Journal of Composites for Construction

An effective method for strengthening existing concrete beams in bending consists in bonding fiber-reinforced composite laminates on the tension faces of the beams. Laboratory tests on beam specimens show, however, that it is often difficult to develop the full composite strength capacity because of premature failure due to delamination and peeling-off of the laminate. Conditions at the interface between the composite and concrete are not yet fully understood, and studies are required to determine the development lengths needed to achieve the composite's strength capacity. In this paper, we report results of an experimental and theoretical investigation of composite-to-concrete bonded joints. We first present a new and fairly simple experimental apparatus designed and constructed in our laboratory. With this test rig it is possible to investigate the shear conditions between the composite and concrete, to determine the stress and strain distribution profiles in a composite laminate that is bonded to the concrete, and to make direct measurements of the bond strengths. A brief description of the apparatus is given and experimental results are presented. A theoretical analysis for the behavior of such composite-to-concrete bonded joints has also been developed. Predictions of the numerical model are compared to the experimental results.

An analytical tension-stiffening model was developed for fiber-reinforced polymer (FRP) sheets bonded to reinforced concrete based on equilibrium equations and the assumption of linear behavior between bond stress and slip. Experimental measurements were made on small- and large-scale specimens. For the small-scale specimens, the primer coat thickness (adhesive thickness between concrete and FRP sheet) was varied for each specimen. By comparisons of the experimental data to the analytical results, the slip modulus at the concrete/FRP sheet and concrete/steel interface was established at the service load level. From this study, it was found that primer coat thickness has a modest influence on the slip modulus of the repaired member, and stiffening of this interface slightly enhances the stiffness of the specimen. After the formation of secondary and diagonal tension cracks, the stiffness of the specimen dropped rapidly and the linearly elastic slip modulus assumption could not be applied. The slip modulus at the concrete/steel rebar interface is less than at the concrete/FRP sheet interface and the slip modulus is independent of the type of FRP sheet.

Carbon fiber composite connections for hollow-core precast concrete wall panels have been investigated with respect to bond behavior. A panel assembly made up of three precast concrete walls was loaded in the in-plane direction using a cyclic quasi-static load, which was transferred to the wall interface as a shear force. The application of the carbon-fiber-reinforced polymer (CFRP) composite was varied in the covered area, number and orientation of plies, and the concrete surface preparation. An expression, which compares well with the experimental findings, is presented for the effective bond length of carbon fiber composite sheets. The effective bond length of the connection depends on the peel-off shear strength of the concrete, the maximum tensile strain in the composite, the modulus of elasticity of the composite, and thickness of the composite plate.

A numerical procedure for analyzing a common and catastrophic failure mode in pultruded composite material I-beams is presented in this paper. Pultruded wide-flange profiles (often referred to as I-beams) exhibit a number of different failure modes when loaded in flexure or axial compression. The particular failure mode of interest to this paper is that due to the local separation of the flange from the web of the profile following local buckling of the flange. A node-separation technique is used to simulate the progressive failure of the joint between the flange and the web of the wide-flange beam in the postbuckled regime. The procedure has been implemented in NIKE3D, a multipurpose nonlinear implicit finite-element code. The fundamentals of the separation algorithm and the mechanics of the implementation in NIKE3D are described. The results of simulations using the proposed procedure are compared with experimental observations.

A simple procedure is developed for the selection of pultruded structural shapes to be used as beam-columns in structural design. The design equations are then validated by comparison with experimental data gathered during beam-column testing of wide-flange and I-beam pultruded structural shapes. The design procedure accounts for axial load eccentricity and bending action induced by lateral loads and end-moments. The design equations are set in the context of load and resistance factor design, considering both strength and serviceability. This paper addresses the methodology to determine the resistance factors, which should be used with properly selected load-factors accounting for the variability and uncertainty of the loads. The design equations use section-properties, such as the bending stiffness (EI), which must be measured and supplied by industry. It is found that the section-properties used in the design of beams and columns are sufficient for the design of beam-columns. Therefore, the cost and time involved in testing structural shapes are minimized. This paper also addresses the means by which section-properties can be generated effectively and inexpensively.