STR - 64

EXPERIMENTAL STUDY ON LINK SLAB FOR HIGHWAY GIRDERS

Tayagorn Charuchaimontri1

Ekasit Limsuwan2

1Ph.D.Student, Department of Civil Engineering, Chulalongkorn University, Bangkok 10330, Thailand,

Chanwut.C@student.netserv.chula.ac.th 2Professor, Department of Civil Engineering, Chulalongkorn University, Bangkok 10330, Thailand ,

fceels@kankrow.eng.chula.ac.th ABSTRACT : Most of highway constructions in Thailand are prefabricated construction comprising precast members incorporated with cast in-situ slab deck. Continuity of the deck can be treated to eliminate gaps between adjacent spans providing smooth riding with link slab. The link slab would accommodate all movements into the structure by means of axial deformation, rotation and translation can be interaction of structural behavior between adjacent spans. Therefore, in this paper the behavior of link slab with lap reinforcement under cyclic loading was observed for crack distribution, crack width, load-deflection relationship and ultimate strength through experimental work considering variable length of lap reinforcement which can be classified to 3 types of detailing. All specimens failed in shear mode but have different cracking behavior. Also numerical results by means of the 3-dimensional nonlinear finite element method using the microplane model (MASA3) and truss model were compared with those of experimental results. The results would be the principal parameters for further design approach. KEYWORDS : Link slab, Highway girder, Reinforcement detailing, Nonlinear finite element method 1. Inroduction Most of highway construction are now prefabricated construction comprising precast members, fabricated on site and then provide cast-in-situ slab deck on site. The precast girder is particularly advantageous to permit large volume of production in short period, uniformed and systematic quality control, and erection with little or no interference with the traffic. There are many types and sections of precast construction, the most common used in Thailand are I, T, U section and the box-section which normally practise with their appropriate span length and erection equipment. In current practice, each span is usually constructed to separate for individual span and provide expansion joint for each single span. Continuity of bridge deck to eliminate gaps between adjacent spans supports is desirable to improve appearance and riding quality, to reduce vibration and noise problems, to improve long term serviceability, aesthetics, and safety, to prevent drainage of deck surface through joint as for improving durability, to eliminate initial cost of joints and their subsequent maintenance, and to improve overall safety as continuous structure where redistribution of moments can be accented. However, there still be some limitation depended on many factors affecting girder movements due to loads and the serviceability. In recent years, expansion joints have been eliminated to reduce its number throughout total length of overall construction especially for new bridges. The joints can be eliminated by providing continuous deck as which girders can still be simply support and the joint can be eliminated by proper measures to encounter among complicated interactions in the region. The section of the deck

connecting the two adjacent simple-span girders is called link slab as shown in fig 1. The criteria of providing link slab as continuous deck would accommodate all movements into the structure by means of translation, rotation, deformation and settlement as boundary condition counteract with structural behavior of adjacent spans. Those constraints will be affected not only for principal behaviors but also the secondary stresses due to thermal, moisture, settlement, post-tensioning, and some others. Those conditions are complex in the behaviors which are related to the accuracies and its degree of reliability on strength, serviceability and durability. Link slab problems were produced not only by movement of adjacent girders resulting from prestressing force, loads, creep, shrinkage, thermal effect, foundation settlement, and bearing pad effect, but also by link slab such as wheel load. So the theoretical model to study this problem must be modeled the boundary condition and cause of link slab. From these effects, the link slab behavior will be non-linear caused by cracking and material properties which must be taken care in the study. The link slab is assumed to be flexible in comparison with end stiffness of the girders. Both ends of the slab are subjected to vertical and longitudinal movement due to the modulus of elastomeric bearing pad and also subjected to longitudinal and rotational movement due to girder movements. 2. Experimental works The behavior of link slab under cyclic loading was observed for crack distribution, crack width, load-deflection relationship and ultimate strength through

การประชุมวิชาการวิศวกรรมโยธาแหงชาติครัง้ท่ี ๑๐ ชลบุรี ๒ – ๔ พฤษภาคม ๒๕๔๘

STR - 65

experimental work considering variable length of lap reinforcement which can be classified to 3 types of detailing. 2.1 Test Specimen In order to study the influence of reinforcement detailing, 3 RC link slab specimens were tested. The specimens were designed to be possible for construction with a span of 2,000 mm and slab thickness of 200 mm (Fig 2). The support of link slab is designed to be fixed because the stiffness of girder is much more than stiffness of link slab and it is designed to be produced both positive and negative moments in link slab in order to simulate the behavior of link slab which has to resist both moments. The measured dimensions of specimens are summarized in table 1. The length of lap reinforcement was parameters in this test and the specimens’ detailing and name were shown in Fig 2 and table 1. The link slab specimens were tested under point load which is applied using a survo-pulser multi-jack actuators of 100 tons capacity. The load is transmitted through a rectangular steel plate of size 50 cm x 20 cm x 2 cm to the link slab in order to represent the AASHTO HS20-44 standard truck wheel load. 2.2 Materials In this study, ready-mixed-concrete was used and the characteristic compressive strength of concrete is 300 ksc. The used reinforcementS were DB16 and DB20 of which yield strength were 4000 ksc. Twenty eight days strength of concrete was measured using the standard specimen for casting concrete as shown in Table 1. Tests were conducted at 28 days after casting. 2.3 Test procedure and Measurements The overview of the test setup is shown in fig 2. Electrical resistance strain gages of length 0.5 cm are fixed across the mid section and edge section of link slab on both top and bottom reinforcements in order to measure the tensile and compressive strains. Out-of-plane deflections at three points along the mid-section, at one-sixth span and at supports are measured by using linear variable deflection transducers (LVDT). During testing, LVDT, load cell and strain gauges signals were input to a computerized data acquisition system. The specimens were initially loaded gradually up to 1 ton and then the load is released in order to ensure the loading edges remained in proper contact with the specimen. Then the specimens are cyclic loaded in the first range before and after cracking to observed cracking load and tension stiffening behavior. Next, they are loaded from zero to 9.3 tons and released back to zero and repeated the cycle five times in order to establish a base line curve. The 9.3 tons load represents the factored load for the AASHTO standard HS20-44 truck wheel load (7.1 tons + 30% for impact). Finally they are loaded to failure.

2.4 Results The first cracks of all specimens were flexural cracks which occurred at supports and mid span. Fig 4 shows the crack patterns and failure modes of all specimens. Table 2 summarized main test results on cracking loads, crack sizes and failure loads/displacements. The maximum deflections under cyclic factored wheel load were the same. Fig 5 shows the relationship between mid-span load and vertical displacement of link slab specimens. In the figure, it could be confirm that the detailing of link slab affected the stiffness, crack distribution, crack width. The magnitude of the stiffness was increased as the amount of reinforcement in the specimen was increased and force distribution of structure. LS_000_S specimen acted as a cantilever slab after mid-span cracking so the mid-span deflection of this specimen is more than LS_025_S and LS_183_D respectively. LS_025_S specimen acted as semi-cantilever slab which can be observed from the out-of-plane deflections at three points along the mid-section, at one-sixth span and at supports. The stress in reinforcements are not yielded before shear failure. Shear crack was started from mid depth and then crack was immediately propagated through its depth until failure. 3. Numerical results by means of nonlinear finite element[5] The used program MASA3 has been developed by Josko Ožbolt at the Institute of Construction Materials, University of Stuttgart. MASA is an abbreviation of Macroscopic Space Analysis. The finite element code of MASA is based on the microplane model, and possible to apply to 2-dimensional and 3-dimensional analysis of quasi-brittle materials[4]. The smeared crack approach is employed and the constant stiffness method (CSM) is applied as a root finding method. Furthermore, the crack band approach is used in order to be independent of mesh dependency. In the numerical model concrete is modeled with solid elements and reinforcement is modeled with bar and beam element. Taking the symmetry into account, only ¼ of the specimen has to be modeled. The load displacement relationship and crack patterns results from MASA3 is shown in fig 6,7. From the numerical results it shows a good results of crack patterns and failure loads compared to testing results as shown in table 3. 4. Analytical results by means of truss model In reinforced concrete slab without transverse reinforcement, the structural behaviour is satisfactorily explained by the “tooth model”. The shear force is mainly transferred in the cracked tension zone of the member by the combined action of the friction along the crack faces and the dowel force of the longitudinal reinforcement. From the stress field between the cracks is governed by an inclined biaxial tension-compression field in the concrete. It is well represented by the simple truss model[6]. The allowable tensile strength for concrete ties is

between /c0.25 f (MPa) and /

c0.42 f (MPa) depended on

STR - 66

cracking in ties. Using truss model and failure load from testing, the tensile strength for concrete ties are

/c0.26 f (MPa) , /

c0.30 f (MPa) and /c0.28 f (MPa) for

LS_000_S, LS_025_S, LS_189_D respectively which is in the range shown above. 5. Conclusion [1] The different reinforcement detailings of link slab which are possible for construction is affected different force distribution, stiffness, crack distribution and crack width which is served for different action from end girder boundary and different attempt to maintain link slab. [2] However all of the specimens are failed by shear mode, the link slab design is based on serviceability and durability which is deformation control and crack control. [3] Numerical results can estimate very good results of crack pattern and load-deflection curve compared to test result. [4] Ultimate load of link slab is shear strength of slab without transverse reinforcement which can be predicted from FEM and STM. Acknowledgement We would like to thank TRF (Thailand Research Fund) for the financial support of the Ph.D. study in Chulalongkorn University and DAAD (Deutscher Akademischer Austausch Dienst) for financial support of the 6-months study at the University of Stuttgart, Germany.

References [1] Charuchaimontri C., Limsuwan E., May 19-21, 2004. End Movement of Highway Girders, 9th National Convention on Civil Engineering, v.1, pp. invited1-6. [2] Issam Harik, et al., May 23-27, 1999. Static Testing on FRP Bridge Deck Panel, 44th International SAMPE Symposium, pp. 1643-1654. [3] HyungKeun Ryu, SungPil Chang, YoungJin Kim and BongChul Joo, December 16-18, 2003. Experimental Works on Precast Concrete Decks with Loop Joints, 9th EASEC, pp. RCS7-13. [4] Ožbolt, J.: MASA3, Finite element program for 3D nonlinear analysis structures, MASA manual, Institute for Construction Materials, University of Stuttgart, 2002. [5] Charuchaimontri C., Limsuwan E., December 13-15, 2003. Numerical Studies on Detailing of Link Slab for Highway Girder Considering Cracking Behavior, 16th KKCNN symposium on civil engineering, pp. 325-330. [6] Karl-Heinz Reineck, 1991. Ultimate Shear Force of Structural Concrete Members without Transverse Reinforcement Derived from a Mechanical Model, ACI Structural Journal, v.88 No.5: pp.592-602.

Figure 2 Test specimen (LVDT and load arrangement)

Figure 1 Link slab in highway girders

STR - 67

Average thickness (cm)

Level of upper/lower reinforcement (cm) Name

L M R L M R

Average Concrete Strength

(ksc)

Lap Length (cm)

Top reinforcement

Bottom reinforcement

LS_000_S 20.2 20.3 20.2

16.2/3.2

16.8/3.5

16.4/3.2 233 0 DB16@10 DB16@10

LS_025_S 20.0 20.0 20.5

16.5/3.1

16.5/3.8

16.0/3.1 261 25 DB16@10 DB16@10

LS_183_D 20.3 20.5 20.3

16.8/3.0

16.9/3.3

16.9/3.2 333 183 DB20@10 DB16@10

Load(T) at crack width(mm) Failure load (T)/displacement (mm) Specimens Cracking load

(T) 0.2 0.3 0.4 testing FEM

LS_000_S 3.0 6.0 9.0 15.0 63.59 / 10.89 65.52 / 6.86

LS_025_S 11.0 30.0 40.0 70.0 80.96 / 7.28 90.82 / 6.69

LS_183_D 12.0 40.0 70.0 * 77.78 / 3.73 82.47 / 4.52

Table 1 Test specimens

Table 2 Test result, FEM result and STM result

Figure 3 Specimens’ reinforcement detailing: LS_000_S, LS_025_S, LS_183_D

Note: * failure before this crack width

STR - 68

Figure 4 crack pattern and failure mode of LS_000_S, LS_025_S, LS_183_D

Load-Displacement at midspan

0102030405060708090

100

0 2 4 6 8 10 12Displacement (mm)

Load

(Ton)

LS_000_S

LS_025_S

LS_183_D

Figure 5 Load-displacement relationship from testing

STR - 69

Figure 6 Crack pattern before failure (left) and at failure (right) of LS_000_S, LS_025_S, LS_183_D

Load-Displacement at midspan

0102030405060708090

100

0 2 4 6 8 10 12Displacement (mm)

Load

(Ton)

LS_000_S

LS_025_S

LS_183_D

Figure 7 Load-displacement relationship from finite element method