Performance of Porous Slabs Using Recycled Ash

Permeable concrete is an environmentally friendly material that improves water permeability and slip resistance. The manuscript describes a new study aimed at improving the strength of permeable concrete obtained using local materials for the partial replacement of cement with rice and wheat straw ash due to the high amount of silica and pozzolanic characteristics present in the ash. For this purpose, nine concrete mixes were made (Phase I). The mixes were classified into four groups: Group A, with cement/aggregate ratios of 0.23, 0.34, and 0.44 for Mixes 1, 2, and 3, respectively; Group B, with sand added at 10% and 15% to the coarse aggregate for Mixes 4 and 5; Group C, with rice straw ash replacement ratios of 10% and 15% in the cement for Mixes 6 and 7; and, finally, Group D with wheat straw ash replacement ratios of 10% and 15% in the cement for Mixes 8 and 9. For Groups B to D, the water/binder ratio was 0.238. Fresh and hardened concrete tests were conducted. The results showed that Mixes C and D, which contained rice and wheat straw ash, increased the compaction factor due to their spherical shape and higher surface area compared with traditional pervious concrete. Additionally, permeability and porosity increased slightly for the mixes using rice and wheat straw ash. This could be attributed to increasing the interconnected voids. Optimum porosity was reached with 15% rice straw ash. The optimum mix design from Phase I was used in Phase II. Therefore, six pervious concrete slabs, reinforced with different types of reinforcement, were tested under flexural load. With the help of ANSYS, a finite element model was created to verify the results of experiments. The results of the numerical simulation are consistent with the results of the experiment. This article represents a definite step to new knowledge in the field of research of permeable concrete obtained using the partial replacement of cement with rice and wheat straw ash. Hence, this form of concrete can be used for parking lot paving, sludge beds for sewage plants, swimming pool surfaces, bridge walkways, zoo area floors, and animal barns. This concrete can also be used in applications requiring lightweight concrete.


Introduction
Pervious concrete is a concrete category that was developed for reducing the runoff amount of water and improving water flow near sidewalks and parking lots. This type of concrete can be produced by mixing coarse aggregate, cement, and water with/or without admixtures [1]. Pervious concrete is relatively inexpensive due to the lower cement content in the mix. The lack of fine aggregates decreases the surface to be covered with cement paste. Due to the pervious concrete components, there is a thin layer of cement paste and a higher void content in comparison to traditional concrete.
These voids occupy about 15-30% of the total volume [2] and are vital in specifying the characteristics of that concrete, particularly its compressive strength and permeability [3]. As the void content increases, lower compressive strength and higher permeability results [4,5]. Due to its lower compressive strength, this concrete is used as a building material with limited usage [6]. Pervious concrete has a high permeability that is desired, with negligible compressive strength. To improve compressive strength, many investigators have substituted the coarse aggregate with precise fine materials [7][8][9]. These Table 1 illustrate the 9 mix proportions with 162 samples used in this study. The mixes were classified into four groups, Groups A-D. Group A, with cement/aggregate ratios of 0.23, 0.34, and 0.44 for Mixes 1, 2, and 3, respectively; Group B, with sand added at 10% and 15% to the coarse aggregate for Mixes 4 and 5; Group C, with rice straw ash replacement ratios of 10% and 15% in the cement for Mixes 6 and 7; and, finally, Group D with wheat straw ash replacement ratios of 10% and 15% in the cement for Mixes 8 and 9.

Figure 1 and
For Groups B to D, the water/binder ratio was 0.238. The variables were the ratio of cement paste to aggregate, with sand, rice, and wheat straw ash as a partial alternative ratio. All considered ratios were expressed on a dry basis. Fresh concrete tests were conducted to assess the slump, compaction factor index, and density. Hardened concrete tests were performed to obtain compressive strength, splitting tensile strength, flexural strength, permeability, and porosity. The results showed that the porosity increased for mixes using rice and wheat straw ash.   [47] and ASTM C127-88 [48], a sieve analysis was performed. The results of the sieve analysis and the physical property test results are shown in Table 2 and Figure 2.

Coarse Aggregate
A conventional aggregate of crushed limestone with a maximum size of 10 mm was obtained from a local quarry. According to Egypt Standard Specification (ESS) 203/2018 [47] and ASTM C136-84a [49], sieve analysis was performed. The aggregate's mechanical and physical properties are shown in Table 3, and the grading is as shown in Figure 3.  [50]. The physical and chemical properties are given in Tables 4 and 5, respectively.   Rice and wheat straw ash, as forms of agricultural waste, were used as a cement replacement material with unit weights of 840 and 820 kg/m 3 , respectively. RSA and WSA were first ground in a bug milling machine [29,30] and were then passed through different sieves of 20-200 mesh sizes. The chemical compositions for RSA and WSA are given in Table 6, which fulfilled ASTM C618 [51]. 'MasterGlenium 51 superplasticizer was provided by BASF. The superplasticizer density was 1.1 kg/L. The chloride content was less than 0.1%, and the alkaline content was less than 3%, which confirms Bong et al. [52].

Polypropylene Fibers
Used in amounts of 0.1% and 0.2% by cement weight. The physical and mechanical properties are given in Table 7.

Geogrid
Geogrid CE121 polyethylene of two types were used: Type I, with opening dimensions of 50 × 50 mm, and type II, with opening dimensions of 70 × 70 mm.

GFRP Bars
Glass fiber reinforced polymer bars with mechanical properties, as given in Table 8.  Table 1 and Figure 1 present the nine mixes divided into four groups, A-D. The specimens were kept in the molds for 24 h in air and then removed from the molds and immersed in fresh water until the test time. Table 9 presents the slabs mix design and Figure 4 shows the used six group slabs, S1-S6. The slabs were kept in the molds for 24 h in the air, and then removed from molds and cured in freshwater until the test time.

Experimental Tests for Phase I
The tests executed were the slump, compacting factor, density, compressive strength, split tensile strength, flexural strength, permeability, and porosity tests.

Slump Test
A concrete slump test was executed according to ASTM C143 [53]. The slump was measured as the difference between the mold height and the concrete height after removing the mold. For all mixes, nine samples (nine samples/mix) were tested.

Compacting Factor Test
The compacting factor test was done according to ACI Standard 211.3 [54]. Equation (1) was used for calculating the compacting factor value. For all mixes, nine samples (nine samples/mix) were tested.
Cylindrical specimens with dimensions of 150 × 300 mm were used in measuring the density according to ASTM C138 [55] using Equation (2). For all mixes, nine samples (nine samples/mix) were tested.
where D is the density, M is the sample mass, and V is the sample volume.

Compressive Strength
Cubes with dimensions of 150 mm × 150 mm × 150 mm were used in measuring the compressive strength according to ASTM C39 [56] using Equation (3). We used an ELE automatic compression testing machine at 2000 kN. For each mix, three cubes were tested for 7 and 28 days.
where Rc is the compressive strength (MPa), P is the load (kN), and A is the area (mm 2 ).

Split Tensile Strength
Cylindrical specimens with dimensions of 150 × 300 mm were used in measuring the splitting tensile strength according to ASTM C496 [57]. A universal automatic testing machine of 1000 kN was used. For each mix, three cylinders were tested for 7 and 28 days. The maximum tensile stress was calculated using Equation (4).
where P is the vertical load applied, l is the cylinder length, d is the cylinder diameter, and T is the tensile stress.

Flexural Strength
Prism samples of 100 × 100 × 500 mm were used in measuring the flexural strength according to ASTM C78 [58]. A universal automatic testing machine of 1000 kN was used. For each mix, three prisms were tested. The flexural strength is expressed as the rupture modulus R (MPa) using Equation (5).
where R is the rupture modulus, P is the maximum utilized load indicated by the testing machine, L is the span length, b is the specimen width, and d is the specimen depth.

Permeability
Beams of 300 × 300 × 100 m were used in determining the permeability according to ASTM PS 129 [59]. For each mix, three beams were tested. The coefficient of water permeability can be computed using Equation (6).
where Q is the water amount (cc/s); A is the concrete specimen cross-section area (mm 2 ); H = (h1 − h2), h1 is the initial water level (mm) and h2 is the final water level (mm); L is the concrete specimen thickness (mm), and K is the water permeability coefficient (mm/s).

Porosity
Cubes of 150 × 150 × 150 mm were used in determining the porosity according to ASTM D7063/D7063M [60]. For each mix, three cubes were tested. The porosity can be calculated using Equation (7).
where P is total porosity (%), W 1 is the oven dry weight (kg), W 2 is the weight under water (kg), V is the sample volume (m 3 ), and ρ w is the water density at 21 • C (kg/m 3 ).

Experimental Tests for Phase II
Six pervious concrete slabs with 1200 × 1200 × 140 mm dimensions were tested for flexural strength using a 5000 kN universal compression testing machine connected to a data acquisition system. A 70 mm thick stiff rubber layer was placed under the slab to be like the subgrade reaction. A vertical concentrated load was applied at the center of the slab by a 125 mm radius circular plate. The slabs were instrumented to record the measured actual deformations. The deformations were monitored using LVDT connected to the data acquisition system. Table 10 summarizes all the slab details. Table 2 shows the best economical mix design for the pervious concrete slabs obtained in Phase I.

Slump Test Results
The slump test results for each mix are plotted in Figure 5. The slump results ranged from 27 to 90 mm. The slump was 90 mm for mixes M1 and M2 and improved to 27-36 mm for all the other mixes, M3 to M9. This led to a stiff mix of a near-to-zero slump, which agreed with Mazumdar et al. [61].

Density Test Results
The density results are recorded in Figure 7 for all the mixes. M1 had a density of 1710 kg/m 3 , while mix M2 had a density of 1728 kg/m 3 . M3 had a density of 1800 kg/m 3 . For Group B, the mixes M4 and M5 had densities of 2097 and 2151 kg/m 3 , respectively. In Group C, the mixes M6 and M7 had densities of 1900 and 1710 kg/m 3 , respectively. In Group D, the mixes M8 and M9 had densities of 1611 and 1566 kg/m 3 , respectively. The results showed that the Group B mixes, M4 and M5, with fine aggregates, had a higher density due to the reduction in voids compared with the other mixes, which agreed with Mazumdar et al. [61].   The results showed that mixes M4 and M5 in Group B had higher strengths at 7 and 28 days. This could be attributed to the presence of fine aggregate, as shown by Mazumdar et al. [61]. The low difference in compressive strength at 7 and 28 days was due to shear failure in the aggregate, as shown by Mazumdar et al. [61]. The decrease in compressive strength for all other mixes was due to the presence of the air content, as shown in Desmaliana Erma et al. [62]. The compressive strengths measured for the mixes M6-M9 were 42% to 84% and 45% to 81% lower than those of mix M3; this might be because the quantity of RSA and WSA in the mix may be higher than that required to react with the liberated calcium hydroxide resulting from cement hydration, thus leading to excess silica being leached out and causing a deficiency in strength as it replaces part of the cementitious material but does not contribute to strength; this is in agreement with data published by Jankovský et al. [63]. Figure 9 shows that the splitting tensile strength ranged from 0.21 to 1.83 N/mm 2 and from 0.25 to 2.34 N/mm 2 for the results from 7 and 28 days, respectively. As predicted, the splitting tensile strengths at 28 days were higher than those of 7 days. For Group A, the splitting tensile strengths for the mixes M1, M2, and M3 at 7 days were 0.59, 0.87, and 1.3 N/mm 2 ; at 28 days, they were 0.73, 0.92, and 2.07 N/mm 2 , respectively. For Group B, the mixes M4 and M5 were, at 7 days, 1.39 and 1.83 N/mm 2 and, at 28 days, 1.81 and 2.34 N/mm 2 , respectively. For Group C, the mixes M6 and M7, the splitting tensile strengths decreased at 7 and 28 days, where they were 0.88 and 0.54 N/mm 2 as well as 1.37 and 0.57 N/mm 2 , respectively. Finally, for Group D, the mixes M8 and M9, the splitting tensile strengths at 7 days were 0.59 and 0.21 N/mm 2 as well as 0.65 and 0.25 N/mm 2 at 28 days, respectively. The results showed that a similar trend was observed for compressive strength. The splitting strength for the Group B mixes M4 and M5 had higher strengths at 7 and 28 days, as shown by Ali Shagea and Kacha Smit [64]. On the other hand, the low strength for the rest groups was due to the poor bond between the cement paste and aggregate, which agreed with Ali Shagea and Kacha Smit [64]. As explained earlier in the compressive strength, these splitting strengths measured for mixes M6-M9 were 32% to 84% (at 7 days) and 33% to 88% (at 28 days) lower than those of mix M3, agreeing with Jankovský et al. [63]. Figure 10 shows that the 2-day flexural strength of the tested mixes ranged between 0.24 and 1.13 N/mm 2 . Group A flexural strengths were 0.34, 0.52, and 0.92 N/mm 2 for the mixes M1, M2, and M3, respectively. The group B flexural strengths were 0.98 and 1.13 N/mm 2 for the mixes M4 and M5, respectively. In Group C, for the mixes M6 and M7, the flexural strengths were 0.24 and 0.37 N/mm 2 , respectively. In Group D, the mixes M8 and M9, the flexural strengths were 0.28 and 0.30 N/mm 2 , respectively. The results indicated that the mixes M4 and M5 for Group B had higher strengths at 7 and 28 days. This could be attributed to the presence of fine aggregate, as seen in Mazumdar et al. [61].  Figure 11 shows that the permeability for the concrete mixes ranged between 116 and 395 lt.cm 2 /s. The Group A permeabilities were 355, 352, and 319 lt.cm 2 /s for the mixes M1, M2, and M3, respectively. For the Group B mixes M4 and M5, the permeabilities were 128 and 116 lt.cm 2 /s, respectively. Permeability was decreased when compared with Group A. Thus, it is recommended to use sand with a percentage of less than 10% for an accepted permeability. The permeability decreased with an increase in rice straw ash by 10% and 15% in Group C for the mixes M6 and M7 from 336 to 245 lt.cm 2 /s. Similarly, in the Group D mixes M8 and M9, the permeability decreased with an increase in wheat straw ash replacement by 10% and decreased with an increase of 15% from 395 to 309 lt.cm 2 /s. Using the rice and wheat straw ash in Groups C and D, respectively, increased the permeability compared with Group A more than with Group B, which agreed with Narayana et al. [65]. The permeability of the pervious samples increased directly with increases in the interconnecting voids, which agreed with Kia et al. [66].

Porosity Test Results
The porosity of the tested mixes ranged from 15.12% to 29.25%, as shown in Figure 12. Group A showed almost the same porosity, with values of 28.64%, 26.41%, and 25.85% for the mixes M1, M2, and M3, respectively. For the Group B mixes M4 and M5, the porosities were 18.63% and 15.12%, respectively. The Group C porosities were 21.42% and 29.25%, respectively. For Group D, the porosities were 28.26% and 27.54%. The porosity increase was because of the increase in the interconnected voids when the percentage of the RSA and WSA replacement increased, which agreed with Narayana et al. [65]. When the interconnected voids increased, extra water content was able to pass through the concrete, and, therefore, the concrete permeability directly increased, which agreed with Kia et al. [66].

Ultimate Load and Displacements
The ultimate load and displacements for all the slabs are given in Table 11. Figure 13 shows that the ultimate load ranges between 155.30 and 366.10 kN, and the ultimate displacement ranges between 12.10 to 19.40 mm, as indicated in Figure 14. Results show an increase in the ultimate load and displacement for all reinforcement types than control slabs. Using 0.1% and 0.2% polypropylene fibers provided an insignificant increase in the ultimate load by 179.80 and 180.20 kN, respectively. Using geogrid type I and II gave the ultimate loads of 352.80 and 319.00 kN, respectively. GFRP bar slabs exhibited the highest ultimate failure load of 366.10 kN. The impact of RSA with 15% on slab load-carrying capacity is clear from the result. This could be attributed to the presence of RSA, as seen in the study by Mazumdar et al. [61]. Figures 15 and 16 show the ductility ratio and energy absorption for the tested slabs. The ductility ratio ranged from 3.12 to 4.60. The ductility ratio was reduced by using GFRP bars. In contrast, most of the slabs incorporating other reinforcements achieved significant deflection at failure. The energy absorption of S6 was lower than the control S1. The energy absorption was about 139.78% for S2, 134.29% for S3, 366.97% for S4, 292.46% for S5, and 300.31% for S6 relative to control slab S1.

Load-Deflection Relationship
The load-deflection curves are plotted in Figures 17-22 for all slabs. Figure 17 shows the load-deflection curve for the control slab. From these figures, it can be clearly seen that for all slabs, the relationship between the load and deflection can be divided into two stages: Stage 1 is the elastic behavior until first cracking-The load-deflection relationship in this stage is linear; Stage 2, with large plastic deformation. Figures 18 and 19 show the load-deflection curves for slabs S2 and S3, reinforced by polypropylene 0.1% and 0.2%. Figures 20 and 21 show the load-deflection curves for slabs S4 and S5 that are reinforced by geogrid Type I and Type II. Figure 22 shows the load-deflection curve for slab S6, reinforced by GFRP bars.    The results show an increase in ultimate failure load and ultimate displacement for all slabs compared to control slab S1. Slabs S2 and S3 increased the ultimate load by 115.78% and 116.03% compared to S1. Therefore, it can be concluded that the increase in fiber reinforcing causes a noticeable increase in the ultimate load and ultimate displacement, which agrees with Erfan et al. & Roesler et al. [67,68]. In slabs S4 and S5, the ultimate load increased by 227.17% and 205.41%, respectively. Similarly, for slab S6, reinforced by glass fiber rods, the ultimate load increased by 235.74% compared with slab S1, which agreed with Mahroug et al. [69]. The ultimate displacements ranged from 12.10 to 19.40 mm for all tested slabs. Slab S1 had the lowest displacement value, and slab S4 gave the highest displacement. Hence, it can be said that slabs with geogrid type I had more ductile behavior. The impact of RSA with 15% on slab load-carrying capacity is clear from the result.

Crack Patterns
The cracks developed for all slabs under flexural load as vertical cracks near the slab's edges, starting from the bottom. The cracks are enlarged near the compression zone. The slabs are collapsed by diagonal cracks under the loading plate. All slabs exhibited almost the same crack.

NonLinear Finite Element Analysis
A finite element model was created to validate the experimental program using the ANSYS 2019-R1 [70] NLFEA program.

Modeling
Finite element analysis was carried out to investigate the flexural behavior of pervious concrete slabs reinforced with different types of reinforcement, as in Table 12. The ANSYS-2019-R1 finite element model is indicated in Figure 23. The investigated behavior includes the crack pattern, the ultimate load, and deflection of the slabs.

Comparisons between Experimental and NLFEA Results
There was good agreement obtained between experimental and ANSYS 2019-R1 results by applying the finite element model. Comparisons were applied between ultimate load, deflection at ultimate load, the first crack load, and the crack patterns.

Ultimate Load
Table 13 and Figure 24 shows the comparison between the ultimate experimental load and the numerical one. There is a reasonable agreement between the experimental and numerical ultimate loads.  Figure 24. Comparison between Exp. and NLA ultimate loads.

Ultimate Displacement
Table 13 & Figure 25 shows the comparison between the ultimate experimental displacement and the numerical one. The load-displacement curves for experimental and modeled slabs, as shown in Figures 26-31, showed good agreement with respect to control slab deflection. The impact of RSA with 15% on slab load-carrying capacity is clear from the result.        Figure 32 shows the typical cracking patterns of cracks for all reinforced pervious slabs obtained from experimental and nonlinear studies. The tension crack first appeared in slabs, and, with an increase in the load, the cracks widened and extended towards the compression zone. The slab failed by a diagonal crack under the loading plate.

Conclusions
Based on the studies performed on the pervious concrete prepared from normal and recycled materials, the following points were drawn as follows:

1.
Rice and wheat straw ash gave accepted permeabilities and porosities with lower densities compared to the control mixes. This was attributed to the chemical compositions for rice and wheat straw ash.

2.
Rice and wheat straw ash increased the compaction factor due to their spherical shapes and higher surface area compared with the control mixes.

3.
Increasing the cement/aggregate ratio from 0.22 to 0.40 led to increases in the 28-day compressive strength, splitting tensile strength, and flexural strength. 4.
Using sand increased the compressive, splitting tensile and flexural strengths, with a significant decrease in porosity and permeability. This was attributed to the higher density, which reduced the void content. Thus, we recommend the use of sand with a percentage less than 10% for accepted permeability.

5.
The results showed that permeability and porosity increased slightly for mixes using rice and wheat straw ash. This could be attributed to increasing the interconnected voids. Optimum porosity was reached with 15% rice straw ash (Mix 7).

6.
The concrete mixes with higher porosity showed high water permeability and low compressive splitting tensile and flexural strengths. The results proved that porosity and compressive strength were in a reverse relationship. 7.
The void ratio should be controlled by employing polypropylene fibers, such as E300, to control the gap between the particles; consequently, the compressive strength will be increased. 8.
All the studied reinforced pervious concrete slabs provided an increase in ultimate loads and displacements compared to the unreinforced control slab. 9.
Pervious slabs reinforced by GFRP bars gave the highest ultimate load, followed by the slabs reinforced with geogrid, while the lowest ultimate load was for slabs with polypropylene fibers. 10. A nonlinear study was performed to verify the experimental study. The numerical results were in good agreement with the experimental ones.
Finally, the presented study showed that the green recycled pervious concrete provided good economic savings compared with the other types. In addition, this concrete could be used for sustainability purposes.