Arsenic Removal From Ou Water

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

Arsenic Removal from OU Water C Chheem miiccaall E Ennggiinneeeerriinngg-- U Unniivveerrssiittyy ooff O Okkllaahhoom maa R Roossss C Chhaaffffiinn R Raannddyy G Goollll SSaam mii K Kaarraam m R Room maann V Voorroonnoovv

SSpprriinngg 22000033

Page 1 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

Executive Summary In 2000, the EPA mandated that the arsenic level for water distribution systems must be reduced to 10 ppb (parts per billion) by the year 2006. The results of our Senior Capstone Project show how the University of Oklahoma can save several million dollars over a long term horizon, by using treatment methods, and only purchasing water from the City of Norman occasionally at peak demands. In order to comply with the new regulation, our project recommends the construction of an Ion Exchange (IX) treatment facility as opposed to water purchase from the City of Norman. The Net Present Cost (NPC) of purchasing water for the next 20 years, that is the total cost at 2003 dollars, is estimated to be approximately $8,000,000 at the current water price of $1.14/1000gals. In comparison, the NPC of building a treatment facility and treating the water for Arsenic has only a Net Present Cost of approximately $3,000,000, for the same project lifetime. In fact, in order for the Water Purchase (WP) option to become more economically attractive than treatment, the water price charged by the City of Norman would have to be lower than $0.45/1000gals. A price so low could not be offered by the City, since it costs Norman roughly $0.60/1000 gals to produce it. Assuming that the University of Oklahoma borrows money to sponsor the project, the money saved in the first year alone by building an IX treatment facility instead of purchasing water would amount to $150,000. Furthermore, the savings would increase until the loan is paid off and would roughly double after the first ten years, when no more payments are made. While the initial capital investment required for the Ion Exchange facility is substantial, the savings made by implementation of treatment versus water purchase begin in the first year of the project and greatly outweigh the inconvenience of borrowing money. In addition to the substantial economic benefits provided by treatment, construction of a treatment facility would allow the University of Oklahoma to remain independent from the City of Norman, as Norman faces water shortages of its own; and may be forced to purchase water on an emergency basis from Oklahoma City. Conclusion: Over a twenty-year period, the total savings provided by the construction of a treatment facility will be in excess of 4.8 million dollars. At the current commercial price of $1.14 per 1000 gallons, the yearly cost of water purchase will be near $500,000, and will increase with the water demand of the University.

Page 2 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

1.1 Table of Contents 1.3 Table of Figures 1.4 Table of Common Abbreviations

4 5

2. Introduction 2.1 Arsenic Law 2.2 OU Arsenic Situation 2.3 Water Demand Data 2.4 Arsenic and Ion Exchange Chemistry

7 7 10 11

3. Work Previously Done 3.1 Water Treatment Options 3.2 Water Purchase 3.3 CH2M Hill Report 3.4 CE 5244 Arsenic Report

14 15 15 17

4. Treatment Analysis 4.1 Overview 4.2 Ion Exchange 4.3 Coagulation/Filtration 4.4 Nanofiltration 4.5 Comparison of Previous Conclusions

19 20 24 26 31

5. Detailed Ion Exchange Model 5.1 P&ID 5.2 Variations with Capacity 5.3 Safety

33 34 35

6. Planning Model, Results and Recommendations 6.1 Planning Model 6.2 Model Results 6.3 Financial Analysis 6.4 Risk Analysis 6.5 Recommendations

39 42 44 47 50

7. Supporting Documentation 7.1 References 7.2 Appendices

51 52

Page 3 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

1.2 Table of Figures Tables 2.1 3.1 3.2 4.1 4.2 6.1 6.2

OU Well Data Plan alternative cost summary – p. 6-16 of CH2M Hill Report Estimated costs for the Ferric Chloride coagulation O&M Costs, CF Plant Nanofiltration Facility Present Worth Results for a 20 yr Project Lifetime Plant Costs Used in Model Deterministic Results

8 16 17 25 30 41 42

2.1 2.2 2.3 2.4

Norman and Campus well field Arsenic Concentration Gradient Wells and Piping Layout Demand By Month

8 9 10 11

4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9

Targeting of Number of Columns Ion Exchange Capital Costs Ion Exchange O&M Costs Capital Costs for Coagulation/Filtration EPA Preliminary TCI Graph for CF Nanofiltration PFD Capital Costs for Nanofiltration Capital Costs for Nanofiltration, Worst Case NPC Comparison of Previous Solutions

21 22 23 24 26 27 29 30 31

5.1 5.2 5.3 5.4

P&ID Capital Investment vs. Capacity Operating Cost vs. Capacity Bed Volumes treated vs. Breakthrough

33 34 35 36

6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 6.10 6.11

Water Consumption Projection Purchased Water over project lifetime Loan and Repayment Yearly Cost with Loan Yearly Cost without Loan Savings NPC at different Prices of Norman Water Sensitivity of NPC to Water Cost Sensitivity of NPC to Maximum Field Capacity Sensitivity of NPC to Capacity Unit Cost NPC Probability Distribution

41 43 43 44 45 45 46 47 48 49 50

Figures

Page 4 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

1.3 Table of Common Abbreviations

As CE cf CF CI EPA gal gpm IX kWh MCL MGD NF NPC O&M OC OKC OU ppb ppm USGS WP

Arsenic (including all forms) Civil Engineering Cubic Feet Coagulation and Filtration Capital Investment Environmental Protection Agency Gallon gallons per minute Ion Exchange Kilowatt Hours Maximum Containment Level Million Gallons per Day Nanofiltration Net Present Cost Operating and Maintenance Operating Cost Oklahoma City The University of Oklahoma Parts Per Billion (µg / L) Parts Per Million (mg / L) United States Geological Survey Water Purchase

Page 5 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

Introduction • • • •

Arsenic Law OU Arsenic Situation Water Demand Data Arsenic and Ion Exchange Chemistry

Page 6 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

2.1 Arsenic Law____________________________________________ Under the 1996 amendments to the Safe Drinking Water Act, the EPA reviewed the acceptable arsenic levels in drinking water. As of January 22, 2001, the previous levels of 50 ppb will be replaced by a 0 ppb non-enforceable and 10 ppb enforceable levels of Arsenic in all public drinking water systems effective as of January 23, 2006. The reason for this recent change from the old 50 ppb rule, which was established in 1975, is that a March 1999 report by the National Academy of Sciences concluded that the current standard does not achieve EPA's goal of protecting public health. Recent studies have linked long-term exposure to Arsenic in drinking water to cancer of the bladder, lungs, skin, kidney, nasal passages, liver, and prostate. Non-cancer effects of ingesting Arsenic include cardiovascular, pulmonary, immunological, neurological, and endocrine (e.g., diabetes) effects. Short-term exposure to high doses of Arsenic can cause other adverse health effects that are reversible in some cases. (Reference #6).

2.2 OU Arsenic Situation____________________________________ 2.2.1 Well Data With the new Arsenic rule, the OU campus is now faced with the ultimatum to find a solution for its Arsenic problem. Most of the problem wells in the Norman area are located on campus property and are used by the University of Oklahoma to supply on-campus buildings with potable water. With an average Arsenic content of 35 ppb for all the wells being presently used, this automatically makes the whole system non-compliant with the 2006 deadline. Norman is located atop of what is known as the “Garber-Wellington” field. This water aquifer is one of the main sources of drinking water used by the city of Norman and the OU campus. When the Navy donated all the buildings and water wells to help in the formation to what is now present day OU, technology wasn’t available to detect the high levels of Arsenic present in the wells drilled. Figure 2.1 is a USGS map that shows the locations of all of the OU potable water wells. All of the wells are in or around the Westheimer Airpark/Research Facility, which is the location of the old Navy base nestled between I-35 and highway 77, just north of Robinson Street.

Page 7 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

Figure # 2.1 OU Campus well field/Westheimer Airpark

OU currently operates 11 of those wells (FIGURE 2.1) to feed the Norman campus with potable water. The OU well data table (TABLE 2.1) illustrates the flow rate variations for each well; this is due to variation in water consumption on campus. The wells are operated in an automated fashion that regulates the flow accordingly. Also included in that table are the Arsenic levels observed in each well. Table#2.1 OU individual well data (2000-2002) Well #

Average Capacity (g/min)

Maximum Capacity (g/min)

Average Arsenic (ppb)

2 3 4 5 6 7 8 9 12 13 14

110 120 140 --160 120 180 120 160 120 150

125 200 150 --200 160 200 165 200 160 160

94.1 68.7 30.7 --14.8 21.8 22.4 66.2 94.7 32.9 33.9

AVG:

1380

1720

48.02

Page 8 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

The Arsenic average in TABLE #2.1 (48.02 ppb) is not representative of the published average (35ppb) because readings taken at well source and averaged are different than readings taken at the delivery pipeline after all well water has been mixed. Automation cause different wells to be operated at different rates thus affecting the final Arsenic content in the delivered water. 2.2.2 Geological Data The “Gerber-Wellington” well field is located atop a geological formation that favors the precipitation of Arsenic in the water aquifer. The underground water moves in South-West current, which explains why Arsenic concentrations increase in that direction. This is illustrated in the concentration gradient graph (FIGURE 2.2).

Figure#2.2: Arsenic Concentration Gradient “Gerber-Wellington” Field

2.2.3 Current Piping Layout Since the new wells can only be drilled on University-owned land, the spatial restrictions of the drilling locations limit the options available to the University of Oklahoma. Any wells drilled outside campus property belong to the City of Norman. All the operational wells are situated in the Westheimer Airpark North of Robinson Street along HWY 77. A piping network connects them all to one main pipeline going down Berry Street towards the main campus as illustrated in FIGURE 2.3. Some wells (like well #9) are no longer on OU property.

Page 9 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

Figure#2.3: Westheimer Airpark/Research Facility, Wells & Piping Layout

Water Collection Pipelines Well site

As it is seen from Figure #2.3, the geographical restrictions are extremely limiting. This plays an important part in narrowing down the options available for the OU campus, since the University does not have the option of simply drilling new wells in the northeast region of the “Gerber Wellington” field.

2.3 Water Demand Data Though this project is not specifically concerned with the water demand at OU, any analysis of a water treatment facility should include the parameters of the demand that should be met. OU’s maximum water consumption peaks toward the spring semester when residence halls and overall demand increases due to the heat. Currently, peak consumption reaches 1.5 million gallons per day. The maximum capacity of the water grid is around 2.0 million gallons per day, due to water tower limitations.

Page 10 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

2002 Demand by Month 1600

1400

Demand (1000 gal/day)

1200

1000

800

600

400

200

0 Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

Figure #2.4 Water Demand by month

2.4 Arsenic and Ion Exchange Chemistry Arsenic is found naturally occurring in ground water; and can be found in two forms, Arsenic (III, or arsenate, and Arsenic (V), or arsenate. Arsenite (H3AsO3) occurs predominantly in the non-ionic form when the pH of water is less than 8.01. When the pH is above 9.0, arsenate (HAsO42-) in the ionic form is more prevalent in water. The average pH in the OU wells is 8.97, which means that the arsenic in the water is naturally occurring in the non-ionic arsenite state and the ionic arsenate state. The ion exchange process will not work effectively unless the arsenic is in a charged state. Arsenic (III) must be oxidized to arsenic (V). Sodium hypochlorite (NaOCl) will be added for the oxidation reaction to occur. Below is the oxidation reaction that forms arsenate. The added NaOCl will cause the pH of the water to go down, so sodium hydroxide (NaOH) is added after the treatment to bring the pH to normal standards. H3AsO3 + NaOCl ->HAsO42- + NaCl + 2H+ (oxidation reaction arsenite to arsenate)

1

www.hdrinc.com/corporate/JumpStories/WQTCPoster.pdf Page 11 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033 After arsenite is oxidized, arsenate is formed; the water with arsenate anions will be run through the resin bed. The resin bed is made up of polymeric beads that have been pre-saturated with chloride anions on their outer layer. When the water is run through the bed, arsenate ions adsorb to the resin beads; and trades places with the chloride ions, hence the name ion exchange. Below is the reaction where arsenate binds to the resin. 2RCl + HAsO42- -> R2HAsO4 + 2Cl- (reaction where arsenate binds to the resin) As can be seen, the chloride ion (Cl-) is bound to the polymeric bead (R). The polymeric bead has a higher selectivity for arsenate, and knocks the chlorine ion from the polymer. The resin beads exchange with other ions also. In fact the selectivity for sulfate (SO42-) is higher than the selectivity for arsenate. Consequently, the sulfate can overtake the bed causing the arsenic ions to be released back into the water if the ion exchange bed is not monitored properly. If OU chooses to go through with ion exchange, the arsenic content, although still very important, will no longer be the most important measurement taken. Sulfate content will need to be monitored to ensure that sulfate is not overtaking the exchanging of the arsenate ions. After the process is complete, the bed must be regenerated to remove the arsenate ions from the resin bed. The bed is then regenerated with concentrated sodium chloride (NaCl) in the reaction as follows. R2HAsO4 +2NaCl -> 2RCl + Na2HAsO4 (bed regeneration reaction) The regeneration is driven by concentration. The sodium chloride has such a high concentration the arsenate ion is knocked off of the resin beads, and the chloride ions are then replaced to the beads. After this process, the concentrated NaCl with the arsenate ions is sent to a precipitation tank. Sulfuric acid is added to this stream to bring the pH down, so ferric chloride (FeCl3) can be added to precipitate ferric hydroxide Fe(OH)3. The arsenic finally adsorbs to the Fe(OH)3, is dried, and sent to a landfill.

Page 12 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

Work Previously Done • • • •

Water Treatment Options Water Purchase CH2M Hill Report CE 5244 Arsenic Report Findings

Page 13 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

3.1 Water Treatment Options Ion Exchange (IX): Simple IX on a basic media is an attractive solution for arsenic removal, due to the simplicity of the process. Water is run through a resin and Arsenic replaces a nontoxic ion on the resin, resulting in purified water downstream of the resin. After all adsorption sites are filled, the bed must be returned to its initial state through regeneration via pH adjustment and exchanging the arsenic ion on the resin with the initial, non-hazardous ion. Infinite Brine Recycle: One possible way to avoid consistent replacement of the exchange resin is by regeneration process via concentrated sulfate solution. The ‘brine’ may be reused, due to its high concentration and selectivity, and the As may actually remain in the brine for several cycles. Activated Alumina (AA): Ion Exchange involving activated alumina is highly selective to arsenate ions. Cost is also very comparable to Ion Exchange. However, three basic problems limit the use of activated alumina: rapidly diminishing capacity, narrow 5.5-6 pH operation range, hazardous waste with costly disposal. AA appears most promising for small communities, especially those with a high sulfate concentration that would make regular IX difficult. Reverse Osmosis (RO): RO is a membrane separation method allowing the removal of the finest particles, even ions. The membranes employed by this method have pores on the order of 1 A& and are operated in Cross Flow mode. Osmotic pressure must be overcome and permeate flow is driven by high operating pressures (200-400psig). Nanofiltration (NF): NF is a membrane separation method that has proven to achieve Arsenic rejection efficiencies in the range of 80-90%. The membranes employed by this method use the principle of RO to filter out Arsenic and are operated in Cross Flow mode. NF membranes have pores on the order of 1nm. Point of Entry (POE) and Point of Use (POU): POE/POU technology is a strategy that suggests treating only the potable water instead of treating all of the water via a centralized water treatment plant. Coagulation/Filtration (CF): Ferric Chloride reacts with arsenic in water to produce Ferric Hydroxide, and then forms a precipitate. The water is then run through a very large bed of sand and iron fillings where the arsenic is precipitated and filtered out of the water. The large bed of sand and iron that is used as a filter will last for several years. Lime Softening: Calcium hydroxide is first added to drinking water until the drinking water reaches a pH of 10, and then sodium carbonate is used to cause the arsenic to precipitate out of the water. Consequently, Arsenic is removed via simple filtration. Toxicity of the chemicals used in Lime Softening present a safety issue to its usage. Electrodialysis Reversal (ER): Salt in water is dissociated and then passes through ion-selective membranes. Usually used for water desalination, it was found that an 80% efficiency in removing arsenic can be achieved using this method. As with RO, this process generates a waste stream (20-25% of inflow) highly concentrated, which means that part of the treated water has to be disposed of, creating a problem for water starved regions. Page 14 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

3.2 Water Purchase Option__________________________________ To date the University of Oklahoma does not have a solution to the lowering of Arsenic level limits in 2006. The only solution at this point is buying water from the City of Norman. Buying water from Norman may not be the least expensive option available to OU. Currently, it is unclear how much Norman will sell the water for. The City of Norman currently sells water to OU at a rate of $1.14/1000 gallons. Additionally, OU is negotiating a deal with Norman involving sewers, waste, and water purchase to buy water at a rate of $0.85/1000 gallons of water. The rate of $0.85/1000 gallons is the minimum suggested price that the City of Norman could charge for its water, considering that it spends $0.60/1000 gallons to produce it2. The current peak water usage for OU in August is about 1.8 million gallons per day of water. For the economic analysis 1.1 million gallons of water was used. Variations in monthly water usage and growth of the university will be taken into account. At $1.14/1000 gallons and 1.1 million gallons per day the university will spend $475,000 per year and $5,565,000 over the next 20 years. At $0.85/1000 gallons the university will spend about $354,000 per year and $4,150,000 over the next 20 years.

3.3 CH2M Hill Report In 2002, the City of Norman and the University of Oklahoma hired CH2M Hill, an engineering firm based in Colorado, to do an analysis of the water supply in the Norman area; and find the best way to become compliant with the 2006 MCL of 10 ppb. The report treated the well field of Norman and OU as one entity, and the solution given was the best for the community as a whole. The report is public domain, and available upon request from the Norman Public Library. 3.3.1 Options Analyzed Initially, treatments were analyzed separately from the general options, with Ion Exchange, Nanofiltration, Coagulation/Filtration, and Granular Ferric Hydroxide among the examined. Norman city officials examined the five treatments assuming that treatment would be used on a large scale, for all of the contaminated water in the Norman area. Using both economic and social factors, such as waste, Granular Ferric Hydroxide was chosen as the “best” technology. The report then analyzed five options: water purchase from Oklahoma City, blending with water purchase, zonal blending with the creation of new wells, treatment, and a combination of treatment, blending, and new well creation.

2

Source: Don Carter, OU Physical Plant Page 15 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

Table #3.1 Plan alternative cost summary – p. 6-16 of CH2M Hill Report

Plan Alternative Plan A – Water Purchase Plan B – Blending and Water Purchase Plan C – New Wells and Blending Plan D - Treatment Plan E – Treatment, Blending and New Wells

Capital Cost $0.7

Annual O&M $4.1

Present Worth $47

$3.9

$2.8

$37

$9.2

$1.3

$24

$17.2

$2.2

$42

$12.4

$1.7

$32

Costs in millions of dollars

Plan C was recommended by the report because it provides the following3: • • • •

Independence from relying on OKC treated water Non-treatment options for achieving compliance Development of non-potable uses for non-compliant groundwater wells, decreasing demand on OU and City water systems Lowest estimated present worth cost

The report recommends the drilling of fifteen new wells over the next 4 years, located in the far northeastern part of Norman, where current wells display a level of Arsenic below the 2006 MCL. 3.3.2 Limitations The report seeks – and gives – a feasible solution for the Norman area. However, the report assumes that the University of Oklahoma and City of Norman operate on the same water system. This is not the case. The University currently produces its own water. Because the University does not have the land for drilling new wells, the adoption of the CH2M Hill report’s recommendation would likely cause the University to have to purchase water from the City of Norman. When applied to the University alone, only reason two, of the four reasons given above, still holds as advantages to the chosen plan. The plan will cause the University to have to rely on Norman treated water, instead of remaining independent. The University already has its own non-potable water system4, used for irrigation and power supply. Of the five options, water purchase from Oklahoma City had the highest Net Present Cost. It follows that if OU were to have to purchase water from Norman, the cost may also be high. Though the CH2M Hill report is the most rigorous examination of the Arsenic problem in Norman to date, its recommendation does not apply strictly to the University of Oklahoma. A better option for MCL compliance may exist.

3 4

CH2M Hill Arsenic Report, p. 6-16 Source: Don Carter, OU Physical Plant Page 16 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

3.4 CE 5244 Arsenic Report Findings______________________

____

An Arsenic study was submitted by a group of graduate civil engineering students at the University of Oklahoma College of Civil Engineering and Environmental Science, CE 5244 Water and Wastewater Treatment Course. The Civil Engineering report was submitted on April 22, 2001, to the City of Norman under the title of “Arsenic Removal in Groundwater”. 3.4.1 Recommended treatment (Coagulation) Ferric Chloride Coagulation followed by micro-filtration was recommended by the Civil Engineering study to be chosen as the solution to the Arsenic problem for the University of Oklahoma. Table 3.2 is summary of the costs associated with the proposed design processes recommended by the study: Table #3.2: Estimated costs for the Ferric Chloride coagulation

Total

Project lifetime (yr)

Design Flow Rate (MGD)

20

3

Waste Disposal Capital ($) 4,000,000 500,000 4,500,000 Capital Cost ($)

Operational Cost ($/yr) 90,000

Waste Disposal ($/yr)

120,000 210,000

Note – The actual report is available at http://www.soonercity.ou.edu/ce5244/index.htm

Among the considered options for the University of Oklahoma, the Ferric Chloride Coagulation water treatment option was found to be most economical solution. However, it required high initial capital costs. 3.4.2 Drilling New Wells Drilling new deep wells, drilling shallow wells in the Canadian River alluvia, and purchasing water from the City of Norman were three other alternatives considered by the study. The options involving drilling of new wells were rejected on the basis of performed geological research. According to the study, drilling new wells is associated with high risk regarding the concentration of Arsenic in the wells. 3.4.3 Buying from Norman Purchasing water from the City of Norman was also not recommended due to high long term costs, estimated at $1.8 million per year. Moreover, the study mentions a potential water shortage for the City of Norman in the near future, which would force the city to purchase water from elsewhere. This would result in an even higher purchase price of water for the University of Oklahoma.

Page 17 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

Treatment Analysis • • • •

Overview Ion Exchange Coagulation Filtration Nanofiltration

Page 18 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

4.1 Overview The Environmental Protection Agency gives a list of possible technologies for the removal of Arsenic from water. These are detailed in the appendix. After examining these technologies for immediate applicability to the OU water supply in the next several years, Ion Exchange, Coagulation/Filtration using ferric chloride, and Nanofiltration were chosen for further study. The following treatments were not investigated in detail: 4.1.1 Activated Alumina Activated alumina was not chosen for scale up for 3 simple reasons: Longevity, operating pH, and waste disposal. Activated alumina is extremely selective to arsenate ions, but only for a short amount of time. It was decided that a process that had a long life span was necessary. The operating pH for activated alumina is very low. Although this process is very similar to ion exchange, it needs a pH adjustment to lower the pH before the process; and a pH adjustment to raise the pH after the process. This would cause an extra cost that may not be needed for other options such as ion exchange. Waste disposal for activated alumina would be very costly for the simple fact that the waste would be considered hazardous. Longevity, operating pH, and waste are the primary reasons for discounting activated alumina. 4.1.2 Lime Softening Lime softening requires large amounts of calcium hydroxide to be added to raise the pH to 10, and large amounts of sodium carbonate to lower the pH to normal standards. An operator also needs to monitor this system 24 hr/day, because if either of these chemicals get out of control, they can be toxic. Adding the high costs of calcium hydroxide, sodium carbonate, and labor it becomes very obvious that this solution is not economically feasible. 4.1.3 Electrodialysis Reversal Electrodialysis reversal involves water being pumped through a membrane. These membranes have a very large capital cost, and the waste streams are very large. 4.1.4 Reverse Osmosis Reverse osmosis was not used because of the large energy cost that would be cause from the small pores in the membrane. Nanofiltration would be more fitting for this problem because it has a much larger pore diameter. 4.1.5 Pilot Plant Technologies To date there are several Arsenic removal technologies that are still in developmental stages. Some involve surfactants, membranes (Ultrafiltration), or even a constant ion exchange recycle process (Infinite Brine Recycle). These may, in the future, turn out to be potential Arsenic removal options. Unfortunately they have not yet reached full-scale models. It was decided not to consider such options, because the Arsenic problem at the University of Oklahoma calls for a fully developed treatment process with existing designs and infrastructure. Page 19 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

4.1.6 Point of Entry (POE) and Point of Use (POU) Point of Entry (POE) and Point of Use (POU) technology is not an actual large scale treatment, but is instead a different approach to drinking water sanitation. Specifically, it is a strategy that suggests treating only the water that is really "important" instead of treating all of the water via a centralized water treatment plant. Therefore, POE/POU strategies are not considered as a possible solution to the Arsenic problem at the University of Oklahoma, by this study. However, it is recommended to explore these strategies in future studies as potential optimization techniques for the water treatment at OU. 4.1.7 Blending with Norman Water Though not a treatment, it is commonly recommended that Norman’s water simply be blended with water from OU’s existing well field, minimizing water purchased while requiring no capital investment. However, EPA regulations require all water entering a distribution system to be under the MCL. Also, due to the high concentration of arsenic in OU’s water, very little could be used, meaning that the water purchase price would not be reduced that greatly, anyway.

4.2 Ion Exchange The design for an Ion Exchange facility is broken up into many parts. The facility will be housed in a building just north of Robinson Street, between Flood and Berry, on OU’s airport campus. All designs will be based on the CH2M Hill design capacity of 2.08 million gallons per day. This value is approximately 30% higher than the maximum OU demand, but with OU growing in population and water demand annually, using this value near to the maximum flow of OU’s well system is prudent. 4.2.1 Booster Pumping In addition to the pumps already in place at the well sites, pumps must be added to make up for the loss of head caused by the Ion Exchange system. They must be high capacity. The pumps must handle 1448 gallons per minute at full capacity, and provide a 35 ft head. Providing three 750 gpm pumps in parallel will do the trick. These pumps are estimated to cost $24,000 each, with a 50% installation premium. Annually, the pumps require almost 50,000 kWh, easily the highest energy demand in the entire system. This is also the stage where oxidizing chemicals will be added in very small quantities to adjust the pH and oxidize the Arsenic (III) to Arsenic (V), which is more easily removed. Chemical costs are covered in a different section. 4.2.2 Influent Straining The treatment process calls for a small straining process, to eliminate particulate contaminants before entering the Ion Exchange columns. Any large particulate matter will contribute to clogging the column or fouling the resin, reducing efficiency in either case. The cost is small; however, the one 500 micron selfcleaning strainer costs $17,500.

Page 20 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033 4.2.3 Ion Exchange Facilities An Ion Exchange column with a diameter of seven feet was chosen using the targeting procedure shown in Figure 4.1: (Costs are detailed in a supporting spreadsheet) 900000 800000 700000 Cost ($)

600000

Total Cost

500000

Vessel Cost

400000

Resin Cost

300000 200000 100000 0 0

2

4

6

8

10

12

Diameter (ft)

Figure #4.1: Targeting Of Number Of Columns

At seven feet, the resin volume needed and total number of vessels shows a minimum cost. At this value, four vessels are needed, each with a cost of $70,000. Resin is expensive, at $110 per cubic foot of packing. A control system should be installed to keep flow uniform in the columns. Two important practical matters should be discussed. First, there is no reason that this process should not be readily expandable. Placing valves and connections in such a way that another column could be added if demand exceeds the process capacity is prudent, even though it is unlikely that the well field could support more than 2 million gallons per day. Second, with these columns in parallel, the capability exists to exchange the resin and regenerate the columns individually. Thus, the water supply to OU will not be cut off the several times a day that regeneration is required. 4.2.4 Brine Regeneration A brine of Sulfuric Acid is used to strip the resin of the adsorbed Arsenic and return the column to its default state. The brine must be stored, pumped through the column, and disposed of safely to the filter press. Pumps, a mixer and a tank must be purchased. This setup, complete with a 2,100 gallon storage tank, will most likely take up the most space in the facility.

Page 21 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033 4.2.5 Dewatering Facility The waste will be disposed of as a dry sludge. The concentrated brine must be pressed to remove liquid (which will be disposed of in the sewer system). Cost will primarily be for the filter press. 4.2.6 Chemical Storage Iron (III) Chloride and Sodium Hydroxide are used to prepare the feed for Ion Exchange (primarily in oxidizing the small amount of arsenite to arsenate). A much higher quantity of Sulfuric Acid and Sodium Hydroxide salt is required for column regeneration. Feed pumps and storage tanks are required for all three chemicals. 4.2.7 Land and Building It is estimated that 2,600 square feet of area will be needed. At a crude estimate of $100/square foot, the process will cost $260,000 to house. The University of Oklahoma already owns the land, which will be used for the facility.

Capital Cost Breakdown

Chemical Feed Facility Filter Press

21000 130000

Brine Maker

535640 402178

Brine Regen. Facility IX Facility Influent Straining

34000

Booster Pumping

20000

Allow ances

49200 103888

311181

Contengency Building

260000

Figure #4.2: Breakdown of Ion Exchange Capital Costs

Operating Costs 4.2.8 Power Power is required for all mixers, computer systems, pumps, etc. At 60,000 kWh of power per year, power costs will be near $4,200.

Page 22 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033 4.2.9 Chemicals The three chemicals discussed before, plus brine salt, must be acquired regularly. At current prices of $59 per ton of salt, $0.10 per pound of sulfuric acid, $0.13 per pound of sodium hydroxide, and $0.15 per pound of Iron (III) Chloride, the total chemical cost per year for the operation will be near $60,000, a cost that is almost entirely made up of the salt and sulfuric acid cost. If the level of sulfate in the water is significantly higher than estimated, the price to regenerate the column will also become significantly higher, as more frequent regeneration cycles will become necessary. 4.2.10 Labor Labor estimations allow for annual equipment equal to 1% of the capital cost, an operator who spends 40% of their time on the process, and a manager for whom the process is 10% of their responsibilities. Total labor costs, therefore, are close to $35,000. Maintenance and maintenance labor are given as a percentage of total capital cost. 4.2.11 Resin Replacement The resin must be replaced every five years. Even with frequent regeneration cycles, the resin will slowly lose its ability to adsorb Arsenic. The cost of all new resin, roughly $50,000, is linearized over a fiveyear period, adding a $10,000 per year cost. Breakdown of Annual Costs 10120 18671

4144

36664

Power Chemicals (not Salt) Salt Residual Disposal Labor Maintenance Resin Replacement

16500

20759

23904

Figure #4.3: Breakdown of Annual O&M Costs

Page 23 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

4.3 Coagulation/Filtration 4.3.1 Background Coagulation/Filtration is used to remove Arsenic from drinking water by adding a coagulant that precipitates out of water while causing the Arsenic to adsorb to the precipitate. The precipitate is then filtered or gravity drains from the water. The coagulants that are normally used are iron or aluminum ionic compounds. These compounds are added to the water in the liquid form. They then hydrolyze into an insoluble salt that precipitates out and is filtered from the water. The two types of metallic compounds that are used, aluminum and iron, form aluminum and iron hydroxides that are both insoluble in water. The aluminum or iron hydroxide is what the Arsenic molecule adsorbs to. The hydroxide precipitate is then filtered, and discarded in a landfill as long as the Arsenic filled sludge that is produced is dry and has a concentration of less than 5 mg/L. 4.3.2 Costs The estimated Present Worth Analysis for the coagulation filtration system was about $5.741 million. This figure includes initial capital costs to implement this process, and operating and maintenance costs over the first twenty-year period of the implementation of this process. This analysis was done assuming an average production of 2.26 MGD with a design capacity of 2.79 MGD. Today OU has an average water consumption of 1.1 MGD with a peak summer time capacity of 1.5 MGD. This is much less than what was used for the economic analysis. This was done because of expected growth in the university over the next 20 years. However if an analysis were done at the present flow rates the Present Worth Analysis of this process would be much less than that of buying water, but would have a great deal of inaccuracy from not taking into account growth of the university. Below is a graph of the total capital costs, and a table of the operating and maintenance costs for this process. These numbers are largely taken from the CH2M Hill Report for the City of Norman. Capital Costs for Coagulation/Filtration

Pumping Costs Rapid Mixing

$175,000 $560,000

Filtration

$18,000

Thickener

$1,146,600

$480,000

Dewatering Chemical Feed System Building

$400,000 $270,000

$220,000 $101,000

Figure #4.4: CF Capital Costs

Page 24 of 57

Piping, I&C, Electrical, Yard Piping Allowances Contingency 20%

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033 Table #4.1: Operating and Maintenance Costs Reason for Cost Power Cost

Annual Cost $14,000

FeCl3 Cost

$7,700

CO2 Cost Labor Cost Equipment Cost Disposal Cost

$57,500 $16,500 $33,700 $3,000

Total O&M Costs

$132,000

4.3.3 Explanation of Costs The filtration facility is the most expensive of the capital costs. This facility takes up $1.1 million of the estimated $3.4 million of the entire investment. This is a very high because of the elaborateness of the filtration facility. This high cost is not desired, but it is necessary because of the high flow rates and large of amount of precipitate that will be filtered. The next highest capital cost was the 20 % contingency that was implemented to provide an overestimate for this process. Since Coagulation/Filtration produces a fair amount of waste, there is a price of disposing. There is also a need of large pumps to pump water through the system. A vessel to mix the water and ferric chloride is also needed. This vessel contains an impeller to stir the ferric chloride and the water. The water needs a slight pH adjustment to ensure the highest efficiency possible. Carbon dioxide is added to the water for this adjustment. Costs for storage of carbon dioxide, along with ferric chloride need to be taken into account. A building of about 4,000 square feet is needed to house this facility (CH2M Hill Report). Power and operating & maintenance costs need to be taken into account too. 4.3.4 Analysis For preliminary cost estimations of total capital costs and operating & maintenance costs an EPA economic analysis from, Technologies and Costs for Removal of Arsenic from Drinking Water, was used. This graph calculates total capital cost by using the appropriate equation for the desired design flow rate in million gallons per day. The total capital cost (y) is found from the equation y = 1030810x + 1067733 (1 < x < 10 MGD), where x is the flow rate in million gallons per day (see figure 4.5 below). The average design flow rate used in the CH2M Hill report was 2.26 MGD. From this equation 2.26 MGD gives a total capital cost of $3,397,000. This is not far off from the CH2M Hill report of $3,371,000. The EPA correlation is meant to be a rough estimate, and the CH2M Hill report is supposed to be very near accuracy. The fact that these values are within a range of ± $20,000 is evidence that these estimations may be close to accuracy.

Page 25 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

Figure #4.5: EPA Preliminary Total Capital Investment Estimation Graph

According to the CH2M Hill report, the total capital cost consists of pumping facilities, mixing facilities, filtration facilities, thickener facilities, dewatering facilities, chemical storage, I & C electrical, and building facilities. The pumping facilities consist of 3 pumps with a capacity of 724 gpm. The mixing facilities have a mixing vessel of 65 ft3, and an impeller of 1.81 ft. in diameter. The filtration facilities have 6 pressure filters with a vessel diameter of 12 ft. The 4,000-ft.2 building cost close to $400,000. These items bring the total capital cost to an estimated $3,371,000. Using the O & M costs for 20 years, and the total capital costs give a Present Worth Analysis of $5.7 million.

4.4 Nanofiltration Due to a Present Worth larger than that of the other treatment options and due to an uncertainty range of almost $13,200,000 present in the economic evaluations, the Nanofiltration treatment is not recommended as a treatment option to the University of Oklahoma. As a result of profitability evaluation of a Nanofiltration facility the Present Worth for the best and the worst case scenario were calculated to be $4,100,000 and $17,300,000, respectively. During the economic evaluation it was determined that the Present Worth is largely a function of membrane cost. The Present Worth of an Ion Exchange treatment option was used with the membrane, pumping and utility costs specific to the NF processing facility. Costs specific to Ion Exchange were not considered. The manufacturer quoted membrane costs. Pumping and utility costs were obtained from a Pro-II simulation and literature correlations. Page 26 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033 4.4.1 Technology Background Figure #4.6 illustrates a process flow diagram of a Nanofiltration water treatment facility:

Figure #4.6: Nanofiltration Water Treatment Facility PFD. 5

Nanofiltration (NF) is a membrane separation method that has proven to achieve Arsenic rejection efficiencies in the range of 80-90%. The membranes employed by this method use the principle of Reverse Osmosis to filter out Arsenic and are operated in Cross Flow mode (see Appendix A). Larger membrane pores and lower utility requirements set Nanofiltration apart from Reverse Osmosis. Low feed recovery rates and high capital cost are a major concern in NF design. A Nanofiltration facility consists of the membranes themselves, pumps, piping and the controls with instrumentation needed for operation. Nanofiltration presents an appealing option for the University of Oklahoma because at the given conditions it can lower the Arsenic content in water from an average of 30ppb to below 1.0 ppb. Nanofiltration would allow OU to meet the legal Arsenic standards, to be implemented in 2006, by producing water with constant quality and also by removing other impurities. The NF process is simple, fully automated and requires little monitoring. The only chemicals requiring storage for the NF treatment facility are those needed for membrane cleaning, unless it is determined that pretreatment is needed. Another appealing quality of NF treatment is that no onsite residual processing is required. The concentrate would be sent directly to the City of Norman water treatment plant, which has processes to treat Arsenic. 5

http://www.membranes.com/docs/trc/flowcon.pdf Page 27 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033 4.4.2 Economic Analysis Although, the annual operating and maintenance costs of Nanofiltration are generally minimal, this separation method is plagued with very high capital costs. A major part of the capital cost is the equipment associated with Nanofiltration, with the membranes themselves being the most expensive piece of equipment (approximately 15-40% of the capital cost). Therefore, it is assumed that the major differences in Present Value of Nanofiltration and that of other methods come from membrane, pumping and electricity costs. In order to calculate the Present Value for an NF facility, the economic evaluation of an Ion Exchange facility was used with appropriate changes made to three previously mentioned costs. 4.4.3 Membrane Cost Initially, an extensive membrane price research was performed and the membrane with the highest permeate flowrate per dollar of price ratio was isolated from the pool of candidates. A Hydranautics model ESNA1 membrane turned out to have the largest permeate to cost ratio of 13 gal/$. Since low feed recovery is a big concern in Nanofiltration design for scarce water regions (see Appendix A), the ESNA 1 presented the optimal choice of membrane for the University of Oklahoma. According to the manufacturer, a NF facility consisting of ESNA 1 membranes would achieve 95% Arsenic rejection with a 75-80% feed recovery. The total number of membrane units needed for an average production permeate flow of 2.6 MGD was calculated to be 205 units. The price quoted by the manufacturer for the total number of membranes required is approximately $172,000. The lifetime of the membranes was taken to be 4 years, as quoted from the manufacturer. Osmotic and operating pressures of 1.37-4psi and 76psi, respectively, were calculated for the ESNA 1 membrane. 4.4.4 Pumping Costs The work required for a pump in order to raise the feed to 76 psi was estimated by running a Pro-II simulation, assuming 70% pump efficiency. The work was determined to be 142kW. Electricity price used for utility costs was 7 cents per kW hour, which is the same price that was used to evaluate other treatment options. Pump cost for a centrifugal pump was estimated from Peters and Timmerhaus to be $20,000. 4.4.5 Facility Costs All costs specific to the Ion Exchange process were removed from the Present Value evaluation, except for the chemical and chemical storage costs. These costs were assumed to be the same as what was determined for Ion Exchange by CH2M Hill. This is a good assumption because Nanofiltration would require less expensive chemicals, in smaller amounts than what would be needed for Ion Exchange. Therefore, the NF Present Value would not be in danger of underestimation due to this assumption.

Page 28 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033 Figure 4.7 presents the capital cost break down for the NF facility using the Hydranautics ESNA 1 membrane: Capital Cost Breakdown For NF Treatment Facility (Best Case Scenario) $104,000, 10%

$250,000, 23%

Pumping Membrane&Installation Chemical Feed Facility

$241,000, 23%

Building Contingency Allow ances

$233,000, 22%

$50,000, 5% $178,000, 17%

Figure #4.7: Capital Cost Breakdown for NF Treatment Facility for the Best Case Scenario.

In order to observe the effect of the membrane cost on the Present Value, the procedure above was repeated with the most expensive membrane available. An AMI model M-N2521A membrane turned out to have the lowest permeate to cost ratio of 1.33 gal/$. A total of 9040 units would be need with an overall cost of $1,695,000. The life time of the membranes was taken to be 4 years. Figure #4.8 presents the capital cost break down for the NF facility using the AMI M-N2521A membrane. From comparison of Figures #4.7 and #4.8, it is apparent that the membrane cost is always one of the major components of the capital cost.

Page 29 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

Capital Cost Breakdown For NF Treatment Facility (Wost Case Scenario) $2,373,000, 57% Pumping Membrane&Installation

$104,000, 3%

Chemical Feed Facility

$250,000, 6%

Building Contingency

$233,000, 6%

Allow ances

$50,000, 1% $1,137,000, 27%

Figure #4.8: Capital Cost Breakdown for NF Treatment Facility for The Worst Case Scenario.

4.4.6 Conclusion It is apparent from the best and the worst case scenarios for the Present Worth that the membrane cost makes an enormous difference on the profitability of the NF treatment process. This is largely due to the capital cost of the membranes and their annualized replacement. Table 4.2 is a summary of the Present Worth results for the best and worst case scenario of the Nanofiltration facility with a project lifetime of 20 years: Table #4.2 Nanofiltration Facility Present Worth Results for a 20 yr Project Lifetime

Membrane Type Hydranautics model ESNA1 AMI model M-N2521A

Present Value, $ 4,100,000 17,300,000

Uncertainty, $ 13,200,00

Due to the uncertainty range of $13,200,000, it is not recommended to choose Nanofiltration as the treatment option for the University of Oklahoma. A chlorine-based pretreatment will most likely be required in order to convert Arsenite to Arsenate. Chlorine destroys many of the NF membranes available, including the best-case scenario membrane. Therefore, even though a more detailed analysis might reveal a Present Worth that is slightly lower than what was estimated in this study, it is highly unlikely. In fact it is likely to be much higher than the best-case scenario. This treatment option should be investigated in as much detail as possible, if its utilization is to be considered in the future.

Page 30 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

4.5 Comparison of Previous Conclusions Compiling the findings of both the CH2M-Hill report and the CE 5244 Group, a comparison can be made about how the Net Present Cost compares to the finding in this report. Attention should be taken to the fact that this report’s findings were based solely on using wells already owned and operated by OU.

Previous Conclusions $16,000,000

Net Present Cost

$14,000,000 $12,000,000 $10,000,000 $8,000,000 $6,000,000 $4,000,000 $2,000,000 $0 WP

IX CH2M Hill

C/F CE Group

NF

Our Group

Figure #4.9: NPC Comparison of Previous Solutions (at different plant capacities)

As is apparent from Figure 4.9, investigating the OU situation independently from the City of Norman may result in significant savings. Two of the three proposed treating processes, Ion Exchange and Coagulation Micro-filtration, are more economically feasible than simply becoming a customer of the City of Norman. According to the preliminary findings, choosing to build an Ion Exchange treatment facility is the cheapest option available to the University. Based on this analysis, the Ion Exchange design from section 4.2 will be further specified and applied to Norman. It is important to note that the estimates performed by all three groups were made at different plant capacities: Our Group=1.8MGD, CH2M Hill=2MGD, CE Group=3MGD. However, the graph reveals a very important point – at the current OU maximum well capacity (used by Our Group), Ion Exchange is the most economically favorable treatment option. It is also important to note that even though the capacities were different, for different groups, the NPC values were relatively close. In fact, if the estimates for other groups were to be rescaled to the plant capacity used in this study, the NPC values would agree much more favorably.

Page 31 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

Detailed Ion Exchange Model • • • •

P&ID Specifications Variations with Capacity Safety

Page 32 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

5.1 P&ID

Figure 5.1: P&ID of Ion Exchange Treatment Facility

5.1.1

Regular Operation

Regular operation of the Ion Exchange plant is given by the solid blue lines. Under regular operation, water will travel through a sand filter, then mixed with sodium hipochlorite to oxidize the arsenite to arsenate, making it possible to remove the compound in the ion exchange column. The water then runs through the primary columns. Sampling ports are set up to measure concentration of sulfate and arsenic, so that the column may be ‘timed’, and the system assured of no breakthrough of the extra “safety column”. 5.1.2 Regeneration To remove arsenic and sulfate from the IX resin, replacing it with the original chloride ion, a column is shut off, then concentrated NaCl brine is run through the column backwards. This allows the resin to shift, allowing the resin to age evenly. The brine, with sulfate and arsenic in solution, is then mixed with Ferric Chloride, which causes the arsenate to precipitate out. The waste sludge is then dried and sent to a landfill. Periodically, sulfuric acid is introduced to act as a cleaning agent for the resin. Regeneration is shown by the dashed orange lines.

Page 33 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

5.1.3 Controls Only very simple controls are needed. First, to ensure that NaOH is not wasted, a pH meter determines how much the pH is adjusted. Second, the level of the brine settling tank is monitored. Green lines represent the control system.

5.2 Variations with Capacity The capital investment required for an ion exchange plant has both fixed and variable costs. For the purposes of a model that is explained later, all variable costs are assumed to vary linearly with respect to the capacity of the plant. Of the major costs of the facility, only column capacity (and thus the associated allowances and contingency) is particularly sensitive to changes in the flow rate. A zero capacity portion of the plant (everything except the ion exchange columns) is still need, and regeneration and chemical feed facilities must still exist. The cost of this “capacity-less” portion of the plant is 1.1 million dollars. The plant discussed in section 4.2 had a capacity of 2.08 MGD and a Fixed Capital investment of 2.05 million dollars. This gives the relationship between capital investment and capacity seen in figure 5.2. The analysis shows that the capital investment increases $450,000 per MGD of capacity.

Capital Investment ($MM)

Capital Investment and capacity 2.5 2 1.5 y = 0.4567x + 1.1

1 0.5 0 0

0.5

1

1.5

Capacity (MGD) Figure 5.2: Capital Investment vs. Capacity

Page 34 of 57

2

2.5

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033 A similar analysis is done with the operating cost. In this case, the labor and maintenance costs for a zerocapacity plant are the “fixed” values, while the rest (power for pumping, chemicals used, etc) depend directly on the capacity. Therefore, at zero capacity, the operating cost is near $35,000 per year, while at 2.08 MGD, the operating cost is 130,000. Figure 5.3 shows the relationship between operating cost and capacity. The operating cost is increased $43,000 for every MGD of capacity.

Operating Cost ($)

Operating Cost and Capacity 140000 120000 100000 80000 60000 40000 20000 0

y = 45673x + 35000

0

0.5

1

1.5

2

2.5

Capacity (MGD) Figure 5.3: Capital Investment vs. Capacity

5.3 Safety 5.3.1 Arsenic Spiking The Ion Exchange process must be over-designed to ensure that all adsorption sites in the column are never filled. The exchange media is more selective towards sulfate in the water than it is towards Arsenic compounds. Thus, because sulfate is more plentiful, more sites are actually used to adsorb sulfate than Arsenic. As the sites begin to fill, the sulfate begins to replace Arsenic on the exchange media. This actually causes the effluent stream to “peak” in Arsenic, as all the adsorbed Arsenic is pushed from the media by sulfate at once. The Arsenic level, after mixing in OU’s water tower, will most likely return to pre-treatment levels at the Point of Use. Figure #5.4 is taken from sulfate breakthrough on a laboratory scale.

Page 35 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

Figure #5.4: Pilot experiment on sulfate breakthrough as related to Arsenic peaking. (http://www.hdrinc.com/engineering/topics/ionexchange.htm)

All care must be taken to ensure that regeneration cycles take place before Arsenic breaks through, rendering the entire process worthless. It is therefore necessary to sacrifice cost in the form of chemical cost by reducing total bed volumes between regeneration in order to ensure no breakthrough. 5.3.2 Chemical Safety and Storage

Iron Chloride: Iron Chloride is a light green to yellow crystalline powder at normal conditions. It does

not have an odor and is non-flammable. Iron Chloride is severe irritant and is corrosive upon physical contact. Chronic effects of exposure are iron poisoning, liver damage, cytogenic and reproductive effects. Iron Chloride should be stored in a tightly sealed container away from incompatible substances. The container should be kept in a cool, dry, ventilated area.

Sodium Hydroxide: Sodium Hydroxide has an appearance of white, deliquescent pellets or flakes under normal conditions. It is an odorless, nonflammable, poisonous substance. Sodium Hydroxide is severe irritant and is corrosive upon physical contact. Its destructive effects upon tissue include blindness and severe burns. The Permissible Exposure Limit is 2mg/m3. Scarring of tissue and death are possible upon ingestion. Protective coating should be worn and vent hoods should be used when handling Sodium Hydroxide. Sodium Hydroxide should be stored in a tightly sealed container away from sources of heat, moisture and incompatibles. Sulfuric Acid: Sulfuric Acid is a white powder, with no strong odor. It is commonly dissolved in water, in which case it appears as a clear liquid. Burns and tissue necrosis are possible upon contact. Inhalation may be fatal. Prolonged or repeated inhalation may cause kidney and lung damage. Protective goggles, gloves and clothing should be worn when handling Sulfuric acid. Sulfuric acid should be stored in a cool, dry, well-ventilated area away from incompatible substances and combustible materials. Material Safety Data Sheets are found in Appendix C. Page 36 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033 Any arsenic treatment plant that has on-site regeneration facilities (i.e.: Ion Exchange or Activated Alumina) is classified as a large quantity hazardous waste generator (LQG) if waste brine generated exceeds 1000 kg/month. The handling of these arsenic laden residuals must comply with the provisions of Resource Conservation and Recovery Act (RCRA). Arsenic is considered to be at toxic levels if it is found in liquid, sludge and solids with a solid concentration of more than 0.5% (CFR, Part 261.24). An LQG, such as in our case, accumulating hazardous waste on site must: -

Comply with the preparedness and prevention procedures (CFR, part 265, Subpart C) Develop and maintain a contingency plan on site Comply with personnel training requirements Before shipping hazardous waste to an RCRA facility, an LQG must comply with several pretreatment requirements: EPA ID number, preparing a Uniform Hazardous Waste Manifest Comply with Department of Transportation (DOT) requirements

5.3.3 Disposal Options for Arsenic Contaminated Residuals Waste generated will consist of solid, liquid and sludge: -

Allowable discharge of arsenic contaminated waste liquid will be affected by the ability of the receiving stream to assimilate the arsenic without exceeding the arsenic standard of the receiving water Disposal of liquid waste to a water treatment facility, generated sludge has to meet the following requirements: Land disposal 73 mg/kg Land application (clean sludge) 41 mg/kg

An option to discharge the waste brine into a receiving stream is highly unlikely due to high concentration of arsenic and sulfates. Disposal into a sanitary sewer is an option considering it meets the TBLL requirements (Technically Based Local Limits). Treatment of the brine by adding Ferric Chloride will greatly reduce arsenic concentration. Precipitate can be dried and taken to municipal waste since limited data shows that it will pass TCLP (Toxicity Characteristic Leaching Procedure) test. Then with sufficient arsenic removed from the waste brine, disposal into a sanitary sewer is possible considering the high salt content is acceptable locally.

Page 37 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

Planning Model, Results and Recommendations • • • • •

Planning Model Model Results Financial Analysis Risk Analysis Recommendations

Page 38 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

6.1 Planning Model 6.1.1 Background The preliminary calculations for the Water Purchase option were based on the current $1.14/1000gal price of water and the average daily demand of 1.1 MGD. The net present cost for 20 years was calculated to be $5.6 million, based on these figures. In order to make an economic comparison, the net present cost for the Ion Exchange option was calculated making similar assumptions that demand, plant capacity and operating costs remain constant. The net present cost of Ion Exchange was estimated to be $3.6 million. Although the preliminary estimates show the Ion Exchange option to be more economically attractive, the previously listed assumptions do not represent well what occurs in the real world. The water demand for the University of Oklahoma varies from month to month, displaying patterns of increased demand in the hot months of the school year and decreased demand in the winter and midsummer months. Additionally, the water demand displays yearly patterns of growth due to increasing student body of the University and campus expansions. Other important parameters that introduce a significant degree of uncertainty to the problem are variation of water prices, operating costs for the IX facility, unforeseen changes in the well field, inflation, depreciation and the possibility of additions to the existing plant. It quickly becomes obvious that the magnitude and complexity of an economic comparison, more accurate than what was done in the preliminary calculations, quickly gets out of hand and is not feasible to perform without the use of technological tools. The need for greater accuracy calls for a computer based simulation that uses a mathematical model in order to describe the Arsenic situation at the University of Oklahoma. Such high performance numerical simulations are becoming widely used for financial planning and are already considered to be mainstream engineering tools in the industry. 6.1.2 Planning Model Description A mathematical model was developed in order to compare the net present cost of building an Ion Exchange facility to the purchase of water from the City of Norman option. The optimal solution for the Arsenic problem is determined by the planning model based on the economic comparison results. The model was solved using a programming interface called Gams® (www.gams.com), which is a high-level modeling system for mathematical programming problems. The main objective of the mathematical model is to describe mathematically the Arsenic situation at the University of Oklahoma and find the most economically attractive solution by minimizing the expected cost, for a project life time of 20 years: Total Expected Cost = ∑ p s C s where C is scenario cost, p is probability, and s is scenario. s

Two types of models were solved : a deterministic model based on one scenario with constant parameters and a stochastic model based on many scenarios with varying parameters. Aside from minimizing the total expected cost the mathematical model must: a) Choose between building a Treatment Plant or Purchase b) Meet the Water Demand Page 39 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033 c) d) e) f)

Decide When/How Much To Build Expand Capacity as Needed Buy Wholesale on Emergency Borrow/Repay Money

6.1.3 Detailed Explanation a) The model has the capability of choosing between building a treatment facility or purchasing water from the City of Norman. This means that the model can choose to not ever build a plant and purchase water all of the time, or vice versa. The optimum choice between the two options is made based on the option that would give the least total cost for the project lifetime. The model also explores the possibilities of combining the two options (see part b). b) One of the constraints used in the model is that it has to satisfy the projected water demand by any means necessary. This means that the model explores the possibilities of building plants of different capacities, and then either add-on to the existing plant to increase water production or purchase water from the City of Norman, all this while minimizing the total cost of the project. c) If the model makes a decision to build a treatment facility, it has the flexibility of choosing when to build and what the capacity of the plant should be. d) The model has the ability to upgrade the existing facility in order to increase the water production capacity. The upgrade option is available to the model’s discretion in every year, after year one. e) The price of the water purchased depends on the amount that it purchases. A wholesale or a higher, emergency-based price is charged for water purchase, based on the amount of water that the model decides is necessary to purchase. f) Finally, the mathematical model has the finance equations necessary to consider borrowing money from a bank in case the specified budget is too small to cover the expenditures. The model calculates the optimum time to borrow money and the best way to pay off the loan within the project lifetime. Time value of money, as well as price inflation, are taken into account. For a complete set of equations used in the mathematical model and a copy of the GAMS program, see Appendix F. 6.1.4 Parameters A project life time of 20 years was used in the model, with varying monthly water demand. The water demand was projected into the future based on the historical consumption data for the University of Oklahoma and based on 25% growth, over the project lifetime, estimated by the OU Physical plant. Correlations for the capital investment and operating costs for the Ion Exchange treatment plant were developed based on the preliminary calculations performed in the section 4.2. Both capital investment and operating cost correlations depended on a fixed, “must pay” fee and a plant capacity based fee. Table #6.1 summarizes the plant costs used in the Mathematical Model: Table #6.1 Plant Costs used in the Mathematical Model.

CI ($) 563,900.00

CI ($/1000gal) 635.70

OP ($/yr) 38,000.00

Page 40 of 57

OP ($/(yr 1000gal)) 3.72

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033 The maximum well capacity was determined to be 1.8MGD, based on the existing well data provided by the OU Physical Plant. A depreciation rate of 6% and an interest rate of 9% were assumed for the financial purposes of the model. The $1.14/1000gal water price, currently charged by the City of Norman, was used as the wholesale price that would be charged if the amount of water purchased by the model averaged to be over 1,000 gallons per month. Conversely, if the model purchased water sporadically, it charges an Emergency-based price of $3.00/1000 gal. A 2% inflation rate was applied to the water price throughout the project lifetime. A $500,000 per year budget was specified for the model. 6.1.5 OU Water Consumption Projection According to the OU physical plant, the demand for water on the OU campus is expected to increase 25% over the next 20 years. The horizon for this increase is shown in Figure 6.1:

Average Daily Consumption (1000 gal

W a te r C o n s u m p tio n P r o je c tio n 1800 1700 1600 1500

Average C ons um ption

1400

Augus t (P eak) C ons um ption

1300 1200 1100 1000 0

5

10

15

20

Y e ar Figure 6.1: Water Consumption Projection

The error bars in figure 6.1 represent the standard deviation, as we predict that the water demand could vary as much as 5% in either direction. This consumption projection is used later in the report, in order to determine the future water demands upon the Westheimer airport field. It is important to note that the black line gives the horizon for average daily water consumption over a year, but in the peak month of August, the consumption is about 190,000 gallons per day higher.

Page 41 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033 6.1.6 Constraints Physical Constraints: 9 Maximum Capacity - The maximum water production capacity of the treatment plant must be less than or equal to the maximum production capacity of the existing OU wells. 9 Total Capacity – Total capacity for a given year must equal to the sum of the capacities of the existing plant and the previously built additions. 9 Meeting Demand – Total capacity for a given year must be greater than or equal to the demand in any month of the year. Financial Constraints: 9 Water price - Wholesale price must be charged if the amount of water purchased by the model averaged to be over 1,000 gallons per month. Otherwise, the Emergency-based price must be applied. 9 Budget – The total cost for any given year cannot exceed the budget for that year. 9 Debt Payment – The minimum amount of money repaid to the bank must be greater than or equal to twice the interested accumulated from last year’s debt 9 Repay Loan – The debt in the final year of the project must equal to zero.

6.2 Model Results 6.2.1 Deterministic Model Results: As a result of the deterministic simulation, the mathematical model produced an optimal solution for the Arsenic problem at the University of Oklahoma. The model decided to make a one time capital investment, in order to build a large capacity treatment facility. The model determined that building a plant in year one of the project lifetime instead of choosing to purchase water would save the most amount of money for the University of Oklahoma. The detailed results produced by the model are described in Table #6.2: Table #6.2 Deterministic model results.

Total Cost($)

Plant Capacity(MGD)

Area(ft2)

Labor($/yr)

2,600,000

1.6

2,600

20,000

# of Columns 4

Column Dia (ft) 6

Figure #6.2 displays the water purchased by the model during the peak times of the projected demand. It can be seen from the figure that if a 1.6 MGD treatment facility is built in the first year of the project, the University of Oklahoma will not begin purchasing water until approximately the second half of the project lifetime.

Page 42 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

.6 .4

Purchased Water

.2

20

19

18

17

15

14

11

.0

8

Purchased Water (MGD)

Purchased Water Chart

Project Lifetime (yr) Figure #6.2: Purchased Water over lifetime

It is also important to note that the model chose not to build any add-ons to the treatment facility. Purchasing water on Emergency basis turned out to be more economical. In order to finance the project the model decided to take out a loan and repay it in the first five years of the project lifetime. Figure #6.3 displays the financial planning chart produced by the deterministic model. 140 000 0 120 000 0 100 000 0 80 000 0 60 000 0 40 000 0 20 000 0 0 -20 000 0

1

2

3

4

5

6

7

-40 000 0

Figure #6.3: Loan repayment

The loan was repaid in equal amounts of approximately $300,000 per year, for five years and the remaining sum was repaid in year seven of the project. It is important to note that this result assumes a budget of near $400,000 per year. This is not meant to represent the way that OU will pay for the capital investment of the plant, but instead gives one example of how it is possible, and the low price compared to the water purchase cost of at least $450,000 a year.

Page 43 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

6.3 Financial Analysis The results obtained from the mathematical model are shown in Figures 6.4-6.5 in terms of the cost per year under both water purchase and treatment conditions. Figure 6.4, assumes that the capital cost of an Ion Exchange facility is supplied by a loan taken out in year one, and paid off over the next ten years, in equal payments of $350,000 per year. When compared to water purchase at $1.14 per 1000 gallons from Norman, the difference in costs of the two options is apparent in year one of the project lifetime. Initially the difference in costs is on the order of $150,000, but as time goes on the water purchase option becomes increasingly more and more expensive, with yearly costs growing to as much as $800,000 in the second half of the lifetime. The cost of the Ion Exchange treatment option remains constant at $350,000 per year until the loan is paid off. When no more payments are made, the cost of treatment drops to roughly $100,000 per year, in the second half of the project lifetime.

Yearly Cost With Loan $1,000,000

Cost($/yr)

$800,000 $600,000 $400,000 $200,000 $0 1

2

3

4

5

6

7

8

9 10 11 12 13 14 15 16 17 18 19 20

Project Lifetime(yr) Water Purchase

Ion Exchange

Figure #6.4: Yearly cost with a loan

If sufficient capital is made available so the University can build the facility with no loan or bond sale needed, the savings after year one would be substantially greater. Figure 6.5 is a graph analogous to Figure 6.4, specific to this situation.

Page 44 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

Yearly Cost Without Loan $1,800,000 $1,600,000

Cost ($/yr)

$1,400,000 $1,200,000 $1,000,000 $800,000 $600,000 $400,000 $200,000 $1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

Project Lifetime(yr) Water Purchase

Ion Exchange

Figure #6.5: Yearly Cost without A Loan

Therefore, substantial savings are obvious right away. Two price levels have been proposed for water sold by the City of Norman. The first is the current commercial price level, $1.14 per 1000 gallons, upon which most of the previous calculations were based. However it is possible that the University and the City of Norman will agree upon a lower price. The model was given the water price of $0.85 per 1000 gallons as a low margin estimate. Figure 6.6 shows that the savings are substantial from year one even at this low price level.

900000 800000 700000 600000 500000 400000 300000 200000 100000 0 Project Lifetime (yr)

Figure #6.6: Savings At Both Price Levels Page 45 of 57

19

17

15

13

11

9

7

5

3

$1.14/1000 gallons $0.85/1000 gallons

1

Savings ($)

Savings per year Current Dollars

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

As can be seen from Figure 6, the ion exchange (IX) plant option results in significant savings for every year of the project lifetime, when compared to the Water Purchase option. For approximately the first half of the project lifetime the savings start out low, $50,000 for the $0.85 price and $150,000 for the $1.14 price, and increase steadily until the money borrowed for the capital investment is paid off. This occurs approximately midway through the project lifetime, at which point the savings roughly double and stay constant until the last year. Figure 6.7 is a Net Present Cost comparison of the Ion Exchange treatment and the water purchase options. 9 Net Present Cost($MM)

8 7 6 5

Ion Exchange Plant

4

Water Purchase

3 2 1 $1.14

$0.85

$0.45

Price of water ($/1000 gal) Figure #6.7: Comparison Of NPC For Ion Exchange And Water Purchase At Different Price Levels

Assuming that the price of water charged by the City of Norman remains constant at the current value of $1.14/1000 gal, for the next 20 years, the Ion Exchange facility option will result in approximately $5,000,000 of Net Present Cost savings. It is also apparent from Figure 6.7 that the minimum allowable price of water for the WP option to match the economic attractiveness of the IX option is an unrealistic price of $0.45/1000 gals. A price so low could not be offered by the City of Norman, since it costs the city roughly $0.60/1000 gals to produce it. Any price charged by the city of Norman that is above this number, results in loss of money if the WP option is chosen. It should also be taken into consideration that the price of water is not expected to stay constant or drop. Conversely, it is expected to increase, because of Norman’s overwhelming growth that is projected to result in water shortage by year 2010.

Page 46 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

6.4 Risk Analysis The mathematical model is also able to show how the results respond to uncertainty in the values that were used to estimate Net Present Cost. For example, the price of “emergency”based water purchase was varied from the wholesale commercial value of $1.14 per 1000 gallons to $7.00, which represents a value higher than is realistically possible. Figure 6.8 shows how the Net Present Cost of building the plant reacts to such large variations.

Present Cost ($MM)

3.3 3.25 3.2 3.15 3.1 3.05 3 0

1

2

3

4

5

Price of high-rate water ($/1000 gal)

6

7 NPC Linear (NPC)

Figure #6.8: Sensitivity of NPC to water cost

The price of emergency water, as seen, has very little effect on the overall NPC of the plant. Even if the water price were to be doubled, the cost of the plant would only increase by approximately 3%. The well field’s maximum output capability could also vary. Currently, the field’s maximum sustainable output is 1.8 MGD. Figure 6.9 shows how the mathematical model responds to changes in the maximum output, which is essentially an indicator of how NPC changes if a well is added or shut down. The line planes at two points, one at low production, where it becomes more cost effective to simply purchase all the water (at slightly more than .5 MGD) and the point where even increasing the output does not cause the model to want to build a bigger plant (approximately 1600 MGD).

Page 47 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

Net Present Cost ($million)

10 8 6 NPC 4 2 0 0

500

1000

1500

2000

2500

Maximum Field Capacity (1000 gal/day) Figure #6.9: Maximum Field Capacity vs. NPC

Operating cost per unit capacity and capital investment per unit capacity were some of the parameters estimated by this study. It is conceivable that, due to rising chemical prices or unforeseen complications, uncertainty could exist in these numbers. Figure 6.10 plots Operating Cost vs. Capital Investment and shows the ‘threshold’ between Water Purchase and IX Treatment options. This is the point at which raising the values causes the plant to become economically infeasible, causing the model to recommend Water Purchase as a solution. Areas below and to the left of each black line are areas where Treatment is favorable at a given price level, while areas to the right and above the black line are those at which Water Purchase is the best choice. As Figure 6.10 shows, the costs estimated for the Ion Exchange treatment facility (marked by the blue diamond) are well within the area of treatment, at either $0.85 per 1000 gallons or $1.14 per 1000 gallons. In fact, one value would have to be off by 400% for treatment not to be economically favorable when compared to water purchase. Therefore, it is safe to conclude that the Ion Exchange treatment will be less expensive than purchasing water from the city of Norman, even if some uncertainty in the estimates is encountered.

Page 48 of 57

Operating Cost $/1000 gallons capacity

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033 35 30

Area of Water Purchase

25 20 15

$1.14

10

Area of Treatment $0.85

5 0 0

1000

2000

3000

4000

5000

Capital Investment $/1000 gal capacity Figure #6.10: Sensitivity of NPC to capacity unit cost

The probability distribution displayed in Figure 6.11 quantifies the amount of risk involved in building the Ion Exchange treatment facility. This distribution was obtained by running the stochastic model. The average value of the produced Net Present Cost distribution occurs at approximately $3,150,000, which compares favorably with the preliminary estimate of the study and the results previously generated by the model. The shape of the probability distribution is similar to that of the normal distribution, with the exception of its tails. The probability distribution has a short left tail, because there is less probability of the project being cheaper than estimated. This makes sense, since there is a minimum capital investment associated with building a treatment facility. The right tail of the distribution quantifies the probability of the Net Present Cost being higher than what was estimated by the model. The fact that the right tail does not extend far from the mean value signifies that the risk associated with the Net Present Cost estimate is minimal. In fact, there the probability of the Net Present Cost exceeding $3,600,000 is estimated to be less than 5%!

Page 49 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

Net Present Cost Probability Distribution 25 Frequency

20 15 10 5

$2

$2

,7 00

,0 ,8 00 00 $2 ,00 0 ,9 00 $3 ,00 0 ,0 00 $3 ,00 0 ,1 00 $3 ,00 0 ,2 00 $3 ,00 0 ,3 00 $3 ,00 0 ,4 00 $3 ,00 0 ,5 00 $3 ,0 0 0 ,6 00 ,0 00 M or e

0

Net Present Cost

Figure #6.11: NPC Probability Distribution

6.5 Recommendations As a result of the planning model, our group recommends that OU construct an ion exchange plant with a capacity of 1.6 MGD at the Westheimer airport facility. No additions to the capacity of the plant are needed. The plant’s capital investment is expected to be less than $2,000,000, with a total net present cost of near $3,000,000. While this is a substantial investment, the yearly cost of buying water from Norman starts at $500,000 and increases, with an expected 20-year net present cost of $8,100,000. There is little doubt that treatment is the best solution for the University of Oklahoma. However, before implementation, we recommend further study of the ion exchange process itself, in order to more accurately design the treatment facility.

Page 50 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

7.1 References 7.1.1 Online References These References can all be found in appendix G. Reference.1.html Key parameters for Arsenic removal, HDR, inc. Reference.2.ppt Civil Engineering Arsenic Project Reference.3.html EPA Regulations Reference.4.html MIT Synopsis of Arsenic Removal Processes Reference.5.html Online pH monitor Reference.6.html Effects of Arsenic Consumption by Humans Reference.7.doc Document exploring the effects of Arsenic treatment on the environment 7.1.2 Other References Bryan Mitchell, City of Norman Don Carter, OU Physical Plant CH2M Hill: Joe Chwirka, Murray Flemming Trenton Brown, OU Environmental Health and Safety CE: Dr. Sabatini, Daniel Gallo (graduate asst.)

Page 51 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033

7.2 Appendix 1 Mathematical Model Equations6:

TotalCost = ∑ p s C s s

(

)

C s = ∑ C yr =∑ CI yr + OPyr + Pr ice yr * WPyr * (1 + i ) yr −1 − Borrowed yr + Repaid yr * df yr yr

yr

CI yr = a * z yr + b * Cap yr yr

OPyr = α F ∑ zξ + β ∑ Q yr ,mo , s ξ =1

mo

Demand yr ,mo , s = Q yr ,mo , s + WPyr ,mo , s ztot yr = ∑ zξ if ξ ≤ yr ξ

Finance Equations:

Pr ice yr , s = WholeSale Pr ice * y yr , s + Emergency Pr ice * (1 − y yr , s )

Debt yr , s = (1 + i ) * Debt yr −1, s + Borrowed yr , s − Repaid yr , s Constraints: Physical Constraints:

CapTot yr = ∑ (Capξ + CapAdd ξ ) if ξ ≤ yr ξ

CapTot yr ≥ Q yr ,mo , s Cap yr ≤ MaxCap * z yr Finance Constraints:

∑WP

yr , mo , s

− 1000 * y yr , s ≥ 0

mo

C yr , s ≤ Budget Debt 20, s = 0 Repaid yr ,s ≥ 2 * i * Debt yr −1, s

6

Subscripts: s-scenario; yr-year; ζ-year alias; mo-month. Page 52 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033 GAMS Model $Title STOCHASTIC MODEL, BUDGET FIXED FOR S, PRICE FIXED FOR S. $offlisting $offsymxref $offsymlist $offuelxref $offuellist options iterlim=1000000; *DECLARING VARIABLES=================================== Sets yr project lifetime /1*20/ mo month of each year /1*12/ s scenario number /1*1/ * zeta number of facilities in that year? /0*10/; *ENTERING DATA============================================= Scalar *CI=(a+b(capacity))*z a capital cost of facility addon /1163900/ b slope_extra price for capacity /428/ * b slope_extra price for capacity /635.7/ maxcap maximum capacity /1800/ * beta OpCost varying costs fudge /3.72/ alpha OpCost fixed costs fudge /3800/ df depreciation factor /.06/ * price water price each year(CONST) /34.2/ budget budget /500000/ i interest rate /0.09/ aa capital investment of added /300000/ * bb capital investment of added /500/ inflate inflation rate /.02/; Table demand(yr,mo,s) projected monthly water demand for 1.1*1 2.1*1 3.1*1 4.1*1 6.1*1 7.1*1 8.1*1 9.1*1 10.1*1 12.1*1 1 1146 1093 1242 1348 1200 1332 1408 1261 1256 1067 2 1056 1102 1236 1336 1199 1336 1402 1418 1170 973 3 1160 1084 1299 1406 1320 1355 1396 1440 1276 1112 4 1207 1144 1202 1320 1222 1451 1539 1345 1275 1029 5 1297 1148 1318 1314 1153 1249 1434 1293 1304 1129 6 1221 1242 1319 1525 1355 1324 1525 1426 1398 1121

Page 53 of 57

each year 5.1*1 11.1*1 1322 1175 1338 1129 1358 1171 1457 1218 1324 1251 1436 1282

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033 7 1369 1230 8 1391 1216 9 1317 1213 10 1278 1148 11 1343 1249 12 1384 1185 13 1434 1244 14 1429 1397 15 1345 1229 16 1326 1310 17 1446 1220 18 1521 1168 19 1538 1262 20 1461 1327;

1360 1463

1140 1634

1257 1334

1376 1288

1494 1239

1304 1419

1105 1610

1348 1462

1437 1446

1413 1331

1224 1571

1255 1712

1357 1371

1510 1433

1384 1252

1360 1534

1330 1588

1426 1481

1647 1343

1518 1302

1260 1608

1261 1568

1405 1339

1362 1513

1553 1252

1207 1400

1272 1651

1374 1602

1511 1417

1520 1373

1383 1590

1143 1815

1549 1500

1568 1404

1543 1344

1357 1432

1279 1607

1541 1375

1537 1519

1507 1461

1405 1790

1441 1717

1553 1261

1511 1514

1616 1350

1451 1343

1505 1914

1592 1574

1518 1539

1574 1357

1552 1571

1186 1639

1474 1492

1672 1458

1548 1426

1394 1462

1420 1680

1552 1647

1599 1557

1720 1444

1417 1664

1312 1695

1369 1487

1706 1533

1444 1474

1371 1552

1302 1848

1537 1595

1593 1647

1657 1495

*display demand; *VARYING PARAMETERS FOR RISK ANALYSIS BASED ON SCENARIO Parameter beta demands(yr,mo,s) varied scenario specific demand; beta(s) = normal(3.72,0); demands(yr,mo,s)=normal(demand(yr,mo,s),0) *display demands; *DECLARING VARIABLES======================================

Page 54 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033 Variables x (yr) secondary binary y(yr,s) price binary m(yr,s) price dummy varible z(yr) plant binary ztot(yr) z sum capadd(yr) capacity added cap(yr) capacity for given year q(yr,mo,s) water wp(yr,mo,s) water purchased ci(yr) capital investment in given year ciadd(yr) capital investment added op(yr,s) operating cost in given year c TOTAL COST captot(yr) sum of capacities for all years borrowed(yr,s) Money borrowed in given year repayed(yr,s) Money repayed in given year debt(yr,s) debt in given year yrcost(yr,s) Cost in given year ns sum of scenarios scost(s) total cost in scenario s wcost(yr,s) water purchasing cost * price(yr,s) price of water(whole sale or demand based); Binary variable z,y,x; Positive variable cap; Positive variable capadd; Positive variable wp; Positive variable q; Positive variable repayed; Positive variable borrowed; Positive variable debt; Positive variable wcost; *EQUATIONS================================================= alias (yrr, yr) Equations capacity(yr) capacity eqn constraint capinv(yr) capital investment for given year capinvadd(yr) capitalinvestment added opcost(yr,s) operating cost for given year cost Total cost expectation value(BIG CAHUNA) watcost(yr,s) Water cost each month scenariocost(s) total cost for all years for scenario s qcap(yr,mo,s) water capacity in yr&month less than year capacity waterdemand(yr,mo,s) total water demand processed+purchased totfac(yr) total facility captotal(yr) sum of capacities capaddd(yr) capacity added constr(yr) constraint of adding capacity * scenario *FINANCE STUFF yearcost(yr,s)

Page 55 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033 budgetcost(yr,s) budget =g= cost debtfinal(s) last year no debt debteqn(yr,s) defines debt repay(yr,s) pay interest minimum *WATER PRICE STUFF * waterprice(yr,s) defines water price(wholesale or not) * pricedisplay(yr,s) chooseprice(yr,s) pone(yr,s) system of equations that makes price linear operation ptwo(yr,s) -----------"------------"----------pthree(yr,s) -----------"------------"----------pfour(yr,s) -----------"------------"----------; *=====================EQUATIONS=================================== *Scenario Probability *CAPITAL INVESTMENT CALC captotal(yr).. captot(yr) =e= sum(yrr$(ord (yrr) le ord(yr)),cap(yrr)+capadd(yrr)); qcap(yr,mo,s).. captot(yr) =g= q(yr,mo,s); capacity(yr).. cap(yr) =l= maxcap*z(yr); capinv(yr).. ci(yr) =e= (a*z(yr)+b*cap(yr))*(1/(1+df)**(ord(yr)-1)); waterdemand(yr,mo,s).. demands(yr,mo,s) =e= q(yr,mo,s)+wp(yr,mo,s); totfac(yr).. ztot(yr) =e= sum(yrr$(ord (yrr) le ord(yr)),z(yrr)); opcost(yr,s).. op(yr,s) =e= (alpha * ztot(yr) + beta(s) * sum(mo,q(yr,mo,s)))*(1/(1+df)**(ord(yr)-1)); watcost(yr,s).. wcost(yr,s) =e= yrcost(yr,s)-op(yr,s)ci(yr)+borrowed(yr,s)-repayed(yr,s); Capaddd(yr).. capadd(yr) =l= maxcap * x(yr); constr(yr).. x(yr) =l= ztot(yr-1); capinvadd(yr).. ciadd(yr) =e= (aa*x(yr)+1.1*b*capadd(yr))*(1/(1+df)**(ord(yr)-1)); *VARIABLE WATER PRICE chooseprice(yr,s).. sum(mo,wp(yr,mo,s))-1000*y(yr,s) =g= 0; *waterprice(yr,s).. price(yr,s) =e= (34.2*y(yr,s)+45*(1-y(yr,s))); *making price linear pone(yr,s).. m(yr,s)-y(yr,s)*40000 =l=0; ptwo(yr,s).. m(yr,s) =g= 0; pthree(yr,s).. (sum(mo,wp(yr,mo,s))- m(yr,s))-(1-y(yr,s))*40000 =l=0; pfour(yr,s).. sum(mo,wp(yr,mo,s))-m(yr,s) =g=0; *pricedisplay(yr,s).. price(yr,s)=e=(34.2*m(yr,s)+45*(sum(mo,wp(yr,mo,s))-m(yr,s))); *FINANCE BUDGET yearcost(yr,s).. yrcost(yr,s) =e= op(yr,s)+ci(yr)+(1+inflate)**(ord(yr)1)*(25.5*m(yr,s)+90*(sum(mo,wp(yr,mo,s))-m(yr,s)))borrowed(yr,s)+repayed(yr,s); budgetcost(yr,s).. yrcost(yr,s) =l= budget * (1/(1+df))**(ord(yr)-1); debteqn(yr,s).. debt(yr,s) =e= (1+i)*debt(yr-1,s)+borrowed(yr,s)repayed(yr,s); debtfinal(s).. debt('20',s) =e= 0;

Page 56 of 57

AArrsseenniicc R Waatteerr UW OU mO moovvaall ffrroom Reem 55..22..22000033 repay(yr,s).. repayed(yr,s) =g= 2*i*debt(yr-1,s); *repay(yr,s).. repayed(yr,s) =g= Sum(yrr$(ord (yrr) le ord(yr)),i*(1+i)**(20-ord(yrr))*debt(yrr,s)/((1+i)**(20-ord(yrr)))-1); *TOTAL COST scenariocost(s).. scost(s) =e= sum(yr,op(yr,s))+sum(yr,ci(yr))+sum(yr,ciadd(yr))+sum(yr,(1+inflate)**( ord(yr)-1)*(25.5*m(yr,s)+90*(sum(mo,wp(yr,mo,s))m(yr,s)))*(1/(1+df)**(ord(yr)-1))) -sum(yr,borrowed(yr,s)*(1/(1+df)**(ord(yr)1)))+sum(yr,repayed(yr,s)*(1/(1+df)**(ord(yr)-1))); cost.. c =e=sum(s,(1/1)*( sum(yr,op(yr,s))+sum(yr,ci(yr))+sum(yr,ciadd(yr))+sum(yr,(1+inflate)**( ord(yr)-1)*(25.5*m(yr,s)+90*(sum(mo,wp(yr,mo,s))m(yr,s)))*(1/(1+df)**(ord(yr)-1)))) -sum(yr,borrowed(yr,s)*(1/(1+df)**(ord(yr)1)))+sum(yr,repayed(yr,s)*(1/(1+df)**(ord(yr)-1)))); *OUTPUT TO EXCEL *$libinclude XLEXPORT scost.l c:\results.xls c_ /m file out /m:\output.txt/; put out; *put c.l; *CALL PROGRAM model totcost /all/; solve totcost using mip minimizing c; display cap.l, z.l,x.l, ci.l,ciadd.l, beta, wp.l,op.l,ci.l,scost.l,borrowed.l, repayed.l, captot.l, c.l, yrcost.l;

Page 57 of 57