DESIGN OF AXIALLY LOADED COLUMNS

DESIGN OF AXIALLY LOADED COLUMNS

10
1.0 INTRODUCTION

In an earlier chapter, the behaviour of practical columns subjected to axial compressive

loading was discussed and the following conclusions were drawn.

 Very short columns subjected to axial compression fail by yielding. Very long
columns fail by buckling in the Euler mode.
 Practical columns generally fail by inelastic buckling and do not conform to the
assumptions made in Euler theory. They do not normally remain linearly elastic upto
failure unless they are very slender
 Slenderness ratio (/r) and material yield stress (fy) are dominant factors affecting the
ultimate strengths of axially loaded columns.
 The compressive strengths of practical columns are significantly affected by (i) the
initial imperfection (ii) eccentricity of loading (iii) residual stresses and (iv) lack of
defined yield point and strain hardening. Ultimate load tests on practical columns
reveal a scatter band of results shown in Fig. 1. A lower bound curve of the type
shown therein can be employed for design purposes.

c Test data (x) from collapse tests

(Mpa) on practical columns
fy xx
xx
x
xx Euler curve
200 x
x
x
x x
x
xx Design curve
100 x
xx
x
x x

50 100 150
Slenderness  (/r)

Fig. 1 Typical column design curve

© Copyright reserved

Version II 10 - 1
 DESIGN OF AXIALLY LOADED COLUMNS

2.0 HISTORICAL REVIEW

Based on the studies of Ayrton & Perry (1886), the British Codes had traditionally based
the column strength curve on the following equation.
( f y   c ) (  e   c )    e  c (1)

where

fy = yield stress

c = compressive strength of the column obtained from the positive root of the
above equation
 2E
e = Euler buckling stress given by 2 (1a)

 = a parameter allowing for the effect of lack of straightness and eccentricity
of loading.

 = Slenderness ratio given by (/r)

In the deviation of the above formula, the imperfection factor  was based on

y 
 (2)
r2

where y = the distance of centroid of the cross section to the extreme fibre of the
section.

 = initial bow or lack of straightness

r = radius of gyration.

Based on about 200 column tests, Robertson (1925) concluded that the initial bow ()
could be taken as length of the column/1000 consequently  is given by

 y 
   0.001  .  
 r  r


       (3)
r

where  is a parameter dependent on the shape of the cross section.

Version II 10 - 2
 DESIGN OF AXIALLY LOADED COLUMNS

c
(Mpa)
fy

200 Euler curve

Design curve with  = 0.003

100

50 100 150

Fig.2 Robertson’s Design Curve

Robertson evaluated the mean values of  for many sections as given in Table 1:

Table1:  values Calculated by Robertson

Column type  Values

Beams & Columns about the major axis 0.0012

Rectangular Hollow sections 0.0013

Beams & Universal columns about the minor axis 0.0020

Tees in the plane of the stem 0.0028

He concluded that the lower bound value of  = 0.003 was appropriate for column
designs. This served as the basis for column designs in Great Britain until recently. The
design curve using this approach is shown in Fig. 2. The Design method is termed "Perry-
Robertson approach"

3.0 MODIFICATION TO THE PERRY ROBERTSON APPROACH

3.1 Stocky Columns

It has been shown previously that very stocky columns (e.g. stub columns) resisted loads
in excess of their squash load of fy.A (i.e. theoretical yield stress multiplied by the area
of the column). This is because the effect of strain hardening is predominant in low

Version II 10 - 3
 DESIGN OF AXIALLY LOADED COLUMNS

values of slenderness (). Equation (1) will result in column strength values lower than fy
even in very low slenderness cases. To allow empirically for this discrepancy, recent
British and European Codes have made the following modification to equation 3 given by

 = 

In the unmodified form this will cause a drop in the calculated value of column strength
even for very low values of slenderness. Such columns actually fail by squashing and
there is no drop in observed strengths in such very short columns. By modifying the
slenderness,  to ( - 0) we can introduce a plateau to the design curve at low
slenderness values. In generating the British Design (BS: 5950 Part-1) curves

0  0.2  E / fy  was used as an appropriate fit to the observed test data, so that we
obtain the failure load (equal to squash load) for very low slenderness values. Thus in
calculating the elastic critical stress, we modify the formula used previously as follows:

 2E
e  for all values of   0 (4)
  0 2
Note that no calculations for e is needed when   e as the column would fail by
squashing at fy.

c Test data (x) from collapse tests

(Mpa) on practical columns
fy xx
x
x
x
200 x Euler curve
x x
x
xx
x x
xx
100 x x x Robertson’s curve
xx
x x
x x
x

0 50 100 150
[0.2 (E/fy )1/2] 

Fig.3 Strut curve with a plateau for low slenderness values

Version II 10 - 4
 DESIGN OF AXIALLY LOADED COLUMNS

3.2 Influence of Residual Stresses

Reference was made earlier to the adverse effect of locked-in residual stresses on column
strengths (see Fig. 4). Studies on columns of various types carried out by the European
Community have resulted in the recommendation for adopting a family of design curves
rather than a single “Typical Design Curve” shown in Fig. 3. Typically four column
curves are suggested in British and European codes for the different types of sections
commonly used as compression members [See Fig. 5(a)]. In these curves,  =  ( -0)

where 0  0.2 E
fy and the  values are varied corresponding to various sections.

Thus all column designs are to be carried out using the strut curves given in [Fig. 5(a)].

T
T
T T
C T C
C T C
C
C
C C 6000
xx   36.1
C C
166
C
. 1
T T
C T C
T
T T 3000
 yy   58.1
51.6
Rolled beam Welded box Rolled column

Fig. 4 Distribution of residual stresses

a)  = 0.002 (Curve A)
b)  = 0.0035 (Curve B)
300 c)  = 0.0055 (Curve C)
250 d)  = 0.008 (Curve D)
200
fy 150
Compressive 100
strength c
(Mpa) 50
0
0 50 100 150 200 250
Slenderness, 

Fig. 5(a) Compressive strength curves for struts for different values of 

Version II 10 - 5
 DESIGN OF AXIALLY LOADED COLUMNS

The selection of an appropriate curve is based on cross section and suggested curves are
listed in Table 2.

c
(Mpa) 20N/mm2
fy

150

50

50 100 150


Fig. 5(b) Compressive strength of welded sections

Table 2: Choice of appropriate values of 

Sections Axis of buckling

X -X Y-Y
Hot rolled structural hollow sections  = 0.002 (Curve A)  = 0.002 (Curve A)
Hot rolled I section  = 0.002 (Curve A)  = 0.0035 (Curve B)
Welded plate
I section (up to 40 mm thick)  = 0.0035 (Curve B)  = 0.0055 (Curve C)
I section (above 40 mm thick)  = 0.0035 (Curve B)  = 0.008 (Curve D)
Welded Box Section
(Up to 40 mm thick)  = 0.0035 (Curve B)  = 0.0035 (Curve B)
(Over 40 mm thick)  = 0.0055 (Curve C)  = 0.0055 (Curve C)
Rolled I section with Welded cover
plates (Up to 40mm thick)  = 0.0035 (Curve B)  = 0.002 (Curve A)
(Over 40mm thick)  = 0.0055 (Curve C)  = 0.0035(Curve B)
Rolled angle , Channel ,T section
Compound sections - Two rolled Check buckling about ANY axis With
sections back to back, Battened or  = 0.0055 (Curve C)
laced sections

Version II 10 - 6
 DESIGN OF AXIALLY LOADED COLUMNS

Note: For sections fabricated by plates by welding the Value of fy should be reduced by
20 N/mm2 [See Fig. 5(b)].

For computational convenience, formulae linking c and  are required. The lower root of
equation (1) [based on Perry - Robertson approach] represents the strut curves given Fig.
3 and BS: 5950 Part - 1.

 c     2  f y e  f y (5)

f y   1 e
where,   (6 a )
2

and     0  (6b)

3.3 Types of Column Sections

Steel suppliers manufacture several types of sections, each type being most suitable for
specific uses. Some of these are described below. It is important to note that columns
may buckle about X, Y, V or U axis. It is necessary to check the safety of the column
about several axes, so that the lowest load that triggers the onset of collapse may be
identified.

Universal Column (UC) sections have been designed to be most suitable for compression
members. They have broad and relatively thick flanges, which avoid the problems of
local buckling. The open shape is ideal for economic rolling and facilitates easy beam-to-
column connections. The most optimum theoretical shape is in fact a circular hollow
section (CHS) which has no weak bending axis. Although these have been employed in
large offshore structures like oil platforms, their use is somewhat limited because of high
connection costs. Rectangular Hollow Sections (RHS) have been widely used in multi-
storey buildings satisfactorily. For relatively light loads, (e.g. Roof trusses) angle
sections are convenient as they can be connected through one leg. Columns, which are
subjected to bending in addition to axial loads, are designed using Universal beams (UB).
Values of  to be used in all these cases are available in published codes and are typically
0.002, 0.0035, 0.0055 and 0.008 depending on the type of the column section and the
extent of residual stresses present. The values suggested in British Codes are included in
Fig. 5(a) and Table 3 gives the maximum axial compressive stresses using these  values,
for fy equals to 250 M pa.

3.4 Heavily Welded Sections

Although both hot rolled sections and welded sections have lock-in residual stresses, the
distribution and magnitude differ significantly. Residual stresses due to welding are very
high and can be of greater consequence in reducing the ultimate capacity of compression

Version II 10 - 7
 DESIGN OF AXIALLY LOADED COLUMNS

members. The British Code recommends that the design strength values be reduced by
20 N/mm2 and the resulting curve would be adequate to account for the reduced buckling
capacity [See Fig. 5(b)]

4.0 EFFECTIVE LENGTH OF COLUMNS

Pin-ended axially loaded columns are rare in practice. However, available analytical
studies – starting from Euler's Theory – are all based on the assumption of pin-ended
columns. Results of experiments carried out on pin ended struts for over 100 years are
well documented in published literature. Hence we employ the columns hinged at both
ends as the standard case for purposes of comparison and the concept of “effective
lengths” for columns having various boundary conditions was presented in a previous
chapter. By employing the effective lengths (e) appropriate to the column under
consideration, it is usually possible to predict its strength. e may regarded as the
equivalent length of a pin-ended column having the same cross section, which would be
expected to have the same strength and stiffness as the column being designed.

Ideally fixed ended columns are difficult to achieve in practice. BS: 5950 Part-1
recommends the following effective lengths for design purposes

Column pin-ended at both ends e = 1.0

Column pin-ended at one end and fixed at the other e = 0.85

Column fixed at both ends e = 0.7 

Column fixed at one end and on roller support at the other e = 1.2 

Column fixed at one end and free at the other e = 2.0 

(Note: These values are NOT the same as those contained in IS: 800-1984. For the
second and third boundary conditions IS: 800-1984 suggests 0.80  and 0.65 
respectively which are also followed by Australian and American practices)

Designs of columns have to be checked using the appropriate effective length for
buckling in both strong and weak axes. A worked example illustrating this concept is
appended to this chapter.

5.0 STEPS IN THE DESIGN OF AXIALLY LOADED COLUMNS

The procedure for the design of an axially compressed column is as follows:

(i) Assume a suitable trial section and classify the section in accordance with the
classification in "Local Buckling and Section classification" chapter. (If, the
section is slender then apply appropriate correction factor)

Version II 10 - 8
 DESIGN OF AXIALLY LOADED COLUMNS

(ii) Arrive at the effective length of the column by suitably considering the end
conditions.

(iii) Calculate the slenderness ratios ( values) in both minor and major axes direction
and also calculate 0 using the formula given below:

0  0.2 E
fy
(iv) Calculate e values along both major and minor axes from equation 1(a)

(v) Calculate  = [( - 0)], using appropriate  values from Table 2

(vi) Calculate  from equation (6a) and c from equation (5)

(vii) Compute the load that the compression member can resist (c A)

(viii) Calculate the factored applied load and check whether the column is safe against
the given loading. The most economical section can be arrived at by trial and
error, i.e. repeating the above process.

6.0 CROSS SECTIONAL SHAPES FOR COMPRESSION MEMBERS AND

BUILT- UP COLUMNS
Although theoretically we can employ any cross sectional shape to resist a compressive
load we encounter practical limitations in our choice of sections as only a limited number
of sections are rolled by steel makers and there are sometimes problems in connecting
them to the other components of the structure. Another limitation is due to the adverse
impact of increasing slenderness ratio on compressive strengths; this virtually excludes
the use of wide plates, rods and bars, as they are far too slender. It must be specially
noted that all values of slenderness ratio referred to herein are based on the least
favourable value of radius of gyration, so that (/r) is the highest value about any axis.

6.1 Rolled Steel Sections

Some of the sections employed as compression members are shown in Fig. 6. Single
angles [Fig 6(a)] are satisfactory for bracings and for light trusses. Top chord members
of roof trusses are usually made up of twin angles back to back [Fig 6(b)]

Double angle sections shown in Fig. 6(b) are probably the most commonly used members
in light trusses. The pair of angles used has to be connected together, so they will act as
one unit. Welds may be used at intervals – with a spacer bar between the connecting
legs. Alternately “ stitch bolts”, washers and “ring fills” are placed between the angles to
keep them at the proper distance apart (e.g. to enable a gusset to be connected).

When welded roof trusses are required, there is no need for gusset plates and T sections
[Fig 6(c)] can be employed as compression members.

Version II 10 - 9
 DESIGN OF AXIALLY LOADED COLUMNS

Single channels or C-sections [Fig. 6(d)] are generally not satisfactory for use in
compression, because of the low value of radius of gyration. They can be used if they
could be supported in a suitable way in the weak direction.

Circular hollow sections [Fig. 6(e)] are perhaps the most efficient as they have equal
values of radius of gyration about every axis. But connecting them is difficult but
satisfactory methods have been evolved in recent years for their use in tall buildings.

The next best in terms of structural efficiency will be the square hollow sections (SHS)
and rectangular hollow sections, [Fig. 6(f)] both of which are increasingly becoming
popular in tall buildings, as they are easily fabricated and erected. Welded tubes of
circular, rectangular or square sections are very satisfactory for use as columns in a long
series of windows and as short columns in walkways and covered warehouses. For many
structural applications the weight of hollow sections required would be only 50% of that
required for open profiles like I or C sections.

(c) Tee
(a) Single Angle (b) Double Angle

(d) Channel (e) Circular Hollow (f) Rectangular Hollow

Section (CHS) Section (RHS)

Fig 6: Cross Section Shapes for Rolled Steel Compression Members

The following general guidance is given regarding connection requirements:

When compression members consist of different components, which are in contact with
each other and are bearing on base plates or milled surfaces, they should be connected at
their ends with welds or bolts. When welds are used, the weld length must be not less
than the maximum width of the member. If bolts are used they should be spaced
longitudinally at less than 4 times the bolt diameter and the connection should extend to
at least 1 ½ times the width of the member.

Version II 10 - 10
 DESIGN OF AXIALLY LOADED COLUMNS

Single angle discontinuous struts connected by a single bolt are rarely employed. When
such a strut is required, it may be designed for 1.25 times the factored axial load and the
effective length taken as centre to centre of the intersection at each end. Single angle
discontinuous struts connected by two or more bolts in line along the member at each end
may be designed for the factored axial load, assuming the effective length to be 0.85
times the centre to centre distance of the intersection at each end.

For double angle discontinuous struts connected back to back to both sides of a gusset or
section by not less than two bolts or by welding, the factored axial load is used in design,
with an effective length conservatively chosen. (A value between 0.7 and 0.85 depending
upon the degree of restraint provided at the ends).

All double angle struts must be tack bolted or welded. The spacing of connectors must be
such that the largest slenderness ratio of each component member is neither greater than
60 nor less than 40. A minimum of two bolts at each end and a minimum of two
additional connectors spaced equidistant in between will be required. Solid washers or
packing plates should be used in-between if the leg width of angles exceed 125 mm.

For member thickness upto 10 mm, M16 bolts are used; otherwise M20 bolts are used.
Spacing of tack bolts or welds should be less than 600 mm.

The following guide values are suggested for initial choice of members:

(i) Single angle size : 1/30 of the length of the strut  r 150
(ii) Double angle size : 1/35 of the length of strut r 100120
(iii) Circular hollow sections diameter = 1/40 length  r 100

6.2 Built-up or fabricated Compression Members

When compression members are required for large structures like bridges, it will be
necessary to use built-up sections. They are particularly useful when loads are heavy and
members are long (e.g. top chords of Bridge Trusses). Built up sections [illustrated in
Fig. 7(a) and 7(b)] are popular in India when heavy loads are encountered. The cross
section consists of two channel sections connected on their open sides with some type of
lacing or latticing (dotted lines) to hold the parts together and ensure that they act
together as one unit. The ends of these members are connected with “batten plates” which
tie the ends together. Box sections of the type shown in Fig. 7(a) or 7(b) are sometimes
connected by solid plates also (represented by straight lines).

A pair of channels connected by cover plates on one side and latticing on the other
[Fig.7(c)] is sometimes used as top chords of bridge trusses. The gussets at joints can be
conveniently connected to the inside of the channels. Plated I sections or built-up I
sections are used when the available rolled I sections do not have sufficient strengths to

Version II 10 - 11
 DESIGN OF AXIALLY LOADED COLUMNS

resist column loads [Fig 7(d) and 7(e)]. For very heavy column loads, a welded built up
section [See Fig. 7(f)] is quite satisfactory.

(a) Box Section (b) Box Section (c) Box Section

(f) Built-up Box

(d) Plated I Section (e) Built - up I Section
Section

Fig 7: Cross Section Shapes for Built - up or fabricated Compression Members

Built up columns made up of solid webs:

In these columns the webs are solid and continuous [See Fig 7(d), 7(e), and 7(f)]. Flange
plates or channels may be used in combination with rolled sections to enhance the load
resistance of the commonly available sections, which are directly welded or bolted to
each other. For preliminary calculations, approximate values of radii of gyration given in
Fig. 8 for various built-up sections may be employed.

The lateral dimension of the column is generally chosen at around 1/10 to 1/15 of the
height of the column. For purposes of detailing the connection between the flange cover
plates or the outer rolled sections to the flanges of the main rolled section, it is customary
to design the fasteners for a transverse shear force equal to 2.5% of the compressive load
of the column. (Connection Design is dealt later in this resource).

Open Web Columns:

In Fig. 9 the two channel sections of the column are connected together by batten plates
or laces which are shown by dotted lines. A typical lacing or batten plate is shown in Fig.
9. Laced columns (also called latticed columns) generally carry 10% more load than
battened columns for the same area of cross section. This necessitates a 10% increase in
the slenderness ratio for battened columns. All columns should be tied at the ends by tie
plates or end battens to ensure a satisfactory performance.

Version II 10 - 12
 DESIGN OF AXIALLY LOADED COLUMNS

y
y
B
rx =0.29 D
D rx =0.31 D
x ry =0.29 B x
D ry =0.215 B
B
y =B (0.21+0.025)
y y
rx =0.4 D
D
x Dm =mean D
B y
y x x rx =0.21 D
D ry =0.21 B
r =0.25 D rz =0.19 D
y
D x B

y
z
y
rx =0.32 D
x r =0.21 B
x y
Dm D = B (0.19 +0.025)
r =0.35 Dm

z
B
y
y

D rx = ry = 0.31 D
x rz =0.197 D x rx=0.42D
D ry=0.42B
B

B
y
rx =0.29 D
D x ry =0.32 B y
B

By
r3 =0.18(D+B)/2 rx=0.36 D
D x ry=0.45 B
x
D

Fig 8: Approximate radii of gyration

(Continued in next page)

Version II 10 - 13
 DESIGN OF AXIALLY LOADED COLUMNS

y y

x rx=0.36 D x D rx=0.36 D
D ry=0.6 B x ry=0.53 B
x
B
B

y
y

x rx=0.38 D
rx=0.39 D D ry=0.22 B
x D ry=0.55 B

B
B

y y

rx=0.4 D
x x r =0.42 D
D ry=0.21 B D x

Fig 8: Approximate radii of gyration

6.3 Design Considerations for Laced and Battened Columns

The two channel constituents of a laced column, shown in Fig. 9(a) and 9 (b) have a
tendency to buckle independently. Lacing provides a tying force to ensure that the
channels do not do so. The load that these tying forces cause is generally assumed to
cause a shearing force equal to 2.5% of axial load on the column. (Additionally if the
columns are subjected to moments or lateral loading the lacing should be designed for the
additional bending moment and shear). To prevent local buckling of unsupported lengths
between the two constituent lattice points (or between two battens), the slenderness ratio
of individual components should be less than 50 % or 70% of the slenderness ratio of the
built up column (whichever is less).

In laced columns, the lacing should be symmetrical in any two opposing faces to avoid
torsion. Lacings and battens are not combined in the same column. The inclination of
lacing bars from the axis of the column should not be less than 400 nor more than 700.
The slenderness ratio of the lacing bars should not exceed 145. The effective length of
lacing bars is the length between bolts for single lacing and 0.7 of this length for double

Version II 10 - 14
 DESIGN OF AXIALLY LOADED COLUMNS

lacing. The width of the lacing bar should be at least 3 times the diameter of the bolt.
Thickness of lacing bars should be at least 1/40th of the length between bolts for single
lacing and 1/60 of this length for double lacing (both for welded and bolted connections).

(a) Single Lacing (b) Double Lacing (c) Battens

Fig. 9 Built-up column members

In the Western world, it was common practice to “build-up” the required cross sectional
area of steel compression members from a number of smaller sections. Since then, the
increasing availability of larger rolled steel sections and the high fabrication costs have
resulted in a very large drop in the use of built-up compression members. The disrepute
of built-up compression members arises from the unrealistic expectation of many
designers that a built-up member should have the same capacity of a solid member and
also behaves in every respect in an identical manner to the latter. Early designers were
disappointed to discover that this was just not possible. A contributory cause to this
decline was the restrictive clauses introduced in many western design codes, following
the Quebec Bridge Failure in 1907. However there is a continuing use of built-up
members, where stiffness and lightness are required, as in Transmission line Towers.
There is, however, a wide spread use of built up compression members in the developing

Version II 10 - 15
 DESIGN OF AXIALLY LOADED COLUMNS

world, India included, largely because of the non-availability of heavier rolled sections
and the perception that fabrication of members is cheaper.

In practical columns, the battens are very stiff and as they are normally welded to the
vertical members, they can be considered as rigid connectors.

To allow approximately for this behaviour, a modified formula for calculating the
effective slenderness (b) of battened columns has been widely employed. This ensures
that the Perry-Robertson approach outlined earlier could be used with a modified value
for slenderness given by b defined below.

λ b  λ 2f λ 2 (7)

where,

f = lower value of slenderness of the individual vertical members between batten

intervals and
 = slenderness of the overall column, using the radius of gyration of the whole built up
section.

This equation, though an approximation, has been shown by Porter and Williams of
Cardiff University actually to give accurate and safe values over the entire range of
practical parameters for uniform columns with normal depth battens. In calculating the
values of e in the Perry Robertson equation, [equation 1 and 1(a)] b is to be employed
in place of , using slenderness ratio in defined in equation (3). The imperfection ( b) is
calculated from

b = 0.0055(b) (8)

The strength of the battened column is evaluated from

f y  η b  1σ e  f y  η b  1σ e 
2

σc      f y σ e (9)
2  2 

b = effective slenderness with b computed as given in equation (8)

e = calculated using b values given in equation (7)

7.0 BASE PLATES FOR CONCENTRICALLY LOADED COLUMNS

The design compressive stress in a concrete footing is much smaller than it is in a steel
column. So it becomes necessary that a suitable base plate should be provided below the
column to distribute the load from it evenly to the footing below. The main function of
the base plate is to spread the column load over a sufficiently wide area and keep the
footing from being over stressed.

Version II 10 - 16
 DESIGN OF AXIALLY LOADED COLUMNS

For a purely axial load, a plain square steel plate or a slab attached to the column is
adequate. If uplift or overturning forces are present, a more positive attachment is
necessary. These base plates can be welded directly to the columns or they can be
fastened by means of bolted or welded lug angles. These connection methods are
illustrated in Fig. 10.

A base plate welded directly to the columns is shown in Fig. 10(a). For small columns
these plates will be shop-welded to the columns, but for larger columns, it may be
necessary to ship the plates separately and set them to the correct elevations. For this
second case the columns are connected to the footing with anchor bolts that pass through
the lug angles which have been shop-welded to the columns. This type of arrangement is
shown in Fig. 10(b).

Anchor bolts

Concrete Angle
Anchor bolt footing

(a) (b)

Fig. 10 Column base plates

Sometimes, when there is a large moment in relation to the vertically applied load a
gusseted base may be required. This is intended to allow the lever arm from the holding
down bolts to be increased to give maximum efficiency while keeping the base plate
thickness to an acceptable minimum.

A critical phase in the erection of a steel building is the proper positioning of column
base plates. If they are not located at their correct elevations, serious stress changes may
occur in the beams and columns of the steel frame. In many cases, levelling plates of the
same dimensions as the base plate are carefully grouted in place to the proper elevations
first and then the columns with attached base plates are set on the levelling plates.

Version II 10 - 17
 DESIGN OF AXIALLY LOADED COLUMNS

The lengths and widths of column base plates are usually selected in multiples of 10 mm
and the thickness chosen to conform to rolled steel plates. Usually the thickness of base
plates is in the range of 40-50 mm. If plates of this range are insufficient to develop the
applied bending moment or if thinner plates are used, some form of stiffening must be
provided.

Concrete support area should be significantly larger than the base plate area so that the
applied load can disperse satisfactorily on to the foundation. To spread the column loads
uniformly over the base plates, and to ensure there is good contact between the two, it is
customary not to polish the underside of the base plate, but grout it in place.

Columns supporting predominantly axial loads are designed as being pin-ended at the
base. The design steps for a base plate attached to an axially loaded column with pinned
base is explained below.

Procedure for empirical design of a slab base plate for axial load only (pinned
connection)

1. Determine the factored axial load and shear at the column base.
2. Decide on the number and type of holding down bolts to resist shear and tension. The
chosen number of bolts are to be arranged symmetrically near corners of base plate or
next to column web, similar to the arrangement sketched in Fig. 10.
3. Maximum allowable bearing strength = 0.4 fcu (where fcu = cube strength of concrete)
Actual bearing pressure to be less than or equal to 0.4 fcu.
4. Determine base plate thickness t;

For I, H, channel, box or RHS columns

t  2.5w
f yp
a 2
 0 .3 b 2  but not less than the thickness of the flange of the

supported column.

w = pressure in N/mm2 on underside of plate, assuming a uniform distribution.

a = larger plate projection from column [See Fig. 11]
b = smaller plate projection from column
fyp = design strength of plate, but not greater than 250 N/mm2 divided by m

a tf1 a tf1

b b
tf2

tf1

Fig. 11 Base plates subjected to concentric forces

Version II 10 - 18
 DESIGN OF AXIALLY LOADED COLUMNS

5. Check for adequacy of weld. Calculate the total length of weld to resist axial load.
6. Select weld size.
7. Check shear stress on weld.
8. Vector sum of all the stresses carried by the weld must not exceed pw, the design
strength, of the weld.
9. Check for bolt. Check maximum co-existent factored shear and tension, if any, on the
holding down bolts.
10. Check the bolts for adequacy (see a later chapter for bolt design).

7.0 CONCLUDING REMARKS

Design of columns using multiple column curves is discussed in this chapter. Additional
provision required for accounting for heavily welded sections are detailed. Built-up
fabricated members frequently employed (when rolled sections are found inadequate) are
discussed in detail. Design guidance is provided for laced/battened columns. Effective
lengths for various end conditions are listed and illustrative worked examples are
appended. A simple method of designing a base plate for an axially loaded column is
proposed and illustrated by a worked example.

8.0 REFERENCES

1. Owens G.W., Knowles P.R (1994): "Steel Designers Manual", The Steel Construction
Institute, Ascot, England.
2. Dowling P.J., Knowles P.R., Owens G.W (1998): "Structural Steel Design",
Butterworths, London.
3. British Standards Institution (1985): "BS 5950, Part-1 Structural use of steelwork in
building", British Standards Institution, London.

Version II 10 - 19
 DESIGN OF AXIALLY LOADED COLUMNS

Table3: Ultimate Compressive stress (c) values in compression members

( fy = 250 N/mm2 )

 = 0.002  = 0.0035  = 0.0055  = 0.008



15 250 250 250 250

20 249 248 247 245
25 246 243 240 235
30 243 239 233 225
35 240 234 225 216
40 237 228 218 206
45 233 223 210 196
50 229 216 202 187
55 225 210 194 177
60 219 203 185 168
65 213 195 176 160
70 206 187 168 150
75 198 178 159 141
80 189 169 150 133
85 180 160 141 125
90 170 151 133 118
95 159 142 125 111
100 149 133 118 104
110 130 117 104 92
120 114 103 92 82
130 99 91 82 73
140 87 80 73 66
150 77 71 65 59
160 69 64 59 53
170 62 57 53 48
180 55 52 48 44
190 55 47 44 40
200 45 43 40 37

Version II 10 - 20
 DESIGN OF AXIALLY LOADED COLUMNS

Job No: Sheet 1 of 5 Rev

Structural Steel Job Title: AXIALLY COMPRESSED COLUMN
Design Project Worked Example - 1
Made by Date 23-09-99
SSSR
Calculation Sheet Checked by Date 28-09-99
RN

Obtain factored axial load on the column section ISHB400. The height of the
column is 3.0m and it is pin-ended.

[ fy = 250 N/mm2 ; E = 2* 105 N/mm2 ; m = 1.15]

CROSS-SECTION PROPERTIES:

b
T

X d

Y 3.0 m

Version II 10 - 21
 DESIGN OF AXIALLY LOADED COLUMNS

Job No: Sheet 2 of 5 Rev

Structural Steel Job Title: AXIALLY COMPRESSED COLUMN
Design Project Worked Example - 1
Made by Date 23-09-99
SSSR
Calculation Sheet Checked by Date 28-09-99
RN

Flange thickness = T = 12.7 mm

Clear depth between flanges = d = 400 – (12.7 * 2) = 374.6 mm

Thickness of web = t = 10.6 mm

Flange width = 2b = 250 mm

b = 125 mm

Self –weight = w = 0.822 kN/m

Area of cross-section = A = 10466 mm2

rx = 166.1 mm

ry = 51.6 mm

(i) Type of section:

b 125
  9.8  10 
T 12.7

d 374.6
  35.3  41 BS: 5950
t 10.6 Part - 1
Table - 7

250 250
where,    1 .0
fy 250

Hence, cross- section is “COMPACT”

10 - 22
 DESIGN OF AXIALLY LOADED COLUMNS

Job No: Sheet 3 of 5 Rev

Structural Steel Job Title: AXIALLY COMPRESSED COLUMN
Design Project Worked Example - 1
Made by Date 23-09-99
SSSR
Calculation Sheet Checked by Date 28-09-99
RN

(ii) Effective Length:

As, both ends are pin-jointed effective length = x = y = 3.0 m

(iii) Slenderness ratios:

 x 3000
x    18.1
rx 166.1

y 3000
y    58.1
ry 51.6

(iv) Values of ( ) :

For rolled I-sections,

In x – direction x = 0.0020
In y – direction  y = 0.0035

22 200000
o  0.2 E
fy
 0.2 *
7 250

= 17.8
(v) values of :

 =  ( - 0 )

x =  x (x -  0) = 0.002* (18.1 – 17.8) = 0.001

y = y (y -  0) = 0.0035* (58.1 – 17.8) = 0.141

10 - 23
 DESIGN OF AXIALLY LOADED COLUMNS

Job No: Sheet 4 of 5 Rev

Structural Steel Job Title: AXIALLY COMPRESSED COLUMN
Design Project Worked Example - 1
Made by Date 23-09-99
SSSR
Calculation Sheet Checked by Date 28-09-99
RN

(vi) Calculation of maximum compressive stress at failure (c):

We have,
 2E
e 
2

 c     2  f y e  f y

f y  ( 1)  e
where,  
2

In x-direction,

π 2 E π 2 * 200000
σ ex  2
 2
 6025 N/mm2
λx (18.1)

250  (0.001  1)*6025

φx   3140 N/mm2
2

σ cx  3140  3140 2  250* 6025  250

 250 N/mm2

10 - 24
 DESIGN OF AXIALLY LOADED COLUMNS

Job No: Sheet 5 of 5 Rev

Structural Steel Job Title: AXIALLY COMPRESSED COLUMN
Design Project Worked Example - 1
Made by Date 23-09-99
SSSR
Calculation Sheet Checked by Date 28-09-99
RN

In y-direction,
π 2 E π 2 * 200000
σ ey    585 N/mm2
λ y2 (58.1)2

250  (0.141  1) 585.0

φy   459N/mm2
2

σ cy  459  459 2  250* 585  250

 205 N/mm2

Hence, Allowable axial compressive stress, c = 205 N/mm2

Safe axial compressive stress = c/m = 205/1.15 = 178 N/mm2

(vii) Factored Load:

Factored Load = c A/m = 178 * 10466/1000

= 1863 kN

10 - 25
 DESIGN OF AXIALLY LOADED COLUMNS

Job No: Sheet 1 of 2 Rev

Structural Steel Job Title: AXIALLY COMPRESSED COLUMN
Design Project Worked Example - 2
Made by Date 23-09-99
SSSR
Calculation Sheet Checked by Date 28-09-99
RN

Obtain maximum axial load carried by the column shown when ISHB 400 is
employed. The column is effectively restrained at mid-height in the y-
direction, but is free in x-axis. The data is the same as in problem1.

[ fy = 250 N/mm2 ; E = 2.0*105 N/mm2 ; m = 1.15 ]

(i) Type of section:

Section is “COMPACT” from

previous example.

(ii) Effective lengths:

x = 6000 mm
3.0 m y = 3000 mm

(iii) Slenderness ratios:

6000
x   36.1
166.1

 y  3000  58.1
51.6

(iv) Values of ( ) and ( ):

3.0 m In x – direction x = 0.002

In y – direction y = 0.003

and 0 = 17.8

x = 0.002 (36.1 - 17.8) = 0.037

y = 0.0035 (58.1– 17.8) = 0.141

10 - 26
 DESIGN OF AXIALLY LOADED COLUMNS

Job No: Sheet 2 of 2 Rev

Structural Steel Job Title: AXIALLY COMPRESSED COLUMN
Design Project Worked Example - 2
Made by Date 23-09-99
SSSR
Calculation Sheet Checked by Date 28-09-99
RN

(v) Calculation of c :

In x- direction,

 2 E  2 * 200000
 ex    1515 N / mm2
x2
(36.1) 2

250  (0.037  1) * 1515

x   911 N / mm2
2

 cx  911 9112  250*1515  250

 239 N / mm

In y-direction

 2 E  2 * 200000
 cy    585 N / mm2
y2
(58.1) 2

250  (0.141  1) * 585

y   459 N / mm2
2

 cy  459  4592  250 *585  250

 205 N / mm2

(vi) Factored Load:

Factored Load = cA/m = 205/1.15* 10466/1000 = 1863 kN

10 - 27
 DESIGN OF AXIALLY LOADED COLUMNS

Job No: Sheet 1 of 1 Rev

Structural Steel Job Title: BASE PLATE
Design Project Worked Example - 3
Made by Date 20-2-00
SSSR
Calculation Sheet Checked by Date 28-2-00
RN

Design a simple base plate for a ISHB400 @ 0.822 kN/m column to carry a
factored load of 1800 kN.

[fcu = 40 N/mm2 ; fy = 250 N/mm2 ; m = 1.15]

ISHB 400 300 mm

450 mm

Bearing strength of concrete = 0.4fcu = 0.4 * 40 = 16 N/mm2

Area required = 1800*103/16 = 112500 mm2

Use plate of 450 X 300 mm (135000 mm2)

Assuming projection of 25 mm on each side

w = (1800 * 103) / (450 * 300) = 13.33 N/mm2

fyp = 250/1.15 = 217.4 N/mm2

tp 

2.5w a 2  0.3b 2 


2.513.33 252  0.3 252 
 8.2 mm
f yp 217.4

Hence, use 450 X 300 X 10 mm plate.

(Design of connections are discussed in later chapters)

10 - 28