The fiber-reinforced polymer (FRP) plate or sheet debonding or cover delamination (concrete cover separation) failure mode in externally strengthened reinforced concrete beams has attracted a lot of attention. In this paper, a closed-form analytical solution is developed to determine the nonlinear shear stress distribution along the laminate interface and cover area for any load stage assuming a perfect bond. Trilinear moment-curvature and moment-extreme compression fiber strain is assumed to realize the analytical results. By differentiating the FRP axial tension force with respect to position along the beam, closed-form derivatives in terms of curvature and extreme compressive fiber strain are obtained. The results show three distinct regions of constant or stepwise linear shear distribution in each. These correspond to the uncracked, postcracked, and postyielded zones of the shear span. The results are shown to yield an exact match to those numerically obtained by dividing the shear spans into a large number of small segments and applying nonlinear sectional analysis in the middle of each segment. The analytical solution also compares well with the finite-element results using ABAQUS. The analysis of a number of strengthened beams at experimental debonding or cover delamination failure load show that the interface shear stress distribution varies from cases having no cracking at the plate tip (three regions) to those encountering two regions only (postcracked and postyielded) when the FRP plates or sheets extend close to the supports. It also shows that this distribution, at failure, consists of two regions in most of the cases and may only have a postcracked region in beams with relatively shorter plates or sheets. DOI: 10.1061/(ASCE)EM.1943-7889.0000341. (C) 2013 American Society of Civil Engineers.

The current experimental method to determine the transfer length in prestressed concrete members consists of measuring concrete surface strains with a mechanical strain gauge before and after releasing tension. Because this is a time-consuming and tedious process, transfer lengths are seldom measured on a production basis. Furthermore, when transfer lengths are determined using the current method, the times to release tension of the members being measured are often delayed, thereby resulting in artificially higher release strengths for the members evaluated.A rapid, noncontact method for determining transfer lengths in pretensioned concrete members has been developed. The new method uses laser-speckle patterns that are generated and digitally recorded at various points along the prestressed concrete member. The technique was verified against results obtained using the traditional method of adhering stainless steel discs and measuring surface strains with a mechanical strain gauge. The new method has a higher accuracy, requires minimal setup, and can be implemented on a production-based time frame.

This report presents the results from strand end-slip measurements and load tests on 73 pretensioned members that were fabricated at six different Precast/Prestressed Concrete Institute (PCI) producer member plants. Sixty-seven of these beams utilized standard production concrete mixes and placement techniques. As such, the data presented in this report are believed to be representative of standard industry practice between 2005 and 2008. This study revealed that the top-bar effect for pretensioned strands is primarily the result of a small amount of concrete above the strand, rather than a large amount of concrete below the steel. Accordingly, the findings of this investigation indicate that the current design assumptions for bond in pretensioned members are largely un-conservative for members with strands located near the top (as-cast) surface. This phenomenon can result in extremely large transfer lengths for strands located within a few inches of the top surface, including those in shallow members. In addition, the top-cast strand effect typically becomes more pronounced when concrete fluidity is increased. However, these same findings also revealed that the current design assumptions for bond were generally satisfied when strands were located deeper in the members. This was true for both flowable concrete and self-consolidating concrete mixtures.

The current experimental method to determine the transfer length in prestressed concrete members consists of measuring concrete surface strains before and after de-tensioning with a mechanical strain gage. Since this is a time-consuming and tedious process, transfer lengths are seldom measured on a production basis. Furthermore, when transfer-lengths are determined using the current method, the detensioning times of the members being measured are often delayed, thereby resulting in artificially higher release strengths for the members evaluated. A rapid, non-contact method for determining transfer lengths in pre-tensioned concrete members has been developed at Kansas State University. The new method utilizes laser-speckle patterns that are generated and digitally recorded at various points along the prestressed concrete member. The technique was verified against results obtained using the traditional method of adhering stainless-steel discs and measuring surface strains with a mechanical strain gage (Demec or Whittemore type). The new method has a higher accuracy, requires minimal setup, and can be implemented on a production-based time frame.

The prediction of Fiber Reinforced Polymer (FRP) plate and sheet detachment failure in externally strengthened beams is still a subject of interest to develop admissible design models. The majority of studies attribute this failure mode of initiation and sudden propagation of the horizontal plane of separation to the shear stress concentration at the plate tip. Accordingly, several investigators have developed and used anchorage U-wraps to delay or control such separation. While this is one viable failure mode, it is shown in this study to be the less likely immediate cause of separation in a significant number of experiments. A numerical nonlinear analysis procedure is developed for this purpose. The main hypothesis of the design model, established in conjunction with the analysis procedure, is the arresting of plate tip shear crack by internal shear reinforcement followed later by the initiation of a horizontal shear crack branching off of any shear, flexure or shear-flexure crack along the shear span including the tip crack. Depending on the shear stress concentration at the FRP cut off point, the separation failure is classified into two categories. The experimental results of 101 specimens with long plates and 31 specimens with short plates were analyzed using the present procedure. The results, obtained from a plate tip crack initiation model and a maximum plate strain limit model, are evaluated using the experimental data of the 132 tests yielding inconsistent findings. Cracking at the shear stress concentration point is shown to initiate at an early loading stage causing no instantaneous separation failure. The parameters controlling the critical modes of failure are identified. The design model proposed is based on an interface shear stress limit. The design model developed is applicable to test data of long and short plates or sheets.

Results of an extensive experimental program conducted to determine the material, bond characteristics, and time-dependent deformations of a proposed self-consolidating concrete (SCC) mixture for bridge girders are presented. This research program was completed in three phases. The first phase consisted of 15 full-scale, pretensioned SCC flexural specimens tested to evaluate their transfer and development lengths. These specimens included both single-strand and multiple-strand beams, as well as specimens designed to evaluate the so-called “top-strand” effect. The top-strand specimens, with more than 20 in. of concrete below the strand, were tested to evaluate the current American Association of State Highway and Transportation Officials (AASHTO) requirement of a 30% increase in the development length when the concrete below the strand is more than 12 in. Strand end-slip measurements, used to estimate transfer lengths, indicated the proposed SCC mixture meets American Concrete Institute (ACI) and AASHTO requirements. In addition, flexural tests confirmed the proposed SCC mixture also meets current code requirements for development length. The second phase was to evaluate the elastic shortening, creep, and shrinkage properties of the proposed SCC mixture for bridge girders. Four bridge girders with an inverted-T profile were used to measure these time-dependent deformations. In two of the specimens, the strands were tensioned to 75% of the ultimate tensile strength, simulating a girder at service. Strands of the other two specimens were left untensioned to evaluate shrinkage effect of the concrete alone. The shrinkage was then subtracted from the fully tensioned specimens and elastic shortening and creep were isolated after relaxation losses were calculated from code expressions. In addition, the fully tensioned specimens were used to determine transfer lengths of the prestressing strand. The final phase of the program was to record strain measurements of the actual bridge girders used in the field. Elastic shortening, creep, and shrinkage prestress losses of the proposed SCC mixture were compared with current design equations. Instrumentation of seven pretensioned girders in a five-span bridge located in Cowley County, Kansas, was used to measure time-dependent deformations. Three of these girders utilized SCC, while the other four were cast with conventional concrete.

This paper presents the results of an extensive experimental program conducted to determine the material and bond characteristics of a proposed self-consolidating concrete (SCC) mixture for bridge girders. Fifteen full-scale, pretensioned SCC specimens were tested to evaluate the transfer and development lengths of their strand. These specimens included both single-strand and multiple-strand beams, as well as specimens designed to evaluate the so called "top-strand" effect.The top-strand-effect specimens, each with more than 20 in. (500 mm) of concrete below the strand, were tested to evaluate the current American Association of State Highway and Transportation Officials' (AASHTO) requirement. AASHTO requires a 30% increase in the strand development length when there is more than 12 in. (300 mm) of concrete below the strand. Before placing SCC in the beams, the prestressing strand was pre-qualified using the large block pullout test. Strand end-slip measurements, which were used to estimate transfer lengths, indicated that the proposed SCC allows strand to meet the American Concrete Institute and AASHTO transfer-length requirements.In addition, flexural tests on the same specimens confirmed that the SCC also meets the current code requirements for development length. Furthermore, the test results indicated that a 30% increase in development length was not necessary to achieve the full tensile capacity of the strand in the topstrand-effect specimens.

Controlling the prestressing strand-stress range in precracked prestressed concrete girders is critical in the FRP strengthening process to avoid long-term fatigue failures. This paper will address the details of a design procedure that was developed to satisfy target-strengthening requirements while imposing stress range serviceability limits. Two main CFRP flexural strengthening designs were established for use in the experimental program herein. In the first, the amount of CFRP was designed to limit the average strand-stress range to 125 MPa (18 ksi), as per AASHTO requirements, under service live load while maintaining the service-ultimate moment relationship constant. The second design was intended to double the strand-stress range under service live load while keeping the same service-ultimate moment relationship. This was accomplished with iterative cycles of nonlinear sectional analysis to determine the amount of external CFRP reinforcement needed to yield both the targeted stress range and ultimate capacity. The girders were overly reinforced for shear with internal steel stirrups. However, external CFRP stirrups were used to prevent the longitudinal CFRP from premature separation and to develop full flexural capacity. The ACI 318-05 model for shear friction was used for this purpose. The paper also presents analysis results to qualify the experimental behavior of the tested girders. Load-deflection, load-strain, and moment-strand stress variations are seen to have excellent correlation with corresponding experimental curves. CFRP is shown to develop higher strains across cracks relieving strand stresses at these critical locations.

Strengthening concrete girders with fiber-reinforced polymers (FRP) is becoming an increasingly common practice as more research investigations are favorably qualifying the technique. However, important behavioral aspects, such as fatigue in prestressed concrete beams, are yet to be adequately evaluated. An experimental program was conducted to test five pretensioned, prestressed concrete T beams designed for specific prestressing strand stress ranges under live-load conditions. The experimental testing consisted of precracking the beams, strengthening them with carbon FRP, and mechanically loading them to study the effect of increasing the live load on strand fatigue. The beams were either loaded monotonically to ultimate capacity or cyclically fatigued and then loaded monotonically to failure. All the beams were monotonically loaded past their cracking moment at midspan prior to strengthening, to simulate girders in the field. Beam 1 was tested as a control specimen under static loading up to failure. Beams 2 and 3 were strengthened with carbon FRP to have a design stress range of 124 MPa (18 ksi) under service load condition. Beams 4 and 5 were strengthened to have a higher stress range of 248 MPa (36 ksi). For all the strengthened beams, the failure mode observed was FRP rupture. The results favorably qualify the application of FRP strengthening to increase the live load of concrete beams prestressed with straight strands.