Euler Column Buckling Theory;
Effects of Residual Stresses
MORGAN STATE UNIVERSITY
SCHOOL OF ARCHITECTURE AND PLANNING
LECTURE III
Dr. Jason E. Charalambides
What is Residual Stress?
Although Steel is considered to be homogenous material, the
process of fabrication allows portions of an element to form
differently than others.
Rolled shapes may go through the rollers hot or cold
For cold rolled, it is understood that
the steel is exposed to stresses that
bring it into its plastic region to have
permanent deformation. When an
element is stressed to the point that it
deforms and it does not return to its
original form, portion of the energy
that was received remains within it.
That is translated to a stress that is
carried within the structure of that
element. That is residual stress.
2
What is Residual Stress?
Continuing on causes of residual stress,
For hot rolled elements, the residual stresses are not of the same
scale, but they can still be significant, especially on larger
elements. Those are developed by the uneven rate of cooling in
different areas. Areas that cool quicker, such as the middle of the
web or the tips of flanges of a W-Section tend to have residual
compressive stress, whilst areas that cool slower, such as the
intersections of web and flange, develop residual tensile stress.
What is Residual Stress?
Residual stress is addressed as L
drops of the scale of 10-15ksi may λ=
r
have an effect on the reliability of a
design. The greatest reductions in
strength are noticed in columns that
have a “Slenderness ratio” between
70 – 90.
Slenderness ratio is the result of the
division of the effective length “L”
over the radius of gyration “r”
– The former is essentially the
length of the element multiplied
by the “k” factor
– The latter is a factor that can be
found in the AISC User's Manual.
4
Qualities of Different
Shapes
Some shapes are more practical to fabricate
Some shapes have better response to compressive loads
Some shapes handle bending better
Qualities of Different
Shapes
Other advantages / disadvantages:
Round columns have less surface to paint or fireproof
Round columns have constanr “r” and “I” values
They have better torsional resistance and less resistance to
wind loads
Square or round columns are more economical and efficient
unless moments play an important role, especially in larger
structures
Hollow columns are easier to keep clean, but also easier to
be exposed to corrosion over W, S, or T shapes
6
Buckling
Main difference of a
compressive axially loaded
member over a tensile
axially loaded member is
“buckling.”
That is the “loss of compressive
load carrying capacity resulting
from a change in the geometric
formation of a member”
A slight defect, or a slight
eccentricity, may generate the
deflection that will lead to a
column's failure
NY
City
Transit
released
photos
of
Cortlandt
St
sta5on
the
week
of
24
September
2001.
Subway
columns
are
buckled
from
the
impact,
near
the
center
of
the
[Link]
of
this
is
7
[Link]:
hEp://[Link]/~brennan/abandoned/Cort-‐damage-‐[Link]
Buckling
Design equation:
The ultimate axial load is equal
or less than the factored
nominal strength
P u ≤Φ P n
8
Strength of Isolated
Columns
Euler's solution to
theoretical elastic behavior:
– Based on the following
assumptions:
The column is pin connected
P u ≤Φ P n
It is perfectly straight
Load is perfectly axial
Behaves elastically and does
not yield
No residual stresses
Bends and buckles about a
principal axis w/out torsion.
Strength of Isolated
Columns
π2 E I
Euler's elastic buckling: L
2
The buckled shape resembles ½ a
sinusoidal distribution.
The buckling load Pe is proportional to
the Moment of Inertia of the element
The buckling load is inversely
proportional to the square value of the
length of the element ( L 2 )
π2 E I
• The longer the element the more P E=
L2
susceptible to buckling
Buckling is proportional to the Young's
modulus of elasticity but independent of
the yield strength of the material (Fy)
10
Consider the Effects of
Axial Load on a W Shape
X & Y are Principal Axes
I_max = Ix (Strong Axis)
Higher Moment of Inertia
I_min = Iy (Weak Axis)
Lower Moment of Inertia
Buckling will be about the (y)
weak axis, unless the weak
axis is restrained.
11
Consider the Effects of
Axial Load on a W Shape
2
π EI x
2
L
2
π EI y
2
L
12
Addressing Euler's
Buckling Load w.r.t. Stress
Stress can be viewed as Load (P) divided by Area (A)
If we divide both sides of Euler's equation by Area →
PE π2 E I
=
A AL
2
But since the radius of gyration (r)is equal to the square root of
moment of inertia (I) divided by area (A)...
√ I I
r= ΟR r 2=
A A
Thus Euler's elastic buckling stress is:
2
π2 E r2 π E
F E= ΟR F E=
( )
2
L2 L
r
13
Bringing Back the Factor
of Boundary Conditions
π2 E
( )
KL 2
r
14
Load Deflection Behavior
The Length “L” divided by the
radius of gyration “r” is the π2E
slenderness ration of a column
By plotting a graph of the stress ( )
L 2
r
in the ordinates and the
slenderness ratio in the
abscissa...
r minimum corresponds to I
minimum
(L/r) max corresponds to r min
Weaker axis of W section (lower
I) controls in buckling
15
Load – Deflection Behavior
Effects of deflection
π2 E I
The column bends as soon as it is P E=
L2
loaded, i.e. buckling is not an
instantaneous effect.
There is already stress in the column
before loading.
Based on Elastic theory (material does
not yield) P is asymptotic to PE. No
loss of strength due to deflection
In reality material yields, and the
additional bending stress from
deflection causes earlier yielding and
loss of strength
Small deflection → little loss of
strength, and vice versa
16
Effects of Load
Eccentricity
Effects of eccentricity
(eccentrically applied
loading) are identical to the
effects of Δ0
17
Definition and Effects of
Residual Stresses
Residual stress definition
They are developed within a member during manufacturing.
They are self equilibrating (their sum is zero) as they exist in the
absence of any external loading.
They are generated by:
Uneven cooling of hot rolled elements Along
Uneven cooling of welded built up elements entire
length
Cold forming or cambering of members
Punching, shearing, or cutting
Welding at specific points Localized
Effect
Varying behavior at specific points, areas, or along an axis
18
Causes for Residual Stress
Cooling of Rolled Shapes
The uneven rate of cooling of the
cross section.
Member is allowed to coll slowly.
Some portions (e.g. flange tips)
cool quicker because they have
more surface exposed to air
Typically residual stresses are:
Quick cool → Compressive
Slow cool → Tensile
Residual stresses are normal, not
shear stresses ∫ σ res dA=0 ksi ksi
σ res ≈10 tο 15
Residual stresses are higher on
max
welded shapes than rolled shapes
19
The Stump Column Test
Residual stresses reduce
the stiffness of a member
Investigated by testing a
“stub column,” i.e. too short
to buckle.
If there are no residual
stresses, all fibers of the
cross section yield
simultaneously when the
applied load reaches A×F y
• ...i.e. when the applied
stress reaches P
=F y
A
20
The Stump Column Test
Residual stresses reduce the
stiffness of a member
If residual stresses are present
the first parts of the element that
will yield are the tips of the
flanges.
Then the effect will extend further
beyond the tips of the flanges and
the central portion of the web
And eventually the whole section
will yield
Although the maximum load will
still be P =A×F y , the load
deflection curve is not the same
21
Using the Tangent
Modulus
The P vs Δ curve..
Can be replotted in the P
σ avg = =average applied stress
form of average applied A
stress vs strain. Δ
ε= =applied strain
L0
Et is a measure of the
cross section's average
stiffness, considering
that portions of the
cross section are
yielded, while others
are still elastic
22
Effect of Residual Stresses
on Column Strength
Consider a column that is initially perfectly straight
The buckling load can be obtained using the “tangent
modulus theory” that was just discussed.
The buckling load can be computed using Euler's
equation, but replacing E with Et.
2
The resulting buckling load is referred to as P = π E t I
t 2
“Tangent Modulus buckling load” ( K ×L)
Similarly we can define the Pt
2
π Et
“Tangent Modulus buckling stress” F t = =
( )
2
A K ×L
This leads to two classes of buckling: r
• Elastic, and
• Inelastic
23
Elastic / Inelastic Buckling
Elastic
No yielding of the cross section
occurs prior to buckling and Et=E at
buckling
2
π E
2 Pt π Et
F E= F t=
A
( )
2
predicts buckling K× L
( )
2
K ×L
r r
2
Inelastic π E
( )
2
K×L
Yielding occurs on portions of the
r
cross section prior to buckling and
there is loss of stiffness.
2
π ET
FT=
predicts buckling
( )
2
K×L
r
24
Strength of Columns
The discussion that was held until now indicates that the
strength of a column is dependent upon the following:
L
Slenderness λ=
r
End restraint K factor
Eccentricity (loading or form) Must consider variability
Yielding and Residual stresses in these factors
All of the above factors need to be addressed in order to
determine the strength of a real column but there are two
approaches to do that:
Experiments (we shall not engage in this!)
Numerical Analysis
25
Numerical Method of
Analysis
The AISC provides a series of equations that allow us to
compute the column strength:
Nominal compressive strength P n = A g × F cr
where Ag is Area gross, and Fcr is the critical or buckling stress
Design compressive strength ΦP n = Φ× A g × F cr
where Φ is the factor of safety and it is equal to 0.9
Criterion for design P u ≤Φ × P n
26
Numerical Method of
Analysis
The AISC provides a series of equations that allow us to
compute the column strength:
Nominal compressive strength P n =A g ×F cr
where Ag is Area gross, and Fcr is the critical or buckling stress
Design compressive strength ΦP n =Φ× A g ×F cr
where Φ is the factor of safety and it is equal to 0.9
Criterion for design P u ≤Φ×P n
27
Computing the Nominal
Compressive strength
2
The definition is: π E for Euler's buckling Stress
F E=
( )
2
KL
r
√
KL E Fy
When ≤ 4.71 OR ≤ 2.25 → Inelastic Buckling
r Fy Fe
[ ( )] F
Fy
F cr = 0.658
FE (E3-2)
y
√
KL E Fy
When > 4.71 OR > 2.25 → Elastic Buckling
r Fy Fe
F cr =0.877Fe (E3-3)
For elastic buckling we adopt 0.877 times the Euler's formula,
accounting for geometric imperfections.
Note that Fcr is independent of Fy
28
The Red Line for
Elastic / Inelastic
Inelastic Buckling
(( [ )]
Fy
π2 E
F = [ 0.658 ] F = 0.658
( ) KL 2
)
Fy
FE r
cr y Fy
Note: As KL/r→ 0 Fcr→Fy
Elastic Buckling
2
π E
F cr =0.877FE =0.877
( )
2
KL
r
29
The Red Line for
Elastic / Inelastic
Taking the limit: KL
r
=4.71
E
Fy √ for for 36 and 50 grade steel
√
ksi
29,000
4.71 =133.681 → 36 ksi gives 133.7
36ksi
√
ksi
29,000
4.71
50
ksi
=113.432 → 50 ksi gives 113
Take a typical column e.g W12x53, of Lu=12' and r=2.48”.
With K of 1.0 this W section will give KL/r=58. In either of the
grades of steel this column will buckle in the inelastic range
30
Relation of Critical Stress
and Slenderness Ratio
At KL
r
2
= 4.71
E
Fy
2
√
π E π E Fy 2
π E
F E= 2= =0.44F y F E=
( )
2
KL 4.71 E
( )
2
KL
r r
Fy
Thus = 2.25
FE
F cr = 0.877FE = 0.877∗.44 F y= 0.39F y
( KL ) F
2
y
2 .877 F E
0.668 r π E
The transition from
elastic to inelastic
occurs at an applied
axial compression
stress of 0.39Fy
4.71
√ E
Fy
31
Basic Procedures for
Analysis
Given the shape, the K factor, the Length and the type of
steel can we determine the Φpn?
Well, ….. ΦP n = Φ× A g × F cr
Fcr depends uponKL/r and Fy,
The r we chose is the weaker one →(KL/r) max controls
Compute (KL/r)x and (KL/r)y, and larger will govern
32
Wide Flange Shape subjected to axial loading
Problem Statement:
Determine the capacity in axial loading of the given W shape. The element is pinned at top and bottom with
no intermediate bracing, therefore having an unbraced length of 15ft in both [Link] A992 steel
2
Area Ag := 15.8in Young's Modulus of Elasticity E := 29000ksi
Bolt diameter d b := 0.875in
Length: Lu := 15ft
radius of gyration y ry := 2.56in Yield Stress: Fy := 50ksi
radius of gyration x rx := 4.37in
Ultimate Strength: Fu := 65ksi
K factor K := 1
Factor of Safety phi ϕ := 0.9
Solution Method 1: Using Chapter E Equations:
1) Determining the governing slenderness ratio
15ft ⋅ 12in
K⋅ Lu
λx :=
ft
λx = 41.19
rx 4.37in
15ft ⋅ 12in
K⋅ Lu
λy :=
ft
λy = 70.313
ry 2.56in
(
r := min rx , ry ) r = 2.56⋅ in governing radius of gyration
The above was already obvious but it was carried on just to "academicallly" justify the numbers
2) Calculating Euler's Buckling Stress
2 2
π ⋅E 3.14 ⋅ 29000 ksi
FE := FE = 57.894⋅ ksi
2 2
K⋅ Lu 15ft ⋅ 12in
ft
r
2.56in
3) Determining if the buckling will be elastic or inelastic.
K⋅ Lu E
Buckling := if ≤ 4.71 , "Inelastic" , "Elastic" Buckling = "Inelastic"
r Fy
Alternatively we can also follow the process below:
Fy Fy
= 0.864 Buckling := if ≤ 2.25 , "Inelastic" , "Elastic" Buckling = "Inelastic"
FE
FE
4) Calculating the Buckling Stress (Fcr) and the load capacity of the section:
Fy
50ksi
FE
57.89ksi
Fcr := 0.658 ⋅ Fy 0.658 ⋅ 50ksi Fcr = 34.832⋅ ksi
2
ΦPn := ϕ⋅ Ag ⋅ Fcr 0.9⋅ 15.8in ⋅ 34.832ksi ΦPn = 495.314 ⋅ kip
Solution Method 2: Using Table 4-22:
1) Determining the governing slenderness ratio
16ft⋅ 12 in
K⋅ Lu
λy :=
ft
λy = 70.313
ry 2.56in
2) Using table we locate the KL/r value
corresponding to the Fy used for factrized critical
stress:
The value indicated would be between 31.1 and
31.4. Let's take 31.3
ΦFcr := 31.3ksi
Note: From our previous calculations:
Fcr = 34.832⋅ ksi
Therefore:
ϕ⋅ Fcr = 31.349⋅ ksi
3) Calculating the capacity of the element:
ΦPn := ϕ⋅ Ag ⋅ Fcr
2
0.9⋅ 15.8in ⋅ 34.832ksi
ΦPn = 495.314 ⋅ kip
Solution Method 3:
Using Table 4-1 for
W shapes pp 4-12 to
4-23:
Oh you will love this one! All you need
is the unbraced length and the shape: