Professional Documents
Culture Documents
Flange Lateral Bending Stress (F) Under The Wind Pressure: Atorod Azizinamini, Ph.D.,P.E
Flange Lateral Bending Stress (F) Under The Wind Pressure: Atorod Azizinamini, Ph.D.,P.E
Wind Pressure
By
Atorod Azizinamini, Ph.D.,P.E.
1- Introduction and Objectives
The purpose of preparing this document is to evaluate the application
of AASHTO LRFD Bridge Design Specification (Third edition) to
calculate flange lateral bending stress, fl, for a specific design example
and compare the results to detail finite element analysis. Specific
objectives are as follows:
a) calculate the flange lateral bending stress using two and three
dimensional finite element analyses
b) calculate
the
flange
lateral
bending
stress
using
code
recommendations
c) calculate the magnification factor using detail finite element
analysis approach and that recommended by the code
d) Incorporate flange lateral bending stress term in constructability
limit state check for a three span continuous bridge designed
using High performance steel
e) Provide preliminary conclusions with respect to advantage and
shortcomings of the procedures suggested by AASHTO LRFD
Bridge Design Specification to calculate flange lateral bending
stress
2- Brief Summary of the Bridge Configuration
Length
Span No.
(ft)
_________________________________________________________________
1
135
2
175
3
135
No. of Lanes
No. of Girders
Skew Angle
Dist. Curbline To Exterior Girder
=
=
=
=
3
4
0
3
a-
Elastic
Three
Dimensional
Finite
Element
Analysis-
Stress in x-direction
Figure 1 the deflection and stress of the bridge under the wind pressure
Figure 2 Flange stress under the wind pressure, a closer view- First Span
Sec
Sy
(in3 )
Sec.
Mod.
(1)
(2)
53
232
M (kips-in)
Stresses, (ksi)
2D Sap,
2D Sap,
2D Sap,
Constant
specified
Specified
properties
Properties
Properties
(3)
(4)
1092
(x=49.5ft)
2213(x=135ft)
832.5
2941
3D Ansys, One
3D Ansys,
Girder Model
Entire Bridge
(5)
(6)
(7)
15.7
12.1
9.4
(x=66.3 ft)
(x=43.5ft)
(x=45.1ft)
12.7
(x=135ft)
8.4 (x=135ft)
8.4
(x=135ft)
3
49
1232
(x=222.5 ft)
535
10.9
(x=222.5ft)
9.5
(x=222.5ft)
8.2
(x=222.5ft)
0.3 k/ft applied to one girder and resulting stress divided by number of girders
65 lb/ft2 which is equivalent of 0.3 k/ft applied to outside girder only. Analysis
takes care of number of girders
Figure 3 the moment diagram from the 2-D analysis with uniform section
properties along the girder
M
Sy
(1)
Where, Sy is the section modulus about the minor axis of one flange
and M is the wind induced moment per each girder (the moment
shown in Figure 3 is for one girder. It is assumed that total wind load is
resisted by all four girders, which is in accordance with results obtained
from three dimensional finite element analyses).
Table 1 shows summary of the results obtained from two dimensional
analyses in terms of moment and stresses at three locations along one
girder.
d- Calculating the Flange Lateral Bending Stresses using
AASHTO Recommendation
Section 4.6.2.7 of AASHTO LRFD in its commentary provides a
recommendation on how to calculate the resulting moments in the
girders due to wind pressure. The recommendation are as follows:
WL2b WL2
Mw
10
8N b
(2)
Using equation (2), the maximum moment for first span could be
calculated as follows:
Mw
187.3k f
10
8( 4)
(3)
Using equation (2) the maximum moment for middle span could be
calculated as follows:
Mw
303.1k f
10
8( 4)
(4)
flange
lateral
bending
stresses
using
the
AASHTO
LRFD
recommendation.
Table 2 Lateral flange bending stress under the wind pressure based on the
AASHTO recommendations
Se
Sy (in3 )
M (kips-in)
Section
AASHTO LRFD
Modules
Recommendation
53
2247.6
42.4
232
N.A.
N.A.
49
3637.4
74.2
c.
(ksi)
AASHTO LRFD Recommendation
4- Magnification Factor
It is assumed that the compression flange, when subjected to lateral
loads, acts as a beam column. The lateral load causes the flange to
displace in lateral direction. As a result the maximum moment in the
compression flange will increase due to secondary effects. AASHTO
LRFD treats this magnification in a similar way that AISC building code
treats design of beam columns. Appendix A provides derivation of the
moment magnification factor as used in AISC building code.
As indicated in appendix A the general from of the magnification factor
is as follows:
Cm
P
1
Pek
Where P is the applied axial load to the beam column and P ek is the
critical column buckling load. Cm is a factor that accounts for different
loading cases. Appendix A provides more detail discussion of this
magnification factor.
Lb 1.2 L p
C b Rb
f bm
Fyc
(5)
Where
LP rt
E
Fyc
(6)
29000
6.6 ft
50
(7)
Lb 23.1 ft 1.2(6.6)
1.0(1.0)
11.0 ft
25.8
50
(8)
Since actual unbraced length between the cross frames is 23.1 ft and
the limiting value is 11.0 ft, we must then magnify the fl term.
Substituting the numerical values, magnification factor becomes:
MF
0.85
2.29
25.8
1
41
(10)
is
2021
ft-kips.
The
resulting
compressive
stress
in
the
M 1.25( 2021)
3.0ksi
Sx
833
(11)
The limiting value for the unbraced length is calculated using Equation
(5).
Lb 23.1 ft 1.2(6.6)
1.0(1.0)
32.3 ft
3.0
50
(12)
Therefore, for this scenario, there is no need to magnify the flange lateral bending stress
due to wind loads.
c- Magnification factor using Nonlinear Geometric Analysis
Nonlinear finite element analyses were carried out to account for
second order effect directly. ANSYS5.7 was used. Complete three
dimensional model of the bridge was used in the analysis.
Several scenarios were simulated in the nonlinear geometrical finite
element analysis using the full three dimensional model of the bridge.
A) scenario where the dead weight consisted of the weight of the
girders only, b) scenario where dead weight consisted of weight of the
girders plus the weight of the wet concrete before it is hardened and c)
scenario where dead weight consisted of weight of the girder plus
weight of the concrete after concrete is hardened.
For the three cases described above, Table 3 gives the magnification
factors to be used in conjunction with flange lateral bending stresses at
mid-span of the middle span (section 3 in table 2). The magnification
factors reported in Table 3 is simply the ratios between flange lateral
stresses obtained from nonlinear finite element analysis divided by the
corresponding value from linear finite element analysis.
fbm, ksi
2D
Magnification Factor
AASHTO
Analysis
Self weight of Girders
Ansys, 3D
nonlinear
3.0
1.0
1.01
25.8
2.29
1.31
1.0
1.03
only
Wet concrete & Girder
weight
Composite girder
Linear Analysis
Nonlinear Analysis
Figure 4 The longitudinal stress contours in x-direction under wind and steel girder weight
Linear Analysis
Nonlinear Analysis
Figure 5 The longitudinal stress contours in x-direction under wind, steel girder weight and slab
weight
Linear Analysis
Nonlinear Analysis
Figure 6 The longitudinal stress contours in x-direction under wind, steel girder weight and slab
weight in composite condition
Magnification Factor
AASHTO
Ansys, 3D
nonlinear
1.0
1.06
2.29
1.24
1.0
1.03
only
Wet concrete & Girder
weight
Composite girder
5- Summary
Previous sections provided different approaches for calculating the
flange lateral bending stresses. Methods used included
a) Three dimensional model of the entire bridge
b) Three dimensional model of one girder
c) Two dimensional model of girder using uniform sectional
properties
d) Two dimensional model of the girder using specified section
properties
e) AASHTO LRFD code recommendations
Table 5 provides summary of the flange lateral bending stresses for
mid-span of the middle span
Method Used
Flange
Lateral
Bending
Stresses, ksi
3-D Entire bridge
8.2
9.5
25.1
10.9
AASHTO Recommendations
74.2
c) magnification
factor
calculated
using
weight
of
hardened
Dead Load
fbm, ksi
Magnification Factor
Considered
Self weight of
2D
AASHTO
Nonlinear
Analysis
Requirement
Analysis
3.0
1.0
1.01
25.8
2.29
1.31
1.0
1.03
Girders only
Wet concrete &
Girder weight
fbu, ksi
2D Analysis
3.0
only
Wet concrete & Girder
25.8
weight
f bu
(13)
1
f l f Fnc
3
a) Calculations
provisions
(14)
based
on
values
obtained
from
code
The flange lateral bending stress is fl is 74.2 ksi which exceeds the
Fy and therefore code requirement is violated regardless of which
dead loads are considered in the calculations
1
f l f Fnc
3
1
f l f Fnc
3
Appendix A
Derivation of the Moment Magnification
factor Used in AISC Building Code
BEAM-COLUMNS
A beam-column is a member that is subjected to both axial force and
bending moment.
To begin our discussion, let us consider a simply supported beam-column
subjected to an axial force P and uniformly distributed lateral load.
w (intensity)
P
EI = constant
y
w
P
M
o
wl
2
V
x
M0 0
M
wx 2
wl
Py
x0
2
2
M Py
wx 2 wl
x
2
2
But,
M EIy ''
EIy '' Py
EIy
Letting
k2
wx 2 wl
x
2
2
"
wx 2
Py
2
P
EI
y" k
w
2 EI
y yh y p
y h A sin kx x cos K x x
wl
x
2
w
wl
w
x2
x
2
2
2 Elk
2 EIk
Elk 4
w
wl
w
x2
x
2
2
2 EIk
2 EIk
EIk 4
w
EIk 4
wl
2 EI
w
hl
l
y'( ) 0 A
tan
4
2
2
EIk
kl
2
wl 4
16 EIu 4
2ux
2ux
wl 2
tan
u
sin
cos
x (l x )
l
l
8 EIu 2
from which the moment distribution along the length of the member is:
wl 2
2ux
2ux
tan u sin
cos
1
2
1
1
4u
M EIy 2
2
wl 4 1 cos u
wl 4
16 EIu 4 cos u
32 EIu 2
y
5wl 4
384 EI
12( 2 sec u u 2 2)
5u 4
12(2 sec u u 2 2)
5u 4
= yo
where
yo
5wl 4
384 EI
2 EI
). , the value of A.F. approaches infinity. In other words, as
l2
sec u 1
Noting that:
u
kl l
2 2
Pe
ymax =
EI
2
P
Pe
2 EI
l 2
P
P
y o 1 1.003( ) 1.004( ) 2 ...
Pe
Pe
Or approximately
ymax
P P
yo 1
Pe Pe
...
Therefore:
y max y 0
1 P
Pe
M max
wl 2 2(sec u 1)
8
u2
2(sec u 1)
u2
M max M 0
M0 is the moment, that would exist if only lateral load had been
applied. Therefore, the term in brackets represents amplification
factor due to axial load.
Again using the power series, expansion for sec u, ,the expression for
Mmax could be simplified as follows:
kl
u
2
Mmax= M0
or
P l
EI 2
P
Pe
Pe
2 EI
l 2
P
P
1.031
1 1.028
Pe
Pe
EI
1.032
Pe
Pe l 2
...
M max
P
P
P
1 1.003
1.004
M 0 1 1.028
Pe
Pe
Pe
...
or approximately,
P
P P
1
1 1.028
Pe
Pe Pe
Mmax =Mo
M max
P
1
M 0 1 1.028
Pe 1 P
Pe
...
P
1 0.028 P
M0
P
1
Pe
or
M max M 0
1 P
Pe
Q
a
P
EI = constant
L
y
P
M
P
Q L a
L
x
Q (l a )
x Py
l
EIy "
l x
EIy " Qa
For the case of
tan u
u
M max M o
Where
existed)
Mo
Ql
4
l
it
2
for 0 x a
Py
for a x l
and
kl
2
To simplify the expression for maximum moment, we will use the power
series expansion for tan u.
1
2
17 7
tan u u u 3 u 5
u ....
3
15
315
1
2
17 6
M max M o 1 u 2 u 4
u ...
3
15
315
kl
2
2
P
Pe
P
P
0.812
M max M o 1 0.823
P
e
Pe
0.811
Pe
P
P
P
1 0.987
0.985
M 0 1 0.823
Pe
Pe
Pe
P
P P
1
M o 1 0.823
Pe
Pe Pe
P
1 0.18 P
P
1 0.2 P
P
1
Pe
M o
P
P
1 P 0.823 P
e
e
Mo
P
Pe
M max M o
M max
...
P
1
M o 1 0.823
P
e 1 P
Pe
P
1
Pe
...
...
M max
M0
Where
P
Pe
P
1
Pe
M0
QL
and
4
=0.2
So in summary:
M max
P
l
P
Pek
MB
Cm 1
where 0.2
MA
Cm 1
Cm
P
1
Pek
MA
0.4
Cm 0.6 0.4
MB
(Simplified Form)
MA
P
P
EI = Constant
MB
MA
M
MA MB
l
x
Assume M B M A
The D.E. for this case is as follows:
EIy " Py
M
l
M A cos kl M B
2
EIk sin kl
MA
M M
M
cos kx A 2 B A2
2
EIk
lEIk
EIk
2
k
EI
x M
y"
M EIy
".
The moment is the maximum where shear is zero, therefore to calculate the
location where the moment is maximum, we could equate:
dm
V EIy"' to zero.
dx
For this problem, it can be shown that:
M A / M B 2 2( M A ) cos kl 1
MB
sin 2 kl
M max M B
M max M B
M A / M B 2 2 M A / M B cos kl 1
sin 2 kl
21 cos kl
sin 2 kl
M max could
be simply
Primary
Moment
M
+
Secondary
Moment
Total
Moment
Mmax
l
When a beam is subjected to axial load and unequal end moment, calculation
of M max could be simplified using the concept of equivalent moment.
The figure below shows schematically the concept of equivalent moment.
MA
MB
P
Meq
Meq
P
MB>MA
MA
Meq
MB
Mmax
Meq
Mmax
could be found by equating the expressions for M max for two cases
derived:
M eq
M A / M B 2 2 M A / M B cos kl 1
MB
sin 2 kl
M eq
21 cos kl
sin 2 kl
or
M eq
M A / M B 2 2 M A / M B cos kl 1
21 cos kl
MB )
or
M eq C m M B
where C m
MA / M B 2 2 M A / M B cos kl 1
21 Coskl
21 cos kl
sin 2 kl
M max C m M B
21 cos kl
sin 2 kl
(Assuming
kl
2
and
kl
2 kl
cos
2
2
sin 2 kl 4 sin 2
kl
M max C m M B sec
2
kl
u
2
Letting
sec u 1
1 2 5 4 61 6
u
u
u ...
2
24
720
kl
2
2
P
Pe
P
P
sec u 1 1.23
1.268
Pe
Pe
P
P P
1
1 1.23
Pe
Pe Pe
...
1.27
Pe
...
MB MA)
P
1
1 1.23
P
Pe
1
Pe
P
P
1.23
Pe
Pe
P
1
Pe
1 0.23
P
Pe
P
Pe
P
Pe
Therefore:
M max
Note that
Cm M B
P
1
Pe
Cm
P
1
Pe
is an amplification factor.
In all the cases considered, beams were simply supported. In the event that
end conditions are different than simply supported it could be shown that
Pe
2 EI
2 EI
Pe k
would
be
replaced
by:
where k is the effective length
kl 2
l2
M max
Cm
MB
P
.
1
Pe k