Design of Compression

Members
Dr Bijily Balakrishnan
 Introduction

• Classification of Columns Based on Type of Transverse

Reinforcement

tie spiral

longitudinal
bars

(a) tied column (b) spiral column

structural
steel section

(c) composite column

2
 Introduction
• Classification of Columns Based on Type of Loading
ex = Mx/P
P e = M/P P ey = My/P P

centroidal axis

ELEVATION
Dx
D
X
CROSS ey
Dy Y Y
SECTION
e
ex
X

(a) (b) (c)

axial loading loading with loading with
(concentric) uniaxial eccentricities biaxial eccentricities

Different loading situations in columns

 Introduction
• Classification of Columns Based on Slenderness Ratios
• Short columns
• Slender (or long) columns
• Estimation of Effective Length of a Column
• The effective length, le is defined as the length of a hinged-ended column
whose elastic critical load is the same as the theoretical critical load of the
framed member under consideration
• It follows that the effective length of a column for a given plane of
bending is the length between the theoretical points of inflection in the
buckled configuration of the column in that plane
• The effective length depends on the unsupported length l (i.e., distance
between lateral connections, or actual length in case of a cantilever) and
the boundary conditions at the column ends introduced by connecting
beams and other framing members
le= kl
 Introduction
• where k is the effective length ratio (also known as effective length
factor) whose value depends on the degrees of rotational and/or
translational restraints at the column ends
• Unsupported Length
• The Code (Cl. 25.1.3) defines the ‘unsupported length’ l of a column
explicitly for various types of constructions
• In conventional framed construction, l is to be taken as the clear
distance between the floor and the shallower beam framing into the
columns in each direction at the next higher floor level
• When a column is framed in any direction by beams of different
depths on either side, then the unsupported length (with respect to
buckling about a perpendicular axis) shall be considered,
conservatively, with reference to the shallower beam
 Introduction
• For a rectangular column section (width Dy × depth Dx), we may use the
terms, lex = kx lx and ley = ky ly to denote the effective lengths referring to
buckling about the major and minor axes respectively, where lx and ly
denote the corresponding unsupported lengths and kx and ky denote the
corresponding effective length factors
y

x x
Dx (a) plan
Major axis
Dy Dy  Dx
Minor axis Iy  Ix
y

upper floor

slenderness ratios:
k xlx k yl y
,
Dx Dy

lx
ly

top of lower floor (b) section at x-x

Dy
 Introduction
• Effective Length (Idealised Boundary Conditions)
P

P P P

le = 0.7l
le = l l
le = 0.5l l le = kl

P P 0.5 < k < 1.0

P
P

(a) (b) (c) (d)

both ends one end rotationally both ends both ends partially
rotationally fixed fixed, the other free rotationally free restrained
(rotational)
 Introduction
P

P P
P

0.5l

le = kl
l le =  l
l

le = 2l
le = l

P P
P
P

(c) 1<k<
(a) both ends
both ends rotationally free
rotationally fixed (d)
both ends partially
restrained
(b) (rotational)
one end rotationally
fixed, the other free
 Introduction
• Effective Lengths of Columns Braced Against Sideway
• Code recommended values of k
Conditions k
Columns braced against sideway
both ends ‘fixed’ rotationally 0.65 (instead of 0.5)
one end ‘fixed’ and the other ‘pinned’ 0.80 (instead of 0.7)
both ends ‘free’ rotationally (‘pinned’) 1.00
Columns unbraced against sideway
both ends ‘fixed’ rotationally 1.20 (instead of 1.0)
one end ‘fixed’ and the other ‘partially 1.50
fixed’
one end ‘fixed’ and the other free 2.00
 Introduction
Use of Code Charts for realistic estimate of the effective length
• Charts are given in Fig. 26 and Fig. 27 of the Code for determining
the effective length ratios of braced columns and unbraced
columns respectively, in terms of coefficients 1 and 2 which
represent the degrees of rotational freedom at the top and bottom
ends of the column
I h c s

  jt
for braced columns
 I h   0.5(I
jt
c s
jt
b lb )

I h c s

  jt
for unbraced columns
 I h   1.5(I
jt
c s
jt
b lb )

where the notation jt denotes that the summation is to be done for the members framing into
the top joint (in case of β1) or the bottom joint (in case of β2)
 Introduction
• The increased beam stiffness for unbraced columns, compared to
braced columns, is attributable to the fact that in the case of the
latter (braced), the beams are bent in single curvature, whereas in
the case of the former, the beams are bent in double curvature, in
the buckled configuration
• The limiting values β = 0 and β = 1 represent ‘fully fixed’ and ‘fully
hinged’ conditions respectively
Introduction
 Slenderness Limits
• The Code (Cl. 25.3.1) specifies that the ratio of the unsupported
length (l) to the least lateral dimension (d) of a braced column
should not exceed a value of 60
• In the case of ‘unbraced’ columns, it is desirable to adopt a more
stringent limit — l/d < 40
• In case one end of a column is free (i.e., cantilevered column) in any
given plane, the Code (Cl. 25.3.2) specifies that
l  100b2 D
where D is the depth of the cross-section measured in the plane of the
cantilever and b is the width (in the perpendicular direction)
 Minimum Eccentricities
• General case of loading on a compression member is one comprising axial
compression combined with biaxial bending (biaxial eccentric
compression)
• Very often, eccentricities not explicitly arising out of structural analysis
calculations act on the column due to various reasons
• lateral loads not considered in design
• live load placements not considered in design
• accidental lateral/eccentric loads
• errors in construction (such as misalignments)
• slenderness effects underestimated in design
l 500  Dx 30 l 500  Dy 30
ex ,min   whichever is greater e y ,min   whichever is greater
 20 mm  20 mm

For non-rectangular and non-circular cross-sectional shapes  l 300

emin   e whichever is greater
20 mm
 Code Requirements on Reinforcement and Detailing

• Longitudinal Minimum Reinforcement (refer Cl. 26.5.3.1)

• 0.8 percent of the gross area

• The limit is specified:

• to ensure nominal flexural resistance under unforeseen eccentricities
in loading
• to prevent the yielding of the bars due to creep and shrinkage effects,
which result in a transfer of load from the concrete to the steel
• In very large-sized columns by architectural considerations, and not
strength under axial compression: 0.8 percent of the area of
concrete required to resist the direct stress, and not the actual
(gross) area
 Code Requirements on Reinforcement and Detailing

• Cover to reinforcement: minimum clear cover of 40 mm

• Transverse Reinforcement:
• All longitudinal reinforcement in a compression member must be
enclosed within transverse reinforcement, comprising either lateral
ties (with internal angles not exceeding 135o) or spirals
• Required to prevent the premature buckling of individual bars; to
confine the concrete in the ‘core’, thus improving ductility and
strength; to hold the longitudinal bars in position during construction;
and to provide resistance against shear and torsion, if required
• Lateral Ties:  4
tie diameter, t   long , max
6 mm
 D

tie spacing, st   16long , min
300 mm

Code Requirements on Reinforcement and Detailing
clear cover
to links spiral
 300  40 mm
ties

longitudinal
bar
(a) (b)

common
closed tie
 300

(c) (d)

 75 mm  75 mm
 48 t

closed
tie
open
(cross)
ties
(e) (f)
 Code Requirements on Reinforcement and Detailing

• Longitudinal Maximum Reinforcement (refer Cl. 26.5.3.1)

• 6 percent of the gross area
• a reduced maximum limit of 4 percent is recommended in general in the
interest of better placement and compaction of concrete
• Minimum diameter / number of bars and their location
• Longitudinal bars in columns should not be less than 12 mm in diameter
• Should not be spaced more than 300 mm apart (centre-to-centre) along
the periphery of the column
• At least 4 bars (one at each corner) should be provided in a column with
rectangular cross-section, and at least 6 bars (equally spaced near the
periphery) in a circular column
• In columns with T-, L-, or other cross-sectional shapes, at least one bar
should be located at each corner or apex
 Design Strength of Axially Loaded Short Columns

• The maximum compressive strain in concrete under axial loading at the

limit state of collapse in compression is specified as εc = 0.002 by the Code
(Cl. 39.1a)
Po  Cc  C s  fcc Ac  fsc Asc
 Po  fcc Ag  ( fsc  fcc ) Asc
0.870 fy for Fe 250

Puo  0.447fck Ag  ( fsc  0.447fck ) Asc fsc   0.790 fy for Fe 415
0.746 f for Fe 500
 y

• The Code requires all columns to be designed for ‘minimum eccentricities’

in loading
• The calculated minimum eccentricity (in any plane) does not exceed 0.05
times the lateral dimension (in the plane considered), obtained by
reducing Puo by approximately 10 percent (Cl. 39.3)
Puo  0.4fck Ag  (0.67fy  0.4fck ) Asc
Conditions of Axial Loading
 Behaviour of axially loaded tied and spiral columns

Puo B E spiral
columns
A C
D

tied columns
axial
load
Po

o axial shortening

• Based on experimental findings , the Code (Cl. 39.4) permits a 5% increase in

strength beyond Puo for columns with spirals, provided the following requirement
is satisfied by the spiral reinforcement:
 Design of Short Columns
Under Compression With Uniaxial Bending

• This loading condition is statically equivalent to a condition of

uniaxial eccentric compression wherein the factored axial load Pu is
applied at an eccentricity e = Mu/Pu with respect to the centroidal
axis, Mu being the factored bending moment
Distribution of Strains at Ultimate Limit State

cu = 0.0035 – 0.75 c, min for xu  D

 3D 7 
cu  0.002 1   for xu  D
 xu  3D 7

e=∞
e = eb
xu,bal
εy xu,min tension
εc,min e = eD compression
e=0 εcu = 0.002
Pivot
xu = ∞ 3D/7 εcu = 0.035
xu = D
xu > D

Strain Profile
 Stress blocks at Ultimate Limit State
D D highly highly D D
centroidal centroidal
compressed compressed centroidal centroidal
d d
axis
d
axis edge
d edged  d
axis
d
axis
d PuR = Cc + Cs
least least
compressed compressed
edge edge
b b MuR = Mc + Ms
COLUMN COLUMN
SECTION SECTION

neutral neutral ith row th Asi Asi

axis
yi yi of steeli row of steel yi yi
axis (total area = A(total
si) area = Asi)
xu xu xu xu
Cc = a fck bD
neutral neutral

Mc  Cc (D 2  x )
axis axis

FAILURE
si si
FAILURE
si si n
C s    fsi  fci  Asi
STRAIN STRAIN
PROFILE PROFILE 0.002 0.002
PIVOT PIVOT
cu cu i 1
cu = 0.0035 cu = 0.0035 3D/7 3D/7
0.0035 0.0035 n
D/14 D/14 Ms    fsi  fci  Asi yi
Csi Csi STRESS STRESS Csi Csi i 1
RESULTANT RESULTANT
S S

0.447fck 0.447fck 0.447fck 0.447fck

Cc Cc Cc Cc
x x
x x
(a) xu  D (a) xu  D (b) xu > D (b) xu > D
 Stress blocks at Ultimate Limit State

n n
C s    fsi  fci  Asi Ms    fsi  fci  Asi yi Cc = a fck bD Mc  Cc (D 2  x )
i 1 i 1

 where a  stress block area factor

0 if si  0
 x  distance between highly
fci   0.447fck if  si  0.002

compressed edge and the line
0.447fck 2( si 0.002)  ( si 0.002)  otherwise
2
of action of Cc

  xu  D 2  yi   0.362 xu D for xu  D
0.0035   for xu  D a

 si  
 xu  0.447(1  4 g 21) for xu  D
 0.002 1  yi  D 14  for x  D
    0.416xu for xu  D
 xu  3D 7 
u
 x
(0.5  8g 49) D (1  4 g 21) for xu  D

where g  16
(7xu D  3)2
 P-M Interaction Diagram
Pu

Pu0
Pu 0

e = emin
Design Interaction curve
(Pu = PuR, Mu = MuR)
e=0
1 e = eD
e
e < eb Compression failure

Pub Balanced failure

e = eb
e > eb Tension failure

e=∞ Muo Mu = Pue

Mub
 Design Charts (for Uniaxial Eccentric
Compression) in SP: 16
• Rectangular sections with
reinforcement distributed
equally on two sides (Charts
27 – 38):

bending axis
D/2 D/2

NO
INNER b
0.5As 0.5As
ROWS

d d

(a) rectangular section with

“reinforcement distributed equally on two sides”
 Design Charts (for Uniaxial Eccentric
Compression) in SP: 16
• Rectangular sections with
bending axis
reinforcement distributed
D/2 D/2

equally on four sides (Charts

NO
INNER b
0.5As 0.5As
39 – 50): ROWS

d d

(a) rectangular section with

“reinforcement distributed equally on two sides”

bending axis
D/2 D/2

min. 0.3As in each

AT LEAST TWO outermost row
b
INNER ROWS

d
d

(b) rectangular section with

“reinforcement distributed equally on all sides”

bending axis

AT LEAST 6 BARS
(EQUAL DIA)
 D/2 D/2

Design Charts (for Uniaxial Eccentric

0.5As
NO
INNER
ROWS
0.5As
b

d Compression) in SP: 16
d

(a) rectangular section with

“reinforcement distributed equally on two sides”

• Circular column sections bending axis

D/2 D/2

(Charts 51 – 62): min. 0.3As in each

AT LEAST TWO outermost row
b
INNER ROWS

d
d

(b) rectangular section with

“reinforcement distributed equally on all sides”

bending axis

AT LEAST 6 BARS
(EQUAL DIA)

(c) circular column section

 Design Charts (for Uniaxial Eccentric
Compression) in SP: 16
• Corresponding to each of the above three cases, there are as many
as 12 Charts available — covering the 3 grades of steel (Fe 250,
Fe 415, Fe 500), with 4 values of d’/D ratio for each grade (viz., d’/D
= 0.05, .0.10, 0.15, 0.20). For intermediate values of d’/D, linear
interpolation may be done. Each of the 12 Charts of SP : 16 covers a
family of non-dimensional design interaction curves with p/fck values
ranging from 0.0 to 0.26.
• If bars of equal diameter are used, this is equivalent to using 20
bars. While actually providing reinforcement to conform to As
computed using these Charts, some adjustments may be called for
in practice. Providing a greater proportion of reinforcement (more
than 0.3As) on the outermost rows is on the safer side.
Short Columns with Biaxial Eccentricities
Short Columns with Biaxial Eccentricities
1 2
neutral axis
MuR,y

PuR
MuR,x
s1

bending 3 4
axis
Cs1 s2
tension
s3
Cs2

s4
Cs3
compression
Cs4 cu
Cc
failure
stress strains
resultants
Interaction Surface for a Biaxially Loaded
Column Y

y
X X
Pu

x ex
ey
Pu
Puo

MuR,y
PuR

MuR,x  Muy1

Mux1
load
contour

0
Muy = Pu ey

Mu  Mux
2
 Muy
2

Mux = Pu ex
 Code Procedure for Design of
Biaxially Loaded Columns
• The simplified method adopted by the Code (Cl. 39.6) is based on
Bresler’s formulation for the ‘load contour’
n n
 Mux   Muy 
0
0.25 0.50 0.75 1.00 Mux/Mux1       1
 ux 1 
M M
 uy 1 
n
0.25
n = 1.0
2.0

0.50

1.0
0.75
n n Pu/Puz
 Mux   Muy 
    
 1
1.00  Mux1   Muy1  0 0.2 0.8 1.0
n = 2.0

Muy/Muy1
(a) (b)

Puz = 0.45fck Ag + ( 0.75fy – 0.45fck) Asc

 Code Procedure
• Given Pu, Mux, Muy, verify that the eccentricities ex = Mux/Pu and ey =
Muy/Pu are not less than the corresponding minimum eccentricities
• Assume a trial section for the column
• Determine Mux1 and Muy1, corresponding to the given Pu (using
appropriate design aids). Ensure that Mux1 and Muy1 are significantly
greater than Mux and Muy respectively; otherwise, suitably redesign
the section
• Determine Puz, and hence αn
• Check the adequacy of the section; if necessary, redesign the
section and check again.
 Design of Slender Columns
P Pmax Pe P
Pmax OAP short column P
P
P∆ OA  short column
max Pe
e
(material failure) (material failure)
OB  long column
OB  long column
(material failure)
 M= P(e+) (material failure)
Ps e M=P(e+∆)OC M very
= P(e+ ) column
long
l/2
l/2 l/2 A Ps e OC  very long column
Ps (instability failure)
Ps A
(instability failure)
max M
P1 e
B P1 e
e max
max P1 Mmax P1 B
l e
l Mmax1 P11
e 1 P11
P2 e
interactionP2 interaction
C
(failure) C
P-∆moment (failure)
Primary curve curve
primary moment primary
moment 0
P–moment momentM 0 M
P P–moment
P

(a) (b) (a) (c)

(b) (c)
• Variation of Mmax with P is nonlinear, with Mmax increasing at a faster rate as P increases.
The axial thrust P effectively reduces the flexural stiffness of the column, and in the
case of a very slender column, the flexural stiffness can be effectively reduced to zero
(dP/dM = 0), resulting in an instability (buckling) failure
• In the case of a very short column, the flexural stiffness is so high that the lateral
deflection Δ is negligibly small, consequently, the P- Δ moment is negligible, and the
primary moment Mpr alone is of significance
 Braced Slender Columns:
Member Stability Effect
P P P P
M2 M
M22 M2 M2 M22
M M2

max max

max 
P 2
max
P2 P2 P2

P1 P1
l l P1 P1
2 2
1 1

2 2
1 1
M1 M1 M1 M1
M1 M1M1 M1
P P P P

(a) single (a) single curvature

curvature (b) double(b) double curvature
curvature

• A ‘braced column’ is one which is not subject to sidesway, i.e., there is no significant relative
lateral displacement between the top and bottom ends of the column.
• The primary moments M1 and M2 that are applied at the two ends of the column are
determined from a ‘first-order’ structural analysis
• The column may be bent in single curvature or double curvature, depending on the directions
of M1 and M2. The notations M1 and M2 generally refer to the smaller and larger column end
moments, and the ratio M1/M2 is considered positive if the column is bent in single curvature,
and negative if it is bent in double curvature.
 Braced Slender Columns:
Member Stability Effect
• The chances of a given slenderness resulting in a peak design moment
larger than M2 fall off significantly as the ratio M1/M2 drops below about
+0.5 and approaches the limit of –1.0
• The possible amplification in bending moment (over the primary moment
M2) on account of lateral displacements (relative to the chord joining the
column ends) is termed as member stability effect.
• The ACI Code recommends that slenderness effects may be ignored (i.e.,
the column may be designed as a ‘short column’) if, for a braced column,

 34  12  1 
le M
r  M2 

where le is the effective length and r the radius of gyration. Thus, the
slenderness ratio (le/r) limit for short columns lies in the range 22–34 in single
curvature and 34–46 in double curvature.
 Unbraced Slender Column: Lateral Drift Effect
P P


• An ‘unbraced column’ is one which is 2H

B
subject to sideway (or ‘lateral drift’), ‘rigid’ beam 
i.e., there is significant lateral
displacement between the top and
bottom ends of the column. The lateral A

drift may occur due to the action of

lateral loads, or due to gravity loads (a) sway frame

when the loading or the frame is  P

P P
∆B P
asymmetric. Mo
M
Ho H HP∆B M1
B B B B’
• The additional moments at the column Member B lateral
drift
stability
total effect
ends caused by the action of the point of inflection
effect
Lateral drift H Primary
vertical load acting on the deflected effect
A membe
configuration of the unbraced column is H P
M2 P∆A stabilit
A
Mo effect
termed the lateral drift effect. H Mo
P
A
P

(b) swayed (c) forces

 Unbraced Slender Column: Lateral Drift Effect

Generally, for unbraced columns, the moments at the column ends

are maximum, and these are due to the primary moments enhanced
by the lateral drift effect alone.

The moment amplification possible due to lateral drift effect in an

unbraced column is generally much more than that due to member
stability effect in a braced column. Further, the effective length of
an unbraced column is much more than that of a braced column
with the same unsupported length. Hence, columns in unbraced
frames are weaker than similar columns in braced frames.
 Second-order Structural Analysis of
Slender Column Structures
 The Code (Cl. 39.7) broadly recommends that when slender columns are
involved in a reinforced concrete structure, a detailed ‘second-order’
structural analysis should be carried out to determine the bending
moments and axial forces for which the slender columns are to be
designed. Indeed, such a rigorous analysis is particularly desirable for
slender columns in unbraced frames. Such analysis must take into account
all slenderness effects, viz. the influence of column and frame deflections
on moments, effects of axial loads and effects of sustained loads.
Realistic moment-curvature relationships should be made use of.

 The principle of superposition is not valid in second-order analysis, and for

this reason, the load effects due to different load combinations cannot be
obtained by an algebraic summing up (with appropriate load factors);
each load combination should be investigated separately. This requires
substantial computational effort.
 Code Procedures for Design of
Slender Columns
• In routine design practice, only first-order structural analysis is
performed, as second-order analysis is computationally difficult and
laborious.
• In recognition of this, the Code recommends highly simplified
procedures for the design of slender columns, which either attempt
to predict the increase in moments (over primary moments), or,
equivalently, the reduction in strength, due to slenderness effects.
• Strength Reduction Coefficient Method
• This is a highly simplified procedure, which is given in the Code for the
working stress method of design. According to this procedure (B-3.3
of the Code) the permissible stresses in concrete and steel are
reduced by multiplication with a strength reduction coefficient Cr,
given by:
le
Cr  1.25  where d is the least lateral dimension of the column (or
48d diameter of the core in a spiral column)
 Code Procedures for Design of
Slender Columns
• It is recommended in the Explanatory Handbook to the Code that
instead of applying the strength reduction factor Cr to the
‘permissible stresses’, this factor may be directly applied to the
load-carrying capacity estimated for a corresponding short column.
Furthermore, it may be noted that although this method has been
prescribed for WSM, it can be extended to the limit state method
(LSM) for the case of axial loading (without primary bending
moments).
 Additional Moment Method
• The method prescribed by the Code (Cl. 39.7.1) for slender
column design by the limit state method is the ‘additional
moment method’

• According to this method, every slender column should be

designed for biaxial eccentricities which include the P–
moment (“additional moment”) components eax  Max/Pu and
eay  May/Pu :
Mux  Pu  ex  eax   Mux  Max

Muy  Pu  ey  eay   Muy  May

• An alternative method called the ‘moment magnification

method’ is adopted by the ACI and Canadian codes.
 Additional Moment Method
• In the basic formulation, the P– effect in a braced slender column with pin-joined ends
is considered. The ‘additional eccentricity’ ea is equal to max, which is a function of the
curvatures to which the column is subjected. If the maximum curvature is denoted as
max, it can be shown that max lies between maxl2/12 and maxl2/8, the former limit
corresponding to a linearly varying curvature (with zero at the pin joints and a
maximum of max at mid height) and the latter corresponding to a constant curvature
 = M/EI
along the column height 8
1

 max
M l2

φ1 1
l/2 l/2
φ2 2
φmax max max 2, max
∆max l l P

12  max
max l2
max φ2<φ<φ 1<  < 
2 1

M
DEFLECTION CURVATURE
• Taking an average value, ea = max(a) max l2/10 (b) case 1 (c) case 2

• Failure of the column at the ultimate limit state is expected to occur at the section
corresponding to max
 Additional Moment Method
• In the basic formulation, the P– effect in a braced slender column
with pin-joined ends is considered. The ‘additional eccentricity’ ea is
equal to max, which is a function of the curvatures to which the
column is subjected. If the maximum curvature (at mid-height) is
denoted as max, it can be shown that max lies between maxl2/12
and maxl2/8, the former limit corresponding to a linearly varying
curvature (with zero at the pin joints and a maximum of max at mid
height) and the latter corresponding to a constant curvature along
the column height
• Taking an average value, ea = max  max l2/10
• Failure of the column at the ultimate limit state is expected to occur
at the section corresponding to max
 Determination of Curvature from
Failure Strain Profile
• Assuming that cu = 0.0035 and s = 0.002† , d’ 0.1D and
further assuming (rather conservatively) that the additional
moment comprises about 80 percent of the total moment,
0.0035  0.002 1 Pu
max   0.8  e + max
0.9D 200D
ea (l D)2

D 2000

PuDx
 lex Dx 
2
Max  Pu eax 
2000
PuDy d D  d
 ley Dy 
2
May  Pu eay  st
  ( cu   st ) ( D  d  )
2000
cu

†This approximately corresponding to the ‘balanced failure’ condition, whereby

st = y at the cracked section. For deflection calculations, the mean steel strain
should be considered, including the effect of ‘tension stiffening’
 Determination of Curvature from
Failure Strain Profile
• The height l has been replaced by the effective length le to extend
the application of the formulation to the various boundary
conditions (other than the pinned-end condition) that occur in
practical columns including unbraced columns. It is reported that
the use of above expressions has been validated with reference to a
large number of experimental tests
• ea/D ratio increases with the square of the slenderness ratio
le/D; ea/D has a minimum value of 0.072 for le/D =12 (transition
between ‘short column’ and ‘slender column’) and a
maximum value of 0.450 for le/D = 30 (recommended limit for
unbraced columns) and 1.800 for le/D = 60 (braced column).
 Determination of Curvature from
Failure Strain Profile
• The derivation assumes that the column is braced and bent symmetrically in single
curvature: some modification is required when the primary moments applied at
the column ends are unequal and/or of different signs:

Mu = 0.4M1 + 0.6M2  0.4M2

Mu  M2 for braced columns

• Further, it is assumed that the axial load level corresponds approximately to the
‘balanced failure’ condition Pu = Pub; modification for axial load levels is required

• For Pu > Pub, the failure mode is one of ‘compression failure’, and the corresponding
e/D ratio is low. The additional moments Max, May may be reduced by multiplying
factors defined as: P P
kax  uz u
for Pu  Pub ,x
Puz  Pub ,x

Puz  Pu
kay  for Pu  Pub ,y
Puz  Pub ,y
 Determination of Curvature from
Failure Strain Profile
• where Puz is the maximum ‘pure compression’ strength of the column and Pub,x and
Pub,y correspond to the axial strength corresponding to balanced failure with
respect to bending about the major axis and minor axis respectively. Pub from the
interaction curve corresponding to a design tensile stress of fyd = 0.87 fy in the
outermost layer of steel.
• It can be seen that k varies linearly from zero (for Pu = Puz) to unity (for Pu = Pub) and
is a highly simplified formula.
• In the case of unbraced columns, the lateral drift effect (hitherto not considered)
needs to be included. An approximate way of accounting for this is by assuming
that the additional moment Ma acts at the column end where the maximum
primary moment M2 is operational. Hence, for design purposes, the total moment
may be taken as:
Mu  M2  Ma

† It is inadvisable to apply the reduction factor k for unbraced columns.

 Members with Axial Tension and Bending

Tu  Cu
xu  xu , purebending
 cu  0.0035
 st   y
Pu 0(  )  0.87 f y Ast
51
 Axial Tension and Slenderness Effects

• In the case of bending with axial compression, the slenderness of

the member is important because of the P – Δ effect, where the
compressive force P interacting with lateral deflection Δ causes
increase in the primary moments.
• The tensile axial force will straighten the member and P – Δ effect
will only reduce the primary moment.
• As such, slenderness is not a factor in members with axial tension
and there is no distinction for such a member as short or long.
 Axial Tension with Biaxial Bending
• The Pu - Mu interaction diagram on the tension side is very nearly
linear
• Under biaxial bending with axial tension, the interaction surface may
be taken as the linear extension of the trace on the X-Y plane of the
interaction surface to the point with co-ordinates (0, 0, – Pu0(-)).
• The interaction surface will be conical (inverted)
• The interaction surface on the tension side will be a plane
represented by
 Pu()   Mux   Muy 
        1.0
 Puo()   Muxo   Muyo 
Thank you