3-EPA
     United States
     Environmental Protection
     Agency
ULTRAVIOLET DISINFECTION GUIDANCE MANUAL
FOR THE FINAL LONG TERM 2 ENHANCED
SURFACE WATER TREATMENT RULE
Office of Water (4601)
EPA815-R-06-007
November 2006

-------
    U.S. Environmental Protection Agency
           Off ice of Water (4601)
       1200 Pennsylvania Avenue NW
           Washington, DC 20460
             EPA815-R-06-007
http://www.epa.gov/safewater/disinfection/lt2/compliance.html
              November 2006

-------
                                       Purpose:

The purpose of this guidance manual is solely to provide technical information on the application
of ultraviolet light for the disinfection of drinking water by public water systems.  This guidance
is not a substitute for applicable legal requirements, nor is it a regulation itself. Thus, it does not
impose legally-binding requirements on any party, including EPA, states, or the regulated
community.  Interested parties are free to raise questions and objections to the guidance and the
appropriateness of using it in a particular situation.  Although this manual covers many aspects
of implementing a UV disinfection system, it is not comprehensive in terms of all types of UV
systems, design alternatives, and validation protocols that may provide satisfactory performance.
The mention of trade names or commercial products does not constitute endorsement or
recommendation for use.

Authorship:

This manual was developed under the direction of EPA's Office of Water, and was prepared by
Malcolm Pirnie, Inc., Carollo Engineers, P.C.,  The Cadmus Group, Inc., Dr. Karl G. Linden, and
Dr. James P. Malley, Jr. Questions concerning this document should be addressed to:

      Dan Schmelling
      U.S. Environmental Protection Agency
      Mail Code 4607M
       1200 Pennsylvania Avenue NW
      Washington, DC 20460-0001
      Tel: (202) 564-5281
      Fax: (202) 564-3767
      Email: schmelling.dan@epa.gov

Acknowledgements:

This document was prepared by the United States Environmental Protection Agency, Office of
Water.  The Work Assignment Manager was Daniel Schmelling, and the Contract Project Officer
was Jane Holtorf. Technical consultants played a significant role in the development of this
document. The work was conducted jointly by Malcolm Pirnie, Inc and Carollo Engineers, P.C.
under contract with The Cadmus Group, Inc. The primary contributors to the document included:

   .  Daniel Schmelling (USEPA)
   .  Christine Cotton and Douglas Owen (Malcolm Pirnie)
   .  Erin Mackey and Harold Wright (Carollo Engineers)
   .  Karl Linden (Duke University)
   .  James Malley, Jr. (University of New Hampshire)

Other contributors included Laurel Passantino, Amy Samuelson, Matthew Yonkin, and James
Collins (Malcolm Pirnie); Dennis Greene, Trinia Dzurny and Stuart Hurley (Carollo Engineers);
Laura Dufresne (The Cadmus Group, Inc.); and Maureen Donnelly and James Melstad
(previously of The Cadmus Group, Inc.).

The following provided valuable technical information to assist in the development of this
manual:
UV Disinfection Guidance Manual              3                                  November 2006
For the Final LT2ESWTR

-------
       American Water Works Association
       Jim Bolton (Executive Director, International Ultraviolet Association)
       Calgon Carbon Corporation
       David Cornwell (Environmental Engineering and Technology)
       Steven Deem (Washington Department of Health)
       Joel DuCoste (North Carolina State University)
       Tom Hargy (Clancy Environmental Consultants)
       OlufHoyer(DVGW)
       Richard Hubel (West Yost & Associates)
       Chris McMeen (then with Washington Department of Health)
       Alexander Mofidi (Metropolitan Water District of Southern California)
       Mike Monty sko (State of New York)
       Issam Najm (Water Quality and Treatment Solutions, Inc.)
       Ondeo Degremont
       Jeffrey Rosen (Clancy Environmental Consultants, Inc.)
       Richard Sakaji (California Department of Health Services)
       Severn Trent Services
       Regina Sommer (University of Vienna)
       Paul Swaim (CH2M Hill)
       Trojan Technologies, Inc.
       Matthew Valade (Hazen and  Sawyer)
       WEDECO UV Technologies, Inc.
UV Disinfection Guidance Manual             4                                  November 2006
For the Final LT2ESWTR

-------
                               Table of Contents

Table of Contents	i
List of Figures	vii
List of Tables	ix
List of Examples	x
List of Checklists	x
Glossary	xi
List of Units, Abbreviations, and Acronyms	xxi

1. Introduction  	1-1
    1.1   Guidance Manual Objectives	1-2
    1.2   Organization	1-2
    1.3   Regulations Summary	1-3
       1.3.1     Filtered PWSs	1-4
       1.3.2     Unfiltered PWSs	1-6
       1.3.3     PWSs with Uncovered Finished Water Storage Facilities	1-6
    1.4   UV Disinfection Requirements for Filtered and Unfiltered PWSs	1-6
       1.4.1     UVDose and Validation Testing Requirements	1-6
       1.4.2     UV Disinfection Monitoring Requirements [40 CFR 141.720(d)(3)(i)]	1-8
       1.4.3     UV Disinfection Reporting Requirements [40 CFR 141.721(f)(15])	1-8
       1.4.4     Off-specification Operational Requirement for Filtered and Unfiltered Systems
                [40 CFR 141.720(d)(3)(ii)]	1-8
    1.5   Regulations Timeline	1-9
    1.6   Alternative Approaches for Disinfecting with UV Light	1-9

2. Overview of UV  Disinfection	2-1
    2.1   History of  UV Light for Drinking Water Disinfection	2-1
    2.2   UV Light Generation and Transmission	2-2
       2.2.1     Nature of UV Light	2-2
       2.2.2     Propagation of UV Light	2-3
    2.3   Microbial Response to UV Light	2-6
       2.3.1     Mechanisms of Microbial Inactivation by UV Light	2-6
       2.3.2     Microbial Repair	2-8
         2.3.2.1  Photorepair	2-8
         2.3.2.2  Dark  Repair	2-9
       2.3.3     UV Intensity, UV Dose, and UV Dose Distribution	2-9
       2.3.4     Microbial Response (UV Dose-Response)	2-11
       2.3.5     Microbial Spectral Response	2-13
    2.4   UV Disinfection Equipment	2-14
       2.4.1     UV Reactor Configuration	2-15
       2.4.2     UV Lamps	2-16
         2.4.2.1  Lamp Start-up	2-19
         2.4.2.2  Lamp Output	2-19
         2.4.2.3  Lamp Sensitivity to Power Quality	2-21
         2.4.2.4  Lamp Aging	2-22
       2.4.3     Ballasts	2-24
       2.4.4     Lamp Sleeves	2-25

UV Disinfection Guidance Manual              i                                   November 2006
For the Final LT2ESWTR

-------
                                 Table of Contents (Continued)
       2.4.5     Cleaning Systems	2-27
       2.4.6     UV Sensors	2-28
       2.4.7     UVT Analyzers	2-29
       2.4.8     Temperature Sensors	2-30
       2.4.9     UV Reactor Dose-Monitoring Strategy	2-31
   2.5   Water Quality Effects and Byproduct Formation	2-31
       2.5.1     Effect of Water Quality on UV Reactor Performance	2-32
         2.5.1.1 UVT	2-32
         2.5.1.2 Particle Content	2-32
         2.5.1.3 Upstream Water Treatment Processes	2-32
         2.5.1.4 Fouling Potential	2-34
         2.5.1.5 Algal Occurrence and Growth	2-36
       2.5.2     Chlorine Reduction through UV Reactors	2-36
       2.5.3     Byproducts from UV Disinfection	2-37
         2.5.3.1 Trihalomethanes, Haloacetic Acids, and Total Organic Halides	2-37
         2.5.3.2 Biodegradable Compounds	2-37
         2.5.3.3 Nitrite	2-37

3. Planning Analyses for UV Facilities	3-1
   3.1   UV Disinfection Goals	3-1
   3.2   Evaluating Integration of UV Disinfection into the Treatment Process	3-3
       3.2.1     UV Disinfection Effects on Treatment	3-3
       3.2.2     Upstream Treatment Process Effect on UV Disinfection	3-4
   3.3   Identifying Potential Locations for UV Facilities	3-5
       3.3.1     Installation Locations for Filtered Systems	3-5
         3.3.1.1 Combined Filter Effluent Installation (Upstream of the Clearwell)	3-5
         3.3.1.2 Individual Filter Effluent Piping Installation	3-6
         3.3.1.3 UV Disinfection Downstream of the Clearwell	3-8
       3.3.2     Unfiltered System Installation Locations	3-9
       3.3.3     Groundwater System Installation Locations	3-10
       3.3.4     Uncovered Reservoir Installation Locations	3-10
   3.4   Defining Key Design Parameters	3-10
       3.4.1     Off-specification Requirements	3-10
       3.4.2     Target Pathogen Inactivation and Required UV Dose	3-12
       3.4.3     Design Flow Rate	3-12
       3.4.4     Water Quality	3-12
         3.4.4.1 UVT and UVT Scans	3-13
         3.4.4.2 Water Quality Parameters That Affect Fouling	3-18
         3.4.4.3 Additional Water Quality Considerations for Unfiltered Supplies and
                Treatment of Uncovered Reservoir Water	3-19
       3.4.5     Fouling/Aging Factor	3-21
         3.4.5.1 Testing to Determine the Fouling Factor	3-22
         3.4.5.2 Testing to Determine the Aging Factor	3-23
       3.4.6     Power Quality Evaluations	3-25
   3.5   Evaluating UV Reactors, Dose Monitoring Strategy, and Operational Approach	3-26
       3.5.1     Characteristics of LPHO and MP Reactors	3-26
   3.5.2 Dose-monitoring Strategy and Operational Approach	3-27
       3.5.2.1   UV Intensity Setpoint Approach	3-28

UV Disinfection Guidance Manual               ii                                   November 2006
For the Final LT2ESWTR

-------
                                 Table of Contents (Continued)
       3.5.2.2   Calculated Dose Approach	3-29
       3.5.2.3   Advantages and Disadvantages	3-30
   3.6   Assessing UV Equipment Validation Issues	3-31
       3.6.1     Off-site Versus On-site Validation	3-31
       3.6.2     Validation and Installation Hydraulics Recommendations	3-32
       3.6.3     Selection of Validation and Hydraulic Approach	3-33
   3.7   Assessing Head Loss Constraints	3-35
       3.7.1     Eliminating Existing Hydraulic Inefficiencies	3-35
       3.7.2     Modifying Clearwell Operation	3-35
       3.7.3     Modifying Filter Operation	3-36
       3.7.4     Installing Intermediate Booster Pumps	3-36
       3.7.5     Modifying Operation of HSPs	3-36
   3.8   Estimating UV Facility Footprint	3-36
       3.8.1     UV Equipment Constraints	3-36
       3.8.2     Develop UV Facility Layout	3-37
   3.9   Preparing Preliminary Costs and Selecting the UV Facility Option	3-38
   3.10  Reporting to the State	3-38

4. Design Considerations for UV Facilities	4-1
   4.1   UV Facility Hydraulics	4-2
       4.1.1     Inlet and Outlet Piping Configuration	4-3
       4.1.2     Flow Distribution, Control, and Measurement	4-3
         4.1.2.1 Flow Distribution and Control	4-3
         4.1.2.2 Flow Rate Measurement	4-3
       4.1.3     Water Level Control	4-5
       4.1.4     Air Relief and Pressure Control Valves	4-6
       4.1.5     Flow Control and Isolation Valves	4-6
       4.1.6     Installation of Intermediate Booster Pumps	4-6
       4.1.7     Evaluating Existing Pumps and Potential Water Hammer Issues	4-7
       4.1.8     Groundwater System Hydraulic Issues	4-7
       4.1.9     Uncovered Finished Water Reservoir Hydraulic Issues	4-8
   4.2   Operating Approach Selection	4-8
   4.3   Instrumentation and Control	4-9
       4.3.1     UV Reactor Start-up and Sequencing	4-9
         4.3.1.1 UV Reactor Start-up	4-10
         4.3.1.2 UV Reactor Sequencing	4-10
         4.3.1.3 Groundwater Pump Cycling Effects on UV Reactor Start-up	4-11
       4.3.2     UV Equipment Automation	4-11
       4.3.3     Alarms and Control Systems Interlocks	4-12
       4.3.4     UV Reactor Control Signals	4-12
         4.3.4.1 UV Intensity	4-13
         4.3.4.2 UV Transmittance	4-13
         4.3.4.3 Flow Rate Measurement	4-14
         4.3.4.4 Calculated and Validated UV Dose (If Applicable)	4-14
         4.3.4.5 Operational Setpoints	4-14
         4.3.4.6 Lamp Age	4-14
         4.3.4.7 Lamp Power, Lamp Status, and Reactor Status	4-14
         4.3.4.8 UV Reactor Sleeve Cleaning	4-14

UV Disinfection Guidance Manual              iii                                   November 2006
For the Final LT2ESWTR

-------
                                Table of Contents (Continued)
        4.3.4.9  Alarms	4-15
   4.4  Electrical Power Configuration and Back-up Power	4-15
       4.4.1     Considerations for Electrical Power	4-15
       4.4.2     Back-up Power Supply and Power Conditioning	4-15
        4.4.2.1  Back-up Power Supply	4-16
        4.4.2.2  Power Conditioning Equipment	4-16
       4.4.3     Ground Fault Interrupt and Electrical Lockout	4-17
   4.5  UV Facility Layout	4-17
       4.5.1     Additional Considerations for Unfiltered and Uncovered Reservoir UV
                Facility Layouts	4-18
       4.5.2     Additional Considerations for Groundwater UV Facility Layouts	4-19
   4.6  Elements of UV Equipment Specifications	4-19
       4.6.1     UV Equipment  Specification Components	4-19
       4.6.2     Information Provided by Manufacturer in UV Reactor Bid	4-22
   4.7  Final UV Facility Design	4-23
   4.8  Reporting to the State during Design	4-24
5. Validation of UV Reactors	5-1
   5.1  Minimum Requirements for Validation Testing	5-2
   5.2  Overview of the Recommended Validation Protocol	5-3
       5.2.1     Key Steps in Recommended Validation Protocol	5-3
       5.2.2     Alternative Validation Protocols	5-6
       5.2.3     Third Party Oversight	5-6
       5.2.4     Emerging Methods	5-7
   5.3  Selecting the Challenge Microorganism	5-7
   5.4  Equipment Needs for Full-scale Reactor Testing	5-9
       5.4.1     Water Source	5-9
       5.4.2     UV Absorbing Chemical	5-10
       5.4.3     Mixing	5-10
       5.4.4     Sampling Ports	5-11
       5.4.5     Configuration of Inlet and Outlet Piping	5-12
       5.4.6     Accounting for Non-uniform Lamp Aging	5-12
       5.4.7     Lamp Positioning to Address Lamp Variability	5-14
       5.4.8     UV Sensors	5-14
       5.4.9     UV Sensor Port Windows	5-15
   5.5  Accuracy of Measurement Equipment	5-15
       5.5.1     Flow Meters	5-18
       5.5.2     UV Spectrophotometers	5-18
       5.5.3     Power Measurements	5-19
       5.5.4     UV Sensors	5-19
   5.6  Identifying Test Conditions	5-20
       5.6.1     Test Conditions for the UV Intensity Setpoint Approach	5-22
       5.6.2     Test Conditions for the Calculated Dose Approach	5-23
       5.6.3     Test Conditions for Confirming an Existing Validation Equation	5-26
       5.6.4     Quality-control Samples	5-26
   5.7  Guidelines for Conducting Experimental Tests	5-27
       5.7.1     Preparing the Challenge Microorganism	5-27

UV Disinfection Guidance Manual              iv                                 November 2006
For the Final LT2ESWTR

-------
                                Table of Contents (Continued)
       5.7.2     Full-scale UV Reactor Testing	5-28
       5.7.3     Collimated Beam Testing	5-30
   5.8  Analyzing Experimental Data	5-30
       5.8.1     Calculating the Reduction Equivalent Dose (RED)	5-30
       5.8.2     Selecting the Minimum RED for the UV Intensity Setpoint Approach	5-34
       5.8.3     Developing the Dose-monitoring Equation for the Calculated Dose
                Approach	5-34
   5.9  Deriving the Validation Factor (VF)	5-36
       5.9.1     RED Bias Factor	5-37
       5.9.2     Uncertainty in Validation (UVai)	5-38
        5.9.2.1  Calculating USP for the UV Intensity Setpoint Approach	5-41
        5.9.2.2  Calculating UIN for the Calculated Dose Approach	5-41
   5.10 Determining the Validated Dose and Validated Operating Conditions	5-42
       5.10.1    Determining the Validated Dose and Operating Conditions for the UV
                Intensity Setpoint Approach	5-42
       5.10.2    Determining the Validated Dose and Operating Conditions for the Calculated
                Dose Approach	5-43
   5.11 Documentation	5-44
       5.11.1    UV Reactor Documentation	5-44
       5.11.2    Validation Test Plan	5-46
       5.11.3    Validation Report	5-48
   5.12 Guidelines for Reviewing Validation Reports	5-50
   5.13 Evaluating the Need for "Re-validation"	5-54

6. Start-up and Operation of UV Facilities	6-1
   6.1  UV Facility Start-up	6-1
       6.1.1     O&M Manual	6-1
       6.1.2     State Coordination during Start-up	6-3
       6.1.3     Functional Testing	6-3
        6.1.3.1  Verification of UV Equipment Components	6-4
        6.1.3.2  Verification of Instrumentation and Control Systems	6-5
        6.1.3.3  Verification of Flow Distribution and Head Loss	6-6
       6.1.4     Determining Validated Operational Conditions  and Setting Operational
                Controls	6-7
       6.1.5     Performance  Testing	6-10
       6.1.6     Final Inspection	6-12
   6.2  Operation of UV Facilities	6-12
       6.2.1     Operational Requirements	6-13
       6.2.2     Recommended Operational Tasks	6-13
       6.2.3     Start-up and Shut-down of UV Reactors	6-14
        6.2.3.1  Routine Start-up	6-14
        6.2.3.2  Start-up Following Maintenance	6-14
        6.2.3.3  Routine Shut-down	6-15
        6.2.3.4  Shut-down Prior to Maintenance	6-15
        6.2.3.5  Winterization	6-15
   6.3  Maintenance of UV Reactors	6-16
       6.3.1     Summary of Recommended Maintenance  Tasks	6-16
       6.3.2     General Guidelines for UV Reactor Maintenance	6-18

UV Disinfection Guidance Manual             v                                   November 2006
For the Final LT2ESWTR

-------
                                Table of Contents (Continued)
         6.3.2.1  Fouling	6-18
         6.3.2.2   Lamp Output Variability	6-19
         6.3.2.3  Reference UV Sensors	6-20
         6.3.2.4  Electrical Concerns	6-20
         6.3.2.5  UV Reactor Temperature and Water Level	6-21
         6.3.2.6  UV Lamp Replacement	6-21
         6.3.2.7  Lamp Sleeves	6-21
         6.3.2.8  On-line UVT Analyzer	6-22
       6.3.3     Spare Parts	6-22
   6.4   Monitoring and Recording of UV Facility Operation	6-24
       6.4.1     Monitoring and Recording for Compliance Parameters	6-24
         6.4.1.1  Monitoring of Duty UV Sensor Calibration	6-24
         6.4.1.2  Monitoring of UVT Analyzer Calibration	6-29
         6.4.1.3  Off-specification Events	6-31
         6.4.1.4  Monitoring and Recording Frequency of Required Parameters	6-32
       6.4.2     Monitoring and Recording for Operational Parameters Not Related to
                Compliance	6-36
   6.5   UV Facility Reporting to the State	6-37
       6.5.1     Required Reporting	6-38
       6.5.2     Example Reporting Forms and Calculation Worksheets	6-39
   6.6   Operational Challenges	6-48
       6.6.1     LowUV Intensity or Validated Dose Below the Setpoint	6-48
       6.6.2     Low UV Transmittance	6-48
       6.6.3     Failure to Meet UV Sensor Calibration Criterion	6-50
   6.7   Staffing, Training, and Safety Issues	6-50
       6.7.1     Staffing Levels	6-53
       6.7.2     Training	6-53
       6.7.3     Safety Issues	6-53

7. Bibliography	7-1
Appendices

Appendix A.     Preparing and Assaying Challenge Microorganisms	A-1
Appendix B.     UV Reactor Testing Examples	B-l
Appendix C.     Collimated Beam Testing to Develop a UV Dose-response Curve	C-l
Appendix D.     Background to the UV Reactor Validation Protocol 	D-l
Appendix E.     UVLamp Break Issues	E-l
Appendix F.     Case Studies	F-l
Appendix G.     Reduction Equivalent Dose Bias Tables	G-l
UV Disinfection Guidance Manual               vi                                  November 2006
For the Final LT2ESWTR

-------
                                 List of Figures

Figure 1.1. LT2ESWTR Compliance Timeline for Initial Source Water Monitoring and
Treatment Installation	1-10

Figure 2.1. UV Light in the Electromagnetic Spectrum	2-2
Figure 2.2. Refraction of Light	2-4
Figure 2.3. Reflection of Light	2-4
Figure 2.4. Scattering of Light	2-5
Figure 2.5. Structure of DNA and Nucleotide Sequences within DNA	2-7
Figure 2.6. UV Absorbance of Nucleotides (left) and Nucleic Acid (right) at pH 7	2-7
Figure 2.7. Hypothetical Dose Distributions for Two Reactors with Differing Hydraulics	2-10
Figure 2.8. Shapes of UV Dose-Response Curves	2-12
Figure 2.9. Comparison of Microbial UV Action and DNA UV Absorbance	2-14
Figure 2.10. Example of UV Disinfect!on Equipment	2-15
Figure 2.11. Examples of UV Reactors: (a) Closed-channel and (b) Open-channel	2-15
Figure 2.12. Construction  of a UV Lamp	2-19
Figure 2.13. UV Output of LP (a) and MP(b) Mercury Vapor Lamps	2-20
Figure 2.14. UV Lamp Output and its Relationship to the UV Absorbance of DNA	2-21
Figure 2.15. Reduction in UV Output of (a) LPHO and (b) MP Lamps Over Time	2-23
Figure 2.16. Lamp Aging for an MP Lamp	2-23
Figure 2.17. Aged UV Lamp (right) Compared to a New UV Lamp (left)	2-24
Figure 2.18. UVT of Quartz that is  1 mm Thick at a Zero-degree Incidence Angle	2-26
Figure 2.19. Examples of (a) Mechanical Wiper System and (b) Mechanical-chemical Wiper
            System	2-28
Figure 2.20. Example of a Dry UV Sensor Mounted on a UV Reactor	2-29
Figure 2.21. Example UVT Analyzer Design	2-30
Figure 2.22. Example Effect of Ozonation on UV Absorbance if Ozone is Quenched Prior to UV
            Disinfection	2-33
Figure 2.23. Example Effect of UV Disinfect! on on Free Chlorine Residual Loss	2-36

Figure 3.1. Example Flowchart for Planning UV Facilities	3-2
Figure 3.2. Schematic for UV Facility Upstream of the Clearwell	3-5
Figure 3.3. Schematic of Individual Filter Effluent Piping Installation in Filter Gallery	3-6
Figure 3.4. UV Disinfection Downstream of High Service Pumps	3-8
Figure 3.5. Example Cumulative Frequency Diagram for Three Filtered Waters	3-16
Figure 3.6. Example Flow Rate and UVT (at 254 nm) Data	3-17
Figure 3.7. Schematic of Hydraulic Option #1 (90°-Bend,  T-Bend, S-Bend Inlet Piping
          Scenarios)	3-33
Figure 3.8. UV Reactor Validation Options and How They Affect Installation Hydraulics	3-34

Figure 4.1. Flowchart for Planning, Design, and Construction of UV Facilities	4-2
Figure 4.2. Open-channel Flow Distribution Options	4-4

Figure 5.1. Overview of Recommended Validation Protocol	5-4
Figure 5.2. Block Diagram of a Typical Biodosimetry Test Stand for Full-scale Reactor
          Testing	5-10
UV Disinfection Guidance Manual              vii                                November 2006
For the Final LT2ESWTR

-------
                                 List of Figures (Continued)
Figure 5.3. Hypothetical Examples of the Spectral Response of a Germicidal and a
          Non-germicidal UV Sensor	5-18
Figure 5.4. UVAL Decision Tree for the UV Intensity Setpoint Approach	5-39
Figure 5.5. UVAL Decision Tree for the Calculated Dose Approach	5-40
Figure 5.6 UVT of Standard and "Ozone-Free" Quartz Assuming Air-Quartz and Quartz-Water
          Interfaces	5-55

Figure 6.1. Start-up and Operation Flowchart	6-2
Figure 6.2. Example Summary Monthly Report	6-41
Figure 6.3. Example Daily Operating Log for Calculated Dose Approach	6-42
Figure 6.4. Example Daily Operating Log for UV Intensity Setpoint Approach	6-43
Figure 6.5. Example Off-specification Calculation Worksheet	6-44
Figure 6.6. Example Monthly UV Sensor Calibration Check Log	6-45
Figure 6.7. Example UV Sensor CF Calculation Worksheet	6-46
Figure 6.8. Example Monthly UVT Analyzer Calibration Check Log	6-47
Figure 6.9. LowUV Intensity or Low Validated Dose Decision Chart	6-49
Figure 6.10. LowUV Transmittance Decision Chart	6-51
Figure 6.11. Monitoring of UV Sensor Calibration Flowchart	6-52
UV Disinfection Guidance Manual              viii                                November 2006
For the Final LT2ESWTR

-------
                                  List of Tables

Table 1.1. Summary of Microbial and DBF Rules	1-4
Table 1.2. Bin Requirements for Filtered PWSs	1-5
Table 1.3. Requirements for Unfiltered PWSs	1-6
Table 1.4. UV Dose Requirements - millijoules per centimeter squared (ml/cm2) 	1-7

Table 2.1. Typical Mercury Vapor Lamp Characteristics	2-17
Table 2.2. Mercury Vapor Lamp Operational Advantages	2-17
Table 2.3. Typical Start-up and Restart Times for LPHO and MP Lamps	2-22
Table 2.4. Comparison of Magnetic and Electronic Ballasts	2-25
Table 2.5. UV Absorbance Characteristics of Common Water Treatment Chemicals	2-34

Table 3.1. Potential Method to Determine Design Flow	3-12
Table 3.2. Summary of UVT Data Collection and Analysis	3-14
Table 3.3. Summary of Fouling Data Collection and Analysis	3-19
Table 3.4. Summary of Particle and Algal Data Collection and Analysis	3-20
Table 3.5. Power Quality Triggers for UV Reactors	3-25
Table 3.6. Dose-monitoring Approaches - Key Characteristics	3-28
Table 3.7. Advantages and Disadvantages of Single-setpoint and Variable-setpoint Operations
          for the UV Intensity Setpoint Approach	3-29
Table 3.8. Advantages and Disadvantages of Off-site and On-site Validation	3-31

Table 4.1. Comparison of Techniques for UV Facility Flow Rate Measurement for Combined
          Filter Effluent and Post-clearwell UV Facilities	4-5
Table 4.2. Typical Alarm Conditions for UV Reactors	4-13
Table 4.3. Possible Content for UV Equipment Specifications	4-20
Table 4.4. Suggested Information to  Be Provided by UV Manufacturer	4-22

Table 5.1. Summary of LT2ESWTR Validation Requirements	5-2
Table 5.2. UV Sensitivity of Challenge Microorganisms	5-8
Table 5.3. Hypothetical Example of the Spectral Response of a Germicidal UV Sensor	5-16
Table 5.4. Hypothetical Example of the Spectral Response of a Non-germicidal UV Sensor.. 5-17
Table 5.5. Factors to be Considered in Validation Test Design	5-21
Table 5.6. Minimum Test Conditions for the UV  Intensity Setpoint Approach	5-23
Table 5.7. Minimum Test Conditions for the Calculated Dose Approach	5-25

Table 6.1. Functional  Testing of Cleaning Systems	6-5
Table 6.2. Example Monitoring During a Four Week Performance Test	6-12
Table 6.3. Recommended Operational Tasks for the UV Reactor	6-13
Table 6.4. Recommended Maintenance  Tasks	6-16
Table 6.5. Typical Design and Guaranteed Lives  of Major UV Components (Based on
          Manufacturers' Input)	6-23
Table 6.6. Off-specification Examples for Each Monitoring Approach	6-32
Table 6.7. Recommended Recording Frequency for Required Monitoring Parameters	6-33
Table 6.8. Recommended Monitoring Parameters and Recording Frequency	6-37
Table 6.9. Factors Impacting Staffing Needs	6-53
UV Disinfection Guidance Manual              ix                                 November 2006
For the Final LT2ESWTR

-------
                               List of Examples

Example 5.1. Determining Flow Conditions for Validation Testing	5-24
Example 5.2. Calculating RED Using Validation Test Data	5-31
Example 5.3. Validation Testing Using Two Challenge Microorganisms	5-33
Example 5.4. Determining the RED Bias factor	5-38

Example 6.1. Setting Operational Controls for the UV Intensity Setpoint Approach - Single
            Setpoint Operation	6-8
Example 6.2. Setting Operational Controls for the Calculated Dose Approach	6-9
Example 6.3. Duty UV Sensors are Verified using Reference Sensor	6-26
Example 6.4. Duty UV Sensors that Do Not Meet Calibration Criteria	6-28
Example 6.5. UVT Analyzer Calibration Check	6-31
Example 6.6. Routine and Off-specification Recording	6-34
Example 6.7. Off-specification Computation	6-39
                              List of Checklists

Checklist 5.1. UV Reactor Documentation	5-45
Checklist 5.2. Key Elements of the Validation Test Plan	5-47
Checklist 5.3. Key Elements of the Validation Report	5-49
Checklist 5.4. Review for Quality Assurance/Quality Control	5-51
Checklist 5.5. Review for Key Validation Report Elements	5-52
UV Disinfection Guidance Manual              x                                 November 2006
For the Final LT2ESWTR

-------
                                   Glossary
       The following definitions were derived from existing UV literature, standard
physics textbooks, and/or industry standards and conventions. Some concepts have more
than one acceptable term or definition, but for consistency within the document, only one
term is used.

Absorption - the transformation of UV light to other forms of energy as it passes
through a substance.

Ais4 (UV Absorbance at 254 nm)- a measure of the amount of UV light that is absorbed
by a substance at 254 nm.

Action Spectra Correction Factor (CFas) - a correction factor to account for greater
proportional inactivation of a challenge microorganism compared to the target pathogen
that results from differences in action spectra.

Action Spectrum - the relative efficiency of UV energy frequencies at inactivating
microorganisms. Each microorganism has a unique action spectrum.

Bacteriophage - a virus that infects bacterial cells and can be used a microbial surrogate
during validation testing.

Ballast - an electrical device that provide the proper voltage and current required to
initiate and maintain the gas discharge within the UV lamp.

Beer's Law -an empirical equation describing the absorption of light as a function of the
transmitting medium's properties; also know as the Beer-Lambert law.

Bioassay - in the context of this document, an empirical assessment of the inactivation
response of a specific microorganism to a controlled dose of UV light, usually in UV
reactors.  Bioassay has been used in the UV disinfection literature in the same context as
"biodosimetry" (see Biodosimetry).

Biodosimetry - a procedure used to determine the reduction equivalent dose (RED) of a
UV reactor. Biodosimetry involves measuring the inactivation of a challenge
microorganism after exposure to UV light in a UV reactor and comparing the results to
the known UV dose-response curve of the challenge microorganism (determined via
bench-scale collimated beam testing).

Calculated Dose Approach - See Dose-monitoring Strategy

Challenge Microorganism - a non-pathogenic microorganism used in validation testing
of UV reactors.
UV Disinfection Guidance Manual              xi                          November 2006
For the Final LT2ESWTR

-------
                                    Glossary (Continued)
Collimated Beam Test - a controlled bench-scale test that is used to determine the UV
dose-response of a challenge microorganism. Both time and UV light intensity are
directly measured; the UV dose is calculated using the intensity of the incident UV light,
UV absorbance of the water, and exposure time.

Dark Repair - an enzyme-mediated microbial  process that removes and regenerates a
damaged section  of deoxyribonucleic  acid (DNA),  using an existing  complimentary
strand  of DNA.  Dark  repair  refers to all  microbial repair processes not  requiring
reactivating light.

Design Flow Rate - the maximum flow that can be treated at the UV facility.  See
Section 3.4 for potential methods for determining design flow rate.

Design UVT - The minimum UVT that will typically occur at the design flow of the UV
facility. The design UVT and design flow are typically used by the UV manufacturer to
determine the appropriate UV equipment for a target pathogen inactivation. The design
UVT may not necessarily be the minimum operating UVT (see Minimum Operating
UVT).

Diffuse Reflection - that portion of light reflected by a rough surface that radiates in all
directions.

Dose Distribution - see UV Dose.

Dose-monitoring Strategy - the method by which a UV reactor maintains the required
dose at or near some specified value by monitoring UV dose delivery.  Such strategies
must include, at a minimum, flow rate and UV intensity (measured via duty UV
sensor[s]) and lamp status.  They sometimes include UVT and lamp power. Two
common Dose-monitoring Strategies that are discussed in this manual are the UV
Intensity Setpoint Approach and the Calculated Dose Approach.

   •  The UV Intensity Setpoint Approach relies on one or more "setpoints" for UV
       intensity that are established during validation testing to determine UV dose.
       During operations, the UV intensity as measured by the UV sensors must meet or
       exceed the setpoint(s) to ensure delivery of the required dose. Reactors must also
       be operated within validated operation conditions for flow rates and lamp status
       [40  CFR 141.720(d)(2)]. In the UV Intensity Setpoint Approach, UVT does not
       need to be monitored separately. Instead, the intensity readings by the sensors
       account  for changes in UVT. The operating strategy can be with either a single
       setpoint (one UV intensity setpoint is used for all validated flow rates) or a
       variable setpoint (the UV intensity setpoint is determined using a lookup table or
       equation for a range of flow rates).

   •  The Calculated Dose Approach uses a dose-monitoring equation to estimate the
       UV dose based on operating conditions (typically flow rate, UV intensity, and
       UVT). The dose-monitoring equation may be developed by the UV manufacturers
UV Disinfection Guidance Manual               xii                         November 2006
For the Final LT2ESWTR

-------
                                    Glossary (Continued)
       using numerical methods; however, EPA recommends that systems use an
       empirical dose-monitoring equation developed through validation testing. During
       reactor operations, the UV reactor control system inputs the measured parameters
       into the dose-monitoring equation to produce a calculated dose. The system
       operator divides the calculated dose by the Validation Factor (see Chapter 5 for
       more details on the Validation Factor) and compares the resulting value to the
       required dose for the target pathogen and log inactivation level.

Dose-pacing Strategy - the method by which a UV reactor maintains the required dose
at or near some specified value that typically involves adjusting the lamp power  or
turning "on" or "off banks of UV lamps or whole UV reactors to respond to changes in
UVT, lamp intensity, or flow rate. A programmable logic controller (PLC) makes
adjustments using an equation(s) developed during the UV reactor validation process.

Duty UV Sensor (or Duty Sensor) - the duty (on-line) UV sensor installed in the UV
reactor that monitors UV intensity during UV equipment operations.

Emission Spectrum - the relative power emitted by a lamp at different wavelengths.

End-of-Lamp Life - The duration of lamp operations after which the lamp should be
replaced

First-order Inactivation - in the context of this document, inactivation of a
microorganism that is  directly proportional to the UV dose.

Fluence - see the definition for UV Dose.

Fluence Rate - see the definition for UV Intensity.

Fouling/Aging Factor - a site-specific factor (the product of a fouling factor and aging
factor) that is used to account for the decline in UV transmittance through the lamp
sleeve due to fouling (e.g., by water quality parameters) and aging of the lamp and lamp
sleeve. The lamp fouling portion of the factor is the estimated fraction of UV light
passing through a fouled sleeve as compared to a new sleeve. The lamp aging portion of
the factor is the fraction of UV light emitted from aged sleeves and lamps compared to
new sleeves and lamps. It can be estimated by the lamp and sleeve aging characteristics
obtained from the UV  manufacturer.

Gas Discharge - a mixture of non-excited atoms, excited atoms, cations, and free
electrons formed when a sufficiently high voltage is applied across a volume of gas.
Most commercial UV  lamps use mercury gas discharges to generate UV light.

Germicidal Effectiveness - the relative inactivation efficiency of each UV wavelength
in an emission spectrum.  This value is usually approximated by the relative absorbance
of DNA at each wavelength.
UV Disinfection Guidance Manual              xiii                        November 2006
For the Final LT2ESWTR

-------
                                    Glossary (Continued)
Germicidal Range - the range of UV wavelengths responsible for microbial inactivation
in water (200 to 300 nm).

Germicidal Sensor - A UV sensor with a spectral response that peaks between 250 and
280 nm and has less than 10 percent of its total measurement due to light above 300 nm
when mounted on the UV reactor and viewing the UV lamps through the water that will
be treated at the water treatment plant.

Inactivation - in the context of UV disinfection, a process by which a microorganism is
rendered unable to reproduce, thereby rendering it unable to infect a host.

Lamp Burn-in - During the first few hours of mercury-vapor lamp operation, output will
diminish rapidly, then stabilize as the impurities within the lamp are burned off. This
initial "burn-in" period is typically assumed to be complete at 100 hours.

Lamp Envelope - the exterior surface of the UV lamp, which is typically made of
quartz.

Lamp Sleeve - the quartz tube or thimble that surrounds and protects the UV lamp. The
exterior is in direct contact with the water being treated. There is typically an air gap
(approximately 1 cm) between the lamp envelope and the quartz sleeve.

Lamp Status - see UV Lamp Status

Light Pipe - a quartz cylinder that transmits light from the interior of the UV reactor to
the photodetector of a UV intensity sensor.

Lignin Sulfonate - a commercially available  liquid lignin mixture (typically procured
from paper mills) used to adjust the UV transmittance of natural waters during validation
testing.

Low-pressure (LP) Lamp - a mercury-vapor lamp that operates at an internal pressure
of 0.13 to 1.3 Pa (2 x 10"5to 2 x 10"4 psi) and electrical input of 0.5 watts per  centimeter
(W/cm).  This results in essentially monochromatic light output at 254 nm.

Low-pressure high-output (LPHO) Lamp - a low-pressure mercury-vapor lamp that
operates under increased electrical input (1.5 to 10 W/cm), resulting in a higher UV
intensity than low-pressure lamps. It also has essentially monochromatic light output at
254 nm.

Medium-pressure (MP) Lamp - a mercury vapor lamp that operates at an internal
pressure of 1.3 and 13,000 Pa (2 to 200 psi) and electrical input of 50 to 150 W/cm. This
results in a  polychromatic (or broad spectrum) output of UV and visible light at multiple
wavelengths, including wavelengths in the germicidal range.
UV Disinfection Guidance Manual              xiv                        November 2006
For the Final LT2ESWTR

-------
                                    Glossary (Continued)
Microbial Repair - enzyme-mediated microbial process where damaged strands of
deoxyribonucleic acid (DNA) are repaired. Energy for this process can be derived by
light energy (photorepair) or chemical energy (dark repair).

Minimum Operating UVT:  The lowest UVT expected to occur during lifetime of the
UV facility. Understanding the minimum UVT is critical because the UV reactor should
be designed and validated for the range of UVT and flow rate combinations expected at
the WTP to avoid off-specification operation.

Monochromatic - light output at only one wavelength,  such as UV light generated by
low-pressure and low-pressure high-output lamps.

Monitoring Window - a quartz disc that transmits light from the interior of the UV
reactor to the photodetector of a UV sensor.

MS-2 Bacteriophage - a non-pathogenic bacteriophage commonly used as a challenge
organism in UV reactor validation testing.

Non-germicidal Sensor - A UV sensor with a spectral response that is not restricted to
the germicidal range (see "Germicidal Sensor" for  more details).

Off-line  Chemical Clean (OCC) - a process to clean lamp sleeves where the UV reactor
is taken off-line and a cleaning solution (typically a weak acid) is sprayed into the reactor
through a service port. After the foulants have dissolved, the reactor is drained, rinsed,
and returned to service. Also called "flush-and-rinse" systems.

Off-specification - A UV facility  that is operating outside of the validated operating
conditions (e.g., at a flow rate higher than the validated range or a UVT below the
validated range).

On-line Mechanical Clean (OMC) - a process to clean lamp sleeves where an
automatic mechanical wiper (e.g.,  O-ring) wipes the surface of the lamp sleeve at a
prescribed frequency.

On-line Mechanical-Chemical Clean (OMCC) - a process to clean lamp sleeves where
an automatic mechanical wiper (e.g., O-ring) with a chemical solution located within the
cleaning  mechanism wipes the surface of the lamp  sleeve at a prescribed frequency.

Operating Strategy - the strategy used by the PWS to operate the UV equipment with
the UV Intensity Setpoint Approach. Typically, single setpoint or variable setpoint
operation is used.

Petri Factor - a ratio used in collimated beam testing that is equal to the average
intensity  measured across the surface of a suspension in a petri dish divided by the
intensity  at the center of a petri dish.
UV Disinfection Guidance Manual               xv                         November 2006
For the Final LT2ESWTR

-------
                                    Glossary (Continued)
Photodetector - a device that produces an electrical current proportional to the UV light
intensity at the detector's surface.

Photorepair- a microbial repair process where enzymes are activated by light in the near
UV and visible range, thereby repairing UV induced damage. Photoreactivation requires
the presence of light.

Polychromatic - light energy output at several wavelengths such as with MP lamps.

Polychromatic Bias - a potential bias in validation test data resulting from
polychromatic differences between validation and operation of a UV reactor at a water
system. Polychromatic bias can occur in MP reactors when non-germicidal sensors are
used.

Quartz Sleeve - see lamp sleeve.

Radiometer - an instrument used to measure UV irradiance.

Rayleigh Scattering - light scattering by particles smaller than the wavelength of the
light.

Reduction Equivalent Dose (RED) - see UV Dose.

Reduction Equivalent Dose (RED) Bias - a correction that accounts for the difference
between the UV dose measured with a surrogate microorganism and the UV dose that
would be delivered to a target pathogen due to differences in the microorganisms'
inactivation kinetics.

Reference UV Sensor (or Reference Sensor) - a calibrated, off-line UV sensor used to
monitor duty UV sensor calibration and to determine UV sensor uncertainty.

Required Dose - the UV dose required for a certain level  of log inactivation.  Required
doses are set forth by the LT2ESWTR.

Sensor Correction Factor - a correction factor that may need to be temporarily applied
during operations when duty sensor(s) fail a calibration check and can not be
immediately replaced. The sensor correction factor allows the UV facility  to remain in
operation while the problem is resolved.

Setpoint (also called "operational setpoint") - a specific value for a critical parameter,
such as UV intensity, that is related to UV dose. Setpoints are established  during
validation testing.  During operations, the PWS compares the measured parameter to the
setpoint to confirm performance.

Solarization - a change in the structure of a material due to exposure to UV light that
increases light scattering and attenuation.
UV Disinfection Guidance Manual              xvi                         November 2006
For the Final LT2ESWTR

-------
                                    Glossary (Continued)
Spectral Response - A measure of the output of the UV sensor as a function of
wavelength.

State - the agency of the state or Tribal government that has jurisdiction over public
water systems. During any period when a state or Tribal government does not have
primary enforcement responsibility pursuant to section 1413 of the Act, the term "state"
means the Regional Administrator, U.S. Environmental Protection Agency.

Subpart H Systems - public water systems using surface water or ground water under
the direct influence of surface water as a source that are subject to the requirements of
subpart H of 40 CFR Part 141.

Target Log Inactivation - For the target pathogen, the specific log inactivation the PWS
wants to achieve using UV disinfection. The target log inactivation is driven by
requirements of the SWTR, LT1ESWTR, IESWTR, and LT2ESWTR.

Target Pathogen (also called "target microorganism") - For the purposes of this
manual, the target pathogen is defined as the microorganism for which a PWS wants to
obtain inactivation credit using UV disinfection.

UV Absorbance (A) - a measure of the amount of UV light that is absorbed by a
substance (e.g., water, microbial DNA, lamp envelope, quartz sleeve) at a specific
wavelength (e.g., 254 nm). This measurement accounts for absorption and scattering in
the medium (e.g., water).  Standard Method 5910B details this measurement method.
However, for UV disinfection applications, the sample should not be filtered or adjusted
for pH as described in Standard Methods.

UV Absorbance at 254 nm (A2s4) - a measure of the amount of UV light that is
absorbed by a substance at 254 nm.

UV Action Spectrum - the relative efficiency of UV energy at different wavelengths in
inactivating microorganisms. Each microorganism has a unique action  spectrum.

UV Dose - the UV energy per unit area incident on a surface, typically  reported in units
of mJ/cm2 or J/m2.  The UV dose received by a waterborne microorganism in a reactor
vessel accounts for the effects on UV intensity of the absorbance of the water, absorbance
of the quartz sleeves, reflection and refraction of light from the water surface and reactor
walls, and the germicidal effectiveness of the UV wavelengths transmitted.  This
guidance manual also uses the following terms related to UV dose:

   •   UV dose distribution - the probability distribution of delivered UV doses that
       microorganisms receive in a flow-through UV reactor; typically shown  as a
       histogram. An example is shown in Figure 2-8.
UV Disinfection Guidance Manual              xvii                        November 2006
For the Final LT2ESWTR

-------
                                    Glossary (Continued)
   •   Reduction Equivalent Dose (RED) - The UV dose derived by entering the log
       inactivation measured during full-scale reactor testing into the UV dose-response
       curve that was derived through collimated beam testing.  RED values are always
       specific to the challenge microorganism used during experimental testing and the
       validation test conditions for full-scale reactor testing.

   •   Required Dose (Dreq) - The UV dose in units of ml/cm2 needed to achieve the
       target log inactivation for the target pathogen. The required dose is specified in
       the LT2ESWTR and presented in Table 1.4 of this guidance manual.

   •   Validated Dose (Dvai) - The UV dose in units of ml/cm2 delivered by the UV
       reactor as determined through validation testing.  The validated dose is compared
       to the Required Dose (Dreq) to determine log inactivation credit.

   •   Calculated Dose - the RED calculated using the dose-monitoring equation that
       was developed through validation testing.

UV Dose-Response - the relationship indicating the level of inactivation of a
microorganism as a function of UV dose.

UV Equipment - the UV reactor and related components of the UV disinfection process,
including (but  not limited to) UV reactor appurtenances, ballasts, and control panels.

UV Facility -  all of the components of the UV disinfection process, including (but not
limited to) UV reactors, control systems, piping, valves, and  building (if applicable).

UV Intensity - the power passing through a unit area perpendicular to the direction of
propagation. UV intensity is used in this guidance manual to describe the magnitude of
UV light measured by UV sensors in a reactor and with a radiometer in bench-scale UV
experiments.

UV Intensity Setpoint Approach - See Dose-Monitoring Strategy.

UV Irradiance - the power per unit area incident to the direction of light propagation at
all angles, including normal.

UV Lamp Status - a parameter that is monitored during validation testing and long-term
operation of UV reactors that indicates whether a particular UV lamp is on or off.

UV Light - light emitted with wavelengths from 200 to 400  nm.

UV Reactor - the vessel or chamber where exposure to UV  light takes place, consisting
of UV lamps, quartz sleeves, UV sensors, quartz sleeve cleaning systems, and baffles or
other hydraulic controls.  The UV reactor also includes additional hardware for
monitoring UV dose delivery; typically comprised of (but not limited to): UV sensors and
UVT monitors.
UV Disinfection Guidance Manual               xviii                        November 2006
For the Final LT2ESWTR

-------
                                    Glossary (Continued)
UV Reactor Validation - Experimental testing to determine the operating conditions
under which a UV reactor delivers the dose required for inactivation credit of
Cryptosporidium, Giardia lamblia, and viruses.

UV Sensitivity - the resistance of a microorganism to inactivation by UV light,
expressed as ml/cm2 per log inactivation.

UV Sensor - a photosensitive detector used to measure the UV intensity at a point within
the UV reactor that converts the signal to units of milliamps (mA).

UV Transmittance (UVT) - a measure of the fraction of incident light transmitted
through a material (e.g., water sample or quartz).  The UVT is usually reported for a
wavelength of 254 nm and a pathlength of 1-cm.  If an alternate pathlength is used, it
should be specified or converted to units of cm"1.  UVT is often represented as a
percentage and is related to the UV absorbance (A25/t) by the following equation (for a 1-
cm path length): % UVT = 100 x 10'A.

Validated Dose - see UV Dose.

Validation Factor - an uncertainty term that accounts for the bias and uncertainty
associated with validation testing.

Validated Operating Conditions - the operating conditions under which the UV reactor
is confirmed as delivering the dose required for LT2ESWTR inactivation credit.  These
operating conditions must include flow rate, UV intensity as measured by a UV sensor,
and UV lamp status.  Also commonly referred to as the "validated range" or the
"validated limits."

Validation Uncertainty - an uncertainty term that accounts for error in measurements
made during validation testing to develop the UV intensity setpoint(s) (for the UV
Intensity Setpoint Approach) or dose-monitoring equation (for the Calculated Dose
Approach).

Visible Light - Wavelengths of light in the visible range (380 - 720 nm).
UV Disinfection Guidance Manual              xix                         November 2006
For the Final LT2ESWTR

-------
                                       Glossary (Continued)
                            This page intentionally left blank.
UV Disinfection Guidance Manual               xx                           November 2006
For the Final LT2ESWTR

-------
          List of Units, Abbreviations,  and Acronyms

 X             wavelength
 ug            microgram
 ug/L          microgram per liter
 um           micrometer, micron
 A254          ultraviolet light absorbance at 254 nanometers
 ACGIH       American Conference of Governmental Industrial Hygienists
 AIAA         American Institute of Aeronautics and Astronautics
 ANSI         American National Standards Institute
 AOC          assimilable organic carbon
 APHA        American Public Health Association
 ATCC        American Type Culture Collection
 AWA         Australian Water Association
 AWWA       American Water Works Association
 Bp0iy          polychromatic bias
 °C            degree Centigrade
 CCPP         calcium carbonate precipitation potential
 CCWA       Clayton County Water Authority
 CEC          Clancy Environmental Consultants
 CF           correction factor
 CFas          action spectra correction factor
 CFD          computational fluid dynamics
 CFR          Code of Federal Regulations
 cfu           colony forming unit
 cfu/mL       colony forming units per milliliter
 cm           centimeter
 cPEL          ceiling permissible exposure limit
 CT           contact time
 DBF          disinfection byproduct
 DBPR        Stage 2 Disinfectants and Disinfection Byproducts Rule
 DNA          deoxyribonucleic acid
 DOC          dissolved organic carbon
 DReq          required UV dose
 DVGW       Deutsche Vereinigung des Gas- und Wasserfaches
 ENR BCI     Engineering News Record Building Cost Index
 EOLL         end-of-lamp-life
 EPA          U.S. Environmental Protection Agency
 EPRI          Electric Power Research Institute
 °F            degree Fahrenheit
 g             gram
 g/L           gram per liter
 g/mL          gram per milliliter
 GAC          granular activated carbon
 GFI          ground fault interrupter
UV Disinfection Guidance Manual
For the Final LT2ESWTR
xxi
                          November 2006

-------
                      List of Units, Abbreviations, and Acronyms (Continued)
 gpm
 gpm/sf
 GWUDI
 HAA
 HAAS

 HazMat
 hr
 HSP
 HVAC
 Hz
 I&C
 IDLH
 IEEE
 IESWTR
 IFE
 J
 J/m2
 kVA
 kW
 L
 LED
 LID
 log I
 LP
 LPHO
 LRAA
 LSA
 LT IESWTR
 LT2ESWTR
 M
 M'1 cm'1
 m/s2
 mA
 mA/mW
 MCL
 mg
 mg/cm
 mg/L
 mgd
 mg-Hg/mJ
 min
 ml
 ml/cm2
 mL
 mm
3
gallon per minute
gallon per minute per square foot
ground water under the direct influence of surface water
haloacetic acid
five haloacetic acids (monochloroacetic, dichloroacetic, trichloroacetic,
monobromoacetic, and dibromoacetic acids)
hazardous materials
hour
high-service pump
heating, ventilating, and air conditioning system
Hertz
instrumentation and control
Immediately Dangerous to Life or Health
Institute of Electrical and Electronic Engineers
Interim Enhanced Surface Water Treatment Rule
individual filter effluents
joule
joule per meter squared
kilovolt ampere
kilowatt
liter
light emitting diode
light intensity distribution
log Inactivation
low pressure
low pressure high output
locational running annual average
lignin sulfonic acid
Long Term 1 Enhanced Surface Water Treatment Rule
Long Term 2 Enhanced Surface Water Treatment Rule
molar
molar absorption coefficient
meter per second squared
milliampere
milliampere per milliwatt
maximum contaminant level
milligram
milligram per centimeter
milligram per liter
million gallon per day
milligrams mercury per meter cubed
minute
millijoule
millijoule per centimeter squared
milliliter
millimeter
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                              xxn
                                                         November 2006

-------
                      List of Units, Abbreviations, and Acronyms (Continued)
 m-mhos/cm    millimhos per centimeter
 MP            medium pressure
 MS2           male-specific-2 bacteriophage
 mW           milliwatt
 mW/cm        milliwatt per centimeter
 mW/cm2       milliwatt per centimeter squared
 mWs/cm2      milliwatt second per centimeter squared
 NEC           National Electric Code
 NIOSH        National Institute for Occupational Safety and Health
 NIST          National Institute of Standards and Technology
 nm            nanometer
 NOM          natural organic matter
 NPL           National Physical Laboratory
 NSF           National Science Foundation
 NTU           nephelometric turbidity unit
 NWRI         National Water Research Institute
 NYSERDA    New York State Energy Research and Development Authority
 O&M          operation and maintenance
 OCC           off-line chemical cleaning
 OMC          on-line mechanical cleaning
 OMCC        on-line mechanical-chemical cleaning
 ONORM       Osterreichisches Normungsinstitut
 oocysts/L      oocysts per liter
 ORP           oxidation-reduction potential
 OSHA         Occupational Safety and Health Administration
 Pa             pascal
 PAC           powder activated carbon
 PEL           permissible exposure limit
 pfu            plaque forming unit
 pfu/mL        plaque forming units per milliliter
 PLC           programmable  logic controller
 psi             pounds per  square inch
 psig           pounds-force per square inch gauge
 PTB           Physikalisch Technische Bundesanstalt
 PWS           public water system
 PWSID        public water system  identification
 QA/QC        quality assurance/quality control
 RAA           running annual average
 RCRA         Resource Conservation and Recovery Act
 RED           reduction equivalent dose
 RNA           ribonucleic  acid
 s              second
 SCADA       Supervisory Control and Data Acquisition
 SDWA        Safe Drinking Water Act
 SMCL         secondary maximum contaminant level
 SUVA         specific ultraviolet absorbance
UV Disinfection Guidance Manual
For the Final LT2ESWTR
XXlll
                           November 2006

-------
                      List of Units, Abbreviations, and Acronyms (Continued)
 SWTR
 TCU
 THM
 TLV
 TNTC
 TOC
 ISA
 TSB
 TTHM
 UPS
 UV
 UV-A
 UV-B
 UV-C
 UDR
 UIN
 Us
 USP
 UVT
 VF
 VFD
 W
 W/cm
 W/cm2
 W/m2
 W/nm
 WEF
 WTP
Surface Water Treatment Rule
total color unit
trihalomethane
threshold limit value
too numerous to count
total organic carbon
tryptic soy agar
tryptic soy broth
total trihalomethane
uninterruptible power supply
ultraviolet
Uncertainty in Validation
ultraviolet range from 315 to 400 nm
ultraviolet range from 280 to 315 nm
ultraviolet range from 200 to 280 nm
Uncertainty of the Dose-response Fit
Uncertainty in Interpolation
Uncertainty in UV Sensor Measurements
Uncertainty in the Setpoint Value
ultraviolet transmittance
validation factor
variable frequency drive
watt
watt per centimeter
watt per centimeter squared
watt per meter squared
watt per nanometer
Water Environment Federation
water treatment plant
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                        XXIV
                                                   November 2006

-------
                                    1.  Introduction
       Interest in using ultraviolet (UV) light to disinfect drinking water is growing among
public water systems (PWSs)1 due to its ability to inactivate pathogenic microorganisms without
forming regulated disinfection byproducts (DBFs). UV light has proven effective against some
pathogens, such as Cryptosporidium, that are resistant to commonly used disinfectants like
chlorine.

       The United States Environmental Protection Agency (EPA) developed the Long Term 2
Enhanced Surface Water Treatment Rule (LT2ESWTR) to further reduce microbial
contamination of drinking water. The rule requires additional treatment for some PWSs based on
their source water Cryptosporidium concentrations and current treatment practices. UV
disinfection is one option PWSs have to comply with the additional treatment requirements.

       The design, operation, and maintenance needs for UV disinfection differ from those of
traditional chemical disinfectants used in drinking water applications. EPA has developed this
guidance manual to familiarize states2 and PWSs with these distinctions, as well as associated
regulatory requirements in the LT2ESWTR. Particularly  important design and operation
considerations include monitoring, reliability, redundancy, lamp cleaning and replacement, and
lamp breakage. Regulatory requirements include UV dose, UV reactor validation, monitoring,
reporting, and off-specification compliance.

       EPA developed the requirements for UV disinfection in the LT2ESWTR and the
guidance in this manual solely for PWSs using UV light to meet drinking water disinfection
standards established under the Safe Drinking Water Act (SDWA). EPA has not addressed and
did not consider the extension of these requirements  and guidance to other applications,
including point-of-entry or point-of-use devices for residential water treatment that are not
operated by PWSs to meet SDWA disinfection standards.
        Chapter 1 covers:
        1.1   Guidance Manual Obj ectives
        1.2   Organization
        1.3   Regulations Summary
        1.4   UV Disinfection Requirements for Filtered and Unfiltered PWSs
        1.5   Regulations Timeline
        1.6   Alternative Approaches for Disinfecting with UV Light
1  Throughout this document, the terms "PWS" and "water system" are used interchangeably.
2  Throughout this document, the terms "state" and "states" are used to refer to all regulatory agencies, including
  both state and federal, with primary enforcement authority for PWSs.

UV Disinfection Guidance Manual             1-1                                   November 2006
For the Final LT2ESWTR

-------
                                       1. Introduction
1.1    Guidance Manual Objectives

       This manual's objectives are as follows:

       .   Provide PWSs and designers with technical information and guidance on selecting,
          designing, and operating UV installations and complying with the UV disinfection-
          related requirements in the LT2ESWTR.

       .   Provide states with guidance and the necessary tools to assess UV installations during
          the design, start-up, and routine operation phases.

       .   Provide manufacturers with testing and performance standards for UV reactors and
          components intended for treating drinking water.


1.2    Organization

       This manual consists of seven chapters and seven appendices:

       .   Chapter 1 - Introduction. The remainder of this chapter summarizes the microbial
          treatment and UV disinfection requirements of the LT2ESWTR.

       .   Chapter 2 - Overview of UV Disinfection. This chapter describes the principles of
          UV disinfection, dose-response relationships, water quality impacts, and UV reactors.

          Chapter 3 - Planning Analyses for UV Facilities. This chapter discusses planning
          for UV disinfection facilities, including disinfection goals, potential locations, basic
          design parameters, UV reactor evaluation, operational strategies, facility hydraulics,
          pilot- and demonstration-scale testing, and preliminary costs.

       .   Chapter 4 - Design Considerations for UV Facilities. This chapter discusses the key
          design features for UV disinfection facilities and presents some common approaches
          to facility design. Key design features include hydraulics, operational optimization,
          instrumentation and controls,  electrical power considerations, facility layout, and
          specifications.

          Chapter 5 -Validation of UV Reactors. This chapter summarizes the LT2ESWTR
          requirements for validation testing and presents EPA's recommended validation
          protocol.

          Chapter 6 - Start-up and Operation of UV Facilities. This chapter discusses start-
          up and operation issues for UV disinfection facilities, recommended maintenance
          tasks, and monitoring requirements and recommendations.

          Chapter 7 - Bibliography. This chapter lists the references used in Chapters 1
          through 6 and Appendices A through G.

UV Disinfection Guidance Manual                1-2                                November 2006
For the Final LT2ESWTR

-------
                                     1. Introduction
          Seven appendices provide supplemental information to Chapters 1-6.

          Appendix A.  Preparing and Assaying Challenge Microorganisms
          Appendix B.  UV Reactor Testing Examples
          Appendix C.  Collimated Beam Testing to Develop a UV Dose-response Curve
          Appendix D.  Background to the UV Reactor Validation Protocol
          Appendix E.  UV Lamp Break Issues
          Appendix F.  Case Studies
          Appendix G.  Reduction Equivalent Dose Bias Tables
1.3    Regulations Summary

       This section summarizes general microbial treatment and specific UV disinfection
requirements in the LT2ESWTR. The rule applies to all PWSs that use surface water or
groundwater under the direct influence of surface water (GWUDI). It builds on existing
regulations—the Surface Water Treatment Rule (SWTR), Interim Enhanced Surface Water
Treatment Rule (IESWTR), and Long Term 1 Enhanced Surface Water Treatment Rule
(LT1ESWTR)—to improve control of Cryptosporidium and other microbial pathogens.

       EPA has developed a Stage 2 Disinfectants and Disinfection Byproducts Rule (DBPR)
with the LT2ESWTR to address the risk-risk trade off between microbial disinfection and the
DBFs formed by commonly used disinfectants. The Stage 2 DBPR aims to reduce peak DBF
concentrations in the distribution system by modifying the Stage 1 DBPR monitoring
requirements and procedures for compliance  determination.  Consequently,  when a PWS assesses
its disinfection strategy, it must consider both the disinfectant effectiveness against the target
pathogen and the DBFs formed as a result of the disinfectant.

       Table 1.1 highlights microbial treatment requirements and DBF maximum contaminant
levels (MCLs) from the  SWTR, IESWTR, LT1ESWTR, LT2ESWTR, Stage 1 DBPR, and Stage
2 DBPR. See the original regulations or the Code of Federal Regulations (CFR) for complete
requirements. Details on the Stage 2 DBPR can be found in 40 CFR 141.600 - 141.629.
UV Disinfection Guidance Manual                1-3                              November 2006
For the Final LT2ESWTR

-------
                                        1. Introduction
                   Table 1.1. Summary of Microbial and DBF Rules
Surface Water Treatment Rules - Minimum Treatment Requirements1
Regulation
SWTR
IESWTR and LT1 ESWTR
LT2ESWTR
Giardia
3-log removal
and/or inactivation
Virus
4-log removal
and/or inactivation
No change from SWTR
No change from SWTR
Cryptosporidium
Not addressed
2-log removal
0- to 2.5-log additional treatment
for filtered systems2
2- or 3-log inactivation for
unfiltered systems2

DBF Rules - MCLs Based on Running Annual Averages (RAAs) or Locational RAAs (LRAAs)
Regulation
Stage 1 DBPR
Stage 2 DBPR4
Total
Trihalomethanes
(TTHM)
(H9/L)3
80 as RAA
80 as LRAA
Five Haloacetic
Acids (HAAS)
(W/L)3
60 as RAA
60 as LRAA
Bromate
(ng/L)3
10
Chlorite (^g/L)3
1000
No change from Stage 1
1  The term "log" means the order of magnitude reduction in concentration; e.g., 2-log removal equals a 99%
  reduction, 3-log removal equals a 99.9% reduction, and 4-log removal equals a 99.99-percent reduction.
2  Specific requirements for each plant depend on source water monitoring results and current treatment practices
  (40CFR 141.710-141.712).
3  micrograms/liter (ug/L)
4  Monitoring locations for LRAAs are identified from the Initial Distribution System Evaluation.
       The following sections describe LT2ESWTR requirements for filtered and unfiltered
PWSs.
1.3.1  Filtered PWSs

       The LT2ESWTR requires filtered PWSs to conduct source water monitoring3 to
determine average Cryptosporidium concentrations. Based on the monitoring results, filtered
PWSs will be classified in one of four possible treatment bins. A PWS's bin classification
determines the extent of any additional Cryptosporidium treatment requirements. The rule
requires filtered PWSs to  comply with additional treatment requirements by using one or more
management or treatment techniques from a "microbial toolbox" of options (40 CFR 141.711).
UV is one option in the microbial toolbox; see the LT2ESWTR for additional options (40 CFR
141.715).
3 The full monitoring requirements are described in the Source Water Monitoring Guidance Manual for Public
  Water Systems for the Long Term 2 Enhanced Surface Water Treatment Rule (USEPA 2006).
UV Disinfection Guidance Manual
For the Final LT2ESWTR
1-4
November 2006

-------
                                         1. Introduction
       Filtered PWSs are exempt from Cryptosporidium monitoring if the PWS provides, or will
provide, a total of at least 5.5-log Cryptosporidium treatment—the maximum treatment required
by the LT2ESWTR for filtered PWSs4—by the treatment compliance date, which varies,
depending on population (see Section 1.5 for compliance dates) [40 CFR 141.701(d)]. Installing
a UV disinfection system that is validated for the appropriate inactivation credit in addition to
filtration treatment can achieve this objective.

Treatment Bin Classification

       Table 1.2 presents the bin classifications and their corresponding additional treatment
requirements for filtered PWSs (40 CFR 141.711). PWSs with average Cryptosporidium
concentrations of less than 0.075 oocysts per liter (oocysts/L) are placed in Bin 1 where no
additional treatment is required. For concentrations of 0.075 oocysts/L or more, treatment
beyond that required by existing rules is necessary. The additional treatment required for each
bin, specified in terms of log removal, depends on the type of treatment the PWS already uses.
                    Table 1.2. Bin Requirements for Filtered PWSs
Cryptosporidium
Concentration
(oocysts/L)
< 0.075
> 0.075 and < 1.0
> 1.0and<3.0
>3.0
Bin
Classifi-
cation
1
2
3
4
And if the following filtration treatment is operating in full
compliance with existing regulations, then the additional treatment
requirements are2...
Conventional
Filtration Treatment
(includes softening)
No additional
treatment
1log
treatment
2 log
treatment4
2.5 log
treatment4
Direct
Filtration
No additional
treatment
1.5 log
treatment3
2.5 log
treatment4
Slog
treatment4
Slow Sand or
Diatomaceous
Earth Filtration
No additional
treatment
1log
treatment
2 log
treatment4
2.5 log
treatment4
Alternative
Filtration
Technologies
No additional
treatment
As determined
by the state3 5
As determined
by the state4'6
As determined
by the state4 7
1 From 40 CFR 141. 711
   or diatomaceous earth filtration, and a 2.5-log credit for direct filtration plants.
   PWSs may use any technology or combination of technologies from the microbial toolbox.
   PWSs must achieve at least 1 log of the required treatment using ozone, chlorine dioxide, UV light, membranes,
   bag/cartridge filters, or bank filtration.
   Total Cryptosporidium treatment must be at least 4.0 log.
   Total Cryptosporidium treatment must be at least 5.0 log.
   Total Cryptosporidium treatment must be at least 5.5 log.
4 Treatment requirements for filtered PWSs [40 CFR 141.711] are based on a determination that conventional, slow
  sand, and diatomaceous earth filtration plants in compliance with the IESWTR and LT1ESWTR achieve an
  average of 3-log removal of Cryptosporidium. EPA has determined that direct filtration plants achieve an average
  of 2.5-log removal of Cryptosporidium (their removal is less than in conventional filtration because they lack a
  sedimentation process).
UV Disinfection Guidance Manual
For the Final LT2ESWTR
1-5
November 2006

-------
                                      1. Introduction
1.3.2  Unfiltered PWSs

       All existing requirements for unfiltered PWSs remain in effect, including disinfection to
achieve at least 3-log inactivation ofGiardia and 4-log inactivation of viruses. The LT2ESWTR
requires 2- or 3-log inactivation of Cryptosporidium, depending on the source water
concentration of Cryptosporidium, as shown in Table 1.3 [40 CFR 141.712)].

                    Table 1.3. Requirements for Unfiltered PWSs
Average Cryptosporidium Concentration
(oocysts/L)
<0.01
>0.01
Additional Cryptosporidium Inactivation
Requirements
2 log1
Slog1
       1  Overall disinfection requirements must be met with a minimum of two disinfectants [40 CFR
       Unfiltered PWSs are exempt from Cryptosporidium monitoring if the PWS provides, or
will provide, a total of at least 3-log Cryptosporidium inactivation—the maximum treatment
required by the LT2ESWTR for unfiltered systems [40 CFR 141.701(d)]—by the treatment
compliance date. (See Figure 1.1.)  Installing a UV disinfection system that is validated for the
appropriate inactivation credit can achieve this objective.
1.3.3  PWSs with Uncovered Finished Water Storage Facilities

       The LT2ESWTR requires PWSs with uncovered finished water storage facilities to either
cover the storage facility or treat the discharge of the storage facility that is distributed to
consumers to achieve inactivation and/or removal of 4-log virus, 3-log Giardia, and 2-log
Cryptosporidium. UV disinfection is a treatment option that can help water systems meet these
requirements.
1.4    UV Disinfection Requirements for Filtered and Unfiltered PWSs

       The LT2ESWTR has several requirements related to the use of UV disinfection. They
address the UV doses for different levels of inactivation credit, performance validation testing of
UV reactors, monitoring, reporting, and off-specification operation.
1.4.1  UV Dose and Validation Testing Requirements

       EPA developed UV dose requirements for PWSs to receive credit for inactivation of
Cryptosporidium, Giardia, and viruses (Table 1.4). The UV dose values in Table 1.4 are
applicable only to post-filter applications of UV disinfection in filtered systems and to unfiltered
systems.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
1-6
November 2006

-------
                                       1. Introduction
       Unlike chemical disinfectants, UV leaves no residual that can be monitored to determine
UV dose and inactivation credit. The UV dose depends on the UV intensity (measured by UV
sensors), the flow rate, and the UV transmittance (UVT).5 A relationship between the required
UV dose and these parameters must be established and then monitored at the water treatment
plant to ensure sufficient disinfection of microbial pathogens.

                         Table 1.4.  UV Dose  Requirements -
                    millijoules per centimeter squared (mJ/cm2)1
Target
Pathogens
Cryptosporidium
Giardia
Virus
Log Inactivation
0.5
1.6
1.5
39
1.0
2.5
2.1
58
1.5
3.9
3.0
79
2.0
5.8
5.2
100
2.5
8.5
7.7
121
3.0
12
11
143
3.5
15
15
163
4.0
22
22
186
         40CFR141.720(d)(1)
       The UV dose requirements in Table 1.4 account for uncertainty in the UV dose-response
relationships of the target pathogens but do not address other significant sources of uncertainty in
full-scale UV disinfection applications. These other sources of uncertainty are due to the
hydraulic effects of the UV installation, the UV reactor equipment (e.g., UV sensors), and the
monitoring approach.

       Due to these factors, the LT2ESWTR requires PWSs to use UV reactors that have
undergone validation testing. This validation testing must determine the operating conditions
under which the reactor delivers the required UV dose for treatment credit [40 CFR
141.720(d)(2)]. These operating conditions must include flow rate, UV intensity as measured by
a UV sensor, and UV lamp  status. Further, validation testing must meet the following
requirements:

       .   Validated operating conditions must account for UV absorbance of the water, lamp
          fouling and aging, measurement uncertainty of online sensors, UV dose distributions
          arising from the  velocity profiles through the reactor, failure of UV lamps or other
          critical system components, and inlet and outlet piping or channel configurations of
          the UV reactor [40 CFR 141.720(d)(2)(i)].

       .   Validation testing must involve full-scale testing of a reactor that conforms uniformly
          to the UV reactors used by the PWS, and it also must demonstrate inactivation of a
          test microorganism whose dose-response characteristics have been quantified with a
          low-pressure mercury vapor lamp [40 CFR 141.720(d)(2)(ii)].

       Using the above requirements as a basis, Chapter  5 presents EPA's recommended
validation protocol. Water systems are not required to follow this protocol but may follow
alternatives that achieve compliance with the regulatory requirements as long as they are
acceptable to the state. Also, states may have additional requirements than are provided in the
federal rule.
  UV intensity measurements may account for UVT depending on sensor locations.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
1-7
November 2006

-------
                                      1. Introduction
1.4.2  UV Disinfection Monitoring Requirements [40 CFR 141.720(d)(3)(i)]

       The LT2ESWTR requires PWSs to monitor their UV reactors to demonstrate that they
are operating within the range of conditions that were validated for the required UV dose.  At a
minimum, PWSs must monitor each reactor for flow rate, lamp status, UV intensity as measured
by a UV sensor, and any other parameters required by the state. UV absorbance should also be
measured when it is used in a dose-monitoring strategy. PWSs must verify the calibration  of UV
sensors and recalibrate sensors in accordance with a protocol the state approves. Section 6.4.1.2
of this guidance describes recommended frequencies for checking sensors.
1.4.3  UV Disinfection Reporting Requirements [40 CFR 141.721(f)(15])

       The LT2ESWTR requires PWSs to report the following items:

          Initial reporting - Validation test results demonstrating operating conditions that
          achieve the UV dose required for compliance with the LT2ESWTR.

          Routine reporting - Percentage of water entering the distribution system that was
          not treated by the UV reactors operating within validated conditions on a monthly
          basis.
1.4.4  Off-specification Operational Requirement for Filtered and Unfiltered
       Systems [40 CFR 141.720(d)(3)(ii)]

       To receive disinfection credit for UV, both filtered and unfiltered PWSs must treat at
least 95 percent of the water delivered to the public during each month by UV reactors operating
within validated conditions for the required UV dose. EPA views this 95-percent limit as a
feasible minimum level of performance for PWSs to achieve, while ensuring the desired level of
health protection is provided. For purposes of design and operation, PWSs should strive to
deliver the required UV dose at all times during treatment.

       In this manual, operating outside the validated limits is defined as off-specification. Off-
specification compliance is based on the volume of water treated. Guidance for calculating off-
specification is provided in Chapter 6.
UV Disinfection Guidance Manual                1-8                              November 2006
For the Final LT2ESWTR

-------
                                       1. Introduction
1.5    Regulations Timeline

       Figure 1.1 provides a timeline for LT2ESWTR initial source water monitoring and
treatment installation.  Compliance dates vary among the following PWS sizes:

       .   Systems serving 100,000 or more people
       .   Systems serving 50,000 to 99,999 people
       .   Systems serving 10,000 to 49,999 people
          Systems serving fewer than 10,000 people


       Treatment installation dates pertain only to PWSs that are required to provide additional
treatment for Cryptosporidium. Further, the actual duration of the treatment installation phase
will be contingent on a number of PWS-specific factors, including  scope of design (i.e., new
facility or retrofit); scale of design (size of facility); available in-house resources; procurement
methods; and validation testing requirements (discussed in detail in Chapters 3, 4, and 5).
1.6    Alternative Approaches for Disinfecting with UV Light

       This manual provides technical information about using UV disinfection for drinking
water treatment. Although it covers many aspects of implementing a UV installation, it is not
comprehensive in terms of all types of UV installations, design alternatives, and validation
protocols that may provide satisfactory performance. For example, pulsed UV and eximer lamps
are two types of UV technologies not included in this manual, but they may provide effective
disinfection.

       Currently, substantial research is being conducted on UV disinfection and its applications
in various industries. As more information becomes available, UV equipment or methods of
operation, design, and validation will evolve. Water systems are not limited by the information
provided in this guidance manual but must meet the requirements of the LT2 and other drinking
water rules, as well as any state-specific requirements. States may approve alternatives in UV
installation design, operation, and validation that are not described in this manual.
UV Disinfection Guidance Manual                 1-9                               November 2006
For the Final LT2ESWTR

-------
                                   I. Introduction
  Figure 1.1. LT2ESWTR Compliance Timeline for Initial Source Water Monitoring
                           and Treatment Installation

o
tM

O
tM

0

CO
0
tM

CM
O
CM

O
CM


0
0
CM


a>
0
o
CM


00
O
0
CM

l~-
0
0
CM


0
o
CM
01
N
CO
S
£
co







(N
R
^~
D.
<>














CD
O
CM
^/
13
O








5 i
•§ g
sa
"-.a
^^^


c S
0 ~
E B
ra ra
2 M
1- £






=s=.
ryptospori
Monitorin
O

/
100,000 People or More


































CM
O
CM
~0V
U













l^
R
^/
^
<
50,000 to 99,999 People






„, *c
Possibk
Extensio
*^
w

= i
0 ~
E 'ra
3%






5
If
§•*
1°
cT
/







CO
R
t3
u x












CO
o
CM
v-V
Q.
<



10,000 to 49, 999 People



fl) C
•S °
=§'1
S|
D- ><

^

II
£ ^
2 5
1- £




5
If
§ o
ll
t5

/







•^-
R
~Ofc
1^









o
o
CM
r/
'^_
<

CO
o
o
CM
v-V
o
O




Fewer than 10,000 People

o^c
fi'5
in Q
-s
^

« g
Treatme
Installat


^r
Monit.
Extensior

Si:
&'E
5-1
/

^ CM
O ±i
Ui3

/





















































CD
^5
Q
§
CD
"o.
E
o
0
egulatory
a:
»


vements.
2
Q.
E
_ro
'o.
ro
•1

'5
£

CD
•a
sr
£
ro
to
0)
'o
c
g
ti)
o
w
TD
0)
"ro
-a
B
ro
b)
E

r extension ma
ro
0)
|


c
g
^
"ro
1
JD
•a
0)
'5
D"
(D


ro
ro
^
w
•s
w"
•a
OJ
Q)
1
Z3
b
0)
H
•a
OJ
(D
"c
o

onitoring applii
E
uj
concentration greater than 50 E.
c
nual mea
c
ro
es an
to
o
TD
C
D)
^C
_o

c
o
^
0
uj
.>.
0
ra'
w
w
OJ
§
ems is
"co
5T
ig for small
o
loridium moniti
cf
month for one year or

mples pe
S
O
D)
~O
_OJ
"o
O
>*
,_
OJ
'oJ
£
^
'I
o
^;
o
o
'c
o
>,
ro
0)
Q.
o
«
m 10,000 p
£
serving fewer
(f>
It
O CO























two years.
b
pie per month
E
ro
one s
UV Disinfection Guidance Manual
For the Final LT2ESWTR
1-10
November 2006

-------
                         2. Overview of UV Disinfection

       Chapter 2 provides an overview of UV disinfection. This overview includes discussion of
basic chemical and physical principles, the components of UV equipment, and performance
monitoring for UV facilities. The overview material in Chapter 2 is intended to present generally
accepted facts and research results related to UV disinfection. The material is not intended to
provide guidance or recommendations for designing, validating,  or installing UV disinfection
facilities. Some guidance is included in this chapter to enhance the information presented, but
any guidance that appears in this section is also documented in the appropriate subsequent
chapters in this manual.
        Chapter 2 covers:
        2.1  History of UV Light for Drinking Water Disinfection
        2.2  UV Light Generation and Transmission
        2.3  Microbial Response to UV Light
        2.4  UV Disinfection Equipment
        2.5  Water Quality Effects and Byproduct Formation
2.1    History of UV Light for Drinking Water Disinfection

       UV disinfection is an established technology supported by decades of fundamental and
applied research and practice in North America and Europe. Downes and Blunt (1877)
discovered the germicidal properties of sunlight. The development of mercury lamps as artificial
UV light sources in 1901 and the use of quartz as a UV transmitting material in 1906 were soon
followed by the first drinking water disinfection application in Marseilles, France, in 1910. In
1929, Gates identified a link between UV disinfection and absorption of UV light by nucleic acid
(Gates 1929). The development of the fluorescent lamp in the 1930s led to the production of
germicidal tubular lamps. Considerable research on the mechanisms of UV disinfection and the
inactivation of microorganisms occurred during the 1950s (Dulbecco 1950, Kelner 1950, Brandt
and Giese 1956, Powell 1959).

       Although substantial research on UV disinfection occurred during the first half of the 20th
century, the low cost of chlorine and operational problems with early UV disinfection equipment
limited its growth as a drinking water treatment technology. The first reliable applications of UV
light for disinfecting municipal drinking water occurred in Switzerland and Austria in 1955
(Kruithof and van der Leer 1990). By 1985, the number of such installations in these countries
had risen to approximately 500 and  600, respectively. After chlorinated disinfection byproducts
(DBFs) were discovered, UV disinfection became popular in Norway and the Netherlands with
the first installations occurring in 1975 and 1980, respectively.

       As of the year 2000, more than 400 UV disinfection facilities worldwide were treating
drinking water; these UV facilities typically treat flows of less than 1 million gallons per day
(mgd) (USEPA 2000). Since 2000, several large UV installations across the United States have
been constructed or are currently under design. The largest of these facilities includes a 180-mgd

UV Disinfection Guidance Manual                 2-1                              November 2006
For the Final LT2ESWTR

-------
                                2. Overview of UV Disinfection
facility in operation in Seattle, Washington, and a 2,200-mgd facility under design for the New
York City Department of Environmental Protection (Schulz 2004). Because of the susceptibility
of Cryptosporidium to UV disinfection and the emphasis in recent regulations on controlling
Cryptosporidium, the number of public water systems (PWSs) using UV disinfection is expected
to increase significantly over the next decade.
2.2    UV Light Generation and Transmission

       The use of UV light to disinfect drinking water involves (1) generating UV light with the
desired germicidal properties and (2) delivering (or transmitting) that light to pathogens. This
section summarizes how UV light is generated and the environmental conditions that affect its
delivery to pathogens.
2.2.1  Nature of UV Light

       UV light is the region of the electromagnetic spectrum that lies between X-rays and
visible light (Figure 2.1). The UV spectrum is divided into four regions: vacuum UV [100 to 200
nanometers (nm)]; UV-C (200 to 280 nm); UV-B (280 to 315 nm); and UV-A (315 to 400 nm)
(Meulemans 1986). UV disinfection primarily occurs due to the germicidal action of UV-B and
UV-C light on microorganisms. The germicidal action  of UV-A light is small relative to UV-B
light and UV-C light; therefore, very long exposure times are necessary for UV-A light to be
effective as a disinfectant. Although light in the vacuum UV range can disinfect microorganisms
(Munakata et al. 1991), vacuum UV light is impractical for water disinfection applications
because it rapidly  dissipates in water over very short distances. For the purposes of this manual,
the practical germicidal wavelength for UV light is defined as the range between 200 and 300
nm. The germicidal range is discussed further in Section 2.3.1.
                Figure 2.1. UV Light in the Electromagnetic Spectrum
                                    100nm     400 nm

Gamma
Rays
X-ray

UV

Visible
Infrared
                                          254 nm
                                            I
Vacuum UV

UV-C
28C
UV-B
nm
315
UV-A
nm
                100 nm
200 nm
300 nm
400 nm
UV Disinfection Guidance Manual
For the Final LT2ESWTR
          2-2
                          November 2006

-------
                                2. Overview of UV Disinfection
       Typically, UV light is generated by applying a voltage across a gas mixture, resulting in a
discharge of photons. The specific wavelengths of light emitted from photon discharge depend
on the elemental composition of the gas and the power level of the lamp. Nearly all UV lamps
currently designed for water treatment use a gas mixture containing mercury vapor. Mercury gas
is advantageous for UV disinfection applications because it emits light in the germicidal
wavelength range. Other gases such as xenon also emit light in the germicidal range.

       The light output from mercury-based UV lamps depends on the concentration of mercury
atoms, which is directly related to the mercury vapor pressure. In low-pressure (LP) UV lamps,
mercury at low vapor pressure  [near vacuum; 2 x 10"5 to 2 x 10"3 pounds per square inch (psi)]
and moderate temperature [40 degrees centigrade (°C)] produces essentially monochromatic (one
wavelength) UV light at 253.7  nm.  In medium-pressure (MP) UV lamps, a higher vapor pressure
[2 - 200 psi] and higher operating temperature (600 - 900 °C) is used to increase the frequency
of collisions between mercury atoms, which produces UV light over a broad spectrum
(polychromatic) with an overall higher intensity. The characteristics of LP and MP lamps are
discussed in Section 2.4.2 and summarized in Table 2.1.
2.2.2  Propagation of UV Light

       As UV light propagates from its source, it interacts with the materials it encounters
through absorption, reflection, refraction, and scattering. In disinfection applications, these
phenomena result from interactions between the emitted UV light and UV reactor components
(e.g., lamp envelopes, lamp sleeves, and reactor walls) and also the water being treated. When
assessing water quality, UV absorbance or UV transmittance (UVT) is the parameter that
incorporates the effect of absorption and scattering. This section briefly describes both the
phenomena that influence light propagation and the measurement techniques used to quantify
UV light propagation.

       Absorption is the transformation of light to other forms of energy as it passes through a
substance. UV absorbance of a substance varies with the wavelength (X) of the light. The
components of a UV reactor and the water passing through the reactor all absorb UV light to
varying degrees, depending on their material composition. When UV light is absorbed, it is no
longer available to disinfect microorganisms.

       Unlike absorption, the phenomena of refraction, reflection, and scattering change the
direction of UV light, but the UV light is still available to disinfect microorganisms.

       Refraction (Figure 2.2) is the change in the direction of light propagation as it passes
through the interface between one medium and another. In UV reactors, refraction occurs when
light passes from the UV lamp into an air gap, from the air gap into the lamp sleeve, and from
the lamp sleeve into the water. Refraction changes the angle that UV light strikes target
pathogens, but how this ultimately affects the UV disinfection process is unknown.

       Reflection is the change in direction of light propagation when it is deflected by  a surface
(Figure 2.3). Reflection may be classified as specular or diffuse. Specular reflection occurs from
smooth polished surfaces and follows the Law of Reflection (the  angle of incidence is equal to
the angle of reflection). Diffuse reflection occurs from rough surfaces and scatters light in all

UV Disinfection Guidance Manual                2-3                              November 2006
For the Final LT2ESWTR

-------
                                 2. Overview of UV Disinfection
directions with little dependence on the incident angle. In UV reactors, reflection will take place
at interfaces that do not transmit UV light (e.g., the reactor wall) and also at UV transmitting
interfaces (e.g., the inside of a lamp sleeve). The type of reflection and intensity of light reflected
from a surface depends on the material of the surface.

                             Figure 2.2. Refraction of Light
                                               Water
                      Incident Light
                      from UVLamp
                                                          Refracted Light
                                                         >    >
                              Figure 2.3. Reflection of Light
         Incident Light
Reflected Light
Incident Light
                                                                           Reflected Light
                    Specular Reflection
                                  Diffuse Reflection
       Scattering of light is the change in direction of light propagation caused by interaction
with a particle (Figure 2.4). Particles can cause scattering in all directions, including toward the
incident light source (back-scattering). Scattering of light caused by particles smaller than the
wavelength of the light is called Rayleigh scattering. Rayleigh scattering depends inversely on
wavelength to the fourth power (l/?i4) and thus is more prominent at shorter wavelengths.
Particles larger than the wavelength of light scatter more light in the forward direction but also
cause some backscattering that is relatively independent of wavelength.

       UV absorbance (A) quantifies the  decrease in the amount of incident light as it passes
through a water sample over a specified distance or pathlength. UV absorbance at 254 nm (A254)
is a water quality parameter commonly used to characterize the DBF formation potential of the
water (e.g., specific UV absorbance calculations). In UV disinfection applications, A254 is used to
measure the amount of UV light passing through the water and reaching the target organisms.
A254 is measured using a spectrophotometer with 254 nm incident light and is typically reported
on a per centimeter (cm"1) basis.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
         2-4
                      November 2006

-------
                                2. Overview of UV Disinfection
                            Figure 2.4. Scattering of Light
                               Back
                             Scattered
                               Light
                     Incident
                      Light
                                       90° Scattered Light
                                                             Target
                                                            Pathogens
                                                            Forward
                                                            Scattered
                                                             Light
       Standard Method 5910B (APHA et al. 1998) calls for filtering the sample through a
0.45-|j,m membrane and adjusting the pH before measuring the absorbance. For UV disinfection
applications, however, A254 measurements should reflect the water to be treated. Therefore, water
samples should be analyzed without filtering or adjusting the pH. More information on collecting
A254 data is provided in Section 3.4.4.1. Although Standard Methods defines this measurement as
UV absorption, this manual refers to it as UV absorbance because the latter term is widely used
in the water treatment industry.

       UV Transmittance (UVT) has also been used extensively in the literature when
describing the behavior of UV light. UVT is the percentage of light passing through material
(e.g., a water sample or quartz) over a specified distance. The UVT can be calculated using
Beer's law (Equation 2.1):
          UVT = 100* —
                      /n
                                                                             Equation 2.1
       where
           UVT =  UV transmittance at a specified wavelength (e.g., 254 nm) and pathlength
                    (e.g., 1 cm)
           /     =  Intensity of light transmitted through the sample [milliwatt per centimeter
                    squared (mW/cm2)]
           Io    =  Intensity of light incident on the sample (mW/cm2)
UVT can also be calculated by relating it to UV absorbance using Equation 2.2:

       % UVT = 100*10~A
                                                                             Equation 2.2
       where
            UVT =  UV transmittance at a specified wavelength (e.g., 254 nm) and pathlength
                     (e.g., 1 cm)
           A     =  UV absorbance at a specified wavelength and pathlength (unitless)
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                                           2-5
November 2006

-------
                                2. Overview of UV Disinfection
       UVT is typically reported at 254 nm because UV manufacturers and PWSs widely use
A254. This manual assumes UVT is at 254 nm unless specifically stated otherwise.
2.3    Microbial Response to UV Light

       The mechanism of disinfection by UV light differs considerably from the mechanisms of
chemical disinfectants such as chlorine and ozone. Chemical disinfectants inactivate
microorganisms by destroying or damaging cellular structures, interfering with metabolism, and
hindering biosynthesis and growth (Snowball and Hornsey 1988). UV light inactivates
microorganisms by damaging their nucleic acid, thereby preventing them from replicating. A
microorganism that cannot replicate cannot infect a host.

       It is important that the assays used to quantify microorganism inactivation measure the
ability of the microorganism to reproduce (Jagger 1967). For bacteria, assays measure the ability
of the microorganism to divide and form colonies. For viruses, assays measure the ability of the
microorganism to form plaques in host cells. For protozoan cysts, the assays measure the ability
of the microorganism to infect a host or tissue culture. Assays that do not measure a response to
reproduction may result in misleading information on the inactivation of microorganisms using
UV light.

       This section describes how UV light causes microbial inactivation, discusses how
microorganisms can repair the damage, and introduces the concept of UV dose-response.
2.3.1  Mechanisms of Microbial Inactivation by UV Light

       Nucleic acid is the molecule responsible for defining the metabolic functions and
reproduction of all forms of life. The two most common forms of nucleic acid are
deoxyribonucleic acid (DNA) and ribonucleic acid (RNA). DNA and RNA consist of single- or
double-stranded polymers comprising building blocks called nucleotides (Figure 2.5). In DNA,
the nucleotides are classified as either purines (adenine and guanine) or pyrimidines (thymine
and cytosine). In RNA, the purines are the same as in DNA, but the pyrimidines are uracil and
cytosine.

       As shown in Figure 2.6, the nucleotides absorb UV light at wavelengths from 200 to 300
nm. The UV absorption of DNA and RNA reflects their nucleotide composition and tends to
have a peak near 260 nm and a local minimum near 230 nm.

       All purines and pyrimidines  strongly absorb UV light, but the rate  of UV-induced
damage is greater with pyrimidines (Jagger 1967). Absorbed UV light induces six types of
damage in the pyrimidines of nucleic acid (Setlow 1967, Snowball and Hornsey 1988, Pfeifer
1997). The damage varies depending on UV dose. The following three types of damage
contribute to microorganism inactivation:
UV Disinfection Guidance Manual                2-6                              November 2006
For the Final LT2ESWTR

-------
                                 2. Overview of UV Disinfection
        Figure 2.5. Structure of DMA and Nucleotide Sequences within DMA
                              Hydrogen Bonded
                              Nitrogenous
                              Base Pairs (A, T, G, C)
                                 Sugar-
                                 Phosphate
                                 Backbone
                              DMA STRUCTURE
                                               -A-T-G-C-G-A-T-C-
                                                 III   I   I   I  I   I
                                               -T-A-C- G-C-T-A-G-
                              Purines

                            A = Adenine
                            G = Guanine


                         DMA SEQUENCE
                                         Pyrimidines
                                         T= Thymine
                                         C = Cytosine
                 Figure 2.6. UV Absorbance of Nucleotides (left) and
                              Nucleic Acid (right) at pH 7
               1.0-

               o.6H
           rj 2
               0.0'
\  •
\
\\
                      Cytosine
    //*A   '.
.   /: \  ',
                                     Adenine
                                    Guanine
                      '--V A     ^ \
                        v-*  Mhymme vV» \
     1.0

85" 0.8-
c n
|«Jj  0.6H

_§ ^  0.4
rj S2
> H  0.2-
                                                   0.0
                                                                         DMA
                200 220 240 260  280  300
                      Wavelength (nm)
                                  200  220 240 260 280 300
                                       Wavelength (nm)
          Source: Adapted from dagger (1967)
       .  Pyrimidine dimers form when covalent bonds are present between adjacent
          pyrimidines on the same DNA or RNA strand, and they are the most common
          damage resulting from UV disinfection.

       .  Pyrimidine (6-4) pyrimidone photoproducts are similar to pyrimidine dimers and
          form on the same sites.

       .  Protein-DNA cross-links are covalent bonds between a protein and a DNA strand,
          and they may be important for the disinfection of certain microorganisms.

       The other three types of damage do not significantly contribute to UV disinfection:
pyrimidine hydrates occur much less frequently than dimers, and single- and double-strand
breaks and DNA-DNA cross-links occur  only at doses that are several orders of magnitude
higher than the doses typically used for UV disinfection (Jagger 1967).
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                        2-7
                                                        November 2006

-------
                                2. Overview of UV Disinfection
       Pyrimidine dimers are the most common form of nucleic acid damage, being 1000 times
more likely to occur than strand breaks, DNA-DNA cross-links, and protein-DNA cross-links.
Of the three possible pyrimidine dimers that can form within DNA (thymine-thymine, cytosine-
cytosine, and thymine-cytosine), thymine-thymine dimers are the most common. For RNA,
because thymine is not present, uracil-uracil and cytosine-cytosine dimers are formed.
Microorganisms with DNA rich in thymine tend to be more sensitive to UV disinfection (Adler
1966).

       Pyrimidine dimer damage and other forms of nucleic acid damage prevent the replication
of the microorganism. The damage, however, does not prevent the metabolic functions in the
microorganism such as respiration. UV doses capable of causing oxidative damage that prevent
cell metabolism and kill  the microorganism (similar to the damage caused by chemical
disinfectants) are several orders of magnitude greater than doses required to damage the nucleic
acid and prevent replication.
2.3.2  Microbial Repair

       Many microorganisms have enzyme systems that repair damage caused by UV light.
Repair mechanisms are classified as either photorepair or dark repair (Knudson 1985). Microbial
repair can increase the UV dose needed to achieve a given degree of inactivation of a pathogen,
but the process does not prevent inactivation.

       Even though microbial repair can occur, neither photorepair nor dark repair is anticipated
to affect the performance of drinking water UV disinfection, as described below:

       .   Photorepair of UV irradiated bacteria can be prevented by keeping the UV disinfected
          water in the dark for at least two hours before exposure to room light or sunlight.
          Treated water typically remains in the dark in the piping, reservoirs, and distribution
          system after UV disinfection. Most facilities also use chemical disinfection to provide
          further inactivation of bacteria and virus and protection of the distribution system.
          Both of these common practices make photorepair unlikely to be an issue for PWSs.

       .   Dark repair is also not a concern for PWSs because the required UV doses shown in
          Table 1.4 are derived from data that are assumed to account for dark repair.
2.3.2.1   Photorepair

       In photorepair (or photoreactivation), enzymes energized by exposure to light between
310 and 490 nm (near and in the visible range) break the covalent bonds that form the pyrimidine
dimers. Photorepair requires reactivating light and repairs only pyrimidine dimers (Jagger 1967).

       Knudson (1985) found that bacteria have the enzymes necessary for photorepair. Unlike
bacteria, viruses lack the necessary enzymes for repair but can repair using the enzymes of a host
cell (Rauth 1965). Linden et al. (2002a) did not observe photorepair ofGiardia at UV doses
typical for UV disinfection applications (16 and 40 mJ/cm2). However, unpublished data from
the same study show Giardia reactivation in light conditions at very low UV doses (0.5 mJ/cm2,

UV Disinfection Guidance Manual                2-8                               November 2006
For the Final LT2ESWTR

-------
                                2. Overview of UV Disinfection
Linden 2002). Shin et al. (2001) reported that Cryptosporidium does not regain infectivity after
inactivation by UV light. One study showed that Cryptosporidium can undergo some DNA
photorepair (Oguma et al. 2001). Even though the DNA is repaired, however, infectivity is not
restored.
2.3.2.2   Dark Repair

       Dark repair is defined as any repair process that does not require the presence of light.
The term is somewhat misleading because dark repair can also occur in the presence of light.
Excision repair, a form of dark repair, is an enzyme-mediated process in which the damaged
section of DNA is removed and regenerated using the existing complementary strand of DNA.
As such, excision repair can occur only with double stranded DNA and RNA. The extent of dark
repair varies with the microorganism. With bacteria and protozoa, dark repair enzymes start to
act immediately following exposure to UV light; therefore, reported dose-response data are
assumed to account for dark repair.

       Knudson (1985) found that bacteria can undergo dark repair, but some lack the enzymes
needed for dark repair (Knudson 1985). Viruses also lack the necessary enzymes for repair but
can repair using the enzymes of a host cell (Rauth 1965). Oguma et al. (2001) used an assay that
measures the number of dimers formed in nucleic acid to show that dark repair occurs in
Cryptosporidium, even though the mircroorganism did not regain infectivity. Linden et al.
(2002a) did not observe dark repair of Giardia at UV doses typical for UV disinfection
applications (16 and 40 mJ/cm2). Shin et al. (2001) reported Cryptosporidium does not regain
infectivity after inactivation by UV light.
2.3.3  UV Intensity, UV Dose, and UV Dose Distribution

       UV intensity is a fundamental property of UV light and has the units of watts per meter
squared (W/m2) (Halliday and Resnick 1978). UV intensity has a formal definition that is derived
from Maxwell's equations, which are fundamental equations that define the wavelike properties
of light. The total UV intensity at a point in space is the sum of the intensity of UV light from all
directions.

       UV dose is the integral of UV intensity during the exposure period (i.e., the area under an
intensity versus time curve). If the UV intensity is constant over the exposure time, UV dose is
defined as the product of the intensity and the exposure time. Units commonly used for UV dose
are joule per meter squared (J/m2), mJ/cm2, and milliwatt seconds per centimeter squared
(mWs/cm2), with mJ/cm2 being the most common units in North America and J/m2 being the
most common in Europe.5

       In a completely mixed batch system, the UV dose that the microorganisms receive is
equal to the volume-averaged UV intensity within the system. An example of a completely
mixed batch system is the collimated beam study in which a petri dish containing the stirred
5 10 J/m2 = 1 mJ/cm2 = 1 mWs/cm2
UV Disinfection Guidance Manual                 2-9                              November 2006
For the Final LT2ESWTR

-------
                                2. Overview of UV Disinfection
microbial solution is irradiated by a collimated UV light beam (see Appendix C for details). In
this case, the average UV intensity is calculated from the measured UV intensity incident on the
surface of the microbial suspension, the suspension depth, and the UV absorbance of the water
(see Appendix C for details). When using polychromatic light sources (e.g., MP lamps), UV dose
calculations in batch system also incorporate the intensity at each wavelength in the germicidal
range and the germicidal effectiveness at the associated UV wavelengths.

       Dose delivery in a continuous flow UV reactor is considerably more complex than in a
completely mixed batch reactor. Some microorganisms travel close to the UV lamps and
experience a higher dose, while others that travel close to the reactor walls may experience a
lower dose. Some microorganisms move through the reactor quickly, while others travel a more
circuitous path. The result is that each microorganism leaving the reactor receives a different UV
dose. Accordingly, UV dose delivered to the microorganisms passing through the reactor is best
described using a dose distribution (Cabaj et al. 1996) as opposed to a single dose value. A dose
distribution can be defined as a histogram of dose delivery (see Figure 2.7). Alternatively, the
dose distribution can be defined as a probability distribution that a microorganism leaving a UV
reactor will receive a given dose.
            Figure 2.7. Hypothetical Dose Distributions for Two Reactors
                               with Differing Hydraulics
0   16   30
46   76  90
                                                     0   16  30   46   TO  90
       The width of the dose distribution is indicative of the dose delivery efficiency of the
reactor. A narrow dose distribution (Figure 2.7a) indicates a more efficient reactor, and a wider
dose distribution (Figure 2.7b) indicates a less efficient reactor. In particular, the average log
inactivation a reactor achieves with a given microorganism is strongly affected by
microorganisms that receive the lowest UV doses.

       The dose distribution a UV reactor delivers can be estimated using mathematical models
based on computational fluid dynamics (CFD) and the light intensity distribution (LID). CFD is
used to predict the trajectories of microorganisms as they travel through the UV reactor. LID is
used to predict the intensity at each point  within the UV reactor. UV dose to each microorganism
is calculated by integrating the UV intensity over the microorganism's trajectory through the
reactor. Biodosimetry (discussed below) is often used to verify these modeling results.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                                 2-10
                                                     November 2006

-------
                                2. Overview of UV Disinfection
       Currently, dose delivery is measured using a technique termed biodosimetry. With
biodosimetry, the log inactivation of a surrogate microorganism is measured through the UV
reactor and related to a dose value termed the reduction equivalent dose (RED) using the UV
dose-response curve of the surrogate microorganism. Methods for conducting biodosimetry are
presented in Chapter 5. Although alternatives to biodosimetry are being developed (e.g., the use
of actinometric microspheres) for measuring the dose distribution of a reactor, such methods
have not yet been proven for measuring dose delivery in UV reactors.
2.3.4  Microbial Response (UV Dose-Response)

       Microbial response is a measure of the sensitivity of the microorganism to UV light and
is unique to each microorganism. UV dose-response is determined by irradiating water samples
containing the microorganism with various UV doses using a collimated beam apparatus (as
described in Appendix C of this manual) and measuring the concentration of infectious
microorganisms before and after exposure. The microbial response is calculated using Equation
  20
 .3.

                              N
       Log Inactivation = log,0 —-                                            Equation 2.3
                               N

       where
           No = Concentration of infectious microorganisms before exposure to UV light
           N  = Concentration of infectious microorganisms after exposure to UV light

       UV dose-response relationships can be expressed as either the proportion of
microorganisms inactivated or the proportion of microorganisms remaining as a function of UV
dose. Microbial inactivation has a dose-response curve with a positive slope, while microbial
survival has a dose-response curve with a negative slope. This manual presents microbial
response as log inactivation because the terminology is widely accepted in the industry.
Therefore, all dose-response curves presented (log inactivation as a function of dose) have a
positive slope with log inactivation on a logarithmic (base 10) scale and UV dose on a linear
scale.

       Figure 2.8 presents examples of UV dose-response curves. The shape of the UV dose-
response curve typically has three regions. At low UV doses, the UV dose-response shows a
shoulder region where little if any inactivation occurs (e.g., Bacillus subtilis curve, Figure 2.8).
The shoulder region has been attributed to dark repair (Morton and Haynes 1969) and
photorepair (Hoyer 1998).  Above  some threshold dose level, the dose-response shows first-order
inactivation where inactivation increases linearly with increased dose.  In many cases, the dose-
response shows first-order inactivation without a shoulder (e.g., E. coli curve, Figure 2.8). At
higher UV doses, the dose-response shows tailing, a region where the slope of the dose-response
decreases with increased dose (e.g., rotavirus and total coliform curves, Figure 2.8). Tailing has
been attributed to the presence of UV-resistant sub-populations of the microorganism and the
presence of paniculate-associated and clumped microorganisms (Parker and Darby 1995). The
shape of the dose-response curve can affect validation results, and information on how to
account for tailing and shoulders in validation testing is included in Section C.6.
UV Disinfection Guidance Manual                 2-11                              November 2006
For the Final LT2ESWTR

-------
                                2. Overview of UV Disinfection
                  Figure 2.8. Shapes of UV Dose-Response Curves
             re
             ^c
             o
             re
             O)
             o
                   n
             O E co//
             El B. subtilis spores
             V Total coliform-wastewater
             X Rotavirus
                             20
  40       60       80
UVDose (mJ/cm2)
100
           Source: Adapted from Chang et al. (1985)
       Microbial response to UV light can vary significantly among microorganisms. The UV
sensitivity of viruses and bacteriophage can vary by more than two orders of magnitude (Rauth
1965). With bacteria, spore-forming and gram-positive bacteria are more resistant to UV light
than gram-negative bacteria (Jagger 1967). Among the pathogens of interest in drinking water,
viruses are most resistant to UV disinfection followed by bacteria, Cryptosporidium oocysts, and
Giardia cysts.

       UV dose-response is generally independent of the following factors:

       .   UV intensity: In general, UV dose-response follows the Law of Reflectivity over an
          intensity range of 1 - 200 mW/cm2, where the same level of inactivation is achieved
          with a given UV dose regardless of whether that dose was obtained with a high UV
          intensity and low exposure time or vice versa (Oliver and Cosgrove 1975, Rice and
          Ewell 2001). Non-reciprocity has been observed at low intensities where repair may
          compete with inactivation (Sommer et al. 1998, Setlow 1967).

          UV absorbance: UV absorbance of the suspension is considered when calculating
          UV dose. Increasing intensity  or exposure time, however, may be necessary to
          achieve a constant UV dose as the absorbance of a suspension changes.

          Temperature: Temperature effects on dose-response are minimal and depend on the
          microorganism. For male-specific-2 (MS2) bacteriophage, inactivation is not
          temperature-dependent (Malley 2000).  Severin et al. (1983) studied three
          microorganisms to determine the dose required to achieve 2-log inactivation as a
          function of temperature. For E. coli and Candiaparapsilosis, the dose requires
          decreases by less than 10 percent as the temperature increases from 5 to 35 °C, and
UV Disinfection Guidance Manual
For the Final LT2ESWTR
      2-12
          November 2006

-------
                                2. Overview of UV Disinfection
          for f2 bacteriophage, the dose requires decreases by less than 20 percent over the
          same temperature interval (Severin et al. 1983).

          pH: Dose-response is independent of the suspension pH from pH 6 to pH 9 (Malley
          2000).

       Particle association and clumping of microorganisms affects UV dose-response. Small
floe particles can enmesh and protect MS2 bacteriophage, and potentially other viruses, from
exposure to UV light (Templeton et al. 2003). Similarly, the inactivation rate of particle-
associated coliforms is slower than that of non-particle-associated coliforms (Ormeci and Linden
2003). The shielding effect of clumping or particle association can cause a tailing or flattening of
the dose-response curve at higher inactivation levels (Figure 2.8, total coliform curve).

       Several studies have examined the effect of particles on UV disinfection performance.
Research by Linden et al. (2002b) indicated that the UV dose-response of microorganisms added
to filtered drinking waters is not altered by variation in turbidity that meets regulatory
requirements for filtered effluents. For unfiltered waters, source water turbidity up to 10
nephelometric turbidity units (NTU) did not affect the UV dose-response of separately added
(seeded) microorganisms (Passantino et al. 2004, Oppenheimer et al. 2002). The effect of
particle enmeshment on the UV dose-response of seeded microorganisms in water has been
studied by adding clays or natural particles. When coagulating suspensions containing kaolinite
or montmorillonite clay using alum or ferric chloride, no difference was observed in the log
inactivation of the seeded microorganisms  (Templeton et al. 2004, Mamane-Gravetz and Linden
2004). When humic acid particles and a coagulant were added to the suspensions, however,
significantly less inactivation  was achieved (Templeton et al. 2004). Further research is needed
to understand fully the effect of coagulation and particles on microbial inactivation by UV light.
2.3.5  Microbial Spectral Response

       Microbial response varies as a function of wavelength of the UV light. The action
spectrum (also called UV action) of a microorganism is a measure of inactivation effectiveness
as a function of wavelength. Figure 2.9 illustrates the UV action spectrum for three microbial
species and the UV absorbance of DNA as a function of wavelength. Because of the similarity
between the UV action and DNA absorbance spectra and because DNA absorbance is easier to
measure than UV action, the DNA absorbance spectrum of a microorganism is often used as a
surrogate for its UV action spectrum. In Figure 2.9, the scale of the y-axis represents the ratio of
inactivation effectiveness at a given wavelength to the inactivation effectiveness at 254 nm.
UV Disinfection Guidance Manual                2-13                               November 2006
For the Final LT2ESWTR

-------
                                2. Overview of UV Disinfection
     Figure 2.9. Comparison of Microbial UV Action and DMA UV Absorbance
               CD
               o
               CD
               -Q ^
               < m
3.5 -


 3 -


2.5 -


 2
               c JO 1.5
               O CD
               <
                      1 -
                    0.5
                      0
  	MS2- Linden et al. 2001
  -•*- Cryptosporidium - Linden etal. 2001
  -•A- Adenovirus - Malley et al. 2004
  -a-- Herpes Simplex - Linden et al. 2001
  	DMA- Rauth 1965
                       200  210  220  230  240   250   260   270  280  290  300
                                         Wavelength (nm)
             Source: Adapted from Rauth (1965), Linden et al. (2001), and Malley et al. (2004)
       For most microorganisms, the UV action peaks at or near 260 nm, has a local minimum
near 230 nm, and drops to zero near 300 nm, which means that UV light at 260 nm is the most
effective at inactivating microorganisms. Because no efficient way to produce UV light at 260
nm is available and mercury produces UV light very efficiently at 254 nm, however, the latter
has become the standard. Although the action spectrum of various microorganisms is similar at
wavelengths above 240 nm, significant differences occur at wavelengths below 240 nm (Rauth
1965).
2.4    UV Disinfection Equipment

       The goal in designing UV reactors for drinking water disinfection is to efficiently deliver
the dose necessary to inactivate pathogenic microorganisms. An example of UV equipment is
shown in Figure 2.10. Commercial UV reactors consist of open or closed-channel vessels,
containing UV lamps, lamp sleeves, UV sensors, and temperature sensors. UV lamps typically
are housed within the lamp sleeves, which protect and insulate the lamps. Some reactors include
automatic cleaning mechanisms to keep the lamp sleeves free of deposits. UV sensors, flow
meters, and, in some cases, UVT analyzers, are used to monitor dose delivery by the reactor.
This section briefly describes the components of the UV equipment and its monitoring systems.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
2-14
                                                        November 2006

-------
                                 2. Overview of UV Disinfection
                 Figure 2.10. Example of UV Disinfection Equipment
                   Reactor
                   Casing
Temperature
 Sensor
      UV Lamp Housed in
        Quartz Sleevi
                                                 Effluent
                                                  Pipe
            Influent
             Pipe'
                                                    Quartz Sleeve
                                                      Wiper
                                                       Wiper
                                                       Motor
                                        Electrical
                              UV Intensity Connection
                               Sensor    to Lamp
                               Transmittance
                                 Analyzer
                                               Control
                                                Panel
  Source: Courtesy of and adapted from Severn Trent Services
  Note: Not to scale
2.4.1  UV Reactor Configuration

       UV reactors are typically classified as either closed or open channel. Water flows under
pressure (i.e., no free surface) in closed-channel reactors (Figure 2.1 la). Drinking water UV
applications have used only closed reactors to date. Open-channel reactors (Figure 2.1 Ib) are
open basins with channels containing racks of UV lamps and are most commonly used in
wastewater applications.

             Figure 2.11. Examples of UV Reactors: (a) Closed-channel
                                  and (b) Open-channel
                                                 b.
         Source: (a) Courtesy of Calgon Carbon Corporation and (b) Courtesy of WEDECO UV Systems
UV Disinfection Guidance Manual
For the Final LT2ESWTR
           2-15
November 2006

-------
                                2. Overview of UV Disinfection
       UV equipment manufacturers design their UV reactors to provide efficient and cost-
effective dose delivery. Lamp placement, baffles, and inlet and outlet conditions all affect mixing
within a reactor and dose delivery. Individual reactor designs use various methods to optimize
dose delivery (e.g., higher lamp output versus lower lamp output and improved hydrodynamics
through increased head loss).

       The lamp configuration in a reactor is designed to optimize dose delivery. In a reactor
with a square cross-section, lamps are typically placed with lamp arrays perpendicular to flow.
This pattern may be staggered to improve disinfection efficiency. With a circular cross-section,
lamps typically are evenly spaced on one or more concentric circles parallel to flow. However,
UV lamps may be oriented parallel, perpendicular, or diagonal to the flow direction. Depending
on the reactor installation, lamps may consequently be oriented horizontally, vertically, or
diagonally relative to the ground surface. Orienting MP lamps parallel to the ground prevents
overheating at the top of the lamps and reduces the potential for lamp breakage due to
temperature differentials.

       The thickness of the water layer between lamps and between the lamps and the reactor
wall influences dose delivery. If the water layer is too thin, the reactor wall and adjacent lamps
will absorb UV light. If the water layer is too thick, water will pass through regions of lower UV
intensity and  experience a lower UV dose. The optimal spacing between lamps depends on the
UVT of the water, the output of the lamp, and the hydraulic mixing within the reactor.

       The flow through UV reactors is turbulent. Residence times are on the order of tenths of a
second for MP lamps and seconds for LP lamps. In theory, optimal dose delivery is obtained
with plug flow hydraulics through a UV reactor.  In practice, however, UV reactors do not have
such ideal hydrodynamics. For  example, turbulence  and eddies form in the wake behind lamp
sleeves oriented perpendicularly to flow. Some manufacturers insert baffles to improve
hydrodynamics in the reactor. Improvements to the hydraulic behavior of a reactor are often
obtained at the expense of head loss.

       Inlet and outlet conditions can significantly affect reactor hydrodynamics  and UV dose
delivery. For  example, changes in flow direction of 90 degrees at inlets and outlets promote
short-circuiting, eddies, and dead zones within the reactor. Straight inlet configurations with
gradual changes in cross-sectional area will help create flow conditions for optimal dose
delivery.
2.4.2  UV Lamps

       UV light can be produced by the following variety of lamps:

       .   LP mercury vapor lamps

       .   Low-pressure high-output (LPHO) mercury vapor lamps
       .   MP mercury vapor lamps

       .   Electrode-less mercury vapor lamps

       .   Metal halide lamps

UV Disinfection Guidance Manual               2-16                               November 2006
For the Final LT2ESWTR

-------
                               2. Overview of UV Disinfection
          Xenon lamps (pulsed UV)
       .   Eximer lamps
       .   UV lasers
       .   Light emitting diodes (LEDs)

       Full-scale drinking water applications generally use LP, LPHO, or MP mercury vapor
lamps. Therefore, this manual limits discussion to these UV lamp technologies. Table 2.1 lists
characteristics of these lamps, and Table 2.2 lists operational advantages of the lamp types.


               Table 2.1. Typical Mercury Vapor Lamp Characteristics
Parameter
Germicidal UV Light
Mercury Vapor Pressure (Pa)
Operating Temperature (°C)
Electrical Input [watts per
centimeter (W/cm)]
Germicidal UV Output (W/cm)
Electrical to Germicidal UV
Conversion Efficiency (%)
Arc Length (cm)
Relative Number of Lamps
Needed for a Given Dose
Lifetime [hour (hr)]
Low-pressure
Monochromatic at
254 nm
Approximately 0.93
(1.35x10'4psi)
Approximately 40
0.5
0.2
35-38
10-150
High
8,000-10,000
Low-pressure
High-output
Monochromatic at
254 nm
0.18-1.6
(2.6x1 0'5- 2.3x1 0'4psi)
60-100
1.5-10
0.5-3.5
30-35
10-150
Intermediate
8,000-12,000
Medium-pressure
Polychromatic,
including germicidal
range
(200 - 300 nm)
40,000-4,000,000
(5.80- 580 psi)
600 - 900
50 - 250
5-30
10-20
5-120
Low
4,000-8,000
 Note: Information in this table was compiled from UV manufacturer data.
              Table 2.2. Mercury Vapor Lamp Operational Advantages
Low-pressure and Low-pressure High-output
. Higher germicidal efficiency; nearly all output at 254 nm
. Smaller power draw per lamp (less reduction in dose if lamp fails)
. Longer lamp life
Medium-pressure
. Higher power output
. Fewer lamps for a given application
UV Disinfection Guidance Manual
For the Final LT2ESWTR
2-17
November 2006

-------
                               2. Overview of UV Disinfection
       LP, LPHO, and MP lamps consist of the following elements, arranged as shown in
Figure 2.12:

          Lamp Envelope: The envelope of the lamp is designed to transmit germicidal UV
          light, act as an electrical insulator, and not react with the lamp's fill gases. A non-
          crystalline form of quartz, vitreous silica, is often used for the lamp envelope because
          of its high UVT and its resistance to high temperatures. The UVT of the envelope
          affects the spectral output of lamps, especially with MP lamps at lower wavelengths.
          Because of this, lamp envelopes can be made from doped quartz (quartz that is altered
          to absorb specific wavelengths) to prevent undesirable non-germicidal photochemical
          reactions. Envelopes are approximately 1-2 millimeters (mm) thick, and the
          diameter is selected to optimize the UV output and lamp life.

       .   Electrodes: Electrode design and operation are critical for reliable long-term
          operation of lamps. Electrodes promote heat transfer so that lamps can operate at an
          appropriate temperature. The electrodes in LP and LPHO lamps are made of a coil of
          tungsten wire embedded with oxides of calcium, barium, or strontium. In MP lamps,
          electrodes consist of a tungsten rod wrapped in a coil of tungsten wire.

       .   Mercury Fill: The mercury fill present in UV lamps can be in the solid, liquid, or
          vapor phase. Amalgams (alloys of mercury and other metals such as indium or
          gallium in the solid phase) are typically used in LPHO lamps, while LP and MP
          lamps contain liquid elemental mercury. As the lamps heat, the vapor pressure of
          mercury increases. LP and LPHO lamps operate at lower temperatures and have
          lower mercury vapor pressures than MP lamps. In MP lamps,  the concentration of
          mercury in the vapor phase is controlled by the amount of mercury in the lamp. In
          LPHO lamps, an excess of mercury is placed in the lamp, and the amount of mercury
          entering the vapor phase is limited by either a mercury amalgam attached to the lamp
          envelope, a cold spot on the lamp wall, or a mercury condensation chamber located
          behind each electrode.

          Inert Gas Fill: In addition to mercury, lamps are filled with an inert gas (typically
          argon). The inert gas aids in starting the gas discharge and reduces deterioration of
          the electrode. The vapor pressure of the inert gas is typically 0.02 - 1 psi.

       In addition to amalgam LPHO lamps, another method is used to increase the output from
LP lamps. In this application, a standard LP lamp with reinforced filaments is used, allowing for
an increase in current through the lamp. The higher current increases the  output from the lamp.
UV Disinfection Guidance Manual                2-18                              November 2006
For the Final LT2ESWTR

-------
                                 2. Overview of UV Disinfection
                        Figure 2.12. Construction of a UV Lamp
                  LOW-PRESSURE MERCURY LAMP - HOT CATHODE TYPE
                       Tungsten Coil Electrode   Envelope
                      Electrical
                     Connection
                       Mercury & Inert Gas Fill
                                                      X

                                                    End Seal
                  LOW-PRESSURE HIGH-OUTPUT MERCURY LAMP-AMALGAM TYPE
                       Tungsten Coil Electrode   Mercury Ama|gam
                         Electrical
                         Connection
                                                                       I
                     Inert Gas Fill
                                         Envelope
                                                               /
                                                               /
                                                             Seal
                   MEDIUM-PRESSURE MERCURY LAMP
                      Seal                  Electrode - Tungsten Coils on a Tungsten Rod
                 Mercury & Inert Gas Fill
                                         Envelope
                                                     Molybdenum Foil
                         E|ectrica|
                        Connection
2.4.2.1    Lamp Start-up

       As lamps start up, the following series of events occurs to generate an arc (i.e., produce
UV light). First, the electrode emits electrons that collide with the inert gas atoms, causing the
inert gas to ionize. This creates a plasma that allows current to flow, which heats the gas. The
mercury in operating lamps vaporizes in the presence of the hot inert gas, and collisions between
the vapor-phase mercury and high-energy electrons in the plasma cause the mercury atoms to
reach one of many excited states. As the mercury returns from a given excited state to ground
state, energy is released (according to the difference in the state energies) in the wavelength
range of the UV spectrum.
2.4.2.2    Lamp Output

       The light that LP and LPHO lamps emit is essentially monochromatic at 253.7 nm
(Figure 2.13a) in the ultraviolet range and is near the maximum of the microbial action spectrum.
These lamps also emit small amounts of light at 185, 313, 365, 405, 436, and 546 nm due to
higher energy electron transition in the mercury. Lamp output at 185 nm promotes ozone
UV Disinfection Guidance Manual
For the Final LT2ESWTR
2-19
November 2006

-------
                               2. Overview of UV Disinfection
formation. Because ozone is corrosive, toxic, and absorbs UV light, LP and LPHO lamps used in
water disinfection applications are manufactured to reduce the output at 185 nm.

         Figure 2.13. UV Output of LP (a) and MP (b) Mercury Vapor Lamps
1.2 -
 a:
2 •- 0.8 -
O •!-•
a: =
•g f-o.6 -
5-0
O | '
E | 0.2 -
s o.o -
2C
1.2 -]
O
O U>
« §1-0-
> a:
"S .E 0.8
O •!-•
a: =
it0-6'
= £0.4
Q. |
E | 0.2 -
W (0
S 0.0
2C




a. Low-Pressure Lamp










» 250










300
Wavelength





350 400
(nm)
b. Medium-Pressure Lamp





^-^wJ/LJU/V
0 250




ill
AJJUJL






300
Wavelength





I 1





„ A
350 400
(nm)
                  Source: Sharpless and Linden (2001)
       MP lamps emit a wide range of UV wavelengths from 200 to 400 nm (Figure 2.13b). The
combination of free electrons and mercury in the lamp creates a broad continuum of UV energy
below 245 nm. Electron transitions in the mercury cause the peaks in the spectrum.

       All UV lamps also emit light in the visible range. Visible light can promote algal growth
as discussed in Section 2.5.1.5.

       Figure 2.14 shows the output of LP and MP lamps superimposed on the DNA absorption
spectrum. In Figure 2.14, the DNA absorbance is plotted relative to the maximum absorbance in
the range (260 nm), and the lamp outputs are presented on a relative scale. In absolute terms,
however, the intensity and power of LP and MP lamps differ significantly (see Table 2.1 for
more information on lamp operating characteristics).
UV Disinfection Guidance Manual
For the Final LT2ESWTR
2-20
November 2006

-------
                               2. Overview of UV Disinfection
                Figure 2.14. UV Lamp Output and its Relationship to
                             the UV Absorbance of DMA
      250
Wavelength (nm)
                                                          300
         Source: Courtesy of Bolton Photosciences, Inc.


2.4.2.3   Lamp Sensitivity to Power Quality

       A UV lamp can lose its arc if a voltage fluctuation, power quality anomaly, or power
interruption occurs. For example, voltage sags that vary more than 10 - 30 percent from the
nominal voltage for as few as 0.5 - 3 cycles (0.01 - 0.05 seconds) may cause a UV lamp to lose
its arc.

       The most common sources of power quality problems that may cause UV lamps to lose
their arcs are as follows:

       .   Faulty wiring and grounding
       .   Off-site accidents (e.g., transformer damaged by a car accident)
       .   Weather-related damage
          Animal-related damage
       .   Facility and equipment modifications
       .   Starting or stopping equipment with large electrical needs on the same circuit at the
          water plant
       .   Power transfer to emergency generator or alternate feeders

       LP lamps generally can return to full operating status within 15 seconds after power is
restored. LPHO and MP reactors that are more typically used in drinking water applications,
however, exhibit significant restart times if power is interrupted. The start-up time for lamps
should be considered in the design of UV disinfection systems as start-up time can contribute to
off-specification operations (see  Section 3.4.1). The start-up and restart behaviors for LPHO and
MP lamps are summarized  in Table 2.3.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
         2-21
November 2006

-------
                                2. Overview of UV Disinfection
       Table 2.3. Typical Start-up and Restart Times for LPHO and MP Lamps
Lamp Type
LPHO
MP
Cold Start2
total time: 4-7 minutes (min)
(0-2 min warm-up
plus 4-5 min to full power)
total time: 1 - 5 min
(No warm-up or cool down
plus 1 - 5 min to full power4)
Warm Start3
total time: 2-7 min
(0-2 min warm-up
plus 2-5 min to full power)
total time: 4-10 min
(2-5 min cool down
plus 2-5 min to full power4)
           Information shown in table is compiled from Calgon Carbon Corporation, Severn Trent, Trojan, and
           WEDECO. Contact the manufacturer to determine the start-up and restart times for specific equipment
           models.
         2  A cold start occurs when UV lamps have not been operating for a significant period of time.
         3  A warm start occurs when UV lamps have just lost their arcs (e.g., due to voltage sag).
         4  60 percent intensity is reached after 3 min.
         Source: Cotton et al. (2005)
       The effects of temperature can increase or decrease the times listed in Table 2.3 and
should be discussed with the UV manufacturer. Individual manufacturers report that colder water
temperatures (below 10 °C) can result in slower start-ups for LPHO lamps than those listed in
Table 2.3. Conversely, MP manufacturers report shorter restart times with colder temperatures
because the cold water accelerates the condensation of mercury (i.e., cool down), which is
necessary for re-striking the arc.
2.4.2.4   Lamp Aging

       UV lamps degrade as they age, resulting in a reduction in output that causes a drop in UV
dose delivery over time. Lamp aging can be accounted for with the fouling/aging factor
(described in Section 3.4.5) in the design of the UV facility.

       Lamp degradation occurs with both LP and MP lamps and is a function of the number of
lamp hours in operation, number of on/off cycles, power applied per unit (lamp) length, water
temperature, and heat transfer from lamps. The rate of decrease in lamp output often slows as the
lamp ages (Figure 2.15). The reduction in output occurs at all wavelengths across the germicidal
range as shown in Figure 2.16, which is an example of MP lamp output reduction after 8,220
hours of operation.

       Preliminary findings from ongoing research into lamp aging at water and wastewater UV
facilities shows that LPHO and MP lamp aging is non-uniform with respect to axial and
horizontal output and varies greatly from lamp to lamp (Mackey et al. 2005). The lamp aging
study by Mackey et al. is still ongoing, and any future findings from this or other studies should
be evaluated and considered once results are available.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
2-22
November 2006

-------
                              2. Overview of UV Disinfection
  Figure 2.15. Reduction in UV Output of (a) LPHO and (b) MP Lamps Over Time
a. Low-Pressure High Output Mercury Lamps b. Medium-Pressure Mercury Lamps
110% -,
SN 100% -
•4-1
3
5- 90% -
3
o
> 80% -
70% -
C
110% -i
SN 100% -
¥
* « •
* • B- 90% -
» 3
O
> 80% -
7fW

[U i
'hiii
' 'f •• l|"i j(
i nf ii
) 2000 4000 6000 8000 10000 12000 0 2000 4000 6000 8000 10000 12000
Time (Mrs) Time (Mrs)
Source: (a) Adapted from WEDECO, (b) adapted from Linden et al. (2004)
                     Figure 2.16. Lamp Aging for an MP Lamp
         .o
1.6

1.4

1.2

1.0

0.8
         CO
         "CD 0.6
         > 0.4
            0.2
            0.0
                          Ohr
                          8,220 hr
               200      225      250      275      300
                                    Wavelength (nm)
                                                 325
350
       Source: Adapted from Linden et al. (2004)
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                           2-23
November 2006

-------
                                2. Overview of UV Disinfection
       Any deposits on the inner or outer surfaces of the lamp envelope and metallic impurities
within the envelope can absorb UV light and cause premature lamp aging. In LP and LPHO
lamps using UV-transmitting glass, mercury may combine with sodium in the glass to create a
UV-absorbing coating. Electrode sputtering during start-up can also coat the inside surface of the
lamp envelope with tungsten as the lamp ages. The tungsten coating is black, non-uniform,
concentrated within a few inches of the electrode, and can absorb UV light (Figure 2.17). If the
lamps are not sufficiently cooled during operation, electrode material in MP lamps may
evaporate and condense on the inside of the envelope.

                    Figure 2.17. Aged UV Lamp (right) Compared
                                to a New UV Lamp (left)
                               Source: Mackey et al. (2004)
       UV lamp manufacturers can reduce electrode sputtering by designing lamps that pre-heat
the electrode before applying the start voltage, are driven by a sinusoidal current waveform, or
have a higher argon (inert gas) content. Electrode sputtering can be reduced by minimizing the
number of lamp starts during operation.
2.4.3  Ballasts
       Ballasts are used to regulate the incoming power supply at the level needed to energize
and operate the UV lamps. Power supplies and ballasts are available in many different
configurations and are tailored to a unique lamp type and application. UV reactors typically use
magnetic ballasts or electronic ballasts.

       Electronic and magnetic ballasts each have specific advantages and disadvantages. UV
reactor manufacturers consider these advantages and disadvantages when determining what
technology to incorporate into their equipment designs. Electronic and inductor-based magnetic
ballasts can provide almost continuous adjustment of lamp intensity. Most transformer-based
magnetic ballasts, however, allow only step adjustment of lamp intensity. Transformer-based
magnetic ballasts are typically more electrically efficient than inductor-based ballasts but are less
efficient than electronic ballasts. However, higher efficiency and additional features can increase
UV Disinfection Guidance Manual
For the Final LT2ESWTR
2-24
November 2006

-------
                                2. Overview of UV Disinfection
the electronic ballast cost. UV lamps that are powered by magnetic ballasts tend to have more
lamp end-darkening (i.e., electrode sputtering) and have shorter lives compared to lamps
powered by electronic ballasts due to the higher frequencies used by electronic ballasts.
Electronic ballasts are generally more susceptible to power quality problems (Section 2.4.2.3)
compared to magnetic ballasts; however, the power quality tolerances of both ballast types
depend on the electrical design. A comparison of magnetic and electronic ballast technologies is
shown in Table 2.4.
             Table 2.4. Comparison of Magnetic and Electronic Ballasts
               Magnetic Ballast
                Electronic Ballast
  Less expensive
  Continuous power adjustment occurs with
  inductor-based magnetic ballast (but not with
  transformer-based magnetic ballast)
  More resistant to power surges
  Proven technology (in use for nearly 70 years)
  Greater separation distance allowed between the
  UV reactor and control panel
    Continuous power adjustment and ability to
    adjust to lower power levels (e.g., 30 %)
    More power efficient
    Lighter weight and smaller size
    Allows for longer lamp operating life and less
    lamp end-darkening
2.4.4  Lamp Sleeves

       UV lamps are housed within lamp sleeves to help keep the lamp at optimal operating
temperature and to protect the lamp from breaking. Lamp sleeves are tubes of quartz (vitreous
silica) that are open at one or both ends. The sleeve length is sufficient to include the lamp and
associated electrical connections. The sleeve diameter is typically 2.5 - 5.0 cm for LP and LPHO
lamps and 3.5 - 10.0 cm for MP lamps. The distance between the exterior of the lamp and
interior of the lamp sleeve is approximately 1 cm. The positioning of the UV lamp along the
length of the sleeve can vary, depending on reactor configuration. Lamp sleeves absorb some UV
light (Figure 2.18), which may influence dose delivery by the reactor.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
2-25
November 2006

-------
                                2. Overview of UV Disinfection
                    Figure 2.18. DVT of Quartz that is 1 mm Thick
                          at a Zero-degree Incidence Angle
g

D

100 -
90 -
80 -
70 •

60 •
50 -
2(

^^^
f



)0 220 240 260 280 300 320 340 360 380 400
Wavelength (nm)
                      Source: GE Quartz (2004a)
       The lamp sleeve assemblies are sealed to prevent water condensation within the sleeve
and contain any ozone formed between the lamp envelope and lamp sleeve. Components within
the sleeve should withstand exposure to UV light, ozone,  and high temperatures. If the
components are not made of the appropriate material, UV light exposure can cause component
deterioration and off-gassing of any impurities present in  the quartz sleeve. Off-gassed materials
can form UV-absorbing deposits on the inner surfaces of  the lamp sleeve. Off-gassing and ozone
formation are of greater concern with MP lamps because  they operate at a higher temperature
and emit low-wavelength ozone-forming UV light. Off-gassing can be minimized through proper
manufacturing of the lamp  sleeves.

       Lamp sleeves are vulnerable to fractures. Fractures can occur from internal stress and
external mechanical  forces  such as wiper jams, water hammer, resonant vibration, and impact by
objects. Fractures may also occur if lamp sleeves are not handled properly when removed for
manual cleaning. Most lamp sleeves are designed to withstand continuous positive pressures of
at least 120 pounds per square inch gauge (psig) (Roberts 2000, Aquafine 2001, Dinkloh 2001).
However, pressures of negative 1.5 psig have been shown to adversely affect sleeve integrity
(Dinkloh 2001). Section 4.1.4  discusses design considerations to reduce the potential for
pressure-related incidents. If a lamp sleeve fractures while in service, water can enter the sleeve.
The temperature difference between the hot lamp and cooler water may  cause the lamp to break.
Lamp breaks are undesirable due to the potential for mercury release. Appendix E discusses the
lamp sleeve and lamp breaks. The tolerance level of the sleeve depends  on the quality of the
quartz and the sleeve's thickness and length.

       Lamp sleeves can also  foul, decreasing the UVT of the lamp sleeve. Fouling on the
internal lamp sleeve surface arises from the deposition of material from  components within the
lamp or sleeve due to temperature and exposure to UV light. The UV reactor manufacturer can
control internal lamp sleeve fouling through appropriate material selection. For example, some
UV reactors using LP or LPHO lamps have sleeves made of Teflon® or  Teflon-coated quartz.
Teflon sleeves have a lower UVT, however, and their transmittance reduces faster than quartz
UV Disinfection Guidance Manual
For the Final LT2ESWTR
2-26
November 2006

-------
                                2. Overview of UV Disinfection
sleeves without Teflon. Deposition of compounds in the water on the lamp sleeve surface cause
fouling on external surfaces. A combination of thermal effects and photochemical processes
causes the external fouling (Derrick and Blatchley 2005). Some compounds that may contribute
to fouling are discussed in Section 2.5.1. External fouling can be removed by cleaning.

       Solarization can also decrease the UVT of the sleeve. Solarization is photo-thermal
damage to the quartz that increases light scattering and attenuation (Polymicro Technologies
2004). Quartz solarizes if exposed to prolonged high energy radiation such as UV light.
Resistance to this type of Solarization increases as the purity of the quartz increases. Solarization
on quartz can be reversed by heating the quartz to about 500 °C (GE Quartz 2004b).
2.4.5  Cleaning Systems

       UV reactor manufacturers have developed different approaches for cleaning lamp
sleeves, depending on the application. These approaches include off-line chemical cleaning
(OCC), on-line mechanical cleaning (OMC), and on-line mechanical-chemical cleaning (OMCC)
methods.

       For OCC systems, the reactor is shut down, drained, and flushed with a cleaning solution.
Solutions used to clean lamp sleeves include citric acid, phosphoric acid, or a solution the UV
reactor manufacturer provides that is consistent with National Sanitation Foundation
International/American National Standards Institute (NSF/ANSI) 60 Standard (Drinking Water
Treatment Chemicals - Health Effects). The reactor is filled with the cleaning solution for a time
sufficient to dissolve the substances fouling the sleeves (approximately 15 minutes), rinsed, and
returned to operation. The entire cleaning cycle typically lasts approximately 3 hours.
Alternatively, instead of rinsing the UV reactor with a cleaning solution, the sleeves can be
removed and manually cleaned. Some LPHO UV equipment uses OCC systems. The frequency
of OCC can range from monthly to yearly and depends on the site-specific water quality and
degree and frequency of fouling.

       OMC and OMCC systems use wipers that are attached to  electric motors or pneumatic
piston  drives. In OMC systems, mechanical wipers may consist of stainless steel brush collars or
Teflon rings that move along the lamp sleeve (Figure 2.19a). In OMCC  systems, a collar filled
with cleaning solution moves along the lamp sleeve (Figure 2.19b). The wiper physically
removes fouling on the lamp sleeve surface while the cleaning solution within the collar
dissolves fouling materials.
UV Disinfection Guidance Manual                 2-27                              November 2006
For the Final LT2ESWTR

-------
                                2. Overview of UV Disinfection
             Figure 2.19. Examples of (a) Mechanical Wiper System and
                       (b) Mechanical-chemical Wiper System
         Source: (a) Courtesy of Infilco Degremont, (b) Courtesy of Trojan Technologies
       Draining the reactor is unnecessary when mechanical and mechanical-chemical wipers
are used. Therefore, the reactor can remain on-line while the lamp sleeves are cleaned. MP
equipment typically uses OMC or OMCC systems because the higher lamp temperatures can
accelerate fouling under certain water qualities. The cleaning frequency for these OMC and
OMCC systems ranges from 1-12 cycles per hour (Mackey et al. 2004).
2.4.6  UV Sensors

       UV sensors measure the UV intensity at a point within the UV reactor (Figure 2.20) and
are used with measurements of flow rate and, potentially, UVT to indicate UV dose delivery.
The measurement responds to changes in lamp output due to lamp power setting, lamp aging,
lamp sleeve aging, and lamp sleeve fouling. Depending on sensor position, UV sensors may also
respond to changes in UVT of the water being treated. UV sensors comprise optical components,
a photodetector, an amplifier, its housing, and an electrical connector. The optical components
may include monitoring windows, light pipes, diffusers, apertures, and filters. Monitoring
windows and light pipes deliver light to the photodetector. Diffusers and apertures reduce the
amount of UV light reaching the photodetector, thereby reducing the sensor degradation that UV
light causes. Optical filters modify the spectral response such that the sensor responds only to
germicidal wavelengths (i.e., 200 - 300 nm). Verification of sensor performance is described in
Chapter 5.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
2-28
November 2006

-------
                                2. Overview of UV Disinfection
                      Figure 2.20. Example of a Dry UV Sensor
                              Mounted on a UV Reactor
                   Source: Courtesy of WEDECO
       UV sensors can be classified as dry or wet. Dry sensors monitor UV light through a
monitoring window, whereas wet UV sensors directly contact the water flowing through the
reactor. Monitoring windows and the wetted ends of wet sensors can foul over time and may
require cleaning similar to lamp sleeves.
2.4.7  UVT Analyzers

       As stated previously, UVT is an important parameter in determining UV dose delivery.
UVT analyzers are essential if UVT is part of the dose-monitoring strategy (see Section 2.4.9 for
a discussion of dose monitoring approaches). If UVT is not part of the dose-monitoring strategy,
analyzers may be provided for the purpose of monitoring water quality and helping to diagnose
operational problems. Several commercial UV reactors use the measurement of UVT to calculate
UV dose in the reactor and, if necessary, change lamp output or the number of energized lamps
to maintain appropriate UV dose delivery.

       Two types of commercial  on-line UVT analyzers are available. One analyzer calculates
UVT by measuring the UV intensity at various distances from a lamp. This type of analyzer is
schematically displayed in Figure 2.21. In this analyzer, which is external to the UV reactor, a
stream of water passes through a cavity containing an LP lamp with three UV sensors located at
various distances from the lamp. The difference in sensor readings is used to calculate UVT.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
2-29
November 2006

-------
                                2. Overview of UV Disinfection
                     Figure 2.21. Example DVT Analyzer Design
                                                         UV Intensity
                                                          Sensor
                                    Inlet
                                              Outlet
                 Source: Courtesy of Severn Trent Services
       The other type of on-line UVT analyzer is a flow-through spectrophotometer that uses a
monochromatic UV light source at 253.7 nm. The instrument measures the A254 and calculates
and displays UVT.
2.4.8  Temperature Sensors

       The energy input to UV reactors that is not converted to light (approximately 60 - 90
percent, depending on lamp and ballast assembly) is wasted as heat. As it passes through a
reactor, water can absorb the heat, keeping the reactor from overheating. Nevertheless,
temperatures can increase when either of the following events occurs:

       .   Water level in the reactor drops and lamps are exposed to air.
       .   Water stops flowing in the reactor.

       UV reactors can be equipped with temperature sensors that monitor the water temperature
within the reactor. If the temperature is above the recommended operating range, the reactor will
shut off to minimize the potential for the lamps to overheat. Because of the high operating
temperature of MP lamps, dissipating heat can be more difficult than in reactors that use LP or
LPHO lamps. As such, UV reactors with MP lamps typically have temperature  sensors; however,
reactors with LP or LPHO lamps may not because of the lower lamp operating temperature.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
2-30
November 2006

-------
                                2. Overview of UV Disinfection
2.4.9  UV Reactor Dose-Monitoring Strategy

       The dose-monitoring strategy establishes the operating parameters used to confirm UV
dose delivery. This guidance manual focuses on UV reactors that use one of these two strategies,
described below. Other existing dose-monitoring strategies or new strategies developed after this
manual is published, however, may also be suitable for reactor operations provided they meet
minimum regulatory requirements.6
       1.  UV Intensity Setpoint Approach. This approach relies on one or more "setpoints"
          for UV intensity that are established during validation testing to determine UV dose.
          During operations, the UV intensity as measured by the UV sensors must meet or
          exceed the setpoint(s) to ensure delivery of the required dose. Reactors must also be
          operated within validated operation conditions for flow rates and lamp status [40 CFR
          141.720(d)(2)]. In the UV Intensity Setpoint Approach, UVT does not need to be
          monitored separately. Instead, the intensity readings by the sensors account for
          changes in UVT. The operating strategy can be with either a single setpoint (one UV
          intensity setpoint is used for all validated flow rates) or a variable setpoint (the UV
          intensity setpoint is determined using a lookup table or equation for a range of flow
          rates).

       2.  Calculated Dose Approach. This approach uses a dose monitoring equation to
          estimate the UV dose based on the flow rate, UV intensity, and UVT, as measured
          during reactor operations.  The dose monitoring equation may be developed by the
          UV manufacturers using numerical methods; however, EPA recommends that water
          systems use an empirical dose monitoring equation developed through validation
          testing. During reactor operations, the UV reactor control  system inputs the measured
          parameters into the dose monitoring equation to produce a calculated dose. The water
          system operator divides the calculated dose by the Validation Factor (see Chapter 5
          for more details on the Validation Factor) and compares the resulting value to the
          required dose for the target pathogen and log inactivation  level.

       The dose-monitoring strategies are described in more detail in Section 3.5.2. Any dose
monitoring strategy must be evaluated during reactor validation (as described in Section 5.1),
and the outputs measured during validation will be part of the monitoring requirements described
in Section 6.4.1 [40 CFR 141.720(d)].
2.5    Water Quality Effects and Byproduct Formation

       Constituents in the water to be treated can affect the performance of UV disinfection.
Additionally, all disinfectants can form byproducts, and the goal of the overall disinfection
process is to maximize disinfection while controlling byproduct formation. This section
6 At a minimum, water systems must monitor flow rate, lamp status, and UV intensity plus any other parameters
required by the State to show that a reactor is operating within validated conditions [40 CFR 141.720(d)(3)(i)].

UV Disinfection Guidance Manual                2-31                               November 2006
For the Final LT2ESWTR

-------
                                2. Overview of UV Disinfection
discusses water quality characteristics affecting UV disinfection performance and the byproducts
that may form during the UV disinfection process.
2.5.1  Effect of Water Quality on UV Reactor Performance

       UVT, particle content, upstream water treatment processes, constituents that foul reactor
components, and algae affect the performance of UV reactors. These effects can be adequately
addressed through proper design of the UV disinfection equipment. The design guidelines are
discussed in Section 3.4.
2.5.1.1   UVT

       UVT has a strong effect on the dose delivery of a UV reactor. As UVT decreases, the
intensity throughout the reactor decreases, which reduces the dose the reactor delivers. UV
reactors are typically sized to deliver the required UV dose under specified UVT conditions for
the application. Section 3.4.4.1 discusses approaches for selecting the UVT for UV facility
design.

       UV absorbers in typical source waters include soluble and particulate forms of humic and
fulvic acids; other aromatic organics (e.g., phenols); metals (e.g., iron); and anions (e.g., nitrates
and sulfites) (Yip and Konasewich 1972, DeMers and Renner 1992). UV absorbance will vary
over time due to changing concentrations of these compounds and seasonal effects—rainfall,
lake stratification and destratification (turnover),  and changes in biological activity of
microorganisms within the water source.

2.5.1.2   Particle Content

       As described in Section 2.3.4, particle content can also affect UV disinfection
performance. Particles in source waters are diverse in composition and size and include large
molecules, microorganisms, clay particles, algae, and floes. Sources of particles include
wastewater discharges, erosion, runoff, microbial growth, and animal waste. The particle
concentration will vary over time both seasonally and over the short term.  Storm events, lake
turnover, and spring runoff are some events that increase the concentration of particles.
2.5.1.3   Upstream Water Treatment Processes

       Unit processes and chemical addition upstream of UV reactors can significantly affect
UV reactor performance because they can change the particle content and UVT of the water.
Additionally, when UV disinfection is used in combination with another disinfectant, synergistic
disinfection potentially may occur (i.e., the combination of disinfectants may be more effective
than either disinfectant acting alone).

       Water treatment processes upstream of the UV reactors can be operated to maximize
UVT, thereby optimizing the design and costs of the UV reactor (Section 3.2.2). For example,
UV Disinfection Guidance Manual                2-32                               November 2006
For the Final LT2ESWTR

-------
                                2. Overview of UV Disinfection
coagulation, flocculation, and sedimentation remove soluble and particulate material, and
filtration removes particles. Activated carbon absorption also reduces soluble organics.

       Adding oxidants (such as chlorine and ozone) can increase the UVT (APHA et al. 1998)
by degrading natural organic matter, reducing soluble material, and precipitating metals. An
example of the effect ozone has on decreasing UV absorbance is shown in Figure 2.22. Ozone is
also a strong absorber of UV light, however, and will decrease the UVT if an ozone residual is
present in significant concentrations in the water passing through a UV reactor. Quenching
agents that do not absorb UV light (such as sodium bisulfite) can be used to destroy the ozone
residual upstream of the UV reactors. Thiosulfate is not recommended as a quenching agent
because it absorbs UV light and can decrease the UVT.
            Figure 2.22. Example Effect of Ozonation on UV Absorbance
                   if Ozone is Quenched Prior to UV Disinfection
                   0.5
                   0.4 -
                
-------
                                 2. Overview of UV Disinfection
Whether the effects of multiple disinfectants are synergistic (i.e., more inactivation observed
when processes are used in combination than is expected from the sum of
                Table 2.5. UV Absorbance Characteristics of Common
                              Water Treatment Chemicals
Compound1
Ozone (O3) (aqueous)
Ferric iron (Fe3+)
Permanganate (MnO4~)
Thiosulfate (S2O32~)
Hypochlorite (CIO")
Hydrogen peroxide (H2O2)
Ferrous iron (Fe2+)
Sulfite (S032-)
Zinc (Zn2+)
Molar Absorption
Coefficient
(IVT1 cm'1)
3,250
4,716
657
201
29.5
18.7
28
16.5
1.7
Impact Threshold
Concentration2
(mg/L)
0.071
0.057
0.91
2.7
8.4
8.7
9.6
23
187
             The following chemicals were also evaluated in the same study (Bolton et al. 2001)
             and were found to have no significant absorbance: ammonia (NH3), ammonium ion
             (NH4+), calcium ion (Ca2+), hydroxide ion (OH"), magnesium ion (Mg2+), manganese
             ion (Mn2+), phosphate species, and sulfate ion (SO4")
           2 Concentration in mg/L resulting in UVT decrease from 91 % to 90 % (A254 increase
             from 0.041 cm"1 to 0.046 cm"1)
           Source: Adapted from Bolton et al. (2001)


the disinfectants acting alone) is currently under debate. Two studies reported synergistic effects
when using UV disinfection and free chlorine, monochloramine, or chlorine dioxide (Ballester et
al. 2003, Lotierzo et al. 2003), while others did not observe synergism (Coronell et al. 2003,
Oppenheimer et al.  2003). The importance of the sequence of the disinfectants is also a subject of
debate. Ballester et  al. (2003) obtained improved disinfection with UV disinfection followed by
monochloramine addition than with chloramination followed by UV disinfection, while the
sequence of disinfectants did not affect the disinfection effectiveness in the  study by Lotierzo et
al. (2003).
2.5.1.4    Fouling Potential

       Compounds in the water can foul the external surfaces of the lamp sleeves and other
wetted components (e.g., monitoring windows of UV sensors) of UV reactors. Fouling on the
lamp sleeves reduces the transmittance of UV light through the sleeve into the water, thereby
reducing the output from the UV lamp into the water. Also, fouling on the monitoring windows
affects measured UV intensity and dose monitoring. Sleeve fouling can be accounted for with the
fouling/aging factor (described in Section 3.4.5) in the design of the UV facility.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
2-34
November 2006

-------
                                2. Overview of UV Disinfection
       Hardness (as CaCO3), alkalinity, temperature, ion concentration, oxidation reduction
potential (ORP), and pH all influence the rate of fouling and, subsequently, the necessary
frequency of sleeve cleaning. Fouling can occur for the following reasons:

          Compounds for which the solubility decreases as temperature increases may
          precipitate [e.g., CaCO3, CaSO4, MgCO3, MgSO4, FePO4, FeCO3, A12(SO4)3]. These
          compounds will foul MP lamps faster than LP or LPHO lamps because MP lamps
          operate at higher temperatures.

       .   Photochemical reactions that are independent of sleeve temperature may cause sleeve
          fouling (Derrick 2005).

          Compounds with low solubility may precipitate [e.g., Fe(OH)3, A1(OH)3].

       .   Particles may deposit on the lamp sleeve  surface due to gravity settling and
          turbulence-induced collisions (Lin et al. 1999a).

       .   Organic fouling can occur when a reactor is left off and full of water for an extended
          period of time (Toivanen 2000).

          Inorganic constituents can oxidize and precipitate (Wait et al. 2005).

       Fouling rate kinetics has been reported as independent of time following a short induction
period (Lin et al. 1999b). Depending on the water quality and UV lamp type, significant fouling
may occur in hours or take up to several months.

       Pilot studies lasting 5-12 months using UV reactors with LP,  LPHO and MP lamps
found that standard cleaning protocols and wiper frequencies (1 - 12 cleaning cycles per hour)
were sufficient to overcome the effect of sleeve fouling with water that had total and calcium
hardness levels less than 140 milligrams per liter (mg/L) and iron less than 0.1 mg/L (Mackey et
al. 2001, Mackey et al. 2004).

       Inorganic fouling is a complex process, however, and is not related only to hardness and
iron concentrations. The solubility of inorganic constituents depends on whether they are in an
oxidized or reduced state, which can be affected by both the ORP and pH of the water (Wait et
al. 2005). ORP is a measurement of the water's ability to oxidize or reduce constituents in the
water. Both pH and ORP are needed to predict the oxidation state of an inorganic constituent.
Studies have found that fouling rates increase as ORP increases (Collins and Malley 2005, Wait
et al. 2005, Derrick 2005).  In some waters with high ORP, however, fouling rates can be
minimized if the iron and manganese are removed through oxidation, precipitation, and filtration
(Wait et al. 2005, Derrick 2005, Jeffcoat 2005). Although ORP can provide valuable
information, measuring it can be challenging and may not be possible in all instances.

       Ultimately, the fouling potential is difficult to predict, but standard cleaning equipment
can remove fouling and may need to be included. Also, pilot-scale or demonstration-scale testing
can determine  the fouling tendencies and cleaning regime if the PWS is concerned about fouling.
UV Disinfection Guidance Manual                2-35                               November 2006
For the Final LT2ESWTR

-------
                                2. Overview of UV Disinfection
2.5.1.5   Algal Occurrence and Growth

       The presence of algae in the water being treated may reduce UVT and interfere with the
UV disinfection process. Algae also may grow upstream or downstream of UV reactors, which
has been observed in MP pilot studies (Mackey et al. 2004). Visible light emitted from the lamps
is transmitted through water farther than germicidal wavelengths. Algal growth depends on the
concentration of nutrients in the water, hydraulics (i.e., dead spaces), and the amount of visible
light transmitted beyond the reactor.
2.5.2  Chlorine Reduction through UV Reactors

       When UV disinfection is applied to water with a free or total chlorine residual, some
reduction of the residual may occur. The reduction in free chlorine residual is proportional to the
delivered dose and independent of flow rate (Brodkorb and Richards 2004). The reduction in
total chlorine residual is also proportional to the delivered dose (Wilczak and Lai, 2006). The
reduction in chlorine residual further depends on the chlorine species, UV light source, and water
quality characteristics (Ormeci et al. 2005, Venkatesan et al. 2003). An example of the effect of
UV light on the free chlorine residual is shown in Figure 2.23.  In other evaluations, a loss of
about 0.3 mg/L of the free chlorine residual was observed in a WTP at a dose between 80 and
120 mJ/cm2 (Kubik 2005), and a loss of 0.2  mg/L of the total chlorine residual was observed in
bench-scale testing at doses up to 40  mJ/cm2 (Wilczak and Lai 2006).
                  Figure 2.23. Example Effect of UV Disinfection on
                             Free Chlorine Residual Loss
00
.0
j~ n 7
^ u. /
0)
£ n R
«±, u.o
t/>
(/> n c
O u-b
_l
n\ n A
Q) \tA
C
£ n •* -
o u.o
.c
On o
\j.£.
0)
2 n 1 -
i °-1
On






^
**




•^
^
V*
1*



/
/
^



*
/
S







-•-1.72mgd
-•-3.33 mgd
4.83 mgd
5.68 mgd
-*-7.00 mgd
-•-7.93 mgd

.u \
0 50 100 150 200
Calculated MS-2 Reduction Equivalent Dose
(mJ/cm2)
      Source: Brodkorb and Richards (2004)
UV Disinfection Guidance Manual
For the Final LT2ESWTR
2-36
November 2006

-------
                                2. Overview of UV Disinfection
2.5.3  Byproducts from UV Disinfection

       Studies indicate that UV disinfection at UV doses up to 200 ml/cm2 do not change the
pH, turbidity, dissolved organic carbon level, UVT, color, nitrate, nitrite, bromide, iron, or
manganese of the water being treated (Malley et al. 1996). Byproducts from UV disinfection,
however, can arise either directly through photochemical reactions or indirectly through
reactions with products of photochemical reactions. Photochemical reactions occur only when a
chemical species absorbs UV light and the resulting excited state reacts to form a new species.
The resulting concentration of new species depends on the concentration of the reactants and the
UV dose. In drinking water, research on potential byproducts of UV disinfection has focused on
the effect of UV light on the formation of halogenated DBFs after subsequent chlorination, the
transformation of organic material  to more degradable components, and on the potential
formation of other DBFs (e.g., biodegradable compounds, nitrite, mutagenicity, and other
byproducts).

2.5.3.1   Trihalomethanes, Haloacetic Acids, and Total Organic Halides

       Trihalomethanes (THMs) and haloacetic acids (HAAs) are two categories of halogenated
DBFs that EPA currently regulates. UV light at doses less than 400 mJ/cm2 has not been found to
significantly affect the formation of THMs or HAAs upon subsequent chlorination (Malley et al.
1996, Kashinkunti et al. 2003, Zheng et al. 1999, Liu et al. 2002, Venkatesan et al. 2003).
2.5.3.2   Biodegradable Compounds

       Several studies have shown low-level formation of non-regulated DBFs (e.g., aldehydes)
as a result of applying UV light at doses greater than 400 m J/cm2 to wastewater and raw drinking
water sources (Liu et al. 2002, Venkatesan et al. 2003). At the doses typical for UV disinfection
in drinking water (< 140 mJ/cm2), however, no significant change was observed (Kashinkunti et
al. 2003). UV disinfection has not been found to significantly increase the assimilable organic
carbon (AOC) of drinking water at UV doses ranging from 18 - 250 mJ/cm2 (Kruithof and van
der Leer 1990, Akhlaq et al. 1990, Malley et al. 1996).
2.5.3.3   Nitrite

       The conversion of nitrate to nitrite is possible with MP lamps that emit at wavelengths
below 225 nm [von Sonntag and Schuchmann (1992), Mack and Bolton (1999), IJpelaar et al.
(2003), Peldszus et al. (2004)]. Sharpless and Linden (2001) reported a conversion rate from
nitrate to nitrite of approximately 1 percent. Therefore, the nitrate-to-nitrite conversion is
unlikely to be a significant issue for PWSs under current regulations. The nitrate levels would
have to be higher than the nitrate MCL of 10 mg/L for the nitrite MCL of 1 mg/L to be exceeded.
UV Disinfection Guidance Manual                2-37                              November 2006
For the Final LT2ESWTR

-------
                                   2. Overview of UV Disinfection
                                This page intentionally left blank.
UV Disinfection Guidance Manual                  2-38                                 November 2006
For the Final LT2ESWTR

-------
                    3.  Planning Analyses for UV Facilities
       This chapter provides information on the elements that should be addressed during the
UV disinfection planning or preliminary design phase.
        Chapter 3 covers:
        3.1   UV Disinfection Goals
        3.2   Evaluating Integration of UV Disinfection into the Treatment Process
        3.3   Identifying Potential Locations for UV Facilities
        3.4   Defining Key Design Parameters
        3.5   Evaluating UV Reactors, Dose Monitoring Strategy, and Operational
             Approach
        3.6   Assessing  UV Equipment Validation Issues
        3.7   Assessing  Head Loss Constraints
        3.8   Estimating UV Facility Footprint
        3.9   Preparing Preliminary Costs and Selecting the UV Facility Option
        3.10 Reporting  to the State
       The planning for any UV facility is site-specific. Given the wide range of possible
treatment scenarios, a guidance document such as this one cannot address or anticipate all
possible treatment conditions. The information presented here should be used within the context
of sound engineering judgment and applied appropriately on a case-by-case basis. Appendix F
presents case studies that illustrate how various public water systems (PWSs) have implemented
UV disinfection in their water systems. Additionally, this manual was written with the
understanding that UV technology will continue to expand and evolve, so the information
presented is current only as of the publication date. Furthermore, unless otherwise stated,
throughout Chapter 3 the water to be disinfected is assumed to be from surface water systems
[(i.e., filtered water, an unfiltered source water, or groundwater under the direct influence
(GWUDI)], meeting applicable regulatory requirements that pre-date the Long Term 2 Enhanced
Surface Water Treatment Rule (LT2ESWTR).

       The process of planning and designing a UV facility is presented in Figure 3.1. Once the
design parameters are defined and the implementation issues are identified, they are incorporated
into the detailed design phase, which is discussed in Chapter 4.
3.1    UV Disinfection Goals

       The first step in planning a UV disinfection facility is to define the goals for the facility
as part of a comprehensive disinfection strategy for the entire treatment process. Additionally,
the target pathogen(s), target log-inactivation, and corresponding required UV dose should be
identified.

UV Disinfection Guidance Manual             3-1                                  November 2006
For the Final LT2ESWTR

-------
                               . Planning Analyses for UV Facilities
              Figure 3.1. Example Flowchart for Planning UV Facilities
     Define the goals of
     the UV facility and
     identify the target
     mi croorgani sm(s)
     and log-
     inactivation
     Section 3.1
Evaluate
integration of UV
disinfection into
the treatment
process
Section 3.2
     Report to
     Primacy
     Agency
     Section 3.10
Define the key design
parameters
  • Water quality
  • Fouling/aging factor
  • Flow rate
  • Power quality
Section 3.4
                                                                           T
                                                                 Evaluate potential UV
                                                                 equipment and dose-
                                                                 monitoring strategy
                                                                 Section 3.5
                                     Evaluate UV reactor
                                     validation issues
                                     Section 3.6
                                                                 Determine head loss
                                                                 constraints
                                                                 Section 3.7
                                                                 Estimate UV facility
                                                                 footprint for all
                                                                 potential locations and
                                                                 UV reactor
                                                                 configurations
                                                                 Section 3.8
                                                                  Calculate costs for
                                                                  installation options,
                                                                  compare options, and
                                                                  select facility location
                                                                  Section 3.9
                                                                 Design UV facility
                                                                 Chapter 4
UV Disinfection Guidance Manual
For the Final LT2ESWTR
             3-2
             November 2006

-------
                               . Planning Analyses for UV Facilities
          Comprehensive Disinfection Strategy: A comprehensive disinfection strategy
          provides multiple barriers to reduce microbial risk, while minimizing disinfectant
          byproduct (DBF) formation. UV disinfection is a tool that can contribute to a
          comprehensive disinfection strategy by providing a cost-effective method of
          inactivating pathogens that are more resistant to traditional disinfection methods.
          Also, UV disinfection can replace chemicals for primary disinfection of chlorine-
          resistant pathogens (e.g., Cryptosporidium and Giardia), thereby reducing DBF
          formation. Note that PWSs that plan to significantly change their disinfection process,
          including adding UV disinfection, must prepare a disinfection benchmark1 (40 CFR
          141.708) and consult with the state before making any changes. Further, PWSs must
          continue to provide 2-log Cryptosporidium removal  by meeting filtered water
          turbidity requirements  (40 CFR 141.173 for PWSs serving at least 10,000 people and
          40 CFR 141.551 for PWSs serving fewer than  10,000 people) unless they meet the
          filtration avoidance criteria.

          Target Pathogen and Log Inactivation: The required UV doses for
          Cryptosporidium  and Giardia inactivation are lower than those needed to inactivate
          viruses. (See Table 1.4.) Accordingly, the capital and operational costs for
          inactivating Cryptosporidium and Giardia should be lower than for viruses. One
          study estimated capital costs for Cryptosporidium and Giardia inactivation by UV
          disinfection on a log removal basis to be about half the cost associated with the UV
          inactivation of viruses  (Cotton et al. 2002). Additionally, most viruses can be easily
          inactivated with chlorine so UV disinfection for virus inactivation may not be
          necessary. The target log inactivation also should be considered because higher target
          inactivation requires higher UV doses that will affect the design  and cost of the UV
          facility. Therefore, the target microorganism(s) and their log-inactivation level should
          be determined early in the planning process.
3.2    Evaluating Integration of UV Disinfection into the Treatment Process

       When installed, UV disinfection will typically be one of several treatment processes to
help meet water quality goals. Accordingly, UV disinfection should be evaluated in the context
of the complete treatment process, and the impacts on UV disinfection on other treatment
processes should be considered. These issues are summarized in this section.
3.2.1  UV Disinfection Effects on Treatment

       Typically, UV disinfection cannot entirely replace chemical disinfectants used in the
treatment process. Some of the reasons are listed below.

           Surface water systems must maintain a disinfectant residual in the distribution system
           (40 CFR 141.72).
1  More information on completing a disinfection benchmark can be found in Disinfection Profiling and
  Benchmarking Guidance Manual (EPA 1999).

UV Disinfection Guidance Manual             3-3                                   November 2006
For the Final LT2ESWTR

-------
                               . Planning Analyses for UV Facilities
       .   UV disinfection is not as efficient in inactivating viruses as more traditional, chlorine-
          based disinfection processes.

          Chemical disinfectants may also be needed to oxidize other constituents present in the
          water (e.g., iron, manganese, or taste- and odor-causing compounds).

       .   Some water systems apply chlorine to reduce algal growth in sedimentation basins.

       Consequently, some level of chlorine-based disinfectant (chlorine or chloramines) usually
will be needed even when UV disinfection is implemented. Therefore, any reduction in chlorine-
based disinfectants should be evaluated in the context of other water quality and treatment goals.

       When UV disinfection is applied to water having a chlorine residual, some chlorine
residual reduction may  occur, depending on the UV dose, chlorine species, UV light source, and
water quality characteristics (Brodkorb and Richards 2004, Ormeci et al. 2005, Venkatesan et al.
2003). Brodkorb and Richards (2004) reported chlorine residual reduction between 0.1 and
0.7 milligrams per liter (mg/L) at a wide range of UV doses (described in Section 2.5.2).
Significant chlorine reduction could occur inadvertently if the UV equipment cannot provide
enough power modulation capacity and actually operates at much higher doses than designed.
Two options are available to avoid chlorine reduction by UV disinfection:

       1.  Consider moving the chlorine addition point to after the UV facility if possible,
          especially when targeting viruses (because their required UV doses are higher).

       2.  Procure the UV equipment that has adequate power modulation to prevent overdosing
          and subsequent chlorine reduction.

       In addition, UV disinfection of water having a chlorine residual, which results in a higher
oxidation-reduction potential  (ORP), could result in sleeve fouling (Section 2.5.1.4) if iron or
manganese are present even at low levels and a proper cleaning system is not in place (Malley et
al. 2001). Several studies have shown that fouling occurs at iron levels below the secondary
maximum contaminant level (SMCL) when the water has a high oxidation-reduction potential
(ORP) (Collins and Malley  2005, Derrick 2005, Wait et al. 2005). Again, moving the point of
chlorination to after the UV facility can possibly reduce sleeve fouling (Section 3.4.4.2).
Alternatively, oxidation and removal of iron and manganese (e.g., by adding potassium
permanganate upstream of the sedimentation basin) reduces the fouling potential.
3.2.2  Upstream Treatment Process Effect on UV Disinfection

       Water treatment processes upstream of the UV reactors can be operated to maximize the
ultraviolet transmittance (UVT), thereby optimizing the design and costs of the UV equipment
(Section 3.4.4.1). For example, coagulation, flocculation, and sedimentation remove soluble and
particulate material, and optimizing coagulation for organics removal will increase the UVT,
which could reduce the UV facility costs. Also, upstream chemicals may affect UV disinfection
performance as described in Sections 2.5.1.3 and 3.4.4.1.
UV Disinfection Guidance Manual             3-4                                   November 2006
For the Final LT2ESWTR

-------
                               . Planning Analyses for UV Facilities
3.3    Identifying Potential Locations for UV Facilities

       The UV dose tables (see Table 1.4) in the LT2ESWTR apply to post-filter applications of
UV disinfection in filtration plants and to unfiltered systems that meet filtration avoidance
criteria. In general, installing UV disinfection prior to filtration in conventional water treatment
plants (WTPs) is not recommended because of the potential particle interference in raw and
settled waters. As such, only post-filter locations are discussed for filtered systems in this
section.

       After the potential locations are identified,  design criteria, hydraulics, validation issues,
and footprint estimations should be evaluated at each location to identify which location is most
feasible for the UV facility. These evaluations are  described in subsequent sections.
3.3.1  Installation Locations for Filtered Systems

       In conventional WTPs, the three most common installation locations are downstream of
the combined filter effluent (upstream of the clearwell), on the individual filter effluent piping
(upstream of the clearwell), and downstream of the clearwell.
3.3.1.1   Combined Filter Effluent Installation (Upstream of the Clearwell)

       A combined filter effluent installation is defined as the application of UV disinfection to
the filtered effluent after the effluent from individual filters has been combined (as opposed to
applying UV disinfection to the individual filter effluents) and ahead of the clearwell, as shown
in Figure 3.2. For retrofits on  existing WTPs, these installations are usually housed in a separate
building.
           Figure 3.2. Schematic for UV Facility Upstream of the Clearwell
                 Source
                 Water
                              Rapid
                              Mix


~^-


*--

/v
                                      Sedimentation
                                        Basin
             Filters
                         UV
                      Disinfection
Clearwell
           To
        Distribution
         System
       This type of design and installation has several advantages:

       .   The UV reactor operation is largely independent of the operation of individual filters,
          which provides flexibility for design and operation.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
3-5
     November 2006

-------
                                . Planning Analyses for UV Facilities
           If the entire UV facility failed, a WTP can continue to disinfect by adding a chemical
           disinfectant to the clearwell. (Note that backup chemical disinfection will likely not
           provide Cryptosporidium inactivation.)

           Surge and pressure fluctuations typically are not a concern for this installation
           location unless membrane filtration, pressure filters, or intermediate booster pumps
           are used.

       .   Because this type of UV facility is typically constructed in a new building, there may
           be greater flexibility to maintain the recommended inlet and outlet hydraulic
           conditions for the UV reactors (Section 3.6.2).

       The primary disadvantages of this type of installation are:

       .   An additional building and space may be necessary.

           The piping and fittings may result in greater head loss than alternative configurations,
           which may result in the need for intermediate booster pumps.
3.3.1.2    Individual Filter Effluent Piping Installation

       Individual filter effluent piping installations are defined as UV reactors installed on each
filter effluent pipe (Figure 3.3). This type of installation is typically located within an existing
filter gallery.
        Figure 3.3. Schematic of Individual Filter Effluent Piping Installation
                                      in Filter Gallery
            Source
            Water
                         Rapid     Flocculation
                          Mix     Sedimentation
                                    Basin
                                                                              HSPs
                                       To
                                                                      Clearwell  Distribution
                                                                               System
UV Disinfection Guidance Manual
For the Final LT2ESWTR
3-6
November 2006

-------
                               . Planning Analyses for UV Facilities
       The primary advantages of this type of installation are:

       .   A new building is not necessary, which will decrease construction costs.

           The hydraulic effect of the UV facility is less because the only additional head loss is
           from the UV reactors (most necessary valves and appurtenances are already present in
           the filter gallery).

           If the UV reactors fail, a WTP can continue to disinfect by adding a chemical
           disinfectant to the clearwell. (Note that backup chemical  disinfection likely will not
           provide Cryptosporidium inactivation.)

       This installation location, however, has several disadvantages:

       .   Many filter galleries have insufficient space within existing effluent piping to
           accommodate the UV reactors.

       .   Sufficient space is needed in the filter gallery or nearby for the control panels and
           electrical equipment.

           Access to existing equipment may be impeded by the UV reactor, and access to UV
           reactor components for maintenance may be more restricted than for a combined filter
           effluent installation.

       .   Environmental conditions (e.g., moisture) in the filter gallery may not be appropriate
           for the installation of the UV reactors, associated control panels, and electrical
           equipment. This situation would necessitate improvements to the heating, ventilating,
           and air conditioning (HVAC) system.

       .   The existing piping may constrain how the UV reactor is validated because of the
           unique inlet and outlet conditions that may be present (Section 3.6.2).

           Surge and pressure fluctuations would need to be investigated if UV reactors are
           installed directly downstream of pressure filters  or membrane filtration because water
           hammer can damage lamp sleeves.

       Additionally, the individual filter effluent installation may also complicate treatment
plant operations and limit operational flexibility, as described below:

           In general, this option increases the number of UV reactors required compared to a
           combined filter installation because the number of filters dictates the number of UV
           reactors. More reactors may increase operation and maintenance costs.

           The head loss of the UV reactors may affect the  operation of the filters and the
           clearwell.

       .   The operations of the UV reactor and the filter are closely related. If one reactor or
           one filter is off-line, the other process may not be operable.

UV Disinfection Guidance Manual             3-7                                   November 2006
For the Final LT2ESWTR

-------
                               . Planning Analyses for UV Facilities
          When a UV reactor goes off-line, the corresponding filter also should be taken off-
          line to minimize off-specification operation.

          The filter backwash cycle can complicate UV reactor operation.

          -   Lamps that remain energized during a backwash may require cooling water
              because some lamps should not be energized in stagnant water. The designer
              should consult the UV manufacturer to determine whether the UV reactor
              requires cooling water during start-up.

          -   If a UV reactor is off-line during a backwash, the UV reactor may be operating
              outside of its validated limits (i.e., off-specification—discussed in Section 3.4.1)
              if water is being treated during lamp warm-up. If the piping configuration permits,
              energizing the UV reactors during the filter-to-waste period and having the filter-
              to-waste water pass through the reactors during the warm-up period would cool
              the lamps and reduce the volume of the off-specification water.
3.3.1.3   UV Disinfection Downstream of the Clean/veil

       A WTP may be able to locate the UV facility downstream of the clearwell, either
upstream or downstream of the high-service pumps (HSPs), as shown in Figure 3.4. In many
WTPs, the HSPs pump water directly from the clearwell, which limits space and the availability
of suitable piping for installing the UV facility upstream of the HSPs. Installation downstream of
the HSPs may provide greater space and flexibility in locating the UV facility.
          Figure 3.4. UV Disinfection Downstream of High Service Pumps
              Source
              Water
                          Rapid   Flocculation
                           Mix    Sedimentation
                                    Basin
Filters
          Clearwell
                        UV
                     Disinfection
Distribution
 System
       The primary advantage of this type of installation is that UV reactor installation is
possible even if the space or available head is insufficient to allow installation of the UV
equipment between the filters and the clearwell. However, these options have significant
disadvantages:

       .   UV facilities located downstream of the clearwell may experience greater fluctuations
          in flow rate because the flow rate is more closely related to demand changes.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                             November 2006

-------
                               . Planning Analyses for UV Facilities
          Accommodating flow rate fluctuations may necessitate increasing the UV reactor size
          or number of UV reactors.

          Post-clearwell installation locations are more prone to water hammer because of their
          proximity to the HSPs and subsequent high pressures, and water hammer could
          damage lamp sleeves and the lamps. Hydropneumatic tanks or pressure-relief valves
          may be needed to avoid water hammer.

          In the event of a lamp break, post-clearwell installations may have less ability to
          contain mercury and quartz resulting from the break in a low-velocity collection area
          (depending on the distribution system configuration).

          In post-HSP installations, the water is at distribution system pressure.  The UV reactor
          housing may need reinforcement to accommodate high pressure, which would
          increase the cost of the UV reactors.

          A UV facility located after the HSPs will reduce the discharge pressure to the
          distribution system, and a UV facility located between the clearwell and HSPs will
          reduce the suction head available for the pumps. As a result, discharge pressures and
          storage utilization could be affected at these two locations unless the HSPs are
          upgraded to account for the UV facility hydraulic needs.

          When UV disinfection is applied to water with a free or total  chlorine residual, some
          reduction of the residual may occur, which may necessitate increasing the chlorine
          dose in the clearwell or moving the chlorination point to downstream of the UV
          facility.
3.3.2  Unfiltered System Installation Locations

       In an unfiltered system, UV facilities can be located either before or after a storage
reservoir. If the storage is covered, UV disinfection facilities can be installed in either location. If
the storage reservoir is uncovered, however, the PWS is subject to the uncovered reservoir
requirements of the LT2ESWTR and as such should install UV disinfection on the discharge side
of the reservoir to provide the necessary treatment. Most unfiltered systems flow to the
distribution system by gravity; however, water hammer may still be a concern if the facility is
located near HSPs (if applicable). This installation location is similar to installations downstream
of the clearwell, and as such, the items described in Section 3.3.1.3 also apply to this location.

       More debris may be present in the influent to UV reactors in unfiltered applications than
in post-filter applications. Debris entering the UV reactor with sufficient momentum can cause
the lamp and sleeve to break. The mass and size of an object that might cause damage are
installation-specific and depend on UV reactor configuration (e.g., horizontal versus vertical
reactor orientation) and water velocity through the reactor. Methods of addressing debris are
described in Section 4.5.1, and additional information on lamp breakage is presented in
Appendix E.
UV Disinfection Guidance Manual             3-9                                   November 2006
For the Final LT2ESWTR

-------
                               . Planning Analyses for UV Facilities
3.3.3  Groundwater System Installation Locations

       For groundwater applications of UV disinfection, UV facilities may be installed either at
each well in a production system or at a centralized facility. If installed at or near well pumps,
the hydraulic and water hammer considerations described in Section 3.3.1.3 will also apply. An
engineering cost analysis can be conducted to compare centralized versus wellhead UV
disinfection treatment, as well as any other treatment needs, such as removing iron, manganese,
or sulfides.
3.3.4  Uncovered Reservoir Installation Locations

       The LT2ESWTR requires PWSs with uncovered finished water storage facilities to either
cover the storage facility or treat the discharge of the storage facility that is distributed to
consumers to achieve inactivation and/or removal of 4-log virus, 3-log Giardia, and 2-log
Cryptosporidium [40 CFR 141.714(c)]. When applying UV disinfection to uncovered reservoirs,
the UV facility should be on the outlet of the uncovered reservoir. In some cases, the inlet and
outlet to the uncovered reservoir is the same pipe, and the UV facility should be designed so it
operates when the water flows from the uncovered reservoir to the customer. Water from most
uncovered reservoirs flows by gravity to the distribution system; however, water hammer may
still be a concern if the UV reactors are located close to HSPs. As such, the items described in
Section 3.3.1.3 also apply to this location.
3.4    Defining Key Design Parameters

       Off-specification requirements (see Section 3.4.1 below), target pathogen inactivation,
flow, water quality, the fouling/aging factor, and power quality affect the sizing of the UV
reactors and associated support facilities. Specifically, UV manufacturers use the design flow,
design UVT, the range of UVT expected, and the fouling/aging factor to determine the
appropriate number of UV reactors to achieve the required UV dose.

       Pilot- and demonstration-scale testing for UV disinfection systems can be helpful in
determining key design parameters but typically are unnecessary. For example, pilot- or
demonstration-scale testing may be warranted when bench-scale analysis cannot determine the
design criteria (e.g.,  prediction of fouling/aging factor in waters with high inorganic
constituents). This section also describes some pilot- or demonstration-scale testing that can be
used to determine key design criteria if deemed necessary by the PWS or design engineer.
3.4.1  Off-specification Requirements

       The LT2ESWTR requires validation of UV reactors to demonstrate that they achieve the
required UV dose [40 CFR 141.720(d)]. Validation testing establishes the conditions under
which the UV reactors must be operated to ensure the required UV dose delivery [40 CFR
141.720(d)].
UV Disinfection Guidance Manual             3-10                                  November 2006
For the Final LT2ESWTR

-------
                                . Planning Analyses for UV Facilities
       Receiving log inactivation credit to meet the treatment requirement of the LT2ESWTR
requires that at least 95 percent of the water delivered to the public during each month is treated
by UV reactors operating within validated limits [40 CFR 141.720(d)(3)]. In other words, the
UV reactors cannot be operated outside of their validated limits for more than 5 percent of the
volume of water that is treated each month. Operating outside of the validated limits is defined in
this manual as off-specification operation.

       Determining the appropriate design criteria related to flow, water quality (UVT and
fouling), the fouling/aging factor, and power quality is important to comply with LT2ESWTR
off-specification requirements. These design criteria also define the conditions under which the
UV reactors must be validated and then operated. If the design parameters are not sufficiently
conservative, the UV reactors may often operate off-specification and be out of compliance.

       The UV reactors are off-specification when any of the following conditions occur:

           The flow rate is higher than the validated range.

       .    The UVT is lower than the validated range [if the Calculated Dose Approach is used
           (see Section 3.5.2)].

           The UV intensity is below the validated setpoint [if the UV Intensity Setpoint
           Approach  is used (see Section 3.5.2)].

           The validated dose 2 is less than the required UV dose at  a given flow rate [if the
           Calculated Dose Approach is used (see Section 3.5.2)].

       .    One or more lamps are not energized unless the UV reactor was validated with these
           lamps off.

       .    All UV lamps are off because of a power interruption or power quality problem, and
           water is flowing through the reactors.

           One or more UV sensors are not within calibration criteria, and the remedial actions
           are not taken. (See Section 6.4.1.1).

           A UVT analyzer is needed for the dose-monitoring strategy; the UVT  analyzer is out
           of calibration; and a corrective action was not taken. (See Section 6.4.1.2.)

       .    The UV equipment includes installed or replaced components (or both) that are not
           equal to or better than the components used during validation testing unless the UV
           equipment was re-validated.  (See Section 5.13.)
2 For the purposes of this manual, the "Validated Dose" is the UV dose in units of mJ/cm2 delivered by the UV
  reactor as determined through validation testing. The validated dose is compared to the required dose to determine
  log inactivation credit. For the Calculated Dose Approach, the validated dose equals the calculated dose from the
  dose-monitoring equation, divided by the Validation Factor. The Validation Factor accounts for key uncertainties
  and biases resulting from validation testing.
UV Disinfection Guidance Manual             3-11                                   November 2006
For the Final LT2ESWTR

-------
                              . Planning Analyses for UV Facilities
3.4.2  Target Pathogen Inactivation and Required UV Dose

       As described in Section 3.1, the UV facility design criteria should include the target
pathogen, log inactivation level, and corresponding required UV dose. The required UV dose
(DReq) for the various pathogens and inactivation are shown in Table 1.4; however, the PWS may
consider increasing the required dose beyond those listed in Table 1.4 by 10 to 20 percent to
provide flexibility and conservatism. Similar approaches are commonly used by many PWSs
with chlorine disinfection where they provide higher chlorine residuals and contact times (CT)
than required.
3.4.3  Design Flow Rate

       The UV facility design criteria should identify the average, maximum, and minimum
flow rates that the UV reactors will experience. Methods for determining the design flow rate for
the installation locations described previously are listed in Table 3.1.
               Table 3.1. Potential Method to Determine Design Flow
Installation Location
Combined Filter Effluent
Individual Filter Effluent
Downstream of the Clean/veil
Unfiltered Application
Groundwater Application
Uncovered Reservoir Application
Design Flow Basis
Combined rated capacity of all duty filters1
Rated design flow for individual filter
Rated capacity of the HSP station
Rated capacity of the treatment facility
Rated capacity of the well pump or well field
Maximum reservoir outflow
      1 Does not include redundant filters


3.4.4 Water Quality

      As highlighted in Chapter 2, the following water quality parameters and issues affect UV
dose delivery and should be considered in UV facility planning:

      .   UVT at 254 nanometers (nm)

      .   UV transmittance scan from 200 - 300 nm (i.e., germicidal range)

      .   Sleeve and UV sensor window fouling, including
          -   Calcium
          -   Alkalinity
          -   Hardness
          -   Iron
          -   Manganese
          -   pH
UV Disinfection Guidance Manual
For the Final LT2ESWTR
3-12
November 2006

-------
                               . Planning Analyses for UV Facilities
          -   Lamp temperature
          -   ORP

       .   Particle content and algae (unfiltered and uncovered reservoir applications)

       Water quality data should be collected from locations that are representative of the
potential UV facility location(s). The duration of sampling, numbers of samples collected, and
data analyses used to evaluate water quality for UV disinfection are similar to the approaches
used for other water treatment technologies. The data collection should capture typical water
quality and any water quality variation due to storm events, reservoir turnover, seasonal changes,
source water blends,  and variations in upstream treatment. The data collection frequency should
be based on flow rate variability, the consistency of the source and treated water qualities, and
the potential for obtaining cost and energy savings by refining the design criteria. The extent of
water quality data to  be collected and the data analysis should be left to the discretion of the
PWS and the design  engineer based on experience and professional judgment.

       Water quality information should be communicated to the UV manufacturers, so they can
determine the applicable UV reactors for the target pathogen inactivation. This section provides
more details on the data collection and analysis recommendations.
3.4.4.1   UVT and UVT Scans

       The most important water quality characteristic affecting UV facility design is UVT3'4
because the UVT of the water directly influences UV dose delivery, as discussed in Chapter 2.
Overly conservative design UVT values (i.e., low UVT) can result in over-design and increased
capital costs. Conversely, inappropriately high design UVT values can result in frequent UV
reactor off-specification operation, which could violate LT2ESWTR requirements.

       Quantifying both a design UVT and the full range of UVT expected during operation is
essential. Understanding the full range of UVT is critical because the UV reactor should be
validated for the range  of UVT and flow combinations expected at the WTP to avoid off-
specification operation. Specifying a matrix of flow and UVT conditions for the UV reactors to
meet the required UV dose may be appropriate. Also, the UV manufacturers may use the UVT
range at the WTP to help determine the turndown (i.e., power modulation) needs of the UV
reactors.

       This section discusses the issues with using existing UVT data and describes the data
collection, UVT measurement, and data analysis that can be used to determine design UVT and
UVT range. Table 3.2 summarizes the recommendations for collecting and analyzing UVT data.
3 UVT in this section implies UVT measurement specifically at 254 nm and 1 cm pathlength unless otherwise noted.
4         (UVT(%)}
 4** = -log   .~~   I
          I  100  I
UV Disinfection Guidance Manual             3-13                                  November 2006
For the Final LT2ESWTR

-------
                                . Planning Analyses for UV Facilities
              Table 3.2. Summary of UVT Data Collection and Analysis1
Issue
Water Quality Events to
Capture in Data Collection
Water Quality Sampling
Locations
Sample Type for Various
Installation Options2
Collection Frequency and
Period
Existing Data for Potential
Use
Recommended Data
Analysis
Recommended Data to
Provide to UV
Manufacturer
Recommendation
• Typical/average water quality conditions
• Rainfall effects on source water
• Reservoir turnover
• Seasonal variations
• Possible water quality blends if multiple source waters are used
• Variation in upstream water treatment
Locations that are representative of potential UV facility location(s)
• Composite samples from operating filters or grab samples from the
combined filtered water header should be collected for combined filter
effluent installations
• Grab samples from representative filter(s) for individual filter piping
effluent installations
• Grab samples from any locations downstream of clean/veil under
consideration
• Weekly for 1 - 2 months if water quality is stable
• Weekly3 for 6 - 12 months (or more) if water quality changes seasonally
A254 is often collected in filtered waters to determine the specific UV
absorbance (SUVA), and these measurements could be used in the data
analysis. However, ultraviolet light absorbance at 254 nm (A2§4) is typically
filtered for the SUVA calculation, which would bias the A2§4 low (high UVT).
Therefore, such data should only be used with this understanding.
• Cumulative frequency analysis
• UVT occurrence with flows
• Matrix of flows with corresponding UVTs
• Target pathogen(s) and log inactivation
• Design UVT4 (corresponding to design flow)
• Range of operating UVTs
  Existing A254 or UVT data may be available, which would reduce the sampling and analysis needed.
  ! The potential installation locations are described in detail in Section 3.3.1.
  ! More frequent samples may be needed to capture a water quality event (e.g., storm events).
  1 The design UVT is the UVT that will typically occur at the location of the facility.
Availability of Existing UVT Measurements

       UVT data collection may not be necessary if sufficient filtered water UVT data are
available to perform the recommended data analysis described subsequently. Additionally,
filtered water A254 is often collected to determine the SUVA, and these measurements could be
used in the data analysis. However, the water sample is typically passed through a 0.45-
micrometer (|J,m) filter for the A254 measurement needed for the SUVA calculation, which may
bias the A254 low (high UVT). If the only available A254 measurements are on water that has been
passed through a 0.45- |j,m) filter, they can still provide input to the planning process, but
additional UVT data collection may be necessary to understand the magnitude of the bias.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
3-14
November 2006

-------
                               . Planning Analyses for UV Facilities
Data Collection

       UVT measurements should be collected from locations that are representative of the
potential facility location(s). UVT data can be collected using grab or composite samples,  and
the type of sample collected depends on the potential UV facility locations under consideration.
For example, composite samples from operating filters or a grab sample from a combined  filter
effluent header should be collected for combined filter effluent UV facilities. For individual filter
effluent pipe installations, grab samples from representative filters at the beginning and the end
of filter runs are recommended. Grab  samples from any location(s) downstream of the clearwell
under consideration should be collected.

       As with most engineering designs, the larger the data set, the more refined the design
UVT can be. If UVT data are not available, weekly UVT measurement is recommended, but the
duration  of the sampling period depends on the source water quality. For example, a PWS with
very stable UVT measurements may need only one or two months of data. A PWS that
experiences  seasonal changes, however, would benefit from more frequent data collection during
seasonal  events and over a longer period (6 to 12 months or more). If seasonal UVT decreases
occur regularly, increased sampling frequency (e.g., daily) during these periods will better
capture the magnitude and duration of the decreases. The possible effect of upstream processes
on UVT  should be assessed by collecting UVT data during the various operating conditions (e.g.,
a range of alum doses). If different sources or combinations of sources are used during the year,
the UVT of the potential source water blends should be characterized to properly identify the
representative water quality conditions.

UVT Measurement

       UVT can be measured with a bench-top spectrophotometer or can be continuously
measured by an on-line UVT analyzer. During planning, UVT is typically measured using a
spectrophotometer and is  typically reported as a percent. The wavelength of the
spectrophotometer should be set to 254 nm, and the pathlength of the quartz cuvette used to
measure UVT is usually 1 centimeter  (cm). If the UVT is high, however, longer pathlengths can
be used to improve measurement resolution. When longer pathlengths are used, the A254
measured on the spectrophotometer should be normalized by the specific pathlength to calculate
the A254 on a per cm basis, and then the UVT should be calculated based on the A254 with the
converted 1-cm pathlength. Because particles can affect the absorbance  of UV light, samples  for
UVT should not be passed through a 0.45-|j,m filter before analysis. The sample pH also should
not be adjusted.

Data Analysis

       A cumulative frequency diagram of the UVT data can help the PWS determine its  design
UVT value and will also illustrate the UVT range. Cumulative frequency diagrams can be
prepared by  ranking UVT results from lowest to highest and then calculating the percentile for
each value. Figure 3.5 presents an example cumulative frequency diagram for three filtered
waters; the cumulative frequency percentile (x-axis) shows the percentage of the dataset that is
less than a given value of UVT over the data collection period. For example, if the 90th percentile
UVT is 91 percent, then 90 percent of the measurements are greater than 91 percent, and 10
percent of the UVT  measurements are less than 91 percent.

UV Disinfection Guidance Manual             3-15                                  November  2006
For the Final LT2ESWTR

-------
                              . Planning Analyses for UV Facilities
       In Figure 3.5, the UVT data for Filtered Waters 1, 2, and 3 display different
characteristics. Filtered Water 1 has a relatively stable UVT, while Filtered Waters 2 and 3 have
gradually increasing cumulative frequency slopes that indicate greater variability. Selection of an
appropriate UVT design value for these waters should consider the variability in UVT and flow
values and the maximum allowable volume of off-specification finished water at different UVT
design levels. The water supply's preferred level of conservatism should also be taken into
account in this comparison.
   Figure 3.5. Example Cumulative Frequency Diagram for Three Filtered Waters
i nnn/
QS% *=!
*



N

o> r
u Q0% 	

g
£ 85%

































*
k






*















(C























4
*








A






X
J














X






4

I






K
































X













+

X







)
A































i












4


j































X







a












+


X







X







X







X
fc


+ Filtered Water 1 Data Collected from May 1991 to July 2001

x Filtered Water 2 Data Collected from January 1991 to March 1998
* Filtered Water 3 Data Collected from February 1993 to July 2001

























X
A







4




+4


A!l
s












i
:fc

X
i
\
T









K

Si









&

»'








-H
*s
















%
1
KF
1
s


;


t





0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Cumulative Frequency Percentile

       Additionally, the minimum operating UVT may not correspond to the period with the
highest flow rates. The relationship between seasonal flow rates and UVT data should be
considered when selecting a design UVT value and the matrix of UVT and flow conditions to be
defined for the UV manufacturer. Figure 3.6 presents flow rate and UVT variations and seasonal
patterns for Filtered Water 3. For this example WTP, the low UVT typically occurs in September
and October and not during the high flow rate period in the summer. In this example, the
following conditions for UVT and flow could be communicated to the UV manufacturers, so
they can determine the applicable UV reactors for the required UV dose:

A 90th-percentile design UVT value of 86 percent at the design 220-million gallons per day
(mgd) capacity
Minimum UVT of 83 percent coupled with a flow of 140 mgd
UV Disinfection Guidance Manual
For the Final LT2ESWTR
3-16
November 2006

-------
                               . Planning Analyses for UV Facilities
              Figure 3.6. Example Flow Rate and UVT (at 254 nm) Data
                    Filtered Water 3
                    WTP Capacity: 220 mgd
                                                                         79
Upstream Treatment Chemicals Effect on UVT

       As described in Section 2.5.1.3 and Bolton et al. (2001), the following chemicals alone
will not significantly affect UVT under typical filtered water conditions: alum, aluminum,
ammonia, ammonium, zinc, phosphate, calcium, hydroxide, ferrous iron (Fe+2), hypochlorite
(CIO"), ferric iron (Fe+3), and permanganate. However, ozone residual affects UVT, as described
below. If other chemicals of concern are present, the effect of water treatment chemicals on UV
absorbance can be assessed by preparing solutions of various concentrations and measuring their
UV absorbance using a standard spectrophotometer.

       If ozone is added before UV disinfection, the UVT of the water can be increased
measurably, thereby improving the efficiency of UV disinfection. Ozone also absorbs UV light,
however, so if residual ozone enters the UV reactor, the resulting decrease in UVT can be
significant and should be considered when determining the design UVT. To address this issue,
PWSs can monitor the ozone residual and add an ozone-reducing chemical prior to the UV
reactor to maintain the ozone residual below a specified setpoint value. Several  chemicals can
quench ozone, but some (such as sodium thiosulfate) also have a high UV absorbance value and
can decrease UVT. Such chemicals should not be used prior to UV disinfection unless their
application causes no residual concentration. Sodium bisulfite is an alternative to sodium
thiosulfate that does not significantly affect UVT.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
3-17
November 2006

-------
                               . Planning Analyses for UV Facilities
UVTScans

       If MP lamps are being considered, measuring the UVT at the wavelengths in the
germicidal range (in addition to 254 nm) may also be important. A UVT scan is used to
determine the UVT of the water over 200 - 300 nm (i.e., germicidal range). In a UVT scan, the
absorbance at each wavelength is measured and converted to UVT using Equation 2.2 (%UVT
= 100 x 10"A). The UV absorbance of water typically decreases with increasing wavelength over
the germicidal range. Thus, the UV light attenuation in a UV reactor and the corresponding
disinfection performance depend  on the absorbance at each emitted wavelength. Some UV
manufacturers use site-specific UVT scans in their UV dose monitoring and control systems.
UVT scans can also vary seasonally; therefore, UVT scans could be measured at different times
during the year to account for this variation. Also, the UVT scans can be used to determine the
appropriate UV-absorbing chemical for validating the UV reactors that will be installed.
3.4.4.2   Water Quality Parameters That Affect Fouling

       Water quality can affect the amount and type of lamp sleeve fouling that occurs in UV
reactors. The factors that affect fouling pertain to all UV equipment.

       Fouling is typically caused by precipitation of compounds on the lamp sleeve, as
described in Section 2.5.1.4. The rate of fouling and the consequent frequency of sleeve cleaning
depend on ORP, hardness, alkalinity, lamp temperature, pH, and the presence of certain
inorganic constituents (e.g., iron and calcium). If significant seasonal shifts in any of the
parameters or coagulant doses are expected, the duration of the monitoring period should be
sufficiently long to capture the variations.

       Although fouling should not be a significant problem for most PWSs, the water quality
parameters listed below should be monitored before the UV facility is designed, unless adequate
water quality data are available. A summary of the data collection and analysis related to fouling
parameters is provided in Table 3.3. Providing these data to UV manufacturers is recommended
to help them qualitatively assess the fouling potential for their UV reactors and to assist
designers in determining whether a particular cleaning system should be specified. These data
will also help determine the fouling/aging factor, which is discussed in Section 3.4.5. (Note that
ORP can be challenging to measure, so the data collected may have limited value.)

        •   Calcium

        •   Alkalinity

        •   Hardness
        •   Iron

        •   Manganese
        •   pH

        •   ORP
UV Disinfection Guidance Manual             3-18                                  November 2006
For the Final LT2ESWTR

-------
                               . Planning Analyses for UV Facilities
            Table 3.3. Summary of Fouling Data Collection and Analysis
Issue
Collection Location
Collection Frequency2 and Period
Recommended Data Analysis
Recommended Data to Provide to
UV Manufacturer
Fouling Parameters1
Locations that are representative of potential UV
facility location(s)
• Monthly for 1 -2 months if water quality is stable
• Monthly for 6 - 1 2 months (or more) if water quality
changes seasonally
Based on design engineer's and PWS' best
professional judgment
Median and maximum values
        Fouling parameters include calcium, alkalinity, hardness, iron, manganese, pH, and ORP.
        More frequent samples may be necessary to capture a water quality event (e.g., storm events).
       Pilot tests of waters with total hardness levels less than 140 mg/L and iron less than 0.1
mg/L found that standard cleaning protocols and wiper frequencies (one sweep every 15-60
minutes) addressed the effect of sleeve fouling at the sites tested (Mackey et al. 2001, Mackey et
al. 2004). Recent research has shown, however, that the addition of a chemical oxidant directly
upstream of UV reactors (i.e., downstream of filters) will increase the ORP and potential for
fouling (Derrick 2005, Wait et al. 2005). Therefore, moving the chemical oxidation point from
immediately upstream of the UV reactors to downstream of the UV reactors should be
considered to reduce the potential for fouling. It should be noted that if oxidation and filtration
occur prior to UV disinfection, the iron and manganese are typically oxidized and then filtered
out prior to the UV reactor, and fouling will be minimal (Derrick 2005, Wait et al. 2005, Jeffcoat
2005).

       If the ORP, pH, and inorganic constituent concentrations are low, fouling is not likely to
be an issue, and a cleaning system may not be necessary. However, a cleaning system should be
considered if iron and manganese are present. Also, if the chemical oxidation point cannot be
moved from immediately upstream of the UV equipment and iron and manganese are present,
pilot testing (Section 3.4.5.1) may be necessary to determine the fouling rate and effectiveness of
sleeve  cleaning.
3.4.4.3   Additional Water Quality Considerations for Unfiltered Supplies and
          Treatment of Uncovered Reservoir Water

       Water supplies are susceptible to variable water quality, turbidity spikes, reservoir
turnover, and seasonal algal blooms. Typically, water treatment processes at filtered WTPs
dampen the effects of such variations on UV disinfection. Unfiltered supplies, however,
generally do not have upstream treatment that mitigates these variations. Specifically, the
presence of particles and algae may affect UV dose delivery, and water quality and UVT may
UV Disinfection Guidance Manual
For the Final LT2ESWTR
3-19
November 2006

-------
                               . Planning Analyses for UV Facilities
fluctuate more in unfiltered supplies and thus should be a consideration in the water quality data
analysis.
       Uncovered reservoirs have similar water quality issues as unfiltered supplies. In most
cases, however, the problems are less severe because the water has been treated before it enters
the uncovered reservoir and the operation of uncovered reservoirs is more controlled (e.g.,
smaller volumes, storm water control, concrete lining,  and bird control). One exception is that
algal blooms may be more prevalent in uncovered reservoirs than in unfiltered supplies if
phosphate-based corrosion inhibitors are added at the WTP. Phosphates can promote algal
growth.

       Issues that should be considered in the water quality data analysis for unfiltered supplies
and uncovered reservoirs are described in this section and summarized in Table 3.4.

       Table 3.4. Summary of Particle and Algal Data Collection and Analysis
Issue
Collection Location
Collection Frequency1 and Period
Recommended Data Analysis
Recommended Data to Provide to
UV Manufacturer
Particles and Algae
Locations that are representative of potential UV facility
location(s)
• Monthly for 1 - 2 months for an Unfiltered PWS
• Bi-weekly for the summer months2 for Uncovered
Reservoirs
Based on design engineer's and PWS' best professional
judgment
Median and maximum values
       More frequent samples may be needed to capture a water quality event (e.g., storm events).
      1 Algal blooms often occur in summer months in uncovered reservoir supplies.
Water Quality Fluctuations from Reservoir Turnover

       Reservoir turnover in unfiltered supplies and uncovered reservoirs may cause water
quality changes that affect UV disinfection. The UVT and parameters that affect fouling should
be monitored over a complete reservoir cycle to account for these issues in the design criteria.
For example, reservoir turnover can cause increased iron levels, which is a factor that should be
considered when assessing fouling potential. If the potential for increased iron levels is not
assessed, the appropriate sleeve cleaning technology may not be installed, and UV dose delivery
may be affected.

Particle Content and UVT Variability

       For unfiltered systems, the Surface Water Treatment Rule (SWTR) allows turbidity up to
5 nephelometric turbidity units (NTU) immediately prior to the first point of disinfection
application (40 CFR 141.71). Storm-related turbidity spikes are more prevalent in unfiltered
supplies than in filtered supplies because no upstream treatment is available to remove the
particles. Particles in water absorb and scatter UV light to varying degrees based on their size
and composition. Particles affect the disinfection process in two ways:
UV Disinfection Guidance Manual
For the Final LT2ESWTR
3-20
November 2006

-------
                               . Planning Analyses for UV Facilities
       1.  Particles can decrease the UVT of water and thereby affect UV dose delivery.

       2.  Microorganisms can associate with particles and be shielded from UV light, thereby
          changing the characteristics of the UV dose-response curve that is obtained using
          collimated beam studies.

       Several studies have found that the effects of turbidity up to 10 NTU on UV disinfection
can be accounted for in the UVT measurements (Passantino et al. 2004, Christensen and Linden
2002). However, the most commonly used spectrophotometer (bench-top direct reading) may
underestimate the UVT of water with turbidity greater than 3 NTU (Christensen and Linden
2002). To reduce this underestimation, all unfiltered systems and uncovered reservoir
applications should use a bench-top UV spectrophotometer with an integrating sphere to provide
more accurate UVT measurements for planning purposes.

       For unfiltered waters susceptible to turbidity fluctuations, the UVT sampling should
occur during these events and be accounted for in the  design UVT and UVT range. If the design
UVT is appropriate, the UV reactor will be able to respond to changes in UVT that arise due to
particles.

       As described previously, particle content and UVT variability will probably be less
prevalent in uncovered reservoirs compared to unfiltered supplies. The UVT sampling, however,
should be conducted during a period sufficient to include seasonal events (e.g., rainstorms and
runoff) that will affect the design UVT and the UVT range.

Algae

       Previous research with male-specific-2 bacteriophage (MS2) has shown that algal counts
up to 70,000 cells/mL do not affect disinfection performance (Wobma et al. 2004). Whether
algal counts greater than 70,000 cells/mL affect the UV disinfection process is unknown.
Therefore, for both unfiltered supplies and uncovered reservoirs, UVT sampling should be
conducted during algal blooms to enable their effects  on UVT to be assessed. At high algal
concentrations, bench-, pilot-, or demonstration-scale  testing may be warranted to  determine if
UV disinfection is significantly affected.
3.4.5  Fouling/Aging Factor

       Sleeve fouling, sleeve aging, lamp aging, and UV sensor window fouling (if applicable)
affect long-term UV reactor performance, as described in Sections 2.4.2 and 2.4.4. The
fouling/aging factor accounts for these issues.

       An acceptable fouling/aging factor and guaranteed lamp life should be determined based
on experience and professional judgment. Alternatively, pilot- or demonstration-scale testing can
be used to estimate the fouling factor and aging factor if deemed necessary by the PWS, as
described in Sections 3.4.5.1 and 3.4.5.2, respectively.

       The lamp-fouling portion of the factor (i.e., fouling factor) is the estimated fraction of
UV light passing through a fouled sleeve as compared to a new sleeve. A lamp sleeve can

UV Disinfection Guidance Manual            3-21                                   November 2006
For the Final LT2ESWTR

-------
                               . Planning Analyses for UV Facilities
become fouled when inorganics (e.g. iron) precipitate onto a lamp sleeve and reduce the UV
transmittance of the sleeve. Water quality parameters that affect fouling are described in Section
3.4.4.2.

       The lamp aging portion of the factor (i.e., aging factor) is the fraction of UV light emitted
from aged sleeves and lamps compared to new sleeves and lamps and can be estimated by the
lamp and sleeve aging characteristics obtained from the UV manufacturer. The lamp aging factor
is important because as UV lamps age, the output of the lamps decrease.

       The fouling/aging factor is calculated by multiplying the fouling factor by the aging
factor and typically ranges from 0.4 (NWRI 2003) to 0.9. The fouling/aging factor is typically
used in validation testing to ensure the UV equipment can meet the required dose in a fouled
and/or aged condition. (See Equation 3.1.)

 UV Dose with Clean Lamps * Fouling Factor * Aging Factor > Required UV Dose Equation 3.1

       When purchasing a pre-validated reactor, the PWS should determine if validation testing
was conducted under conditions of reduced lamp output (e.g., 70 percent) that is equal to or less
than reduced lamp output expected for fouled/aged conditions at its water treatment plant (e.g.,
0.75, or 75 percent). If the site-specific fouling/aging factor is lower (e.g., 0.5, or 50 percent)
than considered during validation testing, adjustments in validation test results or additional
testing should be considered.

       Selection of a fouling/aging factor coupled with a guaranteed lamp life is a trade-off
between maintenance costs (the frequency of lamp replacement or chemical cleanings necessary)
and capital costs (the size of the UV reactors). Both a fouling/aging factor and a guaranteed lamp
life should be selected because doing so will guarantee that the fouling/aging factor will not be
exceeded within the guaranteed lamp life. Lamps for a UV reactor with a lower fouling/aging
factor will require less frequent replacement because the UV reactors are designed with more or
higher powered lamps to achieve the necessary UV output at the guaranteed lamp life. This
strategy, however, may necessitate an increase in the size of the UV reactor and facility.
Conversely, the use of an insufficiently conservative factor may underestimate the reduction in
the lamp output and potentially result in off-specification operation or more frequent lamp
replacement.
3.4.5.1   Testing to Determine the Fouling Factor

       The specific fouling rate and optimal cleaning protocol for any given application cannot
be predicted with existing empirically-proven, mathematical equations. A proper cleaning
protocol and sleeve-fouling factor, however, can be adequately estimated for most water sources
without pilot- or demonstration-scale testing and then adjusted during normal operation.

       Alternatively, fouling rates can be evaluated on a site-specific basis through pilot- or
demonstration-scale testing or during UV reactor start-up. Testing could consist of the following
test elements:
UV Disinfection Guidance Manual            3-22                                  November 2006
For the Final LT2ESWTR

-------
                              . Planning Analyses for UV Facilities
          Test setup: The UV sensors, lamp and sleeve type, power system, and cleaning
          system tested in a pilot- or demonstration-scale system should be identical to the full-
          scale reactor. Differences in lamp and lamp sleeve geometry can lead to erroneous
          conclusions based on pilot data alone.

       .   Flow and UV equipment conditions: Water should flow through the reactor at the
          minimum flow rate, and the lamps should be operated at maximum power.

       .   Establishment of cleaning settings: UV equipment with on-line chemical cleaning
          (OCC) systems should be operated for a prescribed length of time (e.g., 2 weeks)
          without a chemical cleaning to evaluate fouling. With water systems using on-line
          mechanical cleaning (OMC) and on-line mechanical-chemical cleaning (OMCC), the
          cleaning systems should be operated at the manufacturer's recommended frequency
          to assess fouling. One sleeve should be unwiped, however, for the  entire testing
          period to serve as a control to verify that fouling is occurring.

       .   Assessment of fouling factor: Fouling is assessed by placing a new lamp inside a
          fouled sleeve, igniting it, and measuring the UV intensity. The UV intensity should be
          compared to a similar measurement made using a new, clean sleeve. The ratio of
          these two measurements (UV light passing through the fouled sleeve to that passing
          through the new sleeve) is the fouling factor.

       .   Evaluation of sleeve cleaning efficiency: A sleeve cleaning assessment can also be
          performed to determine if more frequent cleaning could reduce the fouling factor.

       .   Sensor window fouling (if applicable): To assess fouling on the UV  sensor
          windows, the windows should be cleaned with phosphoric or citric acid at varying
          time intervals, and the change in UV sensor readings recorded. The fouling rate of the
          lamp sleeves is likely to be greater than the fouling rate of the sensor windows
          because the sleeves are hotter than the windows, and higher temperatures accelerate
          fouling.

          Quality assurance: The fouled sleeve should be manually cleaned, which should
          restore the sleeve UV intensity value to very near that of a new, clean  sleeve after the
          fouling factor has been determined. If not, the inside of the sleeve should be manually
          cleaned and the UV intensity measured again. If the UV intensity is still low, the
          sleeve material has likely  degraded, and the test should be performed with a new
          sleeve to ensure that the test results indicate fouling only and not sleeve degradation.

       The fouling factor data can be analyzed to determine the water system's preferred fouling
factor under the observed sleeve cleaning efficiencies.
3.4.5.2   Testing to Determine the Aging Factor

       The aging factor is the fraction of UV light emitted from aged sleeves and lamps
compared to the fraction emitted from new sleeves and lamps. The lamp aging factor is typically
between 0.5 and 0.8. In most cases, the aging factor can be determined from manufacturer data

UV Disinfection Guidance Manual             3-23                                  November 2006
For the Final LT2ESWTR

-------
                               . Planning Analyses for UV Facilities
with existing empirically proven, mathematical equations. The PWS, however, may desire
testing to better understand lamp aging characteristics. Lamp aging tests assess the reduction and
variance in lamp germicidal output over time under defined worst-case operating conditions.
Factors to consider in designing the test(s) include lamp batch, lamp assembly, electrical
characteristics of the ballasts, heat transfer from the lamps to the water, and lamp operation.
Because lamps are manufactured in batches, lamps from several different lots should be
evaluated to ensure that collected data are representative.

       Lamp age can be tested with either a pilot- or demonstration-scale UV reactor or a test
stand designed to simulate the UV lamp aging in full-scale operation. For either setup, lamps
should be operated in an environment that reflects conditions expected when the UV equipment
is installed at a WTP (e.g., use lamp sleeves, ballasts, and cleaning systems that will be used in
the final application).

       During testing, the following activities should be considered:

       .   Monitor the UV intensity, UVT, electrical power delivered to the ballast, electrical
          power delivered to the lamp, and water temperature over the lamp life.

          Visually inspect the lamp sleeves at regular intervals to document any degradation of
          the lamp assembly, including electrodes and seals, and any darkening of the lamp
          envelope.

       .   Document any fouling on the internal surfaces of the lamp sleeves.

       .   Using either a radiometer equipped with a germicidal filter or a reference UV sensor,
          measure the germicidal output of the lamp under fixed conditions of ballast operation
          (e.g., power setting); heat transfer (e.g., lamp sleeves); and environment (water
          temperature and transmission). The following procedure should be used:

          -   Take one measurement with lamps that have been aged 100 hours ("new").

          -   Measure the output from various positions along the lamp based  on visual
              inspection (i.e., the pattern of darkening on the lamp).

          -   Measure lamp output as a function of lamp power setting if lamp power is
              variable.

          -   Assess the output from lamps of different lots.

       The lamp output measured under fixed operating conditions can be plotted over time and
fit to estimate the mean expected performance for various lamp ages. To determine the aging
factor,  measure the output of a new lamp and the output at the end-of-lamp life.  The aging factor
is the ratio of the output at the guaranteed lamp life to new lamp output and is expressed as a
fraction.
UV Disinfection Guidance Manual             3-24                                  November 2006
For the Final LT2ESWTR

-------
                               . Planning Analyses for UV Facilities
       Although it does not impact reactor design, studies have shown that non-uniform lamp
aging can occur. Non-uniform lamp aging should be considered during validation testing. (See
Section 5.4.6)


3.4.6  Power Quality Evaluations

       UV lamps can turn off if a voltage fluctuation, power quality anomaly, or a power
interruption occurs. Power quality tolerances depend on the UV equipment design and vary
significantly among UV manufacturers (Table 3.5). The UV manufacturer should be contacted to
determine the power quality tolerance and the length of time for the equipment to reach full
power after a power quality event. (See Section 2.4.2.3.)


                 Table 3.5. Power Quality Triggers for UV Reactors1
Power Quality Event
Voltage
Sag/Swell
Tolerance
Power
Interruption
Tolerances4
Voltage2
Duration3
Duration3
LPHO
Manufacturer #1
± 20%
2 seconds (s)
> 0.05s
LPHO
Manufacturer #2
±10%
> 0.03s
> 0.03s
MP
Manufacturer #1
± 30%
> 0.02 s
> 0.009s
MP
Manufacturer #2
± 20%
2s
> 0.05s
   Information shown in the table is compiled from Calgon Carbon Corporation, Trojan Technologies, and
   WEDECO.
  2 Percent of line voltage. For example, a 10-percent voltage loss is when the voltage is at 90% of the line
   voltage.
  3 1 cycle is 0.017s.
   Power interruption assumes total voltage loss.
  Source: Cotton et al. (2005)
       Studies have shown that the typical industrial power user experiences an average of eight
power quality events per month (Grebe et al. 1996). Accordingly, power quality problems alone
likely will not cause UV reactors to exceed the maximum off-specification requirements even
though UV reactors are sensitive to power quality (Cotton et al. 2005). Therefore, a power
quality assessment is probably necessary only when the installation site is (1) known to have
power quality problems (e.g., 30  power interruptions and/or brownouts per month); or (2) located
in a remote area and the power quality is unknown.

       If power quality may be a problem at the intended installation location, a power quality
assessment can be performed to quantify and understand the potential for off-specification
operation, which consists of the following five steps:

       1.  Estimate the power quality at the potential location(s) of the UV facility. Local power
          suppliers often can provide data on power quality and reliability and should be the
          first source of information. Other sources of information are operating records of
          power quality incidents (if available), power interruptions, or Supervisory Control
          and Data Acquisition  (SCADA) information for the existing plant.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
3-25
November 2006

-------
                               . Planning Analyses for UV Facilities
       2.  Understand the power quality tolerance of the UV equipment under consideration by
          contacting the UV manufacturer or consulting published data.

       3.  Contact the UV manufacturer to determine how long it will take their equipment to be
          functioning at full power after a power quality event.

       4.  Estimate the off-specification time for the potential UV equipment-based information
          gathered in Steps 1 through 3. Examples of how to estimate off-specification based on
          this information are presented in Cotton et al. (2005).

       5.  Determine if backup power or power conditioning equipment is needed to reduce off-
          specification time or to improve UV equipment reliability.

       Generally, personnel with a working knowledge of electrical supply and installation will
be able to review power supply data and determine if power quality problems exist. More
advanced assessments can include the installation of power quality monitors  or the retention of
an outside consultant to conduct a detailed power quality assessment.
3.5    Evaluating UV Reactors, Dose Monitoring Strategy, and Operational
       Approach

       Selecting the appropriate UV reactor depends on the installation locations under
consideration and the design parameters discussed in Section 3.4. The UV reactor manufacturer
is a valuable resource for such evaluations and can determine what UV reactors are most
appropriate for the installation locations under consideration. Evaluating the available UV
reactors in the planning process is important because each manufacturer's UV reactors are
unique and proprietary, and installation needs (e.g., power requirements) differ. UV reactors can
generally be characterized based on lamp type with low-pressure high-output (LPHO) lamps and
medium-pressure (MP) lamps applicable to most WTPs. This section discusses the general
characteristics of LPHO and MP reactors and describes the various control strategies. UV
manufacturers should be contacted directly to gain a better understanding of the available and
appropriate UV reactors.
3.5.1  Characteristics of LPHO and MP Reactors

       The fundamental difference between LPHO and MP reactors is the lamp intensity output
(which influences the UV reactor configuration and size), lamp life and replacement, power use,
power modulation capabilities, and sleeve cleaning.

          UV reactor configuration and size: Several UV reactor configurations are available.
          Reactors can be in-line (i.e., shaped like a pipe), S-shaped, or U-shaped, depending
          on the UV manufacturer and the site constraints of the specific installation location.
          Typically, LPHO reactors have a larger footprint than MP reactors because more UV
          lamps are needed to deliver the same required UV dose. MP reactor footprints will
          also vary, depending on lamp orientation (e.g., parallel versus perpendicular to flow).

UV Disinfection Guidance Manual            3-26                                  November 2006
For the Final LT2ESWTR

-------
                              . Planning Analyses for UV Facilities
       .   Lamp life and replacement: Lamp life also varies between LPHO and MP reactors.
          Most manufacturers provide warranties of 8,000 - 12,000 hours for LPHO lamps and
          4,000 - 8,000 hours for MP lamps. Although the lamp life for LPHO reactors is
          greater than that for MP reactors, more lamps are needed for an LPHO reactor. The
          actual number of lamps replaced during  a given period, therefore, may be less for MP
          reactors.

          Power use: Even though LPHO reactors typically have more lamps, they require less
          power input than similarly sized MP reactors because LPHO lamps are more efficient
          in converting the power to germicidal UV light for disinfection. This decreased
          energy efficiency results in higher power needs and increases in overall power
          consumption for MP reactors compared  to LPHO reactors.

       .   Power modulation capabilities: The ability of the UV equipment to adjust lamp
          power or number of UV lamps energized will affect the energy use. Unlike the other
          issues described, power modulation capabilities depend on the UV equipment design
          and not the lamp type.

          Sleeve Cleaning: The lamp sleeve cleaning systems for LPHO and MP reactors  can
          also differ.  LPHO reactors typically have OCC systems, and MP reactors typically
          have OMC systems. Although OCC systems tend to be more labor intensive than
          OMC systems, OMC systems typically have more parts to replace. The extent of
          fouling will determine the amount of maintenance (labor and parts) that is needed on
          a routine basis and  will affect the overall maintenance costs.

       As described, the PWS should evaluate the differences between LPHO and MP  reactors
and determine any preferences based on the different characteristics  and site-specific constraints.
If one technology is precluded, it should not be evaluated further in the planning analyses.
3.5.2  Dose-monitoring Strategy and Operational Approach

       The dose-monitoring strategy establishes the operating parameters used to confirm UV
dose delivery. It affects how a reactor is validated, how instrumentation and controls are
designed, and how the reactor is operated. In the planning phase, the water system should
evaluate the various dose-monitoring strategies to determine whether a particular approach is
preferable based on the ease of integration into their existing operation and control system. If a
particular dose-monitoring strategy is preferred, the water system should select a UV equipment
that has been validated for that strategy. The  effect of the dose-monitoring strategy on the
instrumentation and controls design is described in Section 4.3.

       UV manufacturers commonly design  their reactors to operate using either:

       •   The UV Intensity Setpoint Approach or

       •   The Calculated Dose Approach
UV Disinfection Guidance Manual             3-27                                 November 2006
For the Final LT2ESWTR

-------
                                . Planning Analyses for UV Facilities
       This guidance manual focuses on the design, validation, and operation of UV reactors
that use one of these two approaches. Another existing dose-monitoring strategy or a new
strategy developed after this manual is published, however, may also be suitable for reactor
operations as long as they meet minimum regulatory requirements.5 Alternative strategies should
be considered on a case-by-case basis.

       Table 3.6 summarizes key characteristics of the two dose-monitoring approaches
discussed in this manual. The next two sections provide an overview of how the approaches
operate. Advantages and disadvantages of each are discussed in Section 3.5.2.3, and Section 6.4
provides additional guidance on monitoring frequency and reporting requirements for these
control strategies.
           Table 3.6. Dose-monitoring Approaches - Key Characteristics
Dose-monitoring
Strategy
UV Intensity Setpoint
Approach
Calculated Dose
Approach
Parameter Used as the
Operational Setpoint
UV Intensity
Calculated or Validated dose 1
Parameters Monitored During
Operations to Confirm Dose Delivery
Flow rate
Lamp status
UV intensity
Flow rate
Lamp status
UV intensity
UVT
   As noted in Section 3.4.1, the calculated dose is estimated using a dose-monitoring equation. For the Calculated
   Dose Approach, the validated dose is equal to the calculated dose divided by a Validation Factor, which
   accounts for biases and experimental uncertainty.
3.5.2.1    UV Intensity Setpoint Approach

       As indicated by its name, the UV Intensity Setpoint Approach relies upon one or more
"setpoints" for UV intensity that are established during validation testing. During operations, the
UV intensity, as measured by UV sensors, must meet or exceed the setpoint(s) to ensure delivery
of the validated dose. Importantly, reactors must also be operated within the validated range of
flow rates and lamp statuses (i.e., the "validated operating conditions") [40 CFR 141.720(d)(2)].

       One key  characteristic of the UV Intensity Setpoint Approach is that water systems need
not monitor UVT during operations to confirm dose delivery. Instead, the approach relies on UV
intensity readings by UV sensors to account for changes in UVT. In order for UV sensors to
efficiently monitor dose delivery, they should be as close as possible to the "ideal"  location. This
means that they should be positioned so that the UV intensity is proportional to the UV dose,
irrespective of changes in UVT and lamp output. If the sensor is too close to the lamp, changes in
lamp output will disproportionately impact the measured UV intensity. If the sensor is too far
from the lamp, changes in UVT of the water will disproportionately impact the measured UV
  Systems must monitor flow rate, lamp status, and UV intensity, plus any other parameters required by the state at
  a minimum to show that a reactor is operating within validated conditions [40 CFR 141.720(d)(3)(i)].
UV Disinfection Guidance Manual
For the Final LT2ESWTR
3-28
November 2006

-------
                               . Planning Analyses for UV Facilities
intensity. Water systems can check if sensors are in the ideal location by reviewing validation
test data. (See Chapter 5.)

       The recommended validation protocol in Chapter 5 will produce conservatively high UV
intensity setpoint(s) under many water quality and lamp output conditions if the sensor is not in
the ideal location, resulting in overdosing during operations. In some cases, UV manufacturers
have developed modifications to the UV Intensity Setpoint Approach to account for non-ideal
sensor placement.

       Water systems can use one of the following operating strategies for the UV Intensity
Setpoint Approach: single-setpoint operation or variable-setpoint operations. Table 3.7 describes
these operating strategies and summarizes the advantages and disadvantages of each.
    Table 3.7. Advantages and Disadvantages of Single-setpoint and Variable-
            setpoint Operations for the UV Intensity Setpoint Approach
Operating Strategy
Single-setpoint
Variable-setpoint 1
Description
One UV intensity
setpoint is used for all
flow rates that were
validated
The UV intensity setpoint
is determined using a
lookup table or equation
for a range of flow rates
Advantages
Simplest to operate
and control
Lamp output can be
reduced at low flow
conditions to reduce
energy costs
Disadvantages
When flow rate is variable,
not energy efficient under
most conditions because
reactor is overdosing at low
flow rates
More validation data are
needed. More complex
operation compared to
single-setpoint approach.
Necessitates more
advanced UV reactor
monitoring and control.
1 For the purposes of this guidance manual, variable-setpoint operations refers to variations based on flow
  rate only, as this is the most common application. In theory, multiple setpoints could also be established for
  different lamp statuses and UVT ranges.
3.5.2.2   Calculated Dose Approach

       The Calculated Dose Approach uses a dose-monitoring equation to estimate the UV dose
based the parameters measured during reactor operations.  The most common operational
parameters in dose-monitoring equation are:

        •   Flow rate,

        •   UV intensity, and
        •   UVT

Some manufacturers also consider lamp status as a variable in the dose-monitoring equation.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
3-29
November 2006

-------
                               . Planning Analyses for UV Facilities
       UV manufacturers may develop a theoretical dose-monitoring equation using numerical
models (e.g., computational fluid dynamics [CFD]). Although the theoretical equations can be
used as a starting point, EPA strongly recommends that water systems use an empirical dose-
monitoring equation developed through validation testing. To generate the empirical dose-
monitoring equation, validation tests are performed over a wide range of flow rates, UVT values,
and lamp power combinations. Regression analysis is used to fit the observed validation data to
an equation. Chapter 5 of this manual provides detailed guidance on how to derive an empirical
dose-monitoring equation through validation testing.

       During reactor operations, the UV reactor control system (i.e., the internal reactor
electronics) typically inputs the measured parameters into the dose-monitoring equation to
produce a calculated dose. The system operator divides the calculated dose by a Validation
Factor that accounts for uncertainties and biases to determine the validated dose.6 The operator
compares the validated dose to the required dose for the target pathogen and log inactivation
level.
3.5.2.3   Advantages and Disadvantages

       The principal operating advantage of the UV Intensity Setpoint Approach compared to
the Calculated Dose Approach is that UVT monitoring is not needed to confirm dose delivery.
Another important advantage is that the UV Intensity Setpoint Approach, single-setpoint
operation is straightforward and simple to control with one operational setpoint and one
maximum value for flow rate. For these reasons, EPA believes this option is good for small
water systems. Other advantages are that the UV Intensity Setpoint requires fewer validation
tests than the Calculated Dose Approach and data analyses are relatively straightforward. Data
analyses to develop the dose-monitoring equation for the Calculated Dose Approach can be
complex.

       Water systems may favor the Calculated Dose Approach over the UV Intensity Setpoint
Approach because it offers significant flexibility to reduce operating costs by manipulating lamp
power (e.g., turning off banks of lamps or powering down lamps when the UVT increases and/or
the flow rate decreases). This process is also called "dose pacing." Another potential advantage
is that operations are more intuitive because the calculated dose, adjusted for uncertainties and
biases, can be directly compared to the required dose for the target pathogen and log inactivation.

       Manufacturers may favor the Calculated Dose Approach because they have more
flexibility in UV sensor positioning (i.e., because internal analyzers monitor UVT during
operations instead of relying on sensors to respond to changes in UVT, positioning sensors as
close as possible to the "ideal" location offers no advantages). As noted in Section 3.5.2.1, UV
Intensity  Setpoint Approach operations will be more efficient if the UV sensors are at or near the
ideal location.
6 In some cases, the UV reactor control system will perform this step as well, outputting the validated dose
  automatically.
UV Disinfection Guidance Manual            3-30                                  November 2006
For the Final LT2ESWTR

-------
                                . Planning Analyses for UV Facilities
3.6    Assessing UV Equipment Validation Issues

       For disinfection credit, the LT2ESWTR requires UV reactors to be validated [40 CFR
141.720(d)]. A water system's approach to UV reactor validation and to UV facility design is
interrelated. The issues to consider are whether equipment will be validated on-site or off-site
and the hydraulic conditions of the UV reactor validation and installation. This section describes
how these issues affect the design and the relationship between the validation and hydraulic
installation approaches. Chapter 5 details the UV reactor validation guidelines.


3.6.1  Off-site Versus On-site Validation

       UV reactors can be validated either off-site or on-site. With off-site validation, the UV
reactors are validated before installation (i.e., pre-validated), typically at a third-party validation
test center or a UV manufacturer facility. With on-site validation, the UV reactors are validated
at the PWS after they have been installed. Many PWSs will use off-site validation to meet the
LT2ESWTR requirements. In some cases, however, on-site validation may be appropriate (e.g.,
when the full UVT range was not tested in off-site validation). The advantages and
disadvantages of off-site and on-site validation are presented in Table 3.8.


    Table 3.8. Advantages and Disadvantages of Off-site and On-site Validation
                      Advantages
                                                       Disadvantages
        Broader ranges of flow and water quality are
        tested so a reactor can be validated for more
        than one application
        Installation hydraulics are general, allowing
        for installation at most WTPs
        Process is simpler for utilities because testing
        is conducted at a remote location
        Cost is usually lower
        Reactor performance is known before facility
        is designed and constructed
                                           Re-validation or additional on-site validation
                                           testing may be necessary if site-specific
                                           hydraulics and water quality are not within
                                           the tested ranges
                                           Water quality and hydraulics may not match
                                           the installation location, potentially resulting
                                           in less efficient operation
  0)
  .*;
  c
  O
Exact hydraulics of the installation are used
Water quality tested is specific to the
installation
Having provisions for on-site testing (e.g.,
feed and sample ports and static mixers)
enables flexibility for future testing to optimize
performance
Facility may be designed and constructed
before reactor performance is verified
Water quality is limited to the highest UVT at
the facility during the testing period
Testing logistics can be complex, including
isolation of the test reactor, assessment of
additive mixing, and challenge
microorganism stability
Cost may be higher
Disposal of test water may require special
permits
       The PWS should determine whether off-site or on-site validation will be used to meet the
LT2ESWTR requirements. If on-site validation is preferred, the UV facility design should be
adapted to enable testing. The UV reactor design should incorporate feed and sample ports, static

UV Disinfection Guidance Manual             3-31                                    November 2006
For the Final LT2ESWTR

-------
                               . Planning Analyses for UV Facilities
mixers, space for tanks near the UV facility for adding the challenge microorganism and UV
absorbing chemical, and a method to discharge the validation test water. If off-site validation is
preferred, the UV facility need not incorporate provisions for on-site validation testing.

       If pre-validated reactors that were validated off-site are chosen, the PWS should confirm
that the validation hydraulic recommendations in Section 3.6.2 can be met without additional on-
site validation or PWS-specific off-site validation.


3.6.2  Validation and Installation  Hydraulics Recommendations

       The inlet and outlet piping to the UV reactor in the UV facility should result in a UV dose
delivery that is equal to or greater than the UV dose delivered when the UV reactor was
validated. If off-site validation is used, the three preferred options for meeting this condition are
presented below.

       1.  Minimum five pipe diameters of straight pipe upstream of UV reactor: The
           length of straight pipe upstream of each UV reactor at the UV facility is the length of
           straight pipe used in the validation testing plus a minimum of five (5) pipe diameters.
           During validation testing, the inlet piping to the reactor consists of either a single 90-
           degree bend, a "T" bend, or an "S"  bend, followed by a length of straight pipe if
           necessary. See Figure  3.7 for validation and installation configuration options.

       2.  Identical inlet and outlet conditions: Inlet and outlet conditions  used during
           validation match those used at the WTP for at least ten (10) pipe diameters upstream
           and five (5) pipe diameters downstream of the UV reactor.

       3.  Velocity profile measurement: Velocity of the water measured at evenly spaced
           points through a given cross-section of the flow upstream and downstream of the
           reactor is within 20 percent of the theoretical velocity with both the validation test
           stand and the WTP installation (NWRI 2003). The theoretical velocity is defined as
           the flow rate divided by the cross-sectional area.

       Jetting and swirling flow will impact the assumptions for Options 1 and 3. To avoid
jetting flow, the inlet piping should have no expansions for at least ten (10) pipe diameters
upstream of the reactor. Also, any valves located in that length of straight pipe should always be
fully open during UV reactor operation. To avoid swirling flow, the validation piping should not
include two out-of-plane 90°-bends in series.

       The most suitable validation option depends  on the site-specific layout and piping
constraints and on the validation data. Option 1 is more generally applicability for validation and
installation of UV reactors. For example, the inlet and outlet piping configuration for
installations in  a new building could be designed based on how the procured UV reactor was
validated. Option 2 is most applicable when unique piping configurations are needed or if the
inlet and outlet conditions validated in Option  1 cannot be achieved because of site constraints.
For example, Option 2 may be the only validation option for an individual filter effluent location,
which likely will not have 5 diameters of straight pipe before the UV reactors (Option 1) because
UV Disinfection Guidance Manual             3-32                                   November 2006
For the Final LT2ESWTR

-------
                               . Planning Analyses for UV Facilities
of existing site constraints. Option 3 also provides flexibility but may have the practical
limitation of measuring the velocity through a cross-section at the installation.
      Figure 3.7. Schematic of Hydraulic Option #1 (90°-Bend, T-Bend, S-Bend
                                 Inlet Piping Scenarios)
                        Validation Options
UV Facility Installation Options

 Q
                                                    Q
                                                              1—I
                         Q-*,
                                    JED
                   (x > 0 where x is a multiple of the pipe diameter "D" and [   ] is the reactor.
       If available, the validation report for pre-validated UV equipment under consideration
should be reviewed to determine what the inlet/outlet conditions were during validation, which
will help determine if Option 1 is feasible. The method for meeting these recommended
inlet/outlet constraints should be determined in the planning stage and considered when
developing the UV facility layout (Section 3.8.2).

       CFD modeling and CFD-based UV dose modeling of inlet and outlet conditions may be
used to assess whether UV dose delivery at the WTP installation is better than UV dose delivery
achieved during validation for given conditions of flow rate, UVT, and lamp output. The state
should approve such models and their reliability should be properly evaluated before their results
are accepted. Appendix D provides guidance on evaluating CFD models.


3.6.3  Selection of Validation and Hydraulic Approach

       Whether or not the UV reactor was pre-validated off-site affects the inlet/outlet piping
options for the UV facility. Completing on-site validation provides more inlet/outlet piping
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                           November 2006

-------
                                 . Planning Analyses for UV Facilities
flexibility, but on-site validation means additional design considerations and testing at the water
treatment plant. If the selected UV equipment is not pre-validated, the PWS can choose either
off-site or on-site validation based on their site-constraints and preferences. These options are
described in Figure 3.8.


           Figure  3.8. UV Reactor Validation Options and How They Affect
                                   Installation Hydraulics
                                     All UV reactors must be
                                       validated [40 CFR
                                           141.720(d)]
               PWS purchases a
               pre-validated UV
               reactor that was
               validated off-site
   A. PWS installs
   validated UV reactor
   in accordance to
   installation hydraulic
   Option 1 or 3
     (Section 3.6.2). On-
     site validation is
     not necessary for
     this option.
B. PWS installs a
pre-validated UV
reactor, develops a
validation test plan,
and conducts on-site
validation because
either:
1. The UV facility
has unique hydraulic
conditions (i.e.,
installation
hydraulic  Options 1
and 3 are not
feasible)1
2. The PWS wants
to refine the
validated conditions
to closely  match
their operating
conditions

Hydraulic Option 2
is used with on-site
validation by default
                                     PWS purchases a
                                    UV reactor that has
                                      NOT been pre-
                                        validated
C. PWS develops
detailed validation
test plan for off-site
validation and has
flexibility of using
any hydraulic
installation option.
D. PWS develops
a validation test
plan and conducts
on-site validation
and uses hydraulic
Option 2 by
default.
  1  PWS could contract with an off-site validation center to perform validation testing with specific hydraulic
    conditions rather than perform on-site validation.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
               3-34
                               November 2006

-------
                               . Planning Analyses for UV Facilities
3.7    Assessing Head Loss Constraints

       When selecting a feasible location for UV reactors, the hydraulic requirements should be
met. Head loss through a UV reactor is specific to the equipment and flow rate and generally
varies from 0.5-3 feet (UV reactor only). Characteristic head loss data should be obtained from
the UV manufacturer(s) for all candidate UV reactors. In addition to the head loss associated
with the UV reactor itself, the head loss associated with piping, valves, flow meters, and flow
distribution devices (e.g., baffles) should be considered when assessing the feasibility and
location of the installation. When selecting a reactor that has been validated off-site (Options A
of Figure 3.8), the UV reactor inlet/outlet piping used to  estimate the head loss through the
facility should be consistent with the validation recommendations described in Section 3.6.2. The
head loss through the entire UV facility (i.e., piping, valves, joints, and UV reactors) can be
between 1 and 8 feet.

       If the head loss through the UV facility is greater than the available head, the plant design
or operation, or both, may require modification. Some potential modifications, alone or in
combination, that may be considered to address hydraulic limitations are listed below, and details
for each are provided in the sections that follow:

       .   Eliminating existing hydraulic inefficiencies within the facility to improve head
          conditions (e.g., replacing undersized or deteriorated piping and valves)

       .   Modifying the operation of the clearwell

       .   Modifying the operation of the filters

       .   Installing intermediate booster pumps

       .   Modifying the operation of the HSPs
3.7.1  Eliminating Existing Hydraulic Inefficiencies

       Replacing undersized piping and valves with larger diameter piping and valves may
increase the available head for the proposed UV facility. Older piping can also produce excessive
head loss if the inner pipe surface is pitted or scaled or if the pipe material has a high coefficient
of friction. Slip-lining the interior of existing pipe with material having a lower coefficient of
friction (e.g., high-density polyethylene) is one method of reducing friction losses. Re-lining the
existing pipe interior with a smooth coating will also reduce head loss.
3.7.2  Modifying Clearwell Operation

       A PWS may increase head available to a UV facility by lowering the surface water level
of the clearwell. This strategy, however, decreases the storage volume available to meet peak
demands, reduces the contact time available in the clearwell for chemical disinfectants, and may
affect the pump discharge head and distribution system pressure. Evaluating any potential

UV Disinfection Guidance Manual             3-35                                   November 2006
For the Final LT2ESWTR

-------
                               . Planning Analyses for UV Facilities
reduction in disinfection credit is important if contact time in the clearwell is used for calculating
chlorine disinfectant requirements (i.e., CT). The UV facility, however, may reduce the Giardia
CT requirements sufficiently to offset the reduction in CT.
3.7.3  Modifying Filter Operation

       A treatment facility can alter the operation of its filters (e.g., increase the water elevation
above the filters) to increase the head available for the UV facility. This approach, however, can
reduce filter run times and reduce unit filter run volumes, resulting in a need for more frequent
backwashing. If conditions upstream of the filters are such that additional freeboard and
hydraulic head are available, a second option is to increase the water surface elevation above the
filters to help minimize the reduction in head as the water is filtered.
3.7.4  Installing Intermediate Booster Pumps

       When modifications to the existing facility or operations do not provide adequate head
for the UV reactors, intermediate booster pumps can be installed. Booster pumping increases
flexibility in locating the UV reactors. Installing booster pumps, however, increases facility
operation and maintenance costs and space requirements. The reliability of the pumps should
also be considered in the evaluation because they become a critical operating component. More
information on intermediate booster pumps is presented in Section 4.1.6.
3.7.5  Modifying Operation of HSPs

       When UV disinfection is installed close to the HSPs (e.g., after the clearwell in a
filtration plant or after an unfiltered reservoir), one option to increase the head available for the
UV facility is to modify the pumping operation of the HSPs. Modifications may not be practical,
however, if they change the distribution system pressure.
3.8    Estimating UV Facility Footprint

       The process footprint should be estimated in the planning phase to help determine
feasible UV facility locations. The critical components for estimating the UV facility footprint
are UV equipment constraints and UV facility layout.
3.8.1  UV Equipment Constraints

       The UV equipment constraints that affect the footprint estimation are the number of UV
reactors needed to meet the design criteria, the UV reactor orientation, and the control panel
location constraints.

          Number of UV reactors: The number of UV reactors depends on the redundancy
          chosen and the power modulation capabilities of the UV reactor. UV reactor

UV Disinfection Guidance Manual             3-36                                 November 2006
For the Final LT2ESWTR

-------
                               . Planning Analyses for UV Facilities
          redundancy should be determined using sound engineering approaches similar to
          those used for other major equipment (e.g., capacity to provide full treatment with the
          largest UV reactor out-of-service). The ability of the UV equipment to modulate lamp
          power or change the number of lamps energized also should be considered, so that
          energy efficient operation is possible at the operating range of flows and UVTs
          expected for the UV reactors. The UV manufacturer should be contacted to determine
          a particular UV reactor's power modulation capabilities.

          UV reactor orientation: UV reactors can be oriented either parallel or perpendicular
          to the ground. Two advantages of vertical orientation (i.e., flow perpendicular to the
          ground) are that (1) the footprint will be smaller and (2) the potential for lamp breaks
          due to debris may be reduced (as described in Appendix E).

          Control panel location constraints: Maximum allowable  separation distance
          between the UV reactors and electrical controls should be considered in the UV
          facility layout and footprint estimation. This information is unique to each UV reactor
          and should be obtained from the UV manufacturer.

          Validation hydraulic restrictions: Section 3.6.2 describes how the validation piping
          configuration can dictate the possible UV facility piping configurations.
3.8.2  Develop UV Facility Layout

       The UV facility layout is dictated by site constraints and the UV equipment constraints
described in the previous section. The following items should be considered when developing the
UV reactor and piping configuration and estimating the UV facility footprint in the planning
phase:

       .   Number, capacity, dimensions, and configuration of the UV reactors (including
          redundancy and connective piping)

          Vertical or horizontal orientation of the UV reactor

       .   Maximum allowable separation distance between the UV reactors and electrical
          controls if distance limitations apply

       .   Adequate distance between adjacent reactors to afford access for maintenance tasks
          (e.g., lamp replacement)

          Configuration of the connection piping and the inlet/outlet piping necessary before
          and after each UV reactor, based on validated hydraulic conditions (see Section 3.6.2)
          and UV manufacturer recommendations

          Space and piping for booster pumps and wetwells (if necessary)

       .   Space for electrical equipment, including control panels, transformers, ballasts,
          backup generators, and possible uninterruptible power supplies

UV Disinfection Guidance Manual            3-37                                  November 2006
For the Final LT2ESWTR

-------
                               . Planning Analyses for UV Facilities
       .   Room for storing spare parts and chemicals (if needed)

       .   Lifting capability for heavy equipment

       .   Provisions for on-site validation (if applicable)

       The dimensions of UV reactors and associated electrical equipment vary depending on
the UV manufacturer. Installation footprint and layout should therefore be estimated for all UV
manufacturers being considered. Once the UV facility footprint is estimated, feasible site
locations can be determined based on the available land and buildings.


3.9    Preparing Preliminary Costs and Selecting the UV Facility Option

       The amount of analysis necessary to determine the appropriate application point for a UV
facility is site-specific. Some options clearly will be infeasible, while others may necessitate a
more detailed comparison of the installation options. Once feasible alternatives are identified,
development of life-cycle costs  and consideration of the non-monetary factors (e.g., ease of UV
facility operation) can be useful in selecting among alternatives.

       Preliminary life-cycle cost estimates should include capital costs and operation and
maintenance (O&M) costs.  Capital costs include the cost of the UV reactors; building (if
necessary); piping; pumping (if necessary); electrical  and instrumentation provisions; site work;
contractor overhead and profit; pilot-testing (if necessary); validation costs; and engineering,
legal, and administrative costs. The O&M costs  should include the estimated labor, energy, and
equipment replacement costs. The LPHO equipment and MP equipment have different  O&M
needs (Section 3.5.1) that should be considered in the O&M costs.

       Selection of the best option should be based on the disinfection and design objectives and
consideration of the following and other PWS-specific criteria:

       •   Cost-effectiveness and ability to meet the water system's disinfection and design
           objectives

       •   Ease of installation (where applicable)

       •   Operational  flexibility and reliability

       •   Specific maintenance needs

       •   Flexibility for future treatment expansion (if applicable)


3.10  Reporting to  the State

       Interaction with the  state throughout the  planning and design phases is recommended to
ensure that the objectives of both the PWS and the state are met. This interaction may require
several months and can have a significant effect on the implementation schedule, particularly
when the state requires modifications. Given the relatively limited use of UV disinfection in the

UV Disinfection Guidance Manual             3-38                                  November 2006
For the Final LT2ESWTR

-------
                                . Planning Analyses for UV Facilities
United States to date, drinking water treatment, and the unique technical characteristics of this
technology, state agencies may not have developed approval requirements specifically for UV
disinfection. As such, PWSs are urged to consult with their state early in their UV disinfection
planning process to understand the approvals and documentation that will be required for the use
of UV disinfect!on.

       The state may require that a preliminary design report be submitted that summarizes the
decision logic used to identify, evaluate, and select UV disinfection. The following items may be
addressed in the preliminary design report:

        •    Disinfection objectives (target organism and inactivation)

        •    Overall disinfection strategy
        •    Summary of reasons for incorporating UV disinfection

        •    Description of the overall process train

        •    Description of the proposed UV reactors

        •    Water quality data

        •    Design criteria
        •    Validation Test Plan (if performing on-site or off-site validation- See Section 5.11
            for guidance on developing a Validation Test Plan)
UV Disinfection Guidance Manual              3-39                                   November 2006
For the Final LT2ESWTR

-------
                                  . Planning Analyses for UV Facilities
                                This page intentionally left blank.
UV Disinfection Guidance Manual              3-40                                     November 2006
For the Final LT2ESWTR

-------
                  4. Design Considerations for UV Facilities
       This chapter presents the key factors that should be considered during the detailed design
phase and is written under the premise that the necessary planning and evaluation work discussed
in Chapter 3 has been completed. This chapter focuses primarily on the design of UV
disinfection applications for filtered surface water. Most of the information presented, however,
also applies to unfiltered systems, groundwater under the direct influence (GWUDI), and
uncovered finished water reservoirs.  Additional design issues specifically associated with
unfiltered, GWUDI, and uncovered finished water reservoir systems are also described.
        Chapter 4 covers:
        4.1   UV Facility Hydraulics
        4.2   Operating Approach Selection
        4.3   Instrumentation and Control
        4.4   Electrical Power Consideration and Back-up Power
        4.5   UV Facility Layout
        4.6   Elements of UV Equipment Specifications
        4.7   Final UV Facility Design
        4.8   Reporting to the State during Design
       In the United States, most public water systems (PWSs) purchase or select the UV
equipment before the UV facility design is complete. Pre-purchase or pre-selection of the UV
equipment enables the designer and the UV manufacturer to coordinate during the detailed, final
design phase to consider manufacturer-specific design recommendations. Sometimes the
equipment is pre-selected and the UV equipment manufacturer is included in the construction
contract. Other procurement methods (e.g., base-bid and contractor selection of equipment) are
also used, but these methods are less common.

       The process for designing a UV facility is presented as a flowchart in Figure 4.1. The
illustrated process is based on pre-purchasing or pre-selecting the UV equipment using a
traditional design-bid-build approach. Any of the equipment procurement and contractor
selection approaches currently available within the industry, however, can be used to build UV
facilities. The PWS and the engineer are responsible for selecting the most appropriate approach
for their specific project. The order of the steps for other procurement approaches may differ
from that shown in Figure 4.1, but the analyses completed are likely to be very similar. The steps
described in this chapter follow the order presented in Figure 4.1. Some states may have design
and plan review requirements that could impact the timing or sequence of steps shown in
Figure 4.1. The appropriate state regulatory agency should be contacted early in the design
process to discuss specific design requirements, plan review fees, and review scheduling.
UV Disinfection Guidance Manual             4-1                                  November 2006
For the Final LT2ESWTR

-------
                             4. Design Considerations for UV Facilities
   Figure 4.1. Flowchart for Planning, Design, and Construction of UV Facilities1
   Planning analyses
   for UV facilities
   completed
   Chapter 3
Design UV
facility
hydraulics
Section 4.1
Determine
operating
approach
Section 4.2
Design
instrumentation
and controls
Section 4.3
  Report to Primacy Agency
  Section 3.10 (Planning)
  Section 4.8 (Design)
                                                                      Design electric
                                                                      power systems
                                                                      Section 4.4
       CONSTRUCTION  •«-
             Finalize UV facility
             design, drawings,
             and specifications
             and bid UV facility
             Section 4.7
                                                                       Complete UV
                                                                       facility layout
                                                                       Section 4.5
                        Develop
                        specifications
                        and procure
                        equipment
                        Section 4.6
    Flowchart is based on pre-purchase of UV reactors that have undergone validation testing and equipment
    installation using a traditional design-bid-build approach.
    Additional state coordination may be necessary.
4.1    UV Facility Hydraulics

       After the facility location and UV equipment are selected during the planning phase, a
more detailed evaluation of system hydraulics for the UV facility layout developed in Section 3.!
should be conducted. In most cases, the UV facility will be designed with multiple, parallel UV
reactor trains of the same capacity. Each train consists of the lateral piping, UV reactor, valves,
and flow meter (if applicable) and is joined to the other trains by the distribution and
recombination channel or manifold. The hydraulic evaluation should include upstream and
downstream processes for free water surfaces, the inlet/outlet piping configuration, flow control
and distribution, flow rate measurement, level control, air and pressure controls, valving, and,
where applicable, intermediate booster pumps.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                4-2
                              November 2006

-------
                            4. Design Considerations for UV Facilities
4.1.1       Inlet and Outlet Piping Configuration

       The recommended inlet and outlet conditions for validation and for the UV facility are
described in detail in Section 3.6.2. If validation is conducted at an off-site testing facility, the
designer should refer to the validation report to determine the validated inlet and outlet
conditions, and then use the recommendations in Section 3.6.2 to determine the recommended
inlet and outlet piping for the UV facility. If on-site validation or custom off-site validation is
planned, the inlet and outlet hydraulics should be designed according to manufacturer
recommendations and to accommodate any site-specific constraints. In addition, to avoid jetting
flow, the inlet piping should have no expansions for at least ten (10) pipe diameters upstream of
the reactor.
4.1.2       Flow Distribution, Control, and Measurement

       Regulations specify flow rate, UV intensity, and lamp status as the minimum operating
conditions that a PWS must routinely monitor [40 CFR 141.720(d)(3)]. Accordingly, proper flow
distribution and measurement are essential for compliance monitoring of the UV reactors. This
section discusses various methods for designing proper flow distribution and measurement
through the UV reactors.
   4.1.2.1 Flow Distribution and Control

       The lateral piping for each UV reactor train should be sized and configured to provide
approximately equal head loss through each UV reactor train over the range of flow rates.
Importantly, flow rate through each reactor must conform to the validated operating conditions,
[40 CFR 141.729(d)] as described in the validation report.

       Two approaches for flow distribution and control  are generally used. The first is active
flow control and distribution,  in which a dedicated flow meter and modulating control valve are
installed for each UV reactor. Active flow control provides the greatest hydraulic control in
applications with widely varying flow rates. The second method is passive flow distribution. For
the passive approach, equal flow split is monitored with flow meters.

       For PWSs that use distribution and recombination channels (instead of influent and
effluent manifolds), designers typically have two basic choices to achieve passive flow
distribution (Figure 4.2): (1) a series of individual weirs set at the same elevation or (2) a series
of orifices submerged in the individual UV reactor laterals.
   4.1.2.2 Flow Rate Measurement

       The method of flow rate measurement selected should be based on the variability in plant
flow rate, the type of flow split used, and any state requirements. Selection of the flow rate
measurement method should be at the discretion of the PWS and the design engineer based on
experience and professional judgment. Generally, each UV reactor should have a dedicated flow
meter (as described in Table 4.1) to confirm that the reactor is operating within the validated

UV Disinfection Guidance Manual                4-3                                November 2006
For the Final LT2ESWTR

-------
                             4. Design Considerations for UV Facilities
flow rate. The state, however, may approve other methods (e.g., one flow meter coupled with
pressure differential measurements).

                 Figure 4.2. Open-channel Flow Distribution Options
                                                 Plan
                                                Section
                       A. Flow Splitting Weirs


I




































Zi-


f











1












|













'



1


'


Flow


xxx



xvx


^





•












1



I


1


^_^
. k











                                                 Plan
                                                 Section
                       B. Sit>merged Orifices
UV Disinfection Guidance Manual
For the Final LT2ESWTR
4-4
November 2006

-------
                            4. Design Considerations for UV Facilities
   Table 4.1. Comparison of Techniques for UV Facility Flow Rate Measurement for
              Combined Filter Effluent and Post-clearwell UV Facilities 1
Flow Rate
Measurement
Method
Individual UV
Reactor Flow Rate
Measurement
Individual UV
Reactor Flow Rate
Measurement and
Control
Flow
Control
Method
Passive flow
control such
as a weir or
an orifice
Individual flow
control (valve)
for each UV
reactor
Advantages
. Measures individual UV reactor flow
rates accurately
• Measures individual UV reactor flow
rates accurately
• Does not rely on passive flow
distribution
Disadvantages
. May have unequal flow distribution
• Cannot control the UV reactor flow rate
• Increases capital cost
. May increase facility footprint due to
hydraulics of UV reactor, meter, and
valves
  For individual filter effluent installations, the flow rate from the filters can be used to determine the flow rate through the
  UV reactors.
       When selecting a flow meter, the flow meter's effect on the inlet/outlet hydraulics of the
UV reactor should be considered. Magnetic or other types of flow meters (such as Doppler) that
do not protrude into the flow path exert the least effect on the velocity profile, which minimizes
the potential effect on reactor inlet or outlet hydraulics.
4.1.3
Water Level Control
       The UV lamps in the UV reactor should be submerged at all times to prevent overheating
and UV equipment damage. This is accomplished by installing the UV reactors at an elevation
below the hydraulic grade line elevation. Two common methods for keeping the UV lamps
submerged are to:

       1.  Install a flow control structure (e.g., weir or orifice) immediately downstream of the
          UV reactor or at another location that ensures full pipe conditions through the UV
          reactors.

       2.  Use flow control valves to monitor and maintain the hydraulic grade line.

       Damage to UV lamps caused by operation in air is specific to each lamp type and size.
Low-pressure (LP) lamps can typically operate in air for up to 24 hours with minimal damage.
Low-pressure high-output (LPHO) lamps will begin experiencing damage as a result of
dislodged amalgam or mercury adsorption to the inner surface of the lamps  in 6 - 12 hours
(Lawal 2006). Medium-pressure (MP) lamps can experience advanced aging or solarization in
fewer than 6 hours  and can break (see Appendix E).
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                              4-5
November 2006

-------
                            4. Design Considerations for UV Facilities
4.1.4       Air Relief and Pressure Control Valves

       UV reactors should be kept free of air to prevent lamp overheating. Negative gauge
pressures or surge effects within the UV reactors should also be prevented to avoid damage to
the lamps and lamp sleeves. Quartz sleeves are designed to accommodate continuous positive
pressures of at least 120 pounds per square inch gauge (psig) but have been shown to break at
negative pressures  of 1.5 (Roberts 2000, Aquafme 2001, Dinkloh 2001). Negative pressures can
result from line breaks or accidental dewatering of the reactor. The use of air release valves,
air/vacuum valves, or combination air valves may be necessary to prevent air pockets and
negative gauge pressure conditions. The UV manufacturer should be contacted to determine any
equipment-specific air release and pressure control valve needs. The valve locations will be
dictated by the specific configuration of the facility and should be determined during design.
4.1.5       Flow Control and Isolation Valves

       Each UV reactor should be capable of being isolated and removed from service. Isolating
or shutting down a UV reactor will require valves, gates, or similar devices upstream and
downstream of the UV reactor. Valves are recommended because they provide a tighter seal.
During design, the inlet and outlet valve configuration should be discussed with the UV
manufacturer to ensure that UV reactor performance will not be adversely affected and that the
required inlet conditions used during validation are met, as discussed in Section 3.6.

       If the isolation valves are  also used for flow control, the flow control valve should be
located downstream of the UV reactor to limit the disturbance of the flow entering the UV
reactor. Valves downstream of the UV reactor can be equipped with an actuator to open or close
automatically on a critical alarm occurrence and to enable start-up sequencing.

       Valve seats and other in-pipe seals and fittings within the straight pipe lengths adjacent to
the UV reactors should be constructed of materials that are resistant to UV light and chemicals
that may be used for reactor cleaning. Resistant materials will help avoid valve degradation.
4.1.6       Installation of Intermediate Booster Pumps

       A detailed evaluation and design of a booster pumping system is recommended if head
constraints indicate a pumping system is necessary. Pumps common in water treatment plants
(i.e., vertical turbine, end-suction centrifugal, and split-case centrifugal pumps) tend to have
higher discharge pressures than needed for intermediate pumping applications  and are generally
not appropriate. Mixed- or axial-flow pumps with high-flow and low-head operating
characteristics are usually better choices for intermediate pumping applications because typically
only 1-8 feet of additional hydraulic head is needed to overcome the head loss through the UV
facility.

       Pumps can be installed before or after the UV reactors, allowing more flexibility in the
UV facility's design elevations and the location of the UV reactors. Regardless of pump location,
some form of wetwell should be provided upstream of the pump station. Existing clearwells,
recombination channels, or dedicated pump wetwells may be used.

UV Disinfection Guidance Manual                4-6                               November 2006
For the Final LT2ESWTR

-------
                            4. Design Considerations for UV Facilities
       Booster pump operation may be controlled by the water level within the upstream
wetwell. The use of variable frequency drives or a rate-of-flow controller with a modulating
valve to dampen flow rate peaks is recommended, especially if the pump station is upstream of
the UV reactors. By minimizing hydraulic peaks, the UV reactors can be sized to more closely
match the flow rate through the water treatment plant (WTP).
4.1.7       Evaluating Existing Pumps and Potential Water Hammer Issues

       In some WTPs, the most feasible location for installing the UV reactors may be
immediately upstream or downstream of existing high-service pumps (HSPs) (Section 3.3.1.3).
The HSP discharge curves should be analyzed to determine the effect of the increased head loss
through the UV reactors and whether HSP modifications are necessary.

       If pumps are located adjacent to the UV reactors, the impact of surge conditions should
be evaluated. Of particular concern is the potential for surge if the pumps are operating and
power is lost. Pump start-up procedures should be carefully defined, including procedures for
pump control valves. Control of individual UV reactor isolation valves should be coordinated
with pump starts and stops and with pump control valves where appropriate. Likewise, the
warm-up time associated with the start-up of the UV reactors should be taken into account with
the sequencing of the pump  operation.
4.1.8       Groundwater System Hydraulic Issues

       Common hydraulic issues associated with groundwater systems include high operating
pressures, air entrainment, and the potential for water hammer events.

       Lamp sleeves are designed to resist high external operating pressures. Before selecting
equipment, however, the designer should determine the maximum expected operating pressure,
which may occur during a failure event (e.g., downstream valve closes), and confirm that the
proposed equipment can withstand that pressure.

       Pressure surge events (water hammer) near the UV reactor may be more likely with
groundwater systems than surface water systems because of the UV reactor's proximity to the
well pumps. Surge events can cause positive or negative pressure transients in the well discharge
piping and potentially break the sleeves and lamps. A surge analysis is recommended to
determine if surge protection is necessary. Many well sites and distribution systems are already
equipped with surge control tanks to dampen surge effects. These tanks may provide sufficient
protection for the UV reactors, depending on their location relative to the UV reactors.

       Air binding can interfere with the UV disinfection process or cause the lamps to overheat.
UV reactors should be located downstream of any existing or planned air removal equipment  (if
necessary). Otherwise, the UV facility design should include a means for automatically releasing
air prior to the UV reactor. The UV reactor may have air release valves or valve ports, or air
release valves can be installed in the inlet piping.
UV Disinfection Guidance Manual                 4-7                              November 2006
For the Final LT2ESWTR

-------
                            4. Design Considerations for UV Facilities
4.1.9       Uncovered Finished Water Reservoir Hydraulic Issues
       Many uncovered finished water reservoirs serve as distribution storage and are directly
affected by the water system demand. Others may be used solely as an emergency supply or may
function as both distribution storage and emergency supply. The specific hydraulic
considerations that a PWS and designer should consider will vary depending on the function of
the uncovered finished water reservoir. Regardless of reservoir function, however, specific
hydraulic issues that should be considered when designing UV facilities at uncovered finished
water reservoirs include widely varying flow rates, bi-directional flow (under certain piping
configurations), and the effect a UV facility will have on system pressure.

       .   Variable flow rate: The methodology described in Section 3.4 should be followed to
          determine the flow rate and UVT that are used to design the UV facility. Most UV
          facilities at uncovered finished water reservoirs should be designed to handle the peak
          instantaneous demand that must be met by the reservoir. The instrumentation and
          control (I&C) design must consider how the PWS will sequence the UV reactors with
          highly variable flow conditions, especially warm-up times for UV lamps (Section
          2.4.2.3).

       .   Bi-directional flow: In some cases, the inlet and outlet to the uncovered finished
          water reservoir is the same pipe, and the UV facility should be designed so that
          disinfection continues when the water is flowing from the uncovered finished water
          reservoir. The PWS may also consider operating the UV reactors at a minimum level
          as the water flows into the reservoir so that the UV lamps are energized and ready for
          UV disinfection if the flow direction changes suddenly. The necessity for this latter
          approach depends on the number of directional changes per day in the context of
          meeting off-specification requirements.

       .   UV facility effect on system pressure: As discussed in Section 3.7, head loss
          through a UV reactor generally varies from 0.5 to 3 feet, with the overall head loss
          through a UV facility typically about 1 to 8 feet. This head loss could affect the
          distribution system pressure. As discussed in Chapter 3, a hydraulic assessment
          should be completed to determine if head loss constraints occur for the UV facility or
          if booster pumping is needed.
4.2    Operating Approach Selection

       The operating approach is the method of operating a UV reactor based on the dose-
monitoring strategy (Section 3.5.2) and validation report data and should be determined
before the I&C design is complete. The operating conditions for each UV reactor must be based
on validation testing results [40 CFR 141.720(d)(3)].

       As described in Section 3.5.2, this guidance manual focuses on two dose-monitoring
strategies: UV Intensity Setpoint Approach and Calculated Dose Approach. The UV Intensity
Setpoint Approach can be used with a single or variable setpoint operation; variable setpoint
UV Disinfection Guidance Manual                 4-8                              November 2006
For the Final LT2ESWTR

-------
                            4. Design Considerations for UV Facilities
operation allows for some energy savings. The Calculated Dose Approach typically uses a single
setpoint (e.g., the required dose), and the UV equipment automatically compensates based on the
UVT, UV sensor measurements, and flow rate, which increases energy savings.

       When considering the dose-monitoring strategy and operating approach, the operational
complexity should be compared to the potential for energy savings. The UV manufacturer should
be contacted to determine the potential energy savings with the available dose-monitoring
strategies and operating approaches. For small water systems, the UV Intensity Setpoint
Approach with a single setpoint may be the best option because the energy savings with a more
complex operating approach may not be worth the additional operational needs. Detailed
examples of how to determine the operational setpoints from validation reports for these
operating strategies are described in Section 6.1.4.
4.3    Instrumentation and Control

       The necessary level of I&C depends on the selected techniques for flow control and
distribution, flow rate measurement, and the operating approach. For example, passive flow
distribution with the UV Intensity Setpoint Approach that uses a single setpoint is simple and
demands limited I&C but may result in reduced operating flexibility and energy efficiency. More
complex control strategies, such as the use of dedicated flow meters and flow control valves with
the Calculated Dose Approach, necessitate a higher level of I&C, but improve operating
flexibility and enable optimization of disinfection performance. The control system complexity
and operating flexibility should be balanced to meet the needs of the PWS.

       Most of the manufacturers' equipment has similar I&C attributes and alarm  conditions
incorporated in the UV reactor designs. The designer should identify the

       .   Elements that are preprogrammed in the UV reactor control panel

       .   Necessary supplemental controls to coordinate the operations of the UV reactor trains

       .   Actions necessary for each alarm condition.

       At a minimum, UV lamp intensity, flow rate, and lamp status must be monitored (40 CFR
141.720(d)(3)). The final I&C design can be modified as needed after UV equipment is selected.
The following sections describe the elements that should be considered in I&C design.


4.3.1       UV Reactor Start-up and Sequencing

       This section describes the typical UV reactor start-up protocol, strategies for sequencing
the start-up of multiple UV reactors, and considerations for groundwater UV facility start-up.
UV Disinfection Guidance Manual                 4-9                               November 2006
For the Final LT2ESWTR

-------
                            4. Design Considerations for UV Facilities
   4.3.1.1 UV Reactor Start-up

       The UV reactor start-up sequence depends on whether the UV reactor requires cooling
water while the UV lamps warm up. Some reactors require cooling water (Leinberger 2005,
Larner 2005) and some do not (Larner 2005, Bircher 2005). Without water flow, some UV lamps
may heat the water above the safe operating temperatures of 30 - 49 °C in 2.5 - 15 minutes,
causing the reactor's internal safety devices to shut the reactor off (Leinberger 2005, Bircher
2005). LP and LPHO reactors typically do not and some MP reactors do not need cooling water
as the UV lamps are warming up (Haubner 2005). UV lamp breaks (discussed in detail in
Appendix E) can occur if the lamps become overheated because of no flow or minimal cooling
water flow. The designer should consult the UV manufacturer to determine whether the UV
reactor requires cooling water during start-up.

       The potential start-up sequences for UV reactors that do and do not need cooling water
and are starting cold (i.e., previously off as opposed to shut down for a very short period) are
summarized below:

       .   UV reactors that do not require cooling water:  The potential control sequence will
          ignite the lamps, get the UV reactor to its validated conditions, and open the isolation
          valves. With this strategy, the UV reactor will be "on" for some time when no water
          is flowing through it. Flow should be established in the UV reactor within an hour to
          prevent fouling of the quartz sleeves.

          UV reactors that do require cooling water: The potential control sequence will
          open the isolation valves to allow the minimal cooling water flow, ignite the lamps,
          get the UV reactor to its validated conditions, and then fully open the isolation valves
          to allow the full flow through the UV reactor. The I&C  should be designed to reduce
          the amount of off-specification water by providing the minimal flow necessary to
          keep the lamps cool during start-up. If the amount of off-specification water should
          be limited, methods are available to design the UV facility piping to minimize off-
          specification water (e.g., cooling water being diverted to waste).

       For facilities that do not operate continuously, the designer  should discuss the specific
operating schedule with the manufacturer to identify any special provisions that should be
included in the design or the operating procedures (e.g., automatic cleaning before each start-up,
draining for extended periods of downtime).
   4.3.1.2 UV Reactor Sequencing

       UV facilities with multiple UV reactors should develop two types of UV reactor start-up
sequences in I&C loop descriptions:

          Routine operation: The UV reactor sequencing should be developed based on the
          validated conditions and the operational approach.

          Start-up after a power quality event: The control system should monitor the power
          input to the UV reactors and the UV reactor status. LPHO and MP reactors have
UV Disinfection Guidance Manual                4-10                              November 2006
For the Final LT2ESWTR

-------
                            4. Design Considerations for UV Facilities
          different start-up characteristics after a power quality event (Section 2.4.2.3) and
          should have different start-up sequences to minimize warm-up and corresponding off-
          specification time.

          -  LPHO reactors - UV reactors that were on-line (i.e., operating) before the power
             quality event and shut-down should be restarted first after normal power is
             restored.

          -  MP reactors - UV reactors that were off-line before a power quality event that
             shuts down UV rectors should be started first when normal power is restored.
   4.3.1.3 Groundwater Pump Cycling Effects on UV Reactor Start-up

       Groundwater well cycling can adversely affect UV reactor performance, causing an
increase in off-specification water. An analysis should be performed to estimate off-specification
volume based on the current well cycling frequency. The well cycling approach may need to be
changed if off-specification requirements cannot be met under current well cycling frequency.
Two approaches that can minimize the effects of well cycling, depending on whether the UV
reactors require cooling water, are discussed below.

       .   UV reactors that do not require cooling water: A time delay can be incorporated in
          the I&C loops that prevents the well pump from starting until the UV reactor reaches
          its validated conditions. As described in Section 4.3.1.1, the UV reactor will be "on"
          for some time when no water is flowing through it.

       .   UV reactors that do require cooling water: The I&C programming would supply
          the minimum water flow through the UV reactor until the reactor reaches validated
          conditions. Then, the groundwater flow can be increased to meet system demand. If
          desired, the cooling water can be discharged to waste if site conditions permit.
4.3.2       UV Equipment Automation

       UV equipment operation can range from manual to fully automatic, depending on the
reactor's size and complexity. Manual operation includes manually initiating lamp start-up and
shut-down, and activating the appropriate valves. Various levels and types of automation are
typically part of the internal UV equipment controls and can be added to the manual sequence. A
first level of automation includes the sequencing of lamp start-up and valve actuation to bring
individual UV reactors on-line after manual initiation. Further levels of automation include
starting UV reactors, activating rows of lamps, or making lamp intensity adjustments based on
UV intensity, UVT, or flow rate. Automatic UV reactor shut-down under critical alarm
conditions (e.g., high temperature, lamp or sleeve failure, loss of flow) is essential for all
operating approaches, including manual operation.
UV Disinfection Guidance Manual                4-11                              November 2006
For the Final LT2ESWTR

-------
                            4. Design Considerations for UV Facilities
4.3.3       Alarms and Control Systems Interlocks

       Many UV reactor signals and alarms are specific to the UV facility and the level of
automation used. Alarms may be designated as minor, major, or critical, depending on the
severity of the condition being indicated.

       .   A minor alarm generally indicates that a UV reactor requires maintenance but that
          the UV reactor is operating in compliance. Minor alarms also can be set for
          conditions just short of failure conditions so that major alarm conditions are not
          reached. For example, a minor alarm would occur when the UVT is within 1 percent
          UVT of the minimum allowed UVT or when the end-of-lamp-life based on hours of
          operation is reached, indicating the possible need for lamp replacement.

       .   A major alarm indicates that the UV reactor requires immediate maintenance (e.g.,
          the UV sensor value has dropped below the validated setpoint) and that the unit may
          be operating off-specification. Based on the water system's disinfection objectives,
          this condition may also be handled as a critical alarm.

       .   A critical alarm typically shuts the unit down until the cause of the alarm condition
          is remedied. An example of a critical alarm is the UV reactor's temperature exceeding
          a pre-determined maximum value, resulting in automatic shut-down to prevent
          overheating and equipment damage.

       The same alarm condition may represent a different level of severity depending on the
validated conditions, the type of UV reactor, the operating approach, and the disinfection
objectives of the PWS. For example, if a UV reactor was validated with one lamp out of service,
a single lamp failure alarm may trigger a minor alarm. Had the reactor been validated with  all
lamps in operation, a single lamp failure may trigger a major alarm. Table 4.2 summarizes
typical UV reactor monitoring and alarms that would likely be integral to the UV reactor control
panel.
4.3.4       UV Reactor Control Signals

       The designer should coordinate with the UV manufacturer to determine what elements of
the control system are integral to the UV reactor and what elements should be addressed with
supplemental controls and equipment (i.e., supervisory control and data acquisition or SCAD A). For
installations with multiple UV reactors, a common, master control panel may be necessary to
optimize UV reactor operations. Typically, each UV reactor has a dedicated control panel, and
the plant's SCADA system receives control signals from each control panel to control the entire
UV facility.  The SCADA system also monitors and records the process parameters.
Recommended monitoring and recording frequencies are provided in Chapter 6, and the designer
should coordinate with the state to determine  if expected frequencies differ. This section
describes the control signals that could be transferred from each reactor's control panel  to the
SCADA system.
UV Disinfection Guidance Manual                4-12                              November 2006
For the Final LT2ESWTR

-------
                             4. Design Considerations for UV Facilities
   4.3.4.1 UV Intensity

       Signals from UV sensors should be displayed locally on the UV reactor control panel and
in the plant's SCADA system screen (if applicable).
                Table 4.2. Typical Alarm Conditions for UV Reactors1
Sensor
Lamp Age
Calibration Check of UV sensor
Low UV Validated Dose
Low UV Intensity
Low UV Transmittance
High Flow Rate (Not Integral to UV
Reactor — Relies on Flow Meters)
Mechanical Wiper Function Failure (If
Applicable)
Lamp/Ballast Failure
Low Liquid Level
High Temperature
Alarm Type
Minor alarm
Minor alarm
Major alarm
Major alarm
Major alarm
Major alarm2
Major alarm
Major alarm
Critical alarm
Critical alarm
Critical alarm
Purpose/Description
Run-time for lamp indicates end of defined
operational lamp life.
UV sensor requires calibration check based
on operating time.
Indicated validated UV dose (based on UV
reactor parameters, i.e., flow rate, UV
intensity, and UVT) falls below required UV
dose.
Intensity falls below validated conditions.
UVT falls below validated conditions.
Flow rate falls outside of validated range.
Wipe function fails.
Single lamp/ballast failure identified.3
Multiple lamp/ballast failures identified.
Liquid level within the UV reactor drops and
potential for overheating increases.
Temperature within the UV reactor or ballast
exceeds a setpoint.
  Alarm conditions and relative severity shown above may vary depending on the specific validated conditions, type
  of UV reactor, manufacturer, dose-monitoring strategy, and disinfection objectives of the PWS.
  Based on measurement from dedicated flow meters or calculated based on total flow rate divided by number of UV
  reactors operating.
  Coordinate with UV manufacturer to determine if a lamp/ballast failure could indicate a sleeve and lamp break,
  which should be classified as a critical alarm.
   4.3.4.2 UV Transmittance

       If the Calculated Dose Approach is used, the UVT must be known to ensure that it is
within the validated range. An on-line UVT analyzer or a bench-top spectrophotometer may be
used to monitor UVT. Output from an on-line UVT analyzer can be input directly into a control
loop for most UV reactors, a SCADA system, or both. Results from a bench-top
spectrophotometer can be manually input into a SCADA system or UV reactor control panel(s).
UV Disinfection Guidance Manual
For the Final LT2ESWTR
4-13
November 2006

-------
                           4. Design Considerations for UV Facilities
   4.3.4.3 Flow Rate Measurement

       To maintain regulatory compliance, the flow rate through a UV reactor must be known to
ensure that it is within the validated range [40 CFR 141.720 (d)(2)]. Section 4.1.2 discusses flow
rate measurement and control options. The flow rate signal should be displayed locally or be
input directly into a control loop for the UV reactor, a SCADA system, or both.
   4.3.4.4 Calculated and Validated UV Dose (If Applicable)

       If the Calculated Dose Approach is used, the calculated and validated doses should be
displayed locally and transmitted to the SCADA system. The validated dose is equal to the
calculated dose divided by the Validation Factor (See Section 5.10 for details).
   4.3.4.5 Operational Setpoints

       The operational setpoints should be displayed locally and remotely in the SCADA
system. These setpoints will depend on the specific dose-monitoring strategy, operating approach
(Section 4.2), and the validation data, and may include UV intensity, UVT, flow rate, calculated
dose, and validated dose.
   4.3.4.6 Lamp Age

       The operating time of each lamp should be monitored, displayed locally, and transmitted
to the SCADA system to facilitate O&M and lamp replacement, as discussed in Section 6.3.2.6.
   4.3.4.7 Lamp Power, Lamp Status, and Reactor Status

       Water systems must monitor lamp status to verify that UV reactors are operating within
validated conditions [40 CFR 141.720(d)(3)]. Lamp status refers to whether the lamp is "on" or
"off." The operating power level should also be monitored and displayed at the control panel and
remotely in the SCADA system. Each reactor's on-line or off-line status should also be
monitored and indicated locally and remotely, which can be accomplished by monitoring power
and valve status.
   4.3.4.8 UV Reactor Sleeve Cleaning

       Sleeve cleaning information should be displayed locally and communicated between the
local control panels and the SCADA system. This information should include the date and time
of the last cleaning for off-line chemical cleaning (OCC) systems and the wiping frequency for
on-line mechanical cleaning (OMC) or on-line mechanical-chemical cleaning (OMCC) systems.
UV Disinfection Guidance Manual                4-14                             November 2006
For the Final LT2ESWTR

-------
                           4. Design Considerations for UV Facilities
   4.3.4.9 Alarms

       At a minimum, alarm conditions should be displayed locally. The use of visual or audible
alarms is also recommended. If the UV facility will frequently be unstaffed or a SCADA system
is already in place, provisions should also be included in the design to allow remote monitoring
and display through the SCADA system.
4.4    Electrical Power Configuration and Back-up Power

       The electrical power configuration should take into account the power requirements of
the selected equipment, the disinfection objectives, and power quality issues, if applicable. (See
Section 3.4.6.)
4.4.1       Considerations for Electrical Power

       The proper supply voltage and total load requirements should be coordinated with the UV
manufacturer, considering the available power supply. In addition, the power needs for each UV
reactor component may differ. For example, the UV reactor may require 3-phase, 480-volt
service, while the on-line UVT analyzer may need single-phase, 110-volt service. Excluding
high service pumping, the electrical load from UV reactors will typically be among the larger
loads at the WTP.

       Due to the varying nature of UV reactor loads, current and voltage harmonic distortion
can be induced. Such disturbances can cause electrical system problems, including overheating
of some power supply components and can affect other critical systems, such as variable
frequency drives (VFDs), programmable logic controllers (PLCs), and computers. Selection of
the UV reactors should be based on a thorough analysis of the potential for the equipment to
induce harmonic distortion. Additionally, the UV facility design and UV equipment should meet
the Institute of Electrical and Electronic Engineers (IEEE) 519 Standard that addresses
harmonics.

       One method for controlling harmonics is to use a transformer with Delta Wye
connections to isolate the UV reactor from the remainder of the WTP power system. The Delta-
connected primary feed can be designed and sized to trap and moderate any induced harmonics.
The Wye-connected secondary should be solidly grounded so that the ballasts are powered from
a grounded source in accordance with electrical code requirements. If a separate transformer for
the UV reactors is impractical, harmonic filters  can be added to the UV reactor power supply to
control distortion.
4.4.2       Back-up Power Supply and Power Conditioning

       The continuous operation of the UV reactor is highly dependent on the power supply and
its quality (Section 3.4.6). If the power reliability requirements and, consequently, the
disinfection objectives cannot be met by relying solely on the commercial power supply, the use
of back-up power, power conditioning equipment, or both may be necessary.

UV Disinfection Guidance Manual               4-15                             November 2006
For the Final LT2ESWTR

-------
                            4. Design Considerations for UV Facilities
   4.4.2.1 Back-up Power Supply

       A simple backup power supply (e.g., generator) may be sufficient if power quality issues
are infrequent. If an existing backup power supply is in place, its load capacity should be
assessed to determine whether it can accept the additional load associated with the UV facility.
The time necessary for switching from the primary power supply to a backup power supply and
how that time affects compliance with the allowable off-specification operation should be
determined.
   4.4.2.2 Power Conditioning Equipment

       Power conditioning equipment may be necessary if the power quality analysis reveals
frequent events (Section 3.4.6) that cause the UV facility not to meet disinfection objectives. A
site-specific analysis should be completed to determine the most appropriate power conditioning
approach (Cotton et al. 2005). Consideration should include off-specification compliance, quality
of the power supply, the cost of power conditioning equipment, and site constraints (e.g., land
availability).

       .   Uninterruptible Power Supply (UPS) systems provide continuous power in the
          event of voltage sag or power interruption. The battery capacity is large enough to
          supply power to all connected equipment until a generator starts. UPS systems can
          either be on-line or off-line:

          On-line UPS: The unit and batteries are installed in series between the incoming
          power feed and all critical  equipment. The incoming power feed charges the UPS
          batteries, and the batteries  supply the electrical load. In this situation, the power feed
          is completely separated from the electrical load. This alternative is the most costly
          and has the largest footprint.

          Off-line UPS: The unit is installed in parallel with the connection from the incoming
          power feed to the critical equipment. During normal operations, the electrical load
          receives power directly from the power feed. When the off-line UPS senses a voltage
          fluctuation greater than or  less than 10  percent of the nominal voltage, the  load
          transfers to the UPS until the electrical feed stabilizes or the generator starts. Off-line
          UPS systems are less costly and have a smaller footprint than on-line UPS systems.

       .   Active Series Compensators protect electrical equipment against momentary voltage
          sags and interruptions. These devices boost the voltage by injecting a voltage in series
          with the remaining voltage during a sag condition. The corrected response time is a
          fraction of a cycle, preventing the equipment from experiencing a voltage sag. Active
          series compensators are well  suited for instantaneous sags and interruptions; however,
          they cannot correct sustained sags or interruptions. Active series compensators are the
          lowest cost and smallest power conditioning option.
UV Disinfection Guidance Manual                4-16                              November 2006
For the Final LT2ESWTR

-------
                            4. Design Considerations for UV Facilities
4.4.3       Ground Fault Interrupt and Electrical Lockout

       Proper grounding and insulation of electrical components are critical for protecting
operators from electrical shock and protecting the equipment. When combined with effective
lockout/tagout procedures, the risk of electrical shock is further minimized. Ground fault
interrupt (GFI) is another important safety feature for any electrical system in contact with water,
including UV reactors. All UV reactor suppliers should provide GFI circuits for their lamps,
which should be included in the specifications developed for equipment procurement. For a GFI
to function properly, the transformer in the UV reactor ballast must not be isolated from the
ground. If the UV reactor ballast isolates the output from the ground, ground faults will not be
properly detected, and safety can be compromised.

       Provisions enabling the UV reactors to be isolated and locked out for maintenance, both
hydraulically and electrically, should be included in the design. Control of all lockout systems
should remain local; however, when appropriate, the status of local lockouts could be monitored
remotely. In all cases, the design must comply with electrical code and policy requirements for
equipment lockout.
4.5    UV Facility Layout

       Site layout for a UV facility is generally similar to the layout for any treatment process.
Access for construction, operation, and maintenance should be considered. Typically, a
preliminary layout is developed during project planning (Section 3.8.2). This preliminary layout
may be modified to address space constraints or special installation conditions that result from
the final equipment selection or based  on more extensive site information gathered during
detailed design. In addition to those items identified in Section 3.8.2, this section describes the
items to be considered in the more detailed layout developed in the design phase.

       Components of the UV reactors are typically located inside a building for protection from
the weather and to provide a clean, convenient area for maintenance. The UV reactors
themselves, associated electrical components and controls, and electrical support equipment
should be enclosed. In some installations, UV reactors and control panels are uncovered. Before
designing an uncovered facility, however, the state and UV manufacturer should be consulted.
Exposed equipment and control panels should be rated for the anticipated environment, and
appropriate site security should be in place to restrict public access.

       The piping, valve, and flow meter design developed in the hydraulic evaluation
(Section 4.1) should be considered in the UV facility configuration. For example, the length of
straight-run piping before and after each flow meter to achieve the proper hydraulic conditions
for accurate and repeatable flow rate measurement should be considered in the piping layout,
depending on the flow control and measurement technique used (Section 4.1.2).

       The location of the power and control panels associated with UV reactors should allow
adequate space for panel doors to be opened without interference, and to allow unhindered
access to the UV reactors when the doors  are open. In selecting the location of the power and
control panels, UV manufacturer cable length limitations should not be exceeded. The maximum
allowable cable length is UV manufacturer-specific and may be less than 30 feet. If power
UV Disinfection Guidance Manual                4-17                               November 2006
For the Final LT2ESWTR

-------
                            4. Design Considerations for UV Facilities
quality is a concern, room for power conditioning equipment should be provided. Such
equipment may be located adjacent to the UV reactors or in a separate control room.

       When allotting space for maintenance activities, adequate space to remove the lamps and
the lamp wiper assembly should be provided. In some cases, access may be needed on both sides
of the UV reactor. In addition, provisions should be included to collect and convey water that is
discharged during maintenance activities.

       Certain UV reactors need maintenance involving an OCC procedure in which a UV
reactor is taken off-line, isolated, drained, filled with a cleaning solution, cleaned, flushed, and
returned to service. The OCC equipment is typically self-contained and the cleaning chemical is
recirculated. If applicable, sufficient space  should be maintained around the UV reactors to
provide access for the OCC procedure. Also, the OCC solution often has specific handling
requirements. Appropriate drains, storage, and health and safety equipment (e.g., emergency
eyewash station) should be provided as recommended by the chemical manufacturer.

       Sample taps in the lateral pipe are recommended upstream and downstream of each UV
reactor. The sample taps may be used for collecting water quality samples or  during validation
testing, if on-site validation is necessary. If on-site validation will be conducted, the number and
location of sample and feed ports should be coordinated with the UV manufacturer or third-party
oversight entity to comply with the recommendations of the selected validation protocol.
Additional details on the locations of sample taps and other validation-related appurtenances are
provided in Section 5.4.

       Drain valves or  plugs should be located on each lateral between the two isolation valves.
In many cases, the UV manufacturer may have already incorporated a drain into the UV reactor
design. Drain valves should also be provided at one or more low points in the UV facility to
enable the UV reactor and entire lateral to be fully drained for maintenance activities. These
drains should be large enough to drain the reactor and adjacent piping in a reasonable amount of
time.

       Additionally, the UVT analyzer installation (if necessary) should be considered in the
layout. The specific size and operating characteristics of the UVT analyzer will vary depending
on the UV manufacturer. If an on-line UVT analyzer is included in the design, adequate space
and access to an electrical supply for monitor installation should be provided  and appropriate
sample taps and drains for withdrawing and discharging sample water should be included in the
design. The sample line should be equipped with a valve to isolate the UVT analyzer. A sample
pump (e.g., peristaltic) should be installed if insufficient pressure is available in the system. The
UVT analyzer should be in a location that minimizes the likelihood of air bubbles (which can
cause erroneous readings) passing through  the monitor.
4.5.1  Additional Considerations for Unfiltered and Uncovered Reservoir UV
       Facility Layouts

       Site issues that should be considered with unfiltered systems are generally consistent with
those for filtered surface water systems. The most significant difference is the increased
opportunity for debris to be present in the inflow to UV reactors in unfiltered applications. To

UV Disinfection Guidance Manual                4-18                              November 2006
For the Final LT2ESWTR

-------
                            4. Design Considerations for UV Facilities
address the increased potential for debris, UV facility designs for unfiltered applications should
incorporate features that prevent potentially damaging objects from entering the UV reactor. The
optimal approach is site-specific. Such features could include screens, baffles, or low-velocity
collection areas. Another option is to install the UV reactors vertically with the inlet closest to
the ground, following a low-velocity zone. This arrangement will decrease the momentum of
larger debris and reduce the risk of lamp breakage. The effects of lamp breakage and methods for
minimizing it are discussed in Appendix E.
4.5.2       Additional Considerations for Groundwater UV Facility Layouts

       Site issues that should be considered with groundwater systems are generally consistent
with those for post-filtration surface water systems; the most significant difference is access of
the site and potential sand particles affecting the UV reactor. Because well sites can be located in
remote areas and may be more accessible to the general public or unauthorized individuals, the
UV reactor should be installed within a building to protect sensitive equipment. The need to
enclose the UV facility will ultimately be based on the manufacturer's recommendations, local
regulatory and code requirements, state requirements, environmental conditions, and site-specific
constraints. Site security should be appropriate to prevent tampering with the equipment and
water supply and to protect people from injury (e.g., electrocution).

       In addition, sand or debris flowing through the UV reactor may scratch the lamp sleeves
or cause the sleeve wiping mechanisms to jam. Larger sand and debris could break the lamp
sleeves and lamps. (See Appendix E for lamp breakage issues.) Intermittently used wells may
accumulate sand or other particles; this initial concentration of particles should be discharged
before operation and should bypass the UV reactor to avoid scratching the quartz sleeves. A
sand/debris trap or other removal equipment prior to UV disinfection may be necessary if
evidence suggests that the well pump will pull any sand or particles through the screen during
normal well operation.
4.6    Elements of UV Equipment Specifications

       When procuring the UV reactors, the UV facility layout and UV reactor specification are
typically provided to the UV manufacturer. This section describes the potential elements
included in a UV reactor specification and outlines the information that could be requested from
the UV manufacturer.
4.6.1       UV Equipment Specification Components

       Table 4.3 summarizes the factors that should be considered when developing equipment
specifications for the UV equipment. The information included in Table 4.3 is not exhaustive and
should be modified to meet the specific needs of the PWS and the requirements for the UV
facility.
UV Disinfection Guidance Manual                4-19                              November 2006
For the Final LT2ESWTR

-------
                          4. Design Considerations for UV Facilities
           Table 4.3. Possible Content for UV Equipment Specifications
                         (Table Spans Pages 4-20 - 4-22)
Item
Flow rate
Target Pathogen (s) and Log
Inactivation
Required UV Dose
Water Quality and Environment
Operating Flow and UVT Matrix
Operating Pressure
UV Sensors
Redundancy
Hydraulics
Specification Content
Maximum, minimum, and average flow rates should be clearly
identified. The maximum flow rate must be within the validated
range documented in the validation report [40 CFR 141 .720
(d)(3)]. The minimum flow rate may be important to avoid
overheating with MP reactors. One method for determining the
maximum flow rate is described in Section 3.4.3.
The log inactivation for the target pathogen(s)
The required UV dose for the target microorganism and log
inactivation that must be verified by the validation process.
Additional detail is provided in Chapter 5.
The following water quality criteria should be included:
- Influent temperature - pH
- Turbidity - Iron
- UV transmittance at 254 nm - Calcium
- UVT scan from 200 - 300 - Manganese
nm (MP reactors only) - ORP
- Total hardness
For some parameters, a design range may be most appropriate.
Appropriate matrix of paired flow and UVT values based on flow
and UVT data (Section 3.4.4.1).
The expected operating pressures, including the maximum and
minimum operating pressure to be withstood by the lamp sleeves
and UV reactor housing.
A germicidal spectral response should be specified (Section 5.4.8).
A minimum of one UV sensor should be specified per UV reactor.
The actual number should be identical to the UV reactor that was,
or will be, validated.
The uncertainty of the UV sensors used during validation should
meet the criteria described in Section 5.5.4.
The uncertainty of the duty UV sensors during operation should
meet the criteria described in Section 6.4.1 .1 .
Reference UV sensors should be calibrated against a traceable
standard. For example, the following standards are currently being
used by UV manufacturers:
- National Physical Laboratory (NPL)
- National Institute of Standards and Technology (NIST)
- Deutsche Vereinigung des Gas- und Wasserfaches (DVGW)
- Osterreichisches Normungsinstitut (ONORM)
The reactor redundancy determined in Section 3.8.1 .
The following hydraulic information should be specified:
- Maximum system pressure at the UV reactor
- Maximum allowable head loss through the UV reactor
- Special surge conditions that may be experienced
- Hydraulic constraints based on site-specific conditions and
validated conditions (e.g., upstream and downstream straight
pipe lengths).
UV Disinfection Guidance Manual
For the Final LT2ESWTR
4-20
November 2006

-------
                              4. Design Considerations for UV Facilities
            Table 4.3. Possible Content for UV Equipment Specifications
                             (Table Spans Pages 4-20 - 4-22)
               Item
                   Specification Content
Size/Location Constraints
Any size constraints or restrictions on the location of the UV
reactor or control panels (e.g., space constraints with individual
filter effluent installation).
Validation
The range of operating conditions (e.g., flow, UVT) that must be
included in the validation testing, and submittal of a validation
report (40 CFR 141.720) should be required. The validation testing
should be completed in accordance to the procedures and data
analysis described in detail in Chapters.
Dose-Monitoring Strategy
A description of the preferred dose-monitoring strategy for the UV
reactors.
Operating Approach
A description of the intended operating approach for the UV
reactors, as described in Section 4.2.
Economic and Non-Economic
Factors
The necessary information to thoroughly evaluate the UV
equipment based on the PWS's specific goals. As appropriate, this
information may include both economic (e.g., energy use,
chemical use) and non-economic (e.g., future expansion,
manufacturer experience) factors.
Lamp Sleeves
Lamp sleeves should be annealed to minimize internal stress.
Safeguards
At a minimum, the following UV reactor alarms should be
specified:
  - Lamp or ballast failure
  - Low UV intensity or low validated UV dose (depending on
   dose-monitoring strategy used)
  - High temperature
  - Operating conditions outside of validated range
  - Wiper failure (as applicable)
  - Other alarms discussed in Section 4.3.3, as appropriate.
Instrumentation and Control
At a minimum the following signals and indications should be
specified:
  - UV lamp status
  - UV reactor status
  - UV intensity
  - Lamp cleaning cycle and history
  - Accumulated run time for individual lamps or banks of lamps
  - Influent flow rate.

At a minimum the following UV reactor controls (as applicable)
should be specified:
  - UV dose setpoints, UV intensity setpoints, or UVT setpoints
   (depending on dose-monitoring strategy used)
  - UV lamps on/off
  - UV reactor on/off control
  - UV reactor manual/auto control
  - UV reactor local/remote control
  - Manual lamp power level control
  - Manual lamp cleaning cycle control
  - Automatic lamp cleaning cycle setpoint control.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
           4-21
November 2006

-------
                            4. Design Considerations for UV Facilities
            Table 4.3. Possible Content for UV Equipment Specifications
                           (Table Spans Pages 4-20 - 4-22)
              Item
                                       Specification Content
Performance Guarantee
                    The equipment provided should meet the performance
                    requirements stated in the specification for an identified period or
                    during on-site performance testing (Section 6.1.5). The following
                    specific performance criteria may be included:
                      - Allowable head loss at each design flow rate
                      - Estimated power consumption under the design operating
                        conditions
                      - Disinfection capacity of each reactor under the design water
                        quality conditions
                      - Sensitivity of equipment to variations in voltage or current
                      - Reference UV sensor, duty UV sensor,  and UVT analyzer (if
                        provided) performance compared to specification
Warranties
                    A physical equipment guarantee and UV lamp guarantee should
                    be specified. The specific requirements of these guarantees will be
                    at the discretion of the PWS and engineer. Significant variation
                    from common commercial standards should be discussed with the
                    manufacturer. Lamps should be warranted to provide the lamp
                    intensity under the design conditions for the fouling/aging factor
                    and a minimum number of operating hours. To limit the UV
                    manufacturer's liability, the guarantee could be prorated after a
                    specified number of operating hours.
UVT Analyzer
                     During operation, the difference between the UVT analyzer
                     measurement and the UVT measured by a calibrated
                     spectrophotometer should be less than or equal to 2 % UVT.
4.6.2
Information Provided by Manufacturer in UV Reactor Bid
       The UV manufacturers should provide adequate information when bidding to enable the
designer to conduct a proper, timely review of the proposed equipment. Suggested information to
be obtained from the UV manufacturer is presented in Table 4.4.
       Table 4.4. Suggested Information to Be Provided by UV Manufacturer
Item
Design Parameters
Summary of Design
Reactor Technical
Specifications
UV Equipment
Documentation and
Specifications
UV Manufacturer's
Experience
UV Lamps
Description of Information
Demonstration of an understanding of the design parameters for the UV equipment.
All UV equipment design parameters from the contract documents should be
repeated in the proposed UV equipment submittal information.
A summary of the equipment proposed (number of UV reactors, lamp type) and
specified equipment redundancies.
Ability of proposed UV reactors to meet technical specifications and an explanation
of any exceptions taken.
Documentation that identifies and describes the UV equipment components that
were validated, as described in Section 5.1 1 .1 .1
Information on project experience, including previous facilities and references.
Detailed description of the lamp dimensions and electrical requirements.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                               4-22
November 2006

-------
                            4. Design Considerations for UV Facilities
       Table 4.4. Suggested Information to Be Provided by UV Manufacturer
Item
UV Sensor
Lamp Sleeves
UVT Analyzer
(if applicable)
Validation Report
Upstream and
Downstream
Hydraulic
Requirements
Power
Requirements
Power Quality
Tolerance
Cleaning Strategy
Dose-monitoring
Strategy
Reactor Data
Safeguards
Warranties
Description of Information
Information on the UVsensor(s), including spectral response, acceptance angle,
external dimensions, working range in mW/cm2, measurement uncertainty,
environmental requirements, linearity, and temperature stability.
Data and calculations should be provided showing how the total measurement
uncertainty of the UV sensor used during validation meets the criteria established
in Section 5.5.4.
Data that demonstrate duty UV sensors will meet the criteria described in Section
6.4.1.1 will be met during operation.
Calculations showing the maximum allowable pressure for the lamp sleeves and the
maximum bending stress the lamp sleeves may experience under the maximum
specified flow rate conditions.
Data that prove the UVT analyzer used during validation meets the criterion in
Section 6.4.1 .2 during operation.
UV reactor validation should be provided that includes the elements described in
Section 5.11.3. If on-site validation is proposed, validation data for the UV reactors
from off-site validation (if completed) should be included to provide a baseline
comparison to the proposed conditions.
A statement of the length of straight pipe and hydraulic conditions necessary
upstream and downstream from the UV reactor to ensure the desired flow profile is
maintained and the design conditions are met. If pre-validated equipment is
specified, a description of the hydraulic configuration used during validation testing
should be provided.
The power needs of each UV reactor and which elements, including electrical cable
and wiring, are included as part of the equipment.
The power quality tolerance of the UV equipment for voltage sags, surges, and
interruptions.
The strategy that will be used for cleaning the UV lamps in the UV reactor.
The proposed UV reactor dose-monitoring strategy, including manual and automatic
control schemes and a listing of inputs, outputs, and the types of signals that are
available for remote monitoring and control.
The materials of construction, dimensions of the UV reactors and ancillary
equipment, a list of spare parts, and a sample operations and maintenance manual.
The safeguards built into the UV reactor and accompanying equipment, such as
high temperature protection, wiper failure alarms, and lamp failure alarms.
A statement of the proposed UV reactor guarantees, including the physical
equipment, UV lamps, lamp sleeves, fouling/aging factor, and the system
performance guarantee. Any exceptions should be indicated and explained.
 Key elements of this documentation are also listed in this table.
4.7    Final UV Facility Design

       The UV reactors can be selected after all bids have been carefully reviewed. Once the UV
reactors are selected, the designer can coordinate with the selected UV manufacturer to develop
the final facility design based on the selected UV equipment. The hydraulic design, I&C design,
electrical design, and facility layout should be modified based on the selected UV equipment.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
4-23
November 2006

-------
                            4. Design Considerations for UV Facilities
       Particular emphasis should be given to the integration of the overall dose-monitoring
strategy with the alarms, signals, and interlocks that are integral to the UV reactor design. That
the final design be coordinated with the validation testing results is critical. The validation results
must be sufficient to implement the proposed operations approach and should meet the water
supply's disinfection objectives under the specified operating conditions.
4.8    Reporting to the State during Design

       Interaction with the state throughout the design phases is recommended and increases the
likelihood that the objectives of both the PWS and the state are met. Currently many states have
limited experience in the use of this technology; therefore, the appropriate level of state
involvement during design should be greater than that for more traditional designs. Early
agreement on the specific objectives and requirements of the project can significantly reduce the
potential for conflict or costly design changes later in the project. The level of state involvement
during design, as well as the specific submittal  requirements, will vary by  state and may vary by
project. PWSs are urged to consult with their state early in their UV disinfection design process
to understand what approvals and documentation will be required.
UV Disinfection Guidance Manual                4-24                               November 2006
For the Final LT2ESWTR

-------
                           5. Validation of UV Reactors
       The purpose of validation testing is to determine the operating conditions under which a
UV reactor delivers the validated dose.1 As noted elsewhere in this guidance document, the
validated dose must be greater than or equal to the required dose (presented in Table 1.4) to
receive log inactivation credit for a target pathogen. Validation testing also establishes the
operational setpoints used during reactor operations to confirm delivery of the validated dose.

       This chapter explains the key steps in EPA's recommended validation protocol for UV
reactors. It includes recommendations for selecting test conditions, quality assurance/quality
control (QA/QC) steps, and data analysis procedures.  It provides the rationale for the
recommended steps and cites relevant research studies where appropriate.
       Chapter 5 covers:
       5.1  Minimum Requirements for Validation Testing
       5.2  Overview of the Recommended Validation Protocol
       5.3  Selecting the Challenge Microorganism
       5.4  Equipment Needs for Full-scale Reactor Testing
       5.5  Accuracy of Measurement Equipment
       5.6  Identifying Test Conditions
       5.7  Guidelines for Conducting Experimental Tests
       5.8  Analyzing Experimental Data
       5.9  Deriving the Validation Factor (VF)
       5.10 Determining the Validated Dose and Validated Operating Conditions
       5.11 Documentation
       5.12 Guidelines for Reviewing Validation Reports
       5.13 Evaluating the Need for "Re-validation"
       Several appendices support this chapter:

       .   Appendix A provides recommendations for preparing stock solutions of and assaying
          challenge microorganisms.

       .   Appendix B presents validation testing examples for two hypothetical water systems.

       .   Appendix C provides the recommended procedure for conducting collimated beam
          tests, including test conditions, apparatus design, equipment accuracy, and QA/QC.
          Appendix C also provides guidelines for using collimated beam test data to develop a
          UV dose-response curve.
1  For the purposes of this guidance manual, the validated dose is defined as the UV dose in units of millijoule per
  centimeter squared (mJ/cm2) delivered by the UV reactor as determined through validation testing. The required
  dose is defined as the UV dose needed to achieve log inactivation credit. All UV dose terms are included in the
  glossary at the beginning of this manual.
UV Disinfection Guidance Manual                 5-1                               November 2006
For the Final LT2ESWTR

-------
                                  5. Validation of UV Reactors
       .   Appendix D contains the background theory used to support the recommended
           validation protocol.

       The material in this chapter is intended to help water systems and states understand how
 the validation testing process works. It should be considered as a resource when reviewing
 validation reports or overseeing validation testing activities.  Some of the terms and acronyms
 used in this chapter are unique to UV reactor validation. EPA has included an extensive glossary
 and acronyms list at the beginning of this guidance manual to help the reader keep track of new
 terms.


 5.1    Minimum Requirements for Validation Testing

       Unlike chemical disinfection, UV light leaves no residual that can be monitored to
 determine the delivered dose. UV sensors can measure intensity of UV light, but they cannot
 measure the dose delivered to the microorganisms as they pass through the reactor at different
 trajectories. Therefore, to receive treatment credit for inactivating Cryptosporidium, Giardia, or
 viruses using UV light, the LT2ESWTR requires water systems to use UV reactors that have
 undergone validation testing.

       Section 1.4 of this manual summarizes all LT2ESWTR requirements related to UV
 disinfection, including minimum dose, validation, monitoring, and reporting requirements. For
 easy reference Table 5.1 summarizes the regulatory requirements for validation.
             Table 5.1. Summary of LT2ESWTR Validation Requirements
   Requirement
                 Conditions
       Citation
Validated operating
conditions must
include
Flow rate
UV intensity as measured by a UV sensor
UV lamp status	
 40CFR141.720(d)(2)
Validation testing
must include 1
Full-scale testing of a reactor that conforms uniformly
to the UV reactors used by the water system
Inactivation of a test microorganism whose dose-
response characteristics have been quantified with a
low-pressure mercury vapor lamp	
40CFR141.720(d)(2)(ii)
Validation testing
must account for
UV absorbance of the water
Lamp fouling and aging
Measurement uncertainty of on-line sensors
UV dose distributions arising from the velocity profiles
through the reactor
Failure of UV lamps or other critical components
Inlet and outlet piping or channel configurations of the
UV reactor
40CFR141.720(d)(2)(i)
  The state may approve an alternative approach to validation testing.
 UV Disinfection Guidance Manual
 For the Final LT2ESWTR
                        5-2
         November 2006

-------
                                 5. Validation of UV Reactors
       The LT2ESWTR does not specifically address "re-validation" if the design of a validated
UV reactor changes. If design modifications significantly impact UV dose delivery or
monitoring, however, the UV reactor should be considered a different reactor (with
unsubstantiated performance) than the one previously validated and as such, must be re-validated
[40 CFR 141.720 (d)(2)]. Section 5.13 discusses some of the common types of UV reactor
modifications and provides recommendations for which types of changes necessitate re-
validation.
5.2    Overview of the Recommended Validation Protocol

       EPA's recommended validation protocol uses biodosimetry. Under this approach, the log
inactivation of a challenge microorganism is measured during full-scale reactor testing for
specific operating conditions of flow rate, UV transmittance (UVT),2 and UV intensity. The
dose-response equation for the challenge microorganism (relating UV dose to log inactivation) is
determined using independent, bench-scale testing. Log-inactivation values from full-scale
testing are input into the laboratory derived-UV dose-response relationship to estimate the
Reduction Equivalent Dose (RED). The RED value is adjusted for uncertainties and biases to
produce the validated dose of the reactor for the specific operating conditions tested. The
validated dose is compared to the required dose for compliance purposes.

       The protocol can be described in three main steps, as shown in Figure 5.1 and described
in more detail in Section 5.2.1. Alternative approaches to validation are discussed in Section
5.2.2. Sections 5.2.3 and 5.2.4 present recommendations for third-party oversight and emerging
validation approaches, respectively.
5.2.1  Key Steps in Recommended Validation Protocol

       Consistent with other recommendations in this guidance manual, EPA developed the
validation protocol working closely with industry representatives and experts in the field of UV
disinfection. EPA believes that the approach produces reliable results and can be used to meet
microbial treatment requirements of the LT2ESWTR and encourages water systems to use it
where applicable. Water systems are not required, however, to follow the protocol as long as
they meet the minimum regulatory requirements summarized in Section 5.1. EPA recommends
that water systems contact their states early to discuss any additional state-specific requirements
for reactor validation.
2 In this Chapter, UVT implies UVT measurement specifically at 254 nanometers (nm) and 1 centimeter (cm)
  pathlength unless otherwise noted. UV absorbance at 254 nm (A254) can be related to UVT using the following
  equation:
 .        (UVT(%)}
A,,, = -log	^-L
 254     \  100   J

UV Disinfection Guidance Manual                5-3                               November 2006
For the Final LT2ESWTR

-------
                                    5. Validation of UV Reactors
              Figure 5.1. Overview of Recommended Validation Protocol
              Step 1. Experimental Testing Using a Challenge Microorganism*

               1a: Bench Scale Testing                     1b: Full-Scale Reactor Testing
                         Collimated UV rays
                         Petri dish with challenge
                         microorganism
          Measure the log inactivation for different
          UV doses to develop a UV dose-
          response curve:
         UV dose
         (mJ/cm2)
                                                      Inject challenge
                                                      microorganism
                              Measure UV
                              intensity with
                              a UV sensor.
                  	CT-
                  UV Reactor
                                                    Measure influent
                                                    flow rate, UVT, and
                                                    microorganism
                                                    concentration
                         Measure effluent
                         microorganism
                         concentration,
                         compare to influent to
                         calculate the log
                         inactivation
                   Log inactivation
                  Step 2. Determine the Reduction Equivalent Dose (RED)

                 Input the log inactivation from Step 1b into the dose-response curve from Step
                 1a to estimate RED.
                         UV dose
                         (mJ/cm2)
           •Dose-response curve
            ( from step la)
                                                     Log inactivation (from step Ib)
                          Step 3. Adjust for Uncertainty to Calculate
                                       the Validated Dose

                                     Validated Dose = RED / VF

                          Where VF = Validation Factor that accounts for biases and
                          experimental uncertainty.
            * Simple representations of testing equipment shown.  For more details, see Figures C. 1 and 5.2.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
5-4
November 2006

-------
                                  5. Validation of UV Reactors
Step 1: Conduct Experimental Tests Using a Challenge Microorganism

       Because handling of the target pathogen during validation testing is neither practical nor
in the best interest of public health3, a challenge microorganism whose sensitivity to UV light is
similar to the target pathogen should be used in all experiments. Using a challenge
microorganism instead of the target pathogen, however, introduces uncertainty into the testing
results. This uncertainty is accounted for by applying a validation factor (see Step 3).

       UV reactor validation includes two types of experimental tests, as described below.
Importantly, the same stock solution of challenge microorganisms should be used for both tests
because UV sensitivity for stock solutions can vary.

       la. Bench-scale testing using a collimated beam apparatus. Collimated beam testing
          characterizes the UV dose-response relationship of the  challenge microorganism. In
          these experiments, UV light is directed down a collimating tube to dose a sample of
          challenge microorganisms of a known concentration. After a specified exposure time,
          the sample is analyzed to determine the log inactivation (where log inactivation in
          this situation equals the log concentration prior to UV light exposure minus the log
          concentration after UV light exposure) as a function of UV dose. The UV dose
          delivered to the microorganisms is calculated based on the UV intensity, exposure
          time, and other experimental factors. Figure C.I  in Appendix C illustrates a typical
          collimated beam apparatus.

          Collimated beam tests are performed at a range of doses to generate a VV dose-
          response curve for the specific challenge microorganism.  The functional forms of the
          equations for UV dose-response curves can vary depending on the results (guidance
          on developing UV dose-response curves is provided in Section C.3). A quadratic UV
          dose-response equation is provided below.

           UV Dose = Ax (log inactivation) + B x (log inactivation)2             Equation 5.1

          For this equation type, the coefficients "A" and "B" would be solved for using the
          collimated beam testing data.

       Ib. Full-scale reactor testing. In these experiments, the challenge microorganisms are
          injected upstream of the UV reactor. Samples are analyzed to determine the log
          inactivation (where log inactivation in this situation equals log influent concentration
          minus log effluent concentration) at the test conditions  of flow rate, UVT, lamp
          status, and UV intensity as measured by UV sensors. Full-scale reactor testing can be
          performed on-site at a water treatment plant or off-site at a validation test center (see
          Figure  5.2 in Section 5.4 for a diagram of a typical biodosimetry test stand used for
          off-site validation).
3 Culturing pathogenic microorganisms introduces additional risks in terms of handling, disposal, and cross
  connections. Therefore, the industry regularly uses challenge microorganisms as surrogates for pathogenic target
  microorganisms.
UV Disinfection Guidance Manual                 5-5                                November 2006
For the Final LT2ESWTR

-------
                                 5. Validation of UV Reactors
Step 2: Estimate the Reduction Equivalent Dose

       Step 2 combines results from the two experimental tests in Step 1. The log inactivation of
the challenge microorganism measured during the full-scale testing is entered into the UV dose-
response equation (Equation 5.1 if the relationship is quadratic) to calculate the RED of the
reactor. Another way to conceptualize this step is to consider the RED to be "back-calculated"
using the field-measured log inactivation as the input variable. This approach is the opposite of
most applications in which UV dose is the independent variable and log inactivation is the
dependent variable.

       RED values are always specific to the following:

        .   The challenge microorganism used during experimental testing,

           Validation test conditions during full-scale reactor tests (flow rate, UVT,  lamp status,
           and UV intensity as measured by the UV sensor)

Step 3: Adjust for Uncertainty to Calculate the Validated Dose

       In the last step shown in Figure 5.1, the RED is divided by a Validation Factor (VF) to
produce the Validated Dose. The VF accounts for biases associated with using a challenge
microorganism instead of the target pathogen and for experimental uncertainty (Section 5.9
provides a detailed description of how the VF is derived). The Validated Dose is associated  with
the validation test conditions of flow rate, lamp status, UV intensity as measured by a UV sensor,
and in some cases, UVT.  As noted previously, the validated dose is compared to the required
dose to determine the inactivation credit for the target pathogen.
5.2.2  Alternative Validation Protocols

       The Austrian Standards ONORMM5873-1 andM5873-2 (2001 and 2003, respectively)
and the German Guideline DVGW W294 (2006) define measured flow rate, UV intensity, and
lamp status for a Bacillus subtilis RED of 40 mJ/cm2. Based on the recommended validation
protocol presented in this guidance manual, UV reactors certified by ONORM and DVGW for a
B. subtilis RED of 40 mJ/cm2 should be granted 3-log Cryptosporidium and 3-log Giardia
inactivation credit. Validation by NWRl/AwwaRF Guidelines and NSF Standard 55 should be
evaluated on a case-by-case basis (NWRI2003, NSF 2004).
5.2.3  Third Party Oversight

       EPA recommends that an independent third party provide oversight to ensure that
validation testing and data analyses are conducted in a technically sound manner and without
bias. A person independent of the UV reactor manufacturer should oversee validation testing.
Individuals qualified for such oversight include engineers experienced in testing and evaluating
UV reactors and scientists experienced in the microbial aspects of biodosimetry. Appropriate
individuals should have no real or apparent conflicts of interest regarding the ultimate use of the
UV reactor being tested.

UV Disinfection Guidance Manual                5-6                               November 2006
For the Final LT2ESWTR

-------
                                  5. Validation of UV Reactors
       At a minimum, independent oversight should include observing validation testing to
verify that the individuals performing the validation follow the documented protocol and
reviewing the report for accurate data and results. The independent third party should review the
validation report before its release. When appropriate, the third party should rely on additional
outside experts to review various aspects of UV validation testing, such as lamp physics, optics,
hydraulics, microbiology, and electronics.
5.2.4  Emerging Methods

       In recent years, researchers have been working on alternative approaches to biodosimetry
for UV reactor validation. Potential model-based approaches use computational fluid dynamics
(CFD) to predict microorganism trajectories through a UV reactor, and hence the UV dose
delivered to each microorganism.  Section D.6 in Appendix D describes certain aspects of using
CFD to predict UV dose delivery. A possible approach for verifying and validating hydraulic
CFD models is outlined in the AIAA CFD guide (1998). Another emerging experimental
approach uses microspheres that undergo a chemical reaction when exposed to UV light
(Blatchley et al.  2005). The microspheres are injected upstream of the UV reactor and are
collected downstream. The extent of the UV light-induced chemical reaction within each sphere
is measured and used to calculate the UV dose delivered to that sphere as it traveled through the
reactor.

       Although model and experimental-based approaches clearly have potential for use in
validating UV reactors, they are still subjects of current research. EPA anticipates that these
methods will continue to  develop and improve in the future.
5.3    Selecting the Challenge Microorganism

       For the reasons stated in Section 5.2.1, the disinfection performance of the UV reactor is
measured using a non-pathogenic "challenge" microorganism. Ideally, the challenge
microorganism should have the same sensitivity to UV light (i.e., the same microbial dose-
response) as the target pathogen.4 If medium-pressure (MP) lamps are used, the organism should
display a similar action spectrum, which is the relative sensitivity of the organism at other
wavelengths compared to its sensitivity at 254 nm. In addition, the challenge microorganism
should be:

       .   Easily cultured and enumerated, with repeatable results,

          Culturable to high concentrations, and

       .   Stable over long periods of time (to allow for shipment to and from the laboratory,
          on-site use, and enumeration without loss of viability or change in UV dose-
          response).
4 In this guidance document, the UV sensitivity of the target microorganisms Cryptosporidium, Giardia, and
  viruses is defined by the required UV doses as presented in Table 1.4.
UV Disinfection Guidance Manual                5-7                               November 2006
For the Final LT2ESWTR

-------
                                 5. Validation of UV Reactors
          If the challenge microorganism is a phage, the host bacteria used to assay the phage
          concentration should not be pathogenic to humans.

       Male-specific-2 bacteriophage (MS2) phage and B. subtilis spores historically have been
used for validation testing to receive treatment credit for Cryptosporidium and Giardia. Because
their UV resistance is notably greater than that of Cryptosporidium and Giardia, other, more
sensitive microorganisms such as Tl and T7 phage are gaining favor (Mackey et al. 2006).

       To demonstrate 3- or 4-log inactivation for viruses, validation testing would need to
demonstrate greater than 6-log inactivation of MS2 phage and B. subtilis spores. Such a
demonstration requires an extremely high concentration in the reactor influent to allow for
enumeration of the organisms in the effluent samples. Because of the need for serial dilutions,
these high concentrations are difficult to measure and can introduce  error into the experiment.
Research to find an alternative challenge microorganism for demonstrating virus inactivation is
ongoing.

       Other challenge microorganisms that have been used for validation testing include non-
pathogenic Escherichia coli, Saccharomyces cerevisae, and QP phage. Table 5.2 summarizes the
UV sensitivity of some commonly used and some candidate bioassay microorganisms.
               Table 5.2. UV Sensitivity of Challenge Microorganisms
Microorganism
Bacillus subtilis
MS2 phage
QB phage
PRD-1 phage
B40-8 phage
c|)x174 phage
£. co/;
T7
T1
Reported Delivered UV Dose (mJ/cm'1)
to Achieve Indicated Log Inactivation
1-log
28
16
10.9
9.9
12
2.2
3.0
3.6
c
O
2-log
39
34
22.5
17
18
5.3
4.8
7.5
-10
3-log
50
52
34.6
24
23
7.3
6.7
11.8
-15
4-log
62
71
47.6
30
28
11
8.4
16.6
-20
Reference
Sommeret al. 1998
Wilson et al. 1992
Mackey et al. 2006
Meng and Gerba 1996
Sommer et al. 1998
Sommer et al. 1998
Chang et al. 1985
Mackey et al. 2006
Wright 2006
       Some microorganisms, such as B. subtilis, exhibit shoulders or tailing in their UV dose-
response, meaning that the shape of the UV dose-response curve is flat at either low or high
doses. Shoulders and tailing limit the range of UV doses that can be used to validate the reactor.
See Section C.6 in Appendix C for an example of shoulders and tailing and limitations of using
challenge microorganisms exhibiting this response in developing the UV dose-response curve.

       As noted in Section 5.2.1, the validation test results are adjusted by a VF to account for
bias and experimental uncertainties. A portion of the VF accounts  for the difference in microbial
response between the challenge microorganism and target pathogen. Using a challenge organism
with significantly higher UV resistance than the pathogen of interest (e.g., using MS2 to earn
Cryptosporidium inactivation credit) may result in a high VF. To provide a better estimate of the
UV dose that a UV reactor can deliver to a target pathogen, a challenge microorganism with
similar UV sensitivity to the target pathogen can be used. Alternatively, two challenge
UV Disinfection Guidance Manual
For the Final LT2ESWTR
5-8
November 2006

-------
                                 5. Validation of UV Reactors
microorganisms whose UV sensitivities bracket that of the target pathogen (i.e., one challenge
microorganism is less resistant and the other is more resistant than the target pathogen) can be
selected. One advantage to this second approach is that the factor used to account for the
difference between the  microbial response of the challenge microorganism and target pathogen
can be set to 1.0 (see Section 5.9 for discussion of the RED bias factor).

       If the UV reactor being validated uses MP lamps and a challenge microorganism other
than MS2 Phage or B. subtilis, a correction factor should be applied to the test results to account
for differences in action spectra between the challenge microorganism and the target pathogen.
Section D.4.1 explains  how the correction factor is derived and how it should be applied to
validation testing results.


5.4    Equipment Needs for Full-scale Reactor Testing

       As noted in Section 3.6, full-scale reactor validation can occur on-site at the water
treatment plant or off-site at a third-party validation test center or a UV manufacturer's facility.
If full-scale reactor testing is performed off-site, tests are often conducted using a Biodosimetry
Test Stand as shown in Figure 5.2. Regardless of the testing location, testing equipment should
include (1) injection pumps and ports to introduce the challenge microorganism, the UV-
absorbing compound, and, if needed, a disinfectant residual quenching agent into the feed water,
(2) rate-of-flow control and a flow meter either upstream or downstream of the reactor, and (3) a
strategy to ensure that the water is well mixed before sampling (e.g., static mixers or appropriate
number of pipe lengths with good mixing confirmed, see Section 5.4.3 for details). There should
also be a state-approved plan for wastewater disposal with any associated  required permits.

       The next several sections provide detailed recommendations regarding water source, the
UV-absorbing chemical to be used to simulate reduced UVT, mixing, sampling ports,
configuration of inlet/outlet piping, accounting for non-uniform lamp aging, lamp positioning,
UV sensors, and UV sensor port windows.


5.4.1   Water Source

       When validation testing is conducted off-site, the source water for experiments is usually
a potable water supply  with a high UVT. To protect the potable water source, backflow
prevention should be provided.

       The water passing through the reactor should not contain disinfectant residuals that
inactivate the challenge microorganism during testing. If this is a concern,  a quenching agent  can
be injected into the water upstream of the microorganism injection port. When validating UV
reactors using MP lamps, the quenching  agent should have a minimal impact on the spectral
UVT from 200 to 400 nm. Some common quenching agents like sodium thiosulfate can have  a
significant impact on the UV absorbance spectra if added in high enough concentrations. Testers
should use a quenching agent, such as sodium bisulfite, that does not influence UVT.
UV Disinfection Guidance Manual                5-9                               November 2006
For the Final LT2ESWTR

-------
                                 5. Validation of UV Reactors
        Figure 5.2. Block Diagram of a Typical Biodosimetry Test Stand for
                              Full-scale Reactor Testing
                    UV    Challenge
                  Absorber   Microbe
        Water
        Supply
                Backflow
               Prevention
      Pressure     UV  Pressure
££   ^X  ^ ^   Static
                             Mixer   Valve
                           ~H*1—*
                                       To
                                      Waste
                       Influent    Influent
                      Quenching   Sample
                        Agent       Port
                              Effluent
                              Sample
                                Port
5.4.2  UV Absorbing Chemical

       Typically during validation, a UV-absorbing chemical is injected into the flow to produce
UVT values that span the required range. Common UV-absorbing compounds include the
following:

   .   Coffee,

   .   Lignin sulfonic acid (LSA), and

   .   Humic acids,  such as those derived from leonardite shales (Mackey et al. 2006, Bircher
       2004).


5.4.3  Mixing

       Additives passing through the reactor (e.g., UV-absorbing compound, the challenge
microorganism) should be well mixed through the cross-section of the influent pipe prior to the
reactor influent sampling port. The challenge microorganisms surviving UV disinfection should
be well mixed through the effluent pipe cross-section prior to the reactor effluent sampling port.

       Additives to the influent and effluent can be mixed by either using appropriately sized
and designed static mixers or relying on the turbulent mixing in the lengths of pipe upstream of
the sampling ports. If the water passing through the UV reactor is obtained from a large tank, the
additives can be premixed in the tank to obtain a uniform concentration for testing. If pumps are
used to inject the additives, the mechanism should provide a pulse-free flow rate or have a cycle
time (i.e., time between pulses) an order of magnitude less than the residence time of the reactor.
The flow rate generated by the pump should be stable over the time required to take samples.

       Adequate mixing at the influent sampling port can be confirmed by comparing the UV
absorbance at 254 nm (A25/t) of water samples collected from various locations across the pipe
cross-section. Samples can be collected across a pipe section using a perforated stab tube. The
UV Disinfection Guidance Manual
For the Final LT2ESWTR
     5-10
November 2006

-------
                                 5. Validation of UV Reactors
standard deviation of the A254 values measured at different locations should be less than 5 percent
of the mean A254 value. Another approach is to compare the A254 of water samples collected at
the influent and effluent sampling ports. The average A2s4 values of the influent and effluent
samples should agree within 5 percent and the standard deviation of each should be less than 5
percent of their respective means.

       The mixing at the effluent sampling port can be confirmed by injecting a UV-absorbing
chemical into the flow at a location immediately downstream of the UV reactor and comparing
the A254 of water samples collected from various locations across the pipe cross-section.
Alternatively, the A2s4 of water samples collected at the effluent sampling port and a second
effluent sampling port located five pipe diameters or more downstream of the first can be
compared. The water samples should meet the criteria given above for the influent samples.

       Mixing tests should be done at the minimum and maximum flow rates with the UVT
adjusted to  the minimum value that will be used during testing. If the water samples collected
during the tests do not meet the above criteria for good mixing, the mixing should be increased
and retested.
5.4.4  Sampling Ports

       The sampling points for microorganisms should be located far enough from the UV
reactor that the germicidal UV intensity at the point of sampling is < 0. 1 percent of the
germicidal intensity within the UV reactor. If the outlet sample port is located downstream of a
90° bend (or the inlet sample port is upstream of a 90° bend), incident light is not a concern.

       To estimate intensity at a certain distance from the reactor, the following equation can be
used:
       /(/•) = -e~a*r                                              Equation 5.2
       where:
          PL  =  UV power emitted per unit arc length of the line source (mW/cm)
           r  =  Radial distance from the line source (cm)
          ae  =  Naperian (base e) absorption coefficient of the media for water, (-0.015 cm"1)
         I(r)  =  UV intensity (mW/cm2) at a distance r from the line source

For example, suppose the outlet sample port at a hypothetical UV test facility is located 10 feet
downstream of the last UV lamp in a reactor. The lamp's maximum power per unit arc length is
100 watts per centimeter (W/cm). Using Equation 5.2, the intensity at a radial distance of 10 feet
(305 cm) is calculated to be 5.4 x 10"4 mW/cm. Because this intensity is less than 0. 1 percent of
the intensity within the reactor, the sample port location is acceptable.

       Sample taps may draw water from a single point or simultaneously from multiple points
across the pipe diameter. Samples taken from multiple points within the flow should have the
same concentration of additives and microorganisms (within the measurement error of the
analytical method). If samples from different points in the flow have different concentrations, the
UV Disinfection Guidance Manual                5-11                              November 2006
For the Final LT2ESWTR

-------
                                 5. Validation of UV Reactors
flow at the sampling point is either insufficiently mixed or not at steady state. Sampling from
multiple points at the same time should not be used to compensate for poor mixing.
5.4.5  Configuration of Inlet and Outlet Piping

       Appendix D describes how flow conditions of the water can significantly impact dose
delivery inside a UV reactor. Flow conditions are dependent on the velocity of the water and
configuration of inlet and outlet piping. If the reactor is validated off-site, the inlet and outlet
piping at the water treatment plant should result in a UV dose that is the same or greater than
the dose delivered at the validation test facility. Section 3.6 provides suggestions for inlet and
outlet piping design for water treatment plants and validation testing facilities.

       Computational fluid dynamics (CFD) is a tool that can be used to assess whether the dose
delivery at the treatment plant is the same or greater than the dose delivery at the validation
testing facility. Section D.6 in Appendix D provides guidelines on using CFD for the purposes of
modeling UV dose delivery.
5.4.6  Accounting for Non-uniform Lamp Aging

       As will be discussed in Section 5.6, validation testing of full-scale reactors should
account for decreased UV light transmittance caused by sleeve fouling, sleeve aging, lamp aging,
and UV sensor window fouling. During design, the engineer and UV manufacturer will typically
estimate a "fouling factor" and an "aging factor" for the reactor. The fouling factor is defined as
the fraction of UV light passing through a fouled sleeve as compared to a new sleeve. The aging
factor is the fraction of UV light emitted from aged lamps and sleeves at  the end of the specified
useful life compared to UV light emitted from new lamps and sleeves. The "fouling/aging
factor" is equal to the fouling factor multiplied by the aging factor and typically ranges from 0.4
to 0.9 (NWRI 2003).  See Section 3.4.5 for a more detailed discussion on  determining these
factors.

       Traditionally, lamp power is turned down to simulate aging and fouling during validation
testing. The magnitude of the power reduction (or power turn-down) is determined by calculating
the relative sensor intensity., which is defined as follows:

       Relative sensor intensity = S/S0                                       Equation 5.3

       Where
          S0 =  UV intensity  measured at 100 percent lamp power
           S =  UV intensity  measured at reduced lamp power

For example, if the fouling/aging factor is 0.7, the lamp power would be  reduced until the
relative sensor intensity was 0.7, or 70 percent.

       Recent studies have shown, however, that UV lamps and sleeves  can exhibit significant
non-uniform aging along their length and around their circumference (e.g., Mackey et al. 2005

UV Disinfection Guidance Manual               5-12                              November 2006
For the Final LT2ESWTR

-------
                                 5. Validation of UV Reactors
and 2006). Turning down power during validation testing to simulate aging for lamps that
experience non-uniform aging may result in under-dosing when the reactor is operated at the
treatment plants, particularly at the end of useful lamp life. This guidance manual recommends
that water systems use one of the following three alternatives to account for non-uniform lamp
aging:

       1.  Request data from the manufacturer to verify that the lamps age uniformly. The
          manufacturer can provide such verification by simulating lamp aging in a test bed,
          then measuring lamp  output at different locations along the length of the lamp and
          around the circumference. Results from a recently-completed  AwwaRF study showed
          that output at the lamp ends is usually less than in the middle of the lamp when
          significant non-uniform aging is observed (Mackey et al. 2006). If the manufacturer
          can show that the lamp aging factor either already accounts for non-uniform aging or
          that it is not an issue,  power turn-down can be used to simulate lamp aging during
          validation tests.

       2.  Use aged lamps (i.e.,  lamps that have been operated under similar conditions to the
          end of their specified lamp life) for validation testing. Power turn-down to simulate
          lamp aging during validation tests is not necessary in this approach (although power
          turn-down should still be considered to simulate lamp fouling).

       3.  Conduct experimental testing to determine if lamp aging can be simulated by power
          turn-down:

          a.  Prepare a stock solution of the challenge microorganism.

          b.  Fit the UV reactor with aged lamps and sleeves.

          c.  Pass water through the reactor at a constant UVT, flow rate, and lamp power that
             will be used during challenge testing.

          d.  Inject the challenge microorganism into the flow passing through the reactor
             (ensure they are well-mixed prior to entering the reactor).

          e.  Collect at least five (5) microbiological samples spaced one (1) minute apart from
             the influent and effluent sampling ports for analysis.

          f  Record the UV sensor measurements.

          g.  Fit the UV reactor with new lamps that have undergone 100-hour burn-in and new
             sleeves.

          h.  Operate the UV reactor at the flow rate and UVT used in Step c. Lower the lamp
             power to produce a UV sensor reading equivalent to the reading obtained in Step
             f. Repeat steps e and f.

          If the mean log inactivation achieved with aged lamps is similar to the log
          inactivation achieved with new lamps with reduced lamp power, power turn-down

UV Disinfection Guidance Manual                5-13                              November 2006
For the Final LT2ESWTR

-------
                                 5. Validation of UV Reactors
          can be used to simulate lamp aging. If results show significant differences,
          Alternative 2 (aged lamps and sleeves) should be used during validation testing.
5.4.7  Lamp Positioning to Address Lamp Variability

       Due to manufacturing tolerances and differences in operation and aging, UV output
typically varies from lamp to lamp. If a UV reactor has fewer UV sensors than lamps and the
lamps are randomly distributed in the reactor, the UV sensors may monitor the lamps with the
lowest outputs during validation. If this were to occur, the validation data collected would
typically lead to under-dosing at the treatment plant.

       To prevent underdosing, the lamps with the highest UV output should be placed closest
to the UV sensors during validation testing. Other lamps should be randomly distributed in the
lamp array throughout the reactor.  This approach is unnecessary if the reactor uses one UV
sensor per lamp.

       The lamps with the highest UV output can be identified by measuring UV output using
either a dedicated test stand or the UV reactor. One approach for determining the UV lamp
output using the UV reactor is described below.

       Procedure

       1.  Install a lamp within a lamp sleeve located at the position nearest to one of the
          reactor's UV sensors.

       2.  Pass water through the reactor at a constant flow rate and UVT.

       3.  With only the lamp under evaluation on, record the measured UV intensity.

       4.  Repeat the test for each lamp and rank the results.

       5.  For full-scale reactor testing, install the lamps with the highest output closest to the
          UV sensors. The rest of the lamps should be randomly distributed (with respect to
          lamp intensity).
5.4.8  UV Sensors

       UV sensors are photosensitive detectors that measure UV intensity. UV sensors used in
drinking water UV applications, particularly those with MP or other polychromatic lamps,
should be germicidal. Germicidal sensors are defined as having the following properties:

       A spectral response (i.e., UV intensity measured at various wavelengths) that peaks
       between 250 and 280 nanometers (nm).

   .   Less than 10 percent of its total measurement is due to light above 300 nm when mounted
       on the UV reactor and viewing the UV lamps through the water that will be treated.

UV Disinfection Guidance Manual               5-14                              November 2006
For the Final LT2ESWTR

-------
                                 5. Validation of UV Reactors
       Manufacturers should document the spectral response of the UV sensors. Tables 5.3 and
5.4 provide two examples of spectral response for a hypothetical germicidal sensor and a
hypothetical non-germicidal sensor, respectively. Figure 5.3 graphically depicts the spectral
response for the data in Tables 5.3 and 5.4, respectively.

       Table 5.3 shows that nearly 100 percent of the area under the spectral response curve
between 200 and 400 nm is at wavelengths below 300 nm - almost none of the sensor reading is
due to wavelengths greater than 300 nm. Moreover, the peak spectral response is at 270 nm.
Therefore, this UV sensor is classified as germicidal. Conversely,  Table 5.4 reveals that only 74
percent of the area under the curve is below 300 nm, which means that 26 percent of the area is
measured at wavelengths greater than 300 nm. Because 26 percent is greater than the maximum
allowable 10 percent, this UV sensor is classified as non-germicidal.

       EPA recognizes that, before the publication of this document, some UV reactors using
MP lamps with non-germicidal sensors are in the final design phases, and some have been
installed at water treatment plants. These water systems should apply a correction factor to
validation test data to account for polychromatic bias. Section D.4 in Appendix D explains how
polychromatic bias impacts sensor reading and provides guidelines for deriving the correction
factor. As noted in Chapter 4, facilities installing new UV treatment systems should use reactors
that are equipped with germicidal sensors.
5.4.9  UV Sensor Port Windows

       UV sensors often view the lamps through a UV sensor port window. These windows are
typically made of quartz and have a UVT greater than 90 percent. The UVT of the sensor port
windows should be checked before and after validation testing. If the sensor port windows are
fouled or contaminated, UV sensor readings will be low. If this were to occur during validation
testing, it could lead to under-dosing at the water treatment plant whenever the sensor port
windows are clean. A collimated beam apparatus and a radiometer can be used to measure the
sensor port window UVT either before the reactor is shipped to the test site or during validation
testing.
5.5    Accuracy of Measurement Equipment

       During validation testing, all equipment should be carefully selected and calibrated to
minimize uncertainty. All measurements of flow rate, electrical power consumption, and head
loss5 should be traceable to an independent standard. Moreover, because they are key parameters
that affect UV dose delivery, measurements of UVT and UV intensity should be NIST6-traceable
or equivalent7 with a known measurement uncertainty.
  Although not part of UV validation, headless as a function of flow rate is often measured during validation testing
  as it offers an opportunity to gather such design data on the system
  National Institute of Science and Technology
7 For example, the German national testing and standards agency, Physikalisch Technische Bundesanstalt (PTB), or
  the United Kingdom's National Weights and Measures Laboratory.
UV Disinfection Guidance Manual                 5-15                              November 2006
For the Final LT2ESWTR

-------
                                 5. Validation of UV Reactors
           Table 5.3. Hypothetical Example of the Spectral Response of a
                                Germicidal UV Sensor
Wavelength
(nm)
200
210
220
225
230
235
240
245
250
255
260
265
270
275
280
285
290
295
300
310
320
330
340
350
360
370
380
390
400
Spectral Response1
(mW/cm2)
0.11
0.21
0.30
0.35
0.40
0.48
0.58
0.72
0.88
1.03
1.15
1.23
1.30
1.21
0.30
0.19
0.08
0.03
0.02
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
Area
Under the Curve
Between Readings
-
1.600^
2.550 J
1.625
1.875
2.200
2.650
3.250
4.000
4.775
5.450
5.950
6.325
6.275
3.775
1.225
0.675
0.275
0.125
0.150
0.050
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
Cumulative
Area Under the
Curve
-
1.6
4.15
5.775
7.65
9.85
12.5
15.75
19.75
24.525
29.975
35.925
42.25
48.525
52.3
53.525
54.2
54.475
54.6
54.75
54.8
54.8
54.8
54.8
54.8
54.8
54.8
54.8
54.8
Cumulative Area
as % of Total Area
-
3%
8%
11%
14%
18%
23%
29%
36%
45%
55%
66%
77%
89%
95%
98%
99%
99%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
 UV intensity measured by the UV sensor.
!(0.21 + 0.11) x (210-200)72
' (0.30 + 0.21 )x (220-210)72
 UV Disinfection Guidance Manual
 For the Final LT2ESWTR
5-16
November 2006

-------
                                5. Validation of UV Reactors
          Table 5.4. Hypothetical Example of the Spectral Response of a
                             Non-germicidal UV Sensor
Wavelength
(nm)
200
210
220
225
230
235
240
245
250
255
260
265
270
275
280
285
290
295
300
310
320
330
340
350
360
370
380
390
400
Spectral
Response 1
(mW/cm2)
0.12
0.30
0.47
0.55
0.63
0.71
0.79
0.87
0.94
1.01
1.07
1.12
1.14
1.15
1.12
1.07
1.00
0.91
0.82
0.63
0.48
0.38
0.31
0.23
0.15
0.10
0.06
0.03
0.00
Area
Under the Curve
Between Readings
-
2.100"
3.850 J
2.550
2.950
3.350
3.750
4.150
4.525
4.875
5.200
5.475
5.650
5.725
5.675
5.475
5.175
4.775
4.325
7.250
5.550
4.300
3.450
2.700
1.900
1.250
0.800
0.450
0.150
Cumulative Area
Under the Curve
-
2.1
5.95
8.5
11.45
14.8
18.55
22.7
27.225
32.1
37.3
42.775
48.425
54.15
59.825
65.3
70.475
75.25
79.575
86.825
92.375
96.675
100.125
102.825
104.725
105.975
106.775
107.225
107.375
Cumulative Area as %
of Total Area
-
2%
6%
8%
11%
14%
17%
21%
25%
30%
35%
40%
45%
50%
56%
61%
66%
70%
74%
81%
86%
90%
93%
96%
98%
99%
99%
100%
100%
 UV intensity measured by the UV sensor.
'(0.31 +0.12)x (210-200)72
5 (0.47 + 0.30) x (220-210/2
UV Disinfection Guidance Manual
For the Final LT2ESWTR
5-17
November 2006

-------
                                5. Validation of UV Reactors
          Figure 5.3. Hypothetical Examples of the Spectral Response of a
                    Germicidal and a Non-germicidal UV Sensor
          u
          0)
          (A

          I
          (A
          0)
          U
          0)
          Q.
                                                        x Germicidal
                                                      - •*- Non-germicidal
              0.0
                                        'ft '
                 200
250           300
        Wavelength (nm)
350
400
       Source: Data from Tables 5.3 (germicidal sensor) and 5.4 (non-germicidal sensor).
       The next three sections provide recommendations for verifying measurement uncertainty
for flow meters, UV spectrophotometers, and power consumption. Section 5.5.4 provides the
recommended approach for determining the measurement uncertainty of UV sensors, which the
LT2ESWTR requires. Tests verifying equipment accuracy (particularly UV sensor checks as
described in Section 5.5.4) should be documented in the Validation Report (See Section 5.11 for
guidance).
5.5.1  Flow Meters

       During validation testing, the uncertainty of flow rate measurements should be less than
or equal to 5percent. The measurement uncertainty of the flow meter can be verified by
comparing measured flow rate to a second, calibrated flow meter or a calibrated pitometer.
5.5.2  UV Spectrophotometers

       Spectrophotometer measurements of A254 should be verified using NIST-traceable
potassium dichromate UV absorbance standards and holmium oxide UV wavelength standards.
Many UV spectrophotometers have their own internal QA/QC procedures to verify calibration.
UV absorbance of solutions used to zero the spectrophotometer should be verified using reagent-
grade organic-free water certified by the supplier to have zero UV absorbance.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
           5-18
               November 2006

-------
                                 5. Validation of UV Reactors
       The measurement uncertainty of the spectrophotometer should be 10 percent or less. A
recommended approach for achieving this goal is as follows:

       1.  Verify that the spectrophotometer reads the wavelength to within the accuracy of a
          holmium oxide standard (typically + 0.11 nm at a 95-percent confidence level),

       2.  Verify that the spectrophotometer reads A254 within the accuracy of a dichromate
          standard (e.g., 0.281 + 0.004 cm"1 at 257 nm with a 20 mg/L standard), and

       3.  Verify that the water used to zero the instrument has an A254 value that is within
          0.002 cm"1 of a certified zero absorbance solution.

       When the UVT is greater  than 90 percent, it is recommended that a 4-cm or greater
pathlength cuvette be used (as opposed to the standard 1-cm cuvette). This greatly improves the
accuracy of the UVT measurement at values above 90 percent. Measurements made with a 4-cm
cuvette can be converted to 1-cm UVT measurements using the following equation:

              UrT1_em=UVT^m114                                            Equation 5.4

       When validation testing is performed using unfiltered water, the UV spectrophotometer
should be equipped with an integrating sphere, which will provide more accurate UV absorbance
readings if there are particles in the water.

       If the spectrophotometer provides biased readings, the measurements  should be corrected
to account for that bias, or another instrument with measurement uncertainty  of 10 percent or less
should be used.
5.5.3  Power Measurements

       Voltmeters, ammeters, and power meters used to measure (1) ballast and UV equipment
input voltage, and (2) consumed current and power, should bear evidence of being in calibration
(e.g., have a tag showing that it was calibrated). The accuracy of the measurements can be
verified using a second instrument or a standard measurement. Power meters should provide a
measure of true power as opposed to apparent power in units of kilovolt ampere (kVA).
5.5.4  UV Sensors

       During validation testing, duty UV sensor measurements should be within 10percent of
the average of two or more reference sensor measurements.8 Duty sensors that do not meet this
criterion should be replaced or the measurement uncertainty should be incorporated into the VF
(see Section 5.9).
8  Note that this error range is smaller than recommended for operations (in Section 6.4.1.1, EPA recommends that
  sensor readings be within 20 percent of the average of two or more reference sensors). EPA believes that a 10-
  percent error is easily attainable during validation testing and will help ensure good data quality for developing
  operational setpoints.
UV Disinfection Guidance Manual                 5-19                              November 2006
For the Final LT2ESWTR

-------
                                 5. Validation of UV Reactors
       The following procedure can be used to check the uncertainty of the duty and reference
UV sensors used during validation (this calculation is also illustrated in the examples in
Appendix B).

       1.  Pass water through the reactor at the maximum UVT and the maximum lamp power
          setting to be used during validation testing.

       2.  Using two recently calibrated reference UV sensors (which should agree within the
          calibration certificate-specified measurement uncertainty9), install each sensor on the
          UV reactor at each port and record the measured UV intensity (Sref). Repeat using
          each duty UV sensor (Sduty). If the UV sensors can be rotated, measure the minimum
          and maximum sensor readings across the complete range of rotation.

       3.  Repeat Steps 1 and 2 at either (a) maximum lamp power and UVT decreased to the
          minimum value expected during validation testing, or (b) maximum UVT and lamp
          power decreased to the minimum level expected to occur during validation testing.
          Duty UV sensors can be checked under both conditions, although this is not
          necessary.

       4.  For a given lamp  output and UVT value, the difference between the reference and
          duty UV sensor measurements should follow Equation 5.5:
                c
                °duty  _ ,
                 t.avg
                         <10%                                               Equation 5.5
          where:
            Sduty  =   Intensity measured by a duty UV sensor
        SAvg Ref  =   Average UV intensity measured by all the reference UV sensors in the
                     same UV sensor port with the same UV lamp at the same UV lamp power.

       5.  Duty and reference UV sensors that do not meet this criterion should be replaced.
          Alternatively, measurement uncertainty can be re-stated at the maximum uncertainty
          observed during validation testing and incorporated into the VF (see Section 5.9).

       Duty sensors should be checked prior to full-scale reactor testing to ensure that the data
collected during testing will be useful. Sensors are also often spot-checked during and after full-
scale testing to verify that they are still within the recommended uncertainty limit.
5.6    Identifying Test Conditions

       Numerous combinations of experimental tests can be performed to validate a UV reactor.
The number of tests could range from a few tests to a complex matrix spanning a range of UV
9 If the reference sensors do not agree with the calibration certificate-specified measurement uncertainty, they should
  be sent back to the manufacturer.
UV Disinfection Guidance Manual                5-20                              November 2006
For the Final LT2ESWTR

-------
                                  5. Validation of UV Reactors
dose, flow rate, UVT, ballast power, and lamp status combinations. The test design (i.e., number
of tests and test conditions) depends on several factors, as summarized in Table 5.5.
           Table 5.5. Factors to be Considered in Validation Test Design
Factor
1 . Purpose of validation testing
2. Dose-monitoring strategy of the UV
reactor
3. Operational strategy (for the UV
Intensity Setpoint Approach only)
4. Predicted lamp aging and fouling
5. Target pathogen and target log
inactivation
6. Full operating range of flow rate and
UVT
Examples
Validation of new reactor by water
system vs. validation to confirm an
existing validation equation
UV Intensity Setpoint Approach vs. the
Calculated Dose Approach1
Single-setpoint operations vs. variable-
setpoint operations
Aging factor of 0.8 vs. using aged lamps
used during validation testing (where the
aging factor would equal 1 .0)
2.0 log inactivation of Cryptosporidium
vs. 3.0 log inactivation of viruses
Range of flow = 5-20 mgd, Range of
UVT = 70-90 %
Section of Manual
with Additional
Guidance
3.6
5.6
3.5.2
3.5.2
3.4, 5.4.6
3.1
3.4
  As noted in Section 3.5.2, there are many dose-monitoring strategies for UV reactors. This guidance manual
  focuses on two common strategies, the UV Intensity Setpoint Approach and the Calculated Dose Approach.
       Although all factors in Table 5.5 influence test design, the total number of experiments is
highly dependent on the first three factors. For example, suppose a water system wants to
validate a new UV reactor that uses the UV Intensity Setpoint Approach. The system decides to
use single-setpoint operations, meaning that it will use one UV intensity setpoint for all operating
conditions. Validation testing in this case would be fairly straightforward with a small number of
tests. If another water system selects the same reactor but selects variable-setpoint operations to
allow it to reduce lamp power at low flow rates, that system would conduct more validation tests
to establish different setpoints for the different flow rate ranges. Another common scenario is
when a manufacturer decides to validate a new UV reactor over a wide range of flow rates, UVT
levels, and lamp status  combinations to develop a dose-monitoring equation. This scenario likely
would necessitate many tests.

       As noted in Section 5.4.6, lamp fouling and aging are important factors that should be
accounted for during validation testing. Power turn-down is typically used to simulate lamp
aging and fouling at the end-of-lamp life. Instead of reducing power to simulate lamp aging,
aged lamps can be used during validation testing (although power-turn down would still be
needed to simulate lamp fouling). Section 3.4.5 provides information on how the fouling/aging
factors are estimated, and Section 5.4.6 provides guidance on using new versus aged lamps
during validation testing.

       If a new, un-validated reactor is being tested for a specific water system, the last two
items listed in Table 5.5 can help establish validation test conditions. The target pathogen and
target log inactivation for the water system define the required dose that is the target for
UV Disinfection Guidance Manual
For the Final LT2ESWTR
5-21
November 2006

-------
                                 5. Validation of UV Reactors
validation testing. The full range of operating conditions for flow rate and UVT dictate the flow
rate and UVT conditions used during validation testing.

       Sections 5.6.1 and 5.6.2 discuss the validation test design for the UV Intensity Setpoint
Approach and the Calculated Dose Approach, respectively. Test designs for other dose-
monitoring strategies that use setpoints should be similar to recommendations in Section 5.6.1
and should be developed using professional judgment. Section 5.6.3 provides considerations for
water systems who are confirming an existing dose-monitoring equation (as developed for the
Calculated Dose Approach). Section 5.6.4 lists the types of quality control samples that should
be collected and analyzed during testing. Appendix C provides guidelines for identifying test
conditions for collimated beam testing.

       Experimental test conditions should be documented in a Validation Test Plan.
Section 5.11.2 provides recommendations on what a Validation Test Plan should contain. EPA
recommends including the Test Plan into the final Validation Report (see Section 5.11.3).
5.6.1  Test Conditions for the UV Intensity Setpoint Approach

       For the UV Intensity Setpoint Approach, the purpose of validation testing is to determine
the validated dose corresponding to the UV intensity setpoint for a reactor at a particular flow
rate. Typically, the manufacturer determines the UV intensity setpoint for their reactor. If this is
the case, water systems should work with the manufacturer to ensure that the setpoint is defined
conservatively low enough to account for combined conditions of minimum UVT and maximum
fouling/aging (commonly represented by the fouling/aging factor). If the manufacturer does not
establish the UV intensity setpoint for their reactor, the water system can select a setpoint using
the following procedure:

       1.  Record the UV intensity measurement at conditions of maximum UVT and 100
          percent power (So)w.

       2.  Reduce the lamp power until the measured UV intensity results in the following
          relative sensor intensity (S/S0 per Equation 5.3):

          a.  If aged lamps are used during validation testing, the relative sensor intensity
              should be equal to the fouling factor.

          b.  If new lamps are used during validation testing, the relative sensor intensity
              should be equal to the fouling/aging factor, which is the fouling factor multiplied
              by the aging factor.

       3.  Reduce the UVT of the water to the minimum UVT (see Section 3.4 for guidance on
          determining the  minimum UVT).
10 The impacts of lamp power and UVT on UV sensor readings are not dependent on the specific rate of flow
  traveling through the reactor. Thus, any flow rate can be used for this procedure.
UV Disinfection Guidance Manual                5-22                               November 2006
For the Final LT2ESWTR

-------
                                  5. Validation of UV Reactors
       4.  Record the UV Intensity at reduced power and reduced UVT conditions. This
           intensity is the UV intensity setpoint.

       The UV Intensity Setpoint approach uses two validation test conditions, as specified in
Table 5.6. The first involves reducing UVT until UV intensity measured by the UV sensor is
equal to the UV intensity setpoint. The second involves testing at high UVT but reducing power
until the  UV intensity measured by the sensor is equal to the UV intensity setpoint. Additional
test conditions should be evaluated if the water system will be using variable setpoint operations
(i.e., each test condition in Table 5.6 should be repeated at different flow rates).


   Table 5.6. Minimum Test Conditions for the UV Intensity Setpoint Approach1
Test ID'
1
2
Flow Rate
Design (highest)
Design (highest)
UVT
Lowered to give the
UV intensity setpoint 3
Maximum
Lamp Power
Maximum (100 %)
Lowered to give the UV intensity
setpoint 3
      Minimum test conditions shown are for single-setpoint operations. Additional tests should be conducted
      at different flow rates for variable setpoint operations.
    2 At least three replicate tests with the same stock solution of challenge microorganisms should be
      performed for each test condition.
      The UV intensity setpoint is typically established by the manufacturer. Alternatively, it can be
      established by the water system using the procedure in Section 5.6.1.
       Water systems may decide to use two challenge microorganisms with different UV
sensitivities for validation testing (see Section 5.3 for additional discussion). In many cases,
challenge microorganisms can be tested at the same time if they have been proven not to
interfere with each other.

       The validation approach described herein produces a UV intensity setpoint and Validated
Dose that are independent of UVT. Thus, UVT is not typically monitored during reactor
operations.
5.6.2  Test Conditions for the Calculated Dose Approach

       For the Calculated Dose Approach, the purpose of validation testing is to develop a dose-
monitoring equation relating RED11 to operating parameters such as flow rate, UVT, lamp power
(quantified as relative sensor value), and in some cases lamp status.  For each operating
parameter used in the equation, at least three conditions should be evaluated during validation
testing. Three data points are needed for interpolation of results because the relationship between
RED and operating parameters such as flow rate and UVT is typically non-linear.
11 As a reminder, RED is the reduction equivalent dose, which is determined by inputting the measured log
  inactivation (observed during full-scale reactor testing) into the UV dose-response curve (generated through
  collimated beam testing).
UV Disinfection Guidance Manual                5-23                               November 2006
For the Final LT2ESWTR

-------
                                 5. Validation of UV Reactors
       In many cases, three operating parameters (UVT, flow rate, and lamp power) are used in
the dose-monitoring equation, resulting in a minimum of 27 test conditions (3 x 3 x 3). Fewer
test conditions are needed when the dose monitoring equation is based on fewer than three
parameters, such as when a minimum UVT is assumed for all operating conditions. More than 27
test conditions may be needed when the water system plans to vary lamp status during operations
(e.g., UVT, flow rate, and lamp power are used in the dose monitoring equation and individual
banks of lamps will be turned off and on to conserve power).

       If validation tests are being conducted for a specific water system, the system's operating
range of flow rates, UVT, and the required UV dose for their target pathogen and log inactivation
help establish test conditions. For flow rate, the water system's maximum and minimum flow
rates,  as well as one or more intermediate flow rates, should be selected as test conditions. To
select intermediate flow rates, EPA recommends using a geometric progression (because the
relationship between UV dose and flow is non-linear) using the following equation:

       Qn=QMacP1-"                                                       Equation  5.6

       where:
         Qn   =  nih flow rate to be tested
        QMax   =  Maximum flow rate to be tested
          P   =  Constant with a recommended value between 1.5 and 2.0 to achieve good
                 separation of flow measurements
          n   =  Flow rate test # to be evaluated (must be > 3, if interpolating results)

       The value of |3 should be sufficient to obtain at least three measured data points for
developing the dose-monitoring equation. The value  of n should be selected to span the range of
flow rates. An example using Equation 5.6 is provided  below.
  Example 5.1. Determining Flow Conditions for Validation Testing. A UV reactor using the
  Calculated Dose Approach and operating within the range of 5 - 20 mgd is to be validated. The
  test engineer selects a B value of 1.6, resulting in the following test flow rates:

       n  Q (mgd)
       1    20
       2    12.5
       3    7.8
       4    4.9
       For UVT, test conditions should include the water system's minimum UVT, maximum
UVT, and at least one intermediate value. If the dose-monitoring equation will account for
specific lamps operating either on or off or other power manipulations, validation test design
should include these conditions.
UV Disinfection Guidance Manual                5-24                              November 2006
For the Final LT2ESWTR

-------
                                   5. Validation of UV Reactors
       Table 5.7 summarizes the recommended minimum test conditions for the Calculated
Dose Approach. Table B.9 in Appendix B presents an example test matrix for a hypothetical
water system.
                       Table 5.7 Minimum Test Conditions for the
                               Calculated Dose Approach1
Test IDZ
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
UVT Flow ratej
Maximum
Maximum
Maximum
Maximum
Maximum
Maximum
Maximum
Maximum
Maximum
Intermediate
Intermediate
Intermediate
Intermediate
Intermediate
Intermediate
Intermediate
Intermediate
Intermediate
Minimum
Minimum
Minimum
Minimum
Minimum
Minimum
Minimum
Minimum
Minimum
Flow RateJ
Design
Intermediate
Minimum
Design
Intermediate
Minimum
Design
Intermediate
Minimum
Design
Intermediate
Minimum
Design
Intermediate
Minimum
Design
Intermediate
Minimum
Design
Intermediate
Minimum
Design
Intermediate
Minimum
Design
Intermediate
Minimum
Lamp
Power 4
Maximum
Maximum
Maximum
Minimum expected to
Minimum expected to
Minimum expected to
occur during operations
occur during operations
occur during operations
Intermediate
Intermediate
Intermediate
Maximum
Maximum
Maximum
Minimum expected to
Minimum expected to
Minimum expected to
occur during operations
occur during operations
occur during operations
Intermediate
Intermediate
Intermediate
Maximum
Maximum
Maximum
Minimum expected to
Minimum expected to
Minimum expected to
occur during operations
occur during operations
occur during operations
Intermediate
Intermediate
Intermediate
  Assuming validation on a non-validated UV reactor. Minimum test conditions shown are for all lamps turned on.
  Additional tests should be performed to evaluate other lamp on/off combinations or other power combinations.
  At least three replicate tests with the same stock solution of challenge microorganisms should be performed for each
  test condition.
  See Section 3.4 for guidelines on identifying design flow and minimum and maximum UVT.
  Minimum power should include reduction in lamp output caused by fouling and aging.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
5-25
November 2006

-------
                                 5. Validation of UV Reactors
5.6.3  Test Conditions for Confirming an Existing Validation Equation

       Water systems may decide to perform on-site validation testing to show that the hydraulic
conditions at the water treatment plant result in a UV dose that is the same or greater than the
UV dose delivered at the off-site validation test facility. Test conditions should generally span
the range of operating conditions expected at the treatment plant (e.g., minimum and maximum
UVT, minimum and maximum flow rate). See Section 3.6.2 for additional discussion on
validation testing scenarios.

       EPA cautions water systems on combining on-site validation testing data with off-site
validation data to develop  a new dose-monitoring equation. On-site and off-site testing is often
done under different hydraulic conditions and may produce different results. Combining the
datasets may result in greater noise about the fit for the dose-monitoring equation and, thus, a
higher uncertainty factor (see Section 5.9.2.2 for a discussion of the  uncertainty in interpolation
factor)
5.6.4  Quality-control Samples

       Recommended quality-control samples for full-scale reactor testing are listed below.

       .   Reactor controls - influent and effluent water samples taken with the UV lamps (in
          the reactor) turned off. The change in log concentration from influent to effluent
          should correspond to a change in RED (from the UV dose-response curve) that is
          within the measurement error of the minimum RED measured during validation
          (typically 3 percent or less).

       .   Reactor blanks - influent and effluent water samples taken with no addition of
          challenge microorganism to the flow passing through the reactor. Blanks should be
          collected at least once on each day of testing and the concentration of challenge
          microorganisms should be negligible.

       .   Trip controls - one sample bottle of challenge microorganism stock solution should
          travel with the stock solution used for validation testing from the microbiological
          laboratory to the location of reactor testing and back to the laboratory. The change in
          the log concentration of the challenge microorganism in the trip control should be
          within the measurement error,  (i.e., the change in concentration over the test run
          should be negligible.  This is typically on the order of 3 to 5 percent.
UV Disinfection Guidance Manual                5-26                               November 2006
For the Final LT2ESWTR

-------
                                 5. Validation of UV Reactors
          Method blanks - sample bottle of sterilized reagent grade water that undergoes the
          challenge microorganism assay procedure. The concentration of challenge
          microorganism with the method blank should be non-detectable, according to
          Standard Methods for the Examination of Water and Wastewater (APHA et al. 1998).

          Stability samples - influent and effluent samples collected at low and high UVT that
          are used to assess the stability of the challenge microorganism concentration and its
          UV dose-response over the time period from sample collection to completion of
          challenge microorganism assay. The challenge microorganism concentrations in the
          stability samples should be within 5 percent of each other.
5.7    Guidelines for Conducting Experimental Tests

       Section 5.7.1 provides general guidelines for preparing the challenge microorganism for
testing. Sections 5.7.2 and 5.7.3 provide recommendations for conducting full-scale reactor
testing and collimated beam testing, respectively. Appendix C contains more detail on the
collimated beam testing methods. Importantly, the recommendations in this section and in
Appendix C are not step-by-step procedures, but rather an identification of key steps in the
process. Individuals performing full-scale reactor testing and collimated beam testing should
work closely with the laboratory personnel and experts in the field of validation testing to ensure
that appropriate procedures and QA/QC steps are followed.
5.7.1  Preparing the Challenge Microorganism

       The challenge microorganism used to validate UV reactors should be cultured and
analyzed by a laboratory staffed by professional microbiologists and equipped to perform
microbiological examinations as specified in Standard Methods for the Examination of Water
and Wastewater (APHA et al. 1998). Protocols for culturing the challenge microorganism and
measuring its concentration should be defined and based on published and peer-reviewed
methods.

       The challenge microorganism concentrations should be stable over the holding time
between sampling and completion of the assays. If they are not stable, the data collected will be
unusable because distinguishing the sources of inactivation—exposure to UV light and die-off in
holding—will be impossible. Instability problems with MS2 phage are well documented  in the
literature (Petri et al. 2000, Swaim et al. 2003, Hargy et al. 2004). Factors that can impact MS2
phage stability in water include the presence of chlorine, coagulants, ionic strength, surfactants,
and UV absorbers (Thompson and Yates 1999, Petri et al. 2000, Hargy et al. 2004). Laboratory
methods can also impact the stability of MS2 phage in water (Thompson and Yates 1999).
Microbial stability in the test water should be verified before  experimental testing begins.
Stability verification can help ensure that the bioassay and challenge microorganism samples will
be viable and the data useable.
UV Disinfection Guidance Manual                5-27                              November 2006
For the Final LT2ESWTR

-------
                                  5. Validation of UV Reactors
       Appendix A provides recommended procedures for preparing stock solutions of MS2
phage and B. subtilis spores and assaying their concentrations in water samples. Alternative
procedures and challenge microorganisms can be used if they are acceptable to the state.
5.7.2  Full-scale UV Reactor Testing

       Three key steps comprise full-scale reactor testing: (1) verifying reactor properties, (2)
installing the reactor, and (3) conducting the tests. These steps are summarized below. Note that
key steps are based on UV reactor testing at an off-site validation test facility. Additional steps
may be necessary for on-site validation.

Verifying UV Reactor Properties

       For validation, the UV manufacturer should provide the following:

       .   A UV reactor that matches the provided specifications.

       .   Duty and reference UV sensors that match the provided specifications.

       .   UV lamps that have undergone appropriate burn-in. If new lamps are to be used, the
          recommended burn-in period is 100 hours. If aged lamps are to be used, the
          recommended burn-in period is that which will produce lamp output equivalent to the
          fouling/aging factor. More information on aged lamps is provided in Section 5.4.6.

       .   For UV reactors with more than one lamp per UV sensor, lamps with the highest
          output positioned closest to the sensor. (See Section 5.4.7 for additional guidance on
          sensor positioning to address lamp variability.)

       .   Provisions to reduce lamp output.

       .   Provisions to measure the electrical power delivered to the lamps.

       .   A temperature sensor and safety cut-off switch to prevent overheating if MP lamps
          are used.

Installing the  VV Reactor

       The UV reactor and the reactor inlet and outlet connections should be installed at the test
facility in accordance with the manufacturer's installation and assembly instructions. If reactors
are installed in series, the piping between the reactors should conform to the specifications
provided by the UV reactor manufacturer. The piping should be inspected to ensure compliance
with the manufacturer's specifications. The configuration of inlet and outlet piping to and from
the reactor and its impact on validation testing is discussed in Sections 3.6 and 5.4.5. Good
mixing should be confirmed.
UV Disinfection Guidance Manual                5-28                               November 2006
For the Final LT2ESWTR

-------
                                 5. Validation of UV Reactors
       The physical integrity of the UV reactor and the test train should be verified before
testing. Personnel who operate the UV reactor during all tests should be familiar with its
operation and maintenance manual and with any safety requirements.

Measuring VVDose Delivery

       During full-scale reactor testing, the reactor is operated at each of the test conditions for
flow rate, UVT, and lamp power (in accordance with the Validation Test Plan) as described in
Section 5.6. The following steps should be taken to ensure good results:

       .   Confirm steady-state conditions before injecting the challenge microorganism by
          monitoring the UV sensor measurements and the UVT.

       .   Inject the challenge microorganism, prepared according to Appendix A, into the flow
          upstream of the reactor.

          Collect at least three (3) influent and three (3) effluent samples for each test
          condition. Sample volumes should be sufficient for assessing the challenge
          microorganism concentrations in the influent and effluent (typically  10-15 mL).

       .   Measure and record the flow rate through the reactor, all UV sensor measurements,
          on-line UVT measurements, and any calculated UV dose values both before and after
          the samples are collected.

       .   Measure and record the UVT  as measured by the UV spectrophotometer with each
          influent sample.

       .   Measure and record the electrical power consumed by the lamp ballasts.

       .   Repeat the test if the flow rate, UV intensity, lamp power, or UVT changes by more
          than the recommended error of the measurement over the course of sampling (see
          Section 5.5).

       Sample taps should remain open over the duration of the test. Sample collection should
meet standards of good practice as defined by Standard Methods Section 9060 (APHA et al.
1998). Samples should be collected in bottles that have been cleaned and sterilized and should be
immediately stored on ice, within a cooler, in the dark until analyzed.

       The concentrations of the challenge microorganisms before and after exposure to UV
light should generally be measured within 24 hours of sample collection, unless stability studies
indicate that the samples can reliably be considered stable over longer periods of time. Samples
that are not assayed immediately should be stored in the dark at 4 °C. Exposure  of samples to
visible light should be avoided.
UV Disinfection Guidance Manual                 5-29                              November 2006
For the Final LT2ESWTR

-------
                                 5. Validation of UV Reactors
5.7.3  Collimated Beam Testing

       Collimated beam tests are performed in microbiological laboratories under controlled
conditions. Recommended test procedures are provided in Section C.2.3. Importantly, all
collimated beam testing should be conducted using a water sample collected from the influent
sampling port of the biodosimetry test stand. If the full-scale reactor testing lasts for more than
one day, at least one collimated beam test should be conducted for each day of testing. A
minimum of two collimated beam tests is always recommended, one each at the highest and
lowest UVT values evaluated during full-scale reactor resting.
5.8    Analyzing Experimental Data

       Validation testing of UV reactors produces the following types of data for each
experimental test:

       .   Concentration of the challenge microorganism in the influent and effluent sample
          [e.g., plaque forming units per milliliter (pfu/mL) for MS2 phage, colony forming
          units per milliliter (cfu/mL) for B. subtilis spores]

       .   UVT of water (percent)
       .   Flow rate [gallon per minute (gpm) or mgd]
       .   UV intensity as measured by the UV sensor (mW/cm2)
       .   Lamp power [watt (W) or kilowatt (kW)]
          Status (on/off) for each lamp


All experimental data should be documented, preferably in tabular format, and included in the
Validation Report. (See Section 5.11.3 for additional guidance on the Validation Report and
Appendix B for examples.)

       Section 5.8.1 shows how RED is calculated for each experimental test using a
combination of full-scale reactor testing data and collimated beam results. Additional analyses of
RED data depend on the reactor's UV dose-monitoring strategy. For the UV Intensity Setpoint
Approach, RED results are averaged for each test condition and evaluated to identify the
minimum value. For the Calculated Dose Approach, all RED values and associated test
conditions are used to create a dose-monitoring equation. Sections 5.8.2 and 5.8.3 summarize
recommended next steps for evaluating RED data for the UV Intensity Setpoint Approach and
Calculated Dose Approach, respectively.
5.8.1  Calculating the Reduction Equivalent Dose (RED)

       The RED should be calculated for all full-scale reactor test conditions, individually for
each replicate, using the following method:
UV Disinfection Guidance Manual                5-30                              November 2006
For the Final LT2ESWTR

-------
                                 5. Validation of UV Reactors
       1.  For each test condition replicate (i.e., influent and effluent sample pairs), calculate the
          log inactivation (log I) using Equation 5.7:
           '<*' = '<*
                                  Equation 5.7
          where:
             N0  =   Challenge microorganism concentration in influent sample (pfu/mL or
                     cfu/mL)
              N  =   Challenge microorganism concentration in corresponding effluent sample
                     (pfu/mL or cfu/mL)

       2.  Determine the RED, in ml/cm2 for each test condition replicate pair using the
          measured log inactivation (log I) and the UV dose-response curve developed through
          collimated beam testing (see Appendix C). If individual UV dose-response curves
          cannot be combined, the curve for a given day of testing should be used to determine
          RED for full-scale reactor testing data collected that day. If individual dose-response
          curves developed on the same day of testing cannot be combined, the curve resulting
          in the most conservative (lowest) RED values should be used.

       Note that for the UV Intensity Setpoint Approach, replicates for a given test condition are
averaged. For the Calculated Dose Approach, replicates are evaluated separately to develop the
UV dose-monitoring equation.

       Appendix B shows RED calculations for two hypothetical water systems. Example 5.2
shows the key inputs and results for the hypothetical water system in Section B.I.
Exam]
beamt
dose-n
UVI
Full-sc
and rej
TheRI
jle 5.2. Calculating RED Using Validation Test Data Collimated
esting using a challenge microorganism produces the following UV
;sponse curve:
)ose (ml/cm2) = 2. 18(log I)2 + 15.30(log I) (from Figure B.2)
ale reactor testing produces the following data for each test condition
)licate test:
Test
Condition
1
1
1
2
2
2
Replicate
1
2
3
1
2
3
No
(pfu/mL)
5.94
6.00
5.84
6.01
5.99
6.04
N (pfu/mL)
4.57
4.54
4.56
4.10
4.09
4.06
Log I
1.37
1.46
1.28
1.91
1.9
1.98
RED
(mJ/cm2)
25.1
27.0
23.2
37.2
36.9
38.8
3D values for each test are shown in the last column.

UV Disinfection Guidance Manual
For the Final LT2ESWTR
5-31
November 2006

-------
                                  5. Validation of UV Reactors
       If the UV reactor uses MP lamps and validation testing is performed using a challenge
microorganism other than MS2 phage or B. subtilis, an action spectra correction factor (CFas)
may need to be applied to the RED values to account for differences in the action spectra of the
target pathogen and challenge microorganism. Section D.4.1  in Appendix D describes this
concept and presents the correction factors that should be used for the RED adjustment (i.e.,
divide RED by the correction factor).

       If validation testing is done with two  challenge microorganisms whose UV sensitivities
bracket the UV sensitivity of the target pathogen (i.e., one microorganism is more resistant and
one is less resistant), the following approach can be used to estimate the RED of the target
pathogen for each test condition:

       1. For each test condition, calculate  the UV sensitivity (mJ/cm2 per log I) of the
          challenge microorganism using the following equation:

              UV sensitivity = RED /Log I                                    Equation 5.8

              where:
                RED =  The RED for the test replicate as derived by inputting Log I into the
                        UV dose-response equation
                Log 1=  log inactivation for the test replicate as calculated using Equation 5.7

       2. Create a graph with UV sensitivity on the x-axis and RED (mJ/cm2) on the j/-axis for
          each test condition.

       3. For each challenge microorganism, plot paired UV sensitivity and RED values on the
          graph (2 values).

       4. Draw a straight line between the two points.

       5. Determine the UV sensitivity for  the target pathogen by selecting the UV dose from
          Table 1.4 for 1 log inactivation (log 1=1)

       6. Using the straight line in the graph created in Step 4, read the corresponding RED
          value for the UV  sensitivity of the target pathogen (as determined in Step 5).

       Example 5.3 shows this procedure using hypothetical validation test data. As noted in
Section 5.3, the main advantage of testing two challenge microorganisms whose UV sensitivities
bracket the sensitivity of the target pathogen  is that the factor used to account for challenge
microorganism bias (the RED Bias factor) can be set to 1.0. (See Section 5.9 for discussion of
the RED bias factor.)
UV Disinfection Guidance Manual                5-32                               November 2006
For the Final LT2ESWTR

-------
                                  5. Validation of UV Reactors
  Example 5.3. Validation Testing Using Two Challenge Microorganisms
  Validation testing is performed using MS2 and 9x174. The table below summarizes average
  results for three replicates for one test condition (high UVT).

Challenge
Microorganism
MS2
cpx174
Influent
Cone.
(pfu/mL)
1x10S
1x104
Effluent
Cone.
(pfu/mL)
1x104
0
UV
Sensitivity
(mJ/cm2 per log I)1
20
2


Log I2
2.0
>4.0

RED
(mJ/cm2)3
40
>8.0
         1  As derived from collimated beam testing data for a log inactivation of 2.0 for MS2 and 4.0 for 9x174
           using Equation 5.9.
           Based on measured influent and effluent microorganism concentrations from validation testing.
           Determined by inputting log I  into the UV dose-response equation.

  Paired UV sensitivity and RED values for MS2 and 9x174 were plotted on the graph below.
              45
              40
              35
           ~~ 30
              15
              10
               5 -I
               0
                                      10
                                                 15
                                                            20
                                                                       25
                                UV Sensitivity (mJ/cm per log I)
  Straight-line interpolation between the two points yields the following equation:
         RED = 1.78 x UV Sensitivity + 4.44

  The equation above predicts that the RED delivered to Cryptosporidium, defined with a UV
  sensitivity of 3.9 mJ/cm2 per log inactivation (Table  1.4), is:

         RED = 1.78 x 3.9 + ¥.44 = 11.4 mJ I cm2

  Because the RED represents the dose delivered to Cryptosporidium, the RED Bias Factor is
  equal to 1.0.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
5-33
November 2006

-------
                                5. Validation of UV Reactors
5.8.2  Selecting the Minimum RED for the UV Intensity Setpoint Approach

       Replicate RED values (typically 3-5 values) should be averaged to produce one RED
for each test condition. From these average values, the minimum RED should be selected and
used in the validated dose calculation. If variable-setpoint operations will be used at the water
treatment plant (i.e., different UV intensity setpoints for different flow rate ranges), the minimum
RED value should be identified for each flow rate range.

       Table 5.6 in Section 5.6.1 presents the two test conditions that should be evaluated, at a
minimum, for the UV Intensity Setpoint Approach. If the UV sensor is in the ideal location (i.e.,
a location that gives UV dose delivery proportional to the UV sensor reading), the two test
conditions should yield the same RED. If the sensor is located farther from the lamp than the
ideal location, the minimum RED will be produced under minimum UVT/maximum power
conditions (Test 1). If the sensor is located closer to the lamp than the ideal position, the
minimum RED will be produced under maximum UVT /minimum power conditions (Test 2).
Selecting the minimum RED from these two test conditions accounts for UV reactor designs
where the sensor is not in the ideal location. See Section D.2 in Appendix D for additional
discussion on UV sensor positioning.
5.8.3  Developing the Dose-monitoring Equation for the Calculated Dose
       Approach

       If the reactor uses the Calculated Dose Approach, validation testing results are used to
develop a dose-monitoring equation for RED. The variables in the dose-monitoring equation are
typically flow rate, UVT, UV intensity, or some subset thereof. The number of operating banks
of lamps is also a possible variable for the equation for those water systems that use multiple
banks.

       EPA recommends using multivariate linear regression to fit an equation to the validation
test data. Procedures for multivariate linear regression can be found in standard statistical
textbooks such as Draper and Smith (1998). Software packages, such as Microsoft Excel, can
also be used to perform the regression analysis and determine the goodness-of-fit. Recommended
steps for the analysis are summarized below.

       1.  Fit an equation for RED as a function of the operating parameters of interest
          (using all the replicate inlet-and-outlet pairs) using multivariate linear
          regression. The equation used for interpolating validation data may have various
          forms depending on how it was derived. An empirical equation that can often provide
          a good fit to validation data has the following form (Wright et al. 2005):
               = 10flx454&x| SA I  x( Yn]  xB*                         Equation 5.9


          or in linear form,
UV Disinfection Guidance Manual                5-34                             November 2006
For the Final LT2ESWTR

-------
                                  5. Validation of UV Reactors
                                                      dxlogl  yn\ + exB    Equation 5.10
          where:
                RED   =   The RED calculated with the dose-monitoring equation, also
                            referred to as the "calculated dose" in this guidance manual
                 A254   =   UV absorbance at 254 nm
                    S   =   Measured UV sensor value
                   S0   =   UV intensity at 100 percent lamp power, typically expressed as a
                            function of UVT.
                   Q   =   Flow rate
                   B   =   Number of operating banks of lamps within the UV reactor
            a, b, c, d, e  =   Model coefficients obtained by fitting the equation to the data

          Either the full equation or part of the equation can be used for fitting validation data.
          For example, validation data collected at a constant UVT and lamp power setting can
          be fitted using:
           RED = ax  yo                                                    Equation 5.11
                     \/ x^ j

           or in linear form,


           \og(RED) = log(a) + dlogf V\                                     Equation 5.12
          The exact form of the relationship will depend on the UV reactor and the functional
          relationships between RED and each variable.

          The equation should pass through the origin (0,0) if the RED is calculated as a
          function of measured UV intensity or inverse flow rate. A zero measured dose should
          correspond to a zero calculated dose. A non-zero intercept would introduce a bias.

       2.  Determine the goodness-of-fit. This can be done using procedures found in standard
          statistics books or by reviewing variance tables produced by statistical programs. The
          analysis should determine the p-statistics for the model coefficients. For the fit to be
          acceptable, the p-statistic for each model coefficient should be < 0.05.

          If the p-statistic for a given model coefficient is greater than 0.05, the coefficient is
          not statistically significant. The coefficients are calculated with all the variables
          included. If the p-statistic for any coefficient exceeds 0.05, then, working in reverse,
          the model coefficient with the highest p-statistic should be dropped from the equation
          and the multivariate regression repeated until all p-statistics are less than or equal to
          0.05. Alternatively, the functional form of the equation could be revised to improve
          the relationship between RED and the parameters of interest (e.g., use Equation 5.12
UV Disinfection Guidance Manual                5-35                              November 2006
For the Final LT2ESWTR

-------
                                  5. Validation of UV Reactors
          instead of Equation 5.10).

       3.  Verify that there is no significant bias in the fit. One way to do this is to test for
          randomness in residual values (Draper and Smith 1998). The differences between the
          measured and calculated RED values should be randomly distributed around zero and
          not dependent on flow rate, UVT, or lamp status.

       Because both UVT and UV intensity are part of the dose-monitoring equation, it is not
important that the sensor be in the ideal location. If the UV sensor is in the ideal location,
however, UVT could be removed from the dose-monitoring equation. See Section D.2 for
additional discussion of UV sensor positioning.
5.9    Deriving the Validation Factor (VF)

       Several uncertainties and biases are involved in using experimental testing to define a
validated dose and validated operating conditions. For example, a challenge microorganism may
have a different UV sensitivity than the target pathogen. To determine the validated dose, the
RED (derived in Section 5.8) is divided by a VFto quantitatively account for key areas of
uncertainty. The equation for the VF is shown below.


       VF = BRED x  l + U™                                                  Equation 5.13
       where:
         VF  =  Validation Factor
       BRED  =  RED bias factor
        Uvai  =  Uncertainty of validation expressed as a percentage

       In addition to the RED bias factor, a bias factor to account for the influence of non-
germicidal light on UV sensor readings (referred to as the "polychromatic bias factor") should be
included in Equation 5.13 for MP reactors that meet either of the following criteria:
                                                               12
       •  The MP reactor is equipped with a non-germicidal sensor

       •  The MP reactor is equipped with a germicidal sensor, but the sensor is mounted
          further than 10 cm from the lamp and the water to be treated has a low UVT (< 80%)

Derivation of the polychromatic bias factor and its inclusion in the VF calculation are addressed
in Appendix D, Section DAS.

       The next two sections provide recommendations for calculating the RED bias factor and
uncertainty in validation and determining when each should be applied. Appendix D discusses in
greater detail the basis for the uncertainty and bias terms and how they were derived.
12 EPA recommends that MP reactors be equipped with germicidal sensors to more accurately measure UV light in
  the germicidal range. EPA recognizes, however that reactors with non-germicidal sensors have been installed or
  are about to be installed at water treatment plants prior to the publication of this document.
UV Disinfection Guidance Manual                5-36                               November 2006
For the Final LT2ESWTR

-------
                                 5. Validation of UV Reactors
Appendix B uses two example case studies to illustrate the calculation of the VF using methods
described in this section.

       Some areas of experimental uncertainty are not included in the VF equation. Instead,
EPA recommends that UV reactor monitoring components meet the performance criteria
presented in Chapter 6 and validation test results meet the QA/QC criteria as presented
throughout this chapter and summarized in Section 5.12. Section 5.9.2 includes a method for
checking key areas of experimental uncertainty and determining when factors should be included
in the Uvai calculation.
5.9.1  RED Bias Factor

       The RED bias is a correction factor that accounts for the difference between the UV
sensitivity of the target pathogen and the UV sensitivity of the challenge microorganism. If
validation testing is performed using two challenge microorganisms whose UV sensitivities
bracket those of the target pathogen (i.e., one challenge microorganism is less resistant than the
target pathogen and the other is more resistant than the target pathogen), the RED bias is equal to
1.0 (i.e., it can be corrected for, see Section 5.8.1 for details).

       If the UV sensitivities  of the challenge  microorganism and target pathogen are not the
same, the RED delivered under the same reactor operating conditions will differ. The magnitude
of this difference depends on the following factors:

       .   The dose distribution of the UV reactor
       .   The difference between the inactivation kinetics of the challenge microorganism and
          the target pathogen

       If the challenge microorganism is more resistant to UV light than the target pathogen, the
RED measured during validation will be greater than the RED that would be measured for the
target pathogen. In this case, the RED bias would be greater than 1.0. If the challenge
microorganism is less resistant (more sensitive) to UV light than the target pathogen, the RED
measured during validation will be less than the RED that would be measured for the target
pathogen. In this case, the RED bias should be assigned a value  of 1.0.

       The recommended procedure for determining the RED bias is as follows:

       1.  For the test condition with the lowest UVT, determine the observed  UV sensitivity of
          the challenge microorganism for each test replicate using Equation 5.8.

       2.  Identify the maximum UV sensitivity for all test replicates.

       3.  Use Tables G. 1 - G. 17 (in Appendix G) to find the RED bias for the target pathogen
          and target log inactivation, the maximum UV sensitivity, and the lowest UVT. Note
          that Tables G. 1  - G. 17 are for discreet UVT values of 85 percent, 90 percent, and 95
          percent. RED bias  can be interpolated for intermediate values of UVT.
UV Disinfection Guidance Manual                 5-37                              November 2006
For the Final LT2ESWTR

-------
                                 5. Validation of UV Reactors
       EPA recommends calculating one RED Bias for the UV facility, based on the site-
specific application (i.e., minimum operating UVT and target pathogen log inactivation desired),
which results in a constant VF for all conditions. As an alternative, the RED bias can be defined
as a function of UVT. This alternative may be  advantageous for the Calculated Dose Approach
where UVT is continually  monitored during operations, which means that the VF and the
validated dose would vary along with UVT. The disadvantage of using a variable VF is that the
UV reactor control system would need to be designed and programmed to do these calculations
and that the VF reported to the state will vary (see Section 6.5 for reporting guidance), making
operations and reporting more complex.

       Values in Tables G.I - G.17  are based  on theoretical dose distributions (as determined by
CFD modeling) for several UV reactor designs. Appendix D, Section D.5  provides additional
information on the derivation of values in Tables G.I - G.17. Example 5.4 shows how the RED
bias is determined for hypothetical test conditions.
       Example 5.4. Determining the RED Bias factor. A UV reactor is validated
       using MS2 phage for 3-log Cryptosporidium inactivation credit. The maximum
       MS2 phage UV sensitivity for the validation test condition of lowest UVT (86
       percent) is 18.0 ml/cm2 per log inactivation. The RED bias from Table G.3 is
       1.92
5.9.2  Uncertainty in Validation (Uvai)

       The Uncertainty in Validation (Uyai), also referred to as the experimental uncertainty, has
between 1 and 3 input variables based on how well the validation testing adhered to
recommended QA/QC limits in this guidance manual. At least one input variable, which depends
on the dose-monitoring strategy of the UV reactor, should be used in all cases.

       Figures 5.4 and 5.5 provide decision trees for selecting the appropriate equation for
calculating Uyai and provide a description of the input variables used for the calculation. The
next two sections provide guidance for deriving two of the input variables for Uvai, which are
USP (the uncertainty in the setpoint value, which is always calculated for the UV Intensity
Setpoint Approach) andUiN (the uncertainty in interpolation, which is always calculated for the
Calculated Dose Approach). Us is the uncertainty in UV sensor measurements, expressed as a
fraction (e.g., 15 percent, or 0.15) as described in Section 5.5.4. UDR is the uncertainty of the
dose-response fit at a 95-percent confidence level. Note that if individual UV dose-response
curves cannot be combined and there is more than one UDR value, the maximum value should be
used in the decision tree.  Additional guidelines for estimating UDR are provided in Section C.4.
UV Disinfection Guidance Manual                 5-38                              November 2006
For the Final LT2ESWTR

-------
                                5. Validation of UV Reactors
       Figure 5.4. UVAL Decision Tree for the UV Intensity Setpoint Approach
       Yes
No
                                 Yes
                          No
                     II   = (\ I   2+ I I 2M/2
                     UVal  ^USP    US I
 U   = ((]  2+ U 2+ U   2\1/2
 UVal   ^USP    US    UDR I
                                                            Yes
                                                     No
                                                  UVa, =
                              II   = (\ I  2 + I I   2M/2
                              UVal   ^USP   UDR J
  Where:
  UVa| = Uncertainty of validation (representing experimental
         uncertainty)
  Us  = Uncertainty of sensor value, expressed as a fraction (i.e., if
         sensor uncertainty = 12%, Us = 0.12.  See Section 5.5.4 for
         guidance on determining sensor uncertainty.)
  UDR = Uncertainty of the fit of the dose-response curve, calculated
         using Equation C.6. If there is more than one UDR value, use the
         maximum value
  USP = Uncertainty of setpoint, calculated using Equation 5.14 (See
         Section 5.9.2.1 for additional guidance)

  *lf UDR is calculated using standard statistical methods instead of using equation C.6, is it > 15%?
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                    5-39
November 2006

-------
                                5. Validation of UV Reactors
         Figure 5.5. UVAL Decision Tree for the Calculated Dose Approach
                                 Yes
      No
1
Yes
r


No
i
UVa,= (U,

U = ([} 2+ II 2+ II 2\1/2
UVal ^UIN T US T UDR >
r
2+11 2M/2
N US >>


T
Yes
r
No
1 r
UVa,= U|N

II - (\ | 2 + || 2M/2
UVal ^UIN ^UDR >
    Where:
    UVa| = Uncertainty of validation (representing experimental
          uncertainty)
    Us  = Uncertainty of sensor value, expressed as a fraction (i.e., if
          sensor uncertainty = 12%, Us = 0.12. See Section 5.5.4 for
          guidance on determining sensor uncertainty.)
    UDR = Uncertainty of the fit of the dose-response curve, calculated
          using Equation C.6 If there is more than one UDR value, use the
           maximum  value
    U|N  = Uncertainty of interpolation, calculated using Equation 5.15
          (See Section 5.9.2.2 for additional guidance)
    *lf UDR is calculated using standard statistical methods instead of using equation C.6, is it > 15%?
UV Disinfection Guidance Manual
For the Final LT2ESWTR
5-40
November 2006

-------
                                 5. Validation of UV Reactors
5.9.2.1  Calculating USp for the UV Intensity Setpoint Approach

       The uncertainty in the setpoint value is based on a prediction interval at a 95-percent
confidence level using the following procedure:

       1.  Calculate the average and standard deviation of RED values for each test condition
          (typically at least 3-5 replicate pairs are generated for each test condition).

       2.  Calculate the uncertainty of the setpoint RED using:
           > *P -	/viww/u                                            Equation 5.14
                  RED

          where:
            RED  =  Average RED value measured for each test condition
                   =  Standard deviation of the RED values measured for each test condition
                t  =  t-statistic for a 95-percent confidence level defined as a function of the
                      number of replicate samples using the following:
Number of Samples
3
4
5
t
3.18
2.78
2.57
       3.  Select the highest USP from all test conditions for calculating the VF.


5.9.2.2 Calculating U|N for the Calculated Dose Approach

       For reactors using the Calculated Dose Approach, the uncertainty of interpolation (UIN) is
calculated as the lower bound of the 95-percent prediction interval for the dose-monitoring
equation. This prediction interval reflects the noise in the data about that fit. In non-statistical
terms, the UIN represents the difference between (1) the RED value as derived using measured
log inactivation and the UV dose-response curve, and (2) the RED value as calculated using the
dose-monitoring equation (also referred to as the "calculated dose" in this manual).

          is calculated using the following equation:


                                                                           Equation 5.15
         JJV    RED

       where:
         SD  =  Standard deviation of the differences between the test RED (based on the
                 observed log inactivation and UV dose-response curve), and the RED
                 calculated using the dose-monitoring equation for each replicate
       RED  =  The RED as calculated using the dose-monitoring equation
           t  =  t-statistic at a 95-percent confidence level for a sample size equal to the
                 number of test conditions used to define the interpolation:
UV Disinfection Guidance Manual                5-41                              November 2006
For the Final LT2ESWTR

-------
                                5. Validation of UV Reactors
Number of Data Points
Used to Develop the Dose-
Monitoring Equation
3
4
5
6
7
8
9
10
11
12
13
t
3.18
2.78
2.57
2.45
2.36
2.31
2.26
2.23
2.20
2.18
2.16

Number of Data Points
Used to Develop the
Dose-Monitoring
Equation
14
15
16
17
18
19-20
21
22-23
24-26
27-29
>30
t
2.14
2.13
2.12
2.11
2.10
2.09
2.08
2.07
2.06
2.05
2.04
       The value of UIN depends on the calculated RED (or calculated dose), increasing at low
calculated RED values. EPA recommends that one UIN be selected that represents the most
conservative (largest) uncertainty value calculated for the validated dose operating range (for the
lowest calculated RED). Alternatively, UIN can be expressed as a function of the calculated RED.
5.10   Determining the Validated Dose and Validated Operating Conditions

       As shown in Figure 5.1 in Section 5.2, the last step in the recommended validation
protocol is to adjust the RED results by the VF to determine the Validated Dose for the UV
reactor using the following equation:
       Validated Dose = RED / VF
                                                        Equation 5.16
       Where:
       RED
         VF  =
the Minimum RED for the UV Intensity Setpoint Approach; or the RED as
calculated using the dose-monitoring equation (also referred to as the
calculated dose) for the Calculated Dose Approach
the Validation Factor, as calculated using Equation 5.13
       Because the method and assumptions for this step depend on the dose-monitoring
strategy of the UV reactor, they are discussed separately below.
5.10.1 Determining the Validated Dose and Operating Conditions for the UV
       Intensity Setpoint Approach

       For the UV Intensity Setpoint Approach, Equation 5.16 produces one validated dose for a
given UV intensity setpoint corresponding to the minimum RED. When the UV reactor is
operating at a UV intensity level above the setpoint, the true UV dose delivered to
microorganisms passing through the reactor is always equal to or greater than the validated dose.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                         5-42
November 2006

-------
                                  5. Validation of UV Reactors
The inactivation credit for the target pathogen is determined by comparing the validated dose to
the required dose in Table 1.4.

       Validated operating conditions are as follows:

          The UV intensity measured by UV sensors must be greater than the UV intensity
          setpoint.

          The flow rate must be equal to or less than the flow rate tested.

       .  The lamp status for each lamp (i.e., on/off setting) must be equivalent to the settings
          used during validation testing.
5.10.2 Determining the Validated Dose and Operating Conditions for the
       Calculated Dose Approach

       For the Calculated Dose Approach, the validated dose varies based on operational
parameters. Typically, measured values of UVT, UV intensity, and flow rate are entered into the
dose-monitoring equation to calculate RED. RED is divided by the VF to produce the validated
dose (Equation 5.16). Although EPA recommends using one VF, an equation may be used for
the VF if the RED bias factor is expressed as a function of UVT or if UIN is  expressed as a
function of RED.

       As noted in Section 3.5.2, a key advantage of the Calculated Dose Approach is that water
systems can reduce power when UVT is high and/or the flow rate is low as long as the Validated
Dose is greater than or equal to the required dose for the target pathogen and log inactivation
level. As a reminder, the validated dose must be greater than or equal to the  required dose for the
target pathogen and target log inactivation level to receive treatment credit.

       Validated operating conditions for the Calculated Dose Approach are as follows:

           The operating UVT must be equal to or greater than the minimum UVT  evaluated
           during validation testing.13

           The operating flow rate must not exceed  the flow rate evaluated during validation
           testing (see footnote 13).
    13 If the operating UVT measures higher than the maximum UVT evaluated during validation testing, the
    maximum UVT evaluated during validation testing should be used as the default in the dose-monitoring
    equation.  Similarly, if the operating flow rate measures less than the minimum flow rate evaluated during
    validation testing, the minimum flow rate evaluated during validation testing should be used as the default in the
    dose-monitoring equation. See Section 6.1.4 for guidance on setting operational controls.

UV Disinfection Guidance Manual                5-43                                November 2006
For the Final LT2ESWTR

-------
                                 5. Validation of UV Reactors
5.11   Documentation

       Prior to validation testing, the water system should work with the manufacturers, third
party reviewers, and engineers assisting with or performing validation testing to prepare the
following:

       .   Documentation for the UV reactor

           Validation Test Plan

       Once validation testing and the associated data analyses are complete, the UV reactor
documentation and Validation Test Plan, along with results of validation testing, should be
incorporated into a Validation Report.

       The next several sections provide more detailed recommendations on validation testing
documentation. Water systems purchasing a pre-validated reactor will not be preparing
documentation; however, Sections 5.11.1 through 5.11.3 may be useful as they review validation
documentation from manufacturers and consulting engineers. State personnel may also find these
sections helpful when reviewing validation reports.
5.11.1 UV Reactor Documentation

       Before validation testing, the UV manufacturer should provide the testing party with
documentation identifying and describing the UV equipment. Documentation should include all
reactor and component information that impacts UV dose delivery and monitoring, as described
in Checklist 5.1.
UV Disinfection Guidance Manual                5-44                              November 2006
For the Final LT2ESWTR

-------
                                   5. Validation of UV Reactors
                   Checklist 5.1 UV Reactor Documentation (Page 1 of 2)

                 Does UV reactor documentation contain the following elements?

Yes No
General

D D        Technical description of the reactor's UV dose-monitoring strategy, including the use of
             sensors, signal processing, and calculations (if applicable).

D D        Dimensions and placement of all wetted components (e.g., lamps, sleeves, UV sensors,
             baffles, and cleaning mechanisms) within the UV reactor.

D D        A technical description of lamp placement within the sleeve.

D D        Specifications for the UV sensor port indicating all dimensions and tolerances that impact
             the positioning of the sensor relative to the lamps. If the UV sensor port contains a
             monitoring window separate from the sensor, specifications giving the window material,
             thickness, and UV transmittance should be provided.

Lamp specifications

D D        Technical description
D D        Lamp manufacturer and product number
D D        Electrical power rating
D D        Electrode-to-electrode length
D D        Spectral output of new and aged lamps (specified for 5 nm intervals or less over a
             wavelength range that includes the germicidal range of 250 - 280 nm and the response
             range of the UV sensors)
D D        Mercury content
D D        Envelope diameter

Lamp sleeve specifications

D D        Technical description including sleeve dimensions
D D        Material
D D        UV transmittance (at 254 nm for LP and LPHO lamps, and at 200 - 300 nm for MP lamps
             with germicidal sensors)

Specifications for the reference  and the duty UV sensors

D D        Manufacturer and product number
D D        Technical description including external dimensions
D D        Data and calculations showing how the total measurement uncertainty of the UV sensor is
             derived from the individual sensor properties.  (See Table D. 1 for an example of the
             calculation of UV sensor measurement uncertainty from the uncertainty that arises due to
             each  UV sensor property.)
UV Disinfection Guidance Manual                 5-45                                November 2006
For the Final LT2ESWTR

-------
                                   5. Validation of UV Reactors
                  Checklist 5.1 UV Reactor Documentation (Page 2 of 2)

                Does UV reactor documentation contain the following elements?

Yes No

Sensor measurement properties

D D       Working range
D D       Spectral and angular response
D D       Linearity
D D       Calibration factor
D D       Temperature stability
D D       Long-term stability

Installation and operation documentation:

D D    Flow rate, head loss, and pressure rating of the reactor
D D    Assembly and installation instructions
D D    Electrical requirements, including required line frequency, voltage, amperage, and power
D D    Operation and maintenance manuals that include cleaning procedures, required spare parts, and
         safety requirements. Safety requirements should include information on electrical lockouts, eye
         and skin protection from UV light, safe handling of lamps, and mercury cleanup
         recommendations in the event of lamp breakage.
5.11.2 Validation Test Plan

       A validation test plan should document the key components of UV reactor testing.
Recommended components of a validation test plan are provided in Checklist 5.2. This list is not
meant to be all-inclusive; engineers should document any factors they believe are important for
validation testing in their Validation Test Plan.
UV Disinfection Guidance Manual                5-46                               November 2006
For the Final LT2ESWTR

-------
                                    5. Validation of UV Reactors
            Checklist 5.2 Key Elements of the Validation Test Plan (Page 1 of 1)

                   Does the validation test plan contain the following elements?
Yes No
D D        Purpose of Validation Testing. General description of why the tests are being done and how
             the data will be used.

D D        Roles and Responsibilities. Key personnel overseeing and performing the full-scale reactor
             testing and collimated beam testing, including their qualifications. This section should
             include contact names and telephone numbers.

D D        Locations and Schedule. Location for conducting full-scale reactor testing and collimated
             beam testing. Planned schedule for conducting the tests and performing the data analyses.

D D        Challenge Microorganism Specifications. Specifications for the challenge microorganism
             to be used during validation that include the protocols required for growth and
             enumeration, the expected UV dose-response, and suitability for use in validation testing.

D D        Plan for state review (if applicable).

Design of the Biodosimetry Test Stand/On-site Testing Facilities

D D         Inlet/outlet piping design, including backflow prevention
D D         Mixing
D D         Sample ports
D D         Pumps
D D         Additives (Material Safety Data Sheets for UV-adsorbing chemical, quenching agent)

Collimated Beam Testing Apparatus

D D         Lamp type
D D         Collimating tube aperture
D D         Distance from light source to sample surface
D D         Radiometer make and model

Monitoring Equipment Specifications and Verification of Equipment Accuracy for the following:

D D         Flow meters
D D         UVT analyzers (if used)
D D         UV Spectrophotometers
D D         Power measurement
D D         UV sensors
D D         Radiometer make, model, and calibration certificates

Experimental Test Conditions including, but not limited to:

D D         Number of tests, UVT, flow rate, lamp power, and lamp status for each test condition
D D         Lamp fouling factor, use of new or aged lamps
D D         Influent concentration of challenge microorganisms for each test condition
D D         QA/QCPlan
UV Disinfection Guidance Manual                  5-47                                November 2006
For the Final LT2ESWTR

-------
                                 5. Validation of UV Reactors
5.11.3 Validation Report

       The validation report should provide detailed documentation of all validation testing
results. The report should also include all elements of the Validation Test Plan and a summary of
the field-verified UV reactor properties.

       EPA recommends that the report begin with an executive summary with key information
that can be used by states and water systems to assess inactivation credit for the target
pathogen(s). The executive summary should include, at a minimum,

        .   The validated dose or range of validated doses,
           The log credit achieved for the potential target pathogens by the UV reactor, and
        .   Validated operating conditions (i.e., flow rate, UVT if the Calculated Dose approach
           is used).

If the UV Intensity Setpoint approach is used, the executive summary should provide the UV
intensity setpoint (or setpoints) for the validated dose. If the reactor uses the Calculated Dose
Approach as its dose monitoring strategy, the dose-monitoring equation should be provided.

       In addition to the items listed above, the executive summary should include the
following:

          A brief description of the validated reactor,

       .  The assumed fouling/aging factors for the reactor and indication if new or aged lamps
          were used during validation testing,

       .  A summary of the validation test conditions, including but not limited to the flow
          rate, UVT, and lamp power for each test condition,

       .  Key validation test results used to derive the dose, including but not limited to the
          RED values for each test condition, the UV dose-monitoring equation from
          collimated beam testing, and the VF,

       .  A summary of QA/QC checks and results, including UV sensor and radiometer
          reference checks,

          A description of the validation facilities,

       .  The organizations conducting the validation test, and

       .  Names and credentials of the individuals/organizations providing third party
          oversight.

       Recommended contents for the detailed validation report are listed  in Checklist 5.3. Note
that these recommendations are not intended to be all-inclusive. Engineers should document any
test characteristics or outcomes they believe are important in the Validation Report.
UV Disinfection Guidance Manual                5-48                              November 2006
For the Final LT2ESWTR

-------
                                    5. Validation of UV Reactors
          Checklist 5.3 Key Elements of the Validation Report (Page 1 of 1)

                   Does your validation report contain the following elements?
Yes No
General
D D        Detailed reactor documentation (see Checklist 5.1), including drawings and serial numbers,
             and procedures used to verify reactor properties.
D D        Validation test plan (either a summary of key elements, or the test plan can be attached to
             the validation report along with documentation of any deviations to the original test plan)

Full-scale reactor testing results, with detailed results for each test condition evaluated. Data should
include, but are not limited to:

D D        Flow rate
D D        Measured UV intensity
D D        UVT
D D        Lamp power
D D        Lamp statuses
D D        Inlet and outlet concentrations of the challenge microorganism

Collimated beam testing results, including detailed results for each collimated beam test used to create
the UV dose-response equation:

D D        Volume and depth of microbial suspension
D D        UV Absorption of the microbial suspension
D D        Irradiance measurement before and after each irradiation
D D        Petri factor calculations and results
D D        Calculations for UV dose
D D        Derivation of the UV dose-response equation, including statistical methods and confidence
             intervals (i.e., calculation of UDR)

QA/QC Checks:

D D        Challenge microorganism QA/QC, including blanks, controls, and stability analyses
D D        Measurement uncertainty of the radiometer, date of most recent calibration, results of
             reference checks
D D        Measurement uncertainty of UV sensors and results of reference checks
D D        Measurement uncertainty of the flow meter, UV spectrophotometer, and any other
             measurement  equipment used during full-scale testing

Calculation of the validated dose, log inactivation  credit, and validated operating conditions:

D D        RED for each test condition
D D        Calculation of the VF
D D        Setpoints if the reactor uses the UV Intensity Setpoint Approach
D D        Dose-monitoring equation if the reactor uses the Calculated Dose Approach
D D        Log inactivation credit for target pathogens (e.g.,  Cryptosporidium, Giardia, and viruses)
D D        Validated operating conditions (e.g., flow rate, lamp status, UVT)
UV Disinfection Guidance Manual                 5-49                                 November 2006
For the Final LT2ESWTR

-------
                                 5. Validation of UV Reactors
5.12   Guidelines for Reviewing Validation Reports

       State engineers and water systems purchasing pre-validated reactors should review the
validation report to confirm the following:

       .   Validation testing meets the minimum regulatory requirements as summarized in
          Table 5.1.

       .   EPA's recommended validation protocol was followed and any deviations from the
          protocol are adequately justified.

          Validated doses achieved by the UV equipment meet or exceed the target pathogen
          log inactivation desired.

          QA/QC criteria were met during validation testing.

       Checklist 5.4  summarizes the QA/QC recommendations presented throughout this
chapter and in Appendix C. If a QA/QC plan was prepared prior to validation, reviewers should
request a copy of the  plan and make sure it is consistent with industry standards.

       Checklist 5.5  contains key elements that should be verified by state or water system
personnel when reviewing validation reports. States and systems should keep documentation that
these key validation criteria were met.
UV Disinfection Guidance Manual                5-50                              November 2006
For the Final LT2ESWTR

-------
                                   5. Validation of UV Reactors
      Checklist 5.4 Review for Quality Assurance/Quality Control (Page 1 of 1)

Yes No
Uncertainty in Measurement Equipment (See Section 5.5 and C. 2.2 for more information)

D D        Flow Meter: Is the measurement uncertainty < 5 percent?

D D        UV Spectrophotometer: Is the measurement uncertainty < 10 percent?

D D        UV Sensors: Did duty sensors operate within 10 percent of the average of two or more
             reference sensors? If not, was uncertainty in sensor measurement incorporated into the VF?

D D        Radiometer: (for collimated beam testing only). Do lamp output measurements vary by no
             more than 5 percent over exposure time? Was the accuracy of the radiometer verified with
             another radiometer?

QA/QC ofMicrobial Samples (See Section 5.6.4 for more information)

D D        Reactor controls: For influent/effluent samples taken with the UV reactor lamps turned
             off, does the change in log concentration correspond to a change in RED that is within the
             measurement error of the minimum RED measured during validation (typically < 3 %)?

D D        Reactor blanks: For DAILY influent/effluent samples taken with NO  challenge
             microorganisms injected, are the measured concentrations of the challenge microorganism
             negligible?

D D        Trip Controls: For an UNTESTED sample bottle of challenge microorganism stock
             solution that travels with tested samples between the laboratory and the reactor, is the
             change in the log concentration of the challenge microorganism within the measurement
             error. (I.e., the change in concentration over the test run should be negligible. This is
             typically on the order of 3 to 5%.)

D D        Method Blanks: For sterilized reagent grade put through the challenge microorganism
             assay procedure, is the challenge microorganism concentration non-detectable?

D D        Stability Samples: For influent/effluent samples at low and high UVT, are the challenge
             microorganism  concentrations within 5 percent of each other?

Uncertainty in Collimated Beam Testing Data (See Appendix Cfor more information)

D D        Do the uncertainties in the terms in the UV dose calculation meet the following criteria:
             .   Depth of suspension (d)      <  10 percent
             .   Incidence irradiance (Es)      < 8 percent
             .   Petri factor (Pf)               < 5 percent
             .   L/(d + L)                     < 1 percent
             .   Time (t)                      < 5 percent
             .   (1 - 10-ad)/ad                  < 5 percent

D D        Is the uncertainty in dose-response (UDR), as calculated using equation C.6, less than or
             equal to 30 percent? If not, was UDR incorporated into the VF?
UV Disinfection Guidance Manual                 5-51                                November 2006
For the Final LT2ESWTR

-------
                                    5. Validation of UV Reactors
       Checklist 5.5 Review for Key Validation Report Elements (Page 1  of 2)

Yes No	
D D        Does the validation testing meet QA/QC criteria (see Checklist 5.4)?

D D        For full-scale testing, does the mixing and location of sample ports follow
             recommendations provided in Sections 5.4.3 and 5.4.4, respectively?

D D        If the reactor was validated off-site, do inlet/outlet piping conditions at the water treatment
             plant result in a UV dose-delivery that is the same or greater than the UV dose delivery at
             the off-site testing facility? (See Section 3.6 for recommended inlet/outlet piping
             configurations and Section D.6 for considerations for CFD modeling.)

D D        Were collimated beam tests and full-scale reactor tests performed on the same day for a
             given test condition and using the same stock solution of challenge microorganisms? (See
             Section 5.7 for experimental testing guidelines.)

D D        Is the UV sensitivity of the challenge microorganism and the overall shape of the UV dose-
             response curve consistent with the expected inactivation behavior for that challenge
             microorganism?  See Appendix A of this manual for published UV dose-response curves
             for MS2 and B. subtilis.

D D        Does the validation test design account for lamp fouling and aging, minimum UVT, and
             maximum flow rate expected to occur at the water treatment plant? (See Section 5.6 for
             recommended test design.)

For UV Reactors Using MP Lamps

D D        Is the UV reactor equipped with a germicidal sensor? New UV reactors should have
             germicidal sensors. If an installed reactor uses an MP lamp and a non-germicidal sensor, is
             a polychromatic bias factor incorporated into the derivation of the VF? (See Section D.4.3
             for guidance on the polychromatic bias factor.)

D D        Was validation testing conducted using a challenge microorganism other than MS2 or B.
             Subtilis! If yes, was the need for a correction factor assessed and was that factor applied
             based on the outcome? (See Sections 5.3 and D.4.1 for more information)

For UV Reactors Using the UVIntensity Setpoint Approach

D D        Were the minimum test conditions performed as specified in Section 5.6.1?

D D        Is the UV intensity setpoint low enough to account for combined conditions of minimum
             UVT and maximum lamp fouling/aging at the water treatment plant (See Section 5.6.1 for
             guidance)

D D        Was the minimum RED selected for calculating the validated dose? (See Section 5.8.1 for
             additional guidance.)

D D        Does the VF calculation include both the BRED and USP?  (See Section 5.9 for additional
             guidance.)
UV Disinfection Guidance Manual                  5-52                                 November 2006
For the Final LT2ESWTR

-------
                                    5. Validation of UV Reactors
        Checklist 5.5 Review for Key Validation Report Elements (Page 2 of 2)
 Yes No
 For UV Reactors Using the UV Intensity Setpoint Approach (continued)

 D D       If Us and/or UDR did not meet the QA/QC criteria, were they also included in the VF
             calculation?

 D D       Is the validated dose greater than or equal to the required dose for the water system's target
             pathogen and log inactivation level?

 For UV Reactors Using the Calculated Dose Approach

 D D       Was the minimum number of test conditions evaluated as specified in Section 5.6.2?

 D D       Was the empirical equation developed using standard statistical methods (e.g., multivariate
             linear regression)? (See Section 5.8.2 for additional guidance.)

 D D       Does the validation report include an analysis of goodness of fit and bias for the dose-
             monitoring equation? (See 5.8.2 for additional guidance.)

 D D       Does the VF calculation include both the BRED and UIN? (See 5.9.)

 D D       If Us and/or UDR did not meet the QA/QC criteria, were they also included in the VF
             calculation?

 D D       For the range of UVT values and flow rates expected to occur at the water system, is the
             validated dose greater than or equal to the required dose for the system's target pathogen
	and log inactivation?	
 UV Disinfection Guidance Manual                 5-53                                November 2006
 For the Final LT2ESWTR

-------
                                 5. Validation of UV Reactors
5.13   Evaluating the Need for "Re-validation"

       If a UV reactor is modified in a way that significantly impacts UV dose delivery or
monitoring (e.g., the wetted geometry changes, the lamp technology changes, the UV sensor
characteristics, and/or location change), validation testing should be conducted again (i.e., the
UV reactor has been modified enough to be considered a different reactor with unsubstantiated
performance). This section discusses some common types of UV reactor modifications and
provides guidance on when UV reactors should be "re-validated."

Lamp Assembly

       The relationship between UV dose delivery and monitoring may be impacted by any
design change involving modifications to the following lamp components:

       .   Lamp arc length
          Any reflectors, connectors, and spacers used at the lamp ends
       .   Lamp envelope diameter
       .   Lamp envelope UV transmittance from 185 - 400 nm
       .   Mercury content of the lamp
          Argon content of the lamp

       In many cases, UV  dose delivery and UV sensor modeling can be used to assess the
impacts of changing lamp material and justify the need, or lack of need, for re-validation.

       Changes that will modify the UV output so that emitted intensity is uneven along the
length of the lamp or around its circumference, however, can have a complex impact on UV dose
delivery and would likely warrant re-validation.

Ballasts

       Modifications to lamp ballasts include changing the operating voltage, current,
frequency, and waveform. Modifications to LP lamps will not impact the relationship between
UV dose delivery and UV intensity measurements. With MP lamps, changes in lamp operating
temperature and mercury pressure caused by changes in ballast power will impact the spectral
distribution of emitted light, resulting in a significant impact on UV reactors with non-germicidal
sensors.

       If a water system is using non-germicidal sensors, then EPA recommends that the reactor
be re-validated if there are modifications to the lamp ballasts that change the operating voltage,
current, frequency, and/or waveform.

Lamp Sleeves

       Lamp sleeve design changes include changing the sleeve diameter, thickness, and
material. Changing the sleeve diameter may significantly impact the hydraulics through the
reactor, the measurement of UV intensity, and/or the ideal location of the UV sensors relative to

UV Disinfection Guidance Manual                5-54                              November 2006
For the Final LT2ESWTR

-------
                                5. Validation of UV Reactors
the lamp. Changing the thickness and material of the lamp sleeve will impact its spectral UV
transmittance, thereby impacting both UV dose delivery and UV intensity measurements.

       UV dose delivery and UV sensor modeling may be used to assess the impact of lamp
sleeve design changes. For example, a design change from a standard sleeve to the ozone-free
sleeve described in Figure 5.6 would have a moderate impact on the relationship between UV
dose delivery and UV sensor readings with a non-germicidal sensor and a negligible impact with
a germicidal sensor. Modeling can also be used to show that the UV dose delivery at a given
lamp output, water UVT, and flow rate would be approximately  10 percent greater with the
standard sleeve than with the ozone-free sleeve. If the modeling indicates a change in dose
delivery of greater than 10 percent as a result of lamp sleeve design changes, EPA recommends
that the reactor be re-validated. If it is not possible to model the impact of lamps sleeve design
changes, EPA also recommends the UV equipment be re-validated.
          Figure 5.6 UVT of Standard and "Ozone-Free" Quartz Assuming
                       Air-Quartz and Quartz-Water Interfaces
J100-

S   8° -
c
5   60 -

wi   40 -

£   20 -

^    0
                              Standard Quartz
                                                     Ozone-Tree
                                                       Quartz
                          200    220    240    260    280
                                       Wavelength(nm)
                                      300
320
UV Reactor and Component Dimensions

       Modifications to the wetted dimensions and positioning of the components within the UV
reactor will impact the reactor hydraulics and UV dose delivery. Modifications could also impact
the UV intensity field within the reactor and its measurement. Such changes include altering the
dimensions of the UV reactor, inlet piping, exit piping, baffles,  lamp  sleeves, wipers, and/or UV
sensors. The impact of such modifications on UV dose delivery and UV intensity measurements
can be large or insignificant. Adding a baffle plate will likely have a large impact on UV dose
delivery and a small impact on measured UV intensity. Changing the position of a UV sensor
will likely have a small impact on UV dose delivery and a large impact on the measured UV
intensity.

       UV dose delivery and UV intensity modeling may be used to  assess the impacts of these
modifications. If the modeling indicates a change in dose delivery of greater than 10 percent as a
result of changes to the wetted dimensions of the reactor and/or changes in the positioning of
components, EPA recommends that the reactor be re-validated. If it is not possible to model the
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                     5-55
         November 2006

-------
                                 5. Validation of UV Reactors
impact of modification to the wetted dimensions and positioning of components within the UV
reactor, EPA recommends the UV equipment be re-validated.

UV Sensors

       Modifications to the UV sensors include changes made by the sensor manufacturer to the
sensor itself, its housing and its associated optical components, or installation within the reactor.
Any  modifications that affect the UV sensor response or the flow within the reactor affect should
be evaluated to determine their impacts on dose delivery and dose monitoring. For example, if
the measurement uncertainty of a new sensor is greater than 10 percent, it should be included in
the VF calculations. If the angular response or spectral response of the UV sensor changes,
measurements supported by calculations should be used to evaluate the impact of the change on
UV dose delivery monitoring.
UV Disinfection Guidance Manual                 5-56                              November 2006
For the Final LT2ESWTR

-------
                  6.  Start-up and Operation of UV Facilities

       This chapter describes the start-up activities and routine operational issues associated
with a UV disinfection facility. The start-up discussion focuses on the testing performed during
the start-up process. The rest of the chapter describes requirements and recommendations for
operation, maintenance, monitoring, recording, and reporting for UV facilities. Figure 6.1
illustrates the start-up and routine operation. A detailed description of each activity is provided in
this chapter.

       j Chapter 6 covers:
       {6.1   UV Facility Start-up
       j 6.2   Operation of UV Facilities
       j 6.3   Maintenance of UV Reactors
       | 6.4   Monitoring and Recording of UV Facility Operation
       { 6.5   UV Facility Reporting to the State
       | 6.6   Operational Challenges
       j 6.7   Staffing, Training, and Safety Issues
       The guidelines provided in this manual are based on industry experience and
manufacturers' recommendations. Because of numerous differences among UV facilities and UV
equipment, this document does not address all start-up and operation and maintenance (O&M)
issues that may occur.
6.1    UV Facility Start-up

       For the purposes of this manual, the start-up of the UV facility is considered as the
transition from the construction phase to the operation phase. A start-up plan should be
developed in collaboration with the UV facility designer, plant operations staff, and the UV
manufacturer. The start-up plans should include O&M manual development, state coordination,
functional testing, determination of validated operating parameters, performance testing, and
final inspection.
6.1.1  O&M Manual

       The O&M manual should be site-specific and based on as-built drawings, manufacturer's
shop drawings, operating procedures, operational requirements, recommended maintenance
tasks. If performance testing is completed before the O&M manual is finalized, testing results
should be included in the manual. If possible, the O&M manual should be developed before
performance testing and routine operations. At a minimum, O&M manuals should address the
following items:

       .   Federal and state regulatory requirements and guidelines
          Treatment objectives

UV Disinfection Guidance Manual                 6-1                               November 2006
For the Final LT2ESWTR

-------
                                  6. Start-up and Operation of UV Facilities
                        Figure 6.1. Start-up and Operation Flowchart1
                                               Completion of
                                              Construction and
                                                 Inspection
                   Start-Up Activities
Operations and Maintenance
          Manual
       Section 6.1.1
                                             State Coordination
                                              during Start-up
                                                Section 6.1.2
                                                    z
                                             Functional Testing
                                                Section 6.1,3
                                                    z
                                          Determination of Validated
                                         Operational Conditions and
                                         Setting Operational Controls
                                                Section 6.1.4
                                                    _L
                                            Performance resting
                                                Section 6.1.5
                                                    z
                                              Final Inspection
                                                Section 6.1.6
          Routine Operation
                   Operation of UV
                       Facility
                      Section 6.2
                                                                             z
      Maintenance of
       LV Reactors
        Section 6.3
 Monitoring and
 Recording of UV
Facility Operation
   Section 6.4
                                                                          UV Facility
                                                                        Reporting to the
                                                                            State
                                                                          Section 6.5
                                           Operational Challenges
                                                 Section 6.6
                                        Staffing, Training, and Safety
                                                   Issues
                                                Section 6.7
1 Start-up activities are not necessarily in chronological order.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
           6-2
                    November 2006

-------
                            6. Start-up and Operation of UV Facilities
          General description of UV facility
       .   Relationship to other unit treatment processes
       .   UV reactor design criteria
       .   Validated operational parameters
          Controls and monitoring
          Compliance monitoring, recording, and reporting
       .   Standard operating procedures
       .   Start-up procedures
          Shut-down procedures (manual and automatic)
          Safety issues
       .   Emergency procedures and contingency plan
       .   Alarm response plans
       .   Preventive maintenance needs and procedures
       .   Equipment calibration needs and procedures
       .   Troubleshooting guide
       .   Equipment component summary
          Spare parts inventory
          Contact information for equipment manufacturers and technical services

6.1.2  State Coordination during Start-up
       States should be contacted during construction to determine the state-specific
requirements and submittals. The states may request the record drawings, O&M manual, and an
engineer's certificate of completion. In addition, the state may need to visit the site to approve
the start-up of the UV facility.

6.1.3  Functional Testing
       Functional testing verifies that each component's operation is in accordance with the
specifications in the contract documents. It should include verification of UV equipment
components, instrumentation and control (I&C) systems, and flow distribution and head loss.
Items that are not unique to UV facilities (e.g., valves, flow meters, backup generators, or
uninterruptible power supplies) are not described in this manual; however, their functionality
should still be verified.
UV Disinfection Guidance Manual                6-3                               November 2006
For the Final LT2ESWTR

-------
                            6. Start-up and Operation of UV Facilities
6.1.3.1   Verification of UV Equipment Components

       Most functional testing is completed through simulations of specific operating conditions
and monitoring UV reactor operation and response. Functional testing entails flooding and
energizing UV reactors to confirm the operation of the lamps, ballasts, ballast cooling system,
cleaning system, UV sensors, and UVT analyzers.

       It is strongly recommended that the UV manufacturer inspect the UV facility before the
UV reactors are energized and be present when the UV reactors are first energized.
Manufacturers may require the presence of one of their representatives during these activities as
a condition of their equipment warranty.

UVSemor

       UV sensors must be included in the UV reactor to verify that the reactor is operating
within validated conditions [40 CFR 141.720(d)]. The calibration of the duty and reference UV
sensors should be checked during functional testing using the procedure recommended in Section
6.4.1.1. UV sensors that are not in calibration should be returned to the manufacturer for
replacement or recalibration.

Lamps, Ballasts, and Ballast Cooling System

       The lamps, ballasts, and ballast cooling system operation are verified by energizing the
UV lamps, then verifying lamp and ballast operation via the UV sensor measurements and visual
verification of the ballast cooling fan operation. In addition, the power [kilowatt (kW)] delivered
to the lamps should be verified as the same as documented in the validation report for at least
three power settings.

On-line UVT Analyzer

       If the dose-monitoring strategy of the UV reactor is the Calculated Dose Approach (see
Section 3.5.2 for a description of dose-monitoring  strategies), the UV reactor should be equipped
with a UVT analyzer. Calibration of the on-line UVT analyzer should be verified. A
recommended procedure for verifying calibration is described in Section 6.4.1.2.

Cleaning System

        The necessary functional testing depends on the type of cleaning used, and the
components to be verified for each cleaning system are summarized in Table 6.1. Cleaning
systems are described in Section 2.4.5.
UV Disinfection Guidance Manual                6-4                              November 2006
For the Final LT2ESWTR

-------
                             6. Start-up and Operation of UV Facilities
                 Table 6.1. Functional Testing of Cleaning Systems
      Cleaning System
                   Items to be Verified
     On-line mechanical
     cleaning (OMC)
Smooth movement of the wiper with no jamming or binding of the
wiper on the sleeve
Extension of wiper stroke to the full length of the sleeve with no
impact at the end of travel that could damage or break the sleeve
Proper operation of the wiper drive mechanism and motor with no
slipping or binding
     On-line mechanical-
     chemical cleaning
     (OMCC)
Same as on-line mechanical cleaning (above)
The chemical injection point is accessible
The seal that contains the chemical solution is intact
     Off-line chemical
     cleaning (OCC)
The chemical injection wand should be connected to the chemical
pump to verify that a proper seal is achieved
Outside of the reactor, in a safe location, the chemical pump should
be initiated to ensure that the wand is operating properly and an
appropriate amount of pressure is achieved
The wand should then be connected to the reactor and turned on to
make sure the seal is intact and the wand is functioning properly
6.1.3.2   Verification of Instrumentation and Control Systems

       The amount of testing for the instrumentation and control systems depends on the
complexity of the dose-monitoring strategy and operations approach used. Testing should
include verifying control loops, checking operation functions, and verifying all control actions.
As described below, the UV reactors should be run through a series of simulations that represent
the possible operating scenarios to confirm that the UV reactor responses are appropriate. A
manufacturer representative should be present during the simulations to assist in troubleshooting
and addressing any issues that may result from the packaged UV reactor controls.

       Typically, the packaged UV reactor control panel contains all the components needed to
control and operate the UV reactor. The panel should provide the operating status, lamp status
indicators, diagnostic information, and operator interface capability. The panel may also include
programmable logic controllers (PLC), ballasts, and lamp starters.

       Electronic signal simulations imitate the signals that will be sent  to the control system
during normal operation. The I&C logic programming should be monitored during simulations to
verify the programming is correct. These "dummy" simulations should be used to confirm that
UV reactors and all ancillary equipment and instrumentation, including valves, flow meters, and
UVT analyzers will operate consistent with the I&C programming. The UV reactors should not
operate during these simulations (i.e., water is not flowing and lamps are not energized). As
applicable, the following specific  operating conditions should be electronically simulated, as
well as any other conditions the manufacturer recommends:
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                  6-5
November 2006

-------
                            6. Start-up and Operation of UV Facilities
          Cold start of the UV reactors

          Cool down and restart of the UV reactor

          Sequencing of the UV reactors in multiple-reactor installations
       .   Adjustment of lamp intensity or number of lamps on in response to varying water
          quality and flow rate
       .   Shut-down of the UV reactors

       .   Operation of the UV reactors during line power failure (when back-up generators or
          UPS are available)

       .   Manual override, safety interlocks, and report generation

       .   Operation of the UV reactors through the plant SCAD A system

       .   Incorporation of a sensor correction factor

       In addition to simulating possible operating conditions, each alarm condition and
monitoring function incorporated in the design should be verified. Possible monitoring functions
and alarm conditions are discussed in Section 4.3.3 and may include the following conditions:
       .   Operation outside the validated conditions
          -   Low validated dose or UV intensity
          -   Low UVT
          -   High flow rate
       .   Lamp age
       .   Lamp or ballast failure
       .   Low water level in the UV reactor
       .   High temperature
       .   OMC or OMCC system failure
       .   Loss of control signals

6.1.3.3   Verification of Flow Distribution and Head Loss
       A minimum of three flow rates that span the range of operating conditions should be
tested. If possible, one condition should be the maximum design flow rate through the UV
facility with all duty reactors in operation; the other conditions should consist of combinations of
UV Disinfection Guidance Manual                6-6                                November 2006
For the Final LT2ESWTR

-------
                            6. Start-up and Operation of UV Facilities
the reactors operating at their design flow rates (e.g., two of five 10-mgd reactors operating at a
total UV facility flow rate of 20 mgd). Clamp-on type flow meters can be used for field
verification of the flow split.

       The head loss should be measured at these same test conditions for each reactor and
compared to the head loss specified in the contract documents (if applicable). Pressure
transducers or pressure gauges can be used to measure the head loss.
6.1.4  Determining Validated Operational Conditions and Setting Operational
       Controls

       For each UV reactor, the operating conditions associated with a given level of
inactivation credit must be defined based on validation testing results [40 CFR 141.720(d)(2)].

       Section 5.10.1 and 5.10.2 describe how the validated dose and validated operating
conditions are established for two dose-monitoring strategies, the UV Intensity Setpoint
Approach and the Calculated Dose Approach, respectively. Appendix B supports Sections 5.10.1
and 5.10.2 by providing examples of validation testing data analyses. Examples 6.1 and 6.2
expand on guidance in Section 5.10 and Appendix B by showing how the same hypothetical
water systems in Appendix B established operational alarms to ensure that they operate within
validated conditions.
UV Disinfection Guidance Manual                 6-7                               November 2006
For the Final LT2ESWTR

-------
                            6. Start-up and Operation of UV Facilities
Example 6.1. Setting Operational Controls for the UV Intensity Setpoint Approach - Single
        Setpoint Operation (Corresponds to the Validation Example in Section B.I)

       Background: System X plans to add UV disinfection to its treatment plant to achieve
2.5-log Cryptosporidium inactivation credit. Based on LT2ESWTR UV dose requirements
(summarized in Table 1.4 of this manual), the water system needs to meet a required UV dose of
8.5 ml/cm2 to achieve this level of inactivation. During UV facility planning, the water system
establishes a design flow of 400 gpm and minimum UVT of 90 percent.

       System X selects two low-pressure high-output (LPHO) reactors (one duty and one stand-
by) with eight lamps each that use the UV Intensity Setpoint Approach. Because their flow rate
and UVT do not vary much, System X decided to use the single setpoint approach that applies to
all validated operating conditions.

       Summary of Validation Test Results: Validation testing produced a UV intensity
setpoint of 11.7 mW/cm2 at a maximum flow rate of 394 gpm with a single reactor operating
with all lamps turned on. The validated dose at the  setpoint is 11.3 mJ/cm2, which is greater than
the required dose of 8.5 mJ/cm2. As long as the UV intensity as measured by the UV sensor is
greater than 11.7 mW/cm2, the validated dose is greater than the required dose and the reactor is
operating within the validated limits.

       Operational Controls: As shown in the table below, System X set the UV intensity alarm
at 12.5 mW/cm2 to provide an operational cushion. System X also set a flow rate alarm at 375
gpm. Because the validation testing protocol for the UV Intensity Setpoint Approach (as
described in Chapter 5) accounts for changes in UVT, UVT is not regularly monitored during
operations.
Operating Parameter
UV Intensity as measured by
the UV sensor
Flow rate through the reactor
Validated Operating
Conditions
> 1 1 .7 mW/cnf
< 394 mgd
Major Alarm
Sounds if < 12.5 mW/crn^
Sounds if > 375 gpm
       Although this operating strategy is simple and straightforward, System X could have
improved efficiency by reducing the UV intensity at lower flow rates, which can only be done if
the validation data support UV intensity adjustment with flow. To further improve energy
efficiency using the single setpoint approach, the flow could be maximized through one reactor
before energizing another reactor for multiple reactor systems.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
November 2006

-------
                             6. Start-up and Operation of UV Facilities
       Example 6.2. Setting Operational Controls for the Calculated Dose Approach
                  (Corresponds to the Validation Example in Section B.2)

       Background: System Y plans to add UV disinfection to their treatment plant to achieve
2.0-log Cryptosporidium inactivation credit. Based on the LT2ESWTR UV dose requirements
(summarized in Table 1.4 of this manual), the water system needs to meet a required UV dose of
5.8 ml/cm2 to achieve this level of inactivation. During UV facility planning, System Y
establishes a design flow rate range of 3 to 10 mgd and a minimum operating UVT of 87 percent.

       System Y selects three UV reactors (two duty and one stand-by) with six 8-kW medium-
pressure (MP) lamps each with power settings ranging from 40-100 percent. The reactors have
one germicidal UV sensor monitoring each lamp. The reactors were validated for flow ranges of
2.5 - 10 mgd and use the Calculated Dose Approach.

       Summary of Validation Test Results: Validation testing as described in Appendix B
produced the following dose-monitoring equation (Equation B.14):
log (RED)= -O.S29 -2.519 xlog (A254)+0.166 xlog
                          + 0.409 x log
         where:
            RED  = Calculated dose

       As noted in the validation report, the Validation Factor (VF) is 2.28. The validated dose
is calculated by dividing the calculated dose by the VF. The validated dose must be greater than
the required dose of 5.8 mJ/cm2 for System Y to receive treatment credit for 2.0 log-inactivation
of Cryptosporidium.

       Operational Controls: The table below summarizes the major alarms that System Y
programmed into their PLC to ensure that they operate within validated conditions
Operating Parameter
Validated Dose (equal to the
Calculated Dose/ 2.28)
Flow rate through the reactor
UVT as measured by an on-line
UVT analyzer
Validated Operating
conditions
> 5.8 mJ/cm2
< 10 mgd 1
85 - 95% 2
Major Alarm 3
Sounds if < 6.3 mJ/cm2
Sounds if > 9 mgd
Sounds if <87%
         If the flow rate is less than 2.5 mgd, the PLC will default to 2.5 mgd in the dose-monitoring
         equation.
         If the UVT measured is higher than 95 percent, which is the highest validated UVT, the PLC will
         default the UVT to 95 percent in the dose-monitoring equation.
         Note the major alarms are set at a conservative level compared to the validation conditions to give
         the operators more time to respond to low validated dose, high flows, and low UVTs.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
6-9
November 2006

-------
                             6. Start-up and Operation of UV Facilities
6.1.5  Performance Testing

       Performance testing is intended to assess the operating performance of the UV facility as
a whole and is generally accomplished through extensive monitoring during the early stages of
continuous operation. Note that performance testing is not intended to validate disinfection
performance, which is completed during validation testing (as described in Chapter 5). However,
performance testing can be used to confirm that the actual operating conditions are within the
constraints established during validation testing as described in Section 6.1.4.

       Because performance testing should compare operating conditions to validated
conditions, the lamps should be operated as they were during validation testing. Therefore, UV
lamps should be burned-in before performance testing, which typically takes 100 hours of
continuous operation (Section 5.7.2). The actual required burn-in time should be discussed with
the manufacturer and confirmed through documented operating experience at other UV facilities.

       The scope and duration of performance testing will be project-specific and should be
established by the PWS and designer based on the objectives of the performance testing. The
duration of performance testing should be adequate to demonstrate to the PWS and the state that
the UV facility can continually perform according to specifications in uninterrupted operation.
This could be as little as 48 hours, but may be longer, depending upon the nature of the
installation, the variability of the source, and any specific state and PWS requirements. Similarly,
the scope of the testing may range from an increased monitoring frequency that confirms
operation within validated limits to an extensive testing protocol aimed at optimizing reactor
performance and establishing long-term operating procedures. During performance testing,
treated water may be sent to the distribution system if upstream treatment has not changed, meets
existing regulations, and is approved by the state.

       Performance testing may include the following items:

       .   Operation of each UV reactor in automatic mode to verify that the control system is
           identical to that established during validation testing

       .   Demonstration of UV reactor start-up and switchover sequences that result from
           water quality and/or flow rate changes

           Observation of operation, including periods of off-specification operation that arise
           from alarm conditions and any power quality problems

       .   Measurement of electrical service voltage, current, and power consumption with
           different flow and water quality combinations to optimize energy use within the
           constraints established  during validation

       .   Assessment of the effectiveness of the cleaning system by inspecting sleeve clarity
           and condition at regular intervals throughout the test period
UV Disinfection Guidance Manual                6-10                               November 2006
For the Final LT2ESWTR

-------
                            6. Start-up and Operation of UV Facilities
          Confirmation that the programmed cleaning frequency correlates with the actual
          frequency of cleaning

          Confirmation of duty UV sensor accuracy using reference UV sensors. (See
          Section 6.4.1.1.)

       .   Observation of ballast temperature and cooling system performance

          Verification of the calibration of the on-line UVT analyzer (if applicable). (See
          Section 6.4.1.2.)

          Confirmation of backup generator and/or UPS power transfer to the UV equipment

       The performance testing should be tailored to the specific UV facility. An example
monitoring program for a 4-week performance test is shown in Table 6.2.

       Any off-specification time and flow volume should be recorded during all performance
tests, and these results should be evaluated to verify that off-specification limitations are not
exceeded. Off-specification volume during performance testing does not need to be reported to
the state. Recording of the off-specification time and volume is meant to identify operational
problems to be addressed. During performance testing, any component that is not operating
properly should be corrected and retested to confirm satisfactory operation. This step may
require manufacturer involvement, especially if specifications in the contract documents were
not met. Following performance testing, ongoing monitoring and recording of reactor operation
should continue at a reduced frequency as discussed in Section 6.4 and as required by the state.
UV Disinfection Guidance Manual                6-11                              November 2006
For the Final LT2ESWTR

-------
                            6. Start-up and Operation of UV Facilities
       Table 6.2. Example Monitoring During a Four Week Performance Test
Frequency
Continuous
Weekly
Twice during
testing period
After 4 weeks,
lOOOMCor
OMCC cycles, or
one OCC event
Task
Confirm the validated
setpoint(s)
Develop energy efficient
operation
Log off-specification
occurrences
Monitor UV sensor
calibration
Monitor the on-line UVT
analyzer calibration
Switch to standby
reactor
Switch to standby
power or UPS
Inspect lamp sleeves for
fouling
Notes
Monitor reactor operation to confirm compliance with the
setpoint(s) established during validation (Section 6.1.4).
Monitor the power consumption. Test the automatic
operation and power consumption under the flow and
water quality variations to determine if energy efficiency
improvements can be made within the validation
constraints.
Log alarms and indicate whether the reactor is off-
specification according to validation criteria. Record off-
specification time and volume.
Check the duty UV sensor against a reference UV
sensor, using the recommended protocol (Section
6.4.1 .1) to determine whether the duty UV sensor is in
calibration.
Monitor calibration of the on-line UVT analyzer (Section
6.4.1.2).
Monitor the time necessary to switch to a standby reactor
to determine if operation will be off-specification during
switch-over.
Monitor the time necessary to switch to the standby
power supply to determine if operation will be off-
specification because of power transfer. Test the backup
power supply for a minimum of one hour.
Remove a sleeve from the reactor and inspect as
recommended in Section 6.3.2.1.
  OMC = on-line mechanical cleaning; OMCC = on-line mechanical chemical cleaning; and OCC = on-line
  chemical cleaning.
6.1.6  Final Inspection

       As the last step in the start-up process, a detailed inspection of the UV facility should be
completed. The inspection should include a visual assessment to verify that all components meet
the technical specifications of the UV equipment specification and validation report and that the
UV facility was completed in accordance with the construction documents. All UV facility
components and associated valves and piping should be thoroughly cleaned and disinfected prior
to service.

6.2    Operation of UV Facilities

       The operation of UV facilities will vary based on the UV manufacturer, the UV reactor
configuration, and the dose-monitoring strategy. This section discusses required and
recommended operational and routine start-up and shut-down procedures common to most UV
equipment and is general in nature. The site-specific operational procedures should be  developed
UV Disinfection Guidance Manual
For the Final LT2ESWTR
6-12
November 2006

-------
                            6. Start-up and Operation of UV Facilities
in coordination with the manufacturer, UV facility designer, and facility operators, and should be
described in the O&M manual (Section 6.1.1). Small systems should consider discussing
operations with the state to determine if simplified operations are possible.
6.2.1  Operational Requirements

       To receive inactivation credit, the UV reactors must operate within the validated limits
[40 CFR 141.720(d)]. When a UV reactor is operating outside of these limits, the UV reactor is
operating off-specification as described in Section 3.4.1. Filtered and unfiltered systems that use
UV disinfection to meet the Cryptosporidium treatment requirement of the LT2ESWTR must
demonstrate that at least 95 percent of the water delivered to the public during each month is
treated by UV reactors operating within validated limits [40 CFR 141.720(d)(3)]. Guidance on
determining validated operating conditions is in Section 5.10. The specific monitoring
requirements associated with off-specification operation are described  in Section 6.4.1.
6.2.2  Recommended Operational Tasks

       UV equipment typically uses automatic control systems and does not need significant
manual attention for routine operation. Even when UV equipment is operated manually, the only
parameter that typically can be controlled is lamp power, and some UV reactors can also vary the
number of lamps energized. Therefore, even manual operation does not result in significant
operator interaction. Table 6.3 summarizes recommended operational tasks. Recommended
maintenance tasks are discussed in Section 6.3.
          Table 6.3. Recommended Operational Tasks for the UV Reactor
Frequency
Daily
Weekly
Monthly
Semi-
annually
Recommended Tasks
. Perform overall visual inspection of the UV reactors.
. Confirm that system control is on automatic mode (if applicable).
. Check control panel display for status of system components and alarm status and
history.
. Verify that all on-line analyzers, flow meters, and data recording equipment are
operating normally.
. Review 24-hour monitoring data to confirm that the reactor has been operating within
validated limits during that period.
. Verify that ballast cooling fans are operational and that ballasts are not overheated.
. Initiate manual operation of wipers (if provided) to verify proper operation.
. Check lamp run time values. Consider changing lamps if operating hours exceed
design life.
. Check ballast cooling fans for unusual noise.
. Check operation of automatic and manual valves.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
6-13
November 2006

-------
                            6. Start-up and Operation of UV Facilities
6.2.3  Start-up and Shut-down of UV Reactors

       This section describes start-up procedures, shut-down procedures, and winterization of
the UV reactors.
6.2.3.1   Routine Start-up

       The following routine start-up procedure serves as an example approach. The UV
reactors should be operating within validated conditions once the start-up sequence is complete.

       1.  Initiate the UV reactors' start-up sequence. [Note: Some UV reactors may need
          reduced water flow to cool the lamps during start-up, which would normally be
          initiated  automatically. The cooling water exiting the reactor is not disinfected and is
          considered off-specification unless it is diverted to waste.]

       2.  Check the SCADA panel or other display to verify that the necessary numbers of
          lamps are on and all of the monitoring parameters are being displayed.

       3.  Check and resolve any system alarms being displayed.

       4.  Confirm that all  on-line analyzers (UV sensors and UVT analyzers, if applicable) and
          flow meters are operating within calibration.

       5.  After the lamp warm-up period, increase flow to the validated range (if flow is not
          automatically  adjusted with UV reactor control sequence).

       6.  Verify correct flow split between parallel UV reactors using flow meters and/or
          differential pressure gauges if these devices are available.

       7.  Verify that the UV reactor is operating within validated limits (e.g., flow rate, UV
          intensity, lamp status, validated dose).
6.2.3.2   Start-up Following Maintenance

       The following additional steps should be taken before completing Steps 1-7 described
in the example routine start-up procedure (Section 6.2.3.1) when maintenance has been
performed on the reactor:

       1.  Follow site-specific safety procedures for the power supply and control panel (e.g.,
          removing lockouts and tagouts).

       I.  Confirm that all lamp and ground connections are properly made. Verify that all
          incoming power conductors, including ground conductors, are properly terminated.

       5.  Verify that the lamp ends and all other reactor ports are covered and/or sealed to
          eliminate the potential for operator exposure to UV light.

UV Disinfection Guidance Manual                 6-14                              November 2006
For the Final LT2ESWTR
2.
3.

-------
                             6. Start-up and Operation of UV Facilities
       4.  Confirm the breakers are turned on, and all electrical cabinets and equipment are clear
          and closed.

       5.  Perform Steps 1 - 5 of Section 6.2.3.1.

       6.  Verify that all air is purged from the reactors (i.e., the reactor is completely flooded).
          Check the top of the reactor for heat buildup, which indicates an air pocket.

       7.  Perform Steps 6 and 7 of Section 6.2.3.1.
6.2.3.3   Routine Shut-down

       UV reactors are shut down periodically because of water quality or flow changes. The
main steps involved in shutting reactors down are as follows:

       1.  Throttle the effluent valve (if not part of the control sequence) to close it.

       2.  De-energize the reactor immediately after the effluent valve is closed.


6.2.3.4   Shut-down Prior to Maintenance

       UV reactors are also shut down periodically for maintenance (e.g., cleaning). The
following steps should be taken following the routine shut-down steps (Section 6.2.3.3) to
prepare a reactor for maintenance:

       1.  Follow lockout and tagout procedures for the facility.

       2.  Drain the reactor if necessary for the specific maintenance task.

       3.  Inspect and repair or replace any necessary equipment.

       If extended shut-down time is planned, the reactor should be drained to avoid excessive
fouling. After an extended shut-down period (more than 30 days), the operator should perform a
cleaning and then inspect the lamp sleeves for fouling. Manual or more extensive cleaning may
be necessary before start-up, as described in Section 6.3.2.1.


6.2.3.5   Winterization

       In  most drinking water applications, the UV reactors will be located within a building.
However,  in some instances, the reactors may be located in unheated concrete vaults or outside.
When shutting down a UV reactor for an extended period of time is necessary and damage from
freezing is possible, the UV reactors should be winterized according to the manufacturer's
recommendations.
UV Disinfection Guidance Manual                6-15                               November 2006
For the Final LT2ESWTR

-------
                            6. Start-up and Operation of UV Facilities
6.3    Maintenance of UV Reactors

       No specific regulatory requirements exist for maintaining a UV reactor; however, the UV
reactors should be maintained so that disinfection requirements are met. Poor maintenance may
cause the UV reactors to operate off-specification for extended periods of time. As part of the
maintenance tasks, UV reactor components will need to be replaced; therefore, an inventory of
spare parts is necessary. These tasks are described in this section.
6.3.1  Summary of Recommended Maintenance Tasks

       Table 6.4 summarizes the recommended maintenance tasks and refers to the general
guidelines for those tasks that are discussed in Section 6.3.2. The frequency of performing the
maintenance tasks in this section are recommendations and likely will be specific to the UV
equipment installed. Therefore, the UV manufacturer should be contacted to determine the
appropriate frequency. Items that are not unique to UV facilities (e.g., valves, flow meters,
uninterruptible or backup power supplies) are not described; however, maintenance on such
items should also be performed per the manufacturer's recommendation. Before maintenance is
performed, the operator should wait at least 5 minutes (or as recommended by the UV
manufacturer) for the lamps to cool and the energy to dissipate. Lockout and tagout protocol
should be followed if the main electrical supply to the UV reactors needs to be disconnected for
the maintenance task.
                   Table 6.4. Recommended Maintenance Tasks1
Frequency
Monthly
(no cleaning or
OCC)
Semi-annually
(OMC or OMCC)
Monthly
Bimonthly
(MP lamps)
Quarterly
(LP and LPHO
lamps)
Semi-annually
Task
General Guideline &
Section Reference
Check cleaning
efficiency
Section 6.3.2.1
Check reactor housing,
sleeves, and wiper
seals for leaks
Check intensity of UV
lamps
Section 6.3.2.2
Check cleaning fluid
Action
. Record UV sensor reading.
. Extract one sleeve per reactor (or one sleeve per
bank of lamps) for inspection.
. If fouling is observed on the first sleeve, check
remaining sleeves and all UV sensor windows.
. Manually clean sleeve(s) and UV sensor windows if
fouling is observed.
. Record UV sensor reading after cleaning and
compare to original reading.
Replace housing, sleeve, or wiper seals if damaged or
leaking.
If UV sensors monitor more than one lamp, verify that
the lamp with the lowest intensity value is closest to the
UV sensor by replacing the lamp closest to the UV
sensor with one-forth of the lamps in each row/bank
(minimum of three). Place the lowest intensity lamp next
to UV sensor.
Replenish solution if the reservoir level is low. Drain and
UV Disinfection Guidance Manual
For the Final LT2ESWTR
6-16
November 2006

-------
                              6. Start-up and Operation of UV Facilities
                     Table 6.4. Recommended Maintenance Tasks
                                                                         1
Frequency
(OMCC)
Annually
Annually
When duty UV
sensors fail
calibration
Manufacturer's
recommended
frequency
Lamp/
manufacturer
specific
When lamps are
replaced
Sleeve/
Manufacturer
specific
Manufacturer's
recommended
frequency
Manufacturer's
recommended
frequency
Manufacturer's
recommended
frequency
Task
General Guideline &
Section Reference
reservoir (if provided)
Section 6.3.2.1
Calibrate reference UV
sensor
Section 6.3.2.3
Test-trip GFI
Section 6.3.2.4
Replace or recalibrate
duty UV sensors
Section 6.4.1.1
Check thermometer
and/or water level
indicator
Section 6.3.2.5
Replace lamp
Section 6.3.2.6
Properly dispose of
lamps
Section 6.3.2.6
Replace sleeve
Section 6.3.2.7
Clean UVT analyzer
and replace parts
Section 6.3.2.8
Inspect OMC or OMCC
drive mechanism
Inspect ballast cooling
fan
Section 6.3.2.4
Action
replace solution if the solution is discolored.
Send the reference UV sensor to a qualified facility
(e.g., manufacturer) for calibration. Calibration should
use a traceable standard (e.g., National Institute of
Standards and Technology (MIST), National Physical
Laboratory (NPL), Osterreichisches Normungsinstitut
(ONORM), or Deutsche Vereinigung des Gas- und
Wasserfaches (DVGW)).
Maintain GFI breakers in accordance with
manufacturer's recommendations.
Send the duty UV sensors to a qualified facility (e.g.,
manufacturer) for calibration, or replace the duty UV
sensors.
Visually inspect thermometer and/or water level monitor
and replace at the manufacturer's recommended
frequency.
Replace lamps when any one of the following conditions
occurs:
. Initiation of low UV intensity or low validated dose
alarm (UV intensity or validated dose equal to or less
than setpoint value) after verifying that this condition
is caused by low lamp output.
. Initiation of lamp failure alarm after verifying it is not a
nuisance alarm.
Send spent lamps to a mercury recycling facility or back
to the manufacturer.
Replace sleeve when damage, cracks, or irreversible
fouling significantly decreases UV intensity of an
otherwise acceptable lamp to the minimum validated
intensity level. Adjust the replacement frequency based
on operational experience.
Clean and replace parts according to manufacturer's
recommended procedure.
Inspect and maintain OMC or OMCC drive routinely as
recommended by the manufacturer.
Check the ballast cooling fans for dust buildup and
damage. Replace if necessary. Replace air filters (if
applicable).
   OMC = on-line mechanical cleaning;  OMCC = on-line mechanical  chemical cleaning; and OCC = on-line
   chemical cleaning.
    Maintenance activities should be consistent with manufacturer's instructions.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
6-17
November 2006

-------
                            6. Start-up and Operation of UV Facilities
6.3.2  General Guidelines for UV Reactor Maintenance

       This section describes general guidelines for UV reactor components that relate to
maintenance tasks summarized in Table 6.4.
6.3.2.1   Fouling

       As discussed in Chapters 2 and 3, the lamp sleeves and UV sensors/windows will likely
foul over time, depending on the water quality, lamp type, and cleaning regime. This section
describes possible cleaning techniques and provides some specific recommendations for
addressing fouling issues.

Sleeve and UV Sensor Surface/Window Fouling

       Three types of sleeve cleaning techniques, as discussed in Section 2.4.5, are used: off-line
chemical cleaning (OCC),  on-line mechanical cleaning (OMC), and on-line mechanical-chemical
cleaning (OMCC) methods. The frequency of cleaning is  site-specific. An appropriate sleeve
cleaning frequency (manual or automatic) can be determined based on the rate of fouling during
the start-up period, which can be assessed by monitoring over time the UV sensor measurement
or validated dose (Calculated Dose Approach). For routine operation, the cleaning frequency
should be increased or decreased based on the amount of fouling left on the sleeves determined
from the sleeve inspections and the loss of UV intensity before cleaning.

       Sleeves should initially be inspected for fouling every six months if OMC or OMCC is
used and every month if OCC or no cleaning is used. This frequency should be adjusted, if
necessary, after operating data are available. A decrease in UV intensity or validated dose at a
consistent UVT may indicate sleeve fouling, and sleeves should be inspected if fouling is the
suspected cause of the UV intensity drop (Section 6.6.1). Additionally, the UV sensor windows
(if applicable) should be inspected for fouling and supplemental cleaning should be conducted if
necessary, according to the manufacturer's recommendation.

       For sleeve inspection, one sleeve per reactor (or one  sleeve per bank of lamps for reactors
with multiple rows/banks of lamps) should be inspected. The sleeves should be handled as
described in Section 63.2.7. If damage or fouling is observed, the remaining sleeves should be
inspected. External sleeve fouling can be difficult to identify. Sleeve discoloration is more easily
seen by placing the sleeve  on a clean, white, lint-free cloth next to a new sleeve. The presence of
streaks may indicate that the OMC or OMCC wiper material is worn, damaged, or misaligned;
therefore, the wiper should also be inspected. If fouling is observed, the cleaning frequency
should be increased, and supplemental manual cleaning should be conducted as necessary. The
UV reactors need to be drained for sleeve inspection, and the inside of the UV reactor should
also be inspected. Any algae that has grown on the surface or any other surface fouling that has
occurred should be manually cleaned according to the  UV manufacturer's recommended
procedure.

       Manual cleaning (i.e., beyond routine OCC, OMC, or OMCC cleaning) of lamp sleeves,
if necessary, should be according to manufacturer recommendations and procedures. Abrasive

UV Disinfection Guidance Manual                6-18                               November 2006
For the Final LT2ESWTR

-------
                            6. Start-up and Operation of UV Facilities
cleaners or pads that might scratch the lamp sleeve should not be used. Also, the inside of the
sleeve should be dry before re-installation because water or cleaning solutions could cause a
coating to form during operation. One method of drying the sleeve is to use isopropyl alcohol
and a lint-free cloth; however, no alcohol should remain inside the sleeve after this procedure. As
noted earlier, when the sleeves are re-installed after inspection, the manufacturer's procedure
should be closely followed to avoid over-tightening of the compression nuts.

       If OMC or OMCC cleaning is used, the wipers should be checked for deformation or
degradation at the same time the sleeves are checked. The cleaning solution reservoir in OMCC
systems should be checked every six months to determine whether more solution should be
added. The solution should be replaced if it is discolored or if the OMCC system is not
effectively cleaning the sleeve.

Fouling  While Out-of-service

       When the UV reactors are out-of-service and full of water, the sleeves may foul
(Toivanen 2000). The rate of fouling is site-specific and depends on the water quality. UV
reactors equipped with OMC or OMCC should continue to clean the sleeves, potentially at a
lower frequency, even though the UV reactor is off-line, which should prevent fouling of the
sleeves. For UV reactors that do not include OMC or OMCC, the PWS should consider draining
the UV reactor if it is off-line for more than one week. However, this period could be shorter or
longer, depending on the water quality. After a shut-down period of more than 30  days, the
operator  should perform a cleaning (OCC, OMC,  or OMCC) and then inspect the  lamp sleeves
for fouling. Extraction and manual cleaning of sleeves may be necessary before start-up after
extended periods of standby.
6.3.2.2    Lamp Output Variability

       UV lamp output differs for each lamp, depending on lamp age and lot. As discussed in
Section 2.4.6, a UV sensor measures the changes in UV intensity at its location in the UV
reactor. However, a UV sensor cannot measure lamp output variability unless each lamp has a
UV sensor. PWSs that have UV reactors with a UV sensor monitoring more than one lamp
should assess the UV lamp variability every 2 months for MP lamps or every 3 months for LP
and LPHO lamps. If all the lamps monitored by a UV sensor are close in age (i.e., their age
varies by less than 20 percent), it is not necessary to check the output  of each lamp. In this case,
the oldest lamp should be placed in the position nearest the UV sensor. The recommended
procedure for evaluating the lamp output variability is to:

       1.  Identify the lamps that can be used to evaluate the lamp variability (one-fourth of the
          lamps in each row/bank or a minimum of 3 lamps, which ever is greater)

       2.  Place each evaluation lamp in the position nearest the UV  sensor and record the
          intensity value

       3.  Repeat Steps  1 and 2 until one-fourth of the lamps (3 minimum) have been assessed
UV Disinfection Guidance Manual                6-19                              November 2006
For the Final LT2ESWTR

-------
                             6. Start-up and Operation of UV Facilities
       4.  Place the lamp with the lowest UV intensity value in the position nearest the UV
          sensor for routine operation
6.3.2.3    Reference UV Sensors

        Accurate UV sensors are necessary to verify adequate UV dose delivery during
operation. Two types of UV sensors are available: duty and reference. Duty UV sensors are on-
line sensors that continuously monitor UV intensity. Reference UV sensors are off-line sensors
used to assess the duty UV sensor performance. Both types of UV sensors need to be maintained.
Monitoring of duty UV sensor calibration is described in Section 6.4.1.1.

       The reference UV sensor should be calibrated at least once per year at a qualified facility
(e.g., manufacturer) to confirm that it is calibrated properly.  The reference UV sensor should be
calibrated against a traceable standard. For example, UV manufactures are currently using NIST,
NPL, ONORM, and DVGW standards. The reference UV sensor should be exposed to UV light
for a period no longer than necessary to perform the UV intensity measurement. When not in
use, the reference UV sensor should be stored under conditions that will maintain its integrity
and accuracy as recommended by the manufacturer. Some PWSs may choose to have multiple
reference UV sensors to help determine if one reference UV sensor is out of calibration, as a
replacement reference UV sensor, or to allow multiple duty UV sensors to be checked
simultaneously. Having multiple reference sensors is helpful if the reference and duty sensor
measurements do not match because the operator can easily  determine which one is in error. If
the reference UV sensor is found to be out of calibration, the period between calibrations should
be decreased.
6.3.2.4    Electrical Concerns

       Typically, power to the UV reactors is provided via a distribution transformer, a circuit
breaker, a disconnect switch at the UV reactor, and related wires and conduits. If maintenance on
the control panel is necessary, the main electrical supply should be disconnected and the PWS's
safety procedures should be followed.

       The power to the lamps is typically delivered through individual GFI circuit breakers and
ballasts. The GFI breakers should be test-tripped at least once per year and should be maintained
in accordance with the manufacturer's recommendations. Ballast output should be monitored
through the UV reactor's control panel. Irregularities or instabilities in ballast output may
indicate a problem with  the electrical feed or the ballast itself.

       A ballast cooling system is normally provided with LPHO and MP reactors to maintain
the ballast temperature below the maximum specified limit. LP reactors typically do not need
ballast cooling. This cooling system should be inspected and maintained as recommended by the
manufacturer.
UV Disinfection Guidance Manual                6-20                               November 2006
For the Final LT2ESWTR

-------
                            6. Start-up and Operation of UV Facilities
6.3.2.5   UV Reactor Temperature and Water Level

       The water temperature or water level in the reactor should be monitored because UV
lamps may break if they become overheated (Appendix E). The thermometer and/or water level
monitor should be visually inspected and replaced at the manufacturer's recommended
frequency. Reactor temperature monitoring and/or water level monitoring are typically included
in the packaged control systems for MP reactors, although they may not be included in packaged
control systems for LP and LPHO reactors (due to their much lower operating temperatures).
6.3.2.6   UV Lamp Replacement

       UV lamp output decreases over time. UV lamps therefore should be replaced periodically
to maintain sufficient UV dose delivery. Lamp manufacturers should provide documentation of
lamp output decay characteristics and guaranteed life. This information will help the PWS
determine the lamp replacement frequency.

       The frequency of UV lamp replacement can be based on a PWS-determined schedule,
lamp operating hours, or the UV intensity or validated dose reduction. During replacement, the
lamps and sleeves should be handled in accordance with manufacturer recommendations, using
clean cotton, powder-free latex, or vinyl gloves, because fingerprints can inhibit proper
operation.

       Because spent UV lamps contain mercury, they are usually considered hazardous waste
under Subtitle C of the Resource Conservation and Recovery Act (RCRA) (40 CFR parts 260,
261, 264, and 273). Expended lamps should therefore be sent to a mercury recycling facility
where the mercury is  recovered and lamp components are recycled. Some UV reactor and lamp
manufacturers will accept spent or broken lamps for recycling or proper disposal (Dinkloh 2001,
Lienberger 2002, Gump 2002). PWSs should contact the UV manufacturer to determine if they
accept spent lamps, or contact their state or local agencies for a list of local mercury recycling
facilities.

       Replacement lamps should be identical to those used during reactor validation with
respect to arc length, internal and external diameter, spectral output, and placement within the
quartz  sleeve. If the supplied lamps are not equivalent to the lamps used during validation, the
UV reactor is not operating as validated and will be considered off-specification. The
manufacturer should provide independent data verifying the lamp aging curve over the entire
lamp life to show that the new lamps are equal to or better than the validated lamps. However, if
a PWS replaces the lamps with higher power lamps to receive higher log inactivation credit,
validation testing should be completed to confirm performance.
6.3.2.7   Lamp Sleeves

       Lamp sleeves degrade over time due to solarization (Section 2.4.4) and internal sleeve
fouling, resulting in cloudiness and loss of UV transmittance. Abrasion of the sleeve surface
during handling or mechanical cleaning may also contribute to the loss of UV transmittance.
Reduced sleeve transmittance loss is reflected in the UV sensor reading and, therefore, does not

UV Disinfection Guidance Manual               6-21                              November 2006
For the Final LT2ESWTR

-------
                            6. Start-up and Operation of UV Facilities
have to be monitored. However, a low UV sensor reading may be due to reduced sleeve
transmittance and should be considered when troubleshooting this problem (as discussed in
Section 6.6.2).

       Sleeves should be replaced when damage, cracks, or staining diminish UV intensity to
the point where the minimum validated intensity level or validated dose cannot be met. Sleeves
in MP equipment should typically be replaced every 3 to 5 years, although sleeves in LP or
LPHO equipment may not need to be replaced as frequently. This replacement frequency should
be increased or decreased based on operational experience. Replacement sleeves should be
identical  to the sleeves used during validation in terms of length, inside and outside diameter,
and UV transmittance, and should meet the design and UV manufacturer's material and
construction specifications. If the replacement sleeves differ from those used in validation, UV
dose delivery and UV sensor modeling can be used to assess the impact of the changes as
described in Section 5.13.

       The sleeves should be handled in accordance with manufacturer recommendations, using
clean cotton, powder-free latex, or vinyl gloves because fingerprints can damage the sleeves
during operation. When sleeves are replaced, the manufacturer's procedure should be closely
followed because the lamp sleeve can crack and break from over-tightening of the compression
nuts that  hold it in place.
6.3.2.8   On-line UVT Analyzer

       On-line UVT analyzers should be cleaned and maintained according to the UV
manufacturer recommendation. On-line UVT analyzer calibration is evaluated periodically as
part of compliance monitoring (Section 6.4.1.2).
6.3.3  Spare Parts

       The actual life of a component is a function of many variables, including operating
conditions, maintenance practices, the quality of the construction materials, and fabrication
practices. Consequently, estimating the actual life of every component is impossible. To
overcome the operational impacts of this uncertainty, an adequate inventory of critical spare
parts should be maintained to ensure reliable and consistent performance of the UV equipment
and to avoid the delivery of off-specification water.

       All UV equipment components have both a design life and a guaranteed life. The design
life is the expected duration of operation. The guaranteed life incorporates the risk, assumed by
the manufacturer, to account for the uncertainties associated with the quality of materials used,
production, and operating conditions. Generally, guarantees are conditional and are valid under
specified operating conditions. For example, guaranteed lamp life is normally linked to the lamp
power setting or the number of on/off cycles per 24-hour period. If equipment failure occurs
during the warranty period and if all of the warranty conditions are satisfied, the manufacturer
will typically replace the component and charge the owner a prorated fee for the use of the
replaced component.
UV Disinfection Guidance Manual                6-22                              November 2006
For the Final LT2ESWTR

-------
                            6. Start-up and Operation of UV Facilities
       Table 6.5 provides typical design and guaranteed lives for major UV reactor components.
These represent current industry trends at the time of publication and are likely to change as
more O&M information becomes available and technological advances occur. Manufacturers
should be contacted directly for details specific to their equipment.
            Table 6.5. Typical Design and Guaranteed Lives of Major UV
                   Components (Based on Manufacturers' Input)
Component
Low-pressure Lamps (LP And LPHO)
MP Lamps
Sleeve
Duty And Reference UV Sensors
UVT Analyzer
Cleaning Systems
Ballasts
Design Life1
12,000 hours
8,000 hours
8-10 years
3 -10 years
3-5 years
3-5 years
10-15 years
Guaranteed Life 2
8,000 -12,000 hours
4,000-8,000 hours
1 - 3 years
1 year
1 year
1 - 3 years
1 - 5 years
        Expected duration of operation
        Accounts for variability of material quality, production, and operating conditions


       Following is a suggested minimum inventory of spare parts, expressed as a percentage of
the installed number. The complete list of spare parts will vary depending on the specific
equipment installed and should be coordinated with the UV manufacturer. The number of spare
parts needed depends on the guaranteed life of the specific equipment. For example, a higher
percentage of spare MP lamps may be appropriate compared to LP lamps because they need to
be replaced more frequently.

       .   UV lamps - 10 percent with a minimum of two lamps
       .   Sleeves - 5 percent with a minimum of one sleeve
       .   O-ring seals - 5 percent with a minimum of two seals
       .   OMC or OMCC wipers - 5 percent with a minimum of two wipers
       .   OMC or OMCC wiper drive mechanisms - 2 percent with a minimum of one drive
       .   Ballasts - 5 percent with a minimum of one unit
       .   Ballast cooling fan - 1 unit
       .   Duty UV sensor - minimum of 2 units (adjust number based on operating experience)
       .   Reference UV sensor - 2 units (more may be needed if wet duty UV sensor are used
          as described in Section 6.4.1.1)
          On-line UVT analyzer - 1 unit (if used for dose-monitoring strategy)
UV Disinfection Guidance Manual
For the Final LT2ESWTR
6-23
November 2006

-------
                            6. Start-up and Operation of UV Facilities
6.4    Monitoring and Recording of UV Facility Operation

       This section discusses the required and recommended monitoring and recording activities
for UV facilities. PWSs should always contact their state to identify any state-specific
monitoring and reporting requirements and determine when violations of these reporting
requirements would occur.
6.4.1  Monitoring and Recording for Compliance Parameters

       PWSs must monitor their UV reactors to determine if the reactors are operating within
validated conditions. This monitoring must include UV intensity as measured by a UV sensor,
flow rate, lamp status, and other parameters designated by the state [40 CFR 141.720 (d)(3)]. UV
reactors should also be regularly monitored to diagnose operating problems, determine when
maintenance is necessary, and maintain safe operation. In addition to monitoring operational
parameters, PWSs must verify the calibration of UV sensors in accordance with a protocol that
the state approves [40 CFR 141.720 (d)(3)]. This section describes the requirements for each of
these items.

       Because UVT is a critical parameter for the Calculated Dose Approach, EPA believes
that calibration of UVT analyzers is necessary to determine if reactors are operating within
validated conditions. Therefore, this section also includes a discussion of calibration of UVT
analyzers.
6.4.1.1   Monitoring of Duty UV Sensor Calibration

       Manufacturers will calibrate the UV sensors prior to installation. However, over time the
UV sensors will drift out of calibration. Because UV sensors are vital to assessing disinfection
performance, water systems must verify the calibration of UV sensors with a protocol that the
state approves [40 CFR 141.720 (d)(3)]. If a UV reactor is turned on and the calibration of the
UV sensors has not been verified, the UV reactor is operating off-specification.

       EPA recommends that calibration  of UV sensors be verified with a reference UV sensor
at least monthly. As noted in Section 6.3.2.3, reference UV sensors are off-line UV sensors that
should be at least as accurate as the duty UV sensors and should be constructed identically (with
any exceptions to the reference sensor to make it more accurate).

       Water systems should designate in their protocol whether only the UV sensors in use will
be monitored, or if all duty  and standby sensors will be monitored to confirm calibration.
Verifying calibration of all  duty and stand-by UV reactors has the advantage of rendering all UV
sensors ready for use at any time if they are  needed.

       This section describes the recommended procedure to verify UV sensor calibration and
the options available if the duty UV sensor fails the recommended calibration criterion.
Section 6.6 supports this section by presenting a flowchart of the calibration check procedure to
facilitate decisions if the duty UV sensor fails  the calibration criterion.
UV Disinfection Guidance Manual                6-24                               November 2006
For the Final LT2ESWTR

-------
                              6. Start-up and Operation of UV Facilities
Duty UV Sensor Calibration Evaluation Procedure

       To assess the calibration, the following protocol should be followed:

       1. Because the calibration of the UV sensor is sensitive to the power level of the lamps
          (Swaim et al. 2002), set the lamp power to the level typically used during routine
          operation.

       2. Measure the UV intensity with the duty UV sensor and record the measurement
          result.

       3. Replace the duty UV sensor used in Step 2 with the reference UV sensor in the same
          location (i.e., port).

       4. Measure and record the reference UV sensor measurement.

       5. Calculate the UV sensor calibration ratio (Equation 6.1). If desired, Steps 2-5 can be
          repeated, and a mean calibration ratio can be calculated.

                              f S   ~\
          Calibration Ratio =    D"^                                             Equation 6.1
                              I S^ J

              where:
               Suuty =   Intensity measured with the duty UV sensor (mW/cm2)
                SRef =   Intensity measured with the reference UV sensor (mW/cm2)

       6. Determine if the UV sensor calibration criterion (Equation 6.2) is met for the two UV
          sensor readings or the mean calibration ratio.

          Calibration Ratio < 1.2(see footnote 2)                                     Equation 6.2

       7. If the relationship in Equation 6.2 does not hold true, verify that the reference UV
          sensor is accurate with a different reference UV sensor (i.e., verify that the duty UV
          sensor truly failed the calibration check) by inserting a second reference UV sensor
          and repeating Steps 3-6. If a second reference UV sensor is unavailable, the sensor
          calibration can be checked against two duty sensors (as opposed to another reference
          sensor).

       8. If Step 7 confirms the duty UV sensor is out of calibration, replace the duty UV
          sensor with a calibrated UV sensor or apply a UV sensor correction factor (described
          after Example 6.3).
2 This calibration ratio is higher than the ratio recommended for validation testing (1.1, or 10%, as presented in
  Section 5.5.4). A recommended calibration ratio of 1.2 during operations is based on experience with existing UV
  equipment during routine operations.
UV Disinfection Guidance Manual                  6-25                                November 2006
For the Final LT2ESWTR

-------
                             6. Start-up and Operation of UV Facilities
       9.  If a duty UV sensor was replaced, check the replaced UV sensor one hour later by
          repeating steps 2-6 (or based on UV manufacturer's recommendation) to confirm that
          the replaced duty UV sensor is operating properly.

Issues to Consider when Monitoring UV Sensor Calibration

       The above UV sensor criteria allow the UV facility to operate out of calibration if duty
sensor reads conservatively low values compared to the reference sensor. Operating in this
manner is not energy efficient, however, and the PWS would benefit from having the UV sensor
recalibrated.

       When re-inserting a duty UV sensor, the rotational alignment of the UV sensor within the
UV sensor port can affect its sensitivity. This effect may be due to the UV sensor configuration
(e.g., acceptance angle). The UV sensors should be rotated until the lowest UV intensity reading
is obtained for routine monitoring purposes with the UV sensor completely inserted into the UV
sensor port. This affect may not be an issue if the UV sensor is keyed in place or another method
is used to  prevent adjusting the alignment of the sensor.

       Wet UV sensors are in direct contact with the water; therefore, the water in the UV
reactor needs to be drained before the duty sensors are replaced with reference sensors (Step 3
above). To reduce the number of times the UV reactor needs to be drained, PWSs should own at
least the same number of reference sensors as the duty UV sensors in one UV reactor. For
example, a UV reactor has six duty wet sensors in each UV reactor; therefore, the PWS owns a
minimum of 6 reference UV sensors to reduce the number of times the UV reactor has to be
drained during the calibration check procedure.
            Example 6.3. Duty UV Sensors are Verified using Reference Sensor
                      (Corresponds to Example 6.1 in Section 6.1.4)

       System X has one duty UV reactor and one standby reactor. Each reactor has two banks
of four 200-W LPHO lamps with one germicidal UV sensor per bank (i.e., two UV sensors per
reactor). The data from a monthly calibration check as presented below show that all of the UV
sensors meet the UV sensor calibration criterion. Therefore, the duty UV sensors are in
calibration, and no further action is necessary for the UV sensors this month.
Reactor
Number
1
1
2
2
Bank
Number
forUV
Sensor
1
2
1
2
Duty UV
Sensor
Reading
(mW/cm2)
13.4
12.6
11.9
15.2
Reference UV
Sensor
Reading
(mW/cm2)
14.6
11.8
12.5
13.7
Calibration Ratio
| ^Duty |
UeJ
0.9
1.1
1.0
1.1
Within UV Sensor
Calibration Criterion?
Ac A
SDuty , ~
	 < 1.2
lSRef )
Yes
Yes
Yes
Yes
UV Disinfection Guidance Manual
For the Final LT2ESWTR
6-26
November 2006

-------
                             6. Start-up and Operation of UV Facilities
Use of UV Sensor Correction Factor

       A failed duty UV sensor should be replaced with a calibrated duty UV sensor or the UV
reactor is off-specification (if operated). However, replacement may not be an option if multiple
UV sensors fail and/or no additional UV sensors are immediately available. PWSs that cannot
immediately replace a duty UV sensor that failed the UV sensor calibration criterion
(Equation 6.2) should implement a UV sensor correction factor (CF). In this approach, a CF is
selected and applied to either the intensity setpoint or required dose setpoint (depending on the
dose-monitoring strategy) for the affected UV reactor(s). Operating with a CF is not energy
efficient; however, this method enables the UV facility to remain in operation while the UV
sensor problem is resolved. The selected CF should not be changed until the failed UV sensors
are replaced with factory calibrated UV sensors  This approach is not recommended for long-
term operation, and the UV sensor problem should be resolved as quickly as possible.

       The specific steps for the UV sensor CF approach are summarized below:

       1 . Use the calibration data to determine the correction factor for each failed UV sensor
          (Equation 6.3). Note that twenty percent is subtracted from the calibration ratio to
          account for the acceptable UV sensor error of 20 percent (i.e., Equation 6.2 shows an
          allowable error of 20 percent). For example, if SDuty =  138 W/m2 and SRef= 100
          W/m2, the calibration factor is 1.18.

                        SDu
          Sensor CF =  Duty - 0.2 |                                           Equation 6.3
       2.  Determine the maximum Sensor CF for the failed UV sensors (Equation 6.3)
          (Example 6.4 below presents an example of how to select a Sensor CF).

       3 .  Multiply the UV intensity setpoint or the required dose (depending on the dose-
          monitoring strategy) by the UV sensor CF to determine the corrected setpoint or
          required dose (Equations 6.4 and 6.5) that account for the UV sensor errors.

          Corrected UV intensity setpoint =  UV intensity setpoint x Sensor CF    Equation 6.4

          Corrected required dose = DReq x  Sensor CF                          Equation 6.5

       4.  The sensor CF and the corrected UV intensity setpoint or corrected DReq setpoint
          should be included in the report to the state for the affected reactor(s). These
          corrected setpoints are now the basis for off-specification operation until the UV
          sensor calibration problem is resolved.

       5.  If the failed UV sensor(s) has not been replaced before the next monthly calibration
          check, the UV reactors with the corrected setpoints should use Equation 6.6 to
          evaluate whether any  sensor exceeds the current Sensor CF. The Sensor CF should be
          increased if in any UV sensors fail the previous  month's CF, as described in Equation
UV Disinfection Guidance Manual                6-27                               November 2006
For the Final LT2ESWTR

-------
                              6. Start-up and Operation of UV Facilities
            S-
              Duty
             S-
                     UV Sensor CF+0.2
                                    Equation 6.6
              Ref
       Example 6.4 shows how a hypothetical water system addressed calibration checks that
result in multiple UV sensors that are out of calibration.
IT
sy
Tl
Example 6.4. Duty UV Sensors that Do Not Meet Calibration Criteria
(Corresponds to Example 6.2 in Section 6.1.4)
System Y has two duty reactors and one standby reactor. Each reactor has six germicidal
V sensors. System Y developed a sensor calibration protocol whereby on a monthly basis,
stem operators verify that sensors are calibrated using Equation 6.2 of this guidance manual.
leir protocol was approved by the state.
Data from the UV sensor calibration check for the month of March are presented below:
Reactor
Number
1
1
1
1
1
1
2
2
2
2
2
2
3
3
3
3
3
3
Duty UV
Sensor
Reading
(mW/cm2)
259.4
303.8
284.1
400.5
263.2
258.2
368.7
404.1
287.9
299.8
321.3
265.4
379.6
357.3
258.2
565.5
244.4
238.9
Reference
UV Sensor
Reading
(mW/cm2)
247.8
268.5
303.5
387.1
258.9
266.6
250.6
311.5
314.2
214.9
287.4
347.5
284.6
303.9
281.5
321.3
147.7
268.1
Calibration
Ratio
(^}
(SM)
1.1
1.1
0.9
1.0
1.0
1.0
1.5
1.3
0.9
1.4
1.1
0.8
1.3
1.2
0.9
1.8
1.7
0.9
Within
Calibration?
(? \
D<"y Ui on
10
Yes
Yes
Yes
Yes
Yes
Yes
No
No
Yes
No
Yes
Yes
No
Yes
Yes
No
No
Yes
Correction
Factor
f SDuty Q „ |
I5*/ ' J
NA
NA
NA
NA
NA
NA
1.3
1.1
NA
1.2
NA
NA
1.1
NA
NA
1.6
1.5
NA
Duty UV
Sensor
Replaced?
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
Yes
Yes
No

UV Disinfection Guidance Manual
For the Final LT2ESWTR
6-28
November 2006

-------
                             6. Start-up and Operation of UV Facilities
           Example 6.4. Duty UV Sensors that Do Not Meet Calibration Criteria
                                       (continued)

        Six sensors failed calibration with calibration ratios between 1.3 and 1.8; however,
 System Y has only three spare duty UV sensors. The two worst UV sensors were replaced (i.e.,
 the UV sensors with a calibration ratio of 1.8 and 1.7), and one of the spare UV sensors was
 retained as a back-up, leaving four UV sensors that failed the calibration criterion.  System Y
 applied the UV sensor CF approach to enable their facility to continue operating until the UV
 sensor problem could be resolved with the manufacturer.

        System Y applied the CF to the individual reactors, not the entire UV facility. For
 Reactor 2 the highest calibration ratio was 1.5, resulting in a CF of 1.3 (using Equation 6.3).
 The highest calibration ratio for Reactor 3 (after the two sensors with the calibration ratios of
 1.7 and  1.8 were replaced) was 1.3, giving a CF of 1.1.

        System Y's required dose is 5.8 mJ/cm2 for 2.0 log inactivation ofCryptosporidium.
 The corrected required doses for Reactors 2 and 3 are as follows:

        .   The corrected required dose for Reactor 2 is 7.5 mJ/cm2 (5.8 mJ/cm2 multiplied by a
           CFofl.3).

        .   The corrected required dose for Reactor 3 is 6.4 mJ/cm2 (5.8 mJ/cm2 multiplied by a
           CFof 1.1).

        System Y maintained a validated dose (i.e., the calculated dose from the dose-
 monitoring equation divided by the Validation Factor) of 7.5 mJ/cm2 and 6.4 mJ/cm2for
 Reactors 2 and 3 respectively, until the four duty UV sensors were replaced the following week.
 If the validated dose had fallen below the corrected required dose, the reactors would have been
 off-specification. Any off-specification events and the volume of water treated during the event
 must be reported to the state as described in Section 6.5.

        In this example, System Y applied a correction factor for two UV reactors with sensors
 that failed calibration. Another option for System Y would have been to move all of the UV
 sensors that require a CF to one UV reactor (i.e., switching out UV sensors between Reactors 2
 and 3).  In this case, the CF would have only been applied to  one of the UV reactors instead of
 both Reactors 2  and 3.
6.4.1.2   Monitoring of UVT Analyzer Calibration

       Compliance monitoring of UVT analyzer calibration is required only when UVT is an
integral part of the dose-monitoring strategy, such as with the Calculated Dose Approach. If the
UV Intensity Setpoint Approach is used, UVT analyzer calibration checks are not required
because UVT is not used to verify UV dose delivery (Section 6.4.1.4).
UV Disinfection Guidance Manual                6-29                               November 2006
For the Final LT2ESWTR

-------
                             6. Start-up and Operation of UV Facilities
       EPA recommends that on-line UVT analyzers be evaluated at least weekly by comparing
the on-line UVT measurements to UVT measurements using a bench-top spectrophotometer. The
bench-top spectrophotometer should be maintained and calibrated at the frequency required by
the manufacturer. The calibration monitoring frequency should be decreased or increased based
on the performance demonstrated over a one-year period if approved by the state. For example,
the frequency could be reduced to once per month if the UVT analyzer is consistently within the
allowable calibration error for more than a month during the first year of monitoring.

       To monitor the calibration, the following UVT calibration check protocol should be
followed:

       1.  Record the reading of the on-line UVT analyzer (UVTon_iine).

       2.  Collect a grab sample from a location close to the on-line UVT analyzer sampling
          point.

       3.  Measure the UVT of the grab sample on a calibrated bench-top spectrophotometer
          (UVTbench).

       4.  Compare the on-line UVT (UVTon_iine) reading to the bench-top spectrophotometer
          UVT reading using Equation 6.7.

               UVTon - i,ne(%)- UVTbend{%)  < 2 percent UVT3                    Equation 6.7

       5.  Recalibrate the on-line UVT analyzer if Equation 6.7 is not met. If the UVT analyzer
          is not recalibrated, the UV facility is operating off-specification unless mitigation
          steps are taken.

       If recalibration is necessary in four consecutive weeks, water system operators should
check the calibration daily for 1 week to determine the rate of calibration decay (i.e., the amount
the UVT analyzer drifts from the UVTbench per day over the week period). Use these data to
establish a more frequent recalibration frequency that will enable the on-line UVT analyzer to
stay within the acceptable calibration error. If these data indicate that calibration cannot be
maintained for at least 24 hours, water systems should consider one of the two options described
below. The UV facility is off-specification until one of these options is followed or until the
UVT  analyzer meets the criterion shown in Equation 6.7.

          Option 1 - Take manual UVT  measurements with a calibrated bench-top
          spectrophotometer every 4 hours and enter the UVT into the PLC. The UVTbench
          entered should be used for the  following 4 hours in the monitoring strategy.

          Option 2 - Enter the design UVT value into the PLC and verify daily that the design
          UVT does not exceed the actual UVT with a grab sample.
  The absolute value of the difference between the UVT analyzer and bench measurement should be used because
  both conservative and non-conservative UVT errors can cause inaccuracies with the dose monitoring strategy.
UV Disinfection Guidance Manual                6-30                               November 2006
For the Final LT2ESWTR

-------
                            6. Start-up and Operation of UV Facilities
       Although these options allow the UV facility to continue operating if the calibration
error is exceeded while the on-line UVT analyzer is being repaired or replaced, these
options are not intended for long-term operation. These options should not be employed for
longer than six months.

       Example 6.5 shows how a hypothetical water system confirmed that they met the
calibration criteria listed above.
                     Example 6.5. UVT Analyzer Calibration Check
                      (Corresponds to Example 6.2 in Section 6.1.4)

       System Y has a UVT analyzer that shows a UVTon_iine of 93.5 percent. The PWS took a
 grab sample from the influent line to the UVT analyzer and brought it back to the laboratory.
 The sample was analyzed for UV absorbance at 254 nm (A254) using a bench-top
 spectrophotometer that has been properly calibrated. The sample A254 is 0.032 cm"1. The grab
 sample A254 was converted to UVT using Equation 2.2, which yields a UVTbench value of 92.9
 percent. The absolute difference between the on-line reading and bench spectrophotometer
 reading was 0.6 percent UVT. This difference was within the calibration error range of 2
 percent UVT, and the UVT analyzer did not need to be recalibrated.
6.4.1.3   Off-specification Events

       Off-specification operation occurs when the UV facility operates outside of the validated
limits (Section 6.1.4), a UV sensor is not in calibration (Section 6.4.1.1), the UVT analyzer is not
in calibration (Section 6.4.1.2) (and it is part of the dose-monitoring strategy), or UV equipment
is not equivalent or better than the equipment validated.

Validated Parameters

       PWSs must monitor each reactor to determine whether it is operating within validated
conditions [40 CFR 141.720(d)(3)]. The validated parameters to monitor depend on the dose-
monitoring strategy used and the validation results. Table 6.6 presents the monitoring parameters
for the monitoring approaches and their off-specification triggers.

Calibration of VV Sensors

       A UV reactor is producing off-specification water if all three of the following conditions
occur:
       1.  Any of the duty UV sensors did not meet the calibration criteria in the state-approved
          protocol (Section 6.4.1.1) and
       2.  The duty UV sensors were not replaced with calibrated duty UV sensors and
       3.  UV sensor correction factor was not applied.
UV Disinfection Guidance Manual                6-31                              November 2006
For the Final LT2ESWTR

-------
                            6. Start-up and Operation of UV Facilities
       Table 6.6. Off-specification Examples for Each Monitoring Approach
Dose-
monitoring
strategy
UV Intensity
Setpoint
Approach
Calculated Dose
Approach
Parameters Monitored
UV intensity, flow rate,
lamp status
calculated dose, VF,
validated dose, flow rate,
UVT, lamp status
Off -specification Examples
1 . UV intensity below minimum value
2. Flow rate above validated limit
1 . Validated dose below DReq
2. Flow rate above validated limit
3. UVT below minimum value
VF = Validation factor
DReq = Required UV dose (Table 1.4)
Calculated Dose
                        VF

Calibration of UVT Analyzers

       Similarly, the UV facility is off-specification if the UVT analyzer is found to be out of
calibration and the remedial actions described in Section 6.4.1.2 are not completed.

UV Equipment Components

       The LT2ESWTR requires that water systems use reactors that have undergone validation
testing [40 CFR 141.720 (d)(2)]. It follows, therefore, that installed and replaced components
should be equal to or better than the components used during validation testing. If not, the UV
facility is off-specification unless the UV equipment is re-validated. The need for re-validation
and when the UV facility would be off-specification because of UV equipment components is
described in  Section 5.13.
6.4.1.4   Monitoring and Recording Frequency of Required Parameters

       The required dose-monitoring parameters (flow rate, UV intensity, number of banks on,
etc.) should be continuously monitored (i.e., at least every 5 minutes) for each UV reactor, and
these values should be recorded at least once every 4 hours. Very small systems (e.g., systems
serving fewer than 500 people) that cannot record reactor status every 4 hours (e.g., manual
recording is practiced) could consider a reduced recording frequency; however, the frequency
should not be less than once per day and should be discussed with the state.

       All water systems should record off-specification alarms at a minimum of 5-minute
intervals until the alarm condition has been corrected. The off-specification volume will start as
soon as the flow is found to be outside of the validated range. The measurement of off-
specification volume will stop as soon as the flow is shown to be within the validated limits.

       The EPA recognizes that the off-specification event may begin before the off-
specification alarm is monitored. The off-specification event may also end before the off-
speciation alarm is cancelled and recorded. It is assumed that over time the underestimation of
off-specification water before the alarm is activated and the overestimation of off-specification
water before the alarm is cancelled will minimize any errors in the calculation ofoff-
UV Disinfection Guidance Manual
For the Final LT2ESWTR
6-32
November 2006

-------
                             6. Start-up and Operation of UV Facilities
specification water. If a facility monitors more frequently than the minimum recommended 5-
minute intervals, the off-specification volume will start as soon as the reactor is monitored as
operating off-specification and the off-specification volume will stop as soon as the reactor is
monitored as being on-specification. More frequent off-specification alarm monitoring may more
accurately account for the off-specification volume.

       These off-specification alarm records should be used to determine the percentage of flow
volume that is off-specification. The compliance with the off-specification limits is based on the
off-specification percentage for the UV facility, not for individual reactors. The monitoring
guidelines are summarized in Table 6.7, and Example 6.6 illustrates the routine and off-
specification recording recommendations.
                 Table 6.7. Recommended Recording Frequency for
                           Required Monitoring Parameters
Parameter
Off-specification
Alarm
UV Intensity
UVT1
Validated Dose1
Lamp Status
Flow Rate
Production Volume
Calibration of UV
Sensors
Calibration of On-
line UVT Analyzer1
Recommended
Recording Frequency
Minimum of every 5
minutes
Every 4 hours
Every 4 hours
Every 4 hours
Every 4 hours
Every 4 hours
Off-specification events
and monthly total
Monthly
Weekly2
Notes
Recording should continue until the alarm condition
has been corrected.
The UV intensity must be greater than or equal to the
validated setpoint.
The UVT must be greater than or equal to the
minimum UVT validated.
The validated dose must be greater than or equal to
the DReq.
Lamps should be energized if water is flowing through
the UV reactor.
The flow rate should be less than or equal to the
maximum flow tested in validation.
The production volume needs to be recorded so the off-
specification compliance calculation can be completed.
The calibration of the UV sensor should be monitored
as described in Section 6.4.1 .1 .
The calibration of the UVT analyzer should be
monitored as described in Section 6.4. 1. 2. 1
  Required only if necessary for the dose-monitoring strategy (i.e., the Calculated Dose Approach).
 2
  Frequency could be reduced as described in Section 6.4.1.2.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
6-33
November 2006

-------
                            6. Start-up and Operation of UV Facilities
                  Example 6.6 Routine and Off-specification Recording
                      (Corresponds to Example 6.2 in Section 6.1.4)

       This example illustrates System Y's daily monitoring and recording of UV equipment
 operation to verify that it is operating within validated limits. The System Y has two duty
 reactors and one standby. Reactor 1 is used only for part of the day; Reactor 2 is used 24 hours
 a day; and Reactor 3 is off-line in the 24-hour period. System Y monitors the off-specification
 alarms every 5 minutes, which is the minimum recommended.

       At 1:08 PM the flow at System Y went above the validated range because an upstream
 filter was taken off-line for backwashing. At 1:10 PM, the flow through Reactor 2 was
 recorded as being above the validated limit of 10 mgd as an off-specification alarm. This
 resulted in the reactor operating off-specification while the flow split between the reactors was
 adjusted. The flow returned to within the validated range at 1:17 PM when the backwashed
 filter was placed back on-line.  System Y recorded that the off-specification alarm was remedied
 at 1:20 PM

       The off-specification recording started when the first off-specification alarm occurred
 (1:10 PM) and continued at 5-minute intervals until the reactor was monitored as being on-
 specification again (1:20 PM) when the data recording reverted back to every 4 hours. This
 event is illustrated in the table below. During this 24-hour period, no other off-specification
 events occurred. If System Y monitors the  off-specification alarms at 1 minute intervals, the
 off-specification operation would have been more accurately recorded.
UV Disinfection Guidance Manual                6-34                              November 2006
For the Final LT2ESWTR

-------
                                6. Start-up and Operation of UV Facilities
Example 6.6 Routine and Off-specification Recording (continued)
Monitoring
Time
12:00 AM

4:00 AM

8:00 AM

12:00 PM

1:05 PM
1:10PM
1:1 5PM
1:20 PM

4:00 PM

8:00 PM

12:00 AM
Daily total off-
specification
events
Reactor 1
Reactor
Status
Off
Off
Off
Off
On-
specification
On-
specification
On-
specification
On-
specification
On-
specification
On-
specification
On-
specification
On-
specification
On-
specification
On-
specification
On-
specification
On-
specification
On-
specification
On-
specification

Data
Recorded
Off-line
None
Off-line
None
On-
specification
None
On-
specification
None
None
None
None
None
None
On-
specification
None
On-
specification
None
On-
specification
0 events
Reactor 2
Reactor
Status
On-
specification
On-
specification
On-
specification
On-
specification
On-
specification
On-
specification
On-
specification
On-
specification
On-
specification
Off-
specification
Off-
specification
On-
specification
On-
specification
On-
specification
On-
specification
On-
specification
On-
specification
On-
specification

Data
Recorded
On-
specification
None
On-
specification
None
On-
specification
None
On-
specification
None
None
Off-
specification
Off-
specification
On-
specification
None
On-
specification
None
On-
specification
None
On-
specification
1 event
lasting 10
minutes
Reactor 3
Reactor
Status
Off
Off
Off
Off
Off
Off
Off
Off
Off
Off
Off
Off
Off
Off
Off
Off
Off
Off

Data
Recorded
Off-line
None
Off-line
None
Off-line
None
Off-line
None
None
None
None
None
None
Off-line
None
Off-line
None
Off-line
0 events
Note shaded areas indicate data that were recorded.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
6-35
November 2006

-------
                            6. Start-up and Operation of UV Facilities
Example 6.6 Routine and Off-specification Recording (continued)
The off-specification volume of water must be determined for compliance reporting
(Section 6.5). The table below provides the flow and volume monitoring and recording for this
example day. The volume was calculated using a flow totalizer in the PLC programming for
each 5-minute off-specification period.
Monitoring Time
12:00 AM

4:00 AM

8:00 AM

12:00 PM

1:10PM
1:1 5PM
1:20 PM

4:00 PM

8:00 PM

12:00 AM
Total Daily Off-
specification volume (gal)
Total Daily Volume (gal)
Reactor 1
Flowrate1
(mgd)
0

0

7.2

7.4

NRJ
NRJ
NRJ

7

7.2

7.3


Volume2
(gal)
-

-

1,120,254

1,225,897

NRJ
NRJ
NRJ

1,102,564

1,025,951

1,159,951
-
5,634,617
Total Daily
Volume
(gal)
-

-

1,120,254

2,346,151





3,448,715

4,474,666

5,634,617


Reactor 2
Flowrate1
(mgd)
9.2

9

9.3

9.5

12.2
11.3
9.7

9.4

9.2

9.3


Volume2
(gal)
1,526,782

1,358,972

1,534,682

1,510,036


38,452
37,522

1,551,123

1,520,321

1,536,987
75,974
9,012,121
Total Daily
Volume
(gal)
9,144,269

1,358,972

2,893,654

4,403,690

4,870,357
4,908,809
4,946,331

5,954,813

7,475,134

9,012,121


Note shaded areas indicate data that were recorded.
1 Maximum flow rate was recorded to show the flow was within validated limits.
2 Volume was estimated in the PLC using the flow rate.
3 NR indicates that data were not recorded
       Example 6.6 is based on the flow rate's increasing beyond the validated range. Off-
specification recording would follow the same procedure for any problem resulting in off-
specification time (e.g., UV sensor failure or the UVT decreased beyond the validated range).
6.4.2  Monitoring and Recording for Operational Parameters Not Related to
       Compliance

       To minimize operational problems, facilitate regulatory compliance, and evaluate UV
reactor performance, parameters in addition to those required for regulatory compliance should
be monitored. Table 6.8 presents these suggested parameters and the recommended recording
frequency. These parameters and their monitoring frequency should be adjusted based on site-
UV Disinfection Guidance Manual
For the Final LT2ESWTR
6-36
November 2006

-------
                            6. Start-up and Operation of UV Facilities
specific operating experience. For example, if sleeve fouling is a maintenance issue and
supplemental cleaning is frequent (e.g., monthly), the fouling parameters should be monitored
daily as shown in Table 6.7 rather than weekly.
    Table 6.8. Recommended Monitoring Parameters and Recording Frequency
Parameter
Power Draw
Water Temperature
(Only Necessary for MP
Reactors)
UV Lamp On/Off Cycles
Turbidity (In Addition to
Monitoring Otherwise
Required Under Subpart
H)
pH, Iron, Calcium,
Alkalinity, Hardness,
ORP
UVT Analyzer
Calibration1
Operational Age2 of the
Following Equipment:
. Lamp
• Ballast
. Sleeve
. UV Sensor
Calibration of Flow
Meter
Monitoring
Frequency
Continuous
Continuous
Continuous
Daily
Weekly (reduce if
fouling is not
prevalent)
Weekly (reduce if
appropriate based
on operational
experience)
Monthly
Monthly
Recording
Frequency
Every 4 hours
Daily
Weekly (Total
cycles in a week)
Weekly
Weekly
Weekly
Monthly
Monthly
Notes
This information can be used to
determine the most energy efficient
operation strategies.
Monitoring is important to verify that the
high temperature limit is not exceeded
(often part of packaged UV control
system).
The number of on/off cycles can help
assess lamp aging.
Recommended only if chemicals (e.g.,
lime) are added prior to UV disinfection.
Monitoring may not be necessary for
many UV facilities.
These parameters will help assess
fouling issues if necessary.
This information can assist in planning
scheduled maintenance and the O&M
budget.
This information can assist in planning
scheduled maintenance and the O&M
budget.
This information can assist in planning
scheduled maintenance and the O&M
budget.
 Recommended if not being monitored as discussed in Section 6.4.1.2
2 Operational age is the amount of time the equipment has been operated (e.g., lamp hours)
6.5    UV Facility Reporting to the State

       Monthly reports must be prepared and submitted to the state (CFR 141.721). This section
describes the required reporting and provides example reporting forms.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
6-37
November 2006

-------
                            6. Start-up and Operation of UV Facilities
6.5.1  Required Reporting

       The report must include the percentage of off-specification water for the UV facility and
the UV sensor calibration monitoring [CFR 141.721(f)].

       The percentage of off-specification water should be calculated on a volume basis. The
percentage should be calculated by totaling the 5-minute off-specification alarm records and
associated volume released during those periods for each reactor. The volume released during the
off-specification event can be determined by:

       .   Using a flow totalizer that automatically records the volume when an off-
          specification event occurs.

       .   The PLC calculating the volume based on flow rate in one-minute (or shorter)
          intervals during the off-specification event.

       .   The PLC calculating the volume based on the maximum flow during the off-
          specification period multiplied by the length of time of the off-specification event.

       Off-specification time can be used a surrogate for off-specification volume only if the
flow is constant and this method is approved by the state.

       The total off-specification volume for all UV reactors should be divided by the total
volume produced by the UV facility that month and multiplied by 100 percent (See
Example 6.7). PWSs with constant flows may use off-specification time as an indicator for off-
specification volume (i.e., total off-specification divided by time in operation multiplied by 100
percent). SCADA and PLC interfaces can be designed to automatically calculate off-
specification based on the required monitoring, recording, and reporting.
UV Disinfection Guidance Manual                6-38                              November 2006
For the Final LT2ESWTR

-------
                            6. Start-up and Operation of UV Facilities
                       Example 6.7. Off-specification Computation
                     (Corresponds to Example 6.6 in Section 6.4.1.4)

       This example illustrates the computation of monthly percent off-specification operation
for System Y in Example 6.6. System Y had no other off-specification events this month;
therefore, the table in Example 6.6 captures all off-specification volume for the month. To
determine percent off-specification volume, monthly production volume totals were obtained
from the SCADA system. The table below shows the data used for the computation. In this
example, 0.02 percent (75,974 gal / 305,683,189 gal x 100%) of the volume of water System Y
treated was off-specification, which is within the allowable regulatory limit of 5 percent.
        Reactor
          No.
           1
         Totals
                   Monthly Total Off-
                   specification for UV
                        Facility
Time
 (hr)
0.00
                     0.17J
                     0.00
0.1 T
Volume
 (gal)1
           75,974
             0
 75,974
                   Monthly Total Production for
                           UV Facility
                                       Monthly
                                      Percent Off-
                                     specification
Time
 (hr)
             168
             720
 888
  Volume
   (gal)
           17,568,080
           288,115,109
                            0
305,683,189
          Off-specification volumes are from Example 6.6.
          Total monthly off-specification volume divided by total volume produced in the month
          multiplied by 100 percent
          Total monthly off-specification time was shown in Example 6.6 to be 10 minutes (0.17 hr).
       The percentage of UV sensors that were checked for calibration must be reported
monthly. All UV sensors in operation that month should be checked. Additionally, the daily low
validated dose or daily low UV intensity, depending on the dose-monitoring strategy, should be
reported to the state monthly. The state may also have additional reporting requirements and
should be contacted to determine the specific content of the monthly reports and to coordinate
with other reporting requirements.
6.5.2  Example Reporting Forms and Calculation Worksheets

       Example forms and calculation sheets are shown in Figures 6.2 through 6.8. The state
should be contacted to determine whether these forms will be acceptable. The forms are
described in greater detail below. Two calculation worksheets are also provided that can assist
with completing the compliance forms; these forms need not be submitted to the state.

       Figure 6.2 is an example of a summary report that would be completed by the PWS and
submitted to the state on a monthly basis.

       Figures  6.3 and 6.4 are example operating logs that would be completed on a daily basis
for the calculated  dose and UV Intensity Setpoint Approach, respectively. The forms would be
used to record the operating status of the UV equipment and to record the volume of water
discharged during off-specification operation each  day. The state may request that this
information is submitted on a monthly basis.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                      6-39
                                             November 2006

-------
                             6. Start-up and Operation of UV Facilities
       Figure 6.5 is an off-specification calculation worksheet that can assist PWSs with
calculating the off-specification percentage for the daily logs (Figures 6.3 and 6.4); this form
need not be submitted to the state.

       Figure 6.6 is an example duty UV sensor calibration log. This log would be completed
whenever UV sensor calibration checks are performed. The log would be used to record the
results of the calibration testing and to track any UV sensor recalibration or repair work that was
completed. The state may request this information to be submitted on a monthly basis.

       Figure 6.7 is a UV sensor CF calculation worksheet that can help PWSs determine the
appropriate UV sensor CF when the PWS needs to use this approach to stay in compliance. This
form need not be submitted to the state.

       Figure 6.8 is an example on-line UVT analyzer calibration log. This log would be
completed only by those PWSs that have included on-line UVT analyzers as part of their dose-
monitoring strategies. The log would be completed whenever UVT analyzer calibration checks
are performed. The log would be used to record the results of the calibration testing and to track
any recalibration or repair work that was completed.  The state may request this information to be
submitted on a monthly basis.
UV Disinfection Guidance Manual                6-40                               November 2006
For the Final LT2ESWTR

-------
s  <
& a
ft  r
§  o
   §
   o
Reporting Period:
System/Treatment Plant:
PWSID:
Signature of Principal Executive:
Officer or Authorized Agent:







Unit Number

















Total
Total Run Time
(hrs)


















Total Production
(MG)


















Date:
Date:



Off-Specification Data
Number of Off-
Specification Events


















Compliance Certification
Total Volume of Off-Specification Water Produced (MG) [A]
Total Volume of Water Produced (MG) [B]
Total Off-Specification Water Produced (% of Volume of Water Produced) ([A]/[B]*1 00)
Facility Meets Off-Specification Requirement (< 5% of Volume on a Monthly Basis) (Y/N)
Of the sensors, have been checked for calibration and were
The Following Reactors had a Sensor Correction Factor
Reactor Number





Sensor Correction Factor






Total Off-Specification
Volume
(MG)






















3 within the acceptable range of tolerance.
                                                                                                                                                                               (Q
                                                                                                                                                                               £
                                                                                                                                                                                 >
                                                                                                                                                                               m
                                                                                                                                                                               x
                                                                                                                                                                               Q)

                                                                                                                                                                               •g.
                                                                                                                                                                               CD
                                                                                                                                                                               Q)
                                                                                                                                                                               CD
                                                                                                                                                                               •o
                                                                                                                                                                               O
                                                                                                                                                                               a.

-------
s  <
& a
ft  r
§  o
§
o
Reporting Period:
System/Treatment Plant:
PWSID:
UV Reactor:
Process Train:
Operator Signature:
Date:








Operational Data
Day
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Min
Max
Run Time (hrs)

































Total Production
(MG)
































Maximum Validated Flow Rate:
Min mum Validated UVT:
Target Log Inactivation:
Target Pathogen:
Dose Required (Dreq.d):
Validation Factor (VF):
Dose Requirements
Dreq'd
(mJ/cm2)
[A]









































Calculated Dose
Validated Dose 	

VFxCF
Calculated Dose -Dose that is calculated by validated PLC algorithm
VF = Validation factor
CF = UV intensity sensor correction factor.
The CF is only applied if sensors do not meet recommended criteria
(NOTE - a CF will not be needed in most cases)
Data at Daily Minimum Validated Dose
Sensor Correction
Factor2
[B]

































Calculated Dose3
(mJ/cm2)
[C]

































Daily Minimum
Validated Dose4
([C]/[VF]/[B])
(mJ/cm2)
[D]

































Flow Rate
(MGD)

































UVT
(%)

































UV Dose Adequacy
Determination
Validated Dose >
Dreq'd
([D] > [A])
(Y/N)






























^m
Total Off-
Specification
Total Off-
Specification Volume
(MG)






























^H
1 Dreq.a is the dose required for the target log inactivation without a VF or Sensor CF applied and can be found in the UVDGM Table 1 .4.
2 Sensor CFwillbe! isnoCFis used
3 Calculated dose is calculated using the dose algorithm in the PLC.
4 The Validated Dose is the dose based on the calculated dose that is normalized on the Validation Factor and Correct on Factor
5 Off-specification worksheet (Figure 6.5) should be used to calculate daily off-specification volume. If UVT, flowrate, and/or Validated Dose off-specification occur simultaneously, the off-specification time should only be counted once
                                                                                                                                                                                 (Q
                                                                                                                                                                                 C
                                                                                                                                                                                    CD
                                                                                                                                                                                    CO

                                                                                                                                                                                    m
                                                                                                                                                                                    X
                                                                                                                                                                                    Q)


                                                                                                                                                                                    •O.
                                                                                                                                                                                    (D

                                                                                                                                                                                    D
                                                                                                                                                                                    Q)
                                                                                                                                                                                    (D


                                                                                                                                                                                    5'
                                                                                                                                                                                    (Q

                                                                                                                                                                                    O
                                                                                                                                                                                    (Q

                                                                                                                                                                                    31

                                                                                                                                                                                    o
                                                                                                                                                                                    Q)
                                                                                                                                                                                    I-K
                                                                                                                                                                                    (D
                                                                                                                                                                                    Q.

                                                                                                                                                                                    D
                                                                                                                                                                                    O
                                                                                                                                                                                    (/>
                                                                                                                                                                                    (D
                                                                                                                                                                                    5
                                                                                                                                                                                    Q)
                                                                                                                                                                                    O

-------
B-
   C
   <
   a
§  o
   §
   o
Reporting Period:
System/Treatment Plant:
PWSID:
UV Reactor:
Process Train:
Operator Signature:
Date:









Operational Data

Day
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Min
Max
Total
Run
Time
(hrs)


































Total
Production
(MG)


































Flow Rate
Min (mgd)
































^^m
Ave (mgd)
































^^m
Max (mgd)
































^^m
Maximum Validated Flow Rate:
Minimum Validated UVT:
Target Log Inactivation:
Target Pathogen:
Intensity Setpoint:
Intensity Requirements
Intensity Setpoint
(W/m2)
[A]
































^^^H
Sensor Correction
Factor1
[B]
































^^^H
Adjusted Intensity
Setpoint
(W/m2)
([A] * [B])
[C]
































^^^H






Daily Minimum Intensity
Daily Minimum
Intensity
(W/m2)
[D]
































_
Minimum Daily
Intensity >
Adjusted Intensity
Setpoint
([D] > [C])
(Y/N)






























•
Total Flow Off-
Specification
Total Flow Off-
Specification3
(MG)






























™
1 Sensor CF will be 1 is no CF is used.
2 DVT measurements are not requ red but could be useful in addressing operational issues.
3 Off-specification worksheet (Figure 6.5) should be used to calculate daily off-specification volume. If UV intensity or flowrate off-specification occur simultaneously, the off-specification time should only be
counted once
(Q
 £

 (D

 CD
                                                                                                                                                                            m
                                                                                                                                                                            X
                                                                                                                                                                            Q)


                                                                                                                                                                            •O.

                                                                                                                                                                            (D

                                                                                                                                                                            D
                                                                                                                                                                            Q)
                                                                                                                                                                            (D

                                                                                                                                                                            Q)
                                                                                                                                                                            rt-_

                                                                                                                                                                            5'
                                                                                                                                                                            (Q


                                                                                                                                                                            O
                                                                                                                                                                            (Q
                                                                                                                                                                            (D


                                                                                                                                                                            r+
                                                                                                                                                                            <



                                                                                                                                                                            2.

                                                                                                                                                                            o
                                                                                                                                                                            5
                                                                                                                                                                            Q)
                                                                                                                                                                            O

-------
s  <
& a
ft  r
§  o
                                                                                                                                                                                (Q
                                                                                                                                                                                C
   I.
Reporting Period:
System/Treatment Plant:
PWSID:



Operator Signature:
Date:



Off-Specification Calculation Worksheet (Note - This sheet should only be used when an off-specification event occurs):
Date1









Reactor Number









Process Train
Number









Off-Specification Event
Description2









Flow Rate3
(MGD)
[A]









Time
(days)
[B]









Total Off-Specification Flow for the Day3
1 This workseet should only be used for one date and one reactor.
2 This worksheet assumes that the flowrate is constant during the off-specification event. Off-specification volume can also be obtained from a flow totalizer.
3 Off-Specification event can be caused by UVT, flowrate, intensity, or validated dose being out of the validated range.
4 The total off-specification flow should be transferred to Figure 6.3 or 6.4 if any off-specification events occurred.
Off-Specification
Volume
(MG)
([A]*[B])











 O>
 Ol

 m
 x
 Q)

•D_
 CD

 O
 =8
                                                                                                                                                                                CD
                                                                                                                                                                                O
                                                                                                                                                                                5j
                                                                                                                                                                                o'
                                                                                                                                                                                Q)
                                                                                                                                                                                o
                                                                                                                                                                                Q)
                                                                                                                                                                                I-K

                                                                                                                                                                                5'
                                                                                                                                                                                CD
                                                                                                                                                                                CD
   §
   o

-------
§  o
   £.
   I.
   ON


   (Ji
   §
   ft
Reporting Period:
System/Treatment Plant:
PWSID:
CF used (if applicable):
Operator Signature:
Date:






Calibration Ratio =


Sensor Correction Factor =


UV Sensor Calibration Report (Make Additional Copies of Form as Necessary)
Date


















Reactor Number


















Duty Sensor Number


















UV Sensor
Operating Time
(hrs)


















Certification:
Reference Sensor
Serial Number


















Duty UV Sensor
Reading1
[A]


















Number of UV sensor calibrated
Number of UV sensors out of calibration
Reference UV
Sensor Reading1
[B]




















Calibration Ratio
([A]/[B])


















Calibration Ratio < 1.2
(Y/N)


















' S Duty "I
( SR6f J
fSDu*-o.2]
I, «Ref ;
where S is the measured intensity
Sensor
Correction
Factor Used


















If CF is used,
Calibration
Ratio - 0.2 < CF
(Y/N)



















Number of UV sensor(s) sent to manufacturer to be recalibrated as documented below
UV intensity sensors sent to manufacturer for calibration (Add additional rows as necessary):
Sensor Serial Number




Unit No.




Date Sent




Date Received




If three duty UV sensor and reference UV sensor readings are taken, the mean of the reading can be used.



                                                                                                                                                                                          (Q
                                                                                                                                                                                          C
CD
b>

m
x
Q)


<5"
^
o
                                                                                                                                                                                          CD
                                                                                                                                                                                          O
                                                                                                                                                                                          o
                                                                                                                                                                                          3-
                                                                                                                                                                                          CD
                                                                                                                                                                                          O
                                                                                                                                                                                          O
                                                                                                                                                                                          (Q

-------
s  <
& a
ft  r
§  o
   I.
   ON

   ^
   ON
Reporting Period:
System/Treatment Plant:
PWSID:



CF used (if applicable):
Operator Signature:
Date:



Optional Sensor Correction Factor Worksheet:
Date


















Reactor Number


















Duty Sensor
Number1


















W Sensor
Operating Time
(hrs)


















Reference Sensor
Serial Number


















Duty U\/ Sensor
Reading3
[A]


















1 Only failed sensors need to be input.
2 IN Sensor CF should be based on the calibration ratios for the failed sensors and should use the maximum ratio. The CF is reactor specific.
3 If three duty (JV sensor and reference (JV sensor readings are taken, the mean of the reading can be used.
CorrectionFactc
Where:
SDuty is the duty U\/
SRef is the duty W s
Reference U\/
Sensor Reading3
[B]


















Selected UV Sensor
CF2
Ac A
^Duty
)r - y 0.2
l.SRef J
sensor reading
ensor reading
Sensor Correction
Factor
(([A]/[B])-0.2)



















                                                                                                                                                                                       (Q
                                                                                                                                                                                       C
                                                                                                                                                                                       CD
 m
 x
 Q)


•o.
 CD
                                                                                                                                                                                       (D
                                                                                                                                                                                       O
                                                                                                                                                                                       -n

                                                                                                                                                                                       O
                                                                                                                                                                                       Q)
                                                                                                                                                                                       I-K

                                                                                                                                                                                       5'
                                                                                                                                                                                       I
                                                                                                                                                                                       (D
                                                                                                                                                                                       (D
   §
   o

-------
s  <
& a
ft  r
§  o
   I.
§
o
Reporting Period:
System/Treatment Plant:
PWSID:
UVT Analyzer Number:
Operator Signature:
Date:






|UVTon,ine(%) - UVTbench(%)| < 2% UVT
UVT Analyzer Calibration Report (Make Additional Copies of Form as Necessary)
UVT Analyzer Number





Week Number
1
2
3
4
5
Dates





On-line Reading
(%)
[A]





Grab Sample Result
(%)
[B]





Certification:
All calibration checks were within the acceptable tolerance during this month.
Recalibration was required and is documented below.
On-Site Calibration. | | Manufacturer Calibration.
UVT Analyzer Calibration:
UVT Analyzer Number





On-site or
manufacturer
recalibration?





Date Recalibration
Performed





Recalibration
Successful?
(Y/N)





Initials (On-site Calibration
Only)





Dm*™Ce Difference < 2% UVT?
( ' (Y/N)
( |[A]-[B]| ) ("N)





°
                                                                                                                                                                        (Q
                                                                                                                                                                         £
                                                                                                                                                                      p>
                                                                                                                                                                      00

                                                                                                                                                                      m
                                                                                                                                                                      x
                                                                                                                                                                      Q)


                                                                                                                                                                     •g.
                                                                                                                                                                      (D
                                                                                                                                                                        <

                                                                                                                                                                         C
                                                                                                                                                                         <
                                                                                                                                                                         -\



                                                                                                                                                                         SL


                                                                                                                                                                         (D


                                                                                                                                                                         O
                                                                                                                                                                         SL

                                                                                                                                                                         CT

                                                                                                                                                                         s
                                                                                                                                                                         1-h

                                                                                                                                                                         5'


                                                                                                                                                                         O
                                                                                                                                                                         3"
                                                                                                                                                                         (D
                                                                                                                                                                         O
                                                                                                                                                                         O
                                                                                                                                                                        (Q

-------
                            6. Start-up and Operation of UV Facilities
6.6    Operational Challenges

       An excursion from validated operating limits can be caused by low UV intensity, low
validated dose, low UVT, high flow rate, poor UV sensor calibration, poor UVT analyzer
calibration, or a combination of these conditions. These conditions should be resolved quickly to
verify regulatory compliance because they can result in prolonged off-specification operation.
Additionally, the evaluations described in this section should be initiated before validated criteria
are exceeded and off-specification occurs. This section discusses some of the potential
operational challenges and suggests corrective measures.
6.6.1  Low UV Intensity or Validated Dose Below the Setpoint

       Low UV intensity or validated dose may cause a reactor to operate outside of validated
limits. Although UV intensity limits are not explicitly set in the Calculated Dose Approach, a
low UV intensity will reduce the validated dose that is delivered, along with low UVT or high
flow rate. Therefore, approaches for addressing either a low UV intensity or low validated dose
readings are often the same.

       The output of the UV lamps, UV transmittance of the sleeves, status  of the UV sensor,
and fouling of both lamp sleeves and UV sensor windows affect UV sensor readings and
validated dose. Figure 6.9 presents a decision tree for evaluating low UV intensity or low
validated dose. If strategies in Figure 6.9 cannot be implemented or are not successful in getting
the UV intensity or validated dose above the required setpoint, the UV manufacturer or UV
facility designer should be contacted to investigate the problem further. The  PWS should activate
any backup disinfection, shift production to another WTP or source of supply, or consider
shutting down the WTP until the UV intensity or validated dose is within the validated limits.
Any time that the UV intensity or validated dose is lower than the validated  limit, it should be
recorded as off-specification (Section 6.4.1.3).
6.6.2  Low UV Transmittance

       This evaluation of low UVT presumes that either the low intensity evaluation
(Section 6.6.1) has been completed and either (1) the cause of the low UV intensity was low
UVT or (2) the operational staff has observed low UVT. If the reactor uses the Calculated Dose
Approach, it may be programmed to increase lamp output or number of lamps in service to
accommodate a decrease in UVT if the UVT is still within the validated range. If the UV
equipment does not sufficiently compensate, or if the UV reactor cannot adjust lamp output, the
UV intensity or validated dose may fall below the validated limits.
UV Disinfection Guidance Manual                6-48                              November 2006
For the Final LT2ESWTR

-------
                                   6. Start-up and Operation of UV Facilities
          Figure 6.9. Low UV Intensity or Low Validated  Dose Decision Chart
                                            If possible, adjust UV facility
                                            operation to compensate for
                                            low UV intensity if not done
                                            automatically by the control
                                                     system.
                                                                Yes"
   UV Intensity or
  validated dose is
below validated limits.
     Is the UV
intensity or validated
   dose still low?
                                       IstheUVT
                                  low or below validated
                                         limits?
                                        s the U
                                  lamp age beyond the
                                        esign life?
                                Evaluate duly UV sensors'
                            calibration using the recommended
                            procedures (Figure 8.11). Were any
                                V sensors out of calibration?
    Evaluate and repair
     faulty UV sensor.
                                                  Is UV intensity
                                                 or validated dose
                                                    still low?
                               Take out quartz sleeve
                                 and/or UV sensor
                               window and inspect for
                                     fouling.
                                                                 Clean sleeve and or
                                                              UV sensor surface/ window
                                                              Inspect oltier reactors and
                                                                UV sensors for fouling.
                 re the sleeves
         "or UV sensor surface/window
                   fouled?
                                                                     Is tie UV
                                                                 intensity or validated
                                                                   dose still low?
                                   Is trie quartz
                              sleeve's age beyond the
                                   design life?
                                                              Is UV intensity or
                                                            alidated dose still low?
               Replace sleeve
           Contact manufacturer
               or UV facility
                designer to
            investigate this issue
               further. Shift
           production demand to
             another supply or
            WTP if available, or
               consider WTP
                shutdown.
                                                          Check other lamps and/or
                                                          sleeves in other reactors
                                                            to see if they need to
                                                                be replaced.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                                  6-49
               November 2006

-------
                            6. Start-up and Operation of UV Facilities
       The UVT analyzer or bench-top spectrophotometer equipment should be evaluated to
determine if the instruments are operating properly as described in Section 6.4.1.2. If the low
UVT is not due to faulty instruments and is below the validated UVT, the following WTP
operational changes should be considered:

          Vary  the source water blending ratio (if available) to increase UVT.

       .   Where applicable, evaluate whether the coagulation process has been optimized for
          natural organic matter (NOM) removal and whether the coagulant dose should be
          increased. Poor coagulation caused by coagulant under-dosing can lead to increased
          NOM concentration and an associated decrease in UVT.

          Increase the oxidant dose prior to the UV facility  if possible. However, this strategy
          may increase DBF formation or increase fouling,  which should be evaluated before
          this option is used.

          Investigate potential upstream chemical interferences from a process failure or upset.
          For example, if the ozone quenching system failed, the UVT would decrease.

       A decision tree that summarizes the approach for troubleshooting low UVT is shown in
Figure 6.10. If the strategies presented in Figure 6.10 and described above cannot be
implemented or are not successful in correcting the low UVT, the UV manufacturer or UV
facility designer  should be contacted to investigate the problem further. The PWS may consider
shutting down the WTP or activating a backup disinfection system, if available, until the UVT is
within the validated limits. The low UVT condition must be recorded as off-specification
(Section 6.4.1.3) when the UVT is lower than the validated limit and the Calculated Dose
Approach is used.
6.6.3  Failure to Meet UV Sensor Calibration Criterion

       Unreliable UV sensor readings can be due to UV sensor malfunction, condensation in the
UV sensor or between the UV sensor and UV sensor window, lamp malfunction, poor
grounding, degradation of UV sensor electronics, or electronic short-circuits. Monitoring the UV
sensor calibration will identify poor performance.

       The integrated procedure of monitoring multiple UV sensor calibrations and evaluating
failures of the calibration criterion can be complex, especially if multiple UV sensors fail the
calibration criterion. A decision tree (Figure 6.11) can assist with the monitoring of UV sensor
calibration and determining whether UV sensors should be replaced or whether a UV sensor CF
is needed.
6.7    Staffing, Training, and Safety Issues

       To provide consistent and reliable operation of UV reactors, the PWS must have
appropriate staffing, training, and safety measures in place. This section discusses these issues.

UV Disinfection Guidance Manual                6-50                              November 2006
For the Final LT2ESWTR

-------
                                    6. Start-up and Operation of UV Facilities
                      Figure 6.10. Low UV Transmittance Decision Chart
             Is repeat UVT
            similar and low?
                                                            Yes>
Record on-line UVT
measurement and
check grab
sample UVT with
bench-top
spectrophotometer.
i
r
                 Yes
                                                                       Is the
                                                                 UVTon-|ine - UVT bench
                                                                     = 2% UVT?
   Recalibrate or repair
       bench-top
   spectrophotometer.
                Check
          spectrophotometer's
              calibration.
          Is spectrophotometer
          within manufacturer's
           calibration limits?
                                        Is the UV facility
                               operating off-specification because o
                                          low UVT?
                                         Can UVT be
                             increased through source water changes or
                                    WTP operation changes?
                                                                                Contact manufacturer
                                                                               or UV facility designer to
                                                                           investigate this issue further. Shift
                                                                             production demand to another
                                                                             supply or WTP if available, or
                                                                               consider WTP shutdown.
     Is UVT still below the
validation limit after the changes?
                                         s UV intensity
                                       or validated dose
                                       below validation
                                            limits?
                                        See low UV intensity
                                           decision tree
                                           (Figure 6.9).
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                      6-51
November 2006

-------
                                     6. Start-up and Operation of UV Facilities
                Figure 6.11. Monitoring of UV Sensor Calibration  Flowchart
                                                               Start
 rYes-
   Have
 all duty UV
sensors been
  checked?
                                                Perform monthly calibration check of duty
                                                  UV sensor by comparing the duty UV
                                                sensor reading to a reference UV sensor
                                                 reading. Record the readings. If desired
                                                 three duty UV sensor and reference UV
                                               sensor readings can be taken, and a mean
                                                   calibration ratio can be calculated.
                                            Yes
       Determine the appropriate
       UV sensor CF by reviewing
      Section 6.4.1.1 and Example
       6.4. Multiply the validated
      dose or the intensity selpoint
      for the UV reactor by the UV
              sensor CF
   No-
                                                   Does the
                                             calibration ratio meet the
                                         calibration recommended in Eqn.
                                                     6.2?
                                                   (S   1
        Can the UV facility
      be shut down until failed
         UV sensors are
            replaced?
        Shut down the UV
           facility until
         replacement UV
        sensors arrive and
         notify the state.
                           It is recommended
                            that all duty UV
                           sensor calibrations
                           are checked before
                            determining the
                          course of action for a I
                           failed UV sensor.
                          _Yes-
                                Are the
                            fiumber of failed
                           UV sensors greater
                            than the number
                              of spare UV
                               sensors? ,
      Was the duty UV
 'sensor properly calibrated
    with either the second
 .reference UV sensor or botl
          UV sensors?
                                                                 Yes
                          Check the duty UV sensor
                              again by either
                         (1) inserting a 2nd reference
                          UV sensor and repeating
                            the calibration check
                                   or
                          (2) inserting 2 other duty
                          UV sensors and repeating
                           the calibration check to
                         determine that the duty UV
                          sensor is out of calibration
                          and not the reference UV
                                  sensor.
  Replace the reference UV
  sensor with a 2nd calibrated
  reference UV sensor and
  recheck all duty UV sensors
checked with the bad reference
  UV sensor. Recalibrate the
  failed reference UV sensor
                                  When a UV sensor fails the
                                    calibration criteria, the
                                    utility has two options:
                                    (1) replace the bad UV
                                           sensor
                                             or
                                  (2) apply a UV sensor CF to
                                    the UV reactor setpoint.
                                                                           Have
                                                                         all duty UV
                                                                       sensors been
                                                                         checked?
                                                                                                  -No—'
                                                                 Yes
                                                               Are any
                                                *.Yes—
-------
                             6. Start-up and Operation of UV Facilities
6.7.1  Staffing Levels

       During initial start-up operation, more operator attention will be needed to assist with
functional and performance testing and to establish site-specific O&M procedures (described in
Section 6.1.1). However, depending on the level of automation, a typical UV facility requires
minimal operator attention during normal operation. Generally, UV facilities use PLCs to
monitor operating parameters, control the UV reactor, and generate alarms. Increased automation
(e.g., remote monitoring capability) may be incorporated to further reduce operator requirements.
Table 6.9 describes how various site-specific factors affect staffing needs for a UV facility.
                     Table 6.9. Factors Impacting Staffing Needs
Factor
Type of UV Reactor
Instrumentation and Monitoring
Strategy
Water Quality
Impact on Staffing
LP and LPHO reactors may require more maintenance than
an MP reactor because they have more lamps and typically
are cleaned off-line (i.e., OCC cleaning). However, MP lamps
generally need to be replaced more often than LP lamps.
More automated control strategies will result in lower staffing
levels due to enhanced remote operation and monitoring
capability.
Water quality and UV reactor design affect sleeve fouling and
cleaning frequency. These factors, in turn, impact the staffing
needs for manual cleaning for OCC systems and for
maintaining the OMC or OMCC system.
6.7.2  Training

       Training is necessary for all personnel who are associated with the UV facility, including
operators, maintenance workers, instrumentation technicians, electricians, laboratory staff,
custodial staff, engineers, and administrators. The training program should incorporate any state
requirements and should emphasize both normal and emergency operating procedures, safety
issues, process control and alarm conditions, validated operation, monitoring, instrumentation,
and responses to operational issues.

       The UV manufacturer and UV facility designer should provide training on the UV
reactors, UV facility design, and O&M activities. Training should include both classroom
instruction and field training. Additionally, actively involving the operators during start-up will
provide another opportunity to reinforce classroom instructions. Continued training should be
provided when new employees are hired or when a process or control is altered.
6.7.3  Safety Issues

       This section provides some recommended safety precautions for UV reactor operations.
The recommended precautions in this section should be considered in addition to manufacturer's
recommended safety precautions and procedures, Occupational Safety and Health
UV Disinfection Guidance Manual
For the Final LT2ESWTR
6-53
November 2006

-------
                            6. Start-up and Operation of UV Facilities
Administration (OSHA) regulations, and state guidance and regulations for UV reactor
operations.

       In addition to the standards and procedures established for WTP operations, the following
safety issues pertain specifically to UV reactors:
          UV light exposure
          Electrical safety
          Burns from hot lamps or equipment
       •   -ULilllO ilWlll 11WI ICtlllJJO Wl V^LUJJlllVlll

       .   Abrasions or cuts from broken lamps or sleeves

       .   Potential exposure to mercury from broken lamps
       Threshold limit values (TLVs) for UV light apply to occupational exposure to UV
incident on the skin or eyes. The recommended TLVs depend on the lamp wavelengths emitted
and the UV intensity (mW/cm2). The PWS can determine the appropriate TLVs for their UV
reactors, using TLVs for Chemical Substances and Physical Agents and Biological Exposure
Indices (ACGIH 2006). These values are not enforceable standards, but should be considered
when establishing operational procedures. To limit or prevent operator exposure to the UV light,
UV reactors should have interlocks that deactivate the lamps when reactors are accessed.
Viewing ports, if provided, should be fitted with UV filtering windows, or operators should wear
a UV-resistant face shield when working in the UV reactor. To minimize the danger of exposure,
warning signs also should be posted.

       To reduce the risk of electrical shock, the main electrical supply to the UV reactors
should be disconnected and the operator should wait at least 5 minutes for the lamps to cool and
the energy to dissipate before maintenance is performed in areas where electric shock may be a
risk. All safety and operational precautions required by the National Electric Code (NEC),
OSHA, local electric codes, and the UV manufacturer should be followed and include the
following precautions:

       .   Proper grounding
       .   Lockout, tagout procedures
       .   Use of proper electrical insulators
          Installation of safety cut-off switches

       The ballasts and the reactor chamber can also become hot during operation. The
temperatures of these components should be checked before touching them.

       Broken lamps pose two potential safety hazards. The  lamps and sleeves are constructed
of quartz tubing, which can fracture and cause serious cuts or injury, and broken lamps may
release mercury. Operators should be trained in proper mercury cleanup and disposal procedures
to prevent mercury inhalation or absorption through the skin. Appendix E discusses lamp
breakage and cleanup procedures.
UV Disinfection Guidance Manual                6-54                              November 2006
For the Final LT2ESWTR

-------
                                  7. Bibliography
Adams, M.H. 1959. Bacteriophage. New York: Interscience Publishers.

Adler, H.I. 1966. The genetic control of radiation sensitivity in microorganisms. Advances in
       Radiation 2:161'-191.

AIAA. 1998. Guide for the Verification and Validation of Computational Fluid Dynamics
       Simulations. G-077-1998, American Institute of Aeronautics and Astronautics, Reston,
       VA.

Akhlaq, M.S., H.P. Schuchmann, and C. von Sonntag. 1990. Degradation of the polysaccharide
       alginic acid: A comparison of the effects of UV light and ozone. Environmental Science
       and Technology 24:379-3 83.

Altena, F.W., J.B.J. van Overveld, and H. Giller. 2001. Technological advances in disinfection
       lamps leading to more compact UV sources.  Presented at the First International Congress
       on Ultraviolet Technologies, Washington, DC, June 14 - 16.

ACGUL 2006. TLVs and BEIs Based on the Documentation of the Threshold Limit Values for
       Chemical Substances and Physical Agents and Biological Exposure Indices. American
       Conference of Governmental Industrial Hygienists. Cincinnati, OH.

APHA-AWWA-WEF. 1998. Standard Methods for  the Examination of Water and Wastewater,
       20th ed. Washington, DC. American Public Health Association, American Water Works
       Association, Water Environment Federation.

Aquafine Corporation. 2001. CSL series: Installation, maintenance and operation manual. Part
       No. 115-1.

Ballester, N.A., J.P. Malley, J. North, T. Meyers, and M. Duben. 2003. Synergistic disinfection
       of Adenovirus Type 2. Presented at the Second International Congress on Ultraviolet
       Technologies, Vienna, Austria, July 9-11.

Bircher, K. 2004. Large Scale UV Reactor Design and Validation. Proceedings of the American
       Water Works Association Water Quality Technology Conference, San Antonio, TX,
       November 14-18.

Bircher, K. 2005. Calgon Carbon Corporation. Personal communication with James Collins.
       November 23  -28.

Bircher, K., Gaithuma, D.R., Rennecker, J.L., and Wright, H.B. 2005. Using CFD to Optimize
       Off-Site UV Reactor Validation for a  Drinking Water Plant with a Unique Piping
       Configuration. Poster presented at the American Water Works Association 2005  Water
       Quality Technology Conference, Quebec City, Quebec, Canada, November 6-10.
UV Disinfection Guidance Manual                7-1                              November 2006
For the Final LT2ESWTR

-------
                                   Chapter 7. Bibliography
Blatchley, E.R, III. 1997. Numerical modeling of UV intensity: Application to collimated-beam
       reactors and continuous-flow systems. Water Research 31:2205-2218.

Blatchley, E.R., III, C. Shen, Z. Naunovic, L. Lin, D.A. Lyn, J.P. Robinson, K. Ragheb, G.
       Gregori, D.E. Bergstrom, S. Fang, Y. Guan, K. Jennings, and N. Gunaratna. 2005. Dyed
       microspheres for quantification of UV dose distributions: Photochemical reactor
       characterization by Lagrangian actinometry. Proc. of Disinfection 2005, Sharing
       Disinfection Technologies: Water, Wastewater, and Biosolids, Mesa, AZ, February 6-9.

Bolton, J.R. 2000. Calculation of ultraviolet fluence rate distributions in an annular reactor:
       Significance of refraction and reflection. Water Research 34:3315-3324.

Bolton, J.R., and K. Linden. 2003. Standardization of methods for fluence (UV dose)
       determination in bench-scale UV experiments.  Journal of Environmental Engineering
       129(3):209-216.

Bolton, J.R., M.I. Stefan, R.S. Gushing, and E. Mackey. 2001. The importance of water
       absorbance/transmittance on the efficiency of ultraviolet disinfection reactors. Proc. First
       International Congress on Ultraviolet Technologies, Washington, DC, June.

Bolton Photosciences, Inc. 2004. http://www.boltonuv.com/reactors/940/.

Brandt, C.L., and A.C. Giese. 1956. Photoreversal of nuclear and cytoplasmic effects of short
       ultraviolet radiation onParamecium caudatum. Journal of General Physiology 39:735-
       751.

Brodkorb, T., and D. Richards. 2004. UV disinfection  design to avoid chlorine destruction in
       high UVT waters. Presented at Ontario Water Works Association, Niagara Falls, Ontario,
       Canada, May.

Cabaj, A., R. Sommer, W.  Pribil, and T. Haider. 2001. What means "dose" in UV-disinfection
       with medium pressure lamps? Ozone Science and Engineering 23:239-244.

Cabaj, A., R. Sommer, W.  Pribil, and T. Haider. 2002. The spectral UV sensitivity of
       microorganisms used in biodosimetry. Water Science and Technology: Water Supply
       2(3):175-181

Cabaj, A., R. Sommer, andD. Schoenen. 1996. Biodosimetry: Model calculations forU.V. water
       disinfection devices with regard to UV dose distributions. Water Research 30:1003-
       1009.

Cairns, B. 2000.  Comments on USEPA questions about the risk of using UV disinfection
       technologies in the potable water industry. Trojan Technologies Inc., London, Ontario,
       Canada, May 24.

Calgon Carbon Corporation. 2006. http://www.calgoncarbon.com/uv/.
UV Disinfection Guidance Manual                 7-2                               November 2006
For the Final LT2ESWTR

-------
                                   Chapter 7. Bibliography
Chang, J.C.H., S.F. Osoff, D.C. Lobe, M.H. Dorfman, C.M. Dumais, R.G. Quails, and J.D.
       Johnson. 1985. UV inactivation of pathogenic and indicator microorganisms. Applied and
       Environmental Microbiology 49:1361-1365.

Chiu, K.-P., D.A. Lyn, P. Savoye, and E.R. Blatchley, III. 1999. Effect of UV system
       modifications on disinfection performance. Journal of Environmental Engineering
       125(5):459-469.

Christensen, J., and K. Linden. 2002. New findings regarding the impacts of suspended particles
       on UV disinfection of drinking water. Proceedings of the Annual Conference of the
       American Water Works Association, New Orleans, LA, June 16-20.

Clancy, J., Fallen, K., and Hargy, T. 2004. Development of nonpathogenic surrogates for large
       scale UV reactor validation. Proceedings of the Water Quality Technology Conference,
       San Antonio, November 14-18

Clear, R., and S. Berman.  1994. Environmental and health aspects of lighting: Mercury. Journal
       of Illuminating Engineering Society 23 (2): 13 8-15 6.

Collins, J., and J.P. Malley. 2005. The use of native biodosimetry to validate MPHO and LPHO
       UV reactors. Presented at the 3rd International Congress on Ultraviolet Technologies,
       Whistler, BC, Canada, May 24 - 27.

Coronell, O., M. Page, and B. Marinas. 2003. Sequential disinfection strategies with UV, ozone,
       and chlorine for optimum control of viruses, spores, and Cryptosporidium parvum
       oocysts. Presented at the AMERICAN WATER WORKS ASSOCIATION Water Quality
       Technology Conference, Philadelphia, PA, November 2-6.

Cotton, C. 2002. Confidential communication with major water municipality. June.

Cotton, C., R. Cushing, and D. Owens.  2002. The impacts of LT2ESWTR requirements on UV
       disinfection design and operation.  Proceedings of the Annual Conference and Exhibition
       of the American Water Works Association, New Orleans, LA, June 16-20.

Cotton, C.A., L. Passantino, D.M. Owen,  M. Bishop, M. Valade, W. Becker, R. Joshi,  R. Hubel,
       and M. LeChevallier. 2005. Integrating UV Disinfection into Existing Water Treatment
       Plants. American Water Works  Association Research Foundation, Denver, CO.

Crozes, G., R.  Cushing, J. Rennecker, C. Cleveland, and H. Wright. 2003. Synergies of UV
       disinfection and ozone in water  treatment. Proceedings of the 2nd International Congress
       on Ultraviolet Technologies, Vienna, Austria, July 9-11.

DeMers, L.D., and R.C. Renner. 1992. Alternative Disinfection Technologies for Small Drinking
       Water Systems. Chapter 5, Denver, CO, American Water Works Association Research
       Foundation.
UV Disinfection Guidance Manual                7-3                               November 2006
For the Final LT2ESWTR

-------
                                   Chapter 7. Bibliography
Derrick, B. 2005. Inorganic Fouling and Effects on Intensity in Groundwater Applications of
       Ultraviolet Photoreactors. M.S. Thesis, School of Civil Engineering, Purdue University,
       West Lafayette.

Derrick, B., and E.R. Blatchley, III. 2005. Field investigation of inorganic fouling of UV systems
       in groundwater applications. Presented at the 3rd International Congress on Ultraviolet
       Technologies, Whistler, BC, Canada, May 24 - 27.

Dinkloh, L. 2001. WEDECO-ITT (Formerly WEDECO-Ideal Horizons). Personal
       communication with Malcolm Pirnie staff, regarding UV reactors. Tucson AZ, May 10,
       August 10, and August 15.

Downes, A., and T.P. Blunt. 1877.  Researches on the effect of light upon bacteria and other
       organisms. Proceedings Royal Society London 26.

Draper, N., and H. Smith. 1998. Applied Regression Analysis., Third Edition. New York: Wiley.

Dulbecco, R.  1950. Experiments on photoreactivation of bacteriophages inactivated with
       ultraviolet radiation. Journal of Bacteriology 59:329-347.

DVGW. 2006.  UV Disinfection Devices for Drinking Water Supply—Requirements and
       Testing. DVGW W294-1, -2, and -3.  German Gas and Water Management Union
       (DVGW), Bonn,  Germany.

Dzurny, T., Reddy, S.P., and Wright, H.B. CFD Modeling of UV Reactors - A Critical Review.
       Paper presented at the International Ultraviolet Association 2003 Annual Conference,
       Vienna, Austria, July 9-11,  2003.

Gaithuma, D.R. UV Reactor Validation. 2006. Paper presented at the Ultraviolet Association
       Mid-South UV technology  one-day workshop, Rice University,  Houston, TX, May 24.

Gaithuma, D.R. 2006.  Validation Field Testing: From Planning, Sampling, QA/QCing to
       Analyzing: Essential 'ings'  for Getting Good Bioassay Data. Paper presented at the
       American Water Works Association 2006 Annual Conference and Exposition, San
       Antonio, TX, June 11-15.

Gates, F.L. 1929. A study of the  bactericidal action of ultraviolet light. Journal of General
       Physiology 13:231-260.

GE Quartz. 2004a. http://www.gequartz.com/en/tools/trans/transmission.htm.

GE Quartz. 2004b. http://www.gequartz.com/en/help.htm.

Giese, A.C., C.L. Brandt, C. Jacobson, D.C. Shephard, andR.T. Sanders.  1954. The effects of
       starvation on photoreactivation in Colpidium  colpoda. Physiological Zoology 71:71-78.
UV Disinfection Guidance Manual                7-4                              November 2006
For the Final LT2ESWTR

-------
                                  Chapter 7. Bibliography
Glew, D.N., and D.A. Hames. 1971. Aqueous nonelectrolyte solution: Mercury solubility in
       water. Canadian Journal of Chemistry 49:3114-3118.

Grebe, T.E., D.D. Sabin, and M.F. McGranaghan. 1996. An Assessment of Distribution System
       Power Quality. Electric Power Research Institute (EPRI), Project 3908, Final Report.

Greene, D.J., H.B. Wright, T.M. Dzurny, and E.D. Mackey. 2004. UV reactor optimization using
       CFD and UV intensity models. Presented at the AMERICAN WATER WORKS
       ASSOCIATION 2004 Annual Conference & Exposition, Orlando, FL, June 13 - 17.

Gump, D. 2002. Severn Trent. Personal communication with Jennifer Hafer, Malcolm Pirnie,
       Inc., regarding UV reactors. Tucson, AZ, November 6.

Halliday, D., and R. Resnick. 1978. Physics. John Wiley & Sons, New York.

Haas, C.N., and G.P. Sakellaropoulos. 1979. Rational analysis of ultraviolet  disinfection.
       National Conference on Environmental Engineering, Proc. American Society of
       Chemical Engineers (ASCE) Specialty Conf, San Francisco, CA, July 9-11.

Hargy, T., J. Clancy, K. Fallon, H.B. Wright, and K. Bircher.  2004. Stability of MS2 in
       groundwaters. Proceedings of the AMERICAN WATER WORKS ASSOCIATION
       Water Quality Technology Conference. San Antonio, Texas, November 14-18.

Haubner, K. 2005. WEDECO-ITT (Formerly WEDECO-Ideal Horizons).  Personal
       communication from James Collins. November 23 - December 1.

Havelaar, A.H., C.C.E. Meulemans, W.M. Pot-Hogeboom, and J. Koster.  1990. Inactivation of
       bacteriophage MS2 in wastewater effluent with monochromatic and polychromatic
       ultraviolet light.  Water Research 24(2):1387-1393.

Hoyer, O. 1998. Testing performance and monitoring of UV systems for drinking water
       disinfection. Water Supply 16(1-2):424-429.

Hoyer, O. 2002. DVGW Test Facility for UV Disinfection Devices at
       Wahnbachtalsperrenverband Siegelsknippen, Germany, personal communication with
       Harold Wright of Carollo Engineers.

Upelaar, G.F., B. van der Veer, G.J. Medema, and J.C. Kruithof. 2003. By-product formation
   during UV disinfection of a pre-treated surface water. Proceedings of the 2nd International
   Congress on Ultraviolet Technologies, Vienna, Austria, July 9-11.

Infilco Degremont. 2005. http://www.infilcodegremont.com.

Jagger, J. 1967. Introduction to Research in Ultraviolet Photobiology. Englewood Cliffs, NJ:
       Prentice-Hall, Inc.
UV Disinfection Guidance Manual                7-5                              November 2006
For the Final LT2ESWTR

-------
                                  Chapter 7. Bibliography
Jeffcoat, S. 2005. CH2M Hill, Inc. Personal communication with Christine Cotton, Malcolm
       Pirnie, Inc., regarding fouling in Clayton County. Tucson, AZ, December.

Kashimada, K., N. Kamiko, K. Yamamoto, and S. Ohgaki. 1996. Assessment of
       photoreactivation following ultraviolet light disinfection. Water Science & Technology
       33:261-269.

Kashinkunti, R., K.G. Linden, G.  Shin, D.H. Metz, M.D. Sobsey, M. Moran, and A.  Samuelson.
       2003. Achieving multi-barrier inactivation in Cincinnati: UV, byproducts,  and
       biostability. Journal of the American Water Works Association 96(6): 114-127.

Kawar, K., J. Jenkins, B. Srikanth, and A. Shurtleff. 1998. Ultraviolet light deterioration of the
       light transmittance of quartz sleeves due to continuous exposure to UV radiation.
       Ultrapure Water (October) 67-71.

Kelner, A. 1950. Light-induced recovery of microorganisms from ultraviolet radiation injury,
       with special reference to Escherichia coli. Bulletin of the New York Academy of Medicine
       26:189-199.

Knudson, G.B. 1985. Photoreactivation ofW-irradiatedLegionellapneumoplila and other
       Legionella species. Applied and Environmental Microbiology 49:975-980.

Kolch, A. 2001. Assessing the Risk of Mercury Release from Ultraviolet Discharge Lamps in
       Drinking Water. WEDECO AG Water Technology.

Kruithof, J.C.,  and R.C. van der Leer.  1990. Practical experiences with UV-disinfection in The
       Netherlands. Proceedings of the AMERICAN WATER WORKS ASSOCIATION
       seminar on emerging technologies in practice, Annual Conference of the American Water
       Works Association, Cincinnati, OH, June 17-21.

Kubik, D. 2005. Lethbridge WTP. Personal communication with Christine Cotton, Malcolm
       Pirnie, Inc., regarding chlorine residual loss in UV reactors. Tucson, AZ, September.

Larason T.C., and C.L. Cromer. 2001. Source of error in UV radiation measurements. Journal of
       Research of the National Institute of Standards and Technology 106(4):649-656.

Larason, T.C.,  S.S.  Bruce, and A.C. Parr. 1998. Spectroradiometric Detector Measurements.
       NIST Special Publication 250-41. U.S. Government Printing Office, Washington, DC.

Larner, S. 2005. Hanovia Corporation. Personal communication by James Collins  (Malcolm
       Pirnie). November 23-25.

Lawal, O. 2006. WEDECO ITT (Formerly known as WEDECO Ideal Horizon). Personal
       communication by Laurel Passantino. March 23 - April 25.

Leinberger, J. 2002. Trojan Technologies. Personal communication with Jennifer Hafer,
       Malcolm Pirnie, Inc, regarding UV reactors. November 5.

UV Disinfection Guidance Manual                7-6                             November 2006
For the Final LT2ESWTR

-------
                                   Chapter 7. Bibliography
Leinberger, J. 2005. Trojan Technologies. Personal communication with James Collins.
       November 23 - December 1.

Lin, L.S., C.T. Johnston, and E.R. Blatchley, III. 1999a. Inorganic fouling at quartz:water
       interfaces in ultraviolet photoreactors. I: Chemical characterization.  Water Research
       33:3321-3329.

Lin, L.S., C.T. Johnston, and E.R. Blatchley, III. 1999b. Inorganic fouling at quartz:water
       interfaces in ultraviolet photoreactors. II: Temporal and spatial distribution. Water
       Research 33:3330-3338.

Linden, K.G. 2002. Personal email communication with Laurel Passantino,  Malcolm Pirnie, Inc.
       April 24 - May 6.

Linden, K.G., and J.L. Darby. 1998. UV disinfection of marginal effluents:  Determining UV
       absorbance and subsequent estimation of UV intensity. Water Environment Research
       70(2).

Linden, K.G., G.A. Shin, andM.D. Sobsey. 2001. Comparative effectiveness of UV wavelengths
       for the inactivation of Cryptosporidium parvum oocysts in water. Water Science &
       Technology 43:171-174.

Linden, K.G., G.A. Shin, G. Faubert, W. Cairns, and M.D. Sobsey. 2002a. UV disinfection of
       Giardia lamblia cysts in water. Environmental Science and Technology 36:2519-2522.

Linden, K.G., L. Batch, and C. Schulz. 2002b. UV disinfection of filtered water supplies: Water
       quality impacts on MS2 dose-response curves.  Proceedings of the Annual Conference of
       the American Water Works Association, New Orleans, LA, June 16-20.

Linden, K.G., C. Sharpless, P. Neofotistos, B. Ferran, and S. Jin. 2004. MP  lamp  aging: Spectral
       shifts and irradiance changes. Presented at the Annual Conference of the American Water
       Works Association UVDGM stakeholder workshop, Orlando, FL, June 2004.

Liu, W., S.A. Andrews, J.R. Bolton, K.G. Linden, C. Sharpless, andM. Stefan. 2002.
       Comparison of disinfection byproduct (DBF) formation from different UV technologies
       at bench scale. Water Science Technology- Water Supply 2(5-6):515-521.

Logsdon, G., and J.  Symons. 1973. Mercury removal by conventional water-treatment
       techniques. Journal of the American Water Works Association 65:554-562.

Lotierzo, M., C. Feliers, N. Faure, L. LeGrand, S. Saby, and P. Cervantes. 2003. Synergistic
       effects of sequential treatment by UV irradiation and chemical disinfectant for drinking
       water. Proceedings of the  2nd International Congress on Ultraviolet  Technologies,
       Vienna, Austria, July 9-11.
UV Disinfection Guidance Manual                 7-7                               November 2006
For the Final LT2ESWTR

-------
                                   Chapter 7. Bibliography
Mack, J., and J.R. Bolton. 1999. Photochemistry of nitrite and nitrate in aqueous solution: A
       review. Journal of Photochemistry and Photobiology 128:1 - 13.

Mackey, E.D., R.S. Cushing, and H.B. Wright. 2001. Effect of water quality on UV disinfection
       of drinking water. Proceedings of the First International Ultraviolet Association
       Congress, Washington, DC, June 14 - 16.

Mackey, E.D., R.S. Cushing, M.L. Janex, N. Picard, J.M. Laine, and J.P. Malley. 2004. Bridging
       pilot-scale testing to full-scale design of UV disinfection systems. American Water
       Works Association Research Foundation, Denver, CO.

Mackey, E.D., H. Wright, T. Hargy, A.C. Fonseca, and A. Cabaj. 2004. Reducing UV design
       safety factors, optimization of UV reactor validation. Proceedings of the American Water
       Works Association Water Quality Technology Conference, San Antonio, TX, November
       14-18.

Mackey, E.D., H.B. Wright, and A.C. Fonseca. 2005. Improving UV validation:  Mimicking real-
       world conditions. Presented at the Water Environment Foundation Disinfection
       Conference, Phoenix, AZ, February 6-9.

Mackey, E.D., H.B. Wright, T. Hargy, J. Clancy, and A. Cabaj. 2006 (in press). Optimization of
       UV Reactor Validation.  American Water Works Association Research Foundation,
       Denver, CO.

Malley, J. 2000. The state of the art in using UV disinfection for waters and wastewaters in
       North America. Proceedings of ENVIRO 2000 - Australian Water Association (AWA),
       Sydney, Australia, April 9-13.

Malley, J. 2001. Addressing concerns over mercury releases from UV lamps. IUVA News
       3(3):28-29.

Malley, J.P. 2002. Historical perspective of UV use. Presented at the AMERICAN WATER
       WORKS  ASSOCIATION Water Quality Technology Conference, Seattle, WA,
       November 10- 14.

Malley, J.P., J.P. Shaw, and J.R. Ropp.  1996. Evaluation of by-products produced by treatment
       of groundwaters with ultraviolet irradiation. American Water Works Association
       Research Foundation, Denver, CO.

Malley, J., B.A. Petri, G.L. Hunter, D. Moran, M. Nadeau, and J. Leach. 2001. Full-scale
       Implementation of UV in Groundwater Disinfection Systems. American Water Works
       Association Research Foundation Final Report, Denver, CO.

Malley, J.P., N. Ballester, S. Farabaugh, andD. Gaithuma. 2003. Multiple disinfectant barriers
       within a multiple barrier system: A new drinking water paradigm for public health
       protection. Proceedings  of the 2nd International Congress on Ultraviolet Technologies,
       Vienna, Austria, July 9-11.

UV Disinfection Guidance Manual                7-8                              November 2006
For the Final LT2ESWTR

-------
                                   Chapter 7. Bibliography
Malley, J.P., N.A. Ballester, A.B. Margolin, K.G. Linden, A. Mofidi, J.R. Bolton, G. Crozes, B.
       Gushing, E. Mackey, J. M Laine, and M-L. Janex. 2004. Inactivation of Pathogens with
       Innovative UV Technologies. American Water Works Association Research Foundation.

Mamane-Gravetz, H., and K.G. Linden. 2004. Impact of particle aggregated microbes on UV
       disinfection. Presented at the American Water Works Association Water Quality
       Technology Conference, San Antonio, TX, November 14 - 18.

McClean, J. 200la. Hanovia Corporation. Personal communication with Doug Owen and
       Christine Cotton, Malcolm Pirnie, Inc., regarding UV reactors. Slough, England, April 9.

McClean, J. 2001b. Hanovia Corporation. Personal communication with Christine Cotton,
       Malcolm Pirnie, Inc., regarding UV reactors. Tucson, AZ, May 4.

Meng, Q.S., and C.P. Gerba. 1996. Comparative inactivation of enteric adenovirus, poliovirus
       and coliphages by ultraviolet irradiation.  Water Research 30(11):2665-2668.

Merck & Co., Inc. 1983. The Merck Index 10th Edition, edited by M. Windholz. Rahway, NJ.

Meulemans, C.C.E.  1986. The basic principles of UV-sterilization of water. In: Ozone +
       Ultraviolet Water Treatment, Aquatec Amsterdam, Paris: International Ozone
       Association. B.l.l-B.1.13.

Morton, R.A., and R.H. Haynes. 1969. Changes in the ultraviolet sensitivity of Escherichia coli
       during growth in batch cultures. Journal of Bacteriology 97(3): 1379-85

Moss, L. 2002a. Personal communication with Jennifer Hafer, Malcolm Pirnie, Inc., regarding
       UV reactors. Tucson, AZ, April 29.

Moss, L. 2002b. Personal communication with Jennifer Hafer, Malcolm Pirnie, Inc., regarding
       UV reactors. Tucson, AZ, November 18.

Munakata, N., and C.S. Rupert. 1972. Genetically controlled removal  of "spore photoproduct"
       from deoxyribonucleic acid of ultraviolet-irradiated Bacillus subtilis spores. Journal of
       Bacteriology 111:192-198.

Munakata, N., M. Daito, and K. Hieda. 1991. Inactivation spectra of Bacillus subtilis spores in
       extended ultraviolet wavelengths (50 - 300 nm) obtained with Synchrotron radiation.
       Photochemistry andPhotobiology 54:761-768.

Mutti, D. 2004.  Personal communication with Amy Samuelson, Malcolm Pirnie, Inc., regarding
       UV reactors. Indianapolis,  IN, October 26.

Nieuwstad, T.J., and A.H. Havelaar. 1994. The kinetics of batch ultraviolet inactivation of
       bacteriophage MS2 and microbiological calibration of an ultraviolet pilot plant. Journal
       of Environmental Science and Health A29(5): 1993-2007.

UV Disinfection Guidance Manual                 7-9                               November 2006
For the Final LT2ESWTR

-------
                                   Chapter 7. Bibliography
NSF. 2003. NSF/ANSI 60, Drinking Water Treatment Chemicals—Health Effects. Ann Arbor:
       National Sanitation Foundation

NSF. 2004. NSF/ANSI 55, Ultraviolet Microbiological Water Treatment Systems. Ann Arbor:
       National Sanitation Foundation.

NWRI. 2003. Ultraviolet Disinfection Guidelines for Drinking Water and Water Reuse. Second
       Edition. National Water Research Institute, Fountain Valley, CA, in collaboration with
       American Water Works Association Research Foundation.

OSHA. 2006. Mercury Vapor in Workplace Atmospheres.
       http://www.osha. gov/dts/sltc/methods/inorganic/id 140/id 140.html

OgumaK., H. Katayama, H. Mitani, S. Morita, T. Hirata, and S. Ohgaki. 2001. Determination of
       pyrimidine dimers in Escherichia coli and Cryptosporidium parvum during UV light
       inactivation, photoreactivation, and dark repair. Applied and Environmental
      Microbiology 67:4630-4637.

Oguma, K., H. Katayama, and S. Ohgaki. 2002. Determination of pyrimidine dimers in the
       genomic DNA of Escherichia coli during photoreactivation following inactivation by
       medium-pressure UV lamp. Presented at the 2nd IWA International Congress,
       Melbourne, Australia.

Oliver, E.G.,  and E.G. Cosgrove. 1975. The disinfection of sewage treatment plant effluents
       using ultraviolet light. Canadian Journal of Chemical Engineering 53:170-174.

ONORM. 2001. Plants for the disinfection of water using ultraviolet radiation—Requirements
       and testing—Part 1: Low pressure mercury lamp plants. ONORM M 5873-1.
       Osterreichisches Normungsinstitut, Vienna, Austria.

ONORM. 2003. Plants for the disinfection of water using ultraviolet radiation—Requirements
       and testing—Part 2: Medium pressure mercury lamp plants. ONORM M 5873-2.
       Osterreichisches Normungsinstitut, Vienna, Austria.

Oppenheimer, J.A., I.E. Hoagland, J.-M. Laine, J.G. Jacangelo, and A. Bhamrah. 1993.
       Microbial inactivation and characterization of toxicity and by-products occurring in
       reclaimed wastewater disinfected with UV radiation. Water Environment Federation
       (WEF) Specialty Conference: Planning, Design & Operation of Effluent Disinfection
       Systems. Whippany, NJ.

Oppenheimer, J., T. Gillogly, G. Stolarik, and R. Ward. 2002. Comparing the efficiency of low
       and medium pressure UV light for inactivating Giardia muris and Cryptosporidium
      parvum in waters with low and high levels of turbidity. Proceedings of the Annual
       Conference and Exhibition of the American Water Works Association, New Orleans, LA,
       June 16-20.
UV Disinfection Guidance Manual                7-10                              November 2006
For the Final LT2ESWTR

-------
                                   Chapter 7. Bibliography
Oppenheimer, J., M. Heath, and B. Gallagher. 2003. Impact of chemical disinfectant residuals on
       UV inactivation of viruses. Presented at the American Water Works Association Water
       Quality Technology Conference, Philadelphia, PA, November 2-6.

Ormeci, B., and K.G. Linden. 2003. Inactivation and detection of particle-associated
       microorganisms. Proceedings of the 2nd International Congress on Ultraviolet
       Technologies, July 9-11, Vienna, Austria.

Ormeci, B., JJ. Ducoste, and K.G. Linden. 2005. UV disinfection of chlorinated water: Impact
       on chlorine concentration and UV dose delivery. Journal of Water Supply, Research, and
       Technology - Aqua 54(3).

Parker, J.A., and J.L. Darby. 1995. Particle-associated coliform in secondary effluents: Shielding
       from ultraviolet light disinfection. Water Environment Research 67:1065-1075.

Passantino, L., J. Malley, M. Knudson, R.  Ward, and J. Kim. 2004. Effect of low turbidity and
       algae on UV disinfection performance. Journal of the American Water Works Association
       96:128-137.

Peldszus, S., S.A. Andrews, R.  Souza, F. Smith, I. Douglas, J. Bolton, and P.M. Huck. 2004.
   Effect of medium-pressure UV irradiation on bromate concentrations in drinking water, a
   pilot-scale study. Water Research 38:211-217.

Petri, B.M., G. Fang, J.P. Malley, D.C. Moran, and H. Wright. 2000. Ground water UV
       disinfection: Challenges and solutions. Proceedings of the American Water Works
       Association Water Quality Technology Conference. Salt Lake City, UT, November 5-9.

Pfeifer, G.P. 1997. Formation and processing of UV photoproducts: Effects of DNA sequence
       and chromatin environment. Photochemistry & Photobiology 65:270-283.

Phillips, R. 1983. Sources and Applications of Ultraviolet Radiation. New York: Academic Press
       Inc.

Polymicro Technologies. 2004. http://www.polymicro.eom/catalog/2 21.htm

Powell, W.F. 1959. Radiosensitivity as an  index of Herpes simplex virus development. Virology
       9:1-19.
Primarc Limited. 2001. Primarc UV Technology. September 5.  www.primarc.com

Quails, R.G., and J.D. Johnson. 1983. Bioassay and UV dose measurement in UV disinfection.
       Applied and Environmental Microbiology 45(3):872-877.

Rauth, A.M. 1965. The physical state of viral nucleic acid and the sensitivity of viruses to
       ultraviolet light. Biophysical Journal 5:257'-273.

Region of Waterloo. 2004. Personal communication with Amy  Samuelson, Malcolm Pirnie, Inc.,
       regarding UV reactors. Indianapolis, IN, November 3.

UV Disinfection Guidance Manual                7-11                             November 2006
For the Final LT2ESWTR

-------
                                   Chapter 7. Bibliography
Rice, J.K., and M. Ewell. 2001. Examination of peak power dependence in the UV inactivation
       of bacterial spores. Applied and Environmental Microbiology 67:5830-5832.

Roberts, A. 2000. Hanovia Corporation.  Letter to Dan Schmelling, U.S. Environmental
       Protection Agency. April 10.

Schulz, C.R. 2004. Camp, Dresser, & McKee. Personal communication with Laurel Passantino,
       Malcolm Pirnie, Inc., regarding UV installations. Phoenix, AZ, October 22.

Setlow, J.K. 1967. Comprehensive Biochemistry. Chapter 5 in The Effects of Ultraviolet
       Radiation andPhotoreactivation. 27:157-209.

Severin, B.F., M.T. Suldan, and R.S. Englebrecht. 1983. Effects of temperature on ultraviolet
       light disinfection. Environmental Science and Technology 17:717-721.

Severn Trent Services. 2006. http://www.severntrentservices.com.

Sharpless, C.M., and K.G. Linden. 2001. UV photolysis of nitrate: Effects of natural organic
       matter and dissolved inorganic carbon, and implications for UV water disinfection.
       Environmental Science and Technology 35:2949-2955.

Shin, G.A., K.G. Linden, MJ. Arrowood, G. Faubert, and M.D. Sobsey. 2001. DNA repair of
       UV-irradiated Cryptosporidiumparvum oocysts and Giardia lamblia cysts. Proceedings
       of the First International Ultraviolet Association Congress, Washington, DC, June 14 -
       16.

Snowball, M.R., and IS. Hornsey. 1988. Purification of water supplies using ultraviolet light. In
       Developments in Food Microbiology, ed. R.K. Robinson. New York: Elsevier Applied
       Science. 171-191.

Sommer, R., and A. Cabaj. 1993. Evaluation of the efficiency of a UV plant for drinking water
       disinfection. Water Science and Technology 27(7-8):357-362.

Sommer, R., A. Cabaj, D. Schoenen, J. Gebel, A. Kolch, A.H. Havelaar, and F.M. Schets.  1995.
       Comparison of three laboratory devices for UV-inactivation of microorganisms. Water
       Science and Technology 31:147-156.

Sommer, R., A. Cabaj, and T. Haider. 1996. Microbiocidal effect of reflected UV radiation in
       devices for water disinfection. Water Science and Technology 34(5-6): 173-177.

Sommer, R., T. Haider, A.  Cabaj, W. Pribil, and M. Lhotsky. 1998. Time  dose reciprocity in UV
       disinfection of water. Water Science and Technology 38:145-150.

Sommer, R., T. Haider, A.  Cabaj, W. Pribil, and M. Lhotsky. 1998. Time  UV dose reciprocity in
       UV disinfection of water. International Association of Water Quality (IAWQ) Vancouver
       1998 Poster.

UV Disinfection Guidance Manual                7-12                              November 2006
For the Final LT2ESWTR

-------
                                   Chapter 7. Bibliography
Srikanth, B. 200 la. Aquafme Corporation. Letter to James Malley, University of New
       Hampshire, regarding accidental UV lamp breakage and potential scenarios and
       prevention strategies. June 13.

Srikanth, B. 2001b. Aquafme Corporation. Personal communication with Jennifer Hafer,
       Malcolm Pirnie, Inc., regarding UV reactors. Tucson, AZ, June 21.

Stantec Consulting Ltd. 2004. Mannheim UV Emergency Response Procedure. Prepared for the
       Regional Municipality of Waterloo.

Swaim, P.O., M.A. Morine, R.G. Brauer,  M.A. Neher, J.L. Gebhart, and W.D. Bellamy. 2002.
       Operating data from Henderson's  full-scale UV disinfection facility. Proceedings from
       the American Water Works Association Water Quality Technology Conference,
       November 10-14, Seattle, WA.

Swaim, P.D., S.B. Jeffcoat, M.L. Gaylor,  and W.D. Bellamy. 2003. New challenges of on-site
       UV validation at a WTP with multiple disinfection barriers. American Water Works
       Association Water Quality Technology  Conference, Philadelphia, PA, November 2-6.

Taylor, J.R. 1982. An Introduction to Error Analysis. 2nd Edition, University Science Books,
       Sausalito, CA.

Templeton, M., R. Hofmann, R.  Andrews, and E. Whitby. 2003. UV inactivation of floe-
       associated MS-2 coliphage. American Water Works Association Water Quality
       Technology Conference,  Philadelphia, PA, November 2-6.

Templeton, M., R.C. Andrews, and R. Hoffman. 2004. Particle characteristics influencing the
       UV disinfection of drinking water. Presented at the American Water Works Association
       Water Quality Technology Conference, San Antonio, TX, November 14 - 18.

Thompson, S.S., andM.V. Yates. 1999. Bacteriophage inactivation at the air-water-solid
       interface in dynamic batch systems. Applied and Environmental Microbiology
       65(3): 1186-1190.

Toivanen, E. 2000. Experiences with UV disinfection in Helsinki water. IUVA News. 2(6):4-8.

Tramposch, W. 2004. Personal communication with Amy Samuelson, Malcolm Pirnie, Inc.,
       regarding UV reactors. Indianapolis, IN, October 14.

Tree, J.A., M.R. Adams, and D.N. Lees. 1997. Virus inactivation during disinfection of
       wastewater by chlorination and UV irradiation and the efficacy of f+ bacteriophage as a
       viral indicator.  Water Science and Technology 35:227-232.

Trojan Technologies. 2003. http://www.troianuv.com/en/homeframe.htm
UV Disinfection Guidance Manual                7-13                              November 2006
For the Final LT2ESWTR

-------
                                  Chapter 7. Bibliography
USEPA. 1992. Characterization of Products Containing Mercury in Municipal Solid Waste in
       the United States, 1970-2000. EPA 530-R-92-013, U.S. Environmental Protection
       Agency, Office of Solid Waste, Washington, DC.

USEPA. 1996. (2-1)

USEPA. 1991 z.Mercury-Emergency Spill & Release Facts. 540-K-97-004, OSWER 9378.0-
       10FS, PB97-963405, U.S. Environmental Protection Agency, Office of Emergency and
       Remedial Response, Washington, DC.

USEPA. 1997b. Mercury Emissions from the Disposal of Fluorescent Lamps. F-97-FLEA-
       S0002, U.S. Environmental Protection Agency, Office of Solid Waste, Washington, DC.

USEPA. 2003. Proposal Draft Ultraviolet Disinfection Guidance Manual. EPA 815-D-03-007.
       U.S. Environmental Protection Agency, Office of Groundwater and Drinking Water,
       Washington, DC

USEPA. 2006. Source Water Monitoring Guidance Manual for Public Water Systems for the
       Final Long Term 2 Enhanced Surface Water Treatment Rule. EPA 815-R06-005, U.S.
       Environmental Protection Agency, Office of Water, Washington, DC.

Venkatesan,  N., G. Hua, and D.A. Reckhow. 2003. Impact of UV disinfection on DBF formation
       from  subsequent chlorination. Proceedings of the American Water Works Association
       Water Quality Technology Conference, Philadelphia, PA, November 2-6.

von Sonntag, C., and H.P. Schuchmann. 1992. UV disinfection of drinking water and by-product
       formation: Some basic considerations. Journal of Water Supply Research and
       Technology 41:67- 4.

Wait, I, C. Johnston, A. Schwab, and E. Blatchley, III. 2005. The influence of oxidation-
       reduction potential on inorganic fouling of quartz surfaces in UV disinfection systems.
       Presented at the American Water Works Association Water Quality Technology
       Conference, Quebec City, Quebec, Canada, November 6-10.

Walitsky, P.  Philips Lighting. 2001. Personal communication with Jennifer Hafer, Malcolm
       Pirnie, Inc.,  regarding UV reactors. Tucson AZ, May 14.

WEDECO UV Systems. 2003.  http://www.uvsystems.co.uk/water/index.html.

Wilczak, A.  and H.  Lai. 2006. Preliminary bench and  pilot evaluation of UV-irradiation for
       nitrification control. Proceedings of the American Water Works Association Annual
       Conference. June 11-16, San Antonio, TX.

Wilson, B.R., P.P. Roessler, E. Van Dellen, M. Abbaszadegan, and C.P. Gerba. 1992. Coliphage
       MS-2 as a UV water disinfection efficacy test  surrogate for bacterial and viral pathogens.
       Proceedings of the Water Quality Technology  Conference, Toronto, November 15-19,
       pgs. 219-235.

UV Disinfection Guidance Manual               7-14                             November 2006
For the Final LT2ESWTR

-------
                                  Chapter 7. Bibliography
Wobma, P.C., W.D. Bellamy, J.P. Malley, and D.A. Reckhow. 2004. UVDisinfection and
       Disinfection By-poduct Characteristics ofUnfiltered Water. Denver, CO, American
       Water Works Association Research Foundation.

Wright, H. 2005. Personal communication with Christine Cotton, Malcolm Pirnie, Inc, regarding
       lamp breaks. January 24.

Wright, H.B. 2006. Carollo Engineers. Internal communication with Erin Mackey, Carollo
       Engineers.

Wright, H.B., and Y.A. Lawryshyn. 2000. An assessment of the bioassay concept for the UV
       reactor validation. Disinfection 2000: Disinfection of Wastes in the New Millennium,
       New Orleans, LA, March 15-18; Water Environment Federation, Alexandria, VA.

Wright, N.G., and D.M. Hargreaves. 2001.  The use of CFD in the evaluation of UV treatment
       systems. Journal of Hydroinformatics 3:59-70.

Wright, H.B., and E.D. Mackey. 2003. Rational approaches for defining safety factors for UV
       disinfection. Paper presented at the International Ultraviolet Association 2003 Annual
       Conference, Vienna, Austria, July 9-11.

Wright, H.D., and S.P. Reddy. 2003. Factors impacting the precision of CFD-based modeling of
       UV reactor performance. Proceedings of the AMERICAN WATER WORKS
       ASSOCIATION Annual Conference & Exposition, Anaheim, CA, June 15-19.

Wright, H.B., E.D. Mackey, and R.S. Gushing. 2002. Impact of UV intensity sensor properties of
       dose delivery compliance monitoring for monochromatic and polychromatic UV
       disinfection systems. Paper presented at the Water Environment Foundation/American
       Water Works Association 2002 Conference, St. Petersburg, FL, February 17-20.

Wright, H.B., E. Mackey, and P.  White. 2002. UV disinfection compliance monitoring for
       drinking water  applications. Proceedings of the Water Quality Technology Conference
       and Exhibition, Seattle, WA, November 10 - 14.

Wright, H.B., T.M. Dzurny, and D.J. Greene. 2004. Impact of UV sensor properties on dose
       delivery monitoring by UV disinfection systems:  Number of UV sensors in a multi-lamp
       UV reactor. Paper presented at the International Ultraviolet Association 2004 European
       Conference on  UV Radiation, Karlsruhe, Germany, September 22 - 24.

Wright, H.B., T.M. Dzurny, and A.T. Salveson. 2005. Cost and efficiency of dose pacing
       strategies used  by drinking water and wastewater UV systems. Proceedings of the Water
       Environment Federation Disinfection 2005 Conference, Phoenix, AZ, February 6-9.

Wright, H., Gaithuma,  D., Bircher, K., and Tramposch, W. 2005. Analyzing UV Reactor
       Validation Data to Verify UV Dose Delivery Using the UV Disinfection Guidance
UV Disinfection Guidance Manual                7-15                              November 2006
For the Final LT2ESWTR

-------
                                   Chapter 7. Bibliography
       Manual. Proceedings of American Water Works Association 2005 Water Quality
       Technology Conference, Quebec City, Quebec, Canada, November 6-10.

Wright, H.B., D. Gaithuma, T. Dzurny, C. Schulz, A. Cabaj, and A. Schmalweiser. 2006. Design
       and performance guidelines for UV sensor systems. American Water Works Association
       Research Foundation, Denver, CO (in preparation).

Yahya, M.T., T.M. Straub, and C.P.Gerba. 1992. Inactivation of coliphage MS-2 and poliovirus
       by copper, silver, and chlorine. Canadian Journal of Microbiology 3 8(6) :43 0-43 5.

Yip, R.W., and D.E. Konasewich. 1972. Ultraviolet sterilization of water: Its potential and
       limitations. Water and Pollution Control 14:14-18.

Zheng, M., S.A. Andrews, and J.R. Bolton. 1999. Impacts of medium pressure UV on THM and
       HAA formation in pre-UV chlorinated drinking water. American Water Works
       Association Water Quality Technology Conference, October 31 - November 3, Tampa,
       FL.
UV Disinfection Guidance Manual                7-16                              November 2006
For the Final LT2ESWTR

-------
      Appendix A

 Preparing and Assaying
Challenge Microorganisms

-------
This page intentionally left blank.

-------
                   Appendix A. Preparing and Assaying Challenge Microorganisms
       Sections A. 1 through A.4 describe procedures that can be used for preparing stock
solutions of male-specific-2 bacteriophage (MS2 phage) and Bacillus subtilis spores and
assaying the concentration of those microorganisms in water samples. Procedures for preparing
stock solutions can be scaled to provide the volumes needed for UV reactor validation.
Alternative procedures and challenge microorganisms can be used if they are acceptable to the
state.
A.1    MS2 Phage Stock Preparation

       MS2 phage (American Type Culture Collection [ATCC] 15597-B1) can be propagated
using a variety of host bacteria, including Escherichia coll C3000 (ATCC 15597); E. coli Famp
(ATCC 700891); and others (Meng and Gerba 1996, Oppenheimer et al. 1993, NWRI (2003).
The following propagation method was adapted from NWRI (2003).

       Procedure:

       1.  Inoculate sterile tryptic soy broth (TSB) (Difco, Detroit, Michigan) with host bacteria
          transferred from a single colony grown on a nutrient agar plate. Incubate the culture
          with constant stirring at 35 to 37 degrees Centigrade (°C) for 18 to 24 hours.

       2.  Transfer 0.5 milliliter (mL) of the host bacterial culture to 50 mL of fresh TSB and
          incubate at 35 to 37 °C for 4 to 6 hours with continuous shaking at 100 Hertz (Hz) to
          obtain a culture in its log growth phase (~3 x  108 cfu/mL, where cfu = colony
          forming unit).

       3.  Dilute stock MS2 phage using Tri-buffered saline (pH 7.3) to a concentration of
          -100 pfu/mL (pfu = plaque forming unit).

       4.  Add 1 mL of diluted MS2 phage stock solution to the 50-mL volume of E. coli in
          TSB and incubate overnight at 35 to 37 °C.

       5.  Centrifuge the MS2-E. coli culture at 8,000 x G [G = 9.82 meter per second squared
          (m/s2)] for 10 minutes at 4 °C to remove cellular debris.

       6.  Filter the supernatant by passing it through a 0.45-micrometer (um) low protein-
          binding filter.

       7.  Assay the concentration of MS2 phage in the  stock solution as described in
          Section A.2.

       8.  Collect and refrigerate the filtrate at 4 °C, and use within one month.

       Propagation should result in a highly concentrated stock solution of essentially mono-
dispersed phage whose UV dose-response follows second-order kinetics with minimal tailing.
Figure A. 1 presents the UV dose-response of MS2 phage as reported in the literature. Over the
range of reduction equivalent dose (RED) values demonstrated during validation testing, the

UV Disinfection Guidance Manual              A-l                                  November 2006
For the Final LT2ESWTR

-------
                  Appendix A. Preparing and Assaying Challenge Microorganisms
mean UV dose-response of the MS2 phage stock solution should lie within the 95-percent
prediction interval of the mean response in Figure A. 1.  Over a UV dose range of 0 to
120 millijoule per centimeter squared (mJ/cm2), the prediction intervals of the data shown in
Figure A. 1 may be defined using the following equations:
       Upper Bound:  log/ = -1.4x1 (T4 xUV Dose2 +7.6x1 O^xUVDose    Equation A.I
       Lower Bound:  log/ = -9.6 xKT* xUV Dose2 +4.5x10^ xUV Dose    Equation A.2
                   Figure A.1. UV Dose-response of MS2 Phage
         c
         o
         '^
         re
         '•&
         o
         re
         _c
         O)
         o
                       x
    Havelaaretal., 1990
    Meng and Gerba, 1996
    Nieuwstad and Havelaar, 1994
    Oppenheimer et al., 1993
    Sommeretal., 1998
    Treeetal., 1997
    Mean
- - Mean Response 90% Prediction Interval
                       20
          40        60       80

             UV Dose (mJ/cm2)
100
120
A.2   MS2 Phage Assay

      The concentration of MS2 phage (ATCC 15597-B1) in water samples can be assayed
using agar overlay technique with E. coli (ATCC 15597) as a host bacterium [(Adams (1959),
Yahya et al. (1992), Oppenheimer et al. (1993), and Meng and Gerba (1996)]. Each test sample
should be assayed in triplicate and the sample concentration calculated as the arithmetic average
of the three measured values.  The following procedure can be used.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
              A-2
         November 2006

-------
                   Appendix A. Preparing and Assaying Challenge Microorganisms
       Procedure:

       1.  Inoculate sterile TSB (Difco, Detroit, Michigan) with the host bacterium and incubate
          at 35 to 37 °C for 18 to 24 hours to obtain an approximate concentration of
          108 cfu/mL.

       2.  Transfer 1 mL of the host bacterial culture to 50 mL of fresh TSB and incubate at 35
          to 37 °C for 4 to 6 hours with continuous shaking at 100 Hz to obtain a culture in its
          log growth phase.

       3.  Obtain serial  dilutions of the MS2 phage sample using a 0.001-molar (M) phosphate-
          saline buffer or TSB.

       4.  Combine and gently stir 1 mL of host cell solution, 0.1 mL of diluted MS2 phage
          sample, and 2 to 3 mL of molten tryptic soy agar (TSA) (0.7 percent agar, 45 to
          48 °C) (Difco).

       5.  Pour the mixture onto solidified TSA (1.5 percent agar) contained in petri dishes.
          The time between mixing the MS2 phage sample with the E. coli host and plating the
          top agar layer should not exceed 10 minutes. After plating, the agar should harden in
          less than 10 minutes.

       6.  After the top  agar layer hardens, cover and invert the petri dishes, and incubate 16 to
          24 hours at 35 to 37 °C.

       7.  Count the plaques with the aid of a colony counter. Plaques are identified as clear
          circular zones 1 to 5 millimeter (mm) in diameter in the lawn of host bacteria.

       8.  Record the number of plaques per dish and the MS2 phage sample volume and
          dilution.  If individual plaques cannot be distinguished because of confluent growth,
          record the plate counts as "TNTC" (too numerous to count).

       9.  Calculate the MS2 phage concentration in the water samples:

                                 n
          Concentration = ^10^°  l'avg                                       Equation A.3
          where:

          FD   =   Dilution factor
          n;   =   Number of counts on each plate (cfu or pfu)
          V;   =   Volume of diluted sample used with each plate (mL)
UV Disinfection Guidance Manual             A-3                                  November 2006
For the Final LT2ESWTR

-------
                   Appendix A.  Preparing and Assaying Challenge Microorganisms
  Example A.I.  A water sample containing MS2 phage was diluted 10-, 100-, and 1,000-fold
  using a 0.1-mL aliquot dilution of the sample for each. Each dilution was assayed in triplicate
  and the average count from these three plate counts is the challenge microorganism
  corresponding to the applied UV dose. Plaque forming units observed on the plates were 2, 5,
  and 6 for the 1,000-fold diluted sample and 32, 40, and 47 for the 100-fold diluted sample.
  With the 10-fold dilution, plate counts were too numerous to count. The concentration in the
  original sample is calculated as:
                    (2 + 5 + 6) pfu
                        0.1 mL
(32 + 40 + 47) pfu
      0.1 mL
      Cone. = ±	*—^	'- = 4.15 x 104 pfu/mL
A. 3   Bacillus subtilis Spore Preparation

       B. subtilis spores (ATCC 6633) can be propagated using Schaeffer's medium [Munakata
and Rupert (1972), Sommer et al. (1995), and DVGW (2006)]. The following propagation
method was adapted from DVGW (2006).

       Procedure:

       1.  Prepare 1 liter (L) of Columbia agar (Oxoid CM 331) using 23.0 grams (g) special
          peptone (Oxoid L 72), 1.0 g starch, 5.0 g NaCl, and 10.0 g agar (Oxoid L 1 1) in
          phosphate-buffered (to pH 7) water. Autoclave for 15 minutes at 121 °C.
       2.  Prepare 1 L of sporulation medium using 280 milligrams (mg) MgSO^IrkO, 1 . 1 1 g
          KC1, 3. 1 mg FeSO4-7H2O, and 8.9 g nutrient broth (Oxoid CM 67) in phosphate-
          buffered (to pH 7) water. Autoclave for 15 minutes at 121 °C.

       3.  Inoculate Columbia agar (Oxoid CM 331) plates with three smears of B. subtilis and
          incubate 24 hours at 37 °C.

       4.  Inoculate 300 mL of sporulation medium with three colonies collected from the agar
          plates that were prepared in Step 3.

       5.  Incubate the sporulation medium for 72 hours at 37 °C on a shaker operating at 2 Hz.

       6.  Sonicate the resulting culture for 10 minutes at 50,000 Hz and 10 °C.

       7.  Harvest the spores by centrifuging 80-mL aliquots at 5,000 x G and 10 °C for
          10 minutes.
UV Disinfection Guidance Manual             A-4                                 November 2006
For the Final LT2ESWTR

-------
                    Appendix A. Preparing and Assaying Challenge Microorganisms
       8.  Wash the spores 3 times by re-suspending the pellet in 20 mL of distilled water and
          centrifuging at 5,000 x G for 10 minutes at 10 °C.

       9.  Re-suspend the washed spores in 100 mL of 0.001-M phosphate-saline buffer.

       10. Inactivate the vegetative B. subtilis by heating at 80 °C for 10 minutes.

       11. Sonicate the resulting culture for 10 minutes at 50,000 Hz and 10 °C.

       12. Collect the resulting stock solution and assay the B. subtilis spore concentration as
          described in Section A.4.

       13. Refrigerate at 4 °C and use within one month (unless stability over longer periods of
          time can be substantiated).  Sonicate for 10 minutes at 50,000 Hz and 10 °C before
          use.

       Propagation should result in a highly concentrated stock solution of mono-dispersed
B. subtilis spores with a UV dose-response that follows the UV dose-response curves reported in
the literature and presented in Figure A.2. Over the range of RED values demonstrated during
validation testing, the mean UV dose-response of the B. subtilis stock solution should lie within
the 90-percent prediction interval of the mean response provided in Figure A.2.  Over a UV dose
range of 0 to 70 ml/cm2, the prediction intervals of the data shown in Figure A.2 are defined
using the following equations:

Upper Bound:   log / = -2.0 x 1CT5 x UV Dose* + 2.7 x 1CT3 x UV Dose2 - 5.3 x 1CT2 x UV Dose
                                                                            Equation A.4

Lower Bound:  log / = 5.7xl(T4 xUVDose2 +4.3xl(T2 xUVDose            Equation A.5
UV Disinfection Guidance Manual             A-5                                   November 2006
For the Final LT2ESWTR

-------
                  Appendix A. Preparing and Assaying Challenge Microorganisms
               Figure A.2. UV Dose-Response of B. subtilis Spores
             c
             o
            ^
             re
            ^
             o
             re

             O)
             o
 7 -

 6

 5

 4

 3

 2

 1

 0

-1
                       A   DVGW, 1997
                       n   Chang et al,  1985
                       o   Hoyer, 2002
                       +   Sommer and Cabaj, 1993 (curve 1)
                           Sommer and Cabaj, 1993 (curve 2)
                       •   Sommer et al, 1995
                       x   Sommer et al, 1996
                       x   Sommer et al, 1998
                           Mean
                       -  - 90% Prediction Interval of the Mean Response
                          20
                   40
60
80
100
                                       UV Dose (mJ/cm2)
120
140
A.4  Bacillus subtilis Spore Assay


      The concentration of B. subtilis spores (ATCC 6633) in water samples can be assayed
using plate count agar. As with MS2 phage, each test sample should be assayed in triplicate and
the sample concentration calculated as the arithmetic average of the three measured values. The
following procedure was adopted from DVGW (2006).

      Procedure:

      1.  Prepare 1 L of plate-count agar (Oxoid CM 325) using 5.0 g casein peptone (Oxoid L
          42), 2.5 g yeast extract (Oxoid L 21), 1.0 g glucose, and 9.0 g agar (Oxoid L 11) in
          distilled water. Adjust the pH to 6.8 ± 0.2 and autoclave for 15 minutes at 121 ° C.

      2.  Obtain serial dilutions of the B. subtilis spore sample using 0.001-M phosphate-saline
          buffer.

      3.  Vacuum filter  100 mL of diluted sample through a 47-mm 0.45-|im membrane filter.

      4.  Place filter on a petri dish containing hardened agar and cover plates.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                      A-6
                              November 2006

-------
                    Appendix A. Preparing and Assaying Challenge Microorganisms
       5.  Incubate plates 24 ± 2 hours at 37 ± 1 °C.

       6.  Count the number of colonies formed with the aid of a colony counter.

       7.  Record the number of colonies per dish, and the B. subtilis spore sample volume and
          dilution. If individual colonies cannot be distinguished because of confluent growth,
          record the plate counts as TNTC.

       8.  Calculate the B. subtilis spore concentration in the original samples in units of cfu/mL
          using Equation A.3.
UV Disinfection Guidance Manual             A-7                                  November 2006
For the Final LT2ESWTR

-------
                     Appendix A.  Preparing and Assaying Challenge Microorganisms
                                This page intentionally left blank
UV Disinfection Guidance Manual              A-8                                     November 2006
For the Final LT2ESWTR

-------
       Appendix B



UV Reactor Testing Examples

-------
This page intentionally left blank

-------
                           Appendix B. UV Reactor Testing Examples
       This appendix presents two validation data analysis examples. Section B.I presents an
example for the UV Intensity Setpoint Approach using single-setpoint operations. Section B.2
presents a more complex example for a UV reactor that uses the Calculated Dose Approach.
These two examples bracket a wide range of complexities that UV reactor validation testing can
encompass. All information and data are hypothetical but are representative of real validation
data in terms of selection of test conditions and variability in measured values.
B.1    Example 1 - Validation for the UV Intensity Setpoint Approach (a Single
Setpoint and a Single Disinfection Goal)

       System X plans to add UV disinfection to their treatment process to earn 2.5-log
Cryptosporidium inactivation credit. Based on the LT2ESWTR dose requirements as
summarized in Table 1.4 of this manual, System X needs to deliver a minimum required dose of
8.5 ml/cm2 to receive this level of inactivation credit. The hypothetical proposed installation has
the following design specifications:
Design flow rate
Minimum UVT
Lamp aging factor
Fouling factor
Fouling/aging factor
Disinfection goal
400 gpm
90%
80%
85%
68 % (80 % x 85 %)
2.5-log Cryptosporidium inactivation credit
       The water system's engineer selected a UV reactor (illustrated in Figure B. 1) with the
following characteristics:
UV Reactor
UV Dose-Monitoring Approach
-2 banks of lamps
- 4 300-W LPHO lamps per bank
- Rated for flow rates of 50 - 500 gpm
- 1 UV senso^ank positioned equidistant from Lamps 2 & 3
UV Intensity Setpoint Approach with one alarm setpoint
UV Disinfection Guidance Manual
For the Final LT2ESWTR
B-l
November 2006

-------
                           Appendix B. UV Reactor Testing Examples
          Figure B.1 Schematic of Hypothetical UV Reactor for Example 1

                                            UV Sensors
                                          K	X
                                            LPHO
                                           Lamps
B.1.1  Validation Test Plan

       The validation test plan was developed using Checklist 5.2 in Chapter 5. Key elements
defined in the validation test plan include:

          The UV manufacturer provided three reference UV sensors for calibrating the duty
          sensors during validation. These reference sensors had been calibrated by an
          independent, qualified sensor testing laboratory before validation and had a
          documented measurement uncertainty of 10 percent (Section 5.5.4)

       .   Because the number of UV sensors was less than the number of lamps, the highest-
          output lamps were identified prior to testing and were positioned closest to the
          sensors at lamp positions 2  and 3 (Section 5.4.7). New lamps were used during
          validation testing (after a 100 hour burn-in period).

       .   The validation testing was conducted over a one-day period. The UV dose-response
          of the challenge microorganism (measured via a laboratory collimated beam test) was
          evaluated with 1-L influent water samples collected at high and low UVT values
          (Section C.I).

       .   All recommended testing protocols as listed in Section 5.7 and Appendix C were
          followed.

       The UV manufacturer had already identified a target setpoint (11.7 mW/cm2) using
numerical modeling. System X confirmed with the manufacturer that this setpoint is low enough
to account for their combined conditions of minimum UVT and maximum lamp fouling and
aging. The following two UVT-lamp power operating conditions were tested:

       1.  The UVT was lowered to produce the target UV sensor setpoint (in this case, the
           resultant UVT was 89.9%), while the lamp power was kept at 100%.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
B-2
November 2006

-------
                           Appendix B. UV Reactor Testing Examples
       2.  The UVT was raised back to its maximum value (no UV-absorbing chemical added),
           and the lamp power was reduced to produce the same UV sensor value (in this case,
           the resultant lamp power was 66%).

B.1.2  Test Data

       The data collected during validation testing are presented in Tables B.I through B.5 and
are described below:

       .   Table B.I presents the UV dose-response data measured on the influent water
          collected during field validation testing for the laboratory  collimated beam test.

       .   Tables B.2 through B.4 present the data from full-scale reactor testing

          -  Table B.2 presents the flow rate, UVT, lamp power, and UV sensor readings
             measured for each test condition.

             Table B.3 presents the measured challenge microorganism concentrations for the
             influent and effluent samples collected (in triplicate) from the UV reactor for each
             test condition.
             Table B.4 presents the UV output of the eight lamps used during validation testing
             measured at the same lamp location (Lamp #2 in Row #1) adjacent to the same
             UV sensor (#1), which was used to identify the highest output lamps.

       .   Table B.5 presents data comparing the three reference UV sensor measurements to
          the duty UV sensors used during validation.

       Sections B. 1.3 to B. 1.8 show how the data will be used to determine whether QA/QC
criteria are met, to calculate the necessary correction factors, and to determine the validated
operating conditions for the target log inactivation.

         Table B.1 Challenge Microorganism UV Dose-response Measured
                        Using a Collimated Beam Apparatus
90% UVT
UV Dose
(mJ/cm2)
0
10
20
30
40
60
80
100
Replicate #1
N
(pfu/mL)
882329
180120
64217
20622
7257
1274
188
80
Log
N
5.95
5.26
4.81
4.31
3.86
3.11
2.27
1.90
Replicate #2
N
(pfu /mL)
944980
198394
69438
20100
8145
1399
261
90
Log
N
5.98
5.30
4.84
4.30
3.91
3.15
2.42
1.95
97% UVT
UV Dose
(mJ/cm2)
0
10
20
30
40
60
80
100
Replicate #1
N
(v/mL)
1148154
316328
113644
34679
12624
1980
387
80
Log
N
6.06
5.50
5.06
4.54
4.10
3.30
2.59
1.90
Replicate #2
N
(pfu /mL)
1300460
257749
74396
25189
9226
1722
211
100
Log
N
6.11
5.41
4.87
4.40
3.97
3.24
2.32
2.00
UV Disinfection Guidance Manual
For the Final LT2ESWTR
B-3
November 2006

-------
                         Appendix B. UV Reactor Testing Examples
           Table B.2 Flow Rate, DVT, Lamp Power, and UV Sensor Data
                      Measured during Validation Testing
Test ID
1
2
Banks On
1,2
1,2
Flow Rate
(gpm)
394
403
UVT
(%)
89.9
97.0
Relative Lamp Output
(%)
100%
66%
Sduty, #1
(mW/cm2)
11.7
11.6
Sduty, #2
(mW/cm2)
11.7
11.7
       Table B.3 Measured Influent and Effluent Challenge Microorganism
                                Concentrations
Test ID
1
2
Influent Challenge
Microorganism
Log Concentration
Replicate #
1
5.94
6.01
2
6.00
5.99
3
5.84
6.04
Effluent Challenge
Microorganism
Log Concentration
Replicate #
1
4.57
4.10
2
4.54
4.09
3
4.56
4.06
Table B.4 Sensor #1 Measurements with Lamp #2 Operated at 100-percent Ballast
                                    Power
Lamp ID
1
2
3
4
Sduty #1
(mW/cm2)
13.6
14.6
14.2
13.4
Lamp ID
5
6
7
8
Sduty #1
(mW/cm2)
13.9
13.3
14.5
14.3
                     Table B.5 Reference UV Sensor Checks
Before/After
Validation
Testing
Before
Before
Before
Before
After
After
After
After
UVT
(%)
97
97
90
90
97
97
90
90
Relative
Lamp Power
(%)
100
68
100
68
100
68
100
68
Sensor
ID
1
1
2
2
1
1
2
2
Sduty
(mW/cm2)
11.3
5.1
3.7
2.0
11.6
5.1
3.9
1.9
Sref, #1
(mW/cm2)
11.7
5.5
4.0
1.9
11.8
5.4
4.0
1.8
Sref, #2
(mW/cm2)
12.1
5.7
4.1
1.9
12.2
5.6
4.1
2.0
Sref, #3
(mW/cm2)
11.4
5.3
3.8
1.8
11.4
5.3
3.9
1.8
B.1.3 Develop the UV Dose-response Curve from the Collimated Beam Data

      Figures B. l(a) and (b) present the UV dose-response data from Table B. 1 at UVT values
of 90 and 97 percent, respectively. The data have been fitted to quadratic equations that show log
N as a function of UV dose. The fits were used to identify log N0 values of 5.91 and 6.05 (i.e.,
UV Disinfection Guidance Manual
For the Final LT2ESWTR
B-4
November 2006

-------
                             Appendix B. UV Reactor Testing Examples
log No where the curves intersect the y-axes) from Figures B. l(a) and (b), respectively. Using
these values, Table B.6 presents the UV dose-response data defined as UV dose versus log I,
where  log I = log(N0/N).

       The UV dose-response data described in Table B.6 were analyzed to determine if the two
datasets could be combined using the method referred to in Section C.5 (Draper and Smith
1998).l A statistical analysis of the collimated beam data is recommended to determine which
terms are significant, p-value < 0.05, using a standard regression tool. The process is iterative,
and each time the regression tool is used,  one term is dropped until all of the coefficients are
deemed significant (p-values < 0.05). In this example, three iterations were needed.

       The regression analysis showed that the two measured UV dose-response curves were
statistically similar (p < 0.05) and could be combined. Figure B.2 presents the plot of UV dose as
a function of log  inactivation for the combined dataset and the resultant UV dose-response
equation.
1 The datasets should be combined whenever possible to develop one UV dose-response equation for calculating all
  RED values. The inability to combine datasets indicates a problem may have occurred with either the calculation
  or the test. Details are provided in Section C.5.
UV Disinfection Guidance Manual                 B-5                                 November 2006
For the Final LT2ESWTR

-------
                          Appendix B. UV Reactor Testing Examples
           Figure B.1  Log N Versus UV Dose Using the Data in Table B.1

                   at  (a) 90 Percent UVT and (b) 97 Percent UVT
      O)
      o
        O)
        o
7.00



6.00



5.00



4.00



3.00



2.00



1.00



0.00
                        Log N = 0.0002(UV Dose)" - 0.059(UV Dose) + 5.92
                        20
                        40
60
80
 100
                                  UV Dose (mJ/cm2)


                                        (a)
                      Log N = 0.0002(UV Dose)2 - 0.057(UV Dose) + 6.05
                         20
                         40
60
80
100
                                  UV Dose (mJ/cm2)


                                        (b)
UV Disinfection Guidance Manual

For the Final LT2ESWTR
                              B-6
                          November 2006

-------
                          Appendix B. UV Reactor Testing Examples
        Table B.6 Challenge Microorganism UV Dose-response Defined as
                      UV Dose Versus [Log (N0/N)] (i.e., Log I)
90% UVT
UV Dose
(mJ/cm2)
0
10
20
30
40
60
80
100
Replicate
#1 #2
5.92 - Log N
-0.03
0.66
1.11
1.61
2.06
2.81
3.65
4.02
-0.06
0.62
1.08
1.62
2.01
2.77
3.5
3.97
97% UVT
UV Dose
(mJ/cm2)
0
10
20
30
40
60
80
100
Replicate
#1 #2
6.05 - Log N
-0.01
0.55
0.99
1.51
1.95
2.75
3.46
4.15
-0.06
0.64
1.18
1.65
2.08
2.81
3.73
4.05
           Figure B.2 Log I Versus UV Dose Using the Data in Table B.6
            100
                                     RED (mJ/cm2) = 2.18(log I)2 + 15.30(log I)
               0.00        1.00       2.00       3.00
                                         Log I
                                             4.00
5.00
B.1.4 Verify That QA/QC Criteria Are Met

      Checklist 5.4 was used to ensure that recommended QA/QC criteria were met.
Calculations of key uncertainties are provided in the next two subsections.
B.1.4.1
Collimated Beam Data Uncertainty
      The uncertainty in the UV dose calculation using the collimated beam data is calculated
according to Equation C.6, shown below as Equation B.I:
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                           B-7
   November 2006

-------
                           Appendix B. UV Reactor Testing Examples
         DR
           = t-
                  SD
              UV DoseCB
             :100%
Equation B.I
       where
       UDR
       UV DosecB
       SD

       t
        Uncertainty of the UV dose-response fit at a 95-percent confidence level
        UV dose calculated from the UV dose-response curve in Figure B.2
        Standard deviation of the difference between the calculated UV dose-
        response and the measured values from Table B. 1
        t-statistic at a 95-percent confidence level for a sample size equal to the
        number of test condition replicates used to define the dose-response
       In this case, SD = 2.2 at 1-log inactivation and t = 2.04 for 32 test condition replicates as
shown in Table B.6. Equation B. 1 can then be used to determine UDR at various log inactivation
values from Table B.I. The graph below shows the relationship between log inactivation and
UDR. The value of UDR should not exceed 30 percent at the UV dose corresponding to 1-log
inactivation of the challenge organism. In this case, UDR = 25 percent at 1.0-log inactivation.
This value is less than the recommended limit of 30 percent.
                                    0
B.1.4.2
UV Sensor Uncertainty
       Guidance in Section 5.5.4 recommends that the manufacturer's reported sensor
uncertainty be confirmed as being less than 10 percent. Three reference sensors were used to
confirm duty sensor measurements. Data analyses are shown in the table on the next page (using
the data in Table B.5). The two duty UV sensors were within  10 percent of the average readings
from three reference sensors.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                             B-8
November 2006

-------
                           Appendix B. UV Reactor Testing Examples
Before/
After
Validatio
n Testing
Before
Before
Before
Before
After
After
After
After
UVT
97
97
90
90
97
97
90
90
Lamp
Power
(kW)
100
68
100
68
100
68
100
68
Sensor
ID
1
1
2
2
1
1
2
2
Sduty
(W/rrO
11.3
5.1
3.7
2
11.6
5.1
3.9
1.9
SRef,#l
(W/m2)
11.7
5.5
4
1.9
11.8
5.4
4
1.8
SRef,#2
(W/m2)
12.1
5.7
4.1
1.9
12.2
5.6
4.1
2.0
SRef,#3
(W/m2)
11.4
5.3
3.8
1.8
11.4
5.3
3.9
1.8
SRef,avg
(W/m2)
11.7
5.5
4.0
1.9
11.8
5.4
4.0
1.9

-------
                            Appendix B. UV Reactor Testing Examples
       Per guidance provided in Section 5.9.1, The UV sensitivity (RED / Log I) for each test
replicate was calculated for the test with the lowest UVT value (in this case, Test 1). The RED
bias for 2.5-log Cryptosporidium inactivation credit at the minimum UVT of 90 percent was
determined to be 1.79 using Table G.4.  Data for this analysis are summarized below.
Test#
1.1
1.2
1.3
RED
(mJ/cm2)
25.1
27.0
23.2
Log I
1.37
1.46
1.28
UVT1
89.9
89.9
89.9
Sensitivity
(mJ/cm2 per log I)
18.3
18.5
18.1
BRED
1.79
            Note: rounding off to match the number of significant figures in Table G.4,
            to determine the value of BRED the UVT value measured for Test #1
            becomes 90 percent.
B.1.6.2
Calculate the Uncertainty of Validation
       The decision tree in Figure 5.4 was used to identify the correct equation for UVAL. As
shown in Section B. 1.4.1, UDR is less than or equal to 30 percent. As noted in Section B.I.4.2,
Us is less than or equal to 10 percent. Therefore, the equation for UVAL is as follows:
                                                                             EquationB. 3
       USP is defined by Equation 5.14
       u=-
             txSD
                   RED
         SP
                RED
           :100%
                                                      EquationB.4
       where
           t  =
       RED
    t-statistic for the number of replicates
    the standard deviation for the RED calculations (Table B.7)
    the RED at the specific test condition used for the SDRED
The value for t is 3.18 for 3 test replicates. The highest SDRED and associated RED should be
used in this calculation. Using data from Table B.7, data from test condition #1 should be used as
follows:
       USP =
3.18x1.9
  25.1
x 100% = 24.1%
Using Equation B.3, UVAL = 24.1%
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                             B-10
                                                      November 2006

-------
                          Appendix B. UV Reactor Testing Examples
B.1.6.3
Calculate the Validation Factor
       The value of VF can now be calculated using Equation B.2:

       VF = 1. 79 x (1 + 24.1/100) =2.22
B.1.7  Calculate the Validated Dose

       In the step, the minimum RED (in this case, the average RED from test condition #1,
shown in Table B.7) is divided by the VF to calculate the validated dose using Equation 5.16:
                                                                         Equation B. 5
              2.22
B.1.8 Assign Log Inactivation Credit Based for the Validated Dose

      The validated dose must be greater than or equal to the required UV dose (Dreq) to
achieve a given level of pathogen inactivation credit:
              'req
                                                                       EquationB. 6
       In this case, 11.3 ml/cm2 is greater than the required dose of 8.5 mJ/cm2for 2.5-log
inactivation of Cryptosporidium. The UV reactor can receive 2.5-log inactivation credit for an
installation (with adequate inlet/outlet hydraulics, see Section 5.4.5) that operates under the
following criteria (Table B.8):
              Table B.8 Validated Dose and Operating Conditions for
               2.5-log Cryptosporidium Inactivation Credit Using the
                   Hypothetical UV Reactor Tested in Example 1
UV Sensor Setpoint
11.7 mW/cm2
Lamp Status
All lamps should be turned on during
reactor operations
Flow Rate Range
Q<394gpm
Dval
> 1 1 .3 mJ/cm2
B.2    Example 2 - Validation for the Calculated Dose Approach

       System Y plans to add UV disinfection to their treatment plant to earn 2.0-log
Cryptosporidium inactivation credit. Based on the LT2ESWTR UV dose requirements as
summarized in Table 1.4 of this manual, System Y needs a minimum germicidal dose of 5.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                           B-ll
November 2006

-------
                            Appendix B. UV Reactor Testing Examples
ml/cm2 to receive this level of inactivation credit. The hypothetical proposed installation has the
following design specifications:
Design flow range
Design UVT range
Lamp aging factor
Fouling factor
Fouling/ Aging factor
Disinfection goal
3- lOmgd
87 - 93 %
80%
85%
68 % (80 % x 85 %)
2.0-log Cryptosporidium inactivation credit
       The water system's engineer selected a UV reactor with the following characteristics:
UV Reactor
UV Dose-Monitoring Approach
- 1 bank of lamps
- 6 8-kW MP lamps per bank
- Ballast power settings range from 40-100 %
- Rated for flow rates of 2.5 - 10 mgd
- 1 germicidal UV sensor per lamp
Calculated Dose Approach with dose pacing
B.2.1  Validation Test Plan

       The test plan was developed using Checklist 5.2 in Chapter 5 to identify a range of target
RED values at different flow rate-lamp output-UVT combinations for 1.0- to 3.0-log
Cryptosporidium inactivation credit (depending on water quality and  operating conditions). The
UV manufacturer used modeled predictions of UV reactor performance to develop the desired
validation test conditions. The UV manufacturer selected test conditions that target RED values
ranging from approximately 4-43 ml/cm2 at UVT values of 85, 90, and 95 percent. Lamp
power was to be adjusted during testing of the UV reactor to give RED values within the target
range. Test flow rates of 2.5 - 10 mgd were selected in order to test the full design flow range of
the UV reactor. This information is summarized in Table B.9 below.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
B-12
November 2006

-------
                            Appendix B. UV Reactor Testing Examples
                         Table B.9 Validation Test Conditions
Test ID
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
UVT
(%)
95
95
95
95
95
95
95
95
95
90
90
90
90
90
90
90
90
90
85
85
85
85
85
85
85
85
85
Flow Rate
(mgd)
10
5
2.5
10
5
2.5
10
5
2.5
10
5
2.5
10
5
2.5
10
5
2.5
10
5
2.5
10
5
2.5
10
5
2.5
Relative Lamp Output
(%)
100
100
100
70
70
70
40
40
40
100
100
100
70
70
70
40
40
40
100
100
100
70
70
70
40
40
40
Predicted
RED3
(mJ/cm2)
14.1
24.6
42.8
11.8
20.6
35.8
8.9
15.6
27.1
7.9
13.8
24.1
6.6
11.6
20.1
5.0
8.7
15.2
4.5
7.8
13.5
3.7
6.5
11.3
2.8
4.9
8.6
       Other key test elements in the validation test plan include:

       .   The UV manufacturer provided three reference UV sensors for calibrating the duty
          UV sensors during validation. These reference UV sensors had been calibrated
          previously by an independent, qualified sensor testing laboratory and had a
          documented measurement uncertainty of 10 percent (Section 5.5.4).

       .   New lamps were used during validation testing (after a 100 hour burn-in period).

       .   The validation testing was conducted over a two-day period (Test Conditions 1-18
          on Day 1 and Test Conditions 19 - 27 on Day 2). The UV dose-response of the
          challenge microorganism (measured via a collimated beam test) was evaluated with
          1-L influent water samples at 95 percent UVT, collected on Day 1 of testing and at 85
          percent UVT, collected on Day 2 of testing (Section C.I).
  From the numerical model predictions developed by the manufacturer.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
B-13
November 2006

-------
                            Appendix B. UV Reactor Testing Examples
          All recommended testing protocols as listed in Section 5.7 and Appendix C were
          followed.
B.2.2  Test Data

       The data collected during validation testing are presented in Tables B.10 through B.13
and are described below:

       .   Table B. 10 presents the UV dose-response data measured on the influent water
          collected during field validation testing for the laboratory collimated beam test.

          Tables B. 11 presents the flow rate, UVT, lamp power, and UV sensor readings
          measured for Test Conditions 1-9 (95 percent UVT). Testing was also conducted at
          90 and 85 percent UVT. For simplicity, only the results of testing at 95 percent UVT
          are reported here.

       .   Table B. 12 presents measured challenge microorganism concentrations for the
          influent and effluent samples collected (in triplicate) from the reactor for several test
          conditions. A total of 27 tests (one for each condition described in Table B.9) were
          run. For simplicity, only the first tests at UVT measurements of 95, 90, and 85
          percent, respectively, are shown.

          Table B.13 presents data comparing the three reference UV sensor measurements to
          UV Duty Sensor #1 used during validation. Comparisons were made to all six duty
          sensors, but for simplicity only  the results for Sensor #1 are reported here.

       Sections B.2.3 to B.2.8 show how the data will be manipulated to determine whether
QA/QC criteria are met, to calculate the  necessary correction factors, and to determine the
validated operating conditions  for the target log inactivation.
UV Disinfection Guidance Manual                B-14                              November 2006
For the Final LT2ESWTR

-------
                       Appendix B. UV Reactor Testing Examples
       Table B.10 Challenge Microorganism UV Dose-response Measured
                     Using a Collimated Beam Apparatus
95% UVT
UV Dose
(mJ/cm2)
0
10
20
29
39
59

Replicate #1
N
(PFU/mL)
65560
13270
2790
640
159
23

LogN
4.82
4.12
3.45
2.81
2.20
1.36

Replicate #2
N
(PFU/mL)
67000
15000
2400
591
153
19

LogN
4.83
4.18
3.38
2.77
2.18
1.28

85% UVT
UV Dose
(mJ/cm2)
0
10
20
30
40
60

Replicate #1
N
(PFU/mL)
70440
14400
2590
693
173
28

LogN
4.85
4.16
3.41
2.84
2.24
1.45

Replicate #2
N
(PFU/mL)
70000
16120
2560
529
191
20

LogN
4.85
4.21
3.41
2.72
2.28
1.30

         Table B.11 Flow Rate, UVT, Lamp Power, and UV Sensor Data
            Measured during Validation Testing (for UVT = 95% only)
Test ID
1
2
3
4
5
6
7
8
9
Banks On
1
1
1
1
1
1
1
1
1
Flow Rate
(mgd)
10
5
2.5
10
5
2.5
10
5
2.5
UVT
(%)
95
95
95
95
95
95
95
95
95
Lamp Power
(kW)
8.0
8.0
8.0
5.6
5.6
5.6
3.2
3.2
3.2
Lowest Measured
Sdutv
(W/rrT)
303.1
307.9
297.9
183.1
180.2
190.3
91.8
93.5
89.4
  Table B.12 Measured Influent and Effluent Challenge Microorganism Cone, for
      Three Test Conditions (Three UVTs at 10 mgd and 100% Lamp Power)
Test ID
1
10
19
Influent Challenge
Microorganism
Log Concentration
Replicate
#1
4.92
4.88
4.93
#2
4.8
4.89
4.90
#3
4.87
4.83
4.91
Effluent Challenge
Microorganism
Log Concentration
Replicate
#1
3.79
3.94
4.24
#2
3.69
4.01
4.28
#3
3.70
3.94
4.29
UV Disinfection Guidance Manual
For the Final LT2ESWTR
B-15
November 2006

-------
                            Appendix B. UV Reactor Testing Examples
            Table B.13 Reference UV Sensor Checks for Duty Sensor #1
Before/After
Validation
Testing
Before
Before
Before
Before
After
After
After
After
UVT
(%)
95
95
85
85
95
95
85
85
Lamp
Power
(kW)
8
3.2
8
3.2
8
3.2
8
3.2
Sensor
ID
1
1
1
1
1
1
1
1
Sduty#1
(W/m2)
304.5
80.2
100.9
23.2
320.2
69.6
99.4
19.3
Sref, #1
(W/m2)
339.5
90.9
96.1
20.9
301.1
76.5
99.4
20.4
Sref, #2
(W/m2)
330.7
86.2
91.5
21.3
315.0
76.3
93.2
20.4
Sref, #3
(W/m2)
339.9
82.7
91.0
20.9
330.4
78.3
91.3
20.0
B.2.3 Develop the UV Dose-response Curve from the Collimated Beam Data

       Figures B.3(a) and (b) present the UV dose-response data from Table B. 10 at UVT values
of 85 and 95 percent, respectively. The data have been fitted to quadratic equations that show log
N as a function of UV dose. The fits were used to identify log N0 values for the UV dose-
response curves measured at 85 and 95 percent UVT (4.89 and 4.86, respectively). Table B.14
presents the UV dose-response data defined as UV dose versus log I (log [N0/N]).

       The UV dose-response data described in Table B.14 were analyzed to determine if the
two datasets could be combined using the method referred to in Section C.5 (Draper and Smith
1998).4 A statistical analysis of the collimated beam data is recommended to determine which
terms are significant, p-value < 0.05, using a standard regression tool. The process is iterative,
and each time the regression tool is used, one term is dropped until all coefficients are deemed
significant, p-value < 0.05. In this example, three iterations were required.

       The multiple regression analysis showed that the two  measured UV dose-response curves
were statistically similar and could  be combined [i.e., they could each be expressed with only
two variables, as A log I + B(log I)2]5. Figure B.4 presents the plot of UV dose as a function of
log inactivation for the combined dataset and the resultant UV dose-response equation.
4 The datasets should be combined whenever possible to develop one dose-response equation for calculating all
  RED values. The inability to combine datasets indicates a problem may have occurred with either the calculation
  or the test. Details are provided in Section C.5
5 If the regression analysis had shown that the UV dose-response curves could not be combined, separate curves
  would have been used to calculate RED values for data collected on each day of testing (i.e., the curve for 95%
  UVT would be used to calculate RED for full-scale reactor testing data collected on Day 1, and the curve for 85%
  UVT would be used to calculate RED for full-scale reactor testing data collected on Day 2)
UV Disinfection Guidance Manual
For the Final LT2ESWTR
B-16
November 2006

-------
                         Appendix B. UV Reactor Testing Examples
          Figure B.3 Log N Versus UV Dose Using the Data in Table B.10
                   at (a) 85 Percent UVT and (b) 95 Percent UVT
           §> 3.00
              2.00

              1.00

              0.00
       Log N = 0.0003(UV Dose)2 - 0.082(UV Dose) + 4.89
0      10     20     30     40     50
                 UV Dose (mJ/cm2)
                     (a)
                                                             60     70
           g> 3.00
              2.00

              1.00

              0.00
         Log N = 0.0002(UV Dose)2 - 0.079(UV Dose) + 4.86
                         10      20     30     40     50
                                    UV Dose (mJ/cm2)
                                        (b)
                                           60     70
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                     B-17
November 2006

-------
                         Appendix B. UV Reactor Testing Examples
       Table B.14 Challenge Microorganism UV Dose-Response Defined as
                      UV Dose Versus Log(N/N0) (i.e., Log I)
85% UVT
UV Dose
(mJ/cm2)
0
10
20
29
39
59

Replicate
#1 #2
4.89 - Log N
0.04
0.73
1.48
2.05
2.65
3.44

0.04
0.68
1.48
2.17
2.61
3.59

95% UVT
UV Dose
(mJ/cm2)
0
10
20
30
40
60

Replicate
#1 #2
4.86- Log N
0.04
0.74
1.41
2.05
2.66
3.50

0.03
0.68
1.48
2.09
2.68
3.58

          Figure B.4 Log I Versus UV Dose Using the Data in Table B.14
                 0.00
1.00
2.00

Log I
3.00
4.00
B.2.4 Verify That QA/QC Criteria Are Met

      Checklist 5.4 was used to ensure that recommended QA/QC criteria were met.
Calculations of key uncertainties are provided in the next two subsections.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
          B-18
                             November 2006

-------
                           Appendix B. UV Reactor Testing Examples
B.2.4.1
Collimated Beam Data Uncertainty
       The uncertainty in the UV dose calculation using the collimated beam data is calculated
using Equation C.6, shown below as Equation B.7:
       UDR=t
                  SD
              UVDose
           •xlOO%
Equation B.7
                       CB
       where
       UDR
       UV DosecB
       SD

       t
        Uncertainty of the UV dose-response fit at a 95-percent confidence level
        UV dose calculated from the UV dose-response curve in Figure B.4
        Standard deviation of the difference between the calculated UV dose-
         response and the measured value from Table B.10
        t-statistic at a 95-percent confidence level for a sample size equal to the
        number of test conditions replicates used to define the dose-response
       In this case, SD = 1.2 at 1.0-log inactivation and t = 2.06 for 24 test conditions from
Table B.14. Equation B.7 can then be used to determine UDR at various log inactivation values
from Table B. 10. The graph below shows the relationship between log inactivation and UDR. As
shown in the figure below, UDR = 19percent at 1.0-log inactivation. This value is less than the
recommended limit of 30 percent.
                           40%
                           30%
                       UDR 20%
                           10%
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                            B-19
November 2006

-------
                           Appendix B. UV Reactor Testing Examples
B.2.4.2
UV Sensor Uncertainty
       Guidance in Section 5.5.4 recommends that the manufacturer's reported sensor
uncertainty be confirmed as being less than 10 percent. Three reference sensors were used to
confirm duty sensor measurements. Data analyses are shown below.  The two duty UV sensors
were within 10 percent of the average readings from three reference sensors.
Before/
After
Validation
Testing
Before
Before
Before
Before
After
After
After
After
UVT
95
95
85
85
95
95
85
85
Lamp
Power
(kW)
8.0
3.2
8.0
3.2
8.0
3.2
8.0
3.2
Sduty #1
(W/m2)
304.5
80.2
100.9
23.2
320.2
69.6
99.4
19.3
(W/m2)
339.5
90.9
96.1
20.9
301.1
76.5
99.4
20.4
SRef,#2
(W/m2)
330.7
86.2
91.5
21.3
315.0
76.3
93.2
20.4
SRef,#3
(W/m2)
339.9
82.7
91.0
20.9
330.4
78.3
91.3
20.0
SRef,avg
(W/m2)
336.7
86.6
92.9
21.0
315.5
77.0
94.6
20.3
^duty ,
C
0 Ref.avg

10%
7%
9%
10%
1%
10%
5%
5%
   Source of UV sensor data: Table B.I 3
B.2.5  Calculate Log Inactivation and RED

       Table B.I5 shows the measured log inactivation through the UV reactor and the
associated RED values (determined from the UV dose-response equation) for each validation test
condition. The log inactivation values were determined from the field inactivation data
(excerpted in Table B.12).
 Table B.15 Measured Log I and RED Values for Each Test Condition in Table B.11
Test ID
1
2
3
4
5
6
7
8
9
10
11
12
UVT
95.1
94.8
94.8
95.4
94.6
94.8
94.7
94.5
95.5
90.0
89.9
90.1
Log I
Replicate #
1 2 3
1.13
1.36
1.66
1.03
1.28
1.55
0.91
1.16
1.39
0.94
1.13
1.35
1.11
1.35
1.63
1.09
1.27
1.56
0.92
1.17
1.39
0.88
1.11
1.42
1.17
1.34
1.67
1.12
1.35
1.59
0.90
1.20
1.44
0.89
1.14
1.39
RED (mJ/crrO
Replicate #
1 2 3
15.7
19.2
23.9
14.2
18.0
22.2
12.5
16.2
19.7
12.8
15.7
19.1
15.4
19.0
23.5
15.2
17.8
22.4
12.6
16.3
19.7
12.0
15.3
20.1
16.3
19.0
24.1
15.6
19.1
22.9
12.4
16.8
20.4
12.1
15.9
19.6
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                            B-20
November 2006

-------
                           Appendix B. UV Reactor Testing Examples
 Table B.15 Measured Log I and RED Values for Each Test Condition in Table B.11
                                       (cont.)
Test ID
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
UVT
90.1
90.1
89.8
89.6
89.6
90.0
84.6
84.7
84.6
84.7
85.4
85.3
85.1
85.4
85.5
Log I
Replicate #
1 2 3
0.85
1.05
1.30
0.67
0.91
1.21
0.68
0.87
1.17
0.56
0.84
1.07
0.42
0.64
0.93
0.81
1.03
1.29
0.73
0.94
1.17
0.62
0.91
1.17
0.55
0.76
1.03
0.47
0.67
0.94
0.86
1.07
1.34
0.74
0.95
1.22
0.62
0.87
1.18
0.54
0.77
1.07
0.44
0.67
0.95
RED (mJ/cmz)
Replicate #
1 2 3
10.6
13.4
17.1
8.2
11.5
15.7
8.3
10.9
15.1
6.7
10.4
13.7
4.9
7.8
11.7
10.0
13.1
17.0
8.9
11.8
15.2
7.5
11.4
15.1
6.6
9.4
13.1
5.6
8.1
11.8
10.7
13.7
17.7
9.0
12.0
15.9
7.5
10.9
15.3
6.4
9.5
13.7
5.1
8.1
11.9
B.2.6
Develop an Equation to Calculate RED as a Function of the Control
Variables
       To define an equation to calculate RED as a function of the operating conditions, the
validation data were fitted for a 1-bank configuration using Equations 5.8 and 5.10, shown below
as Equations B.8 and B.9 (there is only one bank of lamps, so it is not included as a variable
here):
            = WaxA254b x
                                                            Equation B.8
       Or in linear form,
\og(RED) = a + b
                    + c x log
                                                 d x log
Equation B.9
          where
             RED

             A254
             S
             So
             Q
       a, b, c, d
          The RED calculated with the UV dose-monitoring equation, also
          referred to as the "calculated dose" in this guidance manual
          UV absorbance at 254 nm
          Measured UV sensor value
          UV intensity measured at 100 percent lamp power.
          Flow rate
          Model coefficients obtained by fitting the equations to the data
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                            B-21
                                                                  November 2006

-------
                           Appendix B. UV Reactor Testing Examples
       Remember that the validation goal pursued in Example 2 is somewhat different from that
in Example 1. In Example 1, the simplest possible method (Single Intensity Setpoint Approach)
was desired. In Example 2, UV dose delivery will be paced to the operating conditions, so the
goal is to develop an equation that provides the best fit of the data. In some validations, the user
may try several different equation forms in an effort to find the best fit. The equation forms
presented here were selected because they have been used successfully at full-scale for numerous
UV reactors and  operating conditions (Wright et al. 2005).

       To develop a best-fit equation for RED in the form shown in Equation B.8 (or B.9 in
linear form), a theoretical equation for S0 should first be developed using validation test data to
capture the variation in S0 as a function of UVT. A strong goodness-of-fit as determined through
statistical analysis allows the equation for S0 to be used in development of the best-fit equation
for RED.

       1. Develop an expression for S0.
       The term S0 is the UV sensor measurement made with a new lamp operating at
100 percent power in a new, unfouled sleeve being monitored by a calibrated UV sensor through
a new, unfouled UV sensor port window. S0 varies with UVT. This relationship can be measured
during validation or determined from the validation test conditions. In this example, S0 is
determined by fitting the validation data using the following relationship. As with Equations B.8
and B.9, this approach has proven successful for several validation tests (Wright 2005):
       S = efegxUVTPh

       Equation B.10 can be expressed as:
                                 EquationB. 10
                                                                           EquationB.il
where P is the lamp power in units of kW and/ g, and h are model coefficients to be determined
in the subsequent analysis. Fitting this equation to the data in Table B. 11 at a 95-percent
confidence level (UVT and P) using the regression tool within spreadsheet software gives the
following values:
Term
f
9
h
Value
-8.402
0.115
1.578
p-Statistic
2.15 x 10""
1.61 x 10'IJ
6.97 x 10"ID
       Inputting these calculated values into Equation B.10 results in the following relationship
(8 kW is the UV reactor's maximum power setting, so S = S0 when P is equal to 8 kW):
         = e-8.402e0.115xC7fTpl.578
                                 Equation B.I2
UV Disinfection Guidance Manual
For the Final LT2ESWTR
B-22
November 2006

-------
                            Appendix B. UV Reactor Testing Examples
       and
              -8.402  0.115 xt/PTol.578
                     Q           o
                                 Equation B.I3
       The goodness of fit was evaluated by determining the p-statistic for the model
coefficients in Equation B.I3 (see Draper and Smith 1998 or similar for procedure).  The p-
statistic of each model coefficient was determined and found to be < 0.05.

       2. Calculate log (S/S0).

       By defining S as the measured UV sensor readings in Table B. 11, Equation B. 13 is then
used to predict S0 and to produce the data that will be fit to Equation B.9. The UVT (measured as
A254 values and converted to UVT units), Sduty and Q values are from the measured data (Tables
B.9 and B.15).

       3. Calculate an expression for RED.

       As with the UV dose-response equations (shown in Figure B.4) and the sensor equations
(Equations B.12 and B.13), the interpolation equation (Equation B.9) is fitted to the data in Table
B.I5 using a regression tool in spreadsheet software (at a 95-percent confidence level).  The
goodness-of-fit was again evaluated by determining the p-statistic for the model coefficients in
Equation B.14 (Draper and Smith, 1998). The p-statistic of each model coefficient was
determined to be < 0.05 (i.e., all were significant). The results of this analysis are shown below
with the following results:
Term
a
b
Value
-0.829
-2.519
p-Statistic
7.76 x 10'16
3.71 x 10'42
Term
c
d
Value
0.166
0.409
p-Statistic
1.21 x 10'19
9.35 x 10'42
       Inputting these calculated values into Equation B.9 results in the following:
       log (RED) = -0.829 - 2.519 x log (Xl254)+0.166 x log ( s/s  \ + 0.409 x log
                                                                           Equation B.I4

       Equation B.14 can be used to calculate RED values as a function of the operating
conditions (measured UVT, flow rate, and UV intensity) provided S0 is calculated using
Equation B.13.

       This equation can be programmed within the UV reactor's program logic controller
(PLC) to calculate the delivered RED as a function of the current operating conditions for UV
dose-monitoring. The  equation can be used for interpolation over the validated range of flow
rates (2.5 - 10 mgd), UVT values (85 - 95 percent), and RED values (8.5 - 24.1 ml/cm2).  If the
flow rate falls below 2.5 mgd, the PLC should default to 2.5 mgd in the dose-monitoring
UV Disinfection Guidance Manual
For the Final LT2ESWTR
B-23
November 2006

-------
                           Appendix B. UV Reactor Testing Examples
equation. If the UVT reads above 85 percent, the PLC should default to 95 percent in the dose-
monitoring equation.
B.2.7  Determine the Validation Factor

       The Validation Factor (VF) is defined according to Equation 5.13, shown below as
EquationB.15:

       VF = BRED x (l + UVal 1100)                                           Equation B. 15
B.2.7.1      Calculate the RED Bias

       Per guidance provided in Section 5.9.1, the UV sensitivity for the test condition with the
lowest UVT (85 percent) was calculated to range fromlS - 14 ml/cm2 per log inactivation. From
Table G.5 (for a 2.0-log Cryptosporidium inactivation credit), the RED bias for the maximum
UV sensitivity of 14 mJ/cm2 is 2.01.
B.2.7. 2      Calculate the Uncertainty of Validation

       The decision tree in Figure 5.5 was used to determine the correct equation for UVAL- As
shown in B.2.4.1, UDR is less than or equal to 30 percent. As noted in B.2.4.2, Us is less than or
equal to 10 percent.  Therefore, the equation for UVAL is as follows:

       UK«/ = Uw                                                          Equati on B . 1 6

       UIN is defined by Equation 5.15:

             ,   O7~~\
       UIN = — — x 1 00%                                                 Equation B . 1 7
         IN   RED

       where
         SD =  Standard deviation of the differences between the test RED (based on the
                 observed log inactivation and UV dose-response curve), and the RED
                 calculated using the dose-monitoring equation for each replicate
       RED =  The RED as calculated using the dose-monitoring equation
       The value of Uvai (i.e., UIN) can be expressed in one of two ways:

       1 .  As a single value, the most conservative (largest) uncertainty value calculated for the
          validated range. This is typically based on the lowest calculated RED value.

       2.  As a function of the calculated RED, that is, as a variable number.
UV Disinfection Guidance Manual                B-24                              November 2006
For the Final LT2ESWTR

-------
                           Appendix B. UV Reactor Testing Examples
In this example, more than 30 test measurements were made for RED, so the t-statistic is 2.04.
The standard deviation was determined from the RED values recorded during testing and the
RED values calculated using the dose-monitoring equation (equation B.14). The value of SD was
determined to be 0.97.

       The water system does not plan to operate below a calculated RED value of 15 mJ/cm2,
so a single value of UIN can be calculated as (2.04 x 1.0) /15 = 0.136. The following  equation
can be used if UIN is calculated as a function of RED at another location (of the same UV
reactor):

                  ,^0/  2.04x0.97   1.98                                   „    ..    0 10
                 x 100% =	=	                                  Equation B. 18
                                                                           M
            RED          RED     RED

       where
          RED = the RED calculated from the dose-monitoring equation (equation B.14)


B.2.7.3          Calculate the Validation Factor

       If the user prefers to use one UIN value, a single VF is determined using the following
equation:


       VF = 2.01xf l + —1 = 2.28                                             Equation B.I9
                I   15 J

       If a user at another water system preferred to use UIN as a function of the calculated RED,
the following equation would be used for calculating the VF:


                                                                          Equation B.20
                    100            RED        RED
B.2.8  Calculate the Validated Dose

       In the last step, the calculated RED associated with the operating conditions is divided by
the VF to produce the validated dose:

             RED
       Dvai  =	                                                       EquationB. 21
               VF

       where
             RED  = RED calculated from the dose-monitoring equation (Equation B.14)
             VF = Validation Factor (either Equation B.19 or Equation B.20)

       To calculate DVAL using a point estimate for the VF, use  Equation B.22:
UV Disinfection Guidance Manual                B-25                              November 2006
For the Final LT2ESWTR

-------
                           Appendix B. UV Reactor Testing Examples
     RED

      2.28
                                                                          EquationB.22
       To calculate a general expression with VF as a function of RED, to determine DVai,
Equation B.20 would be substituted into Equation B.21:
                RED
             2.01 + -
           3.98
          RED"
                                                                          Equation B.23
B.2.9  Assign Log Inactivation Credit for the Validated Dose(s)

       The validated dose must be greater than or equal to the required UV dose in Table 1.4
(Dreq) to achieve a given level of pathogen inactivation credit:
       D
         Val
    >D
                req
Equations.24
       or
       DVai >5.8 ml/cm2
'Val
EquationB. 2 5
       System Y is in the validated range when the calculated RED from the dose monitoring
equation is greater than or equal to 5.8 mJ/cm2 x 2.28 or 13.4 ml/cm2. (A similar calculation
would follow if a different water system preferred VF to vary with RED.)

       The UV reactor can receive 2.0-log Cryptosporidium inactivation credit for an
installation (with adequate inlet/outlet hydraulics, see Section 5.4.5) that operates under the
criteria outlined in Table B.I8:
              Table B.18 Validated Dose and Operating Conditions for
    2.0-log Cryptosporidium Inactivation Credit Using the Hypothetical Reactor
                                 Tested in Example 2
Validated Conditions
Flow Rate Range1
<10 mgd
UVT Range'
> 85%
Lamp Power Range
3.2 - 8 kW
REDJ
> 13.4 mJ/cm2
            1 At flow rates below 2.5 mgd, this value (2.5 mgd) should be used as the
            default value in the RED calculation.
            2 At UVT values above 95 %, this value (95% UVT) should be used as the
            default value in the RED calculation.
            3 Calculated using equations B.13 and B.14.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                                 B-26
 November 2006

-------
           Appendix C

Collimated Beam Testing to Develop a
      UV Dose-response Curve

-------
This page intentionally left blank.

-------
              Appendix C. Collimated Beam Testing to Develop a UV Dose-response Curve
       The LT2ESWTR requires that validation testing be conducted using "a test
microorganism whose dose-response characteristics have been identified with a low pressure
mercury vapor lamp" [40 CFR 141.720 (d)(2)(ii)]. To accomplish this, EPA recommends using a
collimated beam study of the test (or challenge) microorganism, as described in this appendix.
The procedure involves placing sample water with the challenge microorganism in an open
cylindrical container (e.g., a petri dish) and exposing the sample to collimated UV light for a
predetermined amount of time. The UV dose is calculated using the measured intensity of the
UV light, UV absorbance of the water, and exposure time. The measured concentration of
microorganisms before and after exposure provides the "response," or log inactivation of the
microorganisms from exposure to UV light. Regression analysis of measured log inactivation for
a range of UV doses produces the dose-response curve (sometimes expressed as a "dose-
response equation").

       This appendix describes the recommended collimated beam testing procedure and
recommended data analyses for developing the UV dose-response curve. Section C.I provides
guidelines for identifying test conditions. Section C.2 discusses all aspects of experimental
testing for the collimated beam study. Data analyses are discussed in Section C.3, followed by a
discussion of data uncertainty in Section C.4. Specific recommendations for combining dose-
response curves and limitations on applying results when the challenge microorganism exhibits a
shoulder or tailing are presented in Sections C.5 and C.6, respectively. Documentation of test
conditions and all results should be included in the Validation Report (see Section 5.11  for
guidance).
C.1.   Identifying Test Conditions

       At least two water quality conditions should be tested by collimated beam analysis:

       1.  The highest UV transmittance (UVT) used in the full-scale reactor test

       2.  The lowest UVT used in the full-scale test

Because UVT is accounted for in the UV dose calculation, the test conditions should produce
similar results that can be combined to produce one UV dose-response curve. (Performing two
tests instead of one test verifies that the UV dose is independent of UVT.)

       UV doses should be selected to cover the range of targeted values, using a minimum of
five data points plus a control [zero (0) UV dose]. The selected UV doses should result in
challenge microorganism inactivation ranging from 0.5-1 log unit higher than the highest log
inactivation to be demonstrated by the UV reactor. Table C. 1 shows a sample test matrix.

       At least one collimated beam test should be conducted on each day of full-scale reactor
testing.
UV Disinfection Guidance Manual                 C-l                               November 2006
For the Final LT2ESWTR

-------
	Appendix C. Collimated Beam Testing to Develop a UV Dose-response Curve	

 Table C.1. Sample Collimated Beam Test Matrix for Target of 2.0-Logs Inactivation
        of Cryptosporidium and MS2 Phage as the Challenge Microorganism
Sample 1
Lowest UVT
Highest UVT
Test Condition2
1
2
3
4
5
6
7
8
9
10
11
12
Log Inactivation
0 (control)
0.5
1.5
2.0
2.5
3.0
0 (control)
0.5
1.5
2.0
2.5
3.0
Target UV Doses
(mJ/cm2) 3
0
10
30
40
50
60
0
10
30
40
50
60
             The sample should represent the influent water used in full-scale reactor tests. One
             sample should reflect the lowest UVT tested, and one should reflect the highest UVT
             tested. Dose-response curves should be developed separately for each water quality
             condition tested and compared to determine if they can be combined (see Section
             C.5).
             Each test condition should be repeated at least twice (resulting in a minimum of two
             test condition replicates), and three test condition replicates will likely improve the
             quality of the fit for the dose-response  curve.
             Based on UV sensitivity of MS2 Phage
C.2   Measuring the UV Dose-response of the Challenge Microorganism

       The challenge microorganism's UV dose-response should be measured using a low-
pressure (LP) collimated beam apparatus (Figure C.I). This apparatus comprises an enclosed UV
lamp and a tube with a non-reflective inner surface. The UV light enters the suspension with a
near zero-degree angle of incidence and is relatively homogenous across the surface area. The
UV dose delivered to the suspension is calculated using measurements of incident UV intensity,
exposure time,  suspension depth, and the absorption coefficient of the suspension.

       Section C.2.1 provides a physical description of the collimated beam apparatus and
recommendations for operational controls. Accuracy of monitoring equipment is addressed in
Section C.2.2. Section C.2.3 provides the recommended test procedure, followed by equations
for UV dose calculations in C.2.4.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
C-2
November 2006

-------
              Appendix C. Collimated Beam Testing to Develop a UV Dose-response Curve
                       Figure C.1. Collimated Beam Apparatus
                   Low-Pressure
                   Mercury Arc Lamp
                   Petri Dish Containing
                   Microbial Suspension
                Lamp Enclosure


              Collimating Tube


            UV Light® 254 nm

               Magnetic Stirrer
       Note: To measure intensity of UV light, a calibrated radiometer is positioned below the column in place of
       the petri dish.
C.2.1  Apparatus Design and Operation

       Because UV dose requirements are based on the pathogen inactivation achieved using
254-nm light, the collimated beam apparatus should use a lamp that emits germicidal UV light
only at 254 nm (i.e., a LP lamp). To prevent ozone formation, lamps that emit 185-nm light
should not be used. The output from the lamp measured using a radiometer should vary by no
more than 5 percent over the exposure time. A stable lamp output can be obtained by driving the
lamp with a constant power source and maintaining the lamp at a constant operating temperature.
If the line voltage is not sufficiently stable, a voltage regulator may be used to obtain a stable
power supply. A stable temperature can be obtained by controlling the airflow around the lamp.

       The UV lamp should be located far enough above the surface of the microbial suspension
so that uniform irradiance is obtained across the sample's surface and UV light enters the
suspension with a near zero-degree angle of incidence (Blatchley 1997). A recommended
minimum distance from the lamp to the suspension is six times the longest distance across the
suspension's surface. In order to vary the UV intensity incident on the suspension, the distance
between the suspension and the lamp  can be adjusted.

       The uniformity of the intensity field across the sample's surface should be assessed by
measuring the "Petri Factor," defined as the ratio of the average irradiance across the suspension
surface to the irradiance measured at the center (Bolton and Linden 2003). The average
irradiance is determined by averaging radiometer measurements taken at each point in a 5-mm
spaced grid across an area defined by the suspension's surface. If the radiometer's sensing
UV Disinfection Guidance Manual
For the Final LT2ESWTR
C-3
November 2006

-------
	Appendix C. Collimated Beam Testing to Develop a UV Dose-response Curve	
window is wider than 5 mm, it should be reduced using a cover slip with a small hole. The
collimated beam apparatus should have a Petri Factor greater than 0.9.

       The lamp and the light path from the lamp to the suspension should be enclosed to protect
the user from exposure to UV light. A box-like enclosure made of aluminum is often used. A
length of pipe is often used to enclose the light path from the lamp to the microbial suspension.
The inside surface of the pipe should have a low UV reflectance and incorporate apertures to
improve UV light collimation (Blatchley 1997). A shutter mechanism is sometimes used to
control the exposure of the suspension to UV light. The exposure times should be measured with
an uncertainty of 5 percent or less. Exposure times less than 20 seconds are not recommended.

       The microbial suspension should be irradiated in an open cylindrical container (e.g., petri
dish). The diameter of the container should be smaller than the diameter of the light beam
incident on the container. Sample depth should be 0.5 - 2 cm. The material of the container
should not adsorb the challenge microorganism enough to impact its measured UV dose-
response.

       Sample volumes irradiated in the container should be sufficient for measuring the
challenge microorganism's concentration after irradiation. The microbial suspension should be
mixed using a stir bar and a magnetic stirrer at a rate that does not induce vortices.  The volume
and diameter of the stir bar should be small relative  to the volume and depth of the sample.

       The irradiance at the center of the suspension's surface before and after exposure to UV
light should be measured using a UV radiometer calibrated at 254 nm. During measurement, the
radiometer's calibration  plane should match the height of the suspension's surface and be
perpendicular to the incident UV. The calibration plane of the radiometer should be specified in
the radiometer's calibration certificate.
C.2.2 Accuracy of Monitoring Equipment

       Similar to the recommended procedures for full-scale reactor testing in Chapter 5,
spectrophotometer measurements of A254 should be verified using NIST ^traceable potassium
dichromate UV absorbance standards and holmium oxide UV wavelength standards. The
measurement uncertainty of the spectrophotometer should be 10percent or less. See Section
5.5.2 for additional guidance on spectrophotometer use and the recommended procedure for
verifying spectrophotometer measurements.

       Radiometers should be calibrated according to the following procedure to ensure that the
UV intensity is measured with an uncertainty of 8percent or less at a 95-percent confidence
level:

       1. The radiometers used in the collimated beam tests  should come from the
          manufacturer with a certified uncertainty of 8 percent or less at a 95-percent
          confidence level at the intervals suggested by the manufacturer.
1 National Institute of Standards and Technology
UV Disinfection Guidance Manual                C-4                              November 2006
For the Final LT2ESWTR

-------
	Appendix C. Collimated Beam Testing to Develop a UV Dose-response Curve	
       2. At minimum, the accuracy of the radiometer used to measure the UV intensity should
          be verified at least at the beginning and the end of each collimated beam test session
          using a second radiometer.

       3. The two radiometers should read within 5 percent of each other. If the two
          radiometers do not read within 5 percent of each other, a third radiometer should be
          used to identify which radiometer is out of specification. The two radiometers with
          readings within 5 percent of each other should be used. If none of the radiometer
          readings match, at least two of them are likely out of calibration.

       If the above criteria are met, the average radiometer measurement can be used in
calculations. Alternatively, the radiometer that provides the lowest reading could be used. If
these criteria are not met, the radiometers should be recalibrated. The radiometers should also be
checked to be sure that the irradiance measurement does not differ by more than 5 percent before
and after UV exposure.
C.2.3 Recommended Collimated Beam Test Procedure

       Researchers should collect a sample from the influent sampling port of the biodosimetry
test stand (or the influent sample for on-site reactor testing to be used for collimated beam
testing) for collimated beam testing. Typically, a 1 liter sample is sufficient. If the testing extends
over more than one day, at least one collimated beam test should be conducted for each day of
testing. If different batches of challenge microorganisms are used, a UV dose-response curve
should be generated for each batch.

       Personnel who perform collimated beam tests should be experienced with the use and
safety requirements of the equipment. Safety goggles and latex gloves should be worn. Skin
should be shielded from exposure to UV light. Personnel should follow recommended
procedures for challenge microorganism preparation and analysis as presented in Appendix A of
this guidance manual or use an alternative peer-reviewed method.

       The following procedure is recommended for collimated beam testing of a water sample
containing challenge microorganisms:

       1.  Measure the A254 of the sample using a spectrophotometer that has a measurement
          uncertainty of 10 percent of less (see guidance on spectrophotometer measurements
          in Sections C.2.2 and 5.5.2).

       2.  Place a known volume from the water sample into a petri dish and add  a stir bar.
          Measure the water depth in the petri dish.

       3.  Measure the UV intensity delivered by the collimated beam with no  sample present
          using a calibrated radiometer (see Section C.2.2 for guidance on calibrating
          monitoring equipment).

       4.  Calculate the required exposure time to deliver the target UV dose using
          Equation C.2 (described in the next section).

UV Disinfection Guidance Manual                C-5                              November 2006
For the Final LT2ESWTR

-------
_ Appendix C. Collimated Beam Testing to Develop a UV Dose-response Curve

       5.  Block the light from the collimating tube using a shutter or equivalent.

       6.  Center the petri dish with the water sample under the collimating tube.

       7.  Unblock the light from the collimating tube and start the timer.

       8.  When the target exposure time has elapsed, block the light from the collimating tube.

       9.  Remove the petri dish and collect the sample for measurement of the challenge
           microorganism concentration. If the sample is not assayed immediately, store in the
           dark at 4 °C. Each sample should be plated in triplicate and the average microbial
           value for the sample calculated from the three plate replicates.

       10. Re-measure the UV intensity and calculate the average of this measurement and the
           measurement taken in Step 3. The value should be within  5 percent of the value
           measured in Step 3 .

       1 1 . Using Equation C. 1 (described in the next section), calculate the UV dose applied to
           the sample based on  experimental conditions (this should be similar to the target
           dose).

       12. Repeat steps 1 through 1 1 for each replicate and target UV dose value (see Table
           C. 1). Repeat all steps for each water test condition replicate


C.2.4 UV Dose Calculation

       The UV dose delivered to the sample is calculated using:

                           T
       DCB = EsPf (1 - R)                   t                                Equation C. 1
                         (d + L) A254d\n(lO)
where:
 DCB  =   UV dose (ml/cm2)
   Es  =   Average UV intensity (measured before and after irradiating the sample) (mW/cm2)
   Pf  =   Petri Factor (unitless)
    R  =   Reflectance at the air-water interface at 254 nm (unitless)
    L  =   Distance from lamp centerline to suspension surface (cm)
    d  =   Depth of the suspension (cm)
 A254  =   UV absorbance at 254 nm (unitless)
     t  =   Exposure time  (s)
Alternatively, given a target UV  dose, the required exposure time may be calculated by  re-
arranging Equation C.I.


UV Disinfection Guidance Manual                C-6                               November 2006
For the Final LT2ESWTR

-------
              Appendix C. Collimated Beam Testing to Develop a UV Dose-response Curve
                 1      d + L A2,,d\n(W)                                   ^    •    „„
       t = D	-r^	j—4                                   Equation C.2
            ESP(\-R)  L        -Ad
       The term L/(d + L) accounts for the divergence of the UV light from the collimated beam
as it passes through the suspension. The reflectance at the air-water interface (R) can be
estimated using Fresnel's Law as 1.000/1.372, or 0.025 (the index of refraction of air divided by
the  index of refraction for water).

       To control for error in the UV dose measurement, the uncertainties of the terms in the UV
dose calculation should meet the following criteria:

       Depth of suspension (d)           < 10%
       Average incident irradiance (Es)   < 8%
       Petri Factor (Pf)                  < 5%
       L/(d + L)                        <1%
       Time (t)                         < 5 %
       (1 - 10-ad)/ad
The uncertainty in incident irradiance can be determined by the procedure for evaluating
uncertainty of radiometer measurements as presented in Section C.2.2. The remaining
uncertainties listed above should be estimated by laboratory personnel and documented in the
Validation Report (See Section 5.11 for guidance).
C.3   Developing the UV Dose-response Curve

       Collimated beam tests produce the following types of experimental data:

       .   UV Dose in units of mJ/cm2,

       .   Concentration of microorganisms in the petri dish prior to UV exposure (N0) in units
          of pfu/mL, and

          Concentration of microorganisms in the petri dish after UV exposure (N) in units of
          pfu/mL.

One UV dose-response curve should be developed for each UVT condition tested (typically high
and low).  If full-scale reactor testing spans more than one day, at least one UV dose-response
curve should be developed for each day of testing.

       EPA recommends using regression analysis to develop each UV dose-response curve
using the following steps:
UV Disinfection Guidance Manual                C-7                              November 2006
For the Final LT2ESWTR

-------
       	Appendix C. Collimated Beam Testing to Develop a UV Dose-response Curve	
       1.  For each test condition and replicate, plot log N vs. UV dose to identify a common N0
          as the intercept of the curve at UV dose = 0 (an example is illustrated in Figure C.2).2
              Figure C.2. Fitting Effluent Concentration vs. UV Dose to
                 Determine a Common Influent Concentration Value
                                             log N = 0.00010(UV Dose)' - 0.053(UV Dose) + 5.37
                           20
40
60
                                           UV Dose (mJ/cm )
80
100
120
       2.  Calculate log I for each measured value of N (including zero-dose) and the common
           N0 identified in Step 1 using the following equation:
                                                                              Equation C.3
          where:
             No   =   The common N0 identified in Step 1 (pfu/mL)
              N   =   Concentration of challenge microorganisms in the petri dish after
                       exposure to UV light (pfu/mL)

          The log inactivation for each replicate should be averaged to produce one value of log
          I per test condition.

       3.  Plot UV dose as a function of log I for each test condition.

       4.  Use regression analysis to derive an equation that best fits the data, forcing the fit
          through the origin. The equation will have different forms depending on the data. For
          challenge microorganisms exhibiting first-order kinetics, a linear equation should be
          used:
  If the measured value of N0 is used for this calculation, any experimental or analytical error in the measured value
  is carried to all the data points, adding an unrelated bias to each measurement. Therefore, using the y-intercept of
  the curve is recommended.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                                        November 2006

-------
               Appendix C. Collimated Beam Testing to Develop a UV Dose-response Curve
              UV Dose = A x log /
                                 Equation C.4
          See Figure C.3 for an example of a linear UV dose-response curve.
                 Figure C.3. Typical E. coli UV Dose-response Curve
                     UV Dose [mJ/cm2] = 1.83 log I
                            R2 = 0.994
                                   2          3
                                         log Inactivation
          A quadratic equation can also be used, as illustrated in the example in Figure C.4:
              UV
                                 Equation C.5
                  Figure C.4. Typical MS2 UV Dose-response Curve
            120
            100 -
          o
             80 -
          
-------
	Appendix C. Collimated Beam Testing to Develop a UV Dose-response Curve	
       5.     Evaluate the equation's goodness-of-fit—the differences between the measured
              UV dose values and those predicted by the equation should be randomly
              distributed around zero and not be dependent on UV dose. The goodness of the fit
              can be examined by standard statistical tests, such as examining the p-statistics for
              the regression coefficients.

Note that the resulting equation should not be used for extrapolation outside of the measured
range of UV dose.
C.4   Collimated Beam Data Uncertainty

       As noted in Section C.3, collimated beam data will often be fit to a linear or a polynomial
regression. The 95-percent confidence interval (UDR) can be calculated using standard statistical
methods, such as those described in Draper and Smith 1998, or can be conservatively estimated
using Equation C.6.
                   SD
               UV DoseCB
                           : 100%
                                                 Equation C.6
       where:
               UDR

        UV DosecB

                SD

                  t
Uncertainty of the UV dose-response fit at a 95-percent confidence
level
UV dose calculated from the UV dose-response curve for the
challenge microorganism
Standard deviation of the difference between the calculated UV dose-
response and the measured value
t-statistic at a 95-percent confidence level for a sample size equal to
the number of test condition replicates used to define the dose-
response3
Number of Data Points Used to Develop
the Dose-Response Equation
10
11
12
13
14
15
16

t
2.23
2.20
2.18
2.16
2.14
2.13
2.12










Number of Data Points Used to Develop
the Dose-Response Equation
17
18
19-20
21
22-23
24-26
27-29
>30
t
2.11
2.10
2.09
2.08
2.07
2.06
2.05
2.04
       If UV dose-response curves can be combined (as described in the next Section, C.5), the
combined dataset should be used to calculate UDR. If individual dose-response curves cannot be
combined, UDR should be calculated separately for each curve.
3 For example, one test condition evaluated twice (two test condition replicates) with five UV dose points each
  would have a total often points.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                   C-10
November 2006

-------
	Appendix C. Collimated Beam Testing to Develop a UV Dose-response Curve	

       EPA recommends that the value of UDR (calculated by Equation C.6) not exceed
30percent at the UV dose corresponding to 1-log inactivation of the challenge organism (e.g.,
18 ml/cm2 for MS2). If the 95-percent confidence interval is calculated using standard statistical
methods, UDR should not exceed 15percent at the UV dose corresponding to 1-log inactivation
of the challenge organism.4 If there is more than one estimate of UDR (i.e., UV dose-response
curves cannot be combined), the maximum UDR should be used to determine if it meets this
criterion.

       If the UDR value calculated by Equation C.6 is greater than 30 percent (15 percent if the
standard statistical method is used), it should be added to the total uncertainty of validation [e.g.,
Uvai = (UiN2 + UoR2)   , see Section 5.9.2].This allows for a validation plan that is sufficiently
flexible to continue using the dose-response curve at low values, but will increase the Uyai
accordingly. Similarly, if UV dose-response curves cannot be combined and one or more of the
individual curves exhibits a UDR value greater than 30%, the maximum value should be used in
                                                     7      7 1/2
calculating the total uncertainty of validation [UVai = (UjN  + UDR )   ].
C.5   Combining UV Dose-response Curves

       Analysis of regression coefficients indicates whether or not UV dose-response curves
developed using different water samples can be combined. In order for the UV dose-response
curves to be combined, differences between the regression coefficients should not be statistically
significant at a 95-percent confidence level.5 If differences in the coefficients are statistically
significant, the reason for this difference should be documented in the Validation Report.
Differences between measured UV dose-response curves for different water samples could
indicate one or more of the following:

       1.  The UV dose-responses of different batches of the challenge microorganism differ. In
           this case, the UV dose-response curve specific to each cultured batch of the challenge
           microorganism should be used to assess UV dose delivery for the validation test
           conditions using that batch.

       2.  Interferences due to water quality, such as coagulation or inactivation of the challenge
           microorganism. In this case, mitigate the cause of the interference or account for the
           interference when assessing UV dose delivery for the validation test conditions.

       3.  Errors calculating the UV dose delivered by the collimated beam apparatus. Mis-
           measuring the incident UV intensity or the UV absorbance of the water sample could
           introduce such errors.

       If differences between UV dose-response curves cannot be resolved, a single curve
corresponding to one day's worth of full scale reactor testing can be used to calculate RED
values for that day (i.e., there will be one UV dose-response curve per day of full-scale reactor
validation testing). If two or more UV doe-response curves from the same day of testing cannot
be combined,  the curve resulting in the most conservative (lowest) UV dose should be used for
4 This criterion (15 percent) differs from the criterion of 30 percent applied to Equation C.6 due to simplifications
  incorporated into Equation C.6.
5 A good description for performing this test is provided in Draper and Smith, 1998.

UV Disinfection Guidance Manual                C-ll                               November 2006
For the Final LT2ESWTR

-------
	Appendix C. Collimated Beam Testing to Develop a UV Dose-response Curve	
calculating RED values. If different curves are used for RED calculations, the UV sensitivity of
the challenge microorganism and shape of each UV dose-response curve should be consistent
with expected inactivation behavior for that challenge microorganism.
C.6   Using Challenge Microorganisms with Shoulders or Tailing

       In the case of a challenge microorganism with a shoulder or tailing in the UV dose-
response, the UV sensitivity should be defined as the sensitivity over the region of linear log
inactivation that occurs between the shoulder and the onset of tailing.  The shoulder of the UV
dose-response is  defined as the point of intersection of the exponential region with the UV dose
axis (see Figure C.5). The UV dose-response of the challenge microorganism should not
demonstrate a shoulder at a UV dose beyond 50 percent of the demonstrated (measured) RED
range, and should not demonstrate tailing until at least one log inactivation beyond the
demonstrated (measured) inactivation range.
                 Figure C.5. UV Dose-response of B. subtilis Spores
                 c
                 o
                 +J
                 nf
                    2  -\
                       ~
                     1  \
                    o  4
               y = 0.0886*- 1 .4881j
Dose axis intercept
  = 1fi5m.l/rrn2
                        shoulder
                                  20         40         60
                                     UVDose(mJ/cm2)
                                        80
               (Adapted from Sommer et al. 1998)
     Example C.I. The UV dose-response of B. subtilis spores has a shoulder at low UV
     dose values (Figure C.5). Because the measured UV dose-response has a shoulder of
     16.5 ml/cm2, the B. subtilis spores should only be used to demonstrate RED values
     greater than or equal to 2 x  16.5 ml/cm2 = 33 ml/cm2.	
UV Disinfection Guidance Manual
For the Final LT2ESWTR
              C-12
November 2006

-------
                 Appendix D



Background to the UV Reactor Validation Protocol

-------
This page intentionally left blank.

-------
                    Appendix D. Background to the UV Reactor Validation Protocol
       This appendix provides background material for the validation protocol given in
Chapter 5. The background material is organized into the following six sections.

       .   UV dose delivery by UV reactors. Section D.I describes why a correction factor
          (termed the "RED bias") should be applied in the Validation Factor calculation to
          account for systematic errors that arise if the challenge microorganism is more
          resistant to UV light than the target pathogen.

          UV dose monitoring. Section D.2 provides background information on the impact of
          UV sensor placement on UV dose monitoring (whether it is at, closer to, or farther
          from the lamp than the ideal position). It provides a rationale for defining test
          conditions to validate UV reactors using a given UV dose-monitoring approach and
          explains why sensor position is important.

       .   UV sensors.  Section D.3 provides the basis for the UV sensor calibration criterion
          recommended in Chapter 5. It describes the properties of UV sensors, how those
          properties impact the sensor's measurement uncertainty, and how that measurement
          uncertainty can be determined.

       .   Polychromatic considerations. Section D.4 describes systematic errors that can
          occur with the validation of UV reactors that use medium-pressure (MP) UV lamps
          (1) equipped  with non-germicidal UV sensors and/or (2) validated with a challenge
          microorganism  that has a UV action spectrum significantly different from that of the
          target pathogen. This section provides a rationale for assessing those errors.

          Uncertainty  of validation. Section D.5 provides a rationale for defining a validation
          factor that accounts for the random uncertainty associated with UV reactor validation
          and monitoring.

          CFD modeling. Section D.6  provides guidance on using Computational Fluid
          Dynamics (CFD) to model UV dose delivery.
D.1    UV Dose Delivery by UV Reactors

       UV dose delivery by UV reactors to be used at water treatment plants (WTPs) is
currently measured using biodosimetry (Quails and Johnson 1983). With biodosimetry,
inactivation of a challenge microorganism passed through the UV reactor is measured and
related to a single dose value based on the known UV dose-response of that microorganism. This
dose is termed the "reduction equivalent dose," or RED.
D.1.1  Using RED to Demonstrate Target Pathogen Inactivation

       If the UV dose-response of the challenge microorganism does not match the target
pathogen's, and the UV dose distribution of the UV reactor is not known, biodosimetry can only
be used to estimate the target pathogen inactivation within a range bounded by the inactivation
expected assuming "ideal" and "worst-case" hydraulics. Figure D.I provides a comparison of the
UV Disinfection Guidance Manual                 D-l                               November 2006
For the Final LT2ESWTR

-------
                    Appendix D. Background to the UV Reactor Validation Protocol
UV dose distributions of reactors with ideal and worst-case hydraulics to a UV dose distribution
that might be seen with a real reactor.
              Figure D.1. UV Dose Distributions of Ideal, Realistic, and
                               Worst-case UV Reactors
'robability

Ideal

0 50 100
UV Dose (mJ/cm2)
&
TO

8
Q.
                                               Reality
                                             50
                                                      100
                                      UV Dose (mJ/cm )
>
+*
Probabi

Worst Case
1




0 Infinity
UV Dose (mJ/cm2)
D.1.1.1   Ideal Reactor Hydraulics

       A UV reactor with ideal hydraulics delivers the same UV dose to all the microorganisms
passing through the reactor. Its UV dose distribution is represented by a single value. Examples
of a UV reactor with ideal hydraulics include the stirred suspension irradiated during the
measurement of UV dose-response with a collimated beam device and an ideal plug-flow
reactor. In both cases, the delivered dose is the product of the average UV intensity within the
reactor and the residence time. Accordingly, with an ideal reactor, the RED measured with a
challenge microorganism is a measure of the RED delivered to all microorganisms that pass
through the reactor because all the microorganisms receive the same dose.
D.1.1.2   Worst-case Hydraulics

       For a reactor with worst-case hydraulics and a measurable RED, an infinite UV dose is
delivered to one fraction of the flow rate, and zero UV dose is delivered to the other fraction (i.e.,
one of two UV dose values is delivered to each respective microorganism). The net inactivation
achieved is constant, equal to the fraction receiving the infinite UV dose, and hence independent
of the microorganism's inactivation kinetics. With a worst-case UV reactor, the measured
inactivation is that which would occur with any microorganism regardless of its UV sensitivity.
D.1.1.3   Real-world Hydraulics

       Using the above definitions of an ideal and a worst-case UV reactor, the log inactivation
of a pathogen estimated from biodosimetry results will have a value between log(N0jC/Nc) and
[RED/Dp],1 that is, a "real" UV reactor will have a UV dose distribution that falls somewhere
between ideal and worst-case (Wright and Lawryshyn 2000).
 N0 c is the influent challenge organism (c) concentration and Nc is the effluent challenge microorganism
  concentration. Dp is the UV sensitivity of the pathogen (p) in units of mJ/cm2 per log inactivation.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
           D-2
November 2006

-------
                    Appendix D. Background to the UV Reactor Validation Protocol
       If the inactivation of the pathogen must be known with absolute confidence, the lower
bound of that range should be used (assume worst-case hydraulics). When the challenge
microorganism is more resistant to UV light than the target pathogen, the lower bound is the log
inactivation of that challenge organism, log [N0;C/NC]. In other words, if a UV reactor is validated
with a challenge microorganism that is less sensitive to UV light than the target microorganism,
one cannot know with certainty that the reactor achieved more than the log inactivation
demonstrated during validation. For example, if 2-log MS2 inactivation is measured, one can
conclude only that the water system attained > 2-log inactivation for any organism less resistant
to UV light.

       If the challenge microorganism is less resistant to UV light than the target pathogen, the
lower bound is the RED measured with the challenge organism divided by the sensitivity of the
target pathogen, [RED/Dp]. For example, if one measures a (px-174 RED of 12 ml/cm2,
corresponding to 4-log inactivation, one cannot assume that the reactor achieved 4.0-log
inactivation of Cryptosporidium; one can assume only that the water system attained
[12 mJ/cm2] / [4.0 ml/cm2 per log  I] = 3.0-log inactivation.

       But both of these assumptions are extreme. For this reason, the "RED Bias" correction
factor used in the VF is based on the UV dose distribution of a defined "real-world worst-case"
UV reactor. The RED delivered to a pathogen by a given UV reactor can be estimated from the
measured RED of the challenge microorganism using Equation D.I:
        r>r-n   nm
       RED =RED  x - - — = - -                                     Equation D.I
                      REDcwc    BRED

       where:
       REDp    =  Pathogen RED estimated for the UV reactor of interest (mJ/cm2)
       REDC    =  Challenge microorganism RED measured during biodosimetry (mJ/cm2)
             wc =  Pathogen RED estimated from Figure D.2, the "worst-case" UV reactor
                   (mJ/cm2)
       REDC wc =  Challenge microorganism RED estimated from Figure D.2, the "worst-
                   case" UV reactor (mJ/cm2)
       BRED     =  RED Bias, the ratio of the RED of the pathogen to the RED of the
                   challenge microorganism for a given set of operating conditions

The development of this factor (RED Bias, or BRED) is discussed below.

Defining a Realistic Conservative VVDose Distribution

       Because UV manufacturers strive to optimize the hydraulic design of their UV reactors,
using the worst-case UV dose distribution represented in Figure D.I to define the lower bound of
pathogen inactivation is overly conservative. An alternative approach is to use the UV dose
distribution of a commercial UV reactor that is representative of plausibly poor UV reactor
hydraulics.
UV Disinfection Guidance Manual                 D-3                               November 2006
For the Final LT2ESWTR

-------
                    Appendix D. Background to the UV Reactor Validation Protocol
              Figure D.2. The UV Dose Distributions Used to Determine
                      RED Bias Values Tabulated in Appendix G
           0.25

           0.20 -\
         ~ 0.15
         ro
         o 0.10 -
           0.05 -

           0.00 -
                      20
                              40
                                     60
                                            80
                        UV Dose (mJ/cm2)
                                     UVT = 98 %
                                     UVT = 95 %
                                     UVT = 90 %
                                     UVT = 85 %
                                                                20
                                                                       40
                                                                               60
                                                                                      80
                      UV Dose (mJ/cm )
           0.035 -,
           0.030 -
         >, 0.025 -
         1 0.020 -
         -g 0.015 -
         £ 0.010 -
           0.005 -
           0.000 -
               0
                                                    1.000
      Jl  0.800 -
      i-
      %  0.600 -

      c  0.400 -

      |  0.200 -
      u_
         0.000 -
                         50
                                  100
                                           150
                                                                  50
                                                                           100
                                                                                     150
                        UV Dose (mJ/cm2)
                      UV Dose (mJ/cm )
       UV dose modeling based on CFD was used to predict dose distributions for commercial
low-pressure high-output (LPHO) and medium-pressure (MP) UV reactors. Details on the
approach are provided in Wright and Reddy (2003) and Dzurny et al. (2003). CFD was used to
predict the trajectories of approximately 3,000 microbes through the UV reactors. UV intensity
fields within the reactor were modeled using the methods described by Bolton (2000). The UV
dose delivered to each microbe was predicted by integrating the total UV dose delivered over its
trajectory through the reactor. The REDs delivered to the target pathogens were calculated
assuming first-order kinetics with a UV sensitivity defined as the required dose in Table 1.4
divided by the associated log inactivation credit. The dose distributions were scaled to give
pathogen REDs equal to the required dose for a given level  of log inactivation credit plus an
uncertainty factor of 25  percent (i.e., for 3-log Cryptosporidium inactivation credit, dose
distribution was scaled to give a Cryptosporidium RED = 12+ [0.25  x 12] = 15 mJ/cm2). This
approach assumes that the UV reactor uses dose pacing2 to  deliver the required RED without
overdosing.  REDs were estimated for test microbes of various UV sensitivities assuming first-
order kinetics. The RED bias was calculated as the ratio of the test microbe RED to the pathogen
RED (BRED = REDtestmicrobe/REDpathogen)-

       The commercial reactor that resulted in the most conservative RED bias values was used
to develop the RED Bias values in Appendix G. Figure D.2 shows the scaled UV dose
  The UV reactor maintains the delivered dose at or near the target value by adjusting the lamp power or turning
  "on" or "off" banks of UV lamps or whole UV reactors to respond to changes in UV absorbance, lamp intensity,
  and/or flow rate.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
D-4
November 2006

-------
                  Appendix D. Background to the UV Reactor Validation Protocol
distributions used to estimate the RED bias for 3-log inactivation credit with Cryptosporidium.
Figure D.3 shows predictions of RED and log inactivation for those dose distributions as a
function of the test microbe's UV sensitivity. Figure D.4 shows the RED bias for 3-log
Cryptosporidium credit as a function of the test microbe's UV sensitivity obtained using the data
in Figure D.2.
      Figure D.3. RED as a Function of Microorganism UV Sensitivity for the
                      UV Reactor Represented in Figure D.2
             o>  — .
>
             O
             '.5
             U
             %
             
-------
                   Appendix D. Background to the UV Reactor Validation Protocol
 Example D.I. A UV reactor is challenged using MS2 with a UV sensitivity of 18 ml/cm2 per
 log inactivation. The UVT of the water is 85 percent. Two-log inactivation is measured,
 corresponding to an MS2 RED of 2-log x 18 mJ/cm2-log I = 36 ml/cm2. These results are used
 to estimate the log inactivation of two pathogens, one with a UV sensitivity of 10 mJ/cm2 per
 log inactivation and the other with a UV sensitivity of 25 mJ/cm2 per log inactivation.

       In Figure D.3, the RED delivered to the microorganisms with a UV sensitivity of 10,  18,
 and 25 mJ/cm2 per log inactivation would be 20, 25, and 28 mJ/cm2, respectively. The RED
 Bias values for MS2 relative to the first pathogen is 25/20 =  1.25 while the RED Bias for MS2
 relative to the second pathogen is 25/28 = 0.89. Assuming the reactor has a UV dose
 distribution that is better than the dose distribution used to develop Figure D.2, the RED of the
 first pathogen has a  value between 36 and 36/1.25 = 29 mJ/cm2 and the RED  of the second
 pathogen has a value between 36 and 36/0.89 = 40 mJ/cm2.	
D.2    The Impact of UV Sensor Positioning on UV Dose Monitoring

       This guidance manual focuses on two commonly used UV dose-monitoring strategies, the
UV Intensity Setpoint Approach and the Calculated Dose Approach, which are summarized
below. Sections D.2.1 and D.2.2 discuss the impact of UV sensor positioning for the UV
Intensity Setpoint Approach and Calculated Dose Approach,  respectively.

       1.  UV Intensity Setpoint Approach. UV dose delivery is indicated by the measured
          flow rate and UV intensity. Minimum UV dose delivery is verified when the
          measured UV intensity is above an alarm (minimum) setpoint value defined as a
          function of the flow rate through the reactor. In a variation of this method, the
          minimum UV dose can be verified when the measured relative UV intensity
          (calculated as a function of UVT) is above an alarm (minimum) setpoint value
          defined as a function of the flow rate through the reactor.

       2.  Calculate Dose Approach. Minimum UV dose delivery is verified when the
          calculated UV dose (using an equation dependent on flow rate,  relative UV intensity,
          UVT, and sometimes other parameters such as lamp status) is above an alarm
          (minimum) setpoint value.
D.2.1  UV Intensity Setpoint Approach

       With the UV Intensity Setpoint Approach, dose monitoring is impacted by UV sensor
positioning (Wright et al. 2002). To illustrate this impact, Figure D.5 presents the relationship
between UV dose and measured UV intensity for a simple annular reactor containing a single
low-pressure (LP) lamp. UV intensity was calculated using a radial UV intensity model and UV
dose was calculated assuming ideal hydraulics (Haas and Sakellaropoulos 1979). UV intensity
and dose were calculated for a fixed flow rate of 140 gpm. Simulated UVT values ranged from
70 to 98 percent, and simulated relative lamp outputs (characterized by relative sensor values)
ranged from 20 to 100 percent. In each figure, the data are presented as plots of UV dose as a
function of the UV  sensor reading for a range of UVT values. Each point at a given UVT
represents, in order of increasing UV dose, operation at 20, 40, 60, 80, and 100 percent relative
UV Disinfection Guidance Manual                D-6                              November 2006
For the Final LT2ESWTR

-------
                   Appendix D. Background to the UV Reactor Validation Protocol
lamp power. The differences between these figures are due to differences in sensor-to-lamp
distance (i.e., UV sensor placement).

       1.  The UV Sensor Is Located at the "Ideal Position"

       Figure D.5a presents the relationship between delivered UV dose and sensor reading
obtained when the UV sensor is located at the sensor-to-lamp distance where the relationships
between UV dose and measured UV intensity at different UVTs overlap. Because these
relationships overlap, a given UV intensity can be related to a specific level of dose delivery.
 Example D.2. The UV reactor characterized in Figure D.5a is used in a disinfection application
 where the target dose is 20 ml/cm2. A UV sensor value S of 18 mW/cm2 is used as an alarm
 setpoint to indicate the UV reactor delivers a dose of 20 mJ/cm2 across the entire operating
 range—the ideal placement of the sensor ensures that an alarm setpoint value of 18 mW/cm2
 will indicate a dose of 20 mJ/cm2.
       2.  The UV Sensor Is Located Closer to the Lamp Than the "Ideal Position"

       Figure D.5b presents the relationship between delivered UV dose and sensor reading
when the UV sensor is placed closer to the lamp than the ideal position (i.e., a smaller sensor-to-
lamp distance than in Figure D.5a). Because the sensor views the lamp through a relatively thin
water layer, its response to changing UVT is small compared to that in Figure D.5a.
Accordingly, the relationship between dose delivery and measured UV intensity cannot be
described by a single relationship for all values of UVT. Unlike the situation depicted in Figure
D.5a, the delivered dose will  decrease at lower UVTs for a given UV sensor reading.
Accordingly, the measured UV intensity should only be  used to indicate dose delivery at the
lower end of that range, which occurs under conditions of maximum lamp power and
reduced UVT
 Example D.3 The UV reactor characterized in Figure D.S.b is used in an application where the
 target dose is 20 mJ/cm2. The UV manufacturer states that a UV sensor value S of 80 mW/cm2
 will indicate a dose of 20 mJ/cm2 under design conditions of 85% UVT and 60% relative lamp
 output.  However,  as shown in Figure D.S.b,  a UV intensity of 80 mW/cm2  corresponds to a
 dose ranging from 10 mJ/cm2 (for 70% UVT)  to 37 mJ/cm2 (for 98% UVT). For a UV intensity
 alarm setpoint to ensure a delivered dose of 20 mJ/cm2 under all possible conditions of water
 UVT and lamp output,  a sensor setpoint value  S' of 157 mW/cm2 would need to be used.	
UV Disinfection Guidance Manual                D-7                              November 2006
For the Final LT2ESWTR

-------
                 Appendix D. Background to the UV Reactor Validation Protocol
 Figure D.5. Relationship between UV Dose and Intensity for a UV Sensor Located
   (a) at the "Ideal Position," (b) Close to the Lamp, and (c) Far from the Lamp
                      70
                      60
                      50
              UV Dose 40
              (mJ/cm2) 30
                      20
                      10
                       0
140 gpm
            UVT
           254 nm
           — 70%
           - 80%
           — 85%
             90%
           — 94%
           — 98%
                                  20
                40
        60
                                 Sensor (mW/cm )
                                     (a)
                                                           UVT
                                                          254 nm
                                                         — 70%
                                                          -80%
                                                         — 85%
                                                            90%
                                                         — 94%
                                                         — 98%
                         0  20  40  60  80  100 120 140 160
                                 Sensor (mW/cm2)
                                     (b)
                                                           UVT
                                                          254 nm
                                                          — 70%
                                                          - 80%
                                                          — 85%
                                                            90%
                                                          — 94%
                                                          — 98%
                                10
           20
30
40
                                 Sensor (mW/cm )
                                     (c)
UV Disinfection Guidance Manual
For the Final LT2ESWTR
         D-8
                   November 2006

-------
                    Appendix D. Background to the UV Reactor Validation Protocol
       3.  The UV Sensor Is Located Farther from the Lamp Than the "Ideal Position"

       Figure D.5c presents the relationship between delivered UV dose and sensor reading
when the UV sensor is located farther from the lamp than the ideal position (i.e., a greater
sensor-to-lamp distance than in Figure D.5a). Because the sensor views the lamp through a
relatively thick water layer, its response to changing water transmittance is greater at this
position than at either the ideal or closer-than-ideal positions. Again, the relationship between
UV dose delivery and measured UV intensity cannot be described by a single relationship for
different values of UVT. However, unlike Figure D.5b, the UV dose delivered at a given
measured UV intensity increases as UVT decreases. Thus, the measured UV intensity should
only be used to indicate UV dose delivery at the lower end of that range, which  occurs
under  conditions of reduced lamp power and maximum UVT (the opposite of what is
observed with the closer-than-ideal UV sensor position).
Example D.4. The UV reactor characterized in Figure D.5c is used in an application where the
target dose is 20 ml/cm2. A UV intensity alarm setpoint value S of 10 mW/cm2 is proposed
based on the UV intensity measured under design conditions of 85 percent UVT and 60 percent
relative lamp output. However, a sensor value of 10 mW/cm2 indicates a UV dose ranging from
15 to 32 mJ/cm2. To ensure a delivered dose of 20 mJ/cm2 under all possible conditions of
water UVT and lamp output, a setpoint value 5" of 14 mW/cm2 would need to be used.	
       The manufacturer of the UV reactor selects the location of the UV sensor within a UV
reactor. If the UV reactor uses the UV Intensity Setpoint Approach for UV dose monitoring, it is
to the manufacturer's advantage to optimize the UV sensor's location to obtain overlapping
relationships between UV dose delivery and measured UV intensity for different UVT values,
similar to the example given in Figure D.5a.

       If the UV manufacturer does not optimize the UV sensor's location, a given UV intensity
will correspond to a range of UV dose values as opposed to a single value. While this does not
prevent the UV reactor from using the UV Intensity Setpoint Approach, the monitoring approach
will be significantly less efficient than with an ideally located UV sensor because the UV reactor
will be overdosing at many UVT-lamp power combinations that give rise to operation at the
setpoint.  When this occurs, the manufacturer may opt to supplement measurements of UV
intensity  with measurements of UVT to enable more efficient UV dose monitoring (this is
sometimes referred to in the literature as the "UV Intensity-UVT Setpoint Approach"). The UV
reactor is verified to be delivering the required UV dose when both the measured UV intensity
and UVT are above the minimum validated setpoint values (S/S0j UVT setpoint, and UVTsetpoint),
both defined for a specified range of flow rates. With this approach, there are no requirements for
UV sensor positioning.
D.2.2  Calculated Dose Approach

       Measurements of flow rate, UV intensity, and UVT can be incorporated into theoretical,
empirical, or semi-empirical calculations of UV dose delivery. For example, the relationships
represented in Figure D.5a - c could be defined experimentally and used in an empirical manner

UV Disinfection Guidance Manual                 D-9                               November 2006
For the Final LT2ESWTR

-------
                    Appendix D. Background to the UV Reactor Validation Protocol
to calculate UV dose (e.g., Equation 5.8). Relationships could also be defined using advanced
numerical modeling approaches to relate measured intensity to UV dose delivery as a function of
flow rate and UVT.

       In theory, the UV dose calculation does not necessitate that the UV sensor be placed at
any one location within the reactor. However, if the UV sensor were placed at the ideal position
(a location that gives UV dose delivery proportional to the UV sensor reading), the UV dose
calculation would not require UVT as an input parameter.
D.3    UV Sensors

       UV sensors are photosensitive detectors that are used to indicate UV dose delivery by
providing information related to UV intensity at different points in the UV reactor. Reference
UV sensors are used to check that the measurements made by the on-line, or "duty" sensors are
valid.

       UV sensors include the following components, arranged as shown in Figure D.6:

          Monitoring windows and light pipes deliver light to the photodetector. Monitoring
          windows are typically quartz discs and light pipes are cylindrical probes made of
          quartz (quartz silica probe).

          Diffusers and apertures reduce the UV light incident on the photodetector to slow
          UV sensor degradation. Diffusers also modify the UV sensor's angular response.

       .   Diffusers and apertures reduce the UV light incident on the photodetector to slow
          UV sensor degradation. Diffusers also modify the UV sensor's angular response.

       .   Filters limit the light delivered to the photodiode, typically restricting it to germicidal
          UV wavelengths (-200 - 300 nm).

          Photodetectors are solid-state devices that produce a current proportional to the
          irradiance on the detector's active surface. The responsiveness of a typical
          photodetector to UV light is on the order of 0.1 - 0.4 mA/mW.

       .   Amplifiers convert the output of the photodetector from a low-level current to  a
          standardized output proportional to the incident UV intensity.

       .   The housing of the UV sensor protects the components from the external
          environment. The housing should be electrically  grounded to shield the photodetector
          and amplifier, thereby reducing electrical noise and bias.
UV Disinfection Guidance Manual                D-10                              November 2006
For the Final LT2ESWTR

-------
                    Appendix D. Background to the UV Reactor Validation Protocol
                     Figure D.6. Interior UV Sensor Schematics3
      a.
Digital k
Signal
Output
Signal
Output
        Signal Amplifier
     Printed Circuit

         Photodetector"

            UVC Filter-

              Housing

         Quartz Silica Probe
                            UV Light
                                                                  ^-Housing
                                   Photodetector

                                   Filter
                                                                      Quartz
                                                                      Window
                                                         UV Light
D.3.1  UV Sensor Properties

       The UV sensor should detect germicidal UV radiation and produce a standardized output
signal proportional to the incident UV irradiance (e.g., 4-20 mA). A UV sensor may or may not
measure the UV light through a monitoring window that is separate from the sensor body.
Monitoring windows should have a high UVT over the sensor's spectral response range.

       UV sensor properties that impact the measurement of UV intensity and dose delivery
monitoring include angular response, acceptance angle, spectral response, working range,
detection limit and resolution, linearity, temperature response, long term drift, calibration factor,
and measurement uncertainty. An ideal UV sensor would  have the following properties:

          A linear response to incident UV light, independent of water temperature and stable
          over time.

       .   A fixed angular response and a wavelength response that mimics the germicidal
          response of the target microorganism(s).
  Figure courtesy of (a) Aquionics and (b) WEDECO UV Technologies; UVC light is 200 - 280 nm (the germicidal
  range).
UV Disinfection Guidance Manual
For the Final LT2ESWTR
       D-ll
             November 2006

-------
                    Appendix D. Background to the UV Reactor Validation Protocol
       .   Has zero measurement noise and bias.

       .   Responds only to germicidal UV light.

       .   Has zero measurement uncertainty.

       The properties of an ideal UV sensor are presented here to illustrate the benchmark UV
sensor manufacturers strive to approximate as closely as possible, that is, zero measurement
error.

       Angular response is a plot of the UV sensor measurement as a function of the incident
angle of light on the sensor's window. Angular response is affected by the UV sensor's aperture
size, the size of the photodetector's active surface, the distance between the aperture and the
active surface, and the impact of any diffusers and reflecting surfaces within the sensor. An ideal
UV sensor has a "cosine response" (Equation D.2), which results in an accurate measure of the
light incident on the surface of the photodetector. In practice, UV sensors deviate from the cosine
response; some potential responses are shown in Figure D.7.
          = S, cos#
                                                             Equation D.2
       where:
       9   =
Intensity measured by the UV sensor's photodetector [watt per centimeter
squared (W/cm2)]
Intensity incident on the UV sensor's photodetector's surface (W/cm2)
Incident angle at the UV sensor's photodetector surface (°)
                 Figure D.7. Angular Response of Two UV Sensors
                        Relative to the Ideal Cosine Response
                                                             Cosine
                                                       - -  -  Sensor 1
                                                       — — Sensor 2
                                    -60   -30    0     30    60
                                    Incident Angle (degrees)
                                                90
       The opening or acceptance angle of the UV sensor is the angle over which the sensor
detects UV light. The opening angle is typically measured by either the threshold detection of
UV light or detection at some percentage of the maximum value (e.g., 50 percent). The
acceptance angle is a characteristic of the sensor but does not affect its performance.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                           D-12
November 2006

-------
                    Appendix D. Background to the UV Reactor Validation Protocol
       The spectral response is a measure of the output of the UV sensor as a function of
wavelength. It depends on the response of the photodetector and filters and the UV transmittance
of the monitoring windows, light pipes, and filters.

       The working range of the UV sensor is the intensity range that the sensor is able to
measure. The low end of the working range is defined by the detection limit of the measurement.
The high end of the working range is limited by the saturation of the photodetector and the
amplifier. Saturation is the point at which the UV sensor can no longer respond to an increase in
intensity.

       The detection limit of the UV sensor is the lowest UV intensity that can be detected and
quantified at a known confidence level. The detection limit is calculated based on repeated
measurements of low intensity UV light,  usually at a  specific percentage confidence interval.

       The measurement resolution is the smallest difference in UV intensity that can be
differentiated at a given confidence level. The detection limit and the resolution depend on the
measurement noise and on any digitalization of the analog output from the UV sensor by the
system's electronics. Measurement bias and noise of a photodetector are increased by
electromagnetic fields within the UV reactor if the sensor is not properly shielded and grounded.

       An ideal UV sensor (a sensor with an ideal cosine response) responds proportionally to
the intensity incident on the sensor (Figure D.7). The linearity of the UV sensor is a measure of
the adherence of the sensor response to that proportional relationship. It is reported as the ratio of
the measured response to the known incident intensity, usually at a specific confidence level.
Linearity is affected by bias and saturation.
D.3.1.1   Calibration and Quantification of UV Sensor Properties

       UV sensors used to monitor monochromatic lamps are often calibrated using the
substitution method of Larason et al. (1998). With this approach, the intensity of a collimated
beam of UV light at 254 nm is measured using the UV sensor; the measured value is then
compared to that made using a standard measurement, such as a NIST4-traceable UV sensor or
chemical actinometer. The ratio of the standard measurement to the UV sensor output is the
calibration factor. With UV sensors designed to measure the output of MP lamps, the sensor can
be calibrated at 254 nm, calibrated as a function of wavelength, or calibrated using
polychromatic light from an MP lamp with a known spectral output.

       UV sensor linearity is determined by comparing the sensor output as a function of
incident irradiance to standard measurements of that irradiance. UV sensor temperature
response is determined by measuring the dependence of sensor output on the sensor's operating
temperature with the sensor measuring a constant irradiance. The angular response of a UV
sensor is determined by measuring the dependence of the UV sensor reading on the incident
angle of a beam of fixed-intensity, collimated UV light.
4 National Institute of Science and Technology, Boulder, Colorado.
UV Disinfection Guidance Manual                D-13                              November 2006
For the Final LT2ESWTR

-------
                    Appendix D. Background to the UV Reactor Validation Protocol
       The spectral response of a UV sensor is determined by measuring the dependence of the
sensor output on the wavelength of monochromatic light of known irradiance incident on the
sensor. Spectral response is typically presented as a plot of the ratio of sensor output to incident
irradiance as a function of the wavelength of light.

       The measurement accuracy of UV sensors changes over time due to mechanical wear
and environmental exposure. Temperature cycling, exposure to UV light, mechanical vibration,
and other factors will impact the linear, spectral, angular, and temperature response of a sensor.
Long-term sensor stability is best determined using field data, but may be estimated using
accelerated life-cycle testing.
D.3.1.2   Recommendations for Calibration and Quantification of UV Sensor
          Properties

       UV sensors provided by the manufacturer should be individually calibrated. The
manufacturer should determine linearity and temperature-response over the expected operation
range of lamp intensity and water temperature expected during operation at WTPs. Because it
may be affected by infrared transmission of glass filters and fluorescence of diffusers that are
part of the UV sensor (Larason and Cromer 2001), the sensor spectral response should be
evaluated from 200 to 1,000 nm. The sensor response should be "germicidal" (see Section 5.4.8
for the definition of a "germicidal" UV sensor response).

       UV sensor manufacturers should conduct regular testing on their UV sensors to develop a
database on the effect of long-term use on sensor properties. While some UV sensor properties
may be measured with each sensor (e.g., calibration), other properties, such as long-term stability
and angular and spectral response, can practically be measured only on a representative sample
from a lot. The UV sensor manufacturer should have available for inspection the following
information:

       .   A description of the measured UV sensor properties.

          A description of the system used to measure each property.

       .   A description of the measurement standards used.

          The documented uncertainty associated with each measurement.

       .   A description of the QA/QC procedures used to ensure that the measurements were
          traceable to a standard.

       .   Data that demonstrates that the properties of the manufactured UV sensors are within
          specifications over time.
UV Disinfection Guidance Manual                D-14                              November 2006
For the Final LT2ESWTR

-------
                    Appendix D. Background to the UV Reactor Validation Protocol
D.3.2  UV Sensor Measurement Uncertainty

       UV sensor measurement uncertainty quantifies how the UV intensity value measured
with a duty UV sensor (mounted on the UV reactor) compares to the true value. For the purposes
of this manual, UV  sensor uncertainty should be determined by summing the uncertainties that
arise from calibration, linearity, angular and spectral response, temperature response, and long-
term stability (see Table D.I for an example of this calculation):

       .   Uncertainty in the UV sensor calibration arises from the uncertainties associated with
          the standards and instrumentation used to calibrate the sensor (e.g., voltmeters and
          amplifiers).

       .   Uncertainty in the UV sensor's linearity and temperature response arises because
          sensor calibration factors, determined at one temperature and UV irradiance, are used
          over a range of temperatures and irradiances during operations at a WTP.

       .   Uncertainty in angular response arises because UV sensors are used in UV reactors to
          measure UV light impacting from different directions but are calibrated with
          collimated light (i.e., light is incident to the surface from only one angle).

       .   Variability in spectral and angular response from UV sensor-to-UV sensor results in
          an additional measurement uncertainty not accounted for in calibration. The impact of
          spectral and angular response variability on UV sensor measurement uncertainty can
          be determined either by calculation or by measurement.

          In the calculation approach, UV sensor spectral and angular response, measured on a
          representative sample from a lot, is used as an input to a numerical model that
          predicts  sensor readings in a reactor. The variability in the readings predicted by the
          model is used to define an uncertainty term that is included in the calculation of the
          total sensor uncertainty. In the measurement approach, the variability in
          measurements made by  a representative number of UV sensors mounted on the
          reactor is used to define the uncertainty.

       .   Uncertainty in spectral response arises in MP systems because sensors, calibrated at a
          fixed wavelength, are used in UV reactors  equipped with polychromatic lamps.

       .   Additional uncertainty arises from long-term UV sensor drift.

       The information described above should be provided by the manufacturer for each duty
sensor as part of the UV reactor documentation. The purpose of this information is to indicate the
ability of the manufacturer to quantify the uncertainty for important sensor properties and to
demonstrate whether the sensor can meet sensor specifications prepared by the system purchaser.
This information should not be used to verify sensor performance during validation testing or
operations. Instead, sensor uncertainty should be field-verified by comparing duty sensor
measurements to calibrated reference sensors, as described in Sections 5.5.4  and 6.4.1.1.
UV Disinfection Guidance Manual                D-15                              November 2006
For the Final LT2ESWTR

-------
                   Appendix D. Background to the UV Reactor Validation Protocol
                         Table D.1. Example of a UV Sensor
                         Uncertainty Calculation Datasheet
Property
Spectral response
Angular response
Linearity
Calibration
Temperature response
Long term drift
Total Uncertainty1
Uncertainty (%)
4
3
3
5
3
12
15
Total uncertainty is calculated as:
(!2+32+32+52+32+102)1/2= 15%.
 Example D.7. A UV sensor manufacturer calibrates each UV sensor at 20°C with an
 uncertainty of ±1 percent. Linearity, temperature response, angular response, and spectral
 response are evaluated on every tenth sensor manufactured. Linearity ranges from 1-5 percent
 over the measurement range of the sensor. Temperature response ranges from 0.1 - 0.2 percent
 per °C—an uncertainty of 5 percent over the temperature range 0-40 °C. Models predict that
 the variability in angular and spectral response from sensor-to-sensor will cause uncertainties of
 8 percent and 3 percent, respectively. A laboratory evaluation of UV sensors returned from the
 field indicates that the long-term drift over a one-year period is 11  percent. The measurement
 uncertainty of the UV sensors is calculated as the square root of the sum of the squares of the
 individual percent uncertainties:
 Measurement uncertainty = VI2 + 52 + 52 + 82 + 32 +112 = 16°/
         16%
D.3.3  Number of UV Sensors

       Lamp-to-lamp variability in UV output impacts both UV dose delivery and monitoring
(Wright et al. 2004). If a lamp has a lower output than the other lamps in a UV reactor, it will
deliver lower UV doses to microorganisms passing in its vicinity, thereby shifting the UV dose
distribution to lower values and reducing the net performance (UV dose delivery) of the reactor.
The shift in the UV dose distribution will be more pronounced in a reactor with fewer lamps.

       If the number of UV sensors is less than the number of lamps and the sensors do not
monitor the lamps with the lowest output, the monitoring system will overestimate UV dose
delivery. Sections 6.3.2.2 and 5.4.7 provide guidance on dealing with this issue in operations and
validation, respectively.
D.4    Polychromatic Light Considerations

       LP and LPHO lamps are monochromatic, with UV output at a single wavelength, 254
nm. MP lamps are polychromatic, with UV output at multiple wavelengths. UV dose delivery
and monitoring in MP reactors involves UV light from 200 to 320 nm. The output from the UV
UV Disinfection Guidance Manual
For the Final LT2ESWTR
D-16
November 2006

-------
                    Appendix D. Background to the UV Reactor Validation Protocol
sensor is an integrated response to UV light over wavelengths spanning the sensor's spectral
response. If the spectral properties of the UV reactor that influence UV dose delivery and
monitoring during operation at a WTP are the same as during validation, then the characterized
UV dose delivery will occur at the WTP. However, if the spectral properties are significantly
different, UV dose delivery at the WTP can differ substantially from UV dose delivery measured
during validation for the same measured operating values. The following spectral properties may
differ:

          Action spectra of the challenge microorganism and of the target pathogen.
          Spectral UV absorbance of the water used during validation and at the WTP.
       .  UV output of the lamps during validation and at the WTP (see 5.4.6 for details).
       .  UVT of the lamp sleeves during validation and at the WTP (see 5.4.6 for details).


       Section D.4.1 describes approaches for assessing the impact of differences in microbial
action spectrum properties. Section D.4.2 describes an approach for developing a correction
factor for polychromatic bias for MP reactors when the challenge microorganism is something
other than MS2 or Bacillus subtilis. Derivation of the polychromatic bias factor for MP UV
reactors with non-germicidal sensors is presented in Section DAS.
D.4.1  Impact of Microorganism UV Action Spectra Differences

       The dependence of microorganism inactivation kinetics on wavelength can be described
using an action spectrum - the UV inactivation sensitivity of a microorganism as a function of
wavelength (Figure D.8). Ideally, the action spectrum of the challenge microorganism used to
validate a polychromatic UV reactor would either match that of the target microorganism or
provide a conservative estimate of inactivation.
UV Disinfection Guidance Manual                D-17                              November 2006
For the Final LT2ESWTR

-------
                   Appendix D. Background to the UV Reactor Validation Protocol
               Figure D.8. Action Spectra for Various Microorganisms
                  2.5 n

                  2.0

            Action! 5
            relative
              to
                                     -8— Herpes simplex
                                     -B-MS2, R-17,fr, 7-S
                                     -*-<|>x-174
                                     -A-T2
            254 nm
                  1.0
                  0.5
                  0.0
                     225   235   245   255   265   275   285   295   305
                                       Wavelength (nm)
           (Adapted by H. Wright from Rauth 1965.)
       The impact of various action spectra on UV dose delivery may be estimated by
calculating the germicidal lamp output using Equation D.3:
             320
                                                                           Equation D.3
            1=200
       where:
       PG   =
       A/I   =
Germicidal output of the MP lamp (W/cm)
Wavelength (nm)
Lamp  output  at wavelength \ measured over 1-nm increments [watt per
nanometer (W/nm)]
Relative UV sensitivity of the microorganism at wavelength
1-nm wavelength increment (nm)
                                                                       (cm"1)
       Using the published action spectra of fourteen microorganisms (Cabaj et al. 2002, Linden
et al. 2001, Rauth 1965), Table D.2 presents the germicidal lamp output calculated for a
commercial MP lamp and the ratio of that output to that of Cryptosporidium. A ratio greater than
one (1) indicates that the microorganism receives more germicidal output compared to
Cryptosporidium. If a challenge microorganism with a ratio greater than  1.05 is used to validate
a MP reactor for Cryptosporidium inactivation, the ratio should be used as a correction factor
(called the "action spectra correction factor," or CFas) to account for the greater proportional
inactivation of the challenge microorganism that arises from the differences in the two action
spectra. In the case of MS2 and B. subtilis, the ratio is close to one (1) and the correction is small
(< 0.06). However, based on the data in Table D.2, if cpx!74 was used to  show Cryptosporidium
inactivation, an action spectra correction factor of 1.16  would be needed with MP reactors. In
other words, the cpx!74 RED would be divided  by 1.16 to determine the RED used to calculate
the validated dose (see Section 5.8).
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                           D-18
November 2006

-------
                   Appendix D. Background to the UV Reactor Validation Protocol
   Table D.2. Germicidal Output Delivered to 14 Microorganisms by an MP Lamp
Microorganism
Cryptosporidium oocysts
Vaccinia
B. subtilis spores
VSV
MS-2, R-17, fr, 7-S
T2
EMC
cpx-174
Polyoma
Herpes simplex
Reovirus-3
Type / Nucleic acid
(SS = single strand,
DS = double strand)
Protozoa / DS DMA
Animal virus / DS DMA
Aerobic spore / DS DMA
Animal virus / RNA
Bacteriophage / SS RNA
Phage / DS DMA
Animal virus / SS RNA
Bacteriophage / DS DMA
Animal virus / DS DMA
Human virus / DS DMA
Animal virus / DS RNA
Germicidal
Output
(W/cm)
5.64
5.46
5.58
5.53
5.78
6.05
5.98
6.53
6.74
7.00
7.46
Germicidal Output
Relative to
Cryptosporidium
(Action Spectra
Correction Factor)
1.00
0.98
0.99
0.99
1.04
1.07
1.07
1.16
1.18
1.26
1.32
       The germicidal output of the MP lamp calculated using the action spectra of B. subtilis
spores and MS2 is equal to or less than that of most of the 14 microorganisms listed in Table
D.2. Thus, it is reasonable to assume that these microorganisms are acceptable as challenge
microorganisms for many pathogens whose action spectra are not known, like adenovirus and
Giardia. However, if an alternative challenge microorganism is to be used, its action spectra
should be assessed for suitability.

       As an alternate approach to measuring the action spectrum, the correction factor can  also
be estimated by comparing the UV dose-response of the challenge microorganism to that of MS2
measured with a LP and a MP lamp. The correction factor would be defined as:
       CF  =1.04
                                                            EquationD.4
       Where:
       CFas =
       1.04 =
Correction factor for the difference in action spectra between the challenge
microorganism and MS2 (unitless)
Slope of the UV dose-response measured with a MP collimated beam (cm2/mJ)
Slope of the UV dose-response measured with a LP collimated beam (cm2/mJ)
Germicidal output of MS2 relative to Cryptosporidium, from Table D.2
(unitless)
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                           D-19
November 2006

-------
                    Appendix D. Background to the UV Reactor Validation Protocol
 Example D.8. A UV reactor is validated with the virus (px-174 (germicidal output relative to
 Cryptosporidium is 1.14). The reactor uses the UV Intensity Setpoint Approach and measures
 an RED of 25 ml/cm2 at the setpoint. In calculating the validated dose for Cryptosporidium or
 Giardia inactivation credit, the measured RED used to calculate the Validated Dose (discussed
 in Section 5.8.2) would have to be adjusted to 25 mJ/cm2/1.14 = 22 ml/cm2.	
       NOTE: The correction factor described in this section is applicable only to MPreactors.
It should be used if CFas > 1.06. The correction factor that accounts for differences in the action
spectra is not the same correction factor that accounts for differences in the microorganism UV
sensitivities described in Section D.I  (UV dose-distribution impacts). The correction factor
described in Section D.I, the "RED Bias," applies to all UV reactors regardless of lamp type.
D.4.2  Water Absorption of UV Light

       During UV reactor validation, a UV-absorbing chemical is added to the bulk flow passing
through the reactor in order to simulate high-UV absorbance (low-UVT) events that could occur
at a WTP. Common UV-absorbing chemicals currently in use for validation testing include
lignin sulfonate, sodium thiosulfate, fluorescein, coffee, concentrated humic acids, tea, and
parahydroxybenzoic acid. Ideally, the spectral absorption of the water used to validate UV
reactors equipped with MP lamps should match the spectral absorption of the water at the WTP
over the wavelength range associated with UV dose delivery and monitoring. Figure D.9
illustrates UV spectra measured in waters at several different WTPs.

       Figure D. 10 compares the UV absorbance spectra of coffee and lignin sulfonate to those
of two drinking water sources ("Water A" and "Water B"). For a given UVT, the UV absorption
at wavelengths above and below 254 nm is greater with coffee and lignin sulfonate than with the
drinking water sources. If those chemicals are used during validation of a MP reactor, the RED
and UV intensity values measured at a given flow rate, lamp output, and water UVT will be
lower during validation than at the WTP (Wright et al. 2002).
UV Disinfection Guidance Manual                D-20                               November 2006
For the Final LT2ESWTR

-------
                   Appendix D. Background to the UV Reactor Validation Protocol
           Figure D.9. Spectral UV Absorption of Water at Various WTPs
                   0.50
               
-------
                   Appendix D. Background to the UV Reactor Validation Protocol
D.4.3  Determining the Polychromatic Bias Factor (BPoiy)

       The term "polychromatic bias" refers to polychromatic differences between validation
and operation of a UV reactor. UV reactors with MP lamps that were installed prior to the
publication of this document may use non-germicidal sensors and, thus, may exhibit
polychromatic bias. To account for polychromatic bias during validation testing, a polychromatic
bias factor (Bpoiy) should be incorporated into the Validation Factor:
                                '100
                                EquationD.5
See Section 5.9 for a complete discussion of the Validation Factor.

       Tables D.2 and D.3 should be used to estimate BPoiy using the following validation testing
information:

          The UV-absorbing compound (coffee or LSA)

       .   The minimum UVT tested (for all validation tests)

       .   The lamp sleeve-to-sensor distance (i.e., water layer)


Note that Tables D.2 and D.3 are for discreet values of UVT and lamp sleeve-to-sensor distance.
The Polychromatic Bias Factor can be interpolated for intermediate values.
 Example D.9. An MP UV reactor with a non-germicidal sensor located 5 cm from the lamp
 sleeve is validated using coffee as a UV-absorbing chemical. The UV reactor is validated at a
 minimum UVT value of 85%. Using Table D.4, the polychromatic bias values  at 85% UVT
 values is 1.12.
        Table D.3. Polychromatic Bias Values for an MP UV Reactor Using a
                 Non-germicidal UV Sensor and Validated with LSA
Water Layer
(cm)
2
5
10
15
20
25
Polychromatic Bias Values for a UVT of:
70%
1.00
1.22
1.74
2.28
2.76
3.19
80%
1.00
1.08
1.38
1.71
2.07
2.41
85%
1.00
1.04
1.25
1.48
1.74
1.99
90%
1.00
1.00
1.13
1.27
1.42
1.58
95%
1.00
1.00
1.05
1.12
1.18
1.25
98%
1.00
1.00
1.02
1.05
1.07
1.10
     Note: water layer = sensor to lamp distance
UV Disinfection Guidance Manual
For the Final LT2ESWTR
D-22
November 2006

-------
                   Appendix D. Background to the UV Reactor Validation Protocol
        Table D.4. Polychromatic Bias Values for an MP UV Reactor Using a
               Non-Germicidal UV Sensor and Validated with Coffee
Water Layer
(cm)
2
5
10
15
20
25
Polychromatic Bias Values for a DVT of:
70%
1.01
1.57
3.70
9.42
24.6
64.3
80%
1.00
1.22
1.99
3.42
6.11
11.0
85%
1.00
1.12
1.56
2.25
3.34
5.11
90%
1.00
1.05
1.29
1.61
2.04
2.61
95%
1.00
1.01
1.11
1.22
1.35
1.50
98%
1.00
1.00
1.04
1.08
1.12
1.16
    Note: water layer = sensor to lamp distance
       In addition to the polychromatic bias that can occur from the use of non-germicidal
sensors, polychromatic bias can occur when a germicidal sensor in an MP UV reactor is farther
away from the lamp than the ideal location. In this case, the water layer can act as an optical
filter, preferentially absorbing lower wavelength light and introducing polychromatic bias. The
polychromatic bias exhibited by germicidal UV sensors that are further away from the lamp than
the ideal location is not expected to be significant as long as the sensor is 10 cm or closer to the
lamp (or further if the water being tested exhibits a UVT greater than 90%). This criterion is met
for most MP reactors on the market at the time of manual publication and thus is not addressed in
this manual.

D.5    Analytical Foundation for UV Dose Monitoring and  Defining
       Uncertainty

       UV installations should be sized and operated in a manner that accounts for the
measurement uncertainty associated with UV dose delivery monitoring. The objective of UV
dose delivery monitoring is to indicate the level of inactivation of the target pathogen. This
section derives a measurement equation for UV dose monitoring (Wright and Mackey 2003).
This equation is used in this manual as the analytical foundation for defining the uncertainty of
UV dose monitoring.

       Consider a UV installation operating at a WTP. Assuming first-order kinetics, the log
inactivation of a target pathogen achieved by the UV reactor at some point in time can be
expressed using Equation D.6:
log Ip =


where:
log Ip  =
REDp  =
REDp
 A^"
                                                                          Equation D.6
                Log inactivation of the pathogen
                RED of the pathogen (ml/cm2)
                UV sensitivity of the pathogen (mJ/cm2 per log I)
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                                   D-23
                                                           November 2006

-------
                    Appendix D. Background to the UV Reactor Validation Protocol
       If the UV reactor delivers a UV dose distribution, the log inactivation of the pathogen is
related to the inactivation of a challenge microorganism by substituting Equation D.I into
EquationD.6:
                 1
       log/  = -- -                                                  Equation D. 7
                Ft    D
               °RED Mo p

       where:
       REDC  =  RED of the challenge microorganism (ml/cm2)

       Assuming the challenge microorganism RED is proportional to the measured UV
intensity (REDC QC S), the log inactivation of the pathogen can be expressed according to
EquationD.8:

                 1   aS
       log /„ = --                                                   Equation D.8
                ft    D
               DRED Mo^

       where:
       S  =  UV intensity measured at the WTP with a duty UV sensor (mW/cm2)
       a  =  Constant relating challenge microorganism inactivation to measured intensity
              (mJ/mW)

       The constant a is determined during validation as the ratio of the measured RED of the
challenge microorganism to the measured UV intensity (REDC/S). Assuming that inactivation is
proportional to flow rate (log Np oc Q), Equation D.9 can be used:
       log ip =          C__^                                            Equation D.9
                BRED DwP  Sv Q

       where:
       Sv =   UV intensity measured during validation (W/cm2)
       Qv =   Flow rate measured during validation (mgd)
       Q =   Flow rate measured at the WTP (mgd)

       Assuming the UV dose-response of the challenge microorganism follows first-order
kinetics, the challenge microorganism RED during validation is calculated using the log
inactivation of the challenge microorganism measured through the reactor:
                                                                           Equation D. 10


       where:
       DIOC =  UV sensitivity of the challenge microorganism (mJ/cm2 per log inactivation)
       N0>c  =  Challenge microorganism concentration measured at the reactor influent
               (organisms/mL)
UV Disinfection Guidance Manual                D-24                              November 2006
For the Final LT2ESWTR

-------
                    Appendix D. Background to the UV Reactor Validation Protocol
       Nc   =  Challenge microorganism concentration measured at the reactor effluent
               (organisms/mL)

       The UV sensitivity of the challenge microorganism (Dioc) can be calculated according to
Equation D. 1 1 from the UV dose-response measured using the collimated beam apparatus:


       DWc = — ^—                                                      Equation D. 1 1
              log 4

       where:
       DCB   =  UV dose delivered by the collimated beam apparatus (mJ/cm2)
       log /c =   Log inactivation of the challenge microorganism observed with a UV dose of
                 DCB

       Substitution of Equations D.10 and D.I 1 into equation D.9 gives the equation for dose
monitoring using the UV Intensity Setpoint Approach:
    log/, =                                                               EquationD.12
        '   BRED Dwp Sv Q     log/c

       The uncertainty of dose monitoring arises from the uncertainties associated with each
term in the measurement equation (Wright and Mackey 2003, Taylor 1982) and is accounted for
by application of the validation factor (VF) in Section 5.9 and the application of quality
assurance/quality control during operation of the reactor at the WTP.
D.6    Considerations for CFD Modeling

       CFD UV dose modeling of the impact of UV reactor inlet and outlet conditions on RED
could be used in conjunction with one of the approaches outlined in Section 3.6. to assess
whether UV dose delivery at the WTP installation is equal to or better than UV dose delivery
achieved during validation. However, several issues with a CFD-based approach should be
considered:

       .   There is little agreement on appropriate procedures for assessing the credibility of
          CFD models.

          CFD models for prediction of UV dose delivered by a reactor comprise coupled sub-
          models for turbulent flow, microbial transport, UV intensity,  and microbial
          inactivation. Many options and approaches are available for each sub-model.
          Currently, no consensus has been reached for which approaches are most suitable for
          predicting UV dose delivery in a full-scale reactor.

       .   CFD modeling of UV dose delivery requires a multi-disciplinary approach.
          Knowledge of fluid mechanics, light physics, microbial inactivation,  numerical
          modeling, and UV process engineering is essential for credible CFD modeling of UV

UV Disinfection Guidance Manual                D-25                              November 2006
For the Final LT2ESWTR

-------
                    Appendix D. Background to the UV Reactor Validation Protocol
          dose delivery. The pool of this type of integrated expertise is currently limited, which
          presents a challenge for states tasked to review CFD modeling reports.

       A generalized modeling approach for predicting UV dose delivery involves the
following:

       1.  Construct a 3-D computational model of the UV system, including all major
          components that influence the flow patterns in the reactor. This includes resolution of
          all wetted surfaces in the reactor and the upstream/downstream piping systems.

       2.  Perform a steady-state CFD simulation by solving governing flow equations
          (i.e., Navier-Stokes and turbulence equations). This results in a prediction of point
          velocities across the interior of the UV system for the specified inlet flow rate.

       3.  Perform a UV intensity simulation for the UV system using a UV light intensity
          model. This results in a prediction of point UV intensity values across the interior of
          the UV system for specified values of UV lamp intensity and UVT.

       4.  Perform a particle tracking simulation using the combined numerical flow/UV
          intensity field. A random walk or particle physics model may be employed. Hundreds
          of numerical particles are randomly "injected" at the model inlet, and their x,y,z-
          coordinates are predicted as a function of time. The result is a predicted path line for
          each injected particle, which represents a random microbial path through the reactor.

       5.  Calculate the estimated UV dose for each injected particle by summing the
          cumulative UV dose at a series of points along the predicted particle path. The result
          is a UV dose distribution.

       6.  Determine the log inactivation and RED for a microorganism with known UV
          inactivation kinetics based on the UV dose distribution calculated in Step 5.

       If CFD is applied for simulation of UV dose delivery, it should adhere to the following
guidelines:

       1.  Only a qualified party with appropriate expertise should develop a CFD-based
          hydraulic or full UV reactor performance model. Such parties could include a
          professional engineer with extensive modeling experience, a CFD consulting firm, or
          a manufacturer with review by an independent CFD consultant.

       2.  The same overall modeling approach and sub-models should be used for both the
          validation site model and the WTP model. At a minimum, the following QA/QC
          procedures should be used during CFD model development and execution:

              The density of the numerical grid and size of the time step used in simulations
              affect CFD results. In general, results become more accurate as the grid becomes
              finer and the time step becomes smaller. Grid and time-step convergence  analysis
              should be performed to verify that grid and time-step sizes are sufficiently
              resolved such that smaller grid and time step sizes do not change predicted results.

UV Disinfection Guidance Manual               D-26                              November 2006
For the Final LT2ESWTR

-------
                    Appendix D. Background to the UV Reactor Validation Protocol
              Procedures for this analysis are presented in the Guide for the Verification and
              Validation of Computational Fluid Dynamics Simulations (AIAA 1998).

          .   Numerical convergence and consistency of the CFD models should be verified
              and documented. Procedures for this analysis are presented in the above-
              referenced AIAA guide.

              A sensitivity analysis of the major parameters that affect UV-dose prediction
              should be conducted. Examples include (but  are not limited to ) boundary
              conditions for lamp UV output and reactor wall reflection, number of particles
              used in a microbial transport simulation, and UV dose-response inactivation
              constants.

       3.  CFD models should not be calibrated with experimental RED data for the purposes of
          obtaining agreement between model predictions  and field measured values.
          Calibration to RED data for a limited set of conditions does not necessarily improve
          the accuracy of future predictions, particularly because hydraulic conditions can
          greatly differ between the validation site and the WTP installation.

       4.  Error estimates and confidence intervals for the CFD model predictions should be
          developed for both the validation site and the WTP installation. This could be
          performed by comparing CFD model predictions and experimental  data for the
          validation site, then  assuming the same level of error for the CFD model prediction
          for the WTP installation.

       Following the above guidelines, CFD can be used to predict the relative difference in
RED between a validation site and a WTP installation. If analysis indicates that UV dose
delivery is better at the WTP, RED credit should only be granted for  the experimentally
measured RED from the validation site.

       CFD-based UV dose modeling  should not be used in lieu of validation for prediction of
the actual RED magnitude as a  means of granting pathogen  inactivation credit. As discussed
previously, CFD is still an emerging technology, and CFD models for UV dose delivery  are
complex. Uncertainty and error ranges  for these models are  not known. CFD-based UV dose
delivery models would need to undergo a formal industry-wide verification and validation
process before they could be considered suitable for extrapolation of data for establishing
inactivation credit. A possible approach for verification and validation of hydraulic CFD models
is outlined in the AIAA CFD guide (1998).

       It is anticipated that CFD models for UV dose prediction will develop and improve in the
future. This manual is not intended to be the final word on CFD modeling for UV disinfection.
Engineers, regulators, and manufacturers should also consult with the AIAA manual and future
CFD guidance that may arise in the water industry.
UV Disinfection Guidance Manual                D-27                              November 2006
For the Final LT2ESWTR

-------
                      Appendix D. Background to the UV Reactor Validation Protocol
                                This page intentionally left blank.
UV Disinfection Guidance Manual                  D-28                                 November 2006
For the Final LT2ESWTR

-------
     Appendix E



UV Lamp Break Issues

-------
This page intentionally left blank.

-------
                              Appendix E. UV Lamp Break Issues
       The lamps in most UV reactors designed for water disinfection contain mercury or an
amalgam of mercury and another element, such as indium or gallium. Mercury has properties
that allow for cost-effective generation of UV light. These properties include a sufficient vapor
pressure at ambient temperatures to provide for efficient production of resonance radiation and a
low ionization energy to facilitate starting a lamp (Phillips 1983). Lamp manufacturers are
continuing to reduce the mercury content of UV lamps (USEPA 1997b; Walitsky 2001).
However, mercury-free lamps, such as pulsed UV lamps containing xenon, are not widely used
for water disinfection at present.

       The mercury contained within a UV lamp is isolated from exposure to water by the lamp
envelope (referred to as the "lamp" in this appendix for simplification) and a surrounding lamp
sleeve (Figure 2.13). However, breakage of a UV lamp creates the risk of exposure to mercury,
which can cause adverse health effects (USEPA 2006).

       Although UV disinfection utilizes UV lamps with mercury, UV disinfection is an
important disinfection technology that provides additional public health protection. To date,
there have been few lamp breaks at existing UV facilities. The risk to human health and the
environment from the mercury in UV lamps used in the treatment of drinking water is very
small. It can be addressed through engineering and administrative methods used to prevent UV
lamp breaks and exposure to mercury if breaks do occur, as described in this appendix.

       This appendix discusses the issues associated with breaks of UV lamps used for drinking
water disinfection. Lamp breaks are divided into off-line and on-line breaks.  Off-line breaks
occur when the lamps are not installed in the reactor or when the reactor is not in operation. On-
line lamp breaks occur when the lamp and lamp sleeve break during reactor operation.

       Sections E.I.I, E.2.1, and E.2.2 address potential causes of lamp breaks (including
known occurrences) and corresponding preventive measures. Sections E.2.3 and E.2.4 address
containment of mercury after a break and suggest components of a lamp-break response plan.
Regulatory issues associated with lamp breaks, including lamp disposal, are discussed in
Section E.3. Mercury in UV disinfection facilities and documented mercury reactions in PWS, is
discussed in Sections E.4 and E.5. A summary of the information presented in this appendix is
located in Section E.6. References for this appendix are presented in Chapter 7.
E.1    Off-line Lamp Breaks

       Off-line breaks occur when a lamp breaks during shipping, handling, storage, or
maintenance. Off-line breaks also can occur when the lamp and the lamp sleeve break in a UV
reactor that is not in operation. Because water is not flowing through the reactor, off-line breaks
do not pose a hazard to the water consumer but may be a hazard to operators or employees in the
vicinity of the break.
UV Disinfection Guidance Manual                 E-l                               November 2006
For the Final LT2ESWTR

-------
                             Appendix E. UV Lamp Break Issues
E.1.1  Potential Causes of Off-line Lamp Breaks and Corresponding Prevention
       Measures

       Off-line lamp breaks are caused by improper handling. The UV manufacturer should
train operators in proper handling and maintenance of UV lamps. In addition, lamps should be
stored horizontally in individual packaging to reduce the potential for lamp breaks. Lamps
should not be stacked unpackaged or propped vertically in corners (Dinkloh 2001).
E.1.2  Off-line Mercury Release Cleanup Procedures

       Water systems should have a lamp break response plan for containing and cleaning the
off-line spills. The local poison control center, fire department, or public health board can assist
in determining appropriate responses for different spill sizes and in developing a plan.

       Small spills can be contained and collected with commercially available  mercury spill
kits. Small spills are defined as the amount of mercury in a broken thermometer, or less than 2.25
grams (g) (USEPA 1992, USEPA 1997a). Given that the mercury content in a single UV lamp
typically ranges from 0.005 - 0.4 g (discussed in Section E.4.2), a single lamp break and
multiple lamp breaks that result in release of less than 2.25 g are categorized as small spills.

       Mercury and materials used during the cleanup procedure are regulated as hazardous
wastes and should be disposed of properly as described in Section E.3.3. EPA's  Office of
Superfund Remediation and Technology Innovation (formerly the Office of Emergency and
Remedial Response) recommends that"... [in] the event of a large mercury spill (more than a
broken thermometer's worth), immediately evacuate everyone from the area, seal off the area as
well as possible, and call your local authorities for assistance..." (USEPA 1997a).
E.2    On-line Lamp Breaks

       On-line lamp breaks occur when a lamp and lamp sleeve break while water is flowing
through the reactor. These breaks may have the potential to pose a hazard to the water consumer,
as well as to operators or employees in the vicinity of the break. This section discusses potential
causes of on-line lamp breaks, prevention measures, and documented occurrences of on-line
lamp breaks and mercury release.
E.2.1  Potential Causes of On-line Lamp Breaks and Corresponding Prevention
       Measures

       Lamp breaks can be caused by debris in the water, improper UV reactor orientation,
water temperature variations, exceeding positive or negative pressure limits (water hammer),
electrical surges, or improper maintenance. Lamps may also break as a result of manufacturing
defects in the lamp or improper selection of the lamp or sleeve material.
UV Disinfection Guidance Manual               E-2                              November 2006
For the Final LT2ESWTR

-------
                              Appendix E. UV Lamp Break Issues
E.2.1.1   Debris

       Debris may originate from the raw water or from treatment process equipment. Although
most UV reactors will be installed after the filters in the treatment train, upstream equipment
may release parts or fragments such as nuts or bolts that can break the lamp sleeves or UV
lamps. Unfiltered systems may be more prone to debris because there is minimal upstream
treatment to UV disinfection, which should be considered in the UV facility design.
Groundwater systems have the potential to pull stones or  gravel from wells that can enter UV
reactors and break lamps (Malley 2001, Roberts 2000).

       The consideration of prevention measures may be beneficial if debris historically has
occurred prior to installation of the UV facility. Placement of screens, baffles, or low velocity
collection areas upstream of UV reactors or vertical orientation (i.e., vertical flow of water
through the UV reactor) may reduce the risk of debris entering the reactor (Cairns 2000, Malley
2001, McClean 200Ib). Note that the lamps should be oriented horizontally relative to the
ground even if the UV reactor is installed with water flowing vertically as discussed in
Section E.2.1.2. The extent of containment these safety measures provide is unknown. Water
systems and designers should determine the applicability  of these techniques on a site-specific
basis.
E.2.1.2   Improper UV Lamp Orientation

       The orientation of UV lamps within a UV reactor can also increase the potential for lamp
breaks. Orienting lamps perpendicular to the ground can result in differential heating of the lamp
and the sleeve, which can lead to eventual cracking of the lamp and sleeve. As such, regardless
of the direction of water flow relative to the ground, UV lamps should be oriented parallel and
not perpendicular to the ground (Figure E. 1).
              Figure E.1. Example of Proper Horizontal and Improper Vertical
              UV Lamp Orientation in Reactors Relative to the Ground

A
^Installation
Piping


1
k
0
LJ_
Lamp^^


1

.
0
LJ_



******* Horizontal Vertical
Orientation Orientation
(recommended) (not recommended)
UV Disinfection Guidance Manual
For the Final LT2ESWTR
E-3
November 2006

-------
                              Appendix E. UV Lamp Break Issues
E.2.1.3   Loss of Water Flow and Temperature Increases

       UV lamps are designed to operate within a specific temperature range to maximize UV
light output. Without flowing water to cool the lamp, the lamp temperature can rise above the
maximum operating temperature and break (Dinkloh 2001, Malley 2001, Srikanth 200la,
Srikanth 2001b). There are two conditions that may cause overheating of lamps:

       .   Operating UV lamps while there is no water in the UV reactor (i.e., the lamp is in air)

          Operating UV lamps while water is not flowing through the UV reactor (i.e., the
          water in the UV reactor is stagnant)

       Overheating occurs much faster in air than in stagnant water and is more likely to occur
with medium-pressure (MP) than low-pressure (LP) lamps (due to lamp operating temperatures).
If UV lamps are energized in  air, the lower temperature water entering the reactor may cause the
lamp sleeve and the lamp to break due to temperature differentials (Dinkloh 2001, Malley 2001),
even if upper temperature levels are not exceeded. Lamp overheating and temperature
differentials could, therefore,  break all the lamps within the affected reactor.

       Operating a UV lamp  in stagnant water can also cause lamps to overheat and break.
Water flow during UV start-up (i.e., cooling water) cools the lamps and prevents lamps from
overheating and breaking. However, cooling water may not be necessary with low-pressure
high-output (LPHO) lamps (Haubner 2005). Whether cooling water is needed depends on the
specific MP reactor manufacturer, and the manufacturer should be contacted to determine this
(Leinberger 2005, Larner 2005, Bircher 2005).

       To prevent lamp breaks, operating procedures should ensure that the following conditions
are met:

       .   The lamps are not operating while the reactor is not full (i.e., while air is in the
          reactor).

          If recommended by the UV manufacturer, water should be flowing through the UV
          reactor if the UV lamps are operating.

       Hydraulics should be designed so that lamps are submerged at all times during reactor
operation. UV facility designs should also incorporate low flow alarms, air relief valves, or other
devices to ensure that lamps are operating only when the reactor is completely flooded and water
is flowing. UV equipment should and typically does include temperature sensors and alarms that
automatically shut down the reactor before critical temperatures are exceeded (Leinberger 2005,
Larner 2005, Bircher 2005, Dinkloh 2001, Malley 2001, Srikanth 200Ib).
E.2.1.4   Pressure-related Events

       Hydraulic pressures that exceed the operating limits of the lamp sleeves may break them.
Although breaking the lamp sleeve does not automatically break the lamp, the lamp is more
vulnerable when the sleeve has been damaged, potentially allowing the hot lamp to come into
direct contact with colder surrounding water.
UV Disinfection Guidance Manual                E-4                              November 2006
For the Final LT2ESWTR

-------
                              Appendix E. UV Lamp Break Issues
       Most lamp sleeves are designed to withstand continuous positive pressures of at least
120 pounds-force per square inch gauge (psig) (Roberts 2000, Aquafme 2001, Dinkloh 2001,
Srikanth 2001a, Srikanth 2001b). However, negative gauge pressures below -1.5 psig have been
shown to adversely affect lamp sleeve integrity (Dinkloh 2001). The pressure tolerance of the
quartz lamp sleeve varies and depends on the quality, thickness, and length of the sleeve.
Positive and negative pressures that exceed these levels, such as those associated with water
hammer, may cause the lamp sleeve to crack or break.

       Thus, water hammer can potentially break all lamps within an affected reactor. The water
system should perform a surge analysis to determine if water hammer is a potential problem, and
the UV facility designer should specify the pressure and flow ranges expected. The manufacturer
should provide lamp sleeves with the appropriate material, thickness, geometry,  and seals for the
specified pressure and should provide the water system with the lamp sleeve pressure tolerances.
E.2.1.5   Handling and Maintenance Errors

       A lamp or lamp sleeve damaged by improper off-line handling or maintenance may break
when the UV reactor is returned to service. For example, over-tightening compression nuts when
securing the lamp sleeve can cause a fracture of the lamp sleeve or a leak around the sleeve or
compression nut cavity (Aquafme 2001, Dinkloh 2001, Srikanth 200la, Srikanth 200Ib, Swaim
et al. 2002).  This problem may not become apparent until after start-up of the UV reactor and
may cause a lamp break. Operation and maintenance training can help prevent these types of
lamp breaks.
E.2.1.6   UV Reactor Manufacturing Problems

       The UV reactor manufacturer should design the UV reactor to reduce the possibility of
lamp breaks. This section describes UV manufacturing problems that may cause lamp breaks if
not properly addressed. Addressing these UV reactor manufacturing issues is typically the
responsibility of the manufacturer. However, some causes of lamp breaks can be mitigated
during the design of the UV facility.

Electrical Considerations

       If the UV facility electrical support system is improperly designed (e.g., inadequate
circuit breakers and ground fault indicator circuits), electrical surges can cause short-circuiting
and lamp socket damage (Srikanth 200la, Srikanth 200Ib). In addition, system electronics that
can produce voltages exceeding lamp ratings (overdriving lamps) may also cause the lamp to
break (Malley 2001).

       To reduce the likelihood of these problems, the UV facility designer should specify
circuit/ground fault interrupters (GFI) in the UV facility electrical design. In addition,
replacement UV lamps should be electrically compatible with the UV equipment.
UV Disinfection Guidance Manual                E-5                              November 2006
For the Final LT2ESWTR

-------
                              Appendix E. UV Lamp Break Issues
Cleaning Mechanism Considerations

       The cleaning mechanism may break the lamp sleeve and lamp if the mechanism is not
aligned properly. Although the cleaning mechanism closely surrounds the lamp sleeve for
cleaning, manufacturers should ensure that the mechanism is flexible and able to adjust to minor
misalignment of the lamp sleeves.

       At high lamp temperatures, the cleaning mechanism in some UV reactors may fuse to the
lamp sleeve when not in use. As a result, during the next cleaning event, the lamp sleeve may
crack when the cleaning mechanism is activated or when the cleaning mechanism passes back
over the residual left on the lamp sleeve (Dinkloh 2001). Some UV reactors are not subject to
this problem because the wipers rest away from the lamp sleeve when not in use and an alarm
sounds when the wiper stops along the lamp sleeve.

        Once the UV facility is in operation, the operators should perform routine inspections of
the inside of the UV reactors to ensure that the cleaning mechanism is not fused to the sleeve.

Thermal Expansion and Contraction

       Other potential causes of lamp breaks include improper matching of lamp materials with
respect to thermal expansion characteristics. Manufacturers should use compatible materials
within the lamp to avoid stress and damage from thermal expansion and contraction differences
between materials that can occur under various operating, shipping, or handling conditions
(Cairns 2000). In addition, improper seal design or lamp swelling can cause water leaks around
the seals that can result in electrical shorts and cracking of lamps (Cairns 2000).

       The UV facility designer should specify the temperature ranges likely to be encountered
during  shipping, storage, and lamp operation in the UV equipment procurement documents so
the manufacturer can select the appropriate materials.
E.2.1.7   Summary of Potential Causes and Methods of Prevention of On-line UV
          Lamp Breaks

       Table E. 1 summarizes the potential causes of on-line lamp breaks and briefly describes
the preventive measures that UV facility designers and operators can implement to reduce each
risk. Documented cases of on-line lamp breaks are discussed in Section E.2.2.
E.2.2  On-line Lamp Break Incidents

       Relatively few incidents of on-line lamp breaks with mercury release have been
documented. A literature review was conducted to compile information on UV lamp breaks in
operating UV facilities. Several facilities were contacted for more information about the
incidents. Although all documented lamp breaks involved MP lamps, some of the causes
reported for the lamp breaks are independent of the lamp type (Malley 2001). The documented
lamp break incidents, categorized according to the cause of the incident, are summarized in
Table E.2.

UV Disinfection Guidance Manual                E-6                              November 2006
For the Final LT2ESWTR

-------
                            Appendix E. UV Lamp Break Issues
                  Table E.1. Summary of Potential Causes and
                Methods of Prevention of On-line UV Lamp Breaks
Potential Cause
Debris
Lamp Orientation
Loss of Water
Flow and
Temperature
Increases
Pressure-related
Events
Maintenance and
Handling Errors
UV Reactor
Manufacturing
Problems
Description
. Physical impact of debris on lamp
sleeves may cause lamp breaks.
. Vertical installation relative to the
ground may cause overheating and
lamp breaks.
. Lamps may overheat and break.
. The temperature differential
between stagnant water or air and
flowing water (upon resumption of
flow) may cause lamp breaks.
. Excessive positive or negative
pressures may exceed lamp sleeve
tolerances and break the lamp
sleeve.
. Improper handling or maintenance
may compromise the integrity of
the lamp sleeve and/or lamp.
. Electrical surges can cause short-
circuiting and lamp socket damage.
. Applying power that exceeds
design rating of lamps can cause
lamps to burst from within.
. Misaligned or heat-fused cleaning
mechanism may break or damage
the lamp sleeve and lamp.
. Thermally incompatible materials
do not allow for expansion and
contraction of lamp components
under required temperature range.
Preventive Measure
. Installation of screens, baffles, or low
velocity collection areas upstream of UV
reactors or vertical installation of UV
reactors will help prevent debris from
entering the reactor.
. Install reactors with lamps oriented parallel
to the ground to reduce differential heating.
. Reactors should always be completely
flooded and flowing during lamp operation.
Temperature and flow sensors that are
linked to an alarm and automatic shutoff
system can be used to indicate irregular
temperature or flow conditions.
. A surge analysis should be completed
during design to determine the occurrence
of water hammer.
. Pressure relief valves or other measures
can be used to reduce pressure surges.
. Applicable pressure ranges should be
specified for lamp sleeves.
. Operators and maintenance staff should be
trained by the manufacturer.
. Adequate circuit breakers/GFI should be
specified to prevent damage to the reactor.
. Replacement lamps should be electrically
compatible with reactor design.
. Operators and maintenance staff should
perform routine inspection and maintenance
according to manufacturers'
recommendations.
. Designers should specify temperature
ranges likely to be encountered during
shipping, storage, and operation of lamps to
aid the manufacturer in the selection of
thermally compatible materials.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
E-7
November 2006

-------
                               Appendix E. UV Lamp Break Issues
          Table E.2. Mercury Release Incidents Involving UV Lamp Breaks
    Identified
     Cause
Number of
 Incidents
Description of Incident
 Debris
            (4) Stones entered the reactors and struck the lamps.1
            (1) Gravel entered the reactor through the booster pump and struck
            the lamp.2
 Loss of Water
 Flow and
 Temperature
            (2) Lamps were left on and allowed to reach high temperatures
            [600 degrees Centigrade (°C)] in empty non-operating reactors.1
            Restoration of flow caused cooler water (20 °C) to break the lamps.
 Operator Error
            (1) Forklift collided with on-line reactor.
 Manufacturer
 Design
            (1) Applied power exceeded the tolerances of the lamp, causing the
            lamp to burst from within.1
            (2) Vertical orientation of lamps in the reactor resulted in differential
            heating and eventual cracking of the lamp and sleeve because heat
            accumulated at the tops of the lamp and sleeve.1
            (1) High operating temperatures resulted in deformation of the lamp
            sleeve. The lamp sleeve sagged  and on contact with the lamp, both
            the lamp and lamp sleeve broke.4
            (1) Manufacturing defect. Lamps exploded after approximately 300
            hours of operation.5
            (2) Contaminated quartz material used by lamp manufacturer.6
 1 Survey of European water and domestic wastewater and hazardous waste treatment systems (Malley 2001)
 2 European drinking water systems (Roberts 2000)
 3 European brewery (Roberts 2000)
 4 UV-peroxide groundwater remediation reactor (Moss 2002a)
 5 Drinking water system (Region of Waterloo 2004)
 6 Drinking water system (Wright 2005)
       Impacts from debris caused five of the documented lamp breaks. In four of the five
incidents reported, UV lamps were oriented perpendicular to the flow of water, indicating that
lamps in this orientation may be more vulnerable to lamp breaks. However, the lamps in one
instance were parallel to the flow of water, so orientation alone will not prevent lamp breaks.

       An additional incident involving debris occurred when a bolt from the filter underdrain
broke a lamp sleeve. The lamp was not broken by the bolt, and mercury was not released because
the UV equipment was immediately shut down to respond to the sleeve break (McClean 200la).
Because no mercury was released, the incident is not included in Table E.2; however, this
incident indicates that equipment debris can also be hazardous.

       Loss of water flow and the resulting increase in lamp temperature caused two of the
documented lamp breaks. In these cases, the operating lamps reached extremely high
temperatures (> 600 °C) in air. When water flow resumed, the cooler water (20 °C) caused the
lamps to break (Malley 2001).  These  incidents can be prevented if UV equipment has a safety
mechanism that will shut down the UV lamps if flow decreases or lamp temperature  significantly
increases (Malley 2001).
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                            Eo
                           -O
                            November 2006

-------
                              Appendix E. UV Lamp Break Issues
       Operator error caused one of the documented lamp breaks. A forklift was driven into an
operating reactor and physically damaged the UV reactor. The event activated an alarm and
pneumatic valve closure, which contained the mercury release (Roberts 2000).

       The seven remaining lamp breaks are attributed to improper manufacturer design. In one
of the lamp breaks, 30-kilowatt (kW) power was specified for the application. However, a
manufacturing error resulted in a higher power being applied and caused the lamp to burst from
within (Malley 2001).

       Another manufacturer design problem that resulted in two breaks was vertical orientation
of the lamps within the UV reactor. The vertical orientation allowed heat to accumulate at the
tops of the lamp and sleeve, which caused them to break (Malley 2001). It is worth noting that
modern UV reactors do not mount lamps vertically, even in vertically oriented reactors such as
those discussed in Section E.2.1.2.

       Another lamp break attributed to a manufacturer design flaw resulted from deformation
of the lamp sleeve at operating temperatures. The incident occurred in a UV-peroxide reactor
designed for well-head treatment of tetrachloroethene-contaminated groundwater (Moss 2002a).
The UV reactor was positioned between the groundwater extraction pump and the distribution
system booster pumps. The 7-foot long MP lamp sleeve sagged and came into contact with the
lamp.  The lamp and lamp sleeve broke, releasing mercury. The lamp failure triggered an alarm,
shutting down both the groundwater extraction and distribution system booster pumps. Liquid
mercury was found on the bottom of the reactor. Water samples taken at a nearby fire hydrant
were positive for mercury but were below the maximum contaminant level (MCL) of
2 micrograms per liter (|ig/L) (Moss 2002a, Moss 2002b).

       Similar to the prior incident, a manufacturing defect in an MP lamp caused several lamps
at a WTP to explode after approximately 300 hours of operation,  releasing mercury into the
water (Tramposch 2004). The break occurred in a large 47-inch diameter horizontal reactor. The
lamp failure alarm triggered closure of the reactor isolation valves within 90 seconds and
initiated automatic flushing of the clearwell. The quartz fragments and 64 percent of the mercury
were recovered in the bottom of the reactor. The baffles within the reactor appear to have
prevented the mercury and quartz from leaving the reactor after the break. The flushed clearwell
water was sent to an on-site holding tank where it was tested for mercury, which was not
detected. Mercury was also not detected in the piping between the UV reactors and the clearwell
and in the UV reactor drain water. The water system believes that the remainder of the mercury
was fused to the reactor walls because mercury was not  discovered downstream, and mercury
vapor was detected in the UV reactor when the hazardous materials (HazMat) contractor was
cleaning the UV reactor.

       Twenty four hours after the reactor was drained in response to the lamp break, mercury
vapor concentrations within the reactor exceeded health and  safety limits (Section E.3.2),
although mercury vapor concentrations in the ambient air surrounding the UV reactor were not
above these safety limits. During all stages of the cleanup operation, the HazMat contractor
ensured that the area was well-ventilated and monitored for mercury vapor. As a result of this
incident, the water system observed that cleaning the reactor quickly is imperative because
mercury vapor can accumulate in  the reactor (Region  of Waterloo 2004).
UV Disinfection Guidance Manual                E-9                              November 2006
For the Final LT2ESWTR

-------
                              Appendix E. UV Lamp Break Issues
       The two final documented lamp breaks occurred because of the use of contaminated
quartz material by the lamp manufacturer. The contamination weakened the protective quartz
sleeve and the lamp, resulting in breaks at two water treatment plants (WTP).
E.2.3  Design Considerations for Containment after a Lamp and Sleeve Break

       This section briefly describes potential methods to contain mercury from a lamp break.
However, the extent of containment provided by these measures is unknown. Water systems and
designers should determine the applicability of these isolation techniques on a site-specific basis
and include the specific steps to be taken in the water system's response plan.

       To isolate the mercury in the reactor or downstream, water systems may install spring-
return actuated valves with a short closure time on the reactor inlet and outlet piping (McClean
2001b). Given the short residence time of many MP reactors, the outlet-side valve should be
located far enough downstream so that the valve has time to close and isolate the mercury
upstream. UV facility designers should evaluate valve closure times with respect to the potential
for water hammer.

       Condensed mercury and quartz fragments may be contained and collected in areas  of low
water velocity such as the bottom of a shut-down reactor, sumps, or a clearwell. To prevent
quartz fragments from  entering the water system, a strainer can be installed on the reactor  outlet
piping  (McClean 200Ib, Srikanth 200la, Srikanth 200Ib). Another option is to include a
mercury trap in the design (Figure E.2). A mercury trap could include a tee fitting  after the UV
reactor. Flow will  enter the tee and flow upward. The tee may also include an elbow that is
sealed but accessible. If the water velocity in the tee fitting and following pipe is low enough,
some of the mercury and quartz fragments may settle out in bottom of the elbow (Mutti 2004).
The head loss associated with such measures should be considered in the hydraulic profile.
Designers should also consider installation of drains and piping to allow disposal of potentially
contaminated water from the reactor or trap to a waste container or truck. The effectiveness of a
strainer and mercury trap has not been evaluated and is unknown.
E.2.4  On-line Lamp Break Response Plan

       On-line lamp breaks should be preventable with appropriate design and operation of UV
reactors. However, water systems should develop a written lamp break response plan in case an
on-line UV lamp break occurs. Water systems should coordinate with their state when
developing the following plan components:

       .   Identification of a lamp break
       .   Site-specific containment measures
       .   Mercury sampling and compliance monitoring
          Site-specific cleanup procedures
       .   Reporting requirements
       •   Public notification requirements

UV Disinfection Guidance Manual                E-10                              November 2006
For the Final LT2ESWTR

-------
                               Appendix E. UV Lamp Break Issues
                      Figure E.2. Example of a Potential Mercury Trap*
                              UV Reactor
                                       Water
                                        Flow
                  Blind flange allows for
                   the removal of the
                   mercury and quartz
                                   Isolation
                                   Valve
                             Mercury and quartz
                              may collect in the
                             bottom of the elbow*
                      The effectiveness of the mercury trap has not been evaluated and is unknown.
Identification of a Lamp Break

       UV reactors should be equipped with alarms that are activated when a lamp fails. A lamp
failure alarm may be due to a lamp break or to another problem. Because alarms associated with
lamp failure and GFIs may be due to lamp and sleeve breaks, the UV equipment should be shut
down, isolated in response to these alarms, and inspected to determine whether a lamp break
occurred.

Site-specific Containment Measures

       In the event of a lamp failure alarm, the UV reactor should be immediately shut down,
and operators should assume a lamp break has occurred and implement the procedures to contain
the mercury while determining the cause of the alarm. The containment procedures should be
outlined in detail in the water system's response plan based on the specific UV facility and any
containment measures included in the design.

Mercury Sampling

       Mercury sampling should be implemented after an on-line UV lamp break. Sampling
procedures should specify sample locations, frequencies, and analysis methods. Sampling
frequencies should consider flow rate, detention time, and travel time to the first potential
consumer. Sample locations should be chosen based on where the mercury may settle (e.g., low
velocity areas) and where mercury vapor may accumulate (e.g., a drained UV reactor). Table E.3
lists  some possible sample locations (Region of Waterloo 2004,  Stantec 2004).
UV Disinfection Guidance Manual
For the Final LT2ESWTR
E-ll
November 2006

-------
                               Appendix E. UV Lamp Break Issues
                        Table E.3. Mercury Sampling Locations
     Media
           Location
              Purpose
   Water
Reactor drain
Piping downstream of the UV
reactor, including the distribution
system entry point at a minimum
Low velocity areas, such as
clean/veils
Assess the extent of mercury
contamination and identify areas requiring
cleanup.
   Air1
Reactor or other locations where
mercury vapor may collect
Ambient air
Assess whether it is safe to access
mercury-contaminated equipment and
piping for cleanup. The UV reactor interior
may be accessible through an air vent.
Assess whether adequate ventilation  is
provided to safely proceed with mercury
cleanup.
     Methods for air sampling are available from the Occupational Safety and Health Administration (OSHA) at
     http://vwvw.osha.gov/dts/sltc/methods/inorganic/id140/id140.html.
Site-specific Cleanup Procedures

       Site-specific cleanup procedures should be incorporated into the water system's response
plan. Issues to consider are assessment of mercury contamination in the air, water, or on
surfaces, disposal of any isolated or condensed mercury, potential disposal or treatment of
contaminated water, cleanup responsibilities (by water system staff or contracted hazardous
materials team), and Federal or state cleanup or disposal requirements.

       An example of a currently operating UV facility's site-specific clean-up procedures is
summarized below. The procedure includes the following major steps (Stantec 2004):

       1.   Hydraulically isolate the UV reactor.

       2.   Ventilate the area and shut down ventilation equipment that circulates air to other
           parts of the building.

       3.   Wear personal protective equipment, including gloves, eye protection, suits, shoe
           covers, and breathing protection.

       4.   Drain water from the reactor through a mesh filter into a tank for disposal.

       5.   Measure the mercury vapor concentration within the reactor and ensure that it is at an
           acceptable level (limits shown in Section E.3.2).

       6.   Open the reactor and remove quartz and  mercury from the reactor using a mercury
           spill kit.

       7.   Perform a mass balance to assess how much mercury has been recovered.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                            E-12
                            November 2006

-------
                              Appendix E. UV Lamp Break Issues
Reporting and Public Notification Requirements

       The water system should determine any reporting and public notification requirements by
coordinating with the state. If reporting or public notification is required, the response plan
should include the information that must be reported to the state and the notification procedures.
Reporting requirements may include a description of the release, estimated quantity of the
release, shut-down or containment procedures, cleanup or disposal methods, and sampling
procedures (including sampling locations, frequencies, and results).
E.3    Regulatory Review

       This section presents a review of regulations that may apply if UV lamp breaks occur at a
WTP.


E.3.1  Safe Drinking Water Act

       Under the Safe Drinking Water Act (SOWA), EPA established a primary MCL of 2 |ig/L
for inorganic mercury [40 CFR 141.62(b)] and the associated monitoring requirements. The limit
was designed to protect against mercury contamination in the source water and not a transient
event like lamp breaks. Consequently, the water system should contact the state to determine
whether additional mercury monitoring will be required in response to lamp breaks.


E.3.2  Operator Health and Safety - Exposure Limits

       Mercury exposure to employees in WTPs falls under the regulatory authority of OSHA.
The exposure limits set by OSHA focus on exposure through inhalation. OSHA regulations have
established permissible exposure limits (PELs) for mercury compounds and organo alkyls
containing mercury. A PEL is a time-weighted average concentration that is not to be exceeded
for an 8-hour workday during a 40-hour workweek. When a PEL is designated as a ceiling level
(cPEL), the concentration cannot be exceeded during any part of the workday. PELs and cPELs
are enforceable standards. The National Institute for Occupational Safety and Health (NIOSH)
also publishes Immediately Dangerous to Life or Health (IDLH) concentrations for a variety of
compounds. IDLH concentrations represent the maximum concentrations that one could escape
within 30 minutes without symptoms of impairment or  irreversible health effects. These values
are not enforceable, but can be used as guidance for safety procedures. Table E.4 lists the PELs,
cPELs, and IDLHs for mercury compounds and organo alkyls containing mercury.

In the event of a spill, the volatilization and the resultant mercury vapor concentration depends
on air currents, temperature, surface area/dispersion of  mercury droplets, and time. If a mercury
spill is not cleaned up promptly, the levels in Table E.4 may be exceeded where mercury vapor
collects (e.g., drained UV reactor). For example, in the  lamp break described in Section E.2.2,
these limits were exceeded within the reactor 24 hours after the reactor was drained.  However,
prompt response and proper cleanup procedures (e.g., ventilation and other measures described
in Section E.2.4) should prevent exposure levels over these  standards.
UV Disinfection Guidance Manual                E-13                              November 2006
For the Final LT2ESWTR

-------
                              Appendix E. UV Lamp Break Issues
      Table E.4. Health and Safety Standards for Mercury Compounds in Air
Compound
Mercury compounds
Organo alkyls containing
mercury
OSHA PEL
(mg-Hg/m3)1
NR
0.01
OSHA cPEL
(mg-Hg/m3)
0.1
0.04
NIOSH IDLH
(mg-Hg/m3)
10
2
         NR - not reported
         1 milligrams mercury per meter cubed
E.3.3  UV Lamp Disposal Regulations

       UV lamps must be disposed of properly, as described in Section 6.3.2.6 and should be
recycled. Some UV reactor and lamp manufacturers will accept spent or broken lamps for
recycling or proper disposal (Dinkloh 2001, Leinberger 2002, Gump 2002). Alternatively, water
systems should contact their state primacy agency or other local or state resource agencies for a
list of local mercury recycling facilities.
E.3.4  Clean Water Act

       Mercury discharges to water bodies in the United States are regulated under the Clean
Water Act. Mercury-contaminated water from a lamp break should not be discharged to the
environment through storm sewers or other means; discharges should be coordinated with the
state and the local wastewater authority for proper treatment and disposal.
E.4    Mercury in UV Disinfection Facilities

       Understanding the type of mercury and amount of mercury present in UV disinfection
facilities can help determine the potential dispersion and transport of mercury through a WTP.
However, the fate and transport of mercury after a lamp break has not been assessed by the
drinking water industry.
E.4.1  Type of Mercury in UV Disinfection Facilities

       Characterizing the form of mercury in an operating lamp is important because this form
represents the starting point for mercury dispersion, speciation, and reaction chemistry in the
water following a lamp break. Mercury in an LP or MP UV lamp is pure elemental mercury
while LPHO lamps use a mercury amalgam, which typically is an alloy with indium.

       Elemental mercury is usually a liquid at ambient temperature and pressure. However,
given its vapor pressure (Table E.5), elemental mercury can vaporize at ambient temperatures.
Other physical and chemical properties of elemental mercury that affect its fate and transport are
given in Table E.5.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
E-14
November 2006

-------
                              Appendix E. UV Lamp Break Issues
                   Table E.5. Physical and Chemical Properties of
                     Elemental Mercury (Merck & Co., Inc. 1983)
Property1
Density (g/mL1 at 25 °C)
Solubility (g/L2 at 25 °C)
Vapor pressure (mm Hg at 25 °C)
Value
13.534
0.062
0.002
                     1
                     2  Further information regarding mercury solubility in water
grams per milliliter; grams per liter
Further information regarding mercL
can be found in Glew and Names (1971).
       In operating lamps, elemental mercury (from pure or amalgamated mercury) is vaporized
in the presence of an inert gas. The concentration of mercury in the vapor phase is controlled
predominantly by temperature. At typical LP and LPHO lamp operating temperatures, only a
small portion of the liquid (pure) or solid (amalgam) mercury is vaporized. However, at typical
MP lamp temperatures (600 to 900 °C; Table 2.1), mercury is present primarily in the vapor
phase due to the high operating temperatures (Phillips 1983).

       The relative proportion of mercury in the liquid/amalgam phase and in the vapor phase in
an operating lamp may affect the fate of the mercury. (See Section E.5.) Liquid-phase elemental
mercury is considerably denser than water (density = 13.5 g/mL; Table E.5).

       As the UV lamp is operating, mercury-containing compounds can be formed on the
internal lamp surface (Altena et al. 2001). After a break, these deposits may dissolve in water,
releasing mercury into the water (Merck & Co. 1983).

       Figure E.3 illustrates the expected forms of mercury in an operating lamp. Note that
much of the elemental mercury will volatilize in an operating MP lamp and that amalgams are
only used in LPHO lamps.
E.4.2  Amount of Mercury in UV Disinfection Facilities

       The amount of mercury in a UV disinfection facility is site-specific and can be calculated
using the amount of mercury per lamp, the number of lamps per reactor, and the number of
reactors in the facility. This section contains information on the amount of mercury in UV lamps
and uses this information in example calculations showing the amount of mercury contained in
hypothetical UV facilities. This information is provided as an order of magnitude range of
mercury levels that could be present in UV facilities.

       Mercury content within lamps depends on type (LP, LPHO, or MP), length, and power
rating.  Although mercury content data are specific to manufacturer and lamp, longer lamps and
lamps with higher pressures  and power ratings typically contain more mercury. Table E.6
summarizes the quantities of elemental mercury added to lamps during manufacturing based on
information provided by manufacturers and published literature values.
UV Disinfection Guidance Manual                E-15                              November 2006
For the Final LT2ESWTR

-------
                               Appendix E. UV Lamp Break Issues
                   Figure E.3. Mercury Speciation in Operating UV Lamps
                           Hg-contauiiiig deposits, e.g., HgO,
                            Hg1
                               o
                               (liquid)
OR    He1
                                                    o
           (amalgam)
                      Source of vapor-phase mercury
                          in LP or MP lamps
     Source of vapor-phase mercury
         in LPHO lamps
                 Table E.6. Elemental Mercury Content in UV Lamps
Lamp
Type
LP
LPHO
MP
Electrical
Power Rating
[Watt (W)]
15-70
120-260
400
1000
1 - 25 kW
Mercury Content (mg per lamp)
Phillips
(1983)
"a single drop"(1)
NR
NR
NR
1.4-14.5
mg/cm(5)
Clear and
Berman(1994)
20(2)
26,(3) 36(4)
75.5
250
NR
Manufacturer
Survey
5-50
150
NR
NR
200-400,
0.3 - 7.9 mg/cm length
       1  Phillips (1983)
       2  75-W mercury vapor lamp
       3  175-W mercury vapor lamp
       4  250-W mercury vapor lamp
       5  milligram per centimeter (mg per cm) of lamp length, reported lamp lengths are 6 - 300 cm
         (Primarc Limited 2001)
       NR - Not Reported
       The amount of mercury in a UV facility can be estimated using the values in Table E.6 as
a guide. In order to develop these estimates, two UV reactor manufacturers established design
parameters for three treatment flowrates [0.18, 3.5, and 210 million gallons per day (mgd)] with
a specified water quality and required UV dose and validation reduction equivalent dose (RED)
target (Table E.7). Design parameters included the number of lamps needed to obtain an MS2
phage RED of 40 millijoule per centimeter squared (mJ/cm2)1 during validation testing and the
total number of reactors for each of the three design flows. Calculations assume 50,  150, and 400
milligrams (mg) of mercury per LP, LPHO, and MP lamp, respectively. When determining the
amount of mercury at a specific UV facility, water systems should contact the lamp manufacturer
for updated information because mercury content varies with lamp type and manufacturer.
1 Corresponds approximately to a 3 log Cryptosporidium inactivation (depending on validation testing and
associated Validation Factor)	
UV Disinfection Guidance Manual
For the Final LT2ESWTR
 E-16
November 2006

-------
                              Appendix E. UV Lamp Break Issues
               Table E.7. Mercury Quantity in Example UV Facilities1'2
Design
Flow
(mgd)
0.18
3.5
210
Average
Flow
(mgd)
0.054
1.4
120
Lamp
Type
LP
LPHO
MP
LPHO
MP
LPHO
MP
Average
Number of
Reactors
1
1
1
1
1
6
6
Average Number
of Lamps
(per reactor)
2
1
1
30
4
72
7
Total Hg in UV
Facility3
(g)
0.1
0.2
0.4
4.5
1.6
64.8
16.8
        1 Target MS2 phage RED of 40 mJ/cm2, which corresponds approximately to a 3 log
         Cryptosporidium inactivation (depending on validation testing and associated Validation Factor)
        2 Water quality criteria: Ultraviolet transmittance (UVT) = 89% (A254= 0.05 cm"1),  Turbidity =
         0.1 nephelometric turbidity units (NTU),
         Alkalinity = 60 mg/L as CaCO3, Hardness = 100 mg/L as CaCO3
        3 Values given represent the amount of elemental mercury added to lamps during manufacturing.
E.5    Documented Mercury Reactions in PWSs

       Currently the fate of mercury following a lamp break has not been experimentally
determined. This section describes documented mercury reactions in water systems.

       Liquid elemental mercury and solid mercury amalgams have high densities (Table E.5)
and will probably settle in areas of low water velocity, providing an opportunity for containment
and removal.  In prior cases when liquid mercury was released from water treatment equipment,
such as manometers, flow instrumentation, or pump seals, mercury was found to have settled in
the clearwell, but whether all of the released mercury was recovered is not known (Cotton 2002).
Smaller particles from the vapor phase mercury may be transported farther or be more readily
dissolved in water than liquid elemental mercury and solid mercury amalgams. However, in
sampling following a recent on-line MP lamp break (described in Section E.2.2), mercury was
not detected in any of the downstream sample locations, which could indicate that the mercury
was contained by the UV reactor. The water system theorized  that the remaining mercury was
potentially  attached to the UV reactor walls.

       Liquid-phase elemental mercury does not readily dissolve in water. Kolch (2001)
monitored the mercury concentrations in a 50-L batch reactor  following the destruction of one
LPHO lamp (containing approximately 150 mg Hg). Mercury  concentrations reached
approximately 2.5 |ig/L in the batch reactor water, and amalgamated mercury was found settled
on the bottom of the reactor (Dinkloh 2001). The low concentration of dissolved mercury in the
water is likely an indication that little, if any, of the mercury amalgam dissolved into the water.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
E-17
November 2006

-------
                              Appendix E. UV Lamp Break Issues
E.6    Summary and Conclusions

       UV disinfection is an important disinfection technology that provides additional public
health protection. To date, there have been few lamp breaks at existing UV facilities. The risk to
human health and the environment from the mercury in UV lamps used in the treatment of
drinking water is very  small. Procedures and actions can be taken to reduce the chances of a
lamp break and mitigate mercury release that UV lamp breaks cause. In addition, monitoring of
mercury after known lamp breaks indicates that most of the mercury is contained, and
concentrations in the water downstream of the UV reactor do not exceed the SDWA MCL.
However, more research is needed to understand the fate of mercury in a drinking water
environment following a UV lamp break and to evaluate the dispersion and transport of mercury
through a WTP and distribution system.

       Lamp breaks are divided into off-line and on-line breaks. Off-line lamp breaks typically
occur during storage, handling, or maintenance and cause small spills. Small spills should be
contained, cleaned  up, and disposed of properly. Monitoring of mercury vapor concentration in
the ambient air is important to protect personnel during the clean-up procedures.

       On-line breaks occur when the lamp sleeve and lamp break while the UV reactor is in
operation. Incidents have been reported of on-line UV lamp breaks associated with impact from
debris, improper UV reactor orientation, loss of water flow, temperature differentials, faulty UV
equipment design, procedural errors, and manufacturing defects. However, on-line lamp breaks
are largely preventable with appropriate design, operation,  maintenance, and operator care. The
following engineering and administrative methods may help prevent UV lamp breaks:

       .  Screens, baffles,  or low-velocity collection areas prior to the reactor influent to
          prevent entrance of debris

       .  UV reactor installation with lamps oriented parallel to the ground to reduce
          differential heating

          Temperature and flow sensors and alarms to detect critical conditions and to shut
          down the UV reactors and water flow

       .  Surge analysis to determine if water hammer may be a potential problem or whether
          pressure relief values need to be installed

       .  Comprehensive operator training and UV equipment maintenance program

          Adequate circuit breakers/GFIs should be specified to prevent damage to the reactor.

       .  Operators and maintenance staff should perform routine inspection and maintenance
          according to manufacturers' recommendations.

       .  Designers should specify temperature ranges likely to be encountered during
          shipping, storage, and operation of lamps to aid the manufacturer in the selection of
          thermally compatible materials.
UV Disinfection Guidance Manual                E-18                              November 2006
For the Final LT2ESWTR

-------
                               Appendix E. UV Lamp Break Issues
       In the event of a mercury release, the following engineering controls are additional
precautions that may aid in the containment and collection of mercury:

       .   Strainers and low velocity collection areas downstream of the reactor

          Isolation valves activated by an alarm to attempt to isolate potentially contaminated
          water

       The extent of containment and prevention these measures provide is unknown. Water
systems and designers should consider the applicability of these isolation techniques on a site-
specific basis. Water systems should prepare a lamp break response plan in preparation for a
potential UV lamp break and mercury release. This plan should address sampling and cleanup
procedures as well as compliance with the SDWA, OSHA health and safety standards, and Clean
Water Act. Water systems are encouraged to recycle or return all mercury-containing lamps to
mercury re-generating facilities or the lamp manufacturer.
UV Disinfection Guidance Manual                E-19                               November 2006
For the Final LT2ESWTR

-------
                                 Appendix E. UV Lamp Break Issues
                               This page intentionally left blank.
UV Disinfection Guidance Manual                 E-20                                November 2006
For the Final LT2ESWTR

-------
 Appendix F



Case Studies

-------
This page intentionally left blank.

-------
                                   Appendix F. Case Studies
       This appendix provides examples of how various utilities have implemented UV
disinfection in their water systems. The UV facilities described in the following case studies
were selected because they represent a broad range of UV facility conditions. They represent
medium-pressure (MP) and low-pressure high-output (LPHO) reactor installations; on-site and
off-site validation testing; installation on filtered water, unfiltered water, and an uncovered
reservoir; and other varying goals and design issues.

       The purpose of this appendix is to provide an overview of the manner in which UV
disinfection has recently been implemented for drinking water disinfection in North America.
The case studies describe issues and approaches used to implement UV disinfection technology.
Specific step-by-step procedures for selecting design criteria and validation are not described.
Rather, each case study provides a summary of the reasons for implementing UV disinfection,
the design issues that were considered, and how implementation was approached. They are
meant to be instructive as examples of how UV disinfection can be applied across a range of
source waters, equipment types, and retrofit locations. It is important to follow the specific step-
by-step guidance and examples provided in the previous sections of this manual to ensure that
the final guidance is appropriately applied to  any new installations.

       The organization of each case study generally follows the organization of this manual.
Each study provides introductory information about the water system and a discussion of the
planning, design, validation, and operation and maintenance steps completed by each public
water system (PWS). The first two case studies (Albany, New York and Weber Basin Water
Conservancy District, Utah) feature in-depth  descriptions, and the remaining three case studies
contain briefer summaries of similar information.

       When reading these case studies, it is important to recognize that these facilities were
implemented before the Long Term 2 Enhanced Surface Water Treatment Rule (LT2ESWTR)
and the guidance provided in this manual were finalized. Although drafts of this manual were
available, some of the guidance has changed over time. In particular, the validation approaches
and testing programs have changed since these projects were implemented.

       Following are some of the highlights from each case study:

       •   Section F.I- Albany, New York.  MP reactors installed on an uncovered finished
          water reservoir that experiences bi-directional flow.

       •   Section F.2 - Weber Basin Water Conservancy District, Utah. LPHO reactors
          validated off-site at the Portland Validation Facility.

       •   Section F.3 - Clayton County Water Authority, Georgia. LPHO reactors in which
          challenge microorganism die-off problems were resolved to allow for  on-site
          validation to proceed.

       •   Section F.4 - Newark, Ohio. MP reactors installed at a lime-softening facility on
          individual filter effluent pipes.

       •    Section F.5 - Winnipeg, Manitoba. MP reactors installed on an unfiltered source.

UV Disinfection Guidance Manual                F-l                              November 2006
For the Final LT2ESWTR

-------
                                   Appendix F. Case Studies
F.1    Albany, New York - MP Facility on a Finished Water Reservoir with
       On-site Validation

       The City of Albany (City) owns and operates a 32-million gallons per day (mgd)
conventional surface water treatment plant (WTP), serving over 100,000 people. The Feura Bush
WTP operates at a relatively constant treatment rate (typically about 20 mgd). The Loudonville
Reservoir, a finished water reservoir on the opposite side of the City, floats on the distribution
system, filling and emptying throughout the day as distribution system demand fluctuates.

       Loudonville Reservoir has two functions—distribution storage and emergency/backup
supply. The reservoir is a 200-million gallon, uncovered, finished water storage facility,
consisting of three basins. The reservoir has two inlets/outlets to the distribution system, and
reservoir effluent water is automatically rechlorinated before delivery to the distribution system.
In addition to rechlorinating the water as it re-enters the City's distribution system, the City
periodically batch chlorinates the reservoir to maintain water quality.

       The City expanded its water quality enhancement program at the Loudonville Reservoir,
which consisted of a series of water system improvements, including UV disinfection, being
made under the direction of Albany's Mayor Gerald Jennings to ensure that  customers receive
the best possible water quality at all times.

       Early in the project planning phase, the City and its consultant determined that UV
disinfection offers the most flexible and holistic solution for improving the reservoir water
quality. UV disinfection provides the City with an additional disinfection barrier that is compact,
relatively simple to operate, free from regulated disinfection byproducts (DBF), and effective
against chlorine-resistant pathogens.

       Few data were available on the water quality  at the reservoir. Therefore, the reservoir
water quality was assumed to be similar to the Feura Bush WTP finished water quality, which is
summarized in Table F. 1.
F.1.1  Planning

       This section discusses key planning decisions made for Albany's UV facility. Figure F.I
is a timeline of the process the City used to implement UV disinfection.
UV Disinfection Guidance Manual                F-2                               November 2006
For the Final LT2ESWTR

-------
                                      Appendix F. Case Studies
       Table F.1. Summary of Feura Bush WTP Finished Water Quality (2000)
Parameter
UVAbsorbance(1)
UV Transmittance(1)
Turbidity
PH
Alkalinity
Temperature
Total Hardness
Iron
Manganese
Aluminum
Specific Conductance
Units
cm"1
percent
NTU(2)
-
mg/L(3)-CaC03
°C
mg/L-CaCO3
mg/L
mg/L
mg/L
m-mhos/cm(4)
Average
0.03
93
0.23
8.40
40.9
10.4
54.2
<0.03
<0.03
0.07
176
Minimum
0.011
88
0.12
7.80
35.7
1.1
50.0
<0.03
<0.03
0.07
148
Maximum
0.054
98
0.54
9.20
48.3
20.0
58.0
0.03
<0.03
0.07
211
         1 Data collected January 2001 -September2001.
         2 nephelometric turbidity units
         3 milligrams per liter
         4 millimhos per centimeter
                     Figure F.1. Albany UV Implementation Timeline
                           2001
                                             2002
                                                              2003
                                                                                2004
        Monitoring
        Planning
             UV
           Disinfection
            Planning
        Design
UV disinfection
system bids
received Nov.
2001. Equipment
pre-purchased.
                                         Design
                                                       Construction contracts
                                                       awarded Aug. 2002.
        Construction.
        Validation &
        Start-Up
 Bids received for piping and
 valves. Materials pre-
 purchased April 2002.
                                       J
A Validation Testing,
~ Oct. 2003.
F.1.1.1    UV Disinfection Goals
        The City chose a multiple-barrier approach for disinfection at the reservoir, incorporating
both UV disinfection and chlorination, to provide a greater level of protection. The City's UV
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                           F-3
               November 2006

-------
                                  Appendix F. Case Studies
light and chlorine disinfection systems provide more effective inactivation of viruses and
chlorine-resistant pathogens than the former chlorine-only system, while minimizing DBF
formation. Additionally, the UV facility provides an additional level of protection to facilitate the
City's compliance with the LT2ESWTR inactivation requirements for uncovered storage
(Section 1.3.3).
F.1.1.2   UV Retrofit Location

       The UV facility is located at the reservoir rather than at the WTP. Flow from each of the
three reservoir basins is routed through the UV disinfection facility (Figure F.2) before it enters
the City's distribution system.
              Figure F.2. UV Retrofit Location at Loudonville Reservoir
                                                       City
                                                   Distribution
                                                     System
F.1.1.3   Key Design Parameters

       Water quality, the fouling/aging factor, and flow rate are key parameters to be considered
during the planning phase. Table F.2 summarizes key preliminary design parameters for the UV
facility design.
                Table F.2. UV Facility Preliminary Design Parameters
Criterion
UV Transmittance
Fouling/Aging Factor
Peak Flow Rate
Unit
percent
percent
mgd
Value
88
60
40
UV Disinfection Guidance Manual
For the Final LT2ESWTR
F-4
November 2006

-------
                                  Appendix F. Case Studies
Water Quality

       Several water quality parameters affect UV dose delivery and, therefore, UV equipment
design (Table F. 1). The most important is ultraviolet transmittance (UVT), which is calculated
from A254 as described in Section 3.4.4.1. Reservoir UVT data were collected for approximately
3 months prior to design, but long-term UVT data were not available for the reservoir. The
minimum UVT of 88 percent (A254 of 0.054) measured in the WTP finished water, therefore, was
used to conservatively estimate UVT at the reservoir (Figures F.3 and F.4).
              Figure F.3. UVT Data for Feura Bush WTP Finished Water
        102
        100
         98
   S
   E
    s
   >
-•—Finished Water
 A Basin A
 • Basin B
 » Basin C
Fouling/Aging Factor

       A fouling/aging factor of 0.6 was selected for Albany's UV equipment to be
conservative. Because of the low hardness, iron, and manganese concentrations, fouling was not
considered a significant issue. Nevertheless, the selected fouling/aging factor does reduce the
necessary frequency of lamp replacement.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                    F-5
November 2006

-------
                                  Appendix F. Case Studies
       Figure F.4. Cumulative DVT Data for Feura Bush WTP Finished Water
                                                          Design UVT
                                                          (Conservative value chosen
                                                          because few reservoir data were
                                                          available)
                          20%
                                      40%          60%
                                    Cumulative Frequency Percentile
                                                                          100%
       1 Note: 190 samples collected between January 2001 and November 2001.
Flow Rate

       The UV facility was sized for 40 mgd (10 mgd per unit) for emergency or backup
conditions when the reservoir and UV facility must be able to satisfy the entire system demand
(30 mgd maximum). Under normal operating conditions, the UV facility maximum flow is
10 mgd.

Power Quality

       The electric service provider for the proposed location of the UV facility was contacted
regarding the availability and quality of power at the site.  It was determined that high quality
power was available, and power conditioning equipment was therefore unnecessary.
F.1.1.4   Equipment and Monitoring Strategy

       Both MP and LPHO UV equipment were considered during the planning phase. MP UV
equipment was selected for the design because of the smaller footprint and because, at that time,
there was more experience with MP equipment in the United States. Another benefit of MP
equipment was the use of the calculated dose-monitoring strategy, which allowed the City to
address the anticipated variability in flow rate and direction.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
F-6
November 2006

-------
                                   Appendix F. Case Studies
F.1.1.5    UV Equipment Validation Options

       The City of Albany chose on-site validation testing because no UV validation centers
were operating in the United States at that time and because on-site testing would allow for:

           Testing under the exact piping configuration of the UV facility.
       .    Optimizing UV facility operations throughout the life of the facility.

       Space requirements, injection and sample ports, and other elements required for on-site
validation were coordinated with the UV equipment supplier and included in the UV facility
design.
F.1.1.6    Hydraulics

       No modifications to existing hydraulics were required for the UV facility at the reservoir.
Installation of the UV facility did slightly reduce the full operating capacity of the reservoir.
However, during an extended emergency condition, the UV facility can be bypassed to allow use
of the entire reservoir volume.
F.1.1.7   Selected Configuration

       The selected configuration of the UV equipment is summarized in Table F.3. The target
male-specific-2 bacteriophage (MS2) reduction equivalent dose (RED) to be verified during
validation was 40 millijoules per centimeter squared (mJ/cm2) and was chosen to target high-
level inactivation of various pathogens during emergency operation. The MS2 RED was based
on best practices in North America and Europe at the time.
                        Table F.3. UV Equipment Configuration
Criterion
UV Lamp Type
Target MS2 RED
Number of UV Units (Duty + Standby)
Design Flow Rate per Unit
Number of Lamps per Unit
Lamp Power (Each)
Unit
-
mJ/cm2
number
mgd
number
k\A/3)
Value
MP
40(1)
4 + 0(2)
10
8
10
                   40 mJ/cm is the target MS2 RED to be proven during validation testing
                   and to be used when the reservoir is operating in its emergency mode.
                   The target MS2 RED used during the normal distribution function of the
                   reservoir is 240 mJ/cm2 for virus inactivation.
                   40 mgd is needed for emergency and not normal operation; therefore, a
                   standby unit was not provided.
                   kilowatt (kW)
UV Disinfection Guidance Manual
For the Final LT2ESWTR
F-7
November 2006

-------
                                   Appendix F. Case Studies
F.1.2  Design

       Given the wide range of UV equipment available, pre-purchase of the UV equipment by
the owner was selected as the best way to proceed. Pre-purchase documents were issued in
November 2001. Two suppliers bid on the UV equipment, and the contract was awarded to the
low bidder, Trojan Technologies, Inc., in December 2001. By selecting the equipment early in
the project, the project team was able to work closely with the manufacturer during the design of
the system and support facilities (e.g., instrumentation and control) for the 24-inch UVSwift™
units. The following sections describe Albany's UV facility design.
F.1.2.1   Facility Hydraulics

       Four parallel treatment trains of equal capacity (10 mgd) comprise the UV facility
(Figure F.5). Water enters and exits the UV facility via 48-inch diameter influent and effluent
manifolds. Each parallel treatment train consists of a 24-inch diameter lateral, influent and
effluent modulating isolation valves, strap-on ultrasonic flow meter, and an MP UV unit. The
UV units are installed in vertical piping to minimize the footprint of the UV facility and to
promote settling of debris, if any, in the inlet piping to protect the lamps.
          Figure F.5. UV Disinfection Facility at the Loudonville Reservoir
       Although water hammer and surge conditions were determined to be minimal, a
combination air/vacuum release valve was incorporated into each UV treatment train. The valves
provide protection from adverse pressure conditions and facilitate the release of trapped air
during start-up.

       The UV facility was designed to handle large flow variations. The facility can treat
typical daily flows with one or two 10-mgd units in service. With all four units in service, the
UV Disinfection Guidance Manual
For the Final LT2ESWTR
F-8
November 2006

-------
                                  Appendix F. Case Studies
facility can also treat the City's full system demand during WTP or transmission main shut-
downs or during an emergency condition.

       The UV facility was also designed to handle bi-directional flow through the UV
equipment because water passes through the UV units during both the reservoir fill and the draw
cycles. The bi-directional flow design (described below) enables the City to maintain the UV
facility in a constant state of readiness to deliver disinfected water to the distribution system
whenever the reservoir switches to a draw cycle (i.e., when treated water is sent to customers).

Operation in Fill Mode

       When WTP production exceeds distribution system demand, the reservoir fills (fill
mode). All flow passes through the UV facility to the reservoir, and the UV equipment operates
at minimum intensity because disinfection of the influent water is not needed. The primary
objective of operating the UV facility during the fill mode is to ensure that the UV equipment is
on and ready to provide adequate disinfection when the reservoir switches to draw mode.

Operation in Draw Mode

       When distribution system demand is greater than WTP production, water drains from the
reservoir to the distribution system (draw mode). Because a UV unit is always on, there is no
time delay for disinfection  of outgoing water when the flow direction changes.
F.1.2.2   Operational Strategy, Instrumentation and Control

       The UV equipment was designed so that at least one UV train is in service at all times to
ensure that a UV unit is ready to disinfect the reservoir water whenever flow into the distribution
system occurs. When system demand matches the WTP production rate, however, very little
flow into or out of the reservoir occurs. To prevent high lamp temperature (and automatic shut-
down), a cooling water bypass line was installed downstream of the UV equipment and upstream
of each isolation valve to allow a nominal flow through the unit [approximately 80 gallons per
minute (gpm)]. The cooling water line is equipped with a motor-actuated valve for automatic
opening when the water temperature exceeds a set value [90 degrees Fahrenheit (°F)] or when a
start-up or shut-down signal is received. During start-up of a UV unit, the cooling water flow is
discharged to waste. Following start-up, all flow enters the distribution system.

       The UV equipment is controlled by a central programmable logic controller (PLC) using
the calculated dose-monitoring strategy. The central PLC uses flow rate and direction data from
each of the four treatment trains to control the overall operation of the UV equipment and to
sequence the operation of individual UV units. Input for controlling the UV equipment is
provided by a strap-on flow meter on each UV treatment train, two on-line UVT analyzers in the
piping header, and eight UV sensors in each  UV unit. The individual control panel for  each UV
unit adjusts the lamp power and calculated dose of each UV unit in response to the flow rate,
UVT, and UV intensity, to ensure an appropriate level of disinfection.

       As the distribution system demand increases, the central PLC initiates start-up of the next
UV unit once the flow rate through the first unit reaches a manually entered percentage of its
rated capacity. After the second unit has warmed up (approximately 5 minutes), the central PLC
UV Disinfection Guidance Manual                 F-9                              November 2006
For the Final LT2ESWTR

-------
                                  Appendix F. Case Studies
opens the modulating valve for that train and brings the unit on-line. Flow is then split between
the two active UV trains. This scenario continues with the other two UV units as necessary based
upon distribution system demand.
F.1.2.3   Electrical Power Configuration

       Power quality was not expected to be an issue at the reservoir. Therefore, power
conditioning equipment for the UV equipment was not necessary. An uninterruptible power
supply (UPS) was included for the UV control panel to convey alarms and other critical UV
facility information. A backup diesel generator, capable of providing backup power for all
elements of the UV facility, and an automatic transfer switch were also included in the design.
Because the UV equipment is not on a UPS, a brief interruption of the UV disinfection occurs
when the UV facility switches to the backup generator. Disinfection is reinitiated once backup
power is active and the UV lamps have restarted. To minimize the number of power transfers
and resulting UV lamp power interruptions, retransfer of the facility back to grid power is done
manually to allow an operator to determine when conditions are appropriate for a transfer back to
grid power (e.g., when the reservoir is inflowing).
F.1.2.4   Capital Costs

       Bids were received upon completion of the final design, and the construction contract
was awarded. The approximate cost of the UV facility at the time of construction was
$3,125,000, which, when adjusted to 2006 dollars (ENR BCI = 4356), equates to approximately
$3,805,000. Major cost components (in 2006 dollars) included:

       .   $680,000 for the UV equipment
          $2,410,000 for a new UV building and yard piping
          $360,000 for ancillary piping, valves, and controls
       .   $355,000 for electrical.

F.1.3  Validation

       On-site validation of the UV equipment was completed in October 2003, which was prior
to the promulgation of the LT2ESWTR. The validation was based on a previous draft of this
guidance document. Because the guidance has changed, any new installations should follow the
validation protocol described previously in this manual and not follow the example given in this
section.

       Two validation tests were  performed at this facility. Contract validation testing was
conducted to confirm that the equipment met the design criteria specified in the UV equipment
procurement document. Expanded validation testing was also conducted to assess whether
energy efficiency could be improved by modifying the lamp operating strategies and UV facility
maintenance.  The expanded validation testing was co-funded by the City of Albany and the New
York State Energy Research and Development Authority (NYSERDA).

UV Disinfection Guidance Manual                F-10                              November 2006
For the Final LT2ESWTR

-------
                                  Appendix F. Case Studies
       Trojan Technologies, Inc. conducted the validation testing with Dr. James Malley, Jr.
(University of New Hampshire, Durham) providing third-party oversight for the contract
validation testing. Christine Cotton of Malcolm Pirnie, Inc., provided third-party oversight for
the NYSERDA validation testing.
F.1.3.1   Contract Validation Conditions

       The validation testing procedures were based on the UV Disinfection Guidance Manual
(UVDGM) Proposal Draft (USEPA 2003). The validation testing procedures in this manual and
the protocol followed for Albany's on-site validation testing did not significantly differ.
However, the data analysis and Validation Factor calculations do differ from those in this
manual.

       The validation testing was conducted at the reservoir. The challenge microorganism used
in the testing was MS2 phage. Dissolved instant coffee, a UV absorber, was used to adjust the
UV transmittance to the desired test conditions. There was no residual chlorine in the water;
therefore, the water did not need to be dechlorinated.

       Following is a summary of the range of validation  conditions:

       .   Lamps were operated at 0 percent (off), 60 percent (to account for lamp aging and
          fouling effects), or 100 percent (full power) of their nominal output during the
          validation testing.
       .   Flow rates ranged from 2.0 - 10.3 mgd.
       .   UVT ranged from 88 - 99 percent.
       .   Target MS2 RED values ranged from 0 to 150  mJ/cm2.

       The test conditions and results for the validation tests are summarized in  Table F.4. The
UV equipment passed the contract validation testing.


F.1.3.2   NYSERDA Validation Conditions

       The validation testing procedures used in the NYSERDA validation testing were the
same as in the contract validation testing. The NYSERDA testing conditions follow:

       .   Four or  six (of eight possible) lamps in the unit were energized.
       .   Lamps were operated at 0 percent, 60 percent,  80 percent, or 100 percent of their
          nominal output.
       .   Flow rates ranged from 2.0 - 10.0 mgd.
       .   UVT ranged from 87-100 percent.
       .   Target MS2 RED values ranged from 0-150 mJ/cm2.
UV Disinfection Guidance Manual                F-ll                              November 2006
For the Final LT2ESWTR

-------
                                   Appendix F. Case Studies
    Table F.4. Validation Testing Conditions and Results for Contract Validation
Run
No.
1
2
3
4
5
6
7
8
9
10
11
Flow
(mgd)
9.8
9.7
9.9
9.8
9.7
9.9
5.0
4.9
2.0
2.0
10.3
Configur-
ation
8 lamps on
8 lamps on
8 lamps on
8 lamps on
8 lamps on
8 lamps on
8 lamps on
8 lamps on
8 lamps on
8 lamps on
8 lamps on
UVT
Modifier
None
None
None
None
Coffee
Coffee
None
Coffee
None
Coffee
None
Test
Organism
None
MS2
MS2
MS2
MS2
MS2
MS2
MS2
MS2
MS2
None
Lamp
Power
(%)
0
0
100
60
60
100
60
60
60
60
0
UVT
(%)
98.5
98.3
98.5
98.6
87.4
87.5
98.5
87.8
98.1
88.0
98.6
Influent MS2(1)
(log
PFU/mL(2))
0
6.16
6.14
6.08
6.19
6.21
5.78
5.66
5.83
5.79
0
Effluent MS21
(log PFU/mL)
3.15
6.25
0
0
4.37
3.28
0
3.10
0.37
1.11
2.99
Log
Reduction
0.00
-0.09
6.14
6.08
1.82
2.93
5.78
2.56
5.44
4.68
0.00
MS2
RED
(mJ/cm2)
0
-9.4
150.3
148.7
39.4
68.0
141.0
58.4
132.5
112.7
0
  The value shown represents the average of three replicate samples.
 ! plaque forming units per milliliter
       The collimated beam dose-response results for an example day of testing (Day 1) are
shown in Figure F.6.
                Figure F.6. Collimated Beam UV Dose-response Curve
      6.0 r
      5.0
      4.0
      -1.0
No UVT Adjustment:
y = 0.040x + 0.308
                                                    UVT Adjusted with Coffee:
                                                       y = 0.038x + 0.250
                                                         R2 = 0.987
                                                            + No UVT Adjustment
                                                            • UVT Adjusted with Coffee
                               40
                                          60
                                                     80
                                                               100
                                                                          120
                                                                                     140
                                         UV Dose (mJ/cm2)
UV Disinfection Guidance Manual
For the Final LT2ESWTR
             F-12
November 2006

-------
                                  Appendix F. Case Studies
       The test conditions and results for the validation tests are summarized in Table F.5. The
implications of the NYSERDA validation testing results on operation and maintenance (O&M)
are described in Section F. 1.3.4.   Additional validation data analysis may be completed in the
future to determine the validation factor and validated dose in accordance with the validation
data analyses described in this manual (Chapter 5).
   Table F.5. Validation Testing Conditions and Results for NYSERDA Validation
Run
No.
1
2
3
4
5
6
7
8
9
Flow
(mgd)
10.0
9.9
9.9
9.9
2.0
2.0
10.1
10.0
10.0
Configur-
ation
6 lamps on
6 lamps on
6 lamps on
6 lamps on
6 lamps on
4 lamps on
4 lamps on
4 lamps on
4 lamps on
UVT
Modifier
None
None
None
Coffee
Coffee
Coffee
None
Coffee
None
Test
Organism
None
MS2
MS2
MS2
MS2
MS2
MS2
None
None
Lamp
Power
(%)
0
0
60
80
60
60
80
100
0
UVT
(%)
99.0
99.2
99.1
87.5
87.2
87.4
99.0
88.7
99.9
Influent MS21
(log PFU/mL)
0
6.25
5.85
5.98
6.55
6.34
5.83
0
0
Effluent MS21
(log PFU/mL)
0
6.18
0.30
3.93
2.40
3.89
0.37
0.602
2.752
Log
Reduction
0.00
0.06
5.59
2.04
4.12
2.45
5.83
0.00
0.00
MS2
RED
(mJ/cm2)
0.00
-4.7
140.8
47.4
102.1
58.1
147.1
0.00
0.00
  The value shown represents the
2 Test microorganisms injected in
  back into the main flow stream.
average of three replicate samples.
prior tests had pooled in a deadspace upstream of the UV reactor and
     later bled
F.1.3.3   Issues Encountered During Validation Testing

       During validation testing, the following issues were encountered:

       •  Ultrasonic flow meter uncertainty. Before the start of the contract validation
          testing, a discrepancy between a flow meter installed on the UV unit to be tested and
          an existing flow meter farther downstream was noted. The meter manufacturer was
          contacted, and a representative was sent to the site. A portable strap-on flow meter
          was installed next to the test unit flow meter; the portable meter was consistent with
          the test unit meter. The downstream flow meter was determined to be in error, having
          been set to the diameter of the casing pipe and not the diameter of the internal carrier
          pipe. Upon resolution of the investigation, the validation testing continued using the
          test unit flow meter to measure the flow rate.

       •  Test organism not injected. In the test plan, Run No. 8 of the NYSERDA validation
          testing should have had organisms injected. However, during the testing, no
          organisms were injected. Instead of re-doing the testing, Run No. 8 was used as an
          additional control.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                F-13
November 2006

-------
                                  Appendix F. Case Studies
F.1.3.4   Validation Implications for Operation and Maintenance

       The NYSERDA validation testing results indicated that the UV equipment can achieve
(and exceed) the 40 ml/cm2 target MS2 RED when operating in a power saving mode with only
4 or 6 (of a possible 8) lamps on and with UVT between 87 and 99 percent. When the data were
analyzed in accordance with the UVDGM Proposal Draft (USEPA 2003) guidelines that were
available at the time, the testing indicated that 3-log Cryptosporidium inactivation and 1.5-log
virus inactivation, if desired, could be achieved under expanded lamp control conditions.
Validation of the alternative lamp operating configuration is expected to result in cost savings
from reduced power usage.
F.1.4  Start-up and Operation of the UV Facility

       Construction of the facility was completed in February 2003. Full-scale operation began
in March 2003.
F.1.4.1   Start-up and Construction Issues

       Although some problems occurred upon initial start-up of the UV equipment, all parties
involved worked to resolve the issues to the City's satisfaction. The problems and resolutions are
briefly discussed below:

       •  Control system. The manufacturer's control system does not calculate the UV dose
          correctly during the "fill mode" conditions. Because flow is bi-directional in the
          inlet/outlet pipe, a negative value for flow was used in the fill condition (a positive
          value for the "draw mode"). It could be corrected by changing the programming to
          use an absolute value of the flow input in the calculation.

       •  UVT analyzers. The on-line UVT analyzers initially reported inconsistent readings.
          Samples were taken at the midpoint and top of the pipe. The samples taken at the top
          of the pipe were found to be occasionally erroneous due to air bubbles in the sample.
          To correct the problem, the sample ports for the on-line UVT analyzers were adjusted
          so that all samples were taken from the midpoint of the pipe.

       •  Cleaning system. The wiper cleaning mechanism originally provided with the UV
          equipment frequently jammed due to grit entering and binding the threads of a wiper
          system rod. The UV manufacturer refined the design and provided a replacement
          wiper drive system with a self-cleaning traversing nut and a rod with a larger thread
          pitch and depth.

       •  Intensity sensors. In several instances, the coating on the intensity sensors degraded.
          The UV manufacturer improved the design and provided new UV sensors for the UV
          reactors. The new sensors were provided prior to validation testing, so re-validation
          was not necessary.
UV Disinfection Guidance Manual                F-14                              November 2006
For the Final LT2ESWTR

-------
                                  Appendix F. Case Studies
          UV lamp failure. At start-up, approximately thirty percent of the lamps failed to
          energize. A similar percentage also failed to start in a second shipment of lamps. The
          UV manufacturer tracked the problem to a batch of lamps with manufacturing
          defects. The manufacturer corrected the problem and installed a new batch of lamps.
F.1.4.2   Operation and Maintenance

       The following operational tasks are regularly performed at Albany's UV facility. These
tasks take approximately one hour per day, seven days per week.

       •   Daily overall visual inspection of the UV equipment.

       •   Daily check of the control system to ensure it is in automatic mode.

       •   Daily check of the control panel display for status of UV equipment components and
          alarms.

       •   Daily check of on-line analyzers, flow meters, and data recording equipment.

       •   Daily review of 24-hour monitoring data to ensure that the equipment has been
          operating properly.

       •   Daily check of cleaning mechanism operation.

       •   Daily check of lamp run time.

       •   Daily check of ballast cooling fans for unusual noise.

       •   Weekly check of valve operation.
       The City of Albany also performs regular maintenance tasks at the UV facility. Due to
budget cut-backs, the original maintenance frequencies for several tasks have been reduced. The
current scheduled maintenance tasks include the following:

       •   Monthly calibration check of UV sensors.

       •   As-needed calibration check of UVT analyzers. Due to the sensitivity of these
          analyzers and re-calibration difficulties described in Section F.I.4.3, the calibration of
          these analyzers is checked only when problems are evident (every few months). The
          calibration was formerly checked weekly.

       •   Quarterly to annual  check of equipment housing, sleeves, and wiper seals for leaks.

       •   As-needed replacement of the duty sensor with a calibrated back-up sensor. To date,
          replacement has been unnecessary.
UV Disinfection Guidance Manual                F-15                              November 2006
For the Final LT2ESWTR

-------
                                  Appendix F. Case Studies
       •   Annual check of the cleaning system efficiency by inspecting and manually cleaning
          the sleeves. This maintenance was previously performed quarterly.

       •   Quarterly check of the cleaning fluid reservoir.

       •   Annual calibration of the reference sensor by the manufacturer.

       •   As-needed replacement of lamps that have broken or are at the end of their lamp life
          (currently replaced after 4,000 hours). Approximately 20 lamps have been replaced in
          the first 2 years of operation.

       •   As-needed replacement of sleeves that have broken or fouled. To date, approximately
          3 sleeves have been broken and replaced.

       •   As-needed cleaning of UVT analyzers.

       •   As-needed inspection of cleaning system drive mechanism.

       •   As-needed inspection of ballast cooling fan.

       Performance of these  tasks is currently estimated to take approximately two hours per
week per unit (8 hours per week total). An additional 8 hours per month is spent on
troubleshooting. Before the cut-backs, performance of these tasks at the recommended frequency
(Sections 6.3.1 and 6.4.1.4) took an estimated eight hours per week per unit (32 hours per week
total).


F.1.4.3   Operational Challenges

       Although the City generally has found the facility to be relatively simple to operate, a few
challenging conditions have been encountered:

       •   UVT analyzers. Due to the very high UVT of the water (typically 95 percent), the
          City's maintenance personnel have difficulty calibrating the UVT analyzers.

       •   Wiping collar maintenance. The City's maintenance personnel have had problems
          re-aligning the cleaning system's wiper collars on the sleeves when the collars have
          been completely removed from the equipment for maintenance. Difficulties with this
          maintenance task  have resulted in several broken sleeves.

       •   UV equipment draining. Inadequately sized drains in the UV unit delay
          maintenance because of the excessive time needed to drain water.
F.1.5  Future UV Facility Plans

       Because very few data were available on the water quality at the reservoir, a conservative
UVT of 88 percent was selected for the design of the UV equipment. However, full-scale
operating data indicate that the UVT of the water at the reservoir is actually much greater
UV Disinfection Guidance Manual                F-16                              November 2006
For the Final LT2ESWTR

-------
                                   Appendix F. Case Studies
(approximately 95 percent), which enables the City of Albany to target virus inactivation with its
UV facility over a larger range of flow rates when the reservoir is functioning as distribution
storage. Albany's validation data indicate that the facility can achieve 1.5-log virus inactivation
and could likely be validated for greater virus inactivation if an appropriate challenge organism
can be identified. Therefore, credit for greater than 1.5-log virus inactivation may be sought in
the future.
F.2    Weber Basin Water Conservancy District - LPHO Facility with Off-site
       Validation

       The Weber Basin Water Conservancy District (District) was established in 1950 to
provide for the conservation and development of the water resources within the District
boundaries and to use these resources to the greatest benefit of the public. The District is
currently a drinking water wholesaler, serving a total population of approximately 400,000
people.

       The District's Weber WTP No. 3 is a 46-mgd conventional WTP with settled water
ozonation for taste and odor control and UV light for enhanced disinfection (Figure F.7). The
plant was expanded to its present capacity and other improvements were made in 2001.
UV Disinfection Guidance Manual                F-17                              November 2006
For the Final LT2ESWTR

-------
                                  Appendix F. Case Studies
                Figure F.7. Weber WTP No. 3 Process Flow Diagram
TI
o
.c
1 1
o >-

 Structure/ -
Flash Mix










- > UV »
" Reactors "






•
W X
1
Chlorine Finished Water
Contact Basin ~* Storage
Reservoir



-------
                                 Appendix F. Case Studies
              Table F.6. Summary of Raw Water Quality (1996 - 1998)
Parameter
Turbidity
PH
Alkalinity
Temperature
Total Hardness
Calcium Hardness
Iron
Manganese
Total Organic Carbon
Units
NTU
-
mg/L as CaCO3
°C
mg/L as CaCO3
mg/L as CaCO3
mg/L
mg/L
mg/L
Average
29.2
7.5
177
10.4
215
N/A
N/A
N/A
3.2(1)
Minimum
0.3
6.6
58
2.0
142
N/A
N/A
N/A
1.1
Maximum
3,800
8.5
268
19.6
264
N/A
N/A
N/A
7.6
        Data collected at Weber WTP No. 3.
            Table F.7. Summary of Filtered Water Quality (2002 - 2004)
Parameter
Turbidity
PH
Alkalinity
Temperature
Total Hardness
Calcium Hardness
Iron
Manganese
Total Organic Carbon
Units
NTU
-
mg/L-CaCO3
°C
mg/L-CaCO3
mg/L-CaCO3
mg/L
mg/L
mg/L
Average
<0.15
7.40
180
10.7
225
N/A
<0.02
Below detection limit
<3
F.2.1  Planning

       This section discusses key planning decisions made for Weber Basin's UV facility.
Figure F.8 is a timeline of the process the District used to implement UV disinfection.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
F-19
November 2006

-------
                                  Appendix F. Case Studies
             Figure F.8. Weber WTP No. 3 UV Implementation Timeline

Monitoring
Planning &
Design
Construction,
Validation &
Start-Up
1996

1997

1998

uv
TransmHance Cryptoa/xiridiiim
Monitoring Monitoring









1999



2000


2001

/- Bid
/ con
Planning A
& Design T


Val

2002

period. Cons
tract awarded
dation testing
rtland),July2
2003

truction
April 2001 .
003. — — St
Ju
2004

art-Lip, ~j
le 2004 /
A A7
Construction ^f w




F.2.1.1   UV Disinfection Goals

       Raw water monitoring by the District indicated that the water system could be classified
as a Bin 2 or Bin 3 system under the LT2ESWTR. (See Section 1.3.1.) For the purposes of the
preliminary  design, it was assumed that the water system could either initially or ultimately be
classified as a Bin 3 system, which would require the UV facility to provide 2.0-log additional
Cryptosporidium inactivation.

       As part of the preliminary design process for the UV facility and other plant
improvements, the District reevaluated its Giardia treatment. Before the 2001 improvements, the
Weber WTP No. 3 used free chlorine for disinfection of Giardia and viruses.  Following the 2001
improvements, the finished water reservoirs had sufficient capacity to continue to provide the
required level of Giardia and virus inactivation. However, the use of free chlorine for 3-log
Giardia inactivation in the finished water reservoirs was discontinued for the  following reasons:

       •   The UV reactors could easily be designed for both Giardia and Cryptosporidium
          inactivation.

       •   Incorporating  Giardia disinfection with the proposed Cryptosporidium disinfection
          process would make a significant portion of the existing finished water reservoirs
          available for operational storage to benefit the distribution system.

       Table F.8 summarizes the treatment goals for the Weber WTP No. 3.
F.2.1.2   UV Retrofit Location

       Because of existing hydraulic constraints, the UV reactors could be installed only
downstream of the filters (see Figure F.7).
UV Disinfection Guidance Manual
For the Final LT2ESWTR
F-20
November 2006

-------
                                  Appendix F. Case Studies
                             Table F.8. Disinfection Goals
Process
Filters
UV Light
Chlorine
Total Provided
Cryptosporidium
3.5(1)
2.0
-
5.5
Giardia
2.5
2.0
1.0
5.5
Virus
2.0
-
4.0
6.0
               1 Combined filter effluent turbidity 0.15 NTU in 95 percent of samples each
                month to provide a second barrier.

F.2.1.3   Key Design Parameters

       Water quality, fouling/aging factor, and flow rate are critical parameters to be considered
during the planning phase. Table F.9 summarizes key preliminary design parameters for the UV
reactor design.

                Table F.9. UV Facility Preliminary Design Parameters
Criterion
UV Transmittance
Fouling/Aging Factor
Flow Rate
Unit
percent
percent
mgd
Value
90
67
15.3
Water Quality

       Several water quality parameters affect UV dose delivery and, therefore, UV reactor
design. (See Table F.7.) The most important is UVT, which is calculated from the UV
absorbance at 254 nm (A254) as described in Section 3.4.4.1. Based on the available UV
absorbance data, a design UVT of 90 percent (A254 of 0.046 cm"1) was selected (Figures F.9 and
F.10).

Fouling/Aging Factor

       A fouling/aging factor of 67 percent was selected during planning. The factor was
incorporated into the design to account for the reduction in lamp output at the end of lamp life
and the reduction in lamp output due to irreversible sleeve fouling.

Flow Rate

       The UV facility capacity was designed to match the 46-mgd WTP capacity with three
units in operation and one unit out of service. Therefore, the design flow rate through each
reactor was  15.3 mgd.

Power Quality

       To ensure operation of the UV equipment, standby power was provided with a new
backup generator. No other power conditioning equipment was needed.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
F-21
November 2006

-------
                                 Appendix F. Case Studies
             Figure F.9. UVT Data for Weber WTP No. 3 Finished Water
100 -i
98
Qfi
0)
(0
•" 94
E
if>
a
ro
^ 92
Z>
90
RR
86


/
• * V »«»
~~ " Average ' _ ^T " ^7.^*1 . ». . T~» "
V* • **« V ** *• ** • ^
v * » / •«*»••»*• * » » V
• / *• • *• • ' * • * * * • **
* • »t* ^ * * * * * ^95% Confidence
* »t ** * * * * Interval
... * . £...•.................
99% Confidence *
Interval
^ ^ <$ ^ ^ ^ ^ ^ 4$ $*









      Figure F.10. Cumulative UVT Data for Weber WTP No. 3 Finished Water
           100% -
            98%
            96%
            94%
         £  92%
                    10%    20%
                                30%    40%   50%    60%
                                     Cumulative Frequency Percentile
                                                         70%    80%    90%    100%
UV Disinfection Guidance Manual
For the Final LT2ESWTR
F-22
November 2006

-------
                                  Appendix F. Case Studies
F.2.1.4   Equipment and Monitoring Strategy

       Because both MP and LPHO reactors were considered for the design, the footprint for the
larger LPHO reactors was used for planning. The District selected variable setpoint operation
based on flow rate for the UV equipment to conserve power.
F.2.1.5   UV Reactor Validation Options

       Off-site validation was selected for the following reasons:

       .   Discharge of validation test water from on-site validation testing was not feasible.
          At the time, UV disinfection was an innovative process without references for potable
          water applications in the United States. Therefore, off-site validation prior to delivery
          was deemed a reasonable validation option.

F.2.1.6   Hydraulics

       Changes to the plant's hydraulic profile were made as part of the overall plant expansion
and improvements in 2001. The UV reactors fit into the plant's new hydraulic profile at the
combined filter effluent without need for additional modifications or intermediate pumping.


F.2.1.7   Selected Configuration

       Only UV reactor manufacturers that had validated a reactor were considered for the
design, but pre-validation of the proposed UV reactor was not required. A bid specification was
written before design of the UV facility. The base bid was for LPHO reactors with alternate bids
allowed for MP reactors. Although the UV equipment was not pre-purchased, bids were
evaluated, and the detailed design was based on the selected manufacturer (WEDECO Inc.).  The
selected configuration of the UV reactors is summarized in Table F.10. The target MS2 RED to
be verified in validation was selected to target high-level inactivation of various pathogens based
on best practices in North America and Europe at the time. The selected configuration has one
sensor for every bank of lamps  and has one UVT analyzer.
                        Table F.10. UV Reactor Configuration
Criterion
UV Lamp Type
MS2 RED1
Number of UV Units (Duty + Standby)
Design Flow Rate per Unit
Number of Banks per Unit
Number of Lamps per Bank
Lamp Power (Each)
Unit
-
mJ/cm2
number
mgd
number
number
W
Value
LPHO
40
3 + 1
15.3
6
12
360
              1 The MS2 RED to be proven during validation testing.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
F-23
November 2006

-------
                                  Appendix F. Case Studies
P.2.2  Design

       The following sections describe the UV facility design in more detail.


F.2.2.1   Facility Hydraulics

       The UV reactors were installed on the combined filter effluent to fit into the plant's
hydraulic profile. Filtered water from a common header is passively divided into four 30-inch
influent pipes to the UV reactors. To compensate for a possible uneven flow split, upstream
isolation valves can be manually throttled. The head loss through the UV reactors, ancillary
piping, and valves is 2.3 feet of water at maximum plant capacity (46 mgd). Effluent weirs are
installed to ensure that the reactors remain submerged.


F.2.2.2   Operational Strategy, Instrumentation and Control

       The UV supplier provided a variable setpoint control strategy based on a UV Intensity
Setpoint Approach. In this approach, the minimum UV intensity values determined during
validation can be set for several flow rate ranges. The UV ballast system has 50 to 100 percent
variable power capabilities, allowing the UV reactor to automatically adjust based on relative
sensor intensity and flow rate to conserve power (see section F.2.3.1 for details). For the UV
reactor to stay in compliance (i.e., to ensure minimum UV dose delivery), the UVT must remain
at or above the minimum value (90 percent), the flow rate through the reactor must be less than
or equal to the maximum validated flow rate, and the UV sensor values must all be at or above
the UV intensity setpoint for that flow rate and number of lamp banks on as determined by the
validation data (See Section F.2.3.).

       Flow meters and flow control valves were not provided for each reactor. However, each
reactor was provided with UV sensors. Additionally, a motorized valve downstream of each
reactor remains closed during start-up  until the reactor is on and warmed up. Therefore, off-
specification water is not delivered to consumers and does not need to be wasted during reactor
start-up.


F.2.2.3   Electrical Power Configuration

       An electrical engineer reviewed power fluctuations and quality at the WTP and
determined that power conditioning equipment was not needed. However, standby power was
provided with a new backup generator to ensure continuous UV equipment operation.

F.2.2.4   Capital Costs

       The UV facility construction was part of a larger expansion and improvement project.
The portion of the construction cost attributed to the UV facility was $2,230,000 in 2006 (ENR
BCI = 4356) dollars. Major cost components included:

       .   $1,210,000 for the UV equipment.

UV Disinfection Guidance Manual                F-24                              November 2006
For the Final LT2ESWTR

-------
                                  Appendix F. Case Studies
       .   $400,000 for a new UV building.
       .   $250,000 for ancillary piping, valves, and controls.
       .   $370,000 for electrical improvements.

F.2.3  Validation

       Off-site validation testing was originally conducted at the Deutsche Vereinigung des Gas-
und Wasserfaches (DVGW) facility in Germany. However, validation testing at the DVGW
facility proved problematic because the strict DVGW requirements do not allow flexibility in
validation or operation setpoints. As such, further validation testing at the Portland Validation
Facility was conducted in July 2003. The second validation was in accordance with the U.S.
guidelines based on the UVDGM Proposal Draft (USEPA 2003).

       The Portland Validation Facility is located at the City of Portland, Oregon Bureau of
Water's Groundwater Pumping Station of the Columbia Southshore Wellfield, Portland, Oregon.
The Columbia Wellfield is a 90-mgd supplemental drinking water supply that the Portland Water
Bureau owns and operates. The wellfield provides up to 43 mgd of continuous flow to the UV
reactor test train. Typical water quality of the groundwater is shown in Table F.I 1
                   Table F.11. Southshore Wellfield Water Quality
                                    Characteristics
Parameter
UVT
Hardness
Alkalinity
PH
Chlorine
Unit
%
mg/L
mg/L CaCO3
unitless
mg/L
Value
96.8 - 98.6 (98.3 average)
38-144
34-169
5.8-8.8
none
       The high UVT allowed testing of the full range of operating UVT conditions, and the
zero chlorine residual eliminated the need to quench the chlorine prior to adding chlorine-
sensitive challenge microorganisms.

       Carollo Engineers conducted the validation testing. Clancy Environmental Consultants
(CEC), St. Albans, Vermont, supervised the injection and sampling of the challenge
microorganism. CEC prepared all stock solutions of the challenge microorganism, measured
challenge microorganism UV dose-response using the collimated beam apparatus, and assayed
challenge microorganism concentrations. WEDECO Inc. (Charlotte, North Carolina) operated
the UV reactor during biodosimetry testing with oversight by Carollo.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
F-25
November 2006

-------
                                  Appendix F. Case Studies
F.2.3.1   Validation Conditions

       The Portland Validation Facility allowed testing conditions (e.g., piping configuration) to
be defined for each validation test, which allowed the testing to be optimized to US guidelines
based on the UVDGM Proposal Draft (USEPA 2003). The Weber WTP No. 3 reactor was
validated with inlet piping that included a 90-degree bend located three pipe diameters upstream
of the reactor and another 90-degree bend less than three pipe diameters downstream of the
reactor. This configuration did not represent the actual piping arrangement at the Weber WTP
No. 3 but, instead, represented the "worst case" flow conditions through the UV reactor.

       The high UVT of the Columbia Wellfield water allowed the full range of UVT conditions
to be tested. Lignin sulfonic acid (LSA), a UV absorber, was used to reduce the UVT as needed.
The challenge microorganism used in the validation testing was MS2 phage. Static mixers were
used to ensure that additives were well mixed upstream of the reactor inlet sampling port and the
reactor exit sampling port. The testing configuration is shown in Figure F. 11.
                    Figure F.11. Validation Testing Configuration
                  2 Million
                   Gallon
                  Reservoir
       100 gpm
       Injection
           Lo°P
                            Backflow
                            Prevention
                                          ; Pump
UV Absorber Addition
Challenge Microbe Addition
                               UV    Static
                             Reactor  Mixer Valve
           90 mgd
            Water
            Supply
,,.
Valve
           Flow
           Meter
Static
Mixer
     Inlet
    Piping
                                              Waste
                                               .     „
                                               Influent
                                               Sample
                                                Port
                                                               Effluent
                                                               Sample
       The UV reactor was tested using a range of flow rate, UVT, and operating lamp
combinations to validate UV dose delivery and UV sensor measurements. The experimental
matrix was designed to validate the vendor's UV intensity setpoint approach with variable
setpoint operation for a range of water quality conditions within the defined design criteria. The
experimental matrix also was intended to enable the PWS to optimize performance (i.e., deliver
the target MS2 RED with a minimal number of lamp banks in operation at a minimum power
level).

       Following is a summary of the range of validation conditions:

       •   All lamps were operated at 67 percent of their nominal output during the validation
          testing to account for lamp aging and fouling effects.

       •   Flow rates ranged from 0.94 - 20 mgd.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
                  F-26
                            November 2006

-------
                                 Appendix F. Case Studies
       •   UVT ranged from 85-95 percent.

       •   The number of lamp banks that were on was 1, 2, 4, 5, and 6 (all banks).

       •   Target MS2 RED values ranged from 20-100 mJ/cm2.

       The test conditions and results for several of the 34 validation tests that were conducted
are summarized in Table F.12.
        Table F.12. Excerpt of Test Conditions for Validation Testing at the
                Portland Validation Facility (Total No. of Tests = 34)
Run No.
12
9
3
16
36
22
38
19
29
Flow Rate
(mgd)
2.07
2.36
0.94
17.61
19.57
20.00
14.49
15.65
12.07
No.
Banks On
2
1
1
2
6
4
5
4
6
UVT
(%)
84.7
95.0
84.7
94.8
94.8
89.9
90.1
85.0
85.3
UVT
Modifier
ISA
ISA
ISA
ISA
ISA
ISA
ISA
ISA
ISA
Test
Organism
MS2
MS2
MS2
MS2
MS2
MS2
MS2
MS2
MS2
Lamp Power
(%)
67
67
67
67
67
67
67
67
67
       The collimated beam dose-response results for an example day of testing are shown in
Figure F.12, and Table F.13 summarizes validation testing results.

       UV transmittance measurements were checked using National Institute of Standards and
Technology (NIST)-traceable UV absorbance standards. Flow measurements were checked by
comparison of manufacturer calibration to internal settings. The UV dose-response of the MS2
phage met bounds described by the UVDGM Proposal Draft (2003) and NWRI (2003).
Biodosimetry and sensor testing and data analysis were based on the June 2003 Draft UVDGM
recommendations (Tier 2 analysis).
UV Disinfection Guidance Manual
For the Final LT2ESWTR
F-27
November 2006

-------
                                 Appendix F. Case Studies
               Figure F.12. Collimated Beam UV Dose-response Data
c r\
Log Inactivation
-^ O -^ N> CO Ji. t
D b b b b b c
_L- I I I I
•
. B
i
i
A
£
^
i


• Sample 15
• Sample 3
A Sam pie 36
Sam pie 24


0 20 40 60 80 100 120
UV Dose (mJ/cm2)
                       Table F.13. Validation Testing Results
Run
No.
12
9
3
16
36
22
38
19
29
Flow
Rate
(mgd)
2.07
2.36
0.94
17.61
19.57
20
14.49
15.65
12.07
No.
Banks On
2
1
1
2
6
4
5
4
6
Lamp
Power
(%)
67
67
67
67
67
67
67
67
67
UVT
(%)
84.7
95.0
84.7
94.8
94.8
89.9
90.1
85.0
85.3
Log
Infl. MS2
(pfu/mL)
5.04
5.29
5.38
4.69
5.32
4.69
4.81
4.76
4.67
Log
Effl. MS2
(pfu/mL)
2.16
0.97
2.41
3.23
2.00
3.45
2.73
3.53
2.57
Log I
2.88
4.32
2.97
1.46
3.32
1.24
2.07
1.23
2.09
MS2 RED
(mJ/cm2)
57.9
96.0
66.9
28.4
76.7
23.6
42.9
23.4
41.6
       The measured relationship between UV sensor measurements and UVT at the 80 percent
intensity turn-down reflecting end-of-lamp-life (EOLL) was described using a power function (A
and B are constants):
       UV Sensor (UVT, ROLL] = eAe>
                                    xUVT
                                 Equation F.I
       The functions describing the UV sensor measurements as a function of ballast power
setting and UVT were obtained using new lamps, sleeves, and UV sensors in a clean UV reactor.
The functions can be compared to measurements made at a WTP to assess the relative output of
the lamps compared to the data measured during validation.

       The test results were evaluated by plotting measured MS2 RED (mJ/cm2) as a function of
number of operating banks divided by flow rate (Q in mgd), again based on a specific UVT value
UV Disinfection Guidance Manual
For the Final LT2ESWTR
F-28
November 2006

-------
                                 Appendix F. Case Studies
and at the intensity turn-down reflecting end-of-lamp-life. Statistical analysis was used to
determine if data sets obtained with different rows could be combined. The relationship was
fitted by a polynomial function (A and B are constants):
       RED (UVT, EOLL) =

                                           0
                                 Equation F.2
       Based on these equations, an automatic UV dose-monitoring strategy was developed that
determined the necessary number of banks/rows in operation and the ballast power, so that a
selected target MS2 RED (e.g., 40 ml/cm2) is met.
F.2.3.2   Validation Implications for O&M

       The validation testing data were used to develop equations that would automatically
determine the needed number of banks in operation and ballast power so that the selected target
MS2 RED (e.g., 40 ml/cm2) could be provided.
F.2.4  Start-Up and Operation of the UV Facility

       Full-scale operation of the facility began in June 2004. Photos of the UV equipment are
shown in Figures F.I3 and F.I4.
                 Figure F.13. UV Reactors at the Weber WTP No. 3
                       (3 duty + 1 standby reactor in parallel)
UV Disinfection Guidance Manual
For the Final LT2ESWTR
F-29
November 2006

-------
                                  Appendix F. Case Studies
        Figure F.14. UV Reactor Electrical Cabinets at the Weber WTP No. 3
                          (the floor above the UV reactors)
F.2.4.1   Start-up and Construction Issues

       Since start-up, the UV facility has been operating as intended (no frequent UV unit shut-
downs, lamp failures, or similar mechanical problems). However, as with any new unit process,
some problems have been experienced, including the following:

       •   UV-monitoring system. Not all of the low-level alarm settings and controls in the
          monitoring system software worked properly in the first version of software provided.
          The vendor subsequently updated the software.

       •   Manganese fouling. Sleeve and sensor fouling was a serious problem when the UV
          equipment was first started. Although dissolved manganese concentrations were not
          measured, the problem began when the District began adding ferric chloride for
          coagulation. Analysis of the foulant indicated that manganese, an impurity present in
          the coagulant, caused the fouling. To control this problem, hypochlorite was fed
          upstream of the filters to oxidize the manganese, which was then removed by the
          filters. The District plans to discontinue the hypochlorite feed once the intermediate
          ozonation system is fully operational.

       •   Cleaning system. The phosphoric acid chemical cleaning system originally intended
          for use with the UV reactors was installed on a cart on the upper level of the UV
          disinfection room. A long hose with a spray-nozzle attachment was to be hand-carried
          to the lower level and inserted into the UV reactors for cleaning. However, the
          cleaning system could not provide enough suction to work properly with the cart
          located on the upper level, and the cart could not be moved to the lower level due to
          mobility constraints. Therefore, a new chemical cleaning system that could provide
          appropriate pumping power was constructed on the lower level adjacent to the UV
          reactor.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
F-30
November 2006

-------
                                  Appendix F. Case Studies
          Control panel. The UV equipment was designed with one transformer for each
          ballast enclosure. During the first summer of operation, the enclosures overheated and
          had to be opened so the transformers could be cooled with box fans. The
          manufacturer recommends that the temperature in the control room not exceed
          100 °F. The temperature in this room was not measured during the first summer of
          operation, so whether the overheating was due to an inadequate or faulty control
          panel cooling system or to high temperatures in the control room is unknown. The
          source of this problem remains under investigation.

          Training. The manufacturer did not provide on-site operator training on UV reactor
          O&M until after the UV equipment had been operating for several months.
F.2.4.2   Operation and Maintenance Requirements

       The UV sensors are calibrated monthly, and no sensor drift has been observed since the
facility has been operational. Similarly, the online UVT analyzer is also checked monthly and no
drift has been observed. No lamps have been replaced since the UV facility began operations,
and no chemical cleaning has been performed. Inspections of some of the lamps, however, reveal
no signs of fouling. Information on the amount of labor required to perform these O&M tasks
was not readily available. No information was readily available on power usage.
F.2.4.3   Operational Challenges

       Except for resolving the start-up and construction issues (Section F.2.4.1), the District has
generally found the facility to be relatively simple to operate.
F.2.5  Future UV Facility Plans

       The District plans to apply to the Utah Department of Environmental Quality for
approval of the UV disinfection system for Cryptosporidium and Giardia credit in the future.

       The District may target different Cryptosporidium and Giardia inactivation levels in the
future to respond to future regulatory requirements. The UV manufacturer has provided curves
showing flow rate versus number of UV lamp banks in operation for three different MS2 RED
(20, 30, and 40 ml/cm2) to enable future operational flexibility.
F.3    Clayton County Water Authority - LPHO Facility with On-site
       Validation

       The Clayton County Water Authority (CCWA) owns and operates three WTPs, which
serve more than 250,000 people in Clayton County, Georgia. The Freeman Road WTP is a 12-
mgd conventional surface WTP. Chlorine dioxide is applied prior to the rapid mix process to
oxidize taste and odor compounds and iron and manganese, and free chlorine is applied to the
filtered water for disinfection. In 2002 the plant was upgraded to include a UV disinfection

UV Disinfection Guidance Manual               F-31                             November 2006
For the Final LT2ESWTR

-------
                                  Appendix F. Case Studies
facility. The filtered water quality characteristics that were the basis for the UV reactor design
are summarized in Table F.14.
                    Table F.14. Summary of Freeman Road WPP
                                 Filtered Water Quality
Parameter
Turbidity
PH
Alkalinity
Iron
Manganese
Total Organic Carbon
Units
NTU
-
mg/L-CaCO3
mg/L
mg/L
mg/L
Design Value
0.18
8.1
29
0.1
0.02
<2
F.3.1  Planning and Design

       This section discusses the key planning and design decisions made for CCWA's UV
facility.


F.3.1.1   UV Disinfection Goals

       The UV equipment was installed at the Freeman Road WTP to provide an additional
pathogen barrier. The basis for the facility design was 2.5-log Cryptosporidium inactivation. UV
disinfection was selected over other disinfection technologies because of its effectiveness against
pathogens and its cost-effectiveness.


F.3.1.2   UV Retrofit Location

       The UV reactors were installed on the combined filter effluent piping in a new stand-
alone building. As part of the UV retrofit, chemical feeds for lime, fluoride, phosphoric acid, and
chlorine were relocated to follow the UV reactors.


F.3.1.3   Key Design Parameters

       The UV reactors for the Freeman  Road WTP were bid and selected before detailed design
of the facility. The bid was open to LPHO and MP reactors. A life cycle cost analysis that
incorporated the capital costs and the anticipated energy and maintenance costs was used to
select the UV reactors. Ultimately, LPHO reactors were selected for the Freeman Road WTP.
After selecting the reactors, one set of plans and specifications was developed for the design. The
UV facility consisted of three WEDECO  Series K reactors. Key design parameters for the UV
reactors are shown in Table F.15. A conservative MS2 RED (50 mJ/cm2) to be verified during
UV Disinfection Guidance Manual
For the Final LT2ESWTR
F-32
November 2006

-------
                                  Appendix F. Case Studies
validation testing was selected because the design was to be completed while the LT2ESWTR
and this manual were still under development.
                     Table F.15. UV Reactor Design Parameters
Criterion
UV Transmittance
Fouling/Aging Factor
Total Flow Rate
Target MS2 RED1
Number of UV Units (Duty + Standby)
Design Flow Rate per Unit
Number of Lamps per Unit
Lamp Power (Each)
Unit
percent
percent
mgd
mJ/cm2
number
mgd
number
W
Value
91
60
12
50
2 + 1
6
30
275
                     1 The MS2 RED to be proven during validation testing.
Facility Hydraulics

       No modifications to the plant hydraulics were required at the Freeman Road WTP
because the available head between the filter effluent control weir and the clearwell was
sufficient for the UV reactors.

Operational Strategy, Instrumentation and Control

       A magnetic flow meter is installed in the piping upstream of each UV reactor to monitor
the flow split between the reactors. Each UV reactor is equipped with three UV sensors, one for
each bank of lamps. Additionally, an on-line UVT analyzer enables the UV reactors to be
operated with either a (1) UV Intensity Setpoint Approach or (2) Calculated Dose Approach.
However, the UV equipment has not yet been validated for operation in calculated dose mode.

Electrical Power Configuration and Power Quality

       A UPS was provided  at the Freeman Road WTP to ensure that  the UV equipment
remains in continuous operation. No power quality or outage issues have been experienced at the
facility.

Capital Cost

       The total capital cost for the UV facility at the Freeman Road WTP was approximately
$2,170,000 in 2006 (ENR BCI = 4356) dollars. The cost includes all elements related to the UV
facility, the building, UV reactors, piping, valves, electrical system, instrumentation and controls,
and other ancillary equipment.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
F-33
November 2006

-------
                                    Appendix F. Case Studies
 F.3.2  Validation

        Validation testing was conducted in February 2003. On-site rather than off-site validation
 was selected to maximize flexibility in selecting the specific operating conditions for testing and
 in allowing potential future testing as EPA requirements are established.

        One of the three reactors was designed to serve as a test reactor for on-site validation.
 The inlet and outlet piping to the test reactor can be isolated, and the outlet piping allows flow to
 be routed to waste. The other two UV reactors and their upstream and downstream piping are
 identical in design  to the test reactor, so the testing was representative of each of the other
 reactors.

        Preliminary testing before validation indicated that nearly complete die-off of the MS2
 phage had occurred in both the influent and effluent samples from the UV reactor. Although the
 chlorine dioxide preoxidation system had been shut down several days before testing, jar test
 results indicated that low levels of chlorate or chlorine dioxide caused the die-off. The jar tests
 also indicated that the effect of the chlorate or chlorine dioxide on the MS2 could be alleviated
 by adding LSA, a compound commonly used during validation to reduce UVT. The problem was
 therefore resolved by spiking the microbial samples with LSA before shipping them to the
 laboratory for analysis.

        The challenge microorganism used in the testing was MS2 phage. A dilute LSA solution
 was used to reduce the UVT as needed in the filtered water and to prevent die-off in the MS2
 samples. The results of the validation testing are shown in Table F.16.
                 Table F.16. Validation Testing Conditions and Results
                            for CCWA's Freeman Road WTP
Run
No.
1F
2F
3F
4F
5F2
Flow
(mgd)
5.41
5.95
6.49
7.29
7.39
Configuration1
3 banks on
3 banks on
3 banks on
3 banks on
3 banks on
UVT
Modifier
LSA
LSA
LSA
LSA
LSA
Test
Organism
MS2
MS2
MS2
MS2
MS2
Lamp
Power
(%)
50
50
50
50
-
UVT
(%)
91.2
90.8
90.7
91.9
-
Influent MS2
(log PFU/mL)
5.35
5.44
5.43
5.48
5.54
Effluent MS2
(log PFU/mL)
2.76
2.96
3.30
3.46
5.57
Log
Reduction
2.59
2.48
2.13
2.02
-0.04
MS2
RED
(mJ/cm2)
57.2
54.1
44.6
41.7
-
1 Each reactor contains 3 banks with 10 lamps per bank.
 Control run
 F.3.3  Start-up and Operation of the UV Facility

        Construction of the UV facility at the Freeman Road WTP was completed in December
 2002, and full-scale operation began in April 2003. At the time of publication, operations and
 maintenance data were not made available for the Freeman Road WTP.
        Although the UV equipment has generally operated well since start-up, an issue requiring
 a minor change did arise in the first year. In October 2003  after several months of operation,
 UV Disinfection Guidance Manual
 For the Final LT2ESWTR
F-34
November 2006

-------
                                  Appendix F. Case Studies
WEDECO replaced the UV lamps in the reactors with a new production model. The new lamps
were tested to ensure that the UV intensity of the replacement lamps (following a 100-hour burn-
in period) was equal to or better than the intensity of the lamps that had been replaced. However,
comparison of the intensity data for the replacement lamps to the data that had been collected
during validation indicated that the intensity of the replacement lamps was less than that of the
previous lamps.

      An investigation of the problem determined that microbubbles in the water passing
through  the UV reactor were causing the measured decrease in UV intensity, not the replacement
lamps. The existing air release valve on each reactor did not sufficiently release entrained air
from the water, particularly in colder months due to  the higher dissolved oxygen concentration in
the water. To alleviate the problem, two additional air release valves were installed on the
influent  header between the filter control weir (the source of the entrained air) and the UV
reactors. The additional air release capability minimized the formation of microbubbles during
periods of low water temperature.
F.4    Newark Water Treatment Plant - MP Reactors on Each Filter Effluent
       Pipe

       The Newark WTP, located in Newark, Ohio, is a 15-mgd surface WTP. The Newark
WTP has an average daily flow rate of approximately 8 mgd and serves a population of more
than 47,500 people. Treatment processes at the Newark WTP include preoxidation with
potassium permanganate and powdered activated carbon for removal of taste- and odor-causing
contaminants, lime softening, sedimentation, recarbonation, rapid sand filtration, and disinfection
with UV light and chlorine.

       The filtered water quality characteristics that were the basis for the UV reactor design are
summarized in Table F. 17.
            Table F.17. Summary of Newark WTP Filtered Water Quality
Parameter
PH
Turbidity
Total Alkalinity
Total Hardness
Calcium Hardness
Iron
Manganese
Temperature
Total Organic Carbon
Units
-
NTU
mg/L-CaCO3
mg/L-CaCO3
mg/L-CaCO3
mg/L
mg/L
°F
mg/L
Average
7.6
0.23
60
120
67.4
0.03
0.02
60
1 -2
Minimum
7.2
0.18
40
90
75
0.01
0.01
33
No data
Maximum
8.2
0.53
100
160
60
0.10
0.03
80
No data
UV Disinfection Guidance Manual
For the Final LT2ESWTR
F-35
November 2006

-------
                                   Appendix F. Case Studies
F.4.1  Planning and Design

       This section discusses the key planning and design decisions made for Newark's UV
facility.


F.4.1.1    UV Disinfection Goals

       UV disinfection was installed at the Newark WTP to provide an additional treatment
barrier against pathogens and to ensure public health protection in the event of high turbidity in
the raw water. The City's water source has historically experienced turbidity spikes following
rainfall events.


F.4.1.2    UV Retrofit Location

       Three locations were considered for the UV reactors. Two locations were on the
combined filter effluent at the plant's chlorine contact basin that is used to achieve chlorine
disinfection requirements. This basin is located prior to the clearwell and finished water pump
station. The third location considered was on each of the ten individual filter effluents (IFE). The
IFE location was selected because both the capital and O&M costs were less than the costs for
the other alternatives. Additionally, the IFE location  provided a high degree of redundancy and
O&M enhancements due to the number of reactors.

       To accommodate the retrofit, filter effluent piping had to be rearranged on four filters to
provide the desired straight piping runs upstream and downstream of the reactor. Also, existing
valves on two filters had to be rotated 90 degrees to provide sufficient clearance to service the
reactors. Figure F. 15 illustrates the UV reactor installation on one of the filter effluent pipes.
                     Figure F.15. UV Reactor at the Newark WTP
UV Disinfection Guidance Manual
For the Final LT2ESWTR
F-36
November 2006

-------
                                  Appendix F. Case Studies
F.4.1.3   Key Design Parameters

       The UV reactor specification allowed MP reactors only because LPHO reactors could not
meet the space constraints of this application. The competitive bid resulted in the selection of
Trojan's 12-inch UVSwift™ reactor because the other bidder could not meet the head loss
requirement. The UV equipment was selected prior to design and purchased as part of the UV
facility construction contract.

       Key design parameters for the UV reactor are shown in Table F. 18. The target MS2 RED
of 40 ml/cm2 to be verified in validation was selected based on best practices in North America
and Europe at the time to inactivate a range of pathogens.
                     Table F.18. UV Reactor Design Parameters
Criterion
UV Transmittance
Fouling/Aging Factor
Total Flow Rate
MS2 RED1
Number of UV Units
Design Flow Rate per Unit
Number of Lamps per Unit
Lamp Power (Each)
Unit
percent
percent
mgd
mJ/cm2
number
mgd
number
kW
Value
85
80
15
37
10
1.5
4
1.26
                        The MS2 RED to be proven during validation testing.
Facility Hydraulics

       The head losses created by the UV reactors and the necessary piping modifications were
less than 6 inches at the maximum flow rate through the reactors. Therefore, no additional
pumping or other hydraulic modifications were required for the addition of the UV reactors.
Each reactor was rated for a maximum flow rate of 1.5 mgd, which corresponded to the
rated maximum filter capacity of 4 gallons per minute per square foot (gpm/sf).

Operational Strategy, Instrumentation, and Control

       The UV equipment operates using the Calculated Dose Approach. The UV reactors
automatically adjust to changing conditions to ensure that the calculated dose does not fall below
the dose setpoint. Each reactor normally operates for 4 days followed by a day  out of operation,
corresponding to the normal filter service times. When UV reactors are returned to operation, an
isolation valve located upstream of the UV reactor is closed, and plant service water flowing at
approximately 20 gpm is used to cool the lamps while the reactor is started up and the lamps
return to full power (approximately  10 to 15 minutes). After passing through the reactor, the
cooling water enters the process train and is sent to the contact time (CT) basin for primary
disinfection and then to the finished water clearwell. Once the lamps  reach full power, the
UV Disinfection Guidance Manual
For the Final LT2ESWTR
F-37
November 2006

-------
                                  Appendix F. Case Studies
upstream isolation valve is opened, and filtered water flows through the UV reactors. The
advantage of using plant service water as cooling water during reactor start-up is that because it
has previously been treated by UV disinfection, the cooling water does not have to be included in
the calculation of off-specification water.

       Each reactor was located downstream of an existing flow control valve and flow meter.
New isolation valves were installed downstream of each reactor. In addition to UV sensors, each
reactor has a level sensor and temperature sensor. The level and temperature sensors protect the
UV reactor from running dry or overheating or both. Each reactor was also provided with an
auxiliary potable water supply connection to maintain a minimum flow rate of 15 gpm to prevent
overheating during reactor start-up and shut-down. Each reactor has a UVT analyzer located on
the filter effluent pipe to provide UVT measurements for dose monitoring.

Electrical Power Configuration and Power Quality

       Power quality at Newark WTP was assessed over a fifteen-month period (March 2003 -
June 2004) as part of an American Water Works Association Research Foundation Project
(Cotton et al. 2005). During this time, 240 power quality events occurred. The frequency and
classification of the power quality events at the WTP for this fifteen-month period are shown in
Table F. 19.
                      Table F.19. Power Quality at Newark WTP
Power Quality Event
Instantaneous Voltage Sag
Momentary Voltage Sag
Temporary Voltage Sag
Instantaneous Swell
Instantaneous Interruption
Momentary Interruption
Temporary Interruption
Sustained Deep Undervoltage
Sustained Interruption
Total Estimated Time (minutes)
Estimated % Off-specification Time
City of Newark
Total
215
6
0
0
0
8
3
1
7
NA
NA
Monthly
Average
14.33
0.4
0
0
0
0.53
0.2
0.07
0.47
168
0.38
Maximum
Month
75
2
0
0
0
4
1
1
4
775
1.74
       Approximately 90 percent of the power quality events were instantaneous voltage sags
(i.e., voltage sags lasting between 0.5 and 30 cycles). The UV reactors were not operational until
May 2004, so the off-specification time shown in Table 4.19 is an estimate calculated by
assuming  10 minutes of off-specification time for each voltage sag lasting more than 2 cycles.
Although  a UPS system was not installed at the Newark WTP, the UV equipment's ballast and
electrical design prevents the UV reactors from losing power in many cases. As a result, the
WTP is not having trouble meeting the off-specification requirements proposed in this guidance
manual.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
F-38
November 2006

-------
                                  Appendix F. Case Studies
Capital Cost

       The capital cost for adding the UV equipment to the Newark WTP was $1,135,000 (2006
dollars - ENR BCI = 4356). The cost includes modifications to the existing building and piping,
UV reactors, piping, valves, instrumentation and controls, and other ancillary equipment.
F.4.2  Validation

       No validation testing had been performed at the time of publication because no
disinfection credit was needed. Newark may choose to validate the reactors in the future.

F.4.3  Start-up and Operation of the UV Facility

       Construction of the UV facility was substantially complete and full-scale operation began
in May 2004. The project was completed in July 2004. Overall, the UV equipment has operated
smoothly, and only minor issues were encountered during start-up.

       Minor issues with the control panels and wiring were resolved by the factory
representative and the contractor. Additionally, the automatic backwash sequence programming
had to be rewritten to accommodate the UV reactor cooling water. During the reprogramming,
problems with existing valve actuators were uncovered that required some actuator limit
switches to be adjusted.

       Operations and maintenance costs were not readily available at the time of publication.
F.5    City of Winnipeg Water Treatment Plant - MP Facility with On-site
       Validation

       The City of Winnipeg's water supply is obtained from a surface water source and is
currently unfiltered. Water is chlorinated, and fluoride and phosphate are also added before it is
distributed to the 630,000 people served by the water system.

       A new WTP is currently under construction for the City of Winnipeg, which will use the
following processes: rapid mix, coagulation, flocculation, dissolved air flotation, ozone,
biological activated carbon filtration, UV disinfection, and chloramination. The UV facility was
constructed before the rest of the treatment plant (scheduled for completion in 2007) to minimize
the risk posed by Cryptosporidium and other waterborne pathogens. The UV facility will be
integrated within the new WTP when it is constructed.

       The raw water quality characteristics that were the basis for the UV reactor design are
summarized in Table F.22.
UV Disinfection Guidance Manual                F-39                              November 2006
For the Final LT2ESWTR

-------
                                  Appendix F. Case Studies
              Table P.22. Summary of Raw Water Quality (1989 - 1994)
Parameter
PH
Turbidity
Total Organic Carbon
Dissolved Organic Carbon
Plankton
Total Alkalinity
Total Hardness
Color (true)
Units
-
NTU
mg/L
mg/L
cells/ml
mg/L-CaCO3
mg/L-CaCO3
TCU1
Average
8.2
1.0
9.3
8.3
39,700
81
83
<5
Minimum
7.4
0.3
5
4
200
72
68
<5
Maximum
9.1
5.3
17
15
666,000
95
97
10
         1 True color units
F.5.1  Planning and Design

       This section discusses the key planning and design decisions made for the City of
Winnipeg's UV facility.
F.5.1.1   UV Disinfection Goals

       The UV reactors were designed to provide 2-log Cryptosporidium inactivation in the
Deacon Reservoir raw water. The goal will remain unchanged when the UV facility is later used
to treat filtered water, even though the facility will be able to treat higher flow rates (at higher
UVT).
F.5.1.2   UV Retrofit Location

       The UV facility currently treats unfiltered raw water. In 2007 when the WTP is expected
to be complete, the UV facility will be located downstream of the combined filter effluent. The
equipment was installed in an existing pump station building on the site.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
F-40
November 2006

-------
                                   Appendix F. Case Studies
F.5.1.3   UV Reactor Selection

       Because of space limitations in the existing building, MP reactors were selected. The MP
reactors were selected in a competitive pre-selection/proposal process prior to completion of the
final design. A cost/benefit model was used to evaluate the two pre-selected UV equipment
alternatives (a typical model summary is shown in Figure F. 16). Benefit scores (example values
shown in stacked bars in Figure F.16) for non-monetary evaluation criteria were developed in
advance of bids for each alternative by assigning relative weights to each criterion and then
scoring each alternative against the criteria.  The present worth costs for each alternative
(example values shown in the line plot in Figure F.16) were then divided by the corresponding
benefit score to calculate the cost/benefit ratio (example values shown in line plot in Figure
F.16). The supplier that had the lowest cost/benefit ratio (i.e., low cost and high benefit), was
then selected.
F.5.1.4   Key Design Parameters

       Six Calgon Sentinel® 48-inch UV reactors comprise the UV facility. Key design
parameters for the UV reactors are shown in Table F.23. The target MS2 RED to be verified in
validation was based on criteria from the UVDGM Proposal Draft (USEPA 2003) for 2-log
inactivation of Cryptosporidium.
       Following construction of Winnipeg's WTP, the design UVT will be increased to
90 percent. The design flow rate per reactor will also be increased; however, the total design
flow rate through the facility will be reduced to 106 mgd as reactors are changed to stand-by and
other measures are taken to improve UV facility redundancy.

         Figure F.16. Cost-Benefit Comparison for Winnipeg's UV Reactors
            2

           1.8

           1.6

           1.4

           1.2

            1

           0.8

           0.6

           0.4

           0.2

            0
       $10.0


       $8.0  -
            »
       $6.0  3
            W)
            o
            O
       $4.0


       $2.0


       $0.0


       -$2.0


       -$4.0


       -$6.0


       -$8.0
  1 Equipment
   characteristics

  1 Maintenance
I  I Manufacturer's
   experience

   Ope ration
   Cost/Benefit ($M)
  -Present Worth ($m)
                Supplier 1     Supplier 2     Suppliers
                      UV Equipment Suppliers
UV Disinfection Guidance Manual
For the Final LT2ESWTR
F-41
                 November 2006

-------
                                   Appendix F. Case Studies
                      Table F.23. UV Reactor Design Parameters
Criterion
UV Transmittance
Fouling/Aging Factor
Total Flow Rate
MS2 RED1
Number of UV Units (All Duty)
Design Flow Rate per Unit
Number of Lamps per Unit
Lamp Power (Each)
Unit
percent
percent
mgd
mJ/cm2
number
mgd
number
kW
Value
75
70
130
28
6
22
9
21.6
                  The MS2 RED to be proven during validation testing.
Facility Hydraulics

       No modifications to the facility hydraulics were required for the addition of the UV
facility at the existing building. Furthermore, the hydraulics of the future WTP will be designed
to incorporate the UV facility.

Operational Strategy, Instrumentation, and Control

       Control of the UV reactors is based on the UV intensity  setpoint (i.e., UV Intensity
Setpoint Approach). A UV sensor was provided for each lamp in each reactor to monitor
performance. Flow meters and modulating valves on each reactor are used to distribute the water
among operating reactors, and isolation valves are located upstream of each reactor.

       As the system demand increases and reactors approach their maximum capacities,
additional reactors are started up as needed. The procedure is followed in reverse as system
demand decreases.

Electrical Power  Configuration and Power Quality

       Power quality problems are not common at the location  of Winnipeg's UV facility, so a
UPS was not provided. For initial (unfiltered) operation, a back-up generation system was not
provided for the UV facility; therefore, the UV facility is not operational during power outages.
A back-up power system will be provided for long-term operation when the WTP is constructed.

Capital Cost

       The total capital cost for the City of Winnipeg's UV facility was approximately
$5,885,000 in 2006 U.S. dollars (ENR BCI = 4356). The cost includes the UV reactors, piping,
valves, instrumentation and controls, and other ancillary equipment.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
F-42
November 2006

-------
                                  Appendix F. Case Studies
F.5.2  Validation

       Although the UV reactors had been validated off-site before installation, the off-site
testing had focused mainly on typical UVT levels, and only a limited number of runs had tested
UVT levels below 80 percent. Therefore, the City of Winnipeg made on-site validation testing a
bidding requirement. The on-site validation testing focused on lower UVT levels (70 -
78 percent), consistent with the raw water to be treated by the UV facility. The on-site testing
included tests at a range of flow rates (6 - 25 mgd) and lamp settings.

       The on-site testing was conducted in February 2005. The challenge microorganism was
MS2, and SuperHume™ (potassium humate salts) was used to adjust the UVT of the test water.
Thirty-eight tests were run, and additional blanks and other quality control samples were also
taken. An excerpt of the validation testing conditions and results are shown in Table F.24.
F.5.3  Start-up and Operation of the UV Facility

       Construction of the UV facility was completed in December 2004. Although validation
has been completed, full-scale operation of the facility will be started up after functional testing
has been completed.
       Table F.24. Excerpt of the Validation Testing Conditions and Results
                               for the City of Winnipeg
Run
No.
1
2
3
4
13
14
16
17
30
31
Flow
(mgd)
24.9
25.1
25.0
24.9
25.3
25.0
12.5
12.6
6.3
6.3
Configur-
ation1
3 banks on
3 banks on
3 banks on
3 banks on
3 banks on
2 banks on
3 banks on
2 banks on
2 banks on
3 banks on
UVT
Modifier
SH
SH
SH
SH
SH
SH
SH
SH
SH
SH
Test
Organism
MS2
MS2
MS2
MS2
MS2
MS2
MS2
MS2
MS2
MS2
UV
Intensity
(W/m2)
138.0
97.0
155.0
63.0
30.0
87.0
117.0
63.0
63.0
48.0
UVT
(%)
74.9
74.9
77.5
77.5
70.2
70.0
77.5
77.8
77.7
74.8
Influent MS2
(log PFU/mL)
4.94
5.02
5.26
5.27
5.35
5.44
5.25
5.26
5.28
5.36
Effluent MS2
(log PFU/mL)
3.18
3.21
3.31
4.27
4.77
4.24
2.84
3.67
3.59
3.22
Log
Reduction
1.76
1.80
1.94
1.00
0.58
1.21
2.41
1.59
1.70
2.14
MS2
RED
(mJ/cm2)
32.6
33.4
36.5
17.3
9.4
20.5
48.3
29.7
31.9
41.9
1 Each reactor contains 3 banks with
SH-SuperHume™
3 lamps per bank.
UV Disinfection Guidance Manual
For the Final LT2ESWTR
              F-43
November 2006

-------
                                      Appendix F. Case Studies
                                This page intentionally left blank.
UV Disinfection Guidance Manual                 F-44                                 November 2006
For the Final LT2ESWTR

-------
            Appendix G



Reduction Equivalent Dose Bias Tables

-------
This page intentionally left blank.

-------
                        Appendix G. Reduction Equivalent Dose Bias Tables
       Tables G.I - G.I 7 present RED Bias as a function of water ultraviolet transmittance
(UVT) and challenge microorganism UV sensitivity for various log inactivation levels (ranging
from 4.0 - 0.5) for Cryptosporidium, Giardia, and viruses. Tables G. 1 - G.8 present RED Bias
values for Cryptosporidium, Tables G.9 - G. 16 present RED Bias values for Giardia, and Table
G. 17 presents RED Bias values for viruses. The RED Bias values for intermediate UVT values
(e.g., UVT between 85 and 90 percent) can be interpolated from the values in the table, if
desired.
                                    Index of Tables
Table No.
G.1
G.2
G.3
G.4
G.5
G.6
G.7
G.8
G.9
G10.
G.11
G.12
G.13
G.14
G.15
G.16
G.17
Table Title
RED Bias Values for 4.0-log Cryptosporidium Inactivation Credit as a Function of UVT and UV
Challenge Microorganism Sensitivity
RED Bias Values for 3.5-log Cryptosporidium Inactivation Credit as a Function of UVT and UV
Challenge Microorganism Sensitivity
RED Bias Values for 3.0-log Cryptosporidium Inactivation Credit as a Function of UVT and UV
Challenge Microorganism Sensitivity
RED Bias Values for 2.5-log Cryptosporidium Inactivation Credit as a Function of UVT and UV
Challenge Microorganism Sensitivity
RED Bias Values for 2.0-log Cryptosporidium Inactivation Credit as a Function of UVT and UV
Challenge Microorganism Sensitivity
RED Bias Values for 1.5-log Cryptosporidium Inactivation Credit as a Function of UVT and UV
Challenge Microorganism Sensitivity
RED Bias Values for 1.0-log Cryptosporidium Inactivation Credit as a Function of UVT and UV
Challenge Microorganism Sensitivity
RED Bias Values for 0.5-log Cryptosporidium Inactivation Credit as a Function of UVT and UV
Challenge Microorganism Sensitivity
RED Bias Values for 4.0-log Giardia Inactivation Credit as a Function of UVT and UV
Challenge Microorganism Sensitivity
RED Bias Values for 3.5-log Giardia Inactivation Credit as a Function of UVT and UV
Challenge Microorganism Sensitivity
RED Bias Values for 3.0-log Giardia Inactivation Credit as a Function of UVT and UV
Challenge Microorganism Sensitivity
RED Bias Values for 2.5-log Giardia Inactivation Credit as a Function of UVT and UV
Challenge Microorganism Sensitivity
RED Bias Values for 2.0-log Giardia Inactivation Credit as a Function of UVT and UV
Challenge Microorganism Sensitivity
RED Bias Values for 1.5-log Giardia Inactivation Credit as a Function of UVT and UV
Challenge Microorganism Sensitivity
RED Bias Values for 1.0-log Giardia Inactivation Credit as a Function of UVT and UV
Challenge Microorganism Sensitivity
RED Bias Values for 0.5-log Giardia Inactivation Credit as a Function of UVT and UV
Challenge Microorganism Sensitivity
RED Bias Values for Virus Inactivation Credit as a Function of UV Challenge Microorganism
Sensitivity
Page
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
UV Disinfection Guidance Manual
For the Final LT2ESWTR
G-l
November 2006

-------
                      Appendix G. Reduction Equivalent Dose Bias Tables
 Table G.1. RED Bias Values for 4.0-log Cryptosporidium Inactivation Credit as a
          Function of UVT and UV Challenge Microorganism Sensitivity
Cryptosporidium log inactivation credit
Required UV dose (mJ/cm2)
Cryptosporidium UV sensitivity (mJ/cm2/log I)
UVT (%)
Challenge UV sensitivity (mJ/cm2/log I)
Lower
0
>2
>4
>6
>8
> 10
>12
>14
>16
>18
>20
>22
>24
>26
>28
>30
>32
Upper
<2
<4
<6
<8
<10
< 12
<14
<16
<18
<20
<22
<24
<26
<28
<30
<32
<34
4.0
22
5.5
>98
>95
>90
>85
>80
>75
>65
RED Bias
1.00
1.00
1.01
1.05
1.08
1.11
1.13
1.15
1.16
1.17
1.18
1.19
1.20
1.21
1.22
1.22
1.23
1.00
1.00
1.02
1.09
1.15
1.20
1.24
1.28
1.31
1.34
1.36
1.38
1.40
1.41
1.43
1.44
1.45
1.00
1.00
1.03
1.12
1.21
1.29
1.37
1.44
1.50
1.55
1.61
1.66
1.70
1.74
1.78
1.81
1.85
1.00
1.00
1.03
1.14
1.25
1.34
1.44
1.53
1.61
1.69
1.77
1.84
1.91
1.98
2.04
2.10
2.16
1.00
1.00
1.03
1.16
1.27
1.38
1.49
1.59
1.69
1.78
1.87
1.96
2.05
2.14
2.22
2.30
2.38
1.00
1.00
1.03
1.17
1.30
1.42
1.53
1.65
1.76
1.87
1.97
2.08
2.18
2.28
2.38
2.47
2.57
1.00
1.00
1.04
1.18
1.33
1.46
1.60
1.73
1.86
1.99
2.11
2.24
2.36
2.48
2.60
2.73
2.84
UV Disinfection Guidance Manual
For the Final LT2ESWTR
G-2
November 2006

-------
                      Appendix G. Reduction Equivalent Dose Bias Tables
 Table G.2. RED Bias Values for 3.5-log Cryptosporidium Inactivation Credit as a
          Function of UVT and UV Challenge Microorganism Sensitivity
Cryptosporidium log inactivation credit
Required UV dose (mJ/cm2)
Cryptosporidium UV sensitivity (mJ/cm2/log 1)
UVT (%)
Challenge UV sensitivity (mJ/cm2/log 1)
Lower
0
>2
>4
>6
>8
> 10
>12
>14
>16
>18
>20
>22
>24
>26
>28
>30
>32
Upper
<2
<4
<6
<8
<10
< 12
<14
<16
<18
<20
<22
<24
<26
<28
<30
<32
<34
3.5
15
4.3
>98
>95
>90
>85
>80
>75
>65
RED Bias
1.00
1.00
1.05
1.09
1.11
1.14
1.15
1.17
1.18
1.19
1.20
1.21
1.22
1.23
1.23
1.24
1.24
1.00
1.00
1.08
1.16
1.22
1.27
1.30
1.33
1.36
1.38
1.40
1.42
1.43
1.45
1.46
1.47
1.48
1.00
1.00
1.11
1.23
1.33
1.43
1.51
1.58
1.64
1.70
1.75
1.79
1.83
1.87
1.91
1.94
1.97
1.00
1.00
1.13
1.27
1.40
1.52
1.63
1.73
1.83
1.92
2.01
2.09
2.16
2.23
2.30
2.36
2.42
1.00
1.00
1.14
1.30
1.44
1.58
1.71
1.84
1.96
2.08
2.19
2.30
2.40
2.50
2.60
2.69
2.78
1.00
1.00
1.16
1.33
1.49
1.64
1.79
1.94
2.08
2.21
2.35
2.48
2.61
2.73
2.86
2.98
3.09
1.00
1.00
1.17
1.36
1.55
1.73
1.90
2.07
2.24
2.41
2.58
2.74
2.90
3.07
3.23
3.38
3.54
UV Disinfection Guidance Manual
For the Final LT2ESWTR
G-3
November 2006

-------
                      Appendix G. Reduction Equivalent Dose Bias Tables
 Table G.3. RED Bias Values for 3.0-log Cryptosporidium Inactivation Credit as a
          Function of UVT and UV Challenge Microorganism Sensitivity
Cryptosporidium log inactivation credit
Required UV dose (mJ/cm2)
Cryptosporidium UV sensitivity (mJ/cm2/log 1)
UVT (%)
Challenge UV sensitivity (mJ/cm2/log 1)
Lower
0
>2
>4
>6
>8
> 10
>12
>14
>16
>18
>20
>22
>24
>26
>28
>30
>32
Upper
<2
<4
<6
<8
< 10
< 12
<14
<16
<18
<20
<22
<24
<26
<28
<30
<32
<34
3.0
12
4.0
>98
>95
>90
>85
>80
>75
>65
RED Bias
1.00
1.00
1.05
1.09
1.12
1.14
1.16
1.17
1.18
1.19
1.20
1.21
1.22
1.22
1.23
1.23
1.24
1.00
1.00
1.10
1.18
1.23
1.27
1.31
1.33
1.36
1.38
1.39
1.41
1.42
1.43
1.44
1.45
1.46
1.00
1.00
1.15
1.27
1.38
1.47
1.55
1.62
1.68
1.73
1.78
1.82
1.85
1.89
1.92
1.95
1.97
1.00
1.00
1.17
1.32
1.47
1.59
1.71
1.82
1.92
2.01
2.10
2.18
2.25
2.32
2.38
2.44
2.50
1.00
1.00
1.19
1.36
1.52
1.68
1.82
1.96
2.09
2.22
2.34
2.45
2.56
2.66
2.76
2.86
2.95
1.00
1.00
1.21
1.40
1.58
1.75
1.92
2.08
2.24
2.39
2.54
2.69
2.83
2.96
3.10
3.23
3.35
1.00
1.00
1.23
1.45
1.66
1.86
2.06
2.26
2.45
2.65
2.84
3.03
3.21
3.40
3.58
3.76
3.94
UV Disinfection Guidance Manual
For the Final LT2ESWTR
G-4
November 2006

-------
                      Appendix G. Reduction Equivalent Dose Bias Tables
 Table G.4. RED Bias Values for 2.5-log Cryptosporidium Inactivation Credit as a
          Function of UVT and UV Challenge Microorganism Sensitivity
Cryptosporidium log inactivation credit
Required UV dose (mJ/cm2)
Cryptosporidium UV sensitivity (mJ/cm2/log 1)
UVT (%)
Challenge UV sensitivity (mJ/cm2/log 1)
Lower
0
>2
>4
>6
>8
> 10
>12
>14
>16
>18
>20
>22
>24
>26
>28
>30
>32
Upper
<2
<4
<6
<8
< 10
<12
<14
<16
<18
<20
<22
<24
<26
<28
<30
<32
<34
2.5
8.5
3.4
>98
>95
>90
>85
>80
>75
>65
RED Bias
1.00
1.02
1.07
1.11
1.13
1.15
1.17
1.18
1.19
1.20
1.21
1.21
1.22
1.23
1.23
1.23
1.24
1.00
1.04
1.14
1.21
1.26
1.30
1.32
1.35
1.37
1.38
1.40
1.41
1.42
1.43
1.44
1.45
1.45
1.00
1.06
1.22
1.36
1.46
1.55
1.63
1.69
1.74
1.79
1.83
1.87
1.90
1.93
1.95
1.97
1.99
1.00
1.07
1.27
1.44
1.60
1.74
1.87
1.98
2.08
2.17
2.26
2.33
2.40
2.47
2.53
2.58
2.63
1.00
1.07
1.30
1.50
1.69
1.87
2.03
2.19
2.34
2.47
2.60
2.72
2.84
2.95
3.05
3.15
3.24
1.00
1.08
1.33
1.56
1.77
1.98
2.18
2.37
2.56
2.74
2.91
3.07
3.23
3.39
3.54
3.68
3.82
1.00
1.09
1.37
1.63
1.89
2.15
2.39
2.64
2.88
3.12
3.35
3.58
3.81
4.03
4.26
4.48
4.70
UV Disinfection Guidance Manual
For the Final LT2ESWTR
G-5
November 2006

-------
                      Appendix G. Reduction Equivalent Dose Bias Tables
 Table G.5. RED Bias Values for 2.0-log Cryptosporidium Inactivation Credit as a
          Function of UVT and UV Challenge Microorganism Sensitivity
Cryptosporidium log inactivation credit
Required UV dose (mJ/cm2)
Cryptosporidium UV sensitivity (mJ/cm2/log 1)
UVT (%)
Challenge UV sensitivity (mJ/cm2/log 1)
Lower
0
>2
>4
>6
>8
> 10
>12
>14
>16
>18
>20
>22
>24
>26
>28
>30
>32
Upper
<2
<4
<6
<8
< 10
< 12
<14
<16
<18
<20
<22
<24
<26
<28
<30
<32
<34
2.0
5.8
2.9
>98
>95
>90
>85
>80
>75
>65
RED Bias
1.00
1.04
1.09
1.12
1.14
1.16
1.17
1.18
1.19
1.20
1.21
1.21
1.22
1.23
1.23
1.23
1.24
1.00
1.08
1.17
1.23
1.27
1.31
1.33
1.35
1.37
1.38
1.39
1.40
1.41
1.42
1.42
1.43
1.44
1.00
1.12
1.30
1.43
1.53
1.62
1.68
1.74
1.78
1.82
1.85
1.88
1.91
1.93
1.95
1.97
1.99
1.00
1.14
1.37
1.57
1.74
1.88
2.01
2.12
2.22
2.30
2.38
2.45
2.51
2.57
2.62
2.67
2.71
1.00
1.16
1.42
1.66
1.88
2.08
2.26
2.43
2.58
2.73
2.86
2.98
3.09
3.20
3.30
3.39
3.48
1.00
1.18
1.47
1.75
2.00
2.25
2.48
2.70
2.91
3.11
3.31
3.49
3.66
3.83
3.99
4.14
4.29
1.00
1.20
1.54
1.87
2.19
2.50
2.80
3.10
3.40
3.69
3.97
4.25
4.53
4.80
5.06
5.33
5.59
UV Disinfection Guidance Manual
For the Final LT2ESWTR
G-6
November 2006

-------
                      Appendix G. Reduction Equivalent Dose Bias Tables
 Table G.6. RED Bias Values for 1.5-log Cryptosporidium Inactivation Credit as a
          Function of UVT and UV Challenge Microorganism Sensitivity
Cryptosporidium log inactivation credit
Required UV dose (mJ/cm2)
Cryptosporidium UV sensitivity (mJ/cm2/log I)
UVT (%)
Challenge UV sensitivity (mJ/cm2/log I)
Lower
0
>2
>4
>6
>8
> 10
>12
>14
>16
>18
>20
>22
>24
>26
>28
>30
>32
Upper
<2
<4
<6
<8
<10
<12
<14
<16
<18
<20
<22
<24
<26
<28
<30
<32
<34
1.5
3.9
2.6
>98
>95
>90
>85
>80
>75
>65
RED Bias
1.00
1.05
1.10
1.12
1.14
1.16
1.17
1.18
1.19
1.20
1.20
1.21
1.21
1.22
1.22
1.22
1.23
1.00
1.10
1.18
1.23
1.27
1.30
1.32
1.33
1.35
1.36
1.37
1.37
1.38
1.39
1.39
1.40
1.40
1.00
1.17
1.34
1.47
1.56
1.63
1.68
1.73
1.77
1.80
1.83
1.85
1.87
1.89
1.90
1.92
1.93
1.00
1.21
1.46
1.66
1.83
1.97
2.09
2.18
2.27
2.35
2.41
2.47
2.53
2.57
2.62
2.66
2.69
1.00
1.24
1.54
1.80
2.04
2.25
2.43
2.60
2.75
2.89
3.01
3.13
3.23
3.33
3.42
3.51
3.59
1.00
1.26
1.60
1.92
2.21
2.49
2.74
2.98
3.21
3.42
3.62
3.80
3.98
4.15
4.31
4.46
4.60
1.00
1.30
1.71
2.10
2.48
2.85
3.21
3.56
3.90
4.24
4.57
4.89
5.21
5.52
5.82
6.12
6.41
UV Disinfection Guidance Manual
For the Final LT2ESWTR
G-7
November 2006

-------
                      Appendix G. Reduction Equivalent Dose Bias Tables
 Table G.7. RED Bias Values for 1.0-log Cryptosporidium Inactivation Credit as a
          Function of UVT and UV Challenge Microorganism Sensitivity
Cryptosporidium log inactivation credit
Required UV dose (mJ/cm2)
Cryptosporidium UV sensitivity (mJ/cm2/log I)
UVT (%)
Challenge UV sensitivity (mJ/cm2/log I)
Lower
0
>2
>4
>6
>8
>10
>12
>14
>16
>18
>20
>22
>24
>26
>28
>30
>32
Upper
<2
<4
<6
<8
< 10
< 12
<14
<16
<18
<20
<22
<24
<26
<28
<30
<32
<34
1.0
2.5
2.5
>98
>95
>90
>85
>80
>75
>65
RED Bias
1.00
1.05
1.09
1.12
1.14
1.15
1.16
1.17
1.18
1.18
1.19
1.19
1.20
1.20
1.20
1.21
1.21
1.00
1.10
1.17
1.21
1.24
1.26
1.28
1.29
1.30
1.31
1.32
1.33
1.33
1.34
1.34
1.35
1.35
1.00
1.18
1.34
1.45
1.52
1.58
1.62
1.66
1.69
1.71
1.73
1.75
1.76
1.78
1.79
1.80
1.81
1.00
1.24
1.49
1.68
1.83
1.95
2.05
2.13
2.20
2.26
2.32
2.36
2.40
2.44
2.47
2.50
2.53
1.00
1.28
1.60
1.87
2.10
2.30
2.47
2.63
2.76
2.88
2.99
3.09
3.18
3.26
3.33
3.40
3.47
1.00
1.32
1.69
2.04
2.35
2.63
2.89
3.12
3.34
3.54
3.73
3.90
4.06
4.22
4.36
4.49
4.62
1.00
1.37
1.83
2.28
2.71
3.12
3.53
3.91
4.29
4.66
5.01
5.36
5.69
6.02
6.33
6.64
6.94
UV Disinfection Guidance Manual
For the Final LT2ESWTR
G-8
November 2006

-------
                      Appendix G. Reduction Equivalent Dose Bias Tables
 Table G.8. RED Bias Values for 0.5-log Cryptosporidium Inactivation Credit as a
          Function of UVT and UV Challenge Microorganism Sensitivity
Cryptosporidium log inactivation credit
Required UV dose (mJ/cm2)
Cryptosporidium UV sensitivity (mJ/cm2/log 1)
UVT (%)
Challenge UV sensitivity (mJ/cm2/log 1)
Lower
0
>2
>4
>6
>8
> 10
>12
>14
>16
>18
>20
>22
>24
>26
>28
>30
>32
Upper
<2
<4
<6
<8
<10
<12
<14
<16
<18
<20
<22
<24
<26
<28
<30
<32
<34
0.5
1.6
3.2
>98
>95
>90
>85
>80
>75
>65
RED Bias
1.00
1.02
1.06
1.08
1.10
1.11
1.12
1.12
1.13
1.14
1.14
1.14
1.15
1.15
1.15
1.16
1.16
1.00
1.04
1.10
1.13
1.15
1.17
1.18
1.19
1.20
1.21
1.21
1.22
1.22
1.23
1.23
1.23
1.23
1.00
1.07
1.19
1.27
1.32
1.36
1.39
1.41
1.43
1.45
1.46
1.47
1.48
1.49
1.50
1.51
1.51
1.00
1.10
1.30
1.44
1.55
1.63
1.69
1.74
1.79
1.83
1.86
1.89
1.91
1.93
1.95
1.97
1.98
1.00
1.13
1.40
1.61
1.79
1.93
2.04
2.15
2.23
2.31
2.38
2.44
2.49
2.54
2.58
2.62
2.66
1.00
1.15
1.49
1.78
2.03
2.24
2.43
2.60
2.75
2.89
3.02
3.13
3.24
3.33
3.43
3.51
3.59
1.00
1.18
1.62
2.04
2.42
2.79
3.13
3.45
3.76
4.06
4.33
4.60
4.86
5.10
5.34
5.56
5.78
UV Disinfection Guidance Manual
For the Final LT2ESWTR
G-9
November 2006

-------
                       Appendix G. Reduction Equivalent Dose Bias Tables
      Table G.9. RED Bias Values for 4.0-log Giardia Inactivation Credit as a
           Function of UVT and UV Challenge Microorganism Sensitivity
Giardia log inactivation credit
Required UV dose (mJ/cm2)
Giardia UV sensitivity (mJ/cm2/log 1)
UVT (%)
Challenge UV sensitivity (mJ/cm2/log 1)
Lower
0
>2
>4
>6
>8
>10
>12
>14
>16
>18
>20
>22
>24
>26
>28
>30
>32
Upper
<2
<4
<6
<8
<10
<12
<14
<16
<18
<20
<22
<24
<26
<28
<30
<32
<34
4.0
22
5.5
>98
>95
>90
>85
>80
>75
>65
RED Bias
1.00
1.00
1.00
1.01
1.05
1.08
1.11
1.13
1.15
1.16
1.17
1.18
1.19
1.20
1.21
1.22
1.22
1.00
1.00
1.00
1.02
1.09
1.15
1.20
1.24
1.28
1.31
1.34
1.36
1.38
1.40
1.41
1.43
1.44
1.00
1.00
1.00
1.03
1.12
1.21
1.29
1.37
1.44
1.50
1.55
1.61
1.66
1.70
1.74
1.78
1.81
1.00
1.00
1.00
1.03
1.14
1.25
1.34
1.44
1.53
1.61
1.69
1.77
1.84
1.91
1.98
2.04
2.10
1.00
1.00
1.00
1.03
1.16
1.27
1.38
1.49
1.59
1.69
1.78
1.87
1.96
2.05
2.14
2.22
2.30
1.00
1.00
1.00
1.03
1.17
1.30
1.42
1.53
1.65
1.76
1.87
1.97
2.08
2.18
2.28
2.38
2.47
1.00
1.00
1.00
1.04
1.18
1.33
1.46
1.60
1.73
1.86
1.99
2.11
2.24
2.36
2.48
2.60
2.73
UV Disinfection Guidance Manual
For the Final LT2ESWTR
G-10
November 2006

-------
                       Appendix G. Reduction Equivalent Dose Bias Tables
      Table G.10. RED Bias Values for 3.5-log Giardia Inactivation Credit as a
           Function of UVT and UV Challenge Microorganism Sensitivity
Giardia log inactivation credit
Required UV dose (mJ/cm2)
Giardia UV sensitivity (mJ/cm2/log I)
UVT (%)
Challenge UV sensitivity (mJ/cm2/log I)
Lower
0
>2
>4
>6
>8
>10
>12
>14
>16
>18
>20
>22
>24
>26
>28
>30
>32
Upper
<2
<4
<6
<8
<10
< 12
<14
<16
<18
<20
<22
<24
<26
<28
<30
<32
<34
3.5
15
4.3
>98
>95
>90
>85
>80
>75
>65
RED Bias
1.00
1.00
1.05
1.09
1.11
1.14
1.15
1.17
1.18
1.19
1.20
1.21
1.22
1.23
1.23
1.24
1.24
1.00
1.00
1.08
1.16
1.22
1.27
1.30
1.33
1.36
1.38
1.40
1.42
1.43
1.45
1.46
1.47
1.48
1.00
1.00
1.11
1.23
1.33
1.43
1.51
1.58
1.64
1.70
1.75
1.79
1.83
1.87
1.91
1.94
1.97
1.00
1.00
1.13
1.27
1.40
1.52
1.63
1.73
1.83
1.92
2.01
2.09
2.16
2.23
2.30
2.36
2.42
1.00
1.00
1.14
1.30
1.44
1.58
1.71
1.84
1.96
2.08
2.19
2.30
2.40
2.50
2.60
2.69
2.78
1.00
1.00
1.16
1.33
1.49
1.64
1.79
1.94
2.08
2.21
2.35
2.48
2.61
2.73
2.86
2.98
3.09
1.00
1.00
1.17
1.36
1.55
1.73
1.90
2.07
2.24
2.41
2.58
2.74
2.90
3.07
3.23
3.38
3.54
UV Disinfection Guidance Manual
For the Final LT2ESWTR
G-ll
November 2006

-------
                       Appendix G. Reduction Equivalent Dose Bias Tables
      Table G.11. RED Bias Values for 3.0-log Giardia Inactivation Credit as a
           Function of UVT and UV Challenge Microorganism Sensitivity
Giardia log inactivation credit
Required UV dose (mJ/cm2)
Giardia UV sensitivity (mJ/cm2/log I)
UVT (%)
Challenge UV sensitivity (mJ/cm2/log I)
Lower
0
>2
>4
>6
>8
> 10
>12
>14
>16
>18
>20
>22
>24
>26
>28
>30
>32
Upper
<2
<4
<6
<8
<10
< 12
<14
<16
<18
<20
<22
<24
<26
<28
<30
<32
<34
3.0
11
3.7
>98
>95
>90
>85
>80
>75
>65
RED Bias
1.00
1.01
1.07
1.10
1.13
1.15
1.16
1.18
1.19
1.20
1.21
1.22
1.22
1.23
1.24
1.24
1.25
1.00
1.02
1.12
1.20
1.25
1.29
1.32
1.35
1.37
1.39
1.41
1.42
1.44
1.45
1.46
1.47
1.47
1.00
1.03
1.18
1.31
1.42
1.51
1.59
1.66
1.72
1.77
1.81
1.85
1.89
1.92
1.95
1.98
2.00
1.00
1.03
1.21
1.37
1.52
1.65
1.77
1.89
1.99
2.08
2.17
2.25
2.32
2.39
2.46
2.52
2.57
1.00
1.03
1.23
1.41
1.58
1.74
1.90
2.04
2.18
2.31
2.43
2.55
2.66
2.77
2.87
2.97
3.06
1.00
1.03
1.25
1.45
1.65
1.83
2.01
2.18
2.34
2.51
2.66
2.81
2.96
3.11
3.25
3.38
3.52
1.00
1.04
1.28
1.51
1.74
1.95
2.17
2.38
2.59
2.79
2.99
3.19
3.39
3.59
3.78
3.98
4.17
UV Disinfection Guidance Manual
For the Final LT2ESWTR
G-12
November 2006

-------
                       Appendix G. Reduction Equivalent Dose Bias Tables
      Table G.12. RED Bias Values for 2.5-log Giardia Inactivation Credit as a
           Function of UVT and UV Challenge Microorganism Sensitivity
Giardia log inactivation credit
Required UV dose (mJ/cm2)
Giardia UV sensitivity (mJ/cm2/log I)
UVT (%)
Challenge UV sensitivity (mJ/cm2/log I)
Lower
0
>2
>4
>6
>8
>10
>12
>14
>16
>18
>20
>22
>24
>26
>28
>30
>32
Upper
<2
<4
<6
<8
<10
<12
<14
<16
<18
<20
<22
<24
<26
<28
<30
<32
<34
2.5
7.7
3.1
>98
>95
>90
>85
>80
>75
>65
RED Bias
1.00
1.03
1.08
1.12
1.14
1.16
1.18
1.19
1.20
1.21
1.22
1.22
1.23
1.23
1.24
1.24
1.25
1.00
1.06
1.16
1.23
1.28
1.32
1.34
1.37
1.39
1.40
1.42
1.43
1.44
1.45
1.46
1.46
1.47
1.00
1.09
1.26
1.40
1.51
1.60
1.67
1.74
1.79
1.83
1.87
1.91
1.94
1.97
1.99
2.01
2.03
1.00
1.11
1.32
1.50
1.67
1.81
1.94
2.06
2.16
2.25
2.34
2.41
2.48
2.55
2.61
2.66
2.71
1.00
1.12
1.35
1.57
1.77
1.96
2.13
2.29
2.45
2.59
2.72
2.85
2.97
3.08
3.18
3.28
3.38
1.00
1.13
1.39
1.63
1.87
2.09
2.30
2.50
2.70
2.88
3.07
3.24
3.41
3.57
3.73
3.88
4.02
1.00
1.14
1.44
1.73
2.00
2.27
2.54
2.80
3.06
3.31
3.56
3.81
4.05
4.30
4.53
4.77
5.00
UV Disinfection Guidance Manual
For the Final LT2ESWTR
G-13
November 2006

-------
                       Appendix G. Reduction Equivalent Dose Bias Tables
      Table G.13. RED Bias Values for 2.0-log Giardia Inactivation Credit as a
           Function of UVT and UV Challenge Microorganism Sensitivity
Giardia log inactivation credit
Required UV dose (mJ/cm2)
Giardia UV sensitivity (mJ/cm2/log I)
UVT (%)
Challenge UV sensitivity (mJ/cm2/log I)
Lower
0
>2
>4
>6
>8
>10
>12
>14
>16
>18
>20
>22
>24
>26
>28
>30
>32
Upper
<2
<4
<6
<8
<10
<12
<14
<16
<18
<20
<22
<24
<26
<28
<30
<32
<34
2.0
5.2
2.6
>98
>95
>90
>85
>80
>75
>65
RED Bias
1.00
1.05
1.10
1.13
1.15
1.17
1.18
1.20
1.21
1.21
1.22
1.23
1.23
1.24
1.24
1.24
1.25
1.00
1.11
1.20
1.26
1.30
1.33
1.36
1.37
1.39
1.40
1.41
1.42
1.43
1.44
1.45
1.45
1.46
1.00
1.17
1.35
1.48
1.59
1.67
1.74
1.79
1.84
1.87
1.91
1.94
1.96
1.98
2.01
2.02
2.04
1.00
1.20
1.44
1.65
1.82
1.97
2.10
2.21
2.31
2.40
2.48
2.55
2.61
2.67
2.72
2.77
2.81
1.00
1.22
1.50
1.76
1.99
2.19
2.39
2.56
2.72
2.87
3.01
3.13
3.25
3.36
3.46
3.56
3.65
1.00
1.24
1.56
1.85
2.13
2.39
2.64
2.87
3.09
3.30
3.51
3.70
3.88
4.06
4.22
4.38
4.54
1.00
1.27
1.64
1.99
2.34
2.67
3.00
3.32
3.64
3.95
4.25
4.55
4.85
5.14
5.43
5.71
5.98
UV Disinfection Guidance Manual
For the Final LT2ESWTR
G-14
November 2006

-------
                       Appendix G. Reduction Equivalent Dose Bias Tables
      Table G.14. RED Bias Values for 1.5-log Giardia Inactivation Credit as a
           Function of UVT and UV Challenge Microorganism Sensitivity
Giardia log inactivation credit
Required UV dose (mJ/cm2)
Giardia UV sensitivity (mJ/cm2/log I)
UVT (%)
Challenge UV sensitivity (mJ/cm2/log I)
Lower
0
>2
>4
>6
>8
>10
>12
>14
>16
>18
>20
>22
>24
>26
>28
>30
>32
Upper
<2
<4
<6
<8
<10
< 12
<14
<16
<18
<20
<22
<24
<26
<28
<30
<32
<34
1.5
3.0
2.0
>98
>95
>90
>85
>80
>75
>65
RED Bias
1.00
1.08
1.13
1.15
1.17
1.19
1.20
1.21
1.22
1.23
1.23
1.24
1.24
1.25
1.25
1.25
1.26
1.00
1.16
1.25
1.30
1.33
1.36
1.38
1.39
1.41
1.42
1.43
1.43
1.44
1.45
1.45
1.46
1.46
1.00
1.28
1.47
1.60
1.69
1.77
1.82
1.87
1.91
1.94
1.96
1.99
2.01
2.03
2.04
2.06
2.07
1.00
1.35
1.63
1.86
2.04
2.19
2.31
2.42
2.51
2.59
2.66
2.72
2.78
2.83
2.87
2.91
2.95
1.00
1.40
1.75
2.05
2.31
2.54
2.75
2.93
3.10
3.25
3.39
3.51
3.63
3.74
3.84
3.93
4.01
1.00
1.45
1.84
2.21
2.55
2.86
3.15
3.42
3.68
3.92
4.14
4.35
4.55
4.74
4.92
5.09
5.25
1.00
1.51
1.99
2.45
2.89
3.32
3.75
4.16
4.56
4.96
5.34
5.72
6.08
6.44
6.80
7.14
7.48
UV Disinfection Guidance Manual
For the Final LT2ESWTR
G-15
November 2006

-------
                       Appendix G. Reduction Equivalent Dose Bias Tables
      Table G.15. RED Bias Values for 1.0-log Giardia Inactivation Credit as a
           Function of UVT and UV Challenge Microorganism Sensitivity
Giardia log inactivation credit
Required UV dose (mJ/cm2)
Giardia UV sensitivity (mJ/cm2/log I)
UVT (%)
Challenge UV sensitivity (mJ/cm2/log I)
Lower
0
>2
>4
>6
>8
>10
>12
>14
>16
>18
>20
>22
>24
>26
>28
>30
>32
Upper
<2
<4
<6
<8
<10
<12
<14
<16
<18
<20
<22
<24
<26
<28
<30
<32
<34
1.0
2.1
2.1
>98
>95
>90
>85
>80
>75
>65
RED Bias
1.00
1.07
1.11
1.14
1.16
1.17
1.18
1.19
1.20
1.20
1.21
1.21
1.22
1.22
1.22
1.22
1.23
1.00
1.14
1.21
1.25
1.28
1.30
1.32
1.33
1.34
1.35
1.36
1.37
1.37
1.38
1.38
1.38
1.39
1.00
1.26
1.42
1.53
1.61
1.67
1.72
1.75
1.78
1.81
1.83
1.84
1.86
1.87
1.89
1.90
1.91
1.00
1.34
1.61
1.82
1.98
2.10
2.21
2.30
2.37
2.44
2.49
2.54
2.58
2.62
2.66
2.69
2.72
1.00
1.40
1.75
2.04
2.29
2.51
2.70
2.86
3.01
3.14
3.25
3.36
3.46
3.55
3.63
3.70
3.77
1.00
1.45
1.86
2.24
2.58
2.89
3.17
3.43
3.67
3.89
4.10
4.29
4.47
4.63
4.79
4.94
5.08
1.00
1.52
2.04
2.53
3.01
3.47
3.91
4.35
4.77
5.17
5.57
5.95
6.32
6.68
7.03
7.38
7.71
UV Disinfection Guidance Manual
For the Final LT2ESWTR
G-16
November 2006

-------
                       Appendix G. Reduction Equivalent Dose Bias Tables
      Table G.16. RED Bias Values for 0.5-log Giardia Inactivation Credit as a
           Function of UVT and UV Challenge Microorganism Sensitivity
Giardia log inactivation credit
Required UV dose (mJ/cm2)
Giardia UV sensitivity (mJ/cm2/log I)
UVT (%)
Challenge UV sensitivity (mJ/cm2/log I)
Lower
0
>2
>4
>6
>8
> 10
> 12
> 14
> 16
> 18
>20
>22
>24
>26
>28
>30
>32
Upper
<2
<4
<6
<8
<10
< 12
< 14
< 16
< 18
<20
<22
<24
<26
<28
<30
<32
<34
0.5
1.5
3.0
>98
>95
>90
>85
>80
>75
>65
RED Bias
1.00
1.03
1.06
1.09
1.10
1.11
1.12
1.13
1.14
1.14
1.15
1.15
1.15
1.16
1.16
1.16
1.16
1.00
1.05
1.11
1.14
1.16
1.18
1.19
1.20
1.21
1.22
1.22
1.23
1.23
1.24
1.24
1.24
1.25
1.00
1.09
1.22
1.29
1.35
1.39
1.42
1.44
1.46
1.48
1.49
1.50
1.51
1.52
1.53
1.53
1.54
1.00
1.14
1.34
1.49
1.59
1.68
1.74
1.80
1.84
1.88
1.91
1.94
1.97
1.99
2.01
2.03
2.04
1.00
1.17
1.45
1.67
1.85
2.00
2.12
2.23
2.32
2.40
2.47
2.53
2.59
2.64
2.69
2.73
2.77
1.00
1.20
1.55
1.85
2.11
2.34
2.54
2.72
2.88
3.02
3.15
3.28
3.39
3.49
3.58
3.67
3.76
1.00
1.24
1.70
2.13
2.54
2.92
3.28
3.63
3.95
4.26
4.56
4.84
5.11
5.37
5.62
5.86
6.09
UV Disinfection Guidance Manual
For the Final LT2ESWTR
G-17
November 2006

-------
                       Appendix G. Reduction Equivalent Dose Bias Tables
           Table G.17. RED Bias Values for Virus Inactivation Credit as a
               Function of UV Challenge Microorganism Sensitivity
Virus log inactivation credit
Required UV dose (mJ/cm2)
Virus UV sensitivity (mJ/cm2/log I)
Challenge UV sensitivity (mJ/cm2/log I)
Lower
> 1
>25
>50
>60
>70
>80
>90
>90
Upper
<25
<50
<60
<70
<80
<90
<100
<100
0.5
39
78
1.0
58
58
1.5
79
53
2.0
100
50
2.5
121
48
3.0
143
48
3.5
163
47
4.0
186
47
RED Bias
1.00
1.00
1.00
1.00
1.00
1.01
1.02
1.02
1.00
1.00
1.00
1.02
1.04
1.05
1.06
1.06
1.00
1.00
1.02
1.04
1.05
1.06
1.07
1.07
1.00
1.00
1.03
1.05
1.06
1.07
1.08
1.08
1.00
1.01
1.03
1.05
1.07
1.08
1.09
1.09
1.00
1.01
1.03
1.05
1.07
1.08
1.09
1.09
1.00
1.01
1.04
1.06
1.07
1.08
1.09
1.09
1.00
1.01
1.04
1.06
1.07
1.08
1.09
1.09
UV Disinfection Guidance Manual
For the Final LT2ESWTR
G-18
November 2006

-------