EXPERIMENTAL STUDY OF HOLLOW FRP SECTIONS UNDER BENDING FOR USE AS BRIDGE DECKING

Abstract

In this thesis, the behaviour of square hollow structural sections made of glass fiber reinforced polymers (GFRP) under four-point bending both independently and in composite with concrete was investigated. The end goal of the research was to determine if the systems tested could be viable for use as bridge decking. Hollow specimens of three different lengths, 1 m, 2 m, and 3 m were originally tested and deemed too flexible to be used in a bridge decking system. The main portion of testing was conducted on 18 FRP specimens with concrete slabs of different thicknesses cast overtop. Three slab thicknesses were used: 35 mm, 55 mm, and 75 mm to simulate three different cases of neutral axis location. Half of the slabs were bonded to the FRP by first applying a layer of epoxy paste to the top face of the FRP and covering it with small aggregate before casting the concrete. The other half had bolts drilled into the specimens at 100 mm intervals along the specimen before concrete was poured. It was found that the addition of concrete decreased overall strength, as it governed failure over the FRP in all slab specimens tested. The concrete gave an increase in stiffness between 62 and 78% for the epoxy and aggregate bonded specimens, and an increase between 54 and 79% for bolt bonded specimens. Two analytical models were created to predict the load-deflection behaviour of the specimens. A linear-elastic model was developed and successfully predicted hollow specimen behaviour. A non-linear model was used to predict behaviour of slab specimens assuming full composite action. It could predict a range of loads and deflections the test data fell into, as specimens acted only partly in composite. Modelling partial composite action was not in the scope of this project. Overall, the epoxy and aggregate bonded specimens are the most promising for use as bridge decking, but more research is needed to confirm this.