November, 2013
NSF13/39/EPADWCTR
EPA/600/R-13/248
Environmental Technology
Verification Report
Reduction of Microbial Contaminants in
Drinking Water by Ultraviolet Technology
ETS UV Technology
ETSUV Model ECP-113-5
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 ECP-113-5
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
List of Tables iv
List of Figures v
Abbreviations and Acronyms vi
Chapter 1 1
Introduction 1
1.1 ETV Program Purpose and Operation 1
1.2 Purpose of Verification 1
1.3 Verification Test Site 3
1.4 Testing Participants and Responsibilities 3
Chapter 2 5
Equipment Description 5
2.1 General Information ETS UV Technology 5
2.2 ETS ModelECP-113-5 UV System Description 5
2.3 ETS UV Model ECP-113-5 Specifications and Information 8
Chapters 10
Methods and Procedures 10
3.1 Introduction 10
3.2 UV Sensors Assessment 11
3.3 Headloss Determination 12
3.4 Power Consumption Evaluation 12
3.5 Feed Water Source and Test Rig Setup 12
3.6 Installation of Reactor and Lamp Burn-in 14
3.7 Collimated Beam Bench Scale Testing 15
3.8 Full Scale Testing to Validate UV dose 19
3.9 Analytical Methods 23
3.10 Full Scale Test Controls 26
3.11 Power Measurements 26
3.12 Flow Rate 26
3.13 Evaluation, Documentation and Installation of Reactor 26
Chapter 4 28
Results and Discussion 28
4.1 Introduction 28
4.2 Sensor Assessment 28
4.3 Collimated Beam Dose Response Data 30
4.4 Development of Dose Response 32
4.5 MS and Operational Flow Test Data 50
4.6 Set Line for a Minimum RED of 40 mJ/cm2 56
4.7 Deriving the Validation Factor and Log Credit for Cryptosporidium 57
4.8 Validated Dose (REDVai) for MS2 as the Target Organism 66
iii
-------
November 2013
4.9 Water Quality Data 68
4.10 Headless 72
4.11 Power Measurement 72
Chapters 73
Quality Assurance/Quality Control 73
5.1 Introduction 73
5.2 Test Procedure QA/QC 73
5.3 Sample Handling 73
5.4 Chemistry Laboratory QA/QC 73
5.5 Microbiology Laboratory QA/QC 73
5.6 Engineering Lab - Test Rig QA/QC 75
5.7 Documentation 76
5.8 Data Review 76
5.9 Data Quality Indicators 78
Chapter 6 80
References 80
Appendices
Attachment 1 Model ECP-113-5 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 8
Table 2-2. Medium Pressure Lamp Information 8
Table 2-3. UVLamp Sleeve Information 8
Table 2-4. UV Sensor Information 9
Table 3-1. Test Conditions for Validation 21
Table 3-2. Analytical Methods for Laboratory Analyses 24
Table 4-1. Sensor Assessment Data First Set of Test Runs (July 2012) 29
Table 4-2. Sensor Assessment Data Second Set of Test Runs (September 2012) 29
Table 4-3. UV Dose Response Data from Collimated Beam Tests at 79% (July 2012) 33
Table 4-4. UV Dose Response Data from Collimated Beam Tests at 95% (July 2012) 35
Table 4-5. UV Dose Response Data from Collimated Beam Tests at 79% (September 2012)....37
Table 4-6. UV Dose Response Data from Collimated Beam Tests at 97% (September 2012)....39
Table 4-7. ETS UV Model ECP-113-5 MS2 Operational Data 50
Table 4-8. ETS UV Model ECP-113-5 MS2 Concentration Results 51
Table 4-9. ETS UV Model ECP-113-5 MS2 Log Concentration for Influent and Effluent
Samples 52
Table 4-10. ETS \3VModelECP-113-5 MS2 Log Inactivation Results 53
iv
-------
November 2013
Table 4-11. ETS UVModelECP-113-5 MS2 Observed RED Results 55
Table 4-12. RED Bias Factor for Each Set Point for Cryptosporidium 58
Table 4-13. Uncertainty of the Validation (Uvai) and BRED Values for Cryptosporidium 62
Table 4-14. Validation Factors and Validated Dose (REDVai) for Cryptosporidium 63
Table 4-15. Validation Factors and Validated Dose (REDVai) based on MS2 67
Table 4-16. Temperature and pH Results 69
Table 4-17. Total Chlorine, Free Chlorine, and Turbidity Results 69
Table 4-18. Iron and Manganese Results 70
Table 4-19. HPC, Total Coliform, and E. co//Results 71
Table 4-20. Headloss Data 72
Table 4-21. Power Measurement Results 72
Table 5-1. Trip Blank Results 75
Table 5-2. MS2 Stability Test Results 75
Table 5-3. Flow Meter Calibration Results 76
Table 5-4. Reactor Control and Reactor Blank MS2 Results 77
Table 5-5. Completeness Requirements 79
List of Figures
Figure 2-1. ETS UV Model ECP-113-5 6
Figure 2-2. ETS UV Model ECP-113-5 configuration drawing 7
Figure 3-1. Schematic of NSF test rig 13
Figure 3-2. Photograph of the Model ECP-113-5 Test Setup 14
Figure 4-1. Collimated beam dose versus log NUVT 79% (July 2012) 41
Figure 4-2. Collimated beam dose versus log N UVT 95% (July 2012) 42
Figure 4-3. Collimated beam dose versus log N UVT 79% (September 2012) 43
Figure 4-4. Collimated beam dose versus log N UVT 97% (September 2012) 44
Figure 4-5. Dose response - log I versus dose UVT 79% (July 2012) 45
Figure 4-6. Dose response - log I versus dose UVT 95% (July 2012) 46
Figure 4-7. Dose response - log I versus dose UVT 79% (September 2012) 47
Figure 4-8. Dose response -log I versus dose UVT 97% (September 2012) 48
Figure 4-9. Set line at 40 mJ/cm2 RED for ETS UV Model ECP-113-5 56
Figure 4-10. Set line for Minimum 3.0 log Cryptosporidium Inactivation for ETS UV Model
ECP-113-5 64
-------
November 2013
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
CPU Colony Forming Units
cm Centimeter
DWS Drinking Water Systems
DVGW Deutscher Verein des Gas- und Wasserfaches e.V. - Technisch -
wissenschaftlicher Verein -German Technical and Scientific Association
for Gas and Water
EPA U. S. Environmental Protection Agency
ETS Engineered Treatment Systems
ETS UV ETS UV Technology - joint venture of ETS and atg
ETV Environmental Technology Verification
°F Degrees Fahrenheit
gpm gallons per minute
in inch(es)
h hours
HPC Heterotrophic Plate Count
L Liter
Ibs pounds
LEVIS Laboratory Information Management System
log I log base 10 Inactivation
LSA Sodium Lignin Sulfonic Acid
LT2ESWTR Long Term 2 Enhanced Surface Water Treatment Rule
m meter
min minute
ml milli-joules
mg Milligram
mL Milliliter
MS2 MS2 coliphage ATCC 15597 Bl
NaOH Sodium Hydroxide
ND Non-Detect
NIST National Institute of Standards and Technology
nm Nanometer
NRMRL National Risk Management Research Laboratory
NSF NSF International (formerly known as National Sanitation Foundation)
NTU Nephelometric Turbidity Unit
ONORM Osterreichisches Normungsinstitut Austria Standard
ORD Office of Research and Development
pfu Plaque Forming Units
Protocol Generic Protocol
vi
-------
November 2013
psig Pounds per Square Inch, gauge
QA Quality Assurance
QC Quality Control
QA/QC Quality Assurance/Quality Control
QAPP Quality Assurance Project Plan
QMP Quality Management Plan
RED Reduction Equivalent Dose
Measured Reduction Equivalent Dose - from test runs
Validated Reduction Equivalent Dose - based on selected pathogen and
uncertainty
RPD Relative Percent Deviation
SM Standard Methods for the Examination of Water and Wastewater
SOP Standard Operating Procedure
SPt Set Point Condition
TQAP Test / Quality Assurance Plan
IDS Total Dissolved Solids
ISA Tryptic Soy Agar
TSB Tryptic Soy Broth
UVT ultraviolet transmittance
ug microgram
jam microns
UVDGM Ultraviolet Disinfection Guidance Manual - 2006
USEPA U. S. Environmental Protection Agency
UDR uncertainty of collimated beam data
USP uncertainty of set point
Us uncertainty of sensor
UVAL 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 ECP-113-5 Water Purification System
(Model ECP-113-5) 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
1
-------
November 2013
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
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 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
(REDvai) 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 Technology (ETS UV) selected flow rates of 50, 75, and 100 gpm as the target flow
rates based on their system design for Model ECP-113-5.
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.
ETS UV also requested an additional set point be tested, at a higher flow rate of 175 gpm. The
purpose of this additional set point was to demonstrate a minimum 3-log inactivation of
Cryptosporidium at the higher flow rate. The goal was to use the additional set point and,
combined with the set points at 50, 75 and 100 gpm, to develop a set line for flow rate and
irradiance conditions that could achieve a minimum 3-log inactivation of Cryptosporidium.
This verification test did not evaluate cleaning of the lamps or quartz sleeves, nor any other
maintenance and operational issues. The automated wiper system was operated before and
during the test in accordance with the operating manual.
-------
November 2013
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 NSF certification of 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.
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.
-------
November 2013
1.4.3 ETS UV Technology
ETS UV Technology 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 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, waste
water and industrial UV treatment applications.
The atg UV is based in the North West of England, serving an international customer base. Since
being founded in 1981 as Willand UV System, atg indicates 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 ECP-113-5 UV System Description
The ETS UV Water Purification System validated in this test is Model ECP-113-5. This unit is
rated by ETS UV to handle 260 gpm for 3 log reduction of Cryptosporidium and 180 gpm to
deliver 40 mJ/cm2 REDmeas based on MS2. The system uses one (1) medium pressure mercury
amalgam lamp and one intensity sensor mounted in a stainless steel flow chamber. Figure 2-1
presents a photograph of the system and a system configuration drawing is shown in Figure 2-2.
Additional specifications for the unit are presented in Section 2.3. The operating manual and
technical information is provided in Attachment 1. The operating manual includes schematics
and tables with parts and dimensions for the reactor, the sensors, the lamps and the quartz
sleeves. All specifications and equipment information was provided by ETS UV in advance of
the actual shipment of the unit to NSF. ETS UV 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. This information is presented in Attachment 2.
NSF performed a normal technical review of the sensor specifications, UV lamp and quartz
sleeve specification, and general review of the reactor chamber and overall system as required by
theGP-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 te
this report for reference.
-------
November 2013
Figure 2-1. ETS \]\ Model ECP-113-5.
-------
November 2013
\7
CHAMBER MUST BE MOUNTED AS SHOW ON THIS DRAWING
1
1
I
i »
1
1
1
f •um-iB&flru
/ \
•
K 1
It
1
I
t
"1
I
n*
t
L_
i_
JL
J L
' I 5
Til
Fu <« |l| UC ZIJISI
Figure 2-2. ETS UV Model ECP-113-5 configuration drawing.
-------
November 2013
2.3 ETS UV Model ECP-113-5 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
Maximum cleaning temperature
Inlet pipe size
ETS UV Technology
ECP-113-5
Cross Flow Medium Pressure UV Disinfection
System
2008 and onwards
260 gpm
421bs
236 cubic inches
2 phase 220 VAC, 60Hz; 20 amp single pole,
earth ground.
1300 W
60psi
40 to 114 degrees °F
180 degrees °F (unit turned off)
Sin
Table 2.2. Medium 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
UV output
Operating lamp power
Lamp current and voltage
Arc length
Medium-pressure
W1501200
1
See Lamp spectral graph in Attachment 1..
4000 hrs
SPECTRA R4.02 SP-A-220 #A18587-X
Magnetic Choke with Igniter
90 W/cm
35 W
1300 W
9.0 A; 160 V
140mm
Table 2.3. UV Lamp Sleeve Information.
Type or model
Quartz material
Pressure resistance
GE 214 Clear Fused Quartz
Clear Fused Quartz
7000 psi
-------
November 2013
Table 2.4. UV Sensor Information.
Manufacturer / model
Measuring field angle
Number of sensors per reactor and placement
Signal output range
Measuring range Output signal
UV-Technik SUV20.1 A2Y2C
160 degree
1
4 -20mA
0 - 100 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
The tests followed the procedures described in the Test/Quality Assurance Plan for The ETS UV
Ultraviolet (UV) Reactor, Medium Pressure Lamps, June 2010 (TQAP). The TQAP was
adapted from the GP-2010 and was updated in 2011. The ETV Generic Protocol was 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 (RED)". The RED value is adjusted for uncertainties
and biases to produce the validated dose of the reactor for the specific operating conditions
tested.
The 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 measured RED
(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.
During testing of the unit, an additional single set point test at a higher flow rate and intensity
was performed which defined an operating condition that could achieve a minimum of 3-log
inactivation of Cryptosporidium. This higher flow rate point was then used with the other set
points to develop of a set line that demonstrated a 3-log inactivation of Cryptosporidium.
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 REDmeas.
6. Adjust the REDmeas 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 50, 75, and 100 gpm as the target flow rates based on their system design for Model
10
-------
November 2013
ECP 113-5. The additional flow rate selected for testing based on ETS UV's request was 175
gpm.
3.2 UV Sensors Assessment
The Model ECP-113-5 duty sensor was evaluated according to the UV sensor requirements in
the 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
M5873-2 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 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.
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]
11
-------
November 2013
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 50 gpm to 200 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
50, 100, 150 and 200 gpm. These data are reported in Section 4.11.
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 4.10.
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
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.
12
-------
November 2013
Fluid Flow
[ Dosing Pump
No:e: Ali plumbing is schedule BO PVC
Figure 3-1 Schematic of NSF test rig(<
13
-------
November 2013
Figure 3-2 Photograph of the Model ECP-113-5 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 to the GP-2011.
14
-------
November 2013
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 ECP-113-5. 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
UV light, UV absorbance of the water, and exposure time. The measured concentration of
microorganisms before and after exposure provides the "response," or log inactivation of the
microorganisms from exposure to UV light. Regression analysis of measured log inactivation for
a range of UV doses produces the dose-response curve.
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. This validation test included three days of testing. The lowered UVT test runs were
performed on the first day (July 18, 2012). The intensity readings at each UVT (79%. 89%, 93%)
were recorded during test runs with full lamp power. Collimated beam tests were run on the
minimum UVT water (79%) with duplicate runs being performed. On the second day (July 19,
2012) using high UVT water (95%), the power was reduced to achieve the same intensity as
measured for each of the lowered UVT waters on day one. Collimated beam tests were run on
day two on the high UVT water (95%) with duplicate runs being performed. Additional testing
was performed on September 11, 2012 (test day three) for the lowest flow rate (50 gpm) for both
the lowered UVT water (79%) and for the lowered power tests. In addition, one medium flow
rate test (75 gpm) at the lowered power setting required a retest as part of the September test
runs. This third day of testing included both lowered UVT water (79%) and the use of high UVT
water (97%) for the lowered power runs. Collimated beam tests were performed in duplicate on
both the 79% and 97% UVT water on the third day of testing. Therefore, for this validation test,
there are four sets of duplicate collimated beam test data, two for the lowest UVT water (79%)
and two for the high UVT water (water not adjusted with LSA).
15
-------
November 2013
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 ml/cm2. The samples are clustered close to the
40mJ/cm2 target dose with two doses above and below the target of 40 ml/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.
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, which is attached to the TQAP for reference. 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.
16
-------
November 2013
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
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) in
units of plaque forming units (pfu)/mL, and
• Concentration of microorganisms in the petri dish after UV exposure (N) in units
of pfu/mL.
17
-------
November 2013
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 (ml/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 three days of testing and there were four
sets of data (low UVT - test day 1; high UVT test day 2; low UVT test day 3; high UVT
test day 3).
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:
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 No, 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
RED for each full scale test sample.
The full set of collimated beam data and all calculations and regression analyses are presented in
Chapter 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.. 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 for the first two days of testing, a single curve corresponding to one day's
worth of full scale reactor testing was used to calculate RED values for that day. The higher
UVT dose response curve was used for the high UVT water (day two) with reduced power and
the lower UVT dose response curve was used for day one when the UVT of the test water was
lowered with LSA. On the third day of testing the low UVT collimated beam results were used
18
-------
November 2013
for the low UVT test runs and the high UVT collimated beam data were used for the high UVT
test runs.
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 (UDR) 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
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 Sections 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 ECP-113-5 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. Basic descriptions of the equipment were presented previously 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 (50, 75, 100 gpm) and UV target intensity levels (80, 90,
105 W/m2) based on the results of screening test performed at NSF prior to the validation tests.
The intensity targets were based on the expected intensity at UVT's of 79%, 89%, and 94%.
Data were also developed during an additional set point (175 gpm, intensity of 105 W/m2) for
validating a dose that would achieve a minimum of 3-log inactivation ofCryptosporidium.
Each set point represents a given flow rate with testing under two conditions, (1) lowered UVT-
max power and (2) high UVT-reduced 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 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 rates - intensity points (50 gpm - 80 W/m2; 75 gpm - 90 W/m2; 100 gpm - 105
19
-------
November 2013
W/m2) were tested for the set line. All conditions were performed in duplicate. The intensity
targets were based on expected intensity at UVT's of 79%, 89%, and 94%.
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 ml/cm2,
additional calculations (adjusting REDmeas for uncertainty and RED bias) were performed to
demonstrate the log inactivation of Cryptosporidium.
An additional fourth set point test at a higher flow rate of 175 gpm, UVT target of 94% and
intensity target of 105 W/m2 was performed to provide additional data for demonstrating
Cryptosporidium log inactivation. These tests were performed with both lowered UVT (with full
power) and reduced power (with high UVT) and were performed in duplicate.
A reactor control test (MS2 injection with the lamp off) was run at the low flow rate (50 gpm)
and with high UVT water, which demonstrated that there was no reduction of MS2 with the
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 it's 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 its duplicate.
20
-------
November 2013
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.
Table 3-1. Test Conditions for Validation with MS2 Phase.
Validation Test
Condition 1
Condition 2
Condition 3
(reactor control)
Condition 4
(reactor blank)
Target Flow Rate
50 gpm
75gpm
100 gpm
175 gpm
50 gpm
75 gpm
100 gpm
175 gpm
50 gpm
50 gpm
UV Transmittance
Target UVT (%)
79%
90%
94%
94%
>95%
>95%
>95%
>95%
>95%
Daily Source water
- ether high or low
UVT
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 microorganism"
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
21
-------
November 2013
per centimeter squared (mJ/crm), the prediction intervals of the data shown in Appendix A of the
UVDGM-2006 are represented by the following equations"
Upper Bound: log / = -1.4 X 10-4 X UV Dose2 + 7.6 X 10-2 X UV Dose
Lower Bound: log /= -9.6 X 10-s X UVDose2+ 4.5 X 10-2 X UV Dose
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 ,
• 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 minutes of continuous
flow at steady conditions. Each set of influent and effluent grab samples were
22
-------
November 2013
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
microbiology laboratory^-aftd-,.
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;
2. 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;
3. A sample for UVT was collected and measured by a UV spectrophotometer for each
influent sample and at least one effluent sample;
4. 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, and
5. The electrical power consumed by system was recorded.
Chapter 4 describes the calculations and presents the data for determining the REDmeas and the
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.
23
-------
November 2013
Table 3-2. Analytical Methods for Laboratory Analyses.
Parameter
Temperature
pH
E. coli 1 Total Conform
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^
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
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 'A254
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
for direct comparison with the on-line analyzer to ensure the data were comparable.
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
24
-------
November 2013
spectrophotometer internal Quality Assurance/Quality Control (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:
1. Verify that the spectrophotometer reads the wavelength to within the accuracy of a
holmium oxide standard (typically ± 0.2 nm at a 95-percent confidence level),
2. Verify that the spectrophotometer reads A254 within the accuracy of a dichromate
standard (e.g., 0.281 ± 0.005 at 257 nm with a 20 mg/L standard), and
3. Verify that the water used to zero the instrument has an A254 value that is within 0.002
cm"1 of a certified zero absorbance solution.
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.
• 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.
25
-------
November 2013
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 logio.
• Reactor blanks - Influent and effluent water samples were collected with no addition of
challenge 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 logio.
• Trip controls - Trip controls were collected to monitor any change in challenge MS2
during transport to the laboratory (in the same building).
• Method blanks - A sample bottle of sterilized reagent grade water was analyzed using the
challenge microorganism assay procedure. The concentration of challenge 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 the challenge
microorganism concentration and its UV dose-response over the time period from sample
collection to completion of challenge microorganism assay. The challenge 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 was were started
within a couple of hours of collection.
3.11 Power Measurements
The voltmeter and ammeter meter used to measure UV equipment had traceable evidence of
calibration. The meters had a tag showing that it was calibrated. Calibrations are performed at
least yearly and all power equipment was calibrated 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 accuracy was within 0.6 to 2.7% 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 ECP-113-5 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
26
-------
November 2013
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
• 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
27
-------
November 2013
Chapter 4
Results and Discussion
4.1 Introduction
ETS UV specified target flow rates of 50, 75, and 100 gpm. The intensity initial targets were 80,
90, and 105 W/m2 based on the expected intensities at UVTs of 79%, 90%, and 94%. These
points were projected to deliver a RED of >40 ml/cm2. An additional set point at 175 gpm with
intensity of 105 W/m2 was tested to demonstrate of 3 log Cryptosporidium inactivation.
The main validation tests were run on two days, July 18 and July 19, 2012. A retest of the lower
flow rate (50 gpm) and one medium flow rate (75 gpm) was performed on September 11, 2012.
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. On the third day of testing, both low UVT water (<79%)
with full power at a flow rate of 50 gpm and high UVT water with reduced power for flow rates
of 50 and 75 gpm were used. The test conditions and detail on the test rig setup, sampling
procedures, and unit operation have boon 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 ECP-113-5 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 medium 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 for both the July and September 2012 test runs, using the
procedure described in the GP-2011 and the UVDGM-2006. Tables 4-1 and 4-2 present the
28
-------
November 2013
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
range of 0.0 to 2.1% at 100% power and 0.0 to 3.5% at 64% power.. The percent differences
were calculated by taking the difference between a given sensor intensity reading and the
average of the two reference sensor readings.
% difference = The absolute value of [(I Ref/1 AvgRef) - 1] XI00
where:
I Ref = Intensity measured by a reference UV sensor (Ref 1 or Ref 2),
I Avg Ref = Average UV intensity measured by the two reference UV sensors in the
same UV sensor port with the same UV lamp at the same UV lamp power.
The power could not be reduced below 64% as the lamp would lose its arc and shut down below
at less than 64% power level (4.1 - 4.2 amps) when the input voltage was 207 V. During the
retest runs at the lowest power setting, power was reduced to approximately 45-50% and the
lamps did not lose its arc when the input voltage was 242 V.
Table 4-1. Sensor Assessment Data First Set of Test Runs (July 2012)
Sensor
Reference # 1
V6154
Reference #2
V6156
Average of
Reference Sensor
Duty Sensor
V6161
Deviation of Duty
Sensor from
Reference
Intensity at
100% power
Before testing
(W/m2)
46.51
46.51
46.51
44.16
5.1%
UVT = 78%
Intensity at
100% power
After testing
(W/m2)
138.36
136.01
137.03
131.38
4.1%
UVT = 97%
Intensity at
64% Power
Before testing
(W/m2)
22.96
22.96
22.96
21.78
5.1%
UVT = 78 %
Intensity at
64% Power
After testing
(W/m2)
67.71
65.35
66.53
64.18
3.5%
UVT = 97%
Table 4-2. Sensor Assessment Data Second Set of Test Runs (September 2012)
Sensor
Reference # 1
V6154
Reference #2
V6156
Average of
Intensity at
100% power
Before testing
(W/m2)
252.59
241.99
247.29
Intensity at
100% power
After testing
(W/m2)
254.94
252.59
253.77
Intensity at
64% Power
Before testing
(W/m2)
137.19
133.65
135.42
Intensity at
64% Power
After testing
(W/m2)
138.36
137.19
137.78
29
-------
November 2013
Reference Sensor
Duty Sensor
V6161
Deviation of Duty
Sensor from
Reference
238.46
3.6%
UVT = 97%
237.28
6.5%
UVT = 97%
128.94
4.8%
UVT = 97 %
127.77
7.3%
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. The steady sensor readings from the start through the end of
the testing at the various UVT-power combinations indicated 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
performed in duplicate at the minimum and maximum UVT test conditions. This validation test
included three days of testing. The lowered UVT test runs were performed on the first day. The
intensity readings at each UVT (78%, 89%, 93%) were recorded during each 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 (95%), the power was reduced
to achieve the same intensity as measured for each of the lowered UVT waters on day one.
Collimated beam tests were run on day two on the high UVT water (95%) with duplicate runs
being performed. Additional testing was required for the lowest flow rate (50 gpm) for both the
lowered UVT water (79%) and for the lowered power tests. In addition one medium flow rate
test (75 gpm) at the lowered power setting required a retest. This third day of testing included
both lowered UVT water (79%) and the use of high UVT water (97%) for lowered power runs.
Collimated beam tests were performed in duplicate on both the 79% and 97% UVT water on the
third day of testing. Therefore, for this validation test, there are four sets of duplicate collimated
beam test data, two for the lowest UVT water (79%) and two for the highest UVT water (water
not adjusted with LSA).
UV doses covered the range of the targeted RED dose, which in this case was >40 ml/cm2. UV
target doses were set at 0, 20, 30, 40, 60 and 80 ml/cm2. As discussed in the RED results
presented later, the actual RED for two test runs exceeded the maximum collimated beam dose
of 80 mJ/cm2. RED cannot be quantitatively determined if the measured RED exceeds the top
range of the collimated beam data. These data are presented as calculated, but any RED 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
30
-------
November 2013
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.
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 4-3 through 4-6. These data were calculated
as the average of the three individual results obtained at each dose level.
31
-------
November 2013
4.4 Development of Dose Response
The development of the UV dose response curves for use with flow tests to establish the RED 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 through 4-4 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 four 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 4-3 through 4-6 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 4-5 through
4-8 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 isare presented in Section 4.5.
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
(P <0.05) for all of the dose response curves. The statistics are shown in Tables 4-3 through 4-6.
A Grubbs' test was also 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. The Grubbs' statistics are shown in Tables 4-3 through 4-6.
A summary of the statistics for uncertainty for the collimated beam dose response data is
presented at the end of Tables 4-2 through 4-6. The uncertainty (UDR) of the collimated beam
results was slightly higher than 30% at 1 log inactivation for the September retest data set for the
high UVT water (33.46%). The UDR for the high UVT water for the first set of data (July 2012)
was 20.74%. The uncertainty for the sets of low UVT water (July and September) was 27.48%
32
-------
November 2013
and 26.99%, respectively. At 2-log inactivation (dose of approximately 40 ml/cm2 RED) the UDR
was between 9.33% and 14.92%.
Figures 4-5 through 4-8 show the results of the UDR calculations plotted on the dose response
curve. Also shown in Figures 4-5 through 4-8 are the QC limits for MS2 taken from the
UVDGM-2006. The results show that the MS2 dose response curves are within the boundaries
established for MS2.
33
-------
November 2013
Table 4-3. UV Dose - Response Data from Collimated Beam Tests at 79% UVT (July 2012)
UVT
(%)
78.9
Rep
1
2
Target
UV Dose
(mJ/cm2)
0
20
30
40
60
80
0
20
30
40
60
80
DRC
A: 15.147
B: 2.5292
Actual
UV Dose
0.00
20.76
31.20
41.49
62.23
82.80
0.00
20.74
31.26
41.37
62.24
82.62
UV
Dose2
0
431
973
1721
3873
6856
0
430
977
1711
3874
6826
Log N0
5.25
Avg
pfu/ml
188,000
15,900
4,900
1,760
257
84
153,000
15,800
3,970
1,150
258
57
Avg
Log(pfu)
5.27
4.20
3.69
3.25
2.41
1.93
5.18
4.20
3.60
3.06
2.41
1.76
Log I
-0.02
1.05
1.56
2.01
2.84
3.33
0.07
1.06
1.66
2.19
2.84
3.50
Log I2
0.000
1.110
2.448
4.038
8.094
11.083
0.005
1.116
2.743
4.814
8.084
12.243
PRED
Dose
-0.29
18.76
29.89
40.65
63.56
78.46
1.08
18.82
32.02
45.41
63.51
83.96
Avg:
SD:
P:
t (95%):
Residual
(mJ/cm2)
0.3
2.0
1.3
0.8
-1.3
4.3
-1.1
1.9
-0.8
-4.0
-1.3
-1.3
0.07
2.18
12
0.05
2.228
G
0.1
0.9
0.6
0.4
0.6
2.0
0.5
0.8
0.4
1.9
0.6
0.7
Outlier?
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
DRC - dose response coefficients
Grubbs' Test for
Outliers
P:
t (90%):
Grubbs'
Statistic
(GCRIT):
0.10
3.691
2.412
34
-------
November 2013
Table 4-3. (continued)
Uncertainty of Dose-Res
Logl
0.001
0.25
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
3.50
Dose
(mJ/cm2)
0.0
3.9
8.2
17.7
28.4
40.4
53.7
68.2
84.0
101.1
84.0
t
2.23
2.23
2.23
2.23
2.23
2.23
2.23
2.23
2.23
2.23
SD
2.18
2.18
2.18
2.18
2.18
2.18
2.18
2.18
2.18
2.18
ponse (UDR)
UDR (%)
123.12
59.19
27.48
17.09
12.02
9.05
7.12
5.78
4.81
5.78
^
(mJ/cnr/Log I)
15.15
15.78
16.41
17.68
18.94
20.21
21.47
22.73
24.00
25.26
24.00
t - student t test factor SD - standard deviation
Dose Response Curve Statistics
Regression Statistics
Multiple R
R Square
Adjusted R
Square
Standard Error
Observations
0.999054
0.998109
0.89792
2.287586
12
ANOVA
Regression
Residual
Total
Df
2
10
12
ss
27620.61
52.33049
27672.94
MS
13810.31
5.233049
F
2639.055
Significance
F
3.46E-13
Intercept
X Variable 1
X Variable 2
Coefficients
0
15.14714
2.529192
Standard
Error
1 .250784
0.435994
tStat
12.11012
5.80098
P-value
2.68E-07
0.000173
Lower
95%
12.36022
1.557737
Upper
95%
17.93406
3.500647
Lower
95.0%
12.36022
1 .557737
Upper
95.0%
17.93406
3.500647
35
-------
November 2013
Table 4-4. UV Dose - Response Data from Collimated Beam Tests at 95% UVT (July 2012)
UVT
(%)
95.3
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
20.67
30.94
41.45
62.06
82.46
0.00
20.75
31.11
41.49
62.26
82.57
DRC
A: 14.898
B: 1.8771
UV
Dose2
0
427
957
1718
3851
6800
0
429
968
1721
3876
6818
Avg
pfu/ml
313000
23700
6000
1990
277
59
251000
22200
6530
1500
225
57.3
Log N0
5.48
Avg
Log(pfu)
5.50
4.37
3.78
3.30
2.44
1.77
5.40
4.35
3.81
3.18
2.35
1.76
Log I
-0.02
1.10
1.70
2.18
3.03
3.71
0.08
1.13
1.66
2.30
3.12
3.72
Log I2
0.000
1.215
2.886
4.745
9.209
13.752
0.006
1.279
2.763
5.294
9.765
13.830
PRED
Dose
-0.27
18.70
30.73
41.36
62.50
81.07
1.16
19.24
29.95
44.22
64.89
81.37
Avg:
SD:
t (95%):
Residual
(mJ/cm2)
0.3
2.0
0.2
0.1
-0.4
1.4
-1.2
1.5
1.2
-2.7
-2.6
1.2
0.07
1.56
12
0.05
2.228
G
0.1
1.2
0.1
0.0
0.3
0.8
0.8
0.9
0.7
1.8
1.7
0.7
Outlier?
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
DRC - dose response coefficients
Grubbs' Test for
Outliers
P:
t (90%):
Grubbs'
Statistic
(GCRIT):
0.10
3.691
2.412
36
-------
November 2013
Table 4-4. (continued)
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
3.72
Dose
(mJ/cm2)
0.0
3.8
7.9
16.8
26.6
37.3
49.0
61.6
75.1
89.6
81.4
t
2.23
2.23
2.23
2.23
2.23
2.23
2.23
2.23
2.23
2.23
SD
1.56
1.56
1.56
1.56
1.56
1.56
1.56
1.56
1.56
1.56
UDR (%)
90.57
43.95
20.74
13.10
9.33
7.10
5.65
4.63
3.88
4.28
°2L
(mJ/cm2/l_og I)
14.90
15.37
15.84
16.78
17.71
18.65
19.59
20.53
21.47
22.41
21.88
t - student t test factor SD - standard deviation
Dose Response Curve Statistics
Regression Statistics
Multiple R
R Square
Adjusted R
Square
Standard Error
Observations
0.999512
0.999025
0.898927
1 .639565
12
ANOVA
Regression
Residual
Total
Df
2
10
12
ss
27540.78
26.88174
27567.66
MS
13770.39
2.688174
F
5122.581
Significance
F
1.76E-14
Intercept
X Variable 1
X Variable 2
Coefficients
0
14.89799
1.877055
Standard
Error
0.822193
0.263522
tStat
18.11982
7.122945
P-value
5.62E-09
3.21 E-05
Lower
95%
13.06603
1.289891
Upper
95%
16.72995
2.464219
Lower
95.0%
13.06603
1.289891
Upper
95.0%
16.72995
2.464219
37
-------
November 2013
Table 4-5. UV Dose - Response Data from Collimated Beam Tests at 79% UVT (September 2012)
UVT
(%)
78.8
Rep
1
2
Target
UV Dose
(mJ/cm2)
0
20
30
40
60
80
0
20
30
40
60
80
DRC
A: 14.413
B: 2.6421
Actual
UV Dose
0.00
20.66
30.74
40.77
61.37
81.42
0.00
20.74
31.05
41.34
61.79
81.97
UV
Dose2
0
427
945
1662
3766
6629
0
430
964
1709
3818
6719
Log N0
5.95
Avg
pfu/ml
1,000,000
51,000
20,500
11,600
1,120
350
930,000
64,000
14,500
9,200
1,230
310
Avg
Log(pfu)
6.00
4.71
4.31
4.06
3.05
2.54
5.97
4.81
4.16
3.96
3.09
2.49
Log I
-0.05
1.25
1.64
1.89
2.91
3.41
-0.01
1.15
1.79
1.99
2.86
3.46
Log I2
0.002
1.555
2.699
3.573
8.441
1 1 .632
0.000
1.319
3.216
3.963
8.206
11.994
PRED
Dose
-0.65
22.08
30.81
36.68
64.18
79.89
-0.20
20.04
34.34
39.16
62.97
81.60
Avg:
SD:
P:
t (95%):
Residual
(mJ/cm2)
0.6
-1.4
-0.1
4.1
-2.8
1.5
0.2
0.7
-3.3
2.2
-1.2
0.4
0.08
2.07
12
0.05
2.228
G
0.3
0.7
0.1
1.9
1.4
0.7
0.1
0.3
1.6
1.0
0.6
0.1
Outlier?
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
DRC - dose response coefficients
Grubbs' Test for
Outliers
P:
t (90%):
Grubbs'
Statistic
(GCRIT):
0.10
3.691
2.412
38
-------
November 2013
Table 4-5. (continued)
Uncertainty of Dose-Res
Logl
0.001
0.25
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
3.46
Dose
(mJ/cm2)
0.0
3.8
7.9
17.1
27.6
39.4
52.5
67.0
82.8
99.9
81.6
t
2.23
2.23
2.23
2.23
2.23
2.23
2.23
2.23
2.23
2.23
SD
2.07
2.07
2.07
2.07
2.07
2.07
2.07
2.07
2.07
2.07
ponse (DDR)
UDR (%)
122.15
58.51
26.99
16.70
11.68
8.76
6.87
5.56
4.61
5.64
°2L
(mJ/cm /Log I)
14.42
15.07
15.73
17.06
18.38
19.70
21.02
22.34
23.66
24.98
23.56
t - student t test factor SD - standard deviation
Dose Response Curve Statistics
Regression Statistics
Multiple R
R Square
Adjusted R
Square
Standard Error
Observations
0.999131
0.998263
0.898089
2.168467
12
ANOVA
Regression
Residual
Total
Df
2
10
12
ss
27022.78
47.02248
27069.8
MS
13511.39
4.702248
F
2873.389
Significance F
2.36E-13
Intercept
X Variable 1
X Variable 2
Coefficients
0
14.41301
2.642101
Standard
Error
1.147736
0.399141
tStat
12.55778
6.619463
P-value
1.9E-07
5.93E-05
Lower
95%
11.8557
1.752759
Upper
95%
16.97033
3.531443
Lower
95.0%
11.8557
1.752759
Upper
95.0%
16.97033
3.531443
39
-------
November 2013
Table 4-6. UV Dose - Response Data from Collimated Beam Tests at 97% UVT (September 2012)
UVT
(%)
97.6
Rep
1
2
DRC
A 15.834
B 2.1804
Target
UV Dose
(mJ/cm2)
0
20
30
40
60
80
0
20
30
40
60
80
Actual
UV
Dose
0.00
20.77
31.05
41.04
61.43
81.79
0.00
20.85
31.27
41.52
62.32
83.18
Log N0
5.88
UV
Dose2
0
431
964
1684
3774
6690
0
435
978
1724
3884
6919
Avg
pfu/ml
850,000
31,300
17,800
8,570
1,150
244
817,000
58,700
35,700
6,300
990
253
Avg
Log(pfu)
5.93
4.50
4.25
3.93
3.06
2.39
5.91
4.77
4.55
3.80
3.00
2.40
Log I
-0.04
1.39
1.63
1.95
2.82
3.50
-0.03
1.12
1.33
2.09
2.89
3.48
Log I2
0.002
1.930
2.671
3.809
7.975
12.231
0.001
1.246
1.774
4.349
8.346
12.121
PRED
Dose
-0.70
26.20
31.70
39.21
62.10
82.04
-0.43
20.39
24.96
42.50
63.94
81.55
Avg:
SD:
p:
t (95%):
Residual
(mJ/cm2)
0.7
-5.4
-0.6
1.8
-0.7
-0.3
0.4
0.5
6.3
-1.0
-1.6
1.6
0.15
2.71
12
0.05
2.228
G
0.2
2.1
0.3
0.6
0.3
0.1
0.1
0.1
2.3
0.4
0.7
0.5
Outlier?
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
DRC - dose response coefficients
Grubbs' Test for
Outliers
P:
t (90%):
Grubbs'
Statistic
(GCRIT):
0.10
3.691
2.412
40
-------
November 2013
Table 4-6. (continued)
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
3.50
Dose
(mJ/cm2)
0.0
4.1
8.5
18.0
28.7
40.4
53.2
67.1
82.1
98.2
82.0
t
2.23
2.23
2.23
2.23
2.23
2.23
2.23
2.23
2.23
2.23
SD
2.71
2.71
2.71
2.71
2.71
2.71
2.71
2.71
2.71
2.71
UDR (%)
147.21
71.24
33.46
21.04
14.92
11.33
8.98
7.34
6.14
7.35
°2L
(mJ/cm /Log I)
15.84
16.38
16.92
18.01
19.10
20.19
21.28
22.37
23.46
24.56
23.46
t - student t test factor SD - standard deviation
Dose Response Curve Statistics
Regression Statistics
Multiple R
R Square
Adjusted R
Square
Standard Error
Observations
0.998529
0.997061
0.896767
2.842009
12
ANOVA
Regression
Residual
Total
Df
2
10
12
ss
27401.4
80.77013
27482.17
MS
13700.7
8.077013
F
1696.258
Significance
F
2.52E-12
Intercept
X Variable 1
X Variable 2
Coefficients
0
15.83368
2.180374
Standard
Error
1.475888
0.50681
tStat
10.72824
4.302155
P-value
8.32E-07
0.001556
Lower
95%
12.5452
1.051131
Upper
95%
19.12217
3.309616
Lower
95.0%
12.5452
1.051131
Upper
95.0%
19.12217
3.309616
41
-------
November 2013
2 -
1 -
Log N as a function of UV Dose
\
^
\
^
v =0.0
002x2 - 0
058* + 5
.2549
R2 = 0.9964
10 20 30 40 50 60
UV Dose (mJ/cm2)
70
80
Figure 4-1 Collimated beam dose versus log N UVT 79% (July 2012)
42
-------
November 2013
4 -
D. 3H
2 -
Log N as a function of UV Dose
y = 0.0002X2 - 0.0607X + 5.4773
R2 = 0.9981
10 20 30 40 50 60
UV Dose (mJ/cm2)
70
SO
90
Figure 4-2 Collimated beam dose versus log N UVT 95% (July 2012)
43
-------
November 2013
5 -
3 -
2 -
1 -
Log N as a function of UV Dose
^
^\
*\
^
y = O.OOC
^\
»2x2 - 0.0
R2 = 0.9
^__
596x + 5.
944
^-^_
9546
^
10 20 30 40 50 60
UV Dose (mJ/cm2)
70
80 90
Figure 4-3 Collimated beam dose versus log N UVT 79% (September 2012)
44
-------
November 2013
Log N as a function of UV Dose
3 -
\.
~\
^\,
^
o^^\
o
-------
November 2013
100
Dose-Response Curve
150
120
234
Log Inactivation
CB Data UVDGM QC Limits
•Udr(%)
Dose-Response Curve
Figure 4-5 Dose response - log I versus dose UVT 79% (July 2012)
46
-------
November 2013
100
80 -
o 60 -
en
O
40 -
20 -
Dose-Response Curve
150
1.8771x2 + 14.898x ,
R2 = 0.9969 /
120
234
Log Inactivation
CB Data
UVDGMQC Limits
•Udr(%)
Dose-Response Curve
Figure 4-6 Dose response - log I versus Dose UVT 95% (July 2012)
47
-------
November 2013
100
80 -
(SI
o 60
0
v>
O 40
20 -
Dose-Response Curve
y =2.6421x2 + 14.413x /
R2 = 0.9945
150
120
- 90
60
30
234
Log Inactivation
CB Data UVDGM QC Limits
•Udr(%)
Dose-Response Curve
Figure 4-7 Dose response - log I versus dose UVT 79% (September 2012)
48
-------
November 2013
100
80 -
(N
o 60
o>
w
O 40
20 -
Dose-Response Curve
y =2.1804x2 + 15.834x
R2 = 0.9907
150
120
90
60
30
01 23456
Log Inactivation
CB Data UVDGM QC Limits Udr(%) Dose-Response Curve
Figure 4-8 Dose response - log I versus dose UVT 97% (September 2012)
49
-------
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 4-7. 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 4-8.
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 4-9. Table 4-10 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 4-5 through 4-8). 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 4-11.
All of the flow rate tests at 50, 75, and 100 gpm, with feed water at 78%, 89%, and 93% UVT or
the equivalent reduced power tests, achieved a minimum REDmeas of 40 ml/cm2. The results
from the additional flow test at 175 gpm and the minimum REDmeas, standard deviation (SDRED)
and the uncertainty of the set point (Usp) shown in Table 4-11 were used in the example validated
dose calculation for Cryptosporidium shown in Section 4.7.
The REDmeas for two of the test runs exceeded the maximum collimated beam dose of 80
mJ/cm2. These runs showed calculated RED between 90.9 and 93.2 ml/cm2. The RED 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 two RED values above 80 mJ/cm2 should be
considered as estimates only.
50
-------
November 2013
ETS UV Model ECP-113-5 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
22
23
4
5
6
7
24
25
26
15
16
17
%of
Full
Power(1)
100
100
100
100
100
100
45
45
74
74
83
83
UVT
(%)
78.3
78.4
89.3
89.3
93.4
93.4
97.9
97.9
97.9
96.8
97.2
97.2
Flow
(gpm)
50
51
75
75
101
100
50
51
76
75
100
101
Intensity
(W/m2)
82
82
84
84
104
105
81
81
89
87
105
105
High Flow Rate Test for 3-log Cryptosporidium inactivation demonstration
Lowered UVT - Full Power (SPt 4)
Lowered UVT - Full Power Duplicate (SPt 4)
Lowered Power - High UVT (SPt 4)
Lowered Power - High UVT Duplicate (SPt 4)
8
9
18
19
100
100
45
45
93.5
93.4
97.3
97.4
175
175
176
175
104
103
105
105
Table 4-7.
(1) % of full power less than 100% estimated based on measured amperage for the system, where amperage at reduced power
is divided by sensor intensity at full power.
SPt = Set Point Condition
51
-------
November 2013
Table 4-8. ETS UV Model ECP-113-5 MS2 Concentration Results
Test Condition
Lowered UVT - Full
Power (SPt1)
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
22
23
4
5
6
7
24
25
26
15
16
17
Influent (pfu/mL)
Rep1
5.14E+05
5.51E+05
7.33E+04
9.00E+04
1.48E+05
3.30E+05
8.17E+05
1.08E+06
5.77E+05
2.84E+05
1.89E+05
2.46E+05
Rep 2
4.01 E+05
5.75E+05
8.77E+04
1.07E+05
1.27E+05
2.86E+05
6.13E+05
1.43E+06
4.67E+05
2.91 E+05
1.97E+05
2.00E+05
Rep 3
4.34E+05
4.80E+05
7.67E+04
7.20E+04
1.53E+05
3.20E+05
7.10E+05
6.63E+05
4.07E+05
2.85E+05
1.98E+05
2.39E+05
Rep 4
4.06E+05
3.23E+05
5.20E+04
9.40E+04
1.43E+05
3.10E+05
7.93E+05
7.13E+05
6.13E+05
3.62E+05
1.90E+05
2.42E+05
Rep 5
3.80E+05
5.46E+05
5.37E+04
9.13E+04
2.14E+05
3.30E+05
6.73E+05
1.04E+06
4.63E+05
2.68E+05
1.99E+05
2.56E+05
Effluent (pfu/mL)
Rep1
8.47E+02
9.80E+02
2.03E+02
4.03E+02
9.50E+02
8.63E+02
1.30E+02
1.61E+02
1.22E+03
5.63E+02
3.57E+02
5.63E+02
Rep 2
1.07E+03
9.97E+02
1.30E+02
4.03E+02
6.67E+02
8.63E+02
1.51E+02
1.12E+02
1.35E+03
4.77E+02
2.90E+02
6.87E+02
Rep 3
1.11E+03
8.97E+02
1.52E+02
3.03E+02
8.63E+02
no data
1.21E+02
1.42E+02
1.08E+03
4.37E+02
4.37E+02
6.07E+02
Rep 4
8.27E+02
8.73E+02
1.32E+02
5.13E+02
6.93E+02
9.17E+02
9.70E+01
1.02E+02
9.80E+02
4.83E+02
3.53E+02
5.30E+02
Rep 5
9.57E+02
1.17E+03
1.20E+02
3.73E+02
7.73E+02
9.67E+02
1.08E+02
1.71E+02
7.50E+02
5.83E+02
4.47E+02
5.30E+02
High Flow Rate Test for 3-log Cryptosporidium inactivation demonstration
Lowered UVT - Full
Power (SPt 4)
Lowered UVT - Full
Power Dup (SPt 4)
Lowered Power - High
UVT (SPt 4)
Lowered Power - High
UVT Dup (SPt 4)
8
9
18
19
2.30E+05
2.83E+05
2.23E+05
1.31E+05
2.31 E+05
2.57E+05
1.78E+05
1.90E+05
2.85E+05
2.45E+05
2.38E+05
1.25E+05
2.97E+05
2.27E+05
2.65E+05
7.87E+04
2.77E+05
2.33E+05
2.41 E+05
1.19E+05
3.07E+03
7.90E+03
1.37E+03
3.90E+03
5.37E+03
4.80E+03
1.49E+03
3.99E+03
3.17E+03
6.90E+03
1.15E+03
2.85E+03
5.13E+03
5.03E+03
2.50E+03
3.02E+03
2.67E+03
3.57E+03
2.78E+03
3.76E+03
SPt = Set Point Condition
52
-------
November 2013
Table 4-9. ETS UV Model ECP-113-5 MS2 Log Concentration for Influent and Effluent Samples
Test Condition
Lowered UVT - Full Power
(SPt1)
Lowered UVT - Full Power
Dup(SPfl)
Lowered UVT - Full Power
(SPt2)
Lowered UVT - Full Power
Dup(SPt2)
Lowered UVT - Full Power
(SPt 3)
Lowered UVT - Full Power
Dup(SPtS)
Lowered Power- High UVT
(SPt1)
Lowered Power- High UVT
Dup(SPfl)
Lowered Power- High UVT
(SPt 2)
Lowered Power- High UVT
Dup(SPt2)
Lowered Power- High UVT
(SPt 3)
Lowered Power- High UVT
Dup( SPt 3)
Run
22
23
4
5
6
7
24
25
26
15
16
17
Log Influent Concentration
Rep1
5.71
5.74
4.87
4.95
5.17
5.52
5.91
6.03
5.76
5.45
5.28
5.39
Rep 2
5.60
5.76
4.94
5.03
5.10
5.46
5.79
6.16
5.67
5.46
5.29
5.30
Rep 3
5.64
5.68
4.88
4.86
5.18
5.51
5.85
5.82
5.61
5.46
5.30
5.38
Rep 4
5.61
5.51
4.72
4.97
5.16
5.49
5.90
5.85
5.79
5.56
5.28
5.38
Rep 5
5.58
5.74
4.73
4.96
5.33
5.52
5.83
6.02
5.67
5.43
5.30
5.41
Log Effluent Concentration
Rep1
2.93
2.99
2.31
2.61
2.98
2.94
2.11
2.21
3.09
2.75
2.55
2.75
Rep 2
3.03
3.00
2.11
2.61
2.82
2.94
2.18
2.05
3.13
2.68
2.46
2.84
Rep 3
3.05
2.95
2.18
2.48
2.94
no
data
2.08
2.15
3.03
2.64
2.64
2.78
Rep 4
2.92
2.94
2.12
2.71
2.84
2.96
1.99
2.01
2.99
2.68
2.55
2.72
Rep 5
2.98
3.07
2.08
2.57
2.89
2.99
2.03
2.23
2.88
2.77
2.65
2.72
High Flow Rate Test for 3-log Cryptosporidium inactivation demonstration
Lowered UVT - Full Power
(SPt 4)
Lowered UVT - Full Power
Dup(SPt4)
Lowered Power- High UVT
(SPt 4)
Lowered Power- High UVT
Dup(SPt4)
8
9
18
19
5.36
5.45
5.35
5.12
5.36
5.41
5.25
5.28
5.46
5.39
5.38
5.10
5.47
5.36
5.42
4.90
5.44
5.37
5.38
5.08
3.49
3.90
3.14
3.59
3.73
3.68
3.17
3.60
3.50
3.84
3.06
3.46
3.71
3.70
3.40
3.48
3.43
3.55
3.44
3.57
SPt = Set Point Condition
53
-------
November 2013
Table 4-10. ETS UV Model ECP-113-5 MS2 Log Inactivation 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
22
23
4
5
6
7
24
25
26
15
16
17
Log Inactivation
Rep1
2.78
2.75
2.56
2.35
2.19
2.58
3.80
3.83
2.67
2.70
2.72
2.64
Rep 2
2.57
2.76
2.83
2.42
2.28
2.52
3.61
4.11
2.54
2.79
2.83
2.46
Rep 3
2.59
2.73
2.70
2.38
2.25
no data
3.77
3.67
2.58
2.82
2.66
2.60
Rep 4
2.69
2.57
2.60
2.26
2.31
2.53
3.91
3.84
2.80
2.87
2.73
2.66
Rep 5
2.60
2.67
2.65
2.39
2.44
2.53
3.79
3.78
2.79
2.66
2.65
2.68
High Flow Rate Test for 3-log Cryptosporidium inactivation demonstration
Lowered UVT - Full Power (SPt 4)
Lowered UVT - Full Power Dup (SPt 4)
Lowered Power - High UVT (SPt 4)
Lowered Power- High UVT Dup (SPt 4)
8
9
18
19
1.88
1.55
2.21
1.53
1.63
1.73
2.08
1.68
1.95
1.55
2.31
1.64
1.76
1.65
2.02
1.42
2.02
1.81
1.94
1.50
SPt = Set Point Condition
54
-------
November 2013
Table 4-11. ETS UV Model ECP-113-5 MS2 Observed RED 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
22
23
4
5
6
7
24
25
26
15
16
17
RED
(mJ/cm2)
Rep1
60.58
59.61
55.27
49.52
45.39
55.98
91.60
92.52
57.95
53.96
54.53
52.42
Rep 2
54.60
59.93
63.09
51.57
47.65
54.25
85.53
101.78
54.26
56.05
57.23
48.11
Rep 3
55.11
58.99
59.44
50.25
46.85
no data
90.63
87.45
55.26
56.82
52.84
51.31
Rep 4
57.92
54.44
56.38
47.22
48.60
54.49
95.33
93.10
61.32
58.32
54.68
52.88
Rep 5
55.30
57.29
57.92
50.61
52.07
54.60
91.48
91.14
61.16
52.97
52.63
53.49
Average
56.70
58.05
58.42
49.84
48.11
54.83
90.91(1)
93.20(1)
57.20
55.62
54.38
51.64
SD(RED)
2.52
2.26
3.05
1.64
2.51
0.78
3.51
5.28
3.16
2.16
1.85
2.13
USP
12.34
10.82
14.49
9.11
14.46
4.53
10.73
15.71
15.34
10.80
9.43
11.45
High Flow Rate Test for 3-log Cryptosporidium inactivation demonstration
Lowered UVT - Full Power (SPt 4)
Lowered UVT - Full Power Dup (SPt 4)
Lowered Power - High UVT (SPt 4)
Lowered Power - High UVT Dup (SPt 4)
8
9
18
19
37.31
29.64
42.15
27.09
31.52
33.76
39.04
30.28
39.27
29.58
44.54
29.54
34.54
31.96
37.85
24.86
40.82
35.81
35.91
26.61
36.69
32.15
39.90
27.68
3.72
2.69
3.45
2.22
28.18
23.23
23.97
22.22
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.
55
-------
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 REDmeas of 40
mJ/cm2, which was the target minimum REDmeas for developing the set line. Figure 4-9 shows
the set line. The unit is validated for a minimum REDmeas of 40 mJ/cm2 for any combination of
flow rate and intensity above and to the left of the set line. The maximum flow rate demonstrated
was 100 gpm. A UV system cannot operate above the highest validated flow rate and claim a 40
mJ/cm2 REDmeas. The lowest intensity demonstrating a REDmeas of 40 mJ/cm2 was 82 W/cm2. A
UV system cannot operate below the lowest validated irradiance and claim a 40 mJ/cm2 RED.
Set Point 1 - 50 gpm; 82 W/m2
Set Point 2-75 gpm; 89 W/m2
Set Point 3-100 gpm; 105 W/m2
160
40
20
20
40 60 80
Flow Rate (gpm)
100
120
Figure 4-9. Set line for 40 mJ/cmz REDmeas for ETS UV Model ECP-113-5.
56
-------
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 4-12 shows the data for the replicates at each set point. The highest
57
-------
November 2013
RED bias at each set point is used in the validation factor calculations shown later in Section
4.7.3.
Table 4-12. RED Bias Factor for Each Set Point for Cryptosporidium
Sample
Number
22-1
22-2
22-3
22-4
22-5
23-1
23-2
23-3
23-4
23-5
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
24-1
24-2
24-3
24-4
24-5
25-1
25-2
25-3
25-4
25-5
26-1
26-2
Test
Run
22
22
22
22
22
23
23
23
23
23
4
4
4
4
4
5
5
5
5
5
6
6
6
6
6
7
7
7
7
7
24
24
24
24
24
25
25
25
25
25
26
26
UVT
%
78.3
78.3
78.3
78.3
78.3
78.4
78.4
78.4
78.4
78.4
89.3
89.3
89.3
89.3
89.3
89.3
89.3
89.3
89.3
89.3
93.4
93.4
93.4
93.4
93.4
93.4
93.4
93.4
93.4
93.4
97.9
97.9
97.9
97.9
97.9
97.9
97.9
97.9
97.9
97.9
97.9
97.9
Sensitivity
(mJ/cm2 per Log I)
RED Log I Sensitivity
60.58
54.60
55.11
57.92
55.30
59.61
59.93
58.99
54.44
57.29
55.27
63.09
59.44
56.38
57.92
49.52
51.57
50.25
47.22
50.61
45.39
47.65
46.85
48.60
52.07
55.98
54.25
N/A
54.49
54.60
91.60
85.53
90.63
95.33
91.48
92.52
101.78
87.45
93.10
91.14
57.95
54.26
2.78
2.57
2.59
2.69
2.60
2.75
2.76
2.73
2.57
2.67
2.56
2.83
2.70
2.60
2.65
2.35
2.42
2.38
2.26
2.39
2.19
2.28
2.25
2.31
2.44
2.58
2.52
N/A
2.53
2.53
3.80
3.61
3.77
3.91
3.79
3.83
4.11
3.67
3.84
3.78
2.67
2.54
21.8
21.2
21.3
21.5
21.3
21.7
21.7
21.6
21.2
21.5
21.6
22.3
22.0
21.7
21.9
21.1
21.3
21.2
20.9
21.2
20.7
20.9
20.8
21.0
21.3
21.7
21.5
N/A
21.5
21.6
24.1
23.7
24.1
24.4
24.1
24.2
24.8
23.8
24.2
24.1
21.7
21.4
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
1.84
1.84
1.77
1.77
1.77
1.77
1.77
1.77
1.77
1.61
1.61
1.61
1.61
1.61
1.61
1.61
N/A
1.61
1.61
1.55
1.55
1.55
1.55
1.55
1.55
1.55
1.55
1.55
1.55
1.36
1.34
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
1.75
1.75
1.75
N/A
1.75
1.75
1.70
1.70
1.70
1.70
1.70
1.70
1.70
1.70
1.70
1.70
1.40
1.38
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
1.78
1.78
1.78
N/A
1.78
1.78
1.73
1.73
1.73
1.73
1.73
1.73
1.73
1.73
1.73
1.73
1.39
1.38
58
-------
November 2013
Sample
Number
26-3
26-4
26-5
15-1
15-2
15-3
15-4
15-5
16-1
16-2
16-3
16-4
16-5
17-1
17-2
17-3
17-4
17-5
Maximum
8-1
8-2
8-3
8-4
8-5
9-1
9-2
9-3
9-4
9-5
18-1
18-2
18-3
18-4
18-5
19-1
19-2
19-3
19-4
19-5
Maximum
Test
Run
26
26
26
15
15
15
15
15
16
16
16
16
16
17
17
17
17
17
BRED
High Flow
8
8
8
8
8
9
9
9
9
9
18
18
18
18
18
19
19
19
19
19
BRED
UVT
%
97.9
97.9
97.9
96.8
96.8
96.8
96.8
96.8
97.2
97.2
97.2
97.2
97.2
97.2
97.2
97.2
97.2
97.2
Set Point
Set Point
Set Point
Sensitivity
(mJ/cm2 per Log I) BRED
4 log
RED Log I Sensitivity crypto
55.26
61.32
61.16
53.96
56.05
56.82
58.32
52.97
54.53
57.23
52.84
54.68
52.63
52.42
48.11
51.31
52.88
53.49
50 gpm -
75 gpm -
100 gpm
Rate Test for 3-log
93.5
93.5
93.5
93.5
93.5
93.4
93.4
93.4
93.4
93.4
97.3
97.3
97.3
97.3
97.3
97.4
97.4
97.4
97.4
97.4
Set Point
37.31
31.52
39.27
34.54
40.82
29.64
33.76
29.58
31.96
35.81
42.15
39.04
44.54
37.85
35.91
27.09
30.28
29.54
24.86
26.61
175 gpm
2.58
2.80
2.79
2.70
2.79
2.82
2.87
2.66
2.72
2.83
2.66
2.73
2.65
2.64
2.46
2.60
2.66
2.68
82 W/m2
89 W/m2
- 1 05 W/m2
Cryptosporidium
1.88
1.63
1.95
1.76
2.02
1.55
1.73
1.55
1.65
1.81
2.21
2.08
2.31
2.02
1.94
1.53
1.68
1.64
1.42
1.50
-105 W/m2
21.5
21.9
21.9
20.0
20.1
20.2
20.3
19.9
20.0
20.2
19.9
20.0
19.9
19.9
19.5
19.8
19.9
19.9
inactivation
19.9
19.3
20.1
19.6
20.2
19.1
19.5
19.1
19.3
19.7
19.1
18.8
19.2
18.7
18.5
17.8
18.0
18.0
17.6
17.7
1.34
1.36
1.36
1.34
1.36
1.36
1.36
1.34
1.34
1.36
1.34
1.34
1.34
1.34
1.34
1.34
1.34
1.34
1.97
1.84
1.61
BRED
3.5 log
crypto
1.38
1.40
1.40
1.38
1.40
1.40
1.40
1.38
1.38
1.40
1.38
1.38
1.38
1.38
1.38
1.38
1.38
1.38
2.35
2.01
1.75
BRED
3.0 log
crypto
1.39
1.39
1.39
1.38
1.39
1.39
1.39
1.38
1.38
1.39
1.38
1.38
1.38
1.38
1.38
1.38
1.38
1.38
2.54
2.10
1.78
demonstration
1.55
1.55
1.61
1.55
1.61
1.55
1.55
1.50
1.50
1.55
1.34
1.34
1.34
1.34
1.34
1.31
1.31
1.31
1.31
1.31
1.61
1.70
1.70
1.75
1.75
1.70
1.70
1.70
1.64
1.64
1.70
1.38
1.38
1.38
1.38
1.38
1.36
1.36
1.36
1.36
1.36
1.75
1.73
1.73
1.78
1.73
1.78
1.73
1.73
1.68
1.68
1.73
1.38
1.38
1.38
1.38
1.38
1.36
1.36
1.36
1.36
1.36
1.78
N/A - sample not analyzed so RED and bias not determined
59
-------
November 2013
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 = (USP2 +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
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. Itf 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 Tables 4-1 and 4-2, 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.
60
-------
November 2013
The UDR results are shown in Tables 4-3 and 4-6 for the low and high UVT waters for both the
July 18 and 19, 2012 test runs and the September 11,2012 test runs. The July UDR results for low
and high UVT waters (27.48% and 20.74%, respectively) are less than 30%, and therefore UDR is
not used in calculating Uyai for the test runs corresponding to these days of testing. The
September UDR results for low and high UVT waters were 26.99% and 33.46%, respectively.
Since the UDR was >30% at the UV dose corresponding to 1-log inactivation of the MS2 the
uncertainty of the dose response (UDR) is included in the calculation of uncertainty (Uyai) for the
test runs performed in September. The 75 gpm flow rate test with the power turned down
included one test run in July and one test run in September. The September test run had the
highest UDR of 33.46%. The highest Udr measured in September was applied to both test runs
and was included in determining the uncertainty (Uyai) for both 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 JAEDmeas values for each test condition.
2. Calculate the uncertainty of the set point REDmeas using:
USP = [(t x SDRED) / (JAEDmeas)] 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), except for test run 7 which had four valid replicates so the t
value is 3.182.
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 4-11. In
accordance with the GP-2011, the highest USP of the four test runs at each set point determines
the USP for that set point. The highest USP for each set point is 15.71% (50 gpm set point),
15.34% (75 gpm set point), 14.46% (100 gpm set point) and 28.18% (175 gpm set point).
The uncertainty of the validation is equal to the highest USP at a set point when the UDR is <30%
(July test runs) or is calculated using the highest applicable UDR (33.46%) and the highest USP at
a set point for the September test runs using the equations:
Uyal = USP
UVai = (USP2 +UDR2)1/2
Table 4-13 shows the Uyai values used for determining the uncertainty of the validation at each
set point.
61
-------
November 2013
Table 4-13 Uncertainty of the Validation (Uvai) and BRED Values for Cryptosporidium
Set Point
50 gpm-
75 gpm -
100 gpm
175 gpm
82 W/m2
89 W/m2
- 105 W/m2
- 105 W/m2
Max
UDR
%
33.46
33.46
27.48
27.48
Max
USP
%
15.71
15.34
14.46
28.18
Uval
%
36.96
36.81(1)
14.46
28.18
4.0 log
1.97
1.84
1.61
1.61
Max
BRED
3.5 log
2.35
2.01
1.75
1.75
3.0 log
2.54
2.10
1.78
1.78
(1) The lowered UVT - full power runs were performed in July. UDR for July is <30% for the UVai for those two
replicates is 15.34%. The low power test runs were in both July and September, so the highest UDR applies from July
and September is used and the UVai is equal to 36.81% for the low power test run replicates.
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 4-14 shows the calculated Validation Factors (VF) for various Cryptosporidium log
inactivation levels (3.0, 3.5, and 4.0 log inactivation).
Table 4-14 shows the REDyai for Cryptosporidium for each test run using the validation factors
for the various Cryptosporidium log inactivation levels. Table 4-14 shows the Validated Dose for
each set point and a comparison to the dose required for various levels of inactivation of
Cryptosporidium. As can be seen, the tests for the 75 - 89 W/m2 and 100 gpm - 105 W/m2 set
points show a validated dose for Cryptosporidium that achieves a minimum of 4.0 log
inactivation. The other set point (50 gpm - 82 W/m2) achieved a minimum of 3.5 log inactivation
for Cryptosporidium.
Table 4-14 also shows the REDVai for the additional 175 gpm - 105 W/m2 tests achieved a
validated dose for Cryptosporidium that demonstrates a minimum of 3.0 log inactivation.
Therefore, the higher flow rate set point achieved the objective to meet a 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.
62
-------
November 2013
Table 4-14 Validation Factors and Validated Dose (REDVai) for Cryptosporidium
Condition
Lowered UVT -
(SPt1)
Lowered UVT -
Dup (SPt 1)
Lowered Power
(SPt 1)
Lowered Power
Dup (SPt 1)
Lowered UVT -
(SPt 2)
Lowered UVT -
Dup (SPt 2)
Lowered Power
(SPt 2)
Lowered Power
Dup (SPt 2)
Lowered UVT -
(SPt 3)
Lowered UVT -
Dup (SPt 3)
Lowered Power
(SPt 3)
Lowered Power
Dup (SPt 3)
Lowered Power
(SPt 4)
Lowered Power
Dup (SPt 4)
Lowered UVT -
(SPt 4)
Lowered UVT -
Dup (SPt 4)
„ Flow
Run
Rate
gpm
=ull Power
=ull Power
- High UVT
- High UVT
Full Power
Full Power
- High UVT
- High UVT
=ull Power
=ull Power
- High UVT
- High UVT
- High UVT
- High UVT
Full Power
Full Power
22
23
24
25
4
5
26
15
6
7
16
17
8
9
18
19
50
51
50
51
75
75
76
75
101
100
100
101
High Flow
175
175
176
175
Intensity
W/m2
82
82
81
81
84
84
89
87
104
105
105
105
Rate Test for
104
103
105
105
Validation Factor
4.0 log 3.5 log 3.0 log
2.70
2.70
2.70
2.70
2.12
2.12
2.52
2.52
1.84
1.84
1.84
1.84
3.22
3.22
3.22
3.22
2.32
2.32
2.75
2.75
2.00
2.00
2.00
2.00
3-log Cryptosporidium
2.06
2.06
2.06
2.06
2.24
2.24
2.24
2.24
3.48
3.48
3.48
3.48
2.42
2.42
2.87
2.87
2.04
2.04
2.04
2.04
inactivation
2.28
2.28
2.28
2.28
REDmeas
mJ/cm2
567 mJ/cm
58.1
90.9
93.2
58.4
49.8
57.2
55.6
48.1
54.8
54.4
51.6
demonstration
36.7
32.1
39.9
27.7
4 log
2
22« m
21.0
21.5
33.7
34.5
27.5
23.5
22.7
22.1
26.1
29.8
29.5
28.0
17.8
15.6
19.3
13.4
REDva,
3.5 log
2
3.0 log
2
J/cih5(1) mJ/cm 0)
17.6
12
18.0
28.2
29.0
26.1
25.2
26.8
21.5
24.1
20.8
20.6
20.2
19.9
24.0
19.4
27.4
23.6
27.1
26.9
25.8
2R 7
,..25.3
16.4
14.3
17.8
12.3
16.3
16.7
16.1
14.1
17.5
12.1
(1) Required dose for log inactivation validation per the UVDGM-2006 Appendix G; SPt = Set Point Condition
63
-------
November 2013
The four set point tests demonstrating a minimum of 3 log inactivation for Cryptosporidium were
plotted to form a set line. Figure 4-10 shows the set line.
The four set points are:
Set Point 1 - 50 gpm; 82 W/m2
Set Point 2 - 75 gpm; 89 W/m2
Set Point 3-100 gpm; 105 W/m2
Set Point 4-175 gpm; 105 W/m2
c
Ol
c 60
40
20
20 40 60 80 100 120
Flow Rate (gpm)
140 160
180
200
Figure 4-10. Set Line for Minimum 3-log Cryptosporidium Inactivation for ETS UV Model
ECP-113-5.
64
-------
November 2013
4.7.5 Low Wavelength Medium Pressure Lamp Bias Correction
At the time of this testing, the UV industry was addressing a concern about MS2 susceptibility to
low wavelength emission from medium pressure lamps. MS2 has action spectra at 254 nm and
also at 220 nm and lower wavelengths. The UV industry comprising of manufacturers,
engineers, water utilities and regulators have been conducting research and developing solutions
to correct for the low wavelength bias in existing validations. When the work of the UV industry
is completed, a correction factor will be necessary for the results presented herein. NSF
understands that the NIPH requires a 30% correction factor and so does the California
Department of Public Health.
One way for a MP UV reactor to use a germicidal sensor (250-280 nm), would be to validate the
reactor with a lamp sleeve that does not transmit in the lower wavelengths during validation. So
a sensor set point could be established using only the 250-280 nm wavelength emitted by the MP
lamps. Another validation could occur with a lamp sleeve that does transmit at the other
wavelengths. In this case, the difference in the UV dose could be observed and accounted for in
a control strategy.
In the future, NSF will require all medium pressure lamps (with a polychromatic bias) to use a
quartz sleeve designed to filter out the low wavelength when using MS2 to validate a reactor.
NSF will also consider a challenge organism that demonstrates action spectra only for a small
region near the 254 nm wavelength.
65
-------
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 parks 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 4-15 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 primary set point test runs (flow rates of 50, 75, and 100 gpm) achieved a 40
mJ/cm2 validated dose based on MS2. The higher flow rate test did not achieve a 40 mJ/cm2
REDvai based on MS2. This was expected as this higher flow rate test (175 gpm - 105 W/m2)
was targeted at achieving a minimum 3 log inactivation for Cryptosporidium and 40 mJ/cm2
RED
meas-
66
-------
November 2013
Table 4-15 Validation Factors and Validated Dose (REDVai) based on MS2
Condition
Lowered UVT - Full Power
(SPt 1)
Lowered UVT - Full Power
Dup(SPfl)
Lowered Power- High UVT
(SPt 1)
Lowered Power- High UVT
Dup(SPfl)
Lowered UVT - Full Power
(SPt 2)
Lowered UVT - Full Power
Dup(SPt2)
Lowered Power- High UVT
(SPt 2)
Lowered Power- High UVT
Dup(SPt2)
Lowered UVT - Full Power
(SPt 3)
Lowered UVT - Full Power
Dup(SPtS)
Lowered Power- High UVT
(SPt 3)
Lowered Power- High UVT
Dup(SPtS)
High Flow Rate
Lowered Power- High UVT
(SPt 4)
Lowered Power- High UVT
Dup(SPt4)
Lowered UVT - Full Power
(SPt 4)
Lowered UVT - Full Power
Dup(SPt4)
Run
22
23
24
25
4
5
26
15
6
7
16
17
Test for
8
9
18
19
Flow
Rate
gpm
50
51
50
51
75
75
76
75
101
100
100
101
Intensity
W/m2
82
82
81
81
84
84
89
87
104
105
105
105
3-log Cryptosporidium
175
175
176
175
104
103
105
105
Validation
Factor
(1)
1.37
1.37
1.37
1.37
1.15
1.15
1.37
1.37
1.14
1.14
1.14
1.14
inactivation
1.28
1.28
1.28
1.28
REDmeas
mJ/cm2
56.7
58.1
90.9
93.2
58.4
49.8
57.2
55.6
48.1
54.8
54.4
51.6
demonstration
36.7
32.1
39.9
27.7
REDval
based on MS2
mJ/cm2
41.4
42.4
66.4
68.0
50.6
43.2
41.8
40.7
42.0
47.9
47.5
45.1
28.6
25.1
31.1
21.6
(1) BRED equal to 1.0 as the target organism is MS2 the same as the test organism.;
SPt - Set Point Condition
67
-------
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 4-16 through 4-19.
68
-------
November 2013
Table 4-16. Temperature and pH Results
Test
Reactor Blank
Reactor Blank
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 UVT - Full Power (SPt 4)
Lowered UVT - Full Power Dup (SPt 4)
Reactor Blank
Reactor Control
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)
Lowered Power - High UVT (SPt 4)
Lowered Power - High UVT Dup (SPt 4)
Run#
1
21
22
23
4
5
6
7
8
9
10
11
24
25
26
15
16
17
18
19
Temperature
(°F)
Influent
72.3
71.1
71.0
70.9
72.4
72.3
72.5
72.2
71.9
72.0
72.3
72.2
70.4
70.2
70.0
72.1
72.1
71.9
72.0
71.8
Effluent
72.6
71.4
71.2
71.1
72.6
72.5
71.9
72.1
72.3
72.1
72.6
72.3
70.7
70.4
70.2
72.3
72.3
72.2
72.2
72.0
PH
(S.U.)
Influent
8.47
7.41
7.38
7.37
8.60
8.60
8.55
8.54
8.54
8.54
8.52
8.51
7.84
7.89
7.90
8.50
8.38
8.44
8.45
8.48
Effluent
8.53
7.34
7.42
7.41
8.60
8.58
8.53
8.54
8.55
8.55
8.49
8.48
7.86
7.90
7.89
8.46
8.53
8.54
8.57
8.58
Table 4-17. Total Chlorine, Free Chlorine and Turbidity Results
Test
Blank
Blank
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 UVT - Full Power (SPt 4)
Lowered UVT - Full Power Dup (SPt 4)
Reactor Blank
Reactor Control
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)
Lowered Power- High UVT (SPt 4)
Lowered Power - High UVT Dup (SPt 4)
Run#
1
21
22
23
4
5
6
7
8
9
10
11
24
25
26
15
16
17
18
19
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
<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
<0.03
<0.03
<0.03
<0.03
<0.03
Turbidity
(NTU)
Influent
0.37
0.75
0.80
0.66
0.26
0.37
0.30
0.29
0.27
0.30
0.18
0.19
0.38
0.38
0.40
0.15
0.33
0.26
0.24
0.22
Note: Runs 21-23 with the addition of LSAto lower UVT to 79% showed higher readings for turbidity;
suspect interference due to the LSA
69
-------
November 2013
Table 4-18. Iron and Manganese Results
Test
Reactor Blank
Reactor Blank
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 UVT - Full Power (SPt 4)
Lowered UVT - Full Power Dup (SPt 4)
Reactor Blank
Reactor Control
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)
Lowered Power- High UVT (SPt 4)
Lowered Power - High UVT Dup (SPt 4)
Run#
1
21
22
23
4
5
6
7
8
9
10
11
24
25
26
15
16
17
18
19
Iron
(mg/L)
Influent
<0.02
0.06
0.02
0.09
0.02
0.03
<0.02
0.04
<0.02
<0.02
<0.02
<0.02
0.09
<0.02
<0.02
<0.02
0.02
<0.02
<0.02
<0.02
Manganese
(mg/L)
Influent
0.002
0.009
0.008
0.009
0.001
0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
0.003
0.002
0.002
<0.001
<0.001
<0.001
<0.001
<0.001
UVT(1)
(%)
Influent
78
78
78
78
89
89
93
93
93
93
95
95
97
97
97
96
95
97
97
97
Effluent
78
78
78
78
89
89
93
93
93
93
95
95
97
97
97
96
95
96
97
97
(l)UVT on grab samples, measured in laboratory after tests; Five
sample reported here; In- line UVT meter used for flow test results
influent samples averaged; single effluent
70
-------
November 2013
Table 4-19. HPC, Total Coliform and E. coli Results.
Test
Reactor Blank
Reactor Blank
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 UVT - Full Power (SPt 4)
Lowered UVT - Full Power Dup (SPt 4)
Reactor Blank
Reactor Control
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)
Lowered Power - High UVT (SPt 4)
Lowered Power - High UVT Dup (SPt 4)
Run#
1
21
22
23
4
5
6
7
8
9
10
11
24
25
26
15
16
17
18
19
Total Coliform
MPN/100ml_
Influent
<1
<1
34
18
<1
<1
2
1
3
4
2
435
10
13
8
1
10
7
22
5
Effluent
<1
<1
<1
12
<1
<1
<1
<1
<1
<1
<1
328
<1
<1
<1
<1
<1
<1
<1
<1
£. co//
MPN/100ml_
Influent
<1
<1
<1
<1
<1
<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
<1
<1
<1
<1
<1
HPC
CFU/mL
Influent
6.50E+02
2.83E+03
3.56E+03
4.20E+03
9.55E+02
9.85E+02
3.35E+02
5.30E+02
5.25E+02
5.20E+02
3.09E+03
5.06E+03
3.60E+03
1.98E+03
2.41E+03
6.70E+02
1.63E+03
6.35E+02
1.03E+03
1.04E+03
Effluent
2.60E+01
5.60E+01
5.15E+01
6.00E+01
2.80E+01
7.50E+00
2.00E+00
4.50E+00
2.15E+01
3.50E+00
3.06E+02
5.35E+03
2.11E+02
6.60E+01
8.50E+01
5.30E+01
4.90E+01
2.20E+01
2.65E+01
2.25E+01
71
-------
November 2013
4.10 Headloss
Headless was measured over the flow range of 50 to 200 gpm. Pressure at the inlet and outlet of
the reactor was measured at several flow rates as shown in Table 4-20.
Table 4-20. Headloss Measurement Results.
Flow Rate
50
100
150
200
Inlet (psi)
1.949
2.155
2.506
2.954
Outlet (psi)
1.903
2.025
2.212
2.525
Headloss (psi)
0.046
0.130
0.294
0.429
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 4-21 presents the
power measurements taken during the flow tests.
Table 4-21. Power Measurement Results
Test
Reactor Blank
Reactor Blank
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 UVT - Full Power (SPt 4)
Lowered UVT - Full Power Dup (SPt 4)
Reactor Blank
Reactor Control
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)
Lowered Power - High UVT (SPt 4)
Lowered Power - High UVT Dup (SPt 4)
Run#
1
21
22
23
4
5
6
7
8
9
10
11
24
25
26
15
16
17
18
19
Unit
Volts
(volts)
206.8
242.1
242.1
241.8
206.4
206.5
204.6
205.9
205.6
205.1
207.7
208.1
207.5
207.0
206.5
207.7
206.1
206.2
206.2
206.4
Unit
Amperage
(amps)
6.6
10.64
10.62
10.62
6.4
6.4
6.3
6.4
6.3
6.3
3.3
0.0
4.85
4.78
5.12
4.9
5,5
5.5
5.5
5.5
Unit
Power
(Watts)
1150
1960
1960
1950
1110
1120
1100
1110
1100
1090
560
0.0
800
790
840
820
950
960
940
950
72
-------
November 2013
Chapter 5
Quality Assurance/Quality Control
5.1 Introduction
An important aspect of verification testing is the QA/QC procedures and requirements. 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 241.1 nm, 250.0 nm and 278.1 nm within + 0.2 nm 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.
73
-------
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 procedure
and UVDGM-2006 requirements. These procedures and data were reviewed as part of the NSF
QA department review of the microbiological laboratory data.
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%
Trip blanks are normally performed to show that the phage stock 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.
74
-------
November 2013
Table 5-1. Trip Blank Results
Date
July 18, 2012
July 19, 2012
September 11,2012
Trip Blank
Lab Retained
(PFU/mL MS2)
2.10E+07
5.93E+07
4.98E+08
Log10
7.32
7.77
8.70
Trip Blank Travel to Test
Rig and Returned
(PFU/mL MS2)
2.07E+07
5.67E+07
3.97E+08
Log10
7.32
7.75
8.60
Difference
Log10
0.00
0.02
0.10
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 95%
Influent 0 Hour
Influent 4 Hour
Influent 8 Hour
Influent 24 Hour
High UVT 95%
Effluent 0 Hour
Effluent 4 Hour
Effluent 8 Hour
Effluent 24 Hour
PFU/mL
4.27E+02
1.45E+02
5.67E+02
6.10E+02
Average
3.37E+02
1.12E+02
3.37E+02
7.27E+02
Log10
2.63
2.16
2.75
2.79
Log10
2.53
2.05
2.53
2.86
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
3.47E+02
2.31E+02
7.30E+02
1.02E+03
Average
3.50E+02
9.83E+01
4.33E+02
8.03E+02
Log10
2.54
2.36
2.86
3.01
Log10
2.54
1.99
2.64
2.90
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 50, 75,
100, and 175 gpm covering the range of expected flow rates. The flow meter accuracy fell within
a range of 0.6 to 2.7% of the measured tank draw down rate over the range of test flow rates. The
calibration data for the flow meter are shown in Table 5- 3 and achieved the requirement of +/-
5%.
75
-------
November 2013
Table 5-3. Flow Meter Calibration Results
Meter Flow
Rate Read by
meter
(gpm)
51.3
100.5
177.6
52.8
78.1
Volume from
Tank
(gallons)
399.4
612.1
874.4
184.8
268.3
Run Time
(min:sec:millsec)
7:37:62
5:55:42
4:50:51
3:26:54
3:24:87
Flow Rate
Calculated
(gpm)
52.4
103.3
180.6
53.7
78.6
Percent
Difference
(%)
2.1
2.7
1.7
1.7
0.6
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 5.23 logic and an average effluent concentration of 5.26
logic showing a difference of 0.03 logic through the system with lamps off. This meets the
criteria of less than a 0.2 logic 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-
19.
5.7 Documentation
All laboratory activities were documented using specially prepared laboratory bench sheets and
NSF laboratory reports. Data from the bench sheets and 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.
76
-------
November 2013
Table 5-4. Reactor Control and Reactor Blank MS2 Results
Test Condition
Reactor Blank
Reactor Blank
Reactor Blank
Reactor Control
Test Condition
Reactor Blank
Reactor Blank
Reactor Blank
Reactor Control
Test
Run
1
10
21
11
Test
Run
1
10
21
11
UVT
(%)
78.1
95.1
78.4
95.0
UVT
(%)
78.1
95.1
78.4
95.0
Flow
(gpm)
51.1
52.2
50.5
50.1
Flow
(gpm)
51.1
52.2
50.5
50.1
Intensity
(W/m2)
45
48
82
0.0
Intensity
(W/m2)
45
48
82
0.0
Influent (pfu/mL)
Repl
<1
<1
<1
1.86E+05
Rep 2
<1
<1
<1
1.83E+05
Rep 3
<1
<1
<1
1.46E+05
Influent logio
Rep 1
0.0
0.0
0.0
5.27
Rep 2
0.0
0.0
0.0
5.26
Rep 3
0.0
0.0
0.0
5.16
Effluent (pfu/mL)
Repl
<1
<1
<1
2.08E+05
Rep 2
<1
<1
<1
1.65E+05
Rep 3
<1
<1
<1
1.76E+05
Effluent logio
Repl
0.0
0.0
0.0
5.31
Rep 2
0.0
0.0
0.0
5.22
Rep 3
0.0
0.0
0.0
5.25
77
-------
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 each method 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:
own = known concentration of the measured parameter
= measured concentration of parameter
Accuracy of the bench top chlorine, pH, and turbidity meters were 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 measurement. 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
78
-------
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
^ = 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)x 100
Where:
%C = percent completeness;
V = number of measurements judged valid; and
T = total number of measurements.
One replicate sample for MS2 (influent and effluent) from test run 7 was not useable. The total
number of test run replicates was 120 (not counting blanks and controls) yielding a completeness
of 98.3%. All other scheduled samples and analyses were a hundred percent complete. All
planned testing activities were conducted as scheduled, and all planned samples were collected
for challenge organism and water chemistry analysis.
79
-------
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 M5873-2, Plants for the disinfection of water using
ultraviolet radiation, Requirements and testing, Medium pressure mercury lamp plants
(2003)
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
80
-------
November 2013
Attachment 1
Model ECP-113-5 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 ECP-113-5 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.
------- |