In recent years, a growing amount of movable highway bridges with a timber deck are approaching the end of their service life. Since these bridges have opened, the traffic intensity has increased yearly and environmental effects had the opportunity to deteriorate the bridges slowly. Inspections revealed that the main load-carrying elements are generally in adequate conditions while the deck structure is deteriorated. In an effort to extend the bridge's lifetime, fibre-reinforced polymer (FRP) decks can be used to replace deteriorated members. These decks have gained popularity over the past years due to their high strength-to-weight and stiffness-to-weight ratio’s which enables fast installation. Furthermore, their freedom in design, the ability to be prefabricated and good durability are all characteristics that make FRP bridge decks a worthy competitor for the refurbishment of movable bridge decks. The main challenge of applying FRP bridge decks is connecting it with the underlying steel girders. Fortunately, recent developments of deck-to-girder connections have suggested a slip-resistant demountable shear connection, namely, the injected Steel Reinforced Resin (iSRR) connection. Initial experiments of the iSRR connection concluded that the connection provides excellent fatigue behaviour which is favourable in hybrid FRP-steel bridges. The main objective of this thesis is to investigate the behaviour of bolted deck-to-girder connections in FRP-steel composite bridges with the application of renovating movable highway bridges with timber decks. Using a series of finite element models, the application of a bolted FRP deck is investigated on an existing movable highway bridge, namely, the Beneden Merwede bridge. In order to achieve interaction between the existing steel structure and FRP panels, the slip-resistant iSRR connections are considered. Starting from a small-scale finite element model of the iSRR connection, a process of simplifications and upscaling of the models will lead to a large-scale bridge model. Firstly, a highly detailed model of a push-out test is developed to assess the static performance of the iSRR connector with embedded nuts and a coupler system. Secondly, the mechanical performance of an FRP-steel composite system is investigated in a composite beam model. Using the latter model, different modelling techniques of the deck-to-girder connections are compared in order to propose an efficient technique to determine the forces in the deck-to-girder connections. Finally, a large-scale bridge model is used to determine the forces in the deck-to-girder connections under a number of load cases and to provide a recommendation for the distribution of the connections. The analyses of the iSRR connection with the embedded nuts and coupler show that both configurations have a similar static behaviour characterised by the shear rupture of the connection and limited damage to the FRP. Based on the known long-term behaviour of the iSRR connection with the embedded nuts, it is advised to use this type of configuration as deck-to-girder joint. Applying the connection on the composite beam model every 600 mm, the iSRR connections provide full composite action between the FRP deck and the steel girder and preserves the interaction between the two components until yielding of the steel beam is initiated. The composite beam analysis also concludes that connection forces may be reduced when friction is not considered during the analysis, e.g. during engineering analysis with linear springs (212 kN/mm) in GSA. Investigating the deck-to-girder connections on the bridge level, it is advised to install iSRR connections on the cross- and main-girders to connect the FRP deck. In such manner, the deck is well supported in longitudinal and transverse direction of the bridge and local force concentrations near the cross- to main-girder connection are prevented. When connecting the FRP panel to the cross- and main-girders, the bending moments in the cross- to main-girder connections are significantly reduced compared to the original design of the bridge with a timber deck.

The use of fibre reinforced polymers (FRP) are stimulated in the construction industry for their remarkable corrosion resistance, good thermo-mechanical properties, high strength-to-weight ratio and high stiffness-to-weight ratio. The open-hole strength of FRP is an important parameter in a composite design because it can be representative in simple features of composite components leading to stress concentrations such as cutouts, fastener holes or a flaw in the material. Therefore, an open-hole tension test is simulated in the finite element program ABAQUS to determine the failure load of an [45°/−45°] lay-up using the material AS4/PEEK. In ABAQUS, both static implicit and dynamic explicit analysis are performed. For both analyses, the failure loads, stability and failure mechanisms are examined. The geometry and material properties are chosen to match the properties used in van der Meer and Sluys in order to compare results. The model is build up of 8808 SC8R elements and the interface between the plies is modelled using the surface-based cohesive behaviour technique. The static and dynamic analysis revealed that there’s a large influence of non physical and time related variables on the simulated failure load of the laminate. The results of the dynamic target time step sensitivity analysis didn’t converge to a legitimate value for the failure loads. The same has been concluded for the mesh size sensitivity analysis. Based on the stability of the solutions a reasonable value has been chosen for these variables. Comparing the results of the static and dynamic analysis with the results of van der Meer and Sluys, illustrated that the failure loads differ largely. The failure mechanisms are different due to the homogenisation of the material that is fundamental to continuum models in ABAQUS. At the failure load in the static and dynamic analysis, the matrix fails under tension and shear through the whole thickness of the laminate in the same direction. This unrealistic event differs from the failure mechanism observed in van der Meer and Sluys. In order to overcome the unrealistic failure behaviour in the static and dynamic analyses, XFEM modelling techniques are recommended to bridge the problem with the damage propagation direction parallel to the fibre direction.