Bond characteristics of straight- and headed-end, ribbed-surface, GFRP bars embedded in high-strength concrete

Construction and Building Materials 83 (2015) 283–298

Contents lists available at ScienceDirect

Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat

Bond characteristics of straight- and headed-end, ribbed-surface, GFRP bars embedded in high-strength concrete Sirajul Islam a, Hamdy M. Afefy a,b,⇑, Khaled Sennah a, Hossein Azimi a a b

Civil Engineering Department, Ryerson University, Toronto, ON, Canada Structural Engineering Department, Faculty of Engineering, Tanta University, Tanta, Egypt

h i g h l i g h t s  Bond characteristics of both straight- and headed-end GFRP bars are studied experimentally.  180 pullout tests were conducted to cover 30 parameters.  Empirical equation is proposed for the development length calculation of GFRP bars.  Development lengths based on experimental results are compared with the available design standards.  CSA S6-06 showed the closest development length results to the experimental findings.

a r t i c l e

i n f o

Article history: Received 7 September 2014 Received in revised form 23 February 2015 Accepted 4 March 2015 Available online 19 March 2015 Keywords: Bond stress Bar diameter Concrete cover Glass Fiber Reinforced Polymer (GFRP) bars Headed bars High-strength concrete Slip

a b s t r a c t Glass Fiber Reinforced Polymer (GFRP) bars as a proper substitute for traditional reinforcing steel bars not only eliminate the durability problem due to corrosion of reinforcing steel, but also provide remarkably enhanced capacity due to their high tensile strength compared to that of the steel bars. This paper presents the experimental findings of 180 pullout tests conducted on GFRP bars embedded into highstrength concrete blocks covering different parameters. The studied parameters were bar diameter size (12 or 16 mm), embedment length (4 or 6 times the bar diameter), bar end condition (straight and headed), and concrete cover (1.5, 2.5, and 5 or 7 times bar diameter for straight bars and 8 or 10.5 times bar diameter for headed bars) in addition to a case of no embedment length except the head length for headed-end bars. In total, 30 variables were studied, while each variable was conducted on 6 identical specimens in order to increase the reliability of the results. Based on the results of the parametric study, the bond stress was shown to be inversely proportional to the embedment length and bar diameter as expected. In addition, the smaller concrete cover appeared to have significant effect on bond stress, leading to side blow-out failure rather than bar pullout or concrete splitting in the case of headed-end GFRP bars. In addition, the GFRP bar with headed-end showed significant increase in pullout strength compared to that for the straight-end bars. Finally, an empirical expression was proposed to calculate the development length of GFRP bars with either straight or headed-end, and then compared with the available design standards such as CSA-S806-02, CSA S6-06, ACI 440-1R-06, and JSCE-97. The comparison showed that the results developed by CSA S6-06 standards was the closest to the experimental findings showed about 2% safety margin exceeding the obtained development length by the proposed expression. Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction Fiber Reinforced Polymer (FRP) bars have desirable characteristics which give them more advantages over traditional reinforcing steel bars. These characteristics include high tensile strength, ⇑ Corresponding author at: 34 Ahmed Farouk Ezzat, Smouha, Alexandria, Egypt. Cell: +20 106 177 3174, +1 416 826 7772. E-mail address: [email protected] (H.M. Afefy). http://dx.doi.org/10.1016/j.conbuildmat.2015.03.025 0950-0618/Ó 2015 Elsevier Ltd. All rights reserved.

corrosive resistance, light weight, electric insulation and fatigue resistance [1]. Therefore, in the recent years, FRP bars have been introduced as a competent alternative to traditional reinforcing steel bars for different concrete structures subjected to severe environmental conditions such as waste water treatment and chemical plants, floating decks, sea walls and water structures [2–7]. In addition, it has been found that FRP bars can eliminate durability problem associated with corroded reinforcing bars [8–11]. However, direct replacement of the reinforcing steel bars

284

S. Islam et al. / Construction and Building Materials 83 (2015) 283–298

with the FRP bars has many concerns due to various differences in the manifested behavior of the two materials under different loading conditions. For instance, FRP exhibits linear elastic behavior up to failure which means that it exhibits limited ductility. In addition, FRP bars have anisotropic material properties while steel bars have isotropic properties, which make the bond behavior dubious. Furthermore, higher cost of the FRP bars compared to that of steel bars and lack of familiarity with the new technology resulted in slow adaptation of FRP as concrete reinforcement [12]. As the transfer of stresses between the concrete and the reinforcement is mainly dependent on the quality of bond, the force transfer mechanism is always a serious issue of the structural design regardless of the type of reinforcement [11,13–16]. Hence, the force between the reinforcement and concrete should be transferred efficiently through the bond between the two materials in order to ensure strain compatibility and composite action in reinforced concrete members. The transfer of forces between a reinforcing bar and concrete is attributed to three different mechanisms, namely: (i) chemical adhesion; (ii) friction and (iii) mechanical interlocking arising from the textures on the bar surface as illustrated in Fig. 1(a). The resultant of these forces can be resolved into an outward component (radial splitting force) and a shear component, parallel to the bar that is the nominal bond force as shown in Fig. 1(b). For traditional steel reinforcement, bond failure is attributed to bearing causing side splitting or shearing of concrete. On the other hand, bearing stress of the GFRP bars can exceed the shear strength between the surface deformations and the bars core resulting in a bond failure at this interface as depicted in Fig. 2(a) [17]. For real structures, it is unusual for a pure pullout or pure splitting failure to occur, mostly a combination of the two modes occurs as shown in Fig. 2(b). Generally, bond behavior between concrete and reinforcing steel bar can be assumed constant, however this assumption is less valid for GFRP bars due to their relatively lower stiffness compared to that of steel bars. This results in greater slip values at the loaded end than at the free end [15]. Thus, the free end slip will be reduced to almost zero once the embedment length is greater than the development length as depicted in Fig. 3(a). Adopting the same concept, the bond stress distribution for headed-ended GFRP bars maybe assumed as shown in Fig. 3(b). Previous research showed that bond behavior of FRP bars in concrete is influenced by several geometric and material-related factors [2,15,18–23,11,24–27]. Regression manipulation on different experimental results indicated that good correlation exists between bond strength and the square root of the compressive strength of concrete [12,20,28]. In addition, bond failure mechanism of FRP bars in concrete is influenced by concrete cover around

the reinforcing bar by virtue of its confining effect [18,19]. Bond failure occurs through splitting of the concrete when the member does not have adequate concrete cover [28]. On the other hand, when enough concrete cover is provided, splitting failure is prevented or delayed while the pullout failure is dominating [29]. Experimental investigations revealed that bond strength of FRP bars increases with decrease in the bar diameter, which is the same results obtained for steel bars [2,9,11,12,30]. Hao et al. [1] and Tighiouart et al. [11] verified that when the diameter of the bar is larger, more bleeding water is trapped beneath the bar. Consequently, there is a greater chance of creating voids around the bar which will eventually decrease the contact surface between the concrete and the bar and thus, reduces bond strength. It was also observed that the maximum average bond stress decreased with an increase in the embedment length as exhibited by steel bars [2,10,11,15,23]. Due to the nonlinear distribution of the bond stress along the length of the reinforcing bar, as the embedment length increases, the stress is distributed over a longer length and henceforth, the bond strength decreases. Bond between reinforcement and concrete can be described by means of a constitutive bond stress-slip relationship that can be introduced in the solution of problems, such as the calculation of bar development length [22]. Although numerous existing formulations for steel bars exist and are well-established, FRP bars still require extensive research effort to determine an analytical model of the bond stress-slip constitutive law. Malvar [31] established the first modeling of the bond behavior in the case of FRP bars with various deformation geometries and radial confining stresses. Cosenza et al. [22] investigated the bond stress-slip behavior of FRP bars and proposed a modification to the bond prediction evaluation (BPE) model to account for the FRP characteristics. Diverse efforts were dedicated in order to develop more refined bond-slip model to cover various surface treatments, shear and axial stiffness, bar diameter, bond length, confinement applied to the FRP bars due to concrete shrinkage or external loads, and swelling of FRP bars due to temperature variation and moisture absorption [32–42]. For many years, bond strength was represented in terms of the shear stress at the interface between the reinforcing bar and the surrounding concrete treating bond as a material property [43]. It is now understood that bond, anchorage, development, and splice strength, are structural properties that are dependent on not only the materials, but also on the geometry of the reinforcing bar and the structural member itself. Glass Fiber Reinforced Polymer (GFRP) bars are commonly used in various projects in North America such as bridge deck slabs traffic barrier and parking garages as a substitute of steel reinforcing

Adhesion and friction forces along the bar's surface

T

Reinforcing bar

Adhesion and friction force Bearing force

Radial splitting force Resultant force Bond force

Friction force

Bearing force

(a)

(b) Fig. 1. Bond force transfer mechanism.

285

S. Islam et al. / Construction and Building Materials 83 (2015) 283–298

GFRP bar

GFRP bar

Fig. 2. Pullout failure modes.

P

Embedded length

Embedded length

Unbonded length

Unbonded length

P

(a) Straight-end bar

(b) Headed-end bar

Fig. 3. Bond stress distribution along GFRP bars.

bars. That is due to their lower cost compared to other types of FRP materials. The current paper presents both experimental and analytical investigations conducted in order to verify the effect of different parameters on the bond characteristics between GFRP bars and the surrounding high strength concrete. In order to increase the reliability of experimental findings, test results of each studied parameter was obtained based on the average results of six identical specimens. As a consequence, the pullout test was performed on 180 specimens in order to cover the effect of different parameters on the bond strength such as bar size, concrete cover, embedded length and the state of the bar end, which was either straight or with headed anchorage. 2. Experimental test program 2.1. Description of the specimens The test specimens consisted of two sets of GFRP bars. The first set had a bar diameter (db) of 12 mm, while the second set had a bar diameter of 16 mm. Table 1 summarizes the geometrical properties for both sizes of the GFRP bars as well as the dimensions of the heads implemented in the case of headed- end bars. A total of 180 GFRP bar specimens were tested. Their configurations are shown in Table 2 in accordance with the terminology shown in Fig. 4. Each GFRP bar size had three groups of eccentricity from the specimen edge (concentric, 1.5db eccentric and 2.5db eccentric, where db is the bar diameter) to simulate concrete clear cover to

Table 1 Geometrical properties of the ribbed-surface GFRP bars. Metric bar size 12 16

Core diameter, db (mm) 12 16

External diameter (mm) 13.5 18

Crosssectional area (mm2) 113 201

Head length (mm) 75 100

Head external diameter (mm) 28 40

Table 2 Specimens parameters. db (mm)

Le1 (mm)

Le2 (mm)

4db

6db

0

4db

6db

12 16

48 64

72 96

0 0

48 64

72 96

Head length (mm)

Number of identical samples

75 100

6 6

the bar surface. Straight GFRP bars had two embedment depths, Le1, of 4db and 6db, as shown in Fig. 4(a) and (c). As shown in Fig. 4(b) and (d), the total embedment depth of headed GFRP bars was considered as the head length plus the distance Le2. Three values for Le2 were considered in this study, namely: 0, 4db and 6db. The combinations of the tested parameters resulted in 30 parameters. Accordingly, in order to increase accuracy in experimental results, six identical specimens were constructed for each parameter. Using plywood, wooden formworks were constructed. The embedment lengths on bars were carefully measured and the unbonded length was covered by regular black electric tape in order to avoid concrete to touch the bar at the unbonded length. Lumber was later used to construct a support structure providing stability for bars prior to concrete hardening. Fig. 5 shows view of the wooden formwork prepared for casting concrete blocks of four parameters. Proper vibration was also provided to avoid air entrapment. Concrete block specimens, while still attached to molding, were cured for few days after casing by placing moist blanket on top of the specimens. 2.2. Material properties 2.2.1. Concrete The used concrete was high-strength ready mix concrete with target concrete cylinder strength of 60 MPa. The test specimens were cast in one day. In order to determine the strength of the concrete, six concrete cylinders of 100  200 mm collected from the different locations of the batch were tested after four weeks of curing in the laboratory along with the concrete specimens. The concrete characteristic strength for the concrete was calculated using Eq. (1) [44]. 0 fc

2 " #0:5 3 2 ðkc VÞ  4 5 ¼ 0:9f c 1  1:28 þ 0:0015 n

ð1Þ

286

S. Islam et al. / Construction and Building Materials 83 (2015) 283–298

P

P

P

P

Unbonded length

Unbonded length

Unbonded length

Head Le 2

50 mm

Clear cover

300 mm

200 mm

(a) Concentric straight bars

Le1

Le1

50 mm

Head Le 2

50 mm

50 mm

Unbonded length

Clear cover

300 mm

200 mm

(b) Concentric headed bar

(c) Eccentric straight bar

(d) Eccentric headed bar

Fig. 4. Details of test specimens for concentric and eccentric bars. The anchor head material is a thermo-setting polymeric concrete made from high strength concrete mix and vinylester resin, which is then injected into a two part tool that wraps around the bar end. High temperature (95 degree) and pressure (6 bar) are applied until the anchor head is cured to the bar. The bond between the cast polymeric concrete and the ribbed surface of the bar is achieved through frictional forces and the bar surface as well the bearing force on the ribs as depicted in the schematic diagram in Fig. 1(a). The maximum outer diameter of the end heads is 2.5 times the diameter of the bar. The head lengths of the 16 and 12 mm diameter bars are 100 and 75 mm long, respectively as given in Table 2. It begins with a wide disk which transfers a large portion of the load from the bar into the concrete. Beyond this disk, the head tapers in five steps to the outer diameter of the blank bar. This tapered geometry ensures optimal anchorage forces and minimal transverse splitting action in the vicinity of the head as shown later in the experimental findings. 2.3. Test setup, test procedure and instrumentation Fig. 5. View of formwork with GFRP bars prepared for 4 test parameters.

0 where f c = concrete characteristic strength; f c = average concrete cylinder strength; = coefficient of variation modification factor for concrete = 1.15 for six samples; n = number of concrete cylinder tests = 6; V = coefficient of variation of concrete cylinder strengths. Table 3 summarizes the results from concrete cylinder tests against the characteristic compressive strength of concrete.

2.2.2. GFRP bars GFRP bars with ribbed [45] was used in the current study. Two different bar diameters of 12 and 16 mm were used in this study as illustrated in Fig. 6(a). Table 4 summarizes the material properties of the GFRP bars as provided by the manufacturer. Unlike steel, bending GFRP bars have to be done in the factory prior to curing of the resin since GFRP bars cannot be bent once hardened. It should be noted that the tensile capacity of bent GFRP bars is much less than that for straight bars because the fibers get rearranged and redirected in the bending process. Consequently, more bars are required when designing the structural member with bent bars. Therefore, manufacturers developed headed-end GFRP bars, shown in Fig. 6(b) and (c), to provide an alternative for such bent bars. In other words, reducing the development length for straight bars (and avoiding bending bars) in case of geometric constraints, when there is not enough room to satisfy the development length requirement is the main reason behind manufacturing bar end heads. Behavior of these bars is proven to be elastic up to failure.

The type of pullout testing undertaken in this research is called confined pullout testing, since the top of concrete is under bearing compression from the bearing plate. This would prevent concrete cones to form on top of concrete at failure [46]. The main requirement specified in CSA-S807 standard [47] for this type of testing is that the grip distance should be at least 40db, where db is the bar diameter. This limitation was considered in all test setups. In addition, all pullout tests were performed according to the S806-02 [48] test equipment and requirements. Two test setups were designed and assembled to apply tensile force on the GFRP bars at their free ends. The first setup was designed for concentric specimens, while the second one was designed for eccentric specimens. Views of test setup used for concentric specimens are shown in Fig. 7(a) and (b). The main components of the test setup included one manual hydraulic jack, two potentiometers, one compression/tension ring load cell and the custom-made grip to hold the bar. Steel plates and rubber pads were also used to apply a smooth load transmission from jack to bars. The load cell was placed between the jack and the grip. One potentiometer (POT-1) was instrumented with plate in front of the jack to measure the bar displacement at different loading stages and another potentiometer (POT-2) was instrumented at the end of bar to measure the free end slip. Due to eccentricity of GFRP bars in eccentric specimens, it was not practical to use the same test setup used for concentric bars since bar eccentricity would create rotation of the concrete block specimens, thus exerting non-uniform bearing pressure on top of concrete. The views of test setup used for eccentric specimens are shown in Fig. 7(c) and (d). In order to avoid rotation of concrete blocks, they were secured using four hollow steel sections (HSS), two at each side as shown in Fig. 7(d). These steel sections were tied to the concrete wall of the structures lab by four long steel threaded rods of 12 mm diameter and tightened by nuts and washers at both ends. Two steel channel sections were used at the other side of

Table 3 Characteristic concrete strength. Specimen No.

Cylinder size, diameter  height (mm)

Ultimate load (kN)

Compressive strength (MPa)

Average strength (MPa)

Characteristic strength (MPa)

1 2 3 4 5 6

100  200

545.85 544.28 531.72 534.07 543.5 500.3

69.5 69.3 67.7 68 69.2 63.7

67.9

57.9

287

S. Islam et al. / Construction and Building Materials 83 (2015) 283–298

(a) Straight bars of 16 and 12mm diameters

(b) Headed-end bars

(c) Sliced headed-bar

Fig. 6. Configurations of the used GFRP bars. the concrete wall of the lab to accommodate steel rods and tighten the rods as shown in Fig. 7(e). All parts on top of the HSS sections were the same as the test setup used in testing concentric GFRP bars except the potentiometers. In this test setup, three potentiometers were used. Two were used at the front side of the concrete wall as shown in Fig. 7(d) to measure any rotation as well as any solid movement in the concrete block. Additional potentiometer was used at the bottom of the specimen to measure pure bar slip. This potentiometer was placed carefully on the ground and was secured to several steel plates to prevent its movement as shown in Fig. 7(f). A custom-made grip that was designed for pullout testing of ribbed-surface GFRP bars and was used to perform pullout testing is shown in Fig. 7(g) and (h). It consists of a steel cylinder with inner conical hollow shape. It has two handles to carry it as shown in Fig. 7(g). To grip the bar, three wedges with threads were manufactured for each bar diameter. First, the grip cylinder should be placed while the GFRP bar is projecting from the middle of the cylinder. Then, three wedges were

Table 4 Material properties of the ribbed-surface GFRP bars [45]. Property

Value

Characteristic value short-term tensile strength (MPa) Characteristic value long-term tensile strength (MPa) Design value long-term tensile strength (MPa) Tension modulus of elasticity (GPa) Strain at ultimate limit state (‰) Transverse shear strength (MPa) Density (g/cm3) Critical temperature (°C)

1188 580 445 60 7.4 150 2.2 400

(e)

(a)

(b)

(c)

(d)

(f)

(g)

(h)

Fig. 7. Test setup: (a) schematic diagram of the test setup for the concentric pullout testing; (b) view of the test setup for concentric bar testing; (c) schematic diagram of the test setup for the eccentric pullout testing; (d) view of front side of the test setup for eccentric bar testing; (e) view of back side of the setup for eccentric bar test; (f) view of the potentiometers at back side of the concrete cube for eccentric bar test; (g) views of the steel grip cylinder and (h) view of the wedges used in the anchorage system.

288

S. Islam et al. / Construction and Building Materials 83 (2015) 283–298

inserted in the cylinder. After assembling the test setup, the pullout test was performed by applying the pullout load at a specified rate using the hydraulic jack operated in an open loop control. The load was applied to the GFRP bar at a rate not greater than 22 kN/min, while the free end slip was recorded with an accuracy of 0.001 mm [48]. The data from the load cell and potentiometers were recorded using test control software (TCS) with a data acquisition unit. The data acquisition system recorded the applied load with a precision of 0.01 kN. According to the S806-02 pullout test requirements, the test was terminated when one of the following conditions occurred: (a) the FRP bar ruptured; or (b) FRP bar slipped a distance at least equal to its diameter.

5 and 6 present the test results of the 12 mm diameter GFRP bars for both straight- and headed-end cases, respectively, while Tables 7 and 8 show the opponent results of the 16 mm diameter GFRP bars. In the current study, the characteristic tensile loads are used instead of average failure loads in order to account for the variation of the actual tensile loads for each specimen. According to S807-10 [47], the characteristic load can be calculated using Eq. (2).

2.4. Specimen nomenclature

Ft ¼

Tables 5–8 presents test parameters and associated specimen descriptions. The specimen nomenclature consists of 4 symbols separated by a dash. The first symbol indicates the bar size and type of bar (12P = 12 mm straight, 12H = 12 mm headed end, 16P = 16 mm straight, 16H = 16 mm headed end). The second nomenclature stands for concrete cover configuration (C = concentric, E1.5 = eccentricity or clear concrete cover equal to 1.5 times the bar diameter, E2.5 = eccentricity or clear concrete cover equal to 2.5 times the bar diameter). The third term in bar designation specifies the Le1 and Le2 embedment lengths shown in Fig. 1 in terms of the bar diameter d (e.g. 4d means that Le1 or Le2 are equal to 4 times the bar diameter). The fourth symbol in the designation indicates the number of identical specimen in the group, ranging from 01 to 06 since there are six identical specimens in total. For instance, 12H-C-0d-01 can be interpreted as follows: 12H = 12 mm headed bar; C = the eccentricity type is concentric; 0d = embedment length consists of the head length where Le2 = 0; and 01 = specimen number 1 in the group.

3. Results and discussions Tables 5–8 present the test results of the studied 30 parameters for all specimens including the failure load as well as the manifested mode of failure for each of the 180 test specimens. Tables

1  1:645V pffiffi 1 þ 1:645V n

!  F av

ð2Þ

where Ft = the characteristic failure load; Fav = the average failure load; n = number of identical samples = 6; V = coefficient of variation of failure loads. Tables 5–8 show the characteristic failure load as well as the average failure load for each parameter. It can be noted that the characteristic failure loads range from 0.58 to 0.85 of the average loads. The higher value corresponds to the smaller variation among the failure loads, while the lower value corresponds to larger variation among the failure loads. According to the S708-10, irregular mode of failure can be encountered since it stated that the development length of the FRP bar tested shall be taken as the longer of the embedment lengths of two consecutively tested specimens, one of which failed by FRP rupture and the other by bond slippage or splitting of concrete (S807-10 clause D8.3). In this part, test results are presented and analyzed in details for the pullout capacity of straight- and headed-end 12 and 16 mm GFRP bars for different embedment lengths (4db and 6db for

Table 5 Test result of the 12 mm GFRP bars with straight ends. Parameter #

Test No.

Specimen description

Failure load (kN)

Average failure load (kN) (5)

Coefficient of variation (COV)

Characteristic load, kN based on S807-10 (7)

(7)/(5)

Mode of failure

1

1 2 3 4 5 6

12P-C-4d-01 12P-C-4d-02 12P-C-4d-03 12P-C-4d-04 12P-C-4d-05 12P-C-4d-06

25.12 19.25 16.6 20.24 18.2 NA

19.88

0.162

13.03

0.66

Bar Bar Bar Bar Bar Bar

pullout pullout pullout pullout pullout pullout

2

1 2 3 4 5 6

12P-E1.5d-4d-01 12P-E1.5d-4d-02 12P-E1.5d-4d-03 12P-E1.5d-4d-04 12P-E1.5d-4d-05 12P-E1.5d-4d-06

23.59 20.17 15.8 22.54 13.52 18.25

18.98

0.205

11.05

0.58

Bar Bar Bar Bar Bar Bar

pullout pullout pullout pullout pullout pullout

3

1 2 3 4 5 6

12P-E2.5d-4d-01 12P-E2.5d-4d-02 12P-E2.5d-4d-03 12P-E2.5d-4d-04 12P-E2.5d-4d-05 12P-E2.5d-4d-06

19.48 28.88 19.58 21.63 17.43 23.57

21.76

0.187

13.39

0.62

Concrete block Bar pullout Concrete block Bar pullout Concrete block Concrete block

4

1 2 3 4 5 6

12P-C-6d-01 12P-C-6d-02 12P-C-6d-03 12P-C-6d-04 12P-C-6d-05 12P-C-6d-06

30.47 29.68 25.07 24.15 32.28 23.39

27.51

0.136

19.54

0.71

Bar Bar Bar Bar Bar Bar

5

1 2 3 4 5 6

12P-E1.5d-6d-01 12P-E1.5d-6d-02 12P-E1.5d-6d-03 12P-E1.5d-6d-04 12P-E1.5d-6d-05 12P-E1.5d-4d-06

34.35 27.76 22.02 30.27 24.63 25.3

27.39

0.161

18.15

0.66

Concrete block split Cover failure Bar pullout Concrete block split Bar pullout Bar pullout

6

1 2 3 4 5 6

12P-E2.5d-6d-01 12P-E2.5d-6d-02 12P-E2.5d-6d-03 12P-E2.5d-6d-04 12P-E2.5d-6d-05 12P-E2.5d-6d-06

30.96 29.29 31.48 25.37 40.83 25.82

30.63

0.183

19.05

0.62

Concrete block Concrete block Concrete block Concrete block Bar pullout Bar pullout

split split split split

pullout pullout pullout pullout pullout pullout

split split split split

289

S. Islam et al. / Construction and Building Materials 83 (2015) 283–298 Table 6 Test result for 12 mm GFRP bars with headed ends. Parameter #

Test No.

Specimen description

7

1 2 3 4 5 6

12H-C-0d-01 12H-C-0d-02 12H-C-0d-03 12H-C-0d-04 12H-C-0d-05 12H-C-0d-06

8

1 2 3 4 5 6

9

10

Failure load (kN)

Average failure load (kN) (5)

Coefficient of variation (COV)

Characteristic load, kN based on S807-10 (7)

(7)/(5)

Mode of failure

77.45 60.12 64.14 76.53 66.28 78.74

70.54

0.113

53.36

0.76

Bar Bar Bar Bar Bar Bar

12H-E1.5d-0d-01 12H-E1.5d-0d-02 12H-E1.5d-0d-03 12H-E1.5d-0d-04 12H-E1.5d-0d-05 12H-E1.5d-0d-06

75.15 78.91 60.25 68.39 55.14 64.12

66.99

0.134

47.88

0.71

Side blow-out Bar pullout Cover failure Concrete block split Cover failure Cover failure

1 2 3 4 5 6

12H-E2.5d-0d-01 12H-E2.5d-0d-02 12H-E2.5d-0d-03 12H-E2.5d-0d-04 12H-E2.5d-0d-05 12H-E2.5d-0d-06

76.23 68.55 60.31 68.84 70.4 75.34

69.95

0.082

57.34

0.82

Cover failure Cover failure Cover failure V-shaped cover failure V-shaped cover failure Cover failure

1 2 3 4

12H-C-4d-01 12H-C-4d-02 12H-C-4d-03 12H-C-4d-04

115.04 87.93 82.59 95.19

96.03

0.152

65.28

0.7

Bar Bar Bar Bar

5 6

12H-C-4d-05 12H-C-4d-06

82.6 112.8

11

1 2 3 4 5 6

12H-E1.5d-4d-01 12H-E1.5d-4d-02 12H-E1.5d-4d-03 12H-E1.5d-4d-04 12H-E1.5d-4d-05 12H-E1.5d-4d-06

53.75 67.55 80.3 69.39 72.8 83.5

71.22

0.148

48.96

0.69

Diagonal concrete cover Side blow-out Cover failure Side blow-out Cover failure Cover failure

12

1 2 3 4 5 6

12H-E2.5d-4d-01 12H-E2.5d-4d-02 12H-E2.5d-4d-03 12H-E2.5d-4d-04 12H-E2.5d-4d-05 12H-E2.5d-4d-06

89.61 84.7 91.1 80.2 74.38 89.2

84.87

0.077

70.52

0.83

Cover failure Concrete block split Bar pullout Cover failure Bar pullout Diagonal concrete cover

13

1 2 3 4 5 6

12H-C-6d-01 12H-C-6d-02 12H-C-6d-03 12H-C-6d-04 12H-C-6d-05 12H-C-6d-06

94.53 108.35 90.59 94.15 99.79 102.98

98.4

0.067

83.83

0.85

Bar Bar Bar Bar Bar Bar

14

1 2 3 4 5 6

12H-E1.5d-6d-01 12H-E1.5d-6d-02 12H-E1.5d-6d-03 12H-E1.5d-6d-04 12H-E1.5d-6d-05 12H-E1.5d-6d-06

66.4 96.23 73.59 56.59 67.32 79.29

73.24

0.186

45.23

0.62

Side blow-out Cover failure Bar pullout Diagonal concrete cover Side blow-out Side blow-out

15

1 2 3 4 5 6

12H-E2.5d-6d-01 12H-E2.5d-6d-02 12H-E2.5d-6d-03 12H-E2.5d-6d-04 12H-E2.5d-6d-05 12H-E2.5d-6d-06

103 84.69 84.53 86.76 83.43 114.3

92.79

0.138

65.56

0.71

Cover failure Cover failure Bar pullout Bar pullout Bar pullout Bar rupture

pullout pullout pullout pullout pullout pullout

rupture pullout pullout rupture

Bar pullout Bar rupture

straight-end bars and 0db, 4db and 6db for headed-end bars) and different concrete covers (concentric, 1.5db and 2.5db). Test results are presented along with their associated modes of failure. The load-slip relationships are discussed and a parametric study is conducted in order to investigate the effect of embedment length, concrete cover, and bar diameter on the failure loads. 3.1. Modes of failure In this section, the influence of different parameters on the mode of failure is investigated. Seven types of failure were observed for

rupture rupture rupture rupture pullout rupture

both straight- and headed-end GFRP bars as shown in Fig. 8. The following presents a brief description of these modes of failure: (1) Pullout failure: the bar is pulled out of the specimen without any splitting/cracking in the concrete as illustrated in Fig. 8(a). This mode of failure occurred mainly in case of straight-end bars with concentric configuration or eccentric configuration with sufficient concrete cover. 61 specimens out of the total test specimens developed this mode of failure, which represents 33.9% of the total number of the test specimens.

290

S. Islam et al. / Construction and Building Materials 83 (2015) 283–298

Table 7 Test result for 16 mm GFRP bars with straight ends. Parameter #

Test No.

Specimen description

Failure load (kN)

Average failure load (kN) (5)

Coefficient of variation (COV)

Characteristic load, kN based on S807-10 (7)

(7)/(5)

Mode of failure

16

1 2 3 4 5 6

16P-C-4d-01 16P-C-4d-02 16P-C-4d-03 16P-C-4d-04 16P-C-4d-05 16P-C-4d-06

37.21 41.51 53.84 42.2 50.51 42.61

44.65

0.14

31.45

0.7

Bar Bar Bar Bar Bar Bar

17

1 2 3 4 5 6

16P-E1.5d-4d-01 16P-E1.5d-4d-02 16P-E1.5d-4d-03 16P-E1.5d-4d-04 16P-E1.5d-4d-05 16P-E1.5d-4d-06

26.84 40.07 29.86 27.91 31.13 38.38

32.37

0.171

20.85

0.64

Diagonal concrete cover Bar pullout Concrete block split Concrete block split Concrete block split Bar pullout

18

1 2 3 4 5 6

16P-E2.5d-4d-01 16P-E2.5d-4d-02 16P-E2.5d-4d-03 16P-E2.5d-4d-04 16P-E2.5d-4d-05 16P-E2.5d-4d-06

36.58 44.17 39.75 37.92 33.89 32.97

37.55

0.109

28.69

0.76

Bar Bar Bar Bar Bar Bar

19

1 2 3 4 5 6

16P-C-6d-01 16P-C-6d-02 16P-C-6d-03 16P-C-6d-04 16P-C-6d-05 16P-C-6d-06

50.07 48.38 55.27 67.71 61.97 57.51

56.82

0.128

41.3

0.73

Bar pullout Bar pullout Bar pullout Bar pullout Bar pullout Concrete block split

20

1 2 3 4 5 6

16P-E1.5d-6d-01 16P-E1.5d-6d-02 16P-E1.5d-6d-03 16P-E1.5d-6d-04 16P-E1.5d-6d-05 16P-E1.5d-6d-06

53.14 44.35 51 57.51 47.14 36.58

48.29

0.152

32.85

0.68

Concrete block split Concrete block split Concrete block split Concrete block split V-shaped cover failure Concrete block split

21

1 2 3 4 5 6

16P-E2.5d-6d-01 16P-E2.5d-6d-02 16P-E2.5d-6d-03 16P-E2.5d-6d-04 16P-E2.5d-6d-05 16P-E2.5d-6d-06

50.36 48.65 43.48 39.37 56.83 57.28

49.33

0.145

34.27

0.69

Diagonal Diagonal Concrete Diagonal Concrete Concrete

(2) Concrete block-split: A crack that started in the smallest cover propagates through the whole concrete block as illustrated in Fig. 8(b). This mode of failure occurred mainly in the case of headed-end bars with concentric configuration or eccentric configuration with sufficient concrete cover as well as straight-end bars of 6db embedded length. 51 specimens out of the total test specimens developed this mode of failure, which represents 28.4% of the total number of the test specimens. (3) Bar rupture: the bar ruptured when its tensile strength was exhausted as illustrated in Fig. 8(c). This mode of failure occurred mainly in the case of headed-end bars of 12 and 16 mm diameters when the applied load exceeded the bar tensile capacity. 24 specimens out of the total test specimens developed this mode of failure, which represents 13.3% of the total number of the test specimens. (4) Side blow-out: the concrete part in the smaller cover side was blowed-out as depicted in Fig. 8(d). This mode of failure occurred mainly in the case of headed-end bars with smallest concrete cover, 1.5db. 18 specimens out of the total test specimens developed this mode of failure, which represents 10% of the total number of the test specimens. (5) Cover failure: A crack in the smallest cover was suddenly formed, with immediate loss of bond resistance leading to separation of the GFRP bar from the concrete block as depicted in Fig. 8(e). This mode of failure occurred mainly in the case of headed-end bars with smaller concrete cover.

pullout pullout pullout pullout pullout pullout

pullout pullout pullout pullout pullout pullout

concrete cover concrete cover block split concrete cover block split block split

16 specimens out of the total test specimens developed this mode of failure, which represents 8.9% of the total number of the test specimens. (6) Diagonal concrete cover failure: due to the lack of homogeneity of the concrete, the developed cover cracks deviated in the lateral direction in the case of eccentric configuration as depicted in Fig. 8(f). This mode of failure occurred mainly in the case of straight-end and headed-end bars with smaller concrete cover. 7 specimens out of the total test specimens developed this mode of failure, which represents 3.8% of the total number of the test specimens. (7) V-shaped cover failure: this mode of failure is similar to the pervious one except that failed concrete cover takes the Vshape as depicted in Fig. 8(g). This mode of failure manifested by only 3 specimens which represents 1.7% of the total number of the test specimens. In summary, for headed-end bars, the concrete cover has to be sufficient enough in order to be able to eliminate the concrete cover failure. In addition, bar size can affect or detain the concrete block-split failure mode where the concrete block-split failure was developed mainly by the 16 mm diameter bars. For the 12 mm diameter straight bars, the clear concrete cover of 1.5db was enough to ensure pure pullout failure, whereas for the 16 mm straight-end bars, the same concrete cover (1.5db) was not enough to prevent cracking. However, considering all studied parameters, the main failure mode was pure bar pullout, which represented

291

S. Islam et al. / Construction and Building Materials 83 (2015) 283–298 Table 8 Test result for 16 mm GFRP bars with headed ends. Parameter #

Test No.

Specimen description

Failure load (kN)

Average failure load (kN) (5)

Coefficient of variation (COV)

22

1 2 3 4 5 6

16H-C-0d-01 16H-C-0d-02 16H-C-0d-03 16H-C-0d-04 16H-C-0d-05 16H-C-0d-06

142.67 107.84 123.94 136.87 122.99 104.36

123.11

0.124

23

1 2 3 4 5 6

16H-E1.5d-0d-01 16H-E1.5d-0d-02 16H-E1.5d-0d-03 16H-E1.5d-0d-04 16H-E1.5d-0d-05 16H-E1.5d-0d-06

80.31 83.5 113.7 110.5 97.38 79.59

94.16

0.163

62.13

24

1 2 3 4 5 6

16H-E2.5d-0d-01 16H-E2.5d-0d-02 16H-E2.5d-0d-03 16H-E2.5d-0d-04 16H-E2.5d-0d-05 16H-E2.5d-0d-06

103.27 132.6 120.5 97.07 96.69 129.7

113.31

0.144

25

1 2 3 4 5 6

16H-C-4d-01 16H-C-4d-02 16H-C-4d-03 16H-C-4d-04 16H-C-4d-05 16H-C-4d-06

140.11 147.91 138.09 157.31 148.53 181.04

152.17

26

1 2 3 4 5 6

16H-E1.5d-4d-01 16H-E1.5d-4d-02 16H-E1.5d-4d-03 16H-E1.5d-4d-04 16H-E1.5d-4d-05 16H-E1.5d-4d-06

91.54 109.73 98.21 130.2 107 100.07

27

1 2 3 4 5 6

16H-E2.5d-4d-01 16H-E2.5d-4d-02 16H-E2.5d-4d-03 16H-E2.5d-4d-04 16H-E2.5d-4d-05 16H-E2.5d-4d-06

28

1 2 3 4 5 6

29

30

(7)/(5)

Mode of failure

0.74

Bar rupture Concrete block Concrete block Concrete block Concrete block Concrete block

split split split split split

0.66

Concrete Concrete Concrete Concrete Concrete Concrete

block block block block block block

split split split split split split

78.82

0.7

Concrete Concrete Concrete Concrete Concrete Concrete

block block block block block block

split split split split split split

0.103

118.12

0.78

Bar Bar Bar Bar Bar Bar

106.13

0.127

77.4

0.73

Side blow out Concrete block split Side blow out Side blow out Side blow out Concrete block split

126.6 109.42 122.63 105.51 153.72 154.52

128.73

0.165

84.55

0.66

Concrete block Concrete block Concrete block Concrete block Side blow out Side blow out

16H-C-6d-01 16H-C-6d-02 16H-C-6d-03 16H-C-6d-04 16H-C-6d-05 16H-C-6d-06

161.31 153.53 140.34 145.69 132.5 142.32

145.95

0.07

0.85

Bar Bar Bar Bar Bar Bar

1 2 3 4 5 6

16H-E1.5d-6d-01 16H-E1.5d-6d-02 16H-E1.5d-6d-03 16H-E1.5d-6d-04 16H-E1.5d-6d-05 16H-E1.5d-6d-06

125.46 101.98 142.46 99.45 102.82 144.3

119.41

0.174

76.26

0.64

Side Side Side Side Side Side

1 2 3 4 5 6

16H-E2.5d-6d-01 16H-E2.5d-6d-02 16H-E2.5d-6d-03 16H-E2.5d-6d-04 16H-E2.5d-6d-05 16H-E2.5d-6d-06

156.12 172.99 110.11 117.31 144.54 122.61

137.28

0.179

86.34

0.63

Bar rupture Bar rupture Concrete block Concrete block Concrete block Concrete block

about one third of the total number of the tested specimens. For more clarity, Fig. 9 shows a pie chart for all developed modes of failure. 3.2. Bond stress-slip relationships Fig. 10 shows the relationship between the applied load and the corresponding free-end slip for selected specimens. The selected specimens were chosen to represent the same parameter for both bar sizes, i.e., the same concrete cover condition (concentric case) and the same embedded length (6db). It can be noticed that using head-end resulted in increase in the resisting applied load by about three times and two times their values of straight-end bars for the

Characteristic load, kN based on S807-10 (7) 90.57

123.4

rupture rupture rupture rupture rupture rupture

split split split split

rupture rupture rupture rupture rupture rupture blow blow blow blow blow blow

out out out out out out

split split split split

case of 12 and 16 mm bar diameters, respectively. In addition, the resisting load by the 16 mm diameter bars showed higher load resisting values compared to those of the opponent specimens of 12 mm diameter. The average bond stress along the embedded length for each specimen was calculated using Eq. (3).

s¼

F

pdb he

ð3Þ

where s = average bond stress (MPa); F = applied load (N); db = bar diameter (mm) and he = embedment length considering the head length as given in Table 2 (mm). Fig. 11 shows the bond stress versus free- end slip relationship for the selected specimens. For these

292

S. Islam et al. / Construction and Building Materials 83 (2015) 283–298

(a) Bar pullout

(b) Concrete b lock-split

(e) Concrete cover failure

(c) Bar rupture

(d) Concrete side blow-out

(f) Diaonal concrete cover failure (g) V-shaped concrete cover failure

Fig. 8. Typical modes of failures observed experimentally.

V-shaped Concrete concrete cover cover failure, failure, [VALUE] [VALUE]

Bar rupture, 13.30%

160 12P-C-6d-06 12H-C-6d-03 16P-C-6d-01 16H-C-6d-05

140 120

Bar pullout, [VALUE]

Load, kN

Concrete side blow-out, [VALUE]

[CATEGORY NAME] failure, [VALUE]

100 80 60

Concrete block-split, 28.40%

40 20 0

Fig. 9. Percentage of occurrence of each mode of failure.

0

2

4 6 Slip, mm

8

10

Fig. 10. Pullout force versus free-end slip for selected specimens.

18 12P-C-6d-06 12H-C-6d-03 16P-C-6d-01 16H-C-6d-05

16 14 Bond stress, MPa

specimens, the bond-slip relationships were mainly linear up to approaching the maximum bond stress, and then some nonlinearity was developed up to the maximum bond stress. After reaching the maximum bond stress, according to the exhibited mode of failure, the bond-slip relationship showed smooth decrease in the bond stress for the case of the pullout failure as manifested by specimens 12P-C-6d-06 and 16P-C-6d-01 or a sudden drop in the bond stress due to bar rupture mode of failure as manifested by specimens 12H-C-6d-03 and 16H-C-6d-05. As shown in Fig. 11, despite the specimen 12H-C-6d-03 resisted smaller load than that of specimen 16H-C-6d-05 by about 37%, it exhibited higher bond stress. This can be attributed to the increased embedment length for the 16 mm diameter bars (196 mm) while it is 147 mm for 12 mm diameter bar. Combining this effect with the smaller size of 12 mm diameter bar in the denominator of Eq. (3) can explain such finding. For all specimens, the bond-slip relationships were constructed where the decreasing branch showed different plateaus according to the developed mode of failure and these relationships can be found elsewhere [49].

12 10 8 6 4 2 0 0

2

4

6 Slip, mm

8

10

3.3. Parametric study Fig. 11. Bond-stress versus free-end slip for selected specimens.

Fig. 12 shows comparison among the average bond stresses associated with all studied parameters in general. It can be noticed that the highest bond stress was developed by the headed-end bars of 12 mm diameter in the case of embedment length was represented by the head length only, while the lowest bond stress was

developed by straight-end bars of 12 mm diameter for the case of eccentric bars for 1.5db concrete cover. However, the effects of embedment length, concrete cover, and bar diameter on the average bond stress are discussed in detail in the following subsections.

293

S. Islam et al. / Construction and Building Materials 83 (2015) 283–298

Bond stress, MPa

25 20 15 10 5

he, mm

48 4d

75

6d

0d

Straight

123

147

64

4d

6d

4d

Headed

96

100

6d

0d

12 mm

164

196

4d

6d

8d

2.5d

8d

1.5d

2.5d

8d

Straight

1.5d

2.5d

5d

1.5d

2.5d

5d

1.5d

2.5d

1.5d

2.5d

10.5d

1.5d

2.5d

10.5d

1.5d

2.5d

72

10.5d

7d

1.5d

2.5d

7d

1.5d

2.5d

1.5d

Concrete cover

0

Headed 16 mm

Fig. 12. Bar chart for the average bond stress corresponding to all studied parameters.

Average bond stress, MPa

25 12P-1.5d 12P-2.5d 12P-7d 16P-1.5d 16P-2.5d 16P-5d 12H-1.5d 12H-2.5d 12H-10.5d 16H-1.5d 16H-2.5d 16H-8d

20 15 10 5 0 0

2

4

6 8 Embedment length/ bar diameter

10

12

14

Fig. 13. Relationship between the embedment length/bar diameter for all specimens and the corresponding average bond stress.

3.3.1. Effect of embedment length on bond stress For the straight-end GFRP bars, the embedment length was recognizable to be either 4db or 6db. On the other hand, for headed-end GFRP bars, the embedment lengths consisted of the length Le2 as illustrated in Fig. 4 and their values are given in Table 2 in addition to the head length as given Table 1. Thus, by summation and these two lengths for both bar diameters are then divided by the bar diameter, the embedment lengths for both bar diameters are 6.25db, 10.25db and 12.25db, respectively. Fig. 13 depicts the effect of the embedment length on the average bond stress for both straight- and headed-end bars of both 12 and 16 mm diameter corresponding to different concrete covers. It is worth mentioning that the second symbol in the specimen designation corresponding to the concrete cover as a multiplier of the bar diameter. Fig. 13 shows that the average bond stress decreases with increase in the embedment length for both straight- and headedend bars for either 12 or 16 mm bar diameter. This finding matches the obtained resultsby others [2,10,11,14,15,20,23,27,29,30]. It can be noted from reported results in Table 9 that the average failure load increases by increasing the embedment length, however the corresponding bond stress decreases as illustrated by Eq. (3). 3.3.2. Effect of concrete cover on bond stress Fig. 14 depicts the effect of the concrete cover on the average bond stress for both straight- and headed-end bars of either 12 or 16 mm diameter corresponding to different embedment

lengths. Based on Fig. 14, it is evident that there is an increase in the bond stress with increase in concrete cover from 1.5db to 2.5db for the 12 mm diameter straight-end bars, while the bond stress decreased with increase in the concrete cover from 2.5db to 7db for both embedment lengths 4db to 6db. This can be attributed to the adopted average approach of the results. On contrary to the 12 mm diameter bars, the 16 mm diameter straight-end bars showed increased bond stress with increase in the concrete cover from 2.5db to 5db. For headed-end bars, the 16 mm diameter bars showed increased average bond stress with increase in the concrete cover from 1.5db to 2.5db, while these increases showed lower rate moving to concrete cover of 8db. On the other hand, the 12 mm diameter bars showed slight increase beyond concrete cover of 2.5db. For both straight- and headed-end bars, the average bond stresses corresponding to the lowest embedment length showed the highest values for both 12 and 16 mm bar diameters. It can be concluded that a concrete cover of 2.5db is sufficient in order to provide enhanced bond stress without concrete cover failure since the concrete cover provides confinement to the bars which increases the bond strength [18,19,28,29,43,50–52]. 3.3.3. Effect of bar diameter on bond stress Based on reported results in Table 9, there is an increase in failure load capacity with increase in bar diameter from 12 to 16 mm in all cases of straight-end bars. The failure load of specimens with 6db embedment depth is much more than those with 4db embedment depths. The same findings were obtained for headed-end

294

S. Islam et al. / Construction and Building Materials 83 (2015) 283–298

Table 9 Average failure loads and corresponding bond strength for the tested specimens. Bar diameter (mm)

End type

12

Straight

Embedment length

Headed

16

Straight

Headed

Concrete cover

Average failure Load (kN) (7)

Characteristic load, kN based on S807-10 (8)

(8)/(7)

Average bond strength (MPa)

Average slip at maximum load (mm)

Times d

mm

Times d

mm

4d

48

6d

72

0d

75

4d

123

6d

147

1.5d 2.5d 7d 1.5d 2.5d 7d 1.5d 2.5d 10.5d 1.5d 2.5d 10.5d 1.5d 2.5d 10.5d

18 30 93 18 30 93 18 30 143 18 30 143 18 30 143

18.98 21.76 19.88 27.39 30.63 27.51 66.99 69.95 70.54 71.22 84.87 96.03 73.24 92.79 98.40

11.05 13.39 13.03 18.15 19.05 19.54 47.88 57.34 53.3 48.96 70.52 65.28 45.23 65.56 83.83

0.58 0.62 0.66 0.66 0.62 0.71 0.71 0.82 0.76 0.69 0.83 0.68 0.62 0.71 0.85

6.11 7.40 7.28 6.69 7.02 7.20 16.94 20.29 18.86 10.57 15.21 14.09 8.16 11.83 15.14

0.37 0.25 0.23 0.27 0.19 0.21 0.21 0.16 0.11 0.14 0.12 0.06 0.12 0.11 0.04

4d

64

6d

96

0d

100

4d

164

6d

196

1.5d 2.5d 5d 1.5d 2.5d 5d 1.5d 2.5d 8d 1.5d 2.5d 8d 1.5d 2.5d 8d

24 40 91 24 40 91 24 40 141 24 40 141 24 40 141

32.37 37.55 44.65 48.29 49.33 56.82 94.16 113.31 123.11 106.13 128.73 152.17 119.41 137.28 145.95

20.85 28.69 31.45 32.85 34.27 41.30 62.13 78.82 90.57 77.40 84.55 118.12 76.26 86.34 123.4

0.64 0.76 0.70 0.68 0.69 0.73 0.66 0.70 0.74 0.73 0.66 0.78 0.64 0.63 0.85

6.49 8.92 9.78 6.81 7.11 8.56 12.37 15.69 18.02 9.39 10.26 14.34 7.74 8.77 12.53

0.31 0.20 0.19 0.21 0.14 0.15 0.16 0.11 0.08 0.10 0.09 0.05 0.07 0.06 0.03

25

Bond stress, MPa

20 15 10 5

12P-4d 12H-4d

12P-6d 12H-6d

16P-4d 16H-0d

16P-6d 16H-4d

12H-0d 16H-6d

0 0

2

4

6 Concrete cover/ bar diameter

8

10

12

Fig. 14. Relationship between the concrete cover/bar diameter for all specimens and the corresponding average bond stress.

bars that the increase in bar diameter generally increases the ultimate load. However, the corresponding bond stress decreased with increase in the bar diameter where such increases in the ultimate loads were accompanied by higher embedment length. Even at the same embedment length, the increases in the ultimate loads could not compensate the decreases in the bond stress due to larger bar diameter values in the denominator of Eq. (3). The test results assured that increasing the bar diameter lowers the bond strength of the GFRP bars [1,2,9,11,12,29,30]. 4. Development length of GFRP bars Eq. (3) provides the relationship between the average bond stress and the corresponding embedment length under the effect of the applied load. The development length is defined as the

minimum length required to fully develop the design tensile stress in the GFRP bars. Using the design value of the tensile strength of the GFRP bars as given in Table 4, Eq. (3) can be rewritten as follows:

s¼

Ab f t

ð4Þ

pdb ld

where s = average bond stress (MPa); ft = design tensile strength of the GFRP bars (MPa); db = bar diameter (mm); Ab = cross-sectional area of the bar (mm2) and ld = development length. From Eq. (4) the development length can be calculated as follows:

ld ¼

Ab f t

pdb s

¼

db f t 4s

ð5Þ

295

S. Islam et al. / Construction and Building Materials 83 (2015) 283–298

Since previous studies indicated that good correlation exists between the bond strength and the square root of the compressive strength of concrete [12,20,28] besides the bond stress is inversely proportional to the bar diameter size, the bond stress can be represented by Eq. (6) [1,2,9,11,12,29,30].

s¼A

qffiffiffiffi 0 fc

ð6Þ

db 0

where f c = concrete compressive cylinder strength (MPa) and constant A can be determined based on experimental results. Hence, Eq. (5) can be rewritten as follows:

ld ¼

by the comparison between the embedment length and the development length. For headed-end bars, the actual tensile stresses were higher than the design values. This means the provided bar heads enabled the GFRP bars to use their design tensile stresses where the head length only without additional embedment length was sufficient to reach the design tensile strength of the GFRP bars. Based on the calculated values, the constant B should be selected in such a way that the resulting equation yields a conservative value of the development length. Consequently, constant B used in Eq. (8) can be suggested as follows:

B ¼ 0:275 ðFor straight-end barsÞ

ð10Þ

le2 ðFor headed-end barsÞ B ¼ 0:13 þ 80d

2 db f t

b

qffiffiffiffi 0 4A f c

ð7Þ

Eq. (7) can be rewritten in the following form:

db f ld ¼ B qffiffiffiffit 0 fc

ð8Þ

where

qffiffiffiffi 0 fc db B¼ ¼ 4s 4A

ð9Þ

Table 10 summarizes the development length parameters including the actual tensile stress developed by the GFRP bars at the ultimate limit state along with the recommended design strength by the manufacturers [45]. Besides, constants A and B are included along with their average values corresponding to each embedment length. It can be noticed that the actual stresses for straight-end bars of either 12 or 16 mm diameter are lower than the design values. Theses bars developed pullout modes of failure. Thus, in order to enable these bars to exhaust their design tensile strength, the embedment lengths had to increase as manifested

where le2 = the additional straight part of the bar to the head as a multiplier of the bar diameter. Table 11 shows comparison between the average values of the bond length constant B and the proposed values for all cases. It can be noticed that the proposed values showed conservative results up to 27.5% compared to the experimental results. Thus, the development length for straight- and headed-end GFRP bars embedded in high-strength concrete can be proposed as follows:

db  f ld ¼ 0:275 qffiffiffiffit ðStraight-end barsÞ 0 fc ld ¼ ð0:13 þ

ð11Þ

le2 db :f t Þ: qffiffiffiffi ðHeaded-end barsÞ 0 80db f

ð12Þ

c

In order to verify the rationality of the proposed development length equation, the proposed equation is compared with those stipulated by different design standards. Since different standards stipulated development length equations for straight-end bars, the results of Eq. (11) was only compared with the results of CSA-S806-02 [48], CSA S6-06 [44], ACI 440-1R-06 [28], and

Table 10 Development length parameters for the tested specimens. Specimen

he (mm)

ft,actual (MPa) (3)

ft,design (MPa) (4)

(3)/(4)

s (MPa)

A

Average A

B

Average B

ld (mm) (11)

(2)/(11)

12P-E1.5d-4d 12P-E2.5d-4d 12P-C7d-4d 12P-E1.5d-6d 12P-E2.5d-6d 12P-C7d-6d 12H-E1.5d-0d 12H-E2.5d-0d 12H-C10.5d-0d 12H-E1.5d-4d 12H-E2.5d-4d 12H-C10.5d-4d 12H-E1.5d-6d 12H-E2.5d-6d 12H-C10.5d-6d 16P-E1.5d-4d 16P-E2.5d-4d 16P-C5d-4d 16P-E1.5d-6d 16P-E2.5d-6d 16P-C5d-4d 16H-E1.5d-0d 16H-E2.5d-0d 16H-C8d-0d 16H-E1.5d-4d 16H-E2.5d-4d 16H-C8d-4d 16H-E1.5d-6d 16H-E2.5d-6d 16H-C8d-6d

48 48 48 72 72 72 75 75 75 123 123 123 147 147 147 64 64 64 96 96 96 100 100 100 164 164 164 196 196 196

97.79 118.5 115.31 160.62 168.58 172.92 423.72 507.43 471.68 433.27 624.07 577.7 400.27 580.18 741.86 103.73 142.74 156.47 163.43 170.5 205.47 309.1 392.14 450.6 385.07 420.65 587.66 379.4 429.55 613.93

445

0.22 0.27 0.26 0.36 0.38 0.39 0.95 1.14 1.06 0.97 1.4 1.3 0.9 1.3 1.67 0.23 0.32 0.35 0.37 0.38 0.46 0.69 0.88 1.01 0.87 0.95 1.32 0.85 0.97 1.38

6.11 7.4 7.28 6.69 7.02 7.2 16.94 20.29 18.86 10.57 15.21 14.09 8.16 11.83 15.14 6.49 8.92 9.78 6.81 7.11 8.56 12.37 15.69 18.02 9.39 10.26 14.34 7.74 8.77 12.53

9.63 11.67 11.48 10.55 11.07 11.36 26.71 32 29.74 16.66 23.99 22.21 12.88 18.66 23.87 13.64 18.76 20.57 14.32 14.94 18 26 32.98 37.9 19.74 21.57 30.15 16.28 18.44 26.35

11.93

0.311 0.257 0.261 0.284 0.271 0.264 0.112 0.094 0.101 0.18 0.125 0.135 0.233 0.161 0.126 0.293 0.213 0.194 0.279 0.268 0.222 0.154 0.121 0.106 0.203 0.185 0.133 0.246 0.217 0.152

0.276

218.59 180.34 183.33 199.47 190.13 185.36 78.81 65.8 70.79 126.35 87.75 94.78 163.52 112.81 88.18 274.43 199.46 181.94 261.4 250.46 207.89 143.95 113.48 98.76 189.64 173.5 124.15 229.96 203.03 142.06

0.22 0.27 0.26 0.36 0.38 0.39 0.95 1.14 1.06 0.97 1.4 1.3 0.9 1.3 1.67 0.23 0.32 0.35 0.37 0.38 0.46 0.69 0.88 1.01 0.86 0.95 1.32 0.85 0.97 1.38

10.99

29.48

20.95

18.47

17.66

15.76

32.29

23.82

20.36

0.273

0.102

0.147

0.173

0.233

0.256

0.127

0.174

0.205

he = embedment length; ft,actual = actual developed tensile stress in the bar at the ultimate load level based on characteristic load; ft,design = design tensile strength as recommended by the manufacturer of the GFRP bars; A = constant A; B = constant B; ld = development length; s = average bond stress.

296

S. Islam et al. / Construction and Building Materials 83 (2015) 283–298

Table 11 Comparison between the average bond length parameter and the proposed one for all cases. Bar diameter (mm)

End type

Embedment length

Average B (4)

Proposed B (5)

(5)/(4)

12 12 12

Straight Straight Headed

4d 6d 0d + head length 4d + head length 6d + head length 4d 6d 0d + head length 4d + head length 6d + head length

0.276 0.273 0.102

0.275 0.275 0.130

0.996 1.007 1.275

12

Headed

12

Headed

16 16 16

Straight Straight Headed

16

Headed

16

Headed

ld ¼

0.147

0.180

1.224

0.173

0.197

1.139

0.233 0.256 0.127

0.275 0.275 0.130

1.180 1.074 1.024

0.174

0.180

1.034

0.205

0.205

1.000

JSCE-97 [53] standards. First, the development length equations for different standards are presented, then a comparison among them is presented including the proposed equation. According to the CSA-S806-02, the development length, ld, of FRP bars in tension can be calculated according to Eq. (13).

f ld ¼ k1 k2 k3 k4 k5 qFffiffiffiffi db 0 fc

ð13Þ

where k1 = bar location factor = 1 (positive moment); k2 = concrete density factor = 1 (normal density); k3 = bar size factor (0.8 for Ab < 300mm2 ); k4 = bar fiber factor (1 for GFRP bar); k5 = bar surface profile factor (1.05 for ribbed-surface); and fF = stress in FRP reinforcement under specified loads. Substituting the values of the known factors leads to:

f ld ¼ 0:42 qFffiffiffiffi db 0 fc

ðCSA-S806-02Þ

ð14Þ

The Canadian Highway Bridge Design Code (CSA S6-06) modified the available development length equation for steel bars in order to be applicable for FRP bars by multiplying the transverse reinforcement index by the modular ratio of FRP to steel bars. The resulting development length equation is given as follows:

ld ¼ 0:45 h

k1 k4 dcs þ ktr

Efrp Es

i

  f frp Afrp f cr

where the term ½dcs þ ktr

Efrp  Es

ð15Þ

qffiffiffiffi 0 6 2:5db ; f cr ¼ 0:4 f c ; k1 = bar location

factor = 1; k4 = bar surface profile factor = 0.8; and ffrp = stress in FRP reinforcement, MPa. Using mathematical manipulation and substituting the known values leads to:

f frp ld ¼ 0:283 qffiffiffiffi db 0 fc

The American Concrete Institute Code (ACI 440-1R-06) adopted Eq. (17) in order to calculate the development length of FRP bars.

ðCSAS6-06Þ

ð16Þ

a 0:083f frpffiffiffif 0  340 c

13:6 þ dC

db

ð17Þ

b

where a = bar location factor = 1; and

C db

is limited to 3.5.

Substituting the known parameters, the Eq. (17) becomes as follows:

0

1

f fr B C ld ¼ @0:705 qffiffiffiffi  19:88Adb 0 fc

ðACI440-1R-06Þ

ð18Þ

The Japanese Design Recommendation (JSCE, 1997) adopted the same concept used by CSA S6-06 design code where both of them made modifications to equations used to define the required development lengths for steel bars and applied it to FRP bars. The development length, ld, of FRP bars can be calculated by the following equation:

ld ¼ a1

f d db 4f bod

ðJSCE;1997Þ

ð19Þ

where fd = design tensile strength of FRP (MPa); a1 = 1.0 (where kc 6 1.0), 0.9 where 1.0 < kc 6 1.5, 0.8 where 1.5 < kc 6 2.0, 0.7 t Et ; c = the where 2.0 < kc 6 2.5 or 0.6 where 2.5 < kc, kc ¼ dc þ 15A Sd Eo b

b

smaller of: downward cover of main reinforcement or half of the space between the anchored reinforcement (mm); At = area of transverse reinforcement which is vertically arranged to the assumed splitting failure surface (mm2); S = distance between the centers of the transverse reinforcement (mm); Et = Young’s modulus of transverse reinforcement (MPa); E0 = standard Young’s modulus, 200,000 MPa; fbod = design bond strength of concrete (MPa) ;2=3

¼ 0:28a2 f ck =cc 6 3:2; a2 = modification factor for bond strength of FRP = 1.0 where bond strength is equal to or greater than that of deformed steel bars; otherwise value shall be reduced according to test results; fck = compressive strength of concrete (MPa); cc = characteristic value for the concrete compressive strength taken as 1.3. In order to compare the obtained development length based on the proposed equation for straight- end GFRP bars, against the results obtained by the mentioned standards, the concrete cylinder strength along with the design tensile strength of the GFRP bars were considered the same for all equations. The used values were 57.5 and 445 MPa for the concrete compressive strength and the design tensile strength of the GFRP bars, respectively. Table 12 summarizes the results of all development length along with comparisons between the calculated development lengths and the calculated value based on the proposed expression. It can be observed that all adopted standards showed conservative results against the proposed expression. The results of the CSA-S806-02 code showed the highest safety margin with about 51% over the results of the proposed expression, while the CSA S6-06 code exhibited the lowest safety margin with about 2% above the results of the proposed expression.

Table 12 Calculated development length for straight-end GFRP bars based on the proposed equation versus those calculated using available expressions in different standards. db (mm)

12 16

Proposed Eq. (11) (mm)

193 257

CSA-S806-02 [48], Eq. (14)

CSA S6-06 [44], Eq. (16)

ACI 440-1R-06 [28], Eq. (18)

JSCE-07 [53], Eq. (19)

mm

Compared to proposed value

mm

Compared to proposed value

mm

Compared to proposed value

mm

Compared to proposed value

292 389

1.51 1.51

196 262

1.02 1.02

251 334

1.30 1.30

290 387

1.50 1.50

S. Islam et al. / Construction and Building Materials 83 (2015) 283–298

297

5. Conclusions

References

The effect of different parameters on the bond behavior of both straight- and headed-end ribbed-surface GFRP bars embedded in high-strength concrete was studied. An experimental investigation was conducted using 180 pullout test specimens such that the effects of the GFRP bar diameter, the embedment length, and the concrete cover on the bond strength could be determined. The results from the experimental data lead to the following conclusions:

[1] Hao QD, Wang B, Ou JP. Fibre reinforced polymer rebar’s application to civil engineering. Concr J 2006;9:38–40. [2] Benmokrane B, Rahman H. Durability of fiber reinforced polymer (FRP) composites for construction. In: Proc., 1st int. conf. (CDCC ‘98), Québec, Canada; 1998. [3] Saadatmanesh H, Ehsani MR. Fibre composites in infrastructure. In: Proc., 2nd int. conf. on composites in infrastructure, ICCI’98, Tucson; 1998. [4] Dolan CW, Rizkalla, SH, Nanni A. In: Proc., 4th int. sym. fibre reinforcement polymer reinforcement for reinforced concrete structures, FRPRCS-4, ACISP188, Baltimore; 1999. [5] Razaqpur AG. In: Proc., 3rd int. conf. on advanced composite materials in bridges and structures, ACMBS-3, Ottawa, Canada; 2000. [6] Alves J, El-Ragaby A, El-Salakawy E. Durability of GFRP bars’ bond to concrete under different loading and environmental conditions. J Compos Constr 2011;15:249–62. [7] Masmoudi R, Masmoudi A, Ouezdou M, Daoud A. Long-term bond performance of GFRP bars in concrete under temperature ranging from 20 °C to 80 °C. Constr Build Mater 2011;25:486–93. [8] Neale KW, Labossiére P. Advanced composite materials on bridges and structures. In: 1st Int. conf., Canadian Society for Civil Engineering, Sherbrooke, Québec; 1992. [9] Nanni A. Fibre reinforced plastic (GFRP) reinforcement for concrete structures: properties and applications. Developments in Civil Engineering, vol. 42. Elsevier; 1993. pp. 450. [10] Nanni A, Dolan CW. Fibre reinforced plastic reinforcement for concrete structures. In: Int. Symp., SP-138, American Concrete Institute (ACI); 1993, SP-138. [11] Tighiouart B, Benmokrane B, Gao D. Investigation of bond in concrete member with fibre reinforced polymer (FRP) bars. Constr Build Mater 1998;12:453–62. [12] Okelo R, Yuan RL. Bond strength of fibre reinforced polymer rebars in normal strength concrete. J Compos Constr 2005;9:203–13. [13] Chaallal O, Benmokrane B. Pullout and bond of glass fibre rods embedded in concrete and cement grout. Mater Struct 1993;26:167–75. [14] Benmokrane B, Masmoudi R. FRP C-bar as reinforcing rod for concrete structures. In: Proc. of the 2nd int. conf. on advanced composite materials in bridges and structures (ACMBS II). Montreal: Canadian Society of Civil Engineering; 1996. p. 181–8. [15] Pecce M, Manfredi G, Realfonzo R, Cosenza E. Experimental and analytical evaluation of bond properties of GFRP bars. J Mater Civil Eng 2001;13:282–90. [16] Faza SS, GangaRao HV. Bending and bond behaviour of concrete beams reinforced with plastic rebars. Transportation Research Record No. 1290, 1990;2:185–93. [17] Cairns J, Abdullah RB. Bond strength of black and epoxy-coated reinforcement: a theoretical approach. ACI Mater J 1996;93:362–9. [18] Ehsani MR, Saadatmanesh H, Tao S. Bond of GFRP rebars to ordinary-strength concrete. In: ACI int. symp. on non-metallic continuous reinforcement. Canada: Vancouver; 1993. p. 333–45. [19] Galati N, Nanni A, Dharani L, Focacci F, Aiello A. Thermal effects on bond between FRP rebars and concrete. Composites Part A 2006;37:1223–30. [20] Ehsani MR, Saadatmanesh H, Tao S. Bond of hooked glass fibre reinforced plastic (GFRP) reinforcing bars to concrete. ACI Mater J 1995;92:391–400. [21] Cosenza E, Manfredi G, Realfonzo R. Analytical modelling of bond between FRP reinforcing bars and concrete. In: Taerwe L, editor. Proc., 2nd int. RILEM symp. (FRPRCS-2); 1995. [22] Cosenza E, Manfredi G, Realfonzo R. Bond characteristics and anchorage length of FRP rebars. In: El-Badry M, editor. Proc., 2nd int. conf. on advanced composite materials in bridge structures; 1996. [23] Cosenza E, Manfredi G, Realfonzo R. Behaviour and modeling of bond of FRP rebar to concrete. J Compos Constr 1997;1:40–51. [24] Tastani SP, Pantazopouou SJ. Bond of GFRP bars in concrete: experimental study and analytical interpretation. J Compos Constr 2006;10:381–91. [25] Lee J, Kim TY, Kim TJ, Yi CK, Park JS, You YC, et al. Interfacial bond strength of glass fiber reinforced polymer bars in high-strength concrete. Composites Part B 2008;39:258–70. [26] Mazaheripour H, Barros JA, Sena-Cruz JM, Pepe M, Martinelli E. Experimental study on bond performance of GFRP bars in self-compacting steel fiber reinforced concrete. Compos Struct 2013;95:202–12. [27] Xue W, Zheng Q, Yang Y, Fang Z. Bond behavior of sand-coated deformed glass fiber reinforced polymer rebars. J Reinf Plast Compos 2014;33:895–910. [28] ACI Committee 440. Guide for the Design and Construction of Structural Concrete Reinforced with FRP Bars (ACI 440.1R-06). Farmington Hills, MI: American Concrete Institute; 2006. [29] Aly R, Benmokrane B, Ebead U. Tensile lap splicing of fiber-reinforced polymer reinforcing bars in concrete. ACI Struct J 2006;103:857–64. [30] Achillides Z, Pilakoutas K. Bond behaviour of fibre reinforced polymer bars under direct pullout conditions. J Compos Constr 2004;8:173–81. [31] Malvar LJ. Bond stress-slip characteristics of FRP rebars. Report TR-2013-SHR, Naval facilities Engineering Service Center, Port Hueneme, California; 1994. [32] CEB-FIB. Model Code 2010 First complete draft Volume 1. Bulletin 55. FibInternational federation for concrete: Switzerland, Lausanne; 2010. [33] Malvar LJ. Tensile and bond properties of GFRP reinforcing bars. ACI Mater J 1995;92:276–85.

(1) The average failure load increased by increasing the embedment length for both straight- and headed-end bars. However, these increases were more significant in the case of straight-end bars. For headed-end bars, due to different concrete splitting mechanisms, there was not much difference between splitting force obtained for various embedment depths in each case. This is due to the stress concentration around the bars’ head which is similar in specimens with various embedment lengths having the same concrete cover. On the other hand, increasing the embedment length led to a decrease in the corresponding bond stress for both straight- and headed-end bars. (2) As the concrete cover increased, the amount of confinement provided by concrete increased and, as a result, the failure load increased in majority of cases along with increasing the corresponding bond stress. Increasing concrete cover further than 2.5db for both the 12 and 16 mm diameter bars did not result in noticeable increase in bond stress. This means that concrete cover of 2.5db can provide enough confinement to prevent premature concrete cover failure. (3) Increasing bar size showed significant increase in the ultimate load, while the corresponding bond stress decreased in all cases. Although, these increases in the ultimate load increased with increased embedment length, the corresponding bond stresses were decreased. (4) Using headed-end bars showed significant increases in both the ultimate load and the bond stress for all embedment lengths and concrete cover values compared to those of straight-end bars. Considering the case of headed-end bars without additional embedment lengths, the embedment length of the heads were 75 and 100 mm for the 12 and 16 mm diameter bars, respectively. These cases showed enhanced bond stresses by about 168% and 105% compared to the case of straight-end bars with embedment lengths of 6db for the 12 and 16 mm bar diameters, respectively. In sum, due to the configuration of the manufactured bar heads, it provides enhanced bond stress compared to straight-end bar with the same embedment length as the head length. (5) The available equations for calculating the development length for straight-end GFRP bars in different standards (CSA-S806-02, CSA S6-06, ACI 440-1R-06, and JSCE-97) showed conservative results compared to the calculated development length using the proposed development length expression.

Acknowledgements The financial support of Onatrio Centres of Excellence (OCE) in the form of research grant is greatly appreciated. In addition, the authors would like to thank Schoeck Canada Inc. for sponsoring this project.

298

S. Islam et al. / Construction and Building Materials 83 (2015) 283–298

[34] Cosenza E, Manfredi G, Realfonzo R. Analytical modelling of bond between FRP reinforcing bars and concrete. In: Taerwe L, editor. Non-metallic (FRP) reinforcement for concrete structures. E&FN SPON; 1995. [35] Alunno Rossetti V, Galeota D, Gianmatteo MM. Local bond stress-slip relationships of glass fibre reinforced plastic bars embedded in concrete. Mater Struct 1995;28:40–4. [36] Tepfers R, De Lorenzis L. Bond of FRP reinforcement in concrete – a challenge. Mech Compos Mater 2003;39:315–28. [37] Malvar LJ, Cox JV, Cochran KB. Bond between carbon fiber reinforced polymer bars and concrete. I: Experimental study. ASCE J Compos Constr 2003;7:154–63. [38] Caggiano A, Martinelli E. A unified formulation for simulating the bond behaviour of fibres in cementitious materials. Mater Design 2012;42:204–13. [39] RILEM/CEB/FIP. Test of the bond strength of reinforcement of concrete: test by bending, Recommendation RC.5, 5; 1978. [40] Faella C, Martinelli E, Nigro E. Direct versus indirect method for identifying FRP-to-concrete interface relationships. J Compos Constr 2009;13:226–33. [41] Martinelli E, Napoli A, Nunziata B, Realfonzo R. Inverse identification of a bearing-stress-interface-slip relationship in mechanically fastened FRP laminates. Compos Struct 2012;94:2548–60. [42] Pepe M, Mazaheripour H, Barros J, Sena-Cruz J, Martinelli E. Numerical calibration of bond law for GFRP bars embedded in steel fibre-reinforced selfcompacting concrete. Composites Part B 2013;50:403–12. [43] ACI Committee 408. Bond and development of straight reinforcing bars in tension (ACI 408R-03). Farmington Hills, MI: American Concrete Institute; 2003.

[44] CAN/CSA-S6-06. Canadian highway bridge design code. Toronto, Ontario, Canada: Canadian Standards Association; 2006. [45] Schöck Canada Inc., Schöck ComBAR Technical Information. . [46] CAN/CSA-A23.3-04 (R2010). Design of Concrete Structures. Toronto, Ontario, Canada: Canadian Standard Association; 2010. p. 258. [47] S807-10. Specification for fiber-reinforced polymers. Canadian Standard Association; 2010. p. 44. [48] CAN/CSA S806-02. Design and construction of building components with fiber reinforced polymers. Toronto, ON, Canada: Canadian Standard Association; 2002. [49] Islam MS. Experimental investigation on the pullout strength of ribbed-surface glass fiber reinforced polymer bars embedded in high-strength concrete. Master of Engineering Project, Civil Engineering Department, Ryerson University, Toronto, ON, Canada; 2014. [50] Kanakubo T, Yonemaru K, Fukuyama H, Fujisawa M, Sonobe Y. Bond performance of concrete members reinforced with FRP bars. In: Nanni A, Dolon CW, editors. Proc. int. symp. fiber reinforced plastic reinforcement for concrete structures, ACI SP-138; 1993. [51] DeFreese JM, Roberts-Wollmann CL. Glass fiber reinforced polymer bars as top mat reinforcement for bridge decks (2002) Technical Report VTRC 03-CR6, Virginia Center for Transportation Innovation and Research. [52] Aly RS, Benmokrane B, Ebead UA. Tensile lap splicing of bundled CFRP reinforcing bars in concrete. J Compos Constr 2006;10:287–94. [53] Japan Society of Civil Engineers (JSCE). State-of-the-art report on continuous fiber reinforcing materials. Concrete Engineering Series 3, Tokyo, Japan; 1997.