Flexural Behavior of Fiber Reinforced Self-Compacting Rubberized Concrete Beams

The massive growth of the automotive industry and the development of vehicles use lead to produce a huge amount of waste tire rubber. Rubber tires are non-biodegradable, resulting in environmental problems such as fire risks. In this search, the flexural behavior of steel fiber reinforced self-compacting concrete (SFRSCC) beams containing different percentages and sizes of waste tire rubbers were studied and compared them with the flexural behavior of SCC and SFRSCC. Micro steel fiber (straight type) with aspect ratio 65 was used in mixes. The replacement of coarse and fine aggregate was 20% and 10% with chip and crumb rubber. Also, the replacement of limestone dust and silica fume was 50%, 25%, and 12% with ground rubber and very fine rubber, respectively. Twelve beams with small-scale (L=1100mm, h = 150mm, b =100mm) were tested under two points loading (monotonic loading). Fresh properties, hardened properties, loaddeflection relation, first crack load, ultimate load, and crack width were investigated. Two tested reinforced concrete beams from experimental work were selected as a case study to compare with the results from ABAQUS program (monotonic loading). These two reinforced concrete beams were simulated as a parametric study under repeated loading using this finite element program. The results showed that the flexural behavior of SFRSCC beams containing rubber was acceptable when compared with flexural behavior of SCC and SFRSCC beams (depended on load carrying capacity). Cracks width was decreased with the addition of steel fibers and waste tires rubber. An acceptable agreement can be shown between the results of numerical analysis and the results obtained from experimental test (monotonic loading). Insignificant ultimate load differences between the results of monotonic loading and repeated loading


INTRODUCTION
The overstocked of waste materials is an unavoidable stage of all industrial and human activities. These wastes create significant environmental and economic problems around the world. Many advantages can be accomplished through waste recycling in other processes, like reduce energy consuming, solve problems of disposal, minimize the use of natural resources (fine and coarse aggregate) also, decrease the health hazards on human and other vital components, (De Brito and Saikia, 2012). The use of alternative materials in concrete opens a whole new range of possibilities in the construction industry. The behavior of self-consolidating rubberized concrete (SCRC) beamcolumn joints under monotonic loading containing steel fibers and replacing fine aggregate with shredded rubber were investigated by (Ganesan, et al., 2013), the percentage of rubber was15% by volume of fine aggregate. The results showed that the addition of shredded rubber improves the behavior of beam-column joint, such as energy absorption ability and ductility. The results also showed the presence of steel fibers and rubber particles improves resistance of crack and loadcarrying capacity. The behavior of eight beams with intermediate scale (1700mm × 200mm × 100mm) containing waste tire rubbers by using four types of concrete, normal concrete, rubberized concrete (RC), self-compacting concrete (SCC), and self-compacting rubberized concrete (SCRC) was presented by (Najim and Hall, 2014), the replacement of fine aggregate was 14% for RC and 18% for SCRC with crumb rubber. They observed that using crumb rubber decreased the flexural capacity and flexural stiffness. On the other hand, the deformability and absorption of energy increased with the crumb rubber increased. The flexural behavior of SCRC (self-compacting rubberized concrete) beams with fullscale (2440mm×250mm×250mm) was studied by (Ismail and Hassan, 2015), the percentage of crumb rubber ranging from 5% to 15% by volume of sand (fine aggregate). The behavior of the specimens was evaluated by using load-deflection relation. The tests of beams showed that increasing the rubber content decreased the first crack load, stiffness, and density, but the percentage of crumb rubber up to 10% improved deformability, ductility, and toughness of tested beams with slightly reducing in the flexural capacity. Many researchers studied fresh and hardened properties of SCC containing waste tire rubbers. Fine aggregate was replaced with crumb rubber (5, 10, and 15% by weight of fine aggregate). Fresh tests (slump flow, V-shape, L-shape, U-shape, and J-ring) and hardened tests (compressive strength, splitting tensile strength, and flexural strength) were carried out at 7, 28 and 90 days of curing by (Padhi and Panda, 2016). The results showed that the incrementation in the proportion of rubber decreases the workability of SCRC. The results also showed that the replacement of fine aggregate by rubber particles decreases the hardened properties such as compressive strength. The flexural strength increased when the percentage of crumb rubber was 5%. (Matar and Assaad, 2019), investigated the workability and strength of self-compacting concrete containing recycled aggregates and polypropylene fibers. Different SCC mixes containing 25% to 100% recycled aggregate (RCA) and 0.25% to 1.75% polypropylene fiber (PPF) were prepared. The results showed that the fresh properties decreased with RCA, and PPF content increased. The compressive strength increased slightly when PPF increased. The reduction in splitting tensile strength due to RCA can be overcome by using PPF. The use of finite element analysis (FEA) has increased because of the progressing knowledge and capability of computer package and hardware. ABAQUS has two approaches for modeling concrete response; smeared cracking and damaged plasticity. A concrete damaged plasticity model is suitable for various loading conditions such as monotonic loading, cyclic loading, and dynamic loading. This approach takes into account the degradation of the elastic stiffness resulting from plastic straining both in compression and tension. From the reason mentioned above, the damage plasticity model has been used for analysis the steel fiber reinforced self-compacting rubberized concrete beams (Chaudhari and Chakrabarti, 2012). The main aim of the present study is to produce successful mixes of steel fiber reinforced selfcompacting rubberized concrete (SFRSCRC) in fresh and hardened properties (using waste tires rubber). Also investigate the effect of using different percentages of waste tire rubber (20% and 10% by weight of aggregates, 50%, 25%, and 12% by weight of limestone dust and silica fume) on behavior of simply supported beams under monotonic loading with the consideration of the following: 1) Load and deflection at first crack 2) Load and deflection at failure 3) Ductility, flexural stiffness, and residual strength factor. The behavior of simply supported beams under repeated loading by using the finite element program (ABAQUS) is also investigated.

Materials Properties
Ordinary Portland cement (type I) was utilized for all the mixes. Fine aggregate (zone 2) and coarse aggregate (maximum size10 mm) were used according to (IQS. No. 45/1984). The fine and coarse aggregate has specific gravity 2.65 and 2.6, respectively. Grey powder (silica fume) and white fine material (limestone dust) were used as pozzolanic and filler materials. The results of chemical and physical tests of silica fume satisfied with the (ASTM C1240-15) requirements. Viscosity modifying admixture (VMA) with specific gravity 1:1 and PH value 6.5 was used to obtain water reduction, workability, and viscosity. Micro steel fiber with length of 13 mm, aspect ratio 65, volume fraction (Vf = 1.5%), and tensile strength 2600 MPa as shown in Fig. 1. Waste tire rubber was prepared with different sizes, (1.18-9.5) mm was used instead of coarse aggregate (chip rubber), (0.15-4.75) mm was used instead of fine aggregate (crumb rubber), 125 and 2.5 microns were used instead of limestone dust and silica fume (ground rubber and very fine rubber). The different sizes of rubber with specific gravity 1.78 and water absorption 2% were obtained from General Company for Rubber Industries and Tires/ Iraq (Fig. 2). Two sizes of steel reinforcement (Ø6mm and Ø4mm) were used to reinforce the beams with yield tensile stress 520 and 565 MPa, respectively (According to the (ASTM A496-02)).

Tests of Fresh Concrete
The tests of slump flow, V-funnel, and L-box were carried out to obtain successful mixes in fresh properties of SCC (flowability and ability of passing). The tests were carried out according to the European guideline for SCC (EFNARC 2005). Fig. 3 shows the tests of fresh concrete.

Element Type
The element types used for modeling the simply supported beams using ABAQUS software are summarized in the following sections: 3.1.1 Modeling of concrete Beams are modeled using three-dimensional finite elements. Standard 3D stress elements in ABAQUS can be utilized for modeling of concrete. An 8-node linear brick (C3D8R element) is used to model concrete beams. The integration point of the C3D8R element is located in the middle of the element. Fig. 5 shows the 8-node brick element with the integration point.

Input Data
The beams selected from experimental work are LS12 and SF12. The parameters obtained from laboratory tests such as compressive strength and splitting tensile strength were used in this program. For other parameters needed, such as dilation angle, eccentricity, / , , and viscosity parameter, ABAQUS default data were used. The value of fractions between concrete and supports obtained after many trails to reach the number that reduces the difference between the experimental and finite element results. This value was assumed 0.2. The input data of beams LS12 and SF12 are summarized in Table 3. Table 3. Input data of beams LS12 and SF1.

Three Dimensional Finite Element Meshes
The beam was meshed (divided) into a number of small finite elements with maximum size 30 mm as shown in Fig. 8; also the modeling of steel reinforcement is shown in Fig. 9. The beam was simply supported at both ends. One support was modeled as a roller by constraining in Y-direction (UY=zero), another support was modeled as a hinge by constraining in X, Y, and Z-direction (UY=UX=UZ= zero) as shown in Fig. 10. The beam was subjected to 12 mm displacement condition for monotonic loading. For repeated loading the beam was subjected to displacement equal to displacement obtained from monotonic test (experimental program).

Fresh properties
The results of all mixes showed that the diameter of slump flow, time of V-funnel, and the blocking ratio of L-box belong under acceptance criteria for the European guideline. The workability of steel fiber reinforced self-compacting concrete mix (RSCC) decreased because of the presence of micro steel fibers. This reduction in workability due to steel fiber obstructs the movement of the mix and increases the friction between aggregate and fibers. It was also noticed that the workability decreased by adding different percentages of rubber. This diminution in workability is due to the low specific gravity of rubber particles relative to the specific gravity of other materials (coarse aggregate, fine aggregate, limestone dust, and silica fume). This leads to low mobility of mixes. The water to cement ratio (w/c) and VMA dosage increased by adding micro steel fibers and rubber particles. Therefore, all mixes satisfy the acceptance limits of the European guideline. The reduction in workability includes decrease in slump flow diameter, an increase in flow time, and a decrease in blocking ratio (H2/H1). The results of fresh properties are presented in Table 4. Table 4. Results of fresh concrete tests.

Hardened properties
The results showed that the compressive strength was decreased as rubber content increased. The reduction in compressive strength about 37.1%, 11.4%, 41.8%, 14.7%, 16.8%, 39.0%, and 10.5% for mixes CA20, CA10, FA20, FA10, LS50, SF50, and SF25, respectively compared with reference mix (RSCC). The reason for this reduction is due to the soft particle of waste tire rubber compared with particle of aggregates or other materials. The adhesion between rubber and cement paste is weak (poor strength of the interfacial transition zone between the rubber particles and cement paste). Also, the reduction in splitting tensile strength and flexural strength attributed to the same reasons affected the compressive strength. The compressive strength was increased about 14.4%, 15%, and 28% in SF12, LS25, and LS12, respectively.  Table 5. Results of hardened concrete tests.

Flexural test results of beams
The results acquired from the flexural testing of the beams are illustrated in Table 6. The crack pattern of all the beams indicates that the failure mode of the beams is a flexural failure mode (Fig.  11). During the first stage of testing, small perpendicular cracks formed in the mid-span of all the beams. The number of these cracks was increased when the applied load increased. The results observed that the width of the crack reduces by adding steel fiber and rubber particles due to the micro steel fiber block these cracks and restricted their widening (Al-Quraishi, et al., 2017) and (Muhsin and AbdElzahra, 2016), also the capacity of rubber particles to absorb higher energy. The flexural stiffness can be defined as the slope of the load-deflection relation (K= ΔF/Δδ). The addition of micro steel fiber improves the flexural stiffness values and reduces the beam deformability, but these values decreased as rubber contains an increase. The decrease in K value is due to the lower modulus of elasticity of rubber particles (improve deform capacity). The results showed that the addition of steel fiber in B2 improves the failure load by about 31% compared to B1 but by adding rubber particles. This percentage dropped nearly 4% in B3, B6, and B12, also decreased approximately 21.7%, 2%, 32.5%, and 9.8% in B5, B8, B10, and B11, respectively. The results also showed improved failure loads in B9 about 2% compared to B2. The presence of steel fiber and waste tire rubber convert behavior of beams from brittle to ductile as shown in Fig. 12. Ductility can be defined as the ability of the structural member to undergo plastic zone before failure (area under the load-deflection curve between first crack and peak failure load). For all beams the ductility value increase as rubber content increase except in B10, B11, and B12. The addition of steel fiber and silica fume lead to increase ductility (post crack resistance), (Nili and Afroughsabet, 2010). Toughness indices and residual strength factors were also computed according to (ASTM C1018-97). Toughness indices (I5, I10, and I20) increase as rubber content increases due to the lower stiffness (increase flexibility and energy absorption). Residual strength factors, which are derived directly from toughness indices. It means the level of strength retained after first crack. Micro steel fibers and rubber particles increase residual strength factors. Table 7 presents the ductility values, toughness indices, and residual strength factors. Table 7. Ductility, toughness indices, and residual strength factors.

Monotonic loading
The numerical analysis results for beams LS12 and SF12 under monotonic loading are illustrated in Table 8. This table also presents the different percentages between experimental and numerical analysis results. The comparison between the numerical and experimental load-deflection curves are presented in Fig. 13. An acceptable agreement can be noticed from this figure.
The ultimate load obtained from FEA is higher than the ultimate load obtained from experimental work. This is due to the right modeling of steel fibers (tension stiffening). Also, it is due to the rubber effect, which represents increasing compressive strength (49.32 MPa and 42.22 MPa) only without taking into account the change in structure of concrete. The structures of the concrete greatly affect the behavior of the beams in experimental work. The ultimate load obtained from finite element analysis was 32.70 kN with central-deflection 6.14 mm and 31.08 MPa with the central-deflection 6.01 mm for LS12 and SF12, respectively. The different percentages between experimental and numerical analysis results were 7.93% and 12.32% for ultimate load, 8.87% and -2.89% for mid-span deflection for LS12 and SF12, respectively. Fig. 14 and Fig. 15 shows the deformation shape and crack pattern for these beams respectively.

Repeated loading
The load history (cyclic load), as shown in Fig. 16 was applied to the specimens (LS12 and SF12). This load history was recommended by (FEMA, 2007). Figure 16. Load history used in FEA.
The numerical analysis results for beams LS12 and SF12 under repeated loading are illustrated in Table 9. This table also presents the different percentages between monotonic and repeated results (FEA). The load-mid span deflection relation obtained from numerical analysis by ABAQUS under monotonic and repeated loading is presented in Fig. 17 of beams LS12 and SF12. Fig. 18 shows the deformation shape of these beams.

LS12 SF12
Figur18. Deformed shape (FEM). Table 9 and Fig. 17 shows that the differences between the results of monotonic loading and repeated loading (ultimate load) were insignificant differences. The addition of steel fibers improves the load-bearing capacity under repeated loading with large strain before failure, (Jun and Stang, 1998). The load-deflection curve (FEA monotonic test) showed the ductile behavior of beams LS12 and SF12, therefor the ultimate load obtained from repeated loading is close to the ultimate load obtained from monotonic test. The percentage of difference between the ultimate loads obtained from monotonic and repeated loading about 1.83% for LS12 and 3.02% for SF12. Also the different percentages between the deflections at ultimate load about 21.01% for LS12 and 15.26% for SF12.

CONCLUSIONS
These conclusions were obtained on the basis of the results obtained by testing the specimens: The fresh properties of SFRSCRC decreased as rubber content increased. The engineering properties decreased with increased rubber content when rubber used as aggregates but these properties increased when rubber used as filler materials. Flexural stiffness and load at failure decreased when the percentage of waster tire rubber increased, for all beams. The results observed that the deformability, ductility, and toughness indices increased with rubber content increased. The failure mode of the tested beams converts from brittle to ductile by adding rubber particles and micro steel fibers. The optimum percentage of rubber replacement was 10% for coarse and fine aggregate, and 12% for limestone dust and silica fume (depended on the results of fresh properties, hardened properties, and behavior of beams). The results obtained from FEA by using ABAQUS program showed that acceptable percentages of differences between the numerical analysis results and the experimental results. The percentages of differences between beams LS12 and SF12 was 7.93% and 12.32% for ultimate load, 8.87% and -2.89% for deflection at ultimate load (monotonic loading). Insignificant ultimate load differences between the results of monotonic loading and repeated loading. The percentage of difference between the ultimate loads obtained from monotonic and repeated loading was about 1.83% for LS12 and 3.02% for SF12.