As traffic demand on bridges increases, loading cycles on critical components will increase, reducing their service life. Modular bridge expansion joints, which are imperative to allowing the bridge superstructure to move, are susceptible to fatigue damage at their field splice. These splices are used to connect segments of the total joint, during staged construction. Current splice designs are either bolted or welded connections, which allow stress concentrations to induce pre-mature fatigue failure. This thesis examines the use of a hybrid FRP/steel design under fatigue loading for use as a splice detail. The splice detail consists of steel plates bolted to steel beam webs and CFRP pultruded plates adhesively bonded to the underside of the steel beam flanges. Two different moduli of CFRP were examined: Normal Modulus and Ultra High Modulus. Two beams of each modulus were tested under static conditions and six under constant amplitude fatigue loading. A testing rig was used to simulate similar bending moments experienced in bridge joints. In the static tests, slippage of the web plates caused considerable stiffness loss and the slippage load varied drastically between CFRP moduli. For the fatigue tests, the intention was to reach two million cycles at the different constant load ranges. Stiffness degradation was noticed during the fatigue process, and was likely due to bolt pre-tension loss and/or plastic deformation of the adhesive. Specimens that reached two million cycles were monotonically loaded to failure. Once the CFRP had failed, a secondary mechanism was observed for reserve load capacity. Simple beam mechanics were used to create prediction models for the initial spliced beam stiffness and peak CFRP load. Flexural and shear deformations of the spliced system were considered for beam stiffness. For the CFRP failure load prediction, a design peak strain in the CFRP was used to account for shear lag effects in the material and variability of the splice detail. While the model was inaccurate for beam stiffness, it provided a good approximate of the peak CFRP load. Based on the presented test data, the Normal Modulus CFRP hybrid splice detail showed better fatigue performance than conventional steel connection details. / Thesis (Master, Civil Engineering) -- Queen's University, 2013-07-24 11:28:19.728

Bridge expansion joints are directly subjected to traffic load, and thus prone to premature fatigue failure. Replacement of components such as modular bridge expansion joints is typically done in a staggered schedule to minimize traffic blockage. Field splices are used to connect the successively installed segments. These splices typically include a combination of field welding or bolting, and experience has shown that they often fail due to fatigue cracking. This thesis reports the investigation of hybrid FRP/steel splice details that avoid the use of field welding. Two configurations have been examined: A GFRP pultruded square tube section, adhesively bonded to the soffit of the spliced beam, consists the moment resisting component in one configuration, whereas the other takes advantage of two series of FRP plates for this purpose. Bolted steel plates splice the beam through web in both cases. The behaviour of these details has been studied extensively under vertical static loads. The effect of several parameters including bond length, FRP end shape, bond surface treatment, adhesive, etc. for each detail has been investigated. A three-dimensional, non-linear finite element model has been developed for each detail and validated using the experimental results. The bond strength of two adhesives was investigated experimentally using double shear lap splice tests. A new method is proposed to analyze the strength of the splice details. This method is based on the results obtained from shear lap splice tests and the verified finite element model developed for the splice detail. The finite element model could thus be used for further parametric studies. More experiments, however, are statistically required before using this model with confidence. The fatigue behaviour of one of the promising splice details has been investigated both experimentally and numerically. A special fatigue test set-up has been designed and used successfully for this purpose. Two fatigue tests to 1,000,000 cycles were run. One failed at 719, 347 cycles and the other survived 1,000,000 cycles. The predicted fatigue life as per the developed model was 871,840 cycles. More experiments are required to understand the fatigue behaviour of the splice detail under various stress ranges. / Thesis (Ph.D, Civil Engineering) -- Queen's University, 2011-11-30 16:53:07.385