14-EDR_Principles_CDPH_Oct09.pdf

F i n a l R e p o r t

Design Principles of Electrodialysis Reversal and

Basis of Design of the Fort Irwin Water Treatment Plant

Technical Engineering Report Fort Irwin Water Treatment Plant

CIP Project W4

Prepared for

California Department of Public Health, Technical Review Committee

October 2009

Prepared by

U.S Army National Training Center

Post Office Box 10450

Fort Irwin, California 92310-0450

BAO\FIR_DESIGN PRINCIPLES EDR_28OCT09_SRBREV2 iii

Contents

Contents ........................................................................................................................................ iii Acronyms and Abbreviations ................................................................................................... v

1.0 Introduction .................................................................................................................. 1-1

1.1 Existing Water Supply and Quality ...................................................................... 1-1

1.1.1 Groundwater Wells ............................................................................. 1-1

1.1.2 Water Quality Parameters of Concern ............................................. 1-2

1.1.3 Water Quality Goals ........................................................................... 1-3

1.2 EDR Membrane Pilot Testing ................................................................................ 1-3

1.2.1 Operational Modes for EDR Membrane Pilot Studies ................... 1-4

2.0 Design Principles of Electrodialysis Reversal (EDR) ........................................... 2-1

3.0 Basis of Design of the Fort Irwin Water Treatment Plant ................................... 3-1

3.1 Water Recovery ....................................................................................................... 3-1

3.2 FIWTP Treatment Process ...................................................................................... 3-2

3.2.1 Pretreatment ........................................................................................ 3-2

3.2.2 Feed Water Pumps .............................................................................. 3-2

3.2.3 Cartridge Filtration Principles ........................................................... 3-3

3.2.4 EDR Membrane System ..................................................................... 3-3

3.2.5 Product, “Off-Spec”, and Flush Water ............................................. 3-4

3.2.6 Concentrate Recirculation .................................................................. 3-4

3.2.7 Electrode Feed and Waste .................................................................. 3-4

3.2.8 Electrode Clean In Place ..................................................................... 3-6

3.2.9 Clean-In-Place ...................................................................................... 3-6

3.2.10 Antiscalant (Scale Inhibitor) ............................................................ 3-7

3.2.11 Treated Water Storage ...................................................................... 3-7

3.2.12 pH Adjustment and Disinfection .................................................... 3-7

3.2.13 Treated Water Pump Station ........................................................... 3-8

4.0 Operator Safety ............................................................................................................ 4-1

4.1 Installation ................................................................................................................ 4-1

4.2 Operational Hazards ............................................................................................... 4-2

4.2.1 Mechanical Hazards ........................................................................... 4-2

4.2.2 Chemical Hazards ............................................................................... 4-2

4.2.3 Electrical Hazards ............................................................................... 4-2

4.2.4 EDR Membrane Stack Precautions ................................................... 4-3

4.2.5 Other Hazards ..................................................................................... 4-3

4.3 Emergency Shutdown ............................................................................................ 4-4

4.4 Maintenance and Repair ........................................................................................ 4-4

4.4.1 Lockout/Tagout .................................................................................. 4-4

4.4.2 Confined Space Entry ......................................................................... 4-4

4.4.3 Mechanical Hazards ........................................................................... 4-5

4.4.4 Electrical Hazards ............................................................................... 4-5

5.0 References ..................................................................................................................... 5-1

CONTENTS

iv BAO\FIR_DESIGN PRINCIPLES EDR_28OCT09_SRBREV2

Tables Table 1-1 ...................................................................................................................................... 1-2

Fort Irwin Groundwater Production Active Well and Pump Information .......... 1-2 Table 1-2 ...................................................................................................................................... 1-3

Drinking Water Quality Goals for the FIWTP .......................................................... 1-3 TABLE 3-1 ................................................................................................................................... 3-3

Design Filter Criteria for Cartridge Filtration System ............................................. 3-3 Table 3-2 ...................................................................................................................................... 3-5

EDR Stacks ..................................................................................................................... 3-5 TABLE 3-3 ................................................................................................................................... 3-6

Design Criteria EDR Membrane Treatment .............................................................. 3-6 TABLE 3-4 ................................................................................................................................... 3-7

Design Criteria for Treated Water Storage ................................................................ 3-7 TABLE 3-5 ................................................................................................................................... 3-8

Design Criteria for Treated Water Pump Station ..................................................... 3-8

Figures FIGURE 2-1 Effect of a DC Potential on Ions ...................................................................... 2-2 FIGURE 2-2 Continuous Electrodialysis .............................................................................. 2-4 FIGURE 2-3 EDR Membrane Stack Flows ........................................................................... 2-5 FIGURE 2-4 Stack under Negative Polarity ......................................................................... 2-6 FIGURE 2-5 Stack under Positive Polarity ........................................................................... 2-6 FIGURE 2-6 Demineralizing Orientation of Flow Spacer ................................................. 2-9 FIGURE 2-7 Concentrating Orientation of Flow Spacer ................................................. 2-10 FIGURE 2-8 Cross-Section through Stack .......................................................................... 2-11 FIGURE 2-9 Hydraulic and Electrical Staging .................................................................. 2-12 FIGURE 2-10 Three Stage EDR System .............................................................................. 2-13

Appendices A EDR Membrane Installations for Drinking Water Treatment Plants (US Only) B Construction Design Drawings (Process Flow Diagrams and P&IDs)

BAO\FIR_DESIGN PRINCIPLES EDR_28OCT09_SRBREV2 v

Acronyms and Abbreviations g/L micrograms per liter

CIP clean-in-place

DC direct current

DO domestic water

ECIP electrode clean-in-place

ED electrodialysis

EDR electrodialysis reversal

FIWTP Fort Irwin Water Treatment Plant gpm gallons per minute

MCL maximum contaminant level mg/L milligrams per liter mgd million gallons per day

OSHA Occupational Safety and Health Administration

OSP off-specification product ppb parts per billion ppm parts per million psi pounds per square inch

RO reverse osmosis

TDS total dissolved solids us/cm2 microsiemens per square centimeters

VFD variable frequency drive

BAO\FIR_DESIGN PRINCIPLES EDR_28OCT09_SRBREV2 1-1

1.0 Introduction

The purpose of this report is to supplement the design documents (plans and specifications) for the construction of the Fort Irwin Water Treatment Plant (FIWTP). Design documents were provided to the California Department of Public Health, Drinking Water Field Operations Branch, San Bernardino District on July 15, 2009.

The San Bernardino District office requested this report to assist in its technical review of the electrodialysis reversal (EDR) technology. The contents of this report include:

Section 1.0 Introduction

Section 2.0 Design Principles of Electrodialysis Reversal

Section 3.0 Basis of Design of the Fort Irwin Water Treatment Plant

Section 4.0 Operator Safety

Section 5.0 References

Appendix A – EDR Membrane Installations for Drinking Water Treatment Plants (United States only)

Attachment B – Construction Design Drawings (process flow diagrams and P&IDs)

1.1 Existing Water Supply and Quality

Fort Irwin currently has two separate water distribution systems: reverse osmosis (RO) treated water and domestic (DO) water. The RO system water is produced by the existing RO treatment facility and is intended for drinking and cooking uses. The DO system provides disinfected (chlorinated) well water for all other uses such as bathing, irrigation, toilets, and laundry. The majority of residences and offices on post are currently supplied with RO and DO water.

1.1.1 Groundwater Wells

The domestic water is supplied directly from Bicycle Lake, Irwin, and Langford Lake groundwater basins. Water from the individual wells at Bicycle and Langford Lake basins is combined in a raw water tank, disinfected with a sodium hypochlorite solution, and pumped to the distribution system by a high-service pump station. Domestic water may also be pumped from the Irwin Basin (currently only Well I-7).

Fort Irwin has nine active groundwater wells that supply water from three basins, as shown in Table 1-1. The Bicycle Lake well field has five active wells (B-1, B-4, B-5, and B-6) and one booster pump station. Well B-9 supplies DO water to the Bicycle Lake Airfield and is not connected to the DO water supply system within the Cantonment. The Langford Lake Basin has three active wells (L-1, L-2, and L-3) and one booster pump station. The Irwin Basin well field currently has only one active well (I-7).

1.0 INTRODUCTION

1-2 BAO\FIR_DESIGN PRINCIPLES EDR_28OCT09_SRBREV2

TABLE 1-1

Fort Irwin Groundwater Production Active Well and Pump Information Design Principles of Electrodialysis Reversal and Basis of Design of the Fort Irwin Water Treatment Plant Technical Engineering Report, Fort Irwin Water Treatment Plant, CIP Project W4

Groundwater Basin

Well Number

Building Number

Cased Depth

(ft) Capacity

(gpm) Head

(ft) Motor (Hp) Pump Type

Bicycle Lake B-1 5998 430 486 475 60 Vertical turbine

Bicycle Lake B-4 5,996 616 325 477 60 Submersible

Bicycle Lake B-5 5,997 600 777 465 125 Submersible

Bicycle Lake B-6 NA 535 620 414 100 Vertical turbine

Langford Lake

L-1 18,017 500 250 400 40 Submersible

Langford Lake

L-2 18,018 560 675 461 125 Submersible

Langford Lake

L-3 18,019 370 700 403 100 Submersible

Irwin Basin I-7 663 490 1,000 --- 125 Vertical turbine gpm = gallons per minute.

1.1.2 Water Quality Parameters of Concern

The water quality varies from basin to basin and also from well to well within the basin.

Analysis of historical water quality for all wells is discussed below.

1.1.2.1 Fluoride

Bicycle Basin produces water with an average fluoride concentration of 2.3 milligrams per liter (mg/L), Langford Basin produces water with an average fluoride concentration of

6.4 mg/L, and Irwin Basin produces an average fluoride concentration of approximately

9.4 mg/L.

1.1.2.2 Arsenic

Arsenic in Bicycle Lake Basin ranges from approximately 6 to 24 micrograms per liter (g/L), with an average of 15 g/L; Langford Lake Basin arsenic ranges are from 10 to 11 g/L, with an average of 10 g/L; and Irwin Basin arsenic ranges are from 32 to 40 g/L, with an average of 37 g/L.

1.1.2.3 Nitrate

High levels of nitrate historically have been observed in the Irwin Basin. Elevated nitrate levels in parts of the Irwin Basin are suspected to be caused by the historical use of septic systems in the area. The nitrate maximum contaminant level (MCL) is not exceeded by water from any of the wells currently used by Fort Irwin.

BAO\FIR_DESIGN PRINCIPLES EDR_28OCT09_SRBREV2 1-3

1.1.2.4 Other Parameters

Iron and total dissolved solids (TDS) are among other constituents that were targeted for removal by the water treatment plant process. The current secondary MCL for TDS is a range of 500 to 1,000 mg/L. Bicycle Basin produces water with average iron concentration of

0.6 mg/L and an average TDS of 620 mg/L. Langford Basin produces water with an average iron concentration of 0.5 mg/L and an average TDS of around 510 mg/L. Irwin Basin produces water with an average iron concentration of approximately 0.9 mg/L and an average TDS concentration that is generally below 500 mg/L.

1.1.3 Water Quality Goals

The FIWTP must produce water that complies with existing and anticipated future regulations. Fort Irwin established water quality goals for treated water based on current and pending regulations. These goals were established with a sufficient margin of safety to ensure consistent compliance during normal operation at the new FIWTP. The selected treatment process must be able to reliably produce water that meets these water quality goals.

Table 1-2 presents a summary of the federal and state MCL requirements and Fort Irwin’s water quality goals.

TABLE 1-2

Drinking Water Quality Goals for the FIWTP Design Principles of Electrodialysis Reversal and Basis of Design of the Fort Irwin Water Treatment Plant Technical Engineering Report, Fort Irwin Water Treatment Plant, CIP Project W4

Constituent Unit Federal MCL California MCL Fort Irwin’s Drinking Water

Quality Goal

Arsenic µg/L 10 10 7.5

Fluoride mg/L 4 2 1.6

Iron mg/L 0.3 0.3 0.225

TDS mg/L 500-1,000 500-1,000 500-1,000 pH Unit 6.5 to 8.5 6.5 to 7.5 7 to 8

NO3/NO2 mg N/L 10 10 7.5

1.2 EDR Membrane Pilot Testing

Pilot studies using EDR membrane technology were performed from March 14 to June 7, 2006. The specific objectives of pilot testing included:

Assessing EDR performance in reducing fluoride, arsenic, TDS, and nitrate.

Determining fluoride and arsenic reduction correlations at various TDS reduction rates.

Determining the highest practical water recovery for EDR.

Determining the required chemicals and dosages.

1-4 BAO\FIR_DESIGN PRINCIPLES EDR_28OCT09_SRBREV2

EDR pilot testing demonstrated the ability of the process to reduce fluoride and arsenic levels to below California MCLs and the Fort Irwin drinking water quality goals. The Fort Irwin Water Treatment Process Selection Based on Pilot Testing Results (CH2M HILL, 2006) presented the results of this pilot testing.

1.2.1 Operational Modes for EDR Membrane Pilot Studies

From a performance standpoint, three independent “trials” were conducted under varying feedwater quality:

Trial 1 – Ambient Feed Water Testing – (March 14 to May 1, 2006) involved pilot testing using the existing DO water.

Trial 2 – Spiked Feed Water - (May 3 to May 23, 2006) involved testing with feed water spiked to simulated increased concentrations of fluoride.

Trial 3 – Process Optimization - (May 24 to June 7, 2006) involved the optimization phase of testing, where the most advantageous operational scheme determined from Trial 1 and Trial 2 testing was used to meet water quality goals while minimizing electricity and chemical use.

The results contained in the 2006 report indicated the EDR process provided a robust process for treating groundwater from each of the basins, either individually or as blends from the basins. The unit was reliable and easy to operate, with minimal maintenance requirements, once the general operating parameters were established. The inherent flexibility of the process allows adjustment of power use (voltage settings) to easily match contaminant removal to meet the water quality goals and to optimize power consumption.

BAO\FIR_DESIGN PRINCIPLES EDR_28OCT09_SRBREV2 2-1

2.0 Design Principles of Electrodialysis

Reversal (EDR)

This section outlines the scientific principles of operation for the technologies used in the water treatment system. These descriptions are general, and may not reflect the system as designed in all details.

EDR is an advanced water treatment process based on electrodialysis (ED). Both treatment processes are used to remove dissolved solids in the form of mineral ions from solution.

EDR was developed to reduce the need for regular chemical addition and cleaning that exists with ED systems.

Many solids will dissociate in water, producing a solution containing dissolved solids.

Because the electrons in the solids do not attach to the same atoms when they dissolve as they would in elemental form, the dissolved solids include negatively charged particles (anions, with an excess of electrons) and positively charged particles (cations, with a deficit of electrons). Together these are called ions.

Many compounds are ionic in nature and act in this way. These include acids, salts, and minerals. For every salt or mineral, there is a limiting quantity that can be dissolved into water. At some point, the solution will not accept any more dissolved salts (or minerals) and it is said to be saturated. Any more salt or mineral added will simply settle to the bottom.

This saturation point (saturation limit) is affected by both the temperature and the exact mix and concentration of salts and minerals in solution. In most cases, increased temperature increases the saturation point. In general, the more ionic species present, the more total ions will be dissolved, although saturation points for individual compounds are commonly reduced.

A key characteristic of an ionic solution is that it conducts an electrical current. A pure water solution will not conduct a current, and the greater the concentration of ions, the greater the conductance. Thus the conductivity of a solution can be used as a measure of the concentration of dissolved solids.

ED/EDR is based on the principles governing the behavior of an ionic solution when it is subject to direct current (DC) potential. Ions are quite mobile in solution and when subjected to direct current potential, are attracted towards the electrode of opposite charge from the ion. Thus, positively charged ions (or cations) are attracted to the cathode (negative electrode), and negatively charged ions (anions) are attracted to the anode (positive electrode).

In ED, alternating anion and cation permeable membranes (called anion transfer and cation transfer membranes) are placed in layers with an anode on one side of the assembly and a cathode on the other. A membrane permeable to cations only is placed nearest to each of the electrodes. When a current is applied to the system, channels where the cation transfer membrane is towards the cathode and the anion transfer membrane is towards the anode are depleted of ions as they migrate through the membranes. In the adjacent channels, the

2.0 DESIGN PRINCIPLES OF ELECTRODIALYSIS REVERSAL

2-2 BAO\FIR_DESIGN PRINCIPLES EDR_28OCT09_SRBREV2

membranes do not allow migration in the direction the ions are drawn. Thus, alternating channels are formed from deionized product water and ion-rich concentrate stream. This is illustrated schematically in Figure 2-1, where “C” represents a cation transfer membrane and “A” represents an anion transfer membrane.

FIGURE 2-1

Effect of a DC Potential on Ions Design Principles of Electrodialysis Reversal and Basis of Design of the Fort Irwin Water Treatment Plant Technical Engineering Report, Fort Irwin Water Treatment Plant, CIP Project W4

(+) Anode

Cathode (-)

Cl -Na +

C

A

A

C

C

Cl -Na +

Cl -Na +

Cl -Na +

Cl -Na +

Cl -Na +

(e -)

The membranes limit how far ions can travel, and the results in the overall system are as follows:

Compartment 1 contains a metal electrode where reduction occurs. Hydrogen gas (H2) and hydroxyl (OH-) ions are created from dissociation of water.

Compartment 2 will contain purified water as Cl- ions pass through the anion transfer membrane (A) into compartment 3, while Na+ ions move through the cation transfer membrane (C) into compartment 1.

Compartment 3 will contain concentrated ions as Na+ ions cannot move through the anion transfer membrane and remain in compartment 3. The Cl- ions cannot pass through the cation transfer membrane and also remain in compartment 3. Additional ions are added from compartments 2 and 4.

Compartment 4 will contain purified water as Cl- ions pass through the anion transfer membrane into compartment 5, while Na+ ions move through the cation transfer membrane into compartment 3.

BAO\FIR_DESIGN PRINCIPLES EDR_28OCT09_SRBREV2 2-3

Compartment 5 will contain concentrated ions as Na+ ions cannot move through the anion transfer membrane and remain in compartment 5. The Cl- ions cannot pass through the cation transfer membrane and also remain in compartment 5. Additional ions are added from compartments 4 and 6.

At the anode (compartment 6) oxidation occurs; oxygen gas (O2) and hydrogen ions (H+) are produced.

The result is that compartments 2 and 4 are depleted of ions, whereas compartments 3 and 5 are concentrated. The electrode compartments are special cases.

In commercial ED applications, the setup is more complicated than in this simple example.

An anion transfer and cation transfer membrane, separated by a spacer, with a spacer on one side form a cell pair. Three hundred to 500 of these cell pairs are stacked one on top of another to form a membrane stack. In the simplest applications, two electrodes (one anode and one cathode) are located one on either end of this stack. When current is applied to this stack, alternating channels of product water and concentrate are created. The system is made continuous by manifolds that direct-feed water through each space in the system and keep product and concentrate streams separated. Essential parts of the system are the anion and cation transfer membranes and the direct current applied to induce ion migration.

When feed water is introduced through both the demineralizing and the concentrate channels, low recoveries (50 percent or less) are achieved. Modification of the stack to define two distinct flow channels, one for feed water and one for concentrate recycle, allows concentrate to be recycled and only feed water to flow through demineralizing channels.

Using concentrate recycle, typical recoveries are 85 to 95 percent.

The concentration of ions in the concentrate is a limiting factor only to the extent that very high ion concentrations lead to precipitation, which will foul membranes. Therefore, some concentrate is continuously blown down to waste and is made up with feed water.

The commercial continuous ED process shown in Figure 2-2 presents two improvements over the batch-wise purification shown in Figure 2-1. First, by creating a stack that has an inlet and an outlet for each of the spaces between the membranes, a continuous process may be implemented. Second, by taking the concentrate from the concentrating spaces and recycling it, the overall recovery of the process is greatly increased. The demineralized stream makes a single journey through the stack and exits as product water. The concentrate stream is partially recycled to reduce waste.

In continuous ED, the motivating pressure for processing of water is provided by two pumps. The first is the feed pump, which provides the pressure for flow of the feed (or dilute) stream through the stack to exit as product water. Because the concentrate (or brine) is hydraulically separate from the feed, a separate pump, the concentrate pump, is used to circulate the concentrate stream.

The primary limiting factor in continuous ED is the concentration of ions in the concentrate recycle. This is a limiting factor only to the extent that very high ion concentrations lead to precipitation, which will foul membranes. The concentrate recycle may be injected with acid and other chemicals to avoid scaling in the stack, but this does not eliminate the problem. To prevent ion concentrations above the precipitation or scaling limits, some of the recycle is

2-4 BAO\FIR_DESIGN PRINCIPLES EDR_28OCT09_SRBREV2

continuously blown down (concentrate blowdown with high ion concentrations) and replaced by feed water (concentrate makeup with a relatively low ion concentration).

FIGURE 2-2

Continuous Electrodialysis Design Principles of Electrodialysis Reversal and Basis of Design of the Fort Irwin Water Treatment Plant Technical Engineering Report, Fort Irwin Water Treatment Plant, CIP Project W4

Concentrate Make-up

Concentrate Recycle Concentrate Blowdown

Concentrate In

Product

Feed

The feed water to an continuous ED system is actually divided into three streams. The largest flow is the feed stream, which is processed in the ED stack and is discharged as desalted product water. The second flow from the feed stream (concentrate makeup) is fed to the concentrate stream, thereby helping to control the number of times concentrate is recycled and thus controlling the recovery. The third stream (not shown in Figure 2-2) is the electrode compartment feed water, which flushes away gases generated at the electrodes and helps maintain a constant pH at the electrodes.

Continuous ED is made more efficient by concentrate recycle but still has the tendency to foul from deposition of particles on the membranes during operations. When fouling occurs, chemical cleaning is required to remove fouling agents.

The EDR process is a refinement of the ED process, allowing continuous operation without regular chemical cleaning. In EDR, the electrical polarity, and thus the demineralized and concentrate flow passages, are automatically reversed two to four times every hour. This results in a reversal of direction of ion movement, which provides electrical flushing of scale-forming ions and colloidal matter from the membrane surfaces. This electrical flushing controls fouling and scaling of membranes and can eliminate the need for extensive pretreatment of feed water. It also greatly reduces the need for down time for cleaning and the use of acids and chemicals to control scaling.

BAO\FIR_DESIGN PRINCIPLES EDR_28OCT09_SRBREV2 2-5

A schematic of the EDR stack is shown in Figure 2-3.

FIGURE 2-3

EDR Membrane Stack Flows Design Principles of Electrodialysis Reversal and Basis of Design of the Fort Irwin Water Treatment Plant

Cation Transfer Membrane

Feed Inlet

Concentrate Inlet Electrode Feed

(-) Cathode

(+) Anode

Electrode Waste

Electrode Feed

Product Outlet

Concentrate Outlet

Demineralized Flow Spacer

Anion Transfer Membrane

Concentrate Flow Spacer

Top End Plate

Bottom End Plate

In this example, one electrode compartment is located at the top of the membrane stack, and one is located at the bottom of the membrane stack. Typically, the electrodes are platinum plate titanium and are masked near the manifolds to prevent shorting and hydrogen embrittlement of the electrode. Each electrode compartment is bounded by an extra thick spacer and a thick heavy cation transfer membrane.

EDR is an improvement over ED in that scaling and chemical addition are greatly reduced, while maintaining continuous production of demineralized water. To reduce scaling, the polarity of the electrodes is reversed two to four times each hour. The concentrating channels become demineralizing, and the demineralizing channels become concentrating since the polarity reversal also reverses the direction of movement for individual types of ions within the membrane stack. The polarity is determined by the charge of the electrode on the bottom of the stack.

2-6 BAO\FIR_DESIGN PRINCIPLES EDR_28OCT09_SRBREV2

This process is shown in Figure 2-4 and Figure 2-5, which show flows through the stack under the two opposite polarities.

FIGURE 2-4

Stack under Negative Polarity Design Principles of Electrodialysis Reversal and Basis of Design of the Fort Irwin Water Treatment Plant

- - - cathode - - -

+ + + anode + + feed water demineralized demineralized water concentrate anion cation anion demineralized cation demineralized

FIGURE 2-5

Stack under Positive Polarity Design Principles of Electrodialysis Reversal and Basis of Design of the Fort Irwin Water Treatment Plant

- - - cathode - - -

+ + + anode + + + feed water demineralized demineralized water concentrate anion cation anion demineralized cation

It can be seen in Figure 2-4 and Figure 2-5 that when the polarity is reversed, the demineralized stream and concentrate stream compartments are also reversed. This

BAO\FIR_DESIGN PRINCIPLES EDR_28OCT09_SRBREV2 2-7

alternating exposure of membrane surfaces to the product and concentrate streams provides a self-cleaning capability that enables desalting of scaling or fouling waters and recovery of up to 94 percent of the feed water.

At reversal, automatically operated valves switch the two inlet and outlet streams so that the incoming feed water flows into the new demineralizing compartments and the recycled concentrate stream flows into the new concentrating compartments. The effect of this reversal is that the concentrate stream remaining in the stack whose salinity is higher than the feed water must now be desalted. This creates a brief period of time in which the demineralized stream (product water) salinity is higher than the specified level. This slug of water is known as off-specification product (OSP). Conductivity controlled valves shunt the product to waste until specifications are met.

GE-Ionics’ EDR systems minimize the OSP time by using phased reversal. Phased reversal is based on independently controlling the reversal of the inlet diversion valves, the electrical polarity of each stage, and the outlet diversion valves. The sequence initiates with the reversal of the inlet valves. This creates a slug of reversal water (a mix of concentrate and demineralized streams) that then proceeds through the system. As this slug of water proceeds through each subsequent stage in the system, the stage reverses electrical polarity.

Finally, the outlet valves reverse. Thus, the total reversal water is only that contained in a single stage in the system instead of that contained in the entire system. The OSP time is reduced to that required for the water to pass through one stage, regardless of the number of stages in operation, instead of the time required to traverse the complete line of stages.

Because of reversal, no flow compartment in the stack is exposed to high solution concentrations for more than 15 to 30 minutes at a time. Any buildup of precipitated salts is quickly dissolved and is carried away when the cycle reverses. Colloidal particles, which generally have a negative surface charge, will have tended to deposit on the anion transfer membrane. When the polarity reverses, these colloidal particles are pulled away and will be flushed out with the flow.

When reversal occurs, the chemical reactions within the electrode compartments are also reversed since the electrodes change polarity. Since the electrodes are reversing polarity, the old cathodic compartment, with its alkaline environment, will now become anodic with a low pH (acidic) environment. Conversely, the old anodic compartment, with its acidic environment, will now become cathodic with a high pH (alkaline) environment. Valves in the electrode streams automatically switch the flows in the anode and cathode compartments.

The acidic environment created at an anode helps prevent scale formation and can now be used to advantage. To obtain maximum benefit from anodic acid generation, the anode stream may be operated in a non-flow condition that allows the concentration of hydrogen ions to increase and lower the pH to approximately 2 to 3. However, chlorine and oxygen gases are also being generated at the anode. These gases tend to accumulate on the electrode to form a layer of gas that greatly increases the resistance at the electrode. This accumulation of gases is known as gas blanketing. To reduce the effect of this condition, the anode stream is allowed to flow for a brief period (typically one minute out of every four to five) to flush out the gas bubbles. This action is known as a “bump.” While the anode is in the bumping

2-8 BAO\FIR_DESIGN PRINCIPLES EDR_28OCT09_SRBREV2

mode, the cathode flows continuously to minimize hydroxyl concentration since an alkaline environment will increase scaling tendency.

Stack shorting (internal leaking bypassing the membranes) and water transfer (water crossing the membranes with ions) allow some water to flow directly past the membranes.

These water losses are calculated with initial stack calculations and do not affect recovery.

The pressure in the concentrate stream is generally kept lower than that of the dilute stream to ensure that any leakage will be from the dilute to the concentrate.

The recovery is controlled by controlling the blowdown rate for the concentrate stream.

Typical recovery rates are 85 percent of feed water. Typically, maximum salt removal for any hydraulic stage is 55 to 60 percent, with normal design values at 40 to 50 percent.

The reversal process has five major effects on the operation of a membrane stack:

Polarization films are broken up two to four times each hour, thus preventing polarization scale.

Freshly precipitated scale and seeds of scale are broken up and flushed to waste before they can grow or cause damage.

Slime and similar formations on membrane surfaces are broken up by electrically reversing the direction of colloidal particle movement.

The entire complex of practical problems associated with the need for continuous feed of acids or complex chemicals is eliminated.

The electrodes are automatically cleaned with acid formed during anodic operations.

When building EDR stacks, three kinds of transfer membranes are used, as discussed above.

These are the anion and cation transfer membranes, plus special heavy cation transfer membranes (used adjacent to electrode compartments). The heavy cation transfer membranes have all of the properties of the regular cation membrane; however, they are made twice as thick (1.0 mm) as the normal cation (0.5 mm) to withstand a greater hydraulic pressure differential than that of a normal membrane. The heavy cation transfer membrane has only two manifold cutouts (to allow dilute and concentrate streams to pass) in contrast to the four manifold cutouts in a regular membrane.

All membranes are flexible sheets of cloth-reinforced resin. The key properties of the membranes are:

Low electrical resistance.

Insolubility in aqueous solutions.

Semi-rigidity for ease of handling during stack assembly.

Resistance to change in pH from 1 to 10.

Operability at temperatures in excess of 115 F (46 C).

Resistance to osmotic swelling (volume expansion due to water uptake) when cycled between 220 parts per million (ppm) and 30,000 ppm salt solutions.

BAO\FIR_DESIGN PRINCIPLES EDR_28OCT09_SRBREV2 2-9

Long life expectancies.

Resistance to fouling.

Impermeability to water under pressure.

Dimensions of 18 inches by 40 inches (18 inches by 20 inches for Mark-II membranes) with thickness of approximately 0.020 inch (0.5 millimeter) (heavy cation membrane about twice as thick).

Cation transfer membranes are amber and stamped CATION in blue. Anion transfer membranes are pale yellow and stamped ANION in blue.

Spacers are used between all membranes to keep the membranes separated and to channel flow across the membranes. Cutouts in the spacers for inlet and outlet are made to match cutouts in the membranes, and each spacer has a defined flow path. Two orientations of spacers are required to allow for concentrate and demineralized flow streams. The flow streams are usually achieved by using a single spacer which is flipped over.

A typical layout for a spacer is shown in Figure 2-6. Other layouts exist, with different flow paths or with the inlet and outlet manifolds at the center of the spacer. All spacer layouts function in the same fashion as that illustrated In Figure 2-6.

FIGURE 2-6

Demineralizing Orientation of Flow Spacer Design Principles of Electrodialysis Reversal and Basis of Design of the Fort Irwin Water Treatment Plant

A

B

C

D

The spacer illustrated in Figure 2-6 is a demineralizing spacer. In the adjacent concentrating spacers, the spacer would be flipped over top to bottom so that the active inlet and outlet in the demineralizing spacer (A and C) would be over the conduit spaces B and D. Thus, no feed water would flow through the adjoining concentrating spacer; instead, the inlet and outlet manifold in the concentrating spacers would be over channels B and D and would carry concentrate through the spacer. To visually illustrate this concept, the concentrating spacer corresponding to the demineralizing spacer in Figure 2-6 is shown in Figure 2-7.

When a membrane stack is properly assembled, the manifold cutouts in the membranes and spacers form vertical pipes in the stack.

2-10 BAO\FIR_DESIGN PRINCIPLES EDR_28OCT09_SRBREV2

FIGURE 2-7

Concentrating Orientation of Flow Spacer Design Principles of Electrodialysis Reversal and Basis of Design of the Fort Irwin Water Treatment Plant

A

B

C

D

Close examination of the spacer reveals an array of strands of netting. These strands force the water to flow in an over-and-under manner. This type of flow causes turbulence in the stream, thereby allowing higher electric current flow per unit area and more efficient use of the membrane area in the stack. The U-shaped screen spacer causes much more turbulence than the tortuous-path spacer, which enables much more efficient use of the membrane area in the stack. This means that a higher salt removal per stage can be achieved. Typically, the U-shaped spacer stacks require one less stage for an application than was required with the tortuous-path spacer stacks. This orientation is shown in Figure 2-8.

Figure 2-8 shows a cross section of a stack illustrating the flow around the strands of netting.

This diagram depicts three spacers sandwiched between four membranes to form flow compartments. Turbulence in the stream created by the strands of netting promotes mixing, which aids transfer of ions to the surface of the membranes. Turbulence also breaks up boundary layers and slimes at the membrane surface.

The manner in which the membrane stack array is arranged is called staging. The purpose of staging is to provide sufficient membrane area and retention time to remove a specified fraction of salt from the demineralized stream. Staging is the process of adding additional passes through an EDR stack for each increment of water processed. The ultimate goal of staging is to increase product water purity. Two types of staging may be used: hydraulic staging and electrical staging. In electrical staging, another pair of electrodes is added to the system and feed water passes between more than one set of electrodes, which can be set at different currents and voltages. In hydraulic staging, water passes through a separate set of hydraulic channels. Additional hydraulic stages can be created without additional electrical stages, but that an additional electrical stage necessarily also introduces a new hydraulic stage.

BAO\FIR_DESIGN PRINCIPLES EDR_28OCT09_SRBREV2 2-11

FIGURE 2-8

Cross-Section through Stack Design Principles of Electrodialysis Reversal and Basis of Design of the Fort Irwin Water Treatment Plant

Anion-Exchange Membrane

Cation-Exchange Membrane

Anion-Exchange Membrane

Cation-Exchange Membrane

Concentrate Flow

Demineralized Product Flow

Concentrate Flow

- - - Cathode (-) - - -

+ + + Anode (+) + + +

Screen Spacers

Turbulent Flow Paths

The membrane stack discussed above is an example of a one hydraulic, one electrical stage stack. Each increment of water, upon entering the stack, makes one pass across the membrane surface between one pair of electrodes and exits. In GE-Ionics’ membrane stack, water flows in multiple parallel paths across the membrane surfaces; a single pass consists of flowing through one water flow spacer between two membranes and exiting through the outlet manifold.

A numerical example will further illustrate the effect of staging. If the inlet water is 2,000 ppm and the desired product water salinity is 250 ppm, three stages would be required, assuming 50 percent salt removal per stage. The salt removed by each stage would be as follows:

Stage 1 – 2,000 ppm in, 1,000 ppm out (50 percent removal) Stage 2 – 1,000 ppm in, 500 ppm out (50 percent removal) Stage 3 – 500 ppm in, 250 ppm out (50 percent removal)

Electrical and hydraulic staging can occur within a single EDR stack or can be spread across multiple stacks. Each stack must have at least one hydraulic and one electrical stage. The decision to use a single stack or multiple stacks is based on the overall capacity required from the system, the final product water quality required, and physical limitations of staging. In general, Aquamite II and Aquamite V systems use a single stack with hydraulic and electrical stages within the stack. Aquamite X and larger systems generally use multiple stacks that each has a single electrical and hydraulic stage.

The simplest application of staging is addition of hydraulic stages within a single stack by adding heavy cation transfer membranes as interstage membranes to create separate flow paths. However, since salt removal is directly proportionate to current and inversely proportionate to flow rate, electrical staging also becomes desirable. For example, with the

2-12 BAO\FIR_DESIGN PRINCIPLES EDR_28OCT09_SRBREV2

three stages postulated above with three hydraulic stages, if these were within one electrical stage (e.g., one stack), the flow rate per cell pair would have to be doubled for each successive stage to obtain the same removal as in the example, and the number of cell pairs in each following stage would have to be halved. This has the secondary effect of greatly increasing the pressure drop with each stage.

Electrical staging is accomplished by inserting additional electrode pairs into a membrane stack. This gives flexibility in system design, thereby providing maximum salt removal rates while avoiding polarization and hydraulic pressure limitations. Each electrical stage allows the use of an independently controlled current (I) to the cell pairs within that stage. In the example above, if two electrical stages are used, there would be two independent currents:

(I1) and (I2).

An example of electrical staging combined with hydraulic staging in a single stack is shown in Figure 2-9.

FIGURE 2-9

Hydraulic and Electrical Staging Design Principles of Electrodialysis Reversal and Basis of Design of the Fort Irwin Water Treatment Plant Technical Engineering Report, Fort Irwin Water Treatment Plant, CIP Project W4

Hydraulic Stage 1 200 Cell Pairs

Hydraulic Stage 2 100 Cell Pairs

Hydraulic Stage 3 100 Cell Pairs

Electrical Stage 1

Feed Inlet Concentrate Inlet

(+) Anode

Product Outlet

Concentrate Outlet

(-)Cathode

(+) Anode

(-) Cathode

Electrical Stage 2

Heavy Cation Interstage Membrane

Hydraulic stages 1 and 2 are now in electrical stage 1, and hydraulic stage 3 is also electrical stage 2. With the ability now to introduce a lower current (I2) for electrical stage 2/hydraulic

BAO\FIR_DESIGN PRINCIPLES EDR_28OCT09_SRBREV2 2-13

stage 3, additional cell pairs can be added to reduce the pressure drop (dP) to an acceptable level. By using 100 cell pairs in hydraulic stage 3, the pressure drop across the entire stack will be within acceptable limits.

The more hydraulic and electrical stages in a single stack, the lower the water production from the stack as the available cell pairs are split into smaller groups. In larger systems, additional hydraulic (and electrical) stages may be created by simply adding more stacks in series to achieve the desired water purity, as shown in Figure 2-10. Each stack also forms a separate electrical stage.

FIGURE 2-10

Three Stage EDR System Design Principles of Electrodialysis Reversal and Basis of Design of the Fort Irwin Water Treatment Plant

Stage

Stage

Demineralized Inlet

Demineralized Inlet

Concentrate Inlet

Concentrate Inlet

Stage

Demineralized Outlet

Demineralized Outlet

Concentrate Outlet

Concentrate Outlet

The concept of staging leads to great flexibility in system design with standard components.

If 75 percent demineralization is required, then sufficient stages are installed to give the required 75 percent. If more stages should ever be needed, even in a plant that has been operating for some time, then additional stages can be added. This may be desired, for example, if a brackish aquifer becomes more saline with time, if the system is used with a more saline aquifer at the same or a different location, or if a higher-purity product is desired.

In systems where greater production volume is required, additional lines of treatment may be used. These lines will be connected in parallel. Generally, each line in an installation will contain the same number of hydraulic and electrical stages.

An EDR unit includes the membrane stacks, the reversal unit, the concentrate recycle/waste system, and electrical control modules. Associated systems include the electrode clean-in-place (ECIP) system, clean-in-place (CIP) system, and the chemical addition systems. The CIP and ECIP are discussed below with the remainder of the EDR system.

Under normal operations, two flows enter the EDR system as a whole, and three flows leave the system. The incoming flows are feed water and brine acid feed. The exiting flows are product, brine blowdown, and electrode waste. Within the EDR system, these steams are split and combined to provide the flow required through the EDR stacks.

The inlet and outlet flows for the EDR system as a whole are driven by the input and output flows from the EDR stacks. Each of the EDR stacks has four inlet and four outlet flows. Two

2-14 BAO\FIR_DESIGN PRINCIPLES EDR_28OCT09_SRBREV2

of these flows are through the top and bottom electrode compartments (and center, if present), which switch between anode and cathode with reversal. The other two flows are through the X and Y streams of the stacks, which switch between concentrating and demineralizing with reversal. The source, components, and destination of each flow are as follows:

The feed (dilute)/product comes from the feed inlet, flows through the stacks where ions are removed, and then flows to the product piping. Small streams are diverted from the feed for brine make-up and electrode feed. The product flow is sent to product storage or diverted to OSP , depending on whether it meets specifications. The product flow is in the Y stream under positive polarity and in the X stream under negative polarity.

The concentrate is recycled within the EDR system. Because ions are continually being added, a portion of the stream is continually blown down to maintain concentrations below precipitation levels. A volume equal to the blowdown is added back as makeup from the feed stream to maintain a constant concentrate volume. Acid from the brine acid feed may be intermittently added to this stream to prevent precipitation. The concentrate flows through the X stream under positive polarity and through the Y stream under negative polarity.

The cathode flow is continual. The electrode feed for this flow is a side-stream from the feed stream. Electrode waste, flowing off of the cathode, is sent through the degasifier before being sent to waste.

The anode flow is intermittent. The electrode feed for this flow is a side stream from the feed stream. Electrode waste, flowing off the anode, is sent through the degasifier before being sent to waste.

The hydraulic piping in the EDR system is designed to direct and control the flows listed above.

In addition, during cleaning activities, additional flows may enter and leave the system.

ECIP occurs intermittently while the system is operating. When ECIP occurs, acid is added to the electrode feed streams to clean the electrode compartments. This typically occurs one hour out of every 24. CIP procedures for the membrane stacks occur periodically as needed.

In order for CIP to occur, the system must be shut down. During CIP, the ECIP tank and pump are connected to the feed and concentrate streams, all other flows are shut off, and the stacks are flushed with chemical solutions for cleaning.

BAO\FIR_DESIGN PRINCIPLES EDR_28OCT09_SRBREV2 3-1

3.0 Basis of Design of the Fort Irwin Water

Treatment Plant

The EDR demonstration pilot plant was carefully modeled to that of a larger EDR system (up to 1.5 million gallons per day [mgd] per unit) using the GE-Ionics-developed WATSYS computer program. The product water quality and the DC power setting required to produce a predicted water quality were correlated between the demo and the full-size unit.

The WATSYS runs usually predict on the conservative side. The WATSYS computer model has been developed based on 25+ years of corrected input from actually operating EDR systems. In the case of Fort Irwin, the balance of anions in the groundwater favored the reduction of fluoride above that in almost all other EDR fluoride reduction applications.

This is due to the ratio of bicarbonate and sulfate, with low chlorides in Fort Irwin source water.

With the…