Construction and Building Materials 47 (2013) 397408

Contents lists available at SciVerse ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Review

Flexural behavior of self-compacting concrete reinforced with different types of steel bers
M. Paja k a,, T. Ponikiewski b
a b

The Silesian University of Technology, Faculty of Civil Engineering, Department of Structural Engineering, Akademicka 5, 44-100 Gliwice, Poland The Silesian University of Technology, Faculty of Civil Engineering, Department of Building Materials and Processes Engineering , Akademicka 5, 44-100 Gliwice, Poland

h i g h l i g h t s
 The load-deection curve depends on properties of bers in SFR-SCC as for SFRC.  The exural tensile strength of SFR-SCC can be predicted as for SFRC.  The relationship between the post-peak parameters of SFR-SCC was analyzed.  The equation to predict deection-CMOD relationship for SFR-SCC was proposed.  The GF of SFR-SCC was investigated and the formula to describe it was presented.

a r t i c l e

i n f o

a b s t r a c t
The aim of the present work is to investigate the exural behavior of self-compacting concrete reinforced with straight and hooked end steel bers at levels of 0.5%, 1.0% and 1.5% and compare it to normally vibrated concrete (NVC). The laboratory tests were determined according to RILEM TC 162-TDF recommendation. The exural behavior of SCC appeared to be comparable to NCV, where the increase of bers volume ratio cause the increase in prepeak and post-peak parameters of SCC. Nevertheless, the type of steel bers inuences much this dependency. Based on the performed analysis, the exural tensile strength of SFR-SCC can be described with the formulas used for SFRC. The equation proposed to predict the deection-CMOD relationship of SCC ts into the coefcient of variation of the formula for NVC. However, the SCC achieves the maximum crack mouth displacement for lower deections than NVC. The article contains also the analysis of the fracture energy of SCC with a formula to describe it. 2013 Elsevier Ltd. All rights reserved.

Article history: Received 17 October 2012 Received in revised form 24 March 2013 Accepted 4 May 2013

Keywords: Self-compacting concrete Steel bers Flexural tensile strength Fracture energy Deection-CMOD relationship

Contents 1. 2. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Experimental program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Materials and mix design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Test results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Test results of fresh and hardened SFR-SCC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Flexural tests results of SFR-SCC and general discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Maximum flexural tensile strength on notched beams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Equivalent and residual tensile strengths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5. Fracture energy GF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6. Deflection-CMOD relationship . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 398 398 398 398 399 399 399 402 402 405 406 407 407

3.

4.

Corresponding author. Tel./fax: +48 32 237 22 88.

E-mail address: malgorzata.pajak@polsl.pl (M. Paja k). 0950-0618/$ - see front matter 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.conbuildmat.2013.05.072

398

M. Paja k, T. Ponikiewski / Construction and Building Materials 47 (2013) 397408

1. Introduction Self-compacting concrete (SCC) was dened by Okamura [1] as concrete that is able to ow in the interior of the formwork, lling it in a natural manner and passing through the reinforcing bars and other obstacles, owing and consolidating under the action of its own weight. These properties enable the SCC to be an excellent material for constructions with complicated shapes and congested reinforcement. One of the main advantages in using SCC is the minimization of skilled labor needed for placing and nishing the concrete. All these benets decrease the costs and reduce the time of the building process over constructions made from traditionally vibrated concrete [2]. However, hardened self-compacting concrete is still as brittle as normal concrete and has a poor resistance to crack growth. To improve the post-peak parameters of SCC the steel bers are added. As generally known, steel bers continue to carry stresses after matrix failure. The addition of steel bers does not change considerably the compressive strength and the modulus of elasticity of concrete but has noteworthy effects on the residual tensile strength and exural strength . Therefore, in case of SFRC analyzing the post-peak behavior of tested specimens is crucial [37]. The main recommendations to analyze post-peak bending behavior of steel ber reinforced concrete are: JSCE method [8], RILEM TC 162-TDF recommendation [9] and based on it EN 14651 [10], and (withdrawn in 2006) ASTM C1018 [11]. The detailed information about the post-peak behavior of SFRC can be found in [4]. The mechanical properties and post-peak behavior of FRC depend mainly on the characteristics of the concrete matrix but also on the type, geometry, content and orientation of bers [1214]. However, effective work in concrete matrix is only one factor that has the inuence on the efciency of the currently produced bers. Efciency is also based on the simplicity of its production which in turn has a signicant inuence on its price. The ve most popular types of steel bers are: traditional straight, hooked, crimped, with deformed ends (coned, with end paddles or end buttons) and with deformed wire (indented, etched or with roughened surface) [15]. A statistical analysis of the assortment offered by the largest ber producers allows to claim that around 67% of sold ber consist of the hooked type. Other most popular ber types are: straight ber (around 9%), ber with deformed wire (around 9%) and crimped ber (around 8%) [16]. Aspect ratio of all steel ber types offered by main producers are from 20.4 to 152. Fifty percent of the population of all types offered by producers has aspect ratio from 45 to 63.5 [15].

The aim of this article is to investigate the tensile behavior of SCC reinforced with two different types of steel bers and compare it to traditionally vibrated concrete. It is widely recognized that the determination of tensile strength in the test in direct tensile is troublesome. Thus, the authors choose the bending tests on notched beams according to [9,10], which is the best candidate to be a standard test method for analyzing the FRC [17]. Among the most popular types of bers authors choose two types with the comparable aspect ratio but different shape: hooked end and straight. The articles usually deal with the bers with aspect ratio higher than 40. Authors choose steel bers with smaller aspect ratio, because that kind of reinforcement is not widely investigated.
2. Experimental program 2.1. Materials and mix design The studies were performed on the self-compacting concrete reinforced with two types of steel bers (SFR-SCC). The experimental investigation of SFR-SCC, conducted by authors, allows to properly adjust the ingredients of the mix in order to obtain the required workability and stability. The composition of the mix, where the proportions of the ingredients of the matrix are maintained and only the ber content changes, is presented in Table 1. The concretes were prepared using water/cement ratios of all tested mixes equal 0.42. To obtain comparable rheological and mechanical parameters of SFR-SCC, the cement containing 25% of silica y ash was employed. The use of mixed cement eliminates the effect of different composition of Supplementary Cementing Materials (SCM) which causes variation of physical and chemical parameters of SFR-SCC. The authors applied high content of the cement (490 kg/m3), which is widely investigated by other researchers [18,19]. The physical and chemical properties of cement CEM II/B-V 32.5R, are presented in Table 2. The aggregate used in the mix was: natural sand and coarse aggregate, with a maximum size of respectively: 2 mm and 8 mm. The Glenium SKY 592 Superplasicizer and Stabilizer RheoMATRIX were applied in the mix in the amount of respectively: 3.5%, 0.4% of the mass of cement. The authors used mix proportioning system proposed by Okamura and Ozawa [20] which assumes general supply from ready-mixed concrete plants. The coarse and ne aggregate contents are xed so that self-compatibility is achieved easily by adjusting only the water-powder ratio and superplasticizer dosage [21]. The detailed mixing procedure and mixing time is illustrated in Fig. 1. The investigations were carried out on straight and two times bigger hooked steel bers. The geometry and parameters of steel bers are presented in Table 3. Three different volume fractions of steel bers were developed to cover the majority of practically used fractions range: 0.5%, 1.0%, 1.5%, which is the dosage of respectively: 40, 80, 120 kg/m3.

2.2. Tests The bending tests were carried out on 550 mm beams with 150 150 mm cross section. The span of the beams was 500 mm. The bottom side of each specimen was notched with 25 mm depth using 2 mm diamond saw.

Table 1 Composition of SFR-SCC mix. Cement CEM II/B-V 32,5R (kg/m3) C 490 Sand (02 mm) (kg/m3) S 808 Fine aggregate (28 mm) (kg/m3) FA 808 Water (l/ m3) W 201 Steel bers (kg/m3) (% by volume) F 40; 80; 120 (0.5; 1.0; 1.5) Superplasicizer (kg/m3) SP 17 Stabilizer (kg/ m3) ST 2 W/ C 0.42

Table 2 Chemical and physical properties of cement CEM II/B-V 32, 5R. Chemical properties Cement composition (%) SiO2 26.4 CaO 47.9 Al2O3 10.8 Fe2O3 3.7 MgO 1.6 Na2O 0.43 SO3 2.2 K2O 1.42 Cl 0.007 370 3300 3.15 Cement specic surface (m2/kg) Physical properties Fineness (cm2/g) Specic gravity Setting (h: min) Initial set 04:15 Final set 06:00 Compressive strength (MPa) after days 2 13 2 7 28 2 28 38 3

M. Paja k, T. Ponikiewski / Construction and Building Materials 47 (2013) 397408

399

Fig. 1. Mixing procedure for SFR-SCC. Table 3 Properties of steel bres. S HE

Fiber shape Length (mm) Wire diameter (mm) Aspect ratio Tensile strength (MPa) Number of bers per kg

Straight 12.5 1.25 0.40 0.04 31.25 >1250 15%

Hooked end 30 1 0.8 + 0.04/0.03 37.5 1100

The specimen preparation for all specimens was the same. The specimens were kept in moulds under polyethylene sheet for one day. After demoulding all specimens were cured in 20 C water till the time of the test. The experimental program is presented in Table 4, where the designation of mixes used further in the paper is included. Three specimens were tested for each type and volume ratio of steel bers. One of the beams with 1.0% of hooked end steel bers underwent catastrophic failure so it is not presented in the paper. For comparison the plain self-compacting concrete beams were also tested. The three-point bending tests were conducted on servo-controlled electrohydraulic machine (Fig. 2) in accordance to RILEM TC 162-TDF [9] and EN 14651 [10]. The mid-span deection of the specimen increased at a constant rate of 0.2 mm/min. The tests were performed until the beam reached the 5 mm deection. The deection of the two opposite sides of the specimen was measured in the center of the specimen by LVDT sensor placed on the steel frame. The steel frame was supported at the top and on both sides of the specimen with screws. The steel frame was used to help to avoid deformations in overall deection coming from the settlement or twisting of the specimen on the supports, which was widely investigated in [22]. The CMOD (crack mouth opening displacement) was measured, simultaneously with the beam deection, with the clip gauge attached at the knife-edge. The 5 mm knife-edge thickness was chosen according to [23] as an acceptable error level for the crack mouth opening displacement of the specimen of 10%. The compressive tests were performed in 3000 kN hydraulic compression testing machine with a constant strain rate. The compressive strength of plain SCC was established on 6 cubes (150 150 150 mm).

3. Test results and discussion 3.1. Test results of fresh and hardened SFR-SCC In order to characterize the ow and workability properties of SFR-SCC, the traditional slum-ow test was performed according to RILEM TC 145-WSM [24]. The slump-ow diameter (SFD), and time to reach 500 mm spread (T50) of the fresh concrete are presented in Table 5. The Table 5 contains also the test results obtained on hardened SFR-SCC. The compressive strength of SCC was 73.4 MPa. The addition of randomly distributed short steel bers increases the compressive strengths of SCC. 3.2. Flexural tests results of SFR-SCC and general discussion The presented loaddisplacement and load-CMOD curves report the exural response obtained on plain beams and beams reinforced with 0.5%, 1.0% and 1.5% of volume fraction of straight (Fig. 3) and hooked end (Fig. 4) steel bers. Each loaddeection

81,139

7801

Table 4 Test program. Mix 0 0.5_S 1_S 1.5_S 0.5_HE 1_HE 1.5_HE Fiber type KrampeHarex DG 12.5/0.4 KrampeHarex DG 12.5/0.4 KrampeHarex DG 12.5/0.4 Steelbet 30/0.8 Steelbet 30/0.8 Steelbet 30/0.8 Fiber volume fraction (%) 0.5 1.0 1.5 0.5 1.0 1.5 Fiber content (kg/m3) 40 80 120 40 80 120 Number of beams 3 3 3 3 3 2 3

Fig. 2. Testing setup.

400 Table 5 Properties of fresh and hardened SFR-SCC mix. Mix Slump ow test T50 (s) 0 0.5_S 1_S 1.5_S 0.5_HE 1_HEa 1.5_HE
a

M. Paja k, T. Ponikiewski / Construction and Building Materials 47 (2013) 397408

Density (kg/m3)

fcm,28d (MPa)

SFD (mm) 690 680 660 640 680 670 640 2305 2325 2363 2385 2334 2355 2387 73.40 80.10 84.60 87.50 98.20 96.50 88.60

1.5 3 5 6 2 3.5 4.5

Average of two specimens.

curve presents the average load of two LVDT cells at mid-span deection. Fig. 5 shows the loaddeection curves derived for different types but of the same volume ratio of steel bers. To better illustrate the initial parts of the curves, Fig. 5 contains curves up to 5 mm and in the picture inside to 1 mm. The behavior of plain SCC specimens (described with 0 in Figs. 3 and 4) is typical for brittle material with almost linear prepeak parts of the loaddeection curves. In the post-peak parts of the curves the rapid load decrease with increase of deection can be observed.

As expected, in case of beams reinforced with randomly distributed short steel bers the type of steel bers signicantly inuences the exural behavior of tested beams. The maximum load increases with an increase of ber content for both types of steel bers. The specimens reinforced with hooked end steel bers indicate higher maximum load than the straight ones, and this dependency increases as the bers volume content increases (Fig. 5). Applying straight steel bers as beam reinforcement affected in obtaining deection that corresponds to the maximum load comparable to the one for plain SCC (Table 6). The prepeak parts of these loaddeection curves show almost no nonlinearity with well marked rst peak, no matter of the volume fraction of steel bers. On the contrary, the specimens reinforced with hooked end steel bers reached a few times bigger deection at maximum load than the plain SCC specimens. In this case, the ascending parts of the loaddeection curves are characterized by apparent nonlinearity. The biggest deection at the maximum load, equal 0.92 mm, was noted for specimens with 1.0% of bers dosage while for 1.5% of volume ratio the deection at the maximum load is just 0.64 mm. The post-peak part of the loaddeection curve differs much in case of those two bers. The hooked end steel bers reinforcement causes the gradual decrease in load after achieving the peak load while the straight steel bers indicate sudden load drop after the peak load.

Fig. 3. Results from three-point bending tests performed on SCC reinforced with different volume ratio of straight steel bers: (a) loaddeection curves; and (b) load-CMOD curves.

Fig. 4. Results from three-point bending tests performed on SCC reinforced with different volume ratio of hooked end steel bers: (a) loaddeection curves; and (b) loadCMOD curves.

M. Paja k, T. Ponikiewski / Construction and Building Materials 47 (2013) 397408

401

Fig. 5. Comparison of loaddeection curves obtained from SCC reinforced with: (a) 0.5%; (b) 1.0%; and (c) 1.5% of volume fraction of straight and hooked end steel bers.

Table 6 Experimental results of peak and post-peak parameters SFR-SCC. Mix Flexural tensile strength (MPa) 2.45(23.4) 3.66(6.5) 4.27(8.0) 5.42(10.6) 3.8(6.9) 6.14(1.5) 8.31(0.9) Maximum load deection (mm) 0.05 0.05 0.06 0.11 0.15 0.92 0.64 feq,2 (MPa) feq,3 (MPa) fR,1d (MPa) fR,2d (MPa) fR,3d (MPa) fR,4d (MPa) fR,1-CMOD (MPa) fR,2-CMOD (MPa) fR,3-CMOD (MPa) fR,4-CMOD (MPa)

0 0.5_S 1_S 1.5_S 0.5_HE 1_HEa 1.5_HE

a

1.72(14.2) 2.99(3.8) 4.62(13.5) 3.38(12.7) 5.61(3.2) 8.21(1.3)

1.25(18.8) 2.25(3.0) 2.85(5.6) 2.77(12.4) 5.21(0.8) 6.91(3.1)

1.99(14.1) 3.21(2.8) 4.75(8.1) 3.60(10.4) 5.59(2.8) 8.16(0.3)

1.29(21.1) 2.31(3.5) 3.04(12.6) 3.04(11.1) 5.77(0.8) 7.32(1.4)

0.89(19.0) 1.70(9.6) 1.49(22.1) 2.21(30.1) 4.22(6.7) 5.73(9.0)

0.66(20.3) 1.24(15.3) 1.19(11.3) 1.69(30.2) 3.42(5.3) 4.38(20.0)

2.09(14.3) 3.31(2.8) 4.85(7.3) 3.61(11.4) 5.81(1.0) 8.09(0.8)

1.40(21.2) 2.48(3.9) 3.64(0.9) 3.13(10.2) 5.65(3.1) 7.58(1.8)

1.00(20.4) 1.93(4.8) 2.70(2.1) 2.60(14.7) 4.46(0.9) 6.32(3.2)

0.76(18.1) 1.45(14.6) 2.04(1.4) 1.97(25.2) 3.55(6.2) 5.09(9.7)

Average of two specimens.

The exural behavior can be conformed to the Nammans classication of FRC composites based on their tensile response [25]. The deection-hardening response was observed in case of hooked end steel bers and deection-softening exural response for straight steel bers. As generally known, steel bers are most effective after cracking of brittle matrix. According to Naaman [4] the

observed exural behavior can be mainly attributed to the length of the bers and their bond strength. The short straight steel bers bridge the microcracks only during strain localization, so this does not inuence much the post-peak part of the loaddeection curve. On the other hand, long hooked end steel bers provide bridging stresses across the crack, which are the result of coalesce

402

M. Paja k, T. Ponikiewski / Construction and Building Materials 47 (2013) 397408

of microcracks. This bridging stress may arrest further propagation of the macrocracks. As the result, toughness of the material can increase. Additionally, the mechanical bond between the ber and the matrix is improved by the deformed shape of hooked end steel bers. Comparing the test results obtained by authors for reinforced self-compacting concrete to traditionally vibrated concrete it can be concluded that SFR-SCC behaves similarly. As in traditionally vibrated concrete the deection-hardening response was observed for SCC reinforced with hooked end steel bers. However, test results presented in the literature for bers with aspect ratio form 50 to 80 indicate that the SCC reinforced with hooked end steel bers can be deection-hardening [26] or deection-softening [27,28]. 3.3. Maximum exural tensile strength on notched beams The exural tensile strength was obtained on notched beams after other researchers [2933]. In the present work it was calculated using the following equation:

strength is almost linear. Naaman and Reinhardt [25] suggested the equation to calculate the exural tensile strength of FRC:

rcc rmu 1 V f asV f

l d

where l and d are length and diameter of steel bers, respectively;

rmu is the exural tensile strength of unreinforced matrix; Vf is the volume fraction of steel bers; s is the average bond strength at the bermatrix interface; a is a product of several coefcients. When Eq. (2) is rearranged and a and s are replaced with one
coefcient B, Eq. (3) can be achieved:

rcc A rmu B A rmu

 d l Vf l d

This form of Eq. (2) can be simply compared to the equation suggested by Akcay [28] obtained from the experiment on notched beams [28]:

rcc k1 rmu k2 V f

l d

rcc

3Pmax L 2bh a0 2

where Pmax is a maximum load; L, b and h are the span, width and depth of the beam; a0 is a notch depth. The values of exural tensile strength and corresponding ber volume fraction are presented graphically in Fig. 6 and summarized in Table 6. Generally, for both types of steel bers the beams reinforced with higher ber content indicated the higher exural tensile strength. As expected, the shape of steel bers strongly inuences the rcc of SCC. The hooked end steel bers are more effective in improving the exural tensile strength of tested specimens than the straight ones. However, this dependency starts to be pronounced for the specimens reinforced with bers dosage higher than 0.5% by volume. In this case the increase of exural tensile strength was about 1.5 times in comparison to plain SCC. Applying the maximum dosage of hooked end steel bers (1.5% by volume) resulted in obtaining the exural tensile strength about 3.4 times higher than in case of plain SCC while the addition of 1.5% of straight steel bers increased the exural tensile strength of plain SCC about 2.2 times. The presented test results are in a basic agreement with generally observed trend in improving exural tensile strength of concrete by adding steel bers. As the increase of the exural tensile

The authors used Eq. (4) to describe the exural tensile strength of SFR-SCC determined on notched beams (Fig. 6). The value of k1 obtained by other researchers mentioned in [29] was equal to 0.9. Eq. (2) assures that the exural tensile strength of unreinforced beams is the exural tensile strength of matrix, so the coefcient k1 should be equal to 1.0. On the basis of the presented test results k2 was equal 0.1 and 0.06 for hooked end and straight steel bers, respectively. These values are close to the ones presented in [29]. However, Akcay related k2 to the strength of bers. He obtained k2 = 0.11 and k2 = 0.08 for SCC reinforced with high and normal strength hooked end steel bers, respectively. As in the Eq. (2) the parameter k2 should be divided, on the basis of further experiments, into parts dependent on the ber strength, type, bond strength and other parameters. The exural tensile strength obtained on notched beams depends also on the notch depth. Thus, the coefcients k1 and k2 should depend on the ratio of notch depth to the high of the specimen. In article [29] and present work the ratio of notch depth to the high of specimen was 0.4 and 0.17, respectively. As was mentioned above, this Eq. (2) is generally used for SFRC. Kang [30] analyzed ultra high performance concrete (UHPC) reinforced with the straight, short steel bers (l = 13 mm, d = 0.2 mm). Based on the laboratory tests on notched beams with the bers volume ratio from 0% to 5%, author determine Eq. (3) where the coefcient A was equal 1 and B was about 0.307. The value of k2, calculated according to (4), was equal 0.17. The comparison of the test results with the test results obtained on notched beams presented by other researchers [29,30,32,33] is presented in the Fig. 7. The values of exural tensile strength were normalized by dividing it by the exural tensile strength of plain concrete. The basic information about the steel bers (aspect ratio, type) are also presented in the Fig. 7. It can be concluded that the values of exural tensile strength obtained for SFR-SCC do not diverge from the one obtained for conventionally vibrated concrete. Generally, this relationship is strongly affected by the factors connected with ber type, dimensions, strength and bond strength between the ber and the matrix. Thus, Eq. (2) proposed by Naaman and Reinhardt [25] can be also applied to SFR-SCC. 3.4. Equivalent and residual tensile strengths The design parameters of steel ber reinforced concrete can be calculated from the three-point bending test performed on notched beams according to RILEM TC 162-TDF recommendation [9] and EN 14651 [10]. The recommendations allow to obtain the

Fig. 6. Flexural tensile strength of SFR-SCC with respect to the ber volume fraction.

M. Paja k, T. Ponikiewski / Construction and Building Materials 47 (2013) 397408

403

Fig. 7. Comparison of normalized exural tensile strength with respect to the ber volume fraction.

parameters characterizing the post-peak behavior of FRC: exural tensile strength at the limit of proportionality (fct,L), equivalent tensile strength (feq,2,feq,3) (Fig. 8) and residual tensile strength (fR,i) (Fig. 9). Equivalent tensile strengths are evaluated for two deection limits: d2 = dL + 0.65 mm and d3 = dL + 2.65 mm, where dLdeection corresponding to the highest value of load in the interval of 0.05 mm (FL). The four values of residual exural tensile strength (fR,i) can be obtained from a loaddeection curve at mid-span deection (dR,i) or from the load-CMOD curve at a crack mouth opening displacement (CMODi). The equations dening this parameters are given in Fig. 9, where: L and b are span and width of the beam; hsp-the distance from the tip of a notch and top of cross section of the beam. Barr and Lee [34] carried out statistical analysis to determine the applicability and robustness of the RILEM TC 162-TDF recommendation. Five independent testing laboratories conducted research using RILEM TC 162-TDF procedure to obtain design

Fig. 10. Equivalent exural tensile strength with respect to the ber volume ratio: (a) feq,2; and (b) feq,3.

parameters. As a result, this recommendation was found to be very satisfactory in determining the post-peak parameters of SFRC with very low coefcients of variation.

Fig. 8. Denition by RILEM TC 162-TDF [10] of equivalent tensile strengths: (a) feq,2; and (b) feq,3.

Fig. 9. Denition by RILEM TC 162-TDF [9] of residual tensile strengths fR,i: (a) residual from loaddeection curve; and (b) from load-CMOD curve.

404

M. Paja k, T. Ponikiewski / Construction and Building Materials 47 (2013) 397408

Fig. 11. Residual exural tensile strength with respect to the ber volume ratio: (a) fR,1d; (b) fR,2d; (c) fR,3d; and (d) fR,4d.

The mean values and standard deviation of the post-peak parameters evaluated according to RILEM TC 162-TDF [9] and EN 14651 [10] from the loaddeection and load-CMOD curves are presented in Table 6. The inuence of the dosage of steel bers on the post-peak parameters is displayed in Figs. 10 and 11. Analyzing the test results, it can be concluded that, the type of steel ber has a pronounced effect on post-peak parameters. As the aspect ratio of both types is comparable, hooked end steel bers are more effective in improving the post-peak parameters than the straight ones for all tested bers volume ratio. The comparison of specimens reinforced with 0.5% of different types of steel bers are the best example. As was mentioned above, the maximum load obtained for these beams was almost the same, regardless of the ber type. Thus, the exural tensile strength is similar for both types of steel bers. However, the post-peak parameters appear to be signicantly bigger in case of hooked end steel bers. It can be noticed in comparison of exural behavior presented in Fig. 11a. In case of straight steel bers, the post-peak part of the loaddeection curve drops down while for hooked end steel bers it decreases slowly. It can be of course attributed to the bond strength of the matrix and steel ber and the bers length. The deformed shape provides better bond conditions than the short straight bers. Also longer bers can arrest not only microcracks, as in case of short bers, but also macrocracks, which has big inuence on the post-peak strength [4,25]. For both types of steel bers the post-peak parameters increase with the increase of ber dosage (Figs. 10 and 11). This dependency is maintained for hooked end steel bers. In case of straight steel bers, the residual tensile strengths fR,3d and fR,4d are almost the same regardless of the volume of steel bers. Analyzing the loaddeection curves obtained for the highest bers dosage of

straight steel bers (1.5%) it can be noted that it decreases suddenly after achieving the 2 mm deection. The straight steel bers were just pulled-out from the beams. Thus, for straight steel bers such a high dosage of bers does not improve the post-peak parameters, which are evaluated for deections higher than the 2 mm, remembering that fR,3d and fR,4d are evaluated for deections 2.15 mm and 3.0 mm, respectively. Barros et al. [35,36] investigated correlations between the postpeak parameters evaluated on concrete according to RILEM TC 162-TDF [9]. He suggested a linear relationship between equivalent tensile strengths of concrete (feq,2, feq,3). He claims that feq,3 is about 92% of the feq,2. Fig. 12 presents the relationship between feq,2 and feq,3 obtained by the authors. Taking into account the test results presented without the results obtained for tests on beams reinforced with 1.5% of straight steel bers, it can be concluded

Fig. 12. Relationship between equivalent exural tensile strengths: feq,2 and feq,3.

M. Paja k, T. Ponikiewski / Construction and Building Materials 47 (2013) 397408

405

that the feq,3 is only about 84% of the feq,2. This value is close to Barros calculations. The further considerations made by Barros suggested the linear trend also between fR,1d and fR,4d. Based on the authors research the residual exural tensile strength fR,4d is about only 52% of the fR,1d (Fig. 13). Thus, the relationship between residual and equivalent exural tensile strength is different than the one obtained by Barros. The relationship between feq,2 and fR,1d does not match with the relationships between feq,3 and fR,4d (Fig. 14). The model proposed by Barros has not been conrmed for SFR-SCC on the basis of the presented test results. 3.5. Fracture energy GF Authors determine the fracture energy as it is the fundamental parameter used in the practice [37]. The fracture energy is dened as the amount of energy necessary to create a crack of unit surface area projected in a plane parallel to the crack direction [38,39].

Table 7 Fracture energy of SRF-SCC. Mix 0 0.5_S 1_S 1.5_S 0.5_HE 1_HEa 1.5_HE
a

Fracture energy GF (N/m) 296 743 1237 1502 1487 2669 3532

Calculated fracture energy GF (N/m) 296 718 1140 1562 1421 2546 3671

Average of two specimens.

GF GF plain CV f

l d

GF

W 0 mg d BD a

where W0 area under the loaddeection curve, m mass of the specimen, g gravity, B,D beam thickness and depth, d deection at nale fracture, a notch depth. The fracture energy obtained on SFR-SCC are presented in Table 7. As the increase of ber content causes the increase of the maximum, residual and equivalent tensile strengths it increases the fracture energy of tested beams. The fracture energy depends almost linearly on the ber content for given ber type. Based on test results presented in the paper an equation to predict the fracture energy of the notched beams reinforced with the steel bers was proposed:

where Vf , l and d are: dosage of bers, length and diameter of ber, respectively. The parameter C should be determined experimentally. It depends on the parameters like: ber strength and type, the notch to depth of the beam ratio and others. In the present investigations the values of C as equal to C = 27 for straight steel bers and C = 60 for hooked end steel bers. GFplain is the fracture energy of plain SCC. The values of fracture energy calculated according to (6) were compared with the test results in Table 7 and graphically in Fig. 15. It can be noticed that the values of fracture energy calculated according to the above mentioned equation ts well to the test re-

Fig. 15. Fracture energy with respect to the volume ratio of steel bers. Fig. 13. Relationship between residual exural tensile strengths: fR,1d and fR,4d.

Fig. 14. Relationship between residual and equivalent exural tensile strengths.

Fig. 16. Comparison of fracture energy obtained by authors and others researchers.

406

M. Paja k, T. Ponikiewski / Construction and Building Materials 47 (2013) 397408

Fig. 17. Relationship between deection and CMOD for SCC beams reinforced with: (a) straight, and (b) hooked end steel bers.

sults. It is worth pointing out that the compared values of fracture energy were established at deection 3.33 mm (L/150). The regression coefcients obtained are in very small ranges. The fracture energy obtained on SCC was comparable to the fracture energy obtained by other researchers for SCC and normally vibrated concrete (Fig. 16). 3.6. Deection-CMOD relationship Crack mouth opening displacement (CMOD) was measured simultaneously with the deection of the beam, during the entire process of loading. The deection-CMOD curves are presented for both straight (Fig. 17a) and hooked end (Fig. 17b) steel bers. Analyzing the deection-CMOD relationship it can be noted that this dependency is almost linear for both types of steel bers (Fig. 17). The presented test results also indicate that the ber volume ratio does not inuence the relationship between deection and CMOD for SFR-SCC beams. The one exception are specimens reinforced with 1.5% of straight steel bers. In this case the deection-CMOD relationship strongly deviates from linear. A deection of about 2 mm causes almost 5 mm CMOD. To this, for beams reinforced with hooked end steel bers and 0.5% and 1.0% of straight steel bers, the deection of almost 3 mm cause 5 mm CMOD. Thus, the specimens with 1.5% dosage of straight steel bers were not taken into account in further considerations. The authors performed the study on evaluation of the relationship between deection and CMOD of SFR-SCC. According to Ding [40] mid-span deection (dm) is about 88.3% of the CMOD (Cw). The basis of the conclusion presented in [40] was a relationship between CMOD adequate to mid-span deection equal to 3.5 mm. The quotient of mid-span deection and CMOD lead to nding

Table 9

Mean value of the gradients Gm (dm/ Cw) 0.735

Standard deviation sp (mm) 0.054

Coefcient of variation Cv (sp/ G m) 7%

Gm + 2sp

Gm 2sp

0.844

0.626

the value of the gradients (Gm) and its coefcient of variation (Cv). In the present work the gradients were established for the deections: 2.0 mm, 2.5 mm and 3.0 mm and the corresponding CMOD (Fig. 17). In the authors research, with the limit of just 3 mm, because of higher mid-span deection, an instantaneous increase in deection with small increase of CMOD was noted. Further considerations were performed based on the calculations presented by Ding in [40]. The main results are shown in Tables 8 and 9. The equation which helps to predict the relationship between mid-span deection and CMOD can be proposed:

dm 0:735C w 0:109

It is worth pointing out that Eq. (4) can be used to predict the deection-CMOD relationship for both types of steel bers. However, analyzing Fig. 17 it can be noticed that for specimens reinforced with hooked end steel bers authors received bigger scatters than in case of beams with straight steel bers. The comparison of deection-CMOD relationship of SCC and normally vibrated concrete (NVC) is presented in the Fig. 18. It can be observed that the SCC reinforced with both types of steel bers achieves 5 mm CMOD for smaller deection (3.5 mm) than in case of NVC (4.4 mm). However, the behavior of SCC ts into the coefcient of variation of NVC.

Table 8 Fibre dosage (kg/m3) 40 80 120 40 80 120 40 80 120

Mid-span deection dm (mm) 3 3 3 2.5 2.5 2.5 2 2 2

CMOD Cw (mm) 4.31 4.02 4.35 3.46 3.27 3.48 2.74 2.78 2.72

Gradient Gm (dm/Cw) 0.697 0.750 0.695 0.723 0.765 0.718 0.729 0.720 0.735

Standard deviation sp (mm) 0.035 0.056 0.067 0.059 0.058 0.043 0.036 0.071 0.033

Coefcient of variation Cv (sp/Gm) 0.050 0.074 0.113 0.081 0.076 0.060 0.050 0.099 0.045

M. Paja k, T. Ponikiewski / Construction and Building Materials 47 (2013) 397408

407

References
[1] Okamura H. Self compacting high performance concrete ferguson lecture for 1996. Concr Int 1997;19(7):504. [2] Goodier CI. Development of self-compacting concrete. Proc ICE Struct Build 2003;156(4):40514. [3] Brandt AM. Fibre reinforced cement-based (FRC) composites after over 40 years of development in building and civil engineering. Compos Struct 2008;86:39. [4] Naaman AE, Reinhardt HW. High performance ber reinforced cement composites 2 (HPFRCC2). Ann Arbor, USA; 1995 June 1114. [5] Naaman AE. High performance ber reinforced cement composites: classication and applications. CBM-CI international workshop, Karachi, Pakistan; 2007. p. 389401. [6] Giaccio G, Tobes JM, Zerbino R. Use of small beams to obtain design parameters of bre reinforced concrete. Cem Concr Compos 2008;30(4):297306. [7] Nataraja MC, Dhang N, Gupta AP. Toughness characterization of steel berreinforced concrete by JSCE approach. Cem Concr Res 2000;30(4):5937. [8] Japan Society of Civil Engineers. Method of test for exural strength and exural toughness of ber reinforced concrete. Standard SF-4; 1984. p. 5866. [9] RILEM TC 162-TDF. Test and design methods for steel bre reinforced concrete, nal recommendations. Mater Struct 2002;35(9):57982. [10] EN 14651:2005+A1:2007(E). Test method for metallic bre concrete measuring the exural tensile strength (limit of proportionality (LOP), residual) CEN European Commit-439 tee for Standardization 440. [11] ASTM C 1018. Standard test methods for exural toughness and rst crack strength of ber reinforced concrete (using beam with third point loading). ASTM 2002(4):63744. [12] Bentur A, Mindess S. Fiber reinforced cementitious composites. New York: Elsevier Applied Science; 1990. [13] Buratti N, Mazzotti C, Savoia M. Post-cracking behaviour of steel and macrosynthetic bre-reinforced concretes. Constr Build Mater 2011;25:271322. [14] Glinicki MA. Beton ze zbrojeniem strukturalnym. XXV Oglnopolskie Warsztaty Pracy Projektanta Konstrukcji. Szczyrk; 2010. p. 279308 [in Polish]. [15] Katzer J, Domski J. Quality and mechanical properties of engineered steel bres used as reinforcement for concrete. Constr Build Mater 2012;34:2438. [16] Katzer J. Steel bres and steel bre reinforced concrete in civil engineering. Pac J Sci Technol 2006;7(1):538. [17] di Prisco M, Plizzari G, Vandewalle L. Fibre reinforced concrete: new design perspectives. Mater Struct 2009;42(9):126181. [18] Ferrara L, Bamonte P, Caverzan A, Musa A, Sanal I. A comprehensive methodology to test the performance of steel bre reinforced selfcompacting concrete (SFR-SCC). Constr Build Mater 2012;37:40624. [19] Sahmaran M, Yurtseven A, Yaman IO. Workability of hybrid ber reinforced self-compacting concrete. Build Environ 2005;40:16727. [20] Okamura H, Ozawa K. Mix design for self-compacting concrete. Concr Lib JSCE 1995. [21] Okamura H, Ouchi M. Self-compacting concrete. J Adv Concr Technol 2003;1(1):515. [22] Banthia N, Trottier J-F. Test methods for exural toughness characterization of ber reinforced concrete: some concerns and a propositions. ACI Mater J 1995;92(1):4857. [23] Baar BIG, Lee MK. Round-robin analysis of the RILEM TC 162-TDF beam bending test: Part 2 approximation of d from the CMOD response. Mater Struct/Mater et Constr 2003;36:62130. [24] RILEM TC 145-WSM. Workability and rheology of fresh concrete: compendium of tests. In: Bartos, Donebi, Tamimi, editors. Report of RILEM technical committee TC 145-WSM workability of special concretes. Cachan: RILEM Publications; 2002. [25] Namman AE, Reinhardt HW. Proposed classication of FRC composites based on their tensile response. Mater Struct 2006;39(5):54755. [26] Torrijos M, Barragan BE, Zerbino RL. Placing conditions, mesostructural characteristics and post-cracking response of ber reinforced selfcompacting concretes. Constr Build Mater 2010;24:107885. [27] Cai J, Jiang H, Zhu Y, Wang D. Mechanical properties of ber reinforced selfcompacting concrete. Optoelectron Adv Mater Rapid Commun 2010;4(7):10136. [28] Miao B, Chern J-H, Yang C-A. Inuences of ber content on properties of selfcompacting steel ber reinforced concrete. J Chin Inst Eng 2003;26(4):52330. [29] Akcay B, Tasdemir MA. Mechanical behavior and bre dispersion of hybrid steel bre reinforced self-compacting oncrete. Constr Build Mater 2012;28:28793. [30] Kang ST, Lee Y, Park YD, Kim JK. Tensile fracture properties of an ultra high performance ber reinforced. [31] S ahin A, Koksal F. The inuences of matrix and steel bre tensile strengths on the fracture energy of high-strength concrete. Constr Build Mater 2011;25:18016. [32] Hameed R, Turatsinze A, Duprat F, Sellier A. Study on the exural properties of metallic-hybrid-bre-reinforced concrete. Maejo Int J Sci Technol 2010;4(02):16984. [33] Bayramov F, TaS demir C, TasS demir MA. Optimisation of steel bre reinforced concretes by means of statistical response surface method. Cem Concr Compos 2004;26:66575. [34] Baar BIG, Lee MK. Round-robin analysis of the RILEM TC 162-TDF beam bending test: part 1 test method evaluation. Mater Struct 2003;36: 60920.

Fig. 18. Comparison of test results and Eq. (6) of SCC to NVC.

4. Conclusion To assess the exural behavior of reinforced self-compacting concrete (SFR-SCC), tests according to RILEM TC 162-TDF and EN 14651 were performed. Two types of steel bers with comparable aspect ratio were investigated. The hooked end and straight steel bers were used with a dosage of 0.5%, 1.0% and 1.5% of volume ratio. The following general conclusions can be drawn from the study presented in the paper:  The type of steel bers affects much the post-peak behavior of SCC. The deection-hardening response was observed in case of SCC reinforced with hooked end steel bers. On the other hand, the addition of straight steel bers affects the deection-softening response. Thus, the deection at the maximum load for straight steel bers was close to the one derived for plain specimens. In case of hooked end steel bers deection at the maximum load was much bigger.  Generally, SCC indicates similar exural behavior to NVC where the increase of bers volume ratio increases the exural tensile strength. A formula to predict the exural tensile strength on SFRC proposed by Naaman was conrmed for SFR-SCC notched beams.  The fracture energy increases with the increase of ber dosage and is higher for hooked end steel bers than for straight ones. A formula to predict the fracture energy of steel bers reinforced SCC was proposed.  The relationship between the residual and equivalent tensile strengths proposed by Barros for SFRC was not conrmed for SFR-SCC.  The ber volume ratio does not inuence the relationship between deection and CMOD in case of SCC with straight and hooked end steel bers. Thus, the deection-CMOD dependency can be predicted by one equation for two investigated types of steel bers. Generally, the deection-CMOD relationship for SCC does not differ much from the one obtained for normally vibrated concrete. Even though, the SCC achieves the maximum CMOD for lower deections than NVC, the test results of SCC t into the coefcient of variation of NVC.  One more general conclusion about using straight steel bers in SCC can be made. In this case, the increase in ber dosage causes the increase of exural tensile strength but with no increase in post-peak parameters obtained for deection higher than 2 mm. For this value of deection the maximum crack mouth opening deection was noted.

408

M. Paja k, T. Ponikiewski / Construction and Building Materials 47 (2013) 397408 [38] RILEM FMT 89. Size-effect method for determining fracture energy and process zone size of concrete. Mater Struct 1990;23(6):4615. [39] RILEM FMC-50. Determination of the fracture energy of mortar and concrete by means of three point bend tests on notched beams. Mater Struct 1985;18(4):28790. [40] Ding Y. Investigation into the relationship between deection and crack mouth opening displacement of SFRC beam. Constr Build Mater 2011;25(5):243240.

[35] Barros JAO, Cunha VMCF, Ribeiro AF, Antunes JAB. Post-cracking behavior of steel bre reinforced concrete. Mater Struct 2005;38:4756. [36] Barros JAO, Antunes JAB. Experimental characterization of the exural behaviour of steel bre reinforced concrete according to RILEM TC 162-TDF recommendations. In: Schntgen B, Vandewalle L, editors. Proceedings of the RILEM TC 162-TDF workshop; 2003. p. 7789. [37] Beygi MHA, Kazemi MT, Nikbin IM, Amiri JV. The Effect of water to cement ratio on fracture parameters and brittleness of self-compacting concrete. Mater Des 2013;50:26776.