- Email: [email protected]
Ecological and mechanical assessment of lightweight fiber-reinforced concrete made with rubber or expanded clay aggregates
Construction and Building Materials 127 (2016) 692–701
Contents lists available at ScienceDirect
Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat
Ecological and mechanical assessment of lightweight fiber-reinforced concrete made with rubber or expanded clay aggregates Alessandro P. Fantilli ⇑, Bernardino Chiaia, Andrea Gorino Department of Structural, Building and Geotechnical Engineering, Politecnico di Torino, Corso Duca degli Abruzzi 24, 10129 Torino, Italy
h i g h l i g h t s Rubber granulates is used herein to cast an eco-friendly lightweight concrete (RLC). RLC is compared with a traditional lightweight concrete with expanded clay (TLC). The results of the eco-mechanical analyses depend on the chosen functional unit. When compressive strength is the functional unit, TLC shows the best performances. Referring to a structural parameter, the use of RLC plates is more convenient.
a r t i c l e
i n f o
Article history: Received 23 May 2016 Received in revised form 13 September 2016 Accepted 6 October 2016
Keywords: Traditional lightweight concrete (TLC) Rubber lightweight concrete (RLC) Plastic fibers Uniaxial compression test Three point bending test Eco-mechanical analysis Functional unit
a b s t r a c t To reduce the environmental impact of traditional lightweight concrete (TLC), porous aggregates can be substituted by rubber granulates. The mechanical properties of such rubber lightweight concrete (RLC) are investigated and compared with those of TLC made with expanded clay aggregates. Uniaxial compression and three point bending tests were performed for assessing the mechanical and ecological performances of the two mixtures, containing or not plastic fibers. As a result, when compressive strength is the functional unit of the analyses, TLC performs better than RLC. Conversely, fiber-reinforced RLC is the best solution when flexural strength and structural ductility are the required performances. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction According to the definition given by Model Code 2010 (MC2010) [1], the density of lightweight aggregate concrete varies from 800 to 2000 kg/m3. To reduce the mass of normal weight structural concrete, stone aggregates are substituted by cellular structured particles. Such lightweight aggregates are generally produced by heating some raw materials (e.g., shale, clay, slates, fly ashes, etc.) to incipient fusion, and then cooling them in the so-called pyroprocessing method [2]. Lightweight concretes are mainly used to reduce the mass, and consequently the dead loads and the inertial seismic actions, of both new and existing structures, and to facilitate transportation and placement of precast elements. In general, to justify the use ⇑ Corresponding author. E-mail addresses: [email protected] (A.P. Fantilli), [email protected] (B. Chiaia), [email protected] (A. Gorino). http://dx.doi.org/10.1016/j.conbuildmat.2016.10.020 0950-0618/Ó 2016 Elsevier Ltd. All rights reserved.
of lightweight concrete, which is more expensive than normal weight concrete, a lower cost of the project and/or an improved functionality must be attained [2]. This is the case of the precast plates proposed by Fantilli et al. [3] for the sidewalks of an existing bridge. Specifically, a lightweight concrete with expanded clay aggregates was tailored to facilitate the lift of the plates, which were reinforced with plastic fibers in place of the traditional steel rebar. With respect to this solution, a more environmental-friendly lightweight concrete can be realized by using rubber from endof-life tires as non-conventional aggregate [4,5]. Indeed, also the density of rubber granulates (with the dimensions of grains comprised between 0.8 and 20 mm [6]) is lower than that of the stone aggregates. The use of rubber in the concrete industry is also convenient from an environmental point of view. Specifically, hundreds of millions scrap-tires are generated each year worldwide, and their landfilling is becoming unacceptable due to the rapid depleting of the sites and to the associated environmental risks [4].
693
A.P. Fantilli et al. / Construction and Building Materials 127 (2016) 692–701
Notations b EI Elc Elci
Elc1 flc flct,fl h klc l Mcr⁄ MI P Pcr⁄ Pd
width of a plate cross-section ecological index tangent modulus of elasticity experimentally evaluated in lightweight concrete average value of the tangent modulus of elasticity of lightweight concrete estimated in accordance with MC2010 [1] secant modulus from the origin to the peak of stress of lightweight concrete [1] compressive strength of lightweight concrete flexural tensile strength of lightweight concrete depth of a plate cross-section plasticity number of lightweight concrete [1] span of a plate in three point bending bending moment at the effective cracking of a plate mechanical index load applied to a plate in three point bending effective cracking load of a plate in three point bending (Fig. 6) design load acting on a plate in three point bending [1]
Nevertheless, the presence of rubber reduces the compressive strength of concrete, as evidenced in several studies [4,5]. Concrete class reduces with the content of rubber, and sometimes becomes lower than the minimum values required for structural uses. Moreover, compressive strength is generally assumed to be the functional unit of concrete, to which the inputs and outputs of a lifecycle assessment must be referred [7]. As high-strength concretes are in principle better than normal-strength concretes from an ecological point of view [8], traditional lightweight concrete (TLC) should also be more environmental-friendly than rubber lightweight concrete (RLC). Despite that, in several applications, the functional unit must take into account more than the mere compressive strength. For instance, to achieve the best eco-mechanical performances of high-strength but brittle concretes, material ductility needs to be enhanced [9]. In addition, in beams and plates subjected to bending actions, tensile (or flexural) strength and fracture toughness are the most important properties. Without modifying the compressive strength, the bending capacity of these structures, and the fracture toughness of the concrete as well, can be increased if fibers are added to the cementitious matrix [10,11]. Accordingly, a more comprehensive analysis, including the environmental impact and the mechanical behavior of plain and fiber-reinforced concrete, needs to be applied to full-scale structures. Since a direct comparison between the material and structural performances of TLC and those of RLC cannot be found in the current literature, the present paper aims at filling this gap. Specifically, concrete cylinders and full-scale one-way plates were tested in uniaxial compression and three point bending, respectively. Although TLC and RLC mixtures can lead to different compressive strength, they were accurately tailored in order to behave in the same way under bending actions [3]. Hence, the measured eco-mechanical performances of the lightweight concrete mixtures, some reinforced with short plastic fibers, can be compared with respect to different functional units. 2. Experimental investigation 2.1. Concrete mixtures Three TLC mixtures, named TLC_0, TLC_7 and TLC_10 and made with expanded clay aggregates, are taken into consideration (see
Pu ultimate load of a plate in three point bending (Fig. 6) Qf, Qf,min fiber content of a plate in three point bending and its minimum value [13] r flexural/compressive strength ratio of a lightweight concrete W section modulus of a plate a quantity of carbon dioxide (CO2) released by concrete components b quantity of embodied energy used by concrete components cG partial safety factor for permanent actions [1] cQ partial safety factor for variable actions [1] d midspan deflection of a plate in three point bending dcr⁄ midspan deflection associated to Pcr⁄ of a plate in three point bending (Fig. 6) du midspan deflection associated to Pu of a plate in three point bending (Fig. 6) e compressive strain elc1 strain at the maximum stress of lightweight concrete r compressive stress
Table 1). Such concretes, used for the maintenance of a bridge [3], are plain (TLC_0) and fiber-reinforced (TLC_7 e TLC_10). More precisely, a cubic meter of the mixtures TLC_7 and TLC_10 is respectively reinforced with 7 kg and 10 kg of short plastic fibers (diameter = 0.48 mm, length = 54 mm, elastic modulus = 5.75 GPa, and tensile strength > 620 MPa). The fibers are commercially available and are made by a mix of polymers (mainly polypropylene). Due to the selected components reported in Table 1, TLC mixtures have a density of about 1650 kg/m3. With respect to these cement-based composites, new and more environmental-friendly lightweight concrete mixtures are tailored herein. Thus, the content of cement is reduced of about 30%, and rubber granulates substitute a portion of the aggregates. To maintain a density of the RLC lower than 2000 kg/m3, an amount of 240 kg/m3 of rubber is added to the new mixtures (see Table 2). Obviously, due to the higher water/cement ratio (twice than that of TLC) and to the presence of rubber, RLC will show a lower compressive strength, which in turn would not alter significantly the behavior of fiber-reinforced concrete beams and plates in bending [3]. As shown by Table 2, the new mixtures RLC_0, RLC_5, and
Table 1 Material components referred to 1 m3 of traditional lightweight concrete TLC. Components
TLC_0
TLC_7
TLC_10
Water (kg) Cement Type II A-LL 42.5R (kg) Stone aggregate (kg) Expanded clay aggregate 3–8 mm (kg) Superplasticizer (l) Polypropylene fibers (kg)
140 500 700 300 5 0
140 500 700 300 5 7
140 500 700 300 5 10
Table 2 Material components referred to 1 m3 of rubber lightweight concrete RLC. Components
RLC_0
RLC_5
RLC_12
Water (kg) Cement Type II A-LL 42.5R (kg) Stone aggregate (kg) Rubber granulates 3–20 mm (kg) Superplasticizer (l) Polypropylene fibers (kg)
168 352 1131 243 6 0
168 352 1123 241 6 5
168 352 1110 238 6 12
694
A.P. Fantilli et al. / Construction and Building Materials 127 (2016) 692–701
RLC_12 contain 0, 5 and 12 kg/m3 of the same plastic fibers, respectively. Although the density of these mixtures (i.e., 1900 kg/m3) is higher than that of TLC, they can still be considered lightweight concretes in accordance to the definition given by MC2010 [1]. 2.2. Specimens and test setup Three cylindrical specimens, with height of 300 mm and base diameter of 150 mm (Fig. 1a), were cast with the six mixtures introduced in the previous section. Such samples were subjected to uniaxial compressive loads through a MTS testing machine (maximum load capacity = 1000 kN), by imposing a constant velocity (0.037 mm per minute) of the stroke. In addition to the average strain measured on the whole height of the specimens, two LVDTs were placed on the central part of each cylinder to evaluate the local strain on a base of 100 mm (Fig. 1a). Three 1000 1000 100 mm plates (see Fig. 1b) were also cast by using all the TLC and RLC mixtures defined in Tables 1 and 2, respectively. Such plates, which do not contain any steel rebar, correspond to the full-scale structure of a simply supported sidewalk [3]. Accordingly, they were tested in three point bending (Fig. 1b). The load P, distributed along the midsection through a steel beam, was applied by means of a MTS machine with a capacity of 100 kN. Two LVDTs were used to determine the midspan deflection d (depurated by the support displacements), as the stroke increased at a constant velocity of 0.360 mm per minute.
3. Experimental results 3.1. Uniaxial compression tests Fig. 2 illustrates the stress-strain (r e) curves obtained from the compression tests on the concrete cylinders depicted in Fig. 1a. During the ascending branch the strains are computed with the measures of the two LVDTs, whereas in the softening stage the shortening of the whole specimen is assumed to be coincident with the stroke of the loading machine. The main parameters of the r e curves (see Fig. 3) are collected in Table 3. They include the average values of the compressive strength flc (i.e., the peak of the r e curves), the tangent modulus of elasticity Elc (in the origin of the r e curves), the strain at the peak of stress elc1, and the plasticity number klc = Elc/Elc1 (i.e., the ratio between Elc and the secant modulus from the origin to the peak of stress Elc1 = flc/elc1). For the six mixtures investigated herein, the average values of flc, elc1, and klc are also reported in the histograms of Fig. 4. In the same Figures, the mechanical properties of two classes of lightweight concrete, named LC12 and LC8 by MC2010 [1], are also put into evidence. In particular, the strength of LC12 (flc = 20 MPa) is close to the maximum strength measured experimentally, whereas LC8 (with flc = 16 MPa) is the minimum concrete grade for structural applications [1]. According to MC2010 [1], the average values of the tangent modulus of elasticity and of the plasticity number can be estimated for LC12 (Elci = 15,237 MPa, klc = 1.3) and
Fig. 1. The experimental campaign on the lightweight mixtures: (a) Uniaxial compression tests; (b) Three point bending tests.
A.P. Fantilli et al. / Construction and Building Materials 127 (2016) 692–701
695
Fig. 2. The stress-stain (r e) curves obtained from uniaxial compression tests: (a)–(c) TLC mixtures; (d)–(f) RLC mixtures.
LC8 (Elci = 18,756 MPa, klc = 1.3). By means of such values, the strains at the maximum compressive stress elc1 are also evaluated (i.e., 2.02‰ for LC12, and 1.33‰ for LC8) and reported in Fig. 4b. The values of flc, elc1, and klc experimentally measured in the TLC mixtures are in good agreement with those of LC12 [1] (see Fig. 4). On the contrary, neither LC12 nor LC8 can define the mechanical properties of all the RLC. Indeed, they are designed to obtain more environmental-friendly lightweight concretes, by increasing the water/cement ratio of TLC and substituting the expanded clay with
rubber. The average compressive strength of RLC is lower than 16 MPa (which is the average strength of LC8 in Fig. 4a), and therefore it cannot be used in structural applications, especially in compressed elements (e.g., columns). Concerning the plasticity number reported in Fig. 4c, the discrepancy between the measured values and the estimations given by MC2010 [1] can be ascribed to the type of aggregates. Although rubber (density = 900 kg/m3) is more dense than expanded clay (density = 380 kg/m3), the latter exhibits a lower deformability.
696
A.P. Fantilli et al. / Construction and Building Materials 127 (2016) 692–701
Fig. 3. An ideal stress-strain relationship and the main lightweight concrete properties measured with uniaxial compression test.
Table 3 The average values of the concrete properties measured with uniaxial compression tests. Specimens
flc (MPa)
Elc (MPa)
elc1 (‰)
klc
TLC_0 TLC_7 TLC_10 RLC_0 RLC_5 RLC_12
21.51 23.36 22.91 9.84 10.94 10.99
15,070 15,619 18,868 12,326 15,667 15,607
1.8 2.2 1.8 1.6 2.0 1.9
1.3 1.5 1.5 1.8 2.9 2.7
Unlike compressive strength (Fig. 4a), the deformability of RLC is higher than that of TLC (Fig. 4c). In both the lightweight mixtures under investigation, the presence of a fiber-reinforcement does not alter neither the maximum stress nor the deformability of the pre-peak stage of the stressstrain curves (Fig. 2). Conversely, for the same type of concrete (TLC in Fig. 2a–c and RLC in Fig. 2d–f), the higher the fiber content, the higher the ductility of (or the area delimited by) the post-peak branches. In other words, the confinement effect produced by the fiber-reinforcement in normal weight concrete under compression [12], also occurs in lightweight concrete made with rubber or expanded clay aggregates. 3.2. Three point bending tests The results of the bending tests on one-way plates are reported in Fig. 5, where the load-midspan deflection curves (P d) are depicted. In all the tests, the effective cracking load Pcr⁄, at which the softening stage of the P d curve begins, is clearly evident. In the case of unreinforced plates (i.e., TLC_0 and RLC_0 in Fig. 5a and d, respectively), Pcr⁄ is the absolute maximum. On the other hand, the P d curves of fiber-reinforced plates show two relative maximums (Fig. 5b–c for TLC, and Fig. 5e–f for RLC). The first coincides with Pcr⁄, and the second, at the onset of strain localization, corresponds to the ultimate load Pu. Fig. 6 illustrates an ideal P d curve, whereas Table 4 reports the experimental values of Pcr⁄ and Pu, and of the corresponding deflections dcr⁄ and du. Moreover, the average values of the effective cracking load and of the ultimate load are also illustrated in the histograms of Fig. 7a–b. For each type of lightweight concrete plate, Pcr⁄ is remarkably higher than the maximum design load Pd = 8.60 kN,
Fig. 4. The average values of the lightweight concrete properties compared with the estimations given by MC2010 [1]: (a) Compressive strength; (b) Strain at the maximum stress; (c) Plasticity number.
reported with a dashed line in Fig. 7a–b. Such a value is obtained by factorizing the dead loads (equal to 2.0 kN/m2) and the variable load (4.0 kN/m2) with the partial safety factors cG = 1.3 and cQ = 1.5, respectively [1]. The brittle behavior of unreinforced lightweight concrete can be observed in Fig. 5, where the P d curves show a sharp drop after the peak, especially in TLC_0 plates (Fig. 5a). On the contrary, in the
A.P. Fantilli et al. / Construction and Building Materials 127 (2016) 692–701
697
Fig. 5. The load-deflection (P d) curves obtained from the three point bending tests: (a)–(c) TLC plates; (d)–(f) RLC plates.
softening stage of RLC_0 plates (Fig. 5d), residual loads can be detected also for large deflections. If fibers are added to the cementitious matrix, the values of residual loads increase in both mixtures. In particular, Pu grows with the fiber content (see Fig. 5b–c for TLC, and Fig. 5e–f for RLC), whereas Pcr⁄ is nearly constant (see Table 4). In some cases (i.e., the plates TLC_10 in Fig. 5c and RLC_12 in Fig. 5f), a very ductile response, with Pu Pcr⁄, can be observed.
To quantify the post-peak behavior of the plates under investigation, the Ductility Index (DI) introduced by Fantilli et al. [13] can be used. Specifically, a lightly reinforced structural element in bending behaves in a ductile manner when the following ratio DI is positive:
DI ¼
Pu Pcr P cr
ð1Þ
698
A.P. Fantilli et al. / Construction and Building Materials 127 (2016) 692–701 Table 4 The average values of the structural properties measured with the three point bending tests on one-way plates. Plates
Pcr⁄ (kN)
dcr⁄ (mm)
Pu (kN)
du (mm)
DI
r
TLC_0 TLC_7 TLC_10
16.01 22.32 21.59
0.27 0.36 0.56
/ 17.95 22.93
/ 6.25 4.82
/ 0.20 0.06
0.10 0.13 0.13
RLC_0 RLC_5 RLC_12
20.24 20.96 20.46
0.34 0.45 0.47
/ 15.84 19.27
/ 2.31 3.64
/ 0.24 0.06
0.28 0.26 0.25
Fig. 7c and d, respectively. Hence, a minimum amount of fibers Qf, min can be easily individuated at the brittle/ductile transition (i.e., when DI = 0). Specifically, Qf,min = 9.2 kg/m3 is determined for TLC (Fig. 7c), whereas 14.2 kg/m3 of fibers are necessary to attain the minimum required ductility in RLC (Fig. 7d).
Fig. 6. Effective cracking and ultimate conditions of an ideal load-deflection curve obtained through a three point bending test.
For the fiber-reinforced plates (i.e., TLC_7, TLC_10, RLC_5 and RLC_12), the average values of DI computed with Eq. (1) are reported in Table 4. A linear increment of DI with the fiber content Qf can be assumed for both TLC and RLC plates [13], as illustrated in
4. Comparative analyses 4.1. Analysis at the material level As reported in Table 3, and illustrated in Fig. 4a as well, flc experimentally measured with RLC specimens is about one half of the compressive strength of TLC. This is due to the different water/cement ratio (see Tables 1 and 2), and also to the different
Fig. 7. Strength and ductility of one-way plates: (a) Average value of the effective cracking load compared with the design load (Pd = 8.60 kN); (b) Average value of the ultimate load compared with the design load (Pd = 8.60 kN); (c) Ductility index of TLC plates as a function of fiber content; (d) Ductility index of RLC plates as a function of fiber content.
A.P. Fantilli et al. / Construction and Building Materials 127 (2016) 692–701
aggregate/cement paste interaction in the two mixtures [14]. However, it is worth noting the large ductility of RLC, which is remarkably higher than that of TLC. This is particularly evident in Fig. 8, where the normalized stress vs. compressive strain curves of both TLC and RLC mixtures are compared. In this Figure, the two lightweight concretes are divided into three groups, depending on the
699
amount of fibers (i.e., TLC_0 and RLC_0 in Fig. 8a, TLC_7 and RLC_5 in Fig. 8b, TLC_10 and RLC_12 in Fig. 8c). As illustrated in Fig. 8a, the fracture toughness of RLC_0, higher than TLC_0, can be compared with that of traditional normal weight concrete containing short fibers randomly dispersed within the cementitious matrix [12]. Thus, it can be argued that the rubber aggregates exert a sort of passive confinement in compression, which is not noticed in the presence of expanded clay aggregates. Obviously, the fracture toughness increases with the fiber content, even if the RLC mixtures always show the higher ductility. As a matter of fact, at the same level of strain, the normalized postpeak residual stresses of RLC always exceed those of TLC, regardless of the fiber content (Fig. 8b and c). 4.2. Analysis at the structural level Despite the lower compressive strength measured in the rubber lightweight concrete, the flexural strength of RLC is comparable with, and sometimes higher than, that of TLC. Thus, the mechanical properties of concrete elements subjected to bending actions can be related to the compressive strength, but not in the same manner for all the concretes. With respect to TLC_0, the higher values of Pcr⁄ measured in RLC_0 plates (see Table 4 and Fig. 7a) is also reported in the traditional normal weight concrete when fibers are added to matrix [15]. This further corroborates the conjecture that rubber aggregates contribute to increase concrete toughness, similarly to a fiber-reinforcement. To quantify the beneficial effect of rubber, the flexural tensile strengths flct,fl of both the lightweight concretes are elastically evaluated from the three point bending tests on plates. In this way, a direct comparison with the corresponding compressive strength can be performed by introducing the following strength ratio:
r¼
f lct;fl M cr 1 ¼ f lc W f lc
ð2Þ
where Mcr⁄ = Pcr⁄ l/4 = effective cracking moment at the midspan cross-section of the plate (where l = 900 mm in Fig. 1b); and W = b h2/6 = section modulus of the plate (where b = 1000 mm, and h = 100 mm in Fig. 1b). According to Eq. (2), the average values of r are herein computed by using the experimental values of Mcr⁄ and flc and then reported in Fig. 9 and Table 4. The positive contribution of rubber aggregates to the flexural strength of lightweight concrete is particularly evident in the histogram of Fig. 9. Indeed, the values of r computed for TLC plates might be more or less doubled to obtain the strength ratio of RLC plates. To be more precise, r = 10–13% in TLC, and r = 25–28% in RLC (Table 4). This is due to the effect produced by the rubber granulates, which is macroscopically similar to the bridging action exerted by the fibers crossing the crack surfaces. 4.3. Ecological and mechanical assessment By introducing the so-called Ecological Index (EI) and Mechanical Index (MI), a more comprehensive comparative analysis, including the environmental impact of the two lightweight mixtures, can be performed [16]. The following equation estimates EI by combining the amount of the released carbon dioxide (i.e., a = carbon footprint) and the energy used in the product lifecycle (i.e., b = embodied energy):
EI ¼ a b Fig. 8. Normalized stress-strain curves of the lightweight concretes in compression: (a) TLC_0 and RLC_0 mixtures; (b) TLC_7 and RLC_5 mixtures; (c) TLC_10 and RLC_12 mixtures.
ð3Þ
Both a and b are measured here for the most relevant concrete components, such as cement, lightweight aggregate (made with
700
A.P. Fantilli et al. / Construction and Building Materials 127 (2016) 692–701
Fig. 9. The average values of the strength ratio r of both TLC and RLC mixtures.
expanded clay or grained rubber from end-of-life tires) and polypropylene fibers. Superplasticizer is not considered in the computation of EI, because it has a negligible environmental impact with respect to the other components of TLC and RLC mixtures. Table 5 collects all the data necessary to compute EI in accordance with Eq. (3) [17–20]. It must be remarked that Eq. (3) takes into account environmental aspects in a simple way. To perform more complex sustainability analyses, including social and economic issues, other methods can be found in the current literature (e.g., MIVES approach [21]). As the mechanical index must be related to a functional unit, three analyses are carried out herein: MI1 = flc at the material level, MI2 = Pmax = max (Pcr⁄; Pu) and MI3 = DI at the structural level (see Table 6). In the first two cases, the ecological index of TLC_0 (Table 6) is used as the upper bound value of the environmental performances (EITLC_0). Whereas, the lower bound values of the mechanical index are assumed to be MI1,TLC_0 = 16 MPa (i.e., the minimum compressive strength for structural applications suggested by MC2010 [1]) and MI2,TLC_0 = Pd = 8.60 kN (i.e., the design load of the plates). Referred to these bounds, two eco-mechanical analyses can be carried out for all the mixtures by means of the non-dimensional diagrams depicted in Fig. 10a–b [16]. With respect to MI1 = flc (Fig. 10a), neither TLC nor RLC mixtures contemporarily fulfil the minimum mechanical and environmental Table 5 The environmental impact of concrete components (referred to 1 kg of material). Components
Carbon footprint (kg CO2)
Embodied energy (MJ)
Cement Expanded clay aggregate Rubber granulates Polypropylene fibers
0.8 0.3 0.2 2.7
5 4 4 100 Fig. 10. Eco-mechanical analysis of TLC and RLC mixtures: (a) Results at the material scale with MI = MI1; (b) Results at the structural scale with MI = MI2; (c) Results at the structural scale for the ideal mixtures having the minimum amount of fibers Qf,min.
Table 6 The ecological and mechanical indexes of the six lightweight concretes. Mixtures
EI (kg CO2 GJ/m3)
MI1 (MPa)
MI2 (kN)
TLC_0 TLC_7 TLC_10
1803 2219 2432
21.51 23.36 22.91
16.01 22.32 22.93
RLC_0 RLC_5 RLC_12
865 1070 1379
9.84 10.94 10.99
20.24 20.96 20.46
requirements. As a matter of fact, fiber-reinforced traditional lightweight concretes satisfy the condition MI > MI1,TLC_0, but EI > EITLC_0 due to the impact of fibers. Conversely, RLC only fits the ecological performances (with and without the fibers), but it does not achieve the minimum strength required for structural use.
A.P. Fantilli et al. / Construction and Building Materials 127 (2016) 692–701
When MI2 = Pmax, TLC only meets the mechanical requirement (i.e., MI > MI2,TLC_0), whereas both EI < EITLC_0 and MI > MI1,TLC_0 are satisfied by all the rubber lightweight concretes investigated in this paper. Among them, RLC_0 seems to be the best cementbased composite to be used in one-way plates. Indeed, EI increases in presence of fibers (see RLC_5 and RLC_12 in Table 6) whereas the values of Pmax are nearly the same for all the RLC plates. Nevertheless, the brittle behavior of RLC_0 plates (see Fig. 5d) cannot be accepted for structural applications. In other words, concrete beams and plates in bending have to show a minimum ductility (i.e., DI P 0) [13]. Thus, the proposed comparative analysis involves the ideal TLC and RLC plates containing the minimum amount of fibers (i.e., Qf,min = 9.2 kg/m3 for TLC, and Qf,min = 14.2 kg/m3 for RLC) graphically evaluated in Fig. 7c and d, respectively. Under this condition, their performances are only compared in terms of EI (see Fig. 10c), because MI3 = DI = 0 for both the ideal plates. As a result, RLC (which shows EI = 1493 kg CO2 GJ/m3) performs better than TLC (with EI = 2380 kg CO2 GJ/m3) when the minimum amount of fibers is added to the mixtures. 5. Conclusions The results of the experimental investigations, concerning the behavior of two different lightweight concrete mixtures, can be summarized by the following points: 1. Due to a higher water/cement ratio and to the presence of rubber granulates, which substitute the expanded clay aggregates, RLC is more environmental-friendly than TLC, but the strength class reduces, regardless of the amount of plastic fibers. Nevertheless, uniaxial compression tests on cylinders reveal the higher ductility of RLC, even in the absence of fiberreinforcement. 2. On the contrary, the flexural strength of TLC and RLC is comparable, as measured in the three point bending tests on unreinforced and fiber-reinforced plates. 3. Like the fibers, rubber granulates exert a sort of passive confinement in compressed elements, and prevent the rapid growth of cracks in the tensile zones of beams and plates. 4. If also the environmental impact is taken into account, the results of an eco-mechanical analysis depend on the functional unit used to compare the concretes. When compressive strength (i.e., a material property) is the reference performance, RLC fits the ecological requirements both with and without the fibers, but it does not achieve the minimum strength for structural uses, especially in compressed elements. 5. On the contrary, if the analyses are referred to structural properties, such as the maximum load Pmax of bending plates, substitution of expanded clay aggregates with rubber granulates is convenient. This is particularly true when the minimum condition of ductility (i.e., DI = 0) is also taken into account. Finally, the above-mentioned differences suggest the necessity of a proper definition of the functional unit to which the inputs
701
and outputs of a life-cycle assessment must be referred. It would be desirable that this function, and the ecological benchmarks of concrete structures as well, are reported by the current building codes. Acknowledgments The grants given by the Italian Laboratories University Network of seismic engineering (ReLUIS) and by the Ecotyre Consortium for the end-of-life tires disposal, and used to develop the present research work, are gratefully acknowledged. References [1] International Federation for Structural Concrete, Model Code 2010 – Final draft. fib, Lausanne, 2012, fib Bulletin 65–66. [2] American Concrete Institute, Guide for Structural Lightweight-Aggregate Concrete, ACI, Farmington Hills, 2003. Committee 213. [3] A.P. Fantilli, A.D. Cavallo, G. Pistone, Fiber-reinforced lightweight concrete slabs for the maintenance of the Soleri Viaduct, Eng. Struct. 99 (2015) 184– 191. [4] R. Siddique, T.R. Naik, Properties of concrete containing scrap-tire rubber – an overview, Waste Manage. 24 (2004) 563–569. [5] K.B. Najim, M.R. Hall, A review of the fresh/hardened properties and applications for plain-(PRC) and self-compacting rubberised concrete (SCRC), Constr. Build. Mater. 24 (2010) 2043–2051. [6] European Committee for Standardization, Materials Produced from End of Life Tyres – Specification of Categories Based on their Dimension(s) and Impurities and Methods for Determining their Dimension(s) and Impurities, CEN, 2010, CEN/TS 14243. [7] B.L. Damineli, F.M. Kemeid, P.S. Aguiar, V.M. John, Measuring the eco-efficiency of cement use, Cement Concr. Compos. 32 (2010) 555–562. [8] G. Habert, N. Roussel, Study of two concrete mix-design strategies to reach carbon mitigation objectives, Cement Concr. Compos. 31 (2009) 397–402. [9] A.P. Fantilli, B. Chiaia, Eco-mechanical performances of cement-based materials: an application to self-consolidating concrete, Constr. Build. Mater. 40 (2013) 189–196. [10] A. Bentur, S. Mindess, Fibre Reinforced Cementitious Composites, Elsevier Applied Science, London and New York, 1990. [11] P. Pujadas, A. Blanco, S. Cavalaro, A. Aguado, Plastic fibres as the only reinforcement for flat suspended slabs: experimental investigation and numerical simulation, Constr. Build. Mater. 57 (2014) 92–104. [12] A.P. Fantilli, H. Mihashi, P. Vallini, B. Chiaia, Equivalent confinement in HPFRCC columns measured by triaxial test, ACI Mater. J. 108 (2011) 159–167. [13] A.P. Fantilli, B. Chiaia, A. Gorino, Fiber volume fraction and ductility index of concrete beams, Cement Concr. Compos. 65 (2016) 139–149. [14] E. Ganjian, M. Khorami, A.A. Maghsoudi, Scrap-tyre-rubber replacement for aggregate and filler in concrete, Constr. Build. Mater. 23 (2009) 1828–1836. [15] A.E. Naaman, Strain hardening and deflection hardening fiber reinforced cement composites, in: Proceedings of the 4th International RILEM Workshop on High Performance Fiber Reinforced Cement Composites, Ann, Abor, 2003, pp. 95–113. [16] A.P. Fantilli, B. Chiaia, The work of fracture in the eco-mechanical performances of structural concrete, J. Adv. Concr. Technol. 11 (2013) 282– 290. [17] International Federation for Structural Concrete, Guidelines for Green Concrete Structures, fib, Lausanne, 2012, fib Bulletin 67. [18] M.F. Ashby, Materials and the Environment: Eco-Informed Materials Choice, Butterworth-Heinemann (Elsevier), Burlington, 2009. [19] Swiss Centre for Life Cycle Inventories, The Ecoinvent Database. Available from:
Contents lists available at ScienceDirect
Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat
Ecological and mechanical assessment of lightweight fiber-reinforced concrete made with rubber or expanded clay aggregates Alessandro P. Fantilli ⇑, Bernardino Chiaia, Andrea Gorino Department of Structural, Building and Geotechnical Engineering, Politecnico di Torino, Corso Duca degli Abruzzi 24, 10129 Torino, Italy
h i g h l i g h t s Rubber granulates is used herein to cast an eco-friendly lightweight concrete (RLC). RLC is compared with a traditional lightweight concrete with expanded clay (TLC). The results of the eco-mechanical analyses depend on the chosen functional unit. When compressive strength is the functional unit, TLC shows the best performances. Referring to a structural parameter, the use of RLC plates is more convenient.
a r t i c l e
i n f o
Article history: Received 23 May 2016 Received in revised form 13 September 2016 Accepted 6 October 2016
Keywords: Traditional lightweight concrete (TLC) Rubber lightweight concrete (RLC) Plastic fibers Uniaxial compression test Three point bending test Eco-mechanical analysis Functional unit
a b s t r a c t To reduce the environmental impact of traditional lightweight concrete (TLC), porous aggregates can be substituted by rubber granulates. The mechanical properties of such rubber lightweight concrete (RLC) are investigated and compared with those of TLC made with expanded clay aggregates. Uniaxial compression and three point bending tests were performed for assessing the mechanical and ecological performances of the two mixtures, containing or not plastic fibers. As a result, when compressive strength is the functional unit of the analyses, TLC performs better than RLC. Conversely, fiber-reinforced RLC is the best solution when flexural strength and structural ductility are the required performances. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction According to the definition given by Model Code 2010 (MC2010) [1], the density of lightweight aggregate concrete varies from 800 to 2000 kg/m3. To reduce the mass of normal weight structural concrete, stone aggregates are substituted by cellular structured particles. Such lightweight aggregates are generally produced by heating some raw materials (e.g., shale, clay, slates, fly ashes, etc.) to incipient fusion, and then cooling them in the so-called pyroprocessing method [2]. Lightweight concretes are mainly used to reduce the mass, and consequently the dead loads and the inertial seismic actions, of both new and existing structures, and to facilitate transportation and placement of precast elements. In general, to justify the use ⇑ Corresponding author. E-mail addresses: [email protected] (A.P. Fantilli), [email protected] (B. Chiaia), [email protected] (A. Gorino). http://dx.doi.org/10.1016/j.conbuildmat.2016.10.020 0950-0618/Ó 2016 Elsevier Ltd. All rights reserved.
of lightweight concrete, which is more expensive than normal weight concrete, a lower cost of the project and/or an improved functionality must be attained [2]. This is the case of the precast plates proposed by Fantilli et al. [3] for the sidewalks of an existing bridge. Specifically, a lightweight concrete with expanded clay aggregates was tailored to facilitate the lift of the plates, which were reinforced with plastic fibers in place of the traditional steel rebar. With respect to this solution, a more environmental-friendly lightweight concrete can be realized by using rubber from endof-life tires as non-conventional aggregate [4,5]. Indeed, also the density of rubber granulates (with the dimensions of grains comprised between 0.8 and 20 mm [6]) is lower than that of the stone aggregates. The use of rubber in the concrete industry is also convenient from an environmental point of view. Specifically, hundreds of millions scrap-tires are generated each year worldwide, and their landfilling is becoming unacceptable due to the rapid depleting of the sites and to the associated environmental risks [4].
693
A.P. Fantilli et al. / Construction and Building Materials 127 (2016) 692–701
Notations b EI Elc Elci
Elc1 flc flct,fl h klc l Mcr⁄ MI P Pcr⁄ Pd
width of a plate cross-section ecological index tangent modulus of elasticity experimentally evaluated in lightweight concrete average value of the tangent modulus of elasticity of lightweight concrete estimated in accordance with MC2010 [1] secant modulus from the origin to the peak of stress of lightweight concrete [1] compressive strength of lightweight concrete flexural tensile strength of lightweight concrete depth of a plate cross-section plasticity number of lightweight concrete [1] span of a plate in three point bending bending moment at the effective cracking of a plate mechanical index load applied to a plate in three point bending effective cracking load of a plate in three point bending (Fig. 6) design load acting on a plate in three point bending [1]
Nevertheless, the presence of rubber reduces the compressive strength of concrete, as evidenced in several studies [4,5]. Concrete class reduces with the content of rubber, and sometimes becomes lower than the minimum values required for structural uses. Moreover, compressive strength is generally assumed to be the functional unit of concrete, to which the inputs and outputs of a lifecycle assessment must be referred [7]. As high-strength concretes are in principle better than normal-strength concretes from an ecological point of view [8], traditional lightweight concrete (TLC) should also be more environmental-friendly than rubber lightweight concrete (RLC). Despite that, in several applications, the functional unit must take into account more than the mere compressive strength. For instance, to achieve the best eco-mechanical performances of high-strength but brittle concretes, material ductility needs to be enhanced [9]. In addition, in beams and plates subjected to bending actions, tensile (or flexural) strength and fracture toughness are the most important properties. Without modifying the compressive strength, the bending capacity of these structures, and the fracture toughness of the concrete as well, can be increased if fibers are added to the cementitious matrix [10,11]. Accordingly, a more comprehensive analysis, including the environmental impact and the mechanical behavior of plain and fiber-reinforced concrete, needs to be applied to full-scale structures. Since a direct comparison between the material and structural performances of TLC and those of RLC cannot be found in the current literature, the present paper aims at filling this gap. Specifically, concrete cylinders and full-scale one-way plates were tested in uniaxial compression and three point bending, respectively. Although TLC and RLC mixtures can lead to different compressive strength, they were accurately tailored in order to behave in the same way under bending actions [3]. Hence, the measured eco-mechanical performances of the lightweight concrete mixtures, some reinforced with short plastic fibers, can be compared with respect to different functional units. 2. Experimental investigation 2.1. Concrete mixtures Three TLC mixtures, named TLC_0, TLC_7 and TLC_10 and made with expanded clay aggregates, are taken into consideration (see
Pu ultimate load of a plate in three point bending (Fig. 6) Qf, Qf,min fiber content of a plate in three point bending and its minimum value [13] r flexural/compressive strength ratio of a lightweight concrete W section modulus of a plate a quantity of carbon dioxide (CO2) released by concrete components b quantity of embodied energy used by concrete components cG partial safety factor for permanent actions [1] cQ partial safety factor for variable actions [1] d midspan deflection of a plate in three point bending dcr⁄ midspan deflection associated to Pcr⁄ of a plate in three point bending (Fig. 6) du midspan deflection associated to Pu of a plate in three point bending (Fig. 6) e compressive strain elc1 strain at the maximum stress of lightweight concrete r compressive stress
Table 1). Such concretes, used for the maintenance of a bridge [3], are plain (TLC_0) and fiber-reinforced (TLC_7 e TLC_10). More precisely, a cubic meter of the mixtures TLC_7 and TLC_10 is respectively reinforced with 7 kg and 10 kg of short plastic fibers (diameter = 0.48 mm, length = 54 mm, elastic modulus = 5.75 GPa, and tensile strength > 620 MPa). The fibers are commercially available and are made by a mix of polymers (mainly polypropylene). Due to the selected components reported in Table 1, TLC mixtures have a density of about 1650 kg/m3. With respect to these cement-based composites, new and more environmental-friendly lightweight concrete mixtures are tailored herein. Thus, the content of cement is reduced of about 30%, and rubber granulates substitute a portion of the aggregates. To maintain a density of the RLC lower than 2000 kg/m3, an amount of 240 kg/m3 of rubber is added to the new mixtures (see Table 2). Obviously, due to the higher water/cement ratio (twice than that of TLC) and to the presence of rubber, RLC will show a lower compressive strength, which in turn would not alter significantly the behavior of fiber-reinforced concrete beams and plates in bending [3]. As shown by Table 2, the new mixtures RLC_0, RLC_5, and
Table 1 Material components referred to 1 m3 of traditional lightweight concrete TLC. Components
TLC_0
TLC_7
TLC_10
Water (kg) Cement Type II A-LL 42.5R (kg) Stone aggregate (kg) Expanded clay aggregate 3–8 mm (kg) Superplasticizer (l) Polypropylene fibers (kg)
140 500 700 300 5 0
140 500 700 300 5 7
140 500 700 300 5 10
Table 2 Material components referred to 1 m3 of rubber lightweight concrete RLC. Components
RLC_0
RLC_5
RLC_12
Water (kg) Cement Type II A-LL 42.5R (kg) Stone aggregate (kg) Rubber granulates 3–20 mm (kg) Superplasticizer (l) Polypropylene fibers (kg)
168 352 1131 243 6 0
168 352 1123 241 6 5
168 352 1110 238 6 12
694
A.P. Fantilli et al. / Construction and Building Materials 127 (2016) 692–701
RLC_12 contain 0, 5 and 12 kg/m3 of the same plastic fibers, respectively. Although the density of these mixtures (i.e., 1900 kg/m3) is higher than that of TLC, they can still be considered lightweight concretes in accordance to the definition given by MC2010 [1]. 2.2. Specimens and test setup Three cylindrical specimens, with height of 300 mm and base diameter of 150 mm (Fig. 1a), were cast with the six mixtures introduced in the previous section. Such samples were subjected to uniaxial compressive loads through a MTS testing machine (maximum load capacity = 1000 kN), by imposing a constant velocity (0.037 mm per minute) of the stroke. In addition to the average strain measured on the whole height of the specimens, two LVDTs were placed on the central part of each cylinder to evaluate the local strain on a base of 100 mm (Fig. 1a). Three 1000 1000 100 mm plates (see Fig. 1b) were also cast by using all the TLC and RLC mixtures defined in Tables 1 and 2, respectively. Such plates, which do not contain any steel rebar, correspond to the full-scale structure of a simply supported sidewalk [3]. Accordingly, they were tested in three point bending (Fig. 1b). The load P, distributed along the midsection through a steel beam, was applied by means of a MTS machine with a capacity of 100 kN. Two LVDTs were used to determine the midspan deflection d (depurated by the support displacements), as the stroke increased at a constant velocity of 0.360 mm per minute.
3. Experimental results 3.1. Uniaxial compression tests Fig. 2 illustrates the stress-strain (r e) curves obtained from the compression tests on the concrete cylinders depicted in Fig. 1a. During the ascending branch the strains are computed with the measures of the two LVDTs, whereas in the softening stage the shortening of the whole specimen is assumed to be coincident with the stroke of the loading machine. The main parameters of the r e curves (see Fig. 3) are collected in Table 3. They include the average values of the compressive strength flc (i.e., the peak of the r e curves), the tangent modulus of elasticity Elc (in the origin of the r e curves), the strain at the peak of stress elc1, and the plasticity number klc = Elc/Elc1 (i.e., the ratio between Elc and the secant modulus from the origin to the peak of stress Elc1 = flc/elc1). For the six mixtures investigated herein, the average values of flc, elc1, and klc are also reported in the histograms of Fig. 4. In the same Figures, the mechanical properties of two classes of lightweight concrete, named LC12 and LC8 by MC2010 [1], are also put into evidence. In particular, the strength of LC12 (flc = 20 MPa) is close to the maximum strength measured experimentally, whereas LC8 (with flc = 16 MPa) is the minimum concrete grade for structural applications [1]. According to MC2010 [1], the average values of the tangent modulus of elasticity and of the plasticity number can be estimated for LC12 (Elci = 15,237 MPa, klc = 1.3) and
Fig. 1. The experimental campaign on the lightweight mixtures: (a) Uniaxial compression tests; (b) Three point bending tests.
A.P. Fantilli et al. / Construction and Building Materials 127 (2016) 692–701
695
Fig. 2. The stress-stain (r e) curves obtained from uniaxial compression tests: (a)–(c) TLC mixtures; (d)–(f) RLC mixtures.
LC8 (Elci = 18,756 MPa, klc = 1.3). By means of such values, the strains at the maximum compressive stress elc1 are also evaluated (i.e., 2.02‰ for LC12, and 1.33‰ for LC8) and reported in Fig. 4b. The values of flc, elc1, and klc experimentally measured in the TLC mixtures are in good agreement with those of LC12 [1] (see Fig. 4). On the contrary, neither LC12 nor LC8 can define the mechanical properties of all the RLC. Indeed, they are designed to obtain more environmental-friendly lightweight concretes, by increasing the water/cement ratio of TLC and substituting the expanded clay with
rubber. The average compressive strength of RLC is lower than 16 MPa (which is the average strength of LC8 in Fig. 4a), and therefore it cannot be used in structural applications, especially in compressed elements (e.g., columns). Concerning the plasticity number reported in Fig. 4c, the discrepancy between the measured values and the estimations given by MC2010 [1] can be ascribed to the type of aggregates. Although rubber (density = 900 kg/m3) is more dense than expanded clay (density = 380 kg/m3), the latter exhibits a lower deformability.
696
A.P. Fantilli et al. / Construction and Building Materials 127 (2016) 692–701
Fig. 3. An ideal stress-strain relationship and the main lightweight concrete properties measured with uniaxial compression test.
Table 3 The average values of the concrete properties measured with uniaxial compression tests. Specimens
flc (MPa)
Elc (MPa)
elc1 (‰)
klc
TLC_0 TLC_7 TLC_10 RLC_0 RLC_5 RLC_12
21.51 23.36 22.91 9.84 10.94 10.99
15,070 15,619 18,868 12,326 15,667 15,607
1.8 2.2 1.8 1.6 2.0 1.9
1.3 1.5 1.5 1.8 2.9 2.7
Unlike compressive strength (Fig. 4a), the deformability of RLC is higher than that of TLC (Fig. 4c). In both the lightweight mixtures under investigation, the presence of a fiber-reinforcement does not alter neither the maximum stress nor the deformability of the pre-peak stage of the stressstrain curves (Fig. 2). Conversely, for the same type of concrete (TLC in Fig. 2a–c and RLC in Fig. 2d–f), the higher the fiber content, the higher the ductility of (or the area delimited by) the post-peak branches. In other words, the confinement effect produced by the fiber-reinforcement in normal weight concrete under compression [12], also occurs in lightweight concrete made with rubber or expanded clay aggregates. 3.2. Three point bending tests The results of the bending tests on one-way plates are reported in Fig. 5, where the load-midspan deflection curves (P d) are depicted. In all the tests, the effective cracking load Pcr⁄, at which the softening stage of the P d curve begins, is clearly evident. In the case of unreinforced plates (i.e., TLC_0 and RLC_0 in Fig. 5a and d, respectively), Pcr⁄ is the absolute maximum. On the other hand, the P d curves of fiber-reinforced plates show two relative maximums (Fig. 5b–c for TLC, and Fig. 5e–f for RLC). The first coincides with Pcr⁄, and the second, at the onset of strain localization, corresponds to the ultimate load Pu. Fig. 6 illustrates an ideal P d curve, whereas Table 4 reports the experimental values of Pcr⁄ and Pu, and of the corresponding deflections dcr⁄ and du. Moreover, the average values of the effective cracking load and of the ultimate load are also illustrated in the histograms of Fig. 7a–b. For each type of lightweight concrete plate, Pcr⁄ is remarkably higher than the maximum design load Pd = 8.60 kN,
Fig. 4. The average values of the lightweight concrete properties compared with the estimations given by MC2010 [1]: (a) Compressive strength; (b) Strain at the maximum stress; (c) Plasticity number.
reported with a dashed line in Fig. 7a–b. Such a value is obtained by factorizing the dead loads (equal to 2.0 kN/m2) and the variable load (4.0 kN/m2) with the partial safety factors cG = 1.3 and cQ = 1.5, respectively [1]. The brittle behavior of unreinforced lightweight concrete can be observed in Fig. 5, where the P d curves show a sharp drop after the peak, especially in TLC_0 plates (Fig. 5a). On the contrary, in the
A.P. Fantilli et al. / Construction and Building Materials 127 (2016) 692–701
697
Fig. 5. The load-deflection (P d) curves obtained from the three point bending tests: (a)–(c) TLC plates; (d)–(f) RLC plates.
softening stage of RLC_0 plates (Fig. 5d), residual loads can be detected also for large deflections. If fibers are added to the cementitious matrix, the values of residual loads increase in both mixtures. In particular, Pu grows with the fiber content (see Fig. 5b–c for TLC, and Fig. 5e–f for RLC), whereas Pcr⁄ is nearly constant (see Table 4). In some cases (i.e., the plates TLC_10 in Fig. 5c and RLC_12 in Fig. 5f), a very ductile response, with Pu Pcr⁄, can be observed.
To quantify the post-peak behavior of the plates under investigation, the Ductility Index (DI) introduced by Fantilli et al. [13] can be used. Specifically, a lightly reinforced structural element in bending behaves in a ductile manner when the following ratio DI is positive:
DI ¼
Pu Pcr P cr
ð1Þ
698
A.P. Fantilli et al. / Construction and Building Materials 127 (2016) 692–701 Table 4 The average values of the structural properties measured with the three point bending tests on one-way plates. Plates
Pcr⁄ (kN)
dcr⁄ (mm)
Pu (kN)
du (mm)
DI
r
TLC_0 TLC_7 TLC_10
16.01 22.32 21.59
0.27 0.36 0.56
/ 17.95 22.93
/ 6.25 4.82
/ 0.20 0.06
0.10 0.13 0.13
RLC_0 RLC_5 RLC_12
20.24 20.96 20.46
0.34 0.45 0.47
/ 15.84 19.27
/ 2.31 3.64
/ 0.24 0.06
0.28 0.26 0.25
Fig. 7c and d, respectively. Hence, a minimum amount of fibers Qf, min can be easily individuated at the brittle/ductile transition (i.e., when DI = 0). Specifically, Qf,min = 9.2 kg/m3 is determined for TLC (Fig. 7c), whereas 14.2 kg/m3 of fibers are necessary to attain the minimum required ductility in RLC (Fig. 7d).
Fig. 6. Effective cracking and ultimate conditions of an ideal load-deflection curve obtained through a three point bending test.
For the fiber-reinforced plates (i.e., TLC_7, TLC_10, RLC_5 and RLC_12), the average values of DI computed with Eq. (1) are reported in Table 4. A linear increment of DI with the fiber content Qf can be assumed for both TLC and RLC plates [13], as illustrated in
4. Comparative analyses 4.1. Analysis at the material level As reported in Table 3, and illustrated in Fig. 4a as well, flc experimentally measured with RLC specimens is about one half of the compressive strength of TLC. This is due to the different water/cement ratio (see Tables 1 and 2), and also to the different
Fig. 7. Strength and ductility of one-way plates: (a) Average value of the effective cracking load compared with the design load (Pd = 8.60 kN); (b) Average value of the ultimate load compared with the design load (Pd = 8.60 kN); (c) Ductility index of TLC plates as a function of fiber content; (d) Ductility index of RLC plates as a function of fiber content.
A.P. Fantilli et al. / Construction and Building Materials 127 (2016) 692–701
aggregate/cement paste interaction in the two mixtures [14]. However, it is worth noting the large ductility of RLC, which is remarkably higher than that of TLC. This is particularly evident in Fig. 8, where the normalized stress vs. compressive strain curves of both TLC and RLC mixtures are compared. In this Figure, the two lightweight concretes are divided into three groups, depending on the
699
amount of fibers (i.e., TLC_0 and RLC_0 in Fig. 8a, TLC_7 and RLC_5 in Fig. 8b, TLC_10 and RLC_12 in Fig. 8c). As illustrated in Fig. 8a, the fracture toughness of RLC_0, higher than TLC_0, can be compared with that of traditional normal weight concrete containing short fibers randomly dispersed within the cementitious matrix [12]. Thus, it can be argued that the rubber aggregates exert a sort of passive confinement in compression, which is not noticed in the presence of expanded clay aggregates. Obviously, the fracture toughness increases with the fiber content, even if the RLC mixtures always show the higher ductility. As a matter of fact, at the same level of strain, the normalized postpeak residual stresses of RLC always exceed those of TLC, regardless of the fiber content (Fig. 8b and c). 4.2. Analysis at the structural level Despite the lower compressive strength measured in the rubber lightweight concrete, the flexural strength of RLC is comparable with, and sometimes higher than, that of TLC. Thus, the mechanical properties of concrete elements subjected to bending actions can be related to the compressive strength, but not in the same manner for all the concretes. With respect to TLC_0, the higher values of Pcr⁄ measured in RLC_0 plates (see Table 4 and Fig. 7a) is also reported in the traditional normal weight concrete when fibers are added to matrix [15]. This further corroborates the conjecture that rubber aggregates contribute to increase concrete toughness, similarly to a fiber-reinforcement. To quantify the beneficial effect of rubber, the flexural tensile strengths flct,fl of both the lightweight concretes are elastically evaluated from the three point bending tests on plates. In this way, a direct comparison with the corresponding compressive strength can be performed by introducing the following strength ratio:
r¼
f lct;fl M cr 1 ¼ f lc W f lc
ð2Þ
where Mcr⁄ = Pcr⁄ l/4 = effective cracking moment at the midspan cross-section of the plate (where l = 900 mm in Fig. 1b); and W = b h2/6 = section modulus of the plate (where b = 1000 mm, and h = 100 mm in Fig. 1b). According to Eq. (2), the average values of r are herein computed by using the experimental values of Mcr⁄ and flc and then reported in Fig. 9 and Table 4. The positive contribution of rubber aggregates to the flexural strength of lightweight concrete is particularly evident in the histogram of Fig. 9. Indeed, the values of r computed for TLC plates might be more or less doubled to obtain the strength ratio of RLC plates. To be more precise, r = 10–13% in TLC, and r = 25–28% in RLC (Table 4). This is due to the effect produced by the rubber granulates, which is macroscopically similar to the bridging action exerted by the fibers crossing the crack surfaces. 4.3. Ecological and mechanical assessment By introducing the so-called Ecological Index (EI) and Mechanical Index (MI), a more comprehensive comparative analysis, including the environmental impact of the two lightweight mixtures, can be performed [16]. The following equation estimates EI by combining the amount of the released carbon dioxide (i.e., a = carbon footprint) and the energy used in the product lifecycle (i.e., b = embodied energy):
EI ¼ a b Fig. 8. Normalized stress-strain curves of the lightweight concretes in compression: (a) TLC_0 and RLC_0 mixtures; (b) TLC_7 and RLC_5 mixtures; (c) TLC_10 and RLC_12 mixtures.
ð3Þ
Both a and b are measured here for the most relevant concrete components, such as cement, lightweight aggregate (made with
700
A.P. Fantilli et al. / Construction and Building Materials 127 (2016) 692–701
Fig. 9. The average values of the strength ratio r of both TLC and RLC mixtures.
expanded clay or grained rubber from end-of-life tires) and polypropylene fibers. Superplasticizer is not considered in the computation of EI, because it has a negligible environmental impact with respect to the other components of TLC and RLC mixtures. Table 5 collects all the data necessary to compute EI in accordance with Eq. (3) [17–20]. It must be remarked that Eq. (3) takes into account environmental aspects in a simple way. To perform more complex sustainability analyses, including social and economic issues, other methods can be found in the current literature (e.g., MIVES approach [21]). As the mechanical index must be related to a functional unit, three analyses are carried out herein: MI1 = flc at the material level, MI2 = Pmax = max (Pcr⁄; Pu) and MI3 = DI at the structural level (see Table 6). In the first two cases, the ecological index of TLC_0 (Table 6) is used as the upper bound value of the environmental performances (EITLC_0). Whereas, the lower bound values of the mechanical index are assumed to be MI1,TLC_0 = 16 MPa (i.e., the minimum compressive strength for structural applications suggested by MC2010 [1]) and MI2,TLC_0 = Pd = 8.60 kN (i.e., the design load of the plates). Referred to these bounds, two eco-mechanical analyses can be carried out for all the mixtures by means of the non-dimensional diagrams depicted in Fig. 10a–b [16]. With respect to MI1 = flc (Fig. 10a), neither TLC nor RLC mixtures contemporarily fulfil the minimum mechanical and environmental Table 5 The environmental impact of concrete components (referred to 1 kg of material). Components
Carbon footprint (kg CO2)
Embodied energy (MJ)
Cement Expanded clay aggregate Rubber granulates Polypropylene fibers
0.8 0.3 0.2 2.7
5 4 4 100 Fig. 10. Eco-mechanical analysis of TLC and RLC mixtures: (a) Results at the material scale with MI = MI1; (b) Results at the structural scale with MI = MI2; (c) Results at the structural scale for the ideal mixtures having the minimum amount of fibers Qf,min.
Table 6 The ecological and mechanical indexes of the six lightweight concretes. Mixtures
EI (kg CO2 GJ/m3)
MI1 (MPa)
MI2 (kN)
TLC_0 TLC_7 TLC_10
1803 2219 2432
21.51 23.36 22.91
16.01 22.32 22.93
RLC_0 RLC_5 RLC_12
865 1070 1379
9.84 10.94 10.99
20.24 20.96 20.46
requirements. As a matter of fact, fiber-reinforced traditional lightweight concretes satisfy the condition MI > MI1,TLC_0, but EI > EITLC_0 due to the impact of fibers. Conversely, RLC only fits the ecological performances (with and without the fibers), but it does not achieve the minimum strength required for structural use.
A.P. Fantilli et al. / Construction and Building Materials 127 (2016) 692–701
When MI2 = Pmax, TLC only meets the mechanical requirement (i.e., MI > MI2,TLC_0), whereas both EI < EITLC_0 and MI > MI1,TLC_0 are satisfied by all the rubber lightweight concretes investigated in this paper. Among them, RLC_0 seems to be the best cementbased composite to be used in one-way plates. Indeed, EI increases in presence of fibers (see RLC_5 and RLC_12 in Table 6) whereas the values of Pmax are nearly the same for all the RLC plates. Nevertheless, the brittle behavior of RLC_0 plates (see Fig. 5d) cannot be accepted for structural applications. In other words, concrete beams and plates in bending have to show a minimum ductility (i.e., DI P 0) [13]. Thus, the proposed comparative analysis involves the ideal TLC and RLC plates containing the minimum amount of fibers (i.e., Qf,min = 9.2 kg/m3 for TLC, and Qf,min = 14.2 kg/m3 for RLC) graphically evaluated in Fig. 7c and d, respectively. Under this condition, their performances are only compared in terms of EI (see Fig. 10c), because MI3 = DI = 0 for both the ideal plates. As a result, RLC (which shows EI = 1493 kg CO2 GJ/m3) performs better than TLC (with EI = 2380 kg CO2 GJ/m3) when the minimum amount of fibers is added to the mixtures. 5. Conclusions The results of the experimental investigations, concerning the behavior of two different lightweight concrete mixtures, can be summarized by the following points: 1. Due to a higher water/cement ratio and to the presence of rubber granulates, which substitute the expanded clay aggregates, RLC is more environmental-friendly than TLC, but the strength class reduces, regardless of the amount of plastic fibers. Nevertheless, uniaxial compression tests on cylinders reveal the higher ductility of RLC, even in the absence of fiberreinforcement. 2. On the contrary, the flexural strength of TLC and RLC is comparable, as measured in the three point bending tests on unreinforced and fiber-reinforced plates. 3. Like the fibers, rubber granulates exert a sort of passive confinement in compressed elements, and prevent the rapid growth of cracks in the tensile zones of beams and plates. 4. If also the environmental impact is taken into account, the results of an eco-mechanical analysis depend on the functional unit used to compare the concretes. When compressive strength (i.e., a material property) is the reference performance, RLC fits the ecological requirements both with and without the fibers, but it does not achieve the minimum strength for structural uses, especially in compressed elements. 5. On the contrary, if the analyses are referred to structural properties, such as the maximum load Pmax of bending plates, substitution of expanded clay aggregates with rubber granulates is convenient. This is particularly true when the minimum condition of ductility (i.e., DI = 0) is also taken into account. Finally, the above-mentioned differences suggest the necessity of a proper definition of the functional unit to which the inputs
701
and outputs of a life-cycle assessment must be referred. It would be desirable that this function, and the ecological benchmarks of concrete structures as well, are reported by the current building codes. Acknowledgments The grants given by the Italian Laboratories University Network of seismic engineering (ReLUIS) and by the Ecotyre Consortium for the end-of-life tires disposal, and used to develop the present research work, are gratefully acknowledged. References [1] International Federation for Structural Concrete, Model Code 2010 – Final draft. fib, Lausanne, 2012, fib Bulletin 65–66. [2] American Concrete Institute, Guide for Structural Lightweight-Aggregate Concrete, ACI, Farmington Hills, 2003. Committee 213. [3] A.P. Fantilli, A.D. Cavallo, G. Pistone, Fiber-reinforced lightweight concrete slabs for the maintenance of the Soleri Viaduct, Eng. Struct. 99 (2015) 184– 191. [4] R. Siddique, T.R. Naik, Properties of concrete containing scrap-tire rubber – an overview, Waste Manage. 24 (2004) 563–569. [5] K.B. Najim, M.R. Hall, A review of the fresh/hardened properties and applications for plain-(PRC) and self-compacting rubberised concrete (SCRC), Constr. Build. Mater. 24 (2010) 2043–2051. [6] European Committee for Standardization, Materials Produced from End of Life Tyres – Specification of Categories Based on their Dimension(s) and Impurities and Methods for Determining their Dimension(s) and Impurities, CEN, 2010, CEN/TS 14243. [7] B.L. Damineli, F.M. Kemeid, P.S. Aguiar, V.M. John, Measuring the eco-efficiency of cement use, Cement Concr. Compos. 32 (2010) 555–562. [8] G. Habert, N. Roussel, Study of two concrete mix-design strategies to reach carbon mitigation objectives, Cement Concr. Compos. 31 (2009) 397–402. [9] A.P. Fantilli, B. Chiaia, Eco-mechanical performances of cement-based materials: an application to self-consolidating concrete, Constr. Build. Mater. 40 (2013) 189–196. [10] A. Bentur, S. Mindess, Fibre Reinforced Cementitious Composites, Elsevier Applied Science, London and New York, 1990. [11] P. Pujadas, A. Blanco, S. Cavalaro, A. Aguado, Plastic fibres as the only reinforcement for flat suspended slabs: experimental investigation and numerical simulation, Constr. Build. Mater. 57 (2014) 92–104. [12] A.P. Fantilli, H. Mihashi, P. Vallini, B. Chiaia, Equivalent confinement in HPFRCC columns measured by triaxial test, ACI Mater. J. 108 (2011) 159–167. [13] A.P. Fantilli, B. Chiaia, A. Gorino, Fiber volume fraction and ductility index of concrete beams, Cement Concr. Compos. 65 (2016) 139–149. [14] E. Ganjian, M. Khorami, A.A. Maghsoudi, Scrap-tyre-rubber replacement for aggregate and filler in concrete, Constr. Build. Mater. 23 (2009) 1828–1836. [15] A.E. Naaman, Strain hardening and deflection hardening fiber reinforced cement composites, in: Proceedings of the 4th International RILEM Workshop on High Performance Fiber Reinforced Cement Composites, Ann, Abor, 2003, pp. 95–113. [16] A.P. Fantilli, B. Chiaia, The work of fracture in the eco-mechanical performances of structural concrete, J. Adv. Concr. Technol. 11 (2013) 282– 290. [17] International Federation for Structural Concrete, Guidelines for Green Concrete Structures, fib, Lausanne, 2012, fib Bulletin 67. [18] M.F. Ashby, Materials and the Environment: Eco-Informed Materials Choice, Butterworth-Heinemann (Elsevier), Burlington, 2009. [19] Swiss Centre for Life Cycle Inventories, The Ecoinvent Database. Available from: