Chapter Three: Compression Members 1

CHAPTER THREE
3. COMPRESSION MEMBERS
3.1. Introduction
Compression members are perhaps the most common structural elements in an ordinary
structure and are variously termed as columns, posts, struts or stanchions, etc. A structural
member is considered to be a compression member if it is designed primarily to resist axial
compression, though some bending may also be present and accounted for in the design. If
the bending action is quite significant, the member is termed as a beam-column and designed
in a different way.

Structural action of columns, stanchions, struts and posts is identical; but due to difference in
their usage different names are used.
 Columns are ordinarily used in buildings, are vertical and transmit some actual load or
beam reaction to another column or foundation.
 Stanchions are steel columns made of rolled steel sections (usually built – up) and
carry heavy loads.
 Struts on the other hand are not necessarily vertical and are used as compression
members in roof trusses and bridge trusses. The term post is loosely used for a column
but the end member of a bridge truss is known as the end-post. Similarly, the main
compression members of a roof truss are known as “rafters”.

Under the general category of compression members could be included columns, compression
members in a trussed structure, component parts of frames such as compression flanges of
beams or plate girders.

Design of Steel and timber structures

 Chapter Three: Compression Members 2

The two main differences between tension and compression members are:

A. Tension members are held

straight by means of tensile
loads, while in the case of
compression members, the
compressive loads tend to bend
the member out of the plane of
loading.

B. For riveted or bolted

connections, the net area will
govern the strength of a tension
member, while for compression
members the rivets are assumed
to fill the holes.

This Chapter will present the assessment

and design of structural members that
are acted upon by pure compression
forces; i.e., direct loads with no
moments acting simultaneously.
The main kinds of compression a. Simple compression members
members are as shown in Fig. 3.1.

b. Tapered members c. Stepped columns

Design of Steel and timber structures

 Chapter Three: Compression Members 3

Fig. 3.1d Built up columns

e. Built up members f. Perforated plate columns

Figure 3.1 Main Kinds of Compression Members

Design of Steel and timber structures

 Chapter Three: Compression Members 4

Compression members can fail by yielding, inelastic buckling, or elastic buckling depending
on the slenderness ratio of the members as well as in local buckling that is usually influenced
by the relative thickness of the component elements that constitute the cross section. Members
with low slenderness ratios generally tend to fail by yielding, whereas members with high
slenderness ratios tend to fail by elastic buckling. Most compression members used in
construction have intermediate slenderness ratios, and so the predominant mode of failure is
inelastic buckling.

Member buckling can occur in one of three different modes: flexural, torsional, and flexural-
torsional.

 Flexural buckling occurs in members with doubly symmetric or doubly anti-

symmetric cross sections such as I and Z sections, and in members with singly
symmetric sections such as C, T, equal-legged L and double L.

 Torsional buckling occurs in members with very thin walls.

 Flexural-torsional buckling occurs in members with singly symmetric cross sections

such as C, T, equal-legged L, double L.

Normally, torsional buckling of symmetric shapes and flexural-torsional buckling of un

symmetric shapes are not important in the design of hot-rolled compression members; either
they do not govern or their buckling strengths do not differ significantly from the
corresponding weak-axis flexural buckling strengths. However, torsional and flexural-
torsional buckling modes may govern for sections that have relatively thin component plates.

In addition to slenderness ratio and cross-sectional shape, the behaviour of compression

members is affected by the relative thickness of the component elements that constitute the
cross section. The relative thickness of a component element is qualified by the width-to-
thickness ratio (b/t) of the element. The width-to- thickness ratios of some selected steel
shapes are shown in Fig. 3.2. If the width-to-thickness ratio falls within a limiting value
stipulated by relevant codes and specifications, local buckling of the component element will
not occur. However, if the width-thickness ratio exceeds these stipulated values,
consideration of local buckling in the design of the compression member is required.

Design of Steel and timber structures

 Chapter Three: Compression Members 5

Fig. 3.2 Dimensions of sections

Figure 3.2 Dimensions and axes of sections

Design of Steel and timber structures

 Chapter Three: Compression Members 6

3.2. Classification of Sections

Classification of sections of compression members depends on their failure modes under load.
Four classes of cross-sections are defined, as follows from EN 1993-1-1: 2005 Section 5.5:
– Class 1 (Plastic) cross-sections are those which can form a plastic hinge with the rotation
capacity required from plastic analysis without reduction of the resistance.
– Class 2 (Compact) cross-sections are those which can develop their plastic moment
resistance, but have limited rotation capacity because of local buckling.
– Class 3 (Semi-Compact) cross-sections are those in which the stress in the extreme
compression fiber of the steel member assuming an elastic distribution of stresses can
reach the yield strength, but local buckling is liable to prevent development of the plastic
moment resistance.
– Class 4 (Thin-walled) cross-sections are those in which local buckling will occur before
the attainment of yield stress in one or more parts of the cross-section.

The classification of a cross-section depends on the width to thickness ratio of the parts
subject to compression. Compression parts include every part of a cross-section which is
either totally or partially in compression under the load combination considered. The various
compression parts in a cross-section (such as a web or flange) can, in general, be in different
classes. A cross-section is classified according to the highest (least favourable) class of its
compression parts. The limiting proportions for Class 1, 2, and 3 compression parts should be
obtained from Table 5.2 of EN 1993-1-1, shown as Table 3.1 here. A part which fails to
satisfy the limits for Class 3 should be taken as Class 4.

Design of Steel and timber structures

 Chapter Three: Compression Members 7

Table 3-1 Classification of Compression Sections According to EN 1993-1-1:2005

(sheet 1 of 3): Maximum width-to-thickness ratios for compression parts

Design of Steel and timber structures

 Chapter Three: Compression Members 8

(sheet 2 of 3): Maximum width-to-thickness ratios for compression parts

Design of Steel and timber structures

 Chapter Three: Compression Members 9

(sheet 3 of 3): Maximum width-to-thickness ratios for compression parts

3.3. Effective Cross-Section Properties of Class 4 Cross-Sections

The properties of Class 4 cross-sections may be established by calculation using the effective
widths of the component elements in compression. The effective cross-section properties of
Class 4 cross-sections should be based on the effective widths of the compression parts. The
effective area, Aeff should be determined assuming that the cross section is subject only to
stresses due to uniform axial compression.

Design of Steel and timber structures

 Chapter Three: Compression Members 10

Gross cross section Effective cross section

G centroid of the gross cross

section
G´ centroid of the effective
cross section
1 centroidal axis of the gross
cross section
2 centroidal axis of the
effective cross section
3 non effective zone

Figure 3.3 Class 4 cross-sections - axial force

G centroid of the gross cross

section
G´ centroid of the effective
cross section
1 centroidal axis of the gross
cross section
2 centroidal axis of the
effective cross section
3 non effective zone

Gross cross section Effective cross section

Figure 3.4 Class 4 cross-sections - Bending Moment

Design of Steel and timber structures

 Chapter Three: Compression Members 11

The effective areas of flat compression elements should be obtained using Table 4.1 of EN
1993-1-5: 2004 for internal elements and Table 4.2 of EN 1993-1-5 : 2004 for outstand
elements. The effective area of the compression zone of a plate with the gross cross-sectional
area A should be obtained from:
Aeff  A 3-1

Where ρ is the reduction factor for plate buckling.

The reduction factor ρ may be taken as follows:

- Internal compression elements


  1 .0 for  p  0.673 3-2(a)

 p  0.055(3   ) 
  1.0 for  p  0.673 where (3   )  0 3-2 (b)
 2
p
- Outstand compression elements:

  1 .0 for  p  0.748 3-3(a)
 
 p  0.188   fy bt
  1 .0 for  p  0.748 where  p   3-3(b)
 2  cr 28.4 k
p

 is the stress ratio determined in accordance with section 4.4 of EN 1993-1-5: 2004

b is the appropriate width to be taken as follows:
bw for webs;
b for internal flange elements (except RHS);
b-3t for flanges of RHS;
c for outstand flanges;
h for equal-leg angles;
h for unequal-leg angles;
kσ is the buckling factor corresponding to the stress ratio  and boundary conditions.

For long plates, kσ is given in Table 4.1 or Table 4.2 of EN 1993-1-5 : 2004.
t is the thickness
σcr is the elastic critical plate buckling stress see equation (A.1) in Annex A.1(2) and Table
4.1 and Table4.2 of EN 1993-1-5: 2004.

For flange elements of I-sections and box girders the stress ratio  used in Table 4.1 or
Table 4.2 of EN1993-1-5:2004 should be based on the properties of the gross cross-sectional
area, due allowance being made for shear lag in the flanges if relevant. For web elements the

Design of Steel and timber structures

 Chapter Three: Compression Members 12

stress ratio  used in Table 4.1 of EN 1993-1-5:2004 should be obtained using a stress
distribution based on the effective area of the compression flange and the gross area of the
web.

Figure 3.5 Definition of Outstand and Internal compression elements

Table 3-2 Internal Compression Elements (Table 4.1 of EN 1993-1-5)

Design of Steel and timber structures

 Chapter Three: Compression Members 13

Table 3-3 Outstand Compression Elements(Table 4.2 of EN 1993-1-5 )

3.4. Effective Buckling Lengths

The effective length of a member will depend on its end conditions. EC3 gives no direct
guidance on calculating the buckling length, therefore it is acceptable to use those given in the
previous EBCS code.
l e  kL 3-4
Where, le is effective buckling length
k is effective length factor
L is actual unbraced length

Design of Steel and timber structures

 Chapter Three: Compression Members 14

Effective Length Factor

The effective length factor, k is a factor which, when multiplied by the actual unbraced length
L of an end-restrained compression member, will yield an equivalent pinned-ended member
whose buckling strength is the same as that of the original end-restrained member.

For a prismatic member, the effective length factor can be determined from different methods
as discussed in following successive pages:

METHOD I: k factor table

The following general recommendations can be used if support condition can be represented
as shown in the following Figure 3.4

Figure 3.6 k factor Table

METHOD II: Using Nomo graphs

The effect of end restraint is quantified by the two end restraint factors G A and GB where the
subscripts A and B refer to the joints at the two ends of the member being considered and G is
defined as:

sum of flexural stiffness of all columns meeting at the jo int

G
sum of flexural stiffness of all beams meeting at the join


 ( EI / L) columns

 ( EI / L) beams

Design of Steel and timber structures

 Chapter Three: Compression Members 15

3-5

Figure 3.7 k factor alignment (Nomo graphs)

Note that if the end of the member is fixed, the theoretical value of G is 0, but a G value of 1
is recommended for use. On the other hand, if the end of the member is pinned, the
theoretical value of G is infinity, but a G value of 10 is recommended for use. The rational
behind the foregoing recommendations is that no support in reality can be truly fixed or
pinned.

Once the G factors are calculated, the effective length factor can be obtained from the
appropriate alignment chart. The chart for sideways-inhibited frames applies to frames that
are braced in such a way that relative displacement between two ends of the member is
negligible. The chart for side sway-inhibited frames applies to frames in which relative
displacement between member ends is not negligible. Although the charts were developed
assuming elastic behaviour for all members, inelasticity in the columns can be accounted for
by multiply the end-restrained factors by the quantity Et/ E, where Et is the tangent modulus.

The alignment charts shown in Fig 3.7 where developed based on a number of simplifying
assumptions; as a result, they do not always give accurate results, especially for members in

L ( P EI )

Design of Steel and timber structures

 Chapter Three: Compression Members 16

frames for which the parameter varies significantly from column to column in a
given story. The alignment charts also fail to give accurate results for frames that contain
leaner columns.

METHOD III: According to EBCS-3,1995

1. For the theoretical models shown in the next pages the distribution factors 1 and  2
is given by,
kc
1  3-6(a)
k c  k11  k12
kc
2  3-6(b)
k c  k 21  k 22
Where, kc is the column stiffness I/L.
kij is the effective beam stiffness coefficient

Figure 3.8 Distribution factors for Columns

Design of Steel and timber structures

 Chapter Three: Compression Members 17

2. For continous columns shown in the next pages the distribution factors 1 and  2 is
given by,
k c  k1
1  3-7(a)
k c  k1  k11  k12
kc  k2
2  3-7(a)
k c  k 2  k 21  k 22

Where, k1 and k2 is the stiffness for the adjacent columns

k11&k12 and k21&k22 are the effective stiffness coefficients for top and bottom beam
respectively

Figure 3.9 Distribution factor for continious columns

Design of Steel and timber structures

 Chapter Three: Compression Members 18

Table 3-4 Effective stiffness coefficients for beams not subjected to axial loads.

a) In Non-sway Mode
I
 0.5  0.14(1   2 )  0.055(1   2 ) 2 3-8(a)
L
Or
I 1  0.145(1   2 )  0.265(1 2 ) 
  3-8(b)
L  2  0.364(1   2 )  0.2471 2 
b) In Sway Mode
0.5
I 1  0.2(1   2 )  0.12(1 2 ) 
  3-9
L  1  0.8(1   2 )  0.601 2 

3.5. Elastic Critical Buckling Load

Ncr is the elastic critical buckling load for the relevant buckling mode based on the gross properties of
the cross section.

N cr  l EI
2
2 3-10
e

3.5.1. Derivation of the elastic critical buckling load

Mathematical Background (Kreyszig, 2011)

Consider second-order homogeneous linear ODEs (Ordinary Differential Equations) whose
coefficients a and b are constant,
y ' ' ay'by  0
And a real general solution of this second-order homogeneous linear ODE is
y  e  ax / 2 ( A cosx  B sin x) (A, B arbitrary)

Design of Steel and timber structures

 Chapter Three: Compression Members 19

A general solution of the ODE,

y ' ' 2 y  0 (  constant, not zero)
is
y  A cosx  B sin x

Derivation of the elastic critical buckling load

Elastic critical buckling load is derived for the simply supported column under axial
compressive load.
From Figure 3.10(c), using equilibrium,
 Fx  0   N  F ( x)  0 → F ( x)   N

 Fz  0   V A  S ( x)  0 → S ( x )  0 since VA = 0.

 M  0 CCW+
A M ( x)  F ( x) w( x)  0

M ( x)  N w( x)  0 (3-10-1)

(a) (b) (c)

Figure 3.10 Representation of: (a) undeflected geometries of Column, (b) deflected
geometries of Column and (c) “cut” column with resultant loads

Design of Steel and timber structures

 Chapter Three: Compression Members 20

The deflection w(x) is governed by the bending relation:

d 2 w M ( x) d 2w
Curvature:    → M ( x)  EI (3-10-2)
dx 2 EI dx 2
Substituting (3-10-2) into (3-10-1):
d 2w
EI  N w( x)  0 (3-10-3)
dx 2
And this can be rewritten as:
N
w ''  w0 (3-10-4)
EI
This is one type of homogeneous linear ODEs (Ordinary Differential Equations) with constant
coefficients. The aim is now to obtain the solution for the ODE in Equation (3-10-4),

N N
Let,  2  → 
EI EI
Now we can write the solution for the ODE in Equation (3-10-4) as:

N N
w( x)  A cos x  B sin x
EI EI
For simply supported column,
w  0
 d 2w
@x  0  2
d w  EI 0
 
2
 M 0 EI dx
 dx 2

w  0
 d 2w
@x  L  d w  EI
2 0
M  0  EI dx 2
 dx 2

N N
w( x)  A cos x  B sin x
EI EI

dw N N N N
 A sin x B cos x recall that (sin x)'  cos x
dx EI EI EI EI

d 2w N N N N
2
 A cos x B sin x
dx EI EI EI EI

Design of Steel and timber structures

 Chapter Three: Compression Members 21

N N
@x  0 w  0  A cos x  B sin x  0 →A = 0.
EI EI

d 2w N N N N N
2
0 A cos x B sin x  0 → B sin x0
dx EI EI EI EI EI

N N N
@x  L w  0  A cos L  B sin L  0 → B sin L0
EI EI EI

d 2w N N N N N N
2
0 A cos L B sin L  0→ B sin L0
dx EI EI EI EI EI EI
N
 B  0  B  0 (trivial solution)
EI

N N
 sin L  0→ L  n where, n is an integer.
EI EI
Buckling occurs in simply supported column if,
n 2 2 EI
N ------Eigen Values
L2
n x
w  B sin -----Eigen modes
L

n=1  2 EI
N1 
L2
n=2 4 2 EI
N2 
L2

n=3 9 2 EI
N3 
L2

Therefore, the elastic critical buckling load is given by,

N cr  l EI
2
2 , where, le is effective (critical) length.
e

Design of Steel and timber structures

 Chapter Three: Compression Members 22

3.6. Limiting Slenderness Ratio

The governing slenderness ratio (le/r) of compression members preferably should not exceed
200.

3.7. Design Criteria for Compression Members: The Eurocode 3

Approach
A number of design checks are required for compression members. In all cases, it is
recommended that the forces and moments in the members are derived from an elastic global
analysis.

In addition to cross-sectional resistance, consideration should be given to overall buckling of

members. Members in compression are susceptible to a number of buckling modes including
local buckling (Class 4 or thin sections only), flexural buckling torsional buckling and
flexural-torsional buckling. The last three modes are overall buckling modes involving the
whole member. Although all modes should be considered, point-symmetric open sections
(such as Z-shapes) are usually more prone to torsional buckling while single-symmetric open
sections to flexural-torsional buckling. Local buckling while single-symmetric open section
to introducing certain modification pertinent to specific code implementation.

To take into account the various imperfections which the Euler formula does not allow for,
the Eurocode uses the Perry-Robertson approach. Table 1 shows the checks required for both
slender and stocky columns:

Table 3-5 Resistance checks required for slender and stocky columns

Slender column Stocky column

 
  0.2   0.2
Cross-section Resistance check, Nc,Rd √ √
Buckling Resistance check, Nb,Rd √

3.7.1. Cross-Section Resistance

EN 1993-1-1 Clause 6.2.4 states that the design value of the Compression force (NEd) must be
less than the design cross-sectional resistance of the sections to uniform compression force
(Nc,Rd).
N Ed
 1.0 3-11
N c, Rd
Cross-section resistance in compression depends on cross-section classification.
Accordingly:-

Design of Steel and timber structures

 Chapter Three: Compression Members 23

For Classes 1, 2 and 3:

Af y
N c, Rd  3-12
 Mo

For Class 4 sections:

Aeff f y
N c , Rd  3-13
 Mo
The partial safety factors are MO = 1.0

Cross-section Resistance Check Summary

1. Determine the design compression force, NEd
2. Choose a section and determine the section classification
3. Determine Nc,Rd, using Equation 3.12 for Class 1,2 and 3 sections, and Equation 3.13
for Class 4 sections.
4. Carry out the cross-sectional resistance check by ensuring Equation 3.11 is satisfied.

3.7.2. Member Buckling Resistance

EN 1993-1-1 Clause 6.3.1 states that the design values of the Compression force (NEd) must be less
than the buckling resistance of the compression member (Nb,Rd).
N
Ed  1.0 3-14
N
b, Rd
Similar to cross-section resistance, buckling resistance is dependent on the cross-section
classification.
For sections with Classes 1, 2 and 3:
Af y
Nb, Rd  3-15
 Mo

For Class 4 sections:

Aeff f y
Nb, Rd  3-16
 M1
Aeff is the effective cross-section for Class 4 cross-sections
A = gross area
 = a reduction factor accounting for buckling
The partial safety factors are MO = 1.0 and M1 = 1.0


Non-dimensional slenderness, 
For sections with classes 1, 2 and 3:

Design of Steel and timber structures

 Chapter Three: Compression Members 24

 Af y  le 1
 or  3-17
N cr i 1
Where i is radius of gyration given by i  I A

For class 4 sections:

Aeff
 Aeff f y  l
 or   e
A 3-18
N cr i 1

1   E f y  93.9 3-19

Imperfection factor, 
 is an imperfection factor, first you will need to determine the required buckling curve from
Table 6.2 and refer to Table 6.1 to get the value of  .
Table 3-6 Imperfection Factor (Table 6.1 of EN 1993-1-1)
Buckling ao A b C d
Curve
Imperfection 0.13 0.21 0.34 0.49 0.76
factor, 

Reduction Factor, 

1
 0.5
 1.0 3-20
 2
   2   
 
Where,
  2 
  0.51   (  0.2)    3-21
 

Alternatively,  may be read from Figure 3.10 (Figure 6.4 of EN 1993-1-1) by using  and
the required buckling curve.

NOTE: Nc.Rd, design compression resistance of the cross-section, taken as a smaller of either
the uniform compression force, Nc,Rd of the gross section from Equation 3.12 (for class 1, 2
and 3 sections) and Equation 3.13 (for class 4 sections) or the design local buckling
resistance, Nb,Rd of the gross section from Equation 3.15 (for class 1, 2 and 3 sections) and
Equation 3.16 (for class 4 sections)

Table 3-7 Selection of buckling curve for a cross-section (Table 6.2 of EN 1993-1-1)

Design of Steel and timber structures

 Chapter Three: Compression Members 25

Design of Steel and timber structures

 Chapter Three: Compression Members 26

Figure 3.11 Buckling curves

Buckling Resistance Check Summary

1. Determine the design axial load, NEd
2. Choose a section and determine the class.
3. Calculate the effective length, le.
4. Calculate Ncr using the effective length, le, and E and I which are section
properties.

5. Calculate 
6. Determine α by first determining the required buckling curve from Table 3-7
and then reading off the required value of α from Table 3.6.

7. Calculate Φ by substituting in the values of α and  .

8. Calculate χ by substituting in the values of Φ and  .
9. Determine the design buckling resistance of the member by using Equation
3.15 or 3.16 and substituting in the value of χ .
10. Make sure that the conditions of Equation 3.14 are satisfied.

Design of Steel and timber structures