November, 2013
NSF13/40/EPADWCTR
EPA/600/R-13/250
Environmental Technology
Verification Report
Reduction of Microbial Contaminants in
Drinking Water by Ultraviolet Technology
ETS UV Technology
ETS UV Model UVL-200-4
Prepared by
NSF International
Under a Cooperative Agreement with
©ERA U.S. Environmental Protection Agency
-------
NOVEMBER 2013
Environmental Technology Verification Report
Reduction of Microbial Contaminants in Drinking Water by
Ultraviolet Light Technology
ETS UV Technology
(A joint venture of Engineered Treatment Systems and atg UV
Technology)
ETS UV MODEL UVL-200-4
Prepared by:
NSF International
Ann Arbor, Michigan 48105
Under a cooperative agreement with the U.S. Environmental Protection Agency
Jeffrey Q. Adams, Project Officer
National Risk Management Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
NOVEMBER 2013
Notice
The U.S. Environmental Protection Agency, through its Office of Research and Development,
funded and managed, or partially funded and collaborated in, the research described herein. It
has been subjected to the Agency's peer and administrative review and has been approved for
publication. Any opinions expressed in this report are those of the author(s) and do not
necessarily reflect the views of the Agency, therefore, no official endorsement should be inferred.
Any mention of trade names or commercial products does not constitute endorsement or
recommendation for use.
11
-------
NOVEMBER 2013
Table of Contents
Verification Statement VS-i
Title Page i
Notice ii
Table of Contents iii
List of Tables iv
List of Figures v
Abbreviations and Acronyms v
Chapter 1 1
Introduction 1
1.1 ETV Program Purpose and Operation 1
1.2 Purpose of Verification 1
1.3 Verification Test Site 2
1.4 Testing Participants and Responsibilities 2
Chapter 2 4
Equipment Description 4
2.1 General Information ETS UV Technology 4
2.2 ETS Model UVL-200-4 UV System Description 4
2.3 ETS UV Model UVL-200-4 Specifications and Information 6
Chapters 8
Methods and Procedures 8
3.1 Introduction 8
3.2 UV Sensors Assessment 9
3.3 Headloss Determination 10
3.4 Power Consumption Evaluation 10
3.5 Feed Water Source and Test Rig Setup 10
3.6 Installation of Reactor and Lamp Burn-in 13
3.7 Collimated Beam Bench Scale Testing 13
3.8 Full Scale Testing to Validate UV dose 18
3.9 Analytical Methods 22
3.10 Full Scale Test Controls 24
3.11 Power Measurements 25
3.12 Flow Rate 25
3.13 Evaluation, Documentation and Installation of Reactor 25
Chapter 4 27
Results and Discussion 27
4.1 Introduction 27
4.2 Sensor Assessment 27
4.3 Collimated Beam Dose Response Data 28
4.4 Development of Dose Response 29
4.5 MS and Operational Flow Test Data 41
4.6 Set Line for a Minimum RED of 40 ml/cm2 47
4.7 Deriving the Validation Factor and Log Credit for Cryptosporidium 48
iii
-------
NOVEMBER 2013
4.8 Validated Dose (REDVai) for MS2 as the Target Organism 55
4.9 Water Quality Data 57
4.10 Headless 61
4.11 Power Measurement 61
Chapters 62
Quality Assurance/Quality Control 62
5.1 Introduction 62
5.2 Test Procedure QA/QC 62
5.3 Sample Handling 62
5.4 Chemistry Laboratory QA/QC 62
5.5 Microbiology Laboratory QA/QC 62
5.6 Engineering Lab - Test Rig QA/QC 64
5.7 Documentation 65
5.8 Data Review 65
5.9 Data Quality Indicators 67
Chapter 6 69
References 69
Appendices
Attachment 1 Model UVL-200-4 Operating Manual and Technical Data
Attachment 2 Sensor Certificates and Sensor Information
Attachment 3 Standard 55 Annex A - Collimated Beam Apparatus and Procedures
Attachment 4 UVT Scans of Feed Water
List of Tables
Table 2-1. Basic UV Chamber Information 6
Table 2-2. Low Pressure Lamp Information 6
Table 2-3. UVLamp Sleeve Information 6
Table 2-4. UV Sensor Information 7
Table 3-1. Test Conditions for Validation 20
Table 3-2. Analytical Methods for Laboratory Analyses 22
Table 4-1. Sensor Assessment Data First Set of Test Runs (June 2012) 28
Table 4-2. UV Dose Response Data from Collimated Beam Tests at 79% UVT (June 2012) ....31
Table 4-3. UV Dose Response Data from Collimated Beam Tests at 97% UVT (June 2012) ....33
Table 4-4. UV Dose Response Data from Collimated Beam Tests at 79% UVT with Outlier
Removed (June 2012) 35
Table 4-5. ETS UV Model UVL-200-4MS2 Operational Data 42
Table 4-6. ETS UV Model UVL-200-4 MS2 Concentration Results 43
Table 4-7. ETS UV Model UVL-200-4 MS2 Log Concentration for Influent and Effluent
Samples 44
Table 4-8. ETS UV Model UVL-200-4 MS2 Log Inactivation Results 45
iv
-------
NOVEMBER 2013
Table 4-9. ETS UV Model UVL-200-4 MS2 Observed RED Results 46
Table 4-10. RED Bias Factor for Each Set Point for Cryptosporidium 49
Table 4-11. Uncertainty of the Validation (Uvai) and BRED Values for Cryptosporidium 52
Table 4-12. Validation Factors and Validated Dose (REDVai) for Cryptosporidium 53
Table 4-13. Validation Factors and Validated Dose (REDVai) based onMS2 56
Table 4-14. Temperature and pH Results 58
Table 4-15. Total Chlorine, Free Chlorine, and Turbidity Results 58
Table 4-16. Iron and Manganese Results 59
Table 4-17. HPC, Total Coliform, and E. coli Results 60
Table 4-18. Headloss Data 61
Table 4-19. Power Measurement Results 61
Table 5-1. Trip Blank Results 64
Table 5-2. MS2 Stability Test Results 64
Table 5-3. Flow Meter Calibration Results 65
Table 5-4. Reactor Control and Reactor Blank MS2 Results 66
Table 5-5. Completeness Requirements 68
List of Figures
Figure 2-1. ETS UV Model UVL-200-4 5
Figure 3-1. Schematic ofNSF test rig 12
Figure 3-2. Photograph of the Model UVL-200-4 Test Setup 13
Figure 4-1. Collimated beam dose versus log N UVT 79% with outlier removed (June 2012)....37
Figure 4-2. Collimated beam dose versus log N UVT 97% (June 2012) 38
Figure 4-3. Dose response - log I versus dose - UVT 79% with outlier removed (June 2012) 39
Figure 4-4. Dose response - log I versus dose - UVT 97% (June 2012) 40
Figure 4-5. Set Line for Model UVL-200-4 For Validated Dose of >40 mJ/cm2 based on MS247
Figure 4-6. Set line for Minimum 3.0 log Cryptosporidium Inactivation for ETS UV Model UVL-
200-4 54
Figure 4-7. Set line for Minimum 40 mJ/cm2 Validated Dose (REDVai) based on MS2 for ETS
UVModelUVL-200-4 56
Abbreviations and Acronyms
A254 Absorbance at wavelength 254 nm
ASTM American Society of Testing Materials
ATCC American Type Culture Collection
ATG atg UV Technology
C degrees Celsius
CFU Colony Forming Units
cm Centimeter
DWS Drinking Water Systems
o
-------
NOVEMBER 2013
DVGW
EPA
ETS
ETSUV
ETV
°F
gpm
in
h
HPC
L
Ibs
LIMS
log I
LSA
LT2ESWTR
m
min
ml
mg
mL
MS2
NaOH
ND
NIST
nm
NRMRL
NSF
NTU
ONORM
ORD
pfu
Protocol
psig
QA
QC
QA/QC
QAPP
QMP
RED
RED
val
Deutscher Verein des Gas- und Wasserfaches e.V. - Technisch -
wissenschaftlicher Verein - German Technical and Scientific Association
for Gas and Water
U. S. Environmental Protection Agency
Engineered Treatment Systems
ETS UV Technology -joint venture of ETS and atg
Environmental Technology Verification
Degrees Fahrenheit
gallons per minute
inch(es)
hours
Heterotrophic Plate Count
Liter
pounds
Laboratory Information Management System
log base 10 Inactivation
Sodium Lignin Sulfonic Acid
Long Term 2 Enhanced Surface Water Treatment Rule
meter
minute
milli-joules
Milligram
Milliliter
MS2 coliphage ATCC 15597 Bl
Sodium Hydroxide
Non-Detect
National Institute of Standards and Technology
Nanometer
National Risk Management Research Laboratory
NSF International (formerly known as National Sanitation Foundation)
Nephelometric Turbidity Unit
Osterreichisches Normungsinstitut Austria Standard
Office of Research and Development
Plaque Forming Units
Generic Protocol
Pounds per Square Inch, gauge
Quality Assurance
Quality Control
Quality Assurance/Quality Control
Quality Assurance Project Plan
Quality Management Plan
Reduction Equivalent Dose
Measured Reduction Equivalent Dose - from test runs
Validated Reduction Equivalent Dose - based on selected pathogen and
uncertainty
VI
-------
NOVEMBER 2013
RPD
SM
SOP
SPt
Tl
T7
TQAP
IDS
ISA
TSB
UVT
ug
jam
UVDGM-2006
USEPA
UDR
USP
Us
UVAL
Relative Percent Deviation
Standard Methods for the Examination of Water and Wastewater
Standard Operating Procedure
Set Point Condition
BacteriophageTl strain
Bacteriophage T7 strain
Test / Quality Assurance Plan
Total Dissolved Solids
Tryptic Soy Agar
Tryptic Soy Broth
ultraviolet transmittance
microgram
microns
Ultraviolet Disinfection Guidance Manual - 2006
U. S. Environmental Protection Agency
uncertainty of collimated beam data
uncertainty of set point
uncertainty of sensor
uncertainty of validation
vn
-------
NOVEMBER 2013
Chapter 1
Introduction
1.1 ETV Program Purpose and Operation
The U.S. Environmental Protection Agency (USEPA) has created the Environmental Technology
Verification (ETV) Program to facilitate the deployment of innovative or improved
environmental technologies through performance verification testing and dissemination of
information. The goal of the ETV Program is to further environmental protection by
accelerating the acceptance and use of improved and more cost-effective technologies. ETV
seeks to achieve this goal by providing high-quality, peer-reviewed data on technology
performance to those involved in the design, distribution, permitting, purchase, and use of
environmental technologies.
ETV works in partnership with recognized standards and testing organizations; with stakeholder
groups consisting of buyers, vendor organizations, and permitters; and with the full participation
of individual technology developers. The program evaluates the performance of innovative
technologies by developing test plans that are responsive to the needs of stakeholders;
conducting field or laboratory testing, collecting and analyzing data; and by preparing peer-
reviewed reports. All evaluations are conducted in accordance with rigorous quality assurance
protocols to ensure that data of known and adequate quality are generated and that the results are
defensible.
The USEPA has partnered with NSF International (NSF) under the ETV Drinking Water
Systems Center (DWS) to verify performance of drinking water treatment systems that benefit
the public and small communities. It is important to note that verification of the equipment does
not mean the equipment is "certified" by NSF or "accepted" by USEPA. Rather, it recognizes
that the performance of the equipment has been determined and verified by these organizations
under conditions specified in ETV protocols and test plans.
1.2 Purpose of Verification
The purpose of the ETV testing was to validate, using the set line approach, the UV dose
delivered by the ETS UV Technology (ETS UV) Model UVL-200-4 Water Purification System
(Model UVL-200-4} as defined by these regulatory authorities and their guidelines and
regulations:
• Water Supply Committee of the Great Lakes-Upper Mississippi River Board of State and
Provincial Public Health and Environmental Managers otherwise known as The Ten
States Standards 2012;
• The Norwegian Institute of Public Health (NIPH) and its guidelines; and
• The New York Department of Health (NYDOH) and its code.
Another purpose was to use the same data set to calculate the log inactivation of a target
pathogen such as Cryptosporidium using the Generic Protocol for Development of Test / Quality
Assurance Plans for Validation of Ultraviolet (UV) Reactors, August 2011 10/01/EPADWCTR
(GP-2011) which is based on Ultraviolet Design Guidance Manual For the Long Term 2
1
-------
NOVEMBER 2013
Enhanced Surface Water Treatment Rule, Office of Water, US Environmental Protection
Agency, November 2006, EPA 815-R-06-007 (UVDGM-2006).
The setline approach was based on validation testing at three set points (a set point is defined as a
single flow rate and irradiance output that delivers the targeted UV dose). The results of the three
set point tests were used to develop a setline that defines the maximum flow rate - minimum
irradiance output required to ensure the UV dose is achieved. The microorganism used for this
validation test was MS2 coliphage virus (MS2). The target UV dose was a measured Reduction
Equivalent Dose (REDmeas) of >40 ml/cm2. This dose was calculated based on the understanding
of dose calculations used internationally and by the Ten States Standards. The REDmeas was then
adjusted based on the uncertainty of the measurements to calculate a MS2 based validated dose
(REDyai) where the RED bias is set equal to one (1.0) in accordance with the unique approach of
the State of New York. The REDmeas data were also adjusted for uncertainty and the
Cryptosporidium RED bias factors from the UVDGM-2006 Appendix G. The data were used to
estimate the log inactivation of Cryptosporidium so that a regulatory agency could grant log
credits under the USEPA's Long Term 2 Enhanced Surface Water Treatment Rule
(LT2ESWTR).
ETS UV selected flow rates of 15, 20, and 25 gpm as the target flow rates based on their system
design for Model UVL-200-4.
Based on the result of the three set points, a setline was developed for this unit. During full-scale
commercial operation, federal regulations require that the UV intensity as measured by the UV
sensor(s) must meet or exceed the validated intensity (irradiance) to ensure delivery of the
required dose. Reactors must be operated within the validated operating conditions for maximum
flow rate - minimum irradiance combinations, UVT, and lamp status [40 CFR 141.720(d)(2)].
Under the UV setline approach, UV Transmittance (UVT) does not have to be measured
separately. The intensity readings by the sensor take into account changes in the UVT and the
setline establishes the operating conditions over a range of flow rates used during the validation
test.
This verification test did not evaluate cleaning of the lamps or quartz sleeves, nor any other
maintenance and operational issues.
1.3 Verification Test Site
UV dose validation testing was performed at the NSF Testing Laboratory in Ann Arbor,
Michigan. The NSF laboratory performs all of the testing activities for NSF certification of
drinking water treatment systems, and pool and spa treatment systems.
1.4 Testing Participants and Responsibilities
The following is a brief description of each of the ETV participants and their roles and
responsibilities.
-------
NOVEMBER 2013
1.4.1 NSF International
NSF is an independent, not-for-profit organization dedicated to public health and safety, and to
protection of the environment. Founded in 1944 and located in Ann Arbor, Michigan, NSF has
been instrumental in the development of consensus standards for the protection of public health
and the environment. The USEPA partnered with NSF to verify the performance of drinking
water treatment systems through the USEPA's ETV Program.
NSF performed all verification testing activities at its Ann Arbor, MI location. NSF prepared the
test/QA plan (TQAP), performed all testing, managed, evaluated, interpreted, and reported on the
data generated by the testing, and reported on the performance of the technology.
Contact: NSF International
789 N. Dixboro Road
Ann Arbor, MI 48105
Phone: 734-769-8010
Contact: Mr. Bruce Bartley, Project Manager
Email: bartley@nsf.org
1.4.2 U.S. Environmental Protection Agency
USEPA, through its Office of Research and Development (ORD), has financially supported and
collaborated with NSF under Cooperative Agreement No. R-82833301. This verification effort
was supported by the DWS Center operating under the ETV Program. This document has been
peer-reviewed, reviewed by USEPA, and recommended for public release.
1.4.3 ETS UV Technology
ETS UV supplied the UV test unit for testing, required reference sensors, detailed specifications
on the equipment, UV lamps, lamp sleeves, and duty sensors, and written and verbal instructions
for equipment operation. ETS UV also provided logistical and technical support, as needed.
Contact: Engineered Treatment Systems, LLC
P.O. Box 392
W9652 Beaverland Parkway
Beaver Dam, Wisconsin
Phone: 1-877-885-4628
Email: info@ets-uv.com
atg UV Technology
Genesis House, Richmond Hill
Pemberton
Wigan, WN5 8AA
United Kingdom
Phone: +44(0) 1942216161
Website: www.atguv.com
-------
NOVEMBER 2013
Chapter 2
Equipment Description
2.1 General Information ETS UV Technology
ETS UV was founded in January 2005 in a joint venture between atg UV Technology (atg) and
Engineered Treatment Systems (ETS) to accommodate the growing demand for ultraviolet
disinfection and photolysis across the US pools and recreational water markets. Systems
are manufactured at the Beaver Dam production facility located in Beaver Dam, Wisconsin.
Production of ultraviolet disinfection systems for the US market began in January 2008. In
2009, the second phase of ETS UV became operational. Based in Ohio, ETS UV Industrial &
Municipal offers low and medium pressure UV systems for municipal drinking water,
wastewater and industrial UV treatment applications.
The atg UV Technology companyis based in the North West of England, serving an international
customer base. Since being founded in 1981 as Willand UV System, atg indicated that they have
served a number of markets including municipal drinking water and wastewater
disinfection, industrial processes and manufacturing, offshore and marine industries and
swimming pool applications.
ETS is based in Beaver Dam Wisconsin. ETS states that it has over three decades of
experience and over 1500 successful case studies in the custom design and production of UV
disinfection systems for a range of applications.
2.2 ETS Model UVL-200-4 UV System Description
The ETS UV Water Purification System that was validated in this test is the Model UVL-200-4.
This unit is rated by ETS UV for a maximum flow rate of 55 gpm. The system uses 1 low-
pressure lamp and one intensity sensor mounted in a stainless flow chamber. Figure 2-1 presents
a picture of the system. Additional specifications for the unit are presented below. ETS UV
provided an operating manual and a technical data book, which included schematics and tables
with parts and dimensions for the reactor, the sensors, the lamps and the quartz sleeves. All
specifications and information were provided to NSF by ETS UV in advance of the testing. ETS
also provided additional information for the UV sensor (spectral data, measuring angle,
measuring range, and output range) and for UV lamps (lamp life, irradiance output, power
requirements, aging data, etc.) as required for the validation test.
NSF performed a normal technical review of the sensor specifications, UV lamp and quartz
sleeve specifications, and general review of the reactor chamber and overall system as required
bytheGP-2011.
The operating manual, technical book and other supplemental specifications for the sensor, lamp,
quartz sleeve, and control system provided by ETS UV are included in Attachments 1 and 2 of
this report for reference.
-------
NOVEMBER 2013
Figure 2-1 ETS UV System UVL-200-4
-------
NOVEMBER 2013
2.3 ETS UV Model UVL-200-4 Specifications and Information
ETS UV has provided the following information about their UV reactor:
Table 2-1. Basic UV Chamber Information
Manufacturer/Supplier
Type or model
Description
Year of manufacture
Maximum flow rate
Net dry weight
Volume
Electrical Power
Operating Power consumption
Maximum Pressure
Ambient water temperature
Max Cleaning Temperature
Inlet pipe size
ETSUV
Model UVL-200-4
Single Lamp Low Pressure UV
System
Disinfection
2008 and onwards
55gpm
52.141bs
0.2637 cubic feet
2 phase 220 VAC, 60Hz; 2 amp
earth ground.
single pole,
< 400 watts
10 bar
32 to 113 degrees °F
158 degrees °F (unit turned off)
2 inch
Table 2-2. Low Pressure Lamp Information
Type
Model
Number of lamps per reactor
UV emission at wavelengths ranging from
240-290 nm
Lamp Life
Power supply unit's name, make and serial
numbers
Ballast
Irradiance @lm (W/cm)
UV Output (W)
Operating Lamp Watts (W)
Lamp Current and Voltage
Arc length (mm)
Low-pressure
LP200-SS19
1
See Lamp spectral graph in Attachment
1.
12,000 hrs
EVG - Ziegler Electronic Devices Gmbh
EVG160-200W/ 2A Electronic Ballast
Magnetic Choke with Starter
100
200
180
2.5 Amps; 90 Volts
1058
Table 2-3. UV Lamp Sleeve Information
Type or model
Quartz material
Pressure resistance (kPa)
GE 214 Clear Fused Quartz
Clear Fused Quartz
7000
-------
NOVEMBER 2013
Table 2-4. UV Sensor Information
Type / model
Measuring field angle
Number of sensors per reactor and placement
Signal output range in mA (mV)
Measuring range- W/m2 Output signal
UV-Technik SUV20.1
A2Y2C
'o' Norm 160 degree
1
4 - 20 mA
0 - 20 W/m2
Additional UV sensor spectral information provided by ETS UV prior to the start of testing
demonstrated the sensor met the requirements of the Generic Protocol for Development of
Test/Quality Assurance Plans for Validation of Ultraviolet (UV) Reactors, NSF International,
7/2010 (GP-2010) and the GP-2011. The GP-2010 and the updated GP-2011 are based on the
USEPA's UVDGM-2006 requirements. The sensor meets the GP-2010 and GP-2011 requirement
that >90% of the response is between 200 - 300 nm. The sensor information is included in
Attachment 2.
-------
NOVEMBER 2013
Chapter 3
Methods and Procedures
3.1 Introduction
A Test Quality Assurance Plan (TQAP) was prepared to detail the experimental design for this
validation work. The experimental design was based on the GP-2010 and GP-2011 as derived
from the USEPA's UVDGM-2006. The TQAP is available from NSF upon request.
The approach used to validate UV reactors is based on biodosimetry which determines the log
inactivation of a challenge microorganism during full-scale reactor testing for specific operating
conditions of flow rate, UV transmittance (UVT), and UV intensity (measured by the duty
sensor). A dose-response equation for the challenge microorganism (MS2 coliphage for this test)
is determined using a collimated beam bench-scale test. The observed log-inactivation values
from full-scale testing are input into the collimated beam derived-UV dose-response equations to
estimate a "Reduction Equivalent Dose (REDmeas)". The REDmeas value can then be adjusted for
uncertainties and biases to produce a validated dose (REDyai) for the reactor for the specific
operating conditions tested.
The methods and procedures were designed to accomplish the primary objective of the validation
test of the Model UVL-200-4, which was to develop a set line based on three set points (each set
point is a specific flow rate-UV intensity combination) that would ensure a REDmeas of at least
40mJ/cm2 based on MS2 as defined by the Ten States Standards 2012. Test procedures were
also designed so that the REDmeas could be adjusted based on the uncertainty of the
measurements to calculate a MS2 based validated dose (REDyai) in accordance with the unique
approach of the State of New York. The REDmeas data were also adjusted for uncertainty and the
Cryptosporidium RED bias factors from the UVDGM-2006 Appendix G.
The GP-2010 required the use of a second less sensitive challenge organism as part of the
validation. The bacteriophage "T7" was initially included in the GP-2010 as a result of research
suggesting it could be a surrogate test microorganism with UV sensitivity similar to the UV
sensitivity of Cryptosporidium (Fallen et.al, JAWWA, 99.3, March 2007). The GP-2010
technical advisory panel had reservations about using any test microorganism other than MS2
which has an excellent record of quality control response for collimated beam regression curves
(Figure A.I in the UVDGM-2006). The ETV GP-2010 technical advisory panel opinion was that
other test microorganisms simply did not yet have the record of quality control limits as did
MS2.
In 2010 during some initial validation studies, NSF attempted to use the bacteriophage T7. The
strain referenced by the JAWWA study (ATCC 11303-B7) was not available through ATCC. In
fact ATCC said verbally that the strain mentioned was not in fact T7 and not available. With the
counsel of the EPA, NSF agreed to try bacteriophage T7 ATCC strain BAA-1103-B38.
Comments in 2011 on the GP-2010 also provided reasons not to specify only T7: "However, T7
cannot be produced at nearly as high a titer as Tl, so in the validation of high-flow reactors,
replacing all the bacteriophage Tl test conditions with T7 test conditions would consume an
unacceptable volume of raw phage stock." Consequently the GP-2010 technical advisory panel
-------
NOVEMBER 2013
recommended the use of any organism other than MS2 will be optional and the use of MS2 will
be mandatory for all types of reactors. The use of a challenge organism other than MS2 will be
determined by the consensus of stakeholders.
For the retesting done for this project, NSF chose to only use MS2 based on the concerns raised
about T7 by reviewers and the changes made in the 2011 ETV UV Protocol. Instead, it was
decided to illustrate how MS2 data were being used to satisfy many different regulatory
requirements while using essentially the same data. The basic biodosimetry data were used to
calculate the log inactivation of Cryptosporidium, the 40mJ/cm2 dose (REDmeas) requirement
found in the Ten States Standards 2012 and the NIPH guidelines, and the "validated" dose
approach (REDVai) based on MS2 used by the NYDOH.
UV reactor validation included:
1. Obtain the technical specifications for the system as provided by ETS UV
2. Assessment of the UV sensors
3. Collimated beam laboratory bench scale testing
4. Full scale reactor testing
5. Calculations to determine the RED
6. Adjust the RED for uncertainty in UV dose and calculate a validated dose for
Cryptosporidium
The target UV dosage validated was a REDmeas of 40 mJ/cm2, based on MS2. ETS UV selected
flow rates of 15, 20, and 25 gpm as the target flow rates based on their system design for Model
UVL-200-4 and the results of screening tests and initial data from 2010.
3.2 UV Sensors Assessment
The Model UVL-200-4 duty sensor was evaluated according to the UV sensor requirements in
the GP-2010 and GP-2011 prior to the verification testing. All UV intensity sensors (the duty
and two reference sensors) were new sensors and specifications provided with the sensors
showed they were designed in accordance with the DVGW guideline W 294 (June, 2006) and the
ONORM M5873-1 standard (June, 2002), respectively. Evidence of calibration of the sensors
within the last 12 months, traceable to a standard of the Physikalisch Technische Bundesanstalt
(PTB) in Braunschweig, was provided by ETS UV as provided to them by the sensor
manufacturer (uv-technik).
The validation testing requires confirmation of the duty sensor spectral response to assess
whether the sensors are germicidal (see UVDGM-2006 Glossary for the definition of germicidal)
with a defined spectral response of at least 90% between 200 and 300 nm. The technical
specifications of the ETS UV sensor and representation of sensitivity to the germicidal
wavelength was provided by ETS UV and found to meet the requirements. The technical
specifications of the ETS UV sensor and representation of sensitivity to the germicidal
wavelength is included in Attachment 2.
-------
NOVEMBER 2013
During validation testing, the duty UV sensor measurement was compared to two reference
sensor measurements to assure the duty sensor was within 10% of the average of the two
reference sensor measurements.
The following steps were used to check the uncertainty of the duty and reference UV sensors.
The sensors were checked before and after the validation testing.
1. Step 1: Water was passed through the reactor at the maximum UV transmittance (UVT)
and the maximum lamp power setting to be used during validation testing.
2. Step 2: Using two recently calibrated (at a minimum annually) reference UV sensors,
each reference sensor was installed on the UV reactor at the sensor port. The UV
intensity was measured and recorded.
Step 2 was repeated using the duty UV sensor.
3. Step 3: Steps 1 and 2 were repeated at maximum UVT and lamp power decreased to the
minimum level expected to occur during validation testing.
4. Step 4: For a given lamp output and UVT value, the difference between the reference and
duty UV sensor measurements were calculated as follows:
The absolute value of [(S duty/ S AvgRef) - 1]
where:
S duty = Intensity measured by a duty UV sensor,
S Avg Ref = Average UV intensity measured by all the reference UV sensors in the
same UV sensor port with the same UV lamp at the same UV lamp power.
3.3 Headloss Determination
Headloss through the unit was determined over the range of expected flow rates, in this case
from 10 gpm to 25 gpm. The inlet pressure near the inlet flange and the outlet pressure near the
outlet flange were measured at several flow rates. Measurements were recorded for flow rates of
5, 10, 15, 20, 25 gpm. These data are reported in Section 4.10.
3.4 Power Consumption Evaluation
The amperage and voltage used by the unit were measured during all reactor test runs.
Power data are presented in Section 4.11.
3.5 Feed Water Source and Test Rig Setup
The water source for this test was City of Ann Arbor Michigan municipal drinking water. The
water was de-chlorinated using activated carbon, as confirmed by testing in the laboratory. For
the lowered UVT conditions, the chemical Sodium Lignin Sulfonic Acid (LSA) was used to
lower the UV transmittance to the UVTs of <79%, <90% and <94%. LSA was added to the
10
-------
NOVEMBER 2013
supply tank before each set of the lowered UVT runs and was well mixed using a recirculating
pump system. UVT was measured continuously using an in-line UVT meter (calibrated daily) to
confirm that proper UVT was attained. UVT measurements were also confirmed by the
collection of samples during each test run and analysis by a bench top spectrophotometer.
NSF used a UV test rig and system setup that is designed to conform to the specifications as
described in the GP-2011 and UVDGM-2006. Figure 3-1 shows a basic schematic of the NSF
test rig and equipment setup. The schematic is reproduced for informational purposes and is
copyright protected. A photograph of the actual setup is shown in Figure 3-2.
The feed water pump to the test unit was a variable speed pump. Flow rate was controlled by
adjusting the power supplied to the pump and by a control valve. A magnetic water flow meter
was used to monitor flow rate. The meter was calibrated and easily achieved the required
accuracy of + 5%. A chemical feed pump (injector pump) was used to inject MS2 coliphage
upstream of an inline static mixer. The inline mixer ensured sufficient mixing of the
microorganism prior to the influent sampling port, which was located upstream of the 90° elbow
installed directly on the inlet to the unit. The effluent sampling port was located downstream of a
90° elbow that was installed directly on the outlet port of the unit and downstream of a second in-
line mixer. This use of an in-line mixer met the UVDGM-2006 requirement to ensure good
mixing of the treated water prior to the effluent sampling port.
11
-------
NOVEMBER 2013
Fluid Flow
Dosing Pump
Note: Al plumbinq is schedule 80 PVC
8" 50 HP
Centrifugal Pump,
VFD Controlled
Figure 3-1 Schematic of NSF test rig(<
12
-------
NOVEMBER 2013
Figure 3-2 Photograph of the Model VVL-200-4 Test Setup
3.6 Installation of Reactor and Lamp Burn-in
The UV reactor and the reactor inlet and outlet connections were installed at the NSF laboratory
in accordance with the ETS UV installation and assembly instructions. Two 90 degree elbows,
one upstream and one downstream of the unit, were used in the test rig setup to eliminate stray
UV light. Figure 3-2 shows a photograph of the test rig setup, which conforms with the GP-2011.
The UV lamp was new and therefore the system was operated for 100 hours with the lamps
turned on at full power prior to the start of the test.
There is one duty sensor and one lamp in the Model UVL-200-4. Therefore, the lamp positioning
check requirements (checking each lamp and placing the lowest output lamp closest to the
sensor) were not required for this validation.
3.7 Collimated Beam Bench Scale Testing
The collimated beam procedure involves placing a sample collected from the test rig and
containing MS2 in a petri dish and then exposing the sample to collimated UV light for a
predetermined amount of time. The UV dose is calculated using the measured intensity of the
13
-------
NOVEMBER 2013
UV light, UV transmittance of the water, and exposure time. The measured concentration of
microorganisms before and after exposure provides the "response," or log inactivation of the
microorganisms from exposure to UV light. Regression analysis of measured log inactivation for
a range of UV doses produces the dose-response curve.
Appendix C of the UVDGM-2006 provides guidance on how to conduct the collimated beam
bench-scale testing and to produce a UV dose-response curve. Based on the UVDGM-2006
guidance, the following sections describe the details of the collimated beam testing.
3.7.1 Test Microorganism (Challenge)
MS2 coliphage ATCC 15597-B1 was used in collimated beam bench scale testing and for the
full-scale reactor dose validation tests. MS2 coliphage ATCC 15597-B1 is a recommended
microorganism for UV lamp validation tests. Further reasons for selecting this microorganism
for UV validation are based on its inter-laboratory reproducibility (UVDGM-2006), ease of use
and culturing, and demonstrated performance of MS2 in validation testing.
3.7.2 Test Conditions
The collimated beam tests were performed in duplicate at the minimum and maximum UVT test
conditions. For this validation the testing occurred over two days. The lowered UVT test runs
were performed on the first day. The intensity readings at each UVT (79%. 90%, 94%) were
recorded during test run with full lamp power. Collimated beam tests were run on the minimum
UVT water (79%) with duplicate runs being performed. On the second day using high UVT
water (97%), the power was reduced to achieve the same intensity as measured for each of the
lowered UVT waters on day one. The Model UVL-200-4 does not have a variable power control
as part of the normal control system. ETS UV provided a variable power controller that allowed
the lamp power to be lowered to achieve the measured intensities in water for the test run. .
Collimated beam tests were run on day two on the high UVT water (97%) with duplicate runs
being performed. Thus, for this validation test, there are two sets of duplicate collimated beam
test data, one at lowest UVT (79%) and one at the high UVT (water not adjusted with LSA).
UV doses covered the range of the targeted RED dose, which in this case is 40mJ/cm2. UV
doses were set at 0, 20, 30, 40, and 60 and 80 mJ/cm2. The samples are clustered close to the
40mJ/cm2 target dose with two doses above and below the target of 40 mJ/cm2.
The collimated beam radiometers were calibrated to ensure that the measured UV intensity met
the criteria of an uncertainty of 8 percent or less at a 95-percent confidence level.
3.7.3 Test Apparatus
NSF uses a collimated beam apparatus that conforms to NSF/ANSI Standard 55 section 7.2.1.2.
and the UVDGM-2006. A description of the apparatus is presented in NSF/ANSI Standard 55®
Annex A, which is presented in Attachment 3.
14
-------
NOVEMBER 2013
3.7.4 Collimated Beam Procedure
NSF collected two (2) one liter samples from the influent sampling port of the test rig for
collimated beam testing. Each bottle was used for one of the replicates for the collimated beam
test. The MS2 spiked water was collected directly from the test rig each day during the test runs.
Therefore, the collimated beam test water and microorganism culture was the same as used in the
full scale reactor tests.
NSF microbiological laboratory personnel followed the "Method for Challenge Microorganism
Preparation, Culturing the Challenge Organism and Measuring its Concentration" in Annex A of
NSF/ANSI Standard 55. Please note that all reproduced portions of NSF/ANSI Standards are
copyright protected.
For collimated beam testing of a water sample containing challenge microorganisms, NSF's
laboratory followed this procedure:
1. Measure the A254of the sample.
2. Place a known volume from the water sample into a petri dish and add a stir bar.
Measure the water depth in the petri dish.
3. Measure the UV intensity delivered by the collimated beam with no sample present
using a calibrated radiometer using a calibrated UV sensor. The UV sensor is placed at
the same distance from the radiometer as the sample surface.
4. Calculate the required exposure time to deliver the target UV dose described in the
next section.
5. Block the light from the collimating tube using a shutter or equivalent.
6. Center the petri dish with the water sample under the collimating tube.
7. Remove the light block from the collimating tube and start the timer.
8. When the target exposure time has elapsed, block the light from the collimating tube.
9. Remove the petri dish and collect the sample for measurement of the challenge
microorganism concentration. Analyze immediately or store in the dark at 4 °C (for up
to 6 hours). Multiple dilutions are used to bracket the expected concentration range
(e.g. sample dilutions of 10X, 100X, 1000X). Plate each dilution in triplicate and
calculate the average microbial value for the dilution from the three plate replicates
that provide the best colony count.
10. Re-measure the UV intensity and calculate the average of this measurement and the
measurement taken in Step 3. The value should be within 5 percent of the value
measured in Step 3. If not, recalibrate radiometer and re-start at Step 1.
11. Using the equation described in the next section, calculate the UV dose applied to the
sample based on experimental conditions. The calculated experimental dose should be
similar to the planned target dose.
12. Repeat Steps 1 through 11 for each replicate and target UV dose value. Repeat all
steps for each water test condition replicate.
The UV dose delivered to the sample is calculated using the following equation:
DCB = Es * Pf * (1-R) * [L* (1-10-A254 '")/(d + L> A254* d * ln(10)] * t
15
-------
NOVEMBER 2013
Where:
DCB = UV dose (ml/cm2)
Es = Average UV intensity (measured before and after irradiating the sample)
(mW/cm2)
Pf = Petri Factor (unitless)
R = Reflectance at the air-water interface at 254 nm (unitless)
L = Distance from lamp centerline to suspension surface (cm)
d = Depth of the suspension (cm)
A254 = UV absorbance at 254 nm (unitless)
t = Exposure time (s)
To control for error in the UV dose measurement, the uncertainties of the terms in the UV dose
calculation met the following criteria:
• Depth of suspension (d) < 10%
• Average incident irradiance (Es) < 8%
• Petri Factor (Pf) < 5%
• L/(d + L)< 1%
• Time(t)< 5%
• (l-10-ad)/ad< 5%
Further details and definitions of these factors are available in the collimated procedure and
technical papers as referenced in the GP-2011 and UVDGM-2006. The QC data for these factors
are presented in Section 5.5.3.
3.7.5 Developing the UV Dose-response Curve
The collimated beam tests produced:
• UV Dose in units of mJ/cm2,
• Concentration of microorganisms in the petri dish prior to UV exposure (No) as
plaque forming units (pfu)/mL, and
• Concentration of microorganisms in the petri dish after UV exposure (N) as
pfu/mL.
The procedure for developing the UV dose response curves was as follows:
1. For each UV test condition (high or low UVT water) and its replicate and for each day of
testing, log N (pfu/mL) was plotted vs. UV dose (mJ/cm2). A best fit regression line was
determined and a common N0 was identified as the intercept of the curve at UV dose = 0. A
separate equation was developed for each UVT condition (lowest and highest) for each day
of testing at that condition. In this test there were two days of testing, so there were two
sets of data.
2. The log inactivation (log I) was calculated for each measured value of N (including zero-
dose) and the common N0 identified in Step 1 using the following equation:
16
-------
NOVEMBER 2013
log I = log(No/N)
Where:
No = The common N0 identified in Step 1 (pfu/mL);
N = Concentration of challenge microorganisms in the petri dish after
exposure to UV light (pfu/mL).
3. The UV dose as a function of log I was plotted for each day of testing and included water
from both high and low UVT test conditions.
4. Using regression analysis an equation was derived that best fit the data, forcing the fit
through the origin. The force fit through the origin is used rather than the measured value
of N0, because any experimental or analytical error in the measured value is carried to all
the data points, adding an unrelated bias to each measurement. Using the y-intercept of the
curve eliminates error carry through. The regression equation was then used to calculate the
REDmeas for each full scale test sample.
The full set of collimated beam data and all calculations and regression analyses are presented in
Sections 4.3 and 4.4.
The regression analysis was used to derive an equation that best fits the data with a force fit
through the origin. Both linear and a polynomial equations were evaluated to determine the best
fit of the data. The regression coefficient, R2, was determined for each trend line and was
considered acceptable if it was 0.9 or greater and for "r" +/- 0.95 or greater. The equation
coefficients for each day were also evaluated statistically to determine which terms were
statistically significant based on the P factor. All coefficients were found to be significant (i.e.
P<0.05).
For this validation a single curve corresponding to one day's worth of full scale reactor testing
was used to calculate REDmeas values for that day. The higher UVT dose response curve was
used for the high UVT test run with reduced power and the lower UVT dose response curve was
used for the test run when the UVT of the test water was lowered with LSA.
3.7.6 Collimated Beam Data Uncertainty
The collimated beam data was fit to a polynomial regression and the uncertainty of the dose
response equation based on a 95% confidence interval (Uon) was calculated as follows:
UoR = t * [SD/ UV DoseCB] * 100%
Where:
UDR = Uncertainty of the UV dose-response fit at a 95% confidence level
UV DoseCB = UV dose calculated from the UV dose-response curve for the
challenge microorganism
SD = Standard deviation of the difference between the calculated UV dose
response and the measured value
17
-------
NOVEMBER 2013
t = t-statistic at a 95% confidence level for a sample size equal to the number of
test condition replicates used to define the dose-response.
The UDR calculations are included in Section 4.4
3.8 Full Scale Testing to Validate UV dose
3.8.1 Evaluation, Documentation and Installation of Reactor
ETS UV provided technical information on Model UVL-200-4 and basic information on the UV
lamps, sensor, and related equipment. An operating manual and a technical specification book
were provided prior to the start of testing. All documentation and equipment data were reviewed
prior to the start of testing. The equipment was described in Section 2. Attachments 1 and 2
include the manuals, specifications, and sensor data provided by ETS UV.
3.8.2 Test Conditions for UV Intensity Set-Point Approach
The purpose of this testing was to determine a REDmeas dose of >40 ml/cm2 at three set points
that were then used to establish a set line based on the three UV intensity and flow rate pairs.
ETS UV specified the target flow rates (15, 20, 25 gpm) and UV target intensity levels (7, 11, 13
W/m2) based on the results of tests performed at NSF prior to the validation tests. The intensity
targets were based on the expected intensity at UVT's of 79%, 90%, and 94%.
During full-scale commercial operation, regulations require that the UV intensity as measured by
the UV sensor must meet or exceed the validated intensity and that the flow rate be at or below
the validated flow rate to ensure delivery of the required dose. Reactors must be operated within
the validated operating condition for maximum flow rate and minimum irradiance. Under the UV
set point approach, UVT does not have to be measured separately. The intensity readings by the
sensor take into account changes in the UVT.
A set point represents a given flow rate with testing with two conditions (lowered UVT-max
lamp power; high UVT-reduced lamp power). The first test condition involved reducing the
UVT until the UV intensity measured by the unit UV sensor equaled the target UV intensity set
point. The second test condition was run with high UVT (unadjusted UVT) and with the power
reduced until the unit UV intensity measured by the sensor was equal to the target UV intensity
set point. Three target flow rate - intensity set points (15 gpm-7.0 W/m2, 20 gpm-11 W/m2, and
25 gpm - 13 W/m2) were tested under the two conditions and each condition was performed in
duplicate. The three set points were then used to develop a set line that defines operating
conditions of flow rate and intensity that achieve a RED of >40 mJ/cm
2
The LT2ESWTR requires validation of UV reactors to determine a log inactivation of
Cryptosporidium or other target pathogen so that States may use the data to grant log credits.
Therefore, in addition to determining the setline to achieve a minimum REDmeas of 40 mJ/cm2,
additional calculations (adjusting REDmeas for uncertainty and RED bias) were performed to
demonstrate the log inactivation of Cryptosporidium.
A reactor control test (MS2 injection with the lamp off) was run at the low flow rate (15 gpm)
and with high UVT water, which demonstrated that there was no reduction of MS2 with the
18
-------
NOVEMBER 2013
lamps off. A reactor blank was also run on each day of testing. The reactor blank was run with
no phage injection at the low flow rate with high UVT water to demonstrate the testing system
was low in MS2 concentration and other microorganisms. Reactor blank and control samples
were collected in triplicate at the influent and effluent sampling locations and submitted for MS2
analyses.
Trip blanks were prepared and analyzed for each day of testing. The microbiology laboratory
took two samples from the challenge solution prepared for one of the test runs. The first sample
remained in the microbiology laboratory and the second sample traveled with challenge solution
to the engineering laboratory and then was returned with the samples collected from the test run.
Both samples were analyzed for MS2 and the results were compared to determine any change
that might have occurred during transport of the samples. As with stability testing, trip blanks are
important when samples must be shipped or carried long distance with the inherent holding time
before delivery to the lab. At NSF the test rig and laboratory are in the same building and the trip
is "down the hall". Therefore travel related impacts are of less concern, but trip blanks were run
as part of the QC plan for these tests.
Table 3-1 shows a summary of the test conditions that were run for the validation test. A Sample
and Analysis Management Program was also prepared and was provided to the NSF engineering
and microbiology laboratories for use during the testing and for setting up the sample and
analysis in the NSF sample management system.
Five sets of samples were collected at the influent and effluent sample ports for MS2 analysis
during each test condition and its duplicate. The delivered dose was calculated for each of the
five samples and then the average of the five results was calculated to determine an average
delivered dose (RED).
Flow rate, intensity, and UVT data (from the NSF in-line UVT monitor) were collected at each
of the five sample collection times for all test runs. These data were averaged to determine the
average flow rate, UVT, and intensity for each test condition and it's duplicate.
In addition, samples for pH, turbidity, temperature, total and residual chlorine, E coli, and HPC
were collected at the influent and effluent sample ports once during each test run. Samples for
iron (Fe) and manganese (Mn) analyses were collected once during each test run at the influent
sample port to provide additional basic water quality data. Samples were also collected at the
influent and effluent for UVT analysis by the chemistry laboratory bench scale
spectrophotometer to confirm the in-line UVT measurements.
Samples of the low and high UVT waters were collected at the influent and effluent locations for
UVT scans. The samples were scanned for UVT measurements in the range of 200 to 400 nm.
19
-------
NOVEMBER 2013
Table 3-1. Test Conditions for Validation with MS2 Pha
Validation Test
Condition 1
Condition 2
Condition 3
(reactor control)
Condition 4
(reactor blank)
Flow Rate
15gpm
20gpm
25gpm
15gpm
20gpm
25gpm
15gpm
15gpm
UV Transmittance
UVT (%)
79%
90%
94%
>95%
>95%
>95%
>95%
Daily Source water
- ether high or low
UVT
?e.
Lamp Power
Maximum
Lowered to
achieved intensity
from Condition 1
Turned off
Full Power
Intensity
Sensor Reading
Record actual
reading
Set to equal
Condition 1 by
lowering lamp
power
Not applicable
Record
Condition 1 and 2 performed in duplicate
Reactor blanks run for each day of testing
UVT scan of feed water with and without UVT adjustment
Trip blanks and method blanks run for each day of testing
3.8.3 Preparation of the Challenge Microorganisms
The challenge microorganism (MS2) used to validate UV reactors was cultured and analyzed by
NSF's microbiology laboratory as specified in Standard Methods for the Examination of Water
and Wastewater. NSF personnel followed the method for "Culture of challenge microorganisms"
in Annex A of NSF/ANSI Standard 55 as presented in Attachment 3.
Propagation resulted in a highly concentrated stock solution of essentially monodispersed phage
whose UV dose-response follows second-order kinetics with minimal tailing. Over the range of
RED values demonstrated during validation testing, the mean UV dose-response of the MS2
phage stock solution was within the 95-percent prediction interval of the mean response in
Figure A.I in Appendix A of the UVDGM-2006. Over a UV dose range of 0 to 120 millijoules
per centimeter squared (mJ/cni2), the prediction intervals of the data shown in Appendix A of the
UVDGM-2006 are represented by the following equations:
Upper Bound: log / = -1.4X10-4
Lower Bound: log / = -9.6X10-s
: UVDosc2+ 7.6X 10-2 X UVDose
: UVDose2+ 4.5 X 10-2 X UVDose
City of Ann Arbor tap water was filtered using activated carbon to remove any residual chlorine
(confirmed by chemical analysis for total chlorine of the test water), organic surfactants and
dissolved organic chemicals that may be UV absorbers. The filtered challenge water was then
tested for the following parameters and found acceptable if the result is non-detectable or as
otherwise indicated below:
• Total chlorine,
• Free chlorine,
• UV254 ,
20
-------
NOVEMBER 2013
• UVT > 95%
• Total iron,
• Total Manganese,
• Turbidity < 0.3 Nephelometric Turbidity Units (NTU);
• Total coliform (<1 cfu/lOOmL),
• Heterotrophic plate count (<100 cfu/mL).
3.8.4 Conduct Testing - Measuring UV Dose
During full-scale reactor testing, the reactor was operated at each of the test conditions for flow
rate, UVT, and lamp power as described in section 3.8.2. The following steps were taken to
assure meeting data quality objectives:
1. Steady-state conditions were confirmed before injecting the challenge
microorganism. Confirmation of steady state involved monitoring UV sensor
measurements and the UVT to assure the test water and reactor met the test
conditions such as UVT reading of 90%. After typically 3-5 minutes of operation and
confirmation that UVT, sensor readings, and flow rate were steady, the injection
pump was started and steady state conditions were achieved by waiting until the
injection pump was at a steady flow rate based on measurements of weight loss of
solution over 15 second time intervals. In all cases, sampling did not start until at
least 2 minutes after the injection pump was started.
2. MS2 was injected into the feed water upstream of the reactor to achieve a greater than
IxlO5 pfu/mL so that a minimum of a 4 log reduction could be measured during the
runs.
3. Sample taps remained open over the duration of the test.
4. Samples were collected in accordance with standards of good practice as defined by
Standard Methods Section 9060.
5. Five (5) sample pairs were collected during approximately ten to twelve minutes of
continuous flow at steady conditions. Each set of influent and effluent grab samples
were collected as close in time as possible. The five sets of samples were spread out
over the 10- to 12-minute continuous flow run.
6. Sample volumes for assessing the challenge microorganism concentrations in the
influent and effluent were collected in 125 mL bottles.
7. Samples were collected in bottles that had been cleaned and sterilized by the NSF
micro laboratory.
8. Collected samples were delivered directly to the microbiological lab located in the
same building after each sampling period. Sample analyses were generally started
immediately, but if samples could be stored in the refrigerator, in the dark, they were
analyzed a couple of hours later. All MS2 analyses were started within 4-6 hours of
the time the sample was collected.
The following measurements and recordings were taken during each test run:
1. The flow rate through the reactor, UV sensor reading and on-line UVT measurements
were recorded when each sample was collected during each run, yielding a minimum
of five measurements for each test run.
21
-------
4.
5.
NOVEMBER 2013
Water chemistry and other microbiological grab samples were collected once per test
condition after one of the challenge organism samples were collected. Samples for
temperature, pH, E. coli, and Heterotrophic Plate Count were collected at the influent
and effluent locations, and samples for iron, manganese, turbidity and residual
chlorine were collected at the influent location.
A sample for UVT was collected and measured by a UV spectrophotometer for each
influent sample and at least one effluent sample.
A sample of the influent and effluent water was collected at the beginning of each test
day and a UVT scan performed over the range of 200 to 400 nm.
The electrical power consumed by the system was recorded.
Chapter 4 describes the calculations and presents the data for determining the REDn
validated dose (REDVai) at a each set point.
3.9 Analytical Methods
All laboratory analytical methods for water quality parameters are listed in Table 3-2.
Table 3-2. Analytical Methods for Laboratory Analyses.
and the
Parameter
Temperature
pH
E. coli 1 Total Coliform
Iron
Manganese
Turbidity
MS2
Absorbance UV 254
Residual chlorine
HPC
Method
SM(2) 2550
SM(2 4500-H+
SM 9223
EPA 200.7
EPA 200.8
SM(2 2130
Top Agar
Overlay
SM5910B
SM 4500-C1 D
SM9215
NSF
Reporting
Limit
-
1CFU
/lOOmL
20 ug/L
lug/L
0.1 NTU
1 pfu/mL
NA
0.05 mg/L
1 CFU/mL
Lab
Accuracy
(%
Recovery)
-
+.0.1 SU
of buffer
-
70-130
70-130
95-105
-
60-140
90-110
-
Lab
Precision
(%RPD (1))
-
+.0.1 SU
-
10%
10%
-
-
<20
<10%
-
Hold
Time
(days)
-
(3)
24 h
180
days
180
days
(3)
24 h(4)
2
(3)
24 h
Sample
Container
-
NA
500 mL
plastic
125 mL
polyethylen
e
125 mL
polyethylen
e
NA
125 mL
plastic
1 L plastic
NA
125 mL
plastic
Sample
Preservation
-
None
1% Tween 80
Nitric acid
Nitric acid
None
1% Tween 80
None
None
1% Tween 80
(1) RPD = Relative Percent Deviation
(2) SM = Standard Methods
(3) Immediate analysis required
(4) h = hours
22
-------
NOVEMBER 2013
3.9.1 Sample Processing, and Enumeration of MS2:
MS2 sample processing and enumeration followed the procedures used in NSF / ANSI Standard
55.
3.9.2 Percent UVT measurements:
The percent UVT for laboratory measurements was calculated from A254. The equation for UVT
using A254 is:
UVT(%)= 100* 10 ~A™
The on-line UVT analyzer provided immediate data throughout all test runs. The on-line
analyzer was calibrated every day of operation. A primary standard was used before the first day
of testing began. Daily calibration was performed on all test days using a certified secondary
standard. Before the start of each day's testing, a sample was taken to the laboratory and analyzed
with the on-line analyzer to ensure the data were comparable and within acceptable quality
control limits for accuracy.
All UVT measurements used a 1-cm path length and are reported on a 1-cm path length basis.
Spectrophotometer measurements of A254 were verified using NIST-traceable potassium
dichromate UV absorbance standards and holmium oxide UV wavelength standards. The UV
Spectrophotometer internal QA/QC procedures outlined in the UVDGM-2006 were used to
verify calibration. UV absorbance of solutions used to zero the Spectrophotometer were verified
using reagent grade organic-free water certified by the supplier to have zero UV absorbance.
The measurement uncertainty of the Spectrophotometer must be 10 percent or less. To achieve
this goal, the following procedures were used to verify:
1. The Spectrophotometer read the wavelength to within the accuracy of a holmium
oxide standard (typically ± 0.2 nm at a 95-percent confidence level.
2. The Spectrophotometer read A254 within the accuracy of a dichromate standard (e.g.,
0.281 ± 0.005 at 257 nm with a 20 mg/L standard), and
3. That the water used to zero the instrument has an A254 value that was within 0.002
cm"1 of a certified zero absorbance solution.
3.9.3 Analytical QA/QC Procedures
Accuracy and precision of sample analyses were ensured through the following measures:
• pH - Three-point calibration (4, 7, and 10) of the pH meter was conducted daily using
traceable buffers. The accuracy of the calibration was checked daily with a pH 8.00
buffer. The pH readings for the buffer were within 10% of its true value. The precision
of the meter was checked daily using duplicate synthetic drinking water samples. The
difference of the duplicate samples was within + 0.1 SU.
• Temperature - The thermometer used to give the reportable data had a scale marked for
every 0.1°C. The thermometer is calibrated yearly using a Hart Scientific Dry Well
Calibrator Model 9105.
23
-------
NOVEMBER 2013
• Total chlorine - The calibration of the chlorine meter was checked daily using a DI water
sample (blank), and three QC standards. The measured QC standard values were within
10% of their true values. The precision of the meter was checked daily by duplicate
analysis of synthetic drinking water samples. The RPD of the duplicate samples was less
than 10%.
• Turbidity - The turbidimeter was calibrated as needed according to the manufacturer's
instructions with formazin standards. Accuracy was checked daily with a secondary
Gelex standard. The calibration check provided readings within 5% of the true value.
The precision of the meter was checked daily by duplicate analysis of synthetic drinking
water samples. The RPD of the duplicate samples was less than 10% or had a difference
of less than or equal to 0.1 NTU at low turbidity levels.
3.9.4 Sample Handling
All samples were labeled with unique identification numbers. These identification numbers were
entered into the NSF Laboratory Information Management System (LIMS), and were used on the
NSF lab reports for the tests. All challenge organism samples were stored in the dark at 4 + 2 °C
and processed for analysis within 4-6 hours.
3.10 Full Scale Test Controls
The following quality-control samples and tests for full-scale reactor testing were performed:
• Reactor controls - Influent and effluent water samples were collected with the UV lamps
turned off. The change in log concentration from influent to effluent should correspond to
no more than 0.2 logic.
• Reactor blanks - Influent and effluent water samples were collected with no addition of
MS2 to the flow passing through the reactor. Blanks were collected once on each day of
testing. The reactor blank is acceptable when the MS2 concentration is less than 0.2 logic.
• Trip controls - Trip controls were collected to monitor any change in MS2 during
transport to the laboratory (in the same building).
• Method blanks - A sample bottle of sterilized reagent grade water was analyzed using the
MS2 assay procedure. The concentration of MS2 in the method blank was non-
detectable.
• Stability samples - Influent and effluent samples at low and high UVT prior to the
introduction of MS2, These samples were used to assess the stability of MS2
concentration and its UV dose-response over the time period from sample collection to
completion of the MS 2 assay. The MS2 were added to achieve a concentration of 1,000
plaque forming units (pfu)/L in the samples containing test water at the lowest and
highest UVT. A sample was analyzed immediately (called time 0) and then 4 hours, 8
hours and 24 hours after time 0. All analyses were performed in triplicate. While stability
samples were performed during the test, they are not directly applicable in this case as all
sample analyses for MS2 were started within a couple of hours of collection.
24
-------
NOVEMBER 2013
3.11 Power Measurements
The voltmeter and ammeter meter used to measure UV equipment had traceable evidence of
being in calibration (e.g., have a tag showing that it was calibrated). Calibrations of meters were
performed at least yearly and within the past year.
3.12 Flow Rate
During validation testing, the QC goal was that the accuracy of flow rate measurements should
be within +5 percent of the true value. Flow meter accuracy was verified by monitoring the draw
down volume in the supply tanks over time. The supply tanks have been calibrated using the
catch and weigh technique. The flow meter was within 1.6% of the true value. Flow meter
calibration data are presented in Section 5.6.
3.13 Evaluation, Documentation and Installation of Reactor
ETS UV provided technical information on the Model UVL-200-4 and basic information on the
UV lamps, sensor, and related equipment. An operating manual was provided prior to the start of
testing. Additional information on the lamp output (confirmation of spectral output) was
provided prior to the start of the validation test. All documentation and equipment data was
reviewed prior to the start of testing. The following documentation was reviewed and found to
conform to the GP-2011 and UVDGM-2006 requirements:
Reactor Specifications
• Technical description of the reactor's UV dose-monitoring strategy, including the use of
sensors, signal processing, and calculations (if applicable^
• Dimensions and placement of all critical components (e.g., lamps, sleeves, UV sensors,
baffles, and cleaning mechanisms) within the UV reactor
• A technical description of lamp placement within the sleeve
• Specifications for the UV sensor port indicating all dimensions and tolerances that impact
the positioning of the sensor relative to the lamps
Lamp specifications
• Technical description
• Lamp manufacturer and product number
• Electrical power rating
• Electrode-to-electrode length
• Spectral output of the lamps (specified for 5 nm intervals or less over a wavelength range
that includes the germicidal range of 250 - 280 nm and the response range of the UV
sensors)
Lamp sleeve specifications
• Technical description including sleeve dimensions
• Material of construction
• UV transmittance at 254 nm
Specifications for the reference and the duty UV sensors
• Manufacturer and product number
• Technical description including external dimensions
Sensor measurement properties
25
-------
NOVEMBER 2013
• Working range
• Spectral and angular response
• Linearity
• Calibration factor
• Temperature stability
• Long-term stability
Installation and operation documentation
• Flow rate and pressure rating of the reactor
• Assembly and installation instructions
• Electrical requirements, including required line frequency, voltage, amperage, and power
• Operation and maintenance manual including cleaning procedures, required spare parts,
and safety requirements.
26
-------
NOVEMBER 2013
Chapter 4
Results and Discussion
4.1 Introduction
ETS UV specified target flow rates of 15, 20, and 25 gpm. The intensity initial targets were 7,
11 and 13 W/m2 based on the expected intensities at UVTs of 79%, 90%, and 94% with the lamp
at full power. These points were projected to deliver a REDmeas and REDyai of >40 ml/cm2.
The main validation tests were run on two days, June 20 and June 21, 2012. All of the results of
the validation set point tests are presented herein. The first day of testing was dedicated to the
test conditions and duplicate runs where the UVT of the feed water was lowered to the target
levels (<79%, <90%, and <94%) and the lamps were operated at full power. The second day of
testing was dedicated to the test conditions and duplicates where high UVT feed water (>95%
target) was used and the lamp power was reduced to achieve the target intensity level. The test
conditions and detail on the test rig setup, sampling procedures, and unit operation are described
in Chapter 3 Methods and Procedures.
All tests were conducted at the NSF laboratory in Ann Arbor, MI, and all analyses were
performed by the NSF microbiological and chemistry laboratories at this location.
4.2 Sensor Assessment
The Model UVL-200-4 duty sensor was evaluated according to the UV sensor requirements in
the EPA's UVDGM-2006 prior to and after the verification testing. All UV intensity sensors
(the duty and two reference sensors) were new sensors and specifications provided with the
sensors showed they were designed in accordance with the DVGW guideline W 294 (June,
2006) and the ONORM M5873-1 standard (June, 2002), respectively. Evidence of calibration of
the sensors traceable to a standard of the Physikalisch Technische Bundesanstalt (PTB) in
Braunschweig, was provided by ETS UV as provided to them by the sensor manufacturer uv-
technik. Certificates are presented in Attachment 2.
The same duty sensor was used for monitoring intensity (irradiance) for all test runs. This sensor
measured the intensity from the single low pressure lamp in the unit. The control panel provided
direct readings of intensity in W/m2. This direct reading was based on converting the 4-20 mA
output signal to intensity based on the calibration certificate provided with the sensor.
Attachment 2 includes the certificates for the two reference sensors and one duty sensor, plus the
spectral data for the sensor.
The duty sensor was compared against two reference sensors to demonstrate that the duty sensor
was within 10% of the average of the two reference sensors. This evaluation was conducted
before and after the validation test runs using the procedure described in the GP-2011 and the
UVDGM-2006. Table 4-1 presents the results of the sensor assessment. These data demonstrate
that the duty sensor was within 10 percent of the average of the two reference sensors. The two
reference sensors showed a variance of 2.7% at 100% power and a range of 2.7% to 3.0% at 60%
power.
27
-------
NOVEMBER 2013
Table 4-1. Sensor Assessment Data First Set of Test Runs (June 2012)
Sensor
Reference #1
W4164
Reference #2
W4166
Average of
Reference
Sensor
Duty Sensor
W4165
Deviation of
Duty Sensor
from Reference
Intensity at
100% power
Before testing
(W/m2)
13.22
13.94
13.58
13.57
<0.1%
UVT = 95.5%
Intensity at
100% power
After testing
(W/m2)
14.56
15.38
14.97
15.08
0.7%
UVT = 97%
Intensity at
-60% Power
Before testing
(W/m2)
4.26
4.52
4.39
4.37
0.5%
UVT = 95. 5%
Intensity at
-60% Power
After testing
(W/m2)
4.81
5.03
4.92
4.94
0.4%
UVT = 97%
The test results shown in the later tables and the sensor assessment data collected before and
after the test were performed to demonstrate the intensity was stable throughout the testing as a
function of ballast power and UVT. This indicates that lamp output was constant and no fouling
occurred to the lamp sleeves and sensor windows.
4.3 Collimated Beam Dose Response Data
Collimated Beam dose response data were generated for both low and high UVT waters in
accordance with the procedures described in Section 3.7.4. The collimated beam tests were run
on the minimum UVT water (79%) and on the high UVT (97%) water used for the minimum
power test runs. All collimated beam tests were performed in duplicate.
UV doses covered the range of the targeted REDmeas dose, which in this case was >40 mJ/cm2.
UV target doses were set at 0, 20, 30, 40, 60 and 80 mJ/cm2. As discussed in Section 4.5, the
actual REDmeas for four test runs slightly exceeded the maximum collimated beam dose of 80
mJ/cm2. REDmeas cannot be quantitatively determined if calculated REDmeas exceeds the top
range of the collimated beam data. These data are presented as calculated, but any REDmeas
values above 80 mJ/cm2 should be used as estimates only.
The collimated beam samples were collected directly from the test rig during the normal testing
runs. A one liter bottle of the seeded influent water (MS2 injection pumping run during the test
run) was collected to provide the two samples for duplicate analyses. Using this approach, the
dose response data reflect the identical conditions to the biodosimetric flow tests for sample
matrix, UVT, and MS2 concentration. The collimated beam samples were irradiated on the same
day as sample collection, and were plated in triplicate along with the flow test samples.
28
-------
NOVEMBER 2013
Therefore analytical conditions for the dose response data were also identical to those for the
flow test samples.
The collimated beam results are presented in Tables 4-2 and 4-3. These data were calculated as
the average of the three individual results obtained at each dose level.
4.4 Development of Dose Response
The development of the UV dose response curves for use with flow tests to establish the REDmeas
is a three step process.
1. For each collimated beam test and its replicate for each day of testing, the log N (pfu/mL)
was plotted vs. UV dose (mJ/cm2). Figures 4-1 and 4-2 show the curves for the low and
higher UVT waters.
2. A separate equation (second order polynomial) was developed for each UVT condition
(low and high). Therefore, there are two sets (low and high UVT) of data with each set
containing collimated test performed in duplicate. A common N0 was identified for each
data set as the intercept of the curve at UV dose = 0.
3. The log inactivation (log I) was then calculated for each day for each measured value of
N (including zero-dose) and the common N0 identified in Step 1 using the following
equation:
log I = log (N0/N)
Where:
No = the common N0 identified in Step 1 (pfu/mL);
N = Concentration of challenge microorganisms in the petri dish after
exposure to UV light (pfu/mL).
Tables 4-2 and 4-3 show the calculated values for log inactivation (LI).
Finally, the UV dose as a function of log I was plotted for each set of data. Figures 4-3 and 4-4
show the curves for dose as a function of log inactivation. Using regression analysis an equation
was derived that best fit the data, forcing the fit through the origin. In each case the equation
was a second order polynomial, which is the most common for MS2 collimated beam data. The
regression equation was then used to calculate the REDmeas for each full scale flow test samples.
REDmeas calculations and full scale data is presented in Section 4.5.
The polynomial equation coefficients for each day were evaluated statistically to determine if the
terms were significant based on the P factor. All coefficients were found to be significant (P
factor <0.05) for all the dose response curves.
A Grubbs' test was run to determine if any replicates should be omitted from the development of
the dose response curve. The Grubbs' test results show that no replicates should be omitted from
the data set. However, one data point, replicate 2 at the 20.88 mJ/cm2 dose for the 79% UVT
water was very close to being a statistical outlier (Grubbs statistics was 2.4 and this data point
had a value of 2.3). The Grubbs's statistics are shown in Tables 4-2 and 4-3.
29
-------
NOVEMBER 2013
A summary of the statistics for uncertainty for the collimated beam dose response data is
presented at the end of Tables 4-2 and 4-3. The dose response uncertainty (UDR) of the
collimated beam results for the high UVT water is less than 30% (UDR = 18.14%) at the 1-log
inactivation level, which is the QC goal. However, the data from the low UVT water shows a
UDR of 38.95%. Further examination of the data shows that the primary cause of the increased
uncertainty of the dose response equation above the objective of <30% is the data point at a
20.88 mJ/cm2 dose for replicate 2. As shown in the Grubbs' test above, this data was very close
to being a statistically significant outlier. The data were analyzed with this data point removed
and the uncertainty of the dose response equation dropped to 26.73%. Given that this data point
has a high Grubbs statistic and by far the highest residual in the statistical tests; the data point
was considered an outlier and removed from the dose response equation. Table 4-4 shows the
recalculated dose response data and equation coefficient with the replicate 2 data point at the
20.88 mJ/cm2 dose removed. The revised dose response equation is the one used for the
calculations of REDmeas in the next section. At 2-log inactivation (a dose of approximately 40
mJ/cm2 RED) the UDR was 12.43% and 8.36%.
It should be noted that if the REDmeas data are calculated using the data with the outlier included,
the REDmeas for these six flow test runs calculated using that set of collimated beam data is still
well above the required target dose of 40 mJ/cm2.
Also shown in Figures 4-3 and 4-4 are the QC limits for MS2 taken from the UVDGM-2006.
The results show that the MS2 results are within the boundaries established for MS2.
30
-------
NOVEMBER 2013
Table 4-2. UV Dose - Response Data from Collimated Beam Tests at 79% UVT (June 2012)
UVT
(%)
79.0
Rep
1
2
Target
UV Dose
(mJ/cm2)
0
20
30
40
60
80
0
20
30
40
60
80
DRC
A: 15.534
B: 1.5796
Actual
UV Dose
0.00
20.76
31.23
41.37
62.14
82.80
0.00
20.88
31.18
41.46
62.18
83.12
UV
Dose2
0
431
975
1711
3861
6856
0
436
972
1719
3866
6909
Log N0
6.56
Avg
pfu/ml
5,430,000
249,000
81,000
33,300
5,130
590
3,600,000
106,000
66,300
22,800
3,010
413
Avg
Log(pfu)
6.73
5.40
4.91
4.52
3.71
2.77
6.56
5.03
4.82
4.36
3.48
2.62
Log I
-0.17
1.17
1.66
2.04
2.85
3.79
0.01
1.54
1.74
2.21
3.09
3.95
Log I2
0.029
1.366 P
2.744
4.172
8.150
14.395
0.000
2.370
3.039
4.871
9.525
15.594
RED Dose
-2.59
20.31
30.06
38.32
57.22
81.67
0.13
27.66
31.88
41.98
62.99
85.97
Avg:
SD:
P:
t (95%):
Residual
(mJ/cm2)
2.6
0.4
1.2
3.1
4.9
1.1
-0.1
-6.8
-0.7
-0.5
-0.8
-2.9
0.13
2.99
12
0.05
2.228
G
0.8
0.1
0.3
1.0
1.6
0.3
0.1
2.3
0.3
0.2
0.3
1.0
Outlier?
OK
OK
OK
OK
OK
OK
OK
OUTLIER
OK
OK
OK
OK
Grubbs' Test for Outliers
p: 0.10
t (90%): 3.691
Grubbs'Statistic
(GCRIT): 2.412
DRC - dose response coefficients
31
-------
Table 4-2. (continued)
NOVEMBER 2013
Uncertainty of Dose-Response (UDR)
Logl
0.001
0.25
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
3.95
Dose
(mJ/cm2)
0.0
4.0
8.2
17.1
26.9
37.4
48.7
60.8
73.7
87.4
86.0
t
2.23
2.23
2.23
2.23
2.23
2.23
2.23
2.23
2.23
2.23
SD
2.99
2.99
2.99
2.99
2.99
2.99
2.99
2.99
2.99
2.99
UDR (%)
167.39
81.67
38.95
24.82
17.83
13.69
10.96
9.04
7.63
7.75
°2L
(mJ/cm /Log I)
15.54
15.93
16.32
17.11
17.90
18.69
19.48
20.27
21.06
21.85
21.77
t - student t test factor SD - standard deviation
Regression Statistics
Multiple R
R Square
Adjusted R Square
Standard Error
Observations
0.99822
0.996444
0.896088
3.140758
12
ANOVA
Regression
Residual
Total
df
2
10
12
ss
27638.73
98.64361
27737.37
MS
13819.36
9.864361
F
1400.939
Significance
F
5.95E-12
Intercept
X Variable 1
X Variable 2
Coefficients
0
15.53382
1 .579603
Standard
Error
1.420807
0.445824
tStat
10.9331
3.54311
P-value
6.98E-07
0.005329
Lower
95%
12.36807
0.586245
Upper
95%
18.69958
2.57296
Lower
95.0%
12.36807
0.586245
Upper
95.0%
18.69958
2.57296
32
-------
NOVEMBER 2013
Table 4-3. UV Dose - Response Data from Collimated Beam Tests at 97% UVT (June 2012)
UVT
(%)
97.0
Rep
1
2
Target
UV Dose
(mJ/cm2
)
0
20
30
40
60
80
0
20
30
40
60
80
Actual
UV
Dose
0.00
20.74
30.96
41.12
61.79
82.51
0.00
20.76
31.14
41.53
62.10
82.62
DRC
A: 15.449
B: 1.4294
UV
Dose2
0
430
959
1691
3818
6808
0
431
970
1725
3856
6826
Avg
pfu/ml
2,030,00
0
120,000
53,300
14,900
1,800
330
2,860,00
0
121,000
48,300
13,300
1,470
247
Log N0
6.36
Avg
Log(pfu
)
6.31
5.08
4.73
4.17
3.26
2.52
6.46
5.08
4.68
4.12
3.17
2.39
Log I
0.05
1.28
1.63
2.19
3.10
3.84
-0.10
1.28
1.68
2.24
3.19
3.97
Log I2
0.003 p
1.641
2.668
4.782
9.639
14.757
0.009
1.631
2.809
5.000
10.193
15.740
RED Dose
0.82
22.13
29.05
40.62
61.74
80.44
-1.48
22.06
29.91
41.69
63.89
83.79
Avg:
SD:
t (95%):
Residua
I
(mJ/cm2
)
-0.8
-1.4
1.9
0.5
0.0
2.1
1.5
-1.3
1.2
-0.2
-1.8
-1.2
0.05
1.37
12
0.05
2.228
G
0.6
1.0
1.4
0.3
0.0
1.5
1.0
1.0
0.9
0.2
1.3
0.9
Outlier
?
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
Grubbs' Test for Outliers
p: 0.10
t (90%): 3.691
Grubbs'Statist!
c(GCRIT): 2.412
DRC - dose response coefficients
33
-------
Table 4-3. (continued)
NOVEMBER 2013
Uncertainty of Dose-Response (UDR)
Logl
0.001
0.25
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
3.97
Dose
(mJ/cm2)
0.0
4.0
8.1
16.9
26.4
36.6
47.6
59.2
71.6
84.7
83.8
t
2.23
2.23
2.23
2.23
2.23
2.23
2.23
2.23
2.23
2.23
SD
1.37
1.37
1.37
1.37
1.37
1.37
1.37
1.37
1.37
1.37
UDR (%)
77.49
37.89
18.14
11.60
8.36
6.44
5.17
4.28
3.62
3.65
°2L
(mJ/cm2/Log I)
15.45
15.81
16.16
16.88
17.59
18.31
19.02
19.74
20.45
21.17
21.12
t - student t test factor SD - standard deviation
Regression Statistics
Multiple R
R Square
Adjusted R Square
Standard Error
Observations
0.999622
0.999244
0.899168
1 .442402
12
ANOVA
Regression
Residual
Total
df
2
10
12
ss
27492.52
20.80525
27513.32
MS
13746.26
2.080525
F
6607.111
Significance F
5.6E-15
Intercept
X Variable 1
X Variable 2
Coefficients
0
15.44899
1 .4294
Standard
Error
0.664129
0.203913
tStat
23.26204
7.009835
P-value
4.88E-10
3.67E-05
Lower
95%
13.96922
0.975052
Upper
95%
16.92877
1 .883747
Lower
95.0%
13.96922
0.975052
Upper 95.0%
16.92877
1.883747
34
-------
NOVEMBER 2013
Table 4-4. UV Dose - Response Data from Collimated Beam Tests at 79% UVT with Outlier Removed (June 2012)
UVT
(%)
79.0
Rep
1
2
DRC
16.10
A 6
1.312
B 9
Target
UV
Dose
(mJ/cm2
)
0
20
30
40
60
80
0
30
40
60
80
Actual
UV
Dose
0.00
20.76
31.23
41.37
62.14
82.80
0.00
31.18
41.46
62.18
83.12
Log N0
6.62
UV
Dose2
0
431
975
1711
3861
6856
0
972
1719
3866
6909
Avg
pfu/ml
5,430,00
0
249,000
81,000
33,300
5,130
590
3,600,00
0
66,300
22,800
3,010
413
Avg
Log(pfu
)
6.73
5.40
4.91
4.52
3.71
2.77
6.56
4.82
4.36
3.48
2.62
Log I
-0.12
1.22
1.71
2.09
2.91
3.85
0.06
1.79
2.26
3.14
4.00
Log I2
0.014 p
1.488
2.916
4.384
8.445
14.786
0.004
3.221
5.100
9.845
16.002
RED Dose
-1.89
21.60
31.33
39.48
57.89
81.35
0.97
33.13
43.07
63.46
85.44
Avg:
SD:
P:
t (95%):
Residua
I
(mJ/cm2
)
1.9
-0.8
-0.1
1.9
4.2
1.5
-1.0
-2.0
-1.6
-1.3
-2.3
0.04
2.06
11
0.05
2.262
G
0.9
0.4
0.1
0.9
2.0
0.7
0.5
1.0
0.8
0.6
1.1
Outlier
?
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
Grubbs'Test for Outliers
p: 0.10
t (90%): 3.751
Grubbs'Statist!
c (GCRIT): 2.355
DRC - dose response coefficients
35
-------
Table 4-4. (continued)
NOVEMBER 2013
Uncertainty of Dose-Response (UDR)
Logl
0.001
0.25
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
4.00
Dose
(mJ/cm2)
0.0
4.1
8.4
17.4
27.1
37.5
48.5
60.1
72.5
85.4
85.4
t
2.26
2.26
2.26
2.26
2.26
2.26
2.26
2.26
2.26
2.26
SD
2.06
2.06
2.06
2.06
2.06
2.06
2.06
2.06
2.06
2.06
UDR (%)
113.32
55.55
26.73
17.17
12.43
9.61
7.74
6.43
5.45
5.45
°2L
(mJ/cm /Log I)
16.11
16.43
16.76
17.42
18.08
18.73
19.39
20.04
20.70
21.36
21.36
t - student t test factor SD - standard deviation
Regression Statistics
Multiple R
R Square
Adjusted R Square
Standard Error
Observations
0.999224
0.998448
0.887164
2.169925
11
ANOVA
Regression
Residual
Total
df
2
9
11
SS
27259.02
42.37717
27301.4
MS
13629.51
4.708575
F
2894.615
Significance F
3.63E-12
Intercept
X Variable 1
X Variable 2
Coefficients
0
16.10622
1.312854
Standard
Error
1 .048756
0.32042
tStat
15.35746
4.097298
P-value
9.19E-08
0.002688
Lower
95%
13.73377
0.588015
Upper
95%
18.47867
2.037694
Lower
95.0%
13.73377
0.588015
Upper
95.0%
18.47867
2.037694
36
-------
NOVEMBER 2013
Log N as a function of UV Dose
y = 0.0001 x2 - 0.0589X + 6.6162
R2 = 0.9948
40 50 60
UV Dose (mJ/cm2)
Figure 4-1 Collimated beam dose versus log N UVT 79% with outlier removed (June 2012)
37
-------
NOVEMBER 2013
Log N as a function of UV Dose
UVDose(mJ/cm2}
y = 0.0002X2 - 0.0605X + 6.3602
R2 = 0.9974
Figure 4-2 Collimated beam dose versus log N UVT 97% (June 2012)
38
-------
NOVEMBER 2013
100
80 -
Dose-Response Curve
1.3129x2 + 16.106x
R2 = 0.9949
150
- 120
90
TI
Z>
- 60
- 30
234
Log Inactivation
0 CBData -- UVDGMQC Limits
Udr (%)
Dose-Response Curve
Figure 4-3 Dose response - log I versus dose - UVT 79% with outlier removed (June 2012)
39
-------
NOVEMBER 2013
Dose-Response Curve
100
80 -
150
y = 1.4294x2 + 15.449x
120
- 90
- 60
- 30
123456
Log Inactivation
CB Data UVDGM QC Limits Udr(%) Dose-Response Curve
Figure 4-4 Dose response - log I versus dose - UVT 97% (June 2012)
40
-------
NOVEMBER 2013
4.5 MS and Operational Flow Test Data
The operational data (flow rate, UVT, lamp power and UV sensor intensity measurements) are
presented in Table 4-5. UVT was monitored continuously by an in-line analyzer. Flow rate,
UVT, and intensity were recorded when each sample was collected, thus providing five data
points for each test run. These values were then used to obtain an average flow rate, UVT, and
intensity for each test run.
The first influent and effluent samples for MS2 determination were taken simultaneously
beginning after approximately 2-3 minutes of steady state operation. Subsequent influent and
effluent samples were collected simultaneously after an additional two to three minutes of
operation, yielding five sets of samples over a ten to twelve minute period. The MS2
concentration data for each test run are shown in Table 4-6.
For each test condition replicate (i.e., each of the five influent and effluent samples), the log
inactivation (log I) was calculated using the following equation:
log/=log(N0/N)
Where:
No = Challenge microorganism concentration in influent sample (pfu/mL);
N = Challenge microorganism concentration in corresponding effluent sample
(pfu/mL).
The log of the influent and effluent concentration is shown in Table 4-7. Table 4-8 shows the
Log Inactivation results. For each test condition replicate, the REDmeas was determined using the
measured log inactivation (log I) and the collimated beam test dose-response curves for each day
of testing (See Figures 4-3 and 4-4). The five replicate REDmeas values were then averaged to
produce one REDmeas for each test run and its duplicate. The calculated REDmeas results in
ml/cm2 are shown in Table 4-9.
All of the flow rate tests at 15, 20, and 25 gpm with feed water at 79%, 90%, and 93% UVT or
the equivalent reduced power tests achieved a minimum REDmeas of 40 ml/cm
2
The REDmeas for four of the test runs exceeded the maximum collimated beam dose of 80
mJ/cm2. These runs showed calculated REDmeas between 82 and 87 ml/cm2. The REDmeas cannot
be quantitatively determined if the measured RED exceeds the top range of the collimated data
and can only be quantified as being >80 mJ/cm2. For informational purposes, these data are
presented as calculated even though they exceeded the maximum collimated beam dose of 80
mJ/cm2 and would normally be reported at >80 mJ/cm2. The four RED values above 80 mJ/cm2
should be considered as estimates only.
41
-------
NOVEMBER 2013
Table 4-5. ETS UV Model UVL-200-4 Operational Data
Test Condition
Lowered UVT - Full Power (SPt 1)
Lowered UVT - Full Power Duplicate (SPt 1)
Lowered UVT - Full Power (SPt 2)
Lowered UVT - Full Power Duplicate (SPt 2)
Lowered UVT - Full Power (SPt 3)
Lowered UVT - Full Power Duplicate (SPt 3)
Lowered Power- High UVT (SPt 1)
Lowered Power- High UVT Duplicate (SPt 1)
Lowered Power - High UVT (SPt 2)
Lowered Power- High UVT Duplicate (SPt 2)
Lowered Power- High UVT (SPt 3)
Lowered Power- High UVT Duplicate (SPt 3)
Run
2
3
4
5
6
7
10
11
12
13
14
15
%of
Full
Power(1)
100
100
100
100
100
100
70
70
80
80
85
85
UVT
(%)
78.5
78.5
89.7
89.7
93.5
93.5
97.0
97.1
97.1
97.0
97.1
97.1
Flow
(gpm)
14.9
14.8
19.9
19.9
24.9
24.9
14.9
14.9
19.9
19.9
24.9
25.0
Intensity
(W/m2)
7.0
7.0
11.0
11.0
12.8
12.8
6.9
6.9
11.0
11.0
12.7
12.7
(1) % of full power estimated based on measured amperage for the system, where amperage at reduced power
is divided by amperage at full power in 97% UVT water.
SPt = Set Point Condition
42
-------
NOVEMBER 2013
Table 4-6. ETS UV Model UVL-200-4 MS2 Concentration Results
Test Condition
Lowered UVT- Full
Power (SPt 1)
Lowered UVT- Full
Power Dup (SPt 1)
Lowered UVT- Full
Power (SPt 2)
Lowered UVT- Full
Power Dup (SPt 2)
Lowered UVT- Full
Power (SPt 3)
Lowered UVT- Full
Power Dup (SPt 3)
Lowered Power -
High UVT (SPt 1)
Lowered Power -
High UVT Dup
(SPt 1)
Lowered Power -
High UVT (SPt 2)
Lowered Power -
High UVT Dup
(SPt 2)
Lowered Power -
High UVT (SPt 3)
Lowered Power -
High UVT Dup
(SPt 3)
Run
2
3
4
5
6
7
10
11
12
13
14
15
Influent (pfu/mL)
Rep1
4.70E+06
1.01E+06
1.45E+06
6.67E+06
8.41 E+05
5.20E+06
4.20E+06
7.27E+05
2.12E+06
1.28E+06
4.73E+06
4.17E+06
Rep 2
3.93E+06
1.44E+06
2.14E+06
6.53E+06
9.03E+05
5.50E+06
2.16E+06
7.87E+05
1.33E+06
9.03E+05
6.67E+06
4.30E+06
Rep 3
3.81 E+06
1.53E+06
2.63E+06
5.03E+06
7.65E+05
3.99E+06
4.13E+06
7.33E+05
1.13E+06
1 .OOE+06
5.47E+06
4.77E+06
Rep 4
4.40E+06
1.44E+06
1.17E+06
5.77E+06
9.50E+05
4.80E+06
4.20E+06
9.47E+05
1.18E+06
5.43E+05
4.73E+06
4.13E+06
Rep 5
4.57E+06
1.31 E+06
3.12E+06
6.03E+06
7.89E+05
4.97E+06
5.03E+06
8.97E+05
2.13E+06
1 .66E+06
7.10E+06
4.37E+06
Effluent (pfu/mL)
Rep1
1.25E+03
2.05E+02
2.97E+02
8.53E+02
7.77E+02
1.22E+03
2.24E+03
2.57E+02
1 .50E+02
3.03E+02
7.60E+02
4.97E+02
Rep 2
9.30E+02
1 .77E+02
3.57E+02
6.97E+02
7.87E+02
8.13E+02
2.06E+03
4.55E+02
1 .27E+02
3.13E+02
5.23E+02
7.20E+02
Rep 3
9.57E+02
6.17E+02
4.80E+02
8.83E+02
8.03E+02
1.11E+03
1.92E+03
7.43E+02
1.39E+02
1.79E+02
8.00E+02
4.63E+02
Rep 4
9.50E+02
6.87E+02
5.63E+02
8.50E+02
7.00E+02
1.07E+03
2.12E+03
8.77E+02
6.63E+01
1 .58E+02
5.37E+02
7.67E+02
Rep 5
9.00E+02
4.40E+02
6.13E+02
9.13E+02
9.63E+02
1.11E+03
2.14E+03
8.23E+02
1 .53E+02
2.28E+02
4.87E+02
4.53E+02
SPt - Set Point Condition
43
-------
NOVEMBER 2013
Table 4-7. ETS UV Model UVL-200-4 MS2 Log Concentration for Influent and Effluent Samples
Test Condition
Lowered UVT - Full Power (SPt 1)
Lowered UVT - Full Power Duplicate
(SPt 1)
Lowered UVT - Full Power (SPt 2)
Lowered UVT- Full Power Duplicate
(SPt 2)
Lowered UVT - Full Power (SPt 3)
Lowered UVT- Full Power Duplicate
(SPt 3)
Lowered Power- High UVT (SPt 1)
Lowered Power- High UVT Duplicate
(SPt 1)
Lowered Power- High UVT (SPt 2)
Lowered Power- High UVT Duplicate
(SPt 2)
Lowered Power - High UVT (SPt 3)
Lowered Power - High UVT Duplicate
(SPt 3)
Run
2
3
4
5
6
7
10
11
12
13
14
15
Log Influent Concentration
Rep1
6.67
6.00
6.16
6.82
5.92
6.72
6.62
5.86
6.33
6.11
6.67
6.62
Rep 2
6.59
6.16
6.33
6.81
5.96
6.74
6.33
5.90
6.12
5.96
6.82
6.63
Rep 3
6.58
6.18
6.42
6.70
5.88
6.60
6.62
5.87
6.05
6.00
6.74
6.68
Rep 4
6.64
6.16
6.07
6.76
5.98
6.68
6.62
5.98
6.07
5.73
6.67
6.62
Rep 5
6.66
6.12
6.49
6.78
5.90
6.70
6.70
5.95
6.33
6.22
6.85
6.64
Log Effluent Concentration
Rep1
3.10
2.31
2.47
2.93
2.89
3.09
3.35
2.41
2.18
2.48
2.88
2.70
Rep 2
2.97
2.25
2.55
2.84
2.90
2.91
3.31
2.66
2.10
2.50
2.72
2.86
Rep 3
2.98
2.79
2.68
2.95
2.90
3.05
3.28
2.87
2.14
2.25
2.90
2.67
Rep 4
2.98
2.84
2.75
2.93
2.85
3.03
3.33
2.94
1.82
2.20
2.73
2.88
Rep 5
2.95
2.64
2.79
2.96
2.98
3.05
3.33
2.92
2.18
2.36
2.69
2.66
SPt - Set Point Condition
44
-------
NOVEMBER 2013
Table 4-8. ETS UV Model UVL-200-4 MS2 Log Inactivation Results
Test Condition
Lowered UVT - Full Power (SPt 1)
Lowered UVT - Full Power Duplicate (SPt 1)
Lowered UVT - Full Power (SPt 2)
Lowered UVT - Full Power Duplicate (SPt 2)
Lowered UVT - Full Power (SPt 3)
Lowered UVT - Full Power Duplicate (SPt 3)
Lowered Power - High UVT (SPt 1)
Lowered Power- High UVT Duplicate (SPt 1)
Lowered Power - High UVT (SPt 2)
Lowered Power - High UVT Duplicate (SPt 2)
Lowered Power - High UVT (SPt 3)
Lowered Power - High UVT Duplicate (SPt 3)
Run
2
3
4
5
6
7
10
11
12
13
14
15
Log Inactivation
Rep1
3.58
3.69
3.69
3.89
3.03
3.63
3.27
3.45
4.15
3.63
3.79
3.92
Rep 2
3.63
3.91
3.78
3.97
3.06
3.83
3.02
3.24
4.02
3.46
4.11
3.78
Rep 3
3.60
3.39
3.74
3.76
2.98
3.56
3.33
2.99
3.91
3.75
3.83
4.01
Rep 4
3.67
3.32
3.32
3.83
3.13
3.65
3.30
3.03
4.25
3.54
3.94
3.73
Rep 5
3.71
3.47
3.71
3.82
2.91
3.65
3.37
3.04
4.14
3.86
4.16
3.98
SPt - Set Point Condition
45
-------
NOVEMBER 2013
Table 4-9. ETS UV Model UVL-200-4 MS2 Observed RED Results
Test Condition
Lowered UVT- Full Power
(SPt 1)
Lowered UVT- Full Power
Duplicate (SPt 1)
Lowered UVT - Full Power
(SPt 2)
Lowered UVT - Full Power
Duplicate (SPt 2)
Lowered UVT- Full Power
(SPt 3)
Lowered UVT- Full Power
Duplicate (SPt 3)
Lowered Power- High
UVT (SPt 1)
Lowered Power- High
UVT Duplicate (SPt 1)
Lowered Power- High
UVT (SPt 2)
Lowered Power- High
UVT Duplicate (SPt 2)
Lowered Power- High
UVT (SPt 3)
Lowered Power- High
UVT Duplicate (SPt 3)
Run
2
3
4
5
6
7
10
11
12
13
14
15
RED
(mJ/cm2)
Rep1
74.36
77.37
77.27
82.60
60.96
75.76
65.88
70.35
88.74
74.81
79.19
82.63
Rep 2
75.66
83.06
79.58
84.68
61.57
80.95
59.71
65.01
85.21
70.57
87.52
78.72
Rep 3
75.00
69.80
78.57
79.01
59.63
73.87
67.36
59.07
82.26
77.96
80.27
85.01
Rep 4
76.68
67.98
67.89
80.99
63.34
76.33
66.47
60.01
91.49
72.50
83.19
77.54
Rep 5
77.71
71.79
77.74
80.68
58.07
76.30
68.33
60.11
88.56
80.99
89.11
84.25
Average
75.88
74.00
76.21
81.59(1)
60.71
76.64
65.55
62.91
87.25(1)
75.37
83.85(1)
81.63(1)
SD(RED)
1.34
6.17
4.74
2.15
1.99
2.61
3.39
4.76
3.57
4.18
4.36
3.33
USP
4.88
23.14
17.25
7.30
9.10
9.46
14.37
21.02
11.36
15.39
14.44
11.34
u
SD - Standard Deviation
SP . Uncertainty of the Set Point {[(Student t * SD)/REDave]*100}
SPt - Set Point Condition
(1) These RED values exceeded the highest dose in the collimated beam tests and therefore should be considered estimates. Since they
are above the maximum dose in the collimated beam test, the results can only truly be quantified as being >80 mJ/cm2.
46
-------
NOVEMBER 2013
4.6 Set Line for a Minimum RED of 40 mJ/cm
The three set point conditions selected for this validation all achieved a minimum RED of 40
ml/cm2, which was the target minimum RED for developing the set line. Figure 4-5 shows the
set line. The unit is validated for a minimum RED of 40 mJ/cm2 for any flow rate - intensity
combination above and to the left of the set line. The maximum flow rate demonstrated was 24.9
gpm. A UV system cannot operate above the highest validated flow rate and claim a 40 mJ/cm2
RED. The lowest intensity demonstrating a RED of 40 mJ/cm2 was 7.0 W/m2. A UV system
cannot operate below the lowest validated irradiance and claim a 40 mJ/cm2 RED.
Set Point 1 - 14.8 gpm; 7.0 W/m2
Set Point 2-19.9 gpm; 11.0 W/m2
Set Point 3 - 24.9 gpm; 12.8 W/m2
16.0
14.0
12.0
10.0
8.0
.JH 6.0
_c
4.0
2.0
0.0
0
10 15 20
Flow Rate {gpm)
25
30
Figure 4-5 Set Line for Model UVL-200-4 For Validated Dose of >40 mJ/cm2 based on MS2
47
-------
NOVEMBER 2013
4.7 Deriving the Validation Factor and Log Credit for Cryptosporidium
4.7.1 Validation Factor Definition
Several uncertainties and biases are involved in using experimental testing to define a validated
dose and validated operating conditions such as challenge microorganism UV sensitivity, and
sensor placement or variability. The validation factor (VF) for Cryptosporidium was determined
quantitatively to account for key areas of uncertainty and variability. The equation for the VF is
shown below.
VF = BRED x[l+(UVai/100)]
Where:
VF = Validation Factor;
BRED = RED bias factor;
Uvai = Uncertainty of validation expressed as a percentage.
The data used for the VF calculations and final results are presented in the following section.
4.7.2 RED Bias (BRED)
The RED bias factor (BRED) is a correction factor that accounts for the difference between the
UV sensitivity of a selected target pathogen and the UV sensitivity of the challenge
microorganism (MS2). If the challenge microorganism is more resistant (less sensitive) to UV
light than the target pathogen, the RED measured during the validation will be greater than the
RED that would be measured for the target pathogen. In this case, the RED bias would be greater
than 1.0. If the challenge microorganism is less resistant (more sensitive) to UV light than the
target pathogen, then RED measured by the validation will be less than the RED that would be
measured for the target pathogen.
A target pathogen must be selected to calculate the RED bias factor. For this test, the target
pathogen Cryptosporidium was selected for use in presenting an example calculation of RED
bias as it is a common pathogen that is evaluated for drinking water applications.
Cryptosporidium was also selected because the EPA's LT2ESWTR requires UV reactors be
validated to demonstrate a log inactivation for Cryptosporidium. A target of 3-log inactivation of
Cryptosporidium was selected as water utilities in the highest risk category or "bin" may need
this maximum level of inactivation. The RED bias tables in Appendix G of the UVDGM-2006
were used for determining the RED bias. The RED bias is determined from the Tables based on
the sensitivity calculated for each test run replicate at a given set point (test condition) and the
UVT of the water. Sensitivity is calculated as:
Sensitivity (mJ/cm2 per log I) = RED/ Log I
Per the GP-2011 and UVDGM-2006, the sensitivity is calculated for each test replicate (five per
test run, 20 samples total per set point). The highest BRED value found among the replicates at a
given set point is then selected for the BRED value for use in the VF calculation per the UVDGM-
2006 requirement. Table 4-10 shows the data for the replicates at each set point. The highest
48
-------
NOVEMBER 2013
RED bias at each set point is used in the validation factor calculations shown later in Section
4.7.3.
Table 4-10. RED Bias Factor for Each Set Point for Cryptosporidium
Sample
Number
2-1
2-2
2-3
2-4
2-5
3-1
3-2
3-3
3-4
3-5
4-1
4-2
4-3
4-4
4-5
5-1
5-2
5-3
5-4
5-5
6-1
6-2
6-3
6-4
6-5
7-1
7-2
7-3
7-4
7-5
10-1
10-1
10-1
10-1
10-1
11-1
11-2
11-3
11-4
11-5
12-1
12-2
Test
Run
2
2
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
6
6
6
6
6
7
7
7
7
7
10
10
10
10
10
11
11
11
11
11
12
12
UVT
%
78.5
78.5
78.5
78.5
78.5
78.5
78.5
78.5
78.5
78.5
89.7
89.7
89.7
89.7
89.7
89.7
89.7
89.7
89.7
89.7
93.5
93.5
93.5
93.5
93.5
93.5
93.5
93.5
93.5
93.5
97.0
97.0
97.0
97.0
97.0
97.1
97.1
97.1
97.1
97.1
97.1
97.1
Sensitivity
(mJ/cm2 per Log I)
RED Log I Sensitivity
74.36
75.66
75.00
76.68
77.71
77.37
83.06
69.80
67.98
71.79
77.27
79.58
78.57
67.89
77.74
82.60
84.68
79.01
80.99
80.68
60.96
61.57
59.63
63.34
58.07
75.76
80.95
73.87
76.33
76.30
65.88
59.71
67.36
66.47
68.33
70.35
65.01
59.07
60.01
60.11
88.74
85.21
3.58
3.63
3.60
3.67
3.71
3.69
3.91
3.39
3.32
3.47
3.69
3.78
3.74
3.32
3.71
3.89
3.97
3.76
3.83
3.82
3.03
3.06
2.98
3.13
2.91
3.63
3.83
3.56
3.65
3.65
3.27
3.02
3.33
3.30
3.37
3.45
3.24
2.99
3.03
3.04
4.15
4.02
20.8
20.9
20.8
20.9
21.0
21.0
21.2
20.6
20.5
20.7
20.9
21.1
21.0
20.5
21.0
21.2
21.3
21.0
21.1
21.1
20.1
20.1
20.0
20.2
19.9
20.9
21.1
20.8
20.9
20.9
20.1
19.8
20.2
20.2
20.3
20.4
20.1
19.7
19.8
19.8
21.4
21.2
BRED
4 log
crypto
1.97
1.97
1.97
1.97
1.97
1.97
1.97
1.97
1.97
1.97
1.77
1.77
1.77
1.77
1.77
1.77
1.77
1.77
1.77
1.77
1.61
1.61
1.61
1.61
1.55
1.61
1.61
1.61
1.61
1.61
1.36
1.34
1.36
1.36
1.36
1.36
1.36
1.34
1.34
1.34
1.36
1.36
BRED
3.5 log
crypto
2.35
2.35
2.35
2.35
2.35
2.35
2.35
2.35
2.35
2.35
2.01
2.01
2.01
2.01
2.01
2.01
2.01
2.01
2.01
2.01
1.75
1.75
1.75
1.75
1.70
1.75
1.75
1.75
1.75
1.75
1.40
1.38
1.40
1.40
1.40
1.40
1.40
1.38
1.38
1.38
1.40
1.40
BRED
3.0 log
crypto
2.54
2.54
2.54
2.54
2.54
2.54
2.54
2.54
2.54
2.54
2.10
2.10
2.10
2.10
2.10
2.10
2.10
2.10
2.10
2.10
1.78
1.78
1.78
1.78
1.73
1.78
1.78
1.78
1.78
1.78
1.39
1.38
1.39
1.39
1.39
1.39
1.39
1.38
1.38
1.38
1.39
1.39
49
-------
NOVEMBER 2013
Sample
Number
12-3
12-4
12-5
13-1
13-2
13-3
13-4
13-5
14-1
14-2
14-3
14-4
14-5
15-1
15-2
15-3
15-4
15-5
Maximum BRED
Test
Run
12
12
12
13
13
13
13
13
14
14
14
14
14
15
15
15
15
15
UVT
%
97.1
97.1
97.1
97.0
97.0
97.0
97.0
97.0
97.1
97.1
97.1
97.1
97.1
97.1
97.1
97.1
97.1
97.1
Set Point
Set Point
Set Point
Sensitivity
(mJ/cm2 per Log I)
RED Log I Sensitivity
82.26
91.49
88.56
74.81
70.57
77.96
72.50
80.99
79.19
87.52
80.27
83.19
89.11
82.63
78.72
85.01
77.54
84.25
14.8 gpm -
19.9gpm -
24. 9 gpm -
3.91
4.25
4.14
3.63
3.46
3.75
3.54
3.86
3.79
4.11
3.83
3.94
4.16
3.92
3.78
4.01
3.73
3.98
7.0 W/m2
1 1 .0 W/m2
12.8 W/m2
21.0
21.5
21.4
20.6
20.4
20.8
20.5
21.0
20.9
21.3
20.9
21.1
21.4
21.1
20.8
21.2
20.8
21.1
BRED
4 log
crypto
1.36
1.36
1.36
1.36
1.36
1.36
1.36
1.36
1.36
1.36
1.36
1.36
1.36
1.36
1.36
1.36
1.36
1.36
1.97
1.77
1.61
BRED
3.5 log
crypto
1.40
1.40
1.40
1.40
1.40
1.40
1.40
1.40
1.40
1.40
1.40
1.40
1.40
1.40
1.40
1.40
1.40
1.40
2.35
2.01
1.75
BRED
3.0 log
crypto
1.39
1.39
1.39
1.39
1.39
1.39
1.39
1.39
1.39
1.39
1.39
1.39
1.39
1.39
1.39
1.39
1.39
1.39
2.54
2.10
1.78
4.7.3 Uncertainty of Validation
The uncertainty of validation (Uvai) addresses many sources of experimental uncertainty. As the
critical source of uncertainty, such as the sensor readings, or the fit of the dose-response curve, is
unknown in advance of the validation testing, the USEPA developed a decision tree to assist in
establishing Uvai. The GP-2011 equations and in accordance with Figure 5.4 of the UVDGM-
2006, which are specific to a UV intensity set point approach, were used to determine Uvai in
calculating the validated dose. Per the GP-2011 and the EPA's UVDGM-2006, any of the
following equations may be used to establish the Uvai:
Uvai =
+UDR2)1/2
Where:
Us = Uncertainty of sensor value, expressed as a fraction;
UDR = Uncertainty of the fit of the dose-response curve;
USP = Uncertainty of set-point;
Uvai = Uncertainty of the validation
50
-------
NOVEMBER 2013
The QC objective for the duty sensor is that the measurements with the duty sensor should be
<10% of the average of two or more reference sensors. It this objective is met, then it eliminates
the need to calculate the Us factor per the GP-2011 and UVDGM-2006, Section 5.4.4. The
sensor met the 10% requirement, as shown in Table 4-1, therefore Us is not used in determining
the uncertainty of validation.
The GP-2011 and UVDGM-2006 in Appendix C Section C4 show the formula and calculations
for the uncertainty of the fit of the collimated beam dose response curve (UDR).
The equation is:
UoR = t * [SD/ UV DoseCB] * 100%
Where:
UDR = Uncertainty of the UV dose-response fit at a 95% confidence level
UV DoseCB = UV dose calculated from the UV dose-response curve for the
challenge microorganism
SD = Standard deviation of the difference between the calculated UV dose
response and the measured value
t = t-statistic at a 95% confidence level for a sample size equal to the number of
test condition replicates used to define the dose-response.
The UDR results are shown in Tables 4-3 and 4-4 for the low and high UVT waters for the test
runs. The UDR results for low and high UVT waters (26.73% and 18.14%, respectively) are less
than 30%, and therefore UDR is not used in calculating Uyai for the test runs.
The uncertainty in the set point value (Usp) is based on a prediction interval at a 95% confidence
level using the following procedure:
1. Calculate the average and standard deviation of REDmeas values for each test condition
2. Calculate the uncertainty of the set point REDmeas using:
USP = [(t x SDRED) / (REDmeas)] x 100%
Where:
REDmeas = Average REDmeas value measured for each test condition;
SDRED = Standard deviation of the REDmeas values measured for each test
condition;
t = t-statistic for a 95% confidence level defined as a function of the number of
replicate samples, in this case 5 replicates were used for testing yielding a t value
of 2.776 (n-1 =4).
3. Select the highest USP from the replicates at each set point for calculating the VF.
The USP results based on the REDmeas and standard deviation are shown in Table 4-9. In
accordance with the GP-2011, the highest USP of the four test runs at each set point determines
51
-------
NOVEMBER 2013
the USP for that set point. The highest USP for each set point is 23.14% (15 gpm set point),
17.25% (20 gpm set point), and 14.44% (25 gpm set point).
The uncertainty of the validation is equal to the highest USp at a set point when the UDR is <30%.
Therefore Uvai is calculated using the equation:
Uval = USP
Table 4-11 shows the Uvai values used for determining the uncertainty of the validation at each
set point.
Table 4-11 Uncertainty of the Validation (Uvai) and BRED Values for Cryptosporidium
Set Point
Set Point 14. 8 gpm -
Set Point 19. 9 gpm -
Set Point 24. 9 gpm -
7.0 W/m2
1 1 .0 W/m2
12.8W/m2
Max
UDR
%
26.73
26.73
26.73
Max
USP
%
23.14
17.25
14.44
Uvai
%
23.14
17.25
14.44
4.0 log
1.97
1.77
1.61
Max
BRED
3.5 log
2.35
2.01
1.75
3.0 log
2.54
2.10
1.78
4.7.4 Validated Dose and Set Line for Cryptosporidium
After establishing the Uvai and the RED bias as described above, the validation factor (VF) is
calculated using the equation:
W = BRED x[l+(UVai/100)]
Where:
VF = Validation Factor;
BRED = RED bias factor for Cryptosporidium
Uvai = Uncertainty of validation expressed as a percentage
The validated dose is then calculated as follows:
Validated dose (REDVai) = REDmeas / VF
Table 4-12 shows the calculated Validation Factors (VF) for various Cryptosporidium log
inactivation levels (3.0, 3.5, and 4.0 log inactivation).
Table 4-12 shows the REDyai for Cryptosporidium for each test run using the validation factors
for the various Cryptosporidium log inactivation levels and a comparison to the dose required for
various levels of inactivation of Cryptosporidium. All of the set points achieved a validated dose
that shows a minimum of a 4-log inactivation for Cryptosporidium. This level of inactivation
exceeds the minimum 3.0 log inactivation of Cryptosporidium, which may be required by the
EPA's LT2ESWTR in cases where a utility is in the highest "bin" or risk category for
Cryptosporidium in their source water.
52
-------
NOVEMBER 2013
Table 4-12 Validation Factors and Validated Dose (REDVai) for Cryptosporidium
Set Point Condition
Lowered UVT - Full Power
(SPt 1)
Lowered UVT - Full Power
Dup(SPtl)
Lowered Power - High
UVT(SPtl)
Lowered Power - High
UVTDup(SPtl)
Lowered UVT - Full Power
(SPt 2)
Lowered UVT - Full Power
Dup(SPt2)
Lowered Power- High
UVT (SPt 2)
Lowered Power- High
UVT Dup (SPt 2)
Lowered UVT - Full Power
(SPt 3)
Lowered UVT - Full Power
Dup (SPt 3)
Lowered Power- High
UVT (SPt 3)
Lowered Power- High
UVT Dup (SPt 3)
Run
2
3
10
11
4
5
12
13
6
7
14
15
Flow
Rate
gpm
14.9
14.8
14.9
14.9
19.9
19.9
19.9
19.9
24.9
24.9
24.9
25.0
Intensity
W/m2
7.0
7.0
6.9
6.9
11.0
11.0
11.0
11.0
12.8
12.8
12.7
12.7
Validation Factor
4.0 log 3.5 log 3.0 log
2.43
2.43
2.43
2.43
2.08
2.08
2.08
2.08
1.84
1.84
1.84
1.84
2.89
2.89
2.89
2.89
2.36
2.36
2.36
2.36
2.00
2.00
2.00
2.00
3.13
3.13
3.13
3.13
2.46
2.46
2.46
2.46
2.04
2.04
2.04
2.04
REDmeas
mJ/cm2
75.88
74.00
76.21
81.59
60.71
76.64
65.55
62.91
87.25
75.37
83.85
81.63
4 log
mJ/cm2
22<1>
31.3
30.5
27.0
25.9
36.7
39.3
42.0
36.3
33.0
41.6
45.5
44.3
REDval
3.5 log
mJ/cm2
15(1)
26.2
25.6
22.7
21.7
32.3
34.6
37.0
32.0
30.3
38.3
41.9
40.8
3.0 log
mJ/cm2
12d)
24.3
23.7
21.0
20.1
31.0
33.1
35.4
30.6
29.8
37.6
41.2
40.1
(1) Required dose for log inactivation validation per the UVDGM-2006 Appendix G; SPt - Set Point Condition
53
-------
NOVEMBER 2013
The three set point tests that achieved a minimum of 3 log inactivation for Cryptosporidium were
plotted to form a set line. Figure 4-6 shows the set line.
The three set points are:
Set Point 1 - 14.8 gpm; 7.0 W/m2
Set Point 2-19.9 gpm; 11.0 W/m2
Set Point 3 - 24.9 gpm; 12.8 W/m2
16.0
14.0
£J 6.0
4.0
2.0
0.0
10 15 20
Flow Rate {gpm)
25
30
Figure 4-6. Set Line for Minimum 3-log Cryptosporidium Inactivation for ETS UV Model
UVL-200-4.
54
-------
NOVEMBER 2013
4.8 Validated Dose (REDVai) for MS2 as the Target Organism
Some regulatory agencies, such as the NYDOH, have established a standard for spray park water
features and other applications based on a validated dose (REDVai) of 40 ml/cm2 based on MS2,
as the pathogen. The calculation of the validation factor for a validated dose based on MS2 is
performed using BRED set equal to 1.0. For MS2 validated dose calculations, BRED is set equal to
1.0 because the pathogen selected, namely MS2, is the same as the test organism, so there is no
bias correction. Therefore, the validation factor will not vary by the log inactivation level.
The Uvai is calculated in the same manner as described in Section 4.7.3.
The validation factor (VF) for evaluating validated dose (REDVai) based on MS2 is calculated
using the same formula as for other pathogens as follows:
W = BRED x[l+(UVai/100)]
Where:
VF = Validation Factor;
BRED = RED bias factor (set equal 1.0)
UVai = Uncertainty of validation expressed as a percentage.
The validated dose is then calculated as follows:
Validated dose (REDVai) = REDobserved / VF
Table 4-13 shows the REDVai based on MS2 for each test run.
Using the VF calculated for each set point, the REDVai based on MS2 was calculated for each
test run. All of the set point test runs achieved a 40 mJ/cm2 validated dose based on MS2.
The three set point tests, which achieved a minimum 40 mJ/cm2 validated dose (REDyai based on
MS2), were plotted to form a set line. Figure 4-7 shows the set line.
The three set points are:
Set Point 1 - 14.8 gpm; 7.0 W/m2
Set Point 2-19.9 gpm; 11.0 W/m2
Set Point 3 - 24.9 gpm; 12.8 W/m2
55
-------
NOVEMBER 2013
Table 4-13 Validation Factors and Validated Dose (REDVai) based on MS2
Set Point Condition
Lowered
Lowered
(SPt1)
Lowered
Lowered
(SPt1)
Lowered
Lowered
(SPt2)
Lowered
Lowered
(SPt2)
Lowered
Lowered
(SPt 3)
Lowered
Lowered
(SPt 3)
UVT-
UVT-
Power
Power
UVT-
UVT-
Power
Power
UVT-
UVT-
Power
Power
Full Power (SPt1)
Full Power Dup
- High UVT (SPt 1)
- High UVT Dup
Full Power (SPt 2)
Full Power Dup
- High UVT (SPt 2)
- High UVT Dup
Full Power (SPt 3)
Full Power Dup
- High UVT (SPt 3)
- High UVT Dup
Run
2
3
12
13
4
5
14
15
6
7
16
17
Flow
Rate
gpm
14.9
14.8
14.9
14.9
19.9
19.9
19.9
19.9
24.9
24.9
24.9
25.0
Intensity
W/m2
7.0
7.0
6.9
6.9
11.0
11.0
11.0
11.0
12.8
12.8
12.7
12.7
Validation
Factor
(1)
1.23
1.23
1.23
1.23
1.17
1.17
1.17
1.17
1.14
1.14
1.14
1.14
REDmeas
mJ/cm2
75.
74.
65.
62.
76.
81.
87.
75.
60.
76.
83.
81.
9
0
5
9
2
6
2
4
7
6
9
6
REDVal
based on
MS2
mJ/cm2
61.6
60.1
53.2
51.1
65.0
69.6
74.4
64.3
53.1
67.0
73.3
71.3
(1) BRED equal to 1.0 as the target organism is MS2 the same as the test organism.; SPt -Set Point Condition
4.O
2.0
0.0
10 15 20
Flow Rate (gpm)
25
30
Figure 4-6. Set Line for Minimum 40 mJ/cm Validated Dose (REDVai) based on MS2 for
ETS UV Model UVL-200-4.
56
-------
NOVEMBER 2013
4.9 Water Quality Data
Samples were collected for general water quality characterization. Influent and effluent samples
were collected during each flow test run and analyzed for temperature, pH, total chlorine, and
free chlorine. An influent sample was collected from each flow test run and analyzed for
turbidity, iron, and manganese.
An influent and effluent sample from each test run was also collected and analyzed for total
coliform, E. coli, and heterotrophic plant count (HPC).
The general chemistry and microbiological results are presented in Tables 4-14 through 4-17.
57
-------
NOVEMBER 2013
Table 4-14. Temperature and pH Results
Test
Blank
Lowered UVT - Full Power
Lowered UVT - Full Power Duplicate
Lowered UVT - Full Power
Lowered UVT - Full Power Duplicate
Lowered UVT - Full Power
Lowered UVT - Full Power Duplicate
Reactor Blank
Reactor Control
Lowered Power- High UVT
Lowered Power- High UVT Duplicate
Lowered Power- High UVT
Lowered Power- High UVT Duplicate
Lowered Power- High UVT
Lowered Power- High UVT Duplicate
Run#
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Temperature
(°F)
Influent
68.0
68.8
68.3
67.6
67.3
67.2
66.8
66.2
66.6
67.1
66.8
67.2
67.0
68.1
67.9
Effluent
68.2
69.0
68.5
67.8
67.3
67.4
67.1
66.4
66.6
67.0
67.0
67.3
66.8
67.9
67.7
PH
(S.U.)
Influent
8.80
8.71
8.88
8.86
8.83
8.89
8.94
9.00
9.03
9.00
8.97
9.00
8.97
8.95
8.98
Effluent
8.82
8.79
8.81
8.86
8.82
8.89
8.87
8.94
9.00
9.01
9.00
9.01
8.96
9.00
8.97
Table 4-15. Total Chlorine, Free Chlorine and Turbidity Results
Test
Blank
Lowered UVT - Full Power
Lowered UVT - Full Power Duplicate
Lowered UVT - Full Power
Lowered UVT - Full Power Duplicate
Lowered UVT - Full Power
Lowered UVT - Full Power Duplicate
Reactor Blank
Reactor Control
Lowered Power- High UVT
Lowered Power- High UVT Duplicate
Lowered Power- High UVT
Lowered Power- High UVT Duplicate
Lowered Power- High UVT
Lowered Power- High UVT Duplicate
Run#
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Total
Chlorine
(mg/L)
Influent
<0.03
<0.03
<0.03
<0.03
<0.03
<0.03
<0.03
<0.03
<0.03
<0.03
<0.03
<0.03
<0.03
<0.03
<0.03
Free
Chlorine
(mg/L)
Influent
<0.03
<0.03
<0.03
<0.03
<0.03
<0.03
<0.03
<0.03
<0.03
<0.03
<0.03
<0.03
<0.03
<0.03
<0.03
Turbidity
(NTU)
Influent
0.54
0.49
0.51
0.40
0.38
0.38
0.31
0.17
0.18
0.22
0.21
0.22
0.20
0.21
0.20
Note: Runs 1-7 with the addition of LSAto lower UVT showed higher readings for turbidity;
suspect an interference due to the LSA
58
-------
NOVEMBER 2013
Table 4-16. Iron and Manganese Results
Test
Blank
Lowered UVT - Full Power
Lowered UVT - Full Power
Duplicate
Lowered UVT - Full Power
Lowered UVT - Full Power
Duplicate
Lowered UVT - Full Power
Lowered UVT - Full Power
Duplicate
Reactor Blank
Reactor Control
Lowered Power- High UVT
Lowered Power- High UVT
Duplicate
Lowered Power- High UVT
Lowered Power- High UVT
Duplicate
Lowered Power- High UVT
Lowered Power- High UVT
Duplicate
Run#
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Iron
(mg/L)
Influent
0.03
0.02
0.02
<0.02
<0.02
0.02
<0.02
<0.02
<0.02
<0.02
<0.02
<0.02
<0.02
<0.02
<0.02
Manganese
(mg/L)
Influent
0.002
0.002
0.002
0.001
0.001
0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
UVT<1)
(%)
Influent
79
79
79
90
90
93
93
97
95
96
96
96
96
96
96
Effluent
79
79
79
90
90
93
93
97
95
97
96
96
94
96
94
(l)UVT on grab samples, measured in laboratory after tests; Five influent samples averaged; single effluent
sample reported here; In-line UVT meter used for flow test results
59
-------
NOVEMBER 2013
Table 4-17. HPC, Total Coliform and E. coli Results.
Test
Blank
Lowered UVT - Full Power
Lowered UVT - Full Power
Duplicate
Lowered UVT - Full Power
Lowered UVT - Full Power
Duplicate
Lowered UVT - Full Power
Lowered UVT - Full Power
Duplicate
Reactor Blank
Reactor Control
Lowered Power- High UVT
Lowered Power- High UVT
Duplicate
Lowered Power- High UVT
Lowered Power- High UVT
Duplicate
Lowered Power- High UVT
Lowered Power- High UVT
Duplicate
Run#
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Total Coliform
MPN/100ml_
Influent
54
31
31
144
115
67
50
15
56
49
37
46
40
28
34
Effluent
<1
3
<1
<1
4
<1
<1
<1
64
3
16
<1
<1
<1
<1
E. co//
MPN/100ml_
Influent
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
Effluent
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
HPC
CFU/mL
Influent
1.30E+04
1 .62E+04
1.34E+04
1 .27E+04
1.34E+04
1.12E+04
1.02E+04
1 .40E+04
1 .27E+04
1 .03E+04
1.51E+04
1 .36E+04
1 .26E+04
1 .48E+04
1 .04E+04
Effluent
1 .45E+02
3.75E+01
4.20E+01
8.05E+01
2.90E+01
7.20E+01
2.35E+01
5.20E+02
1.30E+04
4.25E+01
2.85E+01
5.70E+02
1.72E+02
6.30E+02
1.15E+02
60
-------
NOVEMBER 2013
4.10 Headloss
Headless was measured over the flow range of 5 to 25 gpm. Pressure at the inlet and outlet of the
reactor was measured at several flow rates as shown in Table 4-18.
Table 4-18. Headloss Measurement Results.
Flow Rate
5.2
10.4
15.5
19.9
25.3
Inlet (psi)
13.10
11.95
10.25
8.27
5.12
Outlet (psi)
13.10
11.94
10.18
8.15
4.94
Headloss (psi)
0.00
0.01
0.07
0.12
0.18
4.11 Power Measurement
A power monitoring platform was connected to the unit. This monitoring platform provided
continuous readout of the voltage and amperage being used by the unit for each test run. Volts
and amperes were recorded during each flow test. A series of power measurements were also
made to show the change in intensity at various power down levels. Table 4-19 presents the
power measurements taken during the flow tests.
Table 4-19. Power Measurement Results
Test
Blank
Lowered UVT - Full Power
Lowered UVT - Full Power Duplicate
Lowered UVT - Full Power
Lowered UVT - Full Power Duplicate
Lowered UVT - Full Power
Lowered UVT - Full Power Duplicate
Reactor Blank
Reactor Control
Lowered Power- High UVT
Lowered Power- High UVT Duplicate
Lowered Power- High UVT
Lowered Power- High UVT Duplicate
Lowered Power- High UVT
Lowered Power- High UVT Duplicate
Run#
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Unit
Volts
(volts)
207.8
208.0
207.6
206.1
206.3
206.4
206.3
206.6
206.1
207.1
207.1
208.0
207.3
207.6
207.2
Unit
Amperage
(amps)
0.90
0.90
0.90
0.91
0.91
0.91
0.90
0.91
0.04
0.63
0.63
0.72
0.72
0.77
0.77
61
-------
NOVEMBER 2013
Chapter 5
Quality Assurance/Quality Control
5.1 Introduction
An important aspect of verification is conformance to the QA/QC procedures and requirements
specified in the TQAPP and Generic Protocols. Careful adherence to the procedures ensures that
the data presented in this report is of sound quality, defensible, and representative of the
equipment performance. The primary areas of evaluation were representativeness, accuracy,
precision, and completeness.
Because this ETV was conducted at the NSF testing lab, all laboratory activities were conducted
in accordance with the provisions of the NSF International Laboratories Quality Assurance
Manual.
5.2 Test Procedure QA/QC
NSF testing laboratory staff conducted the tests by following a USEPA-approved test/QA plan(1)
created specifically for this verification. NSF QA Department staff performed an audit during
testing to ensure the proper procedures were followed. The audit yielded no significant findings.
5.3 Sample Handling
All samples analyzed by the NSF Chemistry and Microbiology Laboratories were labeled with
unique identification numbers. All samples were analyzed within allowable holding times.
5.4 Chemistry Laboratory QA/QC
The calibrations of all analytical instruments and the analyses of all parameters complied with
the QA/QC provisions of the NSF International Laboratories Quality Assurance Manual.
The NSF QA/QC requirements are all compliant with those given in the USEPA method or
Standard Method for the parameter. Also, every analytical method has an NSF standard
operating procedure.
The bench top UV spectrophotometer was calibrated with Holmium Oxide with each batch of
samples analyzed and showed peaks at 241. Inm, 250.Onm and 278.1 nm within + 0.2nm of the
actual peak. Bichromate standards were also run with each batch of samples and found to be
within 1% of the true value.
5.5 Microbiology Laboratory QA/QC
5.5.1 Growth Media Positive Controls
All media were checked for sterility and positive growth response when prepared and when used
for microorganism enumeration. The media was discarded if growth occurred on the sterility
check media, or if there was an absence of growth in the positive response check.
62
-------
NOVEMBER 2013
5.5.2 Negative Controls
For each sample batch processed, an unused membrane filter and a blank with 100 mL of
buffered, sterilized dilution water was filtered through the membrane, placed onto the
appropriate media and incubated with the samples as negative controls. No growth was observed
on any blanks.
5.5.3 Collimated Beam Apparatus and QA/QC
The petri dish factor was determined for the collimated beam apparatus prior to the start of the
test program. Radiometers were calibrated and checked in accordance with operating procedures
and UVDGM-2006 requirements. These procedures and data were reviewed as part of the NSF
QA department review of the microbiological laboratory data.
NSF received reviewer comments about the collimated beam data in Draft EPA ETV NSF UV
reports in November of 2011. They identified two issues related to collimated beam data for
NSF to investigate. The two issues were a high degree of uncertainty with replicates in the
collimated beam data and the data trending at or below the lower 95% confidence interval for
MS2 UV sensitivity. The initial investigation revealed no systematic error. However, further
investigation revealed a miscommunication between the company that calibrated NSF's
radiometers and NSF. The radiometers were calibrated at one of two possible settings. As the
company did not inform NSF of which setting it used to calibrate the radiometer, NSF used the
non-calibrated radiometer setting for its CB tests. Hence the MS2 had received about 25% more
UV dose than estimated by the CB data. Therefore, the calculated REDmeas at each set point was
lower than the actual dose delivered by the unit. NSF determined that the best action was to
retest all previous tested units. That testing was done in the summer of 2012 and is reported in
this report. All previous data is not reported herein as it was deemed to be biased low.
The factors used in the collimated test shown below were evaluated against the protocol
requirements and found to meet the QC objectives. The length (distance from the lamp centerline
to the suspension) and the depth of suspension were fixed parameters. These measurements were
made multiple times at the "fixed mark" on the collimated beam apparatus to estimate the
precision of the measurements. The time was checked based on a stop watch with minimal
uncertainty. The petri dish factor was measured several times prior to the start of the test.
Absorbance uncertainty is based on spectrophotometer precision, as is the related reflectance
factor. The average intensity is measured for every collimated beam test, as it is required that
intensity be measured before and after each test.
To control for error in the UV dose measurement, the uncertainties of the terms in the UV dose
calculation met the following criteria:
Estimated Required
• Depth of suspension (d) <5% < 10%
• Average incident irradiance (Es) 2.5% < 8%
• Petri Factor (Pf) 2.1% < 5%
• L/(d + L) 0.7% < 1%
• Time(t) 1.6% < 5%
• (l-10-ad)/ad 1.2% < 5%
63
-------
NOVEMBER 2013
Trip blanks are normally performed to show that the stock phage solution does not change during
shipment to and from the test site. The phage stock solution was delivered from the microbiology
laboratory in the same building as the test rig before each test run and the samples were returned
to the laboratory after each test run. Therefore trip blanks were not required for these tests, as all
stock solution and test samples were received from and delivered to the microbiology laboratory
before/after each test run. No shipping or long holding times was required. However, Trip
Blanks were analyzed for this project to demonstrate that no change was occurring. The results
are shown in Table 5-1.
Table 5-1. Trip Blank Results
Date
June 20, 20 12 Day 1
June 2 1,20 12 Day 2
Trip Blank
Lab Retained
(PFU/mL
MS2)
6.90E+08
5.73E+08
Log10
8.84
8.76
Trip Blank Travel to Test
Rig and Returned
(PFU/mL
MS2)
6.73E+08
5.90E+08
Log10
8.83
8.77
Difference
Log10
0.01
0.01
Stability tests for MS2 are normally performed to show that the phage does not change during
holding times when samples are shipped from the test site to the laboratory and/or held in the
laboratory prior to analysis. However, for these tests, the test rig was located in the same
building as the microbiology laboratory. Samples were delivered to the laboratory after each test
run and the laboratory ran the samples within 4 to 6 hours of sample collection. Stability samples
were run for informational purposes even though the holding time was very short.
Table 5-2. MS2 Stability Test Results
MS2 Stability Test Results
High UVT 97%
Influent 0 Hour
Influent 4 Hour
Influent 8 Hour
Influent 24 Hour
High UVT 97%
Effluent OHour
Effluent 4 Hour
Effluent 8 Hour
Effluent 24 Hour
PFU/mL
2.54E+03
1.38E+03
4.60E+02
4.97E+02
Average
3.93E+03
1.45E+03
4.53E+02
5.40E+02
Log 10
3.40
3.14
2.67
2.70
Log 10
3.59
3.16
2.66
2.73
Low UVT 79%
Influent 0 Hour
Influent 4 Hour
Influent 8 Hour
Influent 24 Hour
Low UVT 79%
Effluent OHour
Effluent 4 Hour
Effluent 8 Hour
Effluent 24 Hour
PFU/mL
3.50E+03
2.02E+03
8.13E+02
8.60E+02
Average
3.93E+03
1.38E+04
6.53E+03
6.60E+03
Log 10
3.54
3.30
2.91
2.93
Log 10
3.59
4.14
3.81
3.82
5.6 Engineering Lab - Test Rig QA/QC
The flow meter for the test rig is part of the NSF tank, pump, and flow control system used for
UV testing and other tests in the engineering laboratory. The flow meter is calibrated by the NSF
QA staff at least annually. Calibration is performed by measuring the draw down volume from
the calibrated feed tank over time. The tank was calibrated by filling with measured volumes of
water and the corresponding depth measured. In addition to the annual calibration, the flow
meter was calibrated prior to the start of these test runs. Calibration was performed at 15, 20 and
25 gpm covering the range of expected flow rates. The flow meter accuracy fell within a range of
64
-------
NOVEMBER 2013
0.5 to 1.6% of the measured tank draw down rate over the range of test flow rates. The
calibration data for the flow meter is shown in Table 5-3 and achieved the requirement of+/- 5%.
Table 5-3. Flow Meter Calibration Results
Meter Flow Rate
Read by meter
(gpm)
15.45
19.90
24.95
Volume from
Tank
(gallons)
270.27
160.97
427.26
Run Time
(min:sec:millisec)
17:24:24
8:00:00
16:50:86
Flow Rate
Calculated
(gpm)
15.53
20.12
25.36
Percent
Difference
(%)
0.5
1.1
1.6
A reactor control and a reactor blank were performed as part of the validation. One reactor
control, with MS2 coliphage injection, and the lamps off, was performed to demonstrate that the
MS2 concentration was not changing as the seeded water passed though the reactor. A reactor
blank was collected to demonstrate that the system was not accumulating or being contaminated
with MS2 at levels that would interfere with the test.
Table 5-4 presents the results of the reactor control and reactor blanks. The reactor control had
an average influent concentration of 5.66 logic and an average effluent concentration of 5.78
logio showing a difference of 0.12 logic through the system with lamps off. This meets the
criteria of less than a 0.2 logio change through the unit with lamps turned off.
The reactor blank results showed no measureable MS2 in the system.
The results for the blank samples for HPC, total coliform, and E. coli were presented in Table 4-
17.
5.7 Documentation
All laboratory activities were documented using specially prepared laboratory bench sheets and
NSF laboratory reports. Data laboratory reports were entered into Microsoft™ Excel®
spreadsheets. These spreadsheets were used to calculate the means and logio reductions. One
hundred percent of the data entered into the spreadsheets was checked by a reviewer to confirm
all data and calculations were correct.
5.8 Data Review
NSF QA/QC staff reviewed the raw data records for compliance with QA/QC requirements. As
required in the ETV Quality Management Plan, NSF ETV staff checked at least 10% of the data
in the NSF laboratory reports against the lab bench sheets.
65
-------
NOVEMBER 2013
Table 5-4. Reactor Control and Reactor Blank MS2 Results
Test Condition
Reactor Blank
Reactor Blank
Reactor Control
Test Condition
Reactor Blank
Reactor Blank
Reactor Control
Test
Run
1
8
9
Test
Run
1
8
9
UVT
(%)
79
97
97.2
UVT
(%)
79
97
97.2
Flow
(gpm)
15.0
15.2
15.0
Flow
(gpm)
15.0
15.2
15.0
Intensity
(W/m2)
7.0
15.0
0.0
Intensity
(W/m2)
7.0
15.0
0.0
Influent (pfu/mL)
Rep 1
<1
<1
2.39E+05
Rep 2
<1
<1
4.83E+05
Rep 3
<1
<1
8.50E+05
Influent logio
Repl
0.0
0.0
5.38
Rep 2
0.0
0.0
5.68
Rep 3
0.0
0.0
5.93
Effluent (pfu/mL)
Repl
<1
<1
6.23E+05
Rep 2
<1
<1
6.53E+05
Rep 3
<1
<1
5.47E+05
Effluent Iog10
Repl
0.0
0.0
5.79
Rep 2
0.0
0.0
5.81
Rep 3
0.0
0.0
5.74
66
-------
NOVEMBER 2013
5.9 Data Quality Indicators
The quality of data generated for this ETV verification is established through four indicators of
data quality: representativeness, accuracy, precision, and completeness.
5.9.1 Representativeness
Representativeness is a qualitative term that expresses "the degree to which data accurately and
precisely represent a characteristic of a population, parameter variations at a sampling point, a
process condition, or an environmental condition." Representativeness was ensured by
consistent execution of the test protocol for each challenge, including timing of sample
collection, sampling procedures, and sample preservation. Representativeness was also ensured
by using each analytical method at its optimum capability to provide results that represent the
most accurate and precise measurement it is capable of achieving.
5.9.2 Accuracy
Accuracy was quantified as the percent recovery of the parameter in a sample of known quantity.
Accuracy was measured through use of both matrix spikes of a known quantity, where
applicable, and certified standards during calibration of an instrument.
The following equation was used to calculate percent recovery:
Percent ReCOVery = 100 X [(Xknown - Xmeasured)/Xknown]
Where:
Xknown = known concentration of the measured parameter
Xmeasured = measured concentration of parameter
Accuracy of the bench top chlorine, pH, and turbidity meters was checked daily during the
calibration procedures using certified check standards. The in-line UVT monitor was calibrated
daily with both a purchased UVT standard and with DI water at 99.9% UVT before the flow
tests.
The NSF Laboratory Quality Assurance Manual establishes the frequency of spike sample
analyses at 10% of the samples analyzed for chemical analyses. Laboratory control samples are
also run at a frequency of 10%. The recovery limits specified for the parameters in this
verification, excluding microbiological analyses, were 70-130% for laboratory-fortified (spiked)
samples and 85-115% for laboratory control samples. The NSF QA department reviewed the
laboratory records and found that all recoveries were within the prescribed QC requirements.
Calibration requirements were also achieved for all analyses.
5.9.3 Precision
Precision refers to the degree of mutual agreement among individual measurements and provides
an estimate of random error. One sample per batch was analyzed in duplicate for the iron and
manganese measurements. At least one out of every ten samples for pH, total chlorine, free
chlorine, temperature, and turbidity was analyzed in duplicate as part of the daily calibration
67
-------
NOVEMBER 2013
process. Precision of duplicate analyses was measured by use of the following equation to
calculate RPD:
RPD =
x200
Where:
S1 = sample analysis result; and
(5*2 = sample duplicate analysis result.
Acceptable analytical precision for the verification test was set at an RPD of 30%. Field
duplicates were collected at a frequency of one out of every 10 samples for each parameter, to
incorporate both sampling and analytical variation to measure overall precision against this
objective. In addition, the NSF Laboratory also conducted laboratory duplicate measurements at
10% frequency of samples analyzed. The laboratory precision for the methods selected was
tighter than the 30% overall requirement, generally set at 20% based on the standard NSF
Chemistry Laboratory method performance.
All RPD were within NSF's established allowable limits for each parameter.
5.9.4 Completeness
Completeness is the proportion of valid, acceptable data generated using each method as
compared to the requirements of the TQAP plan. The completeness objective for data generated
during validation testing is based on the number of samples collected and analyzed for each
parameter and/or method, as presented in Table 5-5.
Table 5-5. Completeness Requirements
Number of Samples per Parameter and/or Method
0-10
11-50
>50
Percent Completeness
80%
90%
95%
Completeness is defined as follows for all measurements:
%C = (V/T)xlOO
Where:
%C = percent completeness;
V = number of measurements judged valid; and
T = total number of measurements.
One hundred percent completeness was achieved for all aspects of this validation. All planned
testing activities were conducted as scheduled, and all planned samples were collected for
challenge organism and water chemistry analyses.
68
-------
NOVEMBER 2013
Chapter 6
References
1. Test/Quality Assurance Plan for The ETS UV Ultraviolet (UV) Reactor, Medium
Pressure Lamps, June 2010
2. Generic Protocol for Development of Test/Quality Assurance Plans for Validation of
Ultraviolet (UV) Reactors, NSF International, 7/2010.
3. Protocol for Development of Test / Quality Assurance Plans for Validation of Ultraviolet
(UV) Reactors August 2011 10/01/EPADWCTR.
4. Ultraviolet Disinfection Guidance Manual For the Long Term 2 Enhanced surface Water
Treatment Rule, Office of Water, US Environmental Protection Agency, November 2006,
EPA815-R-06-007
5. German Association for Gas and Water (DVGW) Technical Standard
Work Sheet W 294-1,2,3 (June 2006)
6. Austrian Standards, ONORM M5873-1, Plants for the disinfection of water using
ultraviolet radiation, Requirements and testing, Low pressure mercury lamp plants (
March, 2001)
7. APHA, AWWA, and WEF (1999). Standard Methods for the Examination of Water and
Wastewater, 20th Edition.
8. NSF International (2007). NSF/ANSI Standard 55 - Ultraviolet Microbiological Water
Treatment Systems.
9. NSF International (2011). NSF/ANSI Standard 50 - Equipment for Swimming Pools,
Spas, Hot Tubs and Other Recreational Water Facilities
10. Water Report 113, Safe, Sufficient and Good Potable Water Offshore: A guideline to
design and operation of offshore potable water systems. 2nd edition. By Eyvind Andersen
and Bj0rn E. L0fsgaard
11. Recommended Standards For Water Works, Policies for the Review and Approval of
Plans and Specifications for Public Water Supplies, 2012 Edition, A Report of the Water
Supply Committee of the Great Lakes-Upper Mississippi River Board of State and
Provincial Public Health and Environmental Managers.
69
-------
NOVEMBER 2013
Attachment 1
Model UVL-200-4 Operating and Technical Manual
Supporting Technical Data
Contact Mr. Bruce Bartley at 734-769-5148 or bartley@nsf.org for a copy of this document.
This attachment is a pdf file.
-------
NOVEMBER 2013
Attachment 2
Model UVL-200-4 Sensor and Lamp Information
Contact Mr. Bruce Bartley at 734-769-5148 or bartley@nsf.org for a copy of this document.
This attachment is a pdf file.
-------
NOVEMBER 2013
Attachment 3
Standard 55 Annex A - Collimated Beam Apparatus
Contact Mr. Bruce Bartley at 734-769-5148 or bartley@nsf.org for a copy of this document.
This attachment is a pdf file.
-------
NOVEMBER 2013
Attachment 4
UVT Scans for Feed Water
High and Low UVT
(with and without LSA)
Contact Mr. Bruce Bartley at 734-769-5148 or bartley@nsf.org for a copy of this document.
This attachment is a pdf file.
------- |