Experimental and finite element analysis of flexural behavior of FRP-strengthened RC beams using cement-based adhesives

Construction and Building Materials 26 (2012) 268–273

Contents lists available at ScienceDirect

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

Experimental and finite element analysis of flexural behavior of FRP-strengthened RC beams using cement-based adhesives S. Hashemi a, R. Al-Mahaidi b,⇑ a b

Department of Civil Engineering, Monash University, Melbourne 3800, Australia Faculty of Engineering and Industrial Science, Swinburne University of Technology, PO Box 218, Melbourne 3127, Australia

a r t i c l e

i n f o

Article history: Received 5 November 2010 Received in revised form 2 May 2011 Accepted 8 June 2011 Available online 16 July 2011 Keywords: FRP Rehabilitation Cement-based adhesives Fire resistance Bonding properties

a b s t r a c t The strengthening and rehabilitation of structures are major issues worldwide. In most situations, strengthening is required when there is an increase in the applied load, human error in the initial construction, a legal requirement to comply with updated versions of existing codes, or as a result of the loss of strength due to deterioration over time. Fiber-Reinforced Polymer (FRP) strengthening systems are enjoying a great deal of popularity as a result of the unique properties of FRPs, namely, their light weight, fatigue resistance non-corrosive characteristics and ease of application. The repair and strengthening technique with epoxy-bonded advanced composites has been applied to a large number of bridges around the world. At elevated temperatures, normally beyond the glass transition temperatures of epoxy adhesive, the mechanical properties of the polymer matrix deteriorate rapidly. It will be very beneficial if they can be replaced by cementitious (mineral)-based bonding agents such as modified concrete, in order to produce fire-resistant strengthening systems. Tests conducted for this paper include the investigation of the flexural behavior of FRP-strengthened reinforced concrete beams using cement-based adhesives. It is concluded that the use of cement-based bonding materials is a promising technique in FRP applications for structures located in hot regions or in danger of fire. Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction The excellent properties of FRP material have encouraged civil engineers over the last two decades to use it for the rehabilitation and strengthening of RC and steel structures. Unfortunately, the FRP application technique currently in use has some drawbacks. The emission of toxic fumes, the danger of skin irritation and eczema, moisture impermeability and flammability are critical issues in using epoxy for FRP strengthening of RC structures [1,2]. Furthermore, the epoxy, a common commercially-available bonding agent, has serious problems at elevated temperature. It loses its strength and stiffness and changes from a stiff material with high mechanical properties including high bond strength, to a viscous material with poor properties when exposed to elevated temperatures exceeding 60–70 °C [3,4]. Structures located in very hot climates or those in danger of fire can easily reach this threshold, especially on their surfaces. As a result, strengthened members using epoxy are vulnerable to high temperature exposure. The fire resistance and functionality of the structure during fire exposure are key roles that the structure should satisfy to maintain

⇑ Corresponding author. Tel.: +61 3 9214 8429. E-mail address: [email protected] (R. Al-Mahaidi). 0950-0618/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.conbuildmat.2011.06.021

its integrity and allow sufficient time for the evacuation of the building. Mineral-based materials are believed to be capable of providing the desired performance at high temperatures. In addition, FRP sheet and textile can be easily embedded in cement-based adhesive. The requirements for the matrix material are as follows:  Sufficient mechanical properties for load transfer.  Correct consistency, good penetration of the reinforcing fabrics, and good bond characteristics for embedded fabrics.  Thermal and chemical compatibility of the fibers and the substrate.  Good thermal and fire resistance.  Workability on site (applicability to large surfaces, long time period for application).  Environmental acceptability. Although some research has been conducted on the application of cementitious bonding material in FRP strengthening, more indepth investigation is needed. Polymer-modified cement-based adhesive has been used as a bonding agent in various studies [2,5–7], and a considerable 30–50% improvement in ultimate capacity has been achieved. However, polymer-modified mortar is believed to be vulnerable to heat exposure. Since the main aim in the application of cementitious adhesive is to maintain

269

S. Hashemi, R. Al-Mahaidi / Construction and Building Materials 26 (2012) 268–273 Table 1 Mix ratios of concrete and mortar.

a b c d e f g h i

Type

Ca (kg)

MCb (kg)

CAc (kg)

FAd (kg)

We (l)

Ff (kg)

SFg (kg)

SPh (kg)

Concrete MSi

306 674.3

– 168.6

1275 –

685 –

177 354

– 716.5

– 84.3

– 75.9

Ordinary Portland cement. Micro-cement. Coarse aggregate. Fine aggregate. Water. Filler (Silica 200G). Silica fume. Superplacticizer (Viscocrete5-500). Cementitious bonding agent.

Table 2 Mechanical properties of concrete and mortar. Type

Concrete substrate

Mortar

Compressive strength (MPa) Modulus of rupture (MPa)

38 3.85

65 5.65

structural integrity under fire conditions, further investigation is needed of the performance of strengthened members at high temperature using this type of adhesive. In another study, the application of FRP sheet impregnated with cementitious mortar was suggested, which appears to be impractical for large-scale projects [8]. The main objective of this paper is to develop a reliable CFRP strengthening technique using cementitious adhesive that can be later adopted in structures in danger of fire.

Fig. 1. Beam layout.

Fig. 2. Strain gauges and loading configuration.

Table 3 Details of specimens.

a

Beam designation

No. of specimens

CFRP material

nf  bf  tfa

Anchorage state

Bonding agent

Control ESF MSF MSR MTF

1 1 1 1 2

N/A Fabric Fabric Fabric Textile

N/A 2  80  0.176 2  80  0.176 2  80  0.176 Equivalent FRP area to that of fabric

N/A

N/A Epoxy Mortar Mortar Mortar

nf  bf  tf: number of layers  sheet width  thickness.

U

270

S. Hashemi, R. Al-Mahaidi / Construction and Building Materials 26 (2012) 268–273 2.1. Loading configuration

Fig. 3. FRP fabric anchorage.

Table 4 Experimental result. Beam designation

Failure mode

Ultimate load (kN)

Control

Yielding of steel followed by secondary compression failure Mid-span and end debond Mid-span and end debond Mid-span debond End and mid-span debond End and mid-span debond

121.2

ESF MSF MSR MTF1 MTF2

161.7 132.1 138.7 151.9 155.2

160

The test results in the form of ultimate load and mode of failure are shown in Table 4. The load–deflection curves are plotted in Fig. 4. A comparison of the control beam and the CFRP-sheet retrofitted beams is shown in Fig. 4a. The CFRP-textile retrofitted beams are compared with the control and epoxy-retrofitted beams in Fig. 4b.

MSR ESF

140

P (KN)

2.2. Results and discussion

2.2.1. Control beam The control beam had a typical failure mode of steel yielding followed by secondary concrete crushing. Several flexural cracks developed in the constant moment zone. The cracks propagated upwards and widened as the load progressively increased. The ultimate failure load was P = 121.2 kN.

(a) 180 120 100

MSF

80

Control

60

Controll beam

40

P/2

MSF MSR

Def.

0

10

20

30

40

Def (mm)

(b) 180 160

ESF

140

2.2.2. ESF beam For the purpose of comparison, one beam was retrofitted with two strips of CFRP fabric using normal epoxy adhesive. The failure was characterized by a combination of mid-span and end debond. The load-carrying capacity was P = 161.7 kN, which is 35% higher than the control beam’s capacity. The capacity increase and failure mode are comparable to the results reported in the literature [10]. The peak recorded strain in the CFRP fabric was 8629 microstrain.

ESF

P/2

20 0

Beam details and the four-point bending configuration are shown in Fig. 1. The beam was tested under a four-point bending test with a loading span of 900 mm, and a total span of 2300 mm. The reinforcement layout is illustrated in Fig. 1. The beam soffit was sand-blasted to achieve a high level of bonding between mortar and concrete. The CFRP material was embedded in a 20 mm overlay cast later. The CFRP was cut 150 mm short, reaching the support. Both types of CFRP material, including fabric and textile, were used in this retrofit. Strain gauges were installed on the CFRP to measure the strain in different locations. They were covered with a special coating to protect them against moisture. Fig. 2 shows the strain gauge layout and the loading configuration. The beam designations and variables are shown in Table 3. A new type of anchorage system was employed to examine its efficiency in MSR beams. The anchorage was achieved by a 30  20 mm groove installed in the concrete substrate along the beam width. Two FRP wires were used to hold down the CFRP sheet, as shown in Fig. 3.

MTF2

2.2.3. MSF beam In this beam, cementitious mortar adhesive was used to attach two layers of CFRP fabric strips to the soffit. As the load was progressively increased, a flexural shear crack developed near the point load. The crack propagated horizontally towards the supports at the mortar–fabric interface. As the load was further increased, most of the fabric was debonded on one side of the beam and the beam started to exhibit a response similar to that of the control beam. The load-carrying capacity was P = 132.1 kN, which is 10% higher than the control beam’s capacity. The peak recorded strain in the CFRP fabric was 2216 microstrain, which is about 25% of that recorded in the ESF beam.

P (KN)

120 100

MTF1

Control

80 60 40

P/2

20 0

Controll beam ESF MTF1 MTF2

P/2 Def.

0

10

20

30

40

2.2.4. MSR beam In this beam, cementitious mortar adhesive was used to attach two layers of CFRP fabric strips to the soffit. In addition, the fabric was anchored at both ends in a manner similar to that used in the single-lap shear tests. As the load was progressively increased, flexural cracks developed in the constant moment zone. The cracks propagated horizontally towards the supports at the mortar fabric interface. As the load was further increased, the horizontal interface cracks propagated towards the supports, but stopped short of reaching the anchorage points. The load carrying capacity was P = 138.7 kN, which is 15% higher than the control beam’s capacity.

Def (mm) Fig. 4. Load–deflection response (a) CFRP sheet and (b) CFRP textile. 2. Experimental study The experimental study included the investigation of the flexural performance of retrofitted reinforced concrete beams. The mortars used as the bonding agents include ordinary Portland cement and micro-cement with a mix proportion of 1:4 by weight. In addition, the mortar included silica fume to improve the mechanical properties of the mortar, filler to reduce cement dosage and superplasticizer to achieve the required workability. Tables 1 and 2 present the mixing ratios and mechanical properties of the concrete substrate and mortar respectively. The compressive strength was measured on 50 mm cubic samples for the mortar, but on ordinary cylinders for the concrete. The mixing proportions were determined in earlier stages of this project by conducting flexural tests on CFRP strengthened concrete beams with different types of mortar [9]. Since the best result was achieved by using the above mixing proportions, the project continued to use this type of mortar.

2.2.5. MTF beams CFRP textile and cementitious mortar were used to retrofit this beam. The peak capacity recorded was 151.9 kN and 155.2 kN in MTF1 and MTF2 respectively, which is nearly 27% higher than that achieved by the control beam. The failure mode was end and mid-span debonding. The vertical flexural cracks that developed in this beam were more distributed in the high moment zones but with smaller crack widths in comparison with the cracks of beams MSF and MSR. The peak recorded strain in the CFRP fabric was 11,469 and 10,019 microstrain in MTF1 and MTF2 respectively, which was higher than that recorded in the ESF beam. This may be due to the non-uniform distribution of strain in the tows across the width of the textile. 2.2.6. Strain distribution The strain variations along the length of the CFRP fabric and textile are shown in Fig. 5. The ESF beam achieved the maximum strain of 8630 microstrain under a load of 160 kN, which is comparable to that reported by Pham [10]. The beams retrofitted with CFRP fabric did not reach high strain values, which is consistent with their load capacity. The MTF1 and MTF2 beams achieved the considerable value of 7610 and

271

S. Hashemi, R. Al-Mahaidi / Construction and Building Materials 26 (2012) 268–273

(a) 10000

2500

40kN

20kN 40kN 60kN 80kN 100kN 120kN 130kN

60kN

8000

80kN

2000

100kN

7000

120kN

6000

140kN

Microstrain

Microstrain

(b)

20kN

9000

160kN

5000 4000

1500

1000

3000 2000

500

1000 0

0

200

400

600

800

0

1000

0

200

9000

40kN

8000

800

20kN 40kN

7000

60kN

6000

80kN

6000

100kN

5000

140kN

120kN

4000 3000

80kN 100kN

5000

120kN 140kN

4000 3000 2000

2000

1000

1000 0

600

60kN

7000

0

250

500

750

1000

0

0

Distance form midspan (mm)

200

400

600

800

Distance from midspan (mm)

(e) 8000

20kN

7000

40kN

6000

80kN

60kN 100kN

Microstrain

Microstrain

(d)

20kN

8000

Microstrain

(c)

400

Distance form midspan (mm)

Distance from midspan (mm)

5000

120kN 140kN

4000 3000 2000 1000 0

0

200

400

600

800

Distance from midspan (mm) Fig. 5. Strain distribution of samples (a) ESF, (b) MSF, (c) MSR, (d) MTF1 and (e) MTF2.

Fig. 6. Shear retention factor.

Fig. 7. Concrete and mortar constitutive model.

272

S. Hashemi, R. Al-Mahaidi / Construction and Building Materials 26 (2012) 268–273

160

Table 5 Material properties defined in FEA.

a b

140

fc0 (MPa)

ft (MPa)

fy (MPa)

fu (MPa)

E (GPa)

Concrete Mortar N12a R10b FRP

38 45 – – –

2.65 3.76 – – –

– – 558 409 –

– – 642 481 –

24 37 202 205 200

120

Load (kN)

Material

100

FEA

80 60

Experiment

40

Longitudinal reinforcement (12 mm). Shear reinforcement (10 mm).

Experiment result

20 0

FEA result

0

10

20

30

40

50

Def. (mm)

4

Bond stress (MPa)

Fig. 10. Loading curves comparison.

3 10000

2

20kN 40kN

8000

1

60kN

0

0.1

0.2

0.3

0.4

Slip (mm) Fig. 8. Bond–slip model for CFRP textile.

Microstrain

80kN

0

6000

100kN 120kN 140kN

4000

140kN-Exp.

6066 microstrain respectively at 140 kN load. This value is comparable to 4818 microstrain in the ESF beam, which confirms the non-uniform distribution of strain in the tows across the width of the textile.

2000

0

3. FE analysis

0

200

400

600

800

1000

Distance from midspan (mm)

3.1. FE model Fig. 11. CFRP strain distribution.

In order to shed further light on the performance of the CFRPtextile strengthened beams using cementitious adhesive, nonlinear finite element analysis was carried out using the software ATENA [11]. Since the best result in the experiments was achieved by the CFRP textile strengthening scheme, the FE model was developed for this composite member only. The fracture model used in ATENA is based on the smeared crack formulation and crack band model. It employs the Rankine failure criterion and exponential softening in tension and variable shear retention factor, as shown in Fig. 6, with a rotated crack model for both the reinforced concrete and the mortar. The concrete and mortar were modeled with 2D plane stress elements with a fracture–plastic constitutive model (Fig. 7), and the reinforcement was modeled based on discrete reinforcement for the longitudinal bars and CFRP material and smeared reinforcement for the stirrups. The measured material properties were used in the finite element model, except for the FRP where the supplier’s values were adopted. Table 5 presents the material properties used for different elements in the model. The steel reinforcement was assumed to

Fig. 12. Failure mode of MTF-2700.

have perfect connection with the concrete; however, the bond–slip model was defined for the FRP. The bond–slip curve was developed by the authors in the early stages of the project [12]. A single-lap

Fig. 9. FE mesh for MTF beam.

S. Hashemi, R. Al-Mahaidi / Construction and Building Materials 26 (2012) 268–273

273

Fig. 13. Crack pattern of FE model.

shear test was applied to develop the bond–slip model, as shown in Fig. 8. The standard Newton–Raphson solver was used. The prescribed deformation was imposed to the model and incremental stepby-step analysis was adopted. Since the beam is symmetric, half of the beam was modeled with appropriate boundary conditions along the line of symmetry. The model was developed for MTF beams. The FE mesh is shown for the aforementioned beams in Fig. 9. 3.2. FE results The experimental and FE load–deflection responses are plotted in Fig. 10. It is evident that the initial stiffness and ultimate strength are predicted with reasonable accuracy. Furthermore, the predicted trends of strain distribution in the CFRP, as shown in Fig. 11, are relatively comparable to the values obtained in the experiments. The peak value obtained from the FE model is about 8707 microstrain, compared to the experimentally-obtained value of 7610 microstrain. In addition, the experimental crack distribution shown in Fig. 12 is comparable to the crack pattern obtained from the FE simulation (Fig. 13). 4. Conclusion  This investigation focused on the development of efficient cement-based bonding agents for the CFRP strengthening of concrete. It can be concluded that considerable composite action can be achieved by using cement-based mortar as an adhesive.  Compared to CFRP fabric, CFRP textile is more compatible as well as more efficient with cement-based mortar. The ultimate load achieved by using CFRP textile-cement mortar is around 80% of what was achieved by using CFRP fabric with epoxy adhesive.

 This strengthening technique is expected to preform well at high temperature. This investigation is still in progress.  The FE analysis showed a good consistency with experimental results, and it can be applied to other problems.

References [1] Kolsch H. Carbon fibre cement matrix (CFCM) overlay system for masonry strengthening. J Compos Constr 1998;2(2):105–9. [2] Täljsten B, Blänksvard T. Mineral-based bonding of carbon FRP to strengthen concrete structures. J Compos Constr 2007;11(120):120–8. [3] Gamage JCPH, Al-Mahaidi R, Wong MB. Bond characteristics of CFRP plated concrete members under elevated temperatures. J Compos Struct 2006; 75(1–4):199–205. [4] fib Bulletin 14. Externally bonded FRP reinforcement for RC structures; 2001. [5] Wiberg A. Strengthening of concrete beams using cementitious carbon fibre composites. Stockholm: KTH Royal Institute of Technology; 2003. [6] Bournas D, Lontou P, Trianfillou T, Papanicolau, C. Textile-reinforced mortar (TRM) versus FRP jacketing for reinforced concrete columns, FRPRCS-8, Patras, Greece; 2007. [7] Bousias S, Spathis A, Fardis M, Traintafilou T, Papanicolau C. Pseudodynamic tests of non-seismically designed RC structures retrofitted with textilereinforced mortar, FRPRCS-8, Patras, Greece; 2007. [8] Wu HC, Sun P. Fibre reinforced cement based composite sheets for structural retrofit. In: Proceedings of the international symposium on bond behaviour of FRP in structures (BBFS 2005); 2005. [9] Hashemi S, Al-Mahaidi R. Cement based bonding material for FRP strengthening of concrete structures. In: Proceedings of the international symposium on fibre reinforced polymer reinforcement for concrete structures (FRPRCS-9), Sydney, Australia; 2009. [10] Pham HB, Al-Mahaidi R. Experimental investigation into flexural retrofitting of reinforced concrete bridge beams using FRP composites. J Compos Struct 2004;66(1–4):617–25. [11] Cervenka V, Jendele L, Cervenka J. Atena Program Documentation. Part 1. Theory. Cervenka Consulting; 2007. [12] Hashemi S, Al-Mahaidi R. Investigation of bond strength and flexural behaviour of FRP strengthened RC beams using cement-based adhesives. In: Proceedings of the structures congress, Orlando (FL); 2010.