Professional Documents
Culture Documents
www.DegreeTutors.com
Introduction
In this document we develop the theory associated with column buckling. In section 1, we
discuss the fundamental concept of buckling stability with reference to an idealised struc-
ture. We develop this further in section 2 to consider more realistic column structures with
distributed stiffness. Finally, in section 3 we consider the impact of realistic imperfections
on our strict mathematical predictions. In particular we focus on the influence of eccentric
loading and initial deformations. These notes are based on a collection of posts on Degree-
Tutors.com,
1
Contents v1.1
Contents
3.2.3 Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
• Both bars are perfectly aligned and the external load P is applied directly through the
longitudinal axis.
• Under ideal conditions the system is in compression and the spring in not under any
load.
Think about what happens if we impose a small lateral displacement on the structure, Fig. 2.
The rotational spring applies a restoring moment MB , equal to the angle through which the
spring rotates multiplied by the spring’s rotational stiffness:
MB = 2θ × βR
The restoring moment acts to decrease the lateral displacement while the axial load, P , tries
to increase it. At this point, two things could happen. If the axial load is ’small’, MB wins and
the structure straightens up returning to its original position/state. In this case the structure
is said to be STABLE.
If the axial load is ’large’, P wins and the lateral displacement continues until the structure
collapses. In this case the structure is said to be UNSTABLE. At some point we must cross
a boundary between the stable and unstable states. This boundary is characterised by the
critical load Pcr .
The critical load, Pcr is the axial load at which a structure under compression
will move from a stable condition or state to an unstable one.
Much of the discussion that follows further below will focus on determining what the critical
load is for a variety of column end conditions.
To determine the critical load for our idealised structure, we can ’cut’ the structure at the lo-
cation of the spring and consider equilibrium of the top half, the resulting free-body diagram
reveals the moment action of the spring at the cut end of the structure, Fig. 3
• Note that we have vertical force equilibrium as a result of the internal axial force re-
vealed by the cut at B.
• We also have moment equilibrium due to the couple creating by both P forces balanc-
ing MB .
• If we assume that the displacement of the system is small, we can approximate the
lateral deflection as:
L
δ≈ ×θ
2
So, taking the sum of the moments about B and assuming clockwise moments are positive
yields,
P × δ − MB = 0
P Lθ
− 2θβR = 0
2
PL
θ − 2βR =0
2
Now since we know θ isn’t equal to zero, the expression in brackets must equal zero. So
solving for P yields:
4βR
Pcr =
L
This is the value of P for which the external axial load is balanced by the restoring moment
of the rotational spring, as such, from what we said above, this is the critical load for our
idealised structure, i.e. the axial load at which it is on the boundary between being stable
and unstable.
It’s important to establish a clear understanding of what we mean when using the terms
equilibrium and stability, particularly in the context of a discussion on buckling.
The axial load is less than the critical load, the idealised structure is straight and in equilib-
rium.
If a small lateral deflection is imposed on the structure, it will return to the straight position.
A good analogy for this is a ball sitting on a concave surface, Fig. 4. The ball will always
return to the same stable state of equilibrium when the disturbance is removed.
When the axial load is larger than the critical load, the idealised structure is straight and in
a state of unstable equilibrium.
The structure is in a precarious (unstable) state of equilibrium. The smallest possible lateral
deflection or disturbance will cause the structure to buckle. From an engineering perspective
we must avoid this state.
The corresponding ball analogy is that of a ball precariously positioned at the top of a convex
surface, Fig. 5. The smallest disturbance will set the ball rolling.
In this case the disturbing influence of the axial load is matched by the restoring influence
of the moment MB .
When P = Pcr , for any ’small’ angle, the structure remains in a state of
NEUTRAL EQUILIBRIUM.
This case is analogous to a ball sitting on a perfectly flat surface. We can visualise the various
states of equilibrium in a graph of axial load versus angle of rotation, Fig. 6.
• Notice that ’in theory’, for an undefined range of θ (angles of rotation), when P = Pcr ,
we have a state of neutral equilibrium and the structure could keep deflecting.
Figure 6: Axial force versus rotation with three states of equilibrium identified.
From a practical point of view, we always aim to maintain axial loads well below the critical
buckling load for a structure. The concept of neutral equilibrium can be thought of as a
theoretical state that is highly unlikely to be observed in practice.
Up to this point we’ve discussed a simplified idealised structure that has all of its resistance
to rotation concentrated at the midpoint in a rotational spring. This has allowed us to get a
feel for some important concepts.
When we refer to a pin support, this is a support that offers no resistance to rotation. Some-
times we come across pin supports in practice that very closely approximate this theoretical
definition. We can see some examples in Fig. 7.
For the purposes of our discussion here, we’ll assume we’re dealing with pin supports like
these, and that they don’t inhibit the rotation of the structure at the support.
In order to analyse this structure, we’re going to use the differential equation of the defection
curve in which M is the internal bending moment at a given cross-section, v represents the
lateral deflection of the column and x the distance along the length of the column. EI is
referred to as the flexural rigidity and is the product of Young’s modulus (E) and the second
moment of area of the cross-section (I).
d2 v
EI =M
dx2
This equation simply models the deflected shape of the column for a given flexural rigidity.
You’re likely to have come across this equation elsewhere in your engineering studies. Our
laterally deflected column behaves in much the same way as a beam. Remember buckling is
a bending failure rather than a failure due to direct compression (crushing). Our task now, is
to use this equation to derive an expression for the critical axial load in a pinned-end column.
Just as for the idealised structure discussed above, our derivation here will also assume all
deflections are small by comparison to the size of the structure.
First consider the structure, its deflected shape and the free-body diagram that results from
cutting the structure at a distance x from the bottom support, Fig. 8. Taking moments about
Figure 8: Pinned-pinned column (a) the structure, (b) the deflected shape, (c) the sub-
structure free-body diagram.
point A yields:
M = −P v
Substituting M back into the differential equation of the deflection curve and rearranging
slightly yields:
d2 v P
2
+ v=0
dx EI
This is a linear, homogeneous, 2nd order differential equation with constant coefficients. It
also happens to be an equation that models the deflected shape of our structure and contains
P , the axial load. Following the same approach as for the idealised structure discussed
previously, if we solve this equation we can determine Pcr , the value of applied load at
which it is balanced by the column’s flexural bending resistance.
So we’re faced with trying to solve a differential equation. At this point, it’s helpful to be able
to point to the general solution of this form of differential equation:
r ! r !
P P
v(x) = C1 sin x + C2 cos x
EI EI
If you want to take a detour at this point to see how the differential equation is solved, watch
me do it in this video. Otherwise, we’ll continue working out the unknown constant in the
next section.
Now that we have the general solution for this format of equation, we need to solve for the
constants of integration using the boundary conditions (pinned-pinned) for our column.
Boundary condition 1:
At x = 0 (base of the column at pin support), v = 0 (the lateral deflection must equal zero.
Imposing these conditions on our general solution yields:
0 = C1 sin(0) + C2 cos(0)
Therefore,
C2 = 0
Boundary condition 2:
At x = L (top of the column at pin support), v = 0 (lateral deflection also zero). Again,
imposing these boundary conditions yields,
r !
P
0 = C1 sin L
EI
So, either C1 equals zero or sin(...) equals zero. If C1 equals zero, the equation is satisfied
for any value of P . This doesn’t really move us forward. We can think of this solution as the
trivial solution. Of more interest is the case where:
r !
P
sin L =0
EI
Knowing what we know about sin functions, this equation can only be true, when,
r
P
L = 0, π, 2π, ...nπ
EI
In this scenario, P would need to equal zero (as all other terms have non-zero values by
definition) and we know this is not the case. So for a pinned-pinned column, our solution is:
r
P
L = nπ
EI
where n = 1, 2, 3... Finally, we can rearrange this equation to make P the subject of the
equation:
n2 π 2 EI
P =
L2
For n = 1, 2, 3,.
This equation represents an infinite series of buckling loads. The lowest one
(n = 1) is the critical buckling load, also known as the Euler Buckling Load
PE .
π 2 EI
PE =
L2
So far, we have established that there is an infinite series of buckling loads and the lowest
one is the critical one and called the Euler Buckling load. This raises the question of what do
the larger buckling loads correspond to?
These are loads that correspond to higher modes of buckling. Each mode of buckling has a
corresponding buckled shape. To explore this further, refer back to our general solution after
we established that C2 = 0:
r !
P
v(x) = C1 sin x
EI
for n = 1, 2, 3,. This equation describes the deflected shape of the column for each value
of the buckling load (n = 1, 2, 3, ...) defined above. Only when the axial load has one of the
values given by the equation (previously defined),
n2 π 2 EI
P =
L2
can the column have a bent shape and be in equilibrium. For any other load, the column
would be straight and in a state of STABLE equilibrium (P < Pcr ) or UNSTABLE equilibrium
(P > Pcr ). We recall this concept from our previous discussion of idealised column struc-
tures, Fig. 6.
Our equation for the deflected or buckled shape of a column contains the parameter C1 ,
left over from our general solution to the differential equation. We now recognise that C1
represents the magnitude of the sine wave or in this context, the maximum lateral deflection
for the deflected/buckled deflection (defined above):
nπx
v(x) = C1 sin
L
If we now recall the states of equilibrium, when the column is in its buckled shape (P = Pcr )
and in a state of neutral equilibrium, the maximum lateral deflection can have any value that
still satisfies our small deflection assumption. We can represent the value of C1 graphically,
Fig. 9
The value of C1 is therefore mathematically undefined, it can have any ’small’ value. For-
tunately this doesn’t cause us any practical difficulty, as mentioned previously, the state of
neutral equilibrium is not something we observe in civil engineering practice. For all intents
and purposed, to a civil or structural engineer, a column experiencing Pcr is considered to
have failed. This theoretical state of neutral equilibrium is also known as a bifurcation point.
Imagine an axially loaded pinned end column that is restrained laterally at its mid-height
point (by a floor slab say). The column would be prevented from buckling under the first
critical (Euler buckling) load due to the lateral restraint. The floor slab restraint literally
holds the column and stops it from buckling.
As a result of this restraint, the column can carry more load, until it reaches the second
buckling load (n = 2). As the column is not restrained against buckling in the second mode,
it would now buckle (or theoretically enter into a state of neutral equilibrium).
We can easily visualise the first three modes of buckling simply by evaluating the load and
deflected shape equations for n = 1, 2, 3.
Mode 1, n=1
π 2 EI
Critical load, Pcr =
L2
πx
Buckled shape, v(x) = C1 sin
L
Mode 2, n=2
4π 2 EI
Critical load, Pcr =
L2
2πx
Buckled shape, v(x) = C1 sin
L
Mode 3, n=3
9π 2 EI
Critical load, Pcr =
L2
3πx
Buckled shape, v(x) = C1 sin
L
For any given mode, Pcr represents the practical upper limit on axial load.
Natural imperfections, for example column ’out of plumb’ or poor load align-
ment will almost certainly induce buckling when the critical load is ex-
ceeded.
So far we’ve looked at the behaviour of a column pinned at both end. This is the typical start-
ing point. But the same process can be followed to determine the corresponding equations
for columns with different types of support conditions. In the videos below, I’ll carry out the
derivations to demonstrate the process for:
• Columns with a base fixed against rotation and free at the top (fixed-free)
• Columns fixed against rotation at the base and pinned at the top (fixed-pinned)
Figure 11 summarises the critical loads for the range of support conditions covered above.
Figure 11: Summary of critical loads and effective lengths for various support conditions.
We recall from the equation for the buckling load that it is a function of I , the second moment
of area of the cross-section,
n2 π 2 EI
P =
L2
So for a given cross-section, a column will always buckle about the axis with the lower sec-
ond moment of area, the ’weaker’ axis. This assumes that both axes have equal restraint.
Consider the case of a universal column (UC) section under compression, Fig. 12,
Since the X-X axis, is the major principle axis (with the largest value of second moment of
area - you might remember this from your study of Mohr’s circle), it is the stronger axis and
so the column will buckle about the minor principle or Y-Y axis first.
When considering the buckling load for a column structure, the cross-section shape plays a
key role, you should evaluate the major and minor principal axes to determine the critical
axis for buckling.
The critical stress is the average axial stress in a cross-section under the
critical load.
Pcr π 2 EI
σcr = =
A AL2
π2E
σcr =
L 2
r
L
λ=
r
The slenderness ratio is a very useful measure of a column’s geometry and susceptibility to
buckling. A high slenderness ratio indicates greater susceptibility to buckling. The slender-
ness ratio should be determined separately for each principle axis. We can get an intuition
for the slenderness ratio by visualising both extremes, Fig. 13.
Figure 13: Columns with high (left) and low (right) slenderness ratios.
The equation for critical stress can therefore be written as a function of the slenderness ratio
as follows,
π2E
σcr =
λ2
Plotting the critical stress versus slenderness ratio for a given value of Young’s modulus
yields an Euler Curve showing the safe range of average axial stress for a given slenderness
ratio (or the safe range of slenderness ratio values for a given axial stress), Fig. 14.
An Euler curve is only valid for critical stresses below the material yield
stress
The next concept we’ll look at is column effective length. The effective length of a column,
Le , is the length between points of inflection on the deflected curve/shape (even if the shape
must be extended until a point of inflection is reached - discussed below). For a column
pinned at both ends, the effective length is simply the full length of the column between pin
restraints, Fig. 15.
Consider a column with a fixed support at its base and no lateral restraint at the top (fixed-
free). At the base, no rotation can occur. In order to obtain two inflection points on the
deflected shape, it must be extended, Fig. 16.
The ’imaginary’ red portion of the deflected shape in Fig. 16 is the extension required to
provide inflection point number 2. Thus the effective length for a fixed-free column is 2L.
Another way to thing about effective length is that it’s the length of an equiv-
alent pinned end column.
Before concluding our discussion of effective length, we can define a general expression for
the critical load as:
π 2 EI
Pcr =
Le 2
If we now introduce the Effective Length Factor, K , such that Le = Kl, we can state the
critical load as a function of the effective length factor:
π 2 EI
Pcr =
(KL)2
So far we have assumed that all compression forces on a column are axial forces that have
their line of action along the column’s longitudinal axis. As a result of this assumption the
column will potentially experience the three states of equilibrium previously discussed:
1. Stable equilibrium
2. Neutral equilibrium
3. Unstable equilibrium
However, in reality a column may not experience such ideal conditions. Compression forces
are often not axial forces but forces applied at some eccentricity from the column’s longitu-
dinal axis. This results in quite different failure behaviour that we will investigate here.
We’ll set the analysis up by considering a column, pinned at both ends, but with an ’outstand’
or cantilever protruding from each end. The cantilever is somewhat unrealistic but it simply
serves here as a way of applying an eccentric compression force. The compression forces
are applied at the ends of these cantilevers as shown below at an eccentricity, e, Fig. 17
The first thing we note is that the load P applied an an eccentricity e, is the same as simul-
taneously applying an axial load P and a moment, Mo ,
Mo = P × e
We can start by following the same procedure as before. We need to determine the differen-
tial equation of the deflection curve. We do this by evaluating the internal bending moment,
Mx at some distance x along the height of the deflected column as shown in the right hand
diagram in Fig. 17.
Mx = Mo + P (−v)
Mx = Mo − P v
Now substituting this expression for Mx into the differential equation of the deflection curve
yields,
d2 v
EI = Mo − P v = P e − P v
dx2
P d2 v
k2 = and = v̈
EI dx2
v̈ + k 2 v = k 2 e
The solution of this equation follows very closely that of the fixed-free column discussed
above. The general solution is given by,
v = C1 sin(kx) + C2 cos(kx) + e
C2 = −e
and,
−e(1 − cos(kL))
C1 =
sin(kL)
We note that for a given value of e and P , the deflection is always defined. In the previous
axially loaded column analyses we considered, we could only determine a buckling mode
shape, but the maximum deflection, when P = Pcr , was undefined. This corresponded to a
state of neutral equilibrium.
From this we conclude that a column with eccentric compression forces has
no neutral equilibrium state and therefore does not exhibit sudden column
buckling behaviour.
For the pinned-pinned column considered in this case, the maximum lateral deflection, δ ,
occurs at the mid-height point, x = L/2,
L
δ = −v
2
π 2 EI
Pcr =
L2
we can rewrite k ,
r
π P
k=
L Pcr
Therefore,
r
P
kL = π
Pcr
We can now substitute this expression back into our equation for the maximum deflection at
the mid-height to yield,
" r ! #
π P
δ = e sec −1
2 Pcr
We can plot this equation to obtain a load versus deflection curve for various values of load
eccentricity, e. This relationship is represented qualitatively below in Fig. 18
Figure 18: Load versus deflection for an eccentrically loaded, pinned-pinned column.
We can observe from this graph that the relationship between lateral deflec-
tion at mid-height and applied load is non-linear. This means we cannot use
the principle of superposition to determine the influence of multiple simul-
taneously applied loads.
The vertical line in the graph above, represents the case when e = 0. In this case we observe
the equilibrium states discussed previously. For values of e > 0, as P approaches the critical
load, the deflection increases and the horizontal line representing the value of critical load
becomes an asymptote for the load-deflection curves.
Remember that the preceding derivation is based on the assumptions of small deflections.
So in reality as the deflections increase, the observed behaviour of the column will deviate
form the strict mathematically predicted behaviour visualised in the graph above.
A key takeaway from this discussion on eccentric loading is that the be-
haviour of a real world column under realistic loading conditions is not likely
to accord with the strict mathematical models we evaluated previously for
perfectly axially loaded columns.
Now we’re going to consider the behaviour of a column that already has an initial lateral
deformation. We will follow the same procedure as before to determine an equation that
Consider the pinned-pinned column below that has an initial lateral deflection, v0 , Fig. 19.
Note that the complete lateral deflection consists of the initial deflection and the buckled
deflection,
Note that for clarity, hereafter we’ll dispense with the (x) as it’s clear that the lateral deflec-
tion is always a function of x. Taking moments about point A yields,
M = −P v
M = −P (v̄ + v0 )
Figure 20: Free-body diagram of the substructure of a pinned-pinned column with an initial
lateral deflection, v0 .
Substituting this expression for M into the differential equation of the deflection curve yields,
d2 v̄
EI = −P v̄ − P v0
dx2
d2 v̄
= v̄¨
dx2
and
P
k2 =
EI
we have,
v̄¨ + k 2 v̄ = −k 2 v0
To proceed with the derivation we need to assume some function of x to describe the ini-
tial deformation of the column. For the purposes of this derivation we can assume that a
simple sine function describes the initial deformation v0 . In this case we have the following
differential equation of the deflection curve,
πx
v̄¨ + k 2 v̄ = −k 2 V0 sin
L
We’ve already seen that the complementary solution (i.e. the solution when the RHS equals
zero) is given by,
This just leaves the particular solution to be calculated. Because the right hand side of the
differential equation contains a sinusoid, we can assume a general sinusoid for the particular
solution, this gives us...
πx πx
v̄p = C3 sin + C4 cos
L L
Remember that the reason we assume a solution is so that we can differentiate it and sub it
back into our governing differential equation, so differentiating yields,
π πx π πx
v̄˙ p = C3 cos − C4 sin
L L L L
π2 πx π2 πx
v̄¨p = −C3 2
sin − C4 2 cos
L L L L
Now substituting these expressions back into our governing differential equation yields,
−π 2 h πx πx i
2
h πx πx i
2
πx
C 3 sin + C 4 cos + k C 3 sin + C 4 cos = −k V 0 sin
L2 L L L L L
Now, if we equate cosine terms on the left hand side with cosine terms on the right hand
side, we get,
−π 2
πx πx
2
C4 cos + k cos =0
L2 L L
Now since we know the term in brackets doesn’t equal zero, we can deduce that C4 = 0.
Now equating sine terms on both sides gives us,
−π 2
πx πx πx
2 2
C3 sin + k sin = −k V 0 sin
L2 L L L
Combining the particular and complimentary solutions gives us the general solution,
k2 πx
v̄ = C1 sin(kx) + C2 cos(kx) + V0 sin
π2
− k2 L
L2
At this point we can apply boundary conditions to determine the remaining unknown con-
stants of integration. At x = 0 (base of the column at pin support), v = 0 (the lateral
deflection must equal zero). Imposing this condition on our general solution we find that
C2 = 0. And so our general solution simplifies to,
k2 πx
v̄ = C1 sin(kx) + V0 sin
π2
− k2 L
L2
The second boundary condition is that at x = L (top of the column at pin support), v = 0
(lateral deflection also zero). Again, imposing this condition yields,
r !
P
C1 sin L =0
EI
From this we can deduce that C1 must equal zero as the sine term cannot equal zero for a
non-trivial solution. Therefore the complete buckling deflection is given by,
P πx
EI
v̄ = V0 sin
π2
− P L
L2 EI
π 2 EI
PE =
L2
P
ρ=
PE
Therefore the total deflection (remember this is the initial deflection plus the buckling de-
flection) is given by,
πx ρ πx
v = V0 sin + V0 sin
L 1−ρ L
V0 πx
v= sin
1−ρ L
or,
πx
v = αV0 sin
L
where,
1
α=
1−ρ
3.2.3 Observations
We can see from the equation derived above that the factor α represents a magnification
factor on the initial displacement. As the axial load, P increases, the magnitude of lateral
deflection increases but the deflected shape remains the same. Note that when P = PE , the
equation breaks down as the magnification factor α goes to infinity.
This is not likely to be a problem because P is not likely to reach PE while the structure
satisfies our small deflection assumption. We can plot the relationship between ρ and α to
visualise the behaviour of the column, Fig. 21.
It’s important to recognise that for a column with an initial deformation, we do not observe
the strict mathematical column buckling behaviour predicted for perfectly loaded perfectly
straight columns. Even so, the Euler load is still an important quantity that has a role in
predicted the lateral deflection of the column via the ratio ρ.