Developed by Scott Civjan

University of Massachusetts, Amherst

Modified and supplemented by

Cem Topkaya
Middle East Technical University
1
 P

COMPRESSION
MEMBER/COLUMN:
Structural member
subjected to axial load

P
2
Examples of Compression Members:

The diagonal braces and the columns of this frame

would typically be idealized as carrying axial load only.
(tension or compression). For loading cases that
place these members in compression, they would
be designed as compression members.

3
Trusses:

Top chord members =

compression members

Rigid Frame
Members subject to
axial compression + bending
treat as a “beam column”

Note: the term “column” is often used

to denote a compression member.
4
Compression member
5
INDIVIDUAL COLUMN

6
Strength design requirements:
Pu  Pn (Pa  Pn/Ω)ASD
Where  = 0.9 for compression
(Ω = 1.67)ASD

7
 Axial Strength

Strength Limit States:

• Yielding (Squash) Load
• Global Buckling
• Local Buckling

8
 Global Local Local
Buckling Flange Web
Buckling Buckling

9
 Squash Load
Fully Yielded Cross Section

10
When a short, stocky column is loaded the strength is limited by
the yielding of the entire cross section.

Absence of residual stress, all fibers of cross-section yield

simultaneously at P/A=Fy.

P=FyA P
L0 D

eyL0 D
P

11
RESIDUAL STRESSES

Self equilibrating stresses that result from the

manufacturing process for a steel member
Sources: uneven cooling of hot rolled or welded shapes
Cold straightening of members

Results in a reduction in the effective stiffness of the

cross section, but the ultimate squash load is unchanged.

Reduction in effective stiffness can influence onset of

buckling.
12
RESIDUAL STRESSES

~0.3 Fy compression

~0.2 Fy tension

~0.2 Fy compression

13
RESIDUAL STRESSES

With residual stresses, flange tips yield first at

P/A + residual stress = Fy
Gradually get yield of entire cross section.

Stiffness is reduced after 1st yield.

No Residual Stress
P=FyA

eyL0 D

14
 = Yielded
RESIDUAL STRESSES Steel

With residual stresses, flange tips yield first at

P/A + residual stress = Fy 1
Gradually get yield of entire cross section.

Stiffness is reduced after 1st yield.

No Residual Stress
P=FyA

P=(Fy-Fres)A
1

eyL0 D

15
 = Yielded
RESIDUAL STRESSES Steel

With residual stresses, flange tips yield first at

P/A + residual stress = Fy 1
Gradually get yield of entire cross section.

Stiffness is reduced after 1st yield.

No Residual Stress 2
P=FyA
2
P=(Fy-Fres)A
1

eyL0 D

16
 = Yielded
RESIDUAL STRESSES Steel

With residual stresses, flange tips yield first at

P/A + residual stress = Fy 1
Gradually get yield of entire cross section.

Stiffness is reduced after 1st yield.

No Residual Stress 2
P=FyA
3
2
P=(Fy-Fres)A
1
3

eyL0 D

17
 = Yielded
RESIDUAL STRESSES Steel

With residual stresses, flange tips yield first at

P/A + residual stress = Fy 1
Gradually get yield of entire cross section.

Stiffness is reduced after 1st yield.

No Residual Stress 2
P=FyA 4
3
2 Effects of Residual
P=(Fy-Fres)A Stress
1
3

eyL0 D

4 18
Euler Buckling

19
 Buckled member

Buckled member

20
 Euler Buckling

Assumptions:
• Column is pin-ended.
• Column is initially perfectly straight.
• Load is at centroid.
• Material is linearly elastic (no yielding).
• Member bends about principal axis (no twisting).
• Plane sections remain plane.
• Small deflection theory.

21
Euler Buckling

P P
Bifurcation Point
π EI
2
PE 
L2

Stable Equilibrium
D D

s E

e
22
Euler Buckling
Dependent on Imin and L2.
Independent of Fy.

For similar unbraced length in each direction,

“minor axis” (Iy in a I-shape) will control strength.

π 2 EI x
PE Major axis buckling
L2
π 2 EI y
L2
Minor axis buckling

L
23
Euler Buckling

Re-write in terms of stress:

π 2
PE = EI
L2
π 2 EI
divide by A, PE/A = AL2 , then with r2 = I/A,

π2 E
 
PE/A = FE = 2
L
r
FE = Euler (elastic) buckling stress
L/r = slenderness ratio
24
Euler Buckling

Buckling controlled by largest value of L/r.

Most slender section buckles first.

Fy π2 E
 
L
r
2

FE

L/r

25
EULER ASSUMPTIONS
(ACTUAL BEHAVIOR)

26
 Initial Crookedness/Out of Straight
D0 = initial mid-span deflection of column

Do Do
M = PDo
P

27
Initial Crookedness/Out of Straight

P
Do= 0

π 2 EI
PE  2
L

Do D

28
Initial Crookedness/Out of Straight

P
Do= 0

π 2 EI
PE  2
L Elastic theory

Do D

29
Initial Crookedness/Out of Straight

P
Do= 0

π 2 EI
PE  2
L Elastic theory

Actual Behavior
Do D

30
Initial Crookedness/Out of Straight
Buckling is not instantaneous.
Additional stresses due to bending of the column,
P/A  Mc/I.
Assuming elastic material theory (never yields),
P approaches PE.
Actually, some strength loss
small D0 => small loss in strength
large D0 => strength loss can be substantial
ASTM limits of D0 = L/1000 or 6 mm in 6000 mm
Typical values are D0 = L/1500 or 4 mm in 6000 mm
31
 Load Eccentricity

e
P P
Do= 0

π 2 EI
PE  2
L
Elastic theory
L D

D D

32
 Load Eccentricity

e
P P
Do= 0

π 2 EI
PE  2
L
Elastic theory
L D

Actual Behavior
D D

33
 Load Eccentricity
Buckling is not instantaneous.
Additional stresses due to bending of the column,
P/A  Mc/I.
Assuming elastic material theory (never yields),
P approaches PE.
Actually, some strength loss
small e => small loss in strength
large e => strength loss can be substantial
If moment is “significant” section must be designed
as a member subjected to combined loads.
34
 End Restraint (Fixed)
Set up equilibrium and solve
similarly to Euler buckling EXAMPLE
derivation.
Determine a “K-factor.” 4π 2 EI π 2 EI
PE  2

L (1 / 2 L) 2
π 2 EI
Pe  KL
 KL 
2 Similar to pin-pin,
with L’ = L/2.
π EI
2
Load Strength =
Fe 
 KL r 
2 4 times as large.

35
End Restraint (Fixed)

Effective Length = KL

Length of equivalent pin ended

column with similar elastic
buckling load,

Distance between points of

inflection in the buckled shape.

36
37
 Inelastic Material Effects

Fy

s ET= Tangent Modulus

(Fy-Fres)
E

e
Test Results from an Axially Loaded Stub Column
38
 Inelastic Material Effects

Elastic Behavior
π E
2
Fe  2
 KL 
 
 r 

KL/r

39
 Inelastic Material Effects

π 2 ET
Fc  2
 KL 
Fy  
 r 

π2E
Fe 
Fy-Fres Inelastic  KL 
2

 
 r 
Elastic
s

KL/r

40
 Inelastic Material Effects

π 2 ET
Fc  2
 KL 
Fy  
 r 

π2E
Fe 
Fy-Fres Inelastic  KL 
2

 
 r 
Elastic
s

KL/r

41
 Inelastic Material Effects

Two classes of buckling:

Elastic Buckling: ET = E
No yielding prior to buckling
Fe  Fy-Fres(max)
Fe = predicts buckling (EULER BUCKLING)

Inelastic Buckling:
Some yielding/loss of stiffness prior to buckling
Fe > Fy-Fres(max)
Fc - predicts buckling (INELASTIC BUCKLING)

42
 Overall Column Strength

Fy
π2E
FE  2
 KL 
 
 r 

Experimental Data

KL/r

43
 Overall Column Strength

Inelastic Material effects

Including Residual Stresses
Fy
π2E
FE  2
 KL 
 
 r 

Experimental Data

KL/r Out of Straightness

44
 Overall Column Strength

Major factors determining strength:

1) Slenderness (L/r).
2) End restraint (K factors).
3) Initial crookedness or load eccentricity.
4) Prior yielding or residual stresses.

The latter 2 items are highly variable between specimens.

45
Chapter E:
Compression Strength

46
 Compression Strength

c= 0.90 (Wc= 1.67)

47
 Compression Strength

Specification considers the following conditions:

CE388 Flexural Buckling

Torsional Buckling
Flexural-Torsional Buckling

48
Compressive Strength

49
 Compression Strength

 The following slides assume:

 Non-slender flange and web sections
 Doubly symmetric members

50
 Compression Strength

Since members are non-slender and doubly symmetric,

flexural (global) buckling is the most likely potential failure
mode prior to reaching the squash load.

Buckling strength depends on the slenderness of the section,

defined as Lc/r. Lc is the effective length of the member.

The nominal strength is defined as

Pn= FcrAg Equation E3-1

51
  Fy

L E then, Fcr  0.658 e  Fy .
F
If, c
 4.71 Equation E3-2
r Fy  
This defines the “inelastic” buckling limit.

Lc E , then F = 0.877F .
If  4.71 cr e Equation E3-3
r Fy
This defines the “elastic” buckling limit
with a reduction factor, 0.877, times the theoretical limit.

2 E
Fe = elastic (Euler) buckling stress, Fe  2
Equation E3-4
 Lc 
 
 r 

52
53
 Inelastic Material Effects

Elastic Behavior
2 E
Fe  2
 Lc 
 
 r 

Lc /r

54
 Inelastic Material Effects

s
2 ET
Fc  2
 Lc 
Fy  
 r 
2 E
Fe  2
Fy-Fres Inelastic  Lc 
 
Elastic  r 

Lc /r

55
 Inelastic Material Effects

s
2 ET
Fc  2
 Lc 
Fy  
 r 
2 E
Fe  2
Fy-Fres Inelastic  Lc 
 
Elastic  r 

Lc /r

56
 Inelastic Material Effects
s  Fy

Fcr  0.658  Fy
Fe

 
Fy
Fcr  0.877 Fe
2 E
Inelastic Fe  2
0.44Fy  Lc 
 
Elastic  r 

E
Lc /r 4.71
Fy

57
Slenderness Criteria

58
 Per Section E.2

It is recommended to
provide Lc /r less than 200

59
Example: For the compression member shown below determine design
compressive strength (Φc Pn) and allowable compressive strength (Pn / Ωc ) for the
two cases;
i. L = 6 m
ii. L = 16 m
HEA300 column S275 Fy = 275 MPa Pinned at top and fixed at bottom no
intermediate bracing. Use recommended K factors.

rx  127 .4 mm weak axis buckling governs

ry  74.9 mm
K  0.8 (Recommended) 200 10 2
4.71  127
K  L 0.8  6000 275
i)   64
r 74.9
HEA300 Ag  11250 mm2
KL E
 4.71 11250
r fy Pn  216.6   2437 kN
1000
 2  200  103
Fe   482 MPa  c Pn  0.9  2437  2193 kN
642
 275
 Pn 2437
Fcr  0.658   275  216.6 MPa
482   1459 kN
W c 1.67
 

60
 K  L 0.8 16000 E
ii)   171  4.71  127 Ag  11250 mm2
r 74.9 fy

 2  200  103
Fe   67.5 MPa
1712

Fcr  0.877  67 .5  59 .2 MPa

11250
Pn  59.2   666 kN
1000

 c Pn  0.9  666  599 kN

Pn 666
  399 kN
W c 1.67

61
Example: For the compression member shown below determine design
compressive strength (Φc Pn) and allowable compressive strength (Pn / Ωc ).

HEM120 column S355 Fy = 355 MPa Pinned at top and bottom. Intermediate
bracing provided at mid height, in weak direction only.

rx  55 .1 mm
2m

ry  32.5 mm

braces weak
2m

direction only

62
 K L 1.0  4000  K L 1.0  2000
    72.6     61.5
 r x 55.1  r y 32.5

Strong axis buckling controls

 2  200  103 200  103

Fe   375 MPa 4.71  111.8
72.62 355

KL E  355

 4.71 Fcr  0.658   355  239 MPa
375
r fy
 

6641
Pn  239   1587 kN
1000

Pn 1587
c Pn  0.9  1587  1428 kN   950 kN
W c 1.67

63
LOCAL BUCKLING

64
Local Buckling is related to Plate Buckling

Flange is restrained by the web at one edge.

Failure is localized at areas of high stress

(maximum moment) or imperfections.
65
Local Buckling is related to Plate Buckling

Flange is restrained by the web at one edge.

Failure is localized at areas of high stress

(maximum moment) or imperfections.
66
Local Buckling is related to Plate Buckling

Flange is restrained by the web at one edge.

Failure is localized at areas of high stress

(maximum moment) or imperfections.
67
 Local Buckling is related to Plate Buckling

Web is restrained by the flanges.

Failure is localized at
areas of high stress
(maximum moment) or
imperfections.

68
 Local Buckling is related to Plate Buckling

Web is restrained by the flanges.

Failure is localized at
areas of high stress
(maximum moment) or
imperfections.

69
 Local Buckling is related to Plate Buckling

Web is restrained by the flanges.

Failure is localized at
areas of high stress
(maximum moment) or
imperfections.

70
 Local Buckling:
 Criteria in Table B4.1
 Strength in Chapter E: Members with Slender Elements

71
Local Buckling Criteria
Slenderness of the flange and web, l, are used as criteria to
determine whether local buckling might control in the elastic or
inelastic range, otherwise the global buckling criteria controls.

Criteria lr are based on plate buckling theory.

For I-Shapes

E
FLB, l = bf /2tf lrf = 0.56
Fy

E
WLB, l = h/tw lrw = 1.49
Fy

72
To qualify as member
with nonslender
elements
l <lr

More detailed
discussions on local
buckling will be
provided in the flexural
members chapter

CE388 covers behavior

of cross sections with
nonslender elements
73
FULL STRUCTURE BEHAVIOR

74
 ALIGNMENT CHART
OR
DIRECT ANALYSIS METHODS

CE388 FOCUS: ALIGNMENT CHART

75
ALIGNMENT CHART

“Traditional Method”

Determine effective length, KL,

for each column.

Basis for design similar to

individual columns.

Does not redistribute restraining

moments into girders/beams.
76
DIRECT ANALYSIS METHOD

Analysis of entire structure interaction.

Include lateral “Notional” loads.

Reduce stiffness of structure.

All members must be evaluated under combined

axial and flexural load.

No K values required.
NOT COVERED IN CE388 77
ALIGNMENT CHART

“Traditional Method”

Determine effective length, KL,

for each column.

Basis for design similar to

individual columns.

Does not redistribute restraining

moments into girders/beams.
78
K-FACTORS FOR END CONSTRAINTS

No Joint Translation Allowed – Sidesway Inhibited

0.5  K  1.0

Joint Translation Allowed – Sidesway Uninhibited

1.0  K  

79
K-FACTORS FOR END CONSTRAINTS

Two categories,
Braced Frames, 0.5  K  1.0
Sway Frames, K ≥ 1.0

Behavior of individual column unchanged

(Frame merely provides end conditions).

80
Sidesway Prevented

Floors do not translate relative

to one another in-plane.

Typically, members are pin

connected to save cost.

81
 Sidesway Prevented

Assume girder/beam
K=0.5 K=0.7 infinitely rigid or
flexible compared to
columns to bound
results.

K=0.7 K=1

82
 Sidesway Prevented
Idealized Equivalent

Shear
Wall

83
 Sidesway Prevented
Idealized Equivalent

Shear
Wall

84
 Sidesway Prevented
Idealized Equivalent

Shear
Wall

85
 Sidesway Prevented

Typically, members are pin-connected to save cost (K = 1).

If members include fixity at connections,

Alignment Chart Method to account for rotational restraint (K < 1).

Typical design will assume K = 1 as a conservative upper bound

(actual K ≈ 0.8 not much difference from K = 1 in design).

86
 Sway Frame

Floors can translate relative to

one another in-plane.

Enough members are fixed to

provide stability.

Number of moment frames

chosen to provide reasonable
force distribution and
redundancy.

87
 Sway Frame

K=1 K=2 Assume girder/beam

infinitely rigid or
flexible compared to
columns to bound
results.

K=2 K=∞

88
 Sway Frame

Moment Frame

89
 Sway Frame

Moment Frame

90
 Sway Frame

Moment Frame

91
 Alignment Charts

Calculate “G” at the top and bottom of the column (GA and GB).

 EI 
  L 
G columns

 EI 
  L 
girders

I = moment of inertia of the members

L = length of the member between joints

G is inversely proportional to the degree

of rotational restraint at column ends.

92
 Alignment Charts

Sidesway Inhibited Sidesway UnInhibited

(Braced Frame) (Sway Frame)

Separate Charts for Sidesway Inhibited and Uninhibited

93
 Alignment Charts

X Gtop

GbottomX
Gtop
X

X
Gbottom

Sidesway Inhibited Sidesway UnInhibited

(Braced Frame) (Sway Frame)

Separate Charts for Sidesway Inhibited and Uninhibited

94
 Alignment Charts

X Gtop
K

GbottomX
Gtop
X
K

X
Gbottom

Sidesway Inhibited Sidesway UnInhibited

(Braced Frame) (Sway Frame)

Separate Charts for Sidesway Inhibited and Uninhibited

95
 Alignment Charts

Only include members RIGIDLY

ATTACHED (pin ended members are
not included in G calculations).

Use the IN-PLANE stiffness Ix if in major axis

direction, Iy if in minor axis.
Girders/Beams are typically bending about Ix
when column restraint is considered.

If column base is “pinned” – theoretical G = ∞.

AISC recommends use of 10.
If column base is “fixed” – theoretical G = 0.
AISC recommends use of 1.
96
 Alignment Charts
ALIGNMENT CHART ASSUMPTIONS:
1) Behavior is purely elastic.
2) All members have constant cross section.
3) All joints are rigid.
4) Sidesway Inhibited (Braced) – single curvature bending of
girders.
5) Sidesway Uninhibited (Sway) – reverse curvature bending of
girders.
6) Stiffness parameter of all columns is equal.
7) Joint restraint is distributed to columns above and below the joint
in proportion to EI/L of the columns.
8) All columns buckle simultaneously.
9) No significant axial compression force exists in the girders.

97
Alignment Charts

Let’s evaluate the assumptions.

98
 Alignment Charts
ALIGNMENT CHART ASSUMPTIONS:
1) Behavior is purely elastic.
2) All members have constant cross section.
3) All joints are rigid.
4) Sidesway Inhibited (Braced) – single curvature bending of
girders.
5) Sidesway Uninhibited (Sway) – reverse curvature bending of
girders.
6) Stiffness parameter of all columns is equal.
7) Joint restraint is distributed to columns above and below the joint
in proportion to EI/L of the columns.
8) All columns buckle simultaneously.
9) No significant axial compression force exists in the girders.

99
 Alignment Charts
If the column behavior is inelastic,

Yielding decreases stiffness of the column.

Relative joint restraint of the girders increases.

G therefore decreases, as does K.

Decrease is typically small.

Conservative to ignore effects.

Can account for effects by using a

stiffness reduction factor (SRF), t, times G.

100
 Alignment Charts
ALIGNMENT CHART ASSUMPTIONS:
1) Behavior is purely elastic.
2) All members have constant cross section.
3) All joints are rigid.
4) Sidesway Inhibited (Braced) – single curvature bending of
girders.
5) Sidesway Uninhibited (Sway) – reverse curvature bending of
girders.
6) Stiffness parameter of all columns is equal.
7) Joint restraint is distributed to columns above and below the joint
in proportion to EI/L of the columns.
8) All columns buckle simultaneously.
9) No significant axial compression force exists in the girders.

101
 Alignment Charts

Only include members RIGIDLY

ATTACHED (pin ended members are not
included in G calculations).

Partial restraint of connections and non-

uniform members effectively change the
rotational stiffness at the connections.

These conditions can be directly accounted

for, but are generally avoided in design.

102
 Alignment Charts
ALIGNMENT CHART ASSUMPTIONS:
1) Behavior is purely elastic.
2) All members have constant cross section.
3) All joints are rigid.
4) Sidesway Inhibited (Braced) – single curvature bending of
girders.
5) Sidesway Uninhibited (Sway) – reverse curvature bending of
girders.
6) Stiffness parameter of all columns is equal.
7) Joint restraint is distributed to columns above and below the joint
in proportion to EI/L of the columns.
8) All columns buckle simultaneously.
9) No significant axial compression force exists in the girders.

103
 Alignment Charts

Calculation of G accounts for rotational stiffness

restraint at each joint based on assumed bending.

For other conditions include a correction factor

“m” to account for actual rotational stiffness of the
girder at the joint.

 EI 
  L 
G columns

 EI 
  m L 
girders

104
 Alignment Charts

Sidesway Inhibited (Braced)

Bending Stiffness =
2EI Assumption: single curvature
L bending of girder.

3EI
Bending Stiffness = Far end pinned
L
m = (3EI/L)/(2EI/L) = 1.5

4EI
Bending Stiffness =
L Far end fixed
m = (4EI/L)/(2EI/L) = 2

105
 Alignment Charts

Sidesway Uninhibited (Sway)

6EI Assumption: reverse curvature
Bending Stiffness =
L bending of girder.

3EI
Bending Stiffness =
L
Far end pinned
m = (3EI/L)/(6EI/L) = 1/2

4EI
Bending Stiffness = Far end fixed
L
m = (4EI/L)/(6EI/L) = 2/3

106
107
 Alignment Charts
ALIGNMENT CHART ASSUMPTIONS:
1) Behavior is purely elastic.
2) All members have constant cross section.
3) All joints are rigid.
4) Sidesway Inhibited (Braced) – single curvature bending of
girders.
5) Sidesway Uninhibited (Sway) – reverse curvature bending of
girders.
6) Stiffness parameter of all columns is equal.
7) Joint restraint is distributed to columns above and below the joint
in proportion to EI/L of the columns.
8) All columns buckle simultaneously.
9) No significant axial compression force exists in the girders.

108
 Alignment Charts

In general, columns are chosen to be a similar size for more

than one story. For each column section this results in sections
with extra strength in upper floors, and close to their strength
in lower floors.

Design typically checks each story

independently, based on these assumptions.

Actual conditions can be directly accounted

for, but are generally ignored in design.

109
Example: Determine design compressive strength (Φc Pn) and allowable
compressive strength (Pn / Ωc ) for IPN380 column. S275 steel Fy = 275 MPa

All connections are rigid except one. Columns are braced out of plane at their ends
(K=1 for all columns out of plane). Use recommended G values for column bases.

IPN340
3.5 m

IPN450 IPN400

A pin connection

IPN380
3.5 m

H H H
5m 4m
H 6m
Section Strong Ix (cm4) Weak Iy (cm4)
IPN380 24010 975
IPN340 15700 674
IPN400 29210
IPN450 1730 110
 975 674  K L 1.0  3500
     23.3
GA  350 350  0.06  r x 150
1730 29210
 1.5 
400 600
 K L 0.7  3500
    81 (Governs)
GB  10 (Recommended)  r y 30.2

 2  200  103
K  0.7 Fe 
812
 301 MPa

rx  150 mm KL E 200  10 3

 4.71   4.71  127
r fy 275
ry  30.2 mm
 275

Fcr  0.658 301   275  188 MPa
 

10700
Pn  188   2011 kN c Pn  0.9  2011  1810 kN
1000

Pn 2011
  1204 kN
W c 1.67
111
Example: Determine design compressive strength (Φc Pn) and allowable
compressive strength (Pn / Ωc ) for IPN380 column. S275 steel Fy = 275 MPa

All connections are rigid except one. Columns are braced out of plane at their ends
(K=1 for all columns out of plane). Use recommended G values for column bases.

IPN340
3.5 m

IPN450 IPN400

A pin connection

IPN380
3.5 m

H H H H
5m 4m 6m
Section Ix (cm4)
IPN380 24010
IPN340 15700
IPN400 29210
IPN450 45850 112
 24010 15700

GA  350 350  0.81 GB  1.0 (Recommended) K  1.3
45850 29210
 0.5 
400 600

 K  L  1.3  3500  K  L  1.0  3500

    30     116 (Governs)
 r x 150  r y 30.2

 2  200 103 KL E 200  10 3

Fe   147 MPa  4.71   4.71  127
1162 r fy 275

 275

Fcr  0.658   275  126 MPa
147

 

10700
Pn  126   1348 kN c Pn  0.9 1348  1213 kN
1000

Pn 1348
  807 kN
Wc 1.67
113