You are on page 1of 4

TECHNICAL NOTES

Steel Girder Stability during Bridge Erection: AASHTO LRFD


Check on L / b Ratios
John S. Hastings, P.E.1; Qiuhong Zhao, P.E., M.ASCE2; and Edwin G. Burdette, P.E., F.ASCE3
Downloaded from ascelibrary.org by Purdue University Libraries on 08/29/19. Copyright ASCE. For personal use only; all rights reserved.

Abstract: The erection of steel plate girders during the construction process of a steel bridge is a complex operation, which is often left
to the contractor and/or the subcontractor to plan and execute. Rules of thumb have been developed through experience to check the
lateral torsional buckling of the steel girder during erection using the maximum L / b 共unbraced length/compressive flange width兲 ratio,
below which no lateral torsional buckling would occur. Although the L / b ratio check has proven to be useful and convenient on-site, it
is necessary to provide a more rational basis for the rules of thumb, and find the maximum L / b ratios by checking the lateral torsional
buckling failure of girders under erection according to the latest AASHTO LRFD code. A series of parametric studies were conducted on
cantilever and simply supported girders under self-weight as well as self-weight plus wind load, in order to: 共1兲 check the rules of thumb
on L / b ratios and 共2兲 determine the effects of girder flange width, flange thickness, web depth, web thickness, and yield strength on the
maximum L / b ratio and girder stability during erection. From the results, rules of thumb were modified for girders with common shapes,
and it was obvious that 共1兲 self-weight plus wind load controls the girder stability during erection in most cases and 共2兲 flange width and
web depth have the most effects on the maximum L / b ratio and girder stability during erection.
DOI: 10.1061/共ASCE兲BE.1943-5592.0000102
CE Database subject headings: Bridges, steel; Bridges, girder; Buckling; Construction.
Author keywords: Steel girder; Erection; Instability; Lateral torsional buckling.

Introduction and Background the girder stability during erection is checked by the L / b ratio
共unbraced length/compressive flange width兲. For a cantilever
When a steel girder is under erection during the construction pro- girder during erection: 共1兲 if the L / b value is less than 30, stabil-
cess of a steel bridge, it may experience instability and lateral ity is guaranteed; 共2兲 if the L / b value is between 30 and 40,
torsional buckling failure if the unbraced length is too large, stability is doubtful; and 共3兲 if the L / b value is over 40, tempo-
which in turn might cause collapse of the whole steel superstruc- rary support will be required. For a simply supported girder, the
ture, as illustrated by the collapse of the State Route 69 Bridge at L / b ratio limits are doubled. Although the rules of thumb on L / b
Clifton, Tennessee on May 16, 1995. ratios are convenient and useful on-site, there is no theoretical
In many cases, the steel girder erection was left to the contrac- background other than experiences, and a more rational basis is
tor and/or the subcontractor to plan and execute, based mainly on urgently needed.
past experiences. Rules of thumb have been used when an erector
needs to conduct a quick stability check or make a decision on
temporary removal of cross frames on-site, since precise stability Checking Stability of Cantilever Girder during
analysis might not be available at the moment. One set of such Erection
rules of thumb have been published in the Handbook for Con-
struction Engineers 共American Bridge Division 1983兲, in which The governing design code AASHTO LRFD Specifications
共AASHTO 2007兲 was adopted to check the rules of thumb on L / b
1
Civil Engineer Manager I, Tennessee Dept. of Transportation ratio. The girders in this study are longitudinally unstiffened,
共TDOT兲, Suite 1100 James K. Polk Building, Nashville, TN 37243-0349, straight, and prismatic with no braces along the span. According
E-mail: john.s.hastings@state.tn.us to AASHTO LRFD, a steel girder will not experience elastic lat-
2
Assistant Professor, Dept. of Civil and Environmental Engineering, eral torsional buckling if f bu ⱕ ⌽ f Fcr, where f bu = external normal
Univ. of Tennessee, 109A Perkins Hall, Knoxville, TN 37996-2010, stress and Fcr = elastic lateral torsional buckling stress. For a cer-
E-mail: qzhao@utk.edu tain girder under erection, f bu increases while Fcr decreases with
3
Professor, Dept. of Civil and Environmental Engineering, Univ. of the span L, therefore, a certain limit Lmax and 共L / b兲max exits,
Tennessee, 115 Perkins Hall, Knoxville, TN 37996-2010. E-mail: beyond which the girder will experience elastic lateral torsional
eburdett@utk.edu
buckling.
Note. This manuscript was submitted on June 6, 2008; approved on
January 19, 2010; published online on January 22, 2010. Discussion pe- Parametric studies were then conducted, in which 共L / b兲max
riod open until April 1, 2011; separate discussions must be submitted for ratios were obtained for a group of cantilever girders with varying
individual papers. This technical note is part of the Journal of Bridge geometric and material parameters, including flange width, flange
Engineering, Vol. 15, No. 6, November 1, 2010. ©ASCE, ISSN 1084- thickness, web depth, web thickness, and material yield strength,
0702/2010/6-759–762/$25.00. as summarized in Table 1. For each girder section, the span L was

JOURNAL OF BRIDGE ENGINEERING © ASCE / NOVEMBER/DECEMBER 2010 / 759

J. Bridge Eng., 2010, 15(6): 759-762


Table 1. Parameters in Analytical Studies
Flange width, b Flange thickness, t f Web depth, D Web thickness, tw Yield strength, Fy
Case 共mm兲 共mm兲 共mm兲 共mm兲 共MPa兲
Cantilever—1 305–914 25–38a 1,016–2,438 13–17b 345
Cantilever—2 914 38–102 1,219–2,438 13–17b 345
Cantilever—3 914 38 1,219–2,438 10–25b 345
Cantilever—4 914 38 1,219–2,438 10–25b 248–483
Simple—1 305–914 25–38a 1,016–2,134 13–14b 345
a
Flange thickness is constant for each given flange width.
b
Web thickness is constant for each given web depth.
Downloaded from ascelibrary.org by Purdue University Libraries on 08/29/19. Copyright ASCE. For personal use only; all rights reserved.

increased in steps, and a spreadsheet was developed to check In summary, for any specific cantilever section, the flange width
f bu ⱕ ⌽ f Fcr for each step. When the equation did not satisfy any and web depth are dominating parameters for girder stability dur-
more, the Lmax and 共L / b兲max values for this girder section were ing erection, and the flange thickness, web thickness, and material
then identified. The AASHTO Load Combination Strength IV yield strength only have negligible effects; the 共L / b兲max ratio
共1.5D兲 and Strength III 共1.25D + 1.40W兲 were considered, respec- could be directly read or interpolated from Fig. 1, beyond which
tively. Purposes of the parametric studies included: 共1兲 check the the girder will lose stability during erection.
rules of thumb on 共L / b兲max ratios for common girder shapes and Comparing the rules of thumb to the AASHTO LRFD results
共2兲 determine the effects of girder flange width, flange thickness, in Fig. 1, it is obvious that the original lower limit of 30 on the
web depth, web thickness, and yield strength on 共L / b兲max ratios L / b ratios, below which stability is guaranteed for any girder
and girder stability during erection. section, turned out to be slightly conservative, since any girder
section with an L / b ratio above 30 but below 35 would still
remain stable. But the lower limit could remain as 30, since not
AASHTO LRFD Load Combination Strength IV all possible sections were checked. On the other hand, the original
„1.5D… upper limit of 40 on the L / b ratios, above which temporary sup-
port would be needed for any girder section, turned out to be
The AASHTO Load Combination Strength IV was considered for
overly conservative, since numerous girders will remain stable
the first set of analyses, which applies to a very high dead load
with a L / b ratio above 40 but below the 共L / b兲max ratios. There-
versus live load case. For Case 1 study, the 共L / b兲max ratios are
plotted in Fig. 1, where each curve represents a group of sections fore, the rules of thumb on the lower and upper limits of L / b
with the same flange width and thickness, while the web depth ratios for cantilever should be modified to 30 and 60, based on the
and thickness vary along the curve. It is obvious that the 共L / b兲max AASHTO LRFD check under only the dead load effects.
ratio will increase with the web depth and thickness, indicating
that a lower girder section with a thinner web would generally
have more tendency for instability during erection compared to a AASHTO LRFD Load Combination Strength III
deeper girder section with a thicker web. From Case 2 to 4 stud- „1.4D + 1.25W…
ies, it turned out that the 共L / b兲max ratio would increase very
slightly with the flange thickness, while essentially not affected The AASHTO Load Combination Strength III was considered for
by the material yield strength. In addition, the 共L / b兲max ratio de- the second set of analyses, which applies to a high wind load
creased very slightly when the web depth was held constant and case. Minor-axis bending is then combined with major axis bend-
the web thickness increased from a minimum value of 150 times ing according to f bu + 共1 / 3兲f l ⱕ ⌽ f Fnc, where f l = external normal
web thickness 共AASHTO 2007兲 to a maximum value of 25 mm, stress due to wind load; f bu = external normal stress due to gravity;
mainly because the external load demand due to girder self- and Fnc = nominal flexural resistance of the compression flange.
weight would increase faster than the girder capacity in this case.
The lateral load also needs to be checked to ensure that minor-
axis failure does not occur before the onset of lateral torsional
buckling by f l ⱕ 0.6Fyc, where Fyc = yield strength of the compres-
sion flange 共AASHTO 2007兲.
The AASHTO code does not specify the magnitude of wind
pressure during erection but only for normal design, for which if
the height of the girder is less than 30 ft above ground, a design
wind pressure of 25.6 pounds per square foot 共psf兲 should be
used, corresponding to a wind velocity of 100 mi/h. However, the
design wind pressure is not suitable for a construction check,
since a contractor is not likely to erect a bridge during periods of
extremely high winds. There might be windy days during various
phases of bridge construction, but the contractor will usually have
multiple girders in place with bracing installed 共i.e., cross frames
and laterals if required兲. Unless something out of the ordinary
occurs, namely an extraordinary event, the steel girder under erec-
tion would not experience a high wind loading without any brac-
Fig. 1. Maximum L / b ratios for cantilevers under self-weight ing along the span. Therefore, the wind pressure experienced by

760 / JOURNAL OF BRIDGE ENGINEERING © ASCE / NOVEMBER/DECEMBER 2010

J. Bridge Eng., 2010, 15(6): 759-762


Downloaded from ascelibrary.org by Purdue University Libraries on 08/29/19. Copyright ASCE. For personal use only; all rights reserved.

Fig. 2. Maximum L / b ratios for cantilevers under self-weight plus Fig. 3. Maximum L / b ratios for simply supported girders under self-
wind load weight plus wind load
girder during erection should be modified for extraordinary
events.
Unfortunately, the AASHTO code does not specify the magni- nation, which turned out to be more critical in most cases. For
tude of wind pressure under extraordinary events either. ASCE Case 1 study, the 共L / b兲max ratios are plotted in Fig. 3, where each
7-05 Minimum Design Loads for Buildings and Other Structures curve represents a group of sections with the same flange width
共ASCE 2005兲 was then referred to, where the wind load factor for and thickness, while the web depth and thickness vary along the
extraordinary events is 1/8 of the wind load factor for normal curve. For any specific simply supported girder section, the
design, which means the wind pressure for extraordinary events is 共L / b兲max ratio could be directly read or interpolated from the
also 1/8 of the wind pressure for normal design. Since the wind curve, beyond which the girder will lose stability during erection.
pressure is proportional to the square of wind velocity, the wind Unlike the cantilevers, the 共L / b兲max ratio would decrease obvi-
velocity for extraordinary events could be calculated as VDZ ously when the web depth and thickness increase, especially for
= 冑共1 / 8兲 ⫻ 100 mi/ h = 35.4 mi/ h. Considering wind gusts in the girders with relatively narrow flanges, since the wind load effects
event of a sudden summer storm, the wind velocity for girder will dominate in those cases.
under erection could be further increased to 45 mi/h, resulting in Comparing the rules of thumb to the AASHTO LRFD results
a wind pressure of qz = 共45 mi/ h / 100 mi/ h兲2 ⫻ 25.6 psf in Fig. 3, it is obvious that the original lower limit of 60 on the
= 5.18 psf. L / b ratios, below which stability is guaranteed, turned out to be
The same groups of parametric studies were conducted using unsafe. All the 共L / b兲max ratios in Fig. 3 are below 60, while any
the Strength III combination. The effects of the web depth and girder section with an L / b ratio below 60 but above 共L / b兲max will
flange width on the 共L / b兲max ratio for Case 1 are shown in Fig. 2, become unstable during erection. In addition, the original upper
which generally looks similar to the Strength IV combination. limit of 80 on the 共L / b兲max ratios, above which temporary support
However, for some of the girders with narrow flanges, the would be needed, turned out to be unconservative as well, since
共L / b兲max ratio would decrease when the web depth and thickness any girder with a L / b ratio above 60 would need temporary sup-
increase, probably because the wind load would affect the girder port already. Therefore, the rules of thumb on the lower and upper
stability more in these cases. From Case 2 to 4 studies, the results limits of L / b ratios for simply supported girder should be modi-
were similar to the Strength IV combination, except that the fied to 45 and 60, based on the AASHTO LRFD check consider-
maximum L / b ratios are slightly higher. ing the combined wind load and dead load effects.
For the same girder section, the 共L / b兲max ratios under the
Strength IV combination 共1.5D兲 and under the Strength III com-
bination 共1.25D + 1.4W兲 were compared together. It turned out Conclusions
that for 73% of all shapes investigated, a lower 共L / b兲max ratio was
obtained under the Strength III combination, which means wind 1. L / b ratio is convenient and useful on-site to check girder
load plus self-weight will control the girder stability during erec- stability during lifting, shipping, erection, and temporary
tion, except for girders with very wide flanges including the lower cross frame removal. A series of graphs for the 共L / b兲max
bound curve. The difference becomes more obvious as the web ratios were created for the convenience of constructors. Note
becomes deeper and the flange becomes narrower, since the wind that the L / b ratio check cannot replace the more rigorous
load effects dominate more. Therefore, it is necessary to consider stability check if deemed needed by the structural engineers.
the combined wind load and dead load effects for most cases, 2. Based on parametric studies using AASHTO LRFD, the
when checking girder stability during erection. The rules of dominating factors for girder stability during erection are the
thumb on the lower and upper limits of L / b ratios for cantilever compression flange width and the web depth.
were then further modified to 30 and 50, based on the AASHTO 3. In most cases, especially for girders with narrow flanges or
LRFD check under the combined wind load and dead load effects. deep webs, it is more critical to consider the combined wind
load and dead load effects when checking girder stability
Checking Stability of Simply Supported Girder during erection.
during Erection 4. Based on parametric studies using AASHTO LRFD, the rules
of thumb on the lower limit of L / b ratios for cantilever were
Analyses were also conducted on simply supported girders to slightly conservative, but could remain as 30; the rules of
determine the 共L / b兲max ratios under the Strength III load combi- thumb on the upper limit of L / b ratios were overly conser-

JOURNAL OF BRIDGE ENGINEERING © ASCE / NOVEMBER/DECEMBER 2010 / 761

J. Bridge Eng., 2010, 15(6): 759-762


vative, and could be modified to 50 considering the com- Fyc ⫽ specified minimum yield strength of
bined wind load and dead load effects. compression flange;
5. Based on parametric studies using AASHTO LRFD, the rules f bu ⫽ largest value of compressive stress throughout
of thumb on the lower and upper limits of L / b ratios for the unbraced length;
simply supported girder were unsafe or unconservative, and f l ⫽ flange lateral bending stress;
should be modified to 45 and 60 considering the combined L ⫽ actual unbraced length;
wind load and dead load effects. qz ⫽ wind stagnation pressure;
VDZ ⫽ design wind velocity;
W ⫽ wind load; and
␾ f ⫽ resistance factor for flexure.
Notation
References
The following symbols are used in this technical note:
Downloaded from ascelibrary.org by Purdue University Libraries on 08/29/19. Copyright ASCE. For personal use only; all rights reserved.

b ⫽ width of compression flange; AASHTO. 共2007兲. LRFD bridge design specifications, 4th Ed., Washing-
D ⫽ dead load due to girder self-weight; ton, D.C.
Fcr ⫽ elastic lateral torsional buckling stress; American Bridge Division. 共1983兲. Handbook for construction engineers,
Fnc ⫽ nominal flexural resistance of a compression United States Steel Corporation, Pittsburgh.
flange; ASCE. 共2005兲. Minimum design loads for buildings and other structures,
Fy ⫽ specified minimum yield strength of steel; Reston, Va.

762 / JOURNAL OF BRIDGE ENGINEERING © ASCE / NOVEMBER/DECEMBER 2010

J. Bridge Eng., 2010, 15(6): 759-762

You might also like