Desalination - Team D
Team D: Final Report
Authors: Thomas Aunins, Robert Cignoni, John Dombrowski, Iris Zhao
Instructors: Fengqi You, David Wegerer
March 11, 2016
Water shortage is one of the foremost and most urgent issues facing the world today, as developing and developed countries alike have struggled with depletion of natural reservoirs and severe droughts. This issue has resulted in the recent rapid development of desalination technology and the construction of desalination facilities. Since the turn of the millennium, the United State alone has increased its desalination capacity from 600 million gallons per day to 1650 million gallons per day, with much more currently being planned. California, in particular, is the focus of a large amount of the United States’ desalination efforts, as its current drought has exposed a discrepancy in water supply contingency and demonstrated a need for non-natural freshwater sources.
This project aims to design a 10 million gallon per day seawater desalination plant on the Southern California Bight--near San Diego--to fill this need. A reverse osmosis system was chosen based on the fact that it is the most rapidly developing and innovating technology in the desalination field, as well as the fact that it has a lower theoretical energy production per gallon of water than the common multi-stage flash purification methods. Our plant will pressurize seawater from subterranean wells off the coast of the bight and send it to our pre-treatment system. There, it will go through a drum screen, multimedia filter, antiscalant addition, and finally ultrafiltration to remove varying size of suspended solids and contaminants, before entering our reverse osmosis system.
The RO system itself is a 2-stage, 6 element per stage process, using Dow SW30XHR-440i membranes and operating at 50% recovery with a feed of 20 million gallons per day. This allows the process to achieve a final dissolved solids concentration of 109 mg/L, far below the California drinking water recommendation of 500 mg/L. This freshwater can then be sent to post-treatment and merged with water of the San Diego County Water Authority’s distribution system. Waste concentrate from the process is sent back into the bay through a long diffuser pipe system that will dilute the brine to necessary levels to avoid environmental damage.
An economic analysis of the process found total capital costs to be slightly more than $600 million, with yearly revenues and operating costs at $25.4 million and $6.2 million, respectively. On a 25 year time scale, this results in a final net present value for the project at -$402.5 million, causing us to conclude that as a commercial venture the project is not viable. We do note, however, that increased demand and decreased supply may cause water prices to rise and create a motivation for government investment in the project in the future. For this reason, we believe that it is possible for this project to become an economically feasible and practically necessary venture in coming years.
Due to drought and the depletion of groundwater, desalination is becoming an increasingly viable source for drinking water in the San Diego, California area. A map of the plant location can be found in Appendix 1. Reverse osmosis appears to be the best route for desalination due to its lower energy costs and high volume of current research efforts. It is also capable of purifying California seawater to the levels recommended by the World Health Organization (WHO) and the state government. The process will separate solids from seawater before subjecting it to a two-stage reverse osmosis unit. Concentrated brine waste will be diluted with seawater before going back into the environment. Permeate streams will be remineralized and disinfected before leaving the facility.
The objective of this process will be to produce fresh drinking-quality water according to standards recommended by the Water Research Foundation. This sets an upper limit for the total dissolved solid in our product at 1000 mg/L, with a non-mandatory guideline of approximately 500 mg/L as an appropriate target. This can be found from in Appendix 2. This encompasses the secondary maximum contaminant levels (MCL) set forth by the State Resources Water Control Board . Additionally, there are guidelines set forth for primary MCLs, which encompasses more dangerous and/or toxic substances in the water. These are a smaller concern for our project because sea water does not naturally contain amounts of these contaminants above the MCLs .
Site Location and Capacity
This project is planned for construction on the Southern California Bight, located just north of San Diego and nearby the San Diego County Water Authority’s (SDCWA) distribution system. This area is of particular interest for seawater desalination projects due to the projected discrepancy between water supply and demand in upcoming years. Statewide in California, the demand for water is expected to increase by 1.2 billion cubic meters per year by 2030, as projections show that population increase of 16% dramatically outstripping water conservation goals.  Southern California in particular has a great need for more freshwater sources, as the lower two-thirds of the state require 80% of California’s water, while the upper third of the state supplies 75% of it. 
Per the aforementioned water scarcity, California’s water demand has become a large part of this growth. There are several large scale desalination plants planned for the area, including large-scale projects at Carlsbad and Camp Pendleton. Each of these plants will be constructed to produce 50 MGD of freshwater to the San Diego area, with the latter expected to expand to 150 MGD within ten years of completion. The construction of these plants, along with other smaller scale plants in the area, indicates an urgent need for desalination capacity. Our plant is being designed to produce 10 million gallons per day (MGD) of fresh water for the San Diego area.
Seawater will be fed from a submerged pipeline off the coast of the Southern California Bight. The subterranean feed inlet will allow for an initial pseudo-filter as the water is pulled through the porous ocean floor, preventing large debris and aquatic life from being pulled into the process intake. Worldwide, seawater salinity averages approximately 35,000 mg/L of total dissolved solids, with the primary salts present being chloride and sodium at 19,000 mg/L and 10,500 mg/L, respectively.  It should be noted that while data on average local seawater composition for Southern California was not available, this area is known to typically have lower total dissolved solids concentrations than average seawater, placing our calculations on the conservative side. Further breakdown of the dissolved ion concentration of our seawater input can be found in Appendix 3.
The objective of this process will be to produce fresh drinking-quality water according to standards set by the California state government and the World Health Organization. Regulations set an upper limit for the total dissolved solid in our product at 1000 mg/L, with a non-mandatory guideline of approximately 500 mg/L as an appropriate target. This encompasses the secondary maximum contaminant levels (MCL) set forth by the State Resources Water Control Board.  Additionally, there are guidelines set forth for primary MCLs, which encompasses more dangerous and/or toxic substances in the water. These are a smaller concern for our project because sea water does not naturally contain amounts of these contaminants above the MCLs. 
Further goals for the permeate composition and quality following post-treatment were taken from recommendations given by the Water Research Foundation on seawater reverse osmosis and from averages taken from San Diego water treatment plants. These can be found in Appendix 2.
Process Flow Diagram, Major Technology, and Alternatives
The process flow diagram (PFD) can be found in Appendix 4. Each stream and piece of equipment is labeled according to which section of the process it pertains to. The final simulation mass balance and stream pressure can be found in Appendix 5. Stream tables can be found in Appendix 6.
100 - Pretreatment
The feed flow rate set to the system is 20 MGD. The seawater intake system proposed for the site is a deep infiltration gallery (DIG) intake system due to the permeable hydrogeology offshore of the proposed location. DIG would be comprised of a series of angled or wells drilled radially and therefore would not supply a large amount of feed water due to low permeability. Therefore, the radial collector wells would be drilled at a downward angle from the barge to the dual-use tunnel, below the loose sand layer. The collector wells act as an infiltration gallery, in that the underground seawater infiltrates into the wells and gravity flows into the annular space of the tunnel, which conveys the feedwater onshore. 
The sea plays host to contaminants that extend well beyond salt. Poor feed quality can lead to short RO membrane lifetime, short periods of operation, and high maintenance costs. Contaminants include suspended solids, dissolved organic contaminants, and sparingly soluble salts. 
First off, a drum screen (F-110) will catch any large solids greater than 0.5 cm that could literally throw a wrench in our operations. A multimedia filter (F-120) captures smaller solids from 1 to 20 µm. The media will consist of anthracite, sand, and gravel, providing a gradient from coarse to fine which creates a media flow pattern necessary to achieve a very low silt density index. 
An antiscalant (T-131) will help us avoid fouling of UF and RO membranes by controlling carbonate scaling, magnesium hydroxide scaling, sulfate scaling, and calcium fluoride scaling. Organophosphates tend to be the most stable antiscalant, as they are not subject to hydrolysis or precipitation like sodium hexametaphosphate or polyacrylates. Alternatives to antiscalants that were investigated were water softening and acidification. Both are not economically favorable compared to antiscalants due to additional post treatment measures required when using these methods. Ultrafiltration (F-140), at 0.01–0.02 µm, will remove much of the remaining biological or particulate matter. This pore size also aids in disinfection, as it excludes viruses. These measures will result in a Silt Density index of less than 2.5. 
Conventional pretreatment methods using chemical coagulants such as ferric chloride in concert with Dissolved Air Flotation or Clarifier units were also considered. The equipment and media are long lasting and require low maintenance, but the chemical usage and disposal costs would be higher. UF membranes will need to be replaced every 5–10 years, so they require a moderate running cost. However, this extensive pretreatment process will help reduce RO operating costs and increase process efficiency downstream.  The selected pretreatment method will decrease our environmental footprint and extend the lifespan of our membranes.
200 - Reverse Osmosis
Seawater Reverse Osmosis Technology
The desalination method for this plant will be through reverse osmosis (RO). This method was chosen for a number of reasons. Firstly, new desalination plants appearing in the United States are increasingly run using reverse osmosis technology. The most notable example is the Carlsbad plant that recently opened up near San Diego which produces up to 50 million gallons per day of fresh water. Furthermore, a thermodynamic analysis was done on different desalination methods including multi-effect distillation (MED) and multistage flash evaporation (MSF).  The analysis found that reverse osmosis has the lowest theoretical energy consumption per unit of fresh water obtained. Due to this, building a reverse osmosis plant likely also has the most security moving forward.
Various membrane technology was investigated for use in this process. Thin film composite polyamide membranes are currently the industry gold standard. These have advantages over asymmetrical cellulose acetate membranes due to their higher permeate fluxes and higher salt rejection. Spiral wound membranes are the current state of the art module and are preferable to hollow fiber and plate and frame modules due to their low fouling which can be attributed to the parallel flow of the feed as opposed to the normal flow regime found in the other modules.
The Dow SW30XHR-440i spiral-wound membrane was chosen because each has the capacity for 6,600 gallons per day of permeate (the maximum available from Dow) and the highest overall dissolved solids rejection fraction at 99.82%. Additionally it is the membrane of choice for plants of a similar scale, such as the plant at Carlsbad, which verifies its practical usefulness for such large-scale operations. Based on this choice, it was determined that a 2-stage, 6 element per stage, single pass process would be necessary to achieve the desired flow rate and recovery for a single unit of our operation. A simplified RO system schematic is shown in Figure 1. [IMAGE] Using equations that were presented by Dow Chemical for designing RO plants, it was found that it was possible to produce 10 MGD of fresh water at a recovery of roughly 50% using 2280 membrane elements. These elements would be arranged in a series of 6 elements per pressure vessel for a total of 380 pressure vessels. The pressure vessels would be arranged in a two stage process with 220 pressure vessels in parallel in the first stage and 160 in the second stage. Detailed composition of pass streams from the reverse osmosis process can be found in Appendix 7.
The energy cost component of seawater RO can be up 70% of the total cost, so reducing the amount of energy consumed by the process was essential to minimizing not only cost, but also environmental impact. Energy use reduction is traditionally achieved through energy recovery devices (ERDs), such as centrifugal devices or isobaric, “pressure-equalizing,” devices.  In all cases, energy from the brine stream is transferred directly a portion of the membrane feed stream, reducing pumping requirements. The PFD and stream table detail how the feed is split, with a portion leading to an ERD before entering a booster pump and rejoining the stream from the high pressure (HP) pump. This significantly reduces the size and energy requirements of the HP pump.  Systems utilizing this technology can realize up to 60% energy reduction compared to those without it. 
Centrifugal ERDs incur lower capital costs, but have limited capacity and efficiency, typically running at a maximum of 82% efficiency. This is because they must transfer hydraulic energy from the brine stream into mechanical energy and then back into hydraulic energy.  Isobaric ERDs are the most efficient ERDs, operating at a maximum net transfer efficiency of up to 97%. Isobaric ERDs can handle increased capacity by being run in parallel, similar to the RO membranes. The PX Pressure Exchanger from Energy Recovery, Inc., requires minimal controls, can operate without periodic maintenance, and use ceramic rotors which do not corrode with seawater.  For that reason, it was selected for our process.
The PX Pressure Exchanger can operate at 96% efficiency for our process, and will require 24 units to handle our capacity. 6900 gpm (49.5%) of the feed stream will be redirected towards the PX Array, where it will be acted upon by the concentrated brine stream before flowing to the booster pump (P-213). The rest of the stream will be served by the HP pump (P-211). Through this technology, our process utilizes 8.9 kWh/kgal in the RO section, compared to 17.4 kWh/kgal without, almost 50% in energy savings. Pumping requirements are summarized in Appendix 8. A diagram portraying the simulation of this process is in Appendix 9.
300 - Posttreatment
After the reverse osmosis process, water will go through post-treatment by adding minerals to prevent corrosion of the distribution pipelines and resemble existing potable water supplies. By adjusting the hardness, alkalinity, and pH of the permeate, the aesthetic water quality will be assured and the distribution pipeline will be protected from corrosion.  The post-treatment will include the addition of sodium bicarbonate (T-311) and calcium chloride (T-312) for remineralization, sodium hydroxide (T-321) for pH adjustment, and sodium hypochlorite (T-331) for disinfection. 
Lastly, the product will be held in a holding tank (T-350) before being blended with municipal stores. This will allow for proper quality analysis of TDS, conductivity, and pH. Afterwards, the product water will blend with existing supplies so that the municipality may maintain consistent water quality for all consumers. Existing water treatment plants will ensure the water is suitable for consumption. The blended water can then be delivered throughout the region from there.
400 - Brine Treatment
There are several possible alternatives for brine treatment in large coastal seawater desalination plants. Possibilities include the use of large evaporation ponds, injection of brine into confined aquifers, and discharge into existing bodies of water. The first two options are largely not viable due to high land costs for evaporation ponds and the requirement of comprehensive land surveys for aquifers. Discharging to the ocean, however, is fairly commonly used as it is a reasonably practical option. 
Some smaller-scale facilities have been able to mix their effluent streams with cooling water effluent from nearby industrial plants or additional seawater as a dilution method to reach the necessary 40 ppt range of dissolved salts.  However, this requires either a conveniently located cooling water source, which our plant cannot assume, or prohibitively high costs to pump in enough seawater to dilute our effluent. Another option, and one that will be used at Camp Pendleton, is an engineered diffuser system on the brine discharge outfall. An engineered diffuser system consists of a long pipeline that will release smaller amounts of the brine over the course of its length and promote mixing to achieve dilution requirements. The Camp Pendleton desalination plant’s plans for this system are shown in Appendix 10 as an example. 
500 - Solids Treatment
Solids separated during the pretreatment process through the drum screen, multimedia filter, and UF membrane will be hauled off-site to a suitable landfill. Since no chemical coagulant, such as ferric chloride, is used in the pretreatment process, the spent backwash water can also be conveyed straight to the brine disposal pipeline and discharged to the ocean because the suspended solids contained will be entirely of marine origin.
Assuming that multimedia filters can support 20 gpm/sq. ft. area,  our process will require 4 multimedia filters, each with 200 sq. ft. of area and priced at $34000. Ultrafiltration modules can operate at 30 gpm,  necessitating 467 UF modules, each priced at $500.
The reverse osmosis system was designed in order to comply with the optimal operating conditions for the aforementioned FilmTec membranes and to achieve the necessary dissolved solids concentration and permeate flow rate for a 10 MGD-scale desalination plant. This resulted in 2280 RO elements in 380 pressure vessels. RO elements are cylinders of length 40.5 inches and diameter 7.9 inches.  A range of prices was found for bulk purchases of the required membrane, which could be conservatively estimated at 700 USD per element. The replacement percentage per year for Dow’s membranes filtering this level of SDI is 13%, which was added to the total capital cost of the system. Pressure vessel capital cost was estimated using the Aspen Economic Analyzer, and were found to cost $6700 each. This equipment capital cost was found to be 4.14 million USD. Additionally, there were significant costs associated with construction and auxiliary RO feed equipment was estimated by scaling the 50 MGD Camp Pendleton budget allocation  according to the following equation:
where Ci refers to equipment and construction cost and Si refers to plant capacity. The value for ‘n’ was set as 0.7 based on guidelines from the Chemical Engineering Design textbook.  This extra capital cost was estimated at 56.3 million USD.
Tunnel materials for the feed intake was calculated to cost $32.1 million, while construction costs were estimated at $48.7 million. The feedwater piping, intake well system, and pump station were estimated to have a total capital cost of $55.4 million. Further details on capital cost can be found in Appendix 11.
Concentrate Return and Dilution Pipelines
Concentrated brine disposal was also modeled after Camp Pendleton. Although their scale of brine disposal is much larger than that of this process, capital cost estimates and sizing were not lowered due to the necessity to dig to the same depth and the use of piping with a similar diameter to return brine concentrate. The brine discharge system was estimated to cost $50.2 million, while the brine discharge pipeline was estimated to cost $9.2 million.
(Requirements summarized in Appendix 8)
Ultrafiltration will require a pump in order to filter our process efficiently. Using guidelines from Dow Chemical,  the optimal operating pressure for each ultrafiltration module is 30 psi, and each filter can operate at 30 gpm. As a result, this process will require 467 UF modules. A pump pressurizing 13889 gpm to 30 psi will require 202.8 kW. Antiscalant will need to be supplied to the process at 1.39 gpm in order to properly prevent fouling in pretreatment membranes. The power required for this would be 3.88*10-5 kW.
Assuming 50% recovery during the RO process, the brine flow rate will be equal to the permeate flow rate, 6945 gpm. The PX Pressure Exchanger requires lubrication for its hydrodynamic bearing, which will be supplied by the high pressure brine stream, leading to the slight loss in efficiency. As a result, 6877 gpm (49.5%) of the inlet stream can be redirected to the PX Array before reaching a booster pump, with the remaining 7012 gpm being served by the high-pressure pump. The booster pump will only need to supply 53 psi of additional pressure compared to the 629 psi required from the high pressure pump. In order to meet minimum discharge pressure required for proper PX operation, it is necessary for the feed streams to be pressurized to 30 psi so that the low pressure brine stream will exit at 15.9 psi.
Post treatment chemicals (sodium hypochlorite, sodium bicarbonate, calcium chloride, sodium hydroxide) are added to the permeate in order to remineralize and pH adjust our water. The pumps used to deliver these chemicals must simply overcome frictional losses in the pipe in order to keep the chemicals moving. All pumps were modeled at 80% efficiency.
Chemical Storage Tanks
Chemicals that are added to the water need to be stored beforehand. Chemical holding tanks were sized according to a day’s worth of chemicals. The holding tank for sodium bicarbonate is quite large and this is cause for concern. This issue could be corrected by introducing the solid chemical directly to the product stream rather than creating a solution, storing that solution and then mixing solutions. The cost of the holding tanks can be found in Appendix 12.
Product Selling Price
The San Diego County Water Authority agreed to pay Carlsbad (a plant of comparable size and location) $2014-2267 per acre foot of water depending on how much is purchased.  Based on this number we estimate that our yearly plant revenue will be roughly $25.4 million.
The San Diego County Water Authority agreed to pay Carlsbad (a plant of comparable size and location) $2014-2267 per acre foot of water depending on how much is purchased.  Based on this number we estimate that our yearly plant revenue will be roughly $25.4 million.
The overall capital costs of our plant are summarized below.
|Project Sector||Capital Cost (MM$)|
|Feedwater Intake and Concentrate Return||195.6|
|Product Water Conveyance||90.4|
|Implementation (Legal, Engineering, Administration)||84.3|
The economic viability of our desalination process was analyzed using a 25 year project lifetime. The cost of capital was set at 12% and the tax rate was set at 35%. A ten-year MACRS depreciation model was used.
The project was found overall to not be economically feasible on a purely commercial level. As revenues would only produce approximately $25.4 million per year the net present value after a project lifetime of 25 years remains extremely low at -$402.5 million. The full economic analysis can be found in Appendix 15.
The primary opportunity for optimization within our process occurs in the management of the reverse osmosis process, with respect to the number of stages and the number of elements per stage used for our filtration. As a starting point we chose to use two stages and six elements per stage, as this is in-line with Dow Chemical Company’s product recommendations for the SW30XHR-440i RO membrane. Additionally, this is the typical configuration for large-scale RO plants using this particular Dow membrane such as the Carlsbad Desalination Plant. 
Using the Dow ROSA software, configurations were evaluated for one, two, and three sequential stages, ranging from four to eight elements per stage. Each of these process conditions was evaluated with respect to the capital costs of equipment as well as the yearly utility cost that would be required. The results of this optimization are summarized in Appendix 16, Table 16.1 and 16.2, with our starting condition and minimum highlighted in each.
This optimization was conducted using a desired recovery of 50%. The number of vessels total and number of vessels per stage were based on the maximum permeate flow for the RO membrane and sizing equations obtained from the Dow RO design guide, respectively. 
These data show a minimal variation between different configurations of our system, particularly compared with more dominant capital and operations cost throughout the remainder of our process. However, this process does find a minimum utility cost for the 1-stage, 8 element per stage configuration, at $80,000 per year less than our current setup. Additionally, the condition of a 2-stage, 4 element configuration has a lower utility cost--$60,000 per year less--with an equivalent capital cost. However, it was also observed in either alternative case that the initial element recovery percentage was 10-11%, rather than the 8% achieved in the original 2 stage, 6 element design. A lower recovery percentage indicates lower fouling rates and thus less frequent replacement of membranes, though this precise economic effect could not be quantified.
Based on the manufacturer and industry standard for reverse osmosis configuration, along with the minimal differences in total costs for alternatives and fouling concerns, the project chose to continue with the 2-stage, 6 element per stage configuration.
Our process was found to be particularly sensitive to three main areas within capital costs, operating, costs, and revenue, which could have significant influence over the final economic analysis if the estimations are off by a significant margin or if the data used for these estimations changes significantly in the coming years.
The construction of the project’s seawater intake/disposal pipeline is priced at approximately $200 million, making up about a third of total capital costs. It may be possible, rather than construct an entirely new water feed and disposal system, to draw used seawater from a nearby large-scale consumer and mitigate these construction costs greatly. For example, it is fairly common for desalination plants to use industrial cooling water effluent for their plants, whether it be as a feed or as a dilution measure, in order to significantly drive down capital costs.
Operating Costs and Revenue
The operating costs, while a small portion of total project expenses, consume more than a quarter of the plant revenue at $6.2 million per year. This is largely governed by the energy costs, which have been estimated at $0.08 per kilowatt-hour. A decrease in this cost would substantially decrease yearly operating cost and allow for greater profit.
Similarly, revenue is governed by the cost of water being paid to the Carlsbad Desalination Plant, at $2260 per acre-foot.  This price is expected to rise as the water shortage becomes more urgent and the population of California increases. A significant increase in this price could greatly help the desalination project to improve economic viability.
Overall, the project designed a 10 million gallon per day seawater desalination plant on the Southern California Bight to fill the need of water shortage. We chose reverse osmosis as a method for desalination. The plant would pressurize seawater from subterranean wells off the coast of the bight. The water is then sent to the pre-treatment system before entering the reverse osmosis system. We decided on a 2 stage, 6 element per stage process using Dow SW30XHR-440i membrane and operating at 50% recovery with a feed of 20 million gallons per day. The system achieves a final dissolved solid concentration of 109 mg/L, which well satisfies the California drinking water recommendation of 500 mg/L of dissolved solids concentration. The fresh water is then sent to post-treatment and merge with existing supplies. Waste concentrate from from the process is sent back into the bay through a long engineered diffuser pipe that can dilute the brine to necessary levels.
Furthermore, we did optimization for our process. The primary opportunity for optimization within our process occurs in the management of the reverse osmosis process, with respect to the number of stages and the number of elements per stage used for our filtration. The result shows a minimal variation between different configurations of our system, particularly compared with more dominant capital and operations cost throughout the remainder of our process. Based on the minimal differences, along with the manufacture and industry standard for reverse osmosis configuration, the project chose to continue with the 2-stage, 6 element per stage configuration.
However, based on the results from the economic analysis, we concluded from the final -$402.5 million net present value, that such an energy intensive process to produce a product that is generally taken for granted is extremely costly. As a commercial venture this project is not viable. On the other hand, with increased demand and decreased supply the water price may rise and become a motivation for the government to invest for the desalination in the future. The project would likely need to taken on by the city of San Diego rather than a private entity. For future development and viability of this project, we recommend to draw used seawater from a nearby large-scale process instead of constructing an entirely new water feed and disposal system. This can mitigate the construction costs greatly.