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.

-------