Hassan Aoude * & Michael Cohen Department of Civil Engineering, University of Ottawa, Ottawa, Canada

Transcription

1 Shear Response o SFRC Beams Constructed with SCC and Steel Fibers Hassan Aoude * & Michael Cohen Department o Civil Engineering, University o Ottawa, Ottawa, Canada * haoude@uottawa.ca ABSTRACT: This paper presents the experimental results rom eleven slender beams constructed with highly workable SFRC tested under our-point loading. In the study sel-consolidating concrete (SCC) was combined with steel ibers to improve workability and concrete placement. The response o the beams in terms o shear versus delection response, crack control and damage tolerance is reported. The results demonstrate that the combined use o SCC and steel ibers in shear-deicient beams results in signiicant improvements in shear resistance and enhancements in lexural ductility. In the second part o the paper a model is proposed or predicting the shear capacity o SFRC beams. The proposed model and various equations proposed in the literature are used to predict the shear capacity o the beams tested in the experimental program. The results demonstrate the need or developing accurate and reliable equations or predicting the shear capacity o SFRC beams constructed with SCC and steel ibers. 1 INTRODUCTION To prevent shear ailure beams are traditionally reinorced with transverse shear reinorcement. One alternative to this reinorcement lies is in the use o steel iber reinorced concrete (SFRC). The addition o ibers improves the diagonal tension capacity o concrete resulting in an increase in shear resistance which can promote lexural ailure and ductility. The structural use o SFRC typically requires the addition o high iber contents which can result in reduced workability and diiculty in concrete placement. The combined use o sel-consolidating concrete (SCC) and steel ibers has been proposed as a solution to this problem and can result in highly workable SFRC that can allow or the higher iber contents required or structural applications. This paper reports the results o an experimental study that investigates the perormance enhancements that can be gained rom the use o highly workable SFRC in beams tested under our-point loading. The response o the beams in terms o shear versus delection response, crack control and damage tolerance is reported. In the second part o the paper a model or predicting the shear resistance o SFRC beams is proposed. Equations proposed in the literature and by the authors are then used to predict the shear capacity o the SFRC beams tested in the experimental program. 2 LITERATURE REVIEW 2.1 Research on shear behavior o SCC beams There is limited published data in the literature on the shear behavior o plain SCC beams. Lachemi et al. (25) tested a series o 18 normal concrete (NC) and SCC beams to examine the inluence o aggregate size and SCC on the shear behavior o beams. The authors ound that a decrease in aggregate size rom 19 mm to 12 mm reduces the ultimate shear capacity o reinorced concrete beams. This reduction in capacity was linked to reduced aggregate interlock when smaller aggregates are used. Similarly a comparison between the behavior o the NC and SCC series showed the beams constructed with SCC had reduced shear capacity when compared to the companion NC beams. This reduction in capacity was linked to reductions in aggregate interlock due to the lesser coarse aggregate content that is typical 71

2 o SCC. In another recent study, Lin and Chen (212) tested 24 beams, which included 8 NC beams and 16 beams constructed with two types o SCC. The "Type I" SCC beams contained a greater amount o coarse aggregate, while the "Type II" SCC beams had lesser aggregate content. The variables included shear-span-to-depth ratio (a/d), spacing and strength o transverse reinorcement. The results demonstrated that the Type I SCC beams had higher shear capacity when compared to the companion NC beams, however in the case o the SCC beams with reduced aggregate content (Type II), structural perormance was inerior with reduced cracking and ultimate strengths. The results rom these two studies demonstrate the need or urther research on the shear behavior o SCC beams and the development o SCC-speciic design equations that can modiy existing "traditional concrete" shear-strength requirements in codes. 2.2 Research on shear behavior o SFRC beams Research over the past three decades has shown that the provision o steel ibers can be used to enhance shear resistance in reinorced concrete beams. Parra-Montesinos (26) presented a large database o SFRC beam test results which included 147 beams rom 16 studies. Analysis showed that all beams in the database having iber content, V.5% exhibited shear stress at ailure greater than.17 'c [MPa], equivalent to Vc as deined in the ACI-318 code (ACI, 211), with beams having V.75% ailing at shear stresses not less than.3 'c [MPa]. Based on the analysis o the database, it was recommended that i added in suicient quantity, steel ibers can be used as a replacement to minimum shear reinorcement in beams subjected to shear orces equivalent to.5 Vc and Vc. 2.3 Research on shear behavior o SCFRC beams The addition o steel ibers in traditional concrete mixes can cause problems in workability and placement, particularly when high iber contents are used. The use o sel-consolidating concrete (SCC) has been proposed to reduce these problems and acilitate placement (Khayat and Roussell, 2). Studies have shown that by utilizing a highly lowable SCC mix, sel-consolidating properties can be maintained at low or moderate iber contents, while adequate workability can be maintained at higher iber contents (with some loss o sel-consolidating properties). Liao et al. (26) presented an excellent review o sel-consolidating iber reinorced concrete (SCFRC) mix designs studied in the literature. While there has been several studies ocusing on the development o SCFRC mix designs, there is limited data in the literature related to the structural use o SCFRC in beams. Greenough and Nehdi (28) perormed a study on 13 slender SCFRC beams having a/d ratios equal to or greater than 3 and having iber contents ranging rom.5-1.%. The authors noted that SCC is particularly well suited or iber addition due to its improved rheological properties. The authors ound that the addition o steel ibers improves the shear behavior o SCC slender beams, with results showing that addition o 1.% iber content results in a 128% increase in shear capacity. You et al. (21) carried out another investigation on two series o SCC beams reinorced with steel ibers. The irst series included 8 beams with an a/d ratio o 3.. Two o the beams were constructed without stirrups while 6 o the beams had transverse reinorcement ratios ranging rom.13%-.22%. The beams in this second series had an a/d ratio o 3.2, and transverse reinorcement ratio ranging rom %.the results showed that the addition o steel ibers increases the shear capacity o SCC beams and that a suicient quantity o ibers can alter the ailure mode rom a brittle shear collapse into a ductile lexural ailure. The authors o this study also demonstrated that combining steel ibers with traditional shear reinorcement in the orm o stirrups can produce a positive hybrid eect and allow or a partial reduction o shear reinorcement without negatively aecting perormance. It is noted that only a moderate iber content o -5 kg/m 3 was used in this study, with the authors noting that 5 kg/m 3 may be an upper limit or a ully selconsolidating iber-reinorced SCC. 3 RESEARCH NEED While there is important research in the literature on the behavior o SFRC beams, there is limited research data on the response o beams constructed with SCC and steel ibers. Given the recent adoption o provisions related to use o SFRC in beams in the ACI 318 code and increased interest in the structural use o SCFRC and highly-workable SFRC, it is o importance to enrich the experimental database related to the shear behavior o beams constructed with these materials. There is also a need to assess 72

3 the ability o existing models proposed or SFRC to predict the shear resistance o beams constructed with SCC and steel ibers. 4 EXPERIMENTAL INVESTIGATION 4.1 Description o test specimens A total o eleven slender SFRC beams constructed with SCC and steel ibers were tested under ourpoint loading. Table 1 summarizes the properties o the beams tested in the experimental program. The beams had cross-sectional dimensions o 125 mm x 25 mm, with all specimens having a clear cover o 3 mm, with two ratios o longitudinal reinorcement. Two 15M bars (d b =16 mm and A s =2 mm 2 ) were placed at the bottom o the seven beams in the M15-series, while two 2M bars (d b =19.5 mm and A s =3 mm 2 ) were used in the our beams in the M2-series. No transverse reinorcement was provided in any o the specimens. All the specimens were simply supported over a span o 24 mm and were subjected to our-point loading with a shear span o 8 mm rom the ace o the support (shear span-to-depth ratio a/d was approximately 3.8), and a constant moment region o 8 mm. The beams rested on roller and pin supports while a load transer beam was used to apply the two point loads. Table 1. Summary o specimen details Beams Concrete Longitudonal Fiber Fiber Type Reinorcement Type content M15-.% ZP-35.% M15-.5% ZP-35.5% M15-1.% KING ZP-35 1.% 2-15M M15-1.5% SCC ZP % (ρ=1.55%) M15-.5%H BP8/3.5% M15-.75%H BP8/3.75% M15-1.5%5D 5D 1.5% M 2-.75% KING ZP-35.75% M 2-1.% SCC 2-2M ZP-35 1.% M 2-1.%A Custom (ρ =2.33%) ZP-35 1.% M 2-1.5%A SCC ZP % Figure 1 shows a typical beam prior to testing. The specimens in the M15 series were cast using sel-consolidating concrete (KING-SCC mix) having dierent iber contents. Beam M15-.% contained no ibers, while Beams M15-.5%, M15-1.% and M15-1.5% were constructed using SCC containing Dramix ZP-35 ibers in a quantity o.5%, 1.% and 1.5% by volume o concrete, respectively. Beams M15-.5%H and M15-.75%H, were built using the same SCC mix but with.5% and.75% o BP 8/3 high-strength steel ibers, respectively. Beam M15-1.5%-5D was built using 1.5% o highstrength 5D ibers (see next section or iber properties). The inal two specimens, M2-1.%A and M2-1.5%A, were constructed in a similar ashion, but a dierent SCC mixture (Custom-SCC) was used, combined with 1.% and 1.5% o Dramix ZP- 35 ibers, respectively. Figure 1. Typical beam specimen prior to testing 4.2 Materials Properties In this test program the combined use o SCC and ibers resulted in highly workable SFRC. Two base SCC mixtures were considered. The KING-SCC mix consisted o a pre-packaged, sel-consolidating concrete mixture with a strength o 5 MPa (MS Sel- Consolidating Concrete, KING Packaged Materials). The mix contained a maximum aggregate size o 1 mm with a sand-to-aggregate ratio o approximately.55 and a water-cement ratio o approximately.42. An air-entraining admixture, a superplasticizer and a viscosity modiying admixture (VMA) are incorporated into the mix in the orm o dry powder. Table 2 lists the composition o this concrete as speciied by the manuacturer. This mix was utilized in nine o the specimens used in the experimental program. The irst "control" batch contained no ibers, while the iber contents o the remaining batches ranged rom.% to 1.5% and included three iber types (Dramix ZP35, Dramix BP8/3 and Dramix 5D ibers). The last two beams in this experimental program were constructed using a customized SCC mixture (Custom-SCC). The objective was to produce a regular-strength SFRC mixture with high workability at high iber contents (1.% and above). This was accomplished by limiting coarse aggregate size, using cement with Fly Ash and the use o 73

4 chemical admixtures. The mixture contained Type 1 cement and class C ly ash. Coarse aggregate size was limited to 12 mm while silica sand was used as ine aggregate. The mix incorporated hooked-end steel ibers at iber contents o 1.% to 1.5%. A superplastizer (ADVA CAST 575) and viscosity modiying admixture (V-MAR 3) were used in order to improve workability and prevent segregation. Table 2 summarizes the inal mix proportions which were determined using a series o trial batches and based on a review o mixtures proposed in the literature (Liao et al. 26). Table 2. Mix design properties Component KING-SCC Custom-SCC Cement 5 (kg/m 3 ) 473 kg/m 3 ) Fly Ash (kg/m 3 ) WC ratio Coarse Agg. 765 (kg/m 3 ) 439 (kg/m 3 ) Fine Agg. 915 (kg/m 3 ) 86 (kg/m 3 ) Mass Density 23 (kg/m 3 ) 24 (kg/m 3 ) Superplastizer VMA In orm o dry powder 292 (ml/1 kg)* 52 (ml/1 kg)** Steel Fibers (%) (%) * ADVA CAST 575, in ml/1 kg o cementitious materials ** V-MAR 3, in ml/1 kg o cementitious materials Three types o steel ibers were used in this experimental program (see Figure 2). All ibers are manuactured by Bekaert. The ZP35 hooked-end ibers have 3 mm length, aspect-ratio (L /d ) o 55 and are made o normal strength steel wire with a tensile strength o 11 MPa. The hooked-end BP8/3 ibers have 3 mm length, aspect-ratio o 8 and a higher tensile strength o approximately 23 MPa. Finally, the 5D ibers have length o 6 mm, aspect-ratio o 65, tensile strength o 23 MPa and have a more optimized hook shape (two bends in the end hooks to increase pullout strength). The 15M bars in the irst series o beams had a yield strength ( y ) o 441 MPa, while the 2M bars which were utilized to reinorce the remaining beams had a yield strength o 474 MPa. The workability o the concrete mixtures was measured by perorming slump low tests on each batch o SCC and SFRC. Table 3 summarizes the average results rom the slump low test. As noted previously the goal o combining SCC and steel ibers in this study was to produce highly workable SFRC. As such, sel-consolidating properties were expected to be lost at the higher iber contents. Nonetheless, all mixtures remained suiciently workable. During casting o the specimens constructed with the KING-SCC mix and normalstrength ibers, no vibration was required at a iber content o.5% and 1.%, while slight vibration was required when casting the specimens with 1.5% ibers. The use o.75% o high-strength ibers signiicantly reduced workability owing to the higher aspect ratio o this iber, and required signiicant vibration during casting. A suiciently workable mix was obtained when using 1.5% o the 5D ibers and only slight vibration was required during casting, although some slight segregation was observed during visual inspection ater the slump low test. In terms o the custom-scc mix, the SFRC showed highly workability allowing or easy concrete placement during construction o the beams. Figure 2. Hooked-end steel ibers considered in this study: (a) ZP35; (b) BP8/3; and (c) 5D ibers. A series o lab cured cylinders and lexural beams were prepared during the mixing o the concrete (three cylinders and one lexural beam or each batch). The specimens were cured with burlap and covered with plastic sheets or seven days, and were then demolded and air-cured in the laboratory until testing. The control samples were tested on the day o the corresponding beam tests (the age o the specimens on the day o testing varied between 95 and 97 days, with the exception o the beam with the 5D ibers which had an age o 45 days at the day o testing). Compressive strengths were obtained by testing standard cylinders having a diameter o 1 mm and a height o 2 mm (average strengths at the day o testing, c are summarized in Table 3). In addition, bending tests perormed according to the ASTM C169 Standard (ASTM, 27) were used to examine toughness. The tests were conducted on samples that were 1 mm by 1 mm by 4 mm in size which were loaded over a span o 3 mm. Average modulus o rupture values, r are summarized in Table 3 and toughness parameters are reported in Table 4 or the various SCC and SFRC mixtures. 74

5 Load (kn) Load (kn) Electronic Journal o Structural Engineering Table 3. Average concrete properties: compressive strength, modulus o rupture, slump low, and sample slump low photos Mix SCC-.% SCC-.5% SCC-.75% SCC-1.% SCC-1.5% c, MPa r, MPa * Slump, mm Sample Slump Photo Mix SCC-.5H% SCC-.75H% SCC-1.%A SCC-1.5%A SCC-1.5%5D c, MPa r, MPa * Slump, mm Sample Slump Photo * No lexural beams were tested or the mixtures with.75% ZP35 ibers and 1.5% 5D ibers Table 4. Results rom lexural toughness tests and comparison to ACI perormance requirements Mixture ASTM C169 Meets ACI requirements? Name r,aci 1 p % 1 75% 1 T 15 v.75% 3 9% % 1 M15-.% M15-.5% M15-1.% M15-1.5% M15-.5%H M15-.75%H M 2-1.%A M 2-1.5%A Load-delection responses are presented in Figure 3 (it is noted that no lexural beam data is available or the specimens with.75% ZP35 ibers and 1.5% 5D ibers). As expected increasing iber content results in improved post-cracking resistance and toughness. For the KING-SCC mix, examination o the results in Table 4 shows that the T 15 toughness parameter, which corresponds to the area under the load-delection curves rom to L/15, increased by 49% or the mix with 1.5% ZP35 ibers when comparing to the mix with 1% ZP35 ibers. Similarly, the use o.5% high-strength BP8/3 ibers resulted in a 61% increase in toughness when comparing to the mix with.5% normal-strength ZP35 ibers. When comparing to the KING-SCC mix, the use o 1% ZP35 ibers in the Custom-SCC mix showed a 63% increase in toughness. The ACI-318 code recently introduced provisions permitting the use o SFRC in lexural members (ACI, 211). In order to qualiy as an alternative to minimum shear reinorcement in beams material, dimensional and loading limits must be met. At the material level, the SFRC must have a minimum iber content o 6 kg/m 3 and meet minimum perormance requirements. The residual strength obtained rom the ASTM C169 test at midspan delections o L/3 and L/15 must be greater or equal to 9% and 75% o the irst-peak strength, respectively (where the irst-peak strength, p, is obtained rom a lexural test but not lesser than the value speciied in Eq. 9-1 o the ACI 318 code) SCC.% SCC.5% SCC 1.% SCC 1.5% Net Delection (mm) (a) %,.5%, 1%, 1.5%; SCC.5%H SCC-.75%H SCC 1.%A SCC 1.5%A Net Delection (mm) (b).5%h,.75%h, 1%A, 1.5%A Figure 3. Sample lexural beam load-delection responses. 75

6 Examining the residual strength values in Table 4 it is seen that or the KING-SCC mix, the perormance requirements speciied by the code are met or the mixtures containing 1.5% normal-strength ibers and.75% high-strength ibers, whereas the requirements are not met or the mixtures with.5% or 1% ibers. For the Custom-SCC mix, the ACI requirements are satisied at iber contents o 1.% & 1.5%. 5 RESULTS AND DISCUSSION 5.1 M15 Series-Eect o iber content To recall, the seven beams in the M15-KING series had identical cross-sectional properties and were cast using a traditional SCC mix. The series included one control beam (Beam M15-.%) and six beams having varying iber contents (ranging rom.5% to 1.5%) and iber types. Table 5 and Figure 4 compare the shear and delections values (at maximum and at ailure) and the load-delection responses o the various beams tested in this series. In terms o shear resistance, as has been reported or traditional SFRC beams, one can see that the addition o steel ibers in SFRC beams constructed with SCC and steel ibers leads to improvements in shear capacity. In addition, as the iber content is increased so does the relative increase in shear strength. For example, as shown in Table 5 and Figure 4(a), the addition o.5% ibers in Beam M15-.5% increased the shear capacity o the beam by 47% when comparing to the strength o the control beam. A urther increase in iber content to 1.% has led to 63% improvement in shear strength when compared to the control beam. The use o 1.5% ibers did not lead to a urther increase in capacity when compared to the companion beam with 1.% ibers. It is noted that these last two beams ailed in lexure and the result shows that addition o steel ibers does not have a signiicant inluence in increasing moment capacity in beams ailing in lexure. 5.2 M15 Series-Eect o iber type In addition to the above specimens, two o the beams in the M15-series were constructed with highstrength BP8/3 steel ibers. A comparison o the results rom Table 5 and Figure 4(b) shows the beneicial impact o using high-strength ibers. Besides the improvement in shear capacity, the use o high-strength ibers has contributed to an enhanced post-peak response and ductility. Just as in the case o the normal strength ibers, the use o highstrength ibers was shown to improve ultimate shear capacity and member response. When compared to the control beam, increases in strength o 54% and 63% or Beams M15-.5%H and M15-.75%H were observed. The response o Beam M15-.5%H shows that the use o high-strength ibers allowed the beam to sustain signiicant delections beore ailure when compared to the control specimen. Similarly, although both Beams M15-.5% and M15-.5%H ailed in shear at similar ailure loads, a comparison between the delections at ailure (see Table 5) and the member responses (see Figure 4(b)) demonstrates an improved level o ductility or the beam reinorced with high-strength ibers (with Beam M15-.5%H experiencing 2% increase in delections beore ailure). The results also show that an increase in high-strength iber content to.75% resulted in a urther increase in strength and was suicient to produce a ductile lexural ailure in Beam M15-.75%H. As noted previously, the use o highstrength ibers resulted in reduced workability and some reduction in compressive strength (possibly due to increased entrapped air). Nonetheless, it is noted that the structural response o Beam M15-.75%H is very similar to the response o Beam M15-1.% in terms o lexural ductility. The results rom specimens M15-.5%H and M15-.75%H indicate the potential beneits rom the use o higher strength steel ibers in beams, however urther research is recommended. Owing to the higher aspectratio o the BP8/3, careul SCC mix proportioning is recommended in order to ensure adequate workability and reliable structural perormance. Figure 4(c) shows the response o Beams M15-1.5% and M15-1.5%5D. Both beams ailed in lexure, however the use o the 5D ibers, which have optimized strength and anchorage properties, resulted in an increase in beam lexural capacity. 5.3 M2 Series-Eect o iber content Beams M2-.75% and M2-1.% were constructed with a traditional SCC mix, normal strength hooked-end steel ibers and iber contents o.75% and 1.% by volume o concrete, respectively. It is noted that no control beam was tested in the M2- series. To allow or a comparison o shear capacities, the nominal shear strength o a reinorced concrete beam having identical properties but without ibers 76

7 was predicted using the general shear design method o the CSA A Standard (CSA, 24); assuming compressive strength, c o 5 MPa and yield strength, y o 474 MPa, a nominal capacity o 34 kn was predicated. Table 5 compares the maximum shear capacities o the two SFRC beams with the calculated nominal capacity and indicates a relative increase in resistance due to addition o steel ibers. The results also show that increasing the iber content rom.75% to 1.% is accompanied by a relative increase in shear resistance. As shown in Figure 5, despite the enhancement in shear capacity, the ailure pattern or both beams in this series was brittle and the beams ailed in shear with a sudden loss in load-carrying capacity at ailure. 5.4 M2 Series-Eect o longitudinal reinorcement The beams discussed in the previous section were reinorced with two 2M bars, whereas all other properties including concrete type remained unchanged rom the M15 test series. A comparison o the responses o Beams M15-1.% and M2-1.% in Figures 4(a) & 5 show that while 1.% ibers was suicient to transorm the brittle shear response in the M15 series beam, the use o 1.% ibers in the beam with 2M reinorcement was not suicient to avoid a brittle shear ailure. The reason returns to the higher lexural capacity in the beams with larger reinorcement ratio. That is to say that the additional shear capacity provided by 1.% steel ibers was not suicient to make up the shear required to reach the beam s lexural capacity in the beam with larger reinorcement ratio. The result puts into evidence the importance o developing accurate equations or predicting the shear resistance o SFRC beams. 5.5 M2 Series-Eect o concrete type In addition to the two beams constructed with traditional SCC, the M2 series also included two other beams constructed with a customized concrete mixture (SCC-Custom Mix). In comparing the results in Table 5 Beams M2-1.%A and M2-1.5%A have shown a 74% and 82% higher shear capacity respectively when compared to the expected nominal shear strength capacity or a beam without ibers. A comparison o the responses o Beam M2-1.% rom the M2-KING series and Beam M2-1.%A allows or an investigation into the eect o concrete type on response. Both beams were reinorced with 1.% normal-strength hooked-end ibers and had identical properties with the exception o concrete mix type. In comparing the results in Figure 5 one can see that while the capacities o the beams was similar and ailure in both cases was in shear, the beam constructed with the customized SCC mix (Beam M2-1.%A) showed an improved level o ductility when compared to the companion beam. This may be linked to the improved concrete tensile properties o this SFRC mixture and the act that yielding was initiated in the longitudinal reinorcement beore the brittle ailure in this specimen. As shown in Figure 5, while 1.% ibers was not suicient to prevent a brittle shear ailure in the previous two beams in this series, the use o 1.5% ibers in Beam M2-1.5%A was adequate to transorm the ailure into a ductile lexural response. Shear (kn) Shear (kn) Shear (kn) M15-.% M15-.5% M15-1.% M15-1.5% Mid-span delection (mm) (a) %,.5%, 1%, 1.5% M15-.5% M15-.5% H 1 M15-.75% H Mid-span delection (mm) (b).5%,.5%h,.75%h Mid-span delection (mm) (c) 1.5%, 1.5%(5D) M15-1.5% M15-1.5%(5D) Figure 4. Results or the M15 series 77

8 Load (kn) Electronic Journal o Structural Engineering Shear (kn) M2-.75% M2-1.% 1 M2-1.% A M2-1.5% A Mid-span delection (mm) Figure 5. Results or the M2 series:.75%, 1%, 1%A, 1.5%A. Table 5. Results or beams tested in the experimental program Beam specimen Max Load (kn) Total Load Shear Load Increase in Shear (%) Delection (mm) at Peak at Failure M15-.% M15-.5% % M15-1.% 96 48* 63% M15-1.5% 93 46* 6% M15-.5%H % 4 4 M15-.75%H 95 48* 63% M15-1.5% 5D 14 52* 73% M 2-.75% %** M 2-1.% %** 2 2 M 2-1.%A %** M 2-1.5%A* * 82%** *Beam ailed in lexure ** comparison to the nominal shear strength o a beam without ibers as predicted using the CSA A23.3 general shear design method 5.6 Eect on crack widths As expected the utilization o steel ibers in the SFRC beams had an important inluence on shear and lexural crack widths (plots or the growth o crack widths or a sample o the beams are shown in Figures 6(a) and 6(b)). For the M15-series, the control specimen (Beam M15-.%) as well as the beams with.5% normal strength and high-strength ibers (Beams M15-.5% and M15-.5%H, respectively) ailed suddenly in shear. For the control beam the ailure was accompanied with very little warning o the eminent shear ailure with the crack width just beore sudden ailure remaining at.2 mm. Beam M15-.5% which also ailed under shear, was able to resist a slightly larger crack width o approximately.4 mm beore ailure. The use o high-strength BP 8/3 ibers in Beam M15-.5%H permitted this specimen to show very large critical shear crack width (exceeding 1. mm) prior to ailure (see Figure 7). This enhancement is linked to the higher tensile strength and increased aspect-ratio o this iber type, which results in enhanced pullout behavior allowing the BP 8/3 ibers to more eectively bridge the critical diagonal shear crack and the act that some lexural yielding had initiated in this beam prior to the brittle shear ailure. Beams M15-1.% and M15-1.5% o the M15-series ailed in lexure. A comparison o the lexure crack widths or these beams demonstrates that increasing the iber content rom 1.% to 1.5% results in a better control o lexural crack widths at equivalent load stages. The results also show that using.75% high-strength BP8/3 ibers results in a better control o crack widths when compared to the companion specimens with 1.% and 1.5% normal strength ibers. In addition to the above observations, examination o the diagonal shear cracks in the SFRC beams reveals that the ibers pullout rather than raction across the cracking plane when shear ailure occurs. An interesting observation is that the hooked-end ibers deorm and straighten during pullout (see Figure 7). Load (kn) M15-.75%H Crack width (mm) M15-.% M15-.5% M15-1.5% M15-.5%H (a) Shear crack-widths as a unction o shear load M15-.75%H Crack width (mm) M15-.% M15-1.% M15-1.5% (b) Flexural crack-widths as a unction o total load Figure 6. Eect o ibers on crack widths Figure 7. Other observations: (let) Control o shear crack up to 1 mm in width prior to ailure in beam M15-.5%H; (right) observed straightening o hooked-end ibers during pullout. 78

9 5.7 Eect on crack patterns In addition to the improvements in controlling crack widths, the use o steel ibers had a noticeable inluence on crack patterns in the beams tested in the experimental program. Figure 8 shows the crack patterns or the various specimens at the end o testing. Comparing Beam M15-.%, M15-.5% and M15-.5%H, it can be seen that use o.5% ibers has reduced the spacing between cracks when compared to the control specimen. The specimen with.5% high-strength ibers shows urther reductions in crack spacing when comparing to the control beam and the companion specimen constructed with normal-strength ibers. Specimens ailing in shear Specimens ailing in lexure M15-.% M15-.5% M15-.5%H M2-.75% M2-1.% M2-1.%A M15-1.% M15-1.5% M15-.75%H M15-1.5%5D M2-1.5%A Figure 8. Comparison o crack patterns in the various beams at the end o testing When comparing the control beam with the specimens that ailed in lexure (Beams M15-1.%, M15-1.5%, M15-.75%H) one can see that the use o higher iber content (greater or equal to 1.%) led to both reductions in crack widths and spacing. In addition, the higher iber contents resulted in a more diused cracking pattern with the ormation o secondary cracks growing out o primary cracks. This is particularly evident in the ailure photo o Beam M15-.75%H which contained high-strength ibers. Similarly the use o 1.5% 5D ibers has also resulted in a more diused cracking pattern when comparing to the beam with 1.5% ZP35 ibers. For the M2 series, it is noted that Beam M2-1.%A, which was constructed the customized SCC mix and 1.% ibers, showed signiicant reductions in crack spacing with multiple lexural and inclined shear cracks despite having ailed in shear. Beam M2-1.5%A which ailed in lexure showed multiple cracking and a more diused cracking pattern when compared to Beam M15-1.5% o the previous test series. 5.7 Comparison to perormance requirements in the ACI 318 code As noted previously, the ACI-318 code now includes provisions permitting the use o SFRC to replace minimum shear reinorcement in beams. In addition to the dimensional and loading limits, the SFRC must have minimum iber content (.75% by volume o concrete) and meet minimum perormance requirements (residual strength limits rom the ASTM C169 lexural toughness test). In this study the use o steel ibers resulted in more ductile behavior, promoting lexural ailure. This tendency increased with higher iber actor (v L /d ) and demonstrates that when the iber actor approaches the minimum residual strength requirements speciied in the ACI 318 code, the SFRC mix becomes more suitable or improving shear resistance and response. 6 ANALYSIS 6.1 Proposed procedure or predicting the shear resistance o SFRC beam This section o the paper describes a procedure or predicting the shear resistance o SFRC beams without web reinorcement. In the proposed model the shear resistance o an SFRC beam is assumed to be equivalent to the expected shear strength o a traditional reinorced concrete beam ( V no) plus the additional shear resistance provided by the steel ibers ( V ib ) as shown in Eq. 1: V n V V (1) no ib Eq. 2 is used to estimate the shear resistance provided by concrete and is the nominal shear strength 79

10 equation o the general shear design method o the CSA A Standard (CSA, 24): V no λ b d (2) c w v In the equation, b w and d v represent the beam width and eective shear depth, is a actor that takes into account the use o low-density concrete, and is a parameter that accounts or the ability o the concrete to transmit tensile stresses between the cracks. The angle is the angle o inclination o the diagonal compressive stresses to the longitudinal axis o the member. In the general design method both these parameters are a unction o the longitudinal strain at mid-depth o the cross-section, x, as shown in Eq. 3 and Eq. 4. The crack spacing parameter, s ze is a unction o the maximum aggregate size (3) x s ze 7 29 (4) x The longitudinal strain at mid-depth o the cross section, x is computed using Eq. 5 or the case o moment and shear: M d V v x (5) 2 Es As In the above expression the shear, V, corresponds to the expected shear resistance o the beam without ibers. The moment, M, represents the corresponding moment at the critical section in the beam (or the beams in this analysis the critical section is taken at a distance d v rom the loading point). A s and E s are the cross-sectional area and modulus o elasticity o the longitudinal tension steel, respectively. The second component o Eq. 1 accounts or the shear strength contribution o the ibers as they pullout across an inclined shear crack (see Figure 9) and is predicted using Eq. 6: V ib d 1.72 N ib.8fp bwd v cot (6) a The eective number o ibers per unit area, N ib, or ibers randomly oriented in three dimensions can be calculated using Eq. 7: v N ib l (7) A Where A is the cross-sectional area o the iber, and where v is the volume raction o ibers in the matrix. The orientation actor,, is used to account or the random orientation o the ibers crossing any arbitrary cracking plane and is taken as 3/8 (Foster, 21). The length actor l is used to account or the variability in the iber embedment length across the cracking plane and is taken as.5 (Aoude, 28). V d d v =.9 d Ө V d V a T Ө V cz F pullout Figure 9. Contribution o ibers to shear resistance with pullout across inclined shear crack Anchorage properties play an important role in improving the pullout resistance o steel ibers. In the case o hooked-end ibers Alwan et al.(1999) showed that the mechanical contribution o the hook is a unction o the cold work needed to straighten the iber as it is being pulled out rom the matrix. It was also demonstrated that this contribution is approximately independent o matrix type and iber embedment length. In the case o hooked-end ibers, the pullout strength can thereore be estimated by adding a hook contribution ( Phook) to the load needed to cause iber debonding as shown in Eq. 8: F p L bond d Phook (8) 2 In the above expression, the parameters d and L represent iber diameter and iber length (taken as the length o the straight portion o the iber). The bond-shear strength, bond, is a unction o the concrete matrix strength and is determined using Eq. 9: as 1.35 bond 1.11 ( ct ) (9) where ct is the tensile strength o concrete taken. Table 6 shows data or average ct '.33 c bond shear stress determined rom single iber pullout tests or a range o concrete matrix strengths as C 8

11 reported by Grünewald (24). Figure 1 shows that Eq. 9 provides a good it to the data in Table 6. Table 6. Range o bond shear strength values Matrix strength class Concrete strength range c (MPa) Bond shear stress bond (MPa) Normal strength Medium strength 5 & High strength The pullout strength in Eq. 8 and 11, F p, is valid or the situation o pure tensile pullout. However in a beam, the ibers will resist tension along the diagonal cracks while undergoing a shear deormation (see Figure 11). To account or reduced pullout eiciency in the situation o combined tension and shear, an empirically determined pullout modiication actor o.8 is applied to Eq. 6 to reduce the pullout resistance o the ibers. The actor was determined based on a regression analysis o a large database o SFRC beam test results (Aoude, 28) Bond shear stress, τ bond (MPa) τ bond = 1.11 ct 1.35 Data rom Table Concrete tensile strength, ct (MPa) Figure 1. Bond shear strength according to Eq. 9 For a typical hooked-end iber, the hook contribution to pullout can be estimated using Eq. 1: 3.5 P hook o o cos45 18 y π d (1) where y is the iber yield strength (in psi) and d is the iber diameter (in inches). For the case o crimped ibers the pullout strength is determined using Eq. 1, where the load needed to cause the debonding o a smooth iber is scaled using a orm actor,, that accounts or improved c pullout resistance in crimped ibers. In the current study the orm actor is estimated to be 2.8 based on regression analysis o data rom single iber pullout tests on crimped ibers (Cohen, 212): F p L c bond d (11) 2 Figure 11. Reduced pullout resistance in the case o combined tension and shear. Several researchers have studied the eect o shear span-to-depth ratio (a/d) on the behavior o SFRC beams. This eect is shown in Figure 12 where test data rom 5 studies shows that a decrease in a/d ratio leads to an increase in beam shear resistance when all other parameters are kept constant. The igure also shows that as the a/d ratio increases beyond 3. the eect o increasing the a/d ratio is present but less pronounced. In the proposed model the eect o a/d ratio is taken into account using the actor "1.72 d/a", determined based on regression analysis o a database o SFRC beam test results rom the literature (Cohen, 212). Normalized Stress v u / c ' Ashour et al.(1992) 1 Ashour et al.(1992) 2 Ashour et al.(1992) 3 Kwak et al.(22) 1 Kwak et al.(22) 2 Kwak et al.(22) 3 Lim et al.(1987) 1 Lim et al.(1987) 2 Lim et al.(1987) 3 Mansur et al.(1986) 1 Mansur et al.(1986) 2 Mansur et al.(1986) 3. Tan et al.(1993) Shear span-to- depth ratio (a/d) Figure 12. Eect o shear span-to-depth (a/d) ratio on shear strength in SFRC beams tested in the literature 6.2 Analysis results - Database o SFRC beams The procedure described in this paper was used to predict the shear capacities o a database o SFRC 81

12 beams tested by other researchers. The database includes results rom 129 SFRC beams having eective depths (d) ranging rom mm, shearspan to depth ratio (a/d) ranging rom , compressive strength ( ' c ) between MPa, longitudinal reinorcement ratios (ρ) ranging rom.4%- 5%, iber volume ractions (v ) ranging rom.25%- 2% (Cohen, 212). Figure 13 compares the experimental shear capacities o the beams to those predicted using the proposed model. The predictions using the proposed model agree reasonably well with the actual capacities or the majority o the beams with means o 1.11 and.94 with standard deviations o.2 and.21 or the case o beams having hooked-end and crimped ibers, respectively. Experimental Shear, V exp. (kn) H C Mean : S.D. :.2.21 Hooked (H) Crimped (C) Predicted Shear, V pred. (kn) Figure 1. Experimental vs. predicted shear capacities or SFRC beams having hooked-end and crimped ibers. 6.3 Analysis results - SFRC beams rom current study In this section the shear resistance o the SFRC specimens rom this experimental program that ailed in shear (Beams M15-.5%, M15-.5%H, M2-.75%, M2-1.%, M2-1.%A) are predicted using the proposed model. Various equations proposed in the literature were also considered in the analysis, including equations proposed by Khuntia et al. (1999), Imam et al. (1997), Ashour et al. (1992), Mansur et al. (1986) and Sharma (1986). Table 7 summarizes the results o the analysis; as can be clearly seen most o the SFRC models rom the literature show wide scatter. Similar observations have been made by other researchers when predicting the shear capacities o traditional SFRC beams (Minelli, 25). Table 7 shows that the proposed equation gives reasonably good predictions, however the shear strengths are generally over-predicted (overprediction is also observed or most o the other models considered in the analysis). As discussed previously, Lachemi et al.(25) ound that SCC beams can have reduced shear capacity when compared to traditional concrete beams due to reduced aggregate size and content, which is typical o SCC. Thus the irst term in Eq. 1 may be over-estimating the concrete contribution to shear resistance in SCC beams with reduced aggregate size and content. Further research is recommended to study the eect o these parameters in SCC and SFRC beams. Overall, the results demonstrate the need o developing reliable and accurate equations or predicting the shear capacity o SFRC beams constructed with SCC and steel ibers. Table 6. Range o bond shear strength values V exp Predictions ( V exp / V pred ) Beam I.D. KN Khuntia et al. (1999) Imam et al. (1995) Mansur et al. (1986) Sharma et al. (1986) Ashour et al. (1992) Proposed Model M15-.5% M15-.5%H M2-.75% M2-1.% M2-1.%A Mean: S.D.: SUMMARY AND CONCLUSIONS This paper presented the experimental results rom eleven slender constructed with SCC and highly workable SFRC tested under our-point loading. The ollowing conclusions are drawn rom this experimental program: (1) The addition o steel ibers to a SCC mix improves the shear capacity o reinorced concrete beams and results in improved crack control and damage tolerance; (2) When added in suicient quantity, provision o steel ibers to a SCC mix can alter the brittle shear ailure mode into a more ductile lexural mode in shear-critical beams; (3) The use o high-strength BP8/3 ibers was shown to improve behavior o the SFRC beams by improving ductility and enhancing crack control, although workability o the SFRC was adversely aected; (4) A procedure which modiies the general method o the CSA A was proposed or predicting the shear capacity o SFRC beams and was 82

13 shown to provide reasonably accurate predictions or a variety o beams tested by other researchers; (4) Existing equations in the literature as well the proposed model were used to predict the shear resistance o the SFRC beams constructed with SCC and steel ibers tested in this study and the analysis demonstrated that there is a need to develop accurate equations or predicting the shear strength o SFRC beams constructed with SCC and steel ibers; (5) Further research is recommended to examine the inluence o reduced aggregate size and content on shear capacity in SCC and SFRC beams. 8 REFERENCES ACI Committee 318, ACI Building code requirements or reinorced concrete and commentary", American Concrete Institute, USA (Farmington Hills), 211. Aoude, H., Structural behaviour o steel ibre reinorced concrete members, PhD thesis, Department o Civil Engineering and Applied Mechanics, McGill University, Canada (Montreal), 27. Ashour, S. A., Hasanain, G. S., and Waa, F. F., Shear Behavior o High-Strength Fiber Reinorced Concrete Beams, ACI Structural Journal, V. 89, No. 2, 1992, pp ASTM, Standard Test Method or Flexural Perormance o Fiber-Reinorced Concrete (ASTM C169/C 169M-7), American Society or Testing and Materials, USA (West Conchohocken), 27, 9 pp. CSA, Design o Concrete Structures (CSA A23.3-4), Canadian Standards Association, Canada (Mississauga), 24, 214 pp. Cohen, M., Structural behaviour o sel consolidating steel iber reinorced concrete beams, Master's thesis, Department o Civil Engineering, University o Ottawa, Canada (Ottawa), 212. Foster, S.J., On behavior o high-strength concrete columns: Cover spalling, steel ibers and ductility, ACI Structural Journal, V. 98, No. 4, 21, pp Greenough T., and Nehdi M., "Shear Behaviour o Fiber- Reinorced Sel-Consolidating Concrete Slender Beams", ACI Materials Journal, V. 15, No. 5, 28, pp Grünewald, S. Perormance-based design o sel-compacting ibre reinorced concrete, PhD thesis, Department o Structural and Building Engineering. Delt University o Technology, Netherlands (Delt), 24 Imam, M., Vandewalle, L., Mortelmans, F. and Van Gemert, D. "Shear Domain o Fibre Reinorced High Strength Concrete Beams" Journal o Engineering Structures, V. 19, No. 9, 1997, pp Khayat K.H., and Roussel Y., "Testing and Perormance o Fiber-Reinorced, Sel-Consolidating Concrete", Materials and Structures, V. 33, No. 6, 2, pp Khuntia, M., Stojadinovic, B. and Goel, S., Shear strength o normal-strength and high-strength ibre reinorced concrete beams without stirrups, ACI Structural Journal, V. 96, No. 2, 1999, pp Lachemi M., Hossain K., and Lambros V., "Shear Resistance o Sel-Consolidating Concrete Beams-Experimental Investigations", Canadian Journal o Civil Engineering, V. 32, No. 6, 25, pp Liao, W.C., Chao, S.H., Park, S.Y., and Naaman A.E., "Sel- Consolidating High Perormance Fiber Reinorced Concrete (SCHPFRC)", Report No. UMCEE 6-2, University o Michigan, Ann Arbor, MI., 26. Lin, C.H., and Chen J.H., "Shear Behaviour o Sel- Consolidating Concrete Beams", ACI Structural Journal, V. 19, No. 3, 212, pp Mansur, M.A., Ong, K.C.G. and Paramsivam, P., Shear strength o ibrous concrete beams without stirrups, Journal o Structural Engineering, V. 112, No. 9, 1986, pp Minelli, F. Plain and iber reinorced concrete beams under shear loading: Structural behavior and design applications, Ph.D. thesis, Department o Civil Engineering, University o Brescia, Starrylink Editrice, Italy (Brescia), 25, 43 pp. Parra-Montesinos, G. J., "Shear Strength o Beams with Deormed Steel Fibers". Concrete International, V. 28, No. 12, 26, pp Sharma, A.K., Shear strength o steel iber reinorced concrete beams, Journal o the American Concrete Institute, V. 83, No. 4, 1986, pp You. Z., Ding Y., and Niederegger C., "Replacing Stirrups o Sel-Compacting Concrete Beams with Steel Fibers". Transactions o Tianjin University, V. 16, No. 6, 21, pp