On the mechanical behavior of a hybrid reinforced concrete for industrial floors

Abstract Civil construction is an industry sector that has been used as an outlet for the reuse of industrial waste. The present work aims to use the residue of Ethylene Vinyl Acetate (EVA) from the footwear industry as a partial substitute for a granulometric range of aggregates, aiming at the production of structural concrete and application to industrial floors. The proposed mixing ratios were evaluated from uniaxial compression, three-point bending, and drying shrinkage tests. The results of the uniaxial compression tests showed that the concrete with EVA addition still has enough strength to be considered structural concrete. In addition, the EVA and polypropylene fiber particles act as stress transfer bridges in the cracked zone, resulting in an increase in residual stresses and, consequently, in the toughness of the concrete in the three-point bending test. Finally, Technical Report 34 was used as a procedure to design an industrial floor based on the compressive strength, Young's modulus, and flexural behavior of the tested composites. The final result showed that even with lower compressive strength, fiber-reinforced concrete with EVA achieves greater structural efficiency for an industrial floor with the same cross-sectional height as ordinary fiber-reinforced concrete.


INTRODUCTION
The industrialization process along with demographic and technological growth brought, in addition to the benefits, an increase in the generation of solid waste, causing environmental problems [1]- [3].Because of this, many industry sectors are valuing the search for eco-friendly solutions to dispose of their waste.In the construction industry, for example, many researchers are concerned about how to reuse demolition waste as inert materials in the mixture of mortar and concrete [4]- [6].
In addition to the demolition waste, advances can be seen in the production of lightweight concrete using other types of waste as lightweight aggregate, which can lead to a decrease in compressive strength, a parameter that directly influences the quality of the cementitious material [7]- [11].In this case, Ethylene Vinyl Acetate (EVA) is a residue that has been widely proposed for research focused on lightweight concrete formulations with permeable characteristics [12], [13] or for bituminous concretes [14], [15].Most applications for this type of material are focused on paving [16], [17], and its use as an addition in structural concrete is a gap in the literature.
The use of plastic materials, such as the EVA mentioned above, has aroused much interest due to the difficulties encountered in disposing of the waste.As highlighted by Lima et al. [2], storage or disposal in landfills implies the availability of large surfaces, and incineration can emit potentially harmful gases and ashes.Because of this, the use of thermoplastic materials has become widespread in various sectors of the industry, as they can be melted in the recycling process [18].However, EVA is not characterized as a thermoplastic material, but as a thermosetting material and, therefore, cannot be melted by heat due to its molecular chains.
In addition to EVA, some other plastic wastes have been destined for use as inert materials in concrete, i.e. polyethylene and polystyrene [19], [20], high-density polyethylene [21], polyethylene terephthalate (PET) [22], polyvinyl chloride (PVC) [23].In general, the increase in the plastic content added to the mixture is inversely proportional to the density and strength of the cementitious material [24]- [27].Li et al. [28] evaluated the variation of the apparent density and the incorporated air content of the cement paste and mortar samples with EVA additions.They noticed that when increasing the EVA content in the cement mixtures by up to 6%, there was a decrease in the apparent density and a significant increase in the content of incorporated air.Furthermore, Li et al. [28] pointed out a direct correlation between the EVA content and the compressive strength, showing that there is a reduction in the compressive strength with the increase of the amount of EVA in the mixture.Plastic aggregates reduce the strength of concrete when compared to ordinary concrete, mainly due to the low Young's modulus of plastic materials [26].Therefore, the deformability of the material prevents the aggregate from functioning as a rigid core, not restricting displacement in its vicinity [27].The impact of this substitution on Young's modulus is significant, ranging from 21 to 42% at levels of substitution up to 100% [27], [29].
On the other hand, some studies point out that the presence of plastic particulates can serve as bridges in cracks, mitigating the propagation of micro cracks and contributing to the toughness of the material [2], [30].The EVAmodified concrete, using addition of EVA powder to the concrete mix, increases the flexural and tensile strength by incorporating up to 16% of EVA powder by weight of cement [31].Thus, despite the impact on the strength of concrete, this material can have suitable applications, such as in situations where crack opening must be controlled or in cases where the structure is subject to impact loads [30].
Brovelli et al. [32] studied the influence of EVA on the rheology of bituminous concrete, comparing it with reference cases (without EVA addition).For all investigated samples, the addition of EVA increased the yield point and the plastic viscosity of the mixture.Furthermore, Ben Fraj et al. [33] performed slump tests comparing samples of ordinary concrete and concrete with partial replacement of aggregates by polyurethane.The results showed that the increase in rheological parameters observed by Brovelli et al. [32] was translated into a significant reduction in a slump test.They attributed this behavior to the fact that polymeric aggregates are quite porous, which facilitates the process of absorbing water from the mixture and cement paste, directly affecting workability.In addition to this mechanism, Alqahtani and Zafar [1] attribute the reduction in a slump to the increase in friction between the plastic particles, which leads to less fluidity.The large surface area of the particles coupled with the lack of uniformity in the geometry of the particles causes this friction [19], [34], [35].
Finally, aiming at the use of concrete with plastic aggregate for structures supported on an elastic base, i.e. industrial floors, it is important to understand the impact of replacement on the abrasion resistance and drying shrinkage of the material.In general, most studies highlight a significant increase in drying shrinkage for replacements of up to 80% [23], [36], [37].This increase is attributed to the low rigidity and compressibility of the plastic material, which results in a low restriction to the shrinkage deformations induced by the cement paste [36], [37].On the other hand, in the case of abrasion, Alqahtani et al. [24] showed that the abrasion resistance of concrete with indirect replacement of aggregates showed an increase of more than 50%.Alqahtani and Zafar [1] highlighted that the low strength and roughness of the plastic aggregate, when compared to the common aggregate, was the main responsible for this increase in abrasion resistance.
Moreover, discrete synthetic fiber addition, such as polypropylene (PP), has been used primarily to reduce shrinkage effects [38].Fibers also mitigate crack propagation and improve impact resistance, flexural and tensile strengths [39]- [41], toughness [41], [42], and abrasion resistance [43].For use in concrete pavements, where high flexural strength is required, PP fibers contribute to reducing the concrete design thickness, and the pavement integrity remains even after the crack formation [44]- [46].
Hybrid EVA particle and PP fiber-reinforced concretes have been also addressed in the literature [47], [48].The flexural performance of PP-EVA hybrid reinforced concrete is improved as the maximum and residual flexural strength increase up to 11% when compared to ordinary concrete, and up to 4% when compared to polypropylene fiber reinforced concrete (PP-FRC) [47].Besides, an improvement of around 40% in the toughness can be seen in the PP-EVA hybrid reinforced concretes compared to a PP-FRC with the same fiber content [47].
Although several studies have already addressed the basic properties of concrete with the replacement of aggregates by recycled plastic aggregates and also the fiber inclusion, few works have sought to replace an existing granulometric range, and none clearly addressed the use in industrial floors, adding a fiber content consistent with the proposed field application.Knowing that the replacement directly affects the strength and elastic properties of the cementitious material, this work proposes the use of polypropylene plastic fibers as a way to improve the control of crack propagation, directly impacting the ability to absorb energy.In addition, a comparison between an industrial floor calculated using the properties of ordinary concrete is compared to the proposed material.

Materials
The cement used was a Portland cement CP V ARI from the manufacturer Lafarge Holcim, which is equivalent to ASTM type III according to Natalli et al. [49].This cement type, according to the Brazilian standard NBR 16697/18 [50], has a minimum compressive strength of 34 MPa at 7 days when mixed in a standard 1:3 mortar, and the results for the used cement batch are presented in Table 1.In addition to the mechanical characteristics, chemical tests were carried out on this batch of cement, and the results are shown in Table 2.  Two coarse granite aggregates were included in the mix design: one with a maximum diameter of 9 mm and the other with 19 mm.In addition, natural quartz sand and mineral stone sand were used as fine aggregates.Natural quartz sand has a maximum characteristic dimension of 4.75 mm and a fineness modulus of 2.74, and mineral stone sand has a maximum dimension of 2.40 mm and a fineness modulus of 2.02.Therefore, both are classified as fine aggregates in the optimal zone according to NBR 7211 [51].In addition, EVA was also used as a mixing material.The EVA used is a waste from the footwear industry in the countryside of the state of Ceará in Brazil, and it is found in a carved condition with a granulometry close to the 9 mm diameter coarse aggregate mentioned above.The granulometry of the aggregates and EVA are shown in Figure 1.Although the EVA particles present a granulometric curve similar to the 9 mm coarse aggregate, their shape is quite irregular and predominantly two-dimensional, as seen in Figure 2. Therefore, evaluation of the shape of the particles with the aid of a Nikon microscope model SMZ800N and through image processing using the Image J software was done.Figure 2 shows the segmentation performed and the binary images generated for the total area of the particles analyzed, with a resolution of 1596x1064 pixels.Then, data related to the analysis of particles was obtained through the segmented image.The data obtained are summarized in Table 3.The results obtained from five EVA particle samples show that the EVA particles have a total area of 78 ± 6.18 mm 2 , which means 46.01 ± 3.61% of the total area considering the image background.Due to the irregularity of the particles, it can be noted that the perimeter varies considerably concerning the calculated average, with values of 39.36 ± 17.44 mm, different from the low standard deviation found for the particle area values.This consideration becomes even clearer when the circularity of the particles is evaluated, which resulted in 0.52.The circularity parameter is a dimensionless measure obtained through a metric used in image processing and analysis techniques, defined by C=4π × (A/P2), where A is the area of the particle and P is the perimeter of the particle.The reference value of the circularity parameter is 1, representing the perfect circle.Therefore, as the value approaches 0, an increasingly elongated shape is indicated.Furthermore, an X-ray Fluorescence analysis (XRF) was performed on the EVA, and the elements and their respective amounts found are presented in Table 4.The polypropylene (PP) fibers used in this work were the 51 mm length Viapol TUF-Strand SF ® with an aspect ratio of 74.The main characteristics and properties of these macro fibers are presented in Table 5.The fibers have twisted geometry, which aims to facilitate the mixing procedure, as well as improve the mechanical anchorage at the fiber-matrix interface.In addition, each fiber is the result of the union of about 3 strands grouped, giving the fiber irregular cross-sections along its length, as seen in Figure 3.

Mixing design
The dosage of the structural concrete matrix used is shown in Table 6, considering the mixture proportion of ordinary concrete (OC), concrete with the volume replacement of coarse aggregate by EVA (EVA-C), PP fiber reinforced concrete (FRC), and the concrete with replacement of coarse aggregate by EVA reinforced with PP fiber (EVA-FRC).The mixing procedure was started by adding the aggregates (quartz sand, mineral sand, coarse aggregates, and EVA) which were mixed with 70% water for one minute.Then, cement was added to the mixture and allowed to mix for another minute.The remaining water was added to the mixture with the superplasticizer chemical admixtures, which was used to give better workability to the concrete.The superplasticizer admixture used was PLASTOL® 4100, produced by the manufacturer VIAPOL, characterized as type SP-II N by the NBR 11768 standard [52].The superplasticizer content was adjusted from the slump test following NBR NM 67 [53] before starting the specimens molding process.For the mixtures containing polypropylene (PP) fibers, a final step was added to the procedure to throw and let the fiber disperse, following the time target stated by Lima et al. [41].All samples were cured in a room with controlled temperature and humidity (23°C and 100%) for 28 days before mechanical testing.On the twenty-eighth day, the specimens were removed from the curing room and taken to the specimen preparation room.Finally, on the twenty-ninth day, the mechanical tests commenced.

Density measurements
The variation of concrete density was measured by weighing cylindrical specimens measuring 10 x 20 cm.For this test, the samples destined for the compressive strength test were weighed before the destructive test.Knowing the weight and volume of the sample, the approximate apparent density of the materials was calculated.It can be seen in Table 7 that the density variation was not so significant, and this is since only the coarse aggregate with a 9 mm maximum particle size was completely replaced by the EVA aggregate.Thus, the coarse aggregate with a maximum particle size of 19 mm and the natural and mineral sands remained as constituent materials of the mixture.Therefore, in terms of weight, there is a 10% weight relief of concrete with EVA (EVA-C) compared to ordinary concrete (OC).

Mechanical tests
The experimental mechanical characterization and testing of concretes evaluated include measurement of uniaxial compressive strength and a 3-point bending test controlled by the crack mouth opening displacement.From the mechanical results obtained, it is possible to verify the application of the material developed for industrial floors, drawing a comparison between the OC and the EVA-C.

Uniaxial compressive strength
Uniaxial compressive strength test and Young's modulus estimation were done at 28 days, according to NBR 5739 [54] and NBR 8522 [55] standards.Cylinders of 100 mm in diameter and 200 mm in height were cast.Before this process, the molds were prepared with a release oil coat for then the concrete to be poured.Its densification was done by external vibration, initially with a rubber hammer and then on a vibrating table.After 24 hours, the specimens were demolded and taken to an environment of controlled temperature and humidity (humid chamber).Approximately 24 hours before the test, the specimens went through the process of gridding the base and the top to ensure the regularity of the surface and parallelism between the faces.It ensures that the transfer of stresses during loading is uniform.
The tests were carried out in a servo-controlled Controls testing machine model MCC8, with a load capacity of 2000 kN and a loading speed of 0.35 MPa/s.For higher precision of Young's modulus, two vertical displacement transducers (LVDT) attached to cylindrical rings were fixed around the specimen to measure the deformation, as can be seen in the scheme of Figure 4.In this way, the value of the specimen's deformation was obtained by the relation between the average value of the measured relative displacement and the initial length of reference given by the distance between the rings.Stress was calculated as the ratio between the applied load and the cross-sectional area of the specimen.Young's modulus corresponds to the slope of the initial region of the stress versus strain curve, up to approximately 30% of the maximum stress.

Three-point bending test
Three-point bending tests were performed on prismatic elements of rectangular sections, following the standard EN 14651 [56].The dimensions of the specimens were 150 x 150 x 550 mm, with a 25 mm high notch in the central region of the lateral face of the prisms made with a 3 mm thick diamond saw.For molding the concrete samples, the mixture was carried out in a concrete mixer with a capacity of 400 l, following the procedure described in section 2.2.
To assemble the test apparatus, support rollers were first positioned.The rollers were 37 mm in diameter, with a 500 mm span, and 25 mm from the edges of the prisms.The rollers had freedom of horizontal movement, however, the left roller had also freedom of transverse movement.This freedom for horizontal displacements prevents the induction of horizontal reactions in the supports and, consequently, the appearance of an eccentric normal force in the structural element, which would cause a change in the stress distribution.
The load application roller was free to move in the transverse direction and had the same diameter as the support rollers.The load application was centered on the top face of the prism.The test equipment used was an MTS model 244.41 servo-hydraulic actuator with a 500 kN loading capacity.To control the piston displacement, a clip gauge and a flex-test 60 controller were used, resulting in a closed-loop system.
The Crack Mouth Opening Displacement (CMOD) values were obtained using an MTS clip gauge model 632.02B-20, fitted to metallic plates fixed to the specimen in the region of the notch Figure 5.The CMOD was adopted as control parameter of the test, carried out at a displacement rate of 0.05 mm/min, to a crack width of 0.1 mm, and then at a rate of 0.2 mm/min to a crack width of 4 mm.Data were acquired at a frequency of 5 Hz.The geometry of the prismatic specimen and the test setup is shown in Figure 5. From the test results, it is possible to plot the stress versus CMOD curves, considering the stress values obtained from Equation 1.The term P refers to the load, S to the test span, b to the width of the specimen, and h sp is the distance between the top of the notch and the top face of the prism.
The advantage of this kind of test is that, due to the presence of the notch in the central region of the specimen, the crack formation is induced to occur in an orientated way, propagating itself along cross-section.In this way, the deformation is always localized, which minimizes energy dissipation throughout the test specimen.Therefore, all the energy absorbed can be directly associated with the fracture along the notch plane and all the energy dissipated can be correlated with the response of the evaluated concrete.
Furthermore, the fib Model Code [57] recommends the determination of residual stresses f R,1 and f R,3 .The stress f R,1 corresponds to a crack width of 0.5 mm and is related to the serviceability limit state.On the other hand, the stress f R,3 corresponds to a crack width of 2.5 mm and is related to the ultimate limit state.It should also be noted that, according to EN 14651 [56], two other residual stresses refer to the crack openings of 1.5 and 3.5 mm, respectively called f R,2 and f R,4 .The last important parameter that must be obtained through the three-point bending tests is the toughness at the 4 mm crack opening (T 4,0 ), calculated as the area under force per CMOD curve.This parameter indicates the material's ability to absorb energy.

Drying shrinkage test
Drying shrinkage tests were performed according to ASTM C157 [58] and ASTM C490 [59] standards, in order to estimate the magnitude of the stress independent strain in the two types of concrete studied (Figure 6).Three prisms measuring 75 x 75 x 285 mm were molded for each of the mixtures.The molding was performed in the same test room, with controlled temperature and relative humidity (20 ± 1•C and 50 ± 5%).Strain readings were taken once a day from the first measurement to 28 days of age.From then on, measurements were taken weekly.

Considerations for numerical approach
Since the main application of fiber-reinforced concrete is aimed at ground-supported industrial floors, we decided to make a comparison between an industrial floor designed for the conventional FRC concrete and another designed for the EVA-FRC.For this analysis, Technical Report 34 [60] was used, and only the case of internal loading was considered, not taking into account the case of loading at the borders of the floor (Figure 7).

Uniaxial compressive strength
According to the test procedure described in Section 3, uniaxial compressive strength tests were performed for ordinary concrete (OC) and concrete with total replacement of the coarse aggregate of smaller diameter by EVA (EVA-C).In addition, the inclusion of fibers was considered in this test stage, evaluating the behavior of the two concretes with the inclusion of 5 kg/m 3 of polypropylene fiber (FRC and EVA-FRC).The results of compressive strength and Young's modulus are shown in Figure 8.It is noteworthy that the results obtained correspond to the average of the results of 5 tests for each specimen group.In the cases of concrete with aggregate replacement by EVA (EVA-C and EVA-FRC), a drop in compressive strength was noticed when compared to the results of ordinary concrete (OC and FRC).This difference may have been caused by the reduction in the compacity of the aggregates, generating critical areas with stress concentrations in the vicinity of the EVA particles that are significantly more deformable than the conventional aggregate [2], [61].It is known that rock aggregates help in the compaction of the material, improving cohesion and friction angle, and in cases where there are shear stresses, the aggregate helps in interlocking [62], [63].In the case of EVA, meshing and interlocking are impaired, but an improvement in material toughness can be expected, a property that will be evaluated with bending tests.
Yang et al. [64] performed compressive strength tests on cement modified with EVA for underground gas storage, varying the percentage of EVA particle inclusion by volume (1%, 2%, 3%, and 4% EVA).The results showed that there is a decrease in the compressive strength of the samples from 3% for the addition of 1% EVA to 14% for the addition of 4% EVA.Li et al. [65] and Lima et al. [2] addressed in their studies a greater inclusion of EVA by volume, considering 10% addition and 50% replacement of aggregate respectively.Consequently, this level of inclusion of EVA in the mixture resulted in a more significant reduction in the compressive strength of the samples, reaching values about 40% and 70% lower than the reference case.In the present work, as the total replacement of only one granulometric range was carried out, the reduction of the compressive strength did not reach the same magnitude found by Lima et al. [2] for the replacement of 50% of the aggregates.Comparing OC and EVA-C the reduction of compressive strength was 33% and comparing FRC and EVA-FRC this reduction was less significant reaching 20%.
As for the fiber-reinforced concrete (FRC) specimen, the compressive strength result was higher than the conventional concrete (OC) strength, and this may have happened due to the contribution of the fibers acting as a confinement agent for the concrete, resulting in a better distribution of the applied stresses [66].However, it is known that the fibers have little influence on the compressive and tensile strength of concrete, being quite effective in the postpeak [67].Therefore, it is likely that this difference of 1 MPa is more correlated to the variability of the results since concrete is a heterogeneous material than to an improvement in strength.
Finally, when comparing only the EVA-FRC and EVA-C concretes, it can be seen that the fibers may be contributing to the improvement of uniaxial compressive strength.It is known that concrete with aggregates with lower Young's modulus, such as EVA, can lose the ability to mesh and interlock, starting to concentrate stress in the vicinity of the EVA [68].Thus, the inclusion of fibers in the mixture can improve the interlocking capacity of the material, serving as a bridge for the stress transfer to the microcracks that may arise.

Three-point bending test
The efficiency of fibers as reinforcement depends mainly on interactions at the interface and includes chemical, frictional adhesion, and mechanical anchoring.Chemical adhesion corresponds to the primary adhesion of the interface, which is the result of chemical reactions from the composition of both the fibers and the matrix [38].To evaluate the efficiency of the reinforcements with polypropylene fiber and the addition of EVA, bending tests were performed with monotonic loading as established in Section 3.
Table 8 and Table 9 present the results of the three-point bending test.For each type of concrete evaluated, three tests were carried out, with the values expressed as the final result corresponding to the average of the results obtained, followed by its standard deviation.P LOP = load corresponding to the limit of proportionality; σ LOP = stress corresponding to the limit of proportionality; CMOD LOP = crack opening corresponding to the limit of proportionality; f R,1 = residual flexural strength corresponding to 0.5 mm CMOD; f R,2 = residual flexural strength corresponding to 1.5 mm CMOD; f R,3 = residual flexural strength corresponding to 2.5 mm CMOD; f R,4 = residual flexural strength corresponding to 3.5 mm CMOD; As recommended by standard EN 14651 [56], the characteristic parameters of concrete were obtained from the tests carried out.Among them, proportionality limit (σ LOP ), that is defined as the stress corresponding to the maximum load within the crack width range from 0 to 0.05 mm.The value of the ultimate load (P U ) was also determined, being the point where the slope of the load versus the CMOD curve is null.In addition, residual stresses were obtained at pre-defined CMOD values, 0.5, 1.5, 2.5, and 3.5 mm, and toughness values were defined as the area under the load versus CMOD curves.For the toughness, the limit point considered for the calculation was the end of the test, that is, a crack opening of 4 mm. Figure 9 and Figure 10 show the stress per CMOD and toughness per CMOD curves, respectively, of all composites evaluated.It is observed that all curves displayed a linear behavior until the appearance of the first crack, followed by a decrease in stress with increasing CMOD.This behavior is called deflection softening and is characterized by the appearance of a single crack, common in composites reinforced with discrete fibers [69]- [71].Therefore, in all cases, the ultimate load value corresponds to the load at the proportionality limit (PLOP).For all composites, the appearance of the first crack is related to the strength of the matrix and, therefore, it can be verified for OC the typical behavior of brittle material with immediate loss of strength after cracking.For FRC, after cracking, the fibers begin to absorb the tensile stresses in the cracked region, making the material behave in a pseudoductile manner.In the case of the three-point bending test, and considering the combinations investigated, the ultimate strength presented an average variation of 5.32% when comparing the fiber-reinforced plain concrete (FRC) and the EVA fiber-reinforced concrete (EVA-FRC).
Although the pre-cracking behavior was similar, the post-cracking behavior varied.The use of fibers improves the mechanical behavior of concrete, especially toughness and post-cracking residual strength.Table 9 presents the relationship between the residual stresses and the ultimate composite stress (fR,i/σLOP), defined as relative residual strength.It is considered that the greater the value of this ratio, the greater the ability of the composite to maintain its strength after the appearance of the first crack.The evaluation of these values allows us to characterize the efficiency of the reinforcement in absorbing the stresses to which they are submitted.In the case of EVA-C, the strength capacity at serviceability state was maintained at approximately 25%.For the case of the FRC, the strength capacity observed was 35% while in the EVA-FRC it resulted in an average strength capacity of 52%.These results indicate that the addition of EVA hybridized with the addition of fibers improves the maintenance of the post-peak strength capacity for the concrete evaluated.
Another way of evaluating the influence of fibers on concrete is through the toughness, which reflects the energy absorption capacity of the material.Thus, it can be seen that both composites resulted in toughness greater than 20 J.The EVA-FRC reached a toughness resultant of about 30 J, which shows that not only the fiber but also the EVA is somehow bridging the crack opened.

Fib Model Code simplified model
Using the simplified model of the fib Model Code (Figure 11) and considering that the ultimate tensile strength is defined by the residual stress fR,3, assuming that the compressive stress resultant is applied on the extrados chord and that the tensile behavior is rigid-linear, it is possible to estimate the moment associated with this ultimate stress and calculate the contribution of EVA in terms of the resistant moment of the section.Thus, calculating the moment equivalent to the CMOD 2.5 mm (CMOD 3 ), the FRC resulted in a resistant moment of 53.90 kN.cm and the EVA-FRC resulted in a resistant moment of 77.34 kN.cm.Therefore, EVA-FRC has a moment resistance of approximately 43.5% higher than conventional FRC, presenting by constitutive law that the EVA plays a role in bridging the fractured section.

Industrial floor design based on Technical Report 34
It should be noted that the constitutive law of the material for this application differs from that described by the fib Model Code since, in this case, the structural element is supported on an elastic media (soil).In this way, Figure 12 presents the constitutive law of the material that now depends on f R,1 and f R,4 and no longer on f R,3 .Hence, the conservatively calculated ultimate moment can be estimated according to Equation 2.  [60].
From the data obtained through the compressive test (compressive strength and Young's modulus) and the threepoint bending test (matrix flexural strength, residual stresses, and toughness), the necessary dimension for an industrial floor was calculated using the two FRCs and considering an elastic base of 0.048 N/mm 3 .As well, it should be noted that the partial safety factor for materials was applied for the ultimate limit state case (γ m,uls =1.5).An isolated point load of 5 tf/pallet end support and 6 tf/axle for forklift traffic were adopted for calculation.The summary of the dimensions and degree of use (M calc /M adm ) considering the loads adopted are shown in Table 10.The design results showed that even with lower compressive strength and a not so significant decrease in the tensile strength of the matrix in bending, the EVA-FRC achieves greater structural efficiency for an industrial floor of the same cross-sectional height (13 cm).Therefore, the carved EVA is functioning as a support to the fibers in the crack bridge.This hypothesis makes sense, as the aggregates that intercept a crack tend to rupture or be completely pulled out at the moment of the matrix rupture.
In the case of carved EVA, as it is a flexible material, it may be stretched and pulled out after the appearance of the crack in the cement matrix, and thus the bending behavior of the concrete with EVA has a certain ductility, further improving the pseudo ductile behavior of fiber-reinforced concrete.
Industrial floors are subject to the stress of loads and potentially from restraint to drying shrinkage, and this combination can cause cracking [44], [45], [72].A realistic evaluation of the combined effects of stresses induced by load and shrinkage is problematic and can produce conservative designs without significantly reducing the risk of cracking.Hence, Technical Report 34 [60] recommends an approach that does not take into account the effect of stresses induced by shrinkage, minimizing shrinkage by focusing on the design of the concrete mix, and constraining shrinkage with careful attention to the design and construction of the sub-base.In addition, Technical Report 34 [60] sets average drying shrinkage of the order of 300 µm/m -450 µm/m.Thus, although the concrete with aggregate replacement by EVA residue presents a significant efficiency for the application on industrial floors, it must be ensured that the concrete drying shrinkage does not exceed recommendations limits.
From a drying shrinkage point of view, the main effect of the aggregate is to restrict the contraction of the cement paste, thus helping to reduce the likelihood of cracking.In general, aggregates with a higher Young's modulus, cubic shape, and rough particle surface textures tend to offer more restraint to concrete shrinkage [73]- [76].However, the EVA used in this work has a low modulus of elasticity and an approximately two-dimensional shape, which can result in a significant increase in drying shrinkage even considering only partial replacement of the aggregates.
Therefore, drying shrinkage tests were performed according to ASTM C157 [58] and ASTM C490 [59] standards.Figure 13 shows the results of the drying shrinkage test for ordinary concrete and concrete with aggregate replacement by EVA, considering three specimens for each group tested.It can be seen from the analysis of the curves that the granular skeleton of the concrete with replacement still considerably restricts the shrinkage and, therefore, even with the presence of EVA in the composition, which is a flexible and easily deformable material, the coarse aggregate 19 mm in diameter and the fine aggregates can contain the deformation of the bulk cement paste.However, considering the shrinkage limit imposed by Technical Report 34 [60], the average shrinkage of EVA-C around 500 µm/m exceeds this limit.Therefore, aiming at the application of this material for industrial floors, chemical expander admixtures must be used in addition to the already widespread wet curing and/or chemical curing processes.

CURRENT STUDY LIMITATIONS AND UNCERTAINTY
From a material point of view, the residue substitution limits need to be addressed, carrying out tests with various residue contents and verifying the design of the mixture.In addition, one should seek to understand the effect of EVA replacement on the rheological properties of concrete and the possible impacts on durability, from a chemical to a mechanical point of view with creep and fatigue tests.In the case of the composite, the work presented only one type of reinforcing fiber and only one fiber volumetric fraction.These points need to be variables for future studies to characterize the contribution of EVA particles to different fibers and different fiber contents.Finally, experimental analysis on a structural scale needs to be carried out to validate the use of this material and verify if the industrial floor design proposed in this work is valid in practice.

CONCLUSIONS
This paper presents valuable information on the influence of EVA particles on the mechanical properties of plain and fiber-reinforced concrete.Density, drying shrinkage, uniaxial compressive strength, and three-point bending tests were performed to draw the following conclusions: • EVA particles have dimensions and granulometric distribution of approximately 9 mm coarse aggregates with 23% lower density.• With the total replacement of the 9 mm stone aggregate by the EVA particles residue, lighter weight concretes (EVA-C and EVA-FRC) were produced, which showed a drop of up to 34% in compressive strength when compared to concretes without replacement of stone aggregates (OC and FRC).• In the three-point bending tests, it can be verified that the appearance of the first crack occurred at similar strength values for all the concretes investigated.After the formation of the first crack, the concrete with the replacement of coarse aggregate by EVA (EVA-C), the fiber-reinforced plain concrete (FRC), and the EVA fiber-reinforced concrete (EVA-FRC) maintained a residual strength of about 25%, 35%, and 52% respectively indicating that the hybridization of EVA with the addition of fibers improves the maintenance of post-peak strength in the investigated concretes.Besides, the toughness values for the FRC were approximately 20 J while the toughness for the EVAFRC was approximately 30 J indicating that the addition of EVA is playing a role in bridging the crack opened.• Applying the results obtained in the three-point bending tests for the FRC and the EVA-FRC in the design of industrial floors according to TR-34 showed that the EVAFRC presented greater structural efficiency than the FRC for a cross-section of the same height (13 cm), considering the application of internal loads.Finally, the EVA particles play an important role in supporting the fiber to bridge the cracks, as could be seen by the constitutive analysis and by the three-point bending tests.Therefore, a pullout mechanism takes place in each EVA particle from the matrix before propagating the crack, improving the behavior of the fiber-reinforced 635 plain concrete.

Figure 2 .
Figure 2. EVA carved particles and microscopic analysis of one EVA particle.(a) EVA carved samples, (b) image of the EVA particle obtained through a Nikon microscope model SMZ800N, and (c) binary image resulting from the segmentation.

Figure 3 .
Figure 3. Polypropylene fiber with 51 mm length.(a) Image presenting the PP twisted geometry and (b) Image of the PP fiber cross section obtained through a Nikon microscope model SMZ800N.

Figure 4 .
Figure 4. Uniaxial compressive strength setup detailing the cylindrical rings and LVDTs positions.

Figure 5 .
Figure 5. Flexural specimens' geometry, setup of the three-point bending test, and positioning of the clip gauge to measure the crack mouth opening displacement (CMOD).

Figure 7 .
Figure 7. Schematic drawing of the type of loading used to calculate the industrial floor (internal loading) according to [60].

Figure 8 .
Figure 8. Uniaxial compressive strength results for the four types of concrete evaluated.(a) Maximum compressive strength and (b) Young's modulus.

Figure 9 .
Figure 9. Stress versus CMOD graphs of the 3-point bending test for the four concretes evaluated (OC, EVA-C, FRC, and EVA-FRC).(a) Ordinary concrete and EVA concrete and (b) Fiber reinforced concretes considering the OC based and EVA based concretes as matrices.

Figure 10 .
Figure 10.Toughness versus CMOD graphs of the 3-point bending test for the four concretes evaluated (OC, EVA-C, FRC, and EVA-FRC).(a) Ordinary concrete and EVA concrete and (b) Fiber reinforced concretes considering the OC based and EVA based concretes as matrices.

Figure 11 .
Figure 11.Fib Model Code simplified model to compute the ultimate tensile strength in uniaxial tension fFTu by means of the residual nominal bending strength fR,3 [57].

Figure 12 .
Figure 12.Fiber reinforced concrete constitutive law for ground supported floors according to Technical Report 34 [60].

Figure 13 .
Figure 13.Schematic drawing of the type of loading used to calculate the industrial floor (internal loading) according to Technical Report 34 [60].

Table 1 .
Compressive tests for the batch of cement used in the research.

Table 2 .
Chemical tests for the cement batch used in the research.

Table 3 .
Summary of geometric measurements of EVA particles accessed by microscopy technique.

Table 4 .
Composition of the EVA generated by the XRF analysis.

Table 5 .
Polypropylene fiber properties provided by the manufacturer.

Table 6 .
Mixture compositions of the concretes used, considering the ordinary concrete (OC), the concrete with replacement of coarse aggregate by EVA (EVA-C), PP fiber reinforced concrete (FRC), and the concrete with replacement of coarse aggregate by EVA reinforced with PP fiber.

Table 7 .
Approximate apparent density for the two analyzed concretes (OC and EVA-C).

Table 8 .
Summary of three-point bending test parameters.

Table 9 .
Peak stress maintenance and toughness results.

Table 10 .
Summary table of the efficiency of FRC and EVA-FRC applied in the design of industrial floors.