Bond between CFRP rod panels and concrete using cementitious mortar

Construction and Building Materials 235 (2020) 117503

Contents lists available at ScienceDirect

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

Bond between CFRP rod panels and concrete using cementitious mortar Akram Jawdhari a,⇑, Amir Fam a, Issam Harik b a b

Department of Civil Engineering, Queen’s University, Kingston, ON K7L 3N6, Canada Dept. of Civil Engineering, University of Kentucky, Lexington, KY 40506, USA

h i g h l i g h t s
44 notched-beam bond tests to evaluate cementitious mortar in bonding CFRP rod panels (CRPs) to concrete.
Mortar resulted in ultimate load (Pult.) 86% that from epoxy, with a ductile rod slipping failure.
Pult. of CRP was 1.17 and 7 times that of flat CFRP plate, for epoxy and mortar, respectively.
A development length and bond strength of 125 mm and 460 kN/m, respectively, were found for mortar-bonded CRP.
Effects of width ratio, rod spacing-to-diameter ratio, and rod diameter, rod surface condition, were investigated.

a r t i c l e

i n f o

Article history: Received 19 July 2019 Received in revised form 1 November 2019 Accepted 5 November 2019

Keywords: Retrofit CFRP rod panels Concrete Bond Cementitious adhesive Mortar Epoxy

a b s t r a c t Carbon-FRP rod panels (CRPs), generated from small diameter rods mounted on a fiberglass mesh, are becoming a viable retrofit option. The gaps between rods enable full encapsulation by adhesive, thereby enhancing bond to existing concrete members, compared to flat plates. Existing studies focused on epoxy adhesive. In this study, 44 notched-beam bond tests, were carried out to investigate the effectiveness of cementitious mortar in bonding CRP to concrete and to examine the effects of a number of material and geometric parameters, comparing CRP to flat plates and mortar to epoxy. Results showed that the mortar was able to achieve a comparable ultimate load (Pult.), 86% that of epoxy, and a much more ductile failure by gradual rod slippage from the mortar. Compared to an equivalent CFRP plate, Pult. of CRP was 1.17 and 7 times, respectively, for epoxy and cementitious mortar. Brittle debonding failure dominated in CRP with epoxy and in CFRP plate with both epoxy and mortar. Pult was found to vary linearly with the bond length of CRP, up to a development length of 125 mm. A value of 460 kN/m can be assumed for bond strength. Rod axial stress (rf) increased by 42%, when CRP panel-to-concrete width (bf /bc) ratio increased from 0.25 to 0.5; decreased linearly by 13% when rod spacing-to-diameter (S/D) ratio increased from 3 to 8; decreased by 76% when rod diameter D increased from 2 to 4 mm. Sand coating the smooth rod resulted in a 45% increase in rf of the 4 mm rods but not the 2 mm rods, although failure shifted from gradual slippage to sudden debonding. Ó 2019 Elsevier Ltd. All rights reserved.

1. Introduction Many structures worldwide, especially those in the transportation sector, have become seriously deteriorated because of aging, environmental effects such as freeze/thaw cycles, human errors in design and/or construction, inadequate maintenance and misuse of facilities among other factors [1–4]. Due to their excellent attributes of high strength, lightweight, resistance to corrosion, minimal change in structure’s geometry, and ease of application, fiber reinforced polymer (FRP) composites have become a mainstream method in retrofit applications [5]. Conventionally, two methods ⇑ Corresponding author. E-mail address: [email protected] (A. Jawdhari). https://doi.org/10.1016/j.conbuildmat.2019.117503 0950-0618/Ó 2019 Elsevier Ltd. All rights reserved.

are used in FRP concrete retrofits; namely, externally bonded reinforcement (EBR) and near surface mounted (NSM) reinforcement [6,7]. These two methods differ in the way FRP is attached to the concrete substrate where the first is externally bonded while the second is inserted in grooves made in concrete cover [6,8]. Several researches have reported some advantages of NSM over EBR, including better bond resistance, protection, and suitability for strengthening negative moment regions of slabs and decks [7–10]. Another FRP retrofitting technique has been developed recently [2,6,11] and applied in several bridge retrofit applications [12]. It consists of small diameter Carbon-FRP (CFRP) rods, typically 1– 5 mm, aligned side by side at a uniform spacing, and is adhesively bonded on one face to a light fiber glass mesh, forming a panel which is commercially known as CRP [2] (Fig. 1(a)). Similar to

2

A. Jawdhari et al. / Construction and Building Materials 235 (2020) 117503

Fig. 1. CFRP rod panel, (a) picture, (b) schematics, (c) two panels connected by finger joint, (d) application on concrete.

EBR, CRPs are externally bonded to concrete (Figs. 1(d), 2(a)); however, the rods are fully embedded in adhesive, which also resembles NSM reinforcement. Several advantages have been reported for CRPs compared with EBR including: increased bond resistance due to full embedment and using small circular rods, hence increasing FRP area in contact with adhesive [13]; ease of application by using short-length panels (typically 1.22 m) connected by a finger joint, instead of continuous (full-length) plates [2]; possible multi-stage application by stopping work at any panel, leaving the finger joint area uncovered with adhesive, and resuming work at another time, which minimizes disruption of structural function, especially in highway bridge rehabilitation [12–14]. Two panels are widely produced and utilized in field applications, namely CRP 070 (with rod diameter (D) of 2 mm and rod spacing (S) of 6.35 mm) and CRP 195 (with D of 4 mm, and S of 9.5 mm). The three-digit number after CRP (e.g. CRP 070) refers to the panel’s tensile strength per unit width in US customary units. For example, CRP 070 can resist 70 kip/ft (or 1037 kN/m, in SI system) of tensile load. Several studies were performed on CRP, and confirmed its effectiveness. In Jawdhari et al. [13,14], bond properties of CRPs were determined through direct shear tests and notched beam flexure tests, respectively. Peiris and Harik [12] established the development length for CRPs with steel substrate. Jawdhari et al. [2,15] carried out flexural tests on fullscale RC beams strengthened with CRP 070 and CRP 195, respectively. Numerical analyses, field applications, and design and con-

struction issues are given in [11–12,16–17]. In all studies thus far on this system, epoxy adhesive has been used. The use of epoxy resin has several limitations including: environmental factors (emission of toxic fumes and steroids); moisture impermeability; inability to install in low temperatures and on wet surfaces; short shelf and pot lives; poor performance at elevated temperatures [3–5]. Cementitious mortars are good alternatives to mitigate some of these problems. Wetting and impregnation of fibers is oftentimes a concern for the bond between the cementitious mortar and FRP; this can be mitigated by adding polymers and silica fume [18–20]. One system that has received great attention is fiber reinforced cementitious mortar (FRCM), which consists of an open mesh of fibers or sheets embedded in the mortar [5,21]. Cementitious mortars are also utilized in NSM technique [4,22]. Multiple research has already confirmed the effectiveness of FRCM and mortar-bonded NSM techniques over epoxy-bonded FRPs in retrofitting RC members at elevated temperatures [5,23– 26]. Hashemi and Al-Mahaidi [3] tested six RC beams under fourpoint bending, examining the effectiveness of cementitious mortar in bonding CFRP fabric or textile to RC beams. The mortar was found to be very effective in bonding CFRP textile, were it resulted in an ultimate load of about 80% that achieved using epoxy adhesive. In another study, cementitious mortar was used to bond NSM CFRP strips to strengthen RC beams in torsion [4]. The mortar resulted in 74–83% ultimate capacity, compared to epoxy resin. FRCM technique was also used to strengthen two-way RC slabs

A. Jawdhari et al. / Construction and Building Materials 235 (2020) 117503

[26], where adding one and four FRCM layers, respectively, resulted in 141 and 135% increase in ultimate load for lowstrength concrete, and 205 and 212% for high-strength concrete. The objective of this study is to investigate for the first time the effectiveness of cementitious mortar in bonding CRP to concrete. A series of notched-beam bond tests were conducted, examining several factors, including comparing CRP to flat CFRP plates and mortar to epoxy. The study also examines the effect of bond length, bond width, rod diameter, rod spacing-to-diameter ratio, and rod surface treatment in terms of smooth or sand coated rods on bond strength and the level of utilization of the axial tensile strength of the rod.

2. Experimental program 2.1. Test method and matrix The bond tests, comprising 44 specimens, were carried out using the small beam test (Fig. 2). Multiple research studies [14,27–28] and engineering standards, such as ASTM [29] and ACI 440 [30], recommend the small beam test because of its simplicity, and ability to facilitate statistical validation and results’ comparison and interpretation [27]. The beam has a cross-section of 102x102 mm and a length of 356 mm, and features a 3 mmwide and 51 mm-deep (half-depth) mid-span notch, which was created in this study by a saw cut. The FRP reinforcement was bonded to the tension face of the beam, and symmetrically aligned on either side of the notch (Fig. 2). Table 1 shows the bond test matrix, designed to evaluate the effectiveness of cementitious mortar in bonding CRP to concrete substrate, and investigate several key parameters. These parameters were: type of FRP reinforcement (CRP, CFRP plate); type of adhesive (mortar, epoxy); CRP bond length (lb); CRP bond width

3

(bf); rod diameter (D); rod spacing-to-diameter (S/D) ratio; rod surface treatment (smooth, sand coated). In addition, control beams, without FRP reinforcement, were also tested as a baseline for FRP-bonded beams. A typical CRP 070 (with D and S of 2 and 6.35 mm, respectively, and S/D ratio of 3) was used in most CRP-bonded specimens. lb was varied from 50 to 150 mm to establish the development length (ld) and bond strength of CRP 070 with cementitious mortar. The effects of CRP bond width were examined by testing two values for bf, 25 and 50 mm, corresponding to CRP-concrete width ratio (bf/bc) of 0.25, and 0.5, respectively. Other investigated parameters are shown in Table 1 and discussed in details in the results section. For most specimens, three replicas were tested as recommended by Gartner et al [27]. Two pilot specimens were tested initially for specimen CRP-cement to evaluate the merit of cementitious mortar for CRP prior to the main testing campaign, resulting in five repetitions for that sample (Table 1). 2.2. Material properties 2.2.1. Concrete The notched-concrete blocks were fabricated from a ready-mix concrete with a maximum aggregate size of 19 mm and a target 28-day compressive strength (f’c) of 35 MPa [28]. An average f’c of 33 MPa was determined from several 150x300 mm cylinder tests, as reported in [28]. Table 2 lists the primary material properties, for the concrete block and other parts, as found from testing, reported by manufacturer, or determined from empirical equations. 2.2.2. CFRP rod panel (CRP) The CRPs used in this study were fabricated from 2 and 4 mmdiameter CFRP rods. The fabrication of CRPs is as follows: (a) cut-

Fig. 2. Small beam bond test; (a) longitudinal view, (b) cross-section [(1) with CRP, (2) with CFRP plate], (c) specimen under testing [dimensions in mm].

4

A. Jawdhari et al. / Construction and Building Materials 235 (2020) 117503

Table 1 Bond test matrix and primary results from the small beams strengthened with CRP and CFRP plate.

1 2 3

4 5 6 7 8

Pult.4 (kN)

Specimen identifier

No. of repetitions

FRP reinf.

Type of adhesive

Additional variables for CRP Variable

Average

Average

Standard deviation

Control1 CRP-cement CRP-epoxy Plate-cement Plate-epoxy CRP-lb 50 CRP-lb 100 CRP-lb 150 CRP-bf 25 CRP-D 4 CRP-S/D 5 CRP-S/D 8 CRP-SC2 CRP-D 4-SC2

3 5 3 3 3 3 3 3 3 3 3 3 3 3

– CRP CRP CFRP plate CFRP plate CRP CRP CRP CRP CRP CRP CRP CRP CRP

– Cement Epoxy Cement Epoxy Cement Cement Cement Cement Cement Cement Cement Cement Cement

– Footnote (3) Footnote (3) – – lb lb lb bf D S/D ratio S/D ratio RST D RST

– Footnote (3) Footnote (3) – – 50 mm 100 mm 150 mm 25 mm 4 mm 5 8 Sand coated 4 mm Sand coated

3.33 23.43 27.27 5.69 23.27 9.53 19.37 24.06 8.22 5.61 15.16 7.60 24.35 10.29

0.10 4.43 0.97 0.49 2.06 0.40 0.85 1.37 1.25 0.57 0.33 1.67 0.44 0.28

Increase

Failure mode

– 7.02 8.17 1.70 6.97 2.85 5.80 7.20 2.46 1.68 4.54 2.27 7.29 3.03

CCN5 RSC6 CDC7 ADC8 CDC7 RSC6 RSC6 RSC6 RSC6 RSC6 RSC6 RSC6 CDC7 CDC7

Concrete beam (no reinforcement). SC = sand coated. Bond length (lb) = 125 mm; bond width (bf) = 50 mm; rod diameter (D) = 2 mm; rod spacing-to-diameter (S/D) ratio = 3; rod surface treatment (RST) = smooth (no treatment). Ultimate (maximum) load. CCN = failure is concrete crack at notch. RSC = failure is rod slipping from cementitious adhesive. CDC = failure is cohesive debonding at concrete-adhesive within concrete. ADC = failure is adhesive debonding at plate-cement interface.

Table 2 Material properties. Property

Elastic modulus, E (GPa) Tensile strength, rt (MPa) Compressive strength, rc (MPa) Bond strength, s (MPa) qffiffiffiffiffi 0 1 =4700 f c , from ACI-318 code [31]. qffiffiffiffiffi 0 2 =0:625 f c , from ACI-318 code [31]. 3 =0:6  rt , according to [32]. 4 Concrete to steel substrate.

Part Concrete block

CFRP rods

CFRP plate

Mortar

Epoxy

271 3.62 33 –

134 2340 14043 –

160 2800 16803 –

21 3.5 40 10

4.5 24.8 59.3 17.94

ting the rods into the required length by a high-shear scissors; (b) placing the rods side-by-side at the intended spacing, using grooved end supports; (c) bonding the self-adhesive fiberglass mesh on the top face (Fig. 3(a)). The manufacturer reported a value of 2,340 MPa for the rod’s tensile strength and 134 GPa for its elastic modulus [33]. 2.2.3. CFRP plate The bond performance of CRP, for both cementitious mortar and epoxy adhesive, was compared to that of a conventional CFRP plate (Table 1). The plate has a thickness (tf) (Fig. 2(b-2)) of 1.4 mm, and a manufacturer specified tensile strength and modulus of elasticity of 2800 MPa and 160 GPa [34], respectively. The plate’s width (bf) is 25 mm, selected based on equating the axial stiffness ((Ef Af), where Ef and Af are the FRP elastic modulus and total area, respectively) of the CFRP plate and CRP.

obtained from the manufacturer datasheet [35], were: a 28-day compressive strength of 40 MPa, a bond strength exceeding 10 MPa, a 28-day splitting tensile strength of 3.5 MPa, and a modulus of elasticity of 21 GPa. 2.2.5. Epoxy adhesive The epoxy in specimens (CRP-epoxy and Plate-epoxy) was a two-part high-modulus structural resin, with a manufacturer reported tensile strength of 24.8 MPa and a tensile modulus of 4,482 MPa, [36]. This epoxy was previously used in several studies to bond CRP to concrete [2,12–15], where it provided an excellent adhesion with CFRP rods and concrete substrate, failing mostly by cohesive debonding in the concrete layer or concrete cover separation, but rarely by interfacial debonding. The same epoxy is also commonly used for bonding CFRP plates. 2.3. Surface preparation and FRP installation

2.2.4. Cementitious mortar A polymer-modified, one-component grout, was selected as the cementitious mortar adhesive used in this study as an alternative to epoxy. This mortar was commercially available in conventional 23 kg (50 lb) bags and required only potable water for preparation [35]. It has the same advantage of adhesion to the surface, like epoxy, which is an important requirement for retrofitting the underside of concrete members. The mechanical properties,

The tension face of the notched-beams was prepared by sandblasting to remove dust, laitance, and other bond inhibiting substances, and provide a rough bonding profile. Prior to applying the FRP reinforcement, the notch was carefully insulated in order to prevent seepage of mortar and epoxy and creating a stiff joint. Installation using cementitious mortar adhesive differed slightly from epoxy, where the former required dampening the concrete

A. Jawdhari et al. / Construction and Building Materials 235 (2020) 117503

5

Fig. 3. FRP installation; (a) prepared CRPs, (b) dampening concrete with water, (c) applying first coat of cementitious mortar, (d) placing CRP, (e) wet burlap curing, (f) standard room curing.

surface with water (Fig. 3(b)), while the latter demanded a dry surface. The application of CRP was as follows: (a) coating the concrete surface with an approximately 2 mm-thick layer of adhesive (mortar or epoxy) using a plastic profile (Fig. 3(c)); (b) placing CRP in position and pressing to force the adhesive to flow around the rods and fill the voids (Fig. 3(d)); (c) applying a second adhesive coat, to fully cover the rods. The nominal thickness of embedded CRP ((ta), Fig. 1(c)) was approximately 4 and 6 mm for rods with D of 2 and 4 mm, respectively, which included the rod diameter, and a nominal 1 mm cover above or below the rods. Manufacturer recommendations were followed in installing CFRP plate to the beam. An approximately 1.5 mm-thick coats of adhesive (mortar or epoxy) were applied first to the concrete and plate surfaces independently, using a triangular profile. The plate was then placed onto the surface and pressed on the concrete with a plastic roller, to remove excessive adhesive and trapped air pockets. A uniform pressure was then placed on the plates and kept for a minimum of 24 h.

Curing method and duration depended on the adhesive type. For the mortar, the beams were first wrapped with wet burlap for a minimum of 24 h (Fig. 3(e)), and then left to cure at standard room conditions (23 °C, 70% relative humidity) for 28 days, as recommended by the manufacturer [33]; whereas for epoxy, specimens were subjected to a standard room curing for 14 days, immediately after FRP application (Fig. 3(f)). 2.4. Test setup and instrumentation The specimens were tested under three-point bending with an effective span of 305 mm (Fig. 2), using a 50 kN capacity universal testing machine at a loading rate of 0.25 mm/min. A 25 mm steel cylinder, with a flat base, was placed above the specimen at midspan and used to distribute the load uniformly (Fig. 2(c)). The load was measured by a load cell integrated within the machine. A linear potentiometer (LP) (Fig. 2(c)), was placed at approximately 10 mm from the notch, and used to estimate the mid-span deflec-

6

A. Jawdhari et al. / Construction and Building Materials 235 (2020) 117503

tion. For each specimen, a 5 mm electric foil resistance strain gage was placed on the centerline of FRP reinforcement at mid-span, and was used to measure the maximum strain in CRP or CFRP plate. The gage was mounted onto the surface of CFRP plate or onto the exterior surface of CRP, on the adhesive. Direct instrumentation of the rods in CRP was impossible due to their small diameter. The strains in the adhesive surface were approximated to correspond to those in the rods for CRP with epoxy, assuming slippage of rods from the epoxy is negligible. However, this assumption cannot be made for CRPs with cementitious mortar adhesive due to the significant rod-mortar slipping that was observed in this study and discussed in the following section. 3. Results and discussions Table 1 gives a summary of the primary results for each specimen, including the ultimate load (Pult.), standard deviation, increase in Pult. relative to the control (un-strengthened) sample, and failure mode. It should be noted that due to the larger number of repetitions and testing in two batches (2 repetitions in pilot test and 3 repetitions in principal test), CRP-cement specimen showed more variation in Pult. than other specimens. However, the coefficient of variation (CV) for the CRP-cement specimen is 18.9% which is within acceptable upper bounds. 3.1. Failure modes 3.1.1. Cementitious mortar Specimens reinforced by CRP and bonded by cementitious mortar, showed a very ductile failure (RSC, in Table 1) by rod slipping from the mortar (Fig. 4(a)). The slipping was preceded by a crack in the mortar and another one in the beam, initiating from the notch’s tip and propagating upward to the compression face (Fig. 4(a)). At this stage, the adhesion between the rod and mortar is likely lost. After Pult. was reached, the cracks became fully developed and the beam was separated into two blocks held only by the rod panel (Fig. 4(a)). As the beam deflects due to the applied load, the panel continues to experience more slipping from the mortar while pivoting around top concrete surface. The slipping continued gradually because of the friction between the rods and mortar, but the test was terminated at large deflections before the panel is completely detached from the mortar (Fig. 4(a)). Specimen (Plate-cement), containing CFRP plate bonded by cementitious mortar, failed suddenly by an adhesive plate debonding (ADC, in Table 1) (Fig. 5(a)). At failure, the mortar on the failed side remained attached to the concrete block – not the CFRP plate – as seen in Fig. 5(a). 3.1.2. Epoxy A cohesive debonding failure (CDC, in Table 1), occurring in a 1– 3 mm thick concrete layer adjacent to the FRP-concrete interface, was observed for both CRP and CFRP plate bonded using epoxy (Figs. 4 (b) and 5(b)). The failure was sudden, and in most cases coupled with a triangular concrete crack in the vicinity of the notch, resulting in separation of a concrete wedge from the block.

mortar, respectively. Most importantly, the cementitious mortar was found to be a very effective bonding agent for CRP, where it resulted in a comparable (86%) Pult. to that of epoxy, and exhibited a much more ductile and gradual failure as discussed above. For CFRP plate, the cement adhesive gave a poor bond performance, where it resulted in a Pult. of 24% that of epoxy, likely because of the low chemical adhesion between the mortar and the flat, rigid plate [19]. The superior bond performance of cementitious mortar with CRP compared to CFRP plate, might be attributed to the difference in the geometric configuration of either reinforcement. The rods in CRP are completely encased in the mortar and thus would resist interfacial shear stresses by two components, friction and cohesion. On the other hand, friction is negligible in the externally bonded plates, as they typically detach from the concrete substrate due to difference in rigidity or effects of self-weight, once adhesion is lost. For that reason, most research and field applications don’t utilize the mortar with pultruded FRP plates, but with bi-directional meshes or textiles because they can be mechanically interlocked and embedded in the mortar, providing better bond resistance. Fig. 7 presents the load (P) versus mid-span deflection (D) curves for selected specimens, showing the effects of FRP type, adhesive type in Fig. 7(a), CRP bond length (lb) in Fig. 7(b), CRP bond width (bf) in Fig. 7(c), rod diameter (D) in Fig. 7(d), rod spacing-to-diameter (S/D) ratio in Fig. 7(e), and rod surface treatment (smooth vs. sand coated) in Fig. 7(f). All specimens containing mortar-bonded CRPs showed a very ductile P-D behavior characterized by significant post-peak softening. On the other hand, specimens utilizing epoxy-bonded CRP or plate, and mortar-bonded plate, failed suddenly, without a softening branch [Fig. 7(a)]. The sudden failure was also observed when the rod surface was sand coated in specimens containing mortar-bonded CRP, as can be seen in Fig. 7(f). Fig. 7(b) shows P-D curves for mortar-bonded CRPs, having three different bond lengths. For lb = 50 and 100 mm, the curve experienced only softening trend after peaking. However, a softening, hardening and then softening trend can be seen for lb = 150, possibly because this length is longer than the development length (ld) of 125 mm, estimated in this study. Further discussion into the definition and effects of ld will be given later. Table 3 lists several key results for CRP and CFRP plates, including the strain at ultimate (emax.); total tensile force (T); maximum tensile stress (rf); stress ratio (rf /rfu), where rfu is the guaranteed tensile strength of reinforcement, and average shear stress (savg.). rf refers to the stress at ultimate, obtained by dividing T by the total FRP area for CRP or CFRP plate. For CRP, the contribution of adhesive (epoxy or mortar) in neglected when calculating rf because the elastic modulus of epoxy is very small compared to modulus of the rods, while the mortar is already cracked at ultimate. The ultimate load (Pult.) was used to calculate the total force (T) in CRP or CFRP plate, using force equilibrium concept in a notched beam as shown in Fig. 8 and assuming linear stresses for concrete at failure, as recommended by Gartner et al. [27] and applied in several studies [14,28]. T was then divided by the FRP bond area (0.5 lb  bf) to determine savg., as follows:

Pult:  L h  b f  lb

3.2. Ultimate strength and effectiveness of cementitious mortar

sav g: ¼ 0:6

The increase in Pult. over control specimen ranged from 1.68 to 8.17 times, with the highest increases being 8.17 (CRP-epoxy); 7.29 (CRP-cement, rods sand coated); 7.20 (CRP-cement, lb = 150 mm); 7.02 (CRP-cement); and 6.97 (plate-epoxy). Fig. 6 plots Pult., in relation to reinforcement type (CRP vs. CFRP plate) and adhesive type (mortar vs. epoxy). As can be seen, CRP resulted in a higher Pult. than the plate; by 1.17 and 7 times, for epoxy and cementitious

The FRP stress ratio (rf/rfu) in CRP, 37% (for cementitious mortar) and 43% (for epoxy), was much higher than that in the CFRP plate, 2.67% (for cementitious mortar) and 10.65% (for epoxy), implying a better utilization of CFRP material in the rod panel. savg was 2.19–13.36 MPa for all specimens, varying based on reinforcement type, adhesive types, and other geometrical parameters for

ð1Þ

A. Jawdhari et al. / Construction and Building Materials 235 (2020) 117503

7

Fig. 4. Failure modes in beams bonded to CRP, (a) cementitious mortar, (b) epoxy.

CRP. Noticeably, the bond strength of cementitious mortar to concrete substrate was 89% that of epoxy with concrete. This observation was made by comparing savg of specimen (CRP-SC) to that of specimen (CRP-epoxy), both failing by cohesive debonding at concrete-adhesive interface (CDC, Table 1). 3.3. Effects of bond length The bond length (lb), for mortar-adhered CRP, was studied using four specimens with lb of 50, 100, 125, and 150 mm, for a typical CRP 070 (with D = 2 mm, S = 6.35 mm, S/D = 3). Other parameters were kept constant; bf = 50 mm and smooth rod surface, as shown in Table 1. Fig. 9 plots the maximum rod tensile stress (rf) versus lb, for the cementitious mortar tested in this study. Also shown for comparison, results for variable bond lengths for epoxy, tested previously by Jawdhari et al. [14] using similar beam bond tests and identical properties for the CRP-concrete joint. Fig. 9 shows that for both adhesive types, rf initially increased linearly with lb, but leveled off to a horizontal plateau after attaining a maximum stress. The length at which the two lines meet has been referred to in literature as the effective bond or development length (ld), and is defined as the length within which most of the interfacial force transfer occurs [14] (i.e. different from development length definition of rebar embedded in concrete). Providing a bond length longer than ld will not increase the bond strength, but ensures a more gradual debonding process.

In Fig. 9, ld was approximated as 80 and 125 mm, for epoxy- and mortar-bonded CRP, respectively. The panel’s bond strength per unit width (sult.) was determined by dividing the total rods force (T) by bf, and was found to be 343 and 460 kN/m for epoxy and cementitious mortar, respectively. Although the mortar required an ld 56% longer than that for epoxy (likely due to rod slipping), it resulted in 34% increase in sult.. It should be noted that the reported ld and sult. values for mortar-bonded CRP were established based on rod slippage failure. Different values might be obtained if other failure modes, such as cohesive debonding in concrete or adhesive debonding in mortar, dominate. In addition, due to the short span of notched-concrete beam, the maximum bond length (lb) that could be experimentally tested was 150 mm, slightly longer than the 125 mm estimated ld. Larger specimen dimensions and longer bond lengths, extending at least two times ld, are recommended for future experimental or numerical study to accurately determine ld for mortar-bonded CRP. 3.4. Effects of bond width The effect of CRP bond width (bf) was examined for mortarbonded panels by testing two values, bf = 25 and 50 mm, corresponding to a CRP-to-concrete width (bf/bc) ratio of 0.25 and 0.5, for a typical CRP 070 (with D = 2 mm, S = 6.35 mm, S/D = 3). Other parameters were kept unchanged; lb = 125 mm and smooth rod surface, as listed in Table 1. Fig. 10 shows the relation between

8

A. Jawdhari et al. / Construction and Building Materials 235 (2020) 117503

Fig. 5. Failure modes in beams bonded to CFRP plate, (a) cementitious mortar, (b) epoxy.

40 CRP/Cement CRP-Cement CRP-Epoxy CRP/Epoxy

Pult. ,(kN)

30

Plate-Cement Plate/Cement Plate-Epoxy Plate/Epoxy

20

10

0 Fig. 6. Ultimate load (Pult.) vs. FRP type (CRP, CFRP plate), and adhesive type (cementitious mortar, epoxy).

maximum rod tensile stress (rf) and bf/bc ratio, showing the effects of adhesive type by comparing specimens with cementitious mortar from current study and those with epoxy from Jawdhari et al. [21]. In that study, three values of (bf/bc) ratio, 0.17, 0.33, and 0.5, comprising two repetitions for each case, were tested. For epoxy, Fig. 10 shows that rf is generally not affected by bf, where it slightly increased then decreased when (bf/bc) ratio varied from 0.17 to 0.5. The insensitivity of rf to variation in (bf/bc) ratio might be related to the failure mode being CDC which is dependent on the shear strength of the concrete block [2]. On the other hand, rf increased by 42% when (bf/bc) ratio varied from 0.25 to 0.5 for the mortar. This interesting trend for the mortar might be related to two factors, crack formation and rod slipping. When increasing (bf/bc) ratio for the same beam size, FRP quantity is also increased, causing a possible delay in cracking in the mortar and in the beam, above the notch; slowing of rod slipping from the mortar and an increase in ultimate load. It should be mentioned that (bf/bc) ratio was limited in the current study to 0.5, anticipating failure to be cohesive debonding in concrete, which is not affected by (bf/bc) ratio as found in [14]. However, with the rod slipping failure and the observed rf trend in Fig. 10 for the mortar, it is recommended

9

A. Jawdhari et al. / Construction and Building Materials 235 (2020) 117503

30

30

10 8

Load, P (kN)

4

20

2 0 0

10 Two specimens each

0 0

Load, P (kN)

6

20

0.5

1

1.5

2

Control CRP-Cement CRP-Epoxy Plate-Cement Plate-Epoxy

5 10 Deflection, Δ (mm)

10

CRP-Cement

0 0

15

(a) FRP and adhesive types 30

15

30 2 mm 4 mm

Load, P (kN)

Two specimens for each bf

Load, P (kN)

5 10 Deflection, Δ (mm) (b) CRP-Cement Bond length (lb)

25 mm 50 mm

20

50 mm 100 mm 125 mm 150 mm

Two specimens for each lb

Control Plate-Cement

Two specimens for each D

20

10

10

CRP-Cement

CRP-Cement

0

0 0

5 10 Deflection, Δ (mm) (c) Bond width (bf)

30

15

Smooth Sand coated

Load, P (kN)

Load, P (kN)

5 10 Deflection, Δ (mm) (d) Rod diameter (D)

30

3 5 8

20

0

15

20

Two specimens for each S/D

Two specimens for each RST

10

10

- CRP-Cement - Rod diameter (D) = 2 mm

CRP-Cement

0

0 0

5 10 Deflection, Δ (mm)

15

(e) Rod spacing-to-diameter ratio (S/D)

0

5 10 Deflection, Δ (mm)

15

(f) Rod surface treatment (RST)

Fig. 7. Load vs. mid-span deflection (P-D) curves, showing effects of various parameters.

to examine (bf/bc) ratios from 0.5 to 1.0 to fully understand the effects of this parameter. 3.5. Effects of (S/D) ratio The rod spacing-to-diameter (S/D), for mortar-adhered CRP, was studied in this study by testing three values for (S), 6.25, 10, and 17 mm with a constant D of 2 mm, thus corresponding to (S/D) ratio of 3, 5, and 8, respectively. Other parameters were kept unchanged; lb = 125 mm, bf = 50 mm, and smooth rod surface (Table 1). In a previous study, Jawdhari et al. [14] also studied this parameter for an identical CRP, bonded to concrete by epoxy, using three (S/D) ratios, 3, 4.75, and 6. Fig. 11 plots the relation between

maximum rod stress (rf) and (S/D) ratio, showing the effects of adhesive type. For epoxy, results show that (rf) is increased by 76% when (S/D) is increased from 3 to 6. In contrast, a decrease of 13% in (rf) is observed for the mortar, as (S/D) is increased from 3 to 8. This difference in response between the two adhesives might be attributed to the effects of FRP stiffness and failure mode. In the case of epoxy, increasing (S/D) results in reducing the number of rods and thus FRP area (Af), for the same bond width (bf). Reducing Af has been reported to reduce the size and extent of concrete wedge failure that typically precedes debonding in epoxyadhered CRP [14], hence resulting in a possible increase in ultimate load. On the other hand, a decrease in Af, for the mortar, might

10

A. Jawdhari et al. / Construction and Building Materials 235 (2020) 117503

Table 3 Key test results from the small beams strengthened with CRP and CFRP plate. Specimen identifier

Strain at ultimate

# of rods N

Max. FRP stress ϭf 2 (MPa)

ϭf /ϭfu3

Avg. shear stress

(kN)

Force per rod T/N (kN)

(%)

(MPa)

21.02 24.47 5.10 20.87 8.55 17.38 21.59 7.38 5.04 13.60 6.82 21.85 9.24

2.63 3.06 – – 1.07 2.18 2.70 1.85 2.52 2.72 2.28 2.74 4.62

871 1013 82 330 354 720 894 611 204 901 753 905 375

37.22 43.32 2.67 10.65 15.14 30.77 38.22 26.12 8.73 38.53 32.20 38.68 16.01

6.73 7.83 3.27 13.36 6.84 6.96 5.92 4.72 4.24 4.36 2.19 6.99 7.78

Total force T

emax.1 CRP-Cement CRP-epoxy Plate-cement Plate-epoxy CRP-lb 50 CRP-lb 100 CRP-lb 150 CRP-bf 25 CRP-D 4 CRP-S/D 5 CRP-S/D 8 CRP-SC CRP-D 4-SC 1 2 3 4 5

0.0082 0.0006 0.0005 0.0020 0.0083 0.0151 0.0053 0.0178 –5 –5 0.0013 0.0014 0.0021

8 8 – – 8 8 8 4 2 5 3 8 2

savg.4

Taken at ultimate load (Pult.). rf is calculated by dividing (T) by the total FRP area for CRP or CFRP plate. rfu is the guaranteed tensile strength in FRP reinforcement, equals to 2,340 MPa and 3100 MPa for CFRP rods, and CFRP plate, respectively. Calculated from Eq. (1). Strain gage malfunction due to crack at mid-span.

Fig. 8. Schematics of notched-beam specimen, showing equilibrium of forces.

Fig. 9. Maximum rod stress (rf) vs. bond length (lb) for CRP bonded specimens, comparing cementitious mortar and epoxy.

Fig. 10. Maximum rod stress (rf) vs. CRP-to-concrete width (bf/bc) ratio, comparing cementitious mortar and epoxy.

3.6. Effects of rod diameter and surface promote cracking in the mortar and the beam at lower loads and expedite rod slipping, which might eventually result in reducing ultimate load.

In Table 1, two rod diameters (D) were investigated for cementbonded CRP, 2 and 4 mm, corresponding to two panels that are widely used in CRP retrofit applications, CRP 070, and CRP 195,

A. Jawdhari et al. / Construction and Building Materials 235 (2020) 117503

Fig. 11. Maximum rod stress (rf) vs. rod spacing-to diameter (S/D) ratio for CRP, comparing cementitious mortar and epoxy.

11

Fig. 13. Maximum rod stress (rf) vs. rod diameter (D) ratio for mortar-bonded CRP, comparing smooth and sand coated rods.

rod surface conditions were examined, smooth and sand coated (Table 1). Noticing the rod slipping failure in specimens containing smooth rods, sand coating was used to roughen the rods’ surface and enhance the bond between rod and mortar, which can result in higher failure load and better utilization of CFRP material. The coating was applied manually by wetting the surface of the smooth rod by a high-viscosity epoxy and embedding it in sand (Fig. 12(a)). Fig. 13 shows the variation in maximum rod stress (rf) in relation to (D) for the smooth and sand coated surfaces. It can be seen that rf is decreased by 76% and 58%, respectively, as D is doubled from 2 to 4 mm. A similar trend was observed in Jawdhari et al. [16] for epoxy-adhered CRP with smooth rods under pure shear loads, where rf was also found to decrease by 32% as D is varied from 2 to 4 mm. The difference in percentage decrease in rf, for the two studies might be attributed to the adhesive type and/or test method. Sand coating caused a shift in failure for mortarbonded CRP, from a ductile rod-mortar gradual slipping to a brittle debonding at the concrete-mortar interface (Fig. 13), similar to that seen in epoxy. In addition, it resulted in a negligible 3.9% increase in rf, for D = 2 mm, and a significant 83% increase in rf for D = 4 mm. A possible additional enhancement to the bond of CRP-concrete joint could have been obtained by roughening the concrete surface by grooving or using exposed aggregate surface, or shear keys. This topic is beyond the scope of this study.

4. Summary and conclusions

Fig. 12. Effects of sand coating treatment, (a) coated rods, (b) debonding in CRP 070 [D = 2 mm], (c) debonding in CRP 195 [D = 4 mm].

respectively. Dimensions of the tested CRP 070 and 195, respectively, were; lb = 125 mm (for both); bf = 50 and 19 mm; number of rods = 8 and 2; Af = 25 mm2 (for both). For each diameter, two

CFRP rod panels (CRPs), composed of small diameter rods, are becoming an efficient, practical and economical method in retrofitting deficient structures and those with limited access. One of the advantages of CRPs is their ease of application, achieved by using short-length panels connected by a finger joint mechanism. The panels are externally bonded, currently by synthetic adhesives such as epoxy. Due to environmental concerns and limitations in s applications, a better replacement option for epoxy is needed for the adhesive. In this study, 44 small beam bond tests were carried out to assess the effectiveness of cementitious mortar in bonding CRP to concrete in tension, and examine the effects of several key parameters. Overall, the study demonstrated an excellent performance of cementitious mortar. The following conclusions were drawn from the experimental program:

12

A. Jawdhari et al. / Construction and Building Materials 235 (2020) 117503

1- The ultimate load, Pult., for CRP bonded using cementitious mortar (CRP-Cement) is 86% of the ultimate load when the CRP is bonded using epoxy resin (CRP-Epoxy). The CRPCement specimens exhibited a more ductile failure by gradual rod slippage from the mortar, compared to the brittle and sudden cohesive debonding for the CRP-Epoxy specimens. 2- Pult. of CRP was equal to 1.17 and 7 times Pult. of conventional CFRP plate, for epoxy and cementitious mortar, respectively. While included in this study for comparison purpose, the mortar is not typically used to bond rigid plates, thereby resulting in low Pult. value. Failure in specimens reinforced by CRP and CFRP plate bonded using epoxy, was cohesive debonding. 3- The achieved tensile stress ratio in CRP rods, as a percentage of the guaranteed strength, was 37% and 43%, for the cementitious mortar and epoxy, respectively. For the CFRP plate, the ratio was 2.7% and 10.7%, for the same respective adhesives. This shows the advantage of CRP in attaining a better utilization of the CFRP material. 4- A development length (ld) of 125 mm, and bond strength (sult.) of 460 kN/m, were established for the mortar-bonded CRP. Initially, the maximum rod tensile stress (rf) varies linearly with bond length (lb), and levels off at ld. A value of 80 mm and 343 kN/m, respectively, were found for ld and sult. for epoxy-bonded CRP. 5- rf increased by 42% when CRP-to-concrete width (bf/bc) ratio for the cementitious mortar varied from 0.25 to 0.5. For epoxy, rf is not affected when (bf/bc) ratio is increased from 0.17 to 0.5. 6- As rod spacing-to-diameter (S/D) ratio increased from 3 to 6, rf decreased by 7.5% for mortar-bonded CRP, and increased by 76% for the epoxy-bonded CRP. 7- Sand coating the rod surface resulted in 83% increase of rf for D = 4 mm but only a negligible 3.9% for the 2 mm rod. It did however change failure mode from rod slippage to debonding. rf decreased by 76 an 58%, for the smooth and sand coated surfaces, respectively, when D is increased from 2 to 4 mm.

Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References [1] B. Täljsten, T. Blänksvard, Mineral-based bonding of carbon FRP to strengthen concrete structures, J. Compos. Constr. 11 (120) (2007) 120–128. [2] A. Jawdhari, A. Peiris, I. Harik, Experimental study on RC beams strengthened with CFRP rod panels, Eng Struct 173 (2018) 693–705, https://doi.org/10.1016/ j.engstruct.2018.06.105. [3] S. Hashemi, R. Al-Mahaidi, Experimental and finite element analysis of flexural behavior of FRP-strengthened RC beams using cement-based adhesives, Constr. Build. Mater 26 (1) (2012) 268–273, https://doi.org/10.1016/ j.conbuildmat.2016.08.095. [4] G. Al Bayati, R. Al-Mahaidi, R. Kalfat, Experimental investigation into the use of NSM FRP to increase the torsional resistance of RC beams using epoxy resins and cement-based adhesives, Constr. Build. Mater 124 (2016) (2016) 1153– 1164. [5] S.M. Raoof, D.A. Bournas, Bond between TRM versus FRP composites and concrete at high temperatures, Compos. Part B: Eng. 127 (2017) 150–165, https://doi.org/10.1016/j.compositesb.2017.05.064. [6] A. Al-Abdwais, R. Al-Mahaidi, Modified cement-based adhesive for nearsurface mounted CFRP strengthening system, Constr. Build. Mater. 124 (2016) 794–800, https://doi.org/10.1016/j.conbuildmat.2016.07.147.

[7] M. Badawi, K. Soudki, Flexural strengthening of RC beams with prestressed NSM CFRP rods – Experimental and analytical investigation, Constr. Build. Mater. 23 (10) (2009) 3292–3300. [8] R. Capozucca, On the strengthening of RC beams with near surface mounted GFRP rods, Compos. Struct. 117 (2014) 143–155. [9] H. Nordin, B. Taljsten, Concrete beams strengthened with prestressed near surface mounted CFRP, ASCE J. Compos. Construct. 10 (1) (2006) 60–80. [10] L. De Lorenzis, A. Nanni, Shear strengthening of reinforced concrete beams with near-surface mounted fiber-reinforced polymer rods, ACI Struct. J. 98 (1) (2001) 60–68. [11] A. Jawdhari, I. Harik, Finite element analysis of RC beams strengthened in flexure with CFRP rod panels, Constr. Build. Mater. 163 (2018) 751–766, https://doi.org/10.1016/j.conbuildmat.2017.12.139. [12] A. Peiris, I.E. Harik, Carbon fiber–reinforced polymer rod panels for strengthening concrete bridges, Adv. Struct. Eng. 21 (4) (2017) 557–570, https://doi.org/10.1177/1369433217732665. [13] A. Jawdhari, A. Peiris, I. Harik, Bond study on CFRP rod panels externally adhered to concrete, J. Compos. Constr. 21 (4) (2016), https://doi.org/10.1061/ (ASCE)CC.1943-5614.0000765. [14] A. Jawdhari, A. Semendary, A. Fam, I. Khoury, E. Steinberg, Bond characteristics of CFRP rod panels adhered to concrete under bending effects, J. Compos. Constr. 23 (1) (2018), https://doi.org/10.1061/(ASCE)CC.1943-5614.0000909. [15] A. Jawdhari, I. Harik, A. Fam, Behavior of reinforced concrete beams strengthened with CFRP rod panels CRP 195, Structures 16 (2018) (2018) 239–253, https://doi.org/10.1016/j.istruc.2018.09.014. [16] A. Jawdhari, A. Fam, I. Harik, Numerical study on the bond between CFRP rod panels (CRPs) and concrete, Constr. Build. Mater. 177 (2018) 522–534, https:// doi.org/10.1016/j.conbuildmat.2018.05.138. [17] A. Peiris, I.E. Harik, Design and construction of CFRP rod panel retrofit for impacted RC bridge girders, J. Compos. Sci. MDPI 2 (3) (2017), https://doi.org/ 10.3390/jcs2030040. [18] R.H. Haddad, N. Al-Mekhlafy, A.M. Ashteyat, Repair of heat-damaged reinforced concrete slabs using fibrous composite materials, Constr. Build. Mater. 25 (2011) 1213–1221. [19] A. Wiberg, Strengthening of Concrete Beams using Cementitious Carbon Fibre Composites Doctoral thesis, Royal Institute of Technology, Structural Engineering, Stockholm, Sweden, 2003. [20] S. Hashemi, R. Al-Mahaidi, Investigation of bond strength and flexural behaviour of FRP strengthened RC beams using cement-based adhesives, Aust. J. Struct. Eng. 11 (2) (2010) 129–139. [21] M.Y. Alabdulhady, L.H. Sneed, C. Carloni, Torsional behavior of RC beams strengthened with PBO-FRCM composite – An experimental study, Eng Struct. 136 (2017) 393–405, https://doi.org/10.1016/j.engstruct.2017.01.044. [22] A. Al-Abdwais, R. Al-Mahaidi, Bond behaviour between NSM CFRP laminate and concrete using modified cement-based adhesive, Constr. Build. Mater. 127 (2016) 284–292, https://doi.org/10.1016/j.conbuildmat.2016.09.142. [23] M. Saafi, Effect of fire on FRP reinforced concrete members, Compos. Struct. 58 (1) (2002) 11–20. [24] S.M. Raoof, D.A. Bournas, TRM versus FRP in flexural strengthening of RC beams: behaviour at high temperatures, Constr. Build. Mater. 154 (2017) 424– 437, https://doi.org/10.1016/j.conbuildmat.2017.07.195. [25] S. Hashemi, R. Al-Mahaidi, Flexural performance of CFRP textile-retrofitted RC beams using cement-based adhesives at high temperature, Constr. Build. Mater. 28 (2012) 791–797, https://doi.org/10.1016/j.conbuildmat.2011. 09.015. [26] G. Loreto, L. Leardini, D. Arboleda, A. Nanni, Performance of RC slab-type elements strengthened with fabric-reinforced cementitious-matrix composites, J. Compos. Constr. 18 (3) (2014), https://doi.org/10.1061/(ASCE) CC.1943-5614.0000415. [27] A. Gartner, E.P. Douglas, C.W. Dolan, H.R. Hamilton, Small beam bond test method for CFRP composites applied to concrete, J. Compos. Constr. 15 (1) (2011) 52–61, https://doi.org/10.1061/(ASCE)CC.1943-5614.0000151. [28] C. McSwiggan, K. Mak, A. Fam, Concrete bond durability of CFRP sheets with bio resins derived from renewable resources, J. Compos. Constr. 21 (2) (2017) 04016082, https://doi.org/10.1061/(ASCE) CC.1943-5614.0000736. [29] ASTM. 2014. New test method for bond strength for FRP bonded to concrete substrate using beam bond test. ASTM WK38604. West Conshohocken, PA: ASTM. [30] ACI (American Concrete Institute). 2017b. Guide test methods for fiber reinforced polymers (FRPs) for reinforcing or strengthening concrete structures. ACI 440.3R. Farmington Hills, MI: ACI. [31] ACI 318-14, ‘‘Building Code Requirements for Reinforced Concrete,” American Concrete Institute, ISO# 193382007E, 2014. [32] A. Jawdhari, A. Fam, Numerical study on mechanical and adhesive splices for ribbed GFRP plates used in concrete beams, Eng. Struct. 174 (2018) 478–494. [33] Diversified Structural Composites. 2016. ‘‘Specialty rods, tubes and shapes.” Accessed October, 2016. https://www.diversified-composites.com/. [34] Sika carboDur plates. Product data sheet. Sika services: Switzerland; 2011. https://usa.sika.com/dms/...get/.../pds-cpd-SikaCarboDur-us.pdf. [35] Sika MonoTop 623. 2019. ‘‘Product data sheet.” Sika Services, Switzerland. https://can.sika.com/dms/getdocument.get/91bef53e.../SikaMonoTop623_pds. pdf. [36] Sikadur 30. 2014. ‘‘Product data sheet.” Sika Services, Switzerland. Accessed September 23, 2014. https://www.sika-distributor.com/pdf/sikadur30/pdscpd-Sikadur30-us.pdf.