Modeling of environmental aspects related to reverse osmosis desalination supply chain

Background: This study aimed to model optimization of strategic environmental management decisions in the operation of reverse osmosis desalination, emphasizing the costs required for the environmental protection during the production of freshwater using reverse osmosis technology. Methods: This analytical research was conducted in five cities of Hormozgan province in Iran for 18 months from February 2018 to September 2019. The research includes eight phases of defining the research problem, data collection, preliminary data analysis and decision criteria, mathematical modeling, model validation, information preparation, analysis and finally discussion, conclusions and suggestions. The main environmental issues were the carbon dioxide (CO2) release rate due to power demand and rejected brine water (RBW) were entered the mathematical model. Results: The desalination plants of Abu Musa, Bandar Abbas, Qeshm, Sirik, and Hormoz with water production flow rate of 2100, 89 000, 5300, 3300 and 1500 m3/d can generate 2360.82, 100053.80, 5958.260, 3709.86 and 1686.30 tons/year of CO2 emissions respectively. This output requires 1.35, 57.47, 3.42, 2.13 and 0.97 million USD for controlling the process, respectively. For reduction of the negative effect of RBW 0.75, 22.79, 1.78, 1.15 and 0.55 million USD respectively, is needed. Conclusion: Recommendations for environmental impacts protection of RBW, for desalination capacity up to 50 000 m3/d, are; (a) for desalination capacity up to 50 000 m3/d; dilution the RBW using raw water before entering into the sea, (b) for capacity of 50 000-100 000 m3/d; dispersing RBW in sea using diffuser, and (c) for capacity more than 100 000 m3/d; hybrid water desalination plants and power plant. Application of power plant cooling water to dilute RBW may reduce cost.


Introduction
Although water accounts for about 71 % of the earth surface, freshwater scarcity is one of the most important worldwide issues. Oceans and seas contain about 97.5% of total water, while freshwater is only 2.5%, of the total, more than 68% is locked up in polar ice and glaciers, and the rest "30%" is groundwater (1). Iran's renewable water resources are estimated to be 110 to 130 billion cubic meters per year. With a population of more than 80 million, the country's per capita renewable water is less than 1700 cubic meters per year. With any of the internationally accepted indices, Iran is in the state of water stress and dehydration which necessitates the use of desalination.
The worldwide capacity for desalination projects increased dramatically from 326 cubic meters per day in 1945 to more than 95.6 million cubic meters per day in 2016 (2). Among the existing desalination technologies, reverse osmosis membrane technology accounts for 66% of the capacity utilized, followed by multi-stage flash and multi-effect distillation with 20% and 7% of the capacity utilized, respectively (3). Seawater accounts for 58% of the world's desalination water feed (4). Environmental impacts of seawater reverse osmosis (SWRO) desalination can be broadly classified into three categories, including energy consumption which releases carbon dioxide (CO 2 ) into the atmosphere, intake and brine discharge (5,6). Desalination effluent results in known environmental effects on seagrass habitats and phytoplankton, invertebrates and fish communities in areas surrounding effluent discharge (7,8). Overall, enclosed and shallow sites with abundant marine organisms are more sensitive to effluent discharge than the offshore sites capable of diluting and dispersing plant rejected water (2,9). The previous studies have shown the variable effects of desalination plants on the salinity of the received water. Based on the studies, the effects of saline effluent discharge can be observed for tens or hundreds of meters (10,11), or in extreme cases, several kilometers from the effluent discharge site (12). Few studies on the environmental issues of intake and brine discharge of SWRO, have shown that proper design of SWRO based on the environmental impact analysis, can minimize the environmental impacts and the costs of environmental protection for desalination plants were analyzed (5,(13)(14)(15)(16). Most of the above-mentioned models focused on optimizing the economic dimension of water supply systems and often overlook the details of the environmental aspects. This research expanded the economic model presented by Al-Nory et al (17) regarding the water supply chain of the desalination plant by emphasizing environmental details. Environmental details include modeling the reduction of salinity and chemical components of the rejected brine water (RBW) from plants by diluting the effluent before being entered the sea according to a standard that permits discharge into the receiving water. Note that, CO 2 emissions have also been modeled as an environmental impact. Desalination supply chain activities include obtaining feed water and chemicals needed for the desalination processes, desalination process systems, water storage and distribution of freshwater to endusers (18). The importance of examining the economic and environmental impacts of water purification using desalination technology and the provision of desalinated water allows decision-makers to examine the system as a whole (19). For instance, any delay in the distribution of water from storage tanks to consumers could disrupt the desalination process and affect the overall performance of the desalination water supply chain. In this study, due to the importance of environmental issues in terms of produced water costs and environmental protection aspects, this subject is evaluated and modeled in Hormozgan province as the center of the desalination of Iran.

Materials and Methods
This analytical research is conducted in 5 cities of Bandar Abbas, Qeshm, Hormoz, Sirik, and Abu Musa in Hormozgan province ( Figure 1 In the present study, the typical SWRO plants are studied. The main parts of typical SWRO plants are included in the Intake section, pretreatment (generally coagulation and granular filter), high-pressure pumps and membrane modules. After that, a post-treatment unit is located to add some minerals to the RO water product ( Figure 2). Modeling fundamental data include the rate of raw water intake, water production, and brine water flow rate, type of discharge into receiving waters, and investment cost and operating costs such as the consumed power used chemicals, manpower, and the other economic and environmental aspects. To evaluate the CO 2 production, since the power generation model in Iran is almost similar to the water production by desalination, the mathematical model of electricity production in Iran is used to estimate the amount of CO 2 emitted from desalination plants (20,21) Since the power generation model in Iran is almost similar to the Portuguese power generation model collected from the literature (21), this model was used to estimate the amount of greenhouse gas CO2 required by the mathematical model.

Model parameters
In this analysis, a mathematical model is employed which its parameters are shown in Table 1. Note that the model is based on AL-Nory et al (17) and Balfaqih et al (1).

Decision variables
Decision variables for the production and investment in desalination plants and transmission lines presented in Table 2. The total cost of water (TWC) is often cited in the literature of the desalination industry as a common comparison between projects. Table 3 shows comparative evaluation of the total cost of the objective functions and their components per cubic meter of freshwater.    Ef CO 2 emission factor (CO 2 kg-e/kwh) opx CO 2 CO 2 cost ($/kg CO 2 )

Pc
The v c l coefficient as the percentage of the total investment cost for the effluent dilution cost Po The v lh o coefficient as the percentage of the total investment cost for the effluent dilution cost

Target function optimization
The objective function specified in Eq (1), minimizes the total investment cost and supply chain operation of both the plant and the transmission line. Furthermore, it minimizes environmental impacts.
The model has a few limitations. Equation (2) represents the net present value (NPV) of the total investment cost (CAPEX) of the plant at the location (l).
Equation (3) represents the NPV of the operation cost at the location (l) in the time of (h).
Equation (4) represents the NPV of the residual value of the plant at the location (l) at the end of the design period.
Equation (5) represents the NPV of the total cost of the plant at the location (l).
Equation (6) represents the NPV of the total investment cost (CAPEX) of the transmission line at the location (l).
Equation (7) represents the NPV of the total operational costs (OPEX) for the transmission line at time (h).
Equation (8) represents the NPV of the total costs for the transmission line.
The binary variable is the establishment or nonestablishment of the plant as Eq (9). Moreover, the binary variable is the establishment or non-establishment of the transmission line as Eq (10). Y t = 0 or 1 (9) Y i = 0 or 1 (10) Equation (11) shows the water produced rate by the plant (t) at the location (l) in the time horizon (h) limited by the plant capacity at the location (l) with the cfl th capacity coefficient. (11) Equation (12) denotes the amount of freshwater that responds to the locations water demand at location (l) in the time period (h), which is equal to or greater than the demand.
Equation (13) shows the amount of fresh water entered into and out of the aggregator; input amount is equal to the output.
The total amount of water produced by the desalination plant (t) at the location (l) enters the transmission line according to Eq (14).
The flow rate of water at the transmission line (i) at the time (h) is limited to the capacity of the transmission line (i) and is as Eq (15).
The CO 2 emission value of the desalination does not exceed the emission limit for CO 2 emissions and is as Eq (16).
The total amount of CO2 emissions at the location (l) is as Eq (17).
The u co2 lt , CO 2 emissions produced by the plant (t) at the location (l) for one cubic meter of water is as Eq (18).

= *
Where Er lt is the energy required by the (t) plant kWh/ m 3 and Ef is the CO 2 emission factor (kg CO 2 -e/kWh). The NPV of the total cost of CO 2 emissions in the period design is (H) and is as Eq (19).
The opx CO 2 is the cost of CO 2 in $/kg CO 2 . The NPV is the cost of diluting the brine discharge to reduce the salinity and effluent chemicals is as Eq (20).
The NPV is the total cost of reducing the environmental impacts of the desalination plant which is expressed in Eq (21).
The cost of operating one cubic meter of water at the opx lt T plant is presented in Eq (22).
Remark that the capacity of the cap T t plant is expressed as m 3 /d. The coefficient capacity of cf lth is considered for the plant of t∈T l at the location of l∈N s at time h∈H. The estimation of the opx i N operating costs per cubic meter of water in the transmission line i is as Eq (23).
Remark that the capacity of the transmission line cap i N is expressed as m 3 /d. The total cost of the water supply chain (TWC) for comparing the objective functions and its components per cubic meter of freshwater is as follows.
To make the target functions understandable and comparable, it is expressed in terms of TWC, the total cost per cubic meter of freshwater (US$/m 3    Based on the previous studies, the total cost of a carbon tax was ranging from $19 per ton (22) to $23 per ton (23). In this study, it is estimated at $23 per ton or $0.023 per kilogram of CO 2 . According to the standards of Iran Environmental Protection Agency (EPA), a maximum of 10% salinity in the receiving water into discharge the effluent is allowed (24). The effluent dilution cost is estimated as a percentage of the total cost of operating and operating for desalination plant. The mathematical model is then coded in MATLAB software and solved using an opti-intlinprog solver from the OPTI Toolbox. By changing the parameters, the sensitivity analysis of the designed model is presented.

Results
The model for the existing water desalination supply chain in the cities of Abu Musa, Bandar Abbas, Dargan Qeshm, Sirik, and Hormuz is solved based on the data collected during 20 years and presented in Table 3. The model output decision variables are presented in Table  4. Comparative evaluation of the total cost of objective functions and their components in terms of cubic meters of fresh water is also illustrated in Figure 3.

Discussion
The function of "environmental target" is investigated in two parts including environmental impacts of CO 2 emissions and environmental effects of saline effluent disposal. For desalination supply chain in Bandar Abbas, Total TWC 5, TWC 3, plant, TWC 4, transmission line TWC 9 CO 2 , TWC 10 Wastewater dilution are 0.5334, 0.3640, 0.0458, 0.0885 and 0.0351 US$/m 3 respectively, presented in Figure 4.
Environmental costs of CO 2 emissions control Increased use of fossil fuels for desalination can increase air pollution caused by CO 2 emissions and cause damage to public health and the environment. For energy consumption of 4 kW/m 3 , the portion of environmental costs related to CO 2 emissions is computed by about 16.59% which equals to 0.0885 US$/m 3 ( Table 5). As shown in Table 6, the Bandar Abbas desalination plant with a nominal and actual capacity of 100 000 and 89 000 m 3 /d respectively, annually produces 10 008, 800 kg CO 2 , assuming 4 kW of energy consumption per a cubic meter fresh water production. Reducing energy consumption leads to a reduction in the amount of CO 2 . With more efficient utilization and energy recovery, energy can  be reduced to some extent to descent CO 2 emissions. However, the main solution is to use renewable energy instead of fossil fuels.
Dilution costs to reduce environmental impacts The desalination plants use significant amounts of chemicals for the pre-treatment of saline and freshwater (25). Excessive salinity of effluent and discharge of large quantities of chemicals into coastal waters results in ecological imbalance and have major effects on receiving waters. To address environmental concerns related to effluent discharge into seawater, the concentration of chemicals and salts in the effluent should be reduced. The target function of "reducing environmental effects due to the effluent salinity" is transformed into a cost function. According to the Table 6 for dilution of different ratios of total dissolved solids (TDS) effluent to TDS receiving water, the percentages of the total cost of wastewater investment and total operating cost are estimated and entered into the model, the results of which are shown in Table 6. As shown in this table, the environmental goals are against environmental costs; hence, by higher costs, it can  TWC1  TWC2  TWC3  TWC4  TWC5  TWC6  TWC7  TWC8  TWC9  TWC10   US$/    meet better environmental quality related to desalination. More accurate cost estimates should be spent on a specific project and its local data. The cost of "wastewater management" is of great interest as the cost of wastewater disposal increases with the production of freshwater. The dilution cost (TWC 10) for the TDS of effluent to receiving water ratio is the US $ 0.0351/m 3 . The decision-maker can attract the attention of the environmental organization according to the cost and reduce the environmental impacts. In case the ratio of TDS effluent to TDS intake water is considered to be in the standard range of 1:1, the share of dilution costs will be 7.89%, according to Figure 5.
To dilute the RBW in seawater, diffuser equipment is required, and to achieve a water flow rate of 5 to 8 meters per second, energy should be consumed. Depending on the site-specific conditions, discharge costs to the seafloor for dilution at sea are significant, typically accounting for 10 to 30 % of the total investment costs of the desalination plants (26). Previously many researches focused on supply chain management coordination and optimization challenges in different industries and circumstances (27)(28)(29)(30). Even some scholars considered environmental and social issues in their studies (31)(32)(33).

Conclusion
This study aimed to model optimization of strategic environmental management decisions in the operation of reverse osmosis desalination; emphasizing the costs required for environmental protection during the production of freshwater using reverse osmosis technology. Due to the relatively high costs of controlling environmental pollutants, unfortunately, many of these desalination plants remain neglected. intake water before entering seawater for desalination up to 50 000 m 3 /d. Eventually for 100 000 m 3 /d water desalination plants and power plants are combined in one place, meaning that using power plant cooling water to dilute the desalination effluent before entering the effluent into the seawater to save the dilution cost. As the pretreatment units are very important determinants in desalination costs, it is suggested that a study should be conducted to evaluate the effects of different pretreatment methods on desalination costs. The limitations of this study were the non-inclusion of small components of investment and operating costs as well as the non-inclusion of the social dimension of sustainability in the mathematical model as it made the model more complicated.