Lateral Buckling Analysis of A Steel Pony Truss
Lateral Buckling Analysis of A Steel Pony Truss
Lateral Buckling Analysis of A Steel Pony Truss
by
Derek Matthies
Committee Members:
Fouad Fanous - Major Professor
Robert Abendroth - Committee Member
Vernon Schaefer - Committee Member
ii
Contents
List of Symbols and Abbreviations................................................................................................ iv
List of Figures ................................................................................................................................. v
1.
2.
Introduction ...................................................................................................................... 1
1.2
Objectives ......................................................................................................................... 2
Background ............................................................................................................................. 3
2.1
Buckling Behavior............................................................................................................ 3
2.2
Euler Buckling.................................................................................................................. 4
2.3
2.4
2.4.1
2.4.2
2.4.3
2.4.4
2.4.5
2.4.6
2.5
3.
4.
3.2
3.3
Discussion and Results of the Analysis of a Pony Truss Top Chord .................................... 23
4.1
4.2
4.2.1
4.2.2
4.2.3
4.2.4
4.2.5
4.2.6
4.3
iii
5.
4.3.1
4.3.2
4.4
4.5
Summary ........................................................................................................................ 38
5.2
Conclusions .................................................................................................................... 39
5.3
Recommendations .......................................................................................................... 39
Appendix A ................................................................................................................................... 40
Appendix B ................................................................................................................................... 43
Appendix C ................................................................................................................................... 45
Appendix D ................................................................................................................................... 47
References ..................................................................................................................................... 50
iv
List of Figures
Figure 1.1
Figure 2.1
Figure 2.2
Figure 2.3
Euler Buckling.........5
Figure 2.4
Figure 2.5
Figure 2.6
Figure 2.7
Figure 2.8
Figure 3.1
Figure 3.2
Figure 3.3
Stress-Strain Curve........22
Figure 4.1
Figure 4.2
Figure 4.3
Figure 4.4
Figure 4.5
Figure 4.6
Figure 4.7
Figure 4.8
Figure 4.9
Figure 4.10
Figure 4.11
Figure 4.12
n Panels
CL
To analyze the compression chord of a pony truss, the chord can be treated as a bar on
elastic supports (Ballio, 1983). This member with intermediate elastic restraints will buckle in
half-waves depending on the stiffness of the elastic restraints. The buckled shape of the bar will
fall somewhere between the extreme limits of a half-wave length of unity and the number of
spans between the end restraints. From the buckled shape, the effective length of the
compression chord can be used to determine the critical load. The method on how to determine
the effective length has long been the focus of compression chord buckling.
The failure of several pony truss bridges at the end of the nineteenth century prompted the
research of compression chord buckling. Engesser (sited in Galambos, 1988) was one of the first
researchers to investigate the problem and develop an approximate formula to determine the
required stiffness for the elastic restraints that corresponds to a specified effective wave length,
k. Engessers approach for determining the stiffness of the elastic restraints and its effects on
the compression chord was based off the assumption that the connection between the web
members and the floor beam is rigid. This theory used the frame consisting of the floor beam
and vertical and diagonal members at each panel point location to provide stiffness for the
compression chord. However, the theory in question is if the idealized structure is a conservative
approach of the actual frame stiffness. In other words, one may argue that investigating the
behavior of the bridge as a three dimensional system may result in a higher stiffness coefficient
of these lateral supports.
To the writers knowledge, all of the research for determining the critical buckling load on a
compression chord with elastic supports is based on Engessers assumption. From this
assumption, others, such as Timoshenko (1936), Bleich (1952) and Holt (1952), provided
methods of solving for the effective buckling length factor, k.
1.2 Objectives
The objective of the work presented herein was to verify the results of the published
solutions for determining the effective length factor using the finite element method. These
objectives were accomplished by performing the following tasks:
1. Conduct a literature search to review available information that is related to the
stability of the top chords in truss structures.
2. Verify the results of analyzing a top chord of a pony truss using the approaches given
in published literature and the results obtained using the finite element method.
3. Recommend the most applicable published analysis technique for determining the
critical load of an unbraced top chord of a truss system.
2. Background
2.1 Buckling Behavior
The failure of an axially loaded bar in compression is defined by limit states which are an
identifying condition of design criteria. Limit states for a structural member include strength
limit states, which may result in yielding or rupture, or serviceability limits states (i.e. deflection,
vibration, slenderness or clearance). Although not a limit state, buckling presents a failure mode
due to high compressive stresses which causes the member to no longer be in equilibrium.
Usually buckling occurs before the column reaches the full material strength. The buckling
strength of compression members has long been studied to relate the empirical methods of
analysis to the actual results. The elastic buckling of an axially loaded column in compression
occurs when a certain critical load is reached causing the member to suddenly bow out. The
deviation of the member axis will result in additional bending that gives rise to large
deformations, which in turn cause the member to collapse. The load at which collapse occurs is
referred to as the buckling load and is thus a design criterion for compression members.
In linear mechanics of deformable bodies, displacements are proportional to the applied
loads. The essence of buckling, however, is a disproportionate increase in displacement resulting
from a small increase in load. For example, Fig. 2.1, from Brush (1975), shows the loaddisplacement relation (referred to equilibrium path) for an axially loaded column. Each point of
this path represents an equilibrium configuration of the structure. However, as the applied load
reaches a critical value, i.e., Euler Load, the equilibrium path will follow the secondary path
shown in Fig. 2.1. Points along the primary (vertical) equilibrium path represent the
configuration of a compressed, perfectly straight column, but as the critical load is reached, a
secondary path is formed representing the bent equilibrium configurations. The critical load is
defined as the minimum load for which the structure remains in equilibrium before instability is
reached and failure occurs. Of course, no real column can be perfectly straight, and hence the
load displacement relation will not follow that shown in the Fig. 2.1 but rather a different load
displacement will be obtained. The load displacement relationship of an imperfect axially loaded
column is shown in Fig. 2.2. When comparing Fig 2.1 and 2.2 for the straight and crooked
columns, the figures show that the equilibrium paths generally converge as the lateral
displacements increase. Analyses for both columns, straight and slightly crooked, lead to large
lateral displacements at the critical load.
P
C1
Secondary
path
Primary
path
-
at x = L/2
at x = L/2
(2.1)
x
y
L/2
L/2
P
Fig. 2.3 Euler Buckling (Thandavamoorthy 2005)
The solution for the linear differential equation above can be written as
= sin + cos
(2.3)
and
0 = sin
(2.4)
0 = sin =
(2.5)
=
(2.6)
By substituting K =
/ into equation 2.5 and solving for P, Euler buckling equation yields
! " #$
%
where m = 0, 1, 2 is referred to as the number of buckling modes. The deformed shapes for
the first three buckling modes are shown in Fig. 2.4.
Pcr
4Pcr
m=1
9Pcr
m=2
m=3
(a)
(c) P
(b)
The analysis proposed by Engesser in the late 1800s can be applied with some reasonable
accuracy to analyze a bar that is pinned at its ends and is supported on equally spaced
intermediate elastic springs provided that the half-wavelength of the buckled shape is at least
1.8 times the spring spacing (See Galambos 1988). However, one must realize that Engessers
solution can only be used as a preliminary design tool and more comprehensive analysis is
needed.
Engesser examined the top-chord buckling problem of pony trusses and summarized his
findings in a paper that was published in 1884. In the following years, he used his work to
explain the failures of pony truss bridges and provide a rational method of design for similar
structures. Engesser developed a simple formula to calculate the required stiffness, Creq, of the
elastic support to reach the desired critical load that is based on a specific buckling length. In
his work, Engesser suggested that one needs to assume an effective length factor, k, of 1.3. The
top chord, including the end posts, is straight and of uniform cross section. Engesser also
provided the following assumptions:
1. Its ends are taken as pin-connected and rigidly supported.
2. The equally spaced elastic supports have the same stiffness and can be replaced by a
continuous elastic medium.
3. The axial compressive force is constant through the chord length.
Engessers solution for the required stiffness of a pony-truss transverse frame which is derived
in Appendix C is
&
'(
)*+ ,
-#$
(2.7)
If Creq is met at each frame location, the chord with the length between panels, l, will
achieve the specified design load, Pcr. Several researchers suggest that one needs to assume a
factor of safety, v, of two when calculating the design load. In other words, the load Pcr, can be
taken as vP, where P is the calculated top chord member load. In addition, once the calculations
show that the stress induced in the member exceeds the limit specified in the design
specifications, the flexural rigidity EI should be modified using the tangent modulus, EtI. By
combining Eulers buckling equation and Eq. 2.7, the required spring constant is
&
'(
= -.*+
,
")
(2.8)
4. Utilize the information calculated above to estimate the ratio, Cl/P. One approach that
can be used to calculate the provided lateral supports stiffness, C, is detailed in
Appendix A.
5. Use the above calculated ratio and the number of panels, n, to calculate the effective
length factor, k, using Table 2.2 from Holt (1956) or Fig. 4.1 citing other authors.
6. Apply k, found in step five, to the equations found in Chapter E of the AISC (2011)
manual to determine the nominal compression capacity of the compression chord.
Similar to the approach for a simply supported column, a column simply supported at both
ends with equally spaced interior elastic supports, as shown in Fig. 2.6, can be defined by an
equation which represents the buckled shape. The total length, L, is the number of bays, n,
multiplied by each bay length, l. Using the energy method, the deflected shape can be defined
by the equation
= / sin
!"
%
(2.9)
where m is the mode number, and the number of modes can be related to n-1. Using the sin
curve, the boundary conditions are y = 0 at x = 0 and y = 0 at x = L.
(a)
L = nl
(b)
To solve for the external work and internal energy of the member, the first and second differential
equations for the line can be solved as
1!"
%
cos
!"
%
(2.10)
10
1! "
%
sin
!"
%
(2.11)
With P as the axial load, the potential energy of the system can be set equal to the external
work, V, plus the internal energy, U to find the critical buckling load. A full derivation of a bar
on two elastic springs using the energy method is presented in Appendix B. Since the energy
method uses an assumed buckled shape for the chord, the solution is obtained by some degree
of approximation. If the assumed shape is properly chosen to satisfy the boundary conditions,
the energy method provides a satisfactory approximation. For the case presented in Fig. 2.6, the
potential energy can be represented as
2 = 3 + 4 =
5)
68
7 +
68
+
&
%
#$
(2.12)
The final term in equation 2.12 represents the energy of the elastic supports as a function of the
spring constant C. By solving for P, the critical buckling can be found as a function of the
mode number.
Timoshenko (1936) extended the work of Engessers to include the effects of a varying
the axial load along the top chord of a truss structure. He assumed the compression load varies
parabolically along the length of the chord with the load equal to zero at the ends then reaching
a maximum value at the center (see Fig. 2.7). In addition, Timoshenkos solution assumed that
the ends were pin connections.
L
P
11
<
4 =
" = #$
(2.13)
-
?@A
?@9 > / +
B,
-
?@A
?@9 /
(2.14)
In his solution, Timoshenko also stated that if the bridge is uniformly loaded, the compressive
forces that are transmitted to the chord by the diagonals are proportional to the distance from the
middle of the bridge span as
C = C8 D1 , G
(2.15)
where x is the distance from the left support in the figure and qo is the maximum force of the
axial load represented by
C8 =
H
)
(2.16)
The external work, V, can then be calculated using the information given above as follows
3 =
,I 68 JK L D G 7
,
(2.17)
Substituting the information given in Eqs. 2.14 and 2.17 into the total potential energy
relationship, one can then obtain the following:
D
(M ,
-
" #$
%
(2.18)
Finally, since the elastic supports are treated as a continuous elastic medium, Eq. 2.19 can be
developed to relate the critical buckling load to the effective length factor by combining Eqs.
2.8, 2.16 and 2.18.
<,
)*
" % 9
= N H D.G
(2.19)
12
2.4.4 Buckling of a Pony Truss Top Chord with Elastic Ends
2.4.4.1 Analysis According to Bleich (1952)
Bleich obtained his solution by using finite difference as an exact approach to quantify
the buckling load of the chord. His solution was based on the ends being pin connections and
equally spaced intermediate supports of equal rigidity. Bleich also assumed the chord had a
constant moment of inertia and constant axial compressive force over the entire length. Later,
Schweda extended Bleichs results to include chords with elastic ends.
vPd Ce
vPd -n
C
vP
l
C
-r
Ce vPd
+r
+n
2nl
The theoretical exact solution proposed by Bleich for a chord supported on rigid ends
was
where vP = Pc and
&=
ST
O) D95 PQR U G1 5 V
,
95
W
ST
XY Z[\ ]U
/ = ^5 R_` ^ />7 a =
^
(2.20)
^J95PQR^L
^5R_`^
(2.21)
To eliminate the tangent modulus, Et, which varies per the axial load, Eulers buckling
equation is substituted into equation 2.21, simplifying the stiffness to
=.
"
(2.22)
For a chord with n spans in equation 2.20, there are n -1 different half-wave buckling
configurations. As n increases, the spring constant required for an infinite number of spans,
C, increases to a limiting value. Bleich (1952) showed that Cn can be replaced by C for any
13
span where n is greater than six which corresponds to an error less than 1%. So, for trusses
with more than six spans, equation 2.16 simplifies to
&
'(
)*
^J95PQR^L
)*
,
(2.23)
where is given in Table 2.1 from Bleich (1952) and is valid for the elastic and plastic range
of buckling.
Table 2.1 Values of in Eq. 2.23 (Bleich, 1952)
1/k
1/k
1/k
1/k
0.3
0.111
0.5
0.309
0.70
0.614
0.90
1.102
0.32
0.126
0.52
0.335
0.72
0.652
0.91
1.138
0.34
0.142
0.54
0.361
0.74
0.692
0.92
1.177
0.36
0.160
0.56
0.388
0.76
0.734
0.93
1.219
0.38
0.179
0.58
0.417
0.78
0.777
0.94
1.264
0.40
0.198
0.60
0.447
0.80
0.822
0.95
1.316
0.42
0.218
0.62
0.478
0.82
0.870
0.96
1.375
0.44
0.239
0.64
0.510
0.84
0.921
0.97
1.444
0.46
0.261
0.66
0.544
0.86
0.976
0.98
1.530
0.48
0.285
0.68
0.578
0.88
1.036
0.99
1.652
1.00
2.000
Since Bleichs theory assumes a constant axial force, which is rarely the case in practice,
the center bay of the chord should be designed with the appropriate k value for the maximum
load and then used for the remaining bays. Although the previous assumption of designing
the bridge for the center span only would yield conservative results, the assumption of rigid
ends can result in unsafe buckling loads when using Bleichs theory. Thus, in order to
continue on Bleichs exact buckling theory, Schweda provides results to determine the
required stiffness for a chord supported elastically on the ends.
The spring constant of the end supports is denoted by Ce and the intermediate supports
by C. Schweda assumed the load was a constant axial force throughout the length of the
14
chord with equally spaced elastic supports similar to Bleichs theory. Figure 2.8 shows the
compression chord with the diagonals extended a length, d, subjected to the compressive force
vPd and pinned at points n and +n. The spring constant must be larger the spring constant of
the chord with rigid end, C0. Thus, the value of C = cC0 where c > 1.1. The required spring
constant with the axial load in the diagonal is
&' =
O)i
+ j&#
(2.24)
where CE is Engessers equation (Eq. 2.8). Schweda calculated as a function of the number
of bays with respect to c and k. These values are listed in Appendix D. If all of the crossframes are identical, Ce in equation 2.24 is equal to C and the equation becomes
i
& = J9 5 kL
O)
(2.25)
Holts research in the 1950s tested pony truss bridges in an attempt to compare the actual
buckling load of the compression chord with the design equations. His research not only tested
the primary constraints mentioned above but also the effects of secondary factors. The
following secondary factors were considered in his research (Holt, 1956):
1. Torsional stiffness of the chord and web members.
2. Lateral support given to the chord by the diagonals.
3. Effect of web-member axial stresses on the restraint provided by them.
4. Effect of non-parallel-chord trusses.
5. Error introduced by considering the chord and end post to be a single straight
member.
The results of Holts analysis proved that the error in determining the critically buckling load by
neglecting the above factors was relatively small. His conclusion stated that the load capacity
of a pony truss bridge would be satisfactorily predicted by previous buckling analyses
mentioned. As shown by Bleich (1952), if the truss has at least ten panels, then the effective
length factor depends only on the stiffness of the transverse frames. Thus, the appropriate
15
effective length factor is a function of Cl/Pc which is shown in the results section. Holt noted
that Bleichs analysis showed adequate results for the entire range of effective length values
where Timoshenkos results show adequate results for k > 2. A summary of Holts results can
be seen in Table 2.2. Based on the results of his research, Holt (1957) also recommended the
following on the design of the end posts
The end post should be designed as a cantilever to carry, in addition to its axial load, a
transverse force of 0.3% of its axial load at its upper end.
1/k
1.00
0.980
0.960
0.950
0.940
0.920
0.900
0.850
0.800
0.750
0.700
0.650
0.600
0.550
0.500
0.450
0.400
0.350
0.300
0.293
0.259
0.250
0.200
0.180
0.150
0.139
0.114
0.100
0.097
0.085
4
3.686
3.352
2.961
2.448
2.035
1.750
1.232
0.121
0
14
3.785
2.771
2.454
16
3.809
2.774
2.479
2.254
2.101
1.968
1.681
1.456
1.273
1.111
0.988
0.878
0.768
0.668
0.537
0.428
0.292
0.183
2.282
2.121
1.981
1.694
1.465
1.262
1.088
0.940
0.808
0.708
0.600
0.500
0.383
0.280
0.187
0
0.135
0.045
0
0.107
0.068
0.103
0.055
0.121
0.053
0.112
0.070
0.017
0
0.031
0.029
0.025
0.003
0
0.010
16
2.4.6 Buckling Load with Initial Out-of-plane Deformations
Initial out-of-plane deformations of the compression chord can reduce the critical buckling
load determined by the previously mentioned methods. There are two primary causes of out-ofplane deformations that need to be taken into consideration. A vehicle load on the floor beams
would cause a displacement of the chord at the location of the load creating initial lateral
displacements in the chord. The chord could also have initial crookedness and unintentional
eccentricities due to manufacturing. Such lateral deflections would reduce the maximum load
capacity of the chord. Lutz and Fisher (1985) addressed this issue in their publication to the
Structural Stability Research Council in 1985. Their work was similar to the stiffness criteria
George Winter proposed in 1960.
Winter (1960) proposed the ideal stiffness, C, needed to fully brace the compression
member over the length, l, is equal to
ideal stiffness
l)*+
,
-)*+
,
compression chord and developed a factor of safety to account for the out-of-plane stiffness.
The stiffness equation by Engesser was
&
'(
= 2.5 %*+
) ,
p
(2.26)
where Le = kl and 2/4 2.5. For trusses with a small l relative to Le equation 2.26 provides an
accurate bracing stiffness for the solution. However, as l increases relative to Le, then equation
2.26 will result in unsafe errors. Thus, Lutz and Fisher proposed the following empirical
equation to determine the required stiffness
- ) ,
(2.27)
Then they extended the applicability of k factors to less than 1.3 going as low as 1.0. The fully
braced case where k = 1 corresponds to Winters stiffness of
-)*+
,
minimum value of stiffness for the compression chord to reach the required critical load
however; there are currently no design procedures available to account for initial imperfections.
Design recommendations by AASHTO only make a note of the design vertical truss members
and the connection to the floor beam.
17
18
19
Fig. 3.1 Finite Element Idealization of the Top Chord as a Bar on Elastic Supports
20
21
The nonlinear solution of the compression chord in ANSYS used the Newton-Raphson
option to converge on the displacements of the solution. An initial load larger than the predicted
buckling load was applied to the chord and ANSYS then uses load steps to continuously apply
the load in small increments to iterate the solution. For each iteration, the program calculates a
new element stiffness matrix based on the element strains in the stress-strain profile provided.
For the nonlinear model in this research, it was assumed that a converged solution was reached
when the difference in displacements between load steps was equal to or less than 0.1%. The
results of analyzing the truss described above using the different available analyses techniques
are summarized in the following chapter.
22
Stress-Strain Curve
40
35
Stress (ksi)
30
25
Et = 7344 ksi
20
15
10
5
E = 29000 ksi
0
0
0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.01
Strain (in/in)
23
1.000
0.800
1/k
Engesser
0.600
Bleich
Timoshenko
0.400
Lutz-Fisher
0.200
0.000
0.000
0.500
1.000
1.500
2.000
2.500
3.000
3.500
4.000
Cl/P
4.500
24
100
Pcr/Pe
80
60
40
20
0
0
500
1000
1500
2000
2500
3000
(CL/Pe)
w
z
f s
xyZ
yz
v r
{888
XM
xM
f
xX~M M
9 8 |
= 6.75 kip/in
25
= 1.50
.98
8
Using the calculated ratio Cl/P in conjunction with Fig 4.1, one can calculate the value of 1/k
according to the different analysis techniques that were previously summarized.
4.2.1 Calculations following Engessers Procedure
1/k = 0.78 (From Fig. 4.1)
k = 1.28
= 50.3
9.
l98
-.8l
Following the design steps that are given in the AISC (2011) manual, one can calculate
the critical buckling stress, Fe, and the nominal buckling load, Pn, using the ASD and
LRFD approaches, respectively as
F =
u
D G
FP = 0.658
{888
J8.NL
]
F
= 113.2 ksi
= D0.658
N]
99N.
G 36
= 31.5 ksi
P` = FP A = 31.5 25.0 =
This procedure is then repeated using the other methods mentioned in chapter two.
4.2.2 Calculations following Bleichs Procedure
1/k = 0.768 (From Fig. 4.1)
k = 1.30
9.N898
-.8l
F =
u
D G
= 51.1
FP = 0.658
{888
J9.9L
]
F
= 109.8 ksi
= D0.658
N]
98{.l G 36
= 31.4 ksi
26
P` = FP A = 31.4 25.0 =
9.98
-.8l
F =
u
D G
= 64.9
FP = 0.658
{888
J-.{L
]
F
= 67.9 ksi
= D0.658
N]
.{ G 36
= 28.8 ksi
P` = FP A = 28.8 25.0 =
9.N{98
-.8l
F =
u
D G
= 54.5
FP = 0.658
{888
J-.L
]
F
= 96.2 ksi
= D0.658
N]
{.
G 36
= 30.8 ksi
P` = FP A = 30.8 25.0 =
9.NN98
-.8l
F =
u
D G
= 52.3
FP = 0.658
{888
J
.NL
]
F
= 104.7 ksi
= D0.658
N]
98-. G 36
= 31.2 ksi
P` = FP A = 31.2 25.0 =
27
4.2.6 Calculations using the Energy Method
P = JL =
u
N---9l.N
J988L
.988
99.l-
= 912.2
= 11.84 kips
PP = P R = 11.84 61.67 =
A summary of these results is shown in Table 4.1. As noted in Galambos (1988) the factor of
safety for the compression chord on elastic supports was 2.0 when determining the Cl/P ratio.
However, it was later addressed that the compression chord in Galambos example was designed
for the maximum panel load using AASHTOs 1983 formula which uses a factor of safety of
2.12 (Ziemian, 2010). In essence, the results should be compared to the design load of
P=360*2.12 = 763 kips. The factor of safety listed in this table was calculated as the ratio
between the estimated critical buckling load and the applied member load of 360 kips.
k
1.28
1.30
1.66
1.39
1.33
-
Pn
788
785
721
770
779
730
F.S
2.19
2.18
2.00
2.14
2.16
2.03
The results above show the methods of analysis in chapter two reasonably predict the critical
load of the compression chord for the pony truss example. However, it may be noted that using
the energy method underestimates the critical buckling load since it is based on an assumed
deformed buckling shape.
28
To verify the critical load calculations of the compression chord, the truss was analyzed in
ANSYS to solve for the critical buckling load. As mentioned in chapter three, the validity of
these tests was checked with both a 2-D and 3-D element model. Using a 2-D beam3 element
with the tangent modulus, Et = 7344 ksi, the critical buckling load was calculated to be 719.213
kips as shown in Fig. 4.3. The 3-D element analysis of the compression yielded the exact same
buckling load of 719.213 kips (Fig. 4.4) validating the model.
The compression chord buckling load was also determined using a nonlinear approach. This
buckling load was found by plotting the Load vs. Displacement in the vertical direction and then
noting the load at which large displacements occur with only a small increase in load. Using Fig.
4.5 the buckling load in the nonlinear model was equal to 720 kips. As seen in the graph, there
are two changes in slope which account for the change in modulus at the proportion limit and the
yielding limit. All of the FEM solutions prove the reasonability of the calculations determined in
the previous section with C = 6.75 and Et = 7344 since the three compression chords analysis all
had approximately the same buckling load of 720 kips. The next step of the analysis was to
check the critical load on the compression chord when the entire truss was modeled.
29
30
The pony truss model was analyzed similar to the compression chords in that a
compression load was applied to the top chord on each side of the truss. In order to only focus
on the lateral displacements of the compression chord, the truss was restrained in a manner to
create two symmetrical sides of the truss. To accomplish this symmetrical model, the truss was
restrained against translation in the z direction and rotations in the x and y directions at the center
of the floor beams. By applying these boundary conditions, the resultant load could be compared
to the compression chord models. The critical buckling load, when modeling the entire pony
truss, increased dramatically to a load of 1121 kips on the compression chord which can be seen
in Fig. 4.6.
It is the opinion of this author that the increase in critical buckling load is due to the effect
of inertia from the compression chord. The spring stiffness, when analyzing the compression
31
chord by previous work, was composed only of the frame stiffness as calculated in Appendix A.
However, the lateral stiffness of the chord itself provides an addition stiffness which explains the
significant increase in load when modeling the entire truss. The goal in the preceding paragraphs
is to analyze this theory using finite element models.
32
33
34
1
2
3
4
5
6
7
8
9
10
11
Displacement, (in)
0.10612
0.07518
0.07249
0.07227
0.07226
0.07226
0.07226
0.07227
0.07249
0.07518
0.10612
Stiffness, C (k/in)
9.426
13.301
13.795
13.837
13.839
13.839
13.839
13.837
13.795
13.301
9.423
35
Displacement, (in)
0.048816
0.052882
0.058840
0.069192
0.100502
0.141938
Stiffness, C (k/in)
13.89
12.98
11.87
10.45
8.08
7.04
6.75 (using Eq. A1.1)
36
compression chord when analyzing the entire truss since the formulas do not account for the
inertia stiffness of the compression chord and the increase in stiffness with respect to k is not a
linear response.
Table 4.4 Critical Load Results with New Stiffness, C
Method
Engesser
Bleich
Timoshenko
Lutz & Fisher
Holt
Energy Method
K
NA
1.02
1.16
1.08
1.01
-
Pn
NA
827
807
819
828
1032
F.S
NA
2.30
2.24
2.28
2.30
2.87
Using ANSYS, the compression chord modeled with the new stiffness provided a large
increase in the compression capacity. Figure 4.11 shows the compression chord modeled with
rigid supports and the interior restraints having a stiffness of 13.8. The buckling load was found
to be 979.03 kips. Although this load is closer to the pony truss analysis of 1121 kips, the end
supports in this model were rigid which does not accurately account for the elasticity at these
supports. The next figure, Fig. 4.12, shows the buckling shape with the elastic supports having
the new stiffness values determined in Table 4.2. This buckled shape essentially shows the
exact buckling load and shape of the compression chord treated as a single member.
37
38
39
5.2 Conclusions
The following are the conclusions that can be attained from the study presented herein:
Current design of a compression chord for a pony truss bridge would be best
accomplished by using the effective length factor, k, provided by Holt to determine
the critical load from Chapter E in the AISC manual. However, such an effective
length factor is dependent on the stiffness of the elastic lateral supports.
The energy method provides a close answer for calculating the critical buckling load
for the truss top chords with rigid supports. However, its application depends on the
number of the provided elastic supports.
The finite element method will provide satisfactory results when using the
appropriate member rigidity, i.e., using the correct tangent modulus of elasticity.
All available procedures require the knowledge of the provided elastic stiffness.
Hence, it is important to be able to calculate this stiffness factor. This means that one
must consider the type of connection between the floor beams and the vertical as well
as with diagonal members when calculating the lateral spring stiffness.
5.3 Recommendations
Past testing on pony truss bridges is limited to Holts work (1957) which focused on the
effective length factor on the compression chord and not the stiffness supplied by the frames.
Empirical testing of a pony truss model could reveal a better understanding of the actual stiffness
supplied by the compression chord inertia. Due to the complexity of a full pony truss and the
necessity of empirical confirmation of any design model, testing of physical models is required
before determining any definitive conclusions.
Test a model to verify the effect of inertia on the compression chord stiffness.
40
Appendix A
Calculations of the stiffness of the lateral support to a pony truss
Ic
Ib
b
The calculation of the stiffness for the lateral supports to the top chord of an unbraced truss can
be calculated using the energy method. This can be accomplished by calculating the force, C,
that is required to induce a unit displacement in the lateral direction at the panel point. For this
purpose, the virtual work method was utilized. The following is the derivation of the spring
constants of the lateral supports if:
1. The connection between the vertical and the floor beam is assumed rigid.
Ch
Ch
Ch
= 6
!
#$
7 = #$ D
& N G + #$ J& a L
41
where, Q, is a virtual load that is applied in the horizontal direction at the points where the
displacement is to be calculated. In these in calculations, a virtual load, Q, of unity was
assumed. Notice that, , is the relative displacement between points A and B. Therefore, to
calculate the elastic constants, C, of the lateral spring, one needs to substitute a value of 2 for the
displacement, , in the equation above. This yields to:
1 2 =
Or;
<Hx
N#$*
& =
<H V
#$W
x
W
s
r f
x*
W
(A1.1)
2. To account for truss diagonals, a term of Ld3/3Id is added in the denominator where the
additional stiffness is an addition to the vertical web.
& =
(A1.2)
W
f
W
x
x* xi
i
ks
B
ks
C
Where;
ks is the joint rotational stiffness. The moment, M, at the vertical-floor beam connection is:
= = & = & /
Following the analysis that was summarized in the section above, the relative displacement
between points A and D can be calculated as:
42
= 2 =
1 =
1 =
1 =
& =
<H
#
<H
#
<H
#
<H
#
| +
H
N$*
| +
|
N$*
H
N$*
N$*
<H V
#$W
$W
#$W
+2
+|
+|
$W
<H V
<H
.
<H
.
+
#
H r f f s
x* W
(A1.3)
Notice that the above equations did not take into account the actual shape of the cross section of
the top chord members. However, in most cases, the cross section of the chord consists of open,
thin-walled sections having only one axis of symmetry, and hence the bending of the chord will
be accompanied by twisting. Thus, the problem of bucking of the chord will be a caused by
flexural and torsion. Therefore, the above relations for calculating the spring constant must be
modified to account for such effects. This problem was studied in detail by Bleich (1952).
Bleich determined that disregarding the torsion on the compression chord would lead to an
unsafe design for members with only one axis of symmetry. Conversely, box sections, which
have a high torsional rigidity, would underestimate the capacity of the chord.
43
Appendix B
Energy Method to determine the buckling load for a bar with 2 springs
1. Initial conditions
EI
L/3
L/3
L/3
2. Displaced shape
x
1
y = a sin
y =
y =
cos
sin
Where: y9 = 9 = a sin
and
3. External Work
V =
V =
5
V =
68 y dx
|68 cos
| +
dx
sin
-
y = = a sin
44
V =
5
-
4. Internal Energy
U = C9
+ C
+ 68
dx
9
U = C9
+ C
+ 68
9
Ca
|sin
U=
U=
+ sin
|sin
= =
+ sin
sin
dx
= = u
sin
-
= = u
-x
5. Combine external work and internal energy and solve for Pcr
= V+U
=
5
-
|sin
+ sin
= = u
-x
= 0.0
0.0 =
5
+ Ca |sin
= C |sin
+ sin
P = |sin
+ sin
+ sin
= = u
x
= = u
x
u
(B1.1)
45
Appendix C
Engessers Method to determine the buckling load for a bar with 2 springs
1. Initial conditions
EI
L/3
L/3
L/3
y = a sin
y =
y=
cos
sin
3. External Work
V =
V =
5
V =
68 y dx
|68 cos
V =
| +
sin
-
5
-
dx
46
4. Internal Energy
U =
68 y
dx +
68
U=
sin
U=
= =
=
sin
dx
= = u
where: = C/(L/3)
sin
= = u
-x
5
-
= = u
-x
= 0.0
5
0.0 =
P =
= = u
x
u
+
(C1.1)
ud
+ ,
ud
+ ,
x
v = c
=
'(
#d $,
<
)*+ ,
-#d ,
)*
O
=0
(C1.2)
47
Appendix D
Schwedas extension of Bleichs analysis
j=
-.
D1
"
9
sinh ;J2>
H
B?PQR
?
1L J2> 1L cosh ;G
Where:
= 2J 2L +
"
-.
=
|1 + cos . l. D. > . G
9
"
"
"
"
= D1 l. G . + l > .
"
"
"
"
2 cosh ; = J + 1L
+
+ J 1L
+
2 cos = J + 1L
+
+ J 1L
+
(D1.1)
48
Tables of Factor :
c
2n=6
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9
2.0
2.1
2.2
2.3
2.4
2.5
1.2
1.00
1.25
1.41
1.39
1.35
1.65
2.01
2.41
2.79
2.98
2.93
2.69
2.27
1.99
1.3
0.78
0.93
1.04
1.06
1.10
1.30
1.54
1.77
1.97
2.12
2.10
1.97
1.76
1.71
1.4
0.65
0.75
0.84
0.87
0.94
1.09
1.25
1.40
1.55
1.64
1.66
1.60
1.48
1.48
1.5
1.6
1.7
1.8
1.9
0.82
0.94
1.06
1.18
1.29
1.36
1.38
1.35
1.27
1.32
0.94
1.03
1.11
1.17
1.19
1.17
1.12
1.17
0.90
0.97
1.02
1.04
1.04
1.00
1.06
0.85
0.90
0.94
0.94
0.91
0.96
0.85
0.83
0.89
1.5
1.6
1.7
1.8
1.9
0.81
0.90
0.99
1.06
1.08
1.08
1.20
1.32
1.46
1.58
0.87
0.92
0.95
0.96
1.05
1.16
1.26
1.36
0.85
0.87
0.94
1.03
1.11
1.20
0.85
0.92
1.00
1.07
0.90
0.97
2n=8
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9
2.0
2.1
2.2
2.3
2.4
2.5
1.2
0.93
1.12
1.27
1.32
1.42
1.69
1.96
2.07
2.04
1.85
2.08
2.43
2.83
3.25
1.3
0.74
0.86
0.96
1.01
1.12
1.28
1.44
1.54
1.54
1.47
1.66
1.89
2.13
2.37
1.4
0.62
0.71
0.80
0.85
0.94
1.06
1.17
1.24
1.26
1.25
1.39
1.55
1.73
1.90
49
2n=10
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9
2.0
2.1
2.2
2.3
2.4
2.5
1.2
0.99
1.06
1.21
1.26
1.43
1.64
1.74
1.72
1.85
2.14
2.42
2.66
2.74
2.65
1.3
0.75
0.83
0.93
1.00
1.11
1.24
1.32
1.34
1.46
1.63
1.80
1.92
1.99
2.00
1.4
0.63
0.70
0.78
0.84
0.93
1.03
1.09
1.13
1.22
1.34
1.45
1.55
1.61
1.63
1.5
1.6
1.7
1.8
1.9
0.81
0.88
0.94
0.98
1.05
1.14
1.23
1.34
1.36
1.39
0.83
0.87
0.93
1.00
1.07
1.14
1.18
1.21
0.83
0.89
0.95
1.01
1.05
1.08
0.86
0.91
0.94
0.97
0.86
0.89
1.5
1.6
1.7
1.8
1.9
0.80
0.87
0.92
0.98
1.05
1.12
1.17
1.20
1.27
1.36
0.82
0.87
0.93
0.99
1.03
1.06
1.12
1.19
0.83
0.88
0.92
0.95
1.00
1.06
0.84
0.87
0.91
0.96
0.82
0.87
2n=12
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9
2.0
2.1
2.2
2.3
2.4
2.5
1.2
0.97
1.06
1.17
1.25
1.41
1.54
1.60
1.73
1.96
2.12
2.17
2.12
2.28
2.57
1.3
0.74
0.82
0.92
0.99
1.10
1.19
1.26
1.35
1.49
1.60
1.65
1.66
1.78
1.94
1.4
0.62
0.69
0.77
0.84
0.93
1.00
1.06
1.12
1.22
1.30
1.37
1.40
1.48
1.60
50
References
American Institute of Steel Construction (AISC). 2011. Steel Construction Manual. American
Institute of Steel Construction, Inc.
American Association of State Highway and Transportation Official (AASHTO). 2007.
AASHTO LRFD Bridge Design Specifications. AASHTO, Washington, D.C.
Ballio, G. and Mazzolani, F. 1983. Theory and Design of Steel Structures. Chapman and Hall.
New York.
Bleich, F. 1952. Buckling Strength of Metal Structures. McGraw-Hill Book Company, Inc.
New York.
Brush, D.O. & Almroth, B.O. 1975. Buckling of Bars, Plates, and Shells. McGraw-Hill Book
Company, Inc. New York.
Galambos, T. 1988. Guide to Stability Design Criteria for Metal Structures 4th Ed. John Wiley
& Sons. New York.
Holt, E. 1951. Buckling of a Continuous Beam-Column on Elastic Supports. Stability of Bridge
Chords without Lateral Bracing, Column Res. Council Rep. No. 1.
Holt, E. 1952. Buckling of a Pony Truss Bridge. Stability of Bridge Chords without Lateral
Bracing, Column Res. Council Rep. No. 2.
Holt, E. 1956. The Analysis and Design of Single Span Pony Truss Bridges. Stability of Bridge
Chords without Lateral Bracing, Column Res. Council Rep. No. 3.
Holt, E. 1957. Tests on Pony Truss Models and Recommendations for Design. Stability of
Bridge Chords without Lateral Bracing, Column Res. Council Rep. No. 4.
Lutz, L., and Fisher, J.M. 1985. A Unified Approach for Stability Bracing Requirements. AISC
Eng. J. Vol. 22, No. 4.
Salmon, C., Johnson, J. and Malhas, F. 2009. Steel Structures Design and Behavior 5 Ed.
Pearson Prentice Hall. New Jersey.
SAP Inc. 2009. ANSYS User's Manual for Release 12.1.
Timoshenko, S. 1936. Theory of Elastic Stability. McGraw-Hill Book Company, Inc. New
York
Thandavamoorthy, T.S. 2005. Analysis of Structures: Strength & Behavior. Oxford University
Press. New York.
51
Winter, G. 1960. Lateral Bracing of Columns and Beams. Trans. ASCE, Vol. 125, pp. 807-845.
Ziemian, R.D. 2010. Guide to Stability Design Criteria for Metal Structures 6th Ed. John Wiley
& Sons. New York.