Fourth Asia-Pacific Conference on FRP in Structures (APFIS 2013) 11-13 December 2013, Melbourne, Australia

© 2013 International Institute for FRP in Construction

MODIFIED CEMENT-BASED ADHESIVE FOR NEAR-SURFACE MOUNTED CFRP STRENGTHENING SYSTEM

A. Al-Abdwais1, R. Al-Mahaidi2, K. Abdouka3

1, 2, 3Civil Engineering Department, Swinburne University of Technology, Australia 1alabdwais@swin.edu.au, 2ralmahaidi@swin.edu.au, 3kabdouka@swin.edu.au

ABSTRACT The near-surface mounted (NSM) strengthening system has emerged as a promising technology to increase the strength of concrete for flexure, shear and torsion. Strengthening by this technique has been applied to a large number of concrete structures worldwide using epoxy adhesives. Due to the rapid deterioration of the mechanical properties of epoxy-based polymer matrix with elevated temperature, and the hazards from toxic fumes, replacing this polymer with a new cement-based bonding agent is essential to enhance the performance of structures in high-temperature environments and reduce the environmental and health hazards. The application of cement-based adhesive has been studied for externally-bonded techniques using CFRP textile. Although, the good bond properties has been achieved, debonding was the critical failure between fiber and concrete substrate. To exploit the advantages of NSM technique, a new modification of cement-based adhesive has been achieved to meet the requirements of NSM application system. To assess the efficiency of this adhesive, an experimental investigation of pull-out testing using a single-lap shear test set-up was conducted to study the bond characteristics between CFRP textile and concret. Four different mixes were used in this investigation. The test results show the efficiency of the NSM technique using modified cement adhesive and its superior bond properties compared to externally-bonded CFRP. KEYWORDS Modified cement-based adhesive, NSM, Strengthening, FRP INTRODUCTION In the civil engineering construction field, the need for the rehabilitation and strengthening of reinforced concrete and steel structures has become a critical issue around the world. This is due to the ageing of structures, overload exposure due to increased service loads, design faults, updating of existing codes, and deterioration due to exposure to aggressive environmental conditions. These factors make structures more vulnerable to sudden catastrophic failure. Civil engineering researchers have recognised that FRP composites with their excellent mechanical properties are vital for the strengthening and rehabilitation of infrastructure projects. Although epoxy adhesive has taken the major role as a bonding agent in this technique as an effective bonding agent for strengthening of structures for both external and NSM techniques, it has some disadvantages. These include hazardous toxic fumes, skin irritation, moisture impermeability and flammability, and the rapid deterioration of strength with elevated temperature higher than the glass transition Tg (60-70) °C (Gamage, and Wong, 2006), (fib Bulletin 14, 2001).which is a critical issue in high-temperature climates. Therefore, replacing the epoxy with cement-based adhesive as bonding agent is essential to improve the performance at high temperature, to lower the cost of material, reduce the toxic hazards to workers, and minimize the environmental impact of toxic fume emission with curing (Taljsten and Blanksvard, 2007), (Kolsch, 1998). Attempts to use cement-based adhesive for strengthening applications have been carried out only for externally-bonded FPR. Polymer-modified mortar has been applied as a cement-based adhesive, achieving 30-50% improvements in ultimate strength compared to unstrengthened samples (Badanoiu, and Holmgren, 2003), (Wiberg, 2003), (Bousias, and Fardis, 2007), (Taljsten and Blanksvard, 2007). Further studies have attempted to improve the density and properties of mortar by adding silica fume (Taljsten and Blanksvard, 2007), (Wu and Sun, 2005). Excellent bond properties have been achieved in the latest reasearch into cement based-adhesive for externally-

high viscosity to provide stability of the cement paste on the surface without dropping or fl

slipperiness to allow ease of penetration of the fibre through the adhesive in the groove by slid

for application. good bond properties to improve the load-carrying capacity.

ing trials with different mixes, it has been found that mixing the cement-based adhesive develchanging the superplasticizer ratio can

ificantly improve the viscosity of the adhesive to be consistent with implementation requirements. Adding

he experimental program involved testing sixteen specimens to study the modification of cement nd ith different mix designs and the workable ages of cement paste (pot-life).

able 1 presents four different mix designs with different ratios of superplasticizer and primer. The required

bonded CFRP textile by (Hashimi and Al-Mahaidi, 2012), who reported 4 MPa of bond stress and an increase 25-30% in ultimate capacity compared to unstrengthened specimens. However, debonding of fibre from the concrete surface was the critical failure in the external bond, which reduced the resulting ultimate bond strength. NSM is a promising new technology for the strengthenimg of concrete structures. The advantages of the NSM technique include better bond properties, no delamination of fibre at the ends being expected to occur, and less preparation of the concrete surface. Despite the excellent bond properties using epoxy resin, no detailed investigations of the NSM technique using cement-based adhesive have been reported. This is due to the low viscosity of cement paste, which causes it to flow away from the groove surface, making it difficult to apply in practice for NSM techniques. The main objective of this investigation is to develop a modified cement-based adhesive which works efficiently for NSM practical applications at normal and elevated temperature environments. DEVELOPMENT OF NEW CEMENT-BASED ADHESIVE FOR NSM APPLICATION In order to efficiently utilise the improvement in bonding properties between FRP reinforcement and concrete substrate using the NSM technique compared to the externally-bonded FRP strengthening technique, a modified cement-based adhesive has been developed for NSM strengthening systems as an alternative to epoxy adhesive. In the case of NSM with thin slots, it is difficult to pour the adhesive and in existing structures the strengthening is mainly applied to the bottom or side surfaces, Therefore, special physical properties of the adhesive required for NSM application including:

• owing away.

• ing the adhesive around fibre surface.

• suitable open-ended time period•

Follow oped by (Hashimi and Al-Mahaidi, 2008) with MBRACE primer and signprimer improves the physical properties with high viscosity to avoid dropping or flowing away and allowing the paste to slide around the fibre to enable ease penetration of fiber through the adhesive. With this development, the adhesive can be applied for velrtical surfaces and work more efficiently for NSM strengthening in practical applications. EXPERIMENTAL PROGRAM T paste aidentify the best bond properties wTlevel of viscosity depends on the primer and superplasticizer ratios. FRP textile strips (three weaves) were embedded in slits cut in the longitudinal direction of the concrete prism 4 mm wide and 18 mm deep. The bond length was fixed at 50 mm for all specimens. The bonding started 50 mm away from edge of the prism to avoid concentration of stresses and premature edge failure. Figure 1 illustrates the specimen details and groove dimensions. The fibres were bonded in the concrete slit using modified cement-based adhesive. Specimen details are shown in Tables 2. The mechanical properties of matrials are presented in Table 3 .The compressive strength represents the average values of three specimens of 100 mm diameter of concrete cylinders and 50 mm diameter of Mortar cylinder tested according to AS 1012.9. The tensile strength of concrete was calculated according to AS1012.11. Splitting tensile test was conducted to find the tensile strength of mortar according AS1012.10. The CFRP textile was supplied by (Fortress). The cross-section dimensions are 1.5 mm thickness and 2.33 mm wide. The tensile test with strain gages used to find the mechanical properties the fibre.

Table 1 Different mix ratios of mortars (grams) Mix type Ca MCb Wc Fd SFe SPf Pg

Mortar Mix-M1 674.3 168.6 354 716.5 84.3 42.1 227.5 Mortar Mix-M2 674.3 168.6 354 716.5 84.3 33.7 151.2 Mortar Mix-M3 674.3 25.3 101.2 Mo 4 6 3 7 8

: Mi 1, 2,a nd cbc

(Viscocret-5-500) (Mbrace primer part A and B)

168.6 354 716.5 84.3 rtar Mix-M 74. 168.6 354 16.5 84.3 16.9 8.6

M1, M2, M3, M4 x Type 3, 4 : Ordinary Portla ement : Micro-cement : Water

d: Filler (Silica 200G) e: Silica fume f: Superplasticizerg: Primer

a) Specimen details

(b) Section A-A of specimen

Figure 1 Specimen details and cross-section

Table 2 Specimen details Specimen disgnation

No of specimens (mm) (3 weaves) (mm2)

ime to fix the fiber after mixing (minutes)

Bond length Fiber area T

TC50-M1 4 50 10.5 5, 10, 20, 30 TC50-M2 4 5,10, 20, 30 50 10.5 TC50-M3 TC50-M4

4 4

50 50

10.5 5, 10, 20, 30 10.5 5, 10, 20, 30

P tex t-50 m ngth pe

SPEC EPARATION AND TEST SET-UP To ev ciency of the m ed cement-based adhes bond fibre and concr ate, pull-out

sts usi e-lap shear test set p were used. The specim f concrete prisms we ned to fit the slit was cut on the surface of each one in a longitudinal

RP textile consisting of weaves of carbon fiber

(a ) Slit in concrete prism (b) FRP textile

c) Mortar in the groove (d) Mortar on the Fiber

Figure 2 Specimen preparation

TC50: FR tile-cemen m bond le , M: mix ty

Tab anica aterialsssive st a) ngth (MPa) ticityMPa

le 3 Mech l properties of m Material Compra rength (MP Tensile stre Modulus of ElasConcrete 41 3.84 32,000

Mortar M4 25,000 86 6.2

FRP Textile --- 1450 135,000

IMEN PR

aluate the effing a singl

odifi-u

ive to ens o

ete substrre desigte

set-up. After 28 days of curing of the concrete prisms a sing a saw, as shown in (Figure 2.a). CFdirection u

longitudinally and kelvar in the transverse direction (Figure 2.b) was applied by placing a thin layer of modified cement adhesive on the surface, ensuring the whole surface of the fibre was covered by adhesive. The groove was filled with the adhesive using a steel blade. The fibre was inserted into the groove to the required depth, then the surface was levelled. The fibre was extended out of each specimen by 100 mm to fix the aluminum plates for gripping. After 21 days of curing, the aluminum plates were glued using Araldite epoxy adhesive, which has high bonding strength to avoid slipping of grips during the test. Figure 3 illustrates the single-lap shear test set-up.

Figure 3 Single-lap shear test set-up

EXPERIMENTAL RESULTS AND DISCUSSION In all specimens, the bond length was f different mix ratios on bond stress.

e test results showed a significant difference of the modified cement-based adhesive as bonding agent ows the test results.

30 min

fixed at 50 mm to identify the effect oThbetween CFRP textile and concrete substrate. Table 4 sh

Table 4 Test results (Ultimate load) (kN)

Specimen designation (a)

5 min (b)

10 min (c)

20 min(d)

TC50-M1 9.72 9.26 8.96 7.85 TC50-M2 7.4

1 1 10.54 8.85 7.81

TC50-M3 TC50-M4

1.51 11.2 0.58 9.62 11.63 11.5 11.27 9.67

In the TC50 f specimens w x M1, four ens were d. It can be se at there is about a 20% drop r 30 minutes e time. Th imum u load achieved by these specimens was 9.72 k shows the force-displacement relationships. lure mode w e debond ng at

terfacial zone between the adhesive and the fibre associated with cracks and peeling of adhesive at the surface,

Figure 4 Force-displacement relationship of series TC50-M1

-M1 series o ith Mi specim teste en thin bond afte of pot-lif e max ltimate N. Figure 4 The fai as th i

inas plotted in Figure 5.

0

2

4

6

8

10

12

0 2 4 6 8 10

Load

 (kN)

Displacement (mm)

TC50‐M1‐aTC50‐M1‐bTC50‐M1‐cTC50‐M1‐d

Figure 5 Failure mode of specimens TC50-M1

In the TC50-M2 series, fo ed by the first specimen after 5 minutes was 10.54 kN. The drop in strength within 30 minuts of time was more than 20%. The failure mode in all specimens was similar to that of series 1. Figures 6 present the force-displacement curves for these four specimens.

Figure 6 Force-displacement relationship of series TC50-M2

In the TC50-M3 series, four specimens were tested. The maximum force achieved by the five specimens was 11.52 kN. It can be seen that the decrease in strength was about 14%, which is less than the previous two series, as illustrated in Figure 7. The mode of failure was the interfacial cracking between adhesive and fibre showed that longitudinal cracks propagated along the adhesive surface as presented in figure 8.

50-M3

ur specimens were tested. The maximum ultimate load achiev

2

4

6

8

10

12

Load

 (kN)

TC50‐M2‐aTC50‐M2‐bTC50‐M2‐cTC50‐M2‐d

00 2 4 6 8 10

Displacement (mm)

Figure 7 Force-displacement relationship of series TC

2

4

6

8

10

12

Load

 (kN)

TC50‐M3‐aTC50‐M3‐bTC50‐M3‐cTC50‐M3‐d

00 2 4 6 8 10

Displacement (mm)

Figure 8 Failure mode of specimens TC50-M3

In the TC50-M4 series, fo The decrease in strength was only 3% within 20 m ad value of the first three specimens in 20 minutes was 11.46 kN. Figure 9 shows he load-displacement curve for all four specimens. The failure mode was similar to that of specimen TC50-M3.

hip of series TC50-M4

The average bond stress with 50 mm bond length is about 10 MPa. The value was simply calculated by dividing the average load of the first three specimens ( ie excluding specimen with 30 minutes pot life) of Mix M4 by fiber surface area bonded with adhesive. To evaluate the test results, Figure 10 illustrates the load versus time (pot life). It can be seen that M4 has the best performance within 20 minutes and provides better bond properties compared to the other mixes.

10 Load-time (pot life) relationship for different mixes

ur specimens were tested. The maximum force was 11.63 kN.inutes and of 14% between 20 o 30 minutes. The average lo t

t

Figure 9 Force-displacement relations

2

4

6

8

10

12

14

Load

 (kN)

TC50‐M4‐a

TC50‐M4‐b

TC50‐M4‐c

TC50‐M4‐d

00 2 4 6 8 10 12

Displacement (mm)

Figure

5

6

5 10 15 20 25 30Time (min)

7

8

9

10

11

12

Load

 (kN)

M1M2M3M4

ONCLUSION The aim of this investigation was to develop an efficient cement-based adhesive for the NSM CFRP strengthening technique. It can be concluded that excellent bond properties can be achieved using modified cement-based adhesive and it works efficiently as a bonding agent with the following findings:

• The best results were achieved by M4 with inconsiderable difference of strength within 20 minutes of

time. It is therefore recommended to use M4 with 20 minutes of pot-life as a cement bonding agent for strengthening with NSM CFRP textile.

• Strengthening with NSM using modified cement adhesive is about 2.5 times more efficient than the externally-bonded CFRP. The average bond stress was about 10 MPa compared to the externally-bonded CFRP textile reported by (Hashimi and Al-Mahaidi, 2012) which achieved only 4 MPa.

• The failure mode was the interfacial zone betwen fibre and adhesive associated with longitudinal cracks in the adhesive surface for all specimens.

• The results showed a co

Bou as, D., Lontou, P., Trianfillou, T. and Papanicolau, C. (2007). “Textile-reinforced mortar versus fiber

f CFRP plated concrete -205.

ahaidi, R. (2008) “Cement Based Bonding Material for FRP”, Proceedings of the eleventh

t. Inorganic-Based Fiber Composites Conference, 5-7 November 2008, Madrid - Span

RP-strengthened RC beams using cement-based adhesives”, Construction and Building Materials,

al of Composites for Construction, 2: pp. 105-109.

concrete

iber Composites”, tockholm, Sweden, KTH Royal Institute of Technology, PhD thesis?

retrofit”, Bond 05): pp. 351-356.

C

nsiderable ductile behaviour in all specimens.

REFERENCES Badanoiu, A. and Holmgren, J. (2003), "Cementations composites reinforced with continuous carbon fibers for strengthening of concrete structures", Cement and Concrete Composites 25(3): 387-394.

rn reinforced polymerconfinement in reinforced concrete columns”, ACI Structural Journal, 104(6): pp.740-748.

s, S., Spathis, A.-L and Fardis, M.N. (2007 Bousia ). “Seismic retrofitting of columns with lap spliced smooth

bars through FRP or concrete jackets”,Journal of Earthquake Engineering,. 11(5): pp. 653- 674. fib Bulletin 14, (2001). “Externally bonded FRP reinforcement for RC structures”.

e, J.C.P.H., Al-Mahaidi, R. and Wong, M.B. (2006)Gamag . “Bond characteristics omembers under elevated temperatures”, Composite Structures, 75(1–4): pp. 199

Hashemi, S., Al-MIn Hashemi, S. and Al-Mahaidi R. (2012). “Experimental and finite element analysis of flexural behavior of F26(1):pp. 268-273.

8). “Carbon fibre cement matrix (CFCM) system for masonry strengthening”,JournKolsch, H., (199

Taljsten, B. and Blanksvard, T. (2007). “Mineral-based bonding of carbon FRP to strengthen

tures”,Journal of Composites for Construction, 11(2): pp. 120-128. struc Wiberg, A. (2003). “Strengthening of Concrete Beams Using Cementitious Carbon FS Wu, H.C. and Sun,(2005). “Fiber reinforced cement based composite sheets for structural

ehaviour of FRP in Structures: Proceedings of the International Symposium (BBFS 20B