AWWA DSS Conference September 2004

1
Modeling Distribution Water Quality and
Validation through Source Tracing

• AWWA DSS Conference ? September 2004
• Alan R. Wong, SFPUC ? Manouchehr Boozarpour,
  SFPUC
• Polly Boissevain, CDM ? Charlotte Smith, CSA ?
  Chi Yu, SFPUC

2
SFPUC Conveyance and Distribution
3
Presentation Outline

• Water Quality Modeling
• Why Source Trace?
• Tracer Basics
• Case Studies
• Wrap Up

4
Key Factors for Modeling Distribution Water
Quality

• Source Water Chemistry
• Water age
• Hydraulic assumptions for reservoirs
• Travel Time
• Flow Velocity
• Source Blending
• Pipe Wall Interactions

20 days after chloramination
5
Water Quality Applications

• Regulatory
• Disinfectant Residuals
• Microorganisms
• Disinfection by-products
• Corrosion
• Daily Operations
• Dirty Water
• Consumer Complaints
• Main Breaks/Leaks

• Contaminant Tracking
• Treatment Event
• Security

6
Tools of the Trade

• Hydraulic Network Models
• e.g. H2ONet, WaterCad
• Reservoir Models
• Theoretical Detention Time Tavg1/Turnover per
  Day
• Maximum Detention Time
• Tmax or T95 Based on
• Computational Fluid Dynamics
• Physical Scale Models

Physical Scale Model of Sunset Reservoir South
Basin, Courtesy of Program Management Bureau,
SFPUC
7
Why Source Trace?

• Distribution tracer studies can help maintain
  current pipe and/or reservoir models, and changes
  due to
• Customer Demands and Diurnal Curves
• Distribution Valving
• Complexities in Reservoir Hydraulics (e.g.
  mixing, detention times, etc.)
• Pipe Wall effects on Hydraulics and Water Quality
  

8
Validate the Model!

• Validating the Model is like completing a Site
  Visit for Design/Construction Projects

Walk the Site!
9
Conducting a Distribution Tracer

• Select Tracer
• Inject Tracer
• Monitor Tracer
• Monitor Operations and collect Data

10
Tracer Basics Chemical Selection

• Selection of Tracer
• Non Reactive
• Cost Effective
• Easy to Measure
• Non-toxic
• Readily available
• May be based on change in Source Water
• gtgte.g Fluoride, Sodium, Calcium Chloride, UV254,
  Chlorine, Chloramine, Conductivity
• Chemical selection and dose may depend on system
  size, monitoring capabilities and regulatory
  constraints. Check your local health department
  for restrictions on use.

11
Tracer Basics - Injection

• Pulse
• Slug Dose
• Instantaneous Application
• Monitor and Calculate Mass
• Step
• Continuous Feed
• Monitor Increase in Concentration
• Can Repeat Tracer after Feeding Chemical

12
Tracer Basics Monitoring Quantity

• How Much Tracer to inject?
• 2-3x background levels
• Beyond error in analytical instrument
• Statistically Significant above background levels
• (Qualitative results may be sufficient)

13
Case Studies
14
Case Study 1 - Chloramine Conversion Objective

• Perform complex operational activities based on
  modeled chloramine arrival at Distribution Point
  of Entry (e.g. chlorine booster station
  shutdown, turnover reservoirs, etc.)

15
Case Study 1 - Chloramine Conversion Tracer Study
Planning

• Tracer Chloramine
• Expected Values
• Background 0.7 mg/L as Free Cl2
• Tracer 3.0 mg/L as Total Cl2
• 4 x background levels
• gt 95th Percentile of background levels
• Measurement
• Colorimeter
• Error - 2
• Frequency Hourly at Distribution Point of Entry
  and downstream at receiving terminal reservoir
  inlet

16
Case Study 1 - Chloramine Conversion Results -
Step Dose Example
17
Case Study 1 - Chloramine Conversion Conclusions
Excellent Modeling Results

• Helped troubleshoot operations

18
Case Study 2 - Treasure IslandObjective

• Verify water age for pressure zones at the far
  end of distribution system

19
Case Study 2 - Treasure IslandModel Results

• Tank Pressure
• Zone Inlet Turnover Outlet Zone
• TI(1MG Tank) 7 days 3-8 days 10-15days 7-15
  days
• YBI(2MG Tank) 10 17-47 27-57 27-57 days
• (based on pipe network model and theoretical
  turnover)
• Will water age goal of 30 days for new
  disinfectant be met?

20
Case Study 2 - Treasure IslandTracer Study
Planning

• Tracer Conductivity
• Due to Change in Source water
• Expected Values
• Background 80 umhos/cm
• Change expected 160 umhos/cm
• 2 x background levels
• gt95th Percentile of background levels
• Measurement
• Error - 2
• Pocket Analyzer
• Frequency Daily at various sites, including tank
  internal (dip)

21
Case Study 2 - Treasure Island Tracer Study
Results

• Zone Modeled Measured
  Measured
• (Tank) (Mains)
• 1MG Tank/TI 10-15 days 26 days 14-26 days
• 2MG Tank/YBI 27-57 31 days 32-37
• Measured water age was higher at TI than
  modeling results due to simplified mixing
  assumptions in reservoirs. Water age expected to
  be much higher than 30 day goal due to other
  operations.

22
Case Study 2 - Treasure Island Conclusions

• Residuals Maintenance in doubt.
• Circulation Improvements Initiated

23
Case Study 3 - Stanford Heights Objective

• Qualitative study of mixing and water age at
  large terminal reservoir.
• Facts
• Stanford Heights Reservoir
• Const. Date 1923
• Capacity 12.9 MG Terminal Storage Reservoir
• Depth 23.6 ft

24
Case Study 3 - Stanford Heights Model Results

• Modeled water age
• Water age, inlet 4 days
• Tavg 4.6 days
• Tmax 9.2 days (2.0 factor)
• (Based on CFD modeling)
• Total 8.6 13.2 days

• Initial Evaluation of Water age appeared good
  based on theoretical turnover.

25
Case Study 3 - Stanford Heights Tracer Study
Planning

• Tracer Free Chlorine
• Expected Values
• Background 0.10 mg/L
• (measured at end of fill cycle)
• Tracer 0.5 mg/l (Influent Stream)
• 0.20 mg/L (Fully mixed)
• 2 - 5x background levels
• Measurement
• Colorimeter
• Error - 2
• Frequency Hourly at 15 monitoring locations
  within reservoir (5 sites, 3 depths each), inlet
  and outlet
• Duration 7 hour fill cycle.

26
Case Study 3 - Stanford Heights Results
Chlorine Residuals (Bottom)
Legend Red lt0.10 mg/L Yellow 0.11 0.2
mg/L Green 0.21 0.3 mg/L Black gt0.3 mg/L
Warning! Mixing only Near Inlet
27
Case Study 3 - Stanford Heights Results
Chlorine Residuals (Top)
Legend Red lt0.10 mg/L Yellow 0.11 0.2
mg/L Green 0.21 0.3 mg/L Black gt0.3 mg/L
Warning Uneven mixing from top to bottom!
28
Case Study 3 - Stanford Heights Results
Warning! Severe Short-Circuiting
29
Case Study 3 - Stanford Heights Investigation
Inlet/Outlet Impacts

• Clearwell, designed to split flow into two
  basins, suspected of breaking inlet velocity and
  causing short-circuiting

• Footing for divider wall also diverted inflow to
  top of reservoir

30
Case Study 3 - Stanford Heights Conclusions

• Placed reservoir on watch list for
  nitrification. Recent chloramine residuals
  reflect water age greater than 9-13 days
  (completely mixed assumption).

31
Conclusions
32
Conclusions

• Distribution tracer studies may be accomplished
  during chemical target changes, source water
  changes, and reservoir fill cycles.
• Tracer verification for high profile, modeled
  conveyance operations is recommended.
• Be wary of simplified reservoir modeling
  assumptions (e.g. completely mixed reservoirs)!

33
The Bottom Line

• Modeling can be invaluable for distribution
  programs
• Capital Improvements
• Master Planning
• Operations
• Regulatory compliance
• (e.g. IDSE)
• Identify actual field conditions to avoid costly
  mistakes and errors due to modeling errors or
  wrong assumptions!

34
Acknowledgements

• SFPUC Divisions
• Water Quality Bureau
• City Distribution Division
• CDD Dive Team
• Program Management Bureau
• San Francisco Water Team
• CDM, Inc.
• Charlotte Smith Associates

• Resources
• Tracer Studies in Water Treatment Facilities A
  Protocol and Case Studies, AWWARF 1996
• Water Quality Modeling of Distribution System
  Storage Facilities, AWWARF 2000
• Advanced Water Distribution Modeling and
  Management, Haestad Methods 2003