- Email: [email protected]
Compressive fatigue performance of fiber-reinforced lightweight concrete with high-volume supplementary cementitious materials
Cement and Concrete Composites 73 (2016) 89e97
Contents lists available at ScienceDirect
Cement and Concrete Composites journal homepage: www.elsevier.com/locate/cemconcomp
Compressive fatigue performance of fiber-reinforced lightweight concrete with high-volume supplementary cementitious materials Se-Jin Choi a, Jae-Sung Mun b, Keun-Hyeok Yang c, *, Si-Jun Kim c a
Department of Architectural Engineering, Wonkwang University, Iksan-si, Jeonbuk, 570-749, South Korea Department of Architectural Engineering, Graduate School, Kyonggi University, Suwon-si, Gyeonggi-do, 443-760, South Korea c Department of Plant Architectural Engineering, Kyonggi University, Suwon-si, Gyeonggi-do, 443-760, South Korea b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 3 June 2015 Received in revised form 29 May 2016 Accepted 9 July 2016 Available online 12 July 2016
As a part of the fundamental investigation to evaluate the compressive fatigue damage of fiber-reinforced lightweight concrete using high-volume supplementary cementitious materials (SCMs), the stressestrain response of the concrete was examined under cyclic loading within maximum stress levels of 0.9, 0.8, and 0.75 fc’ and a minimum stress level fixed at 0.1 fc’, where fc’ is the concrete compressive strength. To produce a high-volume SCM binder, 20% fly ash and 50% ground granulated blast-furnace slag were used as a partial replacement for cement. Amorphous steel (AS) and polyvinyl acetate (PVA) fibers were added individually or in combination considering their conventionally recommended volume fractions. The fatigue tests were carried out by controlling the load between two limits with a sinusoidal variation at a frequency of 1 Hz. Test results showed that PVA fiber was preferable to AS fiber in enhancing the fatigue life of HVS-LWC, whereas the fatigue damage of the PVA fiber concrete was lower than that of the AS fiber concrete. This trend was more notable under a higher maximum stress level. To improve the fatigue life and fatigue damage of concrete, hybridization of both fibers rather than monolithic use is recommended. © 2016 Elsevier Ltd. All rights reserved.
Keywords: Fatigue life Fatigue damage High-volume SCM Fiber Lightweight concrete Stressestrain curve
1. Introduction With increasing interest in the development of sustainable concrete technology, large volumetric reuse of by-products such as fly ash (FA) and ground granulated blast-furnace slag (GGBS) has been practically applied in the concrete industry as a high-volume replacement for ordinary Portland cement (OPC). Furthermore, incinerator bottom ash and FA are high-reuse value materials for producing lightweight aggregates [1,2]. This high-volume recycling of by-products significantly contributes to the reduction of environmental loads, including CO2 emissions, energy consumption, and depletion of natural resources in the concrete industry [3]. The pozzolanic activity of FA and GGBS is also effective for forming a denser matrix, which results in higher strength in terms of longterm aging and better durability of the concrete [4]. The lightweight aggregate concrete saves energy in buildings because of the enhanced thermal insulation capacity through the lower thermal
* Corresponding author. E-mail addresses: [email protected] (S.-J. Choi), [email protected] (J.-S. Mun), [email protected] (K.-H. Yang), [email protected] (S.-J. Kim). http://dx.doi.org/10.1016/j.cemconcomp.2016.07.007 0958-9465/© 2016 Elsevier Ltd. All rights reserved.
conductivity of lightweight aggregates. For these reasons, gradual growth is expected in the large volumetric reuse of by-products in concrete as supplementary cementitious materials (SCMs) and/or alternatives to natural aggregates. There are several structural advantages to using lightweight concrete (LWC) as a structural element. The use of LWC allows for the fabrication of smaller and lighter-weight structural members, which reduces the dead load and improves the seismic resistance capacity of structures. Furthermore, smaller and lighter elements of precast concrete members are preferred because of easier towing, greater buoyancy, and less expensive handling and transporting systems. As a result, the application of LWC to floating structures has gradually attracted great attention [5]. On the other hand, such floating structures are subject to fatigue by wave loading. It is commonly pointed out [6] that the fatigue resistance of LWC is somewhat lower than that of normal-weight concrete (NWC) because of lower aggregate strength and poorer cohesion between pastes and aggregate particles. Furthermore, the lower tensile strength and crack resistance of LWC lead to deterioration of the fatigue life of a concrete member. Hence, enhancing the fatigue strength of LWC is essential for the durable and stable design of
90
S.-J. Choi et al. / Cement and Concrete Composites 73 (2016) 89e97
structures subjected to endless loop loads such as wave and traffic loads. The most important improvements attained by the addition of fibers to concrete are an increase in toughness, ductility, tensile strength, and crack restriction. Owing to improved ductility and a decrease in the unstable crack propagation of fiber concrete, its dynamic properties, including fatigue response, can be enhanced. Hence, it is expected that the addition of fiber would be one of the best alternatives for enhancing the crack resistance and fatigue life of LWC. Most of the fatigue studies [7e10] on fiber-reinforced concrete have mainly focused on determining the flexural fatigue endurance limits for concrete using different types, volume fractions, and aspect ratios of fibers. Jun and Stang [11] reported that the accumulated damage level in fiber-reinforced concrete under fatigue loading was 1e2 orders of magnitude higher than the level recorded in static testing of the same material. Cachim et al. [12] conducted fatigue tests on steel fiber-reinforced concrete in compression and concluded that the superior fatigue life of fiberreinforced concrete as compared to plain concrete depends on the initial imperfections such as microcracks or voids. Thus, the presence of fibers, especially larger fibers, may be an additional cause of imperfections, creating bridges between the aggregates and increasing initial residual stress. Moreover, some fibers remain glued, creating a large fiber ball that accelerates the imperfection and bridge-formation problem. Paskova and Meyer [13] showed that the fatigue life of concrete reinforced with 0.75% steel fiber is smaller than that of concrete with 0.25% steel fiber. Overall, special attention should be taken when the type, aspect ratio, and content of fiber are selected in order to ensure the targeted fatigue life of concrete. However, an understanding on the fatigue performance of fiber-reinforced concrete is very limited and still controversial, especially for LWC. Additionally, reliable test data are needed to for the development of accurate models capable of predicting the fatigue behavior of fiber-reinforced LWC. The objective of the present study is to examine the feasibility of enhancing the fatigue performance of high volume supplementary cementitious material (HVS)-LWC in compression by using different fibers. To produce a high-volume SCM binder, 20% FA and 50% GGBS were used as a partial replacement for OPC. Amorphous steel (AS) fibers and polyvinyl acetate (PVA) fibers were added individually or in combination in order to improve the crack and tensile resistance of concrete. Based on the measured compressive stressestrain curves at each loading cycle, the effect of the added fibers on the deformation capacity of HVS-LWC under fatigue loading was evaluated by comparing the fatigue and residual strain values and the fatigue damage responses of concrete specimens. The fatigue life of fiber-reinforced HVS-LWC was also examined under different maximum stress levels.
2. Experimental program
and a silicon oxide (SiO2)-to-aluminum oxide (Al2O3) ratio by mass of 1.91, as given in Table 1. The loss of ignition (LOI) and 28-day activity coefficient of FA were 0.29% and 89%, respectively. The main chemical composition of GGBS was CaO, SiO2, and Al2O3. The basicity of GGBS calculated from the chemical composition was 2.00. The specific gravity and specific surface area for OPC were 3.15 and 3466 cm2/g, respectively, 2.23 and 3720 cm2/g, respectively, for FA, and 2.91 and 4497 cm2/g, respectively, for GGBS. Artificially expanded clay granules having maximum sizes of 19 mm and 4 mm were used for lightweight coarse and fine aggregates, respectively. The lightweight aggregate particles were spherical, having a closed surface with a slightly rough texture. The core of the particle had a uniformly fine and porous structure, which led to a high water absorption capacity and low strength. The apparent density and water absorption for coarse particles were 1.21 g/cm3 and 18.96%, respectively, and 1.65 g/cm3 and 13.68%, respectively, for fine particles. The apparent density was thus lower in the 19 mm coarse aggregates than in the 4 mm fine aggregates, whereas the water absorption was higher in the coarse aggregates than in the fine aggregates, as given in Table 2. The fineness modulus of the lightweight fine aggregates was slightly higher than the common value (2.0e3.0) of natural sand. The PVA fiber selected had a hydrophilic nature owing to the presence of eOH groups, which led to better dispersion within fresh concrete and improved interfacial bond strength with the cementitious matrix as compared with other synthetic polymers. It is commonly known [15] that PVA fibers provide a positive effect on the crack resistance and tensile strength of concrete or cementitious composites through their good interfacial bond with the cement matrix, high tensile strength, and good affinity with fresh concrete. In contrast, the low lateral stiffness of PVA fibers may lead to premature fiber rupture before being pulled out of the cementitious matrix. A graphical illustration of the monofilament PVA fibers used is shown in Fig. 1(a). The nominal length and diameter of the PVA fiber were 6 mm and 0.015 mm, respectively, producing an aspect ratio of 400, as given in Table 3. The nominal tensile strength and modulus of elasticity of the PVA fiber used were 1269 MPa and 27640 MPa, respectively. AS fiber is a new material that has superior strength and toughness and enhanced corrosion resistance against acidic and aqueous environments as compared with conventional steel fibers. Unlike crystalline metal, amorphous metal forms random and irregular arrays of atoms [16]. Thus, amorphous metal has a densely filled structure, showing liquid characteristics and high flexibility in the solid state, as shown in Fig 1(b). The AS fiber used in this study was in the shape of a very thin sheet with dimensions of 1.6 0.029 30 mm3 (width thickness length), producing an aspect ratio of 18.75. The AS fiber had a smooth surface and no anchorage device at either end. The tensile strength and modulus of elasticity of the AS fiber were 1700 MPa and 140000 MPa, respectively, values that are higher than those of PVA fibers. The AS fiber had lower modulus of elasticity than the conventional steel fiber.
2.1. Materials The prepared cementitious materials were OPC conforming to ASTM Type I, FA belonging to Class F of ASTM C618, and GGBS conforming to ASTM C989 [14]. The FA had low calcium oxide (CaO)
2.2. Specimens and mixture proportions The main parameters selected for the fiber-reinforced HVS-LWC
Table 1 Chemical composition of the cementitious materials (% by mass). Materials
SiO2
Al2O3
Fe2O3
CaO
MgO
K2O
Na2O
TiO2
SO3
LOI
OPC FA GGBS
22.1 53.3 31.55
5.0 27.9 13.79
3.0 7.8 0.53
64.8 6.79 44.38
1.6 1.11 5.2
0.54 0.84 0.4
0.35 0.55 0.18
0.30 e 0.98
2.0 0.82 2.79
0.31 0.29 0.2
S.-J. Choi et al. / Cement and Concrete Composites 73 (2016) 89e97
91
Table 2 Physical properties of lightweight aggregates used. Maximum size (mm)
Unit volume weight (kg/m3)
Apparent density (g/cm3)
Water absorption (%)
Porosity (%)
Fineness modulus
19 4
729 832
1.21 1.65
18.96 13.68
68.17 50.42
6.56 4.34
Fig. 1. Photographs of fibers used.
Table 3 Physical and mechanical properties of fibers used. Type
PVA AS
Physical properties
Mechanical properties
rf (g/cm3)
Lf (mm)
Cross section
Af
Ff (MPa)
Ef (MPa)
1.3 7.2
6 30
Circular section with df ¼ 0.015 mm Rectangular section with bf ¼ 1.6 mm and tf ¼ 0.029 mm
400 18.75
1269 1700
27640 140000
Note. rf , Lf , Af , Ff , and Ef ¼ density, length, aspect ratio, tensile strength, and elastic modulus of fibers, respectively, df ¼ diameter of PVA fiber with circular section, and bf and tf ¼ width and thickness of AS fiber with rectangular section, respectively.
mixtures were the type and volume fraction (Vf ) of the fibers. The PVA and AS fibers were added either monolithically or by hybridization. The volume fractions of the fibers were determined with reference to the commercially recommended amount needed to maintain a certain level of slump for concrete casting and economic efficiency. The conventionally recommended volume fraction (Vf r ) of the AS and PVA fibers per unit weight of concrete are 0.5% and 0.07%, respectively, which results in a unit content of 36 kg/m3 for AS fiber and 0.9 kg/m3 for PVA fiber. Hence, the concrete mixtures were identified using the type and Vf of fiber added as follows: specimens A-R and A-1.5R denote concrete mixtures reinforced with monolithic AS fiber of 1.0 Vf r (0.5% corresponds to 36 kg/m3) and 1.5 Vf r (0.75% corresponds to 54 kg/m3), respectively; specimens P-R and P-1.5R denote concrete mixtures reinforced with monolithic PVA fiber of 1.0 Vf r (0.07% corresponds to 0.9 kg/m3) and 1.5 Vf r (0.105% corresponds to 1.35 kg/m3), respectively; specimens AP-0.5R and AP-0.75R denote concrete mixtures reinforced with a hybrid of AS and PVA fibers at 0.5 Vf r each (incorporation of 0.25% AS fiber and 0.035% PVA fiber) and 0.75 Vf r each
(incorporation of 0.375% AS fiber and 0.053% PVA fiber), respectively. A control fiberless mixture (N) was also prepared. The targeted air content, initial slump, and 91-day compressive strength of the control fiberless concrete were 2.5 ± 0.5%, 200 ± 25 mm, and 30 MPa, respectively. To achieve the designed requirements, a water-to-cementitious material ratio (W=C) of 25% by weight and a fine aggregate-to-total aggregate ratio of 42% by volume were selected for all concrete mixtures. The unit water content was fixed at 155 kg/m3 for all mixtures. A polycarboxylatebased high-range water-reducing agent was added in the amount of 0.5% relative to the unit content of the cementitious materials. The parameter selected to evaluate the fatigue performance of the prepared HVS-LWC mixtures was the constant maximum stress level (Smax ), which varied as 75%, 80%, and 90% of the static uniaxial compressive strength. The constant minimum stress level (Smin ) of all concrete mixtures was fixed at 10% of the static strength, regardless of the maximum stress level. During the preliminary tests, concrete specimens under Smax of 70% static strength tended to converge toward the fatigue limit. Thus, 75% of the static
92
S.-J. Choi et al. / Cement and Concrete Composites 73 (2016) 89e97
concrete compressive strength was selected as a lower bound for the Smax value. 2.3. Casting, curing and testing It is very difficult to maintain the saturated surface dry state (SSD) of lightweight aggregates in the field owing to their high water absorption [2]. In the present ready-mixed commercial mixtures, air-dried (AD) lightweight aggregates, achieved by drying the aggregates in the shade for 24 h after soaking, were used. The water required for the SSD condition was accounted for in the mixture-proportioning procedure. The cementitious materials and aggregates were dry-mixed in a mixer pan for 1 min; water was then added and mixing continued for another minute. PVA fibers previously soaked in mixing water were added to the concrete at the time of water addition. AS fibers were added after wet-mixing of concrete, which was then mixed for another minute for good dispersion of the fibers. Immediately after casting, all specimens were cured at a constant temperature and relative humidity of 23 ± 2 C and 70± 5%, respectively. They were tested after aging for 91 days. All steel molds were removed after 1 day. Moisture conditions during curing and testing were kept constant in all concrete cylinders to minimize the influence of moisture content on the fatigue strength of the concrete. The slump of each fresh concrete mixture was measured immediately after the completion of mixing. Compressive tests were conducted using cylindrical specimens with 100 mm in diameter and 200 mm in height. Immediately before the cyclic tests, monotonic uniaxial tests in compression were carried out to determine the static compressive strength of the concrete. The testing age was selected to be 91 days to minimize the effect of drying shrinkage on the fatigue behavior of the concrete and to investigate the strength of high-volume SCM concrete after longterm aging. The fatigue tests were carried out with a load control between two limits (with a sinusoidal force variation with time at a frequency of 1 Hz). Static and fatigue tests in compression were conducted using a 3000 kN capacity closed-loop servo-hydraulic dynamic testing system. The axial strain at the 100-mm mid-height of cylinders was recorded using two 5-mm-capacity linear variable differential transformers (LVDTs) and two electrical resistance strain (ERS) gauges with effective gauge length of 75 mm. To minimize the eccentricity of the applied load, the loading procedure was repeatedly adjusted within 30% of the axial load capacity of each specimen until the axial strains were in close agreement across all four measurement locations. Thus, the concentration of the applied loads was ascertained from the axial strains recorded at four locations. Note that the average values of strains obtained from axial displacements of two LVDTs were used for the fatigue stressestrain curves. All test data were captured by a dynamic data logger and automatically stored. The test stopped after the specimen failed to maintain the specified maximum stress level. 3. Test results and discussion The mixing of concrete A-1.5R failed because of the segregation and ball-bearing phenomena of AS fibers during the mixing. Thus, cylinder molds for compressive testing were not produced for the A-1.5R mixture. The average values, determined from three cylindrical specimens for each concrete mix, of compressive strength, splitting tensile strength, modulus of elasticity (Ec), and strain at the peak stress are listed in Table 4. The modulus of elasticity was calculated as the slope of the secant line joining zero stress and 40% peak stress, in accordance with ASTM C469 [14]. The control fiberless concrete mix achieved the targeted requirements. In general, the static compressive strength (fc’ ) of the prepared
concrete mixtures was relatively unaffected by the type and volume fraction of fibers added. Meanwhile, the slump of the fresh concrete tended to decrease with increasing volume fraction and length of the fibers. The AS fiber-reinforced concrete had a higher tensile strength and modulus of elasticity than the PVA fiberreinforced concrete. 3.1. Crack propagation and failure under cyclic loadings Initial cracks developed at the mid-height of the specimens and then propagated toward the top and bottom loading points. With the increase in the number (N) of cycles in the cyclic loading, a few vertical macroscopic cracks developed, which generated a wide crack band zone. The rate of crack propagation and the crack band growth tended to be slower in the fiber-reinforced concrete specimens than in the control fiberless concrete. The addition of fiber to concrete restricts crack formation and delays crack growth along the interface between the cementitious matrix and the aggregate particles [16]. Therefore, it is commonly recognized that the unstable growth of cracks during cyclic loading is transformed into slow and controlled growth, depending greatly upon fiber dispersion, the elasto plastic properties of the cementitious matrix, and the effectiveness of the cementitious matrix-lightweight aggregate bond. On the other hand, the developed macrocracks tended to penetrate the lightweight aggregate particles, as shown in Fig. 2. At the failure of the concrete, pull-out of the AS fibers occurred along the macrocracks (Fig. 3), whereas the PVA fibers ruptured before being pulled out of the cementitious matrix (Fig. 3). As a result, during cyclic loading, failure of the fiber-reinforced HVS-LWC primarily occurred by a combination of fracture of the aggregate and progressive pullout or rupture of fibers along the cracks. This implies that the contribution of fibers to the crack restriction would be weakened for LWC as compared with the case of NWC because of weakened aggregate interlocking and macrofiber bridging actions, as shown in Fig. 4. Overall, microfibers are more effective in restricting bond crack propagation along the interface between the cementitious matrix and lightweight aggregates than macrofibers such as AS fibers. Also, crack propagation was more rapid and resulted in more cracks with increasing N for AS fiber-reinforced concrete than for PVA fiber-reinforced concrete. The longitudinal crack propagation in aggregate particles can be more effectively controlled by longer fibers than by shorter fibers, as shown in Fig. 4. The longitudinal cracks in the hybrid fiber-reinforced HVS-LWC propagated more slowly than those in the monolithic fiberreinforced concrete. 3.2. Relationship between maximum stress level and fatigue life Fig. 5 shows a negative linear relationship between Smax and the logarithm of Nf (commonly designated as the SeN curve) measured from each HVS-LWC mixture, where Nf is the fatigue life of the concrete. For the same Smax , Nf of the fiber-reinforced HVS-LWC was generally greater than that of the control fiberless concrete, as given in Table 5. The effect of the fibers on the increase in Nf can be summarized as follows: 1) The PVA fibers were more effective than the AS fibers; 2) Hybridization of both fibers was preferred to monolithic use; and 3) These observations were more notable for higher Smax values. The increasing ratios of Nf of the fibercontaining concrete relative to that of fiberless concrete were 1.37, 2.32, and 8.53 for concrete mixtures A-R, P-R, and AP-0.5R, respectively, when Smax was 0.9, whereas those ratios were 1.39, 1.45, and 3.85, respectively, under Smax of 0.75, as given in Table 5. In terms of restricting and slowing microcrack growth, microfibers are preferable to macrofibers with a larger diameter owing to the better dispersion characteristics of microfibers at the interface
S.-J. Choi et al. / Cement and Concrete Composites 73 (2016) 89e97
93
Table 4 Workability and mechanical properties of concrete mixtures under monotonic tests. Concrete mixture
N A-R P-R P-1.5R AP-0.5R AP-0.75R
Fresh concrete
Mechanical properties of hardened concrete
Air content (%)
Slump (mm)
fsp (MPa)
fc’ (MPa)
Ec (MPa)
ε0
2.6 2.5 2 1.2 1 1.5
206 140 205 200 170 170
3.37 7.54 4.38 5.99 7.28 7.66
35.29 33.14 31.56 32.46 35.42 33.29
17000 26000 21000 22000 22000 28000
0.00266 0.00375 0.00288 0.00301 0.00324 0.00315
1
Smax
0.9 0.8 N P-R AP-0.5R
0.7
A-R P-1.5R AP-0.75R
0.6 100
101
102 103 log (Nf)
104
105
106
Fig. 5. Effect of fibers on Smax e Nf relationship of HVS-LWC in compression. Fig. 2. Typical crack profiles along failure plane of fibreless LWC (specimen N).
Fig. 3. Typical resistance of fibers at the crack planes.
Fig. 4. Local stresses and fiber bridging zones around the lightweight aggregate particles.
zone [7,17]. Furthermore, the presence of macrofibers increases the number of pores and initial flaws in the cementitious matrix [18], such that increasing the macrofiber content can reduce the fatigue resistance of the cementitious matrix. However, at the aggregatebridging zone, a longer fiber is more effective in restricting longitudinal crack propagation in aggregate particles than a shorter fiber. This implies that the stress transfer by the fiber-bridging action along the macrocrack formed at the interface zone between the cementitious matrix and the aggregates is more significant in LWC than in NWC because of the fracture of lightweight aggregate particles. For these reasons, PVA fiber was preferable to AS fiber and hybridization of both fibers was preferable to monolithic use of such fibers in enhancing the composite performance under cyclic loading. To examine the effect of the various parameters (including the amount, aspect ratio, interfacial bond strength, and orientation of the added fiber) on the tensile resistance capacity of fiberreinforced concrete, a fiber-reinforcing index
0.704 0.910 0.837 0.866 0.871 0.921 0.685 0.887 0.811 0.828 0.858 0.881 0.789 1.757 1.701 1.698 1.590 1.879 1.639 2.808 2.201 2.528 2.827 2.990 1.635 2.643 2.118 2.329 2.698 2.698 1.628 2.307 1.931 1.821 2.438 2.168 0.00210 0.00659 0.00490 0.00511 0.00515 0.00592 0.00209 0.00518 0.00390 0.00421 0.00496 0.00423 0.00150 0.00390 0.00257 0.00261 0.00466 0.00423 0.00436 0.01053 0.00634 0.00761 0.00659 0.00942 0.00435 0.00991 0.00610 0.00701 0.00874 0.00850 0.00433 0.00865 0.00556 0.00548 0.00790 0.00683 22522 31358 32800 44670 86780 110676 2121 3100 3580 4950 10221 14160 43 59 100 198 367 655
0.8
0.564 1.040 0.892 0.867 1.438 1.343
0.786 1.381 1.354 1.399 1.531 1.679
0.651 0.862 0.750 0.751 0.825 0.827
0.8 0.9 0.75
Smax Smax Smax
0.9
0.8
0.75
0.9
0.8
N A-R P-R P-1.5R AP-0.5R AP-0.75R
Specimens
Table 5 Summary of fatigue test results.
0.75
0.8
0.75
0.9
0.8 Smax Smax
0.9
Smax
0.9
0.75
Normalized maximum residual strain, ðεr;Nf =ε0 Þ Maximum residual strain at failure, ðεr;Nf Þ Maximum fatigue strain at failure, ðεf ;Nf Þ Fatigue life, (Nf )
Normalized maximum fatigue strain, ðεf ;Nf =ε0 Þ
0.75
S.-J. Choi et al. / Cement and Concrete Composites 73 (2016) 89e97
Fatigue damage at failure, ðDf Þ
94
P (bf ¼ ni¼1 gi ti Vf ;i Af ;i =t0 ) is frequently introduced [19], where gi , ti , and Af ;i are the efficiency factor, interfacial bond strength, and aspect ratio of fiber i, respectively, and t0 (¼ 1 MPa) is the reference bond strength. For the present concrete mixtures, the volume fraction of the AS fiber was higher than that of the PVA fiber, whereas the aspect ratio of the former was smaller than that of the latter, as given in Table 3. As a result, a higher value of bf was obtained for the PVA fiber-reinforced concrete mixtures than for the AS fiber-reinforced concrete mixtures. The bf value of the hybrid fibers was lower than that of the monolithic PVA fiber, although hybridization of AS and PVA fibers was preferable to monolithic use of the PVA fiber in enhancing the fatigue life of the concrete. This indicates that the hybridization benefit of microfibers and macrofibers needs be considered in introducing bf on the SmaxeNf curve of the concrete. 3.3. Fatigue stressestrain curves Typical fatigue stressestrain curves measured from the prepared concrete specimens are shown in Fig. 6. For comparison, monotonic stressestrain curves of those concrete mixtures are plotted in the same figure. The stress is normalized by fc’. The modulus of elasticity (Ec ) of HVS-LWC tended to increase with the addition of the AS fiber, indicating that specimens A-R and AP0.75R had a higher value of Ec than either the fiberless concrete or the PVA-reinforced concrete, as given in Table 4. Furthermore, the strain (ε0 ) at the peak stress of concrete A-R was greater than either that of the fiberless concrete or the PVA-reinforced concrete. This may be attributed to the gradual pullout of AS fibers at the approximation of peak stress. The fatigue strains at Smax and residual strains at Smin gradually increased with increasing N, and thus the stressestrain curves at each loading cycle moved toward the right. The increased fatigue strains intersected the descending branch of the monotonic stressestrain curve and the concrete specimen quickly failed. With increasing N, the fatigue and residual strains increased more rapidly for the AS fiber-reinforced concrete than for the PVA fiber-reinforced concrete. As a result, a greater enclosed area of the curve for each cycle was observed for the AS fiber-reinforced concrete than for the PVA fiber-reinforced concrete. This observation was more notable for concrete with higher Smax values. The concrete with the hybrid fibers had a slower decrease of the descending branch of the monotonic stressestrain curve. This implies that the enhanced ductility of concrete owing to the addition of fibers results in an increase in fatigue life. The incremental fatigue strain (εf ;i ) at Smax after a certain number of cycles under cyclic loading is plotted in Fig. 7. The fatigue strains in each loading cycle are normalized by the ε0 value of each specimen measured under monotonic compression, and the number of loading cycles is normalized by Nf . The schematic pattern of εf ;i =ε0 typically followed the three stages [20]: an initial cyclic creep stage, which rapidly increased owing to flaw initiation; a creepefatigue coupling stage, consisting of slow and progressive growth of the inherent flaws; and a fatigue stage, accompanied by a sufficient number of unstable cracks. The fatigue stage was commonly observed during the final 95% of the fatigue life. The value of εf ;i =ε0 of the fiber-reinforced concrete was higher than that of the fiberless concrete, although the fiberless concrete had the smallest value of ε0 . The difference in the εf ;i =ε0 values according to the addition of fiber was more notable under Smax of 0.75 than under Smax of 0.9. At the same level of N=Nf , the εf ;i =ε0 values of the A-R concrete was commonly higher than those of the other mixtures, although the A-R concrete had the greatest value of ε0 . Along the macrocracks, AS fibers were gradually pulled out from the cementitious matrix; they did not show brittle rupture because of the smooth surface and weakened anchorage capacity of the fiber.
S.-J. Choi et al. / Cement and Concrete Composites 73 (2016) 89e97
95
Fig. 7. Normalized incremental fatigue strain according to number of loading cycles.
Fig. 8. Typical normalized incremental residual strain according to number of loading cycles (Smax ¼ 0.9).
Fig. 6. Fatigue stressestrain curves of the prepared concrete specimens.
Because of this pullout of AS fibers, the AS fiber-reinforced concrete had greater values of ε0 and εf ;i than the other concrete mixtures. The hybrid fiber-reinforced concrete showed slightly greater εf ;i =ε0 values than the PVA-reinforced concrete, regardless of the Smax value. The normalized incremental residual strain (εr;i =ε0 ) at Smin after a certain number of loading cycles is plotted in Fig. 8. The behavior of εr;i =ε0 was very close to that observed for εf ;i =ε0, showing a
similar range of N=Nf during the creepefatigue coupling stage. The effect of the test parameters on the increasing rate of εr;i =ε0 was similar to the trend observed in Fig. 7. This implies that the stresses at reloading and unloading of each loading cycle varied almost linearly up to the axial strain of 0.004, particularly for the PVA fiberreinforced concrete and the hybrid fiber-reinforced concrete, which resulted in the same stiffness at the reloading and unloading branches, as shown in Fig. 6. As compared with the fatigue stressestrain curves obtained from fibreless concrete or AS fiberreinforced concrete, PVA or hybrid fiber-reinforced concrete had steeper slopes at reloading and unloading branches.
3.4. Fatigue damage In fatigue tests with constant stress ranges, the degradation of fatigue modulus (Ef ;i ) is directly associated with the incremental εf ;i
96
S.-J. Choi et al. / Cement and Concrete Composites 73 (2016) 89e97
value. Hence, the fatigue damage (D) in compression is defined as follows [20]:
. D ¼ 1 Ef ;i Ec ¼ 1
"
Smax Ec $εf ;i
# (1)
Fig. 9 shows the typical fatigue damage response of the concrete specimens after a certain number of cycles under cyclic loading. The fatigue damage at failure (Df ) of the concrete specimens is also given in Table 5. The value of D tended to increase with increasing N. The value of D was lower for the fiberless concrete than for the fiber-reinforced concrete mixtures. At the same value of N, the AS fiber-reinforced concrete had a greater D value than the PVA fiberreinforced concrete or hybrid fiber-reinforced concrete. With the increase in Smax , the value of Df decreased and the variation of Df values according to the addition of fibers was more notable. Fig. 10 shows the relationship between Df and Nf measured for each HVS-LWC mixture. The value of Df increased in proportion to the logarithm of Nf . The hybrid fiber-reinforced concrete had higher Df values than the monolithic PVA fiber-reinforced concrete, regardless of the Smax level. Meanwhile, AS fiber-reinforced concrete had higher Df values than the other concrete mixtures, showing that the difference tended to decrease with decreasing Smax level. The gradual pullout of AS fibers along the macrocracks led to the greater incremental εf ;i value, which then resulted in a greater Df value than the other concrete mixtures. Overall, these observations reveal that the effect of fibers on the fatigue damage of concrete is directly related to the restriction of crack propagation and pullout characteristics of fibers along cracks. 4. Conclusions To evaluate the fatigue life and fatigue damage of fiberreinforced high-volume SCM lightweight concrete (HVS-LWC) subjected to uni-axial compression, the stressestrain response was examined under cyclic loading with constant stress ranges. The proposed conclusions given below may be somewhat tentative owing to the numerous influences on the fatigue performance of fiber-reinforced lightweight concrete. Furthermore, the available fatigue test data on fiber-reinforced LWC are very limited because the test procedure is difficult owing to long run-times and uncommon mixing of such concrete. Hence, the effect of fibers on the fatigue response of HVS-LWC needs to be further examined, with a focus on additional parameters such as fiber properties and loading conditions, including load frequency and stress range. From the present elementary investigation, the following conclusions were drawn:
Fig. 10. Effect of fibers on Df e Nf relationship of HVS-LWC in compression.
(1) At failure, amorphous steel (AS) fibers were pulled out along the macrocracks, whereas polyvinyl acetate (PVA) fibers ruptured. Thus, the failure of fiber-reinforced HVS-LWC primarily occurred by the combination of fracture of the aggregate and progressive pullout or rupture of fibers. (2) With increasing loading cycle, a greater number of macrocracks developed more rapidly for HVS-LWC reinforced by AS fiber than for that by PVA fiber. (3) For enhancing the fatigue life of HVS-LWC, PVA fiber was more effective than AS fiber and a hybridization of both fibers was preferred to monolithic fibers. These observations were more notable in concrete mixtures under higher Smax values. (4) The fatigue and residual strains increased more rapidly for the AS fiber-reinforced concrete than for the PVA fiberreinforced concrete, which resulted in a greater enclosed area under the curve for each cycle in the AS fiber-reinforced concrete than the PVA fiber-reinforced concrete. (5) The hybrid fiber-reinforced concrete had higher fatigue damage values than the monolithic PVA fiber-reinforced concrete, regardless of the Smax level. Meanwhile, the AS fiber-reinforced concrete had higher Df values than the other concrete mixtures, showing that the difference tended to decrease with decreasing Smax level. Acknowledgement This research was supported by the Public Welfare & Safety Research Program through the National Research Foundation of Korea (NRF) (2013067519) funded by the Ministry of Science, ICT and Future Planning (2013067519). This work was also supported by Wonkwang University in 2015. Notation
Fig. 9. Typical fatigue damage response of concrete specimens according to the loading cycles (Smax ¼ 0.9).
Af ;i bf df D Df Ec Ef Ef ;i Ff fc’ fsp ft
aspect ratio of fiber i width of AS fiber with rectangular section diameter of PVA fiber with circular section fatigue damage fatigue damage at failure modulus of elasticity of concrete elastic modulus of fibers degradation of fatigue modulus tensile strength of fibers compressive strength of concrete splitting tensile strength of concrete tensile strength of concrete
S.-J. Choi et al. / Cement and Concrete Composites 73 (2016) 89e97
gi Lf N Nf Smax Smin tf Vf Vf r W=C
bf
ε0 εf ;i
rf t0 ti
efficiency factor of fiber i length of fibers the number of cycles in cyclic loadings the fatigue life of concrete maximum stress level minimum stress level thickness of AS fiber with rectangular section volume fraction of fibers conventionally recommended volume fraction of the AS and PVA fibers water-to-cementitious material ratio fiber-reinforcing index strain at the peak stress of concrete incremental fatigue strain density of fibers reference bond strength (¼ 1 MPa) interfacial bond strength of fiber i
References [1] S. Chandra, L. Berntsson, Lightweight Aggregate Concrete: Science, Technology and Applications, Noyes Publications, USA, 2003. [2] ACI 213R-03, Guide for Structural Lightweight Aggregate Concrete, ACI Committee 213, USA, 2003. [3] K. Sakai, T. Noguchi, The Sustainable Use of Concrete, CRC Press, NY, 2012. [4] A.M. Neville, Properties of Concrete, Longman, England, 1995. [5] S.J. Shin, The Development of Bio/ceramic Membrane Pontoon for Marine Purification, 2014, 1st Technical Report, Korea. [6] R. Tepfers, T. Kutti, Fatigue strength of plain, ordinary, and lightweight
97
concrete, ACI J. Proc. 76 (5) (1979) 635e652. [7] G. Batson, C. Ball, L. Bailey, E. Landers, J. Hooks, Flexural fatigue strength of steel fibre reinforced concrete beams, ACI J. 69 (11) (1972) 673e677. [8] C.D. Johnston, R.W. Zemp, Flexural fatigue performance of steel fibre reinforced concrete-influence of fibre content, aspect ratio, and type, ACI Mater. J. 88 (4) (1991) 374e383. [9] X.P. Shi, T.F. Fwa, S.A. Tan, Flexural fatigue strength of plain concrete, ACI Mater. J. 90 (5) (1993) 435e440. sl, Fatigue parameters of cement-based [10] S. Seitl, Z. Kersner, V. Bílek, Z. Kne composites with various types of fibers, Key Eng. Mater 417e418 (2010) 129e132. [11] Z. Jun, H. Stang, Fatigue performance in flexural of fibre reinforced concrete, ACI Mater. J. 95 (1) (1998) 58e67. [12] P.B. Cachim, J.A. Figueiras, P.A.A. Pereira, Fatigue behavior of fiber-reinforced concrete in compression, Cem. Concr. Compos. 24 (2002) 211e217. [13] T. Paskova, C. Meyer, Low-cycle fatigue of plain and fiber-reinforced concrete, ACI Mater. J. 94 (4) (1997) 273e285. [14] ASTM, C618, C989, C469, Annual Book of ASTM Standards, ASTM International, West Conshohocken (PA), 2012. [15] A. Noushini, B. Samali, K. Vessalas, Effect of polyvinyl alcohol (PVA) fibre on dynamic and material properties of fibre reinforced concrete, Constr. Build. Mater. 49 (2013) 374e383. [16] J.P. Won, B.T. Hong, S.J. Lee, S.J. Choi, Bonding properties of amorphous microsteel fibre-reinforced cementitious composites, Compos. Struct. 102 (2013) 101e109. [17] M.K. Lee, B.I.G. Barr, An overview of the fatigue behavior of plain and fibre reinforced concrete, Cem. Concr. Compos. 26 (4) (2004) 299e305. [18] J.J. Beaudoin, Handbook of Fiber Reinforced Concrete: Principles, Properties, Developments and Applications, Noyes Publications, Park Ridge, NJ, 1990. [19] K.H. Yang, Tests on concrete reinforced with hybrid or monolithic steel and polyvinyl alcohol fibers, ACI Mater. J. 108 (6) (2011) 664e672. [20] J. Xiao, H. Li, Z. Yang, Fatigue behavior of recycled aggregate concrete under compression and bending cyclic loading, Constr. Build. Mater. 38 (2013) 681e688.
Contents lists available at ScienceDirect
Cement and Concrete Composites journal homepage: www.elsevier.com/locate/cemconcomp
Compressive fatigue performance of fiber-reinforced lightweight concrete with high-volume supplementary cementitious materials Se-Jin Choi a, Jae-Sung Mun b, Keun-Hyeok Yang c, *, Si-Jun Kim c a
Department of Architectural Engineering, Wonkwang University, Iksan-si, Jeonbuk, 570-749, South Korea Department of Architectural Engineering, Graduate School, Kyonggi University, Suwon-si, Gyeonggi-do, 443-760, South Korea c Department of Plant Architectural Engineering, Kyonggi University, Suwon-si, Gyeonggi-do, 443-760, South Korea b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 3 June 2015 Received in revised form 29 May 2016 Accepted 9 July 2016 Available online 12 July 2016
As a part of the fundamental investigation to evaluate the compressive fatigue damage of fiber-reinforced lightweight concrete using high-volume supplementary cementitious materials (SCMs), the stressestrain response of the concrete was examined under cyclic loading within maximum stress levels of 0.9, 0.8, and 0.75 fc’ and a minimum stress level fixed at 0.1 fc’, where fc’ is the concrete compressive strength. To produce a high-volume SCM binder, 20% fly ash and 50% ground granulated blast-furnace slag were used as a partial replacement for cement. Amorphous steel (AS) and polyvinyl acetate (PVA) fibers were added individually or in combination considering their conventionally recommended volume fractions. The fatigue tests were carried out by controlling the load between two limits with a sinusoidal variation at a frequency of 1 Hz. Test results showed that PVA fiber was preferable to AS fiber in enhancing the fatigue life of HVS-LWC, whereas the fatigue damage of the PVA fiber concrete was lower than that of the AS fiber concrete. This trend was more notable under a higher maximum stress level. To improve the fatigue life and fatigue damage of concrete, hybridization of both fibers rather than monolithic use is recommended. © 2016 Elsevier Ltd. All rights reserved.
Keywords: Fatigue life Fatigue damage High-volume SCM Fiber Lightweight concrete Stressestrain curve
1. Introduction With increasing interest in the development of sustainable concrete technology, large volumetric reuse of by-products such as fly ash (FA) and ground granulated blast-furnace slag (GGBS) has been practically applied in the concrete industry as a high-volume replacement for ordinary Portland cement (OPC). Furthermore, incinerator bottom ash and FA are high-reuse value materials for producing lightweight aggregates [1,2]. This high-volume recycling of by-products significantly contributes to the reduction of environmental loads, including CO2 emissions, energy consumption, and depletion of natural resources in the concrete industry [3]. The pozzolanic activity of FA and GGBS is also effective for forming a denser matrix, which results in higher strength in terms of longterm aging and better durability of the concrete [4]. The lightweight aggregate concrete saves energy in buildings because of the enhanced thermal insulation capacity through the lower thermal
* Corresponding author. E-mail addresses: [email protected] (S.-J. Choi), [email protected] (J.-S. Mun), [email protected] (K.-H. Yang), [email protected] (S.-J. Kim). http://dx.doi.org/10.1016/j.cemconcomp.2016.07.007 0958-9465/© 2016 Elsevier Ltd. All rights reserved.
conductivity of lightweight aggregates. For these reasons, gradual growth is expected in the large volumetric reuse of by-products in concrete as supplementary cementitious materials (SCMs) and/or alternatives to natural aggregates. There are several structural advantages to using lightweight concrete (LWC) as a structural element. The use of LWC allows for the fabrication of smaller and lighter-weight structural members, which reduces the dead load and improves the seismic resistance capacity of structures. Furthermore, smaller and lighter elements of precast concrete members are preferred because of easier towing, greater buoyancy, and less expensive handling and transporting systems. As a result, the application of LWC to floating structures has gradually attracted great attention [5]. On the other hand, such floating structures are subject to fatigue by wave loading. It is commonly pointed out [6] that the fatigue resistance of LWC is somewhat lower than that of normal-weight concrete (NWC) because of lower aggregate strength and poorer cohesion between pastes and aggregate particles. Furthermore, the lower tensile strength and crack resistance of LWC lead to deterioration of the fatigue life of a concrete member. Hence, enhancing the fatigue strength of LWC is essential for the durable and stable design of
90
S.-J. Choi et al. / Cement and Concrete Composites 73 (2016) 89e97
structures subjected to endless loop loads such as wave and traffic loads. The most important improvements attained by the addition of fibers to concrete are an increase in toughness, ductility, tensile strength, and crack restriction. Owing to improved ductility and a decrease in the unstable crack propagation of fiber concrete, its dynamic properties, including fatigue response, can be enhanced. Hence, it is expected that the addition of fiber would be one of the best alternatives for enhancing the crack resistance and fatigue life of LWC. Most of the fatigue studies [7e10] on fiber-reinforced concrete have mainly focused on determining the flexural fatigue endurance limits for concrete using different types, volume fractions, and aspect ratios of fibers. Jun and Stang [11] reported that the accumulated damage level in fiber-reinforced concrete under fatigue loading was 1e2 orders of magnitude higher than the level recorded in static testing of the same material. Cachim et al. [12] conducted fatigue tests on steel fiber-reinforced concrete in compression and concluded that the superior fatigue life of fiberreinforced concrete as compared to plain concrete depends on the initial imperfections such as microcracks or voids. Thus, the presence of fibers, especially larger fibers, may be an additional cause of imperfections, creating bridges between the aggregates and increasing initial residual stress. Moreover, some fibers remain glued, creating a large fiber ball that accelerates the imperfection and bridge-formation problem. Paskova and Meyer [13] showed that the fatigue life of concrete reinforced with 0.75% steel fiber is smaller than that of concrete with 0.25% steel fiber. Overall, special attention should be taken when the type, aspect ratio, and content of fiber are selected in order to ensure the targeted fatigue life of concrete. However, an understanding on the fatigue performance of fiber-reinforced concrete is very limited and still controversial, especially for LWC. Additionally, reliable test data are needed to for the development of accurate models capable of predicting the fatigue behavior of fiber-reinforced LWC. The objective of the present study is to examine the feasibility of enhancing the fatigue performance of high volume supplementary cementitious material (HVS)-LWC in compression by using different fibers. To produce a high-volume SCM binder, 20% FA and 50% GGBS were used as a partial replacement for OPC. Amorphous steel (AS) fibers and polyvinyl acetate (PVA) fibers were added individually or in combination in order to improve the crack and tensile resistance of concrete. Based on the measured compressive stressestrain curves at each loading cycle, the effect of the added fibers on the deformation capacity of HVS-LWC under fatigue loading was evaluated by comparing the fatigue and residual strain values and the fatigue damage responses of concrete specimens. The fatigue life of fiber-reinforced HVS-LWC was also examined under different maximum stress levels.
2. Experimental program
and a silicon oxide (SiO2)-to-aluminum oxide (Al2O3) ratio by mass of 1.91, as given in Table 1. The loss of ignition (LOI) and 28-day activity coefficient of FA were 0.29% and 89%, respectively. The main chemical composition of GGBS was CaO, SiO2, and Al2O3. The basicity of GGBS calculated from the chemical composition was 2.00. The specific gravity and specific surface area for OPC were 3.15 and 3466 cm2/g, respectively, 2.23 and 3720 cm2/g, respectively, for FA, and 2.91 and 4497 cm2/g, respectively, for GGBS. Artificially expanded clay granules having maximum sizes of 19 mm and 4 mm were used for lightweight coarse and fine aggregates, respectively. The lightweight aggregate particles were spherical, having a closed surface with a slightly rough texture. The core of the particle had a uniformly fine and porous structure, which led to a high water absorption capacity and low strength. The apparent density and water absorption for coarse particles were 1.21 g/cm3 and 18.96%, respectively, and 1.65 g/cm3 and 13.68%, respectively, for fine particles. The apparent density was thus lower in the 19 mm coarse aggregates than in the 4 mm fine aggregates, whereas the water absorption was higher in the coarse aggregates than in the fine aggregates, as given in Table 2. The fineness modulus of the lightweight fine aggregates was slightly higher than the common value (2.0e3.0) of natural sand. The PVA fiber selected had a hydrophilic nature owing to the presence of eOH groups, which led to better dispersion within fresh concrete and improved interfacial bond strength with the cementitious matrix as compared with other synthetic polymers. It is commonly known [15] that PVA fibers provide a positive effect on the crack resistance and tensile strength of concrete or cementitious composites through their good interfacial bond with the cement matrix, high tensile strength, and good affinity with fresh concrete. In contrast, the low lateral stiffness of PVA fibers may lead to premature fiber rupture before being pulled out of the cementitious matrix. A graphical illustration of the monofilament PVA fibers used is shown in Fig. 1(a). The nominal length and diameter of the PVA fiber were 6 mm and 0.015 mm, respectively, producing an aspect ratio of 400, as given in Table 3. The nominal tensile strength and modulus of elasticity of the PVA fiber used were 1269 MPa and 27640 MPa, respectively. AS fiber is a new material that has superior strength and toughness and enhanced corrosion resistance against acidic and aqueous environments as compared with conventional steel fibers. Unlike crystalline metal, amorphous metal forms random and irregular arrays of atoms [16]. Thus, amorphous metal has a densely filled structure, showing liquid characteristics and high flexibility in the solid state, as shown in Fig 1(b). The AS fiber used in this study was in the shape of a very thin sheet with dimensions of 1.6 0.029 30 mm3 (width thickness length), producing an aspect ratio of 18.75. The AS fiber had a smooth surface and no anchorage device at either end. The tensile strength and modulus of elasticity of the AS fiber were 1700 MPa and 140000 MPa, respectively, values that are higher than those of PVA fibers. The AS fiber had lower modulus of elasticity than the conventional steel fiber.
2.1. Materials The prepared cementitious materials were OPC conforming to ASTM Type I, FA belonging to Class F of ASTM C618, and GGBS conforming to ASTM C989 [14]. The FA had low calcium oxide (CaO)
2.2. Specimens and mixture proportions The main parameters selected for the fiber-reinforced HVS-LWC
Table 1 Chemical composition of the cementitious materials (% by mass). Materials
SiO2
Al2O3
Fe2O3
CaO
MgO
K2O
Na2O
TiO2
SO3
LOI
OPC FA GGBS
22.1 53.3 31.55
5.0 27.9 13.79
3.0 7.8 0.53
64.8 6.79 44.38
1.6 1.11 5.2
0.54 0.84 0.4
0.35 0.55 0.18
0.30 e 0.98
2.0 0.82 2.79
0.31 0.29 0.2
S.-J. Choi et al. / Cement and Concrete Composites 73 (2016) 89e97
91
Table 2 Physical properties of lightweight aggregates used. Maximum size (mm)
Unit volume weight (kg/m3)
Apparent density (g/cm3)
Water absorption (%)
Porosity (%)
Fineness modulus
19 4
729 832
1.21 1.65
18.96 13.68
68.17 50.42
6.56 4.34
Fig. 1. Photographs of fibers used.
Table 3 Physical and mechanical properties of fibers used. Type
PVA AS
Physical properties
Mechanical properties
rf (g/cm3)
Lf (mm)
Cross section
Af
Ff (MPa)
Ef (MPa)
1.3 7.2
6 30
Circular section with df ¼ 0.015 mm Rectangular section with bf ¼ 1.6 mm and tf ¼ 0.029 mm
400 18.75
1269 1700
27640 140000
Note. rf , Lf , Af , Ff , and Ef ¼ density, length, aspect ratio, tensile strength, and elastic modulus of fibers, respectively, df ¼ diameter of PVA fiber with circular section, and bf and tf ¼ width and thickness of AS fiber with rectangular section, respectively.
mixtures were the type and volume fraction (Vf ) of the fibers. The PVA and AS fibers were added either monolithically or by hybridization. The volume fractions of the fibers were determined with reference to the commercially recommended amount needed to maintain a certain level of slump for concrete casting and economic efficiency. The conventionally recommended volume fraction (Vf r ) of the AS and PVA fibers per unit weight of concrete are 0.5% and 0.07%, respectively, which results in a unit content of 36 kg/m3 for AS fiber and 0.9 kg/m3 for PVA fiber. Hence, the concrete mixtures were identified using the type and Vf of fiber added as follows: specimens A-R and A-1.5R denote concrete mixtures reinforced with monolithic AS fiber of 1.0 Vf r (0.5% corresponds to 36 kg/m3) and 1.5 Vf r (0.75% corresponds to 54 kg/m3), respectively; specimens P-R and P-1.5R denote concrete mixtures reinforced with monolithic PVA fiber of 1.0 Vf r (0.07% corresponds to 0.9 kg/m3) and 1.5 Vf r (0.105% corresponds to 1.35 kg/m3), respectively; specimens AP-0.5R and AP-0.75R denote concrete mixtures reinforced with a hybrid of AS and PVA fibers at 0.5 Vf r each (incorporation of 0.25% AS fiber and 0.035% PVA fiber) and 0.75 Vf r each
(incorporation of 0.375% AS fiber and 0.053% PVA fiber), respectively. A control fiberless mixture (N) was also prepared. The targeted air content, initial slump, and 91-day compressive strength of the control fiberless concrete were 2.5 ± 0.5%, 200 ± 25 mm, and 30 MPa, respectively. To achieve the designed requirements, a water-to-cementitious material ratio (W=C) of 25% by weight and a fine aggregate-to-total aggregate ratio of 42% by volume were selected for all concrete mixtures. The unit water content was fixed at 155 kg/m3 for all mixtures. A polycarboxylatebased high-range water-reducing agent was added in the amount of 0.5% relative to the unit content of the cementitious materials. The parameter selected to evaluate the fatigue performance of the prepared HVS-LWC mixtures was the constant maximum stress level (Smax ), which varied as 75%, 80%, and 90% of the static uniaxial compressive strength. The constant minimum stress level (Smin ) of all concrete mixtures was fixed at 10% of the static strength, regardless of the maximum stress level. During the preliminary tests, concrete specimens under Smax of 70% static strength tended to converge toward the fatigue limit. Thus, 75% of the static
92
S.-J. Choi et al. / Cement and Concrete Composites 73 (2016) 89e97
concrete compressive strength was selected as a lower bound for the Smax value. 2.3. Casting, curing and testing It is very difficult to maintain the saturated surface dry state (SSD) of lightweight aggregates in the field owing to their high water absorption [2]. In the present ready-mixed commercial mixtures, air-dried (AD) lightweight aggregates, achieved by drying the aggregates in the shade for 24 h after soaking, were used. The water required for the SSD condition was accounted for in the mixture-proportioning procedure. The cementitious materials and aggregates were dry-mixed in a mixer pan for 1 min; water was then added and mixing continued for another minute. PVA fibers previously soaked in mixing water were added to the concrete at the time of water addition. AS fibers were added after wet-mixing of concrete, which was then mixed for another minute for good dispersion of the fibers. Immediately after casting, all specimens were cured at a constant temperature and relative humidity of 23 ± 2 C and 70± 5%, respectively. They were tested after aging for 91 days. All steel molds were removed after 1 day. Moisture conditions during curing and testing were kept constant in all concrete cylinders to minimize the influence of moisture content on the fatigue strength of the concrete. The slump of each fresh concrete mixture was measured immediately after the completion of mixing. Compressive tests were conducted using cylindrical specimens with 100 mm in diameter and 200 mm in height. Immediately before the cyclic tests, monotonic uniaxial tests in compression were carried out to determine the static compressive strength of the concrete. The testing age was selected to be 91 days to minimize the effect of drying shrinkage on the fatigue behavior of the concrete and to investigate the strength of high-volume SCM concrete after longterm aging. The fatigue tests were carried out with a load control between two limits (with a sinusoidal force variation with time at a frequency of 1 Hz). Static and fatigue tests in compression were conducted using a 3000 kN capacity closed-loop servo-hydraulic dynamic testing system. The axial strain at the 100-mm mid-height of cylinders was recorded using two 5-mm-capacity linear variable differential transformers (LVDTs) and two electrical resistance strain (ERS) gauges with effective gauge length of 75 mm. To minimize the eccentricity of the applied load, the loading procedure was repeatedly adjusted within 30% of the axial load capacity of each specimen until the axial strains were in close agreement across all four measurement locations. Thus, the concentration of the applied loads was ascertained from the axial strains recorded at four locations. Note that the average values of strains obtained from axial displacements of two LVDTs were used for the fatigue stressestrain curves. All test data were captured by a dynamic data logger and automatically stored. The test stopped after the specimen failed to maintain the specified maximum stress level. 3. Test results and discussion The mixing of concrete A-1.5R failed because of the segregation and ball-bearing phenomena of AS fibers during the mixing. Thus, cylinder molds for compressive testing were not produced for the A-1.5R mixture. The average values, determined from three cylindrical specimens for each concrete mix, of compressive strength, splitting tensile strength, modulus of elasticity (Ec), and strain at the peak stress are listed in Table 4. The modulus of elasticity was calculated as the slope of the secant line joining zero stress and 40% peak stress, in accordance with ASTM C469 [14]. The control fiberless concrete mix achieved the targeted requirements. In general, the static compressive strength (fc’ ) of the prepared
concrete mixtures was relatively unaffected by the type and volume fraction of fibers added. Meanwhile, the slump of the fresh concrete tended to decrease with increasing volume fraction and length of the fibers. The AS fiber-reinforced concrete had a higher tensile strength and modulus of elasticity than the PVA fiberreinforced concrete. 3.1. Crack propagation and failure under cyclic loadings Initial cracks developed at the mid-height of the specimens and then propagated toward the top and bottom loading points. With the increase in the number (N) of cycles in the cyclic loading, a few vertical macroscopic cracks developed, which generated a wide crack band zone. The rate of crack propagation and the crack band growth tended to be slower in the fiber-reinforced concrete specimens than in the control fiberless concrete. The addition of fiber to concrete restricts crack formation and delays crack growth along the interface between the cementitious matrix and the aggregate particles [16]. Therefore, it is commonly recognized that the unstable growth of cracks during cyclic loading is transformed into slow and controlled growth, depending greatly upon fiber dispersion, the elasto plastic properties of the cementitious matrix, and the effectiveness of the cementitious matrix-lightweight aggregate bond. On the other hand, the developed macrocracks tended to penetrate the lightweight aggregate particles, as shown in Fig. 2. At the failure of the concrete, pull-out of the AS fibers occurred along the macrocracks (Fig. 3), whereas the PVA fibers ruptured before being pulled out of the cementitious matrix (Fig. 3). As a result, during cyclic loading, failure of the fiber-reinforced HVS-LWC primarily occurred by a combination of fracture of the aggregate and progressive pullout or rupture of fibers along the cracks. This implies that the contribution of fibers to the crack restriction would be weakened for LWC as compared with the case of NWC because of weakened aggregate interlocking and macrofiber bridging actions, as shown in Fig. 4. Overall, microfibers are more effective in restricting bond crack propagation along the interface between the cementitious matrix and lightweight aggregates than macrofibers such as AS fibers. Also, crack propagation was more rapid and resulted in more cracks with increasing N for AS fiber-reinforced concrete than for PVA fiber-reinforced concrete. The longitudinal crack propagation in aggregate particles can be more effectively controlled by longer fibers than by shorter fibers, as shown in Fig. 4. The longitudinal cracks in the hybrid fiber-reinforced HVS-LWC propagated more slowly than those in the monolithic fiberreinforced concrete. 3.2. Relationship between maximum stress level and fatigue life Fig. 5 shows a negative linear relationship between Smax and the logarithm of Nf (commonly designated as the SeN curve) measured from each HVS-LWC mixture, where Nf is the fatigue life of the concrete. For the same Smax , Nf of the fiber-reinforced HVS-LWC was generally greater than that of the control fiberless concrete, as given in Table 5. The effect of the fibers on the increase in Nf can be summarized as follows: 1) The PVA fibers were more effective than the AS fibers; 2) Hybridization of both fibers was preferred to monolithic use; and 3) These observations were more notable for higher Smax values. The increasing ratios of Nf of the fibercontaining concrete relative to that of fiberless concrete were 1.37, 2.32, and 8.53 for concrete mixtures A-R, P-R, and AP-0.5R, respectively, when Smax was 0.9, whereas those ratios were 1.39, 1.45, and 3.85, respectively, under Smax of 0.75, as given in Table 5. In terms of restricting and slowing microcrack growth, microfibers are preferable to macrofibers with a larger diameter owing to the better dispersion characteristics of microfibers at the interface
S.-J. Choi et al. / Cement and Concrete Composites 73 (2016) 89e97
93
Table 4 Workability and mechanical properties of concrete mixtures under monotonic tests. Concrete mixture
N A-R P-R P-1.5R AP-0.5R AP-0.75R
Fresh concrete
Mechanical properties of hardened concrete
Air content (%)
Slump (mm)
fsp (MPa)
fc’ (MPa)
Ec (MPa)
ε0
2.6 2.5 2 1.2 1 1.5
206 140 205 200 170 170
3.37 7.54 4.38 5.99 7.28 7.66
35.29 33.14 31.56 32.46 35.42 33.29
17000 26000 21000 22000 22000 28000
0.00266 0.00375 0.00288 0.00301 0.00324 0.00315
1
Smax
0.9 0.8 N P-R AP-0.5R
0.7
A-R P-1.5R AP-0.75R
0.6 100
101
102 103 log (Nf)
104
105
106
Fig. 5. Effect of fibers on Smax e Nf relationship of HVS-LWC in compression. Fig. 2. Typical crack profiles along failure plane of fibreless LWC (specimen N).
Fig. 3. Typical resistance of fibers at the crack planes.
Fig. 4. Local stresses and fiber bridging zones around the lightweight aggregate particles.
zone [7,17]. Furthermore, the presence of macrofibers increases the number of pores and initial flaws in the cementitious matrix [18], such that increasing the macrofiber content can reduce the fatigue resistance of the cementitious matrix. However, at the aggregatebridging zone, a longer fiber is more effective in restricting longitudinal crack propagation in aggregate particles than a shorter fiber. This implies that the stress transfer by the fiber-bridging action along the macrocrack formed at the interface zone between the cementitious matrix and the aggregates is more significant in LWC than in NWC because of the fracture of lightweight aggregate particles. For these reasons, PVA fiber was preferable to AS fiber and hybridization of both fibers was preferable to monolithic use of such fibers in enhancing the composite performance under cyclic loading. To examine the effect of the various parameters (including the amount, aspect ratio, interfacial bond strength, and orientation of the added fiber) on the tensile resistance capacity of fiberreinforced concrete, a fiber-reinforcing index
0.704 0.910 0.837 0.866 0.871 0.921 0.685 0.887 0.811 0.828 0.858 0.881 0.789 1.757 1.701 1.698 1.590 1.879 1.639 2.808 2.201 2.528 2.827 2.990 1.635 2.643 2.118 2.329 2.698 2.698 1.628 2.307 1.931 1.821 2.438 2.168 0.00210 0.00659 0.00490 0.00511 0.00515 0.00592 0.00209 0.00518 0.00390 0.00421 0.00496 0.00423 0.00150 0.00390 0.00257 0.00261 0.00466 0.00423 0.00436 0.01053 0.00634 0.00761 0.00659 0.00942 0.00435 0.00991 0.00610 0.00701 0.00874 0.00850 0.00433 0.00865 0.00556 0.00548 0.00790 0.00683 22522 31358 32800 44670 86780 110676 2121 3100 3580 4950 10221 14160 43 59 100 198 367 655
0.8
0.564 1.040 0.892 0.867 1.438 1.343
0.786 1.381 1.354 1.399 1.531 1.679
0.651 0.862 0.750 0.751 0.825 0.827
0.8 0.9 0.75
Smax Smax Smax
0.9
0.8
0.75
0.9
0.8
N A-R P-R P-1.5R AP-0.5R AP-0.75R
Specimens
Table 5 Summary of fatigue test results.
0.75
0.8
0.75
0.9
0.8 Smax Smax
0.9
Smax
0.9
0.75
Normalized maximum residual strain, ðεr;Nf =ε0 Þ Maximum residual strain at failure, ðεr;Nf Þ Maximum fatigue strain at failure, ðεf ;Nf Þ Fatigue life, (Nf )
Normalized maximum fatigue strain, ðεf ;Nf =ε0 Þ
0.75
S.-J. Choi et al. / Cement and Concrete Composites 73 (2016) 89e97
Fatigue damage at failure, ðDf Þ
94
P (bf ¼ ni¼1 gi ti Vf ;i Af ;i =t0 ) is frequently introduced [19], where gi , ti , and Af ;i are the efficiency factor, interfacial bond strength, and aspect ratio of fiber i, respectively, and t0 (¼ 1 MPa) is the reference bond strength. For the present concrete mixtures, the volume fraction of the AS fiber was higher than that of the PVA fiber, whereas the aspect ratio of the former was smaller than that of the latter, as given in Table 3. As a result, a higher value of bf was obtained for the PVA fiber-reinforced concrete mixtures than for the AS fiber-reinforced concrete mixtures. The bf value of the hybrid fibers was lower than that of the monolithic PVA fiber, although hybridization of AS and PVA fibers was preferable to monolithic use of the PVA fiber in enhancing the fatigue life of the concrete. This indicates that the hybridization benefit of microfibers and macrofibers needs be considered in introducing bf on the SmaxeNf curve of the concrete. 3.3. Fatigue stressestrain curves Typical fatigue stressestrain curves measured from the prepared concrete specimens are shown in Fig. 6. For comparison, monotonic stressestrain curves of those concrete mixtures are plotted in the same figure. The stress is normalized by fc’. The modulus of elasticity (Ec ) of HVS-LWC tended to increase with the addition of the AS fiber, indicating that specimens A-R and AP0.75R had a higher value of Ec than either the fiberless concrete or the PVA-reinforced concrete, as given in Table 4. Furthermore, the strain (ε0 ) at the peak stress of concrete A-R was greater than either that of the fiberless concrete or the PVA-reinforced concrete. This may be attributed to the gradual pullout of AS fibers at the approximation of peak stress. The fatigue strains at Smax and residual strains at Smin gradually increased with increasing N, and thus the stressestrain curves at each loading cycle moved toward the right. The increased fatigue strains intersected the descending branch of the monotonic stressestrain curve and the concrete specimen quickly failed. With increasing N, the fatigue and residual strains increased more rapidly for the AS fiber-reinforced concrete than for the PVA fiber-reinforced concrete. As a result, a greater enclosed area of the curve for each cycle was observed for the AS fiber-reinforced concrete than for the PVA fiber-reinforced concrete. This observation was more notable for concrete with higher Smax values. The concrete with the hybrid fibers had a slower decrease of the descending branch of the monotonic stressestrain curve. This implies that the enhanced ductility of concrete owing to the addition of fibers results in an increase in fatigue life. The incremental fatigue strain (εf ;i ) at Smax after a certain number of cycles under cyclic loading is plotted in Fig. 7. The fatigue strains in each loading cycle are normalized by the ε0 value of each specimen measured under monotonic compression, and the number of loading cycles is normalized by Nf . The schematic pattern of εf ;i =ε0 typically followed the three stages [20]: an initial cyclic creep stage, which rapidly increased owing to flaw initiation; a creepefatigue coupling stage, consisting of slow and progressive growth of the inherent flaws; and a fatigue stage, accompanied by a sufficient number of unstable cracks. The fatigue stage was commonly observed during the final 95% of the fatigue life. The value of εf ;i =ε0 of the fiber-reinforced concrete was higher than that of the fiberless concrete, although the fiberless concrete had the smallest value of ε0 . The difference in the εf ;i =ε0 values according to the addition of fiber was more notable under Smax of 0.75 than under Smax of 0.9. At the same level of N=Nf , the εf ;i =ε0 values of the A-R concrete was commonly higher than those of the other mixtures, although the A-R concrete had the greatest value of ε0 . Along the macrocracks, AS fibers were gradually pulled out from the cementitious matrix; they did not show brittle rupture because of the smooth surface and weakened anchorage capacity of the fiber.
S.-J. Choi et al. / Cement and Concrete Composites 73 (2016) 89e97
95
Fig. 7. Normalized incremental fatigue strain according to number of loading cycles.
Fig. 8. Typical normalized incremental residual strain according to number of loading cycles (Smax ¼ 0.9).
Fig. 6. Fatigue stressestrain curves of the prepared concrete specimens.
Because of this pullout of AS fibers, the AS fiber-reinforced concrete had greater values of ε0 and εf ;i than the other concrete mixtures. The hybrid fiber-reinforced concrete showed slightly greater εf ;i =ε0 values than the PVA-reinforced concrete, regardless of the Smax value. The normalized incremental residual strain (εr;i =ε0 ) at Smin after a certain number of loading cycles is plotted in Fig. 8. The behavior of εr;i =ε0 was very close to that observed for εf ;i =ε0, showing a
similar range of N=Nf during the creepefatigue coupling stage. The effect of the test parameters on the increasing rate of εr;i =ε0 was similar to the trend observed in Fig. 7. This implies that the stresses at reloading and unloading of each loading cycle varied almost linearly up to the axial strain of 0.004, particularly for the PVA fiberreinforced concrete and the hybrid fiber-reinforced concrete, which resulted in the same stiffness at the reloading and unloading branches, as shown in Fig. 6. As compared with the fatigue stressestrain curves obtained from fibreless concrete or AS fiberreinforced concrete, PVA or hybrid fiber-reinforced concrete had steeper slopes at reloading and unloading branches.
3.4. Fatigue damage In fatigue tests with constant stress ranges, the degradation of fatigue modulus (Ef ;i ) is directly associated with the incremental εf ;i
96
S.-J. Choi et al. / Cement and Concrete Composites 73 (2016) 89e97
value. Hence, the fatigue damage (D) in compression is defined as follows [20]:
. D ¼ 1 Ef ;i Ec ¼ 1
"
Smax Ec $εf ;i
# (1)
Fig. 9 shows the typical fatigue damage response of the concrete specimens after a certain number of cycles under cyclic loading. The fatigue damage at failure (Df ) of the concrete specimens is also given in Table 5. The value of D tended to increase with increasing N. The value of D was lower for the fiberless concrete than for the fiber-reinforced concrete mixtures. At the same value of N, the AS fiber-reinforced concrete had a greater D value than the PVA fiberreinforced concrete or hybrid fiber-reinforced concrete. With the increase in Smax , the value of Df decreased and the variation of Df values according to the addition of fibers was more notable. Fig. 10 shows the relationship between Df and Nf measured for each HVS-LWC mixture. The value of Df increased in proportion to the logarithm of Nf . The hybrid fiber-reinforced concrete had higher Df values than the monolithic PVA fiber-reinforced concrete, regardless of the Smax level. Meanwhile, AS fiber-reinforced concrete had higher Df values than the other concrete mixtures, showing that the difference tended to decrease with decreasing Smax level. The gradual pullout of AS fibers along the macrocracks led to the greater incremental εf ;i value, which then resulted in a greater Df value than the other concrete mixtures. Overall, these observations reveal that the effect of fibers on the fatigue damage of concrete is directly related to the restriction of crack propagation and pullout characteristics of fibers along cracks. 4. Conclusions To evaluate the fatigue life and fatigue damage of fiberreinforced high-volume SCM lightweight concrete (HVS-LWC) subjected to uni-axial compression, the stressestrain response was examined under cyclic loading with constant stress ranges. The proposed conclusions given below may be somewhat tentative owing to the numerous influences on the fatigue performance of fiber-reinforced lightweight concrete. Furthermore, the available fatigue test data on fiber-reinforced LWC are very limited because the test procedure is difficult owing to long run-times and uncommon mixing of such concrete. Hence, the effect of fibers on the fatigue response of HVS-LWC needs to be further examined, with a focus on additional parameters such as fiber properties and loading conditions, including load frequency and stress range. From the present elementary investigation, the following conclusions were drawn:
Fig. 10. Effect of fibers on Df e Nf relationship of HVS-LWC in compression.
(1) At failure, amorphous steel (AS) fibers were pulled out along the macrocracks, whereas polyvinyl acetate (PVA) fibers ruptured. Thus, the failure of fiber-reinforced HVS-LWC primarily occurred by the combination of fracture of the aggregate and progressive pullout or rupture of fibers. (2) With increasing loading cycle, a greater number of macrocracks developed more rapidly for HVS-LWC reinforced by AS fiber than for that by PVA fiber. (3) For enhancing the fatigue life of HVS-LWC, PVA fiber was more effective than AS fiber and a hybridization of both fibers was preferred to monolithic fibers. These observations were more notable in concrete mixtures under higher Smax values. (4) The fatigue and residual strains increased more rapidly for the AS fiber-reinforced concrete than for the PVA fiberreinforced concrete, which resulted in a greater enclosed area under the curve for each cycle in the AS fiber-reinforced concrete than the PVA fiber-reinforced concrete. (5) The hybrid fiber-reinforced concrete had higher fatigue damage values than the monolithic PVA fiber-reinforced concrete, regardless of the Smax level. Meanwhile, the AS fiber-reinforced concrete had higher Df values than the other concrete mixtures, showing that the difference tended to decrease with decreasing Smax level. Acknowledgement This research was supported by the Public Welfare & Safety Research Program through the National Research Foundation of Korea (NRF) (2013067519) funded by the Ministry of Science, ICT and Future Planning (2013067519). This work was also supported by Wonkwang University in 2015. Notation
Fig. 9. Typical fatigue damage response of concrete specimens according to the loading cycles (Smax ¼ 0.9).
Af ;i bf df D Df Ec Ef Ef ;i Ff fc’ fsp ft
aspect ratio of fiber i width of AS fiber with rectangular section diameter of PVA fiber with circular section fatigue damage fatigue damage at failure modulus of elasticity of concrete elastic modulus of fibers degradation of fatigue modulus tensile strength of fibers compressive strength of concrete splitting tensile strength of concrete tensile strength of concrete
S.-J. Choi et al. / Cement and Concrete Composites 73 (2016) 89e97
gi Lf N Nf Smax Smin tf Vf Vf r W=C
bf
ε0 εf ;i
rf t0 ti
efficiency factor of fiber i length of fibers the number of cycles in cyclic loadings the fatigue life of concrete maximum stress level minimum stress level thickness of AS fiber with rectangular section volume fraction of fibers conventionally recommended volume fraction of the AS and PVA fibers water-to-cementitious material ratio fiber-reinforcing index strain at the peak stress of concrete incremental fatigue strain density of fibers reference bond strength (¼ 1 MPa) interfacial bond strength of fiber i
References [1] S. Chandra, L. Berntsson, Lightweight Aggregate Concrete: Science, Technology and Applications, Noyes Publications, USA, 2003. [2] ACI 213R-03, Guide for Structural Lightweight Aggregate Concrete, ACI Committee 213, USA, 2003. [3] K. Sakai, T. Noguchi, The Sustainable Use of Concrete, CRC Press, NY, 2012. [4] A.M. Neville, Properties of Concrete, Longman, England, 1995. [5] S.J. Shin, The Development of Bio/ceramic Membrane Pontoon for Marine Purification, 2014, 1st Technical Report, Korea. [6] R. Tepfers, T. Kutti, Fatigue strength of plain, ordinary, and lightweight
97
concrete, ACI J. Proc. 76 (5) (1979) 635e652. [7] G. Batson, C. Ball, L. Bailey, E. Landers, J. Hooks, Flexural fatigue strength of steel fibre reinforced concrete beams, ACI J. 69 (11) (1972) 673e677. [8] C.D. Johnston, R.W. Zemp, Flexural fatigue performance of steel fibre reinforced concrete-influence of fibre content, aspect ratio, and type, ACI Mater. J. 88 (4) (1991) 374e383. [9] X.P. Shi, T.F. Fwa, S.A. Tan, Flexural fatigue strength of plain concrete, ACI Mater. J. 90 (5) (1993) 435e440. sl, Fatigue parameters of cement-based [10] S. Seitl, Z. Kersner, V. Bílek, Z. Kne composites with various types of fibers, Key Eng. Mater 417e418 (2010) 129e132. [11] Z. Jun, H. Stang, Fatigue performance in flexural of fibre reinforced concrete, ACI Mater. J. 95 (1) (1998) 58e67. [12] P.B. Cachim, J.A. Figueiras, P.A.A. Pereira, Fatigue behavior of fiber-reinforced concrete in compression, Cem. Concr. Compos. 24 (2002) 211e217. [13] T. Paskova, C. Meyer, Low-cycle fatigue of plain and fiber-reinforced concrete, ACI Mater. J. 94 (4) (1997) 273e285. [14] ASTM, C618, C989, C469, Annual Book of ASTM Standards, ASTM International, West Conshohocken (PA), 2012. [15] A. Noushini, B. Samali, K. Vessalas, Effect of polyvinyl alcohol (PVA) fibre on dynamic and material properties of fibre reinforced concrete, Constr. Build. Mater. 49 (2013) 374e383. [16] J.P. Won, B.T. Hong, S.J. Lee, S.J. Choi, Bonding properties of amorphous microsteel fibre-reinforced cementitious composites, Compos. Struct. 102 (2013) 101e109. [17] M.K. Lee, B.I.G. Barr, An overview of the fatigue behavior of plain and fibre reinforced concrete, Cem. Concr. Compos. 26 (4) (2004) 299e305. [18] J.J. Beaudoin, Handbook of Fiber Reinforced Concrete: Principles, Properties, Developments and Applications, Noyes Publications, Park Ridge, NJ, 1990. [19] K.H. Yang, Tests on concrete reinforced with hybrid or monolithic steel and polyvinyl alcohol fibers, ACI Mater. J. 108 (6) (2011) 664e672. [20] J. Xiao, H. Li, Z. Yang, Fatigue behavior of recycled aggregate concrete under compression and bending cyclic loading, Constr. Build. Mater. 38 (2013) 681e688.