Prestressed FRP Systems

Abstract

This chapter provides an overview on the state-of-the-art in prestressing systems for the structural retrofitting of reinforced concrete (RC) structures using Externally Bonded (EB) Fibre Reinforced Polymers (FRP). Focus is put on flexural strengthening, which currently is the most common application field for composite materials in structural engineering. The manuscript provides information regarding commercially available prestressing systems and their anchorage procedures. In addition to conventional mechanical anchorages, the innovative ‘gradient anchorage’ that lacks any remaining plates or bolts is also presented. Additionally, the authors mention various current prototypes at the laboratory-scale level. Performed experimental investigations, results, and conclusions represent the core content of this chapter. Several studies from various universities and research institutes worldwide are presented and explained. In these research projects, the previously mentioned systems are applied to specific reinforced or prestressed reinforced concrete members for strengthening purposes. Static and/or dynamic loading indicate the efficiency of the retrofitting concept compared to the reference structure. Generally, prestressed FRP will be demonstrated to follow the principle of conventional prestressed concrete by resulting in higher cracking, yielding, and bearing loads. Especially under service loads, the structural behaviour is improved. A special section is dedicated to prestressed near-surface-mounted (NSM) systems. In addition to the experiments section, calculation techniques for designing prestressed FRP for flexural strengthening are also handled. In shear strengthening and column confinement, prestressed FRP has been limited to notably few research applications to date. Nonetheless, an overview is given and future possible employment is discussed. Eventually, examples from real structural retrofitting projects should provide the practitioners with some background to better disseminate the retrofitting technique in question. The concluding section summarizes the actual situation and identifies needs for future research.

Introduction

After pioneering works in the field of unstressed Fibre-Reinforced-Polymer (FRP) strips for structural retrofitting (Meier 1987, 1995; Kaiser 1989) these techniques have been studied for several years. Their application with an initial prestress is more recent and still in development. In structural engineering, the application of externally bonded FRP is currently somewhat limited to carbon. However, the literature also presents applications with Glass Fibre Reinforced Polymer (GFRP) sheets (e.g. Huang et al. 2005). An experimental study on prestressed Aramid tendons is presented in Lees and Burgoyne (1999). The material is in practice mainly used in the form of strips/laminates (flexure), straps (shear), and wraps (confinement). The information given in the present chapter will be limited to the three types of carbon reinforcements. A special section is dedicated to prestressed Near-Surface-Mounted (NSM) reinforcements.

The mounting of an unstressed outer reinforcement (EBR) has the disadvantage of to provide a very limited additional stiffness to the structure under service loads. In addition, a very large number of research activities have clearly demonstrated that ultimate failure at a structural level occurs after the debonding of the composite reinforcement from its concrete substrate. Generally, the FRP is far from reaching its ultimate capacity in tension. Hence, the composite high tensile strength is not exploited; often, only 20–30 % of the material’s capacity is effectively used (Motavalli et al. 2011).

Like externally prestressed steel cables, pretensioned composite reinforcements offer the possibility to act against present dead loads of the respective construction and thus can reduce existing deflections and close existing cracks. Furthermore, prestressing more efficiently exploits the material’s high tensile strength.

The present chapter provides an overview of prestressed FRP for enhancing the structural efficiency in flexure, shear, and column confinement. Special attention is paid to the following topics: prestressing technique, anchorage method, and current knowledge about the system’s performance. Short subsections about shear strengthening and confinement provide information on recent developments in the field. Finally, practical cases complete the overview.

Flexural Strengthening

General Information

Several investigations have documented the structural advantages of prestressed FRP reinforcement in retrofitting (Deuring 1993). A first positive aspect is the possibility to actively act against dead loads and thus reduce the existing deflections and cracks in the structure. With a sufficient amount of initial prestrain in the laminate, the cracking load is considerably increased compared to a reference (unstrengthened) beam. The same is valid for the load at which the inner steel reinforcement begins to yield. The ultimate load carrying capacity is also generally enhanced. However, a decrease of ductility, resulting in a lower deflection when reaching the ultimate load, can be observed when an initial prestress is applied to the laminate. While a structure that is retrofitted with an unstressed strip always exhibits debonding failure at the peak load, tensile failure of the external reinforcement can be obtained if the initial prestrain $\varepsilon_{fp}$ is sufficiently high. Figure 7.1 shows a schematic representation of the moment-curvature $\left( {M - \chi } \right)$ relationship for the three mentioned situations. An enhanced crack, yield, and ultimate load is shown in terms of an increase in the respective bearing moments $\varDelta M_{cr} ,\varDelta M_{y} \;{\text{and}}\;\varDelta M_{u}$.

Fig. 7.1

Schematic Moment-Curvature relationships for an unstrengthened RC element, a strengthened RC element with an unstressed laminate and a strengthened RC element with a prestressed laminate

Prestressing Systems and Anchorage Techniques

Types of Prestressing

An existing structure can generally be prestressed via three methods, which are summarised in El-Hacha et al. (2001) and presented in Fig. 7.2. The first technique, known as the cambered beam system, requires an initial counter-deflection against the dead-loads by means of hydraulic jacks (El-Hacha et al. 2001). Subsequently, the FRP strip is applied and the structure is prestressed due to subsequent releasing of the initially inflicted deflection. A second possibility is the use of an external support construction, which the equipment for prestressing application is supported against. The third and most common category is the prestressing against the structure itself. This measure requires the previous installation of supporting elements, such as anchor bolts that are used to fixate a hydraulic jack. In most cases, these temporary elements are removed after the completion of the retrofitting action.

Fig. 7.2

Different types of prestressing of an existing RC construction (El-Hacha et al. 2001)

Commercially Available Prestressing and Anchorage Systems

Currently, several prestressing and anchorage systems for CFRP strips are available on the market. In general, they foresee a mechanical anchorage at the strip ends.

In most cases, the external reinforcement is prestressed against the existing concrete structure. To do so, a hydraulic jack installed in a frame element is fixed by means of several dowels. The laminate is usually held in a mobile clamping system and pushed towards the structural element ends, which results in prestressing. Figure 7.3 provides an example of such a device. The presented system by S&P Clever Reinforcement Company fixes an aluminium plate that applies compression to the strip. The anchorage system requires a minimum of adhesive curing, which implicates that the prestress force at the ends is generally fully released 1 day after the installation.

Fig. 7.3

a Anchor plate during installation and b reaction frame with hydraulic jack

The system is relatively simple and light, and can be used with CFRP laminates that are up to 100 mm wide. The laminates used with this system are typically between 1.2 and 1.4 mm thick.

Figure 7.4 presents the main steps to be followed in order to prestress CFRP laminates with this system, which include the following:

Fig. 7.4

Main steps for applying the prestress: a Surface preparation; b drilling the holes for the bolts of the anchors and base angles; c placing the clamp unit; d, e introducing the laminate in anchor and fixing it; f placing the aluminium plate; g placing the aluminium frame; h placing the hydraulic jack; i prestressing the laminate and controlling the deformation

• (1) The surface is prepared to remove the concrete laitance and obtain a rough surface while minimising the exposure of aggregates (Fig. 7.4a).

• (2) Holes are drilled to accommodate the bolts to fix the aluminium anchor plates and base angles. The locations of these holes are previously marked on the element to be prestressed (Fig. 7.4b).

• (3) The clamp units are introduced between the base angles and cutting CFRP laminate with the desired length. The epoxy adhesive to bond the laminate to the concrete is then prepared and applied to the laminate. Finally, the laminate is inserted on the element to be strengthened between the clamp units (Fig. 7.4c–e).

• (4) The steel anchor, aluminium frame, and hydraulic jack are successively placed at both extremities (Fig. 7.4f–h).

• (5) The prestress is applied at the predefined load level (Fig. 7.4i).

Strain gauges glued along the laminate can be used to control the predefined prestress to be applied to the laminate via the pressure indicator in the hydraulic pump (manometer). In most cases, visual inspection by measuring the length increase on predefined marks on the laminate and concrete surface is used to double-check the stress level of the strip. After concluding the prestressing application, the main components that comprise the system (including the clamp units) must remain for at least 24 h to assure a minimum cure of the epoxy adhesive. After this period of time, the components of the system, mainly the base angles, clamp units, and aluminium frames, are removed, and the temporary bolts are cut. This system has been successively used worldwide to upgrade buildings and bridges.

Another technique applied on the market is the “Stresshead”-system by Sika and VSL International Ltd., which acts as an external prestressing that only applies force at the strip ends (Berset et al. 2002). The specially developed anchor head is held by a metallic support doweled to the concrete substrate. Figure 7.5 shows a photo of this device.

Fig. 7.5

Stress-Head anchorage for prestressed CFRP strips

One more alternative prestressing and anchoring solution from Sika for CFRP strips is the “Leoba-CarboDur” system, which acts similarly to the S&P Clever Reinforcement technique. The main difference lies in a preinstalled base plate (in the concrete substrate), which acts as a support for the hydraulic jack during the prestressing and curing phase of the epoxy adhesive. Subsequently, another anchor plate is mounted and pressed against the laminate from the other side. The technique is documented in detail in Andrä et al. (2001) and Andrä and Maier (2005).

Professor Urs Meier developed the first and so far only non-mechanical anchorage system for prestressed CFRP strips at Empa. The “gradient-anchorage” foresees a gradual prestress force decrease at both strip ends over a defined length, $\varDelta l_{i}$, to limit the shear stresses that appear in the anchorage zone and avoid premature debonding (Triantafillou et al. 1992). Debonding issues appear when too high prestress forces are transferred from the laminate to the concrete substrate at the ends. The application technique is based on the adhesive’s ability to undergo accelerated curing at high temperatures (Czaderski et al. 2012; Michels et al. 2012). Figure 7.6 schematically presents the force evolution in the gradient anchorage. An intermediate sector with bond length $\varDelta l_{i}$ is chosen for the accelerated adhesive curing. Subsequently, the initial prestress force level F _p in the strip is decreased by $\varDelta F_{i}$ in the jack by transferring a part to the sector with the cured adhesive. This procedure is repeated until a zero force level is reached in the jacks. After the execution, all temporary mechanical components (support bolts, frame, etc.) can be removed to leave a purely concrete-epoxy-strip connection without any dowels and plates, which results in a more appealing appearance of the retrofitting.

Fig. 7.6

Schematic representation of a gradient anchorage (Michels et al. 2013)

Prototype applications at the laboratory scale were presented by Meier and Stöcklin (2005). The used device is shown in Fig. 7.7 (left). A recently completed R&D project (with the industrial partner S&P Clever Reinforcement) at Empa resulted in the development of a heating device suitable for on-site use (Michels et al. 2013), see Fig. 7.7 (right). However, the already existing prestressing devices, i.e. clamps, remained identical to the existing elements used for mechanical anchorage. In addition to the easier handling of the new heating device on the construction site, the heating duration for an accelerated adhesive curing was also optimised (Czaderski et al. 2012; Michels et al. 2013). A total heating time of approximately 2.5 h was necessary to anchor a prestress force F _p of 120 kN (0.6 % of prestrain, about 40 % of the strip’s ultimate strain). Including preparation, one strip can be prestressed and anchored in approximately 4 h. Outside the anchorage zone, the epoxy is cured at room temperature.

Fig. 7.7

First Empa prototype and newest heating device for application on site

Prototypes at Laboratory Level

Several additional types of mechanical anchorages can be found at the laboratory scale. For instance, Wight et al. (2001), El-Hacha et al. (2003), and El-Hacha and Aly (2013) present examples of anchorage techniques that include bar and plate anchors. The systems consist of a dead end and a jacking end anchor at which a fixed hydraulic jack applies the prestressing force. Wight et al. (2001) proposed a multi-layer CFRP sheets technique. Each sheet was separately prestressed using a steel bar and was mounted in a steel frame anchored on the RC member. This solution was only verified in experimental tests and has not been used in practical applications.

Non-metallic, mechanically anchored, and CFRP anchored U-wraps sheets were investigated in Kim et al. (2008d, e). The general idea of this system was an anchorage system based solely on composite materials without any steel or aluminium elements.

A general and very detailed overview of anchorage systems for CFRP reinforcements of all types was presented by Schlaich et al. (2012).

Beam and Girder Strengthening

The following paragraphs present a few literature examples that experimentally investigated the strengthening of reinforced concrete (RC) or prestressed reinforced concrete (PRC) elements. In general, all studies showed a general increase in the cracking load, yielding load, and ultimate load, as well as lower deflections and crack openings than the reference structure at a specific load level when a prestressed CFRP strengthening was applied.

Suter and Jungo (2001) conducted bending tests on beam elements with 6 m span. A mechanical anchorage with a bonded plate as shown in Fig. 7.3 was used. The study suggested prestrain values from 0.6 to 0.8 % to optimise the structural behaviour. The authors also mention an increasing number of problems induced by increasing the initial force in the laminate. A generally known phenomenon is the slipping out of the strip from the clamp during the prestressing procedure.

Similar test procedure was followed by Meier and Stoecklin (2005), Kotynia et al. (2011), Michels et al. (2013, 2014a), which all used a gradient anchorage system with an initial prestrain $\varepsilon_{fp}$ of 0.6 %. For instance, the first research group obtained tensile failure in the strip when using 4 laminates with a cross-section of 50 × 0.6 mm² instead of 2 laminates with a cross-section of 50 × 1.2 mm². This effect was partially due to a reduction in the interfacial shear stress when using a higher number of strips with a smaller cross-section. Strip failure and the force-midspan deflection diagram are shown in Fig. 7.8. Investigations regarding the application of the gradient anchorage on a shotcrete substrate are documented in Michels et al. (2014b).

Fig. 7.8

Tensile failure of a prestressed CFRP laminate (left) and force-deflection diagram of the test series by Meier and Stoecklin (2005) and Kotynia et al. (2011)

In the field of prestressed concrete girder retrofitting Fernandes et al. (2013) tested four prestressed high-strength concrete (HSC) girders with a span of 20 m under four point bending. Two girders were used as reference, while one was externally strengthened with unstressed CFRP laminates and one was externally strengthened with prestressed CFRP laminates. The cross-sectional geometry of the girders is depicted in Fig. 7.9. Two laminates with a rectangular cross-section of 100 × 1.4 mm² per girder were adopted to strengthen the girders. The previously presented commercial system proposed by S&P Clever Reinforcement Company for prestressing and anchoring the laminates was used. A prestrain level of 0.4 % was chosen, which corresponded to approximately 30–35 % of the strip’s ultimate tensile strength. End-anchorage plates were used (see Figs. 7.3 and 7.9 (right)). The strengthening procedure is also interesting from a practical installation point of view, in which lateral steel L-profiles were used to temporarily increase the available surface for the strip clamping.

Fig. 7.9

Cross section (left) and bottom view of the anchorage zone (right) of a prestressed concrete girder retrofitted with two prestressed CFRP laminates with mechanical anchorage (Fernandes et al. 2013)

Figure 7.10 (left) shows the force versus the deflection at the mid span. Based on the presented experimental program, the following main conclusions can be drawn: (a) the load carrying capacity of unstressed and prestressed laminates increased by 22 and 35 %, respectively, compared to the reference girders; (b) the bending stiffness prior to the onset of yielding of the longitudinal steel bars of both retrofitted girders only slightly changed; (c) the benefits of prestressing the CFRP laminates were materialised not only in terms of strength but also in terms of stiffness at different levels: crack initiation, yield initiation of the longitudinal steel bars, and ultimate load; (d) the initially unstressed CFRP laminate reached a strain of 0.65 %, whereas a value of 0.95 % was obtained with an initial prestrain of 0.4 %; (e) the girders failed via concrete crushing at the top flange.

Fig. 7.10

Force-deflection diagram of the experimental program by Fernandes et al. (2013) and Czaderski and Motavalli (2007)

Large-scale prestressed concrete girders taken from an existing bridge in Ticino (CH) were strengthened with four prestressed laminates with a gradient anchorage and subsequently tested under static loading (see Figs. 7.10 (right) and 7.11). The results are documented in Czaderski and Motavalli (2007). Both strengthened girders (with an unstressed and a prestressed CFRP strip reinforcement) showed significant increase in the load-carrying capacity. Conversely, Aram et al. (2008) reported that the structural behaviour of retrofitted prestressed concrete beams did not show any improvement. This lack of improvement was attributed to a premature debonding of the strip due to very high shear stresses in the gradient anchorage zone because of the short beam span.

Fig. 7.11

Large-scale prestressed concrete bridge girder testing at Empa (Czaderski and Motavalli 2007)

Pellegrino and Modena (2009) presented another case of prestressed concrete elements that were strengthened with external prestressed CFRP strips. This study also utilised mechanical anchorage. Five 10 m long real-scale beams (four RC beams and one PRC beam with pretensioned internal strands) of cross-Sect. 300 × 500 mm², were tested at the University of Padova (Italy). A photo of the test setup is given in Fig. 7.12. Regarding the strengthening, unidirectional CFRP pultruded laminates of 1.2 × 100 mm² and 1.2 × 80 mm² were used.

Fig. 7.12

Test setup and force-deflection curve of the beam tests performed by Pellegrino and Modena (2009)

The RC-C (reference) diagram presented in Fig. 7.12 showed the typical flexural behaviour of RC beams with pre-cracked, post-cracked, and post-yielded stages. The RC-N (with EBR—no end anchorage) diagram shows a brittle behaviour due to the sudden strip delamination. The RC-EA (with EBR and end-anchorage) diagram shows similar behaviour but with a higher value of the ultimate load. This was due to the presence of end-anchorage devices. Intermediate delamination of the CFRP occurred in this case, with failure of the end-anchorage device.

Failure of beams RC-PrEA (RC beam with prestressed EBR) and PRC-PrEA (PRC beam with prestressed EBR) was again due to delamination of the CFRP, but prestressing and end-anchorage action delayed complete failure. A relevant increment in the ultimate load and an increment of the load at which the first crack appeared occurred for beams with pretensioned laminates (RC-PrEA, PRC-PrEA) with respect to the control beam (RC-C) due to the compressive axial force transferred by pretensioning and, for beam PRC-PrEA, also by internal strands.

In general, mechanical anchor devices increase the ultimate capacity of the structural element, delaying the end and/or intermediate delamination. The CFRP strengthening system was capable of significantly increasing the load carrying capacity of the structural element at which first cracking occurs. Furthermore, it allows reduction of crack amplitudes, a more uniform distribution of cracks and a better utilization of CFRP material characteristics with strain values near the ultimate.

For additional literature on experimental investigations about prestressed CFRP laminates in bending, the reader is invited to also consult (França and Costa 2007; Kim et al. 2007, 2008a, 2010a, b; Quantrill and Hollaway 1998; Wu et al. 2003; Neubauer et al. 2007).

Influential Parameters

As mentioned, experimental research proved that unstressed strengthening precludes the full utilisation of the CFRP tensile strength. The limited tensile strength of concrete results in the debonding of CFRPs from the concrete surface. A review of the available literature on the strengthening of RC members with unstressed and prestressed laminates shows that the strengthening effect significantly depends on a number of factors, including the type of laminate, its stiffness, the number of layers, and the existing longitudinal and shear reinforcement ratios (Garden and Hollaway 1998; Teng et al. 2002). In addition to previously presented publications, the reader is invited to consult among other the following literature regarding strengthening efficiency of externally bonded unstressed and prestressed CFRP strips: You et al. (2012), Kim et al. (2008d), Pellegrino and Modena (2009), Yu et al. (2008), Wight et al. (2001), Kotynia and Kaminska (2003), Meier and Stoecklin (2005), Kotynia et al. (2011).

The following paragraphs present a study on influence parameters based on database presented in Kotynia et al. (2013b).

The efficiency of strengthening RC members with prestressed CFRP composites has been analysed based on experimental test results. The gathered data include all variable parameters that influence the test results: cross-sectional dimensions (b, h), compressive concrete strength (f _ck ), tensile (yield) strength and elastic modulus of the steel (f _yk , E _s ) and composite (f _fu , E _f ), cross-sectional area of longitudinal steel (A _s ) and composite reinforcement (A _f ) with corresponding reinforcement ratios $\left( {\rho_{s} ,\rho_{f} } \right)$, equivalent composite reinforcement ratio defined by $\rho_{f,eq} = \rho_{f} .\left( {{{E_{f} } \mathord{\left/ {\vphantom {{E_{f} } {E_{s} }}} \right. \kern-0pt} {E_{s} }}} \right)$, initial prestressing strain of the composite $\left( {\varepsilon_{fp} } \right)$, maximal strain of the composite registered during the test $\left( {\varepsilon_{f,max} } \right)$, failure mode, strengthening ratio in terms of load carrying capacity $\left( {\eta_{u} = {{\left( {M_{u} - M_{u0} } \right)} \mathord{\left/ {\vphantom {{\left( {M_{u} - M_{u0} } \right)} {M_{u0} }}} \right. \kern-0pt} {M_{u0} }}} \right)$ and in terms of cracking load $\left( {\eta_{cr} = } \right.{{\left( {M_{cr} - M_{cr0} } \right)} \mathord{\left/ {\vphantom {{\left( {M_{cr} - M_{cr0} } \right)} {M_{cr0} }}} \right. \kern-0pt} {M_{cr0} }}$, where: M _u , M _u0 , M _cr , M _cr0 stand for ultimate bending moment of the strengthened and reference member as well as cracking moment of such members, respectively). All elements were strengthened with CFRP. The main differences between the used materials were the type of the composite (L—laminate, or S—sheet) and the modulus of elasticity (E _f ). To reduce the influence of the CFRP elastic modulus, an equivalent composite reinforcement ratio $\left( {\rho_{f,eq} } \right)$ was introduced, for which the FRP ratio is defined by $\rho_{f} = {{A_{f} } \mathord{\left/ {\vphantom {{A_{f} } {\left( {b \cdot d_{f} } \right)}}} \right. \kern-0pt} {\left( {b \cdot d_{f} } \right)}}\;{\text{and}}\;d_{f}$ is the effective composite reinforcement depth.

The primary classification of the analysed members was based on the failure mode obtained in the test. Three groups were distinguished. The first two include elements that failed due to intermediate crack debonding (IC) of the composite or due to the rupture of fibres (R). In the last group, the composites were not properly applied and utilised, which resulted in the following failure modes: concrete crushing (CC), debonding of the composite’s ends (ED), concrete cover separation (CCS), or failure of the anchorage system (A). Due to the very low efficiency of the strengthening of the members from the third group, these cases were not considered in the analysis (Kotynia et al. 2013b).

The longitudinal steel reinforcement ratio $\left( {\rho_{s} } \right)$ significantly affects the strengthening gain (ratio) defined as $\eta_{u} = {{\left( {M_{u} - M_{u0} } \right)} \mathord{\left/ {\vphantom {{\left( {M_{u} - M_{u0} } \right)} {M_{u0} }}} \right. \kern-0pt} {M_{u0} }}$, which was analysed in terms of ultimate loads and is shown in Fig. 7.13 (Kotynia et al. 2013b). The experimental test results were divided into two groups of different equivalent composite reinforcement ratios $\left( {\rho_{f,eq} } \right)$, which were equal to 0.05 and 0.10 % (Fig. 7.13). The comparison proved that the strengthening efficiency of members with a higher steel reinforcement ratio was lower. The strengthening efficiency of beams with less steel reinforcement $\left( {\rho_{s} = 0.44\,\% } \right)$ was higher $\left( {\eta_{u} = 1.55} \right)$ than that of beams with higher reinforcement ($\rho_{s} = 0.50\,\% \;{\text{and}}\;\rho_{s} = 0.83\,\%$, corresponding efficiency of $\eta_{u} = 0.86\;and\;\eta_{u} = 0.59$, respectively) for lower composite reinforcement ratios.

Fig. 7.13

Influence of steel reinforcement ratio ρ _s on the strengthening efficiency η _u (left), and of composite prestrain ε _fp on the strengthening efficiency η _u (right)

The same observation was made in a group of members with higher composite reinforcement ratio ($\rho_{f,eq} = 0.10\,\%$, see Fig. 7.13) both for strengthening with mechanically anchored and non-anchored CFRP laminates/sheets. In general, when the steel reinforcement ratio increased twofold $\left( {{\text{from}}\;\rho_{s} = 0.44\,\% \;{\text{to}}\;\rho_{s} = 0.89\,\% } \right)$, the strengthening efficiency value decreased more than twofold $\left( {\text{from}\;\eta_{u} = 1.55\;{\text{to}}\;\eta_{u} = 0.59} \right)$.

The different prestressing strains of the composites $( {\varepsilon_{fp} } )$ can explain the difference in the achieved strengthening efficiency for the members of the same steel and composite reinforcement ratio (Fig. 7.13). The same members shown on the graph of the strengthening efficiency as a function of the prestrain (Fig. 7.13 ) demonstrate the positive influence of the prestrain. In both groups $\left( {\rho_{s} = 0.44\% \;{\text{and}}\;\rho_{s} = 0.50\% } \right)$ an increase in the prestrain $\varepsilon_{fp}$ significantly increased the strengthening efficiency. The beneficial effect of the higher CFRP prestressing strain is noticeable for members that failed due to intermediate crack debonding (IC). Otherwise, the prestressing strain did not affect the ultimate loads of the members with anchored CFRP composites, which failed by CFRP fracture.

The relationship between the equivalent composite reinforcement ratio $( {\rho_{f,eq} } )$ and the strengthening efficiency $( {\eta_{u} } )$ for the members that failed due to composite rupture (R) is presented in Fig. 7.14. An increase in the equivalent strengthening ratio resulted in an almost linear increase of the strengthening efficiency.

Fig. 7.14

Influence of composite equivalent reinforcement ratio ρ _f,eq on the strengthening efficiency η _u

The graphic interpretation of the strengthening efficiency is presented in Fig. 7.15 in terms of the ultimate loads ($\eta_{u}$—solid lines on the graph) and cracking loads ($\eta_{cr}$—dashed lines). The FRP prestressing strain significantly affected the slopes of the curves, and the influence of the FRP prestressing strain was significantly more beneficial to the increase of the cracking loads than the load capacity. The increase in the strengthening efficiency in terms of the serviceability conditions $( {\Delta \eta_{cr} } )$ was up to 4 times higher than in terms of the ultimate limit state $( {\Delta \eta_{u} } )$ (see Fig. 7.15).

Fig. 7.15

Influence of the prestressing strain ε _fp on the strengthening efficiency η _u and η _cr

A CFRP prestressing strain equal to $\varepsilon_{fp} = \varepsilon_{fu} - \varepsilon_{f,max}$ is recommended (where $\varepsilon_{fu}$ is the ultimate CFRP strain and $\varepsilon_{f,max}$ should be presumed of 0.7 %, as suggested by the authors) to maximise the strengthening efficiency (with simultaneous CFRP rupture and debonding). However, concrete crushing may occur if the concrete compressive strength is lower.

A limited number of research programs have considered the preloading effect of the RC member prior to the strengthening on the strengthening efficiency. Tests by Kotynia et al. (2013a) proved the high efficiency of flexural strengthening with prestressed CFRP laminates, even for highly loaded structures prior to strengthening. The RC beams exhausted to 25 and 75 % of their ultimate loads and strengthened with the prestressed CFRP laminates indicated optimistic results.

Results indicated minor differences caused by the effect of preloading. The strengthening ratio was shown to be inversely proportional to the steel reinforcement ratio. The preloading level did not affect the ultimate concrete tensile strains.

Adhesion between the CFRP laminate and the concrete significantly affects slab deformation after the steel reinforcement yields. The load-induced strain in the unbonded laminates $( {\varepsilon_{f,max} } )$ ranged from 0.0050 to 0.0069, while the bonded laminates reached strains of 0.0093–0.0069 (Kotynia et al. 2013a). Similarly, the CFRP strain efficiency $( {\eta_{\varepsilon f} })$ ranged from 0.68 to 0.87 for slabs strengthened with bonded laminates and from 0.56 to 0.68 for the slabs with unbonded laminates.

Two-Way Slab Strengthening

In Kim et al. (2008b) large-scale two-way RC slabs were strengthened with prestressed composite strips anchored with a mechanical anchorage system composed of a bolted plate. Similar to applications with unidirectional flexural elements, an enhancement in the cracking, yielding, and ultimate load was noticed. Important reductions of the deflections under service load were observed when a prestressed strengthening system was applied. The ductile failure of the control slab (unstrengthened) was transformed into a pseudo-ductile mode in the post-peak domain after strip delamination.

Kim et al. (2010c) presented a study on the punching shear behaviour of retrofitted RC slabs by means of prestressed CFRP sheets. For additional considerations, more detailed explanations of the related phenomenon are provided in section “Shear strengthening”.

NSM-FRP Prestressed Systems

Introduction

According to the literature review, prestressed FRP systems for the flexural strengthening of reinforced concrete (RC) elements have already been applied successfully using the externally bonded reinforcing technique (EBR). In the context of prestressed EBR, significant improvements are reported in RC elements in serviceability and ultimate limit state conditions, such as increase of the load carrying capacity, durability, and structural integrity. The NSM technique is, however, more effective for the flexural strengthening of RC elements than EBR (Barros et al. 2007). Therefore, some efforts are being done in order to combine the intrinsic benefits of using NSM-CFRP with those derived from the application of prestressed EBR-CFRP. This section resumes the fundamental research carried out in this context.

Currently, there is few work reported in the literature on NSM-based prestress technique that can actually be applied on job site. Most of the tested specimens were strengthened in the sagging region, but the strengthening tasks were performed as if it was a hogging region (Nordin and Täljsten 2006; Rezazadeh et al. 2013), i.e., the elements are initially turned over, strengthened, and finally turned over again to its original position in order to be tested. The scheme in which the hydraulic jacks are being placed (in line with the FRP and beyond the boundaries of the element) is impracticable in real cases. Only Gaafar and El-Hacha (2008) claim to have a system that allows this technique to be applied in job-site, and other author is currently refining the design of a system for the application of prestress to Carbon FRP (CFRP) laminates (Barros 2009). Barros (2009) has proposed an innovative approach for applying prestressed NSM-CFRP laminates in real practice, but the mechanical system subjacent to this technique was not yet available.

Flexurally Strengthened Beams and Plates

Several experimental investigations on beam strengthening can be found in the literature. Nordin and Täljsten (2006) using 10 × 10 mm² prestressed CFRP rods applied according to the NSM technique to strengthen 200 × 300 × 4000 mm³ reinforced concrete beams, realized that the applied stress was efficiently transferred to the surrounding concrete even without the use of any special device to anchor the CFRP. Analysing their results, the loss of ductility in relation to the non-prestressed beams was remarkable. While in the series strengthened with a CFRP of 160 GPa elastic modulus the final deflection of the non-prestressed beam was about 50–55 mm, after the application of 20 % of prestress it was reduced by 40 %, i.e. 33 mm. The same observation can me made based on the results using a CFRP laminate of 250 GPa elastic modulus where the deformation of the 0 % prestress beams ranged between 37 and 41 mm and the reduction observed by the application of 19 % of prestress was nearly 30 %.

Gaafar and El-Hacha (2008) reported the tests performed on 200 × 400 × 5150 mm³ reinforced concrete beams prestressed with two NSM strips of 2 × 16 mm², and verified a considerable increase of the load at cracking and yielding initiation. This increase was, however, followed by a significant reduction of the ductility, since the deflection at failure was dramatically decreased (60 % of prestress conducted to failure at a deflection level of approximately 50 % of the deflection observed in the non-prestressed beam).

Rezazadeh et al. (2013) carried out an experimental program on RC beams flexurally strengthened with passive and prestressed CFRP laminates and verified that a CFRP reinforcement ratio of $\rho_{f} = {{A_{f} } \mathord{\left/ {\vphantom {{A_{f} } {\left( {b \cdot d_{f} } \right)}}} \right. \kern-0pt} {\left( {b \cdot d_{f} } \right)}} = 0.06\,\%$ conducted to an increase of approximately 63 % in the ultimate load carrying capacity of beams with a steel reinforcement ratio of $\rho_{s} = {{A_{s} } \mathord{\left/ {\vphantom {{A_{s} } {\left( {b \cdot d_{s} } \right) = 0.39\,\% }}} \right. \kern-0pt} {\left( {b \cdot d_{s} } \right) = 0.39\,\% }}$, regardless the fact the CFRP laminate is passive or applied with a prestress level of 20 and 40 %. A prestress level of 20 and 40 % conducted to an increase of 47 and 55 % in terms of load carrying capacity at deflection corresponding to the serviceability limit state, while passive CFRP laminate provided an increase of 32 %. Such in the previous experimental programs, they also verified that the load at cracking and steel yielding initiation increased with the prestress level, but the deflection at failure of the RC beams decreased with the increase of the prestress level, since all the strengthened RC beams failed by the rupture of the CFRP laminate.

In the experimental program carried out by Costa (2014), composed of ten RC beams flexurally strengthened with passive and prestressed CFRP laminates (grouped in three series of RC beams), it was verified that the long term losses of prestress have occurred between a couple of days to a couple of weeks to become stabilized. In this experimental program, Costa (2014) has also verified that the load at crack initiation and at steel yield initiation has increased with the prestress level, but the corresponding deflection was not significantly affected. The force-deflection of the series corresponding to the quasi real-scale beams is represented in Fig. 7.16. The prestressing level (percentage of the tensile strength of the CFRP provided by the supplier) is indicated in the designation attributed to each strengthened beam (CFRP laminate of 1.4 × 20 mm² cross-sectional area was applied in each beam, $\rho_{l} = {{28} \mathord{\left/ {\vphantom {{28} {\left( {150 \times 290} \right)}}} \right. \kern-0pt} {\left( {150 \times 290} \right)}}{ = 0} . 0 6 4\,{\text{\%}})$. In all the series of beams it was verified that the prestress level applied to the CFRP laminates has no influence on the ultimate load carrying capacity of the strengthened beams, since failure was in all cases dominated by the CFRP rupture. However, the deflection at failure has significantly decreased with the increase of the prestress level. The total cracked length of the beams has also decreased with the increase of the prestress level. In terms of average crack spacing, it was similar in all the tested beams and equal to the spacing of the steel stirrups.

Fig. 7.16

Load-midspan deflection curve by Costa (2014) for various prestress NSM prestress levels

Based on the bibliographic survey on the topic, it can be concluded that, in general, the deflections measured in the beams decreased with the increase of the applied prestress level and thus, the behaviour of these elements under serviceability limit states significantly improved, although this evidence has as immediate consequence in the reduction of the ductility of the reinforced concrete element.

Unlike the experimental programs previously presented in which the beams were monotonically loaded up to failure, (Badawi and Soudki 2008) performed cyclic tests on beams strengthened with prestressed NSM-FRPs up to 40 and 60 % of their ultimate capacity. According to the obtained results, the application of prestress increased the fatigue capacity of the original reinforced concrete beams. Failure was essentially dominated by the rupture of the tensile reinforcement, mainly due to the accumulation of slippage between CFRP and adhesive that caused an increase of the average stress installed on the steel bars.

Regarding plates, the efficacy of the NSM strengthening technique with passive FRP reinforcements to increase the flexural resistance was assessed by Bonaldo et al. (2008), Dalfré and Barros (2011), and Breveglieri et al. (2012). In fact, NSM CFRP laminates without any prestress level can increase significantly the ultimate load carrying capacity of RC structural elements and high mobilization of the tensile properties of the CFRP can be assured. However, for deflection levels corresponding to the serviceability limit states the benefits of the CFRP are, in general, of small relevance. By prestressing the CFRP, its high tensile capacity is more effectively used, contributing to increase significantly the load carrying capacity of the strengthened elements under both service and ultimate conditions. The prestress can also contribute to close eventual existing cracks, to decrease the tensile stress installed in the existing flexural reinforcement, and to increase the shear capacity of these elements. Thus, prestressing the CFRP seems to be a cost-effective solution to increase both the structural performance and the durability of the strengthened RC structure. The use of FRP prestressed systems for the flexural strengthening of RC slabs is however quite limited.

Hosseini et al. (2014) have carried out an experimental program composed of four RC slabs with the purpose of evaluating the influence of the prestress level in the behaviour of this kind of elements in terms of serviceability and ultimate limit states. The adopted reinforcement systems were designed to assure flexurally failure mode for all the tested slabs (reinforcement yielding). The SREF is the reference slab without CFRP, and the S2L-0, S2L-20 and S2L-40 slabs are those flexurally strengthened using two NSM CFRP laminates with different prestress level: 0 % (S2L-0), 20 % (S2L-20) and 40 % (S2L-40) of the ultimate tensile strength of the CFRP laminates. The CFRP laminates used in the present experimental program have a cross section of 1.4 × 20 mm². The tested slabs had a percentage of longitudinal tensile steel bars $\left( {\rho_{sl} } \right)$ of approximately 0.35 %, while the CFRP strengthening percentage $\left( {\rho_{f} } \right)$ is approximately 0.08 %.

Figure 7.17 shows the relationship between the applied force and the deflection at mid-span $\left( {F - \delta } \right)$, for the tested RC slabs. It has been verified that:

Fig. 7.17

Force-midspan deflection of the tested RC slabs (Hosseini et al. 2014)

• Strengthening RC slabs with prestressed NSM CFRP laminates resulted in a significant increase of the load carrying capacity at serviceability and ultimate limit states. By applying 20 % of prestress in the NSM CFRP laminates, the service and ultimate loads have increased, respectively, 55 and 136 % when compared to the corresponding values of the reference slab, while 40 % of prestress has guaranteed an increase of 119 and 152 %.

• By increasing the prestress level in the NSM CFRP laminates the overall flexural behaviour of the slabs at service and ultimate states has improved, but the deflection at the maximum load and at yield initiation of the steel reinforcement of the slabs has decreased with the increase of the prestress level. However, the deflection at maximum load was more than two times the deflection at yield initiation, with a significant plastic incursion on the steel reinforcement, which assures the required level of deflection ductility for this type of RC structures.

• Regardless the prestress level applied to the CFRP laminates, all the strengthened slabs failed by rupture of the laminates after yielding of the tension steel reinforcement. This failure mode proved the high effectiveness of the NSM technique for the flexural strengthening of RC slabs.

Analytical Models

By using FEM-based advanced constitutive models for the material nonlinear analysis of RC beams flexurally strengthened with passive and prestressed CFRP laminates Rezazadeh et al. (2013) have demonstrated that existing commercial FEM-based software can be used to predict with good accuracy the behaviour of this type of structures, as long as the data for the model parameters is correctly estimated.

Barros et al. (2012) developed a closed form formulation to determine the moment-curvature response of rectangular cross section of RC elements failing in bending that can be strengthened by prestressed FRP systems. Using the moment-curvature relationship predicted by the model and implementing an algorithm based on the virtual work method, a numerical strategy was developed for the prediction of the force-deflection response of statically determinate beams. This approach was extended to statically indeterminate RC elements failing in bending (Barros and Dalfré 2013).

Costa (2014) developed a spreadsheet for the determination of the most significant points of the moment-curvature of reinforced concrete beams using a closed form formulation. This spreadsheet allows the calculation of the cracking, yielding and ultimate curvature, as well as the corresponding bending moment of rectangular reinforced concrete sections with one layer of conventional tensile reinforcement, one layer of conventional compressive reinforcement, and one layer of composite strengthening to which a certain amount of prestress can be applied. The formulation used in this spreadsheet is based on conventional sectional analysis theory according to which the distribution of strain is assumed to be linear along the height of the beam. In this spreadsheet, the behaviour of concrete and steel are assumed to be in accordance with Eurocode 2 (CEN 2004) while the FRP was assumed as having linear elastic behaviour up to failure. This spreadsheet also includes a formulation capable of using the moment-curvature for the evaluation of the mid-span deflection of a simply supported RC element subjected to four-point loading configuration with notable accuracy.

Design Issues

In general, reinforced or prestressed reinforced concrete members strengthened with prestressed CFRP exhibit strip debonding when they reach their ultimate load carrying capacity. In some cases, tensile failure was reported (Meier and Stoecklin 2005). Currently, design codes have not yet been elaborated for prestressed composites.

A semester project at ETH Zürich (Harmanci 2012) revealed that the application of the conventional design rules for debonding in the free length due to excessive interfacial bond shear stress and/or CFRP tensile strain, in this case according to the Swiss SIA 166 (SIA 2004) for externally bonded strip reinforcement, can also be applied to the prestressed case with reasonable precision. Several design codes include a plate-end debonding criterion when unstressed laminates are loaded. For a prestressed system, the end-anchorage needs to be defined in terms of the ultimate load at which the strip would be eventually pulled out of the anchorage. Theoretical considerations regarding the prestressed laminates bonded to a concrete substrate without a mechanical anchorage are presented in Triantafillou and Deskovic (1991). Gradient anchorage design, with different ultimate crack locations, was presented by Czaderski (2012). Harmanci (2013) validated the results by implementing the failure criteria in a numerical code to calculate bonded and unbonded (on the free length) prestressed CFRP strips with gradient anchorage by comparing them to experimental data taken from Czaderski (2012).

The overall ductility of the strengthened system is decreased when the prestrain applied to the laminate increases (Michels et al. 2011). A structural designer should respect common practice and leave sufficient ductility to the structure. For instance, such ductility can be ensured by guaranteeing a significant difference between the curvature at steel yielding and ultimate strip debonding. Detailed investigations that also include energy dissipation concepts are presented in Kim et al. (2008c), Oudah and El-Hacha (2011, 2012).

Shear Strengthening

Prestressed Shear Reinforcement for Concrete

The governing aim for prestressed shear reinforcement is to enhance the shear capacity of concrete. In terms of the concrete web shear performance, the addition of transverse prestress will result in higher cracking loads and steeper crack angles. A prime motivation for early examples of prestressed steel shear reinforcement was to enable the use of deep, thin webbed members, e.g., Freyssinet’s post-war reconstruction of the Esbly bridge in France (Freyssinet 1950).

Prestressed shear reinforcement is not generally used in current practice for new construction. However, this technique has received growing interest for the strengthening and repair of existing reinforced concrete structures. Strengthening options that are external to the structure are likely to be more practical and less disruptive. Therefore, adequate protection against corrosion is a main concern for external steel reinforcements. Meier et al. (1993) patented a method to apply prestressed FRP shear reinforcement for strengthening applications. FRPs are expected to be durable in external environments and have a high strength-to-weight ratio. Thus, FRP cross-sections can be used to deliver strength enhancements similar to those of steel while being thinner and lighter. The strain capacity of FRPs can be relatively high and using prestressed systems represents a more efficient use of the materials (Burgoyne 2001). FRPs with a high strength and stiffness and the ability to sustain stress over the long-term are required for prestressing applications.

FRP Prestressed Shear Strengthening

The force in prestressed shear reinforcement consists of the initial prestress (minus losses), $\varepsilon_{fp}$, plus additional strains $\Delta \varepsilon_{f}$ generated due to the applied load. The compressive prestress relieves the strain in the existing internal transverse steel and increases the crack bridging force for a given shear crack width. The additional strains, $\Delta \varepsilon_{f}$, are small when the concrete is uncracked but will increase as shear cracks open. The FRP strengthening material properties, e.g. stiffness and strength, play a significant role after cracking. Aggregate interlock mechanisms are enhanced because a higher crack bridging force leads to smaller crack widths. In addition, FRPs are elastic and, therefore, they can continue to sustain load after yielding of any internal transverse steel.

While either bonded or unbonded FRP prestressed systems could be used in principle, more research has focused on unbonded systems so far. A non-laminated strap system developed by Winistörfer (1999) has been used for prestressed shear elements. The system uses a thin thermoplastic CFRP tape, and the tape layers can be fusion bonded together to make a closed strap (Lees and Winistörfer 2011). The flexible, self-anchored straps avoid some of the difficulties associated with gripping FRPs and overcome the difficulties of strengthening reinforced concrete with complex geometries. Figure 7.18 also presents a concept for a prestressed shear reinforcement with CFRP straps (Motavalli et al. 2011).

Fig. 7.18

Concept for a prestressed shear reinforcement (Motavalli et al. 2011)

In an unbonded system, the crack opening displacements are averaged over the unbonded length of the reinforcement, which avoids local crack stress concentrations in the FRP. However, unbonded systems experience size effects because the induced strap strain reduces as the beam depth increases for a given crack opening. Stenger (2000) found that the initial strap prestress is an important parameter for the shear resistance in deep beams. Experiments on small scale rectangular shear critical beams strengthened with CFRP straps (Kesse and Lees 2007) identified five different failure modes. The three shear modes were: (1) Strap failure, which eventually leads to global beam shear failure; (2) Shear failure occurs in an unstrengthened concrete region adjacent to the straps, which do not fail; and (3) Extensive concrete crack opening and damage followed by strap failure. The two flexural modes were distinguished by whether the flexural failure was followed by a strap failure with little ductility or whether concrete crushing eventually followed adequate ductility characterised by the yielding of the longitudinal steel. The propensity for a given failure mode depended on the initial prestress, the spacing and the strap stiffness. This system also featured trade-offs, e.g. a certain level of strap prestress was necessary to ensure effective crack bridging. However, the reserve strain capacity to accommodate crack opening prior to strap failure was limited if the prestress was excessive.

Analysis and Design of Prestressed FRP Shear Strengthening Systems

Because FRP materials are linearly elastic, one difficulty in the analysis of prestressed FRP shear reinforcement is the prediction of the FRP force at failure. The force in the FRP is compatible with the transverse expansion of the base structure. Therefore, analysis techniques that consider strain compatibility/crack opening can allow for this relationship. A general consensus about the shear resistance of reinforced concrete is lacking; thus, the inclusion of additional FRP prestressed reinforcement is a further complication. One approach is to add the contribution of the FRP (V _frp ) to concrete, V _c , and steel, V _s , contributions. Chen and Teng (2001) have proposed an expression for V _frp for prestressed FRP straps based on an assumed crack profile, whereas Hoult and Lees (2009) used a shear friction approach and a compatibility relationship between the shear friction and crack opening to determine the force in the FRP. Another approach consists of using a model that considers equilibrium, compatibility and material laws of reinforced concrete cracked in response to shear, such as the modified compression field theory (MCFT) developed by Vecchio and Collins (1988). Unbonded prestressed CFRP straps have been incorporated into the MCFT in an average sense (Lees et al. 2002) and the inclusion of straps with non-uniform spacing has also been investigated (Yapa and Lees 2013). In a two-dimensional finite element analyses of CFRP strap-strengthened T-beams (Dirar et al. 2013) the strap strains were generally underpredicted and the results depended on the shear models and concrete input parameters. A rotating crack model appeared to match experimental results better than a constant or variable shear retention model.

Practical Considerations for Strengthening Applications

Installing external prestressed shear reinforcement on an existing structure presents a number of challenges. Drilling through the structure may be necessary to connect the compression and tension chords to allow for the insertion of the shear reinforcement, e.g. in slab-on-beam or flat slab (see Fig. 7.19) structures. Prestress must also be imparted on the FRP shear reinforcement. Several different installation methods have been investigated for unbonded CFRP strap systems. For a strap installed in situ, a support pad can be inserted under the strap (see Fig. 7.1a) and lifted (Lees et al. 2002) while shims are placed under the pad to lock in the prestress. Another option is to pre-fabricate a flexible CFRP strap with loops at either end (Czaderski et al. 2008; Koppitz et al. 2013; Keller et al. 2013). The strap is placed around the region to be strengthened, supported on saddles and stressed by joining the two ends of the strap together using a turn-buckle or a threaded rod system. Frictional losses around the saddles can reduce the effective prestress (Czaderski et al. 2008), which would need to be considered in the design. Constraints for FRP strap systems include the need for a smooth bearing surface and a minimum radius to avoid failure (Lees and Winistörfer 2011). As with any external system, adequate protection against vandalism must be ensured.

Fig. 7.19

CFRP strap strengthening of a T-beam (©University of Cambridge) and a flat slab (courtesy of Dr. A.U. Winistörfer)

Confinement

Introduction

FRPs can be used to confine the lateral deformation of concrete columns subjected to axial compressive loadings. In a confined concrete column, the confinement is loaded in the hoop direction, while the concrete is loaded in tri-axial compression such that both materials are used to their best advantages. The confinement can greatly enhance both the strength and the ultimate strain of the concrete. Therefore, the ductility of confined concrete is greatly enhanced. However, the advantage of prestressed confinement compared to unstressed confinement is not as obvious. In this section, the benefits of confinement prestressing, namely a smaller decrease in the residual strength and lower deformations, will be presented.

Effect of Prestress on Confined Column Response

Janke et al. (2009) presented an experimental study of concrete cylinders confined with unstressed and prestressed steel and CFRP bands. The steel or CFRP bands were spirally wounded around the cylinders. Neither of the band types was bonded to the concrete surface. To maintain the pretension, steel anchoring clamps were mounted at the end of the concrete cylinders. The prestress and tensile stiffness of the tangential confinement reinforcement were varied. A concentric load was applied to the confined cylinders until failure to obtain data on the ductility and ultimate load of the different specimens. Typical axial load-compressive strain curves are shown in Fig. 7.20.

Fig. 7.20

Typical axial load compressive strain curves of confined columns (Janke et al. 2009)

The slope of the load-deformation curve changed at higher loads when the confinement was prestressed compared to the unstressed confinement for both confinement materials. After this slope change, the curves for the unstressed and prestressed specimens continued in an almost parallel formation. The peak load and axial strain of the prestressed CFRP confined cylinder was lower than that of the unstressed specimen. This difference was likely due to the initial prestrain of the CFRP band, which reduced the usable load strain. The effect of prestressed confinement on the slope change and peak load was the same in experimental studies performed by Zile et al. (2009) and Ciniña et al. (2012). In both campaigns, the columns were confined by unstressed and prestressed basalt (Ciniña et al. 2012) and CFRP (Zile et al. 2009) bands. However, contrary to Janke et al. (2009) the bands were impregnated with an epoxy resin prior to placing them around the cylinders. Therefore, one might argue that prestressed confinement resulted in a small benefit. However, Janke et al. (2009) presented other arguments showing that prestressed confinement is advisable.

Residual Strength of Concrete

External confinement can become ineffective or may need to be removed as a result of fire damage, vandalism, or repair measures. Fires occur frequently after earthquakes. Vandalism, i.e. deliberate destruction, can also be a problem especially for freestanding concrete columns that are externally confined. Corrosion protection measures are a possible reason for the removal and exchange of confinement long after they have been applied. The term ‘confined concrete strength’ loses its meaning for the above mentioned cases, and the residual strength determines the capability of the column to resist the existing loads. Confined concrete strength is a variable of the confined system. In contrast, the residual strength of unconfined concrete, which has previously been part of a confined system, can be considered a material parameter. This parameter is assumed to depend on the loading history in the confined state.

To elaborate, when the confinement is removed after the confined system has been subjected to loads above the unconfined concrete strength (i.e. to a certain overload) the residual strength of the concrete component is significantly diminished due to concrete damage (i.e. microcracking). This effect is especially pronounced for unstressed confinement because the activation of confinement pressure requires considerable lateral strain, which significantly increases the microcrack density and crack width.

Janke et al. (2009) found that prestressing the confinement reinforcement significantly affects the residual strength of columns after an overload. To estimate the residual strength, the concentric load on the confined cylinders was increased only up to a predefined overload above the unconfined load capacity (Fig. 7.21a). The confinement reinforcement was then removed and the residual capacity was determined in another compression test (Fig. 7.21b).

Fig. 7.21

a Load cycle to 140 % of unconfined compressive strength with confinement, b breaking test after removal of confinement (Janke et al. 2009)

In the experiments the axial compressive strain and plastic deformation of the unstressed specimens was significantly higher than that of the prestressed confinement in response to the same overload (Fig. 7.21a). In the example shown in Fig. 7.21 the axial compressive strain of the initially unstressed confined specimen and the prestressed confined specimen were 3.5 times and 1.7 times higher than that of the unconfined reference specimen at peak load, respectively. The residual strength was 56 and 95 %, respectively. Therefore, prestressing the confinement reinforcement significantly reduces damage to the confined concrete under overloads compared to unstressed confinement reinforcement, which was demonstrated by a higher residual capacity (Fig. 7.21b) and lower axial strain (Fig. 7.21a). The latter could be beneficial to serviceability limit states (lower displacements), while the former significantly improves the safety of construction under special circumstances.

Residual Strength Under Cyclic Loading

Prestressed confinement reinforcement was found to maintain the residual strength under cyclic loading compared to the unstressed variant (Janke et al. 2009). Figure 7.22 shows an example of a cyclically loaded specimen with prestressed and unstressed confinement. The unstressed specimen with low confinement stiffness broke prior to the third peak of the cyclic loading process. The peak of the overload was 109 % of the unconfined compressive strength. In contrast, the residual capacity of the prestressed reference cylinder was 95 % after 50 cycles, which demonstrated the effectiveness of prestressed confinement. The residual capacity values of the prestressed specimens were nearly independent of the number of load cycles for the defined level of overload. Therefore, prestressing the confinement is also beneficial for cyclic loading.

Fig. 7.22

Cyclic loading of specimen with low confinement modulus (Janke et al. 2009)

Eccentric Loading of Confined Columns

Unstressed confinement only becomes effective for lateral expansion. Therefore, the effect of such confinement is small under bending stress, because pure bending creates only a slight overall lateral expansion. The typical response of confined specimen subjected to combined bending and axial load is shown in Fig. 7.23. The results show that even moderate confinement prestress is effective in members under bending stress. A significantly smaller residual deviation compared to unstressed confined specimen was found. Janke et al. (2009) related this phenomenon to a lower degree of damage to the concrete. Once more, the beneficial effect of prestressed confinement was shown.

Fig. 7.23

Horizontal deviation of confined specimen at combined bending and axial load (Janke et al. 2009)

Technical Aspects of Confinement Prestressing

In Janke et al. (2009) the steel or CFRP bands were spirally wound around the cylinders using a stationary lathe. This procedure is illustrated in Fig. 7.24. During winding, the slowly rotating cylinder pulled the band from the coil through the guiding device, which served as a friction brake to create axial tension in the band. However, prestressing FRP on site remains technically complex.

Fig. 7.24

Mechanical prestressing procedure (Janke et al. 2009)

Practical Applications

The present section shows several examples of practical applications of flexural strengthening by means of prestressed CFRP strips (Figs. 7.25, 7.26, 7.27, 7.28, 7.29, 7.30 and 7.31).

Fig. 7.25

Bridge box girder strengthening, Rijeka (Croatia), courtesy of S&P Clever Reinforcement AG

Fig. 7.26

Bridge girder strengthening, Bangok (Thailand), courtesy of S&P Clever Reinforcement AG

Fig. 7.27

Office building, Zurich Altstetten (Switzerland), courtesy of S&P Clever Reinforcement AG

Fig. 7.28

Punching shear strengthening in an office building, Nyon (Switzerland), courtesy of S&P Clever Reinforcement AG

Fig. 7.29

Carbo Stress prestressing system, courtesy of Sika Switzerland

Fig. 7.30

Bridge girder strengthening, Winnepeg (Canada), courtesy of Prof. Dr. Y. Kim

Fig. 7.31

Strengthening of the girders of the Battiferro-Navile viaduct (A14 Highway Bologna-Taranto, Italy), courtesy of Prof. Dr. Carlo Pellegrino and Giorgio Giacomin

Durability

Durability is one key aspect in structural engineering. In the field of prestressed FRP reinforcements only few investigations have been presented. El-Hacha et al. (2004a, b) present studies on strengthened reinforced concrete beams with prestressed CFRP sheets at room and low temperatures. However, these investigations have to be qualified as short-term tests. Since 2000, a prestressed CFRP strip (prestrain 0.55 %) with a gradient anchorage is regularly monitored in terms of strain evolution in time (Michels et al. 2013). Up to now, only a slight decrease (0.04 %) with a stable tendency could be noticed. In general, it is presumed that, in addition to concrete degradation phenomena, creep behaviour of both concrete in compression as well as of the epoxy resin play a major role in long-term efficiency of a CFRP prestressing system (Diab et al. 2009). At Empa investigations in the field of durability for bridge retrofitting applications are currently going on. Of special interest is, next to long-term durability, the temperature stability of prestressed CFRP strips with a gradient anchorage when being used as a negative bending moment reinforcement on top of the structure at the moment of the asphalt installation and the related high temperatures.

Fatigue tests are documents in Kotynia et al. (2011), Oudah and El-Hacha (2013a, b).

Concluding Remarks

This chapter presents several aspects of structural concrete strengthening with prestressed FRPs in flexure, shear, and confinement modes. Prestressed FRP systems have already been widely used in the academic community for several years, whereas practical applications in the field are rather limited. However, the presented investigations show a large potential of a much wider use in the future.

The most obvious advantage of additional flexural reinforcement is a reduction of existing deflections and crack openings for use under service loads. Monotonic loading tests have also shown an enhancement of the yielding and ultimate load. In most cases, the debonding of the additional strip reinforcement was the governing failure mode, while a tensile failure of the composite could be observed in isolated cases. The anchorage system is important in terms of practicability: similar to conventional prestressed concrete, adequate anchorage is necessary for load transfer to the structure during and after the installation phase. Currently, several mechanical systems with anchor bolts and plates are available on the market. These systems require drilling into the existing concrete substrate. The gradient anchorage method, as presented in previous sections, leaves a pure strip-epoxy-concrete connection without any remaining mechanical devices. This technique is promising both in terms of durability and aesthetics.

Prestressed carbon composites in shear and confinement modes are currently limited to academic research. Practical applications have been seldom implemented. Such a procedure seems to be highly beneficial and would be worthwhile to investigate and install.

The anchorage system remains a crucial factor. A wider application range will eventually also depend on the practicability of the prestressing system. Flexural strengthening usually features a better access to the structure, while shear strengthening might often be more complicated due to geometric restrictions.

Future research and development should primarily focus on three aspects: (a) the elaboration of design codes to allow structural designers to dispose of a calculation tool, (b) the assessment of durability of the different systems and (c) the development of practical prestressing and anchoring systems.