Journal of Composites for Construction

This paper presents the results of a series of tests conducted on reinforced concrete (RC) beams strengthened in flexure with near surface mounted (NSM) carbon fiber-reinforced polymer (CFRP) strips. As the main focus of the research is on debonding failure mechanisms, the only test variable investigated was the embedment length of the NSM strip and the NSM strip was extensively strain-gauged to monitor its bond behavior. Load-deflection curves, failure modes, strain distributions in the CFRP strip, and local bond stresses at the CFRP–epoxy interface from the tests are all examined in detail and compared with the predictions of a simple analytical model where appropriate. Of the four embedment lengths investigated, all but the shortest one led to a notable increase in the load-carrying capacity and, to a lesser extent, in the postcracking stiffness of the beam. Debonding was found to be the primary failure mode in all cases except for the beam with the longest embedment length. Also reported in this paper are results from preliminary bond tests used to characterize the local bond-slip behavior of the NSM system. Apart from gaining a better understanding of debonding failures in RC beams with NSM FRP strips, the test results reported in the paper should be useful for future verification of numerical and analytical models.

Reinforced concrete beams are now commonly retrofitted using externally bonded (EB) fiber reinforced polymer (FRP) plates as the technique is both inexpensive and unobtrusive. However, tests have shown that EB carbon FRP plates tend to debond at low strains, which can severely limit the ductility or moment redistribution to such an extent that guidelines often preclude moment redistribution. This paper reports the moment redistribution achieved in tests on nine near full-scale two-span continuous reinforced concrete beams that were retrofitted with near-surface mounted (NSM) plates. The plates were either carbon FRP or high yield steel strips which were adhesively bonded within saw grooves cut into the concrete cover on the tension face or sides of the beam. It was found that the debonding strains of these NSM plates were considerably larger than those associated with EB plates and that substantial amounts of moment redistribution occurred. These tests suggest that NSM plates can be used to increase the strength of reinforced concrete structures with little, if any, loss of ductility.

Research on retrofitting reinforced concrete (RC) beams and slabs using externally bonded (EB) fiber reinforced polymer (FRP) or steel plates has reached the stage where the flexural strength can be determined with confidence. Research has also shown that EB plated structures tend to debond at relatively low strains and to such an extent that guidelines often preclude moment redistribution which can severely restrict the use of plating. However, recent research on retrofitting using FRP and steel near surface mounted plates (NSM) has shown that NSM plates tend to debond at high strains which can allow substantial amounts of moment redistribution. A moment redistribution approach has been developed for both NSM and EB plated beams that allows for the wide range of debonding strains that can occur. This allows RC beams to be retrofitted for both strength and ductility which should help expand the use of this convenient and inexpensive form of retrofitting.

The flexural capacity of concrete beams can be efficiently and effectively improved through bonding fiber-reinforced plastic (FRP) plates to the tensile side. Failure of the strengthened member often occurs through debonding of the FRP from the concrete substrate. If the ultimate FRP strain at debonding failure is known, the moment capacity of the member can be obtained through a simple section analysis. In the American Concrete Institute (ACI) Design Guideline, simple empirical equations are proposed to find the ultimate FRP strain in terms of the FRP stiffness alone. However, when the proposed equations are compared to experimental data, a very large scatter is observed, indicating that the effect of other parameters cannot be neglected. In the present investigation, a new empirical approach to obtain the FRP debonding strain is developed. With a comprehensive experimental database of 143 tests, a neural network relating the ultimate FRP strain to various geometric and material parameters is trained and validated. Using the validated network, an empirical design curve and several correction equations are generated to provide a simple means to find the debonding strain in practical design. Through use of the chart and equations, the calculated ultimate failure moments for the 143 tests in our database are found to be in good agreement with experimental results. The applicability of the new empirical approach to the failure prediction of strengthened members is thus demonstrated.

This paper presents the results of an experimental study designed to investigate the effect of fiber-reinforced polymer (FRP) wraps on corrosion activity and concrete cracking in chloride-contaminated concrete cylinders. Thirty-five concrete cylinders, each having $102\mathrm{mm}$ diameter and $204\mathrm{mm}$ height, concentrically reinforced with one steel reinforcing bar, were subjected to accelerated corrosion exposure for $80\text{days}$ . Test parameters included level of applied potential, presence of FRP wraps, and bar diameter. The corresponding current and concrete expansion were continuously monitored throughout the corrosion exposure. At the end of the test, the steel bars were extracted, cleaned of rust, and weighed to determine the actual steel mass loss. The results showed that, for the same applied fixed potential, FRP wraps effectively reduced the corresponding current, the concrete expansion, and the steel mass loss. For the same applied potential, the current density increased as the bar diameter decreased. For the same corrosion depth, the circumferential expansion of the cylinder caused by corrosion decreased as the concrete cover-to-bar diameter ratio $(c∕d)$ increased.

In this paper, a combined analytical and experimental study of dynamic characteristics of honeycomb composite sandwich structures in bridge systems is presented, and a relatively simple and reliable dynamic experimental procedure to estimate the beam bending and transverse shear stiffness is proposed. This procedure is especially practicable for estimating the beam transverse shear stiffness, which is primarily contributed by the core and is usually difficult to measure. The composite sandwich beams are made of E-glass fiber and polyester resins, and the core consists of the corrugated cells in a sinusoidal configuration. Based on the modeling of equivalent properties for the face laminates and core elements, analytical predictions of effective flexural and transverse shear stiffness properties of sandwich beams along the longitudinal and transverse to the sinusoidal core wave directions are first obtained. Using piezoelectric sensors, the dynamic response data are collected, and the dynamic characteristics of the sandwich structures are analyzed, from which the flexural and transverse shear stiffness properties are reduced. The experimental stiffness results are then compared to the analytical stiffness properties, and relatively good correlations are obtained. The proposed dynamic tests using piezoelectric sensors can be used effectively to evaluate the dynamic characteristics and stiffness properties of large sandwich structures suitable for highway bridge applications.

The flexural performance of reinforced concrete-filled glass-fiber reinforced polymer (GFRP) tubes (CFFTs) has been investigated using seven specimens, $220\mathrm{mm}$ in diameter and $2.43\mathrm{m}$ long. Specimens were reinforced with either steel, GFRP, or carbon–fiber reinforced polymer (FRP) rebar of various sizes. Prefabricated GFRP tubes with most of the fibers oriented in the hoop direction were used in five specimens. One control specimen included conventional steel spirals of stiffness comparable to the GFRP tube and the other had no transverse reinforcement. Test results have shown that CFFT beams performed substantially better than beams with a steel spiral. Unlike CFFTs with FRP rebar, CFFTs with steel rebar failed in a sequential progressive manner, leading to considerable ductility. An analytical model capable of predicting the full response of reinforced CFFT beams, including the sequential progressive failure, has been developed, verified, and used in a parametric study. It is shown that laminate structure of the tube affects the behavior, only after yielding of the steel rebar. Steel reinforcement ratio significantly affects stiffness and strength, whereas concrete strength has an insignificant effect on the overall performance.

A composite system consisting of rectangular glass fiber reinforced polymer (GFRP) tubes connected to concrete slabs, using GFRP dowels has been developed. Seven beam specimens have been tested, including hollow and concrete-filled GFRP tubes with and without concrete slabs. Beam–slab specimens had two different shear span-to-depth ratios and one specimen had carbon–fiber reinforced polymer (CFRP)-laminated tension flange for enhanced flexural performance. Additionally, three double-shear GFRP tube-slab assemblies have been tested to assess the shear behavior of GFRP dowels, in both hollow and concrete-filled tubes. Three compression stubs of concrete-filled tubes were also tested by loading them parallel to the cross-section plane, to study GFRP web buckling behavior. The study showed that GFRP dowels performed well in shear and that composite action is quite feasible. While hollow tubes can act compositely with concrete slabs, more slip between the tube and slab would occur, compared to a concrete-filled tube-slab system. Simplified models are proposed to predict critical web buckling load of fiber reinforced polymer (FRP) tubes. Based on the models, a critical shear span-to-depth ratio of 4 was determined, below which web buckling may occur before flexural failure.