Recommended Guide Specification for the Design of Externally Bonded FRP Systems for Repair and Strengthening of Concrete Bridge Elements (2010)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

12 3.1 General This chapter presents the analytical formulations and the experimental data that form the basis upon which the pro- posed Guide Specification is based. In recent years, there have been numerous publications aimed at demonstrating the effectiveness and benefit of exter- nally bonded reinforcement for strengthening reinforced con- crete structural members. However, the vast majority of tests described in these publications reported FRP material prop- erties without sufficient information to enable independent verification of how the data were obtained. In this project only experimental test data, with sufficient documentation to per- mit their direct comparison with results determined from a technically sound structural analysis were considered in the reliability assessment on which the resistance factors in the proposed Guide Specification is based. This approach was fol- lowed throughout the entire project and resulted in small data sets. Experimental data for various limit states governing the behavior of steel-reinforced concrete members with externally bonded FRP reinforcement are summarized in the following sections. 3.2 Flexural Strengthening Experimental and analytical investigations of the behavior of reinforced concrete beams and slabs in flexure have shown that FRP strengthened reinforced concrete members may exhibit, in most cases, one of the following failure modes (Pelvris et al. 1995; Triantafillou 1998; Triantafillou and Antonopoulos 2000): 1. Crushing of concrete in compression (before or after yielding of the tension steel). 2. FRP reinforcement debonding at flexural crack locations. 3. FRP reinforcement end peeling. 3.2.1 Crushing of Concrete in Compression Tests conducted on reinforced concrete members strength- ened with externally bonded FRP composites yielded results that were consistent with prior tests of non-FRP reinforced beams when crushing of concrete in compression occurred first. The experimental data reported by Saadatmanesh and Ehsani (1991); Spadea et al. (1998); Ross et al. (1999); and Almusallam and Al-Salloum (2001) were examined within the context of customary assumptions associated with the analyses of reinforced concrete flexural members in accor- dance with the AASHTO LRFD Bridge Design Specifications. This examination yielded an average experimental to com- puted flexural strength ratio of 1.13 and a coefficient of varia- tion of 11%. The experimental to calculated ratios are shown in Figure 3.1. 3.2.2 Debonding of FRP Reinforcement at Flexural Crack Locations The limit state of FRP plate debonding in steel-reinforced concrete members that also have been reinforced externally with FRP plates occurs when the strain at the concrete/FRP plate interface reaches a limiting value on the order of one- half the ultimate tension strain of the composite materials determined from a standardized direct tension test (e.g., ASTM D3039), as illustrated in Figure 3.2. Laboratory test results (Meier and Kaiser1991; Saadatmanesh and Ehsani 1991; Arduini and Nanni 1997; Spadea et al. 1998; Swamy, et al. 1987; Zureick et al. 2002) have indicated that this limit state is reached when the strain in the FRP reinforcement is between 0.003 and 0.008, as shown in Figure 3.2. In the present analysis, which focuses on flexural strengthening, the strain at the limit state of FRP plate debonding εfrp is set equal to 0.005; lesser values may be appropriate for other limit states involving more brittle failure modes. When the C H A P T E R 3 Interpretation, Appraisal, and Application

13 strain in the FRP system is as low as that shown in the afore- mentioned experiments, the maximum compressive strain in the concrete compression zone invariably is below 0.003. This differs from the customary assumption made in flex- ural analysis of underreinforced concrete beams that the stress in the compression zone can be modeled by a uniform stress equal to 0.85f c´, and a more realistic stress distribution, such as that in Figure 3.3, is necessary for calculating the member flexural strength from the linear distribution of strain. For the development of the Guide Specification in this project, a nonlinear concrete model (Desayi and Krishnan, 1964; Todeschini, et al, (1964) was adopted. The stress- strain relationship for such a model is defined by the follow- ing equations: where εc is the concrete strain, f c´ is the compression strength of the concrete, and ε0 is the strain, corresponding to the maximum stress, computed from: ε0 1 71 3 2= ′ . ( . ) f E c c f f c c c c = ′( )( ) + ( ) 2 0 9 1 3 10 0 2 . ( . ) ε ε ε ε Figure 3.1. Ratios of test to computed flexural strength for crushing of concrete in compression. Figure 3.2. Ratios of test to computed flexural strength for the limit state of debonding of FRP reinforcement.

14 where Ec is the modulus of elasticity for normal weight con- crete. The compressive force in the concrete is obtained by integrating Equation 3.1. Alternatively, for a constant-width compression zone, the compressive force in the concrete can be approximated by an equivalent rectangular stress block having a depth “c” and an average stress of β2(0.9f c´), in which β2 is defined as: Thus the compression force in the concrete is: C bc fc c= ′( )β2 0 9 3 4. ( . ) β ε ε ε ε 2 0 2 0 1 3 3= + ⎛ ⎝⎜ ⎞⎠⎟ ⎡ ⎣⎢ ⎤ ⎦⎥ ⎛ ⎝⎜ ⎞⎠⎟ ln ( . ) c c The center of gravity of the compression zone is k2c from the compression outer edge of the concrete section, where k2 is given in the form: The flexural strength of a rectangular beam or slab exter- nally reinforced with an FRP reinforcement system can be determined from the conditions for equilibrium of forces and compatibility of strains on the cross section, as illustrated in Figure 3.4. 3.2.3 FRP Reinforcement End Peeling The reinforcement end of an externally bonded reinforce- ment system, when subjected to combined shear and flexure, may separate in the form of debonding in three different modes: (1) critical diagonal crack debonding with concrete cover sepa- ration (Yao and Teng 2007) or without concrete cover separa- tion (Oehlers and Seracino 2004); (2) concrete cover separation (Teng et al. 2002); and (3) plate end interfacial debonding (Teng et al. 2002). Critical diagonal crack debonding may occur where the FRP end is located in a zone of high shear force and the amount of steel shear reinforcement is limited. In such a case a major diagonal shear crack forms and intersects the FRP and then propagates toward the end of the FRP rein- forcement. This failure mode is suppressed if the shear strength of the strengthened member remains higher than the flexural strength. k c c c 2 0 0 1 1 2 = − ⎛ ⎝⎜ ⎞⎠⎟ − ⎛ ⎝⎜ ⎞⎠⎟⎡⎣⎢ ⎤ ⎦⎥ ε ε ε ε β ε arctan ε0 2 3 5⎛ ⎝⎜ ⎞⎠⎟ ( . ) Figure 3.4. Strain and force diagrams for a reinforced concrete rectangular section. f c = 2(0.9f' c )(ε/ε0) 1+(ε/ε0)2 0.9f' c ε0 εalt St re ss , f c Strain, ε Figure 3.3. Stress-strain relationship for concrete.

15 In beams with heavy steel shear reinforcement, multiple diagonal cracks of smaller widths dominate the behavior such that concrete cover separation will become the controlling debonding failure mode. Failure of the concrete cover is initi- ated by a crack near the FRP end due to the stress concentra- tion. The crack then propagates to and then along the level of the longitudinal steel tension reinforcement. This mode of failure has been observed in tests on beams with externally bonded steel plates (Jones et al. 1988; Oehlers and Moran 1990) and FRP reinforcement (Malek et al. 1998; Lopez and Naaman 2003; Yao and Teng 2007). Plate-end interfacial debonding is also initiated by high interfacial shear and nor- mal stresses near the end of the FRP that exceed the strength of the weakest element, generally the concrete. Debonding in this case propagates from the end of the FRP towards the middle, near the FRP-concrete interface. This failure mode is only likely to occur when the FRP is significantly narrower than the beam section. In summary, provided that shear failure (in the form of a diagonal shear crack in the concrete) is suppressed (through shear strengthening, if needed), stress concentrations near the FRP reinforcement end may result in debonding through the concrete layer near the level of the longitudinal steel (or, rarely, near the FRP-concrete interface). A wide range of predictive models that include numerical, fracture mechanics, data- fitting, and strength of material-based methods have been developed to address this complex mode of failure (Yao 2004). However, the equations presented in the proposed Guide Spec- ification are based on the approximate analysis of Roberts (1989), because of its simplicity for design purposes. Figure 3.5 shows evaluations of test results conducted on reinforced con- crete beams reinforced with externally bonded FRP reinforce- ment (Yao and Teng 2007) and with externally bonded steel plates (Oehlers and Moran 1990; Swamy et al. 1987). The results of the peeling stress, fpeel, predicted by Robert’s formula were normalized with respect to the interface shear transfer strength of , in which f c´, is in ksi. For the vast majority of tests, the prediction is on the safe side. 3.3 Shear Strengthening Shear strengthening of reinforced concrete members using FRP reinforcement may be provided by bonding the external reinforcement (typically in the form of sheets) with the prin- cipal fiber in the direction (insofar as practically possible) of maximum principal tensile stresses to maximize the effective- ness of FRP reinforcement. For the most common case in which the applied loads acting on a structural member are perpendicular to the member axis (e.g., beams under gravity loads or columns under seismic forces), the maximum prin- cipal stress trajectories in the shear-critical zones form an angle with the member axis which may be taken roughly equal to 45o. However, it is normally more practical to attach the external FRP reinforcement with the principal fiber direction perpendicular to the member axis. Experimental and analytical investigations of the behavior of reinforced concrete members strengthened in shear have revealed the following failure modes: 1. Steel yielding followed by FRP debonding. 2. Steel yielding followed by FRP fracture. τint .= ′0 065 fc Figure 3.5. Normalized calculated peeling stress for externally bonded FRP reinforcement and for steel plates.

16 3. FRP debonding before steel yielding. 4. Diagonal concrete crushing. Depending on the amount of usable steel shear reinforce- ment in the structural element, FRP debonding can occur either before or after steel yielding. The third failure mode is highly unlikely to occur if proper detailing is provided. Diagonal concrete crushing in the direction perpendicular to the tension field can be suppressed by limiting the total amount of steel and FRP reinforcement. Fracture of the FRP reinforcement is highly unlikely to occur because the strain when FRP debonds is substantially lower than that correspond- ing to the FRP fracture strength. 3.3.1 Reinforcing Schemes Typical FRP strengthening schemes for beams and columns are summarized in the following paragraphs. Side bonding is the least effective FRP shear reinforcement scheme due to premature debonding under shear loading and should be avoided if possible (Figure 3.6). Side bonding does not allow for the development of the shear-resisting mecha- nism based on a parallel chord truss model that was first pro- posed by Ritter (1899), due to the lack of tensile diagonals. U-jacketing is the most common externally bonded shear strengthening method for reinforced concrete beams and gird- ers (Figure 3.7). This system is prone to premature debond- ing of the FRP, which may reduce its effectiveness. However, the system is quite popular in practice due to its simplicity. Jacketing combined with anchorage aims to increase the effectiveness of FRP by anchoring the fibers, preferably in the compression zone (Figure 3.8). Properly designed anchors may result in the fibers reaching their tensile capacity, permitting the jacket to behave as if it were completely wrapped. Complete wrapping ensures maximum straining of the fibers and is therefore the most desired reinforcing method, if practically possible (Figure 3.9). Because of the lack of well-documented shear tests in which parameters relevant to the development of these guide specifi- cations have been reported, only data related to U-jackets were examined. A total of 24 reinforced concrete test speci- mens having sufficient information necessary to calculate their nominal shear strengths were selected from the work of Deniaud and Cheng (2001), Deniaud and Cheng (2003), Taerwe et al. (1997), Norris et al. (1997), Leung et al. (2007), and Pellegrino and Modena (2006). The average value of the ratio of the experimental shear strength to that computed value using the equation proposed in the Guide Specification is 1.13 with a coefficient of variation of 28%. The data scatter is shown graphically in Figure 3.10. 3.4 Axially Loaded Members 3.4.1 Axially Loaded Compression Members The most commonly used method of strengthening or upgrading the load-carrying capacity of reinforced concrete columns with FRP reinforcement is to wrap the reinforcement Figure 3.6. Side bonding. Figure 3.7. U-jacketing. Figure 3.8. Jacketing with anchorages. Figure 3.9. Complete wrapping reinforcement.

17 around the section circumference, thus providing confinement that increases both the axial strength and ductility of the col- umn. A review of available work related to strengthening of axially and eccentrically loaded columns with FRP reinforce- ment can be found in Teng et al. (2002). The axial compres- sion strength of a column can be determined directly from Article 5.7.4.4 of the AASHTO LRFD Bridge Design Specifica- tions, in which the compression strength of unconfined con- crete is replaced by the compression strength of the confined concrete. A review of design guidelines for FRP reinforce- ment confining reinforced concrete columns of non-circular cross sections is found in Rocca et al. (2006). In the recommended guide specification, lateral confine- ment pressure for reinforced concrete columns are based on the Canadian guidelines (ISIS 2001). The confinement model is simple enough for adoption in design and yields results that are consistent with the limited test data. 3.4.2 Strengthening Under Axial Loading and Flexure The vast majority of research conducted on strengthening of axially loaded members has been limited to studying the effect of concrete confinement on the axial concentric com- pression strength of short reinforced concrete columns and piers, especially for seismic retrofitting that is necessitated by the inadequacy of transverse reinforcement. A comprehen- sive review of work conducted prior to 2001 was published by Triantafillou (2001). An examination of the 60 papers cited by Triantafillou (2001) shows clearly that there had been no systematic studies that address (1) FRP strengthening of reinforced concrete members under concentric tension and (2) FRP strengthening of reinforced concrete members sub- jected to combined axial loading and bending. The latter issue was recognized by researchers at the Laboratoire Cen- tral des Ponts et Chaussées in Paris, France, who examined two groups of 2.5 m columns of two different concrete strengths that had been externally strengthened with carbon fiber reinforcement and subjected to combined axial com- pression and bending (Quiertant et al. 2004; Quiertant and Toutlemonde, 2005). 3.5 Seismic Retrofitting with Externally Bonded FRP Seismic retrofitting of existing reinforced concrete struc- tural elements may be necessitated by the following: (a) The inadequacy of transverse reinforcement, which may lead to brittle shear failure. This mechanism is associated with inclined cracking (diagonal tension), cover concrete spalling, and rupture or opening of the transverse reinforcement. The shear capacity of sub-standard elements (columns, shear walls, piers, exterior joints, etc.) can be enhanced by providing exter- nally bonded FRPs with the fibers mainly in the hoop direction, in (preferably) closed-type jackets (Figure 3.11a,b). (b) Poor confinement in flexural plastic hinge regions (col- umn ends), where flexural cracking may be followed by cover- concrete crushing and spalling, buckling of the longitudinal reinforcement, or compressive crushing of the concrete. A ductile flexural plastic hinging at the column ends can be Figure 3.10. Scatter of computed strength of reinforced concrete beams with U-jacket FRP reinforcement.

18 achieved through added confinement in the form of FRP jackets with the fibers placed along the column perimeter (Figure 3.11c). This local confinement prevents spalling and delays crushing of concrete; also it delays or even eliminates buckling of longitudinal steel reinforcing bars. (c) Poor detailing in lap splices at the lower ends of columns. The flexural strength of RC columns can only be developed and maintained when debonding of the reinforcement lap splice is prevented. Such debonding occurs once vertical cracks develop in the cover concrete and progresses with increased dilation and cover spalling. The associated rapid flexural strength degradation can be prevented or limited with increased lap confinement, again with fibers along the column perimeter (Figure 3.11d). (a) (b) (c) (d) Figure 3.11. Seismic strengthening examples: (a) shear strengthening of RC column, (b) strengthening of beam-column joint, (c) local confinement in flexural plastic hinge regions, and (d) local confinement at lap splices.