1. LBWD’s Prototype-Scale Testing of NF Membranes for Seawater Desalination Robert C. Cheng, Tai J. Tseng, Kevin L. Wattier February 17, 2011 MSSC National Salinity Summit 1

2. Research Partners   US Bureau of Reclamation   CA Dept. of Water Resources   LA Department of Water and Power   Southern Nevada Water Authority   Tampa Bay Water Authority  UCLA  University of New Hampshire  Clemson University  University of Illinois  Montana State University  University of Central Florida  Virginia Tech  University of Nevada, Reno  University of Iowa AcademiaGovernment Industry  DuPont  Water Research Foundation (AwwaRF)  Black & Veatch  CH2M Hill  MWH

3. Water resources…under stress Los Angeles Aqueduct: reduction California Aqueduct: “Southern California Loses Up to 30 Percent of Its Supplies from Delta Next Year and Possibly Longer” -Business Wire News Colorado River Aqueduct: ~50% reduction …more pressure to act locally to reduce dependence on imports City of Long Beach ~500,000 residents

4. Reliability, at lowest reasonable rates “New” Conjunctive Use Seawater Consumers Potable “Traditional” Groundwater Surface water Conservation 2007 Declaration 16+% decrease in use “100 by 100” Initiative Seawater barrier Irrigation Indirect recharge Reclaimed

5. LBWD’s Resource Mix 2010 2015 Conservation 15% Reclaimed 9% Imports 32% Groundwater 44% Conservation 15% Reclaimed 12% Imports 30% Groundwater 33% Desal 10% 5

6. Shifts in water resources 47% 2007 (declaration of drought) 2010 (shortage allocation) 53% Imports (MWD) (MWD) Groundwater (LBWD) $478 - $574/af Groundwater 53% < $400/af $300/af 47% Imports $740 - $3,134/af

7. Desal research...to lessen risks  Full plant cost will be costly  Other full-scale experiences point out value for research  Substantial interest for accurate operational and cost information  Federal – USBR (federal authorization)  State – CA DWR (CA Prop 50 funding)  Local – LA DWP (research site, power)

8. O&M- Electrical Power 44% Debt (Capital) 37% Membrane Replacement 5% Labor 4% Maintenance & Parts 7% Consumables 3% Federal Roadmap Estimate Power + Debt = 81% Non-energy O&M = 19%

9. Concept - “The Long Beach Method” Two Pass Nanofiltration Nanofiltration  Energy Savings  Lower pressure requirements, lower energy consumption  Quality Protection  Twice the protection of single-pass technology

10. NF membrane for seawater desal  Proof-of-concept  Initiated testing ~2001  verified through 2-yr AwwaRF project, “A Novel Approach to Seawater Desalination Using Dual-Staged Nanofiltration”  Patent application  US patent 7144511, granted 12/5/06  “Two stage nanofiltration desalination system”  Prototype plant construction/operations  2004 - 2010

11. 11 How to integrate seawater into system? Post treatment / Distribution Pretreatment NF 2 or RO A $20 M, 10-year investment A $20 M, 10-year investment Leverage various partnerships for technical input and other support Leverage various partnerships for technical input and other support Federal/State/Local Funds, 50% funding by Reclamation Federal/State/Local Funds, 50% funding by Reclamation Under Ocean Floor Intake and Discharge Prototype UV/ClO 2 Mitigation of WQ impacts due to integration of new source

12. Jul 07 -Jan 04 - Jul 04 - Jan 05 -Jul 05 -Jan 06 -Jul 06 -Jan 07 -Jan 08 - Construction Jan 09 -Jul 08 -Jan 10 -Jul 09 - Jul 10 - Design Desal Prototype Research ClO 2 and UV Under Ocean Floor Desal Site Alternative Study Site Restoration (2012) Post-treatment Design Research schedule Jan 11 -

13. South Train North Train MF Unit Prototype Plant Flow Diagram 13 Intake/ discharge Project site

14. Prototype Plant  300,000 gpd facility, 8-in vessels 14

15. Energy Recovery Operation PX Booster pump PX 15

16. Other Issues  Technical  Water quality met (boron, bromide, etc.)  Blending issues with existing water  Environmental  Impingement/entrainment  Discharge  Public Trust  Sound investment  Transparency  Permitting

17. Research Objectives  Compare NF 2 against RO  Water quality (TDS, boron, bromide), energy, reliability  Optimize NF 2 process  Energy recovery device  Biofouling control method: UV vs. ClO 2  Vary configuration/membranes  Analyze cost for full-scale plant 17

18. Research Approach  Phase I- NF 2 vs. RO  Short tests to determine trends  General WQ and energy recovery monitoring  May ’06 – Dec ‘07  Phase II-NF 2 vs. RO  2+ weeks of selected conditions from Phase I  Detailed WQ analyses  Jan ’08 – Dec ‘08  Phase III and IV-NF 2 optimization test  2+ weeks tests: NF 5 vs. 7, mixed membrane  UV vs. ClO 2  Jan ’09 – Jan ‘10 18

19. NF 2 vs. RO Process MembranesPressure (psi)Recovery (%) NF 2 Pass 1 NF9054039% NF 2 Pass 2 NE9018472% RO Pass 1 SWC3+75640% RO Pass 2 NE9021880% 19 MF Energy Recovery 1 st Pass NF (South Train) Cartridge Filter Combined Effluent Tanks Cartridge Filter Backwash water Influent Tank Energy Recovery 2 nd Pass NF (South Train) 2 nd Pass NF (North Train) 1 st Pass RO or NF (North Train)

20. Water Quality Goal  Total dissolved solids (TDS)  TDS <500 mg/L-secondary WQ standard  Bromide - accelerate disinfectant decay  Bromide <0.4-0.5 mg/L to maintain residual  Boron - toxic to plants at high level  California notification level = 1 mg/L  No “backsliding” of water quality from new source 20

21.  Selecting appropriate base addition location is critical Base Addition Strategy Base Injection Pt Option 1 Base Injection Pt Option 2 More base required to alter pH HIGH potential for fouling Alk = 122 mg/L Ca 2+ = 447 mg/L Alk = 10.4 mg/L Ca 2+ = 11.7 mg/L Less base required to alter pH 97% rejection of Ca 2+. Decreased potential for fouling Pass 1 Pass 2

22. NF 2 vs. RO, Boron 22 CaNL

23. Specific Energy Summary Permeate B <0.8 mg/L 8.0 8.5 9.0 9.5 10.0 10.5 11.0 11.5 12.0 NF2SWRO (2 pass) Specific Energy (kWh/kgal) Maximum Minimum 50% value 75% value 25% value 50 th percentile specific savings = 20%

24. NF 2 vs. RO  Two-pass RO required to meet all water quality objectives  Boron < 1 mg/L  Consistent with USBR DWPR Report 127  NF 2 required less specific energy than RO/NF  NF 2 required 20% less energy (50 th percentile)

25. Mixing Membranes in NF 2 Pass 1 - two stage configuration Stage 1 holds 5 elements/vessel 25 ULP RO Improve flux and water quality by changing membrane types within a vessel

26. NF 2 Optimization Test 26 ConfigurationPressureFlux Prod. TDS kWh/1000 gal Last elem flow Last element pressure Ranking NF-NF-NF-NF-NF 573.974366 13.86 13539.9 NF-NF-NF-NF-ULP5937369914.3213.2536 ULP-NF-NF-NF-NF588.97384314.2212539.9 2 ULP-ULP-NF-NF-NF6047328414.5813.35381 NF-NF-NF-ULP-ULP612.77298714.7914.4499.63 ULP-ULP-ULP-NF-NF6237285715.0413.5535.2 NF-NF-ULP-ULP-ULP642.57233615.5112.6565.5 BW-NF-NF-NF-NF5977388214.4212534.4 BW-BW-NF-NF-NF627.77320015.1513.5522.8 BW-BW-BW-NF-NF667.57251216.1311.9518.2 Source: Trussell, R.S., Sharma, R.R., Trussell, R.R. 2009. Optimization modeling of nanofiltration membranes for seawater desalination: Scale-up from pilot to prototype scale. In AWWA Membrane Technology Conference (Memphis, TN).

27. NF 2 Optimization Test – Energy 27 No UV/ClO 2 UV ClO 2

28. NF 2 Optimization Summary  No clear difference in energy consumption between 5 & 7 elements in series  More ULP membranes in lead position reduced energy consumption 28

29. Cost Analysis  Cost curves  Based on historical information  Cost models  Use location-specific parameters  NF 2 cost model  Based on ADC cost model (funded by Reclamation)  Modified by LBWD, used Prototype data 29

30. NF 2 Cost Model Scenarios  Scenarios tested  Highest overall system recovery  Lowest specific energy (kWh/kgal)  Highest flux (gfd)  Different production rate  50 mgd  5 mgd 30

31. Variables analyzed VariableBaselineVariableEffect Project life30 yrs25 yrsCapital Interest rate5%4%, 6%Capital Membrane life6.5 yrs10 yrsNon-energy O&M Energy cost$0.12/kWh$0.15/kWhEnergy O&M

32. Cost Analysis Two pass RO NFNF 2 Scenario 1Scenario 2Scenario 1Scenario 2Scenario 3 Pass 1 Pass 2 Pass 1 Pass 2 Pass 1 Pass 2 Pass 1 Pass 2 Pass 1 Pass 2 Flux (gfd)6.9115.895.4111.476.6519.556.2915.357.0515.17 Recovery (%)42%82%35%75%46%75%45%78%44%80% Overall recovery (%)34%27%38%41%39% Energy (kWh/kgal)9.31.79.81.36.61.67.02.27.42.3 Optimization parameter Capital, energy Energy Capital 32

33. 50 mgd vs. 5 mgd   Energy O&M independent of size   Capital and Non-Energy O&M   Capacity dependent items Membrane replacements, solid disposal, maintenance, labor Chemical cost   Capital reduction factors varied for 50 mgd vs 5 mgd 20% – process piping, solid disposal, etc. 32% - pumps, chemical systems, etc. 44% – yard piping, site work, etc.   Capacity independent item Permitting - $10M (15% of overall capital cost for 5 mgd, 3% of overall capital cost for 50 mgd)

34. Cost Analysis 34

35. Inflation  Produced water cost can more than double over project life  Historical increase of 3% for goods (www.bls.gov) www.bls.gov  Energy increased by 4% Inflation projection for energy from historical data Inflation projection for energy from historical data 35

36. Cost Analysis  Project life = 30 years  50 mgd RO-NF ~ $3.6 B RO-NF ~ $3.6 B NF 2 ~ $3.2 B NF 2 ~ $3.2 B  5 mgd RO-NF ~ $0.58 B RO-NF ~ $0.58 B NF 2 ~ $0.57 B NF 2 ~ $0.57 B 36

37. Costs  ADC model used  Inputs modified to include research findings  Size of facility has significant impacts  Scaling down from 50 mgd to 5 mgd can increase cost up to 100%  Sensitivity analysis  Membrane life, power, interest rate, project life  Most sensitive to interest rate and power  Cost of desalinated water (2010)  $1,350/AF for NF 2, $1,640/AF for RO/NF (50 mgd)  $2,454/AF for NF 2, $2,496/AF for RO/NF (5 mgd)

38. Next steps ?  Report for USBR in print

39. Research Presentations

40. www.lbwater.org Questions?