16TH CONGRESS OF IABSE, LUCERNE, 2000
Warping Displacement of Skew Reinforced Concrete Box Girder Bridges
Mohamed I. SOLIMAN Aly S. ABDEL FAIAD Osama H. ABDEL WAHED
Professor Assoc. Professor Assoc. Professor
Ain Shams University Ain Shams University Ain Shams University
Cairo, Egypt Cairo, Egypt Cairo, Egypt
Hatem S. EL-BEHAIRY
Engineer
Housing Research Center
Cairo, Egypt
Summary
Skew reinforced concrete multi-cell box girder bridges are commonly used as over crossings in
highway intersections and interchanges, especially in complex intersections and in crowded areas
where lack of space necessitates the use of skew geometry. Due to the unexpected behavior for this
type of bridges, skew box girder bridges are difficult in the analysis and design. Several research
efforts using theoretical as well as experimental approaches have been made over the past 20 years
to understand the actual behavior of this type of bridges when subjected to static and dynamic loads.
The earlier design method for box girder bridges, or hollow girder type bridges, or the multi cell
bridges was conservative and employed large web thickness to neglect the effect of distortion and
warping stresses. However, with the present tendency to use thin web thickness to reduce self-
weight of the structure, the associated distortion and warping effects become significant and must be
taken into consideration during the design procedure.
This paper presents the effect of varying the skew angle on the warping displacement and the
vertical reactions of skew box girder bridges. The theoretical results obtained using three-
dimensional non-linear finite element program for three R.C. box girders are represented. All
girders have the same dimensions and the same reinforcement in both longitudinal and transverse
directions but with different skew angles. The specimens were analyzed under an eccentric static
concentrated load acting at mid span section of the outer web to investigate the effect of varying the
skew angle.
Keywords: Reinforced Concrete Bridges, Box Girder Sections, Multi-cell Bridges, Skew Angle,
Warping Displacement, And Reactions.
1. Introduction
Warping is defined as the out -of-plane movement of points that were initially in the plane of the
cross-section and results from torsion loading. When concrete box girder is subjected to a torsion
loading, it undergoes warping displacements that are composed of two components, the torsion
warping displacements and the dis-torsion warping displacements. Both components result in
longitudinal normal stresses, (warping stresses). Under these conditions, the significance of the
warping stresses depends on the geometry of the structure as well as on the nature of the loading and
support conditions [1].
In the past the practice of the designer engineers was to approximate the skew bridge as a right
bridge and neglect the effect of skew angle. This approximation has been done arbitrary rather than
on a well-founded basis and may be valid for small skew angles [2,3]. For large skew angles, this
approximation may lead to unsafe design. The analysis of skew bridges does not depend only on the
skew angle [φ], but also on the spacing between the longitudinal webs [S] and the span of the bridge
[L] (in case of single box girder bridge S = bridge width). It was recommended that the ratio (S tan
φ/L) to be taken as an appropriate measure of skews, also it is valid to approximate the skew bridge
as a right bridge with (S tan φ/L) up to only 0.05 [4].
16TH CONGRESS OF IABSE, LUCERNE, 2000
Most of the previous papers and researches [5,6] are devoted to calculate the different straining
actions for the box section and ignore the warping displacement, since it needs a powerful three-
dimensional program. So the main purpose of the present research is to investigate the effect of
varying the skew angle on the warping displacement and the support reaction of skew box girder
bridges.
2. Finite Element Method
The finite element method has long been recognized as one of the most accurate and effective
techniques for analyzing complicated structural engineering problems. One of the advantages of this
method is that it provides a convenient and adaptable tool for covering the problems associated with
the analysis of reinforced concrete structures. These problems are cracking of concrete, nonlinear
stress strain response of concrete, shrinkage and creep of concrete, yielding and strain hardening of
reinforcing steel, interaction between concrete and steel (bond-slip), post cracking, shear resistance
due to aggregate interlocking and dowel actions [7].
3. Details of the Analyzed Specimens
The analyzed girders were of the same length and cross sectional dimensions, but with different
skew angles. The external and internal dimensions of the cross section were 40x26cm and 26x12cm
respectively. The experimental results of these girders have been published before in paper ref. [8].
Figure 1 shows the concrete dimensions of the analyzed specimens. All girders have the same
longitudinal and transverse reinforcement. Figure 2 shows the details of reinforcement for the
analyzed girders.
Fig. 1 Dimensions of the girders analyzed (G1, G & G3)
16TH CONGRESS OF IABSE, LUCERNE, 2000
Fig. 2 Reinforcement details for girder G2 (α = 25o)
4. Finite Element Idealization
The used program utilizes a fully nonlinear algorithm analysis. The analysis is based on a linear
iterative secant stiffness formulation and incorporates a fully nonlinear constitutive model for
concrete and steel. Figure 3 shows the finite element idealization used in the program analysis.
Fig. 3 Idealization for the finite element analysis
16TH CONGRESS OF IABSE, LUCERNE, 2000
The different types of elements used in the analysis are stated as follows:
I- Three-dimensional brick element to represent the concrete. [An eight noded regular hexahedron]
II- Three-dimensional truss element to account for the reinforcing steel.
5. Analysis Of The Theoretical Results
The analytical results presented here are concentrated on the warping displacement at the girder
ends as well as the vertical reactions at support line. For more details about other straining actions
refer to [9].
A- Warping Displacement
The warping displacements which represent the horizontal deformations at supports in the X-
direction (i.e. the out of plan displacement, i.e. in the longitudinal direction of the girder) have been
calculated at top and bottom slab corners for the loaded and the unloaded webs at different stages of
loading. Figures 4 to 8 show the warping displacement for the analyzed girders. From the
represented figures, the following remarks could be observed:
1. For all girders as the torsion moment increases the warping displacement increases. This is due to
the reduction in the overall stiffness, which leads to the increase in the warping displacement.
2. For girder G1 (α= Zero), the warping displacements at both ends of the girder were symmetric
about the girder axis. The warping displacements of the loaded web at the bottom slab points
[3,4] were higher than those at the unloaded web points [1,2] by about 13%. While at the top slab
of the unloaded web, the warping displacements at points [1,2] were higher than those at the
loaded web points [3,4] by about 9%.
3. For girder G2 (α=250), by studying the first support zone, the warping displacements at the
bottom and the top slabs of the loaded web points [4,8] were higher than the unloaded web points
[1,5] by about 15%. On the other hand, at the second support, the warping displacements at the
top and the bottom of the unloaded web points [2,6] were higher than those at the loaded web
points [3,7] by about 20%.
4. For girder G3 (α=400), the warping displacements at the loaded and the unloaded webs at the
second support line were higher than those at the first support were. This may be attributed to
that the second support is nearer to the load than the first support.
5. For girder G3 (α=400), by studying the first support zone, the warping displacements at the
bottom of the loaded web point [4] were higher than those of the unloaded web point [1] by
about 44% at torsion moment 100 cm. ton. While the displacements of the top of the loaded web
point [8] were higher than those at the unloaded web point [5] by about 14% at the same
torsional moment level. On the other hand, at the second support, the warping displacements at
the top and the bottom of the unloaded web points [2,6] were higher than those at the loaded web
points [3,7] by about 7%.
6. In general, the warping displacements of the bottom slab points [1,2,3,4] were higher than those
of the top slab points [5,6,7,8] and opposite in directions.
7. Comparison between the three analyzed girders, at the same torsional moment level, the warping
displacements at the bottom of the loaded and the unloaded web at the first support points [1,4]
of girder G2 (α=250) were higher than those of G3 (α=400) than of girder G1 (α=Zero0). While at
the second support, at the bottom of the loaded and the unloaded web points [2,3], the warping
displacements of girder G3 (α=400) were higher than those of girder G2 (α=250) than of girder
G1 (α=Zero0).
16TH CONGRESS OF IABSE, LUCERNE, 2000
LOADED WEB
UNLOADED WEB
PLAN
16TH CONGRESS OF IABSE, LUCERNE, 2000
LOADED WEB
UNLOADED WEB
PLAN
2
16TH CONGRESS OF IABSE, LUCERNE, 2000
B-Vertical Reactions
The vertical reactions at the two supports were obtained at the corner points shown on figure 9
[points: 1,2,3&4] at load equals to 2.0 ton (neglecting the own weight of the girder) using a finite
element analysis program. The results are as shown in Table 1 and figure 9.
Table 1 Reactions at Supports
Reaction At Unloaded web (ton) Reaction At Loaded web (ton)
Girder -1- -2- -3- -4-
G1 0.0232 0.0232 0.601 0.601
G2 -0.2804 0.386 0.2824 0.8367
G3 -0.398 0.6201 0.0332 0.952
From Table 1 and Figure 9 the following remarks could be observed:
1. The vertical reactions at the same load level, at points (1&3) for girder G1 (α= Zero0) were
higher than the reaction for girder G2 (α=250) than that for girder G3 (α=400). [Points (1&3) are
the acute corners of the loaded and the unloaded webs].
2. The vertical reactions at the same load level, at points (2&4) for girder G3 (α=400) were higher
than those for girder G2 (α=250) than those for girder G1 (α= Zero0) at the same load level.
[Points (2&4) are the obtuse corners of the loaded and the unloaded webs].
3. For girders G2 (α=250) and G3 (α=400), the vertical reactions at points (2,4) were higher than the
reaction at points (1,3) respectively. This may be due to that the obtuse support has higher
reaction than the acute support.
4. The max. reactions were at the obtuse support of the loaded web of girder G3 (α=400). This
means that by increasing the skew angle, the vertical reaction at the obtuse support increases
while the acute reaction decreases.
2
/
PLAN
PLAN
2
/!
PLAN
Fig. 9 Vertical support reactions for P = 2 tons
16TH CONGRESS OF IABSE, LUCERNE, 2000
6. Verification Of The Finite Element Analysis
The results of the finite element program were compared with those measured in the experimental
program [9]. Figure 10 shows the comparison between the experimental and the theoretical results
for girder G2 (α=250) which shows a good agreement between the theoretical and the experimental
results.
7. Conclusions
1. For the skew and right girders, the warping displacement of the bottom slab was higher than that
at the top slab. This may be due to the reduction in the girder stiffness.
2. Due to the more rigidity of acute corners than obtuse corners, the warping displacement of the
acute corners was less than that of the obtuse corners. These displacements increase with the
increase in skew angle
3. The warping displacement for the straight girders of the bottom slab at the loaded web was
higher than those at the unloaded web. While for the top slab, the warping displacements at the
unloaded web were higher than that at the loaded web. This may be attributed to the effect of the
warping stresses of the box section at the ends and the effect of the torsional loads.
4. For the straight girders the vertical reactions were transferred mainly by the loaded web, while
for the unloaded web, the vertical reactions are minimum.
5. The vertical reactions at the obtuse supports were higher than those at the acute supports. This
reaction increases with the increase in the skew angle. For the unloaded web, the reaction at the
obtuse corner was higher than that at the acute corner of the loaded web.
6. The support at the acute corner of the unloaded web has a negative reaction, and this reaction
increases with the increase in the skew angle. This support may require a special support
condition to resist the uplift at that corner.
16TH CONGRESS OF IABSE, LUCERNE, 2000
References
[1] Soliman M.L. and Mirza M.S., “ Analysis and Design of Reinforced Concrete Box Girder
Bridges”, Structural Concrete Series No.76-1, Department of Civil Engineering and
Applied Mechanics, McGill University, Montreal, February 1976.
[2] Bakht B. “Analysis Of Some Skew Bridges As Right Bridges” Journal Of Structural
Engineering, Vol. 114 March 26,1987.
[3] Shushkewish K.W. “ Approximate Analysis Of Concrete Box Girder Bridges” Journal Of
Structural Engineering, Vol. 114, No. 7 July 1988.
[4] American Association Of State Highway Officials ASSHTO, “Standard Specifications For
Highway Bridges” Ninth edition, Washington D.C. 1992.
[5] Soliman M.I., El-Hashimy M.M, El-Behairy S.A. and Abdel-Wahed O.H. “ Effect of
diaphragms location on the torsional and distorsional behavior of R.C. curved box girder
bridges” Ph.D. thesis, structural engineering department, Ain shams university, Cairo Egypt.
[6] Soliman M.I., El-Behairy S.A., Abdel Faiad A.S. and Youssef M.A. “ Behavior of R.C. thick
wall girder bridges” Al-Azhar engineering fourth international conference, December 1995.
[7] Soliman M.I., El-Behairy S.A., Abdel Faiad A.S. and El-Mekawy S.M.. “ Non Linear
Behavior Of R.C. Concrete Girder Type Bridges” Ph.D. Thesis, Department of Civil
Engineering, Ain Shams university, 1997.
[8] Soliman M.I., Abdel Faiad A.S., Abdel-Wahed O.H. and El-Behairy Hatem S. “Skew Angle
Effect On The Torsion Behavior Of R.C. Box Girder Bridges” Seventh International
Colloquium On Structural And Geotechnical Engineering, December 1996, Ain Shams
University, Cairo, Egypt.
[9] Soliman M.I., Abdel Faiad A.S., Abdel-Wahed O.H. and El-Behairy Hatem S. “Skew Angle
Effect On The General Behavior Of R.C. Box Girder Bridges” [Link]. Thesis, Structural
Engineering Department, octobor 1998, Ain Shams University, Cairo, Egypt.
