Experimental and theoretical behavior of cementitious plates containing ethylene vinyl acetate reinforced with glass woven fabric under impact load

The workability decreased with the incorporation of EVA into cementitious slabs. Superplasticizer is added to restore workability. The results are in good agreement with those reported in a previous study [15]. Figure 3 shows the unit weight values of the samples depending on the EVA replacement. As shown in Figure 3A, with the increase of EVA replacement, a decrease is determined in the unit weight values of the samples compared to the control concrete, and this decreased value is in the range of 17–46%, which is consistent with the results of studies in the literature [14, 15]. Lima et al. [14] stated that with EVA replacement in the range of 50%, the unit weight values of concrete decreased in the range of 92%, while Zuchetto et al. [15] determined a decrease of 5–63% in concrete unit weight values with 0–100% EVA replacement. The fact that the data obtained in our current study are close to those provided in Strong's study [20] but slightly different from those in Lima et al. [20] can be attributed to two reasons, namely the EVA replacement rate and the EVA particle size employed in the experiments conducted as part of these studies. The first of these, as seen in our study, caused excessive void formation in concrete structures when the EVA replacement exceeded 32%. The other is that due to the smaller particle size, the EVA admixture is mixed more homogeneously in the concrete content while reducing the amount of space in the structure of the concrete and increasing the concrete unit weight value. On the other hand, there is a normal decrease in unit weights of GFRP-coated plate samples up to samples containing 32% replacement EVA. (Figure 3B). The decrease in sample unit weight is more noticeable after the percentage of replacement EVA exceeds 32%. This is because of the amount of gap between the EVA and cement matrix. On the other hand, it is determined that the GFRP coating exhibited a gradual decrease in the concrete unit weight values up to 32% replacement rate due to the increase in the EVA additive, while it showed a dramatic decrease after 32% (Figure 3B). The reason for this is that epoxy can be absorbed in the range of 10–15% due to the reasonable level of void ratios in the concrete up to 32% EVA replacement ratio, while in cases the EVA replacement ratio exceeds 32%, depending on the excessively increased amount of voids in the concrete structure, the epoxy absorption of the concrete may be increase its efficiency up to 30–40% range [25].

Fig. 3

Compressive strength values obtained from EVA-replaced and GFRP-coated samples after EVA replacement, which are subjected to the compressive strength test according to ASTM C349-02 [24] standard, are shown in Figure 4.

Fig. 4

As seen in Figure 4A, it has been determined that the control cementitious sample provides higher compressive strength than the EVA-replaced samples, and this has been reported in many studies in the literature. Moreover, although it has been determined that there is an unexceptional decrease in the compressive strength values due to the increase of EVA replacement, it has also been determined that the compressive strength values show a dramatic decrease after 32% EVA replacement, and the compressive strength value decreases to approximately 1 MPa with 56% EVA replacement. This is due to the fact that the density value (0.55 g/cm³) of the EVA material is lower than that of the aggregate materials (2.7 g/cm³) [14, 15].

The GFRP coating process increased the compressive strength values of both the control cementitious sample and all of the EVA-replaced samples (Figure 4B). This is a result of the mechanical properties of GFRP-added epoxy being superior to concrete [26]. As a matter of fact, in our study, the control cementitious sample exhibited a compressive strength value of 31.43 MPa, while the GFRP-coated sample exhibited a compressive strength value of 42.07 MPa. On the other hand, compressive strength values of EVA-replaced and GFRP-coated samples are lower than those of the control cementitious samples. However, the decrease in compressive strength of EVA-replaced and GFRP-coated samples is less than that of EVA-replaced samples. This shows that GFRP coating provides an improvement in compressive strength to a certain extent. As a matter of fact, after 32% EVA replacement, the compressive strength value in the EVA-replaced samples without GFRP coating is found to be in the range of 0–1 MPa, while the compressive strength value in the samples that are GFRP-coated and contain more than 32% EVA is found to be in the range of 4–23 MPa. In particular, the 22.91 MPa pressure value detected in the 45%-EVA-substituted-and-GFRP-added sample is noteworthy. The increase in the compressive strength value obtained in the sample with more than 32% EVA substitute is a result of the filling of the voids in the structure with resin as a result of the cementitious samples shown in Figure 4C absorbing more resin.

The maximum first peak load values of the samples can be seen in Figure 5. It is clearly seen from Figure 5 that the maximum load decreases with an increasing amount of EVA, and the maximum load increases as the sample thickness increases [19]. It can be also said that samples coated with GFRP have a higher load-carrying capacity than uncoated samples. As the amount of replacement EVA in concrete increases, the peak impact force value decreases. However, when coated with GFRP material, the peak impact force values increase significantly (Figure 5). This is because the EVA material creates voids in the concrete content, while reducing the mechanical properties; on the other hand, during the GFRP coating process, the voids on the surface of the cementitious structures can be filled with resin with higher mechanical properties than cement. The reason for the limited increase in these mechanical properties is that the resin penetrates only the surface of the concrete and cannot reach the cavities in the interior.

Fig. 5

In Figure 6A–C, force-time curves under the impact force of samples containing various proportions of EVA are shown. It is observed that the peak impact force values decrease as the amount of replacement EVA increases in the plates containing PC and replacement EVA. Therefore, PC samples have higher compressive strength than samples containing replacement EVA. As the amount of EVA increases, its compressive strength and impact strength decrease. In addition, the time to peak in the force-time figure of the samples containing 56% EVA increases.

Fig. 6

Depending on the increase of EVA replacement, a decrease in the peak impact force values occurred (Figure 6). In addition, as the thickness of the samples increases, the peak impact force values and the peak impact time values also increase (Figure 6A–C). This indicates that the cementitious samples exhibit a more ductile behavior with the increase in thickness in the samples.

In Figure 7A–C, the force-time curves of the GFRP-coated samples under the impact force are seen. As can be seen from an examination of Figure 7A–C, as the amount of EVA increases, the peak impact force values decrease. The peak impact force also shifts to the right along with the time axis. In addition, the peak impact forces obtained for the different EVA ratios of the GFRP-coated samples are closer to each other than in the uncoated case. While the samples containing EVA are coated by the vacuum infusion method, it is thought that the peak impact force increases due to resin filling of the pores in the EVA concrete. As the amount of EVA increases, the amount of pores increases. Accordingly, the amount of epoxy taken into the gaps also increases. As the amount of EVA increases, the ductility of the plate sample also increases. Therefore, it is thought that the time of the peak impact force formation is shifted to the right in samples containing EVA.

Fig. 7

Considering the force-time curves of the GFRP-coated samples (Figure 7A–C) compared to the un-coated samples (Figure 6A–C), an increase is observed in the peak impact values. In addition, it can be seen that, unlike the uncoated samples (Figure 6A–C), double peak impact values are measured in the force-time curves of the GFRP-coated samples. The increase in the impact peak values is a result of the effect of the superior mechanical properties of GFRP in improving the mechanical properties of cementitious samples [26]. The fact that the samples showed a double peak value (Figure 7A–C) instead of a single peak value (Figure 6A–C) is a result of the fact that the GFRP-added epoxy and the cement core structure have different impact resistance. Double peak pulse (the hill–pit–hill formation) value can be seen in Mubin et al. [19]. It is stated that it is also observed in cementitious samples produced by hand lay-up in Figure 7A–C, and it is stated that the double peak impact value increased the damping time. On the other hand, when the thickness change force-time curves of the EVA substituted GFRP-coated samples are examined (Figure 7A–C), it is seen that the peak impact force values decrease with the increase in the EVA replacement ratio as in the un-coated EVA samples and increase with the increase in the thickness values of the samples. It is seen that the peak impact force values are shifted to the right in the GFRP-coated samples, as in the samples without GFRP coating. This indicates that the addition of the EVA replacement to the samples increases the ductility and is unaffected by the GFRP coating process. Moreover, an increase in thickness resulted in a more evident hill–pit–hill formation. This might be attributed to the fact that under axial impact, the applied force is mainly taken by the cementitious part of the sample [27].

The damage appearances of the samples subjected to the low-velocity impact test under 10.5 kg impact load are shown in Figure 8A–I. As seen in Figure 8, all uncoated samples have failed after the applied load of 10.5 kg, except the PC sample with 50 mm thickness. When the failure mechanisms are examined (Figure 8A, B), it can be seen that the failure mechanism occurred as thin linear fractures in the samples without EVA replacement, while it can be seen that the failure mechanism occurred in the form of more prominent thick linear fractures and local dispersions (surrounding impact zone) in the samples with EVA replaced (Figure 8C–F).

Fig. 8

On the other hand, it is seen that linear fracture did not occur in all of the GFRP-coated samples. The damage of the samples is limited to only the impact zone area, and this type of damage is reported by Cheng et al., [28] who named it “whitening”. The authors defined whitening as the damage of only the GFRP coating in the structure of the cementitious plates, resultant to the effect of applied force. The level of stress decreased with the decrease of deformation; so, whitening formation tended to decrease with increasing sample thickness (Figure 8G–I). Since the thickness of the material increases, the bending strength increases [29].

Peak impact force is an important parameter in structures subjected to impact load to evaluate the impact resistance behavior of the materials [30]. In this section, analytical models are used to approximately forecast the peak impact force of coated GFRP and uncoated GFRP-cementitious plates. Coated and uncoated plates are considered isotropic. In addition, friction and energy losses are neglected.

The mass-plate effect is modeled as an equivalent system, as shown in Figure 9 [30,31,32].

Fig. 9

The initial velocity of the striker hitting the plate element is zero (V = 0). The free-falling striker from the height h₀ suddenly hits the plate with velocity V0=2gh0 {V_0} = \sqrt {2g{h_0}} . The impactor can be made simpler, as a spring (contact stiffness k), a vibration dampener (C), and an impact mass (M). The peak impact force, Pm, is found according to Eqs (1)–(3) [30, 31]. (1) Pm≈(Pms−0.95+Pmd−0.95)−1.053 {P_m} \approx {\left( {{P_{ms}}^{ - 0.95} + {P_{md}}^{ - 0.95}} \right)^{ - 1.053}} (2) Pms=V0kM {P_{ms}} = {V_0}\sqrt {kM} (3) Pmd=CV0 {P_{md}} = {CV_0} where P_ms is the value of peak contact force between the mass-spring system and the rigid plate and P_md is the peak damping force. The damper viscosity (C) is shown in the following equation: (4) C=2π2r0Ghsm0/5 C = 2{\pi ^2}{r_0}\sqrt {G{h_s}{m_0}/5} where r₀ is the radius of impact at the center of the plate under the effect of impact force, r₀ = h/2, and h is the plate thickness. G is the shear modulus of the plate [30]. (5) G=ED2(1+μc) G = {{{E_D}} \over {2\left( {1 + {\mu _c}} \right)}} where E_D and μ_c are the modulus of elasticity and Poisson's ratio obtained by dynamic tests, respectively; h_s is the shear thickness of a solid plate of thickness h.(h_s = h/1.2); and m₀ is the unit weight of the surface of the plate.

The contact stiffness k in Eq. (2) is calculated with the help of Eq. (6), which is the impact model of the hard spherical impactor. (6) k=2πRfc k = 2\pi {Rf_c} where R is the distance from the point where the striker hits the sample to the support and f_c is the dynamic bearing power.

Eqs (1)–(5) are used for the values in Method 1. The value of k is calculated with the help of Eq. (6). Here, f_c, dynamic bearing power, is calculated by increasing the static compressive strength by 20%.

The dynamic elasticity module is used in calculating the shear modulus. The following formulas are used for the dynamic elasticity calculation [33]: (7) ES=0.97W2UW {E_S} = 0.97{W^2}\sqrt {{U_W}} (8) ED=1.16W2UW {E_D} = 1.16{W^2}\sqrt {{U_W}} (9) v=1ES[ 0.16ES+20UW ] v = {1 \over {{E_S}}}\left[ {0.16{E_S} + 20{U_W}} \right] where W (kg/m³) is the density of the concrete, U_W (N/mm²) cubic compressive strength, υ Poisson ratio, and E (kN/mm²) concrete elasticity modules. Here, s index represents static and D index represents dynamic data.

Eqs (1)–(6) are used for the values in Method 2. The static elasticity modulus of the samples is calculated according to ACI 318-95 [34] using the stress values obtained from the pressure tests. The Poisson ratio is taken as υ = 0.2. While calculating the shear modulus, the dynamic elasticity module is used. The following formula is used for Method 2 of dynamic modulus of elasticity (GPa) [35]. (10) ES=57000f(psi) {E_S} = 57000\sqrt f \left( {psi} \right) (11) ED=(ES+4.1)/1.04 {E_D} = \left( {{E_S} + 4.1} \right)/1.04 In this theoretical calculation, the values encompassed in Method 1 are obtained by using the elasticity modules calculated with the help of the formulation defined for LWC. In Method 2, calculations are made with the help of known formulas (Eqs [1,2,3,4,5,6]) for standard concretes.

The data including the comparison of the peak impact force values calculated with the help of Method 1 and Method 2 of the GFRP-coated and uncoated samples and the peak impact force values obtained from the experimental results are shown in Figure 10. When Figure 10 is examined, it is seen that the values obtained from the experimental studies and the theoretical data largely overlap with each other. When the cementitious samples are examined in detail, depending on the modification processes, the following conclusions can be reached.

Fig. 10

When the peak impact force graphs of the coated and uncoated PC samples are examined (Figure 10A, B), it is seen that the data obtained based on Method 2 are closer to the experimental data. Considering the peak impact force graphs of the coated and uncoated samples with 18% EVA replacement are examined (Figure 10C, D), it is determined that the peak impact force values obtained with Method 2 are closer than the experimental data, as in the samples without EVA replacement (Figure 10A, B). The reason for this is that the samples are in the normal weight concrete (NWC) class [29]. In cases where the EVA replacement rates are 32% (Figure 10E, F), 45% (Figure 10G, H), and 56% (Figure 10I, J), the peak impact force values obtained from Method 1 are closer to the experimental values than those from Method 2. The reason for this is that the cementitious samples produced with the increasing EVA replacement rate fall into the LWC from the NWC class. This confirms the unit weight graphs shown in Figure 5.

When the Figure 10 are examined, the calculated peak impact force values are in good agreement with the experimentally obtained peak impact force values. However, there may be some difference between the calculated values and the experimental values, depending on the plate thickness, the amount of replacement EVA, and the amount of epoxy resin. It is apparent that the best agreement between experimental and theoretical results occurs with samples containing 32% EVA. Therefore, in EVA concrete, it is thought that concrete samples behave like EVA when the proportion of EVA exceeds 32%. On examining the effect of the GFRP coating on the low-velocity impact behavior, it is ascertained that the GFRP coating process contributed to better agreement of the experimental results with the theoretical results in all samples without exception.