about which buckling may occur. The values are:

• a = 2.0 - strut curve (a);

• a = 3.5 - strut curve (b);
• a = 5.5 - strut curve (c);
• a = 8.0 - strut curve (d).

Given the initial imperfection model of equation (38) as well as the Perry-Roberston
formula, equation (40), we can plot the four strut curves used in the code:

400
Euler
350 a=2
a = 3.5
Compression Strength (N/mm2)

300 a = 5.5
a=8
250

200

150

100

50

0
0 20 40 60 80 100 120 140 160 180 200
Slenderness λ

Also shown in this plot are the limiting slenderness, and the critical slenderness,
discussed in relation to equation (38).

35 Dr. C. Caprani
Structural Mechanics

The code provides four tables (Table 24(a) to 24(d)) – corresponding to the strut
curves, which are formatted as follows:

py
λ 225 . . . 275
15 . . . . .
.
. Values of pc from equation (40)
.
350 . . . . .

Table 23 of the code allocates the strut curves to different section types and axes:

36 Dr. C. Caprani
Structural Mechanics

Effective Lengths

Table 22 of the code specifies the appropriate effective lengths for columns with
different end conditions, end conditions that occur in practice. The meanings of the
phrases in Table 22 are as follows:

Term Diagram
Effectively held in position, not
restrained in direction.
(pin)

Effectively held in position,

partially restrained in direction.
(rotational spring)

Effectively held in position and

restrained in direction.
(fixed)

Not held in position, and

effectively restrained in direction.
(fixed with sway)

Not held in position, partially

restrained in direction.
(rotational spring with sway)

Not held in position or restrained

in direction.
(free)

37 Dr. C. Caprani
Structural Mechanics

3.2 Column Design Examples

Example 1

Problem
A 5.6 m high column consists of a 203 × 203 × 46 UC section. It is supported along
its x-axis and is pinned at both ends. Find the buckling load.

Solution
Firstly, sketch the column:

38 Dr. C. Caprani
Structural Mechanics

Since the column is supported along its x-x axis, it can only buckle about its y-y axis.
The relevant section properties for a 203 × 203 × 46 UC are:

Cross-Sectional Area 58.8 cm2

Yield stress 275 N/mm2
Modulus of Elasticity 205 kN/mm2
Radius of gyration about the y-y axis 5.12 cm
Robertson Constant for the y-y axis 5.5

Thus:

5600
λ= = 109.6 ≈ 110
51.1

From Table 23 we see that we are using strut curve (c) and so a = 5.5 . Also,
E = 205 kN/mm 2 and p y = 265 N/mm 2 from Table 9. Thus:

The limiting slenderness (equation (39)) is:

π 2E π 2 ⋅ 205 × 103
λ0 = 0.2 = 0.2 = 17.48
py 265

The Perry Factor (equation (38)) is:

a ( λ − λ0 ) 5.5 (110 − 17.48 )

η= = = 0.509
1000 1000

The Euler stress (equation (17) is:

39 Dr. C. Caprani
Structural Mechanics

π 2 E π 2 ⋅ 205 × 103
pE = = = 168 N/mm 2
λ 2
110 2

The modifying stress (equation (33)) is:

p y + (η + 1) pE 265 + ( 0.509 + 1)168

φ= = = 259 N/mm 2
2 2

And so the compressive strength (equations (34)or (40)):

pE p y 168 ⋅ 265
pc = = = 108.6 N/mm 2
φ + φ − pE p y
2
259 + 259 − 168 ⋅ 265
2

Thus the buckling load is:

5880 ⋅ 108.6
P = Ag pc = = 640.4 kN
103

To check this, use Table 24(c), for λ ≥ 110 and p y = 265 N/mm 2 gives:

pc = 108 N/mm 2 and so the capacity is 108 ⋅ 5880 103 = 635 kN , which is similar to
the previous calculation.

40 Dr. C. Caprani
Structural Mechanics

Example 2

Problem
For the column of Example 1, the restraint along the x-x axis has to be removed.
Determine the buckling capacity.

Solution
Again, sketch the column:

41 Dr. C. Caprani
Structural Mechanics

Since the column can now buckle about both its x-x axis and y-y axis, we will need to
determine the buckling capacity of each axis. Of course the buckling capacity about
the y-y axis has already been determined from Example 1, so it remains to find the
capacity about the x-x axis. The relevant section properties are:

Cross-Sectional Area 58.8 cm2

Yield stress 275 N/mm2
Modulus of Elasticity 205 kN/mm2
Radius of gyration about the x-x axis 8.81 cm
Robertson Constant for the x-x axis 3.5

Thus:

5600
λ= = 63.6
88.1

The limiting slenderness is the same as in Example 1 since E nor p y change:

λ0 = 17.48

The Perry Factor (equation (38)) is:

a ( λ − λ0 ) 3.5 ( 63.6 − 17.48 )

η= = = 0.16
1000 1000

The Euler stress (equation (17) is:

π 2 E π 2 ⋅ 205 × 103
pE = = = 500.8 N/mm 2
λ 2
63.6 2

42 Dr. C. Caprani
Structural Mechanics

The modifying stress (equation (33)) is:

p y + (η + 1) pE 265 + ( 0.16 + 1) 500.8

φ= = = 423.3 N/mm 2
2 2

And so the compressive strength (equations (34)or (40)):

pE p y 500.8 ⋅ 265
pc = = = 207.7 N/mm 2
φ + φ − pE p y
2
423.3 + 423.3 − 500.8 ⋅ 265
2

Thus the load to cause buckling about the x-x axis is:

5880 ⋅ 207.7
P = Ag pc = = 1221.5 kN
103

Check this value using Table 24.

Since the buckling capacities are:

• x-x axis: P = 1221.5 kN ;
• y-y axis P = 640.4 kN ;
the column will first buckle about the y-y axis and so its overall buckling capacity is:

P = 640.4 kN

43 Dr. C. Caprani
Structural Mechanics

Example 3

Problem
To increase the capacity of the column of Example 2, the supports in the y-y axis
have been changed to fixed-fixed. Determine the buckling capacity.

Solution
As always, sketch the column:

44 Dr. C. Caprani
Structural Mechanics

We know the buckling capacity about the x-x axis from Example 2, but since the
support conditions for the y-y axis have changed, the buckling capacity about the y-y
axis changes. Again the relevant properties are:

Cross-Sectional Area 58.8 cm2

Yield stress 275 N/mm2
Modulus of Elasticity 205 kN/mm2
Radius of gyration about the y-y axis 5.12 cm
Robertson Constant for the y-y axis 5.5

In this case, since the restraints are fixed-fixed, the effective length is:

LE = 0.7 L = 0.7 ⋅ 5600 = 3920 mm

Thus the slenderness is:

3920
λ= = 76.6
51.2

The limiting slenderness is the same as before, λ0 = 17.48 . The Perry Factor
(equation (38)) is:

a ( λ − λ0 ) 5.5 ( 76.6 − 17.48 )

η= = = 0.32
1000 1000

The Euler stress (equation (17) is:

π 2 E π 2 ⋅ 205 × 103
pE = 2 = = 345.2 N/mm 2
λ 76.6 2

45 Dr. C. Caprani
Structural Mechanics

The modifying stress (equation (33)) is:

p y + (η + 1) pE 265 + ( 0.32 + 1) 345.2

φ= = = 361.2 N/mm 2
2 2

And so the compressive strength (equations (34)or (40)):

pE p y 345.2 ⋅ 265
pc = = = 163.8 N/mm 2
φ + φ − pE p y
2
361.2 + 361.2 − 345.2 ⋅ 265
2

Thus the load to cause buckling about the y-y axis is thus:

5880 ⋅ 163.8
P = Ag pc = = 962.9 kN
103

Check this value using Table 24.

Since the buckling capacities are:

• x-x axis: P = 1221.5 kN - still as per Example 2;
• y-y axis P = 962.9 kN - changed due to new support conditions;
thus, the column will first buckle about the y-y axis and so its overall buckling
capacity is:

P = 962.9 kN

Notice that a change in support conditions has resulted in a nearly 50% increase in
capacity.

46 Dr. C. Caprani
Structural Mechanics

4. Appendix

4.1 Solutions to Differential Equations

The Homogenous Equation

To find the solution of:

d2y
2
+ k2y = 0 (41)
dx

we try y = eλ x (note that this λ has nothing to do with slenderness but is the
conventional mathematical notation for this problem). Thus we have:

dy d2y
= λ eλ x ; 2
= λ 2eλ x
dx dx

Substituting this into (41) gives:

λ 2eλ x + k 2eλ x = 0

And so we get the characteristic equation by dividing out eλ x :

λ2 + k2 = 0

From which:

λ = ± −k 2

Or,

47 Dr. C. Caprani
Structural Mechanics

λ1 = +ik ; λ2 = −ik

Where i = −1 . Since these are both solutions, they are both valid and the expression
for y becomes:

y = A1eikx + A2e − ikx (42)

In which A1 and A2 are constants to be determined from the initial conditions of the
problem. Introducing Euler’s equations:

eikx = cos kx + i sin kx

(43)
e − ikx = cos kx − i sin kx

into (42) gives us:

y = A1 ( cos kx + i sin kx ) + A2 ( cos kx − i sin kx )

Collecting terms:

y = ( A1 + A2 ) cos kx + ( iA1 − iA2 ) sin kx

Since the coefficients of the trigonometric functions are constants we can just write:

y = A cos kx + B sin kx (44)

Which is the solution presented in equation (4).

48 Dr. C. Caprani
Structural Mechanics

Particular Solution of Non-homogenous Equation

We seek here to show the particular solution to the problem:

d2y
2
+ k 2 y = A sin λ x (45)
dx

Is given by:

y = B sin λ x (46)

From (46):

dy d2y
= λ B cos λ x; 2
= −λ 2 B sin λ x
dx dx

Substitution into (45) gives:

−λ 2 B sin λ x + k 2 B sin λ x = A sin λ x (47)

Dividing out the common term:

−λ 2 B + k 2 B = A (48)

And so:

A
B= (49)
−λ + k 2
2

49 Dr. C. Caprani
Structural Mechanics

Thus (46) becomes:

A
y= sin λ x (50)
−λ + k 2
2

Looking at equation (19), we see that:

A ≡ −k 2 a
π
λ≡
L

And so the particular solution of equation (19) becomes:

−k 2 a
y= sin λ x (51)
⎛π ⎞
2

−⎜ ⎟ + k 2
⎝L⎠

Which is given previously as equation (21).

50 Dr. C. Caprani
Structural Mechanics

4.2 Code Extracts

51 Dr. C. Caprani
Structural Mechanics

52 Dr. C. Caprani
Structural Mechanics

53 Dr. C. Caprani
Structural Mechanics

54 Dr. C. Caprani
Structural Mechanics

55 Dr. C. Caprani
Structural Mechanics

56 Dr. C. Caprani
Structural Mechanics

57 Dr. C. Caprani
Structural Mechanics

58 Dr. C. Caprani
Structural Mechanics

59 Dr. C. Caprani
Structural Mechanics

60 Dr. C. Caprani
Structural Mechanics

61 Dr. C. Caprani
Structural Mechanics

4.3 Past Exam Questions

Semester 1 Paper 2007/8

(a) Briefly explain how to calculate the second moment of area of a section comprised of simple shapes. (3 marks)
(b) Describe what is meant by complimentary shear stresses. (4 marks)
(c) Determine the maximum load w that can be sustained by the column BD shown in Fig. Q.4. (18 marks)
The relevant section properties are given in Table Q4 and the Perry-Robertson formula is given.

w kN/m

A B PINNED
C B
CONNECTION

3m
6m

D 3m D

3m 6m
ELEVATION X-X AXIS ELEVATION Y-Y AXIS

FIG. Q4

Table Q4
Relevant Section Properties for 152 × 152 × 23 UC
Cross-Sectional Area 29.2 cm2 Radius of gyration about the x-x axis 6.54 cm
2
Yield stress 275 N/mm Radius of gyration about the y-y axis 3.70 cm
Modulus of Elasticity 205 kN/mm2 Robertson Constant for the x-x axis 5.5
Robertson Constant for the y-y axis 5.5

Perry-Robertson Strut Formula:

pE p y
pc =
φ + φ 2 − pE p y
p y + (η + 1) pE π 2E a ( λ − λ0 ) π 2E
where φ= ; pE = ; η= ; λ0 = 0.2
2 λ2 1000 py

In which the symbols have their usual meanings.

62 Dr. C. Caprani