1

Module 5

DESIGN AND ANALYSIS OF

COMPRESSION MEMBERS :
COLUMNS

Lesson No.1: Euler Column Formula and

Effective Length Factor

I. LEARNING OBJECTIVES
At the end of the lesson, the student will be able to:

1. Know the different names given to structural members carrying compressive

loads.
2. Learn the development of column formula and the derivation of Euler column
formula.
3. Classify of columns according to mode of failures.
4. Determine and calculate the effective length factor, k, for the different types of
column end support.

II. OVERVIEW
There are several types of compression members, the column being the best known.
Among the other types are the top chords of trusses and various bracing members. In
addition, many other members have compression in some of their parts. These include
the compression flanges of rolled beams and built-up beam sections, and members
that are subjected simultaneously to bending and compressive loads. Compression
members are structural elements that are normally subjected only to axial compressive
forces, that is, the loads are applied along the longitudinal axis through the centroid of
the member cross section. The stress which is axial is considered to be uniform over
the entire cross section. However, this ideal state is never achieved in reality because
some eccentricity of the load is inevitable and bending cannot be neglected. This type
of situation is known as “beam-column
Columns are usually thought of as being straight vertical members whose
lengths are considerably greater than their thicknesses (dimension). Short vertical
members subjected to compressive loads are often called struts, or simply,
compression members; however, the terms column and compression member will be
used interchangeably.

Structural members that carry compressive loads are sometimes given names that
identify them as to their function.
a. Columns
▪ vertical compression members that have their length dimension
considerably larger than their least cross-sectional dimension.
b. Knee Brace at the junction of columns and roof trusses.
c. Struts
▪ compression members that serve as bracing
d. Post and Pillars
e. Truss compression member
f. Stanchion
 2

When axially loaded compression member becomes unstable overall (that is, not
locally unstable), there are three general modes by which the columns can fails. These
are flexural buckling, local buckling, and torsional buckling. These modes of buckling
are briefly defined as follows: (See also Figure 1)
1. Flexural buckling (also called Euler buckling): it is the type of buckling
(deflection) caused by bending or flexure about the axis corresponding to the
largest slenderness ratio. This is usually the minor principal axis – the one ith the
smallest radius of gyration. Compression member with any type of cross-
sectional configuration can fail in this way.
2. Torsional buckling (local buckling): this type of failure is caused by twisting
about the longitudinal axis of the member. It can occur only with doubly
symmetric cross section with very slender cross sectional elements – like
members built up from thin plate elements. The susceptibility of a member to
torsional buckling is measured by the width-thickness ratio of the parts of its
cross section.
3. Flexural-torsional buckling : this type of failure is caused by a combination of
flexural buckling and torsional buckling. The member bends and twist
simultaneously This type of failure can occur only with unsymmetrical cross
sections, both those with one axis of symmetric – such as channels, structural
tees, double-shape angle, and equal-leg single angles and those with no
symmetry, such as unequal-leg single angle.

Figure 1. Types of buckling failure of compression members

The tendency of a member to buckle is usually measured by its slenderness ratio. The
longer a column becomes for the same cross section, the greater becomes its
tendency to buckle and the smaller becomes the load it will support. The buckling of the
member is also affected by factors such as
• the types of end connections,
• eccentricity of load application,
• imperfection of column material,
• initial crookedness, and
• residual stress from manufacture.
 3

III. COURSE MATERIALS

1.0 DEVELOPMENT OF COLUMN FORMULA

The use of columns dates to before the dawn of history, but it was not until 1729 that a
paper was published on the subject, by Pieter Van Musschenbroek, a Dutch
mathematician. He presented an empirical column formula for estimating the strength
of rectangular columns

Figure 2:. Test result curve

of a column with various
slenderness ratio

A few years later, in 1757, Leonhard Euler, a Swiss mathematician, wrote a paper of
great value concerning the buckling of columns. He was probably the first person to
realize the significance of buckling.
The testing of columns with various slenderness ratios results in a scattered
range of values, such as those shown by the broad band of dots in the above Figure 2.
The dots will not fall on a smooth curve, even if all of the testing is done in the same
laboratory, because of the difficulty of exactly centering the loads, lack of perfect
uniformity of the materials, varying dimensions of the sections, residual stresses, end
restraint variations, and other such issues.
The magnitudes of the yield stresses of the sections tested are quite important
for short columns, as their failure stresses are close to those yield stresses. For
columns with intermediate slenderness ratios, then yield stresses are of lesser
importance in their effect on failure stresses, nd of no significance for long slender
column. For intermediate range columns, residual stresses have more effect on the
results, while the failure stresses for long slender columns are very sensitive to end
support conditions.

2.0 DERIVATION OF THE EULER FORMULA

The Euler formula is derived for a straight, concentrically loaded, homogeneous, long,
slender, elastic and weightless column with round ends. It is assumed that perfect
column has been laterally deflected as shown in the Figure below.

Figure 3
 4

The 𝑥 and 𝑦 axes are located as shown in the Figure 3. As the bending moment at any
point in the column is – 𝑃𝑦 , the equation f the elastic curve can be written as

𝑑2𝑦
𝐸𝐼 = −𝑃𝑦
𝑑𝑥 2

For convenience in integration, both sides of the equation ate multiplied by 2𝑑𝑦 :

𝑑𝑦 𝑑𝑦
𝐸𝐼2 𝑑 = −2𝑃𝑦𝑑𝑦
𝑑𝑥 𝑑𝑥

𝑑𝑦 2
𝐸𝐼 ( ) = −𝑃𝑦 2 + 𝐶1
𝑑𝑥

When = 𝛿 , 𝑑𝑦⁄𝑑𝑥 = 0, and the value of 𝐶1 will equal 𝑃𝛿 2 and

𝑑𝑦 2
𝐸𝐼 ( ) = −𝑃𝑦 2 + 𝑃𝛿 2
𝑑𝑥

The preceding expression is arranged more conveniently as follows:

𝑑𝑦 2 𝑃
(𝑑𝑥 ) = (𝛿 2 − 𝑦 2 )
𝐸𝐼

𝑑𝑦 𝑃
𝑑𝑥
= √
𝐸𝐼
√𝛿 2 − 𝑦 2

𝑑𝑦 𝑃
= √ 𝑑𝑥
√𝛿 2 − 𝑦 2 𝐸𝐼

Integrating this expression, the result is

𝑦 𝑃
𝑎𝑟𝑐 sin = √ 𝑥 + 𝐶2
𝛿 𝐸𝐼

When 𝑥 = 0 and 𝑦 = 0 , 𝐶2 = 0. The column is bent into the shape of a sne

curve expressed by the equation
𝑦 𝑃
𝑎𝑟𝑐 sin = √ 𝑥
𝛿 𝐸𝐼

When 𝑥 = 𝐿⁄2, 𝑦 = 𝛿, resulting in

𝜋 𝐿 𝑃
= √
2 2 𝐸𝐼

In this expressoon, 𝑃 is the critical bukling load, or the maximum load the
column can support before it becomes unstable . Solving for 𝑃, we have

𝜋 2 𝐸𝐼
𝑃=
𝐿2

2.1 Euler’s Column Elastic Buckling Load

Column buckling theory originated with Leonard Euler (1707-1783). An initially straight
concentrically loaded member, in which all fibers remain elastic until buckling occurs, is
slightly bend as shown in the Figure 4(a).
 5

Figure 4

In Figure 4(a), if axial load P is slowly applied, it will ultimately become large enough to
cause the member to become unstable and assume the shape indicated by the dashed
line. The member has buckled, and the corresponding load is called critical bucking
load. If the member is stockier as shown in Figure (b), a large load will be required to
bring the member to the point of instability. For extremely stocky members, failure may
occur by compressive yielding rather than buckling.
For column to buckle elastically, it will have to be long and slender, and that the
stress just before buckling is below the proportional limit. Its buckling load can be
computed with the Euler Formula. The Euler critical load for a column, homogeneous
with both ends pinned, is

𝜋 2 𝐸𝐼
𝑃= 2
𝐿
This expression is the Euler formula, but usually it is written in a little different form
involving the slenderness ratio. Since 𝑟 = √𝐼 ⁄𝐴 and 𝑟 2 = 𝐼 ⁄𝐴 and = 𝑟 2 𝐴 , the Euler
formula may be written as

𝑃 𝜋2 𝐸
= 2
𝐴 (𝐿⁄𝑟 )

𝑃𝑐𝑟 𝜋2 𝐸
Note that : 𝐹𝑐𝑟 = 𝐹𝑒 = = 2 ≤ proportional limit
𝐴 (𝐿⁄𝑟 )

where: 𝐹𝑒 = 𝐹𝑐𝑟 = elastic or buckling stress

𝐿 = length of member
𝑟 = radius of gyration = √𝐼⁄𝐴
𝐴 = gross cross-sectional area of member

Important: If 𝑃⁄𝐴 𝑜𝑟 𝐹𝑒 exceeds the steel's proportional limit, the elastic Euler formula
is not applicable because column will fail inelastically.

Euler's approach was generally ignored for design because test result did not
agree with it. Columns of "ordinary length" used in the design were not strong would
indicate. Considere and Engesser (1889) independently realized that portions of usual
"length columns" become inelastic prior to buckling. And that a value of 𝐸 should be
used that could account for some of the compressed fibers being strained beyond the
 6

proportional limit. It was thus consciously recognized that in fact "ordinary length"
columns fail by inelastic buckling rather than by elastic buckling.
Complete understanding of the behaviour of concentrically loaded columns was
achieved when Shanley (1946) offered an explanation. He reasoned that it was actually
possible for a column to bend and still have increasing axial compression but that it
begins bend upon reaching what is commonly referred to as the buckling load, which
includes inelastic effects on some or all fibers of the cross section.

3.0 CLASSIFICATION OF COLUMNS ACCORDING TO MODE OF FAILURE

The strength of a column and the manner in which it fails are greatly dependent on its
effective length. A very short, stocky steel column may be loaded until the steel yields
and perhaps on into the strain-hardening range, resulting in a load capacity which is the
same as in tension.
As the effective length of a column increases, its buckling stress will decrease. If
the effective length exceeds a certain value, the buckling stress will be less than the
proportional limit of the steel. Column in this range are said to fail elastically.
Very long columns will fail at loads that are proportional to the bending rigidity of
the column (EI) and independent of the strength of the steel.

It is for this reason that columns are classified as being long, short or intermediate.
(See Figure 5)
1) Long Columns
• the Euler formula predicts very well the strength of long columns where
the axial buckling stress remains below the proportional limit. Such
columns will fail by elastic buckling, where buckling occurs at
compressive stresses within the elastic range.
2) Intermediate Column
• columns will fail by inelastic buckling when localized yielding
occur,initiated at some point of weakness or crookedness. Some of the
fiber stresses for this column will reach the yield stress and some will
not. Most columns fail into those range.
3) Short Column
• short and stocky columns doesn't fail by elastic buckling. It will crush
due to general yielding and compressive stress will be in the inelastic
range.

(a) Long column (b) Intermediate column (c) Short column

Figure 5. Classification of column

 7

4.0 BASIC COLUMN STRENGTH

To determine a basic column strength, certain conditions may be assumed for the ideal
column.
1. With regard to material, it may be assumed
• there are the same compressive stress-strain properties throughout the
section;
• no initial internal stresses exist such are those due to cooling after
rolling and those due to welding.
2. Regarding shape and end conditions, it may be assumed
• the column is perfectly straight and prismatic;
• the load resultant acts through the centroidal axis of the member until
the member begins to bend;
• the end conditions must be determinate so that a definite equivalent
pinned length may be established.
3. Further assumptions regarding buckling may be made, as
• the small deflection theory of ordinary bending is applicable and shear
may be neglected; and
• twisting or distortion of the cross-section does not occur during bending.

Once the foregoing assumptions have been made, it is now agreed that the strength of
a column may be expressed by

𝜋 2 𝐸𝑡
𝑃𝑐𝑟 = 𝐴 = 𝐹𝑐𝑟 𝐴
(𝐾𝐿⁄𝑟)2
where:
𝐸𝑡 = tangent modulus of elasticity at stress 𝑃𝑐𝑟 ⁄𝐴
𝐾 = effective length factor
𝐴 = gross cross-sectional area of member
𝐿 = length of member
𝐾𝐿⁄𝑟 = effective (or equivalent pinned-end) slenderness ratio
𝑟 = radius of gyration
𝐼 = moment of inertia

It is well known that long compression members fail by elastic buckling and that short
stubby compression members may be loaded until the material yields or perhaps even
into the strain-hardening range. However, in the vast majority of usual situations, failure
occurs by buckling after a portion of the cross-section has yielded. This is known as
inelastic buckling.

Figure 6. Typical range of column strength vs slenderness ratio

 8

For many years theoretical determinations of column strength did not agree with test
results. Test results include effects of initial crookedness of the member, accidental
eccentricity of load, end restraint, local or lateral buckling, and residual stress. A typical
curve observed strengths was shown in the Figure 6.
Design formulas, therefore, were base on empirical results. Various straight-line
and parabolic formulas have been used, as well as other complex expressions, in order
to fit the curve of test results in a reasonable accurate and practical manner.

5.0 UNSUPPORTED / UNBRACED LENGTH AND EFFECTIVE LENGTH OF

COLUMNS

The basic information on the behaviour of straight, concentrically loaded slender

column was developed by Euler more than 200 years ago. In generalized form, it states
that such a member will fail by buckling at critical load.

𝜋2 𝐸𝑡 𝐼
𝑃𝑐 = (𝑘𝑙)2

It is seen that he buckling load stress decreases rapidly with increasing slenderness
ratio 𝑘𝑙 ⁄𝑟.
. End restraint and its effect on the load-carrying capacity of the column is very
important. Columns with appreciable rotational and translational end restraint can
support considerably more loads than can those with little rotational end restraint.
The Euler equation is useful only if the end support conditions are carefully
considered. To successfully use the equation for practical columns, the value of “𝑙"
should be the distance between "points of inflection" in the buckled shape. This
distance is referred to as the effective length of column.

5.1 Unsupported Lengths

The length used for calculating the slenderness ratio of a column, 𝑙𝑢 , is its
unsupported length which is equal to the clear distance between slabs, beams, or other
members that provide lateral support to the column.

5.2 Effective Length Factor

The effective length of column is defined as the distance between the points of
zero moment or the distance between its point of inflections. The effective length of a
column is referred to a "𝑘𝑙" where 𝑘 is the effective length factor. The magnitude of 𝑘
depends on the rotational restraint supplied at the ends of the column and upon
resistance to lateral movement provided.
The slenderness of columns is based on their geometry and on their lateral
bracing. As their slenderness increases, their bending stresses increase, and thus
buckling may occur Steel columns generally have small slenderness ratios. As a result,
they can usually be designed as short columns without strength reduction because of
slenderness. However, if they are “slender”, the moment for which the column must be
designed is increased or magnified. Factors involved in the calculation of slenderness
ratio include
a. unsupported column lengths,
b. effective length factors,
c. radii of gyration

Columns with different end conditions have entirely different effective length. Figure 7
below gives the different effective length factors in braced frames. Obviously, the
smaller the effective length of a particular column, the smaller its danger of buckling
and the greater its load-carrying-capacity.
 9

Figure 7. Effective lengths for columns in braced frames (sidesway prevented)

Figure 8 . Columns for unbraced frames

The concept of effective lengths is simply a mathematical method of taking a column –

whatever its end and bracing conditions – and replacing it with an equivalent pinned
end-braced column. A complex buckling analysis could be made for a frame to
determine the critical stress in a particular column. The 𝑘 factor is determined by finding
the pinned end column with and equivalent length that provides the same critical stress.
The 𝑘 factor procedure is a method of making simple solutions for complicated frame-
buckling problems. Steel columns serve as part of frames, and these frames are
sometimes braced and sometimes unbraced. (Figure 8)

5.3 Braced Frame

A braced frame is one in which "lateral stability is provided by diagonal bracings,
shear walls or equivalent means. “The vertical bracing system must be "adequate" as
determined by structural analysis" to prevent buckling of the structure and to maintain
the lateral stability of the structure, including overturning effects of drift, under the
factored loads... "Note that a vertical column in a braced frame would have no
sidesway moment of its top relative to its bottom.
 10

5.4 Unbraced Frame

An unbraced frame is one in which "lateral stability depends upon the bending
stiffness of rigidly connected beams and columns." The buckling of an unbraced frame
is one of sidesway where the top of a column moves to the side relative to the bottom.
Figures 8 (b) shows an unbraced frame having sidesway buckling. The buckled shape
and therefore the effective length of the columns will depend on the stiffness of the
participating members in flexure. The effective length 𝑘𝑙 may be obtained by matching
the buckled shape of a column buckled shape. As shown in the figure, 𝑘𝑙 will always
exceed 𝑙.

For braced frames, 𝑘 values can never be greater than 1.0, but for unbraced frames,
the 𝑘 value will always be greater than 1.0 because of sidesway.

(a) Laterally braced (b) unbraced

Figure 8 . Rigid-frame buckling

6.0 EFFECTIVE LENGTH FACTOR OF COLUMN

6.1 Determination of Effective Length Factor, 𝒌

Table 1 : Effective length factor for centrally loaded columns

 11

In the subject of effective lengths, some suggested K factors in Table 1. These factors were
developed for columns with certain idealized conditions of end restraint, which may be very
different from practical design conditions. The table values are usually quite satisfactory for
preliminary designs and for situations in which sidesway is prevented by bracing. Should the
columns be part of a continuous frame subject to sidesway, however, it would often be
advantageous to make a more detailed analysis.To a lesser extent, this is also desirable for
columns in frames braced against sidesway.
The true effective length of a column is a property of the whole structure, of
which the column is a part. Theoretical mathematical analyses may be used to
determine effective lengths, but such procedures are typically too lengthy and perhaps
too difficult for the average designer.The usual procedure is to consult either Table 1,
interpolating between the idealized values as the designer feels is appropriate, or the
alignment charts.
The most common method for obtaining effective lengths is to employ the charts
shown in Figure 9. They were developed by O. G. Julian and L. S. Lawrence, and
frequently are referred to as the Jackson and Moreland charts, after the firm where
Julian and Lawrence worked. The charts were developed from a slope-deflection and
of the frames that included the effect of column loads. One chart was developed for
columns braced against sidesway and one for columns subject to sidesway. Their use
enables the designer to obtain good K values without struggling through lengthy trial
and-error procedures with the buckling equations.
To use the alignment charts, it is necessary to have preliminary sizes for the
girders and columns framing into the column in question before the K factor can be
determined for that column. Meaning, we have to either estimate some member sizes
or carry out a preliminary design.
When we say sidesway is inhibited, we mean there is something present
other than just columns and girders to prevent sidesway or the horizontal
translation of the joints. That means we have a definite system of lateral bracing, or
we have shear walls. If we say that sidesway is uninhibited, we are saying that
resistance to horizontal translation is supplied only by the bending strength and
stiffness of the girders and beams of the frame in question, with its continuous
joints.
The resistance to rotation furnished by the beams and girders meeting at one
end of a column is dependent on the rotational stiffnesses of those members. The
moment needed to produce a unit rotation at one end of a member if the other end of
the member is fixed is referred to as its rotational stiffness. From our structural analysis
studies, this works out to be equal to 4EI/L for a homogeneous member of constant
cross section. On the basis of the preceding, we can say that the rotational restraint at
the end of a particular column is proportional to the ratio of the sum of the column
stiffnesses to the girder stiffnesses meeting at that joint, or

In applying the charts, the G factors at the column bases are quite variable. It is
recommended that the following two rules be applied to obtain their values:
a. For pinned columns, G is theoretically infinite, such as when a column is
connected to a footing with a frictionless hinge. Since such a connection is not
frictionless, it is recommended that G be made equal to 10 where such nonrigid
supports are used.
b. For rigid connections of columns to footings, G theoretically approaches zero,
but from a practical standpoint, a value of 1.0 is recommended, because no
connections are perfectly rigid
 12

(a) Alignment chart—sidesway inhibited (braced frame).

(b) Alignment chart—sidesway uninhibited (moment frame).

Figure 9. Alignment chart for 𝐺 for (a) Braced Frames, (b) Unbraced Frames
 13

6.2 Frames Meeting Alignment Chart Assumptions

The Jackson and Moreland charts were developed on the basis of a certain set of
assumptions. Among these assumptions are the following:
a. The members are elastic, have constant cross sections, and are connected with
rigid joints.
b. All columns buckle simultaneously.
c. For braced frames, the rotations at opposite ends of each beam are equal in
magnitude, and each beam bends in single curvature.
d. For unbraced frames, the rotations at opposite ends of each beam are equal in
magnitude, but each beam bends in double curvature.
e. Axial compression forces in the girders are negligible.

6.3 Frames Not Meeting Alignment Chart Assumptions As To Joint Rotations

A few comments are presented regarding frames whose joint

rotations (and thus their beam stiffnesses) are not in agreement with
the assumptions made for developing the charts.
It can be shown by structural analysis that the rotation at
point B in the frame of Figure 10 is twice as large as the rotation at
B assumed in the construction of the nomographs. Therefore,
beam BC in the figure is only one-half as stiff as the value assumed Figure 10
for the development of the alignment charts.
The Jackson and Moreland charts can be accurately used for situations in which
the rotations are different from those assumed by making adjustments to the computed
beam stiffnesses before the chart values are read. Relative stiffnesses for situations
other than the one shown in Figure 10 also can be determined by structural analysis.
Table 2 presents correction factors to be multiplied by calculated beam stiffnesses, for
situations where the beam end conditions are different from those assumed for the
development of the charts

Table 2 Multipliers for Rigidly Attached Members

SOLVED PROBLEMS
 14

Problem 1
(a) A W 14x 99 is used as a 9.15-m long pin-connected column. Using the Euler
expression, determine the column’s critical or buckling load. Assume that the
steel has a proportional limit of 250 MPa.
(b) Repeat part (a) if the length is changed to 7.50-m.

Properties of W 14 x 99
A = 18,774 mm2
rx = 156.72 mm
ry = 94.23 mm

Solution:

(a) L = 9.15 m

• Slenderness ratio: r = ry = 94.23 mm

𝐿 9.15 (1000)
= = 97.10
𝑟 94.23

• Elastic or buckling stress

𝜋2 𝐸 𝜋2 (200000)
𝐹𝑒 = = = 209.36MPa
(𝐿 ⁄𝑟 )2 (97.10)2

𝐹𝑒 = 209.36MPa < the proportional limit of 250 MPa

Therefore, the column is in the elastic range

• Fot the elastic buckling load

𝑃𝑐𝑟 = 𝐴 𝑥 𝐹𝑒

𝑃𝑐𝑟 = 18774 x 209.78 = 3930,525 N

(b) L = 7.5 m

• Slenderness ratio: r = ry = 94.23 mm

𝐿 7.5 (1000)
= = 79.6
𝑟 94.23

• Elastic or buckling stress

𝜋2 𝐸 𝜋2 (200000)
𝐹𝑒 = = = 311.53MPa
(𝐿 ⁄𝑟 )2 (79.6)2

𝐹𝑒 = 311.53MPa > the proportional limit of 250 MPa

Therefore, the column is in the inelastic range and Euler equation is

not applicable.

Problem 2
 15

Determine the effective length factor for each of the columns of the frame shown in
Figure 1. if the frame is not braced against sidesway. Use the alignment charts..

Figure 1

Solution:

1. Solve the stiffness factors, I /L:

Note:
a. Modulus of Elasticity E is assumed to be the same for all member (beams
and columns are made of steel) and is therefore neglected in the equation
to calculate G.
b. In the table, the member’s property moment of inertia , I , are taken from
manual handbook.

2. Solve for the G factors for each joint

 16

3. Solve for the effective length factor, K, for each column

Using the chart for frames not braced against sidesway

Alignment chart—sidesway uninhibited (moment frame)

 17

Problem 3
Determine K factors for each of the columns of the frame shown in Figure 2. Here, W
sections have been tentatively selected for each of the members of the frame and their,
I /L, values determined, number in parentheses, and shown in the figure.

Figure 2 .Steel shapes, including their values I / L

Solution:

1. Compute for the G factors for each joint in the frame. In this calculation, the values
for members FI and GJ are multiplied by the appropriate factors.
a. For member FI, the I / L value is multiplied by 2.0, because its far end is
fixed and there is no sidesway on that level.
b. For member, GJ, I / L is multiplied by 1.5, because its far end is pinned and
there is no sidesway on that level.

(hinged support)

(fixed support)

2. Solve for the K factors which are selected from the appropriate alignment chart
 18

(a) Alignment chart—sidesway inhibited (b) Alignment chart—sidesway uninhibited

(braced frame). (moment frame)

Summary

Chart used