## Services on Demand

## Article

## Indicators

## Related links

- Cited by Google
- Similars in Google

## Share

## Journal of the South African Institution of Civil Engineering

##
*On-line version* ISSN 2309-8775

*Print version* ISSN 1021-2019

### J. S. Afr. Inst. Civ. Eng. vol.59 n.2 Midrand Jun. 2017

#### http://dx.doi.org/10.17159/2309-8775/2017/v59n2a2

**TECHNICAL PAPER**

**Fatigue behaviour in full-scale laboratory tests of a composite deck slab with PBL reinforcement**

**P Lu; X Zhan; R Zhao**

**ABSTRACT**

Studies on the fatigue behaviour of composite deck slabs are relatively few. To assess the fatigue performance of a composite deck slab at specific design loads, and to provide a reference for its design in fatigue, two full-scale models A and B of a composite deck slab were developed, comprising steel plates and steel-fibre-reinforced concrete slabs. These models will be a useful reference for experiments and design, and for developing codes. In this study we carried out fatigue experiments and focused on the fatigue performance of a composite deck slab in a column area, and the positive and negative bending moments. For the entire fatigue loading cycle, the overall performance of Models A and B was good, the overall stiffness of the composite deck slab was rarely attenuated, the stress levels in the steel members in relation to the fracture strength were not significant, and the steel member was in the stage of flexible work. Through comprehensive tests of Models A and B it was found that the original design exhibits good fatigue performance and meets the design requirements. The research results provide a basis for the design of a composite bridge deck slab in fatigue.

**Keywords:** steel truss arch bridge, steel-concrete composite deck slab, fatigue performance

**INTRODUCTION**

Perforated-stiffened-plate composite bridge deck slabs are a novel type of composite bridge deck slab. At the bottom sides of the slab, steel plates and concrete are connected through a perforated, stiffened plate that is known as a perfobond (PBL) shear connector. A perforated-stiffened-plate composite bridge deck slab combines the advantages of steel and concrete in a general composite deck slab. For bridge construction, a steel plate is used as permanent formwork. This design saves time during the installation and removal of the scaffolding setup formwork, and reduces the costs in relation to the construction cost of the whole steel beam structure. The steel plate and concrete slab are connected by PBL shear connectors to improve the composite effect of the slab, enhance the slab's stiffness and mechanical performance, and reduce deformation and slippage. The composite bridge deck slab - in conjunction with PBL shear connectors - is widely used in the construction of new bridges, maintenance activities and the reconstruction of existing bridges. Moreover, the composite slab has technical and economic advantages, and improves material performance and fast construction compared to other promising materials. However, despite the increasing use of PBL-shear-connected steel and concrete composite slabs, the bridges in which they are used are not immune to the long-term effects of moving loads. In particular, compared with a stud shear connector, a PBL shear connector is more prone to fatigue damage and failure. Composite deck slab structures are still in their early stages, as they have not reached the ends of their life cycles yet; thus, they have not yet exhibited the effects of fatigue loading. Since the more serious effects of fatigue loading have yet to appear, the fatigue problem has not been sufficiently addressed, nor systematically researched. Fatigue damage and failure decrease the reinforcement provided by the steel in a new bridge deck, which reduces the superstructure stiffness while increasing slipping. These effects seriously affect the vehicle capacity of the bridge structure. Therefore, an assessment of the fatigue performance of a PBL-shear-connected steel and concrete composite slab at highway loads has important academic and design potential, and will serve as a basis for future strategies for preventing or reducing damage due to fatigue loading.

Few studies of perforated-stiffened-plate deck slabs have been published. Ryu *et al *(2007) conducted a full-scale model test of a two-span continuous bridge deck slab to study the crack development of a fatigue-loaded composite deck slab. The results indicated that the cracks in a region with a negative bending moment were controlled within an allowable crack width for a specific fatigue load. The steel shuttering described in their paper was a profiled sheet; compared with steel-concrete bridge deck slabs, the perforated stiffened plates had a different arrangement. To verify the composite effect of the steel shuttering and the perforated steel plate combined with concrete, a test of a full-scale model of a perforated-stiffened plate deck slab was conducted by Kim and Jeong (2006). Their results showed that the perforated and stiffened plate effectively enhanced the composite effect. Their study primarily analysed the effect of the perforated steel plate for the entire composite bridge deck slab. Leitão (2011) developed a numerical model for the dynamic analysis of composite highway bridges using a finite-element method with mesh-refining technology, and achieved satisfactory results when assessing fatigue behaviour. To evaluate cracking in the vertical and horizontal joints of a composite bridge deck, Chang and Shim (2001) performed fatigue behaviour tests of the composite connection points of a continuous composite bridge. Their study also discussed methods of longitudinal pre-stressing. Although they did not conduct a thorough analysis of the effects of fatigue at the connection points on the entire composite bridge deck slab, their results provide a reference for the study of the fatigue behaviour of perforated steel plates, including the junctions of the underside and side steel plates. Allahyari *et al *(2014) investigated the behaviour of bridge decks for a static load applied to the centre of the deck. To evaluate the dynamic properties of the decks, they experimentally investigated the dynamic properties of exodermic bridge decks with alternative PBL shear connectors. Millanes *et al *(2014) investigated the design of a composite steel-concrete deck for a long railway bridge. Gara *et al *(2013) investigated the effectiveness of various casting techniques used to control the tensile stresses in the slab during the construction of continuous steel and concrete composite bridge decks. Leitão *et al *(2013) carried out a fatigue analysis and predicted the lifetime of composite highway bridge decks under traffic loading conditions.

Wang and Jiang (2007) reviewed the fatigue problem of composite structures and the direction of future studies on composite structures. They argued that future studies should continue to focus on the mechanism of fatigue damage accumulation, damage identification, and fatigue reliability. Yang *et al *(2012) performed mechanical performance tests of a fatigue-loaded composite beam of a composite bridge deck. Their results indicated that the fatigue failure mode of the composite beam specimens with a positive bending moment was the crushing of the concrete in the compression zone resulting from fatigue damage to the lowest steel beam. The fatigue life was directly related to the range of fatigue stresses; however, the upper and lower limits of fatigue loading had little effect on the fatigue life. Liu *et al *(2012) proposed a new composite bridge deck system and performed a theoretical analysis including a full-scale model test of a bridge deck. Their results showed that after 2 million cycles of fatigue loading at a load frequency of 5 Hz, no new cracks appeared in the composite deck, nor had the initial cracks expanded. Additionally, the stiffness of the test beam had not decreased, indicating that this composite bridge deck system had satisfactory fatigue properties. Zong and Che (2000) performed a fatigue test on a simply-supported continuous pre-stressed composite beam. They also analysed pre-stressed composite beams constructed from different types of concrete at different magnitudes and orders of pre-stressing. They drew some valuable conclusions by summarising the fatigue test results of the pre-stressed composite beams, discussing the maintenance of the steel members and stud connectors, and proposing a principle for the corresponding fatigue strengths.

For the fatigue of steel-concrete composite bridge decks, the theoretical analyses and experimental results provided by the above-mentioned studies are in agreement regarding fatigue-loaded steel-concrete composite bridge decks. However, the individual behaviour of various combinations of steel-concrete composite bridge decks is quite different, i.e. different combinations of materials and forms lead to different forms of fatigue behaviour. Additionally, fatigue test results are discrete. Therefore, many tests are needed to explore the mechanical behaviour of perforated, stiffened composite slabs. There are relatively few studies of fatigue performance and the corresponding design for novel composite bridge deck slabs. Many theoretical and technical problems still have not been solved, such as the failure mechanism, mechanical behaviour, load-bearing capacity, and deformation of novel composite bridge deck slabs under fatigue-loading conditions. Therefore, considering the variety of steel-concrete bridge decks and fatigue problems, the present study aims to evaluate the fatigue performance of composite bridge decks under highway loading conditions using the Dongping Bridge in Guangdong as a case study.

In this study, we conducted a model test and simulation analysis to evaluate the fatigue performance of a perforated stiffened deck slab at a highway load. We also propose some important indicators of fatigue performance, provide a reference for composite materials and forms, and assess their superior mechanical behaviour to validate the effectiveness of the proposed deck system for bridge applications. Our study contributes to the understanding of fatigue-induced damage or failure of bridge deck structures. The results of this study have important theoretical value and practical value for the optimisation, design and theoretical analysis of composite deck slabs.

**TEST MODEL**

**Engineering background for the fatigue tests**

The main span of the Dongping Bridge (43.5 m + 95.5 m + 300 m + 95.5 m + 43.5 m) in Guangdong is a half-through steel truss arch bridge with a full length of 1 322.2 m. The steel boxes and concrete slabs are connected by PBL shear connectors. The grid beams consist of three main longitudinal girders, secondary longitudinal girders, main beams and secondary beams. Perforated-stiffened-plate composite bridge deck slabs were erected on the grid beams. Each slab has a minimum thickness of 12 cm and a maximum thickness of 20 cm. The overall layout of the Dongping Bridge and the testing zone of the composite bridge deck are shown in Figure 1.

The Dongping Bridge uses a composite deck-binding system with a space grillage design. Under the sustained action of moving loads, the longitudinal shear performance of the perforated and stiffened composite bridge deck slabs will be significantly degraded, and the degradation in the performance of the PBL shear connectors will reduce the composite effect of the composite bridge deck slabs. As a consequence, the load-bearing capacity and stiffness of the composite bridge deck may be reduced, which will affect the mechanical behaviour. Under such conditions, fatigue failure of the composite bridge decks will occur once the fatigue damage has escalated to a certain point. To assess the fatigue performance of the composite decks of the Dongping Bridge, we performed fatigue experiments and a numerical simulation in an area comprising positive and negative bending moments.

**Actual bridge model selection**

Domestic and international research indicates that the stress amplitude Δσ and the number of cycles *N *are the predominant factors affecting fatigue strength. Considering that the bridge deck span in the column area is larger, the most unfavourable girder beam in the bridge deck in the column area was selected for the analytical model of an actual bridge. Owing to the large span of the column area of the bridge, according to Saint-Venant's principle to relax the boundary condition, the longitudinal range of the column area and the transverse range between two main girders were selected as the objects of study for a simplified analysis problem. The focus was the fatigue performance of the composite deck slab in a column area comprising positive and negative bending moments. The elevation and plan of the model structure described above are shown in Figures 2(a) and 2(b).

**Design of the experimental model**

According to the study topics, two types of models - denoted as Models A and B - were constructed. Model A simulated the region of the bridge structure with a negative bending moment, which includes a secondary beam and concrete slabs. Model B simulated the region of the bridge structure with a positive bending moment, which includes the area between two beams. Models A and B were constructed to perform full-scale-model tests of the composite decks of the Dongping Bridge.

The two models were constructed using structural dimensions that were consistent with the actual bridge structure, except for the plate length *L *and plate width B. The geometry of Model B includes *L *and *B *as the objects of study. According to the experimental requirements and conditions, *L *= 5 000 mm and 6 000 mm for Models A and B respectively. The plate widths for Models A and B were selected according to a formula for the effective width of a concrete composite beam slab (JTT 1986; BS 1980). The calculated span of the actual bridge was 8 000 mm, the centre-to-centre distance of the adjacent secondary beams was 3 325 mm, and the top flange of the secondary beams was 706 mm.

The calculated width of the concrete slab was determined in accordance with the comparative results of the following four principles:

■ The first principle is based on the calculated width of the concrete slab, the sum of the width of the top flange and 12 times the thickness of the flange, and the minimum value of one of the three following items: a third of the span, the clear distance between two subbeam plate brackets, or 12 times the thickness of the top flange of the subbeam.

■ The second principle is based on the minimum value of one of the three following items: a third of the span, the centre distance of two subbeams, or the sum of the width of the top flange and 12 times the thickness of the top flange of the subbeam.

■ The third principle is in accordance with the CP117 specification based on the minimum value of the following three values: a third of the span, the centre distance of two subbeams, or the sum of the top width of the plate bracket and 12 times the flange thickness.

■ The fourth principle is in accordance with the BS 5400 specifications, with which the size of the model design is calculated.

According to comparative results of the four principles, the size of Model B was determined to be 2 146 mm. The size of Model A is 2 400 mm x 5 000 mm, and the size of Model B is 2 400 mm x 6 000 mm. The cross-sections of Models A and B and the loading positions are shown in Figures 2(c) and 2(d).

**DESIGN OF THE FATIGUE MODEL TEST**

**Determination and theoretical analysis of fatigue loads**

The stress conditions of the actual concrete plate were the focus of the present study. First, we ensured that the stress conditions of the top concrete plate were consistent with those of an actual situation when we determined the loading scheme of the test model by theoretical calculations. The stress loading was determined by theoretical calculation, hence the stress conditions of the concrete were verified for consistency with the actual situation. To align the stress conditions of the model with the actual structure, the upper limit load *P *and its loading location were determined. Vehicle loads of 20 t, 30 t and 55 t were used to calculate the fatigue vehicle loading with an influence line. Considering the effect of the deadweight of the bridge deck pavement, the magnitude of the stress at the lower limit fatigue load should be consistent with the bridge deck pavement of the actual bridge structure.

The theoretical results of the model test were obtained using an ANSYS finite-element model comprising spatial deck elements and solid elements. The model was based on 180 455 nodes and 207 65 elements. According to Table 1, the applied fatigue vehicle load was determined and the impact coefficient *a *= 0.295 was used. In the transverse direction of the bridge, the most unfavourable loading position was used, and the position of the vertical loading was based on the influence line of the control section of the four-span continuous beams. The finite-element model of Model B consisted of 31 250 nodes and 32 585 elements. The loading scheme of Model B was not determined until the stress conditions of the concrete slab were consistent with the actual bridge conditions. Acc