IABSE Symposium „Metropolitan Habitats and Infrastructure“, Shanghai, 22.-24.09.
2004
Temperature Induced Deformations in Match-Cast Segments
Guenter Axel ROMBACH Ra’ed ABENDEH
Professor Dr.-Ing. MSc.
Univ. Hamburg-Harburg Univ. Hamburg-Harburg
Dept. Concrete Structures Dept. Concrete Structures
Germany Germany
G. Rombach, born 1957, received Rae’d Abendeh, born 1975,
his civil engineering degree from received his Master degree in
Univ. of Karlsruhe, Germany. civil engineering from Irbid
After his Ph.D. in 1991 he worked University, Jordan. Since 2001
as design engineer until he he is research assistant at the
became professor at the Univ. TUHH.
Hamburg-Harburg in 1996. He is
also a chartered engineer.
Summary
Precast segmental hollow box girder bridges have become the preferred constructions method for
many elevated highway projects in the last years. Although many segmental bridges had been built
so far some significant design aspects need still further investigation. One is the bowing of match
cast segments caused by the temperature gradient during hydration of concrete. The resulting gap
between two adjacent elements may significantly reduce the durability and the load bearing
capacity of the structure. A thermo-mechanical finite element model was developed to simulate the
behavior of segments during match-casting. A numerical simulation of a real segmental bridge
consisting of bowed and perfect fitting segments and dry joints will be presented in this paper.
Keywords: segmental bridge; numerical modeling; thermal deformations; thermal stresses;
bowing effect
1 Introduction
In contrast to classical monolithic constructions a segmental bridge consists of small precast
elements, called segments, which are stressed together by external tendons (see Figure 1, 2).
Segmental bridges are often built as a single span structure in order to avoid coupling of the post
tensioning cables and joints near the intermediate supports. Each span consists of three different
segments namely pier, deviator and standard segments. Dry or epoxy joints are used between the
segments. The shear keys at the element faces are designed to carry the shear loads at the joints,
where also the friction between the joint surface takes apart in resisting shear forces.
Fig. 1 Segmental bridge during construction Fig. 2 Match casting of segments
IABSE Symposium „Metropolitan Habitats and Infrastructure“, Shanghai, 22.-24.09.2004
Match casting (Fig. 3) is a common segment production method for segmental post-tensioned
concrete box girder bridges to guarantee an almost perfect fitting . The first segment of a span is
cast between one fixed and one removable bulkhead, as can be seen in Figure 3a. The subsequent
pieces are cast between the fixed bulkhead and the previously cast segment, Figure 3b. The heat of
hydration of the fresh concrete in the new segment causes a thermal gradient in the ‘old’ match cast
segment. This gradient causes a bowing of the match cast segment (see Figure 3c). Thus the new
segments have one straight and one curved side, whereas the old segment returns to its original
shape after cooling to ambient temperature, as can be seen in Figure 3d, where ∆ represents the
deflection value.
b
a
removeable bulkhead 2
1
1
fixed bulkhead fixed bulkhead
d
c
2
∆
1
fixed bulkhead
Fig. 3 Bowing of match cast segments [1]
2 Heat-Transfer Formulations
A study of the heat-transfer during match-casting will allow the determination of the temperature
distribution within a body, the amount of energy moving into or out of the body and the
subsequence thermal stresses and deformations [2].
qx ⋅ A ⋅ dt + Q ⋅ A ⋅ dx ⋅ dt = c( ρ ⋅ A ⋅ dx )dT + qx + dx ⋅ A ⋅ dt + qh ⋅ Ρ ⋅ dx ⋅ dt (1)
qh = h( T − T∞ ) (2)
where qx is the heat conducted (heat flux) into the control volume at surface edge x
qx+dx is the heat conducted out of control volume at the surface edge (x+dx)
t is time
Q is the internal heat source (heat generated per unit time per unit volume)
A is the cross-sectional area perpendicular to heat flow q
c is the specific heat in W◊s/kg◊oC
ρ is the mass density in kg/m3
h is the heat-transfer or convection coefficient, in W/(m2 °C)
T is the temperature of the solid surface, at the solid/fluid interface
T∞ is the temperature of the fluid
P is the perimeter around the constant cross-sectional area A
3 Numerical Simulation of Match-Casting
The finite-element software ANSYS was used to simulate the match-casting of segments. The new
match-cast segment is modeled by eight-node thermal solid elements, the old one by eight-node
thermal and structural field capability solid elements. Link elements are used to model the transfer
of heat across the joint. Figure 4 shows the finite element mesh of match-cast segments and the
resulting deformations.
IABSE Symposium „Metropolitan Habitats and Infrastructure“, Shanghai, 22.-24.09.2004
Link Elements
New Segment
Old
Segment
Fig. 4 Numerical model of the match-cast segments and the deformed shape
4 Simulation of Match-Casting
A thorough verification of such complex finite element models is very important. The field
measurements of the standard segments type I and type III (Fig. 5) of the San Antonio Segmental
Bridge [1] were used for this purpose. The concrete mixes and material properties as specified by
Zeitler [3] were used as only little information about the concrete used for the San Antonio Bridge
was available. Table 1 shows the major characteristic values.
7.31 m 17.06 m
30 25
1.80m
32 40
225 25
5 48
2.44 m 47 1.96 m 4.90 m 5.60 m
a) Segment type I b) Segment type III
Fig. 5 Standard segments and dimensions [1]
Table 1 Characteristic values for the concrete mix
Thermal conductivity coefficient Kxx 2.65 W/m◊oC
Specific heat C 1128 W◊s/kg◊oC
Convection coefficient h 5.8 W/m2◊oC
Concrete density ρ 2340 kg/m3
Modulus of elasticity (28 day-strength) E 34,000 MPa
Thermal expansion coefficient αT 12◊10-6 1/oC
Figure 6 shows one of the deflection measurements conducted in old segment type III in the San
Antonio Project at upper slab and the numerical results obtained from the finite element model. The
maximum measured deflection (1.78 mm) was close to the deflection value obtained numerically
(1.67 mm).
new segment
3
deformations ∆ (mm)
old segment
the deformed
edge
x
2 measured deformations
1 numerical
analysis
Fig. 6 Measured deformations at the
0 upper edge of old segment type III
0 300 600 900 1200 1500 1800 after ten hours of match-cast process.
distance x (m)
IABSE Symposium „Metropolitan Habitats and Infrastructure“, Shanghai, 22.-24.09.2004
Figures 7 and 8 show the deformations of the old match-cast segments for segment types I and III
respectively. It can be seen that the width to length ratio (W/L) of a segment has a significant effect
on the bowing. Type I segment with W/L = 3.0 gives a maximum deflection of 0.31 mm whereas in
the slender type III segments, which has a W/L ratio of 9.33, the resulting greatest deflection was
1.72 mm. Due to the development of the thermal gradient in the old segment, thermal stresses
formed in the old segment as can be seen in Figure 9. It is assumed that the strength is about 20 %
of the 28 day-concrete-strength.
Fig. 7 Deformations (m) in the z-axis direction obtained in the type I old segment
Fig. 8 Deformations (m) in the z-axis direction obtained in the type III old segment
Fig. 9 Thermal stresses σx (MPa) in x-direction of the type III old segment
IABSE Symposium „Metropolitan Habitats and Infrastructure“, Shanghai, 22.-24.09.2004
5 Numerical Model of Segmental Bridge
In the following the influence of the gap between two adjacent segments on the deformation of the
structure and the stresses of a real segmental bridge will be studied. Figure 10 shows the geometry
and the dimensions of what is one of the world’s longest segmental bridge, the Second Stage
Expressway System, which was built in Bangkok, Thailand [4]. The bridge is prestressed by
external tendons and has dry joints.
The hollow box girder beam is modeled by four-node shell elements, the tendons by truss elements.
Special contact elements are used to transfer compression and friction forces across the joints [4].
The joint may open under any tensile stresses. The shear keys between the two webs in the upper
slab of a segment were assumed to have no contact and no frictions with the opposite shear keys in
a adjacent segment due to the existence of a gap between them. The numerical model is verified by
a full-scale test which was carried out by Takebayashi et. al. [7]. The results had been published by
Specker [5].
172.5
50 1022.5 1020 1360 1022.5
No. 3 No. 8
No. 1,2 No. 7
No. 4
No. 5
No. 6
4525cm
(12 segments à 340cm + 2 segments à 172.5cm )
Standard Segment
20
40
20
35
D2 : 18
D3 : 20
<150 variable 75 D2 : 370cm
D3 : 550cm
Fig. 10 Segmental bridge
The stresses in longitudinal direction are significant influenced by the bowing of the segments.
Figure 11 shows the stresses in the top slab for perfect fitting segments and a superstructure with
bowed segments (1 mm gap, maximum value).
A) Top view of the segmental bridge without bowing effect
B) Segmental bridge with segments - 1 mm gap
Fig. 11 Longitudinal stresses σx (MPa) in the top slab of a segmental bridge
IABSE Symposium „Metropolitan Habitats and Infrastructure“, Shanghai, 22.-24.09.2004
The maximum compressive stresses in the web/slab junction is nearly doubled (Fig. 12) due to the
1 mm gap (maximum value) at the outer edge of the deck slab. On the other hand the stress in the
top slab between the webs is nearly decreased up to zero. This may result in serviceability
problems.
Furthermore the maximum deflection of the bridge under full loading is increased by the bowing of
the segments from 0.343 m (‘perfect’ segments) to 0.370 m (bowed segments).
34.3 34.3
segments with
1 mm gap (max)
18.4 18.4
21.4 21.4
14.0 14.0
perfect
fitting
segments
Fig. 12 Comparison of longitudinal stresses σx (MPa) in the joint between perfect fitting segments
and segments having a 1 mm gap (midspan)
6 Conclusions
A finite element model has been developed and verified by experimental data to study the bowing
effect of match cast segments on the behaviour of segmental bridges with dry joints. A significant
increase in the longitudinal stresses had been obtained with segments having a gap of 1 mm only.
Further parametric studies will be conducted. The ongoing investigations will concentrate on the
load bearing capacity of the structure.
7 References
[1] ROBERTS C., BREEN J., KREGER M.: Measurement Based revisions for Segmental Bridge
Design and Construction Criteria. Texas Department of Transportation, Report no.1234-3F,
1993, pp. 303-320
[2] LOGAN D.: A First Course in the Finite Element Method. Second edition, PWS-KENT
Publishing Company, USA, 1992
[3] ZEITLER W.: Untersuchungen zu Temperatur- und Spannungszuständen in Betonbauteilen
infolge Hydratation. Dissertation, Technische Hochschule Darmstadt, 1983, pp. 15-40
[4] ROMBACH G.: Bangkok Expressway - Segmentbrückenbau contra Verkehrschaos. Aus dem
Massivbau und seinem Umfeld (Hilsdorf, Kobler ed.), Schriftenreihe des Institutes für
Massivbau und Baustofftechnologie, Universität Karlsruhe 1995, pp. 645-656
[5] SPECKER A. :Der Einfluss der Fugen auf die Querkraft- und Torsionstragfähigkeit extern
Vorgespannter Segmentbrücken. Thesis, University of Hamburg-Harburg, 2001
[6] ROMBACH G., SPECKER A.: Finite Element Analysis of Externally Prestressed Segmental
Bridges. EM 2000 Conference, Austin, 21.-24.05.2000
[7] TAKEBAYASHI T., DEEPRASERTWONG K., LEUNG Y.: A full-scale destructive test of a
precast segmental box girder bridge with dry joints and external tendons. Proceedings of the
Institution of a Civil Engineering, August 1994, pp. 297-315
