Design of compression members
6.1 INTRODUCTION
When a member is subjected to compressive axial forces, it is
referred to as a compression member. Compression members are found
as:
1. Columns in buildings
2. Piers in bridges
3. Top chords of trusses
4. Bracing members
Compression and tension members differ in the following ways:
1. Slender compression members can buckle.
2. In tension members, bolt holes reduce the effective cross-sectional
area for carrying the loads. In compression members, however, the
bolts tend to fill the holes and the entire area of the cross section is
normally assumed to resist the loads.
In this chapter we cover the design of compression members
subjected to concentric loads. These members are also referred to as
columns. In practice, the dead or live load is usually not centered over an
exterior building column and the line of action of the resultant gravity
Design of compression members 2
load usually is within the inner side of the column. Even in the case of
interior building columns, the dead and live loads may not be centered.
In such cases, the member is subjected to both axial compressive forces
and bending moments due to the eccentric loading. Design of these
members, referred to as beam-columns, is covered in the following
chapter.
6.2 THE EULER FORMULA FOR BUCKLING LOAD OF
COLUMNS
The Euler formula is derived for an ideal or perfect column. That
is, it is assumed that the column is long, slender, straight, homogeneous,
and elastic and is subjected to concentric axial compressive loads. It is
also assumed that the two ends of the columns are hinged. Suppose that
Figure 6.1 Column with hinged ends
Design of compression members 3
this ideal column in equilibrium is laterally distributed as shown in Fig.
6.1. We can write the following differential equation for the lateral
displacement v:
Pv M
dx
v d
EI = =
2
2
(6.1)
where E is the modulus of elasticity, I is the moment of inertia about the
axis of bending in the cross section, P is the axial compressive force, and
M is the bending moment at a distance x from support A. Denoting
2
k
EI
P
= (6.2)
Equation (6.1) can be written as
0
2
2
2
= + v k
dx
v d
(6.3)
This is a second-order homogeneous linear differential equation with
constant coefficients. The solution of this differential equation is
kx C kx C v sin cos
2 1
+ = (6.4)
The integration constants C
1
and C
2
may be found by applying the
following geometric boundary conditions:
At x = 0: ; 0 0
1
= = C v kx C v sin
2
= (6.5)
At x = L: 0 sin 0
2
= = kL C v
(6.6)
Design of compression members 4
Equation (6.6) indicates that either C
2
= 0, which means no lateral
displacement and instability problem, or sin(kL) = 0 with solution
,... 2 , 1 = = n n kL (6.7)
Substituting for k from Eq. (6.7) into Eq. (6.2) and solving for P, we
obtain
,... 2 , 1
2
2 2
= = n
L
EI n
P
(6.8)
The smallest value obtained from Eq.(6.8) is known as critical load,
buckling load, or Euler formula:
2
2
L
EI
P
cr
=
(6.9)
This equation can be interpreted as follows. If we increase the magnitude
of the load P from zero, for small value of P, there is no lateral
displacement and v = 0. But as P approaches the critical value P
cr
, the
lateral displacement becomes
L
x
C v
sin
2
= (6.10)
which is indefinite. In other words, the column shows unstable behavior.
It should be noted that the critical buckling load given by Eq.
(6.9) is independent of the strength of the material (say, F
y
, the yield
stress). The results obtained from Eq. (6.9) compare very well with
experimental results for long slender columns.
Design of compression members 5
Equation (6.9) was found for a column with hinged ends. This
equation can be used for columns with other end conditions, provided
that it is modified as follows:
2
2
) (KL
EI
P
cr
= (6.11)
where KL is the distance between the points of zero moment, or
inflection points. The length KL is known as the effective length of the
column. The dimensionless coefficient K is called the effective length
factor. The values of this factor for different types of columns are given
in Fig. 6.2. These values are ideal values obtained from theoretical
Figure 6.2 Effective length factors for different types of columns.
Design of compression members 6
analyses similar to the one presented in this section. In practice, these
values are modified somewhat. The modified values as recommended by
the ASD code (commentary Sec. C2) are given inside parentheses in Fig
6.2.
Dividing the critical load P
cr
by the cross-sectional area of the
column A, we can find the critical stress f
cr
.
2
2
2
2
) / ( ) ( r KL
E
A KL
EI
A
P
f
cr
cr
= = = (6.12)
where r is the radius of gyration of the cross section about the axis of
bending and KL/r is called the slenderness ratio of the column. Note that
if the column is not restricted to bend in a particular plane, it will buckle
in a plane perpendicular to the minor axis of the cross section. Hence, the
moment of inertia in Eq. (6.11) and the radius of gyration in Eq. (6.12)
Figure 6.3 Critical stress versus slenderness ratio
Design of compression members 7
are with respect to the minor axis of the cross section and therefore the
minimum values for the cross section. The corresponding critical load
and stress are the minimum values causing buckling of the column.
The variation of the critical stress f
c
with the slenderness ratio
KL/r is shown qualitatively in Fig. 6.3. The range of slenderness ratios is
divided into three regions, broadly called short, medium, and long
columns. By short columns, it is simply meant columns with small
slenderness ratios, and so on. Short columns do not buckle and simply
fail by excessive yielding. Long columns usually fail by elastic buckling,
for which Euler buckling stress, Eq. (6.12), is valid. Between short and
long regions, there is an intermediate region in which the failure of the
column occurs through inelastic buckling.
6.3 RESIDUAL STRESSES
Residual stresses due to uneven cooling of standard sections
after hot-rolling, and also welding, can adversely affect the resistance of
columns against buckling. As an example, in an I-section, the outer tips
of the flanges and the middle portion of the web cool more quickly than
the relatively thick portions at the intersection of the flanges and the web.
The result of this uneven cooling is that the areas cooled more quickly
develop residual compressive stresses, while the areas cooled more
slowly develop residual tensile stresses. The magnitude of the residual
stresses can be as large as 1015 ksi.
Design of compression members 8
6.4 ALLOWABLE COMPRESSIVE STRESSES ACCORDING TO
THE ASD CODE
In the ASD code equations for allowable compressive stresses,
various imperfections such as the effect of residual stresses, the actual
end restraint conditions, crookedness, and small unavoidable
eccentricities are empirically taken into account. The ASD code assumes
arbitrarily that the elastic buckling holds valid when the stress in the
column is not greater than one-half of the yield stress (F
y
/2) (Fig. 6.4).
Defining
C
c
= the slenderness ratio dividing elastic from inelastic buckling
Figure 6.4 Variation of critical stress and allowable stress as specified by
the ASD code
Design of compression members 9
we can write
y
c
c
y
F
E
C
C
E
F
2
2
2
2
2
1
= = (6.13)
For columns with slenderness ratio less than C
c
, the ASD code
assumes a parabolic variation and the allowable axial compressive stress
is given by (Fig. 6.4)
c y
c
a
C
r
KL
F
C
r KL
F
=
F.S.
2
) / (
1
2
2
(6.14)
3
3
8
) / (
8
) / ( 3
3
5
Safety of Factor . F.S
c c
C
r KL
C
r KL
+ = = (6.15)
Note that at very small slenderness ratios, KL/r 0 and the F.S. in
compression is 1.67, which is the same as in tension (F
t
= 0.60 F
y
). As
KL/r increases, the F.S. also increases until at KL/r = C
c
it becomes
23/12 = 1.92. This variation of F.S. is rational because the longer the
column, the more sensitive it becomes to the imperfections in the
evaluation of the effective length and the critical stress.
For columns with slenderness ratio greater than C
c
, the Euler
formula [Eq. (6.12)] is employed with a F.S. of 23/12 = 1.92 (Fig. 6.4).
r KL C
r KL
E
F
c a
/
) / ( 23
12
2
2
< =
(6.16)
According to the ASD B7, the slenderness ratio of compression members
preferably should not be greater than 200.
Design of compression members 10
Note that in the design of compression members, local buckling
of plate elements should be prevented. In other words, the allowable
stress values given in this section may not be used unless the width-
thickness ratio of the elements are limited to the values given in the
fourth column of Table 5.1. When the width-thickness ratios of the plate
elements of the section exceed these limiting values, the allowable axial
compressive stress should be modified according to Appendix B of the
ASD code. Since these cases are not encountered frequently in practice,
they are not covered in this chapter.
6.5 EXAMPLES OF DESIGN OF COLUMNS ACCORDING TO
THE ASD CODE
Example 1
The diagonally braced frame shown in Fig. 6.5 should be designed for a
lateral earthquake load of 31 Kips. Three different openings are required
as shown in the figure. The diagonal must fit within the thickness of a 4-
in.-thick wall. Also, the width of the diagonal in the plane of the frame
must not exceed 4 in. Design a rolled section for the diagonal from the
ASD manual, using A36 steel with yield stress of 36 ksi.
Solution
The structure is statically determinate. The diagonal must be
designed as a compression member.
Design of compression members 11
K 84 . 43
45 cos
31
diagonal in the force axial = = =
o
P
The requirement of the problem will be satisfied best by selecting a tube.
Try TUBE 4x4 with wall thickness t = 0.5 in., A = 6.36 in.
2
, and r
min
=
1.39 in.
5 . 146
39 . 1
) 12 )( 2 )( 12 )( 0 . 1 (
= =
r
KL
r
KL
F
E
C
y
c
< = = = 1 . 126
36
) 000 , 29 ( 2 2
2 2
Figure 6.5
Design of compression members 12
ksi 96 . 6
) / ( 23
12
2
2
= =
r KL
E
F
a
Compression capacity of the section =
K 84 . 43 2 . 44 ) 96 . 6 )( 36 . 6 ( = > = = P AF
a
O.K.
Check local flange buckling (column four of Table 5.1):
6 . 39
238
8
5 . 0
4
= < = =
y
F
t
b
O.K.
USE TUBE 4x4 with t = 0.5 in. weight = 21.63 lb/ft
Example 2
The cross section of a built-up column is shown in Fig. 6.6. It consists of
two L8x8x1 angles and is made of A36 steel with yield stress of 36 ksi.
Using the ASD code, find the maximum allowable axial load for this
column for a length of 15 ft. Assume that there exist ideal hinged
conditions at the column ends and that the angles are properly connected
so that they may not buckle individually.
Solution
Properties of L8x8x1: A
1
= area of cross section = 15 in.
2
,
r
z
= 1.56 in. c = 2.37 in. w = 51 lb/ft
Design of compression members 13
We must first calculate the moments of inertia of the built-up section
with respect to x and y (axes of symmetry of the built-up section). From
the geometry we can write (Fig. 6.6)
in. 35 . 3 2 = = c a
in. 06 . 4
2
2
= + = a d
The moment of inertia of one angle about the z-axis shown in Fig. 6.6 is
4 2 2
1 1
in. 50 . 36 ) 56 . 1 )( 15 ( = = =
z z
r A I
Figure 6.6
Design of compression members 14
Thus,
4 2 2
1 1
in. 51 . 567 ] ) 06 . 4 ( 15 50 . 36 [ 2 ] [ 2 = + = + = d A I I
z x
] ) ( ) [( 4 ) ( 4
y ABE y CDE y ABCD y
I I I I = =
Using the relation for moment of inertia of a triangle about its base (Fig.
6.7), we have
] ) ( ) ( [ 4
3
12
1 3
12
1
BH AE CG DE I
y
=
b DE 2 = b CG
2
2
= ) ( 2 t b AE = ) (
2
2
t b BH =
4 4 4
6
1 4 4
6
1
in. 5 . 282 ] ) 1 8 ( 8 [ ] ) ( [ = = = t b b I
y
4
min
in. 5 . 282 = =
y
I I
in. 07 . 3 ) 15 2 /( ) 5 . 282 ( /
min
= = = A I r
y
1 . 126 7 . 58
07 . 3
) 12 )( 15 )( 0 . 1 (
= < = =
c
C
r
KL
From Eq. (6.14):
Figure 6.7
Design of compression members 15
ksi 55 . 17 =
a
F
K 5 . 526 ) 55 . 17 )( 30 ( = = =
a
AF P
K 525
1000
) 51 )( 15 ( 2
5 526 weight = = = . P P
all
Note: Moments of inertia of the built-up section could also be computed
by using rotation of axes and the parallel-axis theorem.
Example 3
A column is built of two tubes 16x8 and one angle L8x8x1
available in a shop. In order to utilize the material efficiently a designer
has proposed the arrangement shown in Fig 6.8. The sections are made of
A36 steel with yield stress of 36 ksi. Using the ASD code, determine the
maximum allowable load for this column for a length of 14 ft. Assume
ideal hinged conditions at the column ends. The sections are properly
connected so that they may not buckle individually. Thickness of the
tubes is t = 0.5 in.
Solution
For L8x8x1:
I
x1
= I
y1
= 89 in.
4
A = 15.0 in.
2
r
z
= 1.56 in. I
z
= 15(1.56)
2
= 36.50 in.
4
For a tube 16x8 (t = 0.5 in.):
Design of compression members 16
A = 22.4 in.
2
C
1
= centroid of the L8x8x1
C: centroid of the built-up section
in. 8.086
2(22.4) 15
) 16 4 )( 4 . 22 ( ) 37 . 2 )( 15 (
=
+
+ +
= x
For one tube (Fig. 6.9):
4
2 2 2 2
2 2
in. 507.92
24
) 15 7(15)(7
-
24
) 16 8 )( 16 ( 8
=
+ +
= =
y x
I I
in. 44 . 11 2 = = x OC
in. 49 . 6 1 4 16
2 2
2
= + = OC
Figure 6.8
Design of compression members 17
deg 04 . 4 1 arctan
16
4
= =
in. 70 . 2 ) 45 cos(
2 1
= = OC OC d
o
in. 48 . 8 ) 45 sin(
2 2
= =
o
OC d
Moment of inertia of the built-up section about the weak axis (x-axis):
4
2 2
min
in. 2360.24
] ) 70 . 2 ( 4 . 22 92 . 507 [ 2 ) 2 37 . 2 44 . 11 ( 15 50 . 36
=
+ + + = =
y
I I
Cross-sectional area of the built-up section:
2
in. 59.8 (2)(22.4) 15 = + = A
in. 28 . 6 8 . 59 / 24 . 2360 / = = = A I r
x x
Figure 6.9
24
)
; 45
12
) cos
2
2 2
d bd(b
d bd(b
+
=
=
+
=
2
x
2 2
x
I
For
sin
I
Design of compression members 18
1 . 126 75 . 26
28 . 6
) 12 14 )( 0 . 1 (
= < =
=
c
C
r
KL
From Eq. (6.14):
F
a
= 20.17 ksi
P = AF
a
= (59.8)(20.17) = 1206.2 K
6.6 LOAD AND RESISTANCE FACTOR DESIGN OF
COMPRESSION MEMBERS
According to the LRFD E2, the design strength of compression
members excluding those with slender compression elements (see section
5.10.1) is equal to
c
P
n
, where
c
= resistance factor for compression = 0.85
P
n
= nominal axial strength = AF
cr
A = gross cross-sectional area of member
y c y
F F
c
) 658 . 0 ( ) 419 . 0 exp(
2
2
=
5 . 1 for
c
F
cr
=
2
/ 877 . 0
c y
F
5 . 1 for >
c
(6.17)
c
=
E
F
r
KL y
(6.18)
In Eq. 6.18, KL/r is the maximum slenderness ratio of the column.
Design of compression members 19
6.7 WEB-BASED INTERACTIVE DESIGN OF COMPRESSION
MEMBERS
The applet presented in this section is for interactive design of
axially loaded columns with various end conditions according to the
ASD and LRFD codes. This applet consists of three panels: Input,
Figure 6.10 Input panel for the applet for design of
compression members
Design of compression members 20
Column Type, and Results, shown in Figs. 6.10, 6.11, and 6.12,
respectively. The user can enter the input for design of axially loaded
columns in the input panel (Fig. 6.10). This input includes design load,
unbraced column length, column type, steel type, section type, nominal
depth of section, and section width (for double sections only).
Column type selection (Fig. 6.13) is based on the columns
effective length factor, K. The user can see the K-value and the end
condition descriptions from the column type panel (Fig. 6.11). Fig. 6.11
shows various column types considered in this applet. Note that the user
can also choose any K-value other than those specified in the applet.
The applet can design six different types of sections: W shape,
HP shape, double W shape, double HP shape, and two different types of
double channel ([ ] and ] [), as noted in Fig. 6.14. When a double section
type is selected, the applet asks the user to enter the width of the section
Figure 6.11 Column type panel
Design of compression members 21
(Fig. 6.15). If the user tries to run the applet without specifying the width
of the section, the warning message shown in Fig. 6.16 will be displayed.
The user may select the nominal depth of the cross section. For
example, if the user selects W12 series among W sections, the applet will
try to come up with the lightest W12 section. If no W12 section available
Figure 6.12 Results panel
Figure 6.13 Column type selection
Design of compression members 22
from the ASD manual satisfies the design requirements, the applet will
automatically move to W14 series and will try to select the lightest W14
section. This process is continued until the applet finds a W section
satisfying all the design requirements. A failure message will be
displayed when no W section is adequate for the given loading condition.
When the user does not select any nominal depth, the applet will come
up with the lightest section within the type of the cross section selected
by the user.
Figure 6.12 shows the minimum weight design for the input
specified in Fig. 6.10 using a W shape without specifying the nominal
depth. Figure 6.17 shows the design for the input specified in Fig. 6.10
when the user selects the double channel ([ ]) with a depth of 10 in. and
Figure 6.14 Section type selection
Figure 6.15 Applet asks the user to enter
the section width when a
double section is selected
Design of compression members 23
section width of 20 in. Figure 6.18 shows the design for the same input
but using a double W shape with a section width of 20 in. without
specifying the nominal depth of the cross section.
Figure 6.16 Applet warns the user if the section
width is not given for a double section.
Figure 6.17
Design of compression members 24
The following load combinations and the corresponding load
factors are considered when the user requires applet to design a column
based on LRFD code.
1.4DL
1.2DL + 1.6LL
where DL is dead load acting on the column and LL is live load acting on
the column.
As an example, the same axially compressed column solved
using input data given in Fig. 6.10. It is subjected to dead and live axial
load of 150K each.
Initially a W10 section was selected. The answer is W10x68,
which is the same as that obtained based on ASD code. Next, the applet
Figure 6.18
Design of compression members 25
was asked to find the lightest available W section. The answer is
W12x65, which is the same as that obtained according to ASD code.
Subsequently, dead load was increased to 200K and live load
was decreased to 100K while the total axial load was kept the same. The
applet was asked to find the lightest W section. The answer is W10x60,
which is lighter than the previous minimum weight design of W12x65.
Finally, the dead load was further increased to 250K and the live
load was decreased to 50K while the total axial load was kept the same.
The applet was again asked to fine the lightest W section. The answer is
W12x58. As the ratio of live load to dead load decreases, it appears that
design on the basis of LRFD code yields a lighter section than that on the
basis of ASD code.
For the sake of brevity, the results are not presented here.
6.8 PROBLEMS
6.1 The column AB shown in Fig. 6.19 is made of a C10x30 and a
C10x20. The two ends of the column are hinged with no lateral
displacement in any direction. At point D, lateral displacement is
prevented in the x direction only. The length of the column is L = 15
ft. Channels are made of A36 steel with yield stress of 36 ksi. (a)
What is the maximum possible allowable axial compressive load for
this column? (b) What is the corresponding minimum distance b
between the channels?
6.2 A column of length 24 ft is subjected to an axial load of 165 kips.
Column ends are hinged. The column is supported at the midheight
Design of compression members 26
only in its weak direction. Using A36 (F
y
= 36 ksi), select the least
weight tube for the column.
6.3 In Problem 6.2, suppose you ought to use a section made of four
L3.5x3.5x5/16 with double lacings as shown in Fig 6.20. Find the
minimum dimensions a and b of the column cross section. Next,
after rounding these minimum dimensions to the nearest inch, design
the double lacing bars. Check all the necessary requirements.
6.4 A column is hinged at one end and fixed at the other end. It is made
of 2 L5x5x and 2 PL5x as shown in Fig. 6.21. The angles and
plates are made of A36 steel with yield stress of 36 ksi. Find the
maximum axial compressive load capacity of this column assuming
Figure 6.19
Design of compression members 27
that angles and plates are connected to each other properly so that
none of them will buckle individually. Use the design value for the
effective length factor.
Figure 6.21
Figure 6.20
Design of compression members 28
6.5 A built-up column consists of an L8x6x and a 0.5 in.-thick plate
welded together as shown in Fig. 6.22. Both angle and plate are
made of A36 steel with yield stress of 36 ksi. The length of the
column is 15 ft. The bottom end of the column is fixed and its top
end is free. Find the maximum axial compressive load capacity of
this column.
6.6 Solve Problem 6.5, using steel with yield stress of 60 ksi.
6.7 Find the lightest column made of A36 steel (F
y
= 36 ksi) with hinged
ends and a length of 10 ft to carry an axial compressive force of 250
K. Use the following shapes:
a. W14
b. WT
c. Square structural tube
d. Pipe
e. Angle with equal legs
6.8 Find the maximum axial compressive load capacity of a column with
a length of 20 ft and the cross section shown in Fig. 6.23. The two
ends of the column are hinged. Assume the following types of steel
Figure 6.22
Design of compression members 29
a. A36 (F
y
= 36 ksi)
b. A441 (F
y
= 50 ksi)
c. A572 (F
y
= 65 ksi)
6.9 Solve Problem 6.8, assuming that the column is an L8x6x1
Figure 6.23
