3.1. Results
Table 3 presents the results of the optimized calculation focused on the limit states of load-bearing capacity. The design was performed according to the ČSN EN 1992-1 standard. The basic thickness of the concrete panel was always optimized for NAC concrete according to all conditions of the load-bearing capacity with an ordinary amount of reinforcement. It means the ordinary diameter and spacing of reinforcement for the considered panel span. Concrete cover
c of the reinforcement was considered as a constant regarding the durability of traditional steel reinforcement. A thickness of 20 mm for reference concrete from a natural aggregate and 35 mm for concrete with a recycled aggregate of concrete cover was considered, based on the results of carbonation resistance according to the ČSN EN 12390-12 standard. For the AR-glass FRP reinforcement, there is no need to propose the reinforcement concrete cover regarding durability, but only regarding the aggregate size and interaction conditions of the reinforcement and concrete, allowing full reinforcement activation. A constant value of 15 mm was selected as the concrete cover of AR-glass FRP reinforcement.
The calculations for the first group in
Table 3, ‘
Steel reinforcement, thickness modification’, for the limit state of load-bearing capacity includes traditional steel reinforcement and adjustment of the concrete thickness to comply with the conditions of the load-bearing capacity for different concrete mixtures and other sub-conditions of the calculation. The amount of reinforcement in the cross-sectional area was constant for this case; the reinforcement diameter was considered as 10 mm with an axial distance of 100 mm. The thickness of the panel was changed according to the measured strength of the concrete mixture.
The next group in
Table 3, ‘
GFRP reinforcement, thickness modification’, for the limit state of load-bearing capacity includes AR-glass FRP reinforcement and adjustment of the concrete thickness. The amount of reinforcement in the cross-sectional area was also constant for this case; the reinforcement diameter was considered as 6 mm thanks to its better tensile strength with an axial distance of 75 mm. The effect of the lower concrete cover was also positive. The thickness of the panel was very slightly changed to achieve the necessary load-bearing capacity. It was also possible to notice a change in the thickness of the panel. This is due to the condition of restricting the position of the neutral axis of the panel in bending.
The group ‘
Steel reinforcement,
reinforcement modification’ in
Table 3 for the limit state of load-bearing capacity presents results with steel reinforcement and the modification of the amount of reinforcement in the cross-sectional area. The diameter of the reinforcement bars was constant and the axis distance was modified to reach the required amount of reinforcement. Changes in the amount of reinforcement were insignificant.
The group ‘
GFRP reinforcement,
reinforcement modification’ in
Table 3 for the limit state of load-bearing capacity presents the modification of the amount of reinforcement in the cross-sectional area with AR-glass FRP reinforcement. The diameter of the reinforcement bars was constant and the axis distance was modified to reach the required amount of reinforcement. Changes in the amount of reinforcement were seriously insignificant. It is also possible to notice one change in the thickness of the panel. This is due to the condition of restricting the position of the neutral axis of the panel in bending.
Next,
Table 4 presents the results of the optimized calculation focused on the limit state of serviceability. The design was performed according to the ČSN EN 1992-1 standard. The decisive parameter is the panel deflection in the worst point—the middle of the panel span. The deflection calculation is the sum of the deflection of the long-term load, the short-term load, and the shrinkage. All these parameters were included in the calculation. The calculation of the limit state of serviceability was included because it is usually decisive in the construction design of horizontal structure elements, and this study presents two materials and their mutual combinations that can negatively affect the result due to the low elastic modulus. It is concrete with recycled aggregate and AR-glass FRP reinforcement.
The basic thickness of the concrete panel was optimized for NAC according to the conditions of serviceability with a limited deflection in the middle of the span of the panel at 24 mm. In addition, conditions of the limit state of load-bearing capacity were of course satisfied, with an ordinary amount of reinforcement in the first step for NAC and steel reinforcement. It means that an ordinary diameter and spacing of reinforcements for the considered panel spanning concrete cover c of the reinforcement was considered, as in the previous case of the design for the limit state of load-bearing capacity. Then, the thickness of the panel or the reinforcement area was modified for different concrete classes with experimentally determined parameters and steel and AR-glass FRP reinforcement. Results were used in the next step to evaluate LCA.
The first group of calculations labeled ‘
Steel reinforcement, thickness modification’ in
Table 4 for the limit state of serviceability includes traditional steel reinforcement and adjustment of the concrete thickness to comply with the conditions of limiting the load-bearing capacity for different concrete mixtures and other sub-conditions of the calculation. The decisive condition of the proposal is the limited deflection in the middle of the span. The amount of reinforcement in the cross-sectional area was constant for this case; the reinforcement diameter was considered as 10 mm with an axial distance of 100 mm, just like in the case of the limit state of load-bearing capacity. The thickness of the panel was changed according to the measured strength of the concrete mixture and is higher in comparison with the same group in
Table 3 for the limit state of load-bearing capacity.
The next group ‘
GFRP reinforcement, thickness modification’ of
Table 4 for the limit state of serviceability includes AR-glass FRP reinforcement and adjustment of the concrete thickness. The amount of reinforcement in the cross-sectional area was also constant for this case; the reinforcement diameter was considered as 6 mm thanks to its better tensile strength with an axial distance of 75 mm, just like in the case of the limit state of load-bearing capacity. The decisive condition of the proposal was clearly the limited deflection in the middle of the span. The thickness of the panel was rapidly changed to achieve the necessarily limited deflection. The negative effect of the low modulus of elasticity of both concrete and, especially, composite reinforcement is visible.
The group labeled ‘
Steel reinforcement,
reinforcement. Modification’ in
Table 4 for the limit state of serviceability presents results with steel reinforcement and the modification of the amount of reinforcement in the cross-sectional area. The diameter of the reinforcement bars was higher, it was 12 mm, and the axis distance was also modified to reach the required amount of reinforcement. Changes in the amount of reinforcement were insignificant. Further, for this group, the decisive condition of the proposal was clearly the limited deflection in the middle of the span. The thickness of the panel had to be changed to 300 mm due to the condition of restricting the position of the neutral axis of the panel when bending and the minimal required amount of reinforcement.
The group labeled ‘
GFRP reinforcement,
reinforcement modification’ in
Table 4 for the limit state of serviceability presents the modification of the amount of reinforcement in the cross-sectional area with AR-glass FRP reinforcement. The diameter of the reinforcement bars was constant at 10 mm and the axis distance was modified to reach the required amount of reinforcement. Changes in the amount of reinforcement were quite significant. It was also possible to notice changes in the thickness of the panel. This is due to the condition of restricting the position of the neutral axis of the panel when bending and the minimal required amount of reinforcement. The negative effect of the low modulus of elasticity of both the concrete and, especially, composite reinforcement is visible, as in the previous case. A higher amount of reinforcement does not have such a great effect on the overall consumption of material and thickness of the concrete panel. From the point of view of the proposal to assess the limit state of serviceability, it can therefore be stated that AR-glass FRP reinforcement seems unsuitable as a reinforcement of horizontal structural elements due to the low stiffness and low modulus of elasticity.
3.2. Discussion
Previously published studies have reported that the use of an RA in concrete cannot lead to significant environmental benefits, which have been mainly limited to the reduction of landfill areas and raw materials extraction [
46]. The probable reason is that although the aggregate represents approximately 70% of the concrete volume, the environmental impact of the production of cement and transportation is too essential, and so the savings from a natural aggregate cannot compete. The majority of studies have also shown the high sensitivity of results related to transportation [
1]. Only a few studies have been published (around 19% of all [
8]) dealing with the environmental assessment of RAC, where the fRA is considered as a possible substitution for natural sand [
23,
24,
27,
28,
47,
48,
49,
50,
51,
52,
53,
54]. In most investigations, it has been reported that the decline of mechanical properties and complications caused by the unknown water–cement ratio (which is related to higher porosity and water absorption) is too essential in comparison with the insignificant reduction of the environmental footprint, mostly related to the reduction of erosion and landscape damage. For this reason, it was argued by Kurda et al. [
54], Dezhampanah et al. [
48], and Roh et al. [
47] that the use of an fRA more essentially influenced the compressive strength of RAC, whereas the environmental impacts are not affected. On the contrary, the studies by Evangelista and De Brito [
17] and Yang et al. [
53] quantified the reductions of the environmental footprint for mixtures with the full replacement of natural sand by an fRA with the same mix design. It obtained a reduction of 27% and 6.8%, respectively. However, the inconsistency of the compressive strength was not considered in both cases. Moreover, Evangelista and de Brito [
17] observed that the environmental impact in the impact categories ADP, GWP, ODP, AP, EP, and POCP decreased by 6–8% when 30% of fRA was used and 19–23% for the full replacement of fine NA. In this case, the 1 m
2 of a solid concrete slab had a thickness of 0.25 m with a constant reinforcement ratio. On the contrary, in this LCA case study, the environmental impact of the reinforcement concrete slab, where the decline of the properties is compensated by the additional slab height or reinforcement level, is analyzed. Although the decline of properties, especially the modulus of elasticity, caused an increase in the slab height or reinforcement level, the normalized and weighted results for concrete show the positive environmental impact of replacement fNA by an fRA (see
Table 7 and
Table 8,
Figure 2 and
Figure 3). This does not correspond with the majority of previously published studies, where the decrease of properties and low reduction of environmental impact has been found [
47,
48,
54]. The divergent results in comparison with previous studies is an insignificant decline of mechanical properties when the natural sand is replaced by the fRA. Furthermore, the production of the recycled aggregate consumes less energy, can be used for demolition for new structures without the necessity of transportation, and, furthermore, brings a positive impact related to the recycling of steel bars which is usually contained to concrete waste.
The results of this case study slightly correspond with the study of Evangelista and de Brito [
17], where a similar structural element was chosen; however, in this study, the fine recycled aggregate concrete was considered without coarse NA and furthermore, the reinforcement by steel bars and the thickness of the slab is considered as constant, without taking into account the decline of mechanical properties. For these reasons, the results of both studies are not comparable.
Generally, the most positive influence can be observed for concrete slabs containing an fRCA, due to the benefit of the steel recycling of parent reinforcement concrete [
28]. However, in comparison with fRMAC, the fRCAC mixtures reached a better mechanical performance, which led to a lower thickness of slab or lower reinforcement level, respectively. Moreover, the results show a better environmental profile of concrete slabs with a constant thickness and various reinforcements (1 NAC, 1 FRMAC, and 1 FRCAC) in comparison with the design optimizing the thickness instead of reinforcement (2 NAC, 2 FRMAC, and 2 FRCAC). This implies that the optimal way is to compensate for the decline of the compressive strength of RAC with a higher reinforcement level and then use more concrete, which increases the cement content. In a comparison of the reinforcement of steel bars and glass fibers, similar results can be observed. Nevertheless, in the assessment of cement, steel, and glass reinforcement, it must be noted that the production of 1 kg of cement is related to 0.84 kg CO
2 eq. while for 1 kg of reinforcement, the result of the climate change indicator is 0.47 for steel and 2.01 kg CO
2 eq. for glass, respectively. Finally, the findings from the comparison of slabs designed for the LBC and SA limit states were not expected. Although the relative decline of the modulus of elasticity was more significant than the relative decrease of compressive strength compared to the reference, a comparison of the environmental impacts based on the limit states between NAC and fRCAC, and fRMAC, is not essential.
A comparison of the reinforcement type indicated a slightly better environmental performance for glass fibers in the case of the LBC limit state; however, the results for the SA limit state showed a completely opposite trend, where the environmental impact of glass fibers reinforcement is significantly higher. The probable reason is the increase in slab thickness correlated with a minimal percentage of glass fiber reinforcement. For this reason, the concrete volume increase leads to higher environmental impacts.
In the case of the climate change indicator describing the effect on global warming (see
Figure 4 and
Figure 5), the results of the environmental assessment show a positive influence of replacing natural sand with an fRCA. On the contrary, the use of an fRMA does not achieve favorable results where it is similar to the normalized and weighted results caused by the decrease of compressive strength, which is essential for a design based on the LBC limit state, and the modulus of elasticity is crucial for a design based on the SA limit state. The decline of mechanical properties is also essential, and the positive influence connected with replacing natural sand is negligible. The reduction of greenhouse gases (GHGs) related to mixtures with the full replacement of natural sand by an fRCA is between 84 kg CO
2 eq and 141 kg CO
2 eq for a design based on the LBC limit state and 140 kg CO
2 eq and 308 kg CO
2 eq for a design based on the SA limit state, which shows a decrease of 23% to 31% and 22% to 35%, respectively. In the case of an fRMA, a similar or slightly higher potential impact on the climate change indicator in comparison with a conventional solution (NAC) can be observed. The maximal increase was 23 kg CO
2 eq which corresponds to 5%.
Generally, from the reinforcement type point of view, the results of the climate change indicator differ for limit states. In the case of the LBC limit state, a more sustainable option seems to be glass fiber reinforcement; on the contrary, in the case of the SA limit state, steel reinforcement can be observed as the more favorable solution. It is possible to observe similar trends for all assessed types of concrete slabs. For this reason, the type of reinforcement is not straightly correlated with the concrete type.
In the case of compensating the properties of an fRAC by the (1) varying thickness of the slab and (2) varying reinforcement level, the results are similar to normalized and weighted results. For all assessed mixtures, the reduction of the climate change impact is more significant when the decline of properties is compensated by more reinforcement.
In addition, concrete mixtures with two amounts of cement were assessed. Concrete (I) contained 260 kg/m3 of OPC and (II) with 300 kg/m3 of OPC. In the comparison, in the perspective of the design oriented to the load-bearing limit state, the result of the climate change indicator is slightly higher for concretes containing more cement where the design of the concrete slab is based on the LBC limit state, and so compressive strength. On the contrary, in the SA limit state point of view, the GWP differs significantly, and it is more favorable for mixtures with higher amounts of cement and, consequently, better mechanical properties, especially the modulus of elasticity.
Considering the results of the climate change indicator, similar conclusions can be drawn based on the results of other environmental indicators, which are presented in the
Supplementary Table S1. The aspects significantly contributing to particular indicators can be discussed based on
Figure 6, which presents indicators representing a broader scale of potential environmental impact chains.
The result of the potential impact in the ozone depletion category is influenced by the use of glass reinforcement production (see
Figure 6a,b), therefore ODP has higher results when glass roving is used.
The results of the ecotoxicity potential indicator are in accordance with results of the climate change indicator (see
Figure 6c,d). The contribution of recycling concrete is negligible. On the other hand, this indicator is affected by transport processes (including diesel production) and the production of raw materials such as cement.
In the case of eutrophication potential (
Figure 6e,f), the production of epoxy resin used for glass reinforcement significantly contributes to the result of the EP indicator. Recycling steel scrap from construction and demolition waste, which is used for the production of recycled concrete aggregate, contributes beneficially to several indicators, not only those that are resource-related (see
Figure 6g,h). The effect of these environmental benefits, which represent the prevention of environmental impacts in the case of steel recycling, can be observed in the results of ADP-min. and ADP-fos. The beneficial impact of steel scrap recycling was presented in a previous study [
28]. While the benefit of this process leads to an overall negative value for results of ADP-min. related to mixtures in which RCA is used, in the perspective of ADP-fos. these benefits result in a rather small decrease of impacts (27–36% in comparison with fRMAC). The ADP-fos. indicator is also affected by the production of cement and fuel for transport (diesel). Another significantly contributing process is the landfilling of disposed concrete slabs, which can reach more than half of the cement production impact. Furthermore, in this indicator, the impact of steel scenarios designated for LBC is slightly higher than the results of slabs with glass reinforcement, which is in accordance with the results of the climate change indicator.
The impact in water depletion potential mainly comes from cement production and landfilling, but an even higher impact is caused by epoxy resin production, which contributes more than the total impacts related to cement production and landfilling. Despite the burden of epoxy resin production, in the comparison of materials for roving, glass and steel reinforcement reach approximately the same result of WDP.
Based on results presented in
Figure 6, use of RCA contributes beneficially to several impact indicators. Moreover, there is a difference between impacts of scenarios with different reinforcement materials, but also production of epoxy resin effects the result of WDP and EP.