Bond durability of basalt fibre-reinforced polymers (BFRP) bars

Bond Durability of Basalt Fibre-Reinforced Polymers (BFRP) bars under freeze-and-thaw conditions Mémoire

Mohamed Amine Ammar

Maîtrise en génie civil Maître ès sciences (M.SC)

Québec, Canada

© Mohamed Amine Ammar, 2014

Résumé Ce mémoire présente les résultats de l’étude de l’adhérence entre les barres en polymères renforcés de fibres de Basalte (PRFB) et le béton. Cinquante-quatre cylindres renforcés par des barres en PRFB et dix-huit cylindres renforcés par des barres en polymères renforcés de fibres de fibres de Verre (PRFV) ont été testés par le test d’arrachement. Les paramètres des tests incluent le type de la barre utilisée, le diamètre de la barre, la longueur d’ancrage et le nombre de cycles de gel-dégel (100 et 200 cycles). Les courbes adhérence-glissement des barres en PRFB et PRFV révèlent la même tendance. Les influences des différents paramètres sur l’adhérence ont été analysées. Les modèles analytiques BPE, BPE-modifié et CMR ont été calibrés pour décrire la relation adhérence-glissement des barres en PRFB. Les résultats montrent la capacité des barres en PRFB à remplacer les barres en PRFV dans le renforcement des structures.

iii

iv

Abstract This thesis presents the test results of a study on the bond behavior of basalt fiberreinforced polymers (BFRP) bars in concrete. Forty-five cylinders reinforced with BFRP bars and eighteen cylinders reinforced with glass fiber-reinforced polymer (GFRP) bars were tested in direct pullout conditions. Test parameters included the FRP material, the bar diameter, the bar’s embedment length in concrete and the number of freeze-and-thaw cycles (100 and 200 cycles). Bond-slip curves of BFRP and GFRP bars revealed similar trend. All BFRP specimens failed in a pullout mode at the bar-epoxy interface. The influence of various parameters on the overall bond performance of BFRP bars is analyzed. The BPE, modified-BPE, and CMR analytical models were calibrated to describe the bondslip relationships of BFRP bars. Results demonstrate the promise of using BFRP bars as an alternative to GFRP bars in reinforcing concrete elements.

v

vi

Table of contents Résumé.................................................................................................................................. iii Abstract ................................................................................................................................... v Table of contents .................................................................................................................. vii List of tables...........................................................................................................................xi List of figures ...................................................................................................................... xiii Acknowledgments ............................................................................................................. xvii Chapter 1:

Introduction ..................................................................................................... 1

1.1

General .................................................................................................................... 1

1.2

Thesis organization ................................................................................................. 2

Chapter 2: 2.1

Literature Review and Research Objectives ................................................... 3

Composite materials ............................................................................................... 3

2.1.1

Fibre .................................................................................................................... 3

2.1.2

Matrix.................................................................................................................. 4

2.1.3

FRP Bars ............................................................................................................. 5

2.2

Basalt Fibres ........................................................................................................... 5

2.3

Bond between FRP bars and concrete .................................................................... 7

2.3.1

Surface Treatment ............................................................................................... 7

2.3.2

Bar diameter and embedment length ................................................................ 10

2.3.3

Bond Durability ................................................................................................ 10

2.4

2.3.3.1

Freeze-and-Thaw cycles ........................................................................... 11

2.3.3.2

Low temperatures ..................................................................................... 12

2.3.3.3

High temperatures ..................................................................................... 12

Bond analytical models ......................................................................................... 13 vii

2.4.1 Malvar model (1994) ........................................................................................ 14 2.4.2 Eligehausen, Popov & Bertero model (BPE Model, 1983).............................. 14 2.4.3 Modified BPE model (mBPE, 1996)................................................................ 15 2.4.4 CMR Model (1995) .......................................................................................... 16 2.5

Conclusions based on previous research .............................................................. 17

2.6

Research objectives .............................................................................................. 17

Chapter 3:

Experimental Program.................................................................................. 19

3.1

Introduction .......................................................................................................... 19

3.2

Materials ............................................................................................................... 19

3.2.1 Concrete ........................................................................................................... 19 3.2.2 FRP Bars .......................................................................................................... 21 3.2.3 Test Specimens ................................................................................................. 21 3.3

Bond Test Program............................................................................................... 23

3.4

Test setup.............................................................................................................. 25

3.4.1 Specimen seating and anchoring ...................................................................... 25 3.4.2 Specimen instrumentation ................................................................................ 27 3.4.3 Specimen testing .............................................................................................. 28 Chapter 4: 4.1

Results and Discussions ............................................................................... 29

Phase I results: Bond behavior of unconditioned specimens ............................... 29

4.1.1 Bond-slip relationships ..................................................................................... 29 4.1.2 Residual Bond Strength .................................................................................... 42 4.1.3 Bond failure mode ............................................................................................ 45 4.1.4 Factors Affecting Bond Behavior of FRP bars ................................................ 49

viii

4.1.4.1

Embedment length.................................................................................... 49

4.1.4.2

Bar type and modulus of elasticity ........................................................... 53

4.1.4.3

Bar diameter .............................................................................................. 53

4.1.4.4

Surface treatment ...................................................................................... 55

4.2

Phase II: Effect of Freeze and Thaw ..................................................................... 56

4.2.1

Effect of freeze-and-thaw on bond-slip response ............................................. 59

4.2.2

Effect of FT cycles on maximum bond stress................................................... 65

4.2.3

Effect of FT cycles on bar adhesion to concrete ............................................... 71

4.2.4

Effect of FT cycles on residual bond stress ...................................................... 73

4.2.5

Effect of FT cycles on the failure mode ........................................................... 76

Chapter 5:

Analytical Models ......................................................................................... 79

5.1

Analysis of unconditioned specimens ................................................................... 80

5.2

Analysis of conditioned specimens ....................................................................... 84

Chapter 6:

Conclusions and recommendations for future work ..................................... 93

6.1

Introduction ........................................................................................................... 93

6.2

Summary of the test results of phase I: unconditioned specimens ....................... 93

6.3

Summary of the test results of phase II: conditioned specimens .......................... 94

6.4

Summary of the analytical phase .......................................................................... 96

6.5

Recommendations for future work ....................................................................... 96

References ............................................................................................................................. 99

ix

x

List of tables Table 1: Typical properties of fibres ....................................................................................... 4 Table 2: Properties of thermosetting resins (ISIS Design Manual 2007) ............................... 4 Table 3: Mechanical properties of FRP bars (Adhikari 2009)................................................ 5 Table 4: Properties of BFRP bars ........................................................................................... 6 Table 5: Composition and characteristics of concrete mix ................................................... 20 Table 6: FRP bars design properties ..................................................................................... 21 Table 7: Specimens’ designation and test parameters of phase I ......................................... 24 Table 8: Specimens’ designation and test parameters of phase II ........................................ 25 Table 9: Results of unconditioned pullout bond specimens ................................................. 30 Table 10: Results of unconditioned and conditioned specimens .......................................... 56 Table 11: Compressive strength of concrete after freeze-and-thaw ..................................... 66 Table 12: Stress at the onset of slip at the loaded end for BFRP and GFRP specimens ...... 72 Table 13: Mean values and coefficient of variation of the models’ parameters for unconditioned BFRP and GFRP specimens ................................................................. 81 Table 14: Mean values and coefficient of variation of the models’ parameters for BFRP specimens exposed to 100 FT cycles ............................................................................ 84 Table 15: Mean values and coefficient of variation of the models’ parameters for BFRP specimens exposed to 200 FT cycles ............................................................................ 84

xi

xii

List of figures Figure 1: Fiber Reinforced Polymer (Nacer 2006) ................................................................. 3 Figure 2: Different FRP bars surfaces used in Cosenza et al. (1997) ..................................... 8 Figure 3: Different types of FRP bars used in Davalos et al. (2008) ...................................... 9 Figure 4: Surface deformations of FRP bars used in Baena et al. (2009) ............................. 10 Figure 5: Cracks in concrete cylinders after freeze-and-thaw cycles used in Wardeh et al. 2011 .............................................................................................................................. 11 Figure 6: Variation of compressive strength with FT cycles (Wardeh et al. 2011) .............. 11 Figure 7: BPE model (Eligehausen et al. 1983).................................................................... 15 Figure 8: mBPE model (Cosenza et al. 1996) ...................................................................... 16 Figure 9: (a) Concrete mixing in the Concrete Laboratory at LU and (b) Slump test .......... 20 Figure 10: Surface deformations of (a) BFRP and (b) GFRP bars ....................................... 21 Figure 11: Centering the bars in the specimen moulds using (a) cylindrical moulds, (b) wooden forms and (c) plastic covers ............................................................................ 22 Figure 12: BFRP bar with PVC tubes dictating the embedment length of the bar ............... 23 Figure 13: Pullout test set up ................................................................................................ 26 Figure 14: Wooden base supported on steel channels .......................................................... 26 Figure 15: Hinge system to minimize the effect of surface irregularities ............................ 27 Figure 16: Steel pipe used to anchor the FRP bar................................................................. 27 Figure 17: Plastic washers to ensure centricity of the bar .................................................... 27 Figure 18: LVDTs fixed at the (a) unloaded ends and (b) loaded ends ................................ 28 Figure 19: Typical bond-slip curves for BFRP bars at unloaded ends: (a) ld = 5d, (b) ld = 7d, (c) ld = 10d, and (d) ld = 15d ......................................................................................... 34 Figure 20: Typical bond-slip curves for BFRP bars at loaded ends: (a) ld = 5d, (b) ld = 7d, (c) ld = 10d, and (d) ld = 15d ......................................................................................... 36 Figure 21: Comparison between bond-slip curves for BFRP and GFRP bars at unloaded ends (a) ld = 5d and (b) ld = 7d, (c) ld = 10d, and (d) ld = 15d....................................... 38 xiii

Figure 22: Ratio of the bond stress to the ultimate strength at slip onset at loaded end ...... 39 Figure 23: Ratio of bond stress to ultimate stress at onset of slip at (a) loaded ends and (b) unloaded ends ............................................................................................................... 41 Figure 24: Ratio of average residual to maximum bond stress for BFRP and GFRP bars at (a) unloaded ends and (b) loaded ends ......................................................................... 44 Figure 25: Typical modes of failure (a) pullout of the bar (specimen B10-100-2), (b) bar rupture (specimen G10-100-2), and (c) concrete splitting (specimen G10-70-2) ........ 45 Figure 26: Pullout mode of failure of (a) BFRP specimen B8-120-2, (b) BFRP specimen B8-56-2, and (c) GFRP specimen G10-70-3 ................................................................ 47 Figure 27: Bond strength versus embedment length for (a), (b), (c) BFRP bars and (d) GFRP bars .................................................................................................................... 51 Figure 28: Average bond strength versus embedment length for (a) BFRP bars and (b) BFRP and GFRP bars of diameter 10 mm ................................................................... 52 Figure 29: Influence of the bar diameter on the average bond strength of BFRP bars ........ 54 Figure 30: Typical Bond-Slip curves for (a), (b) and (c) BFRP and (d) GFRP at unloaded ends............................................................................................................................... 62 Figure 31: Typical Bond-Slip curves for (a), (b) and (c) BFRP and (d) GFRP at loaded ends ...................................................................................................................................... 64 Figure 32: Average bond stress for unconditioned and conditioned specimens .................. 66 Figure 33: Normalized bond strength for unconditioned and conditioned specimens ......... 67 Figure 34: Slip at unloaded ends for unconditioned and conditioned specimens ................ 70 Figure 35: Slip at loaded ends for unconditioned and conditioned specimens .................... 71 Figure 36: Ratio of adhesion to ultimate stress for BFRP and GFRP specimens ................ 72 Figure 37: Average residual bond strength of unconditioned and conditioned BFRP and GFRP bars at the unloaded ends .................................................................................. 74 Figure 38: Average residual bond strength of unconditioned and conditioned BFRP and GFRP bars at the loaded ends ...................................................................................... 74 Figure 39: Failure of concrete at the end of bond test for specimen B12-84-100-2 ............ 76 Figure 40: Pullout Mode of failure of (a) GFRP and (b) BFRP bars ................................... 77

xiv

Figure 41: Experimental versus analytical results using the BPE and mBPE models for both BFRP and GFRP bars with (a) ld = 5d, (b) ld = 10d ...................................................... 82 Figure 42: Experimental versus analytical results using the CMR model for unconditioned (a) BFRP and (b) GFRP bars ........................................................................................ 83 Figure 43: Average values of “α” for unconditioned and conditioned BFRP and GFRP specimens ...................................................................................................................... 85 Figure 44: Average values of “p” for unconditioned and conditioned BFRP and GFRP specimens ...................................................................................................................... 86 Figure 45: Average values of “sr” for unconditioned and conditioned BFRP and GFRP specimens ...................................................................................................................... 87 Figure 46: Experimental versus analytical results using the BPE model for (a) B12-84-1002 and (b) G10-70-100-2 specimens .............................................................................. 88 Figure 47: Experimental versus analytical results using the BPE model for (a) B12-84-2001 and (b) G10-70-200-1 specimens .............................................................................. 89 Figure 48: Experimental versus analytical results using the CMR model for (a) B12-70100-1 and (b) G10-70-100-2 specimens ....................................................................... 90 Figure 49: Experimental versus analytical results using the CMR model for (a) B12-842001- and (b) G10-70-200-1 specimens ....................................................................... 91

xv

xvi

Acknowledgments I would like to express my sincere thanks and gratitude to my parents, my wife and to my family who encouraged and supported me during the completion of this work. I doubt that I will ever be able to convey my full appreciation, but I owe them my endless gratitude. I would also like to express my appreciation and gratitude to my supervisors, Professors Ahmed El Refai and Radhouane Masmoudi from Civil Engineering at Laval and Sherbrooke University, respectively, for their guidance during the research program. They provided me with guidance, technical support, and encouragement. It was their persistence, understanding, and kindness that encouraged me to complete this work. I would also like to acknowledge and thank Mathieu Thomassin, from the Civil Engineering Laboratories at Laval University, for his assistance during the experimental phase of this program. I would also like to thank Nicolas Rouleau and all the technicians from the Civil Engineering Laboratories. Their assistance during the experimental phase of this program is greatly appreciated. I would also like to thank all my colleagues in the Civil Engineering Research Group for their support and valuable discussions. I would also like to express my gratitude to the Research Center on Concrete Infrastructure (CRIB) for their financial support.

xvii

xviii

To my mother, my father, my wife and my sisters

xix

xx

Chapter 1: Introduction 1.1 General Deterioration of reinforced concrete (RC) structures due to corrosion of reinforcing steel bars is a major concern. The use of de-icing materials and the dominance of aggressive environment are in the origin of the corrosion problem. According to Boyle and Karbhari (1994), the cost of replacing a deteriorated bridge deck has been estimated to be twice the original construction cost. The ministry of transport of Québec spends half of the maintenance budget on repairing corrosion-damaged RC structures (El-Salakawi et al. 2003). Fiber-reinforced polymers (FRP) bars, with their anti-corrosive properties, have been used as a viable alternative to steel bars in reinforcing concrete structures. In fact, several experimental studies in addition to several field applications such as marine structures, concrete bridge decks, and concrete pavements, have supported the suitability of the FRP bars for structural use (Aiello et al. 2007; Achillides and Pilakoutas 2004; Baena et al. 2009; Sayed et al. 2011; Tastani et al. 2006; Bilal et al. 2004). Basalt fibres, which have been an essential component in the automotive and sport industries for decades, have been recently used in developing a new FRP composite, referred to as basalt fiber-reinforced polymers (BFRP). Recently, BFRP bars have emerged as a promising alternative to conventional glass FRP (GFRP) bars in reinforcing concrete structures. Preliminary tests conducted have demonstrated the promise of using BFRP bars in reinforcing concrete structures. However, the wide acceptance of their usage necessitates comprehensive investigation of their structural and mechanical performance to ensure their suitability for civil-engineering applications. One of the fundamental aspects of structural behavior of RC members is the bond development between the bar and surrounding concrete. Bond characteristics of the reinforcing bars govern the serviceability, ductility, and capacity of concrete structures. There have been few investigations on the bond behavior of BFRP bars to concrete due to their recent existence in the construction field. Therefore, the present research study is initiated to investigate such behavior. The study 1

reports on the bond durability of the BFRP bars under freeze-and-thaw conditions that characterize the climate in Canada in general and in Quebec in particular. The work consists of both experimental and analytical phases. The experimental phase is conducted on seventy-two pullout specimens reinforced with BFRP and GFRP bars. The analytical phase focused on studying the applicability of the available bond-slip models, initially developed for conventional FRP bars, to describe the bond behavior of the BFRP ones.

1.2 Thesis organization The present research work investigates, both experimentally and analytically, the bond performance of BFRP bars to concrete. A literature review on the bond performance of traditional FRP bars in concrete is presented in Chapter 2. The parameters that influence the bond behavior are discussed. The effects of various environmental conditions on the bond performance of FRP bars to concrete are reported. The chapter concludes with the research needs and the main objectives of the current work. Chapter 3 describes the experimental program followed to achieve the research goals. The material properties and the specimen geometry and fabrication in addition to the test methodology are reported. A full description of the test setup, instrumentation, and loading procedures are also presented. In chapter 4, the main results and observations of the pullout tests are discussed. The bond performance of the bars is described in terms of the measured bond strength and the measured slip in the BFRP bars. Chapter 5 includes the analytical phase of this research study in which available bondslip models are assessed for their suitability to describe the bond performance of BFRP bars. Finally, Chapter 6 summarizes the current study and provides conclusions and recommendations for future work.

2

Chapter 2: Literature Review and Research Objectives In this chapter, a review of the experimental and analytical work that has been done on the bond behavior of FRP bars in concrete is provided. Since BFRP bars are recently introduced in the construction field, the literature review is based on research studies that have been conducted on other types of FRP bars.

2.1 Composite materials A composite is defined as the assembly of two or more distinct materials to achieve new material whose overall performance is higher than the individual ingredients (ISIS Design Manual 2007). The benefits of using composites are: light weight, high tensile strength, and durability if compared to traditional steel reinforcing rebars (Nacer 2006). Fiber reinforced polymers (FRP) are composites that consist of two components as shown in Figure 1: the fibres, which are the load carrying elements, and the matrix, which ensures the cohesion of the fibres, the retransmission of applied loads to the fibres, and the protection of fibres from the external environment.

Figure 1: Fiber Reinforced Polymer (Nacer 2006)

2.1.1 Fibre Aramid, Carbon, and Glass are the most common types of fibres used in civil engineering applications. Table 1 shows typical values of the physical and mechanical 3

properties of these fibres. Aramid fibres are characterized by their good fatigue resistance. However, they are susceptible to damage by ultraviolet radiation. Carbon fibres are known by their high longitudinal tensile strength, their high modulus of elasticity, and their excellent fatigue resistance. On the other hand, two types of glass fibres are commonly used in construction: E-glass and S-glass, with the later having a higher tensile strength. Glass fibres are characterized by their low modulus of elasticity and their low transverse shear resistance. Table 1: Typical properties of fibres

Tensile strength (MPa) Elastic Modulus (GPa) Elongation at break (%) Diameter of filament (mµ) Temperature of application (°C)

E-Glass

S-Glass

Carbon

Aramid

3100~3800 72.5~75.5 4.7 6~21 -50~+380

4020~4650 83~86 5.3 6~21 -50~+300

3500~6000 79.3~93.4 1.5~2 5~15 -50~+700

2900~3400 70~140 2.8~3.6 -50~+290

(Source: Basalt fiber & composite materials technology development CO. LTD; www.basaltfm.com)

2.1.2 Matrix Matrix, or resin, is the bonding agent of FRP composites. There are essentially two types of resins: thermoplastic and thermosetting polymers. The choice of resins during the manufacturing process is crucial because it affects the mechanical properties of composites. Thermoplastic polymers are not used in civil engineering purposes because of their low thermal and creep resistances. However, thermosetting resins, such as epoxies, polyesters and vinyl esters, which are the most used resins, have “good thermal stability and chemical resistance and undergo low creep and stress relaxation” as stated by ISIS Design Manual 2007 and shown in Table 2. Table 2: Properties of thermosetting resins (ISIS Design Manual 2007) Resin Epoxy

Specific Gravity 1.2-1.3

Tensile Strength (MPa) 55-130

Tensile Modulus (GPa) 2.75-4.1

Cure Shrinkage (%) 1-5

Polyster

1.1-1.4

34.5-103.5

2.1-3.45

5-12

4

Vinyl Ester

1.12-1.32

73-81

3-3.35

5.4-10.3

2.1.3 FRP Bars Glass FRP (GFRP), Carbon FRP (CFRP) and Aramid FRP (AFRP) bars have been widely used in construction in the last few decades. Table 3 lists the most commercially available FRP bars and their mechanical properties. Table 3: Mechanical properties of FRP bars (Adhikari 2009) Trade Name

V-rod Aslan Leadline Nefmac V-rod Aslan Nefmac Direction Longitudinal Transverse

Tensile Strength (MPa)

Modulus of Elasticity (GPa)

Ultimate Tensile Strain

Carbon fiber 1596 120 0.013 2068 124 0.017 2250 147 0.015 1200 100 0.012 Glass fiber 710 46.4 0.015 690 40.8 0.017 600 30 0.020 -6 Coefficient of Thermal Expansion (x 10 /°C) Steel GFRP CFRP 11.7 6 to 10 -1 to 0 11.7 21 to 23 22 to 23

AFRP -6 to -2 60 to 80

2.2 Basalt Fibres Unlike conventional glass, carbon, and aramid fibres, basalt fibres are recently introduced in the construction field. Basalt fibres are “green” fibres that are processed from basalt rocks through a melting process similar to that used for glass fibres but with low energy requirement (i.e., low cost and little impact on the environment). Basalt fibres have a high strength-to-weight ratio and high modulus of elasticity if compared to E-glass fibres (Sim et al. 2005; Ramakrishman et al. 2005). They are environmentally safe, non-corrosive, and have excellent fatigue resistance (Wu et al. 2010) and good sound and magnetic insulation properties (Palmieri et al. 2009). They also have an excellent resistance to high temperature and high moisture conditions (Militky et al. 2002), which make them suitable 5

for usage in harsh environments (Palmieri et al. 2009). They maintain their volumetric integrity and 90% of their strength if exposed to temperatures over 600°C. Basalt fibres do not produce toxic substances when subjected to fire, overcoming a serious drawback of conventional fibres (Sim et al. 2005). The fibres also have an outstanding chemical stability (Wei et al. 2010; 2011) and high resistance to alkalinity in the surrounding concrete, thus surmounting a commonly known shortcoming of glass fibres when embedded in concrete (Palmieri et al. 2009; Sim et al. 2005). Basalt fibres are mainly produced in China, Ukraine, and Russia with a large chain of distributors around the world. They are used in manufacturing the basalt fibre-reinforced polymers (BFRP) bars. Due to their recent use in construction and the lack of research studies conducted on BFRP bars, the bars were used in only a few projects as replacement of steel bars. For instance, BFRP bars made by Galen Modern Composites Company of Russia, commercially known as Rockbars, were used to reinforce a 22-meter-long concretedeck section of Thompson’s bridge in Northern Ireland (Engineering News-Record, 2010). Rockbars were also used in reinforcing the Fountain Park in Warsaw, Poland, and part of the transnational highway Europe-Western China. Other types of BFRP bars made by Technobasalt (of Ukraine), commercially known as the Basalt Composite Rebars (BCR), were implemented in other projects around the world especially in bridge construction. Table 4 shows the mechanical properties of the BFRP bars as defined by their manufacturers. It is worth mentioning that the application of BFRP bars in practice was based on a few research studies (Brik 1997; Ramakrishnan and Neeraj 1998; Brik 2003; Parnas et al. 2007; Adhikari 2009) in addition to the characterization tests usually conducted by manufacturers. In fact, most of previous studies that have been conducted on BFRP bars were preliminary in nature, in which several parameters that are known to affect the structural performance of the bars in concrete were not thoroughly investigated. Table 4: Properties of BFRP bars Trade Tensile Modulus of Name Strength (MPa) Elasticity (GPa) Rockbar BCR

1000 1100

50 70

Elongation (%)

Coefficient of Thermal Expansion (x 10-6 /°C)

2.24 2.20

2.0 0.35-0.592

Coefficient of thermal expansion of concrete = 10x10-6 /°C depending on the concrete constituents

6

2.3 Bond between FRP bars and concrete As far as the structural performance of RC members is concerned, bond between FRP reinforcement and concrete is the most significant aspect that controls the structure’s capacity, ductility, and serviceability. In this aspect, bond of GFRP and CFRP bars to concrete has been widely investigated, which resulted in a significant amount of experimental data on their bond performance (ACI 2006; Aiello et al. 2007; Harajli et al. 2010; Alvarez 2004; Tastani et al. 2006). It was established that parameters such as concrete strength, bar diameter, embedment length, and concrete confinement significantly affect the bond performance of FRP bars to concrete (Davalos et al. 2008; Aiello et al. 2007; Achillidis and Pilakoutas 2004; Bilal et al. 2004). Bond development is strongly dependent on the mechanical and physical properties of the surface of the FRP bar and the constituents of the FRP material. It varies widely between different FRP bars due to the unique properties of each bar. In the following sections, the parameters influencing the bond performance of FRP bars to concrete are highlighted.

2.3.1 Surface Treatment Nowadays, many types of FRP bars having different surface treatments and characteristics are available. The FRP bar surface may vary between deformations (ribbed, braided, or indented), or sand coating, or a combination of both. Cosenza et al. (1997) tested GFRP bars with various surface deformations (Figure 2) to compare their bond performance to concrete. The bond strength for the ribbed and indented FRP bars was found to be 11.6 MPa and 10.2 MPa, respectively. These values of bond strength were comparable to those obtained for deformed steel bars (11.9 MPa) but were much lower than the bond strength of sand-coated bars (17.7 MPa). The authors concluded that sand grains glued to the bar surface enhanced its bond strength and that the surface deformation play an important role in developing bond between concrete and the bar’s interface.

7

Figure 2: Different FRP bars surfaces used in Cosenza et al. (1997) Chaallal et al. (1993) compared the bond strength of wrapped sand-coated GFRP and steel bars embedded in 150 x 300 mm concrete cylinders. The authors reported that bond strength of GFRP varied between 11.1 MPa and 15.1 MPa representing 62% to 84% of the bond strength of deformed steel bars. In order to evaluate the bond performance of GFRP in tension, Harajli et al. (2010) used thread wrapped and ribbed bars embedded vertically in concrete cylinders of 150 x 300 mm with an embedded length of 7 d, where d is the bar diameter. The results showed that ribbed bars developed larger average bond strength at failure than the thread wrapped bars. The authors concluded that surface deformations of FRP bars significantly affect their bond strength to concrete and their mode of failure at ultimate limit stat. Similar conclusions were reported by Aiello et al. (2007) who studied the bond performance of various types of FRP bars in pullout tests. A total of 24 prismatic concrete cubes (250 mm) reinforced with AFRP, CFRP and GFRP bars were tested. Ribbed, fine-sanded, coarsesanded, and bars with spirals wound with fibres were used in addition to traditional smooth and ribbed steel bars. Results showed that the maximum bond stress of CFRPsw (spiral wound) was four times larger than that of both fine and coarse sanded CFRP bars. Davalos et al. (2008) tested 12 cylinders (150 x 150 mm) to investigate the effect of FRP bar degradation on the interfacial bond to high strength concrete (60 MPa). Different types of GFRP bars (wrapped GFRP 1, slightly sand-coated GFRP 2, and sand-coated GFRP 3, and sandblasted CFRP) were used in the study (Figure 3). The bars were vertically 8

embedded in concrete for a length of 5 d. The results showed that the sand-coated GFRP3 bars exhibited the highest bond strength (23.42 MPa) compared to the wrapped GFRP1 (19.61 MPa) and the slightly sanded GFRP2 bars (21.38 MPa). CFRP bars exhibited an average bond strength of 22.26 MPa. The authors reported that the surface characteristics of FRP bars not only affect their bond strength to concrete but also affects the post-peak bond stress attained.

Figure 3: Different types of FRP bars used in Davalos et al. (2008) Achillides and Pilakoutas (2004) tested concrete cubes (150 mm) reinforced with four FRP bars (Glass, Carbon, Aramid, and Hybrid) in direct pullout conditions to investigate the effect of the bar surface on the bond behavior to concrete. Rough and smooth bars were used in the tests. Results showed that smooth bars developed only 10–20% of the bond stress of deformed bars and that GFRP and the CFRP bars exhibited similar bond strength (12 MPa and 11.9 MPa respectively). Baena et al. (2009) investigated the effect of six different surface treatments on bond between FRP bars and concrete (Figure 4). The bar specimens included sand-coated GFRP and CFRP bars, textured surface CFRP bars, GFRP bars with and without helical wrapping surface and sand-coating, GFRP bars with grooved surface, and steel bars. A total of 88 concrete cubes of 200 mm with 5 d embedment length were tested. The authors concluded that the sand-coated bars had better chemical bond than other bars, which confirmed the influence of the bar surface treatment on its bond to concrete.

9

Figure 4: Surface deformations of FRP bars used in Baena et al. (2009)

2.3.2 Bar diameter and embedment length Many research studies have reported that bond strength of FRP bars is inversely proportional to the bar diameter (Achillidis and Pilakoutas 2004; Baena et al. 2009; AlMahmoud et al. 2007; Cosenza et al. 1995). Sayed et al. (2011) tested 60 pullout specimens to evaluate the effect of the bar diameter on the bond of CFRP bars to ultra-highperformance fiber-reinforced concrete. Four diameters (8, 10 and 12 mm for smooth bars and 7.5 mm for sand-coated bars) were used with embedment lengths of (5, 10, 15 and 20 d). The authors concluded that specimens with shorter embedment length and smaller bar diameter developed the highest bond strength. Alvarez (2004) reported similar conclusions from testing 72 concrete cylinders (150 x 300 mm) reinforced with four different sand-coated GFRP bars. Four diameters (9.5, 12.7, 15.9 and 19.1 mm) and three embedment lengths (5, 10 and 15d) were investigated under different applied temperatures (20°, 40°, and 60°C) for 4 months. It was concluded that bond strength decreased when the embedment length and the bar diameter increased.

2.3.3 Bond Durability Bond development between FRP bars and concrete is affected by the exposure of the test specimens to various environmental conditions. In this section, a review on the effect of high and low temperatures and freeze-and-thaw (FT) cycles on the bond performance of FRP bars is presented.

10

2.3.3.1 Freeze-and-Thaw cycles Wardeh et al. (2011) reported on the effect of FT cycles on concrete. Concrete cylinders (160 x 320 mm) are tested under different numbers of FT cycles (2 cycles/day between -20°C and 10°C). It was reported that the number of cracks recorded on the concrete surface significantly increased with the increasing number of cycles. The cracks became visible after 60 cycles. After 240 cycles, the concrete was severely deteriorated as can be seen in Figure 5. Concrete cylinders lost 20% and 60% of their compressive strength after 30 cycles and 240 cycles, respectively, as shown in Figure 6.

Figure 5: Cracks in concrete cylinders after freeze-and-thaw cycles used in Wardeh et al. 2011

Figure 6: Variation of compressive strength with FT cycles (Wardeh et al. 2011)

11

Karbhari et al. (2000) exposed concrete cylinders reinforced with CFRP sheets to 200 FT cycles between -20°C and 22.5°C. Test results showed a more sudden rupture between the reinforcement and concrete as compared to the control specimens. However, no significant effect on the cylinder compressive strength was observed. Alvarez (2004) exposed 50 cylinders (150 x 300 mm) to 100 and 200 FT cycles (2 cycles/day between -20°C and 20°C at 70% relative humidity) to evaluate the durability of bond of embedded FRP bars. The authors reported that FT cycles had no effect on the bond strength values except for the 19.1 mm bar diameter that showed a reduction of 15% after 200 cycles. Davalos et al. (2008) exposed 12 cylinders (100 x 200 mm) to 30 FT cycles (1 cycle/day between -20° and 60°C) reinforced with GFRP and CFRP bars. The authors concluded that the free-end slip of the pullout test specimens increased and the bond strength reduced by about 18% after exposure. 2.3.3.2 Low temperatures Limited research has been conducted on the effect of low temperature on the bond strength of FRP bars to concrete. Saiedi et al. (2011) tested five CFRP and two steelprestressed concrete beams (300 x 500 x 4400 mm) under fatigue loading at -28°C. All specimens were monotonically loaded to failure after being cycled. The results showed that low temperature caused an initial slip between the bars and concrete, which occurred at loads ranging between 70% and 90% of the ultimate capacity of the beams. The authors concluded that the initial slip caused by low temperature could initiate a serious bond failure in similar structures subjected to long periods of low temperature. 2.3.3.3 High temperatures Galati et al. (2006) conducted 36 pullout tests on concrete cubes reinforced with GFRP bars. The specimens were tested at an elevated temperature of 70°C to investigate the thermal effects on bond between the bars and concrete. The authors concluded that high temperature decreased the ultimate tensile load with up to 16% of that of the control

12

specimens. This was attributed to the weakness of the FRP matrix (resin-fiber interface) due to high temperature, which has caused the bond failure. Abdallah (2006) investigated the bond of FRP bars embedded in concrete prisms of 50 x 76 x 750 mm. The tests also included 42 pullout specimens. Isorod and C-bar (GFRP), Leadline and CFCC (CFRP), and steel bars were used in the study. The specimens were preheated at 20°, 50° and 100°C before being tested. The results showed that bond strength of the bars decreased with the increasing temperature up to 100°C. The authors concluded that the difference in transverse thermal expansion between GFRP and concrete increased the micro-cracking in concrete and weakened the adhesion between FRP bars and concrete, which resulted in a reduction in the bond strength achieved. Similar conclusion was reported by Davalos et al. (2008) who confirmed that concrete specimens maintained at 60°C for 90 days had bond strength reduction as compared to control specimens. Katz et al. (1999) reported that, at high temperature, FRP bars lose their mechanical properties and their bond strength to concrete. Experimental results showed a reduction in the bond strength of FRP bar between 20% and 40% at 100°C, 75% at 150°C, and 90% at 220°C. Alvarez (2004) tested 72 cylinders (150 x 300 mm) after being exposed to 40°C and 60°C for 4 months. Four diameters (9.5, 12.7, 15.9, and 19.1 mm) of GFRP bars and three embedment lengths (5, 10 and 15d) were used. The results showed a reduction in the bond strength by 18% and 20% after being exposed to 40°C and 60°C, respectively, for specimens with 15.9 mm bar diameter and embedment length of 5d. This reduction in strength increased to 26% at 60°C for the 19.1 mm diameter specimens.

2.4 Bond analytical models In order to predict a bond-slip relationship and determine the structural performance of FRP-reinforced concrete structures, an analytical model that describes the bond behavior of FRP bars is needed. However, given the large number of parameters that affect the bond of FRP bars to concrete, no unique model has been proposed up till now. A few available

13

formulations for FRP bars still have to be validated by experimental tests. In the following, a review of the most recognized models is reported.

2.4.1 Malvar model (1994) The bond-slip ( -s) relationship of GFRP bars was first reported by Malvar (1994) based on an extensive experimental research using GFRP bars with different surface treatments. This model is presented by the following relationship: ( ⁄ (

) ( )( ⁄

)( ⁄ )

( ⁄

) )

(Equation 1)

where τm and sm are the peak bond stress and the corresponding slip; F and G are empirical constants that are determined for each bar type by curve-fitting of the experimental τ-s curves.

2.4.2 Eligehausen, Popov & Bertero model (BPE Model, 1983) The BPE model was proposed by Eligehausen et al. (1983) for deformed steel rods and has been successfully applied to FRP bars (Cosenza et al. 1995; Alunno et al. 1995; Faoro 1992). It describes the ascending branch of the bond-slip (s < sm) relationship as shown in Figure 7 and as described by the following equation:

( )

(Equation 2)

where τ1 = τm = maximum bond stress and s1 = sm = slip at maximum stress. For FRP bars, α is a curve fitting parameter not larger than 1 to be physically meaningful (α = 0.4 for steel bars).

14

Figure 7: BPE model (Eligehausen et al. 1983) As shown in Figure 7, the BPE model presents a second branch with constant bond τ = τm up to a slip s = s2, a linearly descending branch from (s2, τ1) to (s3, τ3), and a horizontal branch for s > s3 at a value of τ = τ3 due to the friction development between the bar and concrete. Values of s2, s3 and τ3 are to be calibrated from the experimental results. It can be noticed that the BPE model offers an initial slope for the ascending curve equal to infinity to accurately model the adhesion between FRP bars and concrete.

2.4.3 Modified BPE model (mBPE, 1996) Comparing the experimental and analytical results obtained by applying the BPE model, Cosenza et al. (1996) noted a disagreement in the descending branch and proposed the modified BPE (mBPE) model shown in Figure 8.

15

Figure 8: mBPE model (Cosenza et al. 1996) The modified model presents the same ascending branch of the BPE model and a softening branch having a slope p from (s1,τ1) to (s3,τ3) given as follows:

(

)

(Equation 3)

2.4.4 CMR Model (1995) Cosenza et al. (1995) proposed the following relationship for the ascending branch of the bond-slip curve as an alternative to the BPE model:

(

⁄

)

(Equation 4)

where sr and β are parameters based on curve fitting of the experimental data. Similar to BPE model, the CMR model presents adhesion between FRP bars and concrete by providing an initial slope for the ascending curve equal to infinity. Various studies have investigated the suitability of these models to describe the bond behavior of FRP bars to concrete. It has been reported that the CMR and the BPE model can provide satisfactory correlation while the Malvar model underestimates the initial stiffness compared to the experimental values (Alvarez 2004; Adhikari 2009).

16

2.5 Conclusions based on previous research A literature review on the bond behavior of FRP bars to concrete was carried out. From the presented literature, the following points may be concluded:  Most of previous bond studies have been conducted on GFRP and CFRP bars as these bars have been widely commercialized in the last few decades. However, a lack of information on the bond performance of BFRP bars is noticed due to their recent use in the construction field.  Many studies were carried out to investigate the effect of various parameters on bond of FRP bars to concrete. Surface deformations, embedment length, and bar diameter are among the most commonly investigated parameters that influence the bond behaviour.  There is lack of experimental data on the bond performance of GFRP and CFRP bars in cold regions, which requires more investigation, not to mention the lack of knowledge on the performance of BFRP bars in such conditions. In this aspect, little information on the effect of freeze-and-thaw cycles and low temperature on FRP bond is available.  Reported results have confirmed that the difference in thermal expansion coefficient between concrete and FRP bars affect the developed bond strength. Further investigation is required to investigate the bond performance of BFRP bars in cold regions.  Analytical models that describe the bond-slip relationships were successfully applied for GFRP and CFRP bars. None of these models has been validated for BFRP bars.

2.6 Research objectives Based on the above conclusions, the main objective of the current study is to investigate the bond behavior of BFRP bars in concrete under the environmental conditions that dominate in Canada in general and in Quebec in particular. The detailed objectives of the current research study are initiated to develop our understanding in this area as follows:  To investigate the bond behavior of BFRP bars as a new reinforcing material in concrete. This includes: 17

 Assessment of the effect of the bar diameter and the embedment length on the bond performance of the bar;  Comparison between the bond performance of BFRP and GFRP bars that have similar surface treatment;  Assessment of the available analytical models and their applicability to describe the bond performance of BFRP bars to concrete.  To investigate the effect of freeze-and-thaw cycles on bond between BFRP bars and concrete. This includes:  Evaluation of the long-term effect of freeze-and-thaw cycles on the BFRP bond performance;  Assessment of the available analytical models and their applicability to describe the bond of BFRP bars after being exposed to freeze-and-thaw cycles.

18

Chapter 3: Experimental Program 3.1 Introduction In this chapter, a description of the materials used (concrete and FRP bars) in the experimental program is presented. The test matrix, the fabrication process of the specimens, and the adopted test procedures are presented. The experimental program is divided into two phases: Phase I: Includes pullout tests on concrete cylinders reinforced with BFRP and GFRP bars at room temperature. The objective of this phase is to study the bond behavior of BFRP bars as compared to GFRP bars. In this phase, bond-slip relationships of both types of bars are developed. Bond failure mode and mechanism and the effect of various parameters on the bond performance of the bars are detailed. Phase II: Includes pullout tests on concrete cylinders reinforced with BFRP and GFRP bars after being exposed to freeze-and-thaw (FT) cycles. The main objective of this phase is to investigate the effect of FT cycles on the bond durability of the bars. Specimens in this group are exposed to 100 and 200 FT cycles. Bond-slip curves and failure modes are reported and compared to those of control specimens.

3.2 Materials 3.2.1 Concrete Concrete mix of compressive strength 50 MPa was used to cast all pullout specimens. This relatively high strength concrete was selected to replicate practical mix design in Canada where freezing environment prevail most of the year. The use of high strength concrete also allows the full utilization of the high tensile capacity of FRP bars. Table 5 shows the constituents of the concrete mix in which locally produced constituents are used. The concrete mix had a maximum aggregate size of 14 mm and a water-cement ratio of 0.45. Admixture of type MasterBuild was used to increase the workability of concrete and 19

to minimize the amount of water cement ratio applied. Ordinary Portland cement was used in the concrete mix. A slump of 110 mm was obtained. All pullout specimens were prepared and cast in the Concrete Laboratory of Laval University (LU) (Figure 9). After being de-molded, specimens were labeled and stored into a curing room at a temperature of 20°C ± 2°C and a humidity of about 95% for 28 days.

(a)

(b)

Figure 9: (a) Concrete mixing in the Concrete Laboratory at LU and (b) Slump test Table 5: Composition and characteristics of concrete mix Components

Quantities

Water (kg/m3) Cement T10 (kg/m3) Sable de la voirie (kg/m3)

211.3 473.7 911.9

Calcaire 2.5-10mm (kg/m3) Calcaire 5-14mm (kg/m3) Micro Air: Master (ml) Superplasticizer: Master Build (ml)

230.5 381.4 85.3 94.7

Characteristics Compressive Strength (MPa) Slump (mm) Entrained air (%) 20

50 110 5.5

3.2.2 FRP Bars BFRP and GFRP bars were used in this study. Prior to pullout testing, tensile tests were carried out on bare bars to characterize their mechanical properties (El Refai, 2013). Test results are listed in Table 6. Bar characteristics reported by the manufacturer are also given in the table as guaranteed properties. As per the manufacturers’ specifications, BFRP and GFRP bars are pultruded using epoxy and vinyl ester resins, respectively. Both FRP bars are sand-coated on their surfaces. Visual inspection of the bars reveals uniform and consistent sand coating on GFRP bar surface whereas the BFRP bar surface shows shallow spiral indentations spaced at 2.75 mm along the bar (Figure 10). Table 6: FRP bars design properties Bar material

Resin type

Surface treatment

Tensile strength (MPa)

Ultimate strain (%)

Guaranteeda Testedb

a b

BFRP

Epoxy

GFRP

Vinyl ester

sandcoated sandcoated

Elastic Modulus (GPa) Guaranteeda Testedb

1168

1017

0.022

50

48

941

986

0.019

54

53

Reported by the manufacturer Tested by El Refai (2013)

(a)

(b)

Figure 10: Surface deformations of (a) BFRP and (b) GFRP bars

3.2.3 Test Specimens Cylindrical moulds of 150 mm diameter and 300 mm height were used to cast the pullout specimens. BFRP and GFRP bars of 1000 mm long were concentrically positioned 21

in the moulds using the wooden frame shown in Figure 11. Circular holes of diameters equal to the bar diameters were made at the top and bottom of the wooden frame. After concrete casting, a plastic cover with holes of equal diameter ensured the centricity of the bar.

(a)

(b)

(c)

Figure 11: Centering the bars in the specimen moulds using (a) cylindrical moulds, (b) wooden forms and (c) plastic covers Prior to casting, FRP bars were properly marked so that the embedment length, ld, would lie in the middle third of the cylindrical mould. PVC tubes were used as bond breakers at top and bottom of the specimen. This arrangement was chosen to prevent compressive stresses that are induced during the pullout testing from influencing the bond behaviour of the bar (Achillides and Pilakoutas 2004). The PVC tubes were glued in position on the bar length before casting. It was ensured that little amounts of glue enough to position the tube in place without affecting the bond performance of the bar was used. Figure 12 shows typical bar preparation before being positioned in the concrete cylinder. Embedment lengths were designed as multiples of the bar diameter to facilitate comparisons between bars of different diameters as will be illustrated.

22

Figure 12: BFRP bar with PVC tubes dictating the embedment length of the bar

3.3 Bond Test Program In each phase of the testing program, the specimens were divided into two groups: (1) group A, which consisted of BFRP-reinforced specimens and (2) group B, the companion group, which consisted of GFRP-reinforced specimens tested for comparison. Phase I: In phase I, a total of 48 specimens reinforced with BFRP and GFRP bars were tested. The test parameters included the bar diameter, the embedment length of the bar in concrete, and the FRP bar material. Three nominal diameters (8, 10, and 12 mm) were used for BFRP bars. Only GFRP bars of diameter 10 mm were tested. Four embedded lengths taken as multiples of the bar diameter (5, 7, 10, and 15 times the diameter) were investigated. For each set of test parameters, three specimen replicates were tested to ensure the reliability of the test results. Specimens are labeled as follows: the first character marks the bar type (B for basalt and G for glass), followed by the bar diameter (8, 10, or 12 mm). The numeral next to the bar diameter refers to the embedment length. Last digit refers to the specimen number in its group. For instance, B12-120-3 refers to a BFRP-reinforced specimen with a bar diameter of 12 mm and an embedment length of 120 mm (equivalent to 10 times the bar diameter). The last digit refers to the third specimen of its set. Table 7 shows the test matrix and the specimens’ designation of phase I.

23

Table 7: Specimens’ designation and test parameters of phase I Specimen’s designation* B8-40-x B8-56-x B8-80-x B8-120-x B10-50-x B10-70-x B10-100-x B10-150-x B12-60-x B12-84-x B12-120-x B12-180-x G10-50-x G10-70-x G10-100-x G10-150-x *

Specimen’s number

3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3

Bar type

Basalt specimens BFRP

BFRP

BFRP

Glass specimens GFRP

db (mm)

ld (mm)

8 8 8 8 10 10 10 10 12 12 12 12

40 56 80 120 50 70 100 150 60 84 120 180

10 10 10 10

50 70 100 150

The letter x in the specimen’s label refers to the specimen number 1, 2, or 3.

Phase II: In this phase, the effect of FT cycles on the bond performance of BFRP and GFRP bars was investigated. Only one embedment length of 7db was considered for all specimens. A total of 24 FRP-reinforced concrete cylinders were tested. The parameters investigated were the FRP material, the bar diameter, and the number of FT cycles (100 and 200 cycles). Table 8 shows the test matrix and the specimens’ designation for phase II. Specimens are labeled similar to those of phase I. The number of FT cycles (100 or 200) is added to the specimen’s label to differentiate between unconditioned and conditioned specimens. For instance, B12-84-100-3 refers to a BFRP-reinforced specimen with a bar diameter of 12 mm and an embedment length of 84 mm that has been subjected to 100 cycles of freezeand-thaw. The last digit refers to the third specimen of its set.

24

Table 8: Specimens’ designation and test parameters of phase II Specimen’s designation

Specimen’s number

B8-56-100-x B10-70-100-x B12-84-100-x B8-56-200-x B10-70-200-x B12-84-200-x

3 3 3 3 3 3

G10-70-100-x G10-70-200-x

3 3

*

Bar type

db (mm)

ld (mm)

Basalt specimens BFRP 8 10 12 BFRP 8 10 12 Glass specimens GFRP 10 10

56 70 84 56 70 84 70 70

# of cycles

100

200

100 200

The letter x in the specimen’s label refers to the specimen number 1, 2, or 3.

3.4 Test setup 3.4.1 Specimen seating and anchoring Pullout arrangement is shown in Figure 13. Concrete specimens were placed in a specifically fabricated steel frame that was positioned in the testing machine as shown in the figure. The frame consisted of a bearing plate 25 mm thick connected to the rigid base of the machine with four rods 20 mm in diameter. Prior to load application, specimens were seated on a wooden plate supported on two horizontal steel beams to allow instrumentation prior to testing as shown in Figure 14. Two circular steel plates and a rubber plate were introduced between the concrete specimen and the bearing plate to secure the contact between the top surface of concrete and the bearing plate (Figure 15). This arrangement was necessary to minimize the effect of surface irregularities on the specimen’s alignment and to prevent accidental lateral movement during testing.

25

Figure 13: Pullout test set up

Wooden plate

Figure 14: Wooden base supported on steel channels For each specimen, a steel pipe of 250 mm long was installed at the end of the FRP bar to serve as end-anchor (Figure 16). Anchors were designed according to the CSA-S806-02 (2002) provisions. In order to ensure concentric alignment of the bar inside the anchor, two plastic washers were used as spacers as shown in Figure 17. Prior to testing, all anchors were grouted by a commercially available expansive mortar known as RockFrac. The grout

26

was allowed to cure for 24 hours after casting to reach its maximum compressive and bearing stresses.

Hinge system Figure 15: Hinge system to minimize the effect of surface irregularities

Figure 16: Steel pipe used to anchor the FRP bar

FRP bar

Plastic washer Figure 17: Plastic washers to ensure centricity of the bar

3.4.2 Specimen instrumentation Each specimen was instrumented by three Linear Variable Displacement Transducers (LVDTs) in order to measure the slip in the bar during load application. Two LVDTs 27

measured the bar movement relative to concrete surface at the free (unloaded) end that extended outside the concrete cylinder (Figure 18-a). Another LVDT placed at the top of specimen measured the bar displacement at the loaded end (Figure 18-b). Holes were made in the upper steel plate and the wooden base in order to allow the FRP bar and the LVDTs to pass through the plates. It is worth mentioning that, in the first tests conducted in this study, four LVDTs were used (two at the unloaded end and two at the loaded end of the bar). However, due to the damage of one of these LVDTs, only three LVDTs were used in the remaining tests.

Free end

Loaded end (a)

(b)

Figure 18: LVDTs fixed at the (a) unloaded ends and (b) loaded ends

3.4.3 Specimen testing The instrumented test specimen was positioned in an MTS universal testing machine of 500 kN capacity. Direct tensile load was applied to the bar in a strain-control mode at a rate of 1.2 mm/min according to CSA-S806-02 (2002) guidelines. Test was halted when the bar could carry no more loads. During the test, a data acquisition system recorded the load and slip readings at a rate of 5 readings/sec. 28

Chapter 4: Results and Discussions This chapter presents the results of the pullout tests carried out on the BFRP and GFRP reinforced concrete cylinders. As described in chapter 2, the FRP material, the bar diameter, the embedment length, and the freeze-and-thaw (FT) conditions were considered in this investigation. In this chapter, the effect of each parameter on the bond behavior of FRP bars is presented. Bond-slip curves and failure mechanism of all tested specimens are discussed.

4.1 Phase I results: Bond behavior of unconditioned specimens In this phase, a total of 48 unconditioned specimens were tested at room temperature. The effect of three main parameters (the bar material, the bar diameter, db, and the embedment length, ld) on bond was studied.

4.1.1 Bond-slip relationships The bond strength, τmax, developed by the bar is defined by equation (5) as follows: (Equation 5) where P is the tensile load, d is the bar diameter, and ld is the bar embedment length. At any stage of loading, bar slip values at the unloaded ends were obtained directly from the LVDTs located at the bottom of the tested specimen. For the loaded ends, the elongation of the bar between the upper LVDT support and the beginning of the bonded zone was subtracted from the LVDT measurements. Table 9 shows the results of pullout tests of the unconditioned specimens.

29

Table 9: Results of unconditioned pullout bond specimens Ascending branch Test

Failure mode

Softening branch τr, le PBPE

τmax

smax, fe

smax, le

τons, fe

τons, le

τr, fe

(MPa)

(mm)

(mm)

(MPa)

(MPa)

(MPa)

(MPa)

(MPa)

BFRP specimens B08-40-1 B08-40-2 B08-40-3 B08-56-1 B08-56-2 B08-56-3 B08-80-1 B08-80-2 B08-80-3 B08-120-1 B08-120-2 B08-120-3

P P P P P P P P NA P P P

16.56 15.81 19.81 14.10 14.01 21.60 18.07 11.80 NA 13.74 15.27 10.65

0.01 0.15 0.10 NA 0.12 NA 0.10 0.01 NA 0.20 0.13 0.16

0.45 0.60 0.98 0.00 0.73 1.20 1.07 0.55 NA 1.87 2.16 1.63

14.22 14.60 19.45 12.18 13.25 19.98 17.93 11.79 NA 13.37 14.96 10.40

0.55 1.40 0.38 NA 0.45 0.80 2.30 2.80 NA 1.30 1.30 0.90

10.80 10.00 14.50 11.90 9.50 13.50 13.50 7.50 NA 10.40 9.39 5.01

15.30 11.92 14.34 9.92 9.00 13.12 13.36 7.74 NA 11.05 12.20 7.50

0.030 0.030 0.016 NA 0.030 NA 0.012 0.003 NA 0.020 0.012 0.022

B10-50-1 B10-50-2 B10-50-3 B10-70-1 B10-70-2 B10-70-3 B10-100-1 B10-100-2 B10-100-3 B10-150-1 B10-150-2 B10-150-3

P P P P P P P P P P P P

14.37 13.71 15.30 18.10 17.11 14.64 13.26 12.76 13.92 10.80 12.71 10.48

0.29 0.23 0.05 0.20 0.20 0.20 0.01 0.01 0.16 0.23 0.23 0.17

0.89 0.00 0.11 0.48 0.85 0.85 1.24 1.44 1.47 1.33 1.58 1.83

12.43 11.30 15.69 10.04 9.24 7.98 12.03 11.78 12.89 10.59 12.55 10.22

0.70 NA 0.60 0.40 0.46 0.40 0.80 0.32 0.32 1.50 1.28 0.25

7.30 9.00 10.00 7.60 7.00 7.00 8.50 8.70 NA 5.30 8.25 6.00

9.22 7.96 12.56 6.61 6.70 6.57 12.14 12.19 9.15 6.40 9.30 8.44

0.055 0.045 0.012 0.040 0.040 0.040 0.002 0.001 0.015 0.022 0.020 0.018

B12-60-1 B12-60-2 B12-60-3 B12-84-1 B12-84-2 B12-84-3 B12-120-1 B12-120-2

P P P P P P P P

16.51 17.02 17.40 17.97 12.35 13.87 14.32 14.42

0.20 0.25 0.00 0.00 0.15 0.13 0.28 0.19

0.84 0.56 0.00 0.75 0.66 0.98 0.00 1.36

12.96 13.37 17.21 9.04 5.99 7.22 12.80 14.17

1.30 0.30 0.27 0.10 NA 0.10 NA 0.97

10.00 12.50 12.00 6.10 4.50 6.00 11.00 10.50

10.17 12.57 12.03 6.00 3.90 5.53 NA 10.70

0.027 0.032 0.000 NA 0.030 0.025 0.027 0.017

30

B12-120-3 B12-180-1 B12-180-2 B12-180-3

NA P P P

NA 10.91 11.68 12.18

NA 0.25 0.30 0.25

NA 2.31 2.37 2.14

NA 10.50 11.15 11.83

NA 0.25 0.40 0.40

NA NA 6.00 6.70

NA NA 9.98 10.20

NA 0.022 0.022 0.017

0.78 0.45 1.80 NA 0.70 0.96 0.16 1.00 NA NA 1.30 NA

4.60 12.50 7.00 8.00 6.50 NA NA NA NA NA NA NA

4.90 14.00 6.73 5.83 5.62 0.34 1.05 NA 2.02 NA NA NA

0.300 NA NA NA NA NA 0.140 NA 0.060 NA NA 0.005

GFRP specimens G10-50-1 G10-50-2 G10-50-3 G10-70-1 G10-70-2 G10-70-3 G10-100-1 G10-100-2 G10-100-3 G10-150-1 G10-150-2 G10-150-3

P P P S S S P R P R R R

20.07 21.03 20.44 17.35 18.22 20.75 17.52 15.21 17.88 12.37 11.56 11.71

0.36 0.00 0.02 0.00 0.00 0.00 0.16 0.01 0.12 NA 0.01 0.01

0.03 1.32 0.00 0.92 1.67 1.13 0.00 0.00 0.78 NA 1.78 0.80

12.13 16.69 NA 17.08 18.09 NA 15.45 15.17 15.16 NA NA NA

Note:P = pullout; S = concrete splitting; R = bar rupture; NA = test data not available; τmax = maximum bond stress; smax, le and smax, fe = slip corresponding at maximum stress at loaded and unloaded (free) ends, respectively; τons, le and τons, fe = bond stress at onset of slip at loaded and unloaded ends, respectively; τr, le and τr, fe = residual bond stress at loaded and unloaded ends, respectively.

Figure 19 and Figure 20 depict typical relationships between bond stress and slip histories for BFRP bars having different diameters and embedment lengths at both the unloaded and loaded ends, respectively. Figure 21 compares the bond-slip curves at the unloaded ends of BFRP and GFRP specimens with bar diameter of 10 mm. It can be noticed that bar slip in some specimens was not fully recorded due to unexpected damage of LVDTs. These specimens can be clearly identified from the plots. As seen from the figures, all BFRP and GFRP curves showed an initial ascending branch up to a maximum stress value, τmax. Increase in bond stress was accompanied by an increase in slip between the bar and the surrounding concrete. Bond-slip curves also showed a falling branch, or softening branch, after maximum bond stress was attained. This portion of the curve was characterized by a significant drop in the bond stress accompanied by an increase in the bar slip.

31

The comparison between bond-slip curves of BFRP and GFRP bars (Figure 21) demonstrates the similarity in the bond-slip trend of both types of bars. Ascending and softening branches can be clearly identified from the plots. BFRP bars of 10 mm diameter developed 71%, 89%, and 79% of the GFRP bond strength for embedment lengths of 5, 7, and 10 d, respectively, with an average value of 80%. Figure 22 compares between average values of maximum stress developed by all specimens. The corresponding slip recorded at unloaded ends was negligible for both types of bars (average of 0.15 mm and 0.06 mm for BFRP and GFRP bars, respectively). At the loaded ends, an average slip of 1.02 mm and 0.81 mm was encountered at maximum stress for BFRP and GFRP bars, respectively. At all stages of loading, bar slip at unloaded ends was significantly smaller than that at loaded ends. In fact, slip initiates at the loaded end almost at the beginning of the test after the chemical adhesion breaks between the bar and concrete. Adhesion of FRP bars to concrete is the principal component that describes the bond performance of the bar at initial loading stages. Once adhesion between the bar and concrete breaks, the loaded end starts to slip and friction between the outer layer of the bar and concrete controls the bond mechanism. Achillides and Pilakoutas (2004) reported that adhesion of GFRP and CFRP bars depends on the bar diameter regardless of its fiber material. Average values of 0.82 MPa and 0.86 MPa for both types of bars, respectively, having 8.5 mm diameter were reported (Achillides and Pilakoutas 2004). In the current study, stress at onset of slip at loaded ends was obtained from the bond-slip curves at the points where initial slope of the ascending branch changed abruptly. These values are listed in Table 3 for all the test specimens. An average value of 1.37, 0.67, and 0.51 MPa was determined for BFRP bars of diameters 8, 10, and 12 mm, respectively, compared to an average value of 1.03 MPa for the GFRP bar of diameter 10 mm. Average stress values are also depicted in Figure 23-a in MPa and as a percentage of the ultimate stress attained. These findings demonstrate that adhesion developed by BFRP bars to concrete is about 65% of that developed by GFRP bars.

32

(a)

(b) 33

(c)

(d) Figure 19: Typical bond-slip curves for BFRP bars at unloaded ends: (a) ld = 5d, (b) ld = 7d, (c) ld = 10d, and (d) ld = 15d 34

(a)

(b)

35

(c)

(d) Figure 20: Typical bond-slip curves for BFRP bars at loaded ends: (a) ld = 5d, (b) ld = 7d, (c) ld = 10d, and (d) ld = 15d

36

(a)

(b) 37

(c)

(d) Figure 21: Comparison between bond-slip curves for BFRP and GFRP bars at unloaded ends (a) ld = 5d and (b) ld = 7d, (c) ld = 10d, and (d) ld = 15d 38

Figure 22: Ratio of the bond stress to the ultimate strength at slip onset at loaded end Adhesion values obtained for BFRP bars in Figure 23-b also confirm the findings of Achillides and Pilakoutas (2004) that adhesion depends on the bar diameter. Test results show that the stress needed to mobilize slip of BFRP bars of small diameters is larger than that needed for bars of large diameters. For instance, the 8 mm diameter BFRP bars slipped at a stress of 1.37 MPa representing 9% of its maximum bond stress, while the 12 mm bars slipped at a stress of 1.03 MPa (6% of its maximum stress). Similar trend was observed at the unloaded ends of BFRP bars. In fact, the slip recorded at the unloaded ends of both BFRP and GFRP bars remained practically zero until the bond stress reached levels close to the bond strength of the bar. The ratio of bond stress at onset of slip of the unloaded ends to ultimate stress is shown in Figure 23-b. Bars with small diameter encountered slip at higher values of stresses than those of large diameters. Bond stress of 8 mm BFRP bar reached 14.77 MPa (95% of its ultimate stress) before the bar started to slip at its unloaded end. This value dropped to 11.74 MPa (81% of maximum stress) for 12 mm diameter bars. On the other hand, the stress that mobilized slip of the 10 mm BFRP and GFRP bars was 11.67 MPa and 14.72 MPa, representing 84% and 87% of 39

the corresponding ultimate stress, respectively. These values are listed in Table 9 for all the test specimens. These values conform well to the values reported by Achillides (1998) and Achillides and Pilakoutas (2004).

40

(a)

(b) Figure 23: Ratio of bond stress to ultimate stress at onset of slip at (a) loaded ends and (b) unloaded ends 41

4.1.2 Residual Bond Strength Residual stress, or post-maximum stress, describes the bond performance of FRP bar after reaching its ultimate bond stress. It indicates the additional resistance of the bar to the applied pullout load along the softening branch. Residual stresses for both BFRP and GFRP bars are determined by visual inspection of the bond-slip curves of all test specimens. It was evaluated as the bond stress recorded on the softening branch before the curve flattens out (e.g. point A in Figure 21-a) or dramatically changes its slope (e.g. point B in Figure 21-b). Values of residual stress at the unloaded and loaded ends are listed in Table 9. Figure 24 shows the residual stress obtained for all BFRP specimens at both ends. Test results showed that all BFRP specimens encountered an average residual stress of 9 MPa and 10 MPa, which represents 61% and 68% of peak stress attained at their unloaded and loaded ends, respectively. Residual stress appears to be independent of the bar diameter. The resistance of the bar post the peak stress is rather attributed to the friction resistance of the embedded length of the bar in concrete, which mainly depends on the surface treatment of the FRP bar, in addition to the resistance of the undamaged part that enters the embedded zone when the bar starts to slip. GFRP bars showed an average residual stress of 7.7 MPa and 5 MPa representing 40% and 30% of the peak stress at unloaded and loaded ends, respectively. This low post peak stress was attributed to the failure of the sand-coated surface and its separation from the bar core. Friction resistance provided by the resulting smooth surface could not provide enough resistance to the pullout force. In fact, GFRP bars showed an abrupt drop in bond stress immediately after reaching the ultimate stress. This was accompanied by a significant energy release and sudden explosion. During this short time, no slip values are recorded. This is shown in Figure 21-b for specimen G10-70-1 where the LVDT recorded a negligible slip value at the ultimate stress of 17.35 MPa before it slips 4.6 mm at a stress value of 6.7 MPa. On the other hand, BFRP specimens showed a smooth transition between the ascending and descending parts of the bond-slip curve. Contrary to GFRP specimens, no energy release was encountered when the maximum stress was attained in the case of 42

BFRP bars. This was attributed to the gradual and partial delamination of the interface between the sand-coated layer and the subsequent layers of the bar, as will be detailed in the following section. Another important observation from the tests is that slip hardening was observed in both types of bars, but was more pronounced in the GFRP ones. In some cases, GFRP specimens restored almost their full or high percentage of their bond strength (Figure 21-b). Slip hardening can be attributed to (a) the development of friction between the embedded portion of the bar and concrete, which was more pronounced in the case of BFRP bars, (b) the contribution of the undamaged part that entered the embedded zone, in addition to (c) the entrapment of concrete and polymer residues between the bar and the surrounding concrete.

43

(a)

(b) Figure 24: Ratio of average residual to maximum bond stress for BFRP and GFRP bars at (a) unloaded ends and (b) loaded ends 44

4.1.3 Bond failure mode Failure modes of all specimens are listed in Table 9. All BFRP bars failed in pullout mode as shown in Figure 25-a for specimen B10-100-2. No visual cracks are observed on the BFRP-reinforced cylinders even for specimens with large embedment length (15d). On the other hand, GFRP specimens with embedded length of 5 and 10 times the diameter failed by pullout of the bars, except one specimen where the bar ruptured before attaining its bond strength (specimen G10-100-2 in Figure 25-b). Failure of this specimen was attributed to the misalignment of the bar during the test.

(a)

(b)

(c)

Figure 25: Typical modes of failure (a) pullout of the bar (specimen B10-100-2), (b) bar rupture (specimen G10-100-2), and (c) concrete splitting (specimen G1070-2) GFRP-reinforced specimens with bar embedment length of 7 and 15 times the diameter exhibited different failure modes. All specimens with ld= 7d failed in a splitting mode where concrete cracked before the bar was pulled out the specimen. This mode of failure is depicted in Figure 25-c for specimen G10-70-2. This finding was inconsistent with the pullout mode of failure of GFRP specimens with longer embedment (e.g. specimens G1045

100-1 and G10-100-3) and was attributed to variation in the concrete strength or to the usual variation in bond tests. On the other hand, GFRP bars with ld= 15d ruptured before bond strength was attained. For specimens with such long embedment length, the pulling force was not capable to overcome the adherence between the bar and concrete. Therefore, the tensile strength of the bar was achieved and the bar ruptured before bond strength was mobilized. This mode of failure of GFRP bars of embedment length 15d indicates that GFRP bars supersede their counterparts in interfacial shear resistance between the grained layer and the core layers of the bar. At the end of each test, the concrete cylinder was split to visually assess the condition of the bar and the surrounding concrete. Figure 26-a and Figure 26-b show the condition of BFRP bars and concrete for specimens B8-120-2 and B8-56-1, respectively. Both specimens failed by pullout of the BFRP bar.

46

(a)

(b)

(c) Figure 26: Pullout mode of failure of (a) BFRP specimen B8-120-2, (b) BFRP specimen B8-56-2, and (c) GFRP specimen G10-70-3 47

As shown from the figure, the bar surface was significantly damaged at the loaded end and the outer layer was entirely peeled off (delaminated) from the subsequent bar layers. White powder was detected close to that end and along the whole embedded length indicating traces of crushed resin. Close to the unloaded end, the surface layer of the bar was partially peeled off. At this end, parts of the sand-coated layer could be seen attached to the bar. No apparent damage to concrete surrounding the bar was observed. These findings suggest that bond failure took place along the interfacial shear surface between the grained layer and the subsequent core layers of the bar. Failure was therefore governed by the shear strength along the fibres interface rather than the shear strength between the bar and concrete. This mode of failure was anticipated in case of high strength concrete used in this study (Tastani et al. 2006; Davalos et al. 2008). Damage concentration at the loaded end suggested that peeling off of the grained layer started at that end and continued along the embedded portion of the bar until failure occurred. High stress that is usually encountered at the loaded ends explains the damage concentration on the surface of the bar at that end. GFRP specimens that failed by pullout of the bar showed a different bond failure mechanism. Contrary to what was observed for BFRP bars, a “uniform” peeling off of the surface layer was noticed along the whole length of the embedded portion of the bar. Figure 26-c depicts this mode of failure for specimen G10-70-3. This observation explains the abrupt nature of failure encountered for GFRP specimens where huge energy release followed the sudden delamination of the grained surface of the bar. Moreover, it clarifies the low residual stress that GFRP bars developed in comparison to that of BFRP bars. The complete delamination of the grained surface resulted in a smooth surface incapable of resisting the increasing pullout force applied. In this case, the residual stress was mostly developed by the undamaged part of the bar that entered the embedded zone in addition to the entrapped concrete and polymer residues between the bar and concrete.

48

4.1.4 Factors Affecting Bond Behavior of FRP bars 4.1.4.1 Embedment length Relationships between the bond strength and the embedment length are shown in Figure 27 for GFRP and BFRP bars of different bar diameters. For each embedment length, ultimate bond stress of three tested specimens is shown. The plot shows the best-fit trend lines obtained by linear regression of the average stress values estimated from the test results. R-squared (R2) of best-fit lines are also shown to describe how well the regression line fits the recorded data. From Figure 27 it can be seen that both BFRP and GFRP bars revealed similar trend. Bond strength of both types of bars is inversely proportional with the embedment length of the bar, i.e. increasing the embedment length results in a decrease in the bond strength of the bar. This observation is valid for all test specimens regardless of the bar diameter. The decrease in bond strength with the bar embedment length is confirmed by many studies (Achillides and Pilakoutas 2004; Sayed et al. 2011) and is thought to be a result of two main factors namely (a) the nonlinear distribution of the bond stress along the embedded portion of the bar, which increases with the increase of the embedment length and (b) the reduction in the bar diameter due to Poisson’s ratio effect, which leads to a reduction in friction along the embedment length. Boyle et al. (1994) reported that this reduction in bond strength is more pronounced for small cross-section reinforcing bars. Tests results of BFRP bars confirmed this finding as illustrated in Figure 28-a. Trend lines of the 8 mm bars showed a steeper slope as compared to that observed for the 10 mm and 12 mm bars.

49

(a)

(b)

50

(c)

(d) Figure 27: Bond strength versus embedment length for (a), (b), (c) BFRP bars and (d) GFRP bars 51

(a)

(b) Figure 28: Average bond strength versus embedment length for (a) BFRP bars and (b) BFRP and GFRP bars of diameter 10 mm 52

A comparison of the influence of the embedment length on the ultimate bond stress of BFRP and GFRP bars is shown in Figure 28-b. Difference in slopes of both trend lines indicates that the influence of bar embedment length is more pronounced in the case of GFRP bars. If the embedded length of the bar increases, GFRP bars of same diameter loose their friction resistance in a much abrupt way than their counterparts of BFRP bars. This trend is not valid at longer embedment lengths where failure modes change dramatically from pullout (e.g. GFRP specimens of ld = 5d) to bar rupture (e.g. GFRP specimens of ld = 15d). 4.1.4.2 Bar type and modulus of elasticity Previous experimental results reported that bond strength significantly varies with the bar type (Achillides and Pilakoutas 2004; Aiello et al. 2007; Wang and Belarbi 2010). The type of fibres and resins used in manufacturing FRP bars affect their mechanical properties and therefore affect their bond to concrete. Influence of the bar type on the bond stress has been demonstrated from the current test results. As previously stated, BFRP bars developed average bond strength of 80% of that developed by GFRP bars (Figure 22). The lower bond strength of BFRP bars can be attributed, in addition to other factors, to the lower modulus of elasticity of BFRP bars (48 GPa for BFRP versus 53 GPa respectively). If other parameters are fixed, FRP bars with high modulus will likely develop high bond strength to concrete. 4.1.4.3 Bar diameter Diameter effect for FRP bars has been demonstrated by other researchers (Benmokrane et al. 2006; Cosenza et al. 1997; Tighiouart et al. 1999; Achillides and Pilakoutas 2004) for G- and C-FRP bars. It was concluded that the larger the diameter of the bar, the less bond strength developed during the tests. This is attributed to the nonlinearity of the stress distribution exerted along the bar, which is more pronounced in case of large diameters as large embedment lengths are needed (Arias et al. 2012). In case of steel reinforcing bars, the loss of bond strength is explained by the high quantity of bleeding water trapped beneath the rebar of large diameters, which leads to the creation of air voids underneath the bar, therefore reducing the contact area between the bar and concrete. 53

Figure 29 compares the bond strength of 8 mm, 10 mm, and 12 mm BFRP bars for different embedment lengths.

Figure 29: Influence of the bar diameter on the average bond strength of BFRP bars It is observed that bars of large diameter developed less average bond strength than those of small diameters. BFRP bars of 8 mm diameter showed bond strength of 16%, 0%, 11%, and 14% higher than the bond strength developed by the 10 mm diameter bars for embedment lengths of 5, 7, 10, and 15d, respectively, with an average increase of 10%. These percentages are 3%, 11.5%, 4%, and 12% more than the bond strength developed for the 12 mm diameter bars for the same embedment lengths, with an average increase of 7.6%. Despite the variation that exists in some of these results, which can be attributed to the variation in concrete strength of the test specimens or to the inconsistency in the mechanical properties of the bar, the general trend conforms with the finding of other

54

researchers that bars of large diameter develop less average bond strength than those of small diameters. 4.1.4.4 Surface treatment Surface deformations play a major role in developing friction forces between FRP bar and concrete. Deformations could consist of just resin, fiber-reinforced resins, or resin containing longitudinal continuous fibres (fib report 2000). BFRP and GFRP bars used in this study are sand-coated on their surface with spiral indentations existing on the outer layer of BFRP bars. Indentations are created during the manufacturing process by winding the bar with separate filaments. Epoxy and vinyl ester resins are used in bonding the fibres in BFRP and GFRP bars, respectively. As previously mentioned, test results indicate that bond strength of BFRP bars was less than that of GFRP bars. Moreover, failure of both types of bars always occurred at the interface between the outer layer of the bar and its core, which suggested that bond was governed by the bar itself irrespective of surrounding concrete. The effect of surface treatment is noticed in the failure mode of both types of bars. GFRP bars coated with a uniform layer of sand showed good performance until the peak bond stress was achieved. The bar then failed abruptly as a result of the sudden detachment of the interface between sand grains and bar. For BFRP bars, spiral indentation created grooved segments at their locations on the bar surface, which prevented the continuity of the sanded coating surface. This explains the smooth mode of failure obtained for BFRP specimens compared to the brittle failure of GFRP ones. In fact, these indentations are supposed to play the role of ribs in FRP ribbed bars or lugs in steel rebars where mechanical interlocking could be activated. In the authors’ opinion, the influence of these indentations on the bond strength of the bar is not realized. Test results showed partial delamination of BFRP bars as shown in Figure 26-a and Figure 26-b. It is thought that no interlocking mechanism has been developed, which is confirmed by the fact that concrete surrounding the bar remained uncrushed. This mode of failure suggests that adhesion and friction remain the principal components in developing the bond strength of the BFRP bar.

55

This result is in agreement with the results reported by Benmokrane et al. (2006) for GFRP bars having helical windings on their surface.

4.2 Phase II: Effect of Freeze and Thaw In this phase, a total of 24 cylinders reinforced with BFRP and GFRP bars were tested. The main parameters were the bar type, the bar diameter, and the number of FT cycles (100 and 200 cycles). Only one embedment length, 7ld, was considered for all specimens. Table 10 shows the results of pullout tests of the conditioned specimens. The results of the relevant unconditioned specimens that have been previously reported in Table 9 are also shown for comparison. Table 10: Results of unconditioned and conditioned specimens Specimens

Failure Mode

τmax (MPa)

smax,fe (mm)

smax,le (mm)

τr, fe (MPa)

τr, le (MPa)

Unconditioned BFRP specimens, ld = 7 d B8-56-1

P

14.10

NA

0.00

11.90

9.92

B8-56-2

P

14.01

0.12

0.73

9.50

9.00

B8-56-3

P

21.60

NA

1.20

13.50

13.12

B10-70-1

P

18.10

0.20

0.48

7.60

6.61

B10-70-2

P

17.11

0.20

0.85

7.00

6.70

B10-70-3

P

14.64

0.20

0.85

7.00

6.57

B12-84-1

P

17.97

0.00

0.75

6.10

6.00

B12-84-2

P

12.35

0.15

0.66

4.50

3.90

B12-84-3

P

13.87

0.13

0.98

6.00

5.53

Conditioned BFRP specimens, 100 cycles B8-56-100-1

P

18.57

0.27

0.90

16.80

16.80

B8-56-100-2

S

16.65

0.06

0.00

12.60

12.50

B8-56-100-3

P

15.90

0.11

0.67

15.43

14.82

B10-70-100-1

P

16.16

0.01

1.00

5.25

5.48

56

B10-70-100-2

P

19.64

0.14

0.74

15.97

15.95

B10-70-100-3

P

18.91

0.18

0.76

14.25

14.42

B12-84-100-1

S

17.62

0.14

0.85

14.00

14.00

B12-84-100-2

S

13.15

0.33

1.22

11.30

11.30

B12-84-100-3

P

19.71

0.15

1.00

15.70

15.70

Conditioned BFRP specimens, 200 cycles B8-56-200-1

P

16.79

0.13

0.84

14.40

14.40

B8-56-200-2

P

22.67

0.17

1.14

20.32

20.32

B8-56-200-3

P

16.53

0.04

0.63

13.30

13.30

B10-70-200-1

P

16.46

0.01

0.84

12.04

12.04

B10-70-200-2

P

17.67

0.09

0.91

4.00

4.00

B10-70-200-3

P

16.39

0.16

0.02

10.82

10.82

B12-84-200-1

P

14.14

0.12

0.55

10.30

10.30

B12-84-200-2

P

16.42

0.18

0.88

11.05

11.05

B12-84-200-3

P

15.32

0.11

0.77

13.47

13.47

Unconditioned and conditioned GFRP specimens G10-70-1

S

17.35

0.00

0.92

8.00

5.83

G10-70-2

S

18.22

0.00

1.67

6.50

5.62

G10-70-3

S

20.75

0.00

1.13

NA

0.34

G10-70-100-1

P

20.25

0.17

0.71

1.96

1.96

G10-70-100-2

P

20.45

0.11

0.60

4.14

4.14

G10-70-100-3

P

19.76

0.14

1.80

2.52

2.52

G10-70-200-1

P

21.53

0.19

1.71

3.27

3.27

G10-70-200-2

P

20.42

0.08

1.68

3.58

3.58

G10-70-200-3

P

20.56

0.08

1.46

2.48

2.48

τmax = maximum bond stress; smax, le and smax, fe = slip corresponding at maximum stress at loaded and unloaded ends, respectively; τr, le and τr, fe = residual bond stress at loaded and unloaded ends, respectively; P = pullout; S = concrete splitting; R = bar rupture; NA = test data is not available.

In the following sections, the effect of FT cycles on the bond performance of the bars is presented. Prior to discussion of the current results, it is worth mentioning that the few studies that have reported on the FT effect on bond of FRP bars to concrete have shown 57

contradictory results. Based on the results obtained from these studies, bond degradation of FRP bars can be explained as follows: a) Concrete deterioration due to FT cycles: When subjected to FT cycles, concrete is saturated with water during the thawing phase. In the freezing phase of the cycle, water expands and causes distress in concrete if there is no space to accommodate the volume expansion. If the stresses induced exceed the tensile strength of concrete, surface and internal cracks appear. Deterioration and spalling of concrete occur depending on the severity of damage associated with the subsequent cycles of FT. Cracking is of great concern when it occurs at the interface of FRP bar and concrete. The presence of cracks causes that some locations along the outer layer of the bar are not in contact with surrounding concrete. With the subsequent cycles of FT, cracks increase in number and width and the expansion process continues, which results in loss of bond and the vulnerability of the bar to be pulled out under low applied forces. b) FRP degradation due to FT cycles: Degradation of FRP bars due to FT cycles is a result of two mechanisms that occur while the bar is embedded in concrete as follows: 1- Moisture absorption during the thawing part of the FT cycle: when FRP bars are subjected to moisture during FT cycles, water diffuses through the FRP matrix. Polyester, vinyl ester, and epoxy are the most commonly used matrices in manufacturing FRP bars. Depending on the resin type, the quality of manufacturing, and the surrounding temperature, deterioration of the resin may occur (Tannous and Saadatmanesh. 1999; Karbhari 2002). Bank et al. (1998) reported severe degradation of GFRP bars with vinyl ester matrix when submerged in tap water for 14 and 84 days. The absorbed water molecules expand during the freezing phase. The increase in volume results in cracks and consequently additional water is absorbed inside the matrix. Cracks that form on the 58

surface layer of the FRP bar in contact with concrete make the bar vulnerable to loss of bond to concrete, as previously mentioned. 2- Temperature cycling: FRP bars have a higher coefficient of thermal expansion (CTE) than concrete in the transverse direction. At high temperatures, the FRP bar embedded in concrete expands causing radial bursting stresses. If these stresses are higher than the tensile strength of concrete, cracks develop and bond between the bar and concrete is negatively affected. When subjected to FT cycles, FRP bar is exposed to an elongated period of freezing temperature that causes higher contraction in the bar than in concrete. Gaps between the bar and concrete develop resulting in loss of bond between the two materials (Masmoudi et al. 2005; Belarbi et al. 2012). It is worth mentioning that bond deterioration can be a result of the combined effect of both mechanisms. The mechanical characteristics of concrete and FRP bar dictate the influence of each factor on the bond degradation of the bar.

4.2.1 Effect of freeze-and-thaw on bond-slip response Figure 30 and Figure 31 show the bond-slip curves of BFRP and GFRP specimens at unloaded and loaded ends, respectively, at room temperature and after being exposed to 100 and 200 cycles of FT. Similar patterns of bond-slip curves are noticed for both unconditioned and conditioned specimens. Unloaded ends of both unconditioned and conditioned BFRP specimens showed almost zero slip until maximum bond stress was attained. This is illustrated in Figure 31 a, b, and c for bar diameters of 8, 10, 12 mm, respectively. The descending curves of the unconditioned and conditioned specimens are also comparable. For both specimens, bar slip increased abruptly post the peak stress and therefore increased gradually with little or no change in the stress applied. For GFRP specimens, FT cycles have not changed the abrupt failure of the specimens and the considerable energy release, which was previously reported for the unconditioned ones. This is seen in the curves of Figure 30-d.

59

Similar bond-slip curves obtained from the tests indicated that FT cycles have a negligible effect on the bond-slip relationship for both BFRP and GFRP bars. This finding has been previously reported by Chen et al. (2007) for GFRP and CFRP-reinforced pullout specimens.

60

25

B8-56-2 B8-56-200-1 B8-56-100-2

Bond stress, MPa

20 15 10 5 0 0

5

10

15

Slip, mm (a) 25

B10-70-100-2 B10-70-200-1

Bond stress, MPa

20

B10-70-1

15

10 5 0 0

5

10

15

Slip, mm (b)

61

25

B12-84-1 B12-84-100-3 B12-84-200-2

Bond Stress, MPa

20 15 10 5

0 0

5

10

15

Slip, mm (c) 25

G10-70-1 G10-70-100-2 G10-70-200-3

Bond stress, MPa

20 15 10 5 0 0

5

10

15

Slip, mm (d) Figure 30: Typical Bond-Slip curves for (a), (b) and (c) BFRP and (d) GFRP at unloaded ends

62

Bond stress, MPa

25

B8-56-100-3 B8-56-200-1 B8-56-2

20 15 10 5 0 0

5

10

15

Slip, mm (a)

Bond stress, MPa

25

B10-70-1 B10-70-100-2 B10-70-200-1

20 15 10 5 0 0

5

10

15

Slip, mm (b)

63

25

B12-84-1 B12-84-100-3

Bond stress, MPa

20

B12-84-200-2

15 10 5 0 0

5

10

15

Slip, mm (c) 25

G10-7-3 G10-7-100-2 G10-7-200-1

Bond stress, MPa

20 15 10 5 0 0

5

10

15

Slip, mm (d) Figure 31: Typical Bond-Slip curves for (a), (b) and (c) BFRP and (d) GFRP at loaded ends 64

4.2.2 Effect of FT cycles on maximum bond stress Table 10 lists the maximum bond stress for all specimens. Figure 32 compares the maximum stress of both unconditioned and conditioned specimens based on the average values obtained for each set of tests. It is worth mentioning that Figure 30 and Figure 31 show the bond-slip curves of representative samples of each set and do not represent average results. From the plot, one can notice the discrepancy in the test results obtained. Results of BFRP specimens with diameters 10 mm and 12 mm showed that specimens exposed to 200 FT cycles exhibited bond strength higher than that of the unconditioned ones and less than that of the specimens conditioned for 100 FT cycles. This contradicts the findings of BFRP specimens of 8 mm diameter and those of the GFRP specimens of 10 mm diameter. The results of the later groups showed that conditioned specimens had higher bond strength than that of the control ones. This discrepancy in results is usually encountered in bond durability studies and has been reported by previous researchers (Chen et al. 2007; Belarbi et al. 2012). It can be attributed to the random nature of the development of the degradation of the FRP bar and concrete (Bank et al. 1998; Belarbi et al. 2012). However, in order to understand this discrepancy in the test results, normalized bond strength of all specimens was calculated. It is normal practice that concrete compressive strength, f’c, increases with longer and better curing that occurred during the exposure to FT cycles. Therefore, the compressive strength for both unconditioned and conditioned specimens was measured.

65

Figure 32: Average bond stress for unconditioned and conditioned specimens Table 11 shows the results of f’c for concrete cylinders tested before and after exposure to FT cycles. It can be noticed that concrete strength increased about 11.5% after being exposed to 200 cycles. In order to eliminate the effect of concrete strength on the maximum bond strength of the bars, the obtained stresses were divided by the square root of f’c. The normalized bond strength values are shown in Figure 32. Table 11: Compressive strength of concrete after freeze-and-thaw

66

Specimens

f’ (before conditioning,

f' (after 100 cycles,

f' (after 200 cycles,

1 2 3

MPa) 49.6 49.0 49.4

MPa) 52.9 53.0 53.1

MPa) 54.8 55.3 54.9

Average

49.3

c

c

53.0

c

55.0

Figure 33: Normalized bond strength for unconditioned and conditioned specimens A comparison between the maximum bond stress shown in Figure 32 and the normalized values of Figure 33 shows that the effect of concrete compressive strength on bond of the bars is negligible. Similar trend can be observed in both plots. For instance, BFRP bars of 10 mm and 12 mm conditioned for 100 FT cycles showed higher bond strength than that of the unconditioned specimens. Similar trend was also observed for GFRP bars where conditioned specimens showed better bond performance than the control ones. This was also the case for BFRP bars of 8 mm diameter where bond enhancement can be shown for specimens conditioned for 200 FT cycles. This suggests that concrete curing during the FT cycles had not contributed to the bond enhancement encountered in the tested specimens. In this aspect, it is worth mentioning that moderately high strength concrete (50 MPa and higher) with low water-to-cement ratio and high density was used in these tests. High strength concrete is known by its stable mechanical properties even in harsh environments (note that the increase in strength after 200 FT cycles was 11.5% of the initial strength). Deterioration of concrete is therefore unlikely to contribute to the loss of bond of the bars. This statement is also supported by the experimental observations of the tested 67

specimens after failure where no signs of concrete deterioration are observed at the interfacial surface between FRP bars and concrete. This finding backs the conclusion that no apparent dependence of bond strength on the variation in concrete strength due to FT cycles. Therefore, it is proposed that variation in the bond strength of the conditioned specimens is attributed to factors related to the FRP bar itself and not to concrete. In fact, this was purposely planned to primarily study the performance of the bar regardless of the concrete compressive strength. The increase in bond strength is therefore attributed to the enlargement in cross section of the bar as a result of water absorption. This finding was previously reported in El Refai 2013 for the same BFRP bars used in the current study. An increase between 5% and 7% in the bar diameter of the bars was encountered after submerging the bars in saline and alkaline solutions. Therefore, it is suggested that the increase in bar diameter has enhanced the friction mechanism of the bar, which improved its bond resistance. The effect of moisture absorption on the bond strength of FRP bars has been reported by previous researchers. Alves et al. (2011) reported an increase of 40% in the bond strength of GFRP pullout specimens subjected to the combined effect of sustained load and FT cycles compared to their counterpart control specimens. The authors concluded that FT conditioning and sustained load have enhanced the bond properties of the specimens. Enhancement of bond strength of FRP bars was attributed to water absorption in the bar. The swelling phenomenon, i.e., the increase in bar diameter after submersion in water or solutions, was also reported by Davalos et al. (2008). The authors observed that swelling have increased the mechanical interlocking of FRP to concrete and therefore increased their bond strength. Based on the manufacturers’ data, the moisture absorption for BFRP bars is 0.47% compared to 0.38% for GFRP bars. This explains the difference in percentage of bond increase that the BFRP and GFRP bars of 10 mm diameter exhibited after being conditioned for 100 FT cycles as compared to their unconditioned counterparts. This percentage of increase was 6% for BFRP bars and 3.7% for GFRP bars. After 200 cycles of FT, the BFRP bars exhibited a 4% loss of bond while the GFRP bars showed a bond stress 68

increase of 5.3%. It seems that, in case of BFRP bars, the detrimental effect of subsequent FT cycles had offset the effect of swelling of the bars, which can be attributed to two combined deterioration mechanisms as follows. The first mechanism is related to the water absorption of the FRP matrix during the thawing part of the FT cycle, which led to the degradation of the fibre-resin interface. This effect became more pronounced in cases of long periods of conditioning. Micro cracks grew as a result of FT cycles, which increased the water absorption into the matrix, thus leading to loss of bond strength. This phenomenon was reported by previous researchers on GFRP bars (Dutta and Taylor 1989; Karbhari et al. 2000; Karbhari 2002). The second mechanism is due to the mismatch of the CTE between FRP bar and concrete, which led to the creation of gaps along the embedded length of the bar. These gaps played the role of air pockets that prevented the bars to be in full contact with surrounding concrete, which resulted in loss of bond between the bar and concrete. This phenomenon was also reported by Davalos et al. 2008. It is worth mentioning that the increase or decrease of the bond strength due to water absorption and/or the variation in CTE between FRP and concrete depends on the severity of damage associated with each phenomenon. In some cases, the enhancement in the bar friction resistance due to its diameter increase countered the effect of micro cracks in the FRP bars due to swelling and the mismatch of CTE, or vice versa. This discrepancy in results is due to the random exposure in most durability studies and the random development of the degradation of the bar as previously mentioned (Bank et al. 1998; Belarbi et al. 2012). However, it can be concluded that factors related to the bar degradation had the dominant effect on the bond variation rather than the concrete deterioration or change in its compressive strength. Figure 34 and Figure 35 show the slip of the unconditioned and conditioned bars at maximum stresses. As seen from the plots, insignificant slip values were encountered at the unloaded ends (less than 0.5 mm) for all specimens. This finding confirms well with the bond strength encountered for conditioned specimens for which it was concluded that FT 69

cycles had a negligible effect on their bond strength. On the other hand, slip values recorded at the loaded ends reveal that slip increased with the increase of the maximum stress attained for 8 mm BFRP bars. Specimens exposed to 200 FT cycles exhibited 0.87 mm slip compared to 0.64 mm for unconditioned specimens. This is consistent with the increase in bond stress encountered for the conditioned specimens (Figure 32 and Figure 33). Results obtained for BFRP bars of 10 mm and 12 mm diameter showed that specimens subjected to 100 cycles of FT encountered an increase in slip of about 14% while those exposed to 200 FT cycles showed 20% less slip. These results are also consistent with the increase in bond stress of the former specimens and the reduction in bond stress for the later specimens. It is therefore suggested that, as mentioned earlier, stiffness of FRP bars exhibited insignificant changes as a result of FT exposure.

Figure 34: Slip at unloaded ends for unconditioned and conditioned specimens

70

Figure 35: Slip at loaded ends for unconditioned and conditioned specimens

4.2.3 Effect of FT cycles on bar adhesion to concrete The bond stress at onset of slip at the loaded ends is the stress threshold at which the chemical adhesion breaks between the bar and concrete. Table 12 summarizes the average adhesion stress as obtained for the unconditioned and conditioned specimens. Normalized values are also listed for comparison between specimens of different concrete strengths. Figure 36 shows the ratio of adhesion to the ultimate stress attained by the specimens. It is worth mentioning that adhesion values are estimated by visual inspection of the bond-slip curves as obtained from the test results. Therefore, high accuracy in slip measurements can hardly be expected especially at the beginning of the tests when slip starts at the loaded ends.

71

Table 12: Stress at the onset of slip at the loaded end for BFRP and GFRP specimens Conditioning Quantity Unconditioned Adhesion, MPa Normalized adhesion, MPa Ratio to ultimate stress, % 100 FT cycles Adhesion, MPa Normalized adhesion, MPa Ratio to ultimate stress, % 200 FT cycles Adhesion, MPa Normalized adhesion, MPa Ratio to ultimate stress, %

B8 0.63 0.09 3.77% 0.30 0.04 1.74% 0.38 0.05 2.02%

B10 0.42 0.06 0.36% 0.32 0.04 0.24% 0.25 0.03 0.20%

B12 0.10 0.01 0.10% 0.12 0.02 0.10% 0.19 0.03 0.16%

G10 0.83 0.12 0.63% 0.23 0.03 0.16% 0.33 0.04 0.22%

Figure 36: Ratio of adhesion to ultimate stress for BFRP and GFRP specimens It can be observed from Table 12 and Figure 36 that FT cycles had some influence on the adhesion stress developed between FRP bars and concrete. For instance, onset slip at the loaded end of unconditioned BFRP bars of 10 mm diameter was encountered at a stress of 0.42 MPa representing 2.53% of the ultimate stress attained by the bar. This stress 72

dropped to 0.32 MPa (1.77% of ultimate stress) after 100 FT cycles. It dropped further to 0.25 MPa (1.46% of ultimate stress) after being exposed to 200 cycles. Another important observation from the tests is that unconditioned FRP bars of smaller diameters developed greater adhesion with surrounding concrete than larger bars. BFRP bars of diameters 8, 10, and 12 mm showed an adhesion stress of 0.63, 0.42, and 0.1 MPa, respectively (Table 12). This confirms the findings of Achillides and Pilakoutas (2004). It can also be observed that bars with smaller diameters encountered the biggest loss of adhesion due to FT cycles. BFRP bars of 8 mm diameter lost more than 50% of its adhesion stress to concrete after 100 cycles of FT, while BFRP bars of 10 mm diameter lost about 30% of their adhesion stress. Moreover, GFRP bars of 10 mm diameter lost more than 70% of their adhesion to concrete after 100 cycles compared to 30% for BFRP bars of equal diameter. This suggests that adhesion is dependent on the fibre materials contrary to what has been reported by Achillides and Pilakoutas (2004).

4.2.4 Effect of FT cycles on residual bond stress As previously mentioned, residual stress describes the bond performance of the FRP bar after reaching its ultimate bond stress. It indicates the additional resistance of the bar to the applied pullout load along the softening branch. Values of residual stresses at the unloaded and loaded ends are listed in Table 10. These values are also shown respectively in the plots of Figure 37 and Figure 38. Figure 37 shows the average value of the residual bond stress obtained for all BFRP and GFRP bars at their unloaded ends before and after being exposed to 100 and 200 cycles of FT. Residual stresses for both BFRP and GFRP bars are determined by visual inspection of the bond-slip curves of all test specimens. It was evaluated as the bond stress recorded on the softening branch before the curve flattens out or dramatically changes its slope.

73

Figure 37: Average residual bond strength of unconditioned and conditioned BFRP and GFRP bars at the unloaded ends

Figure 38: Average residual bond strength of unconditioned and conditioned BFRP and GFRP bars at the loaded ends 74

Test results showed that BFRP bars encountered an increase in the average residual stress attained after being conditioned. This is noticed from the values plotted in Figure 37 and Figure 38 for BFRP bars of 8 mm diameter as a typical example. At the unloaded ends, control specimens showed a residual stress of 11.63 MPa representing about 72% of its maximum stress. This value increased to 14.94 MPa (88% of its maximum stress) and 16.01 MPa (85% of its maximum stress) after exposure to 100 and 200 FT cycles, respectively. The bars also encountered different percentages of increase in their residual stresses developed at their loaded ends. Values encountered for bars of diameter 10 mm and 12 mm showed some inconsistency compared to those of 8 mm diameter. An increase in residual stress after 100 FT cycles is noticed. After exposure to 200 cycles of FT, the percentage of increase in residual stress diminished. In general, the residual stress of conditioned specimens exceeds that of the unconditioned specimens in all BFRP test results. Residual stress appears to be independent of the bar diameter. The resistance of the bar post the peak stress is rather attributed to the friction resistance of the embedded length of the bar in concrete in addition to the resistance of the undamaged part that enters the embedded zone when the bar starts to slip. Friction resistance mainly depends on the surface treatment of the FRP bar. The increase in residual stress after exposure to FT cycles can be attributed to the increase in the bar diameter due to water absorption as previously described, which have increased the friction resistance of the bar with concrete. On the other hand, FT cycles did not affect the trend of GFRP bars in terms of the low residual stresses encountered and the abrupt drop in bond stress immediately after reaching the ultimate stress. Unconditioned GFRP bars showed an average residual stress of 7.25 MPa and 3.93 MPa representing 27% and 22% of the peak stress at unloaded and loaded ends, respectively. FT cycles have negatively affected the residual stress encountered in the GFRP bars. After 100 FT cycles, the bars lost more than 50% of their residual stress at the unloaded end and about 30% at the loaded end. This is consistent with the findings of control specimens. Low post peak stress was attributed to the failure of the sand-coated surface and its separation from the bar core. Friction resistance provided by the resulting 75

smooth surface could not provide enough resistance to the pullout force. This was more pronounced in the conditioned specimens.

4.2.5 Effect of FT cycles on the failure mode As shown in Table 10, all unconditioned BFRP bars failed in pullout mode indicating that concrete provided adequate confinement to the bars, which enabled them to reach their ultimate bond stress. Exposure to FT cycles did not change the mode of failure of the conditioned specimens that also failed in pullout mode. It is worth mentioning that specimens B12-84-100-1 and B12-84-100-2 encountered cracks in concrete close to the end of the test. Figure 39 shows the failure mode of specimen B12-84-100-2.

Loaded end

Unloaded end

BFRP bar

Figure 39: Failure of concrete at the end of bond test for specimen B12-84-100-2 As previously reported, all GFRP unconditioned specimens failed by concrete splitting. This finding was inconsistent with the pullout mode of failure of GFRP specimens with longer embedment (e.g. specimens G10-100-1 and G10-100-3) and was attributed to variation in the concrete strength. On the other hand, all GFRP conditioned specimens failed by pullout of the GFRP bars.

76

An important observation is that pullout of BFRP bars was accompanied by large slip values indicating a ductile mode of failure. However, both unconditioned and conditioned GFRP bars were pulled out of concrete in an abrupt explosive mode associated before the test was halted. This explained the sudden drop in the bond stress observed in the descending branch of the bond-slip history of both unconditioned and conditioned GFRP specimens. It also indicates the negligible effect of FT cycles on the failure mechanism of the bars. At the end of each test, specimens were split to visually assess the condition of the bar and surrounding concrete. Figure 40 shows typical BFRP and GFRP bars and concrete conditions where pullout failure was observed.

Resin and fiber traces Uniform delamination

(a) GFRP specimen G10-84-100-1

(b) BFRP specimen B12-84-200-3

Figure 40: Pullout Mode of failure of (a) GFRP and (b) BFRP bars Similar to what has been reported for the unconditioned specimens, the BFRP bar surface was significantly damaged and the outer layer was partially peeled off (delaminated) from the subsequent bar layers. White powder was detected along the whole embedded length indicating traces of crushed resin. No apparent damage to concrete surrounding the bar was observed. These findings suggest that bond failure took place along the interfacial shear surface between the grained layer and the subsequent core layers of the bar. Failure was therefore governed by the shear strength along the fibres’ interface rather than the shear strength between the bar and concrete. This mode of failure indicates

77

the negligible effect of FT cycles on the bond mechanism and the mode of failure of BFRP specimens. On the other hand, GFRP specimens encountered a different bond failure mechanism. Contrary to what was observed for BFRP bars, a “uniform” peeling off of the surface layer was noticed along the whole length of the embedded portion of the bar. Figure 40 a depicts this mode of failure for specimen G10-84-100-1. This observation explains the abrupt nature of failure encountered for both GFRP unconditioned and conditioned specimens.

78

Chapter 5: Analytical Models Several analytical models that describe bond-slip relationships of FRP bars are available. As previously mentioned, these analytical models have been developed and assessed for conventional FRP bars, such as glass and carbon FRP bars. These models have never been validated for the recently developed BFRP bars, a fact that has motivated this part of the research work. In this study, three models were adopted to model the bond-slip relationship for both unconditioned and conditioned specimens. This includes: a) The BPE model developed by Eligehausen et al. (1983) to describe the ascending branch of the BFRP bond-slip curve, i.e., during the increase of the pullout force up to the peak bond stress. b) The modified BPE model (mBPE) developed by Cosenza et al. (1996) to describe the descending branch of the bond-slip curve, i.e., the post-peak behavior of the FRP specimens. c) The CMR model developed by Cosenza et al. (1995) as an alternative to the BPE model to describe the ascending branch of the FRP bond-slip curve. In this study, the parameters of the BEP and the mBPE models (α and p) were calibrated for both BFRP and GFRP bars using the current experimental data. The parameter α was evaluated by equating the areas underneath the ascending branch, Aτ, of both analytical and experimental bond-slip curves (Cosenza et al. 1997) as follows:

∫

(

)

(Equation 6)

where the above parameters have been previously defined (section 2.4.2). Similarly, the parameter p was obtained by balancing the areas underneath the softening branches. The goal was to calibrate the value of α and p while minimizing the difference between the calculated areas underneath the analytical and experimental curves without compromising the trend of the obtained curve (Cosenza et al. 2002). 79

For the CMR model, parameter β was set at 0.07. This value was reported by Cosenza et al. (1997) for sanded FRP bars. The parameter sr was therefore calibrated for each diameter of BFRP bars by the least-square fit method. The analysis results are presented in the following sections for both unconditioned and conditioned specimens.

5.1 Analysis of unconditioned specimens Table 13 shows the mean values and coefficient of variations (CoV) for unconditioned specimens. Values of α obtained for BFRP ranged between 0.01 and 0.08 with an average value of 0.048. The low values obtained from the tests characterize the ascending branch of the BFRP bars and indicate the high initial stiffness of the bars at this stage of loading. In fact, most specimens develop their bond strength without significant slips at the unloaded ends, as previously mentioned, which explain the obtained low values of α. Values of α for GFRP bars ranged between 0.03 and 0.07 with an average value of 0.05. This value is very close to that of BFRP bars, which depicts similar trends in the bond-slip relationship for both types of bars. This is attributed to the friction-type behaviour observed for both bars. Analytical values of α reported by Cosenza et al. (1997) for sanded GFRP bars show a mean value of 0.067. The parameter p of the softening branch was also obtained for all test specimens. A mean value of 0.023 was reported for BFRP bars compared to 0.167 for GFRP bars. Such difference in values of p explains the variation in the trend of the softening curve of each type of bars. As observed from the test results, GFRP bars showed a sharp and sudden drop in bond stress after reaching the maximum stress of the bar. Figure 41 compares the experimental and analytical bond-slip curves for representative BFRP and GFRP specimens. Good correlation between the obtained curves can be observed for both types of bars. It is worth noting that the calibrated parameters for each specimen were used in representing the ascending and descending branches in lieu of the mean value. An important observation is that the GFRP bars exhibited a horizontal branch

80

after reaching the maximum bond stress. It was therefore found that the original BPE model better represents the GFRP data than the modified model in this part of the bond-slip curve. Table 13: Mean values and coefficient of variation of the models’ parameters for unconditioned BFRP and GFRP specimens

Average Standard deviation Coefficient of variation

BPE and mBPE models BFRP GFRP α P α P 0.048 0.023 0.05 0.167 0.023 0.013 0.023 0.122 0.481 0.577 0.462 0.733

CMR model BFRP GFRP sr sr 0.056 0.050 0.023 0.010 0.462 0.200

For CMR model, values of sr ranged between 0.037 and 0.07 with an average value of 0.056 for BFRP specimens. For GFRP bars, an average value of 0.050 was obtained. As mentioned, a value of β was kept constant at 0.07 for all BFRP and GFRP bars (Cosenza et al. 1997). Good correlation between the analytical and the actual curves can be observed in Figure 42.

81

(a)

(b) Figure 41: Experimental versus analytical results using the BPE and mBPE models for both BFRP and GFRP bars with (a) ld = 5d, (b) ld = 10d 82

30

B12-60-2

Bond stress, MPa

25

CMR model

20 15 10 5 0 0,00

0,10 Slip, mm

0,20

(a) 30

Bond stress, MPa

25 G10-100-3

20

CMR model

15

10 5 0 0,00

0,10 Slip, mm

0,20

(b) Figure 42: Experimental versus analytical results using the CMR model for unconditioned (a) BFRP and (b) GFRP bars 83

5.2 Analysis of conditioned specimens A similar procedure was adopted to model the bond performance of the conditioned specimens. Table 14 and Table 15 show the mean values and coefficient of variations (CoV) for BFRP and GFRP specimens conditioned for 100 and 200 cycles of FT, respectively. Table 14: Mean values and coefficient of variation of the models’ parameters for BFRP specimens exposed to 100 FT cycles

Average Standard deviation Coefficient of variation

BPE and mBPE models BFRP GFRP α P α P 0.041 0.014 0.067 0.070 0.024 0.003 0.000 0.004 0.583 0.184 0.000 0.052

CMR model BFRP GFRP sr sr 0.049 0.040 0.014 0.000 0.288 0.000

Table 15: Mean values and coefficient of variation of the models’ parameters for BFRP specimens exposed to 200 FT cycles

Average Standard deviation Coefficient of variation

BPE and mBPE models BFRP GFRP α P α P 0.040 0.010 0.040 0.035 0.011 0.002 0.000 0.016 0.260 0.223 0.000 0.457

CMR model BFRP GFRP sr sr 0.040 0.037 0.017 0.021 0.433 0.568

Figure 43 compares the average values of α obtained for all specimens. BFRP specimens conditioned at 100 and 200 FT cycles encountered similar values of α equal to 0.040 compared to a value of 0.048 for the unconditioned specimens. This finding indicates that FT cycles had a little influence on the ascending branch of the bond-slip curves. It is worth noting that increasing the value of α indicates a decrease in the initial stiffness of the ascending branch of the bond-slip curve. This is consistent with the test observations where initial stiffness of the BFRP specimens did not vary after exposure to FT cycles. For GFRP

84

specimens conditioned at 100 FT cycles, an average value of α equal to 0.067 was obtained compared to a value of 0.040 for the specimens conditioned at 200 FT cycles. 0,067

0,070

BFRP GFRP

Average value of α

0,060 0,050

0,048

0,050 0,041

0,040 0,040

100 cycles

200 cycles

0,040 0,030 0,020 0,010 0,000 control

Figure 43: Average values of “α” for unconditioned and conditioned BFRP and GFRP specimens The plot in Figure 44 compares the average values of p for all specimens. Average values of 0.014 and 0.010 were obtained for BFRP bars after 100 and 200 cycles, respectively, compared to 0.023 for the unconditioned bars. For GFRP specimens, values of p decreased from 0.167 for the unconditioned bars to 0.070 and 0.035 after 100 and 200 cycles, respectively. The relatively low values of p obtained after FT cycles depict an increase in slip of the bar post the peak stress. Decreasing the value of p means a gradual loss of bond and a more flattened descending curve.

85

0,200 BFRP

0,167

Average value of p

GFRP

0,150

0,100 0,070

0,050

0,035 0,023

0,014

0,010

100 cycles

200 cycles

0,000 control

Figure 44: Average values of “p” for unconditioned and conditioned BFRP and GFRP specimens On the other hand, Figure 45 shows a comparison between values of sr of the CMR model for both unconditioned and conditioned specimens. A value of β equal to 0.07 was assumed for all bars based on Cosenza et al. (1997) recommendations for sanded FRP bars. Values of sr were calibrated against the experimental results. These values ranged between 0.040 and 0.049 for BFRP specimens conditioned for 100 and 200 FT cycles, respectively, compared to 0.056 for unconditioned specimens. For GFRP specimens, these values ranged between 0.040 and 0.037 for specimens conditioned for 100 and 200 FT cycles, respectively, compared to a value of 0.05 for the unconditioned specimens.

86

0,070 BFRP 0,060

0,056

GFRP

Average value of Sr

0,050 0,050

0,049 0,040

0,040

0,040

0,037

0,030 0,020 0,010 0,000

control

100 cycles

200 cycles

Figure 45: Average values of “sr” for unconditioned and conditioned BFRP and GFRP specimens Figure 46 and Figure 47 compare the experimental and analytical curves as obtained by the BPE model for BFRP and GFRP specimens conditioned at 100 and 200 cycles of FT, respectively. Similarly, Figure 48 and Figure 49 compare the experimental and analytical curves as obtained by the CMR model. A good correlation between the curves can be observed.

87

Bond stress, MPa

25

B12-84-100-2 BPE model

20

15 10 5 0

0,00

0,10 Slip, mm

0,20

(a) 25 G10-70-100-2

Bond stress, MPa

20

BPE model

15 10 5 0 0,00

0,10 Slip, mm

0,20

(b) Figure 46: Experimental versus analytical results using the BPE model for (a) B1284-100-2 and (b) G10-70-100-2 specimens 88

Bond stress, MPa

25

B12-84-200-1 BPE model

20 15 10 5 0 0,00

0,10 Slip, mm

0,20

(a)

25

Bond stress, MPa

20 G10-70-200-1

15

BPE model

10 5 0 0,00

0,10 Slip, mm

0,20

(b) Figure 47: Experimental versus analytical results using the BPE model for (a) B1284-200-1 and (b) G10-70-200-1 specimens 89

Bond stress, MPa

25

B12-84-100-1 CMR model

20 15 10 5 0 0,00

0,10 Slip, mm

0,20

(a) 25 G10-70-100-2 CMR model

Bond stress, MPa

20 15 10 5 0 0,00

0,10 Slip, mm

0,20

(b) Figure 48: Experimental versus analytical results using the CMR model for (a) B12-70-100-1 and (b) G10-70-100-2 specimens 90

25 B12-84-200-1

Bond stress, MPa

20

CMR model

15 10 5 0 0,00

0,10 Stress, mm

0,20

(a) 25

Bond stress, MPa

20 G10-70-200-1

15

CMR model

10 5 0 0,00

0,10 Slip, mm

0,20

(b) Figure 49: Experimental versus analytical results using the CMR model for (a) B12-84-2001- and (b) G10-70-200-1 specimens 91

92

Chapter 6: Conclusions and recommendations for future work 6.1 Introduction Tests results of the bond performance and durability of BFRP and GFRP bars have been presented. Pullout tests were conducted in two phases: the first phase investigated the effect of FRP material, bar diameter, d, and bar embedment length in concrete, ld, on the bond performance of the bars. The second phase investigated the effect of freeze-and-thaw cycles on the bond performance of GFRP and BFRP bars having an embedded length of 7 times the bar diameter. The following sections include a summary of the obtained test results.

6.2 Summary of the test results of phase I: unconditioned specimens In phase I, a total of 48 unconditioned specimens reinforced with BFRP and GFRP bars were tested at room temperature. Test results suggested that bond-slip curves of both types of bars had the same trend, with an ascending branch up to the peak stress and a softening branch where slip hardening was noticed. The following concluding remarks were drawn from this phase: 1- Factors known to influence the bond performance of conventional FRP bars have a similar effect on BFRP bars. The larger the bar diameter the less bond strength developed during the tests. Similarly, increasing the embedment length of the bar in concrete resulted in a decrease in the bond strength of the bar. 2- BFRP bars of 10 mm diameter developed 71%, 89%, and 79% of the GFRP bond strength for embedment lengths of 5, 7, and 10 d, respectively, with an average value of 80%. The corresponding slip at unloaded ends was negligible for both types of bars (0.15 mm and 0.06 mm for BFRP and GFRP bars, respectively). At the loaded ends, an average 93

slip of 1.02 mm and 0.81 mm was encountered at maximum stress for BFRP and GFRP bars, respectively. 3- BFRP bars with small diameters showed better adhesion to concrete at initial stages of loading than bars of large diameters (average adhesion stress was 1.37, 0.67, and 0.51 MPa for diameters 8, 10, and 12 mm, respectively). Average adhesion of BFRP bars of 10 mm diameter was 0.67 MPa compared to 1.03 MPa for GFRP bars (65% of that of GFRP bars). 4- BFRP bars exhibited higher residual stress than GFRP bars with an average value of 9 MPa (61% of the maximum stress) and 10 MPa (68% of the maximum stress) at unloaded and loaded ends, respectively, compared to 7.7 MPa (40% of the maximum stress) and 5 MPa (30% of the maximum stress) for GFRP bars at both ends, respectively. 5- All BFRP specimens failed in pullout mode. Failure was governed by the shear interface between the sand-coated layer and the subsequent core layers of the bars. Partial delamination of the outer-grained surface led to a gradual and ductile failure. On the other hand, a more uniform and abrupt peeling off of the outer surface characterized the failure of GFRP bars, which led to significant energy release at failure.

6.3 Summary of the test results of phase II: conditioned specimens In this phase, the effect of FT cycles (100 and 200 cycles) on the bond performance of BFRP and GFRP bars was investigated. The following concluding remarks can be drawn from this phase: 1- A discrepancy in the bond performance of the conditioned BFRP and GFRP specimens was encountered due to the random nature of exposure to FT cycles. 2- Normalized bond stress values suggested that concrete curing during FT cycles had not contributed to the bond enhancement encountered in some specimens after exposure. The enhancement in bond stress attained by BFRP specimens after FT conditioning was 94

attributed to the increase in bar diameter due to water absorption into the bar matrix, which increased the friction and mechanical interlock between the bar and surrounding concrete. 3- In some other specimens, bond enhancement due to bar swelling was countered by (a) the degradation of the outer layer of the bar and (b) the effect of mismatch of the coefficient of thermal expansion between the bar and concrete. This led to a loss of bond with subsequent FT cycles. 4- FT cycles had a detrimental effect on the adhesion of bars to concrete at early stages of loading, which was more pronounced in case of GFRP bars than for BFRP bars. GFRP bars of 10 mm diameter lost more than 70% of their adhesion to concrete after 100 cycles compared to 30% for BFRP bars of equal diameter. 5- Conditioned BFRP specimens encountered an average residual bond stress of 77% and 72% after 100 and 200 FT cycles, respectively, compared to 51% of the peak stress attained at room temperature. For GFRP bars, the residual bond stress decreased from 22% of the peak stress at room temperature to 14% and 15% of the peak stress after 100 and 200 FT cycles, respectively. After 100 FT cycles, GFRP bars lost more than 50% of their residual stress at the unloaded end. 6- FT cycles had a negligible effect on the failure mode or failure mechanism of BFRP and GFRP bars. Comparable modes were encountered for both unconditioned and conditioned specimens. 7- Considering the small variation in bond strengths between the unconditioned and conditioned specimens, one can conclude that FT conditioning had a slight effect on the bond strength of BFRP bars, which make the BFRP bars suitable for use in such harsh environments.

95

6.4 Summary of the analytical phase In this phase, the BPE, mBPE, and the CMR models were calibrated based on the test results. A comparison between the analytical and the experimental bond-slip curves was presented. The following concluding remarks were drawn from this phase: 1- The three models can accurately describe the bond performance of unconditioned BFRP bars considering the following parameters: For BPE and mBPE models (a) α = 0.048 (b) p = 0.023; and for the CMR model (a) β = 0.07 and (b) sr = 0.056. 2- The following design values are proposed to better describe the bond performance of inservice elements subjected to FT cycles: For BPE and mBPE models (a) α = 0.041 (b) p = 0.01. The relatively low value of p depicts an increase in slip of the bar post the peak stress and reflects the gradual loss of bond after exposure. For the CMR model (a) β = 0.07 and (b) sr = 0.04. 3- Considering the above values, a good correlation between the experimental and analytical curves can be obtained.

6.5 Recommendations for future work This study has gone towards enhancing our understanding of the bond performance of BFRP bars. The current findings provide evidence that BFRP bars have demonstrated their promise as reinforcing materials for concrete elements. Further research should be undertaken in the following areas: 1- Establishing the effect of other parameters that have not been included in this study such as concrete strength, concrete type (e.g., fibre-reinforced concrete, self-consolidated concrete, etc.), and concrete confinement on the bond performance of BFRP bars. 2- Widening the research scope on bond of BFRP bars to include other types of commercially available BFRP bars.

96

3- Investigating the bond durability of BFRP bars and their long-term bond performance when exposed to other environmental attacks or extreme loadings. 4- Improving the interface bond between the outer and the core layers of the bar. Bond strengths are directly related to the material properties and the surface characteristics of the bars.

97

98

References

1. Abdalla, H. “Concrete cover requirements for FRP reinforced members in hot climates”, Composite Structures, 2006, 73, 61-69. 2.

Achillides, Z. “Bond behaviour of FRP bars in concrete’’, PhD thesis, University of Sheffield, 1998.

3. Achillides, Z. and Pilakoutas, K. “Bond behavior of fiber reinforced polymer bars under direct pullout conditions”, Journal of Composites for Construction, 2004, 8(2), 173-81. 4. ACI 440.1R-06, ACI Committee. “Guide for the Design and Construction of Structural Concrete Reinforced with FRP Bars”, American Concrete Institute, 2006. 5. Adhikari, S. “Mechanical properties and flexural applications of basalt fiber reinforced polymer (BFRP) bars”, Master thesis, Akron University, 2009. 6. Aiello, M.A., Leone, M. and Pecce, M. “Bond Performances of FRP RebarsReinforced Concrete”, Journal of Materials in Civil Engineering, 2007, 19(3), 205213. 7. Al-Mahmoud, F., Castel, A., François, R. and Tourneur, C. "Effect of surface preconditioning on bond of carbon fibre reinforced polymer rods to concrete". Cement and Concrete Composites, 2007, 29(9), 677-689. 8. Alunno, R.V., Galeota, D. and Giammatteo, M.M. “Local bond stress-slip relationships of glass fibre reinforced plastic bars embedded in concrete”, Materials and Structures/ Matériaux et Constructions, 1995, 28(180), 340-344. 9. Alvarez, V.A. “Temperature effect on bond properties of Fiber Reinforced Polymer (FRP) bars embedded in concrete”, Master theses, Sherbrooke University, 2004.

99

10. Alves, J., El-Ragaby, A. and El-Salakawy, I. “Durability of GFRP Bars’ Bond to Concrete under Different Loading and Environmental Conditions”, Journal of Composites for Construction, 2011, 15(3), 249-262. 11. Arias, J., Vazquez, A., and Escobar, M. M. “Use of sand coating to improve bonding between GFRP bars and concrete.” Journal of Composite Materials, 2012, 46, 2271-2278. 12. Baena, M., Torres, L., Turon, A. and Barris, C. “Experimental study of bond behaviour between concrete and FRP bars using a pull-out test”, Journal of Composites: Part B, 2009, 40, 784-797. 13. Bank, L.C., Puterman, M. and Katz, A. “The effect of material degradation on bond properties of fiber reinforced plastic reinforcing bars in concrete.” ACI Materials Journal, 1998, 95(3), 232-243. 14. Belarbi, A. and Wang, H. “Bond Durability of FRP Bars Embedded in FiberReinforced Concrete”, Journal of Composites for Construction, 2012, 16(4), 371380. 15. Benmokrane, B., Tighiouart, B. and Chaallal, O. “Bond strength and load distribution of composite GFRP reinforcing bars in concrete.” ACI Materials Journal, 2006, 93(3), 246-253. 16. Bilal, S. H., Rteil A., and Khaled, A.S. “Bond Strength of Tension Lap Splices in High-Strength Concrete Beams Strengthened with Glass Fiber Reinforced Polymer Wraps”, Journal of Composites for Construction, 2004, 8(1), 14-21. 17. Boyle, H. C. and Karbhari, V. M. “Investigation of bond behaviour between glass fiber composite reinforcements and concrete”, Polymer-Plastics Technology and Engineering, 1994, 33(6), 733-753.

100

18. Brik, V.B. “Advanced Concept Concrete Using Basalt Fiber/BF Composite Rebar Reinforcement”, Final Report for Highway-IDEA Project 86, Research & Technology Corp., Madison, WI, 2003. 19. Brik, V.B. “Basalt fiber composite reinforcement for concrete”, Final report, Transportation Research Board, National Research Council, Contract NCHRP-96IDO25, 1997. 20. Chaallal, O. and Benmokrane, B. “Pullout and bond of glass-fiber embedded in concrete and cement grout”, Journal Materials and Structures, 1993, 26, 167-175. 21. Chen, Y., Davalos, J.F., Ray, I. and Kim, H.Y. “Accelerated aging tests for evaluations of durability performance of FRP reinforcing bars for concrete structures”, Composite Structures, 2007, 78, 101-111. 22. Cosenza, E., Manfredi, G. and Realfonzo R. “Analytical modeling of bond between FRP reinforcement and concrete.” Proc., 2nd International RILEM Symposium, FRPRCS-2, L. Taerwe, editor, 1995, 164-171. 23. Cosenza, E., Manfredi, G. and Realfonzo R. “Bond characteristics and anchorage length of FRP rebars”, Proceedings, 2nd International Conference on Advanced Composite Materials in Bridge Structures, M. EI-Badry, editor, 1996. 24. Cosenza, E., Manfredi, G. and Realfonzo, R. “Behaviour and modeling of bond of FRP rebars to concrete”, Journal of Composites for Construction, 1997, 1(2), 40-51. 25. Cosenza, E., Manfredi, G. and Realfonzo, R. “Development length of FRP straight rebars”, Journal of Composites Part b: Engineering, 2002, 33, 493-504. 26. CSA-S806-02. “Design and Construction of Building Components with Fibre Reinforced Polymers” Canadian Standard Association, 2002. 27. Davalos, J.F., Chen, Y. and Ray, I. “Effect of FRP bar degradation on interface bond with high strength concrete”, Cement and Concrete Composites, 2008, 30, 722-730. 101

28. Dutta, P.K. and Taylor, S., “A fractographic analysis of graphite-epoxy composites subjected to low temperature thermal cycling”, Proceedings of International Symposium for testing and failure analysis, 1998, Los Angeles, 429-435. 29. El Refai, A. “Durability and fatigue of basalt fiber-reinforced polymer bars gripped with steel wedge anchors”, Journal of Composites for Construction, 17(6), 2013. 30. El-Salakawy, E. and Benmokrane, B. “Fiber reinforced polymer composite bars for the concrete deck slab of Wotton Bridge”, Canadian Journal of Civil Engineering, 2003, 30(5), 861-870. 31. Eligehausen, R., Popov, E.P. and Bertero, V.V. "Local bond stress-slip relationships of deformed bars under generalized excitations." Rep. No. 83/23, Earthquake Engineering Research Center (EERC), University of California, Berkeley, California, 1983. 32. Faoro, M. “Bearing and deformation behaviour of structural components with reinforcements comprising resin bounded glass fibre bars and conventional ribbed steel bars”, Proceedings of International Conference on Bond in Concrete, 1992. 33. fib report. “Bond of reinforcement to concrete: state-of-the-art.” fib bulletin 2000. 34. Galati, N., Nanni, A., Dharani, L., Focacci, F. and Antonietta, A.M. “Thermal effects on bond between FRP rebars and concrete”, Journal of Composites: Part A, 2006, 37, 1223-1230. 35. Harajli, M. and Abouniaj, M. “Bond Performance of GFRP Bars in Tension: Experimental Evaluation and Assessment of ACI 440 Guidelines”, Journal of Composites for Construction, 2010, 14(6), 659-668. 36. ISIS Design Manual No.3. “Reinforcing Concrete Structures with Fibre Reinforced Polymers (FRPs)”, ISIS Canada, 2007.

102

37. Karbhari, V.M. “Response of Fiber Reinforced Polymer Confined Concrete Exposed to Freeze and Freeze-Thaw Regimes”, Journal of Composites for Construction, 2002, 6(1), 35-40. 38. Karbhari, V.M., Rivera, J. and Dutta, P.K. “Effect of short-term freeze-thaw cycling on composite confined concrete”, Journal of Composites for Construction, 2000, 4(4), 191-197. 39. Katz, A., Berman, N. and Bank, L.C. “Effect of high temperature on bond strength of FRP rebars”, Journal of Composites for Construction, 1999, 3(2), 73-78. 40. Malvar, L.J. "Bond stress-slip characteristics of FRP rebars" Rep. TR-2013-SHR, Naval Facilities Engineering Service Center, Port Hueneme, California, 1994. 41. Masmoudi, R., Zaidi, A. and Gérard, P. “Transverse Thermal Expansion of FRP Bars Embedded in Concrete”, Journal of Composites for Construction, 2005, 9(5), 377-387. 42. Militky, J., Kovacic, V. and Rubnerova, J. “Influence of thermal treatment on tensile failure of basalt fibres”, Engineering Fracture Mechanics, 2002, 69, 10251033. 43. Nacer, R. “Évaluation des propriétés physiques et mécaniques de laminés en matériaux composites fabriqués par enroulement filamentaire”, Master thesis, Sherbrooke University, 2006. 44. Palmieri, A., Matthys, S. and Tierens, M. “Basalt fibres: mechanical properties and applications for concrete structures”, Taylor & Francis Group, London, 2009, 165169. 45. Parnas, R., Shaw, M., and Liu, Q. “Basalt fiber reinforced polymer composites.” Report prepared for The New England Transportation Consortium - NETCR63, Project No. 03-7, 2007.

103

46. Ramakrishman, V. and Panchalan, R.K. “A New Construction Material-NonCorrosive Basalt Bar Reinforced Concrete”, ACI Journal Special publication, 2005, 229, 253-270. 47. Ramakrishnan, V. and Neeraj, S. “Performance evaluation of 3-D basalt fiber reinforced concrete and basalt rod reinforced concrete”, Final report, Transportation Research Board, National Research Council, Contract NCHRP-45, 1998. 48. Saiedi, R., Fam, A. and Green, M.F. “Behavior of CFRP-Prestressed Concrete Beams under High-Cycle Fatigue at Low Temperature”, Journal of Composites for Construction, 2011, 15(4), 482-489. 49. Sayed, A.F., Foret, G. and Le Roy R. “Bond between carbon fibre-reinforced polymer (CFRP) bars and ultra high performance fibre reinforced concrete (UHPFRC): Experimental study”, Construction and Building Materials, 2011, 25, 479-485. 50. Sim, J., Park, C. and Moon, D.Y. “Characteristic of basalt fibre as a strengthening material for concrete structures”, Journal of Composites: Part B, 2005, 36(6), 504512. 51. Tannous, E.A. and Saadatmanesh, H. “Durability of AR glass fiber reinforced plastic bars”, Journal of Composites for Construction, 1999, 3(1), 12-19. 52. Tastani, S.P. and Pantazopoulou, S.J. “Bond of GFRP Bars in Concrete: Experimental Study and Analytical Interpretation”, Journal of Composites for Construction, 2006, 10(5), 381-391. 53. Tighiouart, B., Benmokrane, B., Mukhopadhyaya, P. “Bond strength of glass FRP rebar splices in beams under static loading.” Construction and Building Materials, 1999, 13(7), 383-392.

104

54. Wang, H. and Belarbi, A. “Static and Fatigue Bond Characteristics of FRP Rebars Embedded in Fiber-reinforced Concrete”, Journal of Composite Materials, 2010, 44(13), 1605-1622. 55. Wardeh, G., Mohamed, A.S. and Ghorbel, E. “Analysis of concrete internal deterioration due to frost action”, Journal of Building Physics, 2011, 35(1), 54-82. 56. Wei, B., Cao, H. and Song, S. “Degradation of basalt fibre and glass fibre/epoxy resin composites in seawater”, Journal of Corrosion Science, 2011, 53, 426-431. 57. Wei, B., Cao, H. and Song, S. “Tensile behavior contrast of basalt and glass fibres after chemical treatment”, Journal of Materials and Design, 2002, 31, 4244-4250. 58. Wu, Z., Wang, X., Iwashita, K., Sasaki, T., and Hamaguchi, Y. “Tensile fatigue behavior of FRP and hybrid FRP sheets”, Journal of Composites: Part B, 2010, 41(5), 396-402.

105