Potential for Potable Water Savings Due to Rainwater Use in a Precast Concrete Factory

: The objective of this paper was to assess the potential for potable water savings due to rainwater use in a precast concrete factory in southern Brazil. The economic feasibility and the rainwater quality were also assessed. The current water consumption, future water demand, and rainwater demand in the factory were estimated. The future demand considered was two times higher than the current water consumption since there were plans to increase the production. Three scenarios were then simulated using the computer programme Netuno . The ideal rainwater tank capacity, the potential for potable water savings, and the economic feasibility analysis for each scenario were estimated. Samples of rainwater were collected in the factory and tested for quality for manufacturing precast concrete. For a rainwater tank capacity equal to 25,000 L, the potential for potable water savings for the ﬁrst scenario was 55.4%, but the ﬁrst scenario was considered economically unfeasible. For the same tank capacity, the second and third scenarios presented viable results regarding potable water savings and payback. As for the rainwater quality, it was proven to be adequate for manufacturing precast concrete. The main conclusion was that rainwater can be used to manufacture precast concrete in the factory studied herein.


Introduction
Rainwater harvesting systems may be installed in different kinds of buildings, such as houses, commercial and residential buildings, factories, and many others. The economic benefit of the installation of the system is still a matter of concern among specialists. The economic feasibility varies significantly for each region, being highly dependent on the water tariff charged by the local water company, the rainfall, and the adequate sizing of the rainwater tanks [1].
A study carried out by Ghisi and Schondermark [2] about the installation of rainwater harvesting systems in houses in the state of Santa Catarina (Brazil) showed that the economic benefits of these systems depend on the water demand. Although there would be environmental benefits due to the installation of rainwater harvesting systems in all houses, there would be economic feasibility only in those with higher water demands or a higher number of occupants.
The use of rainwater can also be applied to situations that demand low volumes of water. Fernandes et al. [3] projected a rainwater harvesting system in a waste treatment installation close to the city of Mirandela (Portugal). Rainwater would be used to wash cars and other equipment, clean outdoor concrete and asphalt floors, and irrigate green areas. Fernandes et al. [3] pointed out that for installations that demand small amount of water, the water harvesting systems would be efficient even if the tank capacity did not supply 100% of the demand. The tank capacity supplied 90% of the water demand.
After going through adequate treatment, rainwater can be used in different applications that do not require potable water. The Brazilian standard NBR 15527 [4] presents

The Factory
The factory to be assessed is located in São José, southern Brazil. It currently has 65 employees and produces different kinds of concrete elements; the water demand is 15,000 L/d on average. The floor-plan areas of the main roofs of the factory sum 15,346.71 m 2 . Such a large roof area could be used to harvest rainwater. The roofs and their corresponding areas, A, are shown in Figure 1.
Water 2021, 13, x FOR PEER REVIEW 3 of 16 pointed out that there were no noticeable differences between the two samples of fresh concrete, regarding the laying process, workability, and appearance. The Brazilian regulation about water for manufacturing concrete, as well as the regulations of other countries, such as Colombia, allows for the utilisation of rainwater for manufacturing concrete, as long as the water meets the quality criteria established. However, using rainwater for manufacturing concrete is a subject that is still poorly explored globally, and the number of studies is still scarce.
The main objective of this paper was to assess the potential for potable water savings due to rainwater use in a precast concrete factory in southern Brazil. The economic feasibility and the rainwater quality were also assessed.

The Factory
The factory to be assessed is located in São José, southern Brazil. It currently has 65 employees and produces different kinds of concrete elements; the water demand is 15,000 L/d on average. The floor-plan areas of the main roofs of the factory sum 15,346.71 m². Such a large roof area could be used to harvest rainwater. The roofs and their corresponding areas, A, are shown in Figure 1.

Data Collection
In order to perform the analyses, we needed to collect some data, such as daily rainfall data. Furthermore, it was necessary to determine the water demand at the factory and to separate this demand into 2 categories. The first category is the one that requires potable water; the second is the one that does not require potability and therefore can be replaced with rainwater. Moreover, other types of data were essential to obtain the potential for potable water savings, such as the roof areas where the rainwater was collected and their runoff coefficients. . This rainfall station was chosen due to its proximity to the factory (3.5 km) and because it has been operating since 1921.

Current Water Consumption
It was not possible to obtain the total water consumption in the factory with accuracy because currently the water consumed is extracted from an artesian well located in situ, and therefore there is no control of the amount of water pumped daily or monthly. For this reason, water consumption at the factory was estimated. In order to estimate the water consumption, we needed to obtain the number of employees and activities that require water. For each activity, the consumption of water per unit (repetitions, m 3 , etc.) was estimated. The number of repetitions of each process and the volume of concrete manufactured were obtained from the reports kept at the factory.
The number of concrete elements produced varies throughout the year. It depends on the demand, but mainly on the factory planning since a great number of elements are produced to maintain the stock. Therefore, the total water consumption was estimated for each month of the year. The estimated water consumption was divided among the working days of the year since the factory only operates on those days.

Rainwater Demand
Currently, the artesian well is being used close to its maximum capacity. Therefore, in case the demand for water increases, it would be necessary to use water from the local water company, which increases costs. Since the company board is planning on doubling the monthly production of concrete elements, demanding a volume of water that is higher than the one available in the artesian well, the future water demand considered herein is twice as much the current consumption.
The volume of water available in the well varies according to rainfall and other environmental factors. Thus, the volume of water available daily varies throughout the year. For this reason, the average water consumption was considered. The average daily water extraction from the well was considered as 14,000 L. This average figure was given by the factory managers, which obtained it throughout the years of the artesian well exploration.
Consequently, the rainwater demand considered was the figure that exceeds the average limit of exploration of the well. This way, on the days in which the future demand was lower than 14,000 L, there was no demand for rainwater. For the days in which the future demand surpasses the average limit of exploration of the well, the rainwater demand will be the future water demand minus 14,000 L. This was chosen to prioritise the use of water from the artesian well since it does not imply costs for acquisition, only costs with electricity for pumping the water.

Runoff Coefficients
The roofs which were considered to be a part of the rainwater harvesting system are made of different materials, such as ceramic, enamelled, and metallic tiles. The Brazilian standard NBR 15527 [4] presents runoff coefficients equal to 0.8 as the most commonly used. For this reason, 0.8 was adopted as the runoff coefficient for the factory roofs.

Potential for Potable Water Savings
The potential for potable water savings due to the installation of the rainwater harvesting system was estimated using the Netuno computer programme, version 4 [15]. Such savings refer to the savings of potable water supplied by the local water company since water savings from the artesian well are not desired by the company board. In order to estimate the potential for potable water savings using Netuno, we needed to obtain daily rainfall data, as well as some other data such as first flush, rainwater catchment area, total water demand, number of occupants, rainwater demand, and the runoff coefficient of the roofs.
The installation of a lower-level rainwater tank and an upper-level tank was considered. The rainwater collected from the roofs would be stored in the lower-level tank, and then pumped to the upper-level tank, allowing the rainwater distribution for consumption to be done by gravity.
The capacity of the upper-level rainwater tank was also input data for the simulations. As for the lower-level rainwater tank, different capacities were simulated. The upper-level rainwater tank was considered to be installed at roof level, and the lower-level rainwater tank was to be placed at ground level. Different roof areas were also simulated. Three scenarios were considered in order to allow the comparison of different levels of investment in rainwater harvesting according to the roof area used to harvest rainwater. In the first scenario, rainwater was collected from all roofs identified in Figure 1; in the second scenario, all roofs but sheds 1 and 2 were considered; and in the third scenario, all roofs but shed 1 and the refectory were considered.
The Netuno computer programme estimates the potential for potable water savings for each lower-level rainwater tank. Netuno's main output data are potential for potable water savings; daily rainwater consumption; percentage of days in a year in which the rainwater demand is fully met, partially met, and not met; and the ideal capacity for the lower-level rainwater tank. More information about the Netuno computer programme algorithm can be found in Ghisi [16].

Financial Analysis
In order to perform the financial analysis for the three scenarios, we required some estimations. The quantity and the diameters necessary for gutters; piping; pipe connections; rainwater tank; and other accessories such as solenoid valve and float switch, among others, were estimated.
The financial analysis was performed using Netuno. The water and sewage tariffs, the monthly inflation, the period in which water and electricity tariffs increase, the period of analysis, and the minimum acceptable rate of return were some of the input data. Moreover, labour costs and the costs of the rainwater tanks, piping, and accessories were needed. When an upper-level water tank was used, the electricity tariffs and the costs of the upper-level rainwater tank and the motor pump were also necessary. Furthermore, other operational costs may need to be considered, such as costs with maintenance or water treatment, when applicable.
The inflation considered was the average of the monthly inflation between October 2017 and September 2018, obtained from the National Index of Prices for the Wide Consumer (IPCA) from the Brazilian Institute of Geography and Statistics [17]. The water tariffs for consumption in industries were obtained from the local water company, and the sewage tariffs were considered null since the factory has its sewage treatment systems in situ. The electricity tariffs were obtained from the local electricity company. Taxes were also considered. The period of readjustment of water and electricity tariffs was considered as 12 months.
The period of analysis considered was 20 years, which is the average lifespan of the motor pump-the equipment of the system that has the shortest lifespan. The minimum acceptable rate of return considered was the average profitability index of the savings account between October 2017 and September 2018.
The net present value, internal rate of return, and discounted payback were obtained using Netuno. The scenarios were considered economically feasible when the internal rate of return was higher than the minimum acceptable rate of return, and the net present value was positive. The net present value can be estimated using Equation (1) [18].
where NPV is the net present value (R$, which stands for Brazilian Real (BRL)), C n is the net cash flow at the period n, and i is the minimum acceptable rate of return. The period considered for the financial analysis (n) was 240 months (20 years). All costs were considered as negative figures and the savings as positive figures. C 0 was calculated considering the costs of the lower-level rainwater tank, the upper-level rainwater tank, labour, piping, accessories, and pumps. The costs considered for C 1 to C 20 included the electricity tariff and corresponding taxes for the operation of the pump. The payback period represents the time when the invested capital is recovered. It is considered that the investment is recovered when the net present value becomes zero. Therefore, the payback period is the value n in Equation (1) when NPV is equal to zero.
The discount rate that results in a net present value equal to zero, considering the period for the financial analysis, is the internal rate of return. The internal rate of return represents the profitability of the investment and can be obtained using Equation (2) (based on Mishan and Quah [19]).
where IRR is the internal rate of return (% per month); C n is the net cash flow at the period n; and n is the period considered for the financial analysis, which was 240 months.

Rainwater Quality
In this study, rainwater is meant to be used only for manufacturing concrete, cutting pieces, and cleaning lanes. Two rainwater samples were collected in two different days from the same location, i.e., the roof of the office. In the first sample, there was no first flush disposal, and in the second, there was first flush disposal. The objective was to compare the quality of the rainwater collected in 2 different situations. The first sample was collected on the 11 September 2018, after 6 days with no rain. The second sample was collected on the 12 September 2018 and there was a first flush disposal equal to 2 mm. The procedure for collecting and storing the samples complied with NBR 15900 [12].
The rainwater quality was assessed according to NBR 15900 [12]. First of all, it was necessary to perform the preliminary tests. After that, if the water met the requirements established for the preliminary tests, the determination of contaminants (sugars, phosphates, nitrates, lead, and zinc) could be made. In case the water was not approved in the criteria established for the preliminary tests, the concrete setting time and concrete strength tests would be mandatory. If the water being tested did not meet all the criteria established for the preliminary tests, it could still be approved for use in concrete, as long as it was approved in the tests of concrete setting time and concrete strength, chlorides (or meet the requirements of NBR 12655 [20]), sulphate, and alkalis (or meet the requirements of NBR 15577 [21]). We chose to perform the preliminary tests and also the concrete setting time and concrete strength tests regardless of the results of the preliminary tests. By performing these tests, the determination of contaminants would not be necessary.
Afterwards, it was necessary to perform the tests for chlorides and sulphate. The determination of the level of alkalis was not necessary since aggregates not reactive with alkalis are used in the factory. The factory uses aggregates from a single supplier, obtained from the same mine, and the non-reactivity of the material is verified using periodic tests for the alkali-aggregate reaction.
Therefore, the following tests were performed: (i) preliminary tests: oils and fat, detergents, colour, solid material, odour, acids (pH), organic matter; (ii) concrete setting time; (iii) concrete strength; (iv) chlorides; (v) sulphate. The preliminary tests and the tests for the level of chlorides and sulphate were performed in duplicate in the Integrated Laboratory of the Environment (LIMA) in the Department of Sanitary and Environmental Engineering of the Federal University of Santa Catarina (UFSC). The oils and fats, detergents, colour, solid material, and odour tests were performed according to the procedures indicated in the NBR 15900 [12].
The test for acid levels (pH) was performed using a piece of pH-meter equipment, which is an instrument used to measure the pH through the potentiometric method. The organic matter tests were performed using the colourimetric method. The concrete setting time test was performed in the Laboratory of Civil Construction Materials (LMCC) in the Civil Engineering Department of the Federal University of Santa Catarina (UFSC) according to the procedure indicated in the NBR 16607 [22]. The standard NBR 15900 [12] defines that for the water sample to be approved in the concrete setting time criteria, the beginning and the end of the concrete setting time must not differ more than 25% from the beginning and the end of the concrete setting time of paste prepared with distilled or deionised water. The beginning and end of the concrete setting time must also meet the criteria established in the Brazilian regulation for the cement used. In this case, the cement used was the CP IV 32.
The strength tests were performed in the laboratory in the factory. In order to perform the concrete strength tests, we moulded concrete specimens according to the criteria established in NBR 5738 [23]. The specimens were ruptured at 7 and 28 days, according to the procedure stated in NBR 5739 [24]. The concrete mix used for all the specimens was the same, which was the standard mix used in the factory. The cement used was the CP V ARI. The test for chlorides in the water samples was performed using the procedure of titulometry. NBR 15900 [12] establishes limits for the chloride levels in the tested water according to the use intended for the concrete. The possible uses are precast concrete or grout, reinforced concrete, and standard concrete. Precast concrete elements and reinforced concrete elements are manufactured at the factory. Since the maximum level of chloride for precast concrete is the lowest (stricter), the maximum limit of chloride level considered was defined for this category, this being 500 mg/L. The sulphate test was performed using a spectrophotometer with a wavelength of 420 nm. The maximum level of acceptance for sulphate in the water is established in NBR 15900 [12] as 2000 mg/L.

Rainfall Data
Daily rainfall data from September 2001 to July 2018-approximately 17 years of data-were used in the analysis. The annual rainfall over 2002 and 2017 is shown in Figure 2. The average annual rainfall over that period was 1775 mm. The average monthly rainfall, as well as the minimum and maximum for each month, are shown in Figure 3. The month with the highest rainfall was January (223 mm); the lowest was June (85 mm). Although rainfall distribution is not homogeneous along the year, long dry periods were not observed, which indicates that it is possible to collect rainwater during all seasons of the year. riods were not observed, which indicates that it is possible to collect rainwater during all seasons of the year.

Current Water Consumption
The estimates for water consumption in the factory are shown in Table 1. Total water consumption ranged from around 74,000 L to 391,000 L per month.  riods were not observed, which indicates that it is possible to collect rainwater during all seasons of the year.

Current Water Consumption
The estimates for water consumption in the factory are shown in Table 1. Total water consumption ranged from around 74,000 L to 391,000 L per month.

Current Water Consumption
The estimates for water consumption in the factory are shown in Table 1. Total water consumption ranged from around 74,000 L to 391,000 L per month.

Rainwater Demand
The water demand considered for the analysis was twice as high as the current water consumption. The total water consumption in the factory can be divided into two categories: potable water and non-potable water. The non-potable water is related to the uses for manufacturing concrete, cutting elements, and cleaning lanes since these activities do not require potable water. The potable water is used for human consumption. The non-potable water demand is the one that can be replaced with rainwater. The total water demand is shown in Table 2. The volume of water available in the artesian well varies according to the time of the year due to various environmental factors. In order to analyse the potential for potable water savings through the implementation of a rainwater harvesting system, we used an average value provided by the factory managers. The amount of water obtained from the artesian well was 14,000 L/d.
The future water demand will be supplied by three different sources: water from the artesian well, water from the water company, and rainwater. When the daily demand is equal to or lower than 14,000 L, it is fully met by the water in the artesian well. When the daily demand is greater than 14,000 L, the exceeding demand is supplied with rainwater, and when the rainwater cannot fully supply the exceeding demand, the demand is met with water from the water company. Thus, the daily demand for potable purposes (which is 4160 L/d) is fully supplied by the artesian well. Table 3 shows the water demand that can be supplied with rainwater, representing the amount that exceeds the 14,000 L of daily consumption. Figure 4 shows the amount of water that can be supplied by the artesian well, as well as the amount of water that will be supplied with water from the water company and that can be replaced with rainwater.

Potential for Potable Water Savings
After defining the water demand that can be replaced with rainwater, we performed simulations using Netuno to determine the ideal rainwater tank capacity. Simulations were run for three scenarios, as explained in Section 2. The input data for the three scenarios are shown in Table 4. The daily water demand considered was the average daily demand on working days; the demand considered was null over weekends and holidays.

Potential for Potable Water Savings
After defining the water demand that can be replaced with rainwater, we performed simulations using Netuno to determine the ideal rainwater tank capacity. Simulations were run for three scenarios, as explained in Section 2. The input data for the three scenarios are shown in Table 4. The daily water demand considered was the average daily demand on working days; the demand considered was null over weekends and holidays.
It was considered that the factory had only one occupant because the water demand was considered, taking into account the total number of employees and the water demand for all the other uses in the factory. It considered the installation of an upper-level rainwater tank of 5000 L so that the rainwater could be pumped from the lower-level rainwater tank to the upper one and then distributed by gravity. Both tanks were considered as being installed outside of the buildings. The potential for potable water savings was calculated for each scenario considering the ideal lower-level rainwater tank capacity obtained through the simulations and also the potential for potable water savings considering a lower-level tank capacity equal to 25,000 L. The 25,000 L capacity was chosen because it is adequate for any of the three scenarios and it can be easily found on the market. The results obtained from the simulations for the three scenarios are shown in Table 5.  Table 3  The results obtained for the three scenarios showed that even though the roof area variation was significant, it did not greatly impact the ideal lower-level rainwater tank capacity, the potential for potable water savings, or the reliability.

Financial Analysis
The cost of labour was obtained by consulting specialised companies in the region; it was estimated as BRL 75,000 for the first scenario, BRL 18,000 for the second scenario, and BRL 25,000 for the third scenario.
The costs of the rainwater harvesting systems for the three scenarios are shown in Table 6. The cost of accessories refers to equipment such as gutters, piping connections, float switches, solenoid valves, and other necessary items for the system. For the financial analysis, the data shown in Table 7 were used as input data to Netuno. The results for the financial analysis simulated using Netuno are shown in Table 8. In analysing the results, we noticed that the first scenario did not present economic feasibility. This happened because the installation of gutters and pipes was considered in all roofs in this scenario. Thus, the initial costs were much higher than the costs for the two other scenarios. On the other hand, the potential for potable water savings was not much higher than the two other scenarios. The second and third scenarios were economically feasible. Even though the second scenario presented the lowest payback, it was the scenario that presented the lowest potential for potable water savings.

Rainwater Quality
The possibility of using rainwater for the production and cutting concrete pieces and cleaning lanes was considered. In order to consider the use of rainwater to manufacture concrete as viable, we needed to perform some analyses in samples of rainwater collected in situ. The objective of such analyses was to evaluate whether the rainwater was adequate for manufacturing concrete and did not affect its strength or durability.
Tests were performed for four samples of rainwater collected at the same place but on different days. The first collection was performed for two samples with no first flush disposal; meanwhile, the second collection was performed for two other samples considering a first flush disposal equal to 2 mm.
The results obtained from the preliminary tests are shown in Table 9. Average results are shown for the quantity analysis. The rainwater from the second collection was approved in all the requirements established in NBR 15900 [12]. However, the first collection did not meet the criteria on the presence of organic matter, even though it was approved according to the other preliminary analysis criteria. If a sample is not approved in these criteria, it does not mean that such rainwater cannot be used for manufacturing concrete. Such rainwater can still be used, as long as it is approved in the concrete setting time test and the concrete strength test. The results of the concrete setting time and concrete strength tests are shown in Table 10, and the limits for acceptance of the concrete setting time and concrete strength tests are shown in Table 11. The limits for acceptance were calculated on the basis of the results for concrete manufactured using distilled water. The standard NBR 15900 [12] defines that concrete samples can be approved if the beginning and end of setting times present a maximum variation of 25% when compared to concrete produced using distilled water. As for the strength of the concrete specimens, the Brazilian standard defines that they must present a minimum strength of 90% of the strength of the concrete specimens produced using distilled water. By analysing the results, we were able to conclude that both rainwater samples were approved in the concrete setting time and concrete strength criteria, since the results meet the limits shown in Table 11. The water collected from the artesian well, which is used in the factory, has its potability assured by tests performed periodically. The standard NBR 15900 [12] considers that if the water potability is assured, it can be considered adequate for manufacturing concrete without the need for further testing. Thus, even though the concrete manufactured using water from the artesian well did not achieve 90% of the concrete strength of the concrete with distilled water, the water collected from the well can still be considered adequate for manufacturing concrete according to NBR 15900 [12].
The results of the chloride tests are shown in Table 12, and the results of the sulphate tests are shown in Table 13. Two samples from the first rainwater collection were tested, representing the rainwater with no first flush disposal. In order to test the chloride and sulphate levels in the second rainwater collection, which considered a 2 mm first flush disposal, we tested two other samples. The standard NBR 15900 [12] establishes limits for the presence of chloride in the tested water according to the use intended for the concrete. The possible uses for concrete are precast concrete or grout, reinforced concrete, and standard concrete. The factory manufactures precast concrete and reinforced concrete elements. Since the maximum chloride level for precast concrete is lower (stricter), the maximum chloride limit adopted was the one defined for this concrete category, which is 500 mg/L. As for the chloride levels, the rainwater from both the first and the second collection was approved.
The maximum level for acceptance of the tested rainwater for the presence of sulphate established in NBR 15900 [12] is 2000 mg/L. Since the level of sulphate in the rainwater samples from the two collections was lower than the acceptable limit, both samples were approved in this criterion.

Conclusions
This paper investigated if rainwater could be used to manufacture precast concrete in a factory located in southern Brazil. The potential for potable water savings due to rainwater use as well as the economic feasibility analysis and the rainwater quality were assessed.
The budgets for the installation of the rainwater harvesting systems were estimated, and the economic feasibility of three scenarios was evaluated. The first scenario did not present economic feasibility; meanwhile, the payback period for the second scenario was 127 months, and for the third scenario it was 181 months (internal rates of return were 1.07 and 0.73% per month, respectively). For the economically feasible scenarios, the potential for potable water savings would range from 50.2% to 54.6%.
The quality of the rainwater was also assessed through tests using rainwater samples collected from the roof of the factory. The samples met the requirements established by Brazilian standards. Thus, the rainwater collected was considered appropriate for manufacturing concrete and for cutting elements and cleaning lanes in the factory studied herein. However, we acknowledge that the rainwater quality should be assessed over a whole year.
In addition to being economically feasible, rainwater harvesting systems can promote the sustainable use of water and also contribute to increasing the impression of clients about the company. Companies which adopt sustainable strategies, such as the use of rainwater, have more visibility and show their clients that the company is concerned with reducing the impact on the environment.