EPA/600/R-09/123
September 2009
09/32/WQPC-SWP
Environmental Technology
Verification Report
UV Disinfection for Secondary
Effluent and Reuse Applications
Siemens Water Technologies Corp.
Barrier Sunlight H-4XE-HO Open
Channel UV System
Prepared by
NSF International
Under a Cooperative Agreement with
V>EPA U.S. Environmental Protection Agency
-------
Environmental Technology Verification Report
Verification of Ultraviolet (UV) Disinfection for
Secondary Effluent and Reuse Applications
Siemens Water Technologies
H-4XE-HO (2W-1B-1C) Open Channel UV System
Prepared for
NSF International
Ann Arbor, MI 48105
Prepared by
HydroQual, Inc.
Mahwah, NJ 07430
Under a cooperative agreement with the U.S. Environmental Protection Agency
Raymond Frederick, Project Officer
ETV Source Water Protection Pilot
National Risk Management Research Laboratory
Water Supply and Water Resources Division
U.S. Environmental Protection Agency
Edison, New Jersey 08837
September 2009
-------
NOTICE
The U.S. Environmental Protection Agency (USEPA) through its Office of Research and
Development has financially supported and collaborated with NSF International (NSF) under a
Cooperative Agreement. The Water Quality Protection Center, Source Water Protection area,
operating under the Environmental Technology Verification (ETV) Program, supported this
verification effort. This document has been peer reviewed and reviewed by NSF and USEPA
and recommended for public release.
-------
FOREWORD
The U.S. Environmental Protection Agency (EPA) is charged by Congress with protecting the
Nation's land, air, and water resources. Under a mandate of national environmental laws, the
Agency strives to formulate and implement actions leading to a compatible balance between
human activities and the ability of natural systems to support and nurture life. To meet this
mandate, EPA's research program is providing data and technical support for solving
environmental problems today and building a science knowledge base necessary to manage our
ecological resources wisely, understand how pollutants affect our health, and prevent or reduce
environmental risks in the future.
The National Risk Management Research Laboratory (NRMRL) is the Agency's center for
investigation of technological and management approaches for preventing and reducing risks
from pollution that threaten human health and the environment. The focus of the Laboratory's
research program is on methods and their cost-effectiveness for prevention and control of
pollution to air, land, water, and subsurface resources; protection of water quality in public water
systems; remediation of contaminated sites, sediments and ground water; prevention and control
of indoor air pollution; and restoration of ecosystems. NRMRL collaborates with both public
and private sector partners to foster technologies that reduce the cost of compliance and to
anticipate emerging problems. NRMRL's research provides solutions to environmental problems
by: developing and promoting technologies that protect and improve the environment; advancing
scientific and engineering information to support regulatory and policy decisions; and providing
the technical support and information transfer to ensure implementation of environmental
regulations and strategies at the national, state, and community levels.
This publication has been produced as part of the Laboratory's strategic long-term research plan.
It is published and made available by EPA's Office of Research and Development to assist the
user community and to link researchers with their clients.
The following is the final report on an Environmental Technology Verification (ETV) test
performed for NSF International (NSF) and the United States Environmental Protection Agency
(USEPA) by HydroQual, Inc. The verification test for the Siemens Water Technologies H-4XE-
HO (2W-1B-1C) Open Channel UV Disinfection System was conducted from 9/05/08 to
10/07/08 at the Gloversville Johnstown Waste Water Treatment Facility (GJWWTF) located in
Johnstown, New York.
-------
CONTENTS
NOTICE i
FOREWORD ii
FIGURES vi
APPENDICES ix
GLOSSARY AND DEFINITIONS x
ABBREVIATIONS AND ACRONYMS xii
EXECUTIVE SUMMARY 1
1 INTRODUCTION AND BACKGROUND 1-1
1.1 THE ETV PROGRAM 1-1
1.1.1 Concept of the ETV Program 1-1
1.1.2 The ETV Program for Water Reuse and Secondary Effluent Disinfection 1-1
1.1.3 The Siemens Water Technologies ETV 1-2
1.2 MECHANISM OF UV DISINFECTION 1-2
1.2.1 Practical Application of UV Disinfection 1-2
1.2.2 A Comparison of UV and Chemical Disinfection 1-3
1.2.3 Determining Dose Delivery 1-4
1.2.4 Summary of the Biodosimetric Method to Measure Dose 1-5
2 ROLES AND RESPONSIBILITIES OF PARTICIPANTS IN THE VERIFICATION
TESTING 2-1
2.1 NSF INTERNATIONAL (NSF) 2-1
2.2 U.S. ENVIRONMENTAL PROTECTION AGENCY (USEPA) 2-2
2.3 Field TESTING ORGANIZATION (FTO), HYDROQUAL, INC 2-2
2.4 Validation test facility 2-3
2.5 UV TECHNOLOGY VENDOR - Siemens water technologies 2-6
3 TECHNOLOGY DESCRIPTION 3-1
3.1 Siemens water technologies open channel UV DISINFECTION SYSTEM 3-1
3.1.1 Lamps and Sleeves 3-2
3.1.2 Lamp Output Attenuation by Aging and Sleeve Fouling 3-2
3.1.3 Sleeve Cleaning System 3-2
3.1.4 Electrical Controls 3-2
3.1.5 UV Detectors 3-2
3.1.6 Design Operational Envelope 3-3
3.2 UV TEST STAND SPECIFICATIONS 3-3
3.2.1 Test Channel 3-3
3.3 VERIFICATION TEST CLAIMS 3-5
4 PROCEDURES AND METHODS USED DURING VERIFICATION TESTING 4-1
4.1 General Technical Approach 4-1
4.1.1 Site Preparation 4-1
4.1.2 Water Source 4-2
4.1.3 Challenge Water and Discharge Tanks 4-3
4.1.4 Feed Pumps 4-3
4.1.5 Flow Meter 4-3
4.2 DISINFECTION UNIT STARTUP AND CHARACTERIZATION 4-3
4.2.1 100 Hour Lamp Burn-In 4-3
in
-------
4.2.2 Power Consumption and Flow Characterization 4-3
4.3 BActeriophage PRODUCTION AND CALIBRATION 4-5
4.3.1 Bacteriophage Propagation 4-5
4.3.2 Dose-Response Determination 4-6
4.4 BIODOSIMETRIC FIELD TESTS 4-8
4.4.1 Lamp Sleeve Preparation 4-8
4.4.2 Challenge Water Batch Preparation 4-9
4.4.3 Biodosimetric Flow Tests 4-9
4.4.4 Transmittance Measurement 4-9
4.4.5 Bacteriophage Enumeration 4-9
4.4.6 Dose Determination 4-10
4.5 EXPERIMENTAL TEST MATRIX 4-10
5 RESULTS AND DISCUSSION 5-12
5.1 DISINFECTION UNIT STARTUP AND CHARACTERIZATION 5-12
5.1.1 Power Consumption 5-12
5.1.2 Headloss Measurements 5-12
5.1.3 Intensity Sensor Characterization 5-13
5.1.4 Velocity Profile Measurements 5-16
5.2 Bacteriophage DOSE-RESPONSE CALIBRATION CURVES 5-18
5.2.1 Dose-Response Results 5-18
5.2.2 Dose-Response Calibration Curve 5-18
5.2.3 Collimated Beam Uncertainty 5-26
5.3 DOSE-FLOW ASSAYS 5-27
5.3.1 Intensity Attenuation Factor 5-27
5.3.2 Flow Test Data and Results Summary 5-28
5.3.3 Biodosimetric Data Analysis - RED Algorithm 5-31
5.3.4 Sensor Model 5-33
6 QUALITY ASSURANCE/QUALITY CONTROL 6-1
6.1 CALIBRATIONS 6-1
6.1.1 Flow Meter Calibration 6-1
6.1.2 Spectrophotometer Calibration 6-2
6.1.3 UV Intensity Sensors 6-5
6.1.4 Radiometer Calibration 6-5
6.2 QA/QC OF MICROBIAL SAMPLES 6-5
6.2.1 Reactor Controls 6-5
6.2.2 Reactor Blanks 6-7
6.2.3 Trip Controls 6-8
6.2.4 Flow Test Sample Replicates 6-8
6.2.5 Transmittance Replicates 6-9
6.2.6 Method Blanks 6-10
6.2.7 Stability Samples 6-10
6.2.8 Collimated-Beam Apparatus 6-11
6.3 DOSE-RESPONSE DATA 6-12
6.3.1 Excluded Data 6-12
6.3.2 MS2 Compliance with QC Boundaries 6-12
6.3.3 Uncertainty in Dose Response 6-13
IV
-------
CALCULATION OF THE VALIDATION FACTOR FOR RED AND LOG-
INACTIVATION DESIGN SIZING 7-1
7.1 DISINFECTION CREDIT IN ACCORDANCE WITH CURRENT PROTOCOLS 7-1
7.1.1 Validated Dose (Dv) and Targeted Disinfection 7-1
7.2 DETERMINATION OF THE VALIDATION FACTOR ELEMENTS 7-2
7.2.1 RED Bias (BRED) 7-2
7.2.2 Polychromatic Bias (BPOLY) 7-3
7.2.3 Validation Uncertainty (Uvai) 7-3
7.2.4 Calculation of the Validation Uncertainty (Uvai) 7-5
7.3 CALCULATION OF THE VALIDATION FACTOR 7-5
7.3.1 Validation Factor (VF) 7-5
7.4 VALIDATED RED AND LOG INACTIVATION 7-7
EXAMPLE CALCULATIONS FOR SIZING The Siemens H-4XE-HO (2W-1B-1C) 8-1
8.1 DESIGN CONDITIONS FOR EXAMPLE APPLICATIONS 8-1
8.1.1 Application 1 8-1
8.1.2 Application 2 8-4
REFERENCES 9-1
v
-------
FIGURES
Figure Page
Figure E-l. MS2, Tl and QP RED as a Function of UVT and Flow/Lamp E-4
Figure E-2. Sensor Model Prediction as a Function of UVT E-5
Figure E-3. Algorithm Calculated RED versus Observed RED E-7
Figure E-4. Example Solutions for Validation Factor at Fixed Operating Conditions and a
Range of UV Sensitivities E-8
Figure E-5. Credited RED at 50% UVT Across a Range of UV Sensitivities E-9
Figure E-6. Example Fecal Coliforms Dose-Responsive Curve E-12
Figure E-7. Example Calculation of RED as a Function of Flow (55% UVT) for a H-4XE-HO
(2W-1B-1C) Reactor in a Low-Dose Application E-l3
Figure E-8. Example Calculation of RED as a Function of Flow (55% UVT) for a H-4XE-HO
(2W-IB-1C) Reactor in a Reuse Application E-14
Figure 2-1. Project Organization Chart 2-1
Figure 2-2. Aerial View of the Gloversville-Johnstown Joint Wastewater Treatment Facility
and the UV Center 2-4
Figure 2-3. Aerial View of the UV Center Tanks 2-4
Figure 2-4. General Schematic of the UV Test Facility Showing Major Test Stands (Tank 4
not shown) 2-5
Figure 3-1. Plan and Elevation Schematic of H-4XE-HO (2W-1B-1C) Reactor and Channel
Assembly 3-1
Figure 3-2. Influent (Top) and Effluent (Bottom) Piping Configuration for the H-4XE-HO
(2W-1B-1C) Test Unit Channel 3-4
Figure 3-3. Influent Stilling Plate and Effluent Weir Gate for the Siemens H-4XE-HO (2W-
1B-1QUV Unit Test Channel 3-6
Figure 3-4. Photos of H-4XE-HO (2W-1B-1C) Reactor Installed in Channel 3-7
Figure 3-5. Photo of Unit Power Panel 3-8
Figure 4-1. Process Flow Diagram for the H-4XE-HO (2W-1B-1C) UV System Validation
Test Stand 4-2
Figure 4-2. Velocity Profile Measurement Matrix 4-5
Figure 4-3. HydroQual Collimating Apparatus for Conducting Dose Response Tests 4-7
Figure 5-1. Headloss Through a Single H-4XE-HO (2W-1B-1C) Reactor as a Function of
Flow Rate 5-13
Figure 5-2. Relationship of Sensor Output (mA) and PLC Sensor Reading (%) 5-14
Figure 5-3. Sensor Reading (mA) as a Function of UVT 5-15
Figure 5-4. Velocity Profile at 0.2 mgd 5-17
Figure 5-5. Velocity Profile at 0.8 mgd 5-17
Figure 5-6. An Example of NO determination (09/25/08 Dose-Response Data) 5-22
Figure 5-7. An Example of a Dose-Response Regression Analysis for MS2 5-22
Figure 5-8. MS2 Dose-Response Calibration Curves 5-23
Figure 5-9. Tl Dose-Response Calibration Curves 5-24
Figure 5-10. QP Dose-Response Calibration Curves 5-24
Figure 5-11. Example of Dose-Response Curve-Fit Residuals Analysis 5-25
Figure 5-12. Dose-Response Curve-Fit Uncertainty (UDR) 5-27
Figure 5-13. MS2, Tl and Qp RED as a Function of UVT and Flow 5-30
Figure 5-14. MS2, Tl and Qp RED as a Function of UVT and Flow/Lamp 5-31
Figure 5-15. Algorithm-Calculated RED versus Observed RED 5-33
Figure 5-16. Sensor Model Prediction as a Function of UVT 5-34
VI
-------
Figure 6-1. 12-in. Flow Meter Calibration Data and Correction Formula 6-2
Figure 6-2. Dose-Response Data and NWRI QA/QC Boundary Lines 6-13
Figure 7-1. UVal Decision Tree for Calculated Dose Approach 7-3
Figure 7-2. Example Solutions for Validation Factor at Fixed Operating Conditions and a
Range of UV Sensitivity 7-6
Figure 7-3. Credited RED at 50% UVT Across a Range of UV Sensitivities 7-8
Figure 8-1. Example Fecal Coliforms Dose-Response Curve 8-2
Figure 8-2. Example Calculation of RED as a Function of Flow (55% UVT) 8-4
Figure 8-3. Example Calculation of RED as a Function of Flow (55% UVT) for a H-4XE-HO
(2W-1B-1C) Reactor Module in a Reuse Application 8-5
vn
-------
TABLES
Table
Table E-1. H-4XE-HO (2W-IB-1C) Dose-Algorithm Regression Constants E-6
Table E-2. Credited RED Solutions E-10
Table 4-1 Validation Conditions for Siemens H-4XE-HO (2W-1B-1C) UV System 4-11
Table 5-1. Depth Measurements to Compute Headloss 5-12
Table 5-2. Sensor Intercomparison Variance Analysis 5-16
Table 5-3. Dose-Response Data 5-19
Table 5-4. Summary of Dose-Response Curve Regression Parameters 5-25
Table 5-5. MS2 Biodosimetry Tests: Delivered RED and Operations Data 5-29
Table 5-6. Tl Biodosimetry Tests: Delivered RED and Operations Data 5-29
Table 5-7. Q|3 Biodosimetry Tests: Delivered RED and Operations Data 5-30
Table 5-8. H-4XE-HO (2W-1B-1C) Dose-Algorithm Regression Constants 5-32
Table 6-1. 12-in. Flow Meter Calibration 6-1
Table 6-2. Wavelength and Absorbance Checks 6-3
Table 6-3. Comparison of Dual Radiometer Readings for Collimated Beam
Measurements 6-6
Table 6-4. Reactor Control Sample Summary 6-7
Table 6-5. Similarity between Replicate Flow Test Samples 6-7
Table 6-6. Summary of Reactor Blank and Trip Control Sample Analyses 6-7
Table 6-7. Results from Flow Test Replicates 6-9
Table 6-8. Relative Percent Difference for %T Replicates 6-10
Table 6-9. Phage Stability Sample Summary 6-10
Table 7-1. Credited RED Solutions Error! Bookmark not defined.
Vlll
-------
APPENDICES
APPENDIX A Verification Test Plan for the Siemens Water Technologies V-40R-A150 H and
H-4XE-HO (2W-1B-1C) Open Channel UV Systems for Reuse and Secondary
Effluent Applications, V 2.1.
APPENDIX B Operation and Maintenance Manual for the Siemens Water Technologies H-
4XE-HO (2W-1B-1C) Open Channel UV System.
APPENDIX C Master Data - Siemens Water Technologies H-4XE-HO (2W-1B-1C) Reuse
and Secondary Effluent Testing Program.
IX
-------
GLOSSARY AND DEFINITIONS
Accuracy - A measure of the closeness of an individual measurement or the average of a number
of measurements to the true value and includes random error and systematic error.
Bacteriophage - A virus that has a bacterium as its host organism.
Dose - A total amount of germicidal energy deposited into a solution to be disinfected. Units are
usually mJ/cm2 (millijoules per square centimeter).
Effective disinfection zone - The zone in a disinfection lamp assembly where the UV intensity
deposits a disinfecting dose into the solution. This zone is exclusive of mounting hardware on
the end of the lamp sleeves and the submerged ballasts.
End-of-lamp-life (EOLL) - The UV output condition (i.e. intensity) that is present after the
manufacturers recommended maximum life span for the lamps and the maximum fouling on the
quartz sleeves.
Environmental Technology Verification (ETV) - A program initiated by the USEPA to use
objective, third-party tests to quantitatively verify the function or claims of environmental
technology.
Monochromatic - A light output spectrum that consists solely or dominantly of a single specific
wavelength of light.
pfu - Plaque forming units. A single plaque-forming unit is assumed to represent one viable
MS2 bacteriophage organism.
Polychromatic - A light output spectrum containing many specific wavelengths of light or a
continuous spectrum in a range of wavelengths.
Precision - A measure of the agreement between replicate measurements of the same property
made under similar conditions.
Representativeness - A measure of the degree to which data accurately and precisely represent a
characteristic of a population parameter at a sampling point or for a process conditions or
environmental condition.
Survival Ratio - The logio of the ratio of bacteriophage concentration in a UV dosed solution to
an undosed solution. The values are typically negative numbers because the UV dosing reduces
the number of the viable bacteriophage present in the solution.
Test Element - A series of tests designed by the ETV program to validate a group of related
operational characteristics for a specific technology.
Titer - The specific number of viable organisms (e.g., bacteria or bacteriophage) in a given
volume of solution.
x
-------
Field Testing Organization (FTO) - An organization qualified to conduct studies and testing of
UV disinfection equipment in accordance with the Verification Protocol.
UV Demand - UV energy that does not contribute to disinfection because of absorption by the
chemicals in water.
UV or Ultraviolet Radiation - Light energy with a shorter wavelength than that of visible light
in the range of 190nm to 400 nm.
Vendor - A business that assembles or sells UV Disinfection Technology.
Verification - To establish the evidence on the range of performance of equipment and/or device
under specific conditions following an established protocol(s) and test plan(s).
Verification Protocol - A generic written document that clearly states the objectives, goals, and
scope of the testing under the ETV Program and that establishes the minimum requirements for
verification testing and for the development of a verification test plan. A protocol shall be used
for reference during Manufacturer participation in the verification testing program.
Verification Report - A written document that summarizes a final report reviewed and approved
by NSF on behalf of USEPA or directly by the USEPA.
Verification Test Plan (VTP) - A written document that establishes the detailed test procedures
for verifying the performance of a specific technology. It also defines the roles of the specific
parties involved in the testing and contains instructions for sample and data collection, sample
handling and preservation, and quality assurance and quality control requirements relevant to a
given test site.
XI
-------
ABBREVIATIONS AND ACRONYMS
ANSI
AWWARF
°C
CFD
cm
DVGW
Eff
EOLL
USEPA
ETV
ft
FTO
GAC
G-JJWWTF
gal
gpm
hr
I
in.
Inf
ISO
kW
LI
L/min
log
LPHO
LRCM
LSA
m
mm
mA
mgd
mg/L
ml
mL
mW
nm
NIST
NRMRL
NSF
NTU
NYSERDA
NWRI
American National Standards Institute
American Water works Research Foundation (now WRF)
Degrees Celsius
Computational Fluid Dynamics
Centimeter (10"2 meters)
German Technical and Scientific Association for Gas and Water
Effluent
End-of-lamp-life
United States Environmental Protection Agency
Environmental Technology Verification
Feet
Field Testing Organization
Granular Activated Carbon
Gloversville-Johnstown Joint Wastewater Treatment Facility
gallons
gallons per minute
Hour(s)
Intensity
Inch
Influent
International Standards Organization
Kilowatt
Log Inactivation
Liters per minute
Base 10 logarithm
Low-Pressure, High-Output (type of mercury lamp)
Lamp Rack Controller Module
Lignon Sulfonic Acid (Lignon sulfonate)
Meters
Millimeter (10"3 meters)
Micrometer (10"6 meters)
MilliAmp
Million gallons per day
Milligrams per liter
MilliJoule
Milliliters
MilliWatt
Nanometers (10"9 meters)
National Institute of Standards and Technology
National Risk Management Research Laboratory
NSF International
Nephelometric Turbidity Units
New York State Energy Research and Development Authority
National Water Research Institute
xn
-------
ABBREVIATIONS AND ACRONYMS (continued)
O&M
ORD
OSHA
PDC
pfu
pfu/mL
PLC
ppm
Q
QA
QAPP
QC
QMP
RED
RPD
SAG
SOP
SWP
TYBG
%T
UV
uvc
UVDGM
UVS
UVT
V
VF
VO
VR
VTP
W
WQPC
WRF
WW Protocol
Operation and Maintenance
Office of Research and Development, USEPA
Occupational Safety and Health Administration
Power Distribution Center
Plaque forming units
Plaque forming units per milliliter
Programmable Logic Center
Parts per million
Flow rate
Quality Assurance
Quality Assurance Project Plan
Quality Control
Quality Management Plan
Reduction Equivalent Dose
Relative Percent Difference
Stakeholders Advisory Group
Standard Operating Procedure
Source Water Protection Area, Water Quality Protection Center
Tryptone Yeast Extract Glucose Broth
Transmittance
Ultraviolet
Ultraviolet Radiation in the range of 230nm to 280 nm
Ultraviolet Disinfection Guidance Manual
UV Sensitivity in units of dose per log inactivation
UV Transmittance
Volt
Validation Factor
Verification Organization
Verification Report
Verification Test Plan
Watts
Water Quality Protection Center
Water Research Foundation
Wastewater Validation Protocol
Xlll
-------
EXECUTIVE SUMMARY
E.I VALIDATION PROGRAM
E.I.I Validation Protocols for Reuse Water Disinfection
This report documents the testing, data reduction and analysis in conformance with the
recently developed test protocol, "Validation of UV Reactors for Application to the Disinfection
of Treated Wastewaters" (2008, hereafter, referred to as the WW Protocol), which combines and
updates the objectives and methods found in established UV disinfection guidance documents.
The "Ultraviolet Disinfection Guidelines for Drinking Water and Water Reuse" by the National
Water Research Institute and the American Water Works Association Research Foundation
(NWRI/AwwaRF) (2003) was used as an important guidance for this validation report. The
United States Environmental Protection Agency (USEPA) "UV Disinfection Guidance Manual"
(UVDGM, November 2006), and the "Verification Protocol for Secondary Effluent and Water
Reuse Disinfection Applications" by NSF International and the USEPA under the Environmental
Technology Verification Program (ETV, 2000) were also important references.
E.1.2 Barrier Sunlight H-4XE-HO (2W-1B-1C) UV Disinfection System
The Barrier Sunlight H-4XE-HO (2W-1B-1C) UV disinfection system (H-4XE-HO (2W-
1B-1C)) was tested at full-scale in a 21-ft long channel, including a power supply center and a
main control panel. A perforated baffle plate was positioned downstream of the inlet to simulate
reactor inlet flow conditions that are representative of commercial channel design. An adjustable
weir was installed downstream of the reactor to maintain a constant, prescribed water depth
inside the channel. The reactor contains 16, low-pressure, high-output (LPHO) lamps oriented
horizontally and arranged in symmetrical array four lamps across and four lamps deep. Each
lamp is rated for 171 W total power input. The lamps have an arc length of 59 in., and are
enclosed in 28 mm outer diameter quartz sleeves, 70 in. long. The validation was conducted at
full power input, and the UVT was adjusted to simulate a combined fouling and lamp aging
attenuation of 80%. The reactor was equipped with one duty UV intensity sensor (PW-254),
located 2 cm from the quartz surface of the nearest top lamp. The operating strategy for the H-
4XE-HO (2W-1B-1C) uses lamp banks in series and parallel that are brought into service on
demand, based on flow and water quality (UVT).
E.1.3 Validation Test Stand
The Barrier Sunlight H-4XE-HO (2W-1B-1C) UV disinfection unit was installed in a test
channel at the UV Validation and Research Center of New York (UV Center), located in
Johnstown, NY. The test channel was fed through the facility's 12-in. feed pipe test stands,
serviced by up to eight diesel-powered, centrifugal pumps. Flow direction valves, up- and
downstream in-line static mixers, electromagnetic flow meter, and air-relief valves comprise key
elements of the test stand, in conformance with current validation protocols. A pre-mix injection
E-l
-------
system was connected to the test stream to facilitate the addition of challenge microorganisms
and water modifiers.
E.1.4 Verification Test Claims and Objectives
The overall objective of this ETV was to validate the performance of the Siemens Water
Technologies H-4XE-HO (2W-IB-1C) open channel UV disinfection system at water quality
(UVT) and dose (RED) conditions reflective of secondary effluent and reuse applications. The
total attenuation factor of 80% was selected by Siemens as a combined effect of 90% sleeve
fouling factor and 90% of end-of-lamp-life factor. This attenuation was mimicked by lowering
the test water transmittance. Within this goal, specific objectives were fulfilled. Discussions and
presentation of the results that meet these objectives are presented in this Executive Summary,
and detailed in the Verification Report:
1) Verified the performance difference between power turndown and UVT turndown at the
same operating conditions to mimic the total attenuation factor.
2) Verified the flow-dose relationship for the system at nominal UV transmittances of 50%,
65% and 80% for a dose range of 5 to 25 mJ/cm2 using a biological surrogate with a
relatively high sensitivity to UV (Tl coliphage).
3) Verified the flow-dose relationship for the system at a nominal UV transmittance of 50%,
65% and 80% for a dose range of 10 to 40 mJ/cm2 using a biological surrogate with
medium sensitivity to UV (QP coliphage).
4) Verified the flow-dose relationship for the system at a nominal UV transmittance of 50%,
65% and 80% for a dose range of 20 to 80mJ/cm2 using a biological surrogate with
relatively low sensitivity to UV (MS2 coliphage).
5) Adjusted the observed RED performance results by a validation factor in order to account
for uncertainties associated with the verification tests.
6) Verified the power consumption of the unit.
7) Developed a dose-algorithm to control dose-delivery on a real-time basis, based on the
systems primary operating variables.
E.2 VALIDATION TEST RESULTS
Biodosimetric tests were conducted at a simulated total attenuation factor of 80%,
representing the combined effects of the end-of-lamp-life (EOLL) factor and the fouling factor.
The total attenuation factor for the Siemens H-4XE-HO (2W-1B-1C) system was simulated by
lowering the water transmittance. For the three nominal UVT values tested for this validation,
E-2
-------
80%, 65%, and 50%, the actual UVT levels that were needed to include simulation of the 80%
sensor attenuation were 74.5%, 60.4% and 45.8%, respectively.
E.2.1. Biodosimetric Assay Results
A total of 31 acceptable flow tests were conducted for this ETV. Three different
coliphage were used as the challenge organisms: MS2, Q|3 and Tl. The reported reduction
equivalent dose (RED) is based upon the dose-response curve for the collimated beam data from
the same day. The biodosimetric RED data are presented as a function of flow per lamp in
Figure E-l for each challenge phage at their respective nominal UVT levels. The bounds
described by these data represent the validated operating envelope for the UV system:
Flow: 134to866gpm
Flow/lamp 8.37 to 54.14 gpm/lamp
UVT: 50 to 80%
Power: 100 atPLC, or 100% input (2.75 kW/16 lamps, or 171 W/lamp)
E.2.2 Technical Test Results
E.2.2.1 Power Consumption
The power consumption of the Siemens H-4XE-HO (2W-1B-1C) system was
continuously logged when operating. The mean total power input was 2.75 kW, or 171 W/lamp.
E.2.2.2Headloss
Headloss estimates were derived from the hydraulic profile data. Two sample locations
(immediately before and after the unit) were used at eight different flow rates. The influent
depth was held constant by adjusting the downstream weir height. The headloss for the unit can
be estimated from the expression:
Headloss (in. of water} = 3.160 (flow, mgd)2 - 0.938 (flow, mgd) + 0.148
E-3
-------
IOU
1RD
-i /in
itu
ion
cT- U(J
^ mn
^ I UU
E
** pn
Q OU
UJ
Qi en
DU
/in
'tU
on
zu
n
A
4
A
*
A
A
^
<
<
ft
ft
9
1
MS2, Nominal UVT = 80%
MS2, Nominal UVT = 65%
A MS2, Nominal UVT = 50%
T1, Nominal UVT = 50%
QB, Nominal UVT = 80%
*QB, Nominal UVT = 65%
QB, Nominal UVT = 50%
*
0
10 15 20 25 30 35 40 45 50 55 60 65 70
Flow Rate per Lamp (gpm/Lamp)
Figure E-l. MS2, Tl and QP RED as a function of UVT and Flow/Lamp.
E.2.2.3 Velocity Profiles
Cross-sectional velocity measurements were taken at 0.2 (138 gpm) and 0.8 mgd (561
gpm). This maximum is less than the maximum flow tested biodosimerically, and in some cases
may establish the maximum flow rate/lamp. This will depend on guidance relevant to certain
applications (such as in reuse, based on NWRI guidance). Per guidance in the
NWRI/AWWARF Ultraviolet Disinfection Guidelines for Drinking Water and Water Reuse
(2003), the mean velocity at any measured cross-sectional point of a commissioned system
should not vary by more than 20% from the theoretical average velocity (i.e., flow divided by the
cross-sectional area). Further, the commissioned system should exhibit velocity profiles that are
equivalent or better than those exhibited by the validated test unit. This is particularly important
if there is scale-up from the test unit.
The full record of velocity measurements is compiled in Appendix C. Overall, a general
observation is that the velocity profiles were variable at 0.8 mgd. The effluent measurements
tended to be outside the targeted range. The influent measurements a 0.8 mgd were fairly stable,
less variable than the effluent. At 0.2 mgd, velocity profiles were more stable, although the
influent was outside of the variability guideline at the surface. A key observation that can be
E-4
-------
made from these data is that the hydraulic conditions represent a 'worse' case when compared to
minimum full-scale commissioning requirements. As such, the biodosimetry performance data
can be considered conservative.
E.2.2.4 Sensor Model
When commissioned, it is necessary to assure that the same sensor position is maintained
and the same readings are obtained at given operating conditions. To assist with this objective,
sensor measurements were analyzed and a sensor model developed to allow prediction of the
sensor reading in a commissioned system:
3.169
S = 0.000082645 (UVT254)
Where:
S = Sensor reading (%)
UVT254 = UV transmittance at 254nm (%/cm)
Figure E-2 presents the model predictions as a function of the UVT. These data are at a
power setting, P, of 100, which is the normal operating condition for the H-4XE-HO (2W-1B-
1C). As shown, there is good agreement, providing a tool to assess the sensor position and
function for a commissioned system.
100
90
80
2 70
D
Q.
4-i
o
!_
O
(A
C
0)
CO
60
50
40
30
20
10
y = 8.2645E-05x31690
40 45 50 55 60 65
UVT (%/cm)
70
75
80
85
Figure E-2. Sensor model prediction as a function of UVT.
E-5
-------
E.3 CREDITED DOSE-DELIVERY PERFORMANCE
E.3.1 RED Performance Algorithm
A dose algorithm was developed to correlate the observed MS2, Tl and QP RED data
with the reactor's primary operating variables. These are the flow rate per lamp, Q/L, and sensor
reading, S (a function of the lamp output and the UVT). These variables are known on a real-
time basis by the PLC and can be programmed into software to monitor and control the UV
system. Because multiple surrogates were used to test the system, the test results can be
combined to incorporate the sensitivity of each to differentiate their individual reactions at the
specified operating conditions. The commissioned system can then incorporate the sensitivity of
the targeted pathogen (e.g., total or fecal coliform, enterococcus, etc.) when calculating the RED
delivered by the system. The dose algorithm to estimate the RED is:
RED = Wa (Q/L)b Sc UVSd
Where;
Q = Flow rate, gpm
L = Number of lamps
S = Sensor Reading (%)
UVS = UV Sensitivity (mJ/cm2/Log Inactivation)
a, b, c, d = Equation coefficients
The same sensors and their installed conditions, such as model type, position relative to
the lamp, sleeve clarity, etc., must be used to apply this algorithm. Based on a multiple linear
regression analysis in the form of this RED equation, the coefficients were determined and are
summarized in Table E-l. The algorithm-calculated REDs versus the observed MS2, Tl and QP
REDs are plotted in Figure E-3; good agreement is observed between the predicted and observed
RED.
Table E-l. H-4XE-HO (2W-1B-1C) Dose-Algorithm Regression Constants
Coefficient
a
b
c
d
Value
0.950550
-0.609884
0.683241
0.398391
E-6
-------
120
110
100
90
£ 70
Q
m
ra
O
60
30
20
10
10 20 30 40
50 60
70
80 90 100 110 120
Observed RED (mJ/cm2)
Figure E-3. Algorithm Calculated RED versus Observed RED.
E.3.2 Validation Factor
The VF components BRED, BPOLY and Uvai were assessed. The RED bias, BRED, can be
set at 1.0 as long as the sensitivity of the targeted pathogen or pathogen indicator is within the
range of 5 and 20 mJ/cm2/LI, and the sensitivity used in the RED algorithm is equal to or less
than the sensitivity of the targeted microbe. BPOLY is set to 1.0 because the system uses low-
pressure monochromatic lamps.
Within the uncertainty of validation, Uvai, the uncertainties associated with the sensors
(Us) and the collimated beam tests (UDR) can be ignored because QA criteria were met, leaving
only the uncertainty of interpolation, UIN. With its specific elements assessed and defined, the
validation factor for the Barrier Sunlight H-4XE-HO (2W-1B-1C) can be expressed as a function
of the UIN, which reduces to the following expression as a function of the calculated RED:
VF = 1 + (6.017/REDcak)
E-7
-------
Figure E-4 presents a series of solutions for VF at a UVT of 50% and sensitivities
ranging between five and 20 mJ/cm2/LI. VF is shown as a function of flow/lamp under these
specific and fixed operating conditions. Similar calculations can be made at alternate operating
conditions. These calculations are appropriate only when the UVS of the targeted pathogen is
equal to or greater than the sensitivity chosen for the calculations. Thus, if the sensitivity of the
organism of concern is 10 mJ/cm2/LI, then UVS must be 10 or less when conducting the
calculations for the VF. If this is not the case, then an RED bias term, similar to that described
by the UVDGM, would have to be incorporated into the validation factor.
Validation Factor at 50% UVT
UVS = 5 mJ/cm2/LI
UVS = 8 mJ/cm2/LI
UVS = 11 mJ/cm2/LI
UVS = 15mJ/cm2/LI
UVS = 20 mJ/cm2/LI
10 15 20 25 30 35 40 45 50
Flow Rate per Lamp (gpm/Lamp)
55 60 65 70
Figure E-4. Example solutions for Validation Factor at fixed operating conditions and a
range of UV sensitivities.
E.3.3 Credited RED Calculation
Given the validation RED results and the estimate of uncertainty associated with the
experimental effort, one can determine the RED that can be applied, or credited, to the system at
prescribed operating conditions. This credited, or validated RED (REDyai), is calculated as:
VF
Figure E-5 presents solutions for the H-4XE-HO (2W-1B-1C) at a UVT of 50%, across
the same range of UV sensitivities. It is important to note that this assumes the system sensors
have been confirmed to have the same output as in the validation. The solutions for validated, or
credited, RED (REDVai), such as those shown on Figure E-5, would be reported at the PLC of the
Barrier Sunlight H-4XE-HO (2W-1B-1C), based on monitored real-time operating conditions.
Eo
-O
-------
UVS = 5 mJ/cm2/LI
UVS = 8 mJ/cm2/LI
UVS = 11 mJ/cm2/LI
UVS = 15mJ/cm2/LI
UVS = 20 mJ/cm2/LI
Validated RED at 50% UVT
'S
o
Q
LU
o:
a
£
(0
2
"re
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70
Flow Rate per Lamp (gpm/Lamp)
Figure E-5. Credited RED at 50% UVT Across a Range of UV Sensitivities
Table E-2 provides credited RED solutions across a broad range of operating conditions
for the H-4XE-HO (2W-1B-1C) at sensitivities between 5 and 20 mJ/cm2/LI. Figure E-5
displayed those calculations pertinent to the 50% UVT conditions; similar graphical plots can be
generated by the user at alternate conditions.
E-9
-------
Table E-2. Credited RED Solutions
UVT
(%T)
50
50
50
50
50
50
50
50
50
50
50
55
55
55
55
55
55
55
55
55
55
55
60
60
60
60
60
60
60
60
60
60
60
65
65
65
65
65
65
65
65
65
65
65
S
(%)
20.0
20.0
20.0
20.0
20.0
20.0
20.0
20.0
20.0
20.0
20.0
27.
27.
27.
27.
27.
27.
27.
27.
27.
27.
27.
35.7
35.7
35.7
35.7
35.7
35.7
35.7
35.7
35.7
35.7
35.7
46.0
46.0
46.0
46.0
46.0
46.0
46.0
46.0
46.0
46.0
46.0
Q/L
(gpm/L)
8.4
10.0
14.0
18.0
25.0
30.0
35.0
40.0
45.0
50.0
54.0
8.4
10.0
14.0
18.0
25.0
30.0
35.0
40.0
45.0
50.0
54.0
8.4
10.0
14.0
18.0
25.0
30.0
35.0
40.0
45.0
50.0
54.0
8.4
10.0
14.0
18.0
25.0
30.0
35.0
40.0
45.0
50.0
54.0
Credited RED (mJ/cm2) at UVS (mJ/cm2/LI)
5
30.7
27.2
21.4
17.8
13.9
12.1
10.7
9.6
8.8
8.1
7.6
38.8
34.4
27.2
22.7
17.9
15.6
13.9
12.6
11.5
10.6
9.9
47.8
42.5
33.7
28.3
22.4
19.6
17.5
15.9
14.5
13.4
12.6
57.8
51.4
41.0
34.5
27.4
24.1
21.6
19.6
18.0
16.6
15.7
8
37.9
33.7
26.6
22.2
17.5
15.3
13.6
12.3
11.2
10.3
9.7
47.7
42.4
33.7
28.3
22.4
19.6
17.5
15.9
14.5
13.4
12.6
58.6
52.2
41.6
35.1
27.9
24.5
21.9
19.9
18.3
16.9
16.0
70.7
63.1
50.4
42.6
34.0
30.0
26.9
24.5
22.5
20.8
19.7
11
43.7
38.8
30.8
25.8
20.4
17.8
15.9
14.4
13.1
12.1
11.4
54.8
48.8
38.9
32.7
26.0
22.8
20.4
18.5
17.0
15.7
14.8
67.3
60.0
47.9
40.4
32.3
28.4
25.5
23.2
21.3
19.7
18.6
81.0
72.3
57.9
49.0
39.2
34.6
31.1
28.3
26.1
24.2
22.9
15
50.1
44.6
35.4
29.7
23.6
20.7
18.5
16.7
15.3
14.2
13.3
62.7
55.9
44.6
37.6
30.0
26.3
23.6
21.4
19.7
18.2
17.2
76.8
68.5
54.9
46.4
37.1
32.7
29.4
26.7
24.6
22.8
21.6
92.3
82.5
66.2
56.1
45.0
39.8
35.8
32.6
30.1
27.9
26.4
20
56.8
50.5
40.3
33.9
27.0
23.7
21.2
19.2
17.6
16.3
15.4
71.0
63.3
50.6
42.7
34.1
30.1
27.0
24.5
22.6
20.9
19.8
86.8
77.5
62.1
52.6
42.2
37.2
33.5
30.5
28.1
26.1
24.7
104.2
93.1
74.9
63.5
51.1
45.2
40.7
37.2
34.3
31.8
30.2
E-10
-------
Table E-2. Credited RED Solutions (Continued)
UVT S
Q/L
(%T) (%) (gpm/L)
70 58.
70 58.
70 58.
70 58.
70 58.
70 58.
70 58.
70 58.
70 58.
70 58.
70 58.
8.4
10
14
18
25
30
35
40
45
50
54
75 72.3 8.4
75 72.3 10
75 72.3 14
75 72.3 18
75 72.:
75 72.:
75 72.:
75 72.:
75 72.:
75 72.:
75 72.:
5 25
I 30
I 35
I 40
I 45
I 50
5 54
80 88.7 8.4
80 88.7 10
80 88.7 14
80 88.7 18
80 88.7 25
80 88.7 30
80 88.7 35
80 88.7 40
80 88.7 45
80 88.7 50
80 88.7 54
Credited RED (mJ/cm2) at UVS (mJ/cm2/LI)
5
68.7
61.3
49.0
41.3
33.0
29.1
26.1
23.7
21.8
20.2
19.1
80.6
71.9
57.7
48.8
39.0
34.4
31.0
28.2
25.9
24.1
22.8
93.5
83.5
67.1
56.8
45.6
40.3
36.3
33.1
30.5
28.3
26.8
8
83.
74.
60.
50.
40.
36.
32.
29.
27.
25.
23.
98.
87.
70.
59.
48.
42.
38.
34.
32.
29.
28.
113
101
81.
69.
56.
49.
44.
40.
37.
35.
33.
9
9
1
8
7
0
3
4
1
2
8
3
8
6
8
1
5
3
9
2
9
3
.9
.8
9
6
0
6
7
8
7
1
2
11
96.0
85.8
68.9
58.4
46.9
41.4
37.3
34.0
31.4
29.1
27.6
112.3
100.4
80.8
68.6
55.2
48.9
44.1
40.3
37.1
34.5
32.8
130.0
116.3
93.7
79.7
64.3
57.0
51.4
47.0
43.4
40.4
38.4
15
109.3
97.8
78.6
66.7
53.7
47.5
42.8
39.1
36.1
33.6
31.8
127.8
114.4
92.2
78.3
63.2
56.0
50.5
46.2
42.6
39.7
37.7
147.8
132.3
106.8
90.8
73.4
65.1
58.8
53.8
49.8
46.4
44.0
20
123.3
110.3
88.8
75.5
60.8
53.9
48.6
44.4
41.0
38.2
36.2
144.0
128.9
104.0
88.5
71.4
63.4
57.2
52.4
48.4
45.1
42.8
166.5
149.1
120.4
102.5
82.9
73.7
66.6
61.0
56.4
52.6
50.0
E.4 EXAMPLE CALCULATIONS FOR SIZING THE SIEMENS H-4XE-HO (2W-1B-
1C)
An example is given to illustrate the calculations that can be conducted to evaluate the
sizing of the Siemens H-4XE-HO (2W-1B-1C). Consider the following design condition:
Flow Rate:
6000 gpm (8.64 mgd)
E-ll
-------
UVT: 55%
Performance Requirement with two reactors in series:
Application 1: Secondary Effluent, Fecal Coliform < 200 cfu/100 mL (2.3 Log)
Application 2: Reuse, MS2 Dose > 80 ml/cm2
E.4.1 Application 1
This is a "low-dose" application, directed at typical secondary effluents discharged from
wastewater treatment plants. In such cases, collimated-beam measurements would be made to
develop a dose-response (DR) relationship, based on fecal coliform. An example of such data is
provided in Figure E-6, showing the tailing effect due to particulates. Taking the non-
aggregated, linear portion of the curve, the UV sensitivity is estimated to be 6.9 mJ/cm2/LI.
From the DR data, one can observe that the maximum effective dose is in the vicinity of 25
mJ/cm2, beyond which the particulate coliform control and little apparent disinfection occurs. In
order to meet the specification, a lower (than the permit maximum) target fecal coliform is
considered; and the dose is set at 25 mJ/cm2.
6.00 i
E
i 5.00
^
a
~ 4.00
E
| 3.00
o
"ro
« 2.00
1.00
in
9)
0.00
K. Example - Fecal Coliform Dose Response
NsJ
>v Non-Aggregated, Linear Portion
^f.
-------
unit. The UVS in this case is 6.9 mJ/cm2/LI, as shown on Figure E-6 for the site-specific fecal
coliform. The flow/lamp input can be varied to evaluate REDcaic as a function of hydraulic
loading.
Figure E-7 presents solutions for REDcaic as a function of flow/lamp. These must then be
adjusted for the Validation Factor (VF), yielding the validated, or credited, RED. Solutions for
validated RED are also shown on Figure E-7. As shown, a hydraulic loading equal to or less
than 49.7 gpm/lamp will yield an RED of at least 12.5 ml/cm2 in a single reactor; two would be
placed in series for a credited RED of 25 ml/cm2. At the design flow of 6000 gpm, each bank
would require 6000/49.7 = 120 lamps. Note that this analysis is simplified as an example, and
does not address redundancy or other design considerations.
60
55
50
45
u
HI
Calculated Dose
Credited Dose
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70
Flow Rate/Lamp (gpm/Lamp)
Figure E-7. Example calculation of RED as a function of flow (55% UVT)
for an H-4XE-HO (2W-1B-1C) reactor in a low-dose application.
E.4.2 Application 2
In the second application, the performance requirement is to meet an MS2 RED of 80
ml/cm2, a criterion typically found with reuse applications after membrane-filtered secondary
treatment. The approach is the same as discussed above for the "low-dose" application, except
that an MS2 UV sensitivity value is used. This is 20 mJ/cm2/LI. Solutions for calculated and
E-13
-------
credited RED are provided in Figure E-8. In this case, two reactors are placed in series, with a
hydraulic loading of 19.85 gpm/lamp. To meet the design flow of 6000 gpm, approximately 303
lamps are needed for each reactor. This would likely be divided to 2 or three parallel channels.
Note that this is provided as a simplified example; other design aspects such as redundancy are
not considered.
Calculated Dose
Credited Dose
10 15 20 25 30 35 40 45 50 55 60 65 70
Flow Rate per Lamp (gpm/Lamp)
Figure E-8. Example calculation of RED as a function of flow (55% UVT) for a H-4XE-
HO (2W-1B-1C) reactor in a reuse application.
E-14
-------
SECTION 1
INTRODUCTION AND BACKGROUND
1.1 THE ETV PROGRAM
1.1.1 Concept of the ETV Program
The Environmental Technology Verification (ETV) program was created to accelerate
the development and commercialization of environmental technologies through third party
verification and reporting of performance. The goal of the ETV program is to verify
performance characteristics of commercial-ready environmental technologies through the
evaluation of objective and quality assured data so that potential buyers and regulators are
provided with an independent and credible assessment of the technology that they are buying or
permitting.
1.1.2 The ETV Program for Water Reuse and Secondary Effluent Disinfection
This report documents the testing, data reduction and analysis in conformance with the
recently developed test protocol, "Validation of UV Reactors for Application to the Disinfection
of Treated Wastewaters" (2008, hereafter, referred to as the WW Protocol), which combines and
updates the objectives and methods found in established UV disinfection guidance documents.
The "Ultraviolet Disinfection Guidelines for Drinking Water and Water Reuse" by the National
Water Research Institute and the American Water Works Association Research Foundation
(NWRI/AwwaRF) (2003) was used as an important guidance for this validation report. The
United States Environmental Protection Agency (USEPA) "UV Disinfection Guidance Manual"
(UVDGM, November 2006), and the "Verification Protocol for Secondary Effluent and Water
Reuse Disinfection Applications" by NSF International and the USEPA under the Environmental
Technology Verification Program (ETV, 2000) were also important references.
The WW Protocol provides general guidance on the validation of the performance of
commercial UV systems, but is not application-specific, such as for reuse, secondary effluent, or
wet weather flows, as categorized in previously used verification protocols. Instead, a vendor
chooses to conduct validations covering a range of operating conditions (i.e., operating
"envelope") and dose levels to meet their marketing expectations regarding the application of
their respective UV systems.
1-1
-------
1.1.3 The Siemens Water Technologies ETV
This ETV of the Siemens Water Technologies H-4XE-HO (2W-1B-1C) UV disinfection
unit focused on dose-delivery verification at water UV transmittances between 50% and 80%.
The total intensity attenuation factor was 80%, as set by Siemens based on the combined effects
of a sleeve-fouling factor of 90% and lamp aging-factor (end-of-lamp-life factor, or EOLL) of
90%. The unit was operated at full power input under all conditions, and the flow ranged
between 134 and 866 gpm. Biodosimetric testing was accomplished with three test organisms:
coliphages MS2, Tl and Qp.
1.2 MECHANISM OF UV DISINFECTION
Ultraviolet (UV) light radiation is a widely accepted method for accomplishing
disinfection of treated wastewaters. Its germicidal action is attributed to its ability to
photochemically damage links in the DNA molecules of a cell, which prevents the future
replication of the cell, effectively "inactivating" the microorganism. UV radiation is most
effective in the region of the electromagnetic spectrum between 230 and 290 nm (referred to as
the UVC range); this corresponds to the UV absorbance spectrum of nucleic acids. The optimum
germicidal wavelengths are in the range of 255 to 265 nm.
1.2.1 Practical Application of UV Disinfection
The dominant commercial source of UV light for germicidal applications is the mercury
vapor, electric discharge lamp. These are commercially available in "low-pressure" and
"medium-pressure" configurations. The conventional low-pressure lamp operates at 0.007 mm
Hg, and is typically supplied in long lengths (0.75 to 1.5 m), with diameters between 1.5 and 2
cm. The major advantages of the low-pressure lamp are that its UV output is essentially
monochromatic at a wavelength of 254 nm, and it is energy efficient, converting approximately
35 to 38 percent of its input energy to UV light at the 254 nm wavelength. The UV power output
of a conventional low-pressure lamp is relatively low, typically about 25 W at 254 nm for a 70 to
75 W, 1.47-m long lamp. Low-pressure, high-output (LPHO) lamps (-0.76 mm of Hg) have also
been developed using mercury in the form of an amalgam and/or higher current discharges.
LPHO lamps are very similar in appearance to the conventional low-pressure lamps, but have
power outputs 1.5 to five times higher, reducing the required number of lamps for a given
application. LPHO lamps have approximately the same efficiency of conventional low-pressure
lamps.
Medium-pressure lamps operate between 300 to 30,000 mm of Hg, and can have many
times the total UVC output of a low-pressure lamp. Such medium-pressure lamps emit
1-2
-------
polychromatic light, and convert between 10 to 20 percent of the input energy to germicidal UV
radiation, resulting in lower efficiency. However, the sum of all the spectral lines in the UVC
region for a medium-pressure lamp results in three to four times the germicidal output when
compared to low-pressure lamps. Because of the very high UV output rates, fewer medium-
pressure lamps are needed for a given application when compared to low-pressure lamps.
Both low- and medium-pressure germicidal lamps are sheathed in quartz sleeves,
configured in geometric arrays, and placed directly in the wastewater stream. The lamp systems
are typically modular in design, oriented horizontally or vertically, mounted parallel or
perpendicular to flow, and assembled in single or multiple channels and/or reactors.
The key design consideration is directed to efficient delivery of the germicidal UV
energy to the wastewater and to the organisms. The total germicidal effectiveness is quantified
as the "UV dose," or the product of the UV radiation intensity (/, Watts/cm2) and the exposure
time (t, seconds) experienced by a population of organisms. The effective intensity of the
radiation is a function of the lamp output, and of the factors that attenuate the energy as it is
deposited into the water. Such attenuating factors include simple geometric dispersion of the
energy as it moves away from the source, absorbance of the energy by the quartz sleeve housing
the lamp, and the UV absorbance, or UV demand, of the energy by constituents in the
wastewater.
1.2.2 A Comparison of UV and Chemical Disinfection
UV disinfection uses electromagnetic energy as the germicidal agent, differing
considerably from chemical disinfection agents such as chlorine or ozone. The lethal effect of
UV radiation is manifested by the organism's inability to replicate, whereas chemical
disinfection physically destroys the integrity of the organism via oxidation processes.
Germicidal UV radiation does not produce significant residuals, whereas chemical disinfection
results in residuals that may exist long after the required disinfection is complete. Chemical
residuals, such as chlorine or chloramines, may then have a detrimental effect on organisms in
the natural water system to which the effluent is released. An additional, subsequent process,
such as dechlorination, usually ameliorates this detrimental result. This residual effect does not
exist for UV disinfection processes.
Chemical disinfection involves shipping, handling, and storing potentially dangerous
chemicals. In contrast, dangers associated with UV disinfection are minimal. A UV disinfection
system produces high-intensity UVC radiation, which can cause eye damage and skin burns upon
exposure; however, these dangers are easily prevented with protective clothing and goggles, and
by properly enclosing or shielding the UV system. A minor hazard exists because the lamps
1-3
-------
contain very small amounts of liquid or amalgamated mercury requiring that lamps be disposed
of properly. The primary cost associated with operating UV disinfection systems is the
continuous use of significant amounts of electrical power, and routine maintenance, whereas
chemical generation and use is the primary operating expense for chemical disinfection systems.
1.2.3 Determining Dose Delivery
In theory, the delivery of UV radiation to a wastewater can be computed mathematically
if the geometry and hydraulic behavior of the system are well characterized. Ideally, all
elements entering the reactor should be exposed to all levels of radiation for the same amount of
time, a condition described as turbulent, ideal plug flow. In fact, non-ideal conditions exist;
there is a distribution of residence times in the reactor due to advective dispersion and to mixing
in the reactor. The degree to which the reactor strays from ideal plug flow will directly impact
the efficiency of dose delivery in the system. Similarly, the intensity field in the reactor is
variable, a function of the lamp output and spacing, and the UV absorbance of the liquid.
Together, these aspects of UV reactor behavior dictate that some particles (microorganisms) will
receive small UV doses, while other particles will receive larger doses. More generally, it can be
asserted that all UV reactors that are used for water and wastewater treatment in practical
applications are characterized by a UV dose distribution for any given operating condition.
Accurate predictions of UV reactor performance can be developed by integrating the UV
dose distribution with the intrinsic kinetics of the reaction(s) of interest (aha., UV dose-response
behavior). However, the validity of any such prediction relies on the validity of the dose
distribution estimate, as well as the validity of the dose-response information. Purely numerical
simulations were a natural evolution of this modeling approach. These simulations involve
combined applications of computational fluid dynamics (CFD) with intensity field (I) models.
Indeed, CFD-I models have evolved to the point where, in some cases, they now form the basis
for design of new reactors. Manufacturers of UV systems have found that numerical prototyping
is less expensive than physical prototyping, particularly as a means of optimizing reactor
performance for a given application.
While numerical models, such as CFD-I, represent important tools for analysis of UV
reactors, they have not evolved to the point where they can be used for reactor validation.
Several issues can be identified that prevent the application of CFD-I models for validation:
there is no uniform standard for their application; one can expect considerable uncertainty in the
values of some important input variables (e.g., lamp output power); and the models themselves
may ignore or incompletely account for some relevant physical behavior (e.g., reflection and
1-4
-------
refraction of UV radiation). Collectively, these and other factors mean that CFD-I models are
developing, but still need a basis for verification.
Lagrangian actinometry (LA) using dyed microspheres was developed as a method for
direct measurement of the UV dose distribution delivered by a UV reactor for a given set of
operating conditions. Microspheres coated with a photosensitive dye are pass through a reactor,
with each particle fluorescing in proportion to the dose received in its individual trajectory. In
other words, the method allows for dose measurement at the level of an individual particle. This
is an emerging tool that will likely be available for direct validation of a reactor's log
inactivation performance for any pathogen of known dose-response behavior.
1.2.4 Summary of the Biodosimetric Method to Measure Dose
Current practice uses biodosimetric techniques to assess the dose-delivery performance of
UV reactors, whereby the inactivation of a surrogate challenge organism through a reactor is
measured and compared to its dose-response behavior. This results in an estimate of the
reduction equivalent dose (RED) delivered by the UV reactor. The UVDGM presents these
biodosimetric techniques as current state-of-the-art, but recognizes the uncertainties associated
with the selection and analysis of microbiological surrogates, and the potential for widely
divergent dose-distribution characteristics. In order to mitigate the potential impact of such
uncertainties, significant adjustments relating to these uncertainties are made to the observed
RED before a credited inactivation is awarded to a specific reactor installation.
Biodosimetry is a method for determining the germicidal dose delivery to a wastewater
by using an actual calibrated test organism. Put simply, the survival ratio of the organism is
calibrated to a well-controlled UV dose in the laboratory with a dose-response procedure. The
same organisms are then used to field test the actual disinfection system under specified
conditions. Such field tests generate a survival ratio of the organism under specified test
conditions, which can then be converted into an effective delivered dose through the dose-
response calibration curve. This is termed the reduction equivalent dose, or RED, with units of
mJ/cm2. For the tests in this ETV, the bacteriophages MS2, Tl, and QP were used.
The advantages to the biodosimetric method are that the organism records the actual
germicidal dose, the organism can be produced in such large quantities that every milliliter of
test solution contains a statistically significant number of organisms, and there are no
assumptions about the hydraulic behavior or intensity field of the reactor. It is important to
remember that this method is not used to determine the effective germicidal UV dose for any
specific pathogen, it is a method to quantify germicidal dose delivery for a specific microbe.
1-5
-------
SECTION 2
ROLES AND RESPONSIBILITIES OF PARTICIPANTS IN
THE VERIFICATION TESTING
2.1 NSF INTERNATIONAL (NSF)
The Project Organization Chart is provided in Figure 2-1. The ETV Water Quality
Protection Center (WQPC) is administered through a cooperative agreement between the
USEPA and NSF International (NSF). NSF administers the program through the WQPC and
selected a qualified Field Testing Organization (FTO). HydroQual, Inc. (HydroQual) developed
and implemented the Verification Test Plan (VTP) for this ETV. NSF's project responsibilities
included review and approval of the VTP, QA oversight, peer reviews, report approval and
preparation and dissemination of the verification statement.
Operations
Jermey Hill
Lead Field Technician
USEPA
Ray Frederick
NSF International
Thomas Stevens
HydroQual
Director
O. Karl Scheible
UV Center
Chengyue Slien
Project Manager
Field Testing
Chistopher Groth
Engineer II
Siemens Water Technologies |
Ber trand Dussert
HydroQual Laboratory
(Microbiology)
Prakash Patil
Figure 2-1. Project organization chart.
2-1
-------
The key contact at NSF relating to this report is:
Mr. Thomas Stevens, Center Manager
NSF International
789 Dixboro Road
Ann Arbor, MI 48113
(734) 769-5347
stevenst@nsf.org
2.2 U.S. ENVIRONMENTAL PROTECTION AGENCY (USEPA)
The USEPA's National Risk Management Research Laboratory provides administrative,
technical and quality assurance guidance and oversight on all WQPC activities. The USEPA has
review and approval responsibilities through various phases of the verification project. The key
USEPA contact is:
Mr. Ray Frederick
USEPA - NRML Urban Watershed Management Branch
2890 Woodbridge Avenue (MS-104)
Edison, NJ 08837-3679
(732)321-6627
(732) 321-6640 (fax)
Frederick.ray@epa.gov
2.3 FIELD TESTING ORGANIZATION (FTO), HYDROQUAL, INC.
The selected FTO was HydroQual, Inc, which has a well-established expertise in the area
of ultraviolet disinfection technologies. Mr. O. Karl Scheible, Project Director, provided overall
technical guidance for the verification test program. Dr. Chengyue Shen, PE served as the
Project Manager, responsible for day-to-day operations and technical analysis. Dr. Prakash Patil
was the project microbiologist, responsible for all bacteriophage stock preparation and sample
analyses, including collimated beam testing. HydroQual also provided additional in-house staff
as required. HydroQual's responsibilities included development of the VTP, management of the
testing effort, compilation and analysis of the data, and preparation of the verification report.
HydroQual's main office is located in Mahwah, New Jersey and has a staff of over 110. The
mailing address is:
HydroQual, Inc.
1200MacArthurBlvd
Mahwah, New Jersey 07430
(201)529-5151
2-2
-------
(201) 529-5728 (fax)
http ://www.hydroqual. com
Dr. Shen was the primary technical contact person at HydroQual:
Telephone extension: 7191, or
Cell phone: (201) 538-6820, or
Email: cshen@hydroqual.com
Mr. Scheible can be reached at extension 7178 or
Email: kscheible@hydroqual.com
2.4 VALIDATION TEST FACILITY
The ETV tests were conducted at the UV Validation and Research Center of New York
(UV Center), Johnstown, NY, which is operated exclusively by HydroQual. The UV Center was
installed at the wastewater treatment plant site with the support of the New York State Energy
Research and Development Authority (NYSERDA), with direct participation by a number of
manufacturers, including Siemens Water Technologies Corp. Active testing at the validation
facility has been underway since June 2003. The UV Center address is:
HydroQual, Inc.
c/o Gloversville-Johnstown Joint Wastewater Treatment Facility
191 Union Ave Extension
Johnstown, New York 12095
HydroQual On-Site Contact: William Pearson, (646) 235-0382
Figure 2-2 is an aerial view of the Gloversville-Johnstown Joint Wastewater Treatment
Facility. The location of the Test Facility within the plant is circled. Figure 2-3 shows an aerial
view of the tanks and pumps at the UV Center. Up to eight 5500-gpm diesel-powered
centrifugal pumps are available to feed the test systems. The accumulated effluent is slowly
pumped (at rates up to 1500 gpm) back into the wastewater treatment plant for final disposal.
Filtered, high-quality potable water from a surface water supply (90 to 97% UVT at 254 nm) is
provided via a local hydrant. The water is dechlorinated with sodium sulfite. In cases when
higher transmittance waters are needed, the UV Center has granular activated carbon (GAC)
units to polish (and dechlorinate) the water. The UV Center also has access to treated secondary
effluent, which is filtered through 20-micron cloth cartridge filters when filling the source water
tanks.
2-3
-------
Figure 2-2. Aerial view of the Gloversville-Johnstown Joint Wastewater Treatment
Facility and the UV Center.
Reactor Stands
72=24 Inch Piping
30'X 50'Shelter
Source Water Tanks
\ M2.2 Mg
Disposal Tanks
k_|2.2Mgal)
-Control Room
(Pumps, Valves,
Injection)
Pump Galley
(8 x 5500 gpm)
In-line Mixer
I/E, Each Stand
Figure 2-3. Aerial view of the UV Center tanks.
2-4
-------
Figure 2-4 is a general schematic of the test facility. A number of test stands are
available within the facility, ranging from 2-in. to 48-in. diameter feed piping. Typically, test
stands are assembled from these piping systems to accommodate the reactor and preferred
inlet/outlet piping configurations. Flow rate capacities range from five to 45000 gpm. The
facility employs several large concrete tanks that are used to prepare source water for challenge
testing, or to accept testing effluent. The general placement of the two Siemens test units is
shown in Figure 2-4.
12" TEST STANDS
CLEAN WATER
STAGING TANK
0.75 MGAL
CONTROL
STATION
WITH LSA
AND MS2
INJECTION
LOOP
SHELTER
21,' TEST STANDS
PJMP AND BANK MANIFOLD
36" TEST STAND
INJECTION PREMIX LOOP
Figure 2-4. General schematic of the UV Test Facility showing major test stands (Tank 4
not shown).
A laboratory-grade GenTech Model 1901 Double Beam UV/Vis spectrophotometer is
located at the UV Center. In addition, the UV Center provides for pH, turbidity, total chlorine,
and temperature measurements. A diesel-fired generator is used on-site exclusively to power the
UV test units. This allows power conditioning specific to the targeted unit. Other, low-power
electrical requirements are tapped off a local service. Power logging on the input to the UV unit
2-5
-------
power panel is always practiced. Multi-channel data-logging capabilities are available and are
used as needed to record relevant electrical signals, such as flow meter and intensity sensor
outputs.
2.5 UV TECHNOLOGY VENDOR - SIEMENS WATER TECHNOLOGIES
The UV unit was provided by Siemens Water Technologies and represented a
commercial version of their H-4XE-HO (2W-1B-1C) open channel UV disinfection system.
Siemens Water Technologies also provided documentation and calculations necessary to
demonstrate the system's conformity to commercial systems, hydraulic scalability and test
protocol requirements. Siemens Water Technologies UV production operations are located in
New Jersey. Dr. Bertrand Dussert is the primary contact for Siemens. He can be reached at:
Siemens Water Technologies Corp.
1901 West Garden Road
Vineland, NJ 08360
Bertrand Dussert
Bertrand.Dussert@siemens.com
(856) 507-4144
2-6
-------
SECTION 3
TECHNOLOGY DESCRIPTION
3.1 SIEMENS WATER TECHNOLOGIES OPEN CHANNEL UV DISINFECTION
SYSTEM
The O&M manual for the H-4XE-HO (2W-1B-1C) open channel UV disinfection system
is provided as Appendix B. Figure 3-1 provides schematics of the horizontal lamp reactor and its
placement in the channel. Refer to Section 3.2 for a description of the test stand and photos of
the system components and installation.
llWfc Wfc 1 tK - ~^~~*
f-\
\
A \ \ Vf./
L . - v \ 7 pi 3 /
p.
^
24"
*
I
-*
1
V"V\
,-""""'
24"
l_
!!A|
1
\ \ 1
e.
^ t
\ \
~~ ^ \
11P
3-T
J t
\
*' "
«%
_3I
c=
/
,
)
^
^.-"^
"^
2i"
(£
«
4
I
;
>
/
/V 1
/ ^
e
Figure 3-1. Plan and elevation schematic of H-4XE-HO (2W-1B-1C) reactor and channel
assembly.
3-1
-------
3.1.1 Lamps and Sleeves
The H-4XE-HO (2W-1B-1C) UV unit supplied by Siemens utilizes 16 high-output, low-
pressure lamps, oriented horizontally and parallel to the direction of flow (Figure 3-1). The
lamps are housed in two modules, each containing eight lamps. Each lamp has a UV output of
approximately 60 Watts at 254 nm and a total power draw of approximately 170 Watts. The
lamps are approximately 60 in. long and each lamp is housed in a clear fused quartz sleeve to
isolate and protect the lamp from the wastewater. The sleeves have only one open end, which
are sealed with the lamp power cable plug. These quartz sleeves are 70 in. long, have an outer
diameter of 28 mm, a wall thickness of 1.5 mm and a UV transmittance of 91%.
3.1.2 Lamp Output Attenuation by Aging and Sleeve Fouling
The total intensity attenuation factor was set by Siemens at 80%, based on the combined
effects of a sleeve-fouling factor of 90% and a lamp-aging factor (EOLL factor) of 90%. This
aging factor is set at a minimum of 12,000 operating hr.
3.1.3 Sleeve Cleaning System
The H-4XE-HO (2W-1B-1C) UV disinfection system provided by Siemens is equipped
with automatic sleeve wiping systems. The performance of the wipers was not evaluated as part
of this dose-delivery verification. However, the wipers were used to clean the sleeves at the
beginning of each validation test day.
3.1.4 Electrical Controls
The lamps in the H-4XE-HO (2W-1B-1C) unit are powered from electronic ballasts
mounted vertically in a remotely located enclosure. Each ballast powers two lamps in parallel so
that one lamp failure does not cause the peer lamp to turn off. The ballast controls are located in
the control cabinet. The ballast/control panel for the unit allows for lamp power dimming. This
function was used in the verification test in order to simulate the combined attenuation factor, but
dimming the lamps is not usually a control strategy for this commercial unit. The control cabinet
supplied for this ETV validation was powered via an onsite generator that supplied 230V delta
power.
3.1.5 UV Detectors
The disinfection system used for this verification was equipped with a SLS SiC004 UV
intensity sensor certified to DVGW Standards. One sensor was installed on the top cover of the
lamp rack, approximately 2 cm from a lamp sleeve in the top row. The sensor includes a remote,
3-2
-------
dedicated amplifier that operates on a 4-20 mA signal. The sensor has a wavelength selectivity
of 96% between 200 nm and 300 nm and a linear (1%) working range of 0.01 to 20 mW/cm2.
The stability of the sensor is 5% over 10 hr and a range of temperatures from 2 to 30°C.
3.1.6 Design Operational Envelope
The H-4XE-HO (2W-1B-1C) system verified in this ETV was designed to operate at
flow rates up to 868 gpm (1.25 mgd). The intensity monitor can be set for an appropriate reading
depending on the application, and the intensity alarm can be set to activate when a low dose
condition exists. Three common factors can contribute to a low dose condition: attenuation of
UV output by excessive lamp aging, quartz sleeve fouling, or low water transmittance
conditions. The exact setting depends on the specific application requirements. In terms of
intensity reduction due to lamp aging and quartz fouling, the suggested operational protocols
comply with the conditions in this ETV. Quartz fouling of 90% and lamp age intensity reduction
of 90% (at 12000 hr) were simulated during this ETV.
The commercial unit is typically designed to operate at 100% input power (no dimming).
The primary operating variables are the water UVT and flow rate. Within the scope of this ETV,
dose-delivery performance was verified at nominal UVT levels between 50% and 80%
transmittance at 254 nm. The flow rates were varied to yield reduction equivalent doses (REDs)
between approximately 10 and 100 mJ/cm2. In conformance with the WW Protocol and ETV
protocol (2002), a single bank was tested. The single bank is considered additive if placed in
series. The test unit is a 16-lamp version of the H-4XE-HO (2W-1B-1C). Scale-up is limited to
a factor up to 10 times this number of lamps (160) in a single bank (while maintaining the same
liquid velocity). Scale-down (less than 16 lamps/bank) is not considered appropriate.
3.2 UV TEST STAND SPECIFICATIONS
3.2.1 Test Channel
The reactor was housed in an open stainless steel channel, 21 ft-7 in. long, with lamps
oriented horizontally and parallel to the flow direction. The flow from the 12-in. influent pipe
first entered a 24-in. long, 36-in. wide by 32-in. deep influent box (Figure 3-1). From the
influent box, the flow entered a 12-in. wide channel, with a liquid depth of 12-in. The test
reactor was located at the midpoint of this 115-in. long section. At the end of this test section, an
expansion section was provided, discharging to an effluent box equivalent to the influent box.
Flow exited the channel through a 12-in. pipe to the dump tank. Figure 3-2 presents photos of
the influent and effluent piping arrangements. The horizontal-lamp unit channel was the smaller
of the two shown in either photo.
3-3
-------
Figure 3-2. Influent (top) and effluent (bottom) piping configuration for the H-4XE-HO
(2W-1B-1C) test unit channel.
3-4
-------
The test channel had a stilling plate installed at the junction of the influent box and test
section to provide good flow distribution upstream of the test unit. An adjustable weir gate was
installed at the effluent end so that the water level inside the channel could be controlled.
Photographs of the stilling plate and adjustable weir are provided in Figure 3-3. Figure 3-4
shows the reactor installed in the channel. The lamps and quartz sleeves are oriented
horizontally (top photo). Service to the lamps is through cables to the control panel (bottom
photo). The power/control panel is shown in Figure 3-5.
3.3 VERIFICATION TEST CLAIMS
The overall objective of this ETV was to validate the performance of the Siemens Water
Technologies H-4XE-HO (2W-IB-1C) open channel UV disinfection system at water quality
(UVT) and dose (RED) conditions reflective of secondary effluent and reuse applications. The
total attenuation factor of 80% was selected by Siemens as a combined effect of 90% sleeve
fouling factor and 90% of end-of-lamp-life factor. This attenuation was mimicked by lowering
the test water transmittance. Within this goal, seven specific objectives were fulfilled:
1. Verified the performance difference between power turndown and UVT turndown at
the same operating conditions to mimic the total attenuation factor.
2. Verified the flow-dose relationship for the system at nominal UV transmittances of
50%, 65% and 80% for a dose range of 5 to 25 mJ/cm2 using a biological surrogate
with a relatively high sensitivity to UV (Tl coliphage).
3. Verified the flow-dose relationship for the system at a nominal UV transmittance of
50%, 65% and 80% for a dose range of 10 to 40 mJ/cm2 using a biological surrogate
with medium sensitivity to UV (QP coliphage).
4. Verified the flow-dose relationship for the system at a nominal UV transmittance of
50%, 65% and 80% for a dose range of 20 to 80mJ/cm2 using a biological surrogate
with relatively low sensitivity to UV (MS2 coliphage).
5. Adjusted the observed RED performance results by a validation factor in order to
account for uncertainties associated with the verification tests.
6. Verified the power consumption and headloss of the unit.
7. Developed a dose-algorithm to control dose-delivery on a real-time basis, based on
the system's primary operating variables.
3-5
-------
Figure 3-3. Influent stilling plate and effluent weir gate for the Siemens H-4XE-HO (2W-
1B-1C) UV unit test channel.
3-6
-------
Figure 3-4. Photos of H-4XE-HO (2W-1B-1C) reactor installed in channel.
3-7
-------
Figure 3-5. Photo of unit power panel.
3-8
-------
SECTION 4
PROCEDURES AND METHODS USED DURING
VERIFICATION TESTING
4.1 GENERAL TECHNICAL APPROACH
By its nature, the effectiveness of UV is dependent on the upstream processes used for
pretreatment, particularly for solids, oil/grease and organics removal. The UV design basis
typically developed for a UV system application incorporates the characteristics of the
wastewater to be treated, established to reflect a planned level of pretreatment, and the expected
variability in quality and quantity. Finally, the dose required to meet specific target levels is
determined, typically established from direct testing (e.g., collimated-beam dose-response
methods) of the wastewaters or similar wastewaters. Once this "design basis" is established,
independent of the UV equipment, the next step is to select equipment that can meet these
specific dose requirements under the expected wastewater characteristics.
The ETV technical objective is met by demonstrating, or verifying, the ability of a
specific system to deliver an effective dose. This is the Reduction Equivalent Dose (RED)
actually received by the microbes in the wastewater. Direct biodosimetric procedures are used to
estimate the RED for specific reactor configurations, typically as a function of the hydraulic
loading rate and the water UVT. Biodosimetry is a viable and accepted method per current
protocols and has been used successfully for many years, whereby the results are often applied to
qualification requirements in bid documents for wastewater treatment applications.
Biodosimetry uses a known microorganism that is cultured and harvested in the
laboratory and then subjected to a range of discrete UV doses. These doses are applied with a
laboratory-scale, collimated beam apparatus, which can deliver a known, accurately measured,
dose. Measuring the response to these doses (log survival ratio), a dose-response relationship is
developed for the specific organism. A culture of the same organism is then injected into the
large-scale UV test unit, which is operated over a range of hydraulic loadings (thus yielding a
range of exposure times). The response of the organism can then be used to infer, from the
laboratory-based dose response relationship, the reduction equivalent dose that was delivered by
the UV unit.
4.1.1 Site Preparation
The testing for this ETV validation was conducted at the UV Validation and Research
Center of New York (UV Center - refer to Figures 2-2, 2-3 and 2-4). Figure 4-1 presents the test
stand process flow diagram for conducting the dose delivery verification assays, including
4-1
-------
sampling locations. Supporting instrumentation included a flow meter, UV/Vis
spectrophotometer, radiometer with an appropriate UV sensor, turbidity meter, power meter,
powerlogger and dataloggers for other operational parameters.
Challenge Water
Feed Wate r Pu mp s
T_T-,_I, J TAUV1
Influent
Sample
t
1 KIT
/ L
F low Control
Va kf»
/
1
I
/Stilling Plate
/
/ r ' -i
^ESf-'-'-'-'-'-d
^
\.
influent /
Chamber
/
n_
| f~\ Pre-niix
f\T 0Fed Lin^ ^ Sidestream
*\ /
1 2" Row Meter j
j / nJKtion Premie
x.^ Effluent Sample j / charrb^r
1 re
C il'
Effluent Chamber j LJ "
^ ]
Lev
\
e! Control Gate 1
.Additives
injection /K
ISA
o
Phage
M^teriF^
Pumps
Figure 4-1. Process flow diagram for the H-4XE-HO (2W-1B-1C) UV System validation
test stand.
The UV output of the system at 254 nm was measured by a single duty sensor. This was
a SLS SiC004 UV intensity sensor certified to DVGW Standards, connected to a remote,
dedicated amplifier operating on a 4-20 mA signal. The UV transmittance of the test water is
also critical. A laboratory-grade GenTech Model 1901 Double Beam UV/Vis spectrophotometer
was maintained at the UV Center for measuring the UV transmittance of samples. Transmittance
was also verified at the microbiology laboratory with a second GenTech 1901
sp ectrophotometer.
4.1.2 Water Source
Water for cleaning and test purposes was drawn from a local fire hydrant, which is piped
to the source-water tanks. Lignin sulfonate (LSA) was used to adjust the UVT of the challenge
water.
4-2
-------
4.1.3 Challenge Water and Discharge Tanks
The UV Center uses two large concrete storage tanks for the challenge water and two
additional concrete tanks for effluent water storage prior to final discharge to the treatment plant
(Figure 2-3). For this test, challenge water was stored in Tank 1 (0.75 million gallons). Effluent
water was stored in Tank 3 (1.3 million gallons).
4.1.4 Feed Pumps
The challenge waters were pumped to the test unit, or recirculated to the challenge water
tank, with one of eight Godwin centrifugal pumps. Each pump is diesel-powered to provide flow
rates up to 5500 gal/min. The pumps are permanently mounted alongside Tank 1 and Tank 2, as
shown in Figure 2-4.
4.1.5 Flow Meter
Flow to the system was metered using a 12-in. Krone magnetic flow meter, installed in a
12-in. pipeline with a straight run often pipe diameters before and five pipe diameters after the
flow meter to reduce turbulence that could impact meter performance. The flow meter
calibration was regularly checked before testing using a timed volume drawdown method
(Section 6.1.1).
4.2 DISINFECTION UNIT STARTUP AND CHARACTERIZATION
4.2.1 100 Hour Lamp Burn-In
Before dose delivery verification testing began, the lamps were aged for 100 hr to allow
the lamp intensity to stabilize. The lamps were turned on at 100% power with water
recirculating through the channel at a rate of approximately 900 gpm to prevent the lamps from
overheating. The burn-in period spanned five days. Log sheets are provided in Appendix C.2.1.
The powerlogger was not functional during the burn-in period, but manual power measurements
were made to monitor the period.
4.2.2 Power Consumption and Flow Characterization
4.2.2.1 Power Consumption Measurement
For the purposes of this test program, the total system real power consumption was
recorded during the actual bioassay testing and during technical tests that required the unit to be
on. A Mitchell Instruments powerlogger was connected to the control panel to record the power
draw (230V delta phase) from the onsite generator.
4-3
-------
4.2.2.2 Headloss Measurements
Measurements of headless were conducted by attaching staff gauges to the inside of the
reactor channel. The channel was leveled within 0.5 cm before the start of the testing. The zero-
level installation of the staff gauges was established with stationary water in the channel. The
vertical datum was the bottom of the channel under the UV unit, so the measurements represent
the depth of water in the channel
For this verification, the water level was measured at two positions and eight different
flow rates. The flow rates used for measuring the headloss were the minimum and maximum
flow rates used for validation and several intermediate flow rates. The measurement positions
were located approximately 6 in. upstream and 6 in. downstream of the lamp bank.
4.2.2.3 Velocity Profile Measurement
The NWRI/AwwaRF Ultraviolet Disinfection Guidelines for Drinking Water and Water
Reuse (2003) recommends that commissioned systems should have velocity profiles that are
equivalent or better than demonstrated by the validation test unit. The guidance protocol states
that the velocity measurement points should be 6 to 12 cm apart; for reactors larger than 25-cm
wide or diameter, a minimum of nine points should be used for establishing the velocity profile.
As such, a 3 x 3 measurement matrix was designed for the cross-section of the Siemens H-4XE-
HO (2W-1B-1C) UV system channel, as illustrated in Figure 4-2. These measurements were
conducted at flow rates 0.2 and 0.8 mgd. A specially designed frame was used to position the
velocity meter at the desired location inside the channel. At each location, three readings of flow
velocity were recorded. The velocity meter was a Marsh-McBirney Flow-Mate Model 2000.
Each reading was an integrated average recorded by the meter over a period of 7 seconds.
4-4
-------
12
10
£ 8
o
ra
5 4
0> 00
Water Depth (inch)
Figure 4-2. Velocity profile measurement matrix.
4.2.2.4 Shakedown Flows
Two shakedown flow tests were conducted. This allowed an initial calibration run to
determine if power turndown or UVT adjustment would be used to simulate the total intensity
attenuation factor. This also allowed a "test run" to familiarize the technicians with the
equipment operation and sampling scheme. These flow tests were conducted using the
methodology described in Section 4.4.
4.3 BACTERIOPHAGE PRODUCTION AND CALIBRATION
4.3.1 Bacteriophage Propagation
Three different bacteriophages were used for validation testing of the H-4XE-HO (2W-
1B-1C) unit: MS2, Tl and QP. All three of the microorganisms are F-specific RNA
bacteriophages. The MS2 and QP were ATCC 15597-B1 and ATCC 23631-B1, respectively,
and the host E. coll strain was ATCC 23631. Tl originates from an isolate by GAP
Enviromicrobial Services (London, Ontario) Canada. Tl is assayed with E. coll CN13 ATCC
700609 as the host organism.
4-5
-------
The propagation procedure was based on ISO 10705-1 (1995), which was refined to
produce the large volumes needed for biodosimetry. For cultures of all three bacteriophages, the
host strain (E. coif) was grown at 37°C in Trypticase yeast-extract glucose broth until the log-
growth phase was reached. This time was determined by previously completing three growth
curves of the same host-strain working culture. When the optimum log-growth phase was
reached, the stock solutions were pipetted into the bacterial growth cultures to start the infection,
which was allowed to continue overnight. During the following day, the culture media was
filtered through 0.45 and 0.22 um filters to remove cell lysate, and to remove any other bacteria
that may be present. The stock solution was stored over chloroform at 4°C.
4.3.2 Dose-Response Determination
The dose-response behavior of the bacteriophage stocks and seeded influent samples
were determined using a collimated beam apparatus residing in HydroQual's laboratory (Figure
4-3). The lamp housing is a horizontal tube, constructed of an opaque and non-reflective
material, ventilated continuously via a blower for ozone removal and for temperature control.
The collimating tube, also constructed of an opaque non-reflective material, extends downward
from the center of the lamp housing. The housing contains two conventional G64T5L low-
pressure mercury discharge lamps, which emit almost all of their energy at 254 nm. The lamp
temperature was monitored continuously via a digital thermometer with a thermocouple mounted
on the lamp skin. A Petri dish was used to hold the sample for exposure and used a magnetic
mixing system to gently stir the microbial suspension. Typical irradiances were 0.2 mW/cm2 at
the surface of the liquid. The Petri factor was approximately 0.96. A manually operated shutter
was present at the bottom of the collimating tube.
The irradiance or intensity of the collimated beam apparatus was measured using an
International Light IL-1700 radiometer with an SED 240 detector and a NS254 filter. The
radiometers and detectors were calibrated on a regular basis by International Light and are
accompanied by NIST traceable certifications. The calibration interval is approximately three
months, and is usually selected to bracket specific validation work. Per UVDGM guidance and
the WW protocol, a second detector was used to check the duty detector when collimated beam
testing was conducted. The two readings must, and did, agree within 5% of their mean reading.
4-6
-------
Support Frame.
SEE DETAIL -
THIS AREA
BELOW
Air .
Filter
Ir- 4" 0 Copper Pipe
^ ^- UV Lamp
r
*
fo *
CoHimaEor
Sample ttsh ^
Magnetic Stifrer-*'fepii
Platform *
Adjustable
L/ Drive
if for
J Piatfwm
ti
»t i.
Power
Supply
803/8'
6 3/4' i
29'
L = Effective Lamp Are Length
91/8"
29'
63/4'
DETAIL
NTS.
Figure 4-3. HydroQual collimating apparatus for conducting Dose Response tests.
All microbiological samples were exposed in a Petri-type dish, with straight sides and a
flat bottom. The outer perimeter of the sample container was always within the diameter of the
collimator. The intensity was measured at the beginning and at the end of the dose response
series. The dose-delivery calculations were based upon the methods stated in the UVDGM
(USEPA, 2006), Appendix C. The irradiance field of the collimated beam was wide enough to
completely contain the sample dish with an inside diameter of 87 mm. The reflectance of the
4-7
-------
sample surface was 2.5%, and the sample depth was 1.3 cm. In brief, the dose is calculated
using:
Where:
DCB
Es
Pf
R
L
d
%T
t
DCB =E.Pf(\-R)
s /V '
(
(d + L)
UV dose (ml/cm2)
Average incident UV intensity (before and after irradiation) (mW/cm2)
Petri Factor (unitless)
Reflectance at the air-water interface at 254 nm (unitless)
Distance from lamp centerline to suspension surface (cm)
Depth of the suspension (cm)
UV transmittance at 254 nm (cm"1)
Exposure time (s)
For this ETV, a total of 24 dose-response runs were conducted, 12 for MS2, 6 for Tl, and
6 for Qp. All dose response runs were conducted with seeded challenge waters at UVTs ranging
from 43.5% to 80.0 % transmittance. A single dose-response series consisted of a minimum six
doses to achieve a range of inactivation values. For MS2, these doses were typically 0, 10, 20,
40, 60, 80, and 100 mJ/cm2; for Qp, the doses were typically 0, 5, 10, 20, 30, 40, and 60 mJ/cm2;
and for Tl, 5, 10, 15, 20 and 25 mJ/cm2. Extrapolations were not made beyond the minimum
and maximum dose levels actually tested, so in certain instances, higher doses may also have
been analyzed, if necessary.
At least one seeded influent sample was collected from the influent sample port for each
day of flow testing and used for the collimated-beam, dose-response analysis. These were
conducted on the same day that the flow-test samples were enumerated (within 24 hours of
collection). The influent dose-response tests were typically conducted at the minimum UVT
tested on that day. Additionally, one dose-response series was conducted for each challenge
organism with the source water at the highest UVT, unadjusted with lignin sulfonate.
4.4 BIODOSIMETRIC FIELD TESTS
4.4.1 Lamp Sleeve Preparation
Before each flow test series, the lamp sleeves were scrubbed with sponges and an acidic
cleaning solution (e.g., Lime Away). The sleeves were then thoroughly rinsed to remove the
cleaning solution.
4-8
-------
4.4.2 Challenge Water Batch Preparation
Before the start of a series of biodosimetric flow tests, the test stand was prepared. The
source water staging tank (Tank 1) was filled with an adequate amount of dechlorinated (using
sodium sulfite) water, and characterized for pH, temperature, turbidity, and UVT. Samples were
tested to assure that total chlorine was non-detectable at the 0.05 mg/L level. Depending on the
test matrix planned for the day, the UVT of the tank contents were either adjusted on a batch
basis, or "on-the-fly" as each flow test was performed. UVT adjustments were made with a
lignin sulfonate (LSA) solution injected into the test stream. The UVT measurements, made
with the Gentech Model TU-1901 spectrophotometer, are reported with observed REDs. Tests
were conducted with water turbidities consistently less than 2 NTU.
4.4.3 Biodosimetric Flow Tests
Biodosimetric flow tests were conducted by pumping the test water, with the appropriate
injection of coliphage and lignin sulfonate, through the channel at the specified flow rate.
Enough time was allowed for at least five volume changeovers in the lamp assembly, the flow
rate was checked again and sampling commenced. Water that had passed through the test unit
was wasted to Tank 3.
Grab samples were collected in sterile, 120-mL single-use specimen cups. Influent
samples were collected at a sample port located two 90° bends prior to the influent box. Effluent
samples were collected from the sample port located after the effluent box and one 90° bend.
Influent and effluent samples were collected simultaneously and in triplicate, resulting in six
samples for each flow test. The samples were placed in separate influent and effluent closed
(dark) coolers with ice, and transported to the lab the same day. Samples were analyzed the next
day.
4.4.4 Transmittance Measurement
The transmittance of the challenge waters was measured on every influent sample and on
the seeded influent samples used for dose-response analysis. The transmittance was measured in
the field and in the laboratory at 254 nm in a quartz cell with a path-length of 1 cm, using a
GenTech 1901 spectrophotometer at each location. The zero reference was laboratory deionized
reagent water. The instruments were checked periodically with NIST-traceable holmium oxide
and potassium dichromate standards.
4.4.5 Bacteriophage Enumeration
The density or concentration of viable bacteriophage in the flow test and dose-response
samples was determined using ISO 10705-1 and USEPA UVDGM (Appendix C) methods.
4-9
-------
Briefly, samples containing MS2, Tl or QP bacteriophage were serially diluted in peptone-saline
dilution tubes to a dilution determined to be appropriate from experience or from screening runs.
Then, 1 mL of this diluted sample was mixed with 1 mL of host E. coli, and 2.5 mL semi-solid
growth medium. This mixture was plated onto an agar plate and allowed to grow overnight (-16
hours) at 37°C. This double-plating approach employed trypticase yeast-extract glucose broth
(TYGB) as the growth medium. Each sample was plated at two dilutions in duplicate, resulting
in four plates for each sample. Only plates with 30-300 pfu were deemed valid for analysis. The
acceptable data was then averaged geometrically and corrected for the dilution to determine the
bacteriophage concentration (pfu/mL) in the test solution.
4.4.6 Dose Determination
When reducing the dose-response data, the No used for computing inactivation was
estimated by regressing the log of the liters of all dosed and undosed samples versus applied
dose, and taking the y-intercept predicted by a second-order regression equation (UVDGM,
November 2006). This results in a NO value for each dose-response series. Then the inactivation
for each dosed sample was calculated with:
Log Inactivation = log(jV0) - log(TV)
Where:
Inactivation = Coliphage inactivation in log units.
N0 = Titer of undosed sample from y - intercept.
N = Coliphage titer (pfu/mL)in dosed sample.
The dose-response calibration for each bacteriophage was quantified by fitting a second-
order polynomial to the validation dose-response data, thereby generating a relationship where
dose is a function of survival ratio (Section 5.2). All flow test survival ratios were then
converted to Reduction Equivalent Dose (RED) with the use of this relationship (Section 5.3).
4.5 EXPERIMENTAL TEST MATRIX
For reference, the proposed VTP validation matrix for the H-4XE-HO (2W-1B-1C) is
presented in Table 4-1. This was constructed with HydroQual's simplified model. Once data
were collected for the system, the final matrix was modified to assure that the boundary limits
and interpolation points were properly covered.
4-10
-------
Table 4-1. Validation Conditions for Siemens H-4XE-HO (2W-1B-1C) UV System
Lamps
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16 (OFF)
16 (OFF)
16 (OFF)
UVT
(%/cm)
80
80
80
80
80
80
80
80
80
65
65
65
65
65
65
65
65
65
50
50
50
50
50
50
50
50
50
50
50
80
80
80
Flow
(mgd)
0.30
0.40
0.50
0.80
1.00
1.25
1.50
2.00
3.00
0.20
0.30
0.50
0.80
1.00
1.25
1.50
2.00
3.00
0.20
0.30
0.40
0.50
0.60
0.80
1.00
1.40
1.60
2.00
3.00
2.00
2.00
2.00
Predicted RED Tl
(mJ/cm2)
85.96
70.41
60.32
43.54
37.30
31.96
28.16
23.07 1
17.41 1
72.84
54.99
38.59
27.86
23.86
20.44 1
18.01
14.76 1
11.14 1
52.73
39.81
32.61
27.94
24.62
20.17 1
17.28 1
13.68 1
12.47 1
10.68
8.07 1
0.00
0.00 1
0.00
MS2 QB
1
1
1
1
1 1
1
1
1 1
1
1
1 1
1
1 1
1 1
1
1
1 1
1
1 1
1
1 1
1
1
1
1
4-11
-------
SECTION 5
RESULTS AND DISCUSSION
5.1 DISINFECTION UNIT STARTUP AND CHARACTERIZATION
5.1.1 Power Consumption
The power consumption of the Siemens H-4XE-HO (2W-1B-1C) system was
continuously logged when operating. The total attenuation condition was simulated through
UVT adjustment, not power turn-down. This allowed for direct monitoring of total real power
consumption by the H-4XE-HO (2W-1B-1C) unit at the power level (at the PLC) of 100.
Siemens states that this power level is considered the nominal input power rating for this system.
At a PLC setting of 100, the mean total power input was 2.74 kW, or 171 W/lamp.
5.1.2 Headloss Measurements
Headloss estimates were derived from the hydraulic profile data shown in Table 5-1, and
are presented graphically in Figure 5-1. Two sample locations (immediately before and after the
unit) were used at eight different flow rates. Note that the influent depth was held constant by
adjusting the downstream weir height.
Table 5-1. Depth Measurements to Compute Headloss
Flow Rate
(mgd)
0.20
0.30
0.40
0.50
0.61
0.81
1.00
1.26
Water Depth
Influent
12
12
12
12
12
12
12
12.5
(in)
Effluent
12.0
11.9
11.7
11.4
11.1
10.8
9.6
8.5
Headloss
Differential
0.03
0.13
0.31
0.56
0.88
1.25
2.38
4.00
(in)
1 Module
0.03
0.13
0.31
0.56
0.88
1.25
2.38
4.00
For the H-4XE-HO (2W-1B-1C) system the headloss (in. of water) across one 16-lamp
reactor is shown in Figure 5-2 as a function of flow (mgd). Headloss can be approximated by the
relationship:
Headloss (in. of water) = 3.160 (flow, mgd)2 - 0.938 (flow, mgd) + 0.148
5-12
-------
It is important to understand that the headloss was measured within the cited flow range
and should be extrapolated for flow rates outside this range.
4.00
360
"C" -3 on
0) -J-^-u
> 0 On
_ t.OU
o
!i< 9 AD
f
u
.E o nn
o 1 Rn
o
ra
0) 1 on
n Rn
n A.r\
Onn
Headloss (inches of water) = 3.160 (Flow, mgd)2 - 0.938 (Flow, mgd) + 0.148
R2 = 0.995
t
^^
*^"
X
./
s*^
/
/
/
»
/
/
/
f
/
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.<
Flow Rate (mgd)
Figure 5-1. Headloss through a single H-4XE-HO (2W-1B-1C) reactor as a function of
flow rate.
5.1.3 Intensity Sensor Characterization
The output of the duty sensor is a 4-20 mA signal, converted to a percentage at the PLC.
The relationship of mA to Sensor (%) is shown in Figure 5-2, and can be expressed as:
Sensor (%) = 6.25 x (Sensor, mA) -25
5-13
-------
on
so
o ou
_i
Q- 70
is 70
^o en _
*j
3 cn _
Q. OU
*j
Oj.0
o 30
0)
W20
10
n
Sense
r (%) = 6.25 (Sensor, mA) - 25
x
X
X
^x
/^
s
X
X
X
X
X
A
s^
8 10 12
Sensor Output (mA)
14
16
18
20
Figure 5-2. Relationship of sensor output (mA) and PLC sensor reading (%).
5.1.3.1 Sensor Output with UVT and Intensity Attenuation Power Setting
Sensor readings, expressed in mA, are plotted as a function of the UVT in Figure 5-3.
Recall that the Siemens system's equivalent 100% input power is set at the PLC 100 power
setting. At the highest UVT (80%), the mA reading at 100% input power was 18.56 mA. This is
the nominal sensor reading, So. As stated earlier, Siemens has set the combined intensity
attenuation factor at 0.8. This is equivalent to the ratio of the intensity reading at the attenuated
position (I) to the nominal intensity at full input power (Io). Based on the sensor output as a 4 to
20 mA signal, the attenuated sensor reading can be determined:
I/I0=(S-4)/(S0-4)
Where S and So are the sensor mA readings at the attenuated and full power outputs,
respectively. From this relationship, the attenuated sensor output is calculated to be 15.9 mA.
UVT turndown was selected as the method for simulating the attenuation factor when validating
the Siemens H-4XE-HO (2W-1B-1C) system.
5-14
-------
20
18
16
14
< 12
(A
C
0)
CO
Sensor (mA) = 0.0075 (UVT)2 - 0.603 (UVT) + 18.7
i Power Setting =100% |_
30.0 35.0 40.0 45.0 50.0 55.0 60.0 65.0 70.0 75.0 80.0 85.0
UVT (%)
Figure 5-3. Sensor reading (mA) as a function of UVT.
5.1.3.2 Sensor Uncertainty QA Validation
Sensor uncertainty was characterized for the H-4XE-HO (2W-1B-1C) reactor following
UVDGM protocols. The results of the comparisons at high and low UVT are presented in Table
5-2, where the individual sensor signals have been reduced to the appropriate averages for each
test condition. Sensor readings, in mA, were used to calculate the variance between the duty and
reference sensors. Table 5-2 shows that there was no variance observed when comparing the two
reference sensors to the average reference sensor reading (0.0%), nor when comparing the duty
sensor reading to the average reference sensor reading.
The QA requirements are (1) readings by each of two or three reference sensors should
be within 10.0% of the average of the reference sensor readings, and (2) the duty sensor should
be within 10.0% of the average reference sensor. Both of these criteria were met. As will be
discussed in a later section, meeting these QA criteria allows one to ignore sensor uncertainty
when developing the validation factor.
5-15
-------
Table 5-2. Sensor Intercomparison Variance Analysis
Duty Sensor
Reading
Power
(PLC%)
100
100
UVT
(%T/cm)
80.1
49.8
(%of4-
20mA)
91.0
20.0
Reference
mA Sensor ID
18.6 Rl
R2
7.2 Rl
R2
Reference
Reading
(%of4-
20mA)
93.0
95.0
21.0
20.0
Reference
Reading
18.6
18.6
7.2
7.2
Duty
from
Avg
Ref
(mA)
0.0%
0.0%
Duty
Sensor
Pass/Fail
Pass
Pass
Ref
from
Avg
Ref
0.0%
0.0%
0.0%
0.0%
Ref
Sensor
Pass/Fail
Pass
Pass
Pass
Pass
5.1.4 Velocity Profile Measurements
Cross-sectional velocity measurements were taken at 0.2 and 0.8 mgd. Per guidance in
the NWRI/AWWARF Ultraviolet Disinfection Guidelines for Drinking Water and Water Reuse
(2003), the mean velocity at any measured cross-sectional point of a commissioned system
should not vary by more than 20% from the theoretical average velocity (i.e., flow divided by the
cross-sectional area). Further, the commissioned system should exhibit velocity profiles that are
equivalent or better than those exhibited by the validated test unit. This is particularly important
if there is scale-up from the test unit.
The full record of velocity measurements is compiled in Appendix C. Overall, a general
observation is that the velocity profiles were variable at 0.8 mgd. The effluent measurements
tended to be outside the targeted range. The influent measurements a 0.8 mgd were fairly stable
and less variable than the effluent. At 0.2 mgd, velocity profiles were more stable, although the
influent was outside of the variability guideline at the surface. The velocity profiling data are
illustrated in Figure 5-4 and Figure 5-5. These show the average of the horizontal measurements
for each depth location (with the channel floor as the zero datum). The average profiles for the
three measurement locations are shown, as is the mean theoretical velocity (flow/area) and the
+/- 20% band about the theoretical velocity. A key observation that can be made from these data
is that the hydraulic conditions represent a 'worse' case when compared to minimum full-scale
commissioning requirements. As such, the biodosimetry performance data can be considered
conservative.
Note that the maximum flow tested for velocity profiling (561 gpm, or 0.8 mgd) was less
than the maximum flow used for the biodosimetric tests (866gpm, or 1.25 mgd). The
biodosimetric flow rate was imposed to test low Tl RED points that were needed for developing
an RED algorithm. In some applications, protocols may not allow for designing at flows where
5-16
-------
"
-------
the flow exceeds the maximum used for velocity profiles. In many cases, however, it is
acceptable.
5.2 BACTERIOPHAGE DOSE-RESPONSE CALIBRATION CURVES
5.2.1 Dose-Response Results
Biodosimetric testing for the H-4XE-HO (2W-1B-1C) system was carried out on four
different dates in the period September 2008 through October 2008. A seeded influent sample
from each day was used to develop the dose-response relationship for samples collected that day.
All dose-response tests conducted during this ETV were in compliance with the UVDGM and
NWRI/AwwaRF protocols. Calculations follow UVDGM protocols. All raw data are included
in Appendix C.
Data from the dose-response tests conducted on the three bacteriophages used during this
ETV program are summarized in Table 5-3. The delivered doses presented in the table are
calculated using the recommended equation in the UVDGM, as described in Section 4.3.2. The
NO used for computing the inactivation was estimated by regression analysis of the log of the
liters of all dosed and undosed samples versus applied dose, and taking the y-intercept predicted
by a second-order regression equation (UVDGM, November 2006). This results in a No value
for each dose-response series. Figure 5-6 shows an example of the regression analysis used to
determine NO; in this case, MS2 dose-response data are used
5.2.2 Dose-Response Calibration Curve
Once the No for each dose-response series was determined, and the associated log
inactivation (log No/N) at each dose, regression analyses were conducted in the form of a two-
variable second-order equation to yield a dose-response curve in the form:
UVDose = Ax(log^)2+Bxlog(^)
N N
Figure 5-7 presents the regression analysis for 9/25/08 as an example, based on the
calculated No and the observed Log (N/No). The 95%-confidence interval is shown, as are the
MS2 QA boundaries suggested by the UVDGM (November 2006). Each of the three MS2
collimated beam dose-response series are presented in Figure 5-8. As shown, all MS2 dose-
response data generated during the validation test fell within these UVDGM QA bounds.
5-18
-------
Table 5-3. Dose-Response Data
Dose Run
DR1
(09/11/08)
80.0 UVT
DR2
(09/11/08)
80.0 UVT
DR3
(09/11/08)
80.0 UVT
Dose
(mJ/cm2)
0.0
10.0
20.1
30.0
40.0
60.0
80.4
101.2
0.0
10.0
19.9
30.0
40.0
59.8
80.0
100.7
0.0
10.0
20.1
30.0
40.1
59.9
80.2
100.9
Log(N)
6.15
5.61
5.11
4.77
4.18
3.25
2.53
1.75
6.13
5.57
5.08
4.71
4.22
3.46
2.50
1.87
6.16
5.69
5.09
4.78
4.21
3.57
2.82
1.87
Inact1
Log(N0/N)
MS2
No = 6.12
-0.03
0.51
1.01
1.36
1.94
2.88
3.59
4.38
-0.01
0.55
1.04
1.41
1.90
2.67
3.62
4.25
-0.04
0.43
1.03
1.35
1.91
2.56
3.30
4.25
Dose Run
DR1
(09/25/08)
45.1 UVT
DR2
(09/25/08)
45.1 UVT
DR3
(09/25/08)
45.1 UVT
Dose
(mJ/cm2)
0.0
10.0
20.2
30.4
40.5
60.3
81.0
103.8
0.0
9.8
19.7
29.8
39.6
58.9
79.3
101.6
0.0
10.0
20.1
30.2
40.3
60.0
80.7
103.5
Log(N)
6.63
6.11
5.60
5.12
4.73
3.81
2.80
2.09
6.57
6.12
5.58
5.13
4.71
4.02
3.07
2.07
6.81
6.12
5.55
5.09
4.56
3.93
2.85
2.07
Inact1
Log(No/N)
No = 6.63
-0.01
0.52
1.03
1.50
1.90
2.81
3.82
4.54
0.06
0.51
1.05
1.50
1.92
2.61
3.56
4.55
-0.18
0.50
1.08
1.54
2.06
2.69
3.78
4.56
No = 6.53
DR1
(09/30/08)
58.7 UVT
DR1
(09/30/08)
58.7 UVT
0.0
9.9
20.0
40.1
60.0
79.7
102.0
124.3
0.0
9.9
20.1
40.1
60.1
79.7
101.9
124.2
6.54
5.95
5.52
4.50
3.64
2.96
2.09
1.60
6.58
6.05
5.54
4.59
3.95
2.89
2.14
1.57
0.00
0.58
1.01
2.04
2.89
3.58
4.45
4.93
-0.05
0.48
1.00
1.95
2.58
3.64
4.40
4.96
DR1
(09/30/08)
58.7 UVT
0.0
10.0
20.0
40.2
60.1
79.8
102.0
124.3
6.54
5.96
5.47
4.47
3.94
2.96
2.14
1.64
0.00
0.58
1.06
2.06
2.59
3.57
4.39
4.89
1. NO determined from regression of N and dose.
5-19
-------
Table 5-3. Dose-Response Data (Continued)
Dose Run
DR1
(09/25/08)
44.6 UVT
DR1
(09/25/08)
44.6 UVT
Dose
(mJ/cm2)
0.0
2.5
5.0
10.0
15.2
20.0
25.3
0.0
2.5
5.0
9.9
14.9
19.8
24.9
Log(N)
6.74
6.00
5.52
4.54
3.51
2.63
1.54
6.71
5.98
5.50
4.57
3.61
2.49
1.58
Inact1 Dose Run
Log(N0/N)
Tl
No = 6.61
-012 DR1
0.62 (09/25/08)
1.10 44.6 UVT
2.08
3.10
3.98
5.08
-0.09
0.63
1.12
2.04
3.01
4.12
5.04
Dose
(mJ/cm2)
0.0
2.5
5.0
10.0
15.1
20.0
25.4
Log(N)
6.65
6.01
5.70
4.70
3.72
2.65
1.62
Inact1
Log(No/N)
No = 6.84
-0.03
0.60
0.92
1.91
2.89
3.96
5.00
1. NO determined from regression of N and dose.
5-20
-------
Table 5-3. Dose-Response Data (Continued)
Dose Run
DR1
(10/07/08)
45.5 UVT
DR2
(10/07/08)
45.5 UVT
DR3
(10/07/08)
45.5 UVT
Dose
(mJ/cm2)
0.0
4.9
10.1
19.9
30.1
39.8
61.0
0.0
5.1
10.1
20.2
30.4
40.3
61.8
0.0
4.9
10.0
19.9
29.9
39.7
60.9
Log(N)
6.15
5.61
5.12
4.37
3.51
2.60
1.22
6.15
5.52
5.18
4.27
3.49
2.51
1.20
6.11
5.57
5.17
4.31
3.49
2.56
1.13
Inact1
Log(N0/N)
QP
No = 6.11
-0.04
0.49
0.99
1.74
2.60
3.51
4.89
-0.04
0.59
0.92
1.84
2.62
3.60
4.90
0.00
0.54
0.93
1.80
2.61
3.55
4.98
Dose Run
DR1
(10/07/08)
73.4 UVT
DR2
(10/07/08)
73.4 UVT
DR3
(10/07/08)
73.4 UVT
Dose
(mJ/cm2)
0.0
5.0
10.0
20.0
30.1
40.1
60.6
0.0
4.9
9.9
19.9
29.9
39.7
60.2
0.0
5.0
10.1
20.1
30.3
40.3
61.0
Log(N)
6.09
5.63
5.16
4.39
3.49
2.54
1.11
6.17
5.60
5.22
4.41
3.51
2.60
1.27
6.17
5.79
5.27
4.39
3.52
2.47
1.34
Inact1
Log(No/N)
No = 6.17
0.07
0.54
1.01
1.78
2.67
3.63
5.06
-0.01
0.56
0.95
1.76
2.66
3.57
4.90
0.00
0.38
0.90
1.78
2.65
3.70
4.83
1. NO determined from regression of N and dose.
5-21
-------
8.0
7.0 (>
y = O.OOOIx2 - 0.0512x+ 6.6275
20 40 60 80
Dose (mJ/cm )
100
120
Figure 5-6. An Example of No determination (09/25/08 Dose-Response Data)
130
120
110
100
90
0.0
Data
-Curve Fit
-UVDGM QA Boundary
1.0
2.0 3.0 4.0
Inactivation (log(NO) -log(N))
5.0
6.0
Figure 5-7. An example of a Dose-Response regression analysis for MS2
(09/25/08, UVT =45.1%).
5-22
-------
140
120
UVDGMQA
091108
92509
A 93009
0.0
1.0 2.0 3.0 4.0
MS2 Inactivation (Log(N0)-Log(N))
5.0
6.0
Figure 5-8. MS2 Dose-Response calibration curves.
Figure 5-9 and Figure 5-10 present the dose-response data developed for Tl and Q|3
coliphage. The UVDGM does not provide QA bounds for Tl or Q|3, as it does for MS2. Instead,
as described in the VTP, past dose-response date developed by HydroQual's laboratory, outside
of this ETV, were compiled and analyzed to define their 95%-confidence limits, which were then
used to assess the data generated within the project. As shown in Figure 5-11, the behavior
exhibited by the Tl coliphage was consistent with current practice. In the case of Q|3, all but the
highest dose data fell within the QA bounds (Figure 5-10). This may be because the phage stock
used for this ETV contained less particulate, allowing for a more linear behavior at the higher
dose levels. It may also be an artifact of the very limited data set used to develop the confidence
bounds.
5-23
-------
35
30
25
QA Confidence Bound
A 92509
0.0 1.0 2.0 3.0 4.0 5.0
T1 Inactivation (Log(N0)-Log(N))
6.0
7.0
Figure 5-9. Tl Dose-Response calibration curves.
70.0
60.0
50.0
0.0
100708-45UVT
* 100708-74 UVT
QA Confidence Bound
1.0 2.0 3.0 4.0
QP Inactivation (Log(N0)-Log(N))
5.0
6.0
Figure 5-10. QP Dose-Response calibration curves.
5-24
-------
The dose-response regression equation parameters for each day are summarized in Table
5-4. These equations were then used to compute the reduction equivalent dose (RED) for the
field tests collected on their respective days. The residuals resulting from a comparison of the
curve fit prediction with the actual data show no significant trend, supporting the validity of the
curve fit model. An example of the residuals analysis, from the 9/25/08 MS2 data, is shown on
Figure 5-11.
Table 5-4. Summary of Dose-Response Curve Regression Parameters
DR Date
09/11/08
09/25/08
09/25/08
09/30/08
10/07/08
10/07/08
Coliphage
MS2
MS2
Tl
MS2
QP-45UVT
QP-74UVT
A
0.8550
0.7471
0.03143
1.484
0.4763
0.4374
B
19.790
19.162
4.847
17.306
9.9807
9.9642
R2
0.9962
0.9972
0.9974
0.9965
0.9980
0.9968
0)
t/)
o
_ on
E M
TO
OQ
o -in -
15
o
0 n
0)
t/)
Q
-T 1 pi
o
s
(0 on
3 -^U
|D
in
(Z
-^n -
,
»'
f---^
' g
w
-9
/
c
f
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Inactivation (log(N0) - log(N))
4.5
5.0
5.5
6.0
Figure 5-11. Example of Dose-Response curve-fit residuals analysis
(09/25/08 MS2 data).
5-25
-------
5.2.3 Collimated Beam Uncertainty
Specific QA guidance is provided in the protocols for the dose-response collimated beam
tests. Its uncertainty is considered a component of the validation factor, as discussed in a later
section. Using the guidance provided by the UVDGM, the uncertainty of the dose-response
relationship, UDR, is assessed. A standard statistical method described in Draper and Smith
(1998) was followed to determine UDR, expressed as a percentage of the dose response at a
particular log inactivation:
Stat
1 [Log_Inact0-Mean(Log_Inact^ [L(^e DR_, - Dose Calc_,Y
kDR -lOOx x + x
Dosen, JH x-ir, T ,^ ^T T ^^2 I «-2
' Inact -Mean(Log Inact)\
Where:
n = Number of dose-response data points, unitless
Log_Inact; = Biological log inactivation at each dose point, unitless
Log_Inacto = Particular biological log inactivation rate, e.g., 1.0, unitless
Mean (Loglnact) = Average of all dose-response "Log_Inact;" values, unitless
DoseDR-i = Dose applied for each response point, mJ/cm2
Dosecak-i = Calculated dose using dose-response curve for each inactivation point,
mJ/cm2
Dosecaic-o = Calculated dose using dose-response curve for Log_Inacto, mJ/cm2
t_stat = The t statistic of the dose-response data population at 95% confidence
level
The UDR for all dose-response series in this validation is presented in Figure 5-12 as a
function of the phage log-inactivation. Using guidance provided by accepted protocols,
including the UVDGM, the UDR, computed at the 95%-confidence level, should not exceed 15%
at the UV dose corresponding to 1-Log inactivation. As shown in Figure 5-12, this criterion is
met, which means that the UDR does not have to be included in the validation factor.
5-26
-------
09/11/08 Data, UVT = 80.0
09/25/08 Data, UVT = 45.1
09/25/08 Data, UVT = 44.6
09/30/08 Data, UVt = 58.7
10/07/08 Data, UVT = 45.5
10/07/08 Data, UVT = 73.4
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Log Inactivation
3.5
4.0
4.5
5.0
Figure 5-12. Dose-Response curve-fit uncertainty (UDR).
5.3 DOSE-FLOW ASSAYS
5.3.1 Intensity Attenuation Factor
Biodosimetric tests were conducted at a simulated total attenuation factor of 80%,
representing the combined effects of the end-of-lamp-life (EOLL) factor and the fouling factor.
Siemens stated that the PLC power setting of 100 was considered the full or nominal operating
input power for the H-4XE-HO (2W-1B-1C) system. As discussed in Section 5.1.3, the total
attenuation factor for the Siemens H-4XE-HO (2W-1B-1C) system was simulated by lowering
the water transmittance. For testing at the three nominal UVT values, 80%, 65%, and 50%, used
for this validation, the actual UVT levels that were used to simulate 80% sensor attenuation were
determined by direct measurements. These were determined to be 74.6%, 58.5% and 44.7%,
respectively.
5-27
-------
5.3.2 Flow Test Data and Results Summary
A total of 31 acceptable flow tests were conducted for this ETV. The results are
summarized in Table 5-5, Table 5-6 and Table 5-7. Included are the no-dose flow tests that were
conducted with each test organism (Section 6.3.2). All raw data and notes are included in
Appendix C.
Water quality was checked for each day of sampling. Raw total residual chlorine was
typically between 0.4 and 0.5 mg/L. Dechlorination was performed, yielding total residual
chlorine levels always less than 0.05 mg/L (the minimum detection level). Through the full field
test period, the turbidity was between 0.33 and 0.66 NTU; water temperatures ranged between
12.5 and 17.1 °C; and pH was between 6.95 and 7.23. These data are provided in Appendix
C.3.4.
Tables 5-5 to 5-7 present the average values for the operational parameters and the
analytical results for each field test condition (three influent and three effluent samples). The
flow is an average of the flow rate during the sampling period. The reported UVT measurement
is the average of all three influent samples, and the inactivation represents the log difference
between the average of the influent samples and the average of the effluent samples. The
reported reduction equivalent dose (RED) is based upon the dose-response curve for the
collimated beam data from the same day, as presented in Section 5.2.
The biodosimetric RED data are presented as a function of flow (gpm) in Figure 5-13 for
each challenge phage at their respective nominal UVT levels. More specific to design sizing
considerations, Figure 5-14 presents the same RED data as a function of the flow rate per lamp,
which was computed as the observed flow divided by 16 lamps. The bounds described by Figure
5-13 and Figure 5-14 represent the validated operating envelope for the UV system:
Flow: 134 to 866 gpm
Flow/lamp: 8.37 to 54.14 gpm/lamp
UVT: 50 to 80%
Power: 100 at PLC, or 100% input (2.75 kW/16 lamps, or 171 W/lamp
5-28
-------
Table 5-5. MS2 Biodosimetry Tests: Delivered RED and Operations Data
PLC
Power
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
Table
PLC
Power
100
100
100
100
100
s
(%)
17.0
17.0
17.0
0.0
36.0
17.0
78.7
36.0
17.3
79.0
17.7
78.3
36.0
79.0
36.0
79.0
36.0
5-6.
SI
(%)
16.3
17.0
17.0
17.0
0.0
Actual
Power
(kW)
2.7
2.7
2.7
0
2.7
2.7
2.7
2.7
2.7
2.7
2.7
2.7
2.7
2.7
2.7
2.7
2.7
%T
Actual
(%/cm)
45.3
44.9
44.8
45.1
59.1
45.4
74.9
58.8
45.4
74.9
45.9
74.8
58.6
74.8
58.3
74.8
58.5
Tl Biodosimetry Tests:
Actual
Power
(kW)
2.7
2.7
2.7
2.7
0.0
%T
Actual
(%/cm)
44.6
44.9
45.1
44.8
45.1
Flow
(mgd)
0.50
0.79
1.25
1.24
0.19
0.19
0.30
0.31
0.30
0.40
0.40
0.50
0.50
0.60
0.60
0.70
0.80
Flow
(gpm)
346
549
866
864
134
134
208
212
212
281
280
349
350
420
414
484
552
Delivered RED
Flow
(mgd)
0.60
0.79
0.99
1.25
1.24
Flow
(gpm)
415
549
689
866
864
Flow
per
Lamp
(gpm/L)
21.60
34.34
54.14
54.00
8.39
8.37
13.02
13.26
13.23
17.53
17.49
21.81
21.85
26.22
25.88
30.28
34.52
MS2
Inact
(N0-N)
1.56
1.33
0.99
0.00
4.43
2.76
5.92
3.51
2.09
4.75
1.75
4.18
2.53
3.73
2.20
3.35
1.60
and Operations
Flow
per
Lamp
(gpm/L)
25.93
34.34
43.04
54.14
54.00
Tl
Inact
(N0-N)
3.47
2.73
2.54
2.12
-0.01
MS2
RED
(mJ/cm2)
31.7
26.7
19.8
-0.1
105.9
59.0
154.5
79.0
42.7
115.6
34.8
98.2
53.2
85.2
45.4
74.7
31.6
Data
TIRED
(mJ/cm2)
17.2
13.5
12.5
10.4
0.0
5-29
-------
Table 5-7. Q|3 Biodosimetry Tests: Delivered RED and Operations Data
PLC
Power
100
100
100
100
100
100
100
100
100
SI
(%)
17.0
17.0
17.0
36.0
36.0
78.0
36.0
16.7
0.0
Actual
Power
(kW)
2.8
2.8
2.8
2.8
2.8
2.8
2.8
2.8
0.0
%T
Actual
(%/cm)
44.0
44.6
43.7
57.8
57.8
73.8
57.5
43.6
74.3
Flow
(mgd)
0.30
0.40
0.50
0.49
0.60
0.70
0.79
0.79
0.79
Flow
(gpm)
205
277
344
340
414
483
548
551
547
Flow
per
Lamp
(gpm/L)
12.84
17.30
21.48
21.27
25.85
30.18
34.24
34.45
34.19
QP
Inact
(N0-N)
2.84
2.37
2.06
3.68
3.28
4.81
2.67
1.58
-0.1
QPRED
(mJ/cm2)
31.8
26.1
22.4
42.6
37.3
58.0
29.7
16.9
-0.1
1 fin
I DU
1 An
I *\ U
N~* 1 on
g I ZU
^ 1 nn
j I UU
" Rn
Q 8°
LU
An
*fU
on
zu
n
I
»
4
A
A
*
A
4
*
B
*
*
MS2, Nominal UVT = 80%
MS2, Nominal UVT = 65%
A MS2, Nominal UVT = 50%
T1, Nominal UVT =50%
QB, Nominal UVT = 80%
*QB, Nominal UVT = 65%
QB, Nominal UVT = 50%
f
§
A
0 100 200 300 400 500 600 700 800 900 1000
Flow Rate (gpm)
Figure 5-13. MS2, Tl and Qp RED as a function of UVT and flow.
5-30
-------
IOU
1RD
IDU
-i /in
itu
19D
CM '^U
H mn
E
** on
Q ou
LU
* fin
DU
/in
'tU
on
zu
n
A
*
A
D
*
A
^
A
?
^
*
»
9
1
MS2, Nominal UVT = 80%
MS2, Nominal UVT = 65%
A MS2, Nominal UVT = 50%
T1, Nominal UVT = 50%
QB, Nominal UVT = 80%
*QB, Nominal UVT = 65%
DQB, Nominal UVT = 50%
A
0
10 15 20 25 30 35 40 45 50 55 60 65 70
Flow Rate per Lamp (gpm/Lamp)
Figure 5-14. MS2, Tl and QP RED as a function of UVT and flow/lamp.
5.3.3 Biodosimetric Data Analysis - RED Algorithm
A dose algorithm was developed to correlate the observed MS2, Tl and QP RED data
with the reactor's primary operating variables. These are the hydraulic loading rate, as defined
by the flow rate per lamp, Q/L, and sensor reading, S, which reflects the quality of the water
(UVT) and the output of the lamp. These variables are known on a real-time basis by the PLC
and can be programmed into software to monitor and control the UV system.
Because multiple surrogates were used to test the system, it is possible to combine the
test results and incorporate the sensitivity of each in order to differentiate their individual
reactions at the specified operating conditions. The commissioned system can then incorporate
the sensitivity of the targeted pathogen (e.g., total or fecal coliform, enterococcus, etc.) when
calculating the RED delivered by the system. In this context, it is important that the targeted
pathogen (or pathogen indicator) have a UV sensitivity that is within the bounds of those tested.
The three replicates from each operating condition were treated as individual test points
for the dose algorithm development. The dose algorithm to estimate the RED is expressed as:
5-31
-------
RED = 10" (Q/L)b Sc UVSd
Where:
Q = Flow rate, gpm
L = Number of lamps
S = Sensor Reading (%)
UVS = UV Sensitivity (mJ/cm2/Log Inactivation)
a, b, c, d = Equation coefficients
Note that the same sensor and its installed conditions, such as model type, position
relative to the lamp, sleeve clarity, etc., must be used to apply this algorithm (see Section 5.3.4).
This algorithm is valid if sensor readings are confirmed to meet the modeled results as a function
of UVT and power setting.
Based on a multiple linear regression analysis in the form of this RED equation, the
coefficients were determined and are summarized in Table 5-9. The algorithm-calculated REDs
versus the observed MS2, Tl and QP REDs are plotted in Figure 5-15, wherein good agreement
is noted between the predicted and observed RED. This comparison is used in Section 7 to
assess the uncertainty associated with the experimental methods used to generate the RED data.
Table 5-8. H-4XE-HO (2W-1B-1C) Dose-Algorithm Regression Constants
Coefficient Value
a 0.950550
b -0.609884
c 0.683241
d 0.398391
5-32
-------
120
110
100
-. 90
| 80
1 70
Q
£ 60
ra
O
20
10
**
0 10 20 30 40 50 60 70 80 90 100 110 120
Observed RED (mJ/cm2)
Figure 5-15. Algorithm-Calculated RED versus Observed RED.
5.3.4 Sensor Model
The calculated RED results displayed in Figure 5-14 are based on the actual sensor
readings. When commissioned, it will be necessary to assure that the same sensor position is
maintained and the same readings are obtained at given operating conditions. To assist with this
objective, sensor measurements were analyzed and a sensor model developed to allow prediction
of the sensor reading in a commissioned system. This is written:
3.169
Where:
S = 0.000082645 (UVT254)
S = Sensor reading (%)
UVT254 = UV transmittance at 254nm (%/cm)
5-33
-------
Figure 5-16 presents the model predictions as a function of the UVT. These data are at a
power setting, P, of 100, which is the normal operating condition for the H-4XE-HO (2W-1B-
1C). As shown, there is good agreement, providing a tool to assess the sensor position and
function for a commissioned system.
1 UU
QO
SO
S5 70
4-i
3 fin
Q. DU
4-i
3 en
o ou
O 40
-------
SECTION 6
QUALITY ASSURANCE/QUALITY CONTROL
6.1 CALIBRATIONS
6.1.1 Flow Meter Calibration
The 12-in. flow meter installed at the UV Center is periodically checked for accuracy by
measuring the change in level over time while pumping into an accurately measured tank, using
a depth gauge with a resolution of 0.01 ft. The actual flow rate was determined by dividing the
volume change in the tank by the change in time and then comparing it to the average meter flow
reading recorded over the same interval. Several such calibration runs have been conducted
spanning the range of flows normally applied on the 12-in. test stand, and are summarized in
Table 6-1. There is good agreement between the flow meter reading and the flow rate calculated
by water level change. Raw data are included in Appendix C.
Table 6-1. 12-in. Flow Meter Calibration
Date
10/30/07
10/30/07
10/30/07
10/30/07
11/21/07
11/21/07
04/01/08
04/01/08
5/15/08
5/15/08
5/15/08
5/15/08
5/15/08
11/12/08
11/12/08
11/12/08
Actual Flow
Drawdown
(gpm)
132
724
1478
2739
3540
5197
499
1384
1217
830
559
293
101
1577
479
51
Flow meter , , _,
,. Corrected Flow
Reading
(gpm) (gpm)
153
706
1401
2798
3507
5295
512
1423
1231
848
585
296
105
1655
495
50
151
700
1388
2771
3474
5245
507
1410
1219
840
579
293
103
1636
489
49
Average Difference (%)
Difference
(%)
-12.6
3.8
6.7
-1.0
2.1
-0.7
-1.4
-1.6
0.0
-1.1
-3.3
0.3
-1.9
-3.6
-2.2
4.0
-0.77
6-1
-------
Based upon these calibrations, small corrections were applied to the metered flow-rate
data acquired during the validation work. Except where explicitly stated, all of the reported
flow-rate data represent this calibrated flow rate. Table 6-1 shows the "curve-fit flow" that is
predicted by the meter flow and the percent difference with the actual flow. The calibration data
are plotted in Figure 6-1 with the linear correction formula shown in units of gpm. The mean
residual is 0.8% for the meter across the range of flow rates.
12-inch Flowmeter Calibration Curve
5000
_ 4000
E
Q.
5
1 3000
LL
^
"3
S 2000
1000
n
c Meter vs. Actual - 12 inch
v ;' ' ' "
3rl'''i
//'
/." " *
X
/^
,-*
y=1.0116x
R2 = 0.9993
1000
2000 3000 4000
Actual Flow (gpm)
5000
6000
Figure 6-1. 12-in. flow meter calibration data and correction formula.
6.1.2 Spectrophotometer Calibration
Transmittance measurements were made with a GenTech Model 1901 Double Beam
UV/Vis Spectrophotometer. Calibrations were conducted before and after validation testing, and
periodically during testing with a NIST-traceable Holmium Oxide cell for wavelength calibration
(RM-HL S/N 6143), and a NIST-traceable Potassium Bichromate cell with matched reference
for transmittance calibration (RM-02 S/N 5925). One-centimeter path length quartz cells were
used. Table 6-2 presents the NIST-traceable data generated during the validation period. In all
cases, the calibration checks were well within the protocol guidance of 10% measurement
uncertainty.
6-2
-------
Table 6-2. Wavelength and Absorbance Checks
09/03/08
Wavelength
(nm)
235
257
313
350
Certified Wavelength
(nm)
287.66
278.23
250.16
241.24
Certified ABS
(ABS/cm)
0.236
0.277
0.093
0.208
Measured Wavelength
(nm)
287.50
278.00
250.90
241.00
Measured ABS
(ABS/cm)
0.234
0.276
0.092
0.206
Error
(%)
-0.84
-0.36
-1.08
-0.96
09/15/08
Wavelength
(nm)
235
257
313
350
Certified Wavelength
(nm)
287.66
278.23
250.16
241.24
Certified ABS
(ABS/cm)
0.236
0.277
0.093
0.208
Measured Wavelength
(nm)
287.40
278.00
249.90
241.00
Measured ABS
(ABS/cm)
0.235
0.276
0.092
0.206
Error
(%)
-0.42
-0.36
-1.08
-0.96
09/22/08
Wavelength
(nm)
235
257
313
350
Certified Wavelength
(nm)
287.66
278.23
250.16
241.24
Certified ABS
(ABS/cm)
0.236
0.277
0.093
0.208
Measured Wavelength
(nm)
287.40
278.00
250.00
241.00
Measured ABS
(ABS/cm)
0.237
0.277
0.093
0.207
Error
(%)
0.42
0.00
0.00
0.48
6-3
-------
Table 6-2. Wavelength and Absorbance Checks (Continued)
09/30/08
Wavelength
(nm)
235
257
313
350
Certified Wavelength
(nm)
287.66
278.23
250.16
241.24
Certified ABS
(ABS/cm)
0.236
0.277
0.093
0.208
Measured Wavelength
(nm)
287.40
278.00
250.00
241.00
Measured ABS
(ABS/cm)
0.236
0.277
0.093
0.207
Error
(%)
0.00
0.00
0.00
0.48
10/06/08
Wavelength
(nm)
235
257
313
350
Certified Wavelength
(nm)
287.66
278.23
250.16
241.24
Certified ABS
(ABS/cm)
0.236
0.277
0.093
0.208
Measured Wavelength
(nm)
287.40
278.00
249.80
241.00
Measured ABS
(ABS/cm)
0.237
0.278
0.093
0.208
Error
(%)
0.42
0.36
0.00
0.00
10/13/08
Wavelength
(nm)
235
257
313
350
Certified Wavelength
(nm)
287.66
278.23
250.16
241.24
Certified ABS
(ABS/cm)
0.236
0.277
0.093
0.208
Measured Wavelength
(nm)
287.50
278.00
249.90
241.10
Measured ABS
(ABS/cm)
0.235
0.276
0.092
0.207
Error
(%)
-0.42
-0.36
-1.08
-0.48
6-4
-------
6.1.3 UV Intensity Sensors
For validation test purposes, accepted protocols require that the duty sensors are within
10% of the average of two reference sensors, and that the two reference sensors should be within
10% of their individual measurements. These data had been summarized in Table 5-2, and show
that readings are within the 10% QA limits.
6.1.4 Radiometer Calibration
Dose-response data were generated using IL1700 radiometers with an SED240 detector
and 254 filter. Per protocol guidance, the radiometers are regularly factory calibrated to within a
measurement uncertainty less than 8%. Certifications are provided for the radiometers in
Appendix C. Additionally, two radiometers were used for the collimated beam tests, with the
second unit checking the readings of the primary unit. The radiometers should be within 5% of
one another or corrective action is required. Table 6-3 summarizes the comparison of the two
radiometers. The difference from Radiometer 1 to Radiometer 2 ranged between -0.90% and
0.45%, well within the guidance limits. Moreover, the irradiance measurement dose should not
differ by more than 5% before and after UV exposure. According to Table 6-3, the maximum
absolute difference is 3.67% before and after exposure. As such, no adjustments are necessary in
interpreting the dose-response data within runs.
6.2 QA/QC OF MICROBIAL SAMPLES
QA guidance had been provided in the VTP. The field and laboratory measurements
were found to be in general compliance with procedures and results, except as may be noted in
the following discussions. One deviation from the VTP was the fact that duplicate plating was
carried out for each dilution in each coliphage analysis, whereas triplicate plating was cited in the
VTP. This method is accepted within lab standard operating protocols, and had inadvertently
been carried forward for the ETV tests. Although a deviation, it is not believed to have any
impact on the results of the tests, given the strong agreement observed between duplicate plates
and replicate samples. If there was any additional uncertainty caused by having 2 instead of 3
plates, this is accounted for in the overall uncertainty expressed in the Validation Factor (VF).
The VF is discussed in Section 7.
6.2.1 Reactor Controls
Influent and effluent samples were taken with the lamps off and with phage injection.
The equivalent RED of the difference between the influent and effluent liters should be within
the measurement error of the lowest measured RED (cited as less than 3%). Table 6-4
summarizes the results of the "no-dose" control for the H-4XE-HO (2W-1B-1C) system. The
absolute difference between the influent and effluent control samples results in an RED of 0.03
6-5
-------
ml/cm2 for MS2 phage, 0.01 ml/cm2 for Tl phage and 0.4 ml/cm2 for QP corresponding to
0.16% and 0.11% and 4.5% of the minimum observed MS2, Tl, and QP RED value,
respectively. The control sample for QP is more than the suggested 3% of the minimum QP
RED; however, the equivalent RED of the QP control sample corresponds to only -0.04 log in
phage titer change. Such a small difference in phage titer is near the error limits of the test itself.
With this caution, for purposes of this test series, the no-dose differences for each of the phages
are considered within the measurement uncertainty of the phage analysis.
6-6
-------
Table 6-3. Comparison of Dual Radiometer Readings for Collimated Beam Measurements
Radiometer 1
Radiometer 2
Date Plated
9/11/2008
UVT = 80.0%
MS2
9/25/2008
UVT = 45.1%
MS2
9/25/2008
UVT = 44.6%
Tl
9/30/2008
UVT = 5 8. 7%
MS2
10/7/2008
UVT =45. 5%
QP
10/7/2008
UVT = 73 .4%
QP
Run
No.
DR1
DR2
DR3
DR1
DR2
DR3
DR1
DR2
DR3
DR1
DR2
DR3
DR1
DR2
DR3
DR1
DR2
DR3
Initial
(mW/cm2)
0.2470
0.2480
0.2470
0.2170
0.2260
0.2190
0.2210
0.2210
0.2170
0.2330
0.2320
0.2310
0.2250
0.2240
0.2270
0.2240
0.2250
0.2220
Final
(mW/cm2)
0.2480
0.2470
0.2470
0.2200
0.2190
0.2210
0.2210
0.2170
0.2200
0.2320
0.2310
0.2300
0.2240
0.2270
0.2240
0.2250
0.2220
0.2250
Average
(mW/cm2)
0.2475
0.2475
0.2470
0.2185
0.2225
0.2220
0.2210
0.2190
0.2185
0.2325
0.2315
0.2305
0.2245
0.2255
0.2255
0.2245
0.2235
0.2235
Difference
Initial
Final
-0.40
0.40
0.00
-1.37
3.15
-0.91
0.00
1.83
-1.37
0.43
0.43
0.43
0.45
-1.33
1.33
-0.45
1.34
-1.34
Initial
(mW/cm2)
0.2480
0.2490
0.2480
0.2200
0.2280
0.2220
0.2230
0.2260
0.2200
0.2320
0.2310
0.2300
0.2250
0.2230
0.2270
0.2240
0.2240
0.2220
Final
(mW/cm2)
0.2490
0.2480
0.2480
0.2200
0.2220
0.2230
0.2260
0.2200
0.2220
0.2310
0.2300
0.2300
0.2230
0.2270
0.2240
0.2240
0.2220
0.2250
Average
(mW/cm2)
0.2485
0.2485
0.2480
0.2200
0.2250
0.2225
0.2245
0.2230
0.2210
0.2315
0.2305
0.2300
0.2240
0.2250
0.2255
0.2240
0.2230
0.2235
Difference
Initial
Final
-0.40
0.40
0.00
0.00
2.67
-0.45
-1.34
2.69
-0.90
0.43
0.43
0.00
0.89
-1.78
1.33
0.00
0.90
-1.34
Difference
1-2
-0.40
-0.40
-0.40
-0.69
-1.12
-1.14
-1.58
-1.83
-1.14
0.43
0.43
0.22
0.22
0.22
0.00
0.22
0.22
0.00
6-6
-------
Table 6-4. Reactor Control Sample Summary
Phage
MS2
Tl
QP
Date
10/02/08
10/02/08
10/07/08
Control RED
(mJ/cm2)
0.03
0.01
0.4
Minimum RED
(mJ/cm2)
18.9
9.1
8.8
Percentage
(%)
0.16%
0.11%
4.5%
For each reactor control the effluent samples were compared with the average of
the three influent replicates. These data are shown in Table 6-5. The titer differences are
well within the range of similarity for identical samples, reflecting that there are no
extraneous effects on the survival ratios observed during flow tests.
Table 6-5. Similarity between Replicate Flow Test Samples
Date Phage Sample Avg INF EFF Similarity
9/25/2008 MS2
9/25/2008 Tl
10/7/2008 QP
A
B
C
A
B
C
A
B
C
6.26E+00
6.24E+00 6.26E+00
6.21E+00
6.00E+00
5.97E+00 5.98E+00
5.97E+00
5.64E+00
5.57E+00 5.59E+00
5.53E+00
-0.02
-0.02
0.03
-0.03
-0.01
0.00
-0.07
-0.02
0.04
6.2.2 Reactor Blanks
Reactor blanks are daily influent and effluent samples taken when there is no
challenge microorganism injection. Their titer records are summarized in Table 6-6.
Several of the blank samples were noted with measurable liters up to 3-log. This is likely
due to the leakage of residual materials from the phage injection system, which was not
completely disconnected/isolated from the system when the blanks were collected.
However, when these levels are compared to the influent liters of 6-log and above, the
titer in the blanks is less than 0.1% of the influent titer, and can be considered negligible.
Table 6-6. Summary of Reactor Blank and Trip Control Sample Analyses
6-7
-------
Flow
Day
1
2
3
4
5
6
7
8
Date (Phage)
09/11/08 (MS2)
09/16/08 (MS2)
9/18/2008 (MS2)
9/18/2008 (Tl)
9/23/2008 (MS2)
9/23/2008 (Tl)
9/25/2008 (MS2)
9/25/2008 (Tl)
9/30/2008 (MS2)
10/2/2008 (MS2)
10/2/2008 (Tl)
10/07/08 (Q(3)
Phage
Trip
Control
l.OE+12
l.OE+12
l.OE+12
1.20E+06
5.0E+11
7.0E+10
5.0E+11
5.00E+10
7.0E+11
6.5E+11
5.0E+10
1.2E+11
Est.
Diluted
Titer
8.3E+11
8.9E+11
6.9E+11
l.OOE+06
4.6E+11
5.7E+10
3.4E+11
5.58E+10
6.7E+11
4.4E+11
4.9E+10
1.3E+11
Diffin
Log
Cone.
(%)
0.7
0.4
1.3
1.3
0.3
0.8
1.4
-0.4
0.1
1.5
0.1
-0.4
Trip
Blank
(DI
Water)
0
0
0
0
0
0
0
0
0
Influent
Blank
1.1E+01
l.OE+03
l.OE+03
7.2E+00
O.OE+00
l.OE+01
3.3E+02
l.OE+03
l.OE+03
Effluent
Blank
3.3E-01
l.OE+03
l.OE+03
O.OE+00
O.OE+00
1.4E+01
3.3E+02
l.OE+03
1.4E+01
6.2.3 Trip Controls
Trip Controls are samples collected from the challenge phage stocks during the
test days and shipped to the laboratory with the field samples. Any change in the log
concentration of the phage stocks should be less than 3 to 5%. The titer of the stock was
analyzed before shipment to the UV Center, the feed stock was sampled at the Center and
returned to the laboratory. The comparison shown in Table 6-6 shows the measured feed
stock measurement, and the initial feed stock measurement calculated with an equivalent
dilution. As shown on Table 6-6, the differences range from -0.4% to 1.5%.
Additionally, trip blank controls (DI water) were collected on each testing day
and traveled with the samples to assure no contamination happened during the sample
shipment. Table 6-6 shows that this QA check is also satisfied.
6.2.4 Flow Test Sample Replicates
Generally, one influent and one effluent sample were plated in replicate on each
test day for a total of 11 replicate platings. The similarity of these liters allows a
quantitative evaluation of the plating procedure. The liters are compared by calculating
the similarity:
0 , , [ Sample Titer \(pfu I mL}\
Similarity = log ^-Ll{
Sample Titer 2 (pfu/mL)j
6-8
-------
The targeted goal is that these samples should be within the analysis error of 0.2
log. Table 6-7 shows the results of the replicate similarity tests. For the 9 samples plated
in replicate during this ETV validation, all were within the acceptable limit. The
maximum difference was 0.078 log.
Table 6-7. Results from Flow Test Replicates
Date
9/25/2008
9/30/2008
10/7/2008
Flow
868 gpm; 43% T;
100 P
1042 gpm; 43%
OJr '
T; 100 P
347 gpm; 43% T;
or^ * *
100 P
417 gpm; 74% T;
100 P
130 gpm; 45% T;
or^ * *
100 P
278 gpm; 74% T;
100 P
139gpm;60%T;
or^ * *
100 P
347 gpm; 59%;
100 P
347 gpm; 45. 4%;
100 P
Sample
Influent 1
Influent 2
Effluent 1
Effluent 2
Influent 1
Influent 2
Influent 1
Influent 2
Influent 1
Influent 2
Influent 1
Influent 2
Effluent 1
Effluent 2
Influent 1
Influent 2
Effluent 1
Effluent 2
Log Titer
6.06E+00
6.14E+00
5.97E+00
6.00E+00
6.07 E+00
6.14E+00
6.41 E+00
6.40 E+00
6.23 E+00
6.25 E+00
6.31 E+00
6.29 E+00
1.75 E+00
1.81 E+00
6. 10 E+00
6.07 E+00
4.07 E+00
4.05 E+00
Similarity
- 0.078
-0 033
\J . \J ~J ~J
-0.072
0 012
\J . \J _L ^
-0.019
0 029
\J . \J £* -S
-0.062
0 027
\J . \J £* 1
0.026
6.2.5 Transmittance Replicates
During the ETV each influent sample was analyzed for %T at 254 nm at the
laboratory. In 1 1 cases a sample was analyzed in replicate to determine the repeatability
of the transmittance measurement. The samples are compared using the relative percent
difference (RPD):
RpD=
Analysts!- Analysts 2
Average(Analysis)
Table 6-8 shows the RPD of nine transmittance measurements that were
replicated. In all cases, the replicate measurements are in agreement within the 0.5%
allowed by the test plan.
6-9
-------
6.2.6 Method Blanks
Method blanks are used to check the sterilized reagents used for the challenge
virus assay procedure. The challenge microorganism concentration in these blanks was
always non-detectable, as shown in the bench records in Appendix C.
Table 6-8. Relative Percent Difference for %T Replicates
Date Flow UVT1 UVT2 RPD
9/1 1/2008
9/25/2008
9/25/2008
9/30/2008
9/30/2008
9/30/2008
9/30/2008
10/7/2008
10/7/2008
INF-4A
INF-3A
INF-6C
INF-3C
INF-7A
INF- 1 OB
INF-13C
INF-13C
INF-17A
80.7
45.3
44.9
75.0
45.9
74.9
58.3
43.5
57.4
80.6
45.3
44.7
74.9
45.9
74.8
58.3
43.5
57.5
0.12
0.00
0.45
0.13
0.00
0.13
0.00
0.00
-0.17
6.2.7 Stability Samples
Phage stability was checked by comparing the phage concentrations of a sample
plated at 24 hr and at 48 hr after collection. Phage log concentrations of these two
estimates should not differ more than 5% from each other. Table 6-9 summarizes the
stability check results for MS2, Tl and Q|3 phage. In all cases, the phage concentrations
measured at 24 hr and 48 hr did not differ by more than 5%, meeting the criteria.
Table 6-9. Phage Stability Sample Summary
Date Phage Sample UVT 24 hr 48 hr Diff(%)
9/30/2008 MS2
10/2/2008 Tl
A
B
C
A
B
59.1
59.1
59.0
96.0
95.9
6.38
6.28
6.36
5.74
5.83
6.33
6.37
6.38
5.65
5.62
0.74
-1.35
-0.22
1.59
3.58
6-10
-------
10/07/2008
QP
c
A
95.8
73.8
5.77
6.20
5.61
6.26
2.93
1.12
6.2.8 Collimated-Beam Apparatus
The protocol addresses the collimated beam dose calculation and recommends an
examination of the dose-calculation uncertainty. Uncertainty criteria are suggested for
specific terms within the dose calculation. These are summarized in the following
discussions, which present the dose term, the recommended criterion and the estimated
uncertainty associated with the methods used by HydroQual. As shown, the collimating
apparatus used by HydroQual is well within these guidelines:
Depth of Suspension (d): Protocol Requires < 10%
The same Petri dishes are used for holding the test sample, and a constant volume
is added to the sample. This enables one to always achieve the same depth of suspension
from test to test. The error is estimated to be 3.8%.
Average Incident Irradiance (Es): Protocol Requires < 8%
This criterion is similar to that of the radiometer uncertainty and associated
criteria. The radiometer used was periodically calibrated with an uncertainty <8%.
Certifications for the radiometers used by HydroQual are provided in Appendix C.2.
The Protocol recommends that the irradiance measurement should not differ by
more than 5% before and after UV exposure. Additionally, it is required that two
radiometers are used for the collimated beam tests, with the second unit checking the
readings of the primary unit. The radiometers should be within 5% of one another or
corrective action is required. As discussed in Section 6.1.4, this criterion is met.
Petri Factor (Pf): Protocol Requires <5%
The Petri factor is established as the ratio of the average of intensity readings
taken across the sample surface to the intensity at the center of the surface. The Petri
factor is determined using a fixed apparatus, constant grid and dish geometry, and
calibrated detectors. At HydroQual, the Petri factor was typically 0.95, with an error of
approximately 2.2%.
6-11
-------
L/(d+L): Protocol Requires < 1%
The uncertainty of this parameter relates to the measurement of L (distance from
lamp centerline to suspension surface) and d (depth of the suspension). At HydroQual,
the uncertainty was estimated to be approximately 0.12%.
Time (t): Protocol Requires <5%
A timer/stopwatch is used to measure the time of exposure. The minimum
exposure allowed is 30 seconds, although the typical minimum exposure time is 60
seconds. The error estimated for the manually operated shutter at HydroQual is
approximately 1.7%.
O-10'adVad: UVDGM requires <5%
This term accounts for the absorbance through the depth of the water sample.
Absorbance is measured with an estimated uncertainty of 1% at 254 nm.
6.3 DOSE-RESPONSE DATA
All raw data for dose-response analyses are included in Appendix C.
6.3.1 Excluded Data
No dose-response series are excluded from the analysis of this ETV. All dose
response series had plaque counts between the QC boundaries of 30 and 300 on the dosed
samples.
6.3.2 MS2 Compliance with QC Boundaries
The QC criteria for the acceptance of the MS2 dose-response data is described in
the NWRI Verification Protocol (2003) which defines linear boundaries for the data, and
requires greater than 80% of the data to fall between the lines. These QC criteria are
based on the statistical analysis of MS2 dose-response data from several independent
labs. Figure 6-2 shows the linear QC boundaries and the dose-response data for this
ETV. Of the 72 data points from the MS2 dose-responses series (one sample from each
of 3 days, with triplicate exposures) within the bounds of 20 and 130 mJ/cm2, all points
(100%) lie within the specified QC boundary lines, meeting the NWRI criterion.
6-12
-------
o
is 4
o
us
Upper Bound
Log N/N0 = (0.040*Dose) + 0.64
Lower Bound
Log N/N0 = (0.033*Dose) + 0.2
20
40
60
Dose (mJ/cm
80
2.
100
120
140
Figure 6-2. Dose-Response data and NWRI QA/QC boundary lines.
Similar bounds are not available from NWRI or other sources for Q|3 and Tl.
Refer to Section 5.2.2 for alternate presentations of confidence bounds for each of the
three test phages.
6.3.3 Uncertainty in Dose Response
The UVDGM protocols assess the quality of the dose-response data by analyzing
the uncertainties at specific applied dose levels. This analysis was presented in Section
5.2.3 and displayed in Figure 5-13. The uncertainties of the dose-response tests (UDR)
used to estimate MS2, Tl and QP RED for the H-4XE-HO (2W-1B-1C) validation were
always within the quality control criteria, in that the UDR is less than 15% at 1-log of
microbial reduction using standard statistical methods.
6-13
-------
SECTION 7
CALCULATION OF THE VALIDATION FACTOR FOR RED
AND LOG-INACTIVATION DESIGN SIZING
7.1 DISINFECTION CREDIT IN ACCORDANCE WITH CURRENT PROTOCOLS
The wastewater validation protocols set guidelines to account for potential biases and
uncertainties associated with the validation process. Accounting for these uncertainties assures
that the design sizing and operation of the installed system will deliver the targeted dose. In
order to obtain inactivation credit for UV disinfection, the validated dose of the UV system (Dv)
should be equal to or exceed the targeted dose (DT) for a particular drinking water, wastewater or
reuse application, or: Dv > DT
7.1.1 Validated Dose (Dv) and Targeted Disinfection
The overall goal of this validation is to assure that dose and log-inactivation targets can
be safely applied by the Siemens H-4XE-HO (2W-1B-1C) disinfection system in a manner that
is consistent with good design practice. As such, the validation test results described in this
report are decremented by specific experimental uncertainties and potential biases to assure that
a minimum disinfection performance can be confidently maintained. This adjusted RED is
considered the validated dose, which can then be used to determine sizing for specific
performance goals.
The validated dose for a UV system, based on the data generated from full-scale field
testing, is calculated as:
VF
In which:
Dy = Validated dose, in units of ml/cm2
,ic = Dose calculated using the appropriate RED equation (dose algorithm) and
operating conditions. In the case of the H-4XE-HO (2W-1B-1C), the
analysis is based on the combined MS2, Tl and QP RED data.
VF = Validation factor, which quantitatively accounts for certain biases and
experimental uncertainties to assure that a minimum disinfection
performance level can be confidently maintained.
7-1
-------
7.2 DETERMINATION OF THE VALIDATION FACTOR ELEMENTS
The validation factor for the H-4XE-HO (2W-1B-1C) reactor is calculated using the
expression:
Where:
VF = Validation factor.
BRED = RED bias, a dimensionless correction factor that accounts for the difference
between the UV sensitivity of the challenge organism used during the
validation tests to a standardized value for any target organism. Evaluation of
the BRED is explained in Section 7.2.1 below.
Bpoiy = Polychromatic bias, a correction factor that relates to the UV sensor germicidal
wavelength response. For the Siemens H-4XE-HO (2W-1B-1C) system,
Bpoiy=1.0, as explained in Section 7.2.2 below.
Uvai = Experimental uncertainty associated with the validation test.
7.2.1 RED Bias (BRED)
The RED Bias relates to the uncertainty when using a challenge organism that is less
sensitive to UV than the targeted organism. Reuse applications per current California Title 22
requirements, for example, are based on meeting specific MS2 inactivation and RED goals.
These are correlated to targeted viruses. In the case of low-dose secondary effluent applications,
total coliform, fecal coliform, enterococcus or E. coli are usually targeted. The sensitivities of
these classes of microbes are typically similar to the sensitivities of the Tl and QP used in the
validation tests. It is important to note that this assumes use of the linear portion of dose-
response curves developed from actual effluent samples. In the presence of particles (as
measured by the suspended solids analysis), there is often a tailing effect, attributed to the
occlusion of bacteria in the solids and unaffected by UV. One should develop the linear rate for
inactivation in order to determine the log-inactivation or RED that can be accomplished by the
UV system. The particulate bacterial levels would be considered additive to the residual non-
aggregated bacterial densities.
Since this validation used MS2, Tl and QP for application to a broad dose range, the test
microbes can effectively be considered equal or lower in UV sensitivity value (mJ/cm2/LI)
7-2
-------
associated with the targeted pathogens. As such, the BRED does not factor into the calculation of
the Validation Factor, and can be set to 1.0.
7.2.2 Polychromatic Bias (BPOLY)
Since the Siemens H-4XE-HO (2W-1B-1C) system uses monochromatic low-pressure
lamps, the potential bias associated with polychromatic UV sources is not a factor. BPOLY can be
set to 1.0 under such conditions.
7.2.3 Validation Uncertainty (UVai)
The uncertainty of validation (Uvai) in the VF calculation accounts for experimental
uncertainties associated with the major experimental variable. Uvai has between one and three
input variables (described as Us, UDR and UIN below) based on how well the validation test
adhered to recommended QA/QC. The decision tree provided by the validation protocol in
Figure 7-1 gives the associated notes for selection of the appropriate equations for calculating
Uvai.
using
standard statistical
methods?
standard statistical
methods?
Yes No
Yes No
II = (II 2 + M 2 + M 2\1/2
UVal ^UIN US UDR
U,tal = (IL2 + U
Va \~N DR
Figure 7-1. Uvai decision tree for calculated dose approach.
7.2.3.1 Sensor Measurement Uncertainty (Us)
The uncertainties associated with the intensity sensor are presented in Table 5-2. The test
results showed that the maximum variance observed when comparing the reference sensors to
7-3
-------
the average reference sensor reading, and that the maximum variance observed when comparing
the duty sensor reading to the average reference sensor reading were both less than 10%. The
sensor variance criterion was met (Figure 7-1), and Us can be ignored when calculating the
validation factor.
7.2.3.2 Dose-Response Uncertainty (UDR)
With respect to the dose-response uncertainty, the criterion (Figure 7-1) is that the UDR
must be less than 15% at an RED level equivalent to 1-log inactivation. This analysis was
conducted in Section 5 for all MS2, Tl and QP dose-response curves generated during the
validation. As shown in Figure 5-12, this criterion is met at the 95%-confidence level. As such,
the UDR does not have to be included in the validation factor.
7.2.3.3 RED Model Interpolation Uncertainty (Um)
The uncertainty of interpolation, UjN, is evaluated by the following equation:
In which:
Um = Uncertainty of interpolation, expressed as a percentage
tstat = t-statistic, retrieved from standard statistics tables (value dependent upon the
number of validation data points)
SD = Standard deviation of the errors between model-calculated and observed REDs
in the validation data set
REDcaic = Model-calculated RED prediction for any given operation point
The dose-algorithm developed for this reactor is discussed in Section 5.3.5. Refer to
Figure 5-14 for a comparison of the predicted MS2, Tl or QP RED to the observed MS2, Tl or
QP RED. The residuals were determined by comparing the calculated RED (REDCaic) against
the observed RED. With 81 data points, tstat is 1.99 and the standard deviation (SD) of the
residuals was determined to be 3.024 ml/cm2.
The expression for UIN for the Siemens H-4XE-HO (2 W- IB- 1C) becomes:
7-4
-------
1.99x3.0241 I 601.7
REDCalc
um=\ :__ ix 100% =
As noted by the above equation, UIN depends upon the REDcaic value determined for a
specific operating condition. The REDcaic, in turn, is dependent on the sensitivity value being
used for a specific application. An example of the calculated UIN can be shown as a function of
the REDcaic at a UVT of 50%. This can be done for sensitivities associated with Tl (5
mJ/cm2/LI), QP (11 mJ/cm2/LI) and MS2 (20 mJ/cm2/LI). From the dose algorithm and sensor
model presented in Section 5.3.3 and 5.3.4, respectively, at UVT = 50% and flow = 300 gpm:
The REDcaic at a sensitivity equivalent to Tl =22.0 mJ/cm2
The REDcaic at a sensitivity equivalent to QP = 30. 1 mJ/cm2
The REDcaic at a sensitivity equivalent to MS2 = 38.2 mJ/cm2
These REDCaic values are then inserted into the UIN expression:
UIN (Tl) = 601.7/22.0 = 27.4%
UIN (QP) = 601.7/30.1 = 20.0%
UIN (MS2) = 601.7/38.2 = 15.8%
This same calculation would be used at any sensitivity that is associated with a given microbe.
These should fall within the range of sensitivity covered by the validation (5 to 20 mJ/cm2/LI)
7.2.4 Calculation of the Validation Uncertainty (Uvai)
Based on Figure 7-1 and the results of the analyses for Us, UDR and UIN, the value for
Uvai simply becomes UIN, expressed as a percentage:
7.3 CALCULATION OF THE VALIDATION FACTOR
7.3.1 Validation Factor (VF)
With its specific elements assessed and defined, as discussed in Section 7.2, the
validation factor for the Siemens H-4XE-HO (2W-1B-1C) can be expressed as a function of the
UIN:
VF = 1+ (UiN/100)
7-5
-------
Substituting the function for UIN,
VF = 1 + (6.017/REDcak)
If the above examples are carried through this step, the Validation Factors for the given
conditions are computed as:
VF (Tl) = 1+ (6.017/22.0) = 1.274
VF (QP) = 1 + (6.017/30.1) = 1.20
VF (MS2) = 1+ (6.017/38.2) =1.158
Figure 7-2 presents a series of solutions for VF at a UVT of 50% and sensitivities ranging
between 5 and 20 mJ/cm2/LI. VF is shown as a function of flow/lamp under these specific and
fixed operating conditions. Similar calculations can be made at alternate operating conditions.
Note that as RED increases (flow decreases) the VF decreases. These calculations are
appropriate only when the UVS of the targeted pathogen is equal to or greater than the sensitivity
chosen for the calculations. Thus, if the sensitivity of the organism of concern is 10 mJ/cm2/LI,
then UVS must be 10 or less when conducting the calculations for the VF.
3
o
re
u.
c
o
f
re
re
1.5
1.4
1.3
Validation Factor at 50% UVT
UVS = 5 mJ/cm2/LI
- UVS = 8 mJ/cm2/LI
UVS = 11 mJ/cm2/LI
UVS = 15mJ/cm2/LI
UVS = 20 mJ/cm2/LI
5 10 15 20 25 30 35 40 45 50 55 60 65 70
Flow Rate per Lamp (gpm/Lamp)
Figure 7-2. Example solutions for Validation Factor at fixed operating conditions and a
range of UV sensitivity.
7-6
-------
7.4 VALIDATED RED AND LOG INACTIVATION
As discussed earlier, the validated RED (REDVai), is calculated as:
The calculation of VF was presented in Section 7.3. Thus, if the same examples are
carried, the validated, or credited, RED can be determined:
At the lower UVS (5 mJ/cm2/LI):
REDvai = 20.2/1.274 = 15.9 ml/cm2
At the middle UVS (1 1 mJ/cm2/LI):
REDvai = 30.1/1.20 = 25.1 ml/cm2
At the upper UVS (20 mJ/cm2/LI):
REDVai = 38.2/1. 158 = 33.0mJ/cm2
Figure 7-3 presents solutions at a UVT of 50% across the same range of UV sensitivity.
It is important to note that this assumes the system sensors have been confirmed to meet the
sensor model described in Section 5.3.6. Hereto, it is important to note that the UVS used for the
RED calculation is equal to or less than the UVS of the targeted pathogen. The solutions for
validated RED (REDV), such as those shown on Figure 7-3 as a function of flow/lamp, can be
reported on the PLC of the H-4XE-HO (2W-1B-1C), based on monitored real-time operating
conditions.
Table 7-1 provides credited RED solutions across a broad range of operating conditions
for the unit (S, UVT and Q/L), at sensitivities between 5 and 20 mJ/cm2/LI. Figure 7-3 displayed
those calculations pertinent to the 50% UVT conditions. Similar plots can be generated by the
user at alternate conditions.
7-7
-------
's
o
Q
LU
o:
a
£
(0
15
"re
UVS = 5 mJ/cm2/LI
UVS = 8 mJ/cm2/LI
UVS = 11 mJ/cm2/LI
UVS = 15mJ/cm2/LI
UVS = 20 mJ/cm2/LI
Validated RED at 50% UVT
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70
Flow Rate per Lamp (gpm/Lamp)
Figure 7-3. Credited RED at 50% UVT across a range of UV sensitivities.
7-8
-------
Table 7-1. Credited RED Solutions
UVT
(%T)
50
50
50
50
50
50
50
50
50
50
50
55
55
55
55
55
55
55
55
55
55
55
60
60
60
60
60
60
60
60
60
60
60
65
65
65
65
65
65
65
65
65
65
65
S
(%)
20.0
20.0
20.0
20.0
20.0
20.0
20.0
20.0
20.0
20.0
20.0
27.
27.
27.
27.
27.
27.
27.
27.
27.
27.
27.
35.7
35.7
35.7
35.7
35.7
35.7
35.7
35.7
35.7
35.7
35.7
46.0
46.0
46.0
46.0
46.0
46.0
46.0
46.0
46.0
46.0
46.0
Q/L
(gpm/L)
8.4
10.0
14.0
18.0
25.0
30.0
35.0
40.0
45.0
50.0
54.0
8.4
10.0
14.0
18.0
25.0
30.0
35.0
40.0
45.0
50.0
54.0
8.4
10.0
14.0
18.0
25.0
30.0
35.0
40.0
45.0
50.0
54.0
8.4
10.0
14.0
18.0
25.0
30.0
35.0
40.0
45.0
50.0
54.0
Credited RED (mJ/cm2) at UVS (mJ/cm2/LI)
5
30.7
27.2
21.4
17.8
13.9
12.1
10.7
9.6
8.8
8.1
7.6
38.8
34.4
27.2
22.7
17.9
15.6
13.9
12.6
11.5
10.6
9.9
47.8
42.5
33.7
28.3
22.4
19.6
17.5
15.9
14.5
13.4
12.6
57.8
51.4
41.0
34.5
27.4
24.1
21.6
19.6
18.0
16.6
15.7
8
37.9
33.7
26.6
22.2
17.5
15.3
13.6
12.3
11.2
10.3
9.7
47.7
42.4
33.7
28.3
22.4
19.6
17.5
15.9
14.5
13.4
12.6
58.6
52.2
41.6
35.1
27.9
24.5
21.9
19.9
18.3
16.9
16.0
70.7
63.1
50.4
42.6
34.0
30.0
26.9
24.5
22.5
20.8
19.7
11
43.7
38.8
30.8
25.8
20.4
17.8
15.9
14.4
13.1
12.1
11.4
54.8
48.8
38.9
32.7
26.0
22.8
20.4
18.5
17.0
15.7
14.8
67.3
60.0
47.9
40.4
32.3
28.4
25.5
23.2
21.3
19.7
18.6
81.0
72.3
57.9
49.0
39.2
34.6
31.1
28.3
26.1
24.2
22.9
15
50.1
44.6
35.4
29.7
23.6
20.7
18.5
16.7
15.3
14.2
13.3
62.7
55.9
44.6
37.6
30.0
26.3
23.6
21.4
19.7
18.2
17.2
76.8
68.5
54.9
46.4
37.1
32.7
29.4
26.7
24.6
22.8
21.6
92.3
82.5
66.2
56.1
45.0
39.8
35.8
32.6
30.1
27.9
26.4
20
56.8
50.5
40.3
33.9
27.0
23.7
21.2
19.2
17.6
16.3
15.4
71.0
63.3
50.6
42.7
34.1
30.1
27.0
24.5
22.6
20.9
19.8
86.8
77.5
62.1
52.6
42.2
37.2
33.5
30.5
28.1
26.1
24.7
104.2
93.1
74.9
63.5
51.1
45.2
40.7
37.2
34.3
31.8
30.2
7-9
-------
Table 7-1. Credited RED Solutions (Continued)
UVT S
Q/L
(%T) (%) (gpm/L)
70 58.
70 58.
70 58.
70 58.
70 58.
70 58.
70 58.
70 58.
70 58.
70 58.
70 58.
75 72.:
75 72.:
75 72.:
75 72.:
8.4
10
14
18
25
30
35
40
45
50
54
] 8.4
5 10
5 14
5 18
75 72.3 25
75 72.3 30
75 72.3 35
75 72.3 40
75 72.3 45
75 72.3 50
75 72.3 54
80 88.7 8.4
80 88.7 10
80 88.7 14
80 88.7 18
80 88.7 25
80 88.7 30
80 88.7 35
80 88.7 40
80 88.7 45
80 88.7 50
80 88.7 54
Credited RED (mJ/cm2) at UVS (mJ/cm2/LI)
5
68.7
61.3
49.0
41.3
33.0
29.1
26.1
23.7
21.8
20.2
19.1
80.6
71.9
57.7
48.8
39.0
34.4
31.0
28.2
25.9
24.1
22.8
93.5
83.5
67.1
56.8
45.6
40.3
36.3
33.1
30.5
28.3
26.8
8
83.
74.
60.
50.
40.
36.
32.
29.
27.
25.
23.
98.
87.
70.
59.
48.
42.
38.
34.
32.
29.
28.
113
101
81.
69.
56.
49.
44.
40.
37.
35.
33.
9
9
1
8
7
0
3
4
1
2
8
3
8
6
8
1
5
3
9
2
9
3
.9
.8
9
6
0
6
7
8
7
1
2
11
96.0
85.8
68.9
58.4
46.9
41.4
37.3
34.0
31.4
29.1
27.6
112.3
100.4
80.8
68.6
55.2
48.9
44.1
40.3
37.1
34.5
32.8
130.0
116.3
93.7
79.7
64.3
57.0
51.4
47.0
43.4
40.4
38.4
15
109.3
97.8
78.6
66.7
53.7
47.5
42.8
39.1
36.1
33.6
31.8
127.8
114.4
92.2
78.3
63.2
56.0
50.5
46.2
42.6
39.7
37.7
147.8
132.3
106.8
90.8
73.4
65.1
58.8
53.8
49.8
46.4
44.0
20
123.3
110.3
88.8
75.5
60.8
53.9
48.6
44.4
41.0
38.2
36.2
144.0
128.9
104.0
88.5
71.4
63.4
57.2
52.4
48.4
45.1
42.8
166.5
149.1
120.4
102.5
82.9
73.7
66.6
61.0
56.4
52.6
50.0
7-10
-------
SECTION 8
EXAMPLE CALCULATIONS FOR SIZING THE SIEMENS H-
4XE-HO(2W-1B-1C)
8.1 DESIGN CONDITIONS FOR EXAMPLE APPLICATIONS
An example is given to illustrate the calculations that can be conducted to evaluate the sizing
of the Siemens H-4XE-HO (2W-1B-1C). Consider the following design condition:
Flow Rate: 6000 gpm (8.64 mgd)
UVT: 55%
Performance Requirement with two reactors in series:
Application 1: Secondary effluent, Fecal Coliform < 200 cfu/100 mL (2.3 Log)
Application 2: Reuse, MS2 dose > 80 ml/cm2
8.1.1 Application 1
This is a "low-dose" application, directed at typical secondary effluents discharged from
wastewater treatment plants. In such cases, collimated-beam measurements would be made to
develop a dose-response (DR) relationship, based on fecal coliform. An example of such data is
provided in Figure 8-1, showing the tailing effect due to particulates. Taking the non-
aggregated, linear portion of the curve, the UV sensitivity is estimated to be 6.9 mJ/cm2/LI.
From the DR data, one can observe that the maximum effective dose is in the vicinity of 25
mJ/cm2, beyond which the particulate coliform control and little apparent disinfection occurs. In
order to meet the specification, a lower target fecal coliform is considered; and the dose is set at
25 mJ/cm2.
8-1
-------
6.00
i 5.00
M
~ 4.00
E
o
| 3.00
O
2.00
1.00
I/)
o
0.00
^
V
Example - Fecal Coliform Dose Response
N. Non-Aggregated, Linear Portion
^* (UVS = 6.9 m J/cm2/LI)
V ^
NV .^^ Particulate Coliform
txr s
T t /
V1
\
0.0 10.0 20.0 30.0 40.0
Dose (mJ/cm2)
50.0
60.0
Figure 8-1. Example Fecal Coliform Dose-Response curve.
As noted, sizing should consider having two reactors in series to meet this targeted dose.
Since dose is additive, each reactor would need to deliver at least 12.5 mJ/cm2 at the design flow
and UVT. From Table 7-1, at a UVT of 55%, the value of S is 27.07 (this uses the calculation
shown in Section 5.3.4). Using the dose algorithm, compute the REDcaic as a function of
flow/lamp.
From Section 5.3.3, the RED algorithm is:
RED = Wa (Q/L)b
Where:
Q = Flow rate, gpm
L = Number of lamps
S = Average Sensor Reading (%)
UVS = UV Sensitivity (mJ/cm2/Log Inactivation)
a, b, c, d = Equation coefficients
8-2
-------
Coefficient Value
a 0.950550
b -0.609884
c 0.683241
d 0.398391
S has been determined at 27.07, reflecting the same placement of the sensor as in the
validation unit. The UVS in this case is 6.9 mJ/cm2/LI, as shown on Figure 8-1 for the site-
specific fecal coliform. The flow/lamp input can be varied to evaluate REDcaic as a function of
hydraulic loading. For example, at 30 gpm/lamp, the REDcaic is:
,0.39839
RED = 100950550 x (3o)-°609884 x (27.07)°68324 x (6.9)
RED= 18.3mJ/cm2
Figure 8-2 presents solutions for REDcaic as a function of flow. These must then be
adjusted for the Validation Factor (VF).
As discussed in Section 7.3, the VF is:
VF = 1 + (6.071/REDcak)
Therefore, at 30 gpm/lamp, the credited RED is:
REDvai = 18.37(1 + (6.071/18.3)) = 13.74 ml/cm2
Solutions for credited RED are also shown on Figure 8-2. As shown, a hydraulic loading
equal to or less than 49.7 gpm/lamp will yield an RED of at least 12.5 mJ/cm2 in a single reactor;
two reactors would be placed in series for a credited RED of 25 mJ/cm2. At the design flow of
6000 gpm, each bank would require 6000/49.7 = 120 lamps. This may be accommodated by two
parallel channels, with at least 60 lamps per reactor. This analysis is simplified as an example
and does not address redundancy or other design considerations.
8-3
-------
60
55
50
45
Calculated Dose
Credited Dose
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70
Flow Rate/Lamp (gpm/Lamp)
Figure 8-2. Example calculation of RED as a function of flow (55% UVT)
for an H-4XE-HO (2W-1B-1C) Reactor module in a low-dose application.
8.1.2 Application 2
In the second application, the performance requirement is to meet an MS2 RED of 80
ml/cm2, a criterion typically found with reuse applications after membrane-filtered secondary
treatment. The approach is the same as discussed above for the "low-dose" application, except
that an MS2 UV sensitivity value is used. This is 20 mJ/cm2/LI. Solutions for calculated and
credited RED are provided in Figure 8.3. In this case, two reactors are placed in series, with a
hydraulic loading of 19.85 gpm/lamp. To meet the design flow of 6000 gpm, approximately 303
lamps are needed for each reactor. This would likely be divided to two or three parallel
channels. Note that this is provided as a simplified example - other design aspects such as
redundancy are not considered.
8-4
-------
60
55
Calculated Dose
Credited Dose
10 15 20 25 30 35 40 45 50 55 60 65 70
Flow Rate per Lamp (gpm/Lamp)
Figure 8-3. Example calculation of RED as a function of flow (55% UVT) for an H-
4XE-HO (2W-1B-1C) Reactor module in a reuse application.
8-5
-------
SECTION 9
REFERENCES
1. HydroQual, INC., (January 2002). "Generic Verification Protocol for Secondary Effluent
and Water Reuse Disinfection Applications" version 3.4. Prepared for NSF International and
the U.S. Environmental Protection Agency under the Environmental Technology Verification
Program, Source Water Projection Pilot. Mahwah, NJ
2. ISO 10705-1: International Standards Organization (ISO). (1995). "Water Quality-
Detection and Enumeration of Bacteriophage. Part I: Enumeration of F-Specific RNA
Bacteriophage." Switzerland: International Standards Organization
3. National Water Research Institute (NWRI)/AWWA Research Foundation (AwwaRF)
Ultraviolet Disinfection Guidelines for Drinking Water and Water Reuse, Second Edition.
Fountain Valley, CA, 2003.
4. USEPA - Environmental Technology Program, Verification Protocol for Secondary Effluent
and Water Reuse Disinfection Applications, NSF International Water Quality Center,
October 2002
5. USEPA Ultraviolet Disinfection Guidance Manual for the Final Long Term 2 Enhanced
Surface Water Treatment Rule, United States Environmental Protection Agency, Office of
Water, EPA-815-R-06-007, November, 2006
9-1
------- |