May 2013
                            NSF13/38/EPADWCTR
                              EPA/600/R-13/096
Environmental Technology
Verification Report

Inactivation of Microbiological
Contaminants in Drinking Water by
Ultraviolet Technology

NeoTech Aqua Solutions Inc.
Ultraviolet Water Treatment System
NeoTech D438™

                Prepared by
             NSF International

          Under a Cooperative Agreement with
         U.S. Environmental Protection Agency

-------
                            FINAL
          Environmental Technology Verification Report
Inactivation of Microbiological Contaminants in Drinking Water by
                   Ultraviolet Light Technology

                   NeoTech Aqua Solutions, Inc.
                Ultraviolet Water Treatment System
                         NeoTech D438™
                             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

-------
                                         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

-------
                                   Table of Contents
Title Page	i
Notice	ii
Table of Contents	iv
List of Tables	v
List of Figures	vi
Abbreviations and Acronyms	vii
Verification Statement	viii

_Chapter 1 Introduction	1
   1.1    ETV Program Purpose and Operation	1
   1.2    Purpose of Verification	1
   1.3    Verification Test Site	2
   1.4    Testing Participants and Responsibilities	2
Chapter 2 Equipment Description	4
   2.1    NeoTech General Information	4
   2.2    NeoTech UV System Description	4
   2.3    NeoTech D438™ Specifications and Information	7
   2.4    NeoTech Ultraviolet Treatment System Standard Features	8
Chapters Methods and Procedures	9
   3.1    Introduction	9
   3.2    UV Sensor Assessment	10
   3.3    Headloss Determination	11
   3.4    Power Consumption Evaluation	11
   3.5    Feed Water Source and Test Rig Setup	11
   3.6    Installation of Reactor and Lamp Burn-in	15
   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 QA Controls	25
   3.11   Power Measurements	26
   3.12   Flow Rate	26
   3.13   Evaluation, Documentation and Installation of Reactor	26
Chapter 4 Results and Discussion	28
   4.1    Introduction	28
   4.2    Sensor Assessment	28
   4.3    Collimated Beam Dose Response Data	29
   4.4    Development of Dose Response	29
   4.5    MS and Operational Flow Test Data	42
   4.6    SetLineforREDmeasOf40mJ/cm2	48
   4.7    Deriving the  Validation  Factor and Log Inactivation for Cryptosporidium	49
   4.8    Deriving the  Validation  Factor and Log Inactivation for Giardia	56
   4.9    Validated Dose (REDVai) for MS2 as the Target Organism	61
                                          iii

-------
  4.10   Water Quality Data	63
  4.11   Power Measurement	66
  4.12   Headless	66
Chapter 5 Quality Assurance/Quality Control	67
  5.1    Introduction	67
  5.2    Test Procedure QA/QC	67
  5.3    Sample Handling	67
  5.4    Chemistry Laboratory QA/QC	67
  5.5    Microbiology Laboratory QA/QC	67
  5.6    Engineering Lab - Test Rig QA/QC	70
  5.7    Documentation	73
  5.8    Data Review	73
  5.9    Data Quality  Indicators	73
Chapter 6 References	76
                                    Attachments

Attachment 1  NeoTech Technical Manual and Documentation
Attachment 2  Sensor and Lamp Information
Attachment 3  NSF Collimated Apparatus® from NSF Std 55® Annex A
Attachment 4  UVT Scans of Influent and Effluent Water at High and Low UVT

                                    List of Tables

Table 2-1. Basic UV Chamber Information	7
Table 2-2. Low Pressure Lamp Information	7
Table 2-3. Lamp Sleeve Information	7
Table 2-4. UV Sensor Information	7
Table 3-1. Test Conditions for Validation	21
Table 3-2. Analytical Methods for Laboratory Analyses	23
Table 4-1. Sensor Assessment Data	29
Table 4-2. Response Data from Collimated Beam Tests at 91% UVT	32
Table 4-3. Response Data from Collimated Beam Tests at 97% UVT 	34
Table 4-4. Response Data from Collimated Beam Test at 97% UVT outlier removed	36
Table 4-5. Flow Test Operational Data NeoTech D438™ 	43
Table 4-6. Flow Test MS2 Concentrations NeoTech D438™	44
Table 4-7. MS2 Log Concentration for Influent and Effluent NeoTech D438™ 	45
Table 4-8. MS2 Log Inactivation Results NeoTech D438™ 	46
Table 4-9. MS2 RED Results NeoTech D438™ 	47
Table 4-10. RED Bias Factor for Each Set Point for Cryptosporidium	50
Table 4-11. Uncertainty of the Validation (Uvai) and BRED Factors for Cryptosporidium	53
Table 4-12. Validation Factors and Validated Dose for Cryptosporidium Inactivation	55
Table 4-13. RED Bias Factor for Each Set Point for Giardia	57
Table 4-14. Uncertainty of the Validation (Uvai) and BRED Factors for Giardia	59

                                         iv

-------
Table 4-15. Validation Factors and Validated Dose for Giardia Inactivation	60
Table 4-16. Validation Factor and Validated Dose based on MS2	62
Table 4-17. Temperature and pH Results 	64
Table 4-18. Total Chlorine, Free Chlorine, and Turbidity Results	64
Table 4-19. Iron and Manganese Results	65
Table4-17. HPC, Total Coliform, andE. co//Results	66
Table 4-20. Power Measurement Results	67
Table 4-21. Headloss Measurement Results	67
Table 5-1. Stability Results	71
Table 5-2. Trip Blank Results	71
Table 5-3. Flow Meter Calibration	72
Table 5-4. Reactor Control and Reactor Blank MS Results	73
Table 5-5. Completeness Requirements	76
                                   List of Figures

Figure 2-1.  NeoTechD438™	5
Figure 2-2.  NeoTech D438™ Basic Dimensions	6
Figure 3-1.  Schematic of NSF Test Rig	13
Figure 3-2.  Photograph of the Test Unit Setup	14
Figure 4-1. MS2 Collimated Beam Dose versus Log N Day 1 08-02-2012 UVT 91%	38
Figure 4-2. MS2 Collimated Beam Dose versus Log N Day 2 08-03-2012 UVT 97%	39
Figure 4-3. Dose Response Log I versus Dose UVT 91% Day 1 08-02-2012	40
Figure 4-4. Dose Response Log I versus Dose UVT 97% Day 2 08-03-2012  	41
Figure 4-5. Set Line for 40 ml/cm2 REDmeas NeoTech D438™	48
Figure 4-6. Set Line for 3-log Cryptosporidium Inactivation for NeoTech D438™ 	54

-------
                            Abbreviations and Acronyms

A254         Absorbance at 254 nm
ATCC       American Type Culture Collection
°C           Degrees Celsius
CFU         Colony Forming Units
cm           Centimeter
DWS        Drinking Water Systems
EPA         U. S. Environmental Protection Agency
ETV         Environmental Technology Verification
°F           Degrees Fahrenheit
GP           Generic Protocol
gpm         Gallons per minute
h            Hours
HPC         Heterotrophic Plate Count
L            Liter
Ibs           Pounds
LEVIS        Laboratory Information Management System
Log I        Log inactivation
LT2ESWTR  Long Term 2 Enhanced Surface Water Treatment Rule
m            Meter
min         Minute
ml           Millijoules
mg           Milligram
mL          Milliliter
MS2         MS2 coliphage ATCC 15597 Bl
NeoTech     NeoTech Aqua Solutions, Inc.  (formerly Ultraviolet Sciences, Inc.)
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
ORD         Office of Research and Development
PFU         Plaque Forming Units
psi           Pounds per Square Inch
QA          Quality Assurance
QC          Quality Control
QA/QC      Quality Assurance/Quality Control
QAPP       Quality Assurance Proj ect Plan
QMP        Quality Management Plan
RED         Reduction Equivalent Dose
REDmeas      Measured Reduction Equivalent Dose - from test runs
REDvai      Validated Reduction Equivalent Dose - based on selected pathogen and
             uncertainty
RPD         Relative Percent Deviation
                                         vi

-------
SM          Standard Methods for the Examination of Water and Wastewater
SOP         Standard Operating Procedure
TQAP       Test /Quality Assurance Plan
IDS         Total Dissolved Solids
ISA         Tryptic Soy Agar
UVT         UV transmittance
TSB         Tryptic Soy Broth
ug           Microgram
jam          Microns
USEPA      U. S. Environmental Protection Agency
UVDGM     UV Design Guidance Manual 2006
UDR         Uncertainty of collimated beam data
Us           Uncertainty of sensor
USP          Uncertainty of set point
UVAL         Uncertainty of validation
                                         vn

-------
                                       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 ultraviolet light
(UV) dose  delivered by  the  NeoTech Aqua  Solutions  Inc. (NeoTech) Ultraviolet Water
Treatment System Model D438 (NeoTech D438™)  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 and Giardia using the Generic Protocol for Development of
Test  / Quality Assurance Plans for  Validation of  Ultraviolet  (UV) Reactors,  August 2011

                                           1

-------
11/01/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 as a
single flow rate and irradiance output that delivers the targeted UV dose). The results of the three
set point tests were used to develop a setline that defines the maximum flow rate - minimum
irradiance output required to ensure the UV dose is achieved. The  microorganism used for this
validation test was MS2 coliphage virus (MS2).  The target UV dose was a measured Reduction
Equivalent Dose (REDmeas) of >40 ml/cm2. This  dose was calculated based on the understanding
of dose calculations used internationally and by the Ten States Standards. The REDmeas was then
adjusted based on the uncertainty of the measurements to calculate  a MS2 based validated dose
(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 and Giardia RED bias factors from UVDGM-2006 Appendix G.  The data
were used to estimate the log inactivation of Cryptosporidium and  Giardia so that a regulatory
agency could grant log  credits under the Long Term 2 Enhanced Surface Water Treatment Rule
(LT2ESWTR).  NeoTech selected flow rates of  150, 250,  and 435  gpm as the target flow rates
based on their design for NeoTech D438™.

Based  on the results 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, UV transmittance (UVT), and lamp status [40
CFR  141.720(d)(2)]. Under the UV setline approach, UVT does not have to be measured
separately. The intensity readings by the sensor  take into account changes in the UVT and the
setline establishes the operating conditions over a range of flow rates used during the validation
test.

This verification test did not evaluate cleaning  of the lamps or quartz sleeves, nor any  other
maintenance and operation.

1.3    Verification Test Site
UV dose  validation testing was performed at  the NSF Testing  Laboratory in Ann Arbor,
Michigan.  The NSF laboratory is used for all of the testing activities for NSF  certification of
drinking water treatment systems, and pool and spa treatment systems.

1.4    Testing Participants and Responsibilities
The following  is a brief description  of each  of the ETV participants  and their roles and
responsibilities.

1.4.1   NSF International
NSF is an independent, not-for-profit organization dedicated to  public health and safety, and to
protection of the environment.  Founded  in 1944 and located in  Ann Arbor, Michigan, NSF has

-------
been instrumental in the development of consensus standards for the protection of public health
and the environment.  The USEPA partnered with NSF to verify the performance of drinking
water treatment systems through the USEPA's ETV Program.

NSF performed all verification testing activities at its Ann Arbor, MI location. NSF  prepared the
test/QA plan (TQAP), performed all testing, managed, evaluated, interpreted, and reported on the
data generated by the testing, and reported on the performance of the technology.

Contact:  NSF International
          789 N. Dixboro Road
          Ann Arbor, MI 48105
          Phone: 734-769-8010
          Contact: Mr. Bruce Bartley, Project Manager
          Email: bartley@nsf.org

1.4.2   U.S. Environmental Protection Agency
USEPA, through its Office of Research and Development (ORD), has financially supported and
collaborated with NSF under Cooperative Agreement No. R-82833301. This verification effort
was supported by the DWS Center operating under the ETV Program. This document has been
peer-reviewed, reviewed by USEPA, and recommended for public release.
1.4.3   NeoTech Aqua Solutions, Inc. (formerly Ultraviolet Sciences, Inc.)
NeoTech 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. NeoTech also provided logistical and technical support, as needed.

Contact:   NeoTech Aqua Solutions, Inc.
          5893 Oberlin Drive, Suite 104
          San Diego, California 92121
          Phone: 1-858-571-6590 or 1-888-718-5040
          Email:  info@neotechaqua.com
          Website: neotechaqua.com

-------
                                     Chapter 2
                                Equipment Description

2.1    NeoTech General Information
NeoTech, headquartered in San Diego, California, designs and manufactures UV water treatment
systems  for disinfection, TOC reduction, and chlorine reduction  purposes. NeoTech water
treatment products are designed for industrial and commercial applications; pharmaceutical,
microelectronics,  beverage, pools/spas, hospitality, water reclamation, and  small municipal
drinking water delivery systems (<10 MGD).

The NeoTech team of scientists and engineers work in  collaboration with a group of universities
to develop  more efficient UV  systems for water purification applications. NeoTech has taken
these innovations and applied them to a new product line of UV reactors.

NeoTech developed  a highly  reflective UV  treatment chamber (US  Patent 7,511,281) that
maximizes  the use of the UV light emitted by conventional mercury amalgam lamps. NeoTech
has stated the following regarding the new treatment chamber:

   The 99.8% reflective surface keeps the UV light inside the treatment chamber, reducing
   the amount of light energy necessary to  achieve a proper  UV dose. This reflective
   surface  encapsulates the  entire flow channel  ensuring an  even  UV dose exists
   throughout the treatment chamber. No  complicated internal baffling and mixing systems
   are necessary. This efficiency gain is unique to the NeoTech product line, resulting in
   smaller UV systems with fewer  UV lamps to achieve a given  UV dose. As a result
   NeoTech water treatment systems are more compact in design and require significantly
   less power to operate, saving the end user up to 90% in operating costs.

The NeoTech product line of treatment chambers was launched commercially in March 2009
after nearly two years of field trials.

2.2    NeoTech UV System Description

The NeoTech Ultraviolet Water Purification System validated in this test was the largest flow
rated unit, NeoTech D438™. This unit is rated by NeoTech to handle 500 gpm. The system uses
two low-pressure mercury amalgam lamps and one intensity  sensor mounted in a stainless steel
flow chamber. Figure 2-1 presents a picture of the system and basic dimensions of the system are
shown in Figure 2-2. NeoTech provided  an  operating manual  (Attachment 1 of this report),
which included additional schematics and drawings with parts and dimensions of the reactor, the
sensors, the lamps and the quartz  sleeve placement.   NeoTech has also provided additional
information for the UV sensor (spectral data, measuring angle, measuring range, and output
range) and  for UV lamps (lamp life, irradiance output, power requirements, aging data, etc.) as
required for the validation test. This information is presented in Attachment 2.

-------
Figure 2-1. NeoTech D438T
                       M

-------
                 40.55
                 1030
                 36.95
                 938.6
                                     6.00
                                     152.4
                                       A
3.08
78.2
                 29.53
                 750.1
  4X  .40 10.1  THRU
                            7.90
                           200.7
         f
       7.96
       202.2
         i
                                 11.40
                                 289.5
Figure 2-2. NeoTech D438™ Basic Dimensions.

-------
2.3   NeoTech D438™ Specifications and Information
NeoTech provided the following information about their UV reactor:
                      Table 2.1. Basic UV Chamber Information
Manufacturer/Supplier
Type or model
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
NeoTech Aqua Solutions, Inc.
D438
2009
500 gpm
63 pounds
278.38 cu inches
120 VAC, 50/60Hz; 15 A max.
300 watts
150 psi static and transient
35 °F (2 °C) min.; 80 °F (27 °C) max.
Steam sterilize with pure steam up to 257 °F
for up to 90 minutes
3 inches
                     Table 2.2. Low Pressure Lamp Information
Type
Model
Number of lamps per reactor
UV emission at wavelengths ranging from
240-290 nm
Lamp life
Power supply unit's name, make and serial
numbers
Irradiance @lm
UV output
Operating lamp watts
Lamp current and voltage
Low-pressure
Light Sources, Inc. Model M1-4Y-01
NeoTech Part Number LK 38
2
> 90% irradiance at 253-256 nm
9000 hrs - Aging data in Attachment 2
Fulham WHS or WH7 ballast 120 V;
Power supply by NeoTech,
334 (mW/cm2)
35 W
111 W
1.3 A; 86 V
                       Table 2.3. UV Lamp Sleeve Information
Type or model
Quartz material
Pressure resistance
Heraeus Quartz;
Supersil310
lOOOkPa
                          Table 2.4. UV Sensor Information
Type / model
Measuring field angle
Number of sensors per reactor and placement
Signal output range
Measuring range output signal
UVIM-3 - 1660-002
180 degrees with cosine
degrees without cap
corrector cap, 64.7
1
4 - 20 mA
0-160 mW/cm2

-------
UV sensor spectral information has been provided by NeoTech to demonstrate the sensor meets
the basic requirements of the Generic Protocol for Development of Test/Quality Assurance Plans
for Validation of Ultraviolet (UV) Reactors, 7/2010 (GP-2010).  The GP-2010 and updated GP-
2011  are based on the EPA's UVDGM-2006. These data are presented in Attachment 2. Sensor
calibration information was provided by NeoTech prior to the start of the test. Sensor calibration
was also checked with reference sensors as part of the test procedures.  NeoTech has provided
information on the calibration  (NIST traceable) of the  equipment used to calibrate the sensors.
These data are provided Attachment 2.

2.4    NeoTech Ultraviolet Treatment System Standard Features
NeoTech  has provided the  following information  on the features of the NeoTech series of
reactors.

Standard Features of all NeoTech systems include:

    •   316L and 304 stainless steel construction;
    •   Ra-15 finish for all wet  contact surfaces;
    •   Low Pressure Amalgam lamps;
    •   Static Operation up to 1 hour (no flow);
    •   NIST traceable UV intensity monitor with LED output on control panel;
           -   provides real-time lamp intensity information;
              - 4-20mA signal for data logging of UV intensity;
    •   High operating pressure (150 psig);
    •   NEMA 4X Control panel with remote monitoring and control capability;
              4-20mA scaled analog output;
              - Remote Shut off;
              - Alarms for lamp out and power off;
    •   Sanitary fittings and Viton gaskets;
    •   UV lamp replacement requires no tools required; and
    •   UL and CE approved. TUV verified.

-------
                                       Chapter 3
                                Methods and Procedures
3.1    Introduction
The tests followed the procedures described in the Test/Quality Assurance Plan for  The
Ultraviolet Sciences Inc. Ultraviolet (UV) Water Purification System Model UVXS438S Reactor,
August 2010 (TQAP).  The TQAP was adapted from GP-2010. The GP-2010 and updated GP-
2011 version are based on the EPA'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, 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 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 NeoTech D438™, 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".  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 (REDvai)  in accordance with the
unique approach of the  State of New York.  The REDmeas  data were also adjusted for uncertainty
and the Cryptosporidium and Giardia  RED bias factors from the UVDGM-2006 Appendix G.
These validated RED data  can be used by  states to evaluate  applicable log  credits under the
LT2ESWTR.

The GP-2010  requires  the  use of  a second less sensitive challenge organism as  part of the
validation. T7 was initially  included in the ETV UV Generic Protocol in the 2010 version as a
result of research suggesting it could  be a surrogate test microorganism with UV sensitivity
similar to the UV sensitivity of  Cryptosporidium  (Fallen et.al, JAWWA, 99.3, March 2007).
The GP-2010 technical  advisory panel had reservations about using any test microorganism other
than MS2 which has  an  excellent record  of quality control response for  collimated beam
regression curves (Figure  A.I in the  UVDGM-2006). The ETV GP-2010 technical advisory
panel opinion was that  other test microorganisms simply did not yet have the record of quality
control limits as did MS2.

In 2010 during some initial validation  studies, NSF attempted to use T7.  The strain referenced
by the JAWWA study (ATCC 11303-B7) was not available through ATCC.  In fact, ATCC said
verbally that the strain  mentioned was not T7 and was not available.  With the counsel of the
EPA, NSF agreed to try using T7 ATCC BAA-1103-B38.

-------
Comments in 2011 on the GP-2010 also found reasons not to specify only T7: ". However, T7
cannot be produced at nearly as high a titer as Tl, so in the validation of high-flow reactors,
replacing  all the Tl  test conditions with T7 test conditions would consume an unacceptable
volume of raw phage stock." Consequently the GP-2010 technical advisory panel recommended
the use of any organism  other than MS2 will be optional and the use of MS2 will be mandatory
for all  types of reactors.  The use of a challenge organism other than MS2 will be determined by
the consensus of stakeholders.

For the retesting done for this project, NSF chose to only use MS2 based on the concerns raised
about T7 by reviewers and the changes made in the 2011 ETV UV Protocol (GP-2011). Instead,
it was  decided to illustrate how MS2 data was being used to satisfy many different regulatory
requirements while using essentially the same data. The basic biodosimetry data was used to
calculate the log inactivation of two target pathogens: Cryptosporidium and Giardia. The data
was also used to calculate the 40mJ/cm  dose (REDmeas) requirement found in the "Ten States
Standards" and the NIPH guidelines, and the "validated" dose approach (REDvai)  used by the
NYDOH.

UV reactor validation followed these steps:

       1.  Obtain the technical specifications for the system as provided by NeoTech;
       2.  Assessment of the UV sensors;
       3.  Collimated beam laboratory bench scale testing;
       4.  Full scale reactor testing;
       5.  Calculate the  REDmeas; and
       6.  Adjust the REDmeas  for uncertainty in UV  dose  and calculate a  validated dose
          (REDyai) for Cryptosporidium and Giardia to show the log inactivation.
	                                                 r\                    	
The target UV dosage was a REDmeas of > 40 ml/cm , based on MS2. NeoTech selected flow
rates of 150, 250, and 435 gpm as the target flow rates based on their system design for NeoTech
D438™ and screening and initial data from 2010.

3.2    UV Sensor Assessment
The NeoTech test unit duty sensor was evaluated according to the UV sensor requirements in the
UVDGM-2006 prior to and following the verification testing. All UV intensity sensors (the duty
and two reference sensors) were new sensors. Evidence of calibration of the sensors, traceable to
NIST,  was provided by NeoTech.

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 NeoTech  sensor  and  representation of sensitivity   to  the  germicidal
wavelength were provided by NeoTech  and found to meet the requirements. The technical
specifications of the  NeoTech UV sensor and representation of sensitivity to the germicidal
wavelength are included  in Attachment 2.
                                           10

-------
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 UVT and the maximum
       lamp power setting to be used during validation testing.

   2.   Step  2: Using two  calibrated  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 [Sduty/ SAvgRef- 1]

                Where:
                    S duty = Intensity measured by a duty UV sensor,
                    S Avg Ref = Average UV intensity measured by  all the reference UV sensors
                    in the same UV sensor port with the same UV lamp at the same UV lamp
                    power.

3.3    Headloss Determination
Headloss through the unit was determined over the range of expected flow rates, in this case
from 150 gpm to 450 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 150, 250,  350 and 450 gpm. These data are reported  Section 4.12.

3.4    Power Consumption Evaluation

The amperage and voltage used by the unit were measured during all reactor test runs.

Power data are presented in Section 4.11.

3.5    Feed Water Source and Test Rig Setup
The water source for this test was City of Ann Arbor, Michigan municipal drinking water. The
water was de-chlorinated using activated carbon and confirmed by laboratory testing.  For the
lowered UVT conditions,  the chemical Lignosulfonic Acid  (LSA)  was used to lower the UV
transmittance to a level that achieved a duty sensor reading at the selected UV intensity set point.

                                          11

-------
LS A 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.

NSF used a UV  test rig and system setup that is designed to conform to  the  specifications
described in Sections 5.4.3 and 5.4.4 of the 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.  Figure 3-2 shows a photograph of the actual equipment and piping
setup in the laboratory.

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 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 the test unit, downstream
of a 90° elbow attached directly to the unit outlet and downstream of a second inline static mixer.
This ensured good mixing of the treated water prior to the effluent sampling port. The  90°
elbows prevented  stray UV light from exiting the unit.

A power platform that measures amperage, volts, watts, and power factor was used to monitor
power use by the test unit. The unit was wired into the platform and power consumption was
recorded for each  test run.
                                          12

-------
     Fluid Flow
                       Dosing Pump
No'.e: A! plumbing is snedule du PVC
     Figure 3-1. Schematic of NSF Test Rig
                         13

-------
Figure 3-2. Photograph of the Test Unit Setup.




                    14

-------
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 NeoTech installation and assembly instructions.  Two 90°  elbows, one
upstream and one downstream of the unit, were used in the test rig setup to eliminate stray UV
light. Figure 3-1 shows a schematic of the test rig setup and Figure 3-2 shows a photograph of
the actual equipment setup. The UV lamps were new and therefore the system was  operated for
100 hours prior to the start of the tests to provide proper burn-in of the lamps.

There  was one  duty sensor  and  two lamps in the NeoTech system. Therefore,  the lamp
positioning check requirements (i.e. checking each lamp and placing the lowest  output lamp
closest to the sensor) were required for this validation. The sensor readings were basically the
same before and after the lamp positions were switched. The lamp positions that gave the highest
sensor  reading were used for all of the test runs. This provided the most conservative approach.

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 produce a UV dose-response curve. The following sections  describe the
details  of the collimated beam testing as performed by NSF.

3.7.1   Test Microorganism (Challenge)

MS2 coliphage (ATCC  15597-B1) was used in collimated beam bench scale testing and  for the
full-scale reactor dose validation tests.

MS2 coliphage ATCC 15597-B1 is a recommended microorganism  for UV lamp validation tests.
Further reasons  for selecting this  microorganism for UV validation are based on its inter-
laboratory  reproducibility (UVDGM-2006), ease  of use and culturing,  and demonstrated
performance of MS2 in validation testing.

3.7.2   Test Conditions
The collimated beam tests were performed in duplicate at the minimum and maximum UVT test
conditions. For this validation the testing spanned a period of two (2) days for the MS2 test runs,
with Day 1 being the lowered UVT water tests and Day 2 the high UVT water lowered power
tests. Collimated beam tests were run in duplicate on the minimum UVT water (90-91%) on Day
1. Collimated beam tests were run in duplicate on the maximum UVT water (97-98%) on  Day 2.
Thus, for this validation test, there are two sets of duplicate collimated beam test data for MS2,
one at low UVT and one at high UVT.


                                          15

-------
For MS2, 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, 60 and 80 ml/cm2.  The samples are clustered  close to the
        9                                                            9
40mJ/cm target dose with two doses above and below the target of 40mJ/cm .

The collimated beam radiometers were calibrated to ensure that the measured UV intensity met
the criteria of an uncertainty of 8% or less at a 95% confidence level.

3.7.3   Test Apparatus
NSF uses a collimated beam apparatus that conforms to NSF/ANSI Standard 55c section 7.2.1.2.
Attachment  3  includes  a description  of the apparatus  and is reproduced  for informational
purposes and is copyright protected.

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.
The collimated beam test water and microorganism  culture were the same as used in the full
scale reactor tests.

NSF microbiological laboratory personnel followed the "Method for Challenge Microorganism
Preparation, Culturing the Challenge Organism and Measuring its Concentration" in  Annex A of
NSF/ANSI Standard 55®.  Please note that all reproduced  portions of NSF/ANSI Standards are
copyright protected.

For collimated beam testing of a water sample containing challenge  microorganisms,  NSF's
laboratory followed this procedure:

       1. Measure the A254 of 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 a sample.
       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
                                           16

-------
         calculate the average microbial value for the dilution from the three plate replicates
         that provide the best colony count.
       10. Re-measure the UV intensity and calculate the average of this measurement and the
          measurement taken in Step 3. The value should be within 5% of the value measured
          in Step 3. If not, recalibrate radiometer and re-start at step 1.
       1 1 . 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 1 1 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:
       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); and
             t = Exposure time (s).

To control for error in the UV dose measurement, the uncertainties of the terms in the UV dose
calculation meet the following criteria:

   •   Depth of suspension (d) 
-------
              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
              ofPFU/mL.
The procedure for developing the UV dose response curves was as follows:

   1. For each UV test condition (high or low UVT water) and its replicate and for each day of
     testing, log N (PFU/mL) was plotted vs. UV dose (mJ/cm2). A best fit regression line was
     determined and a common N0 was identified as the intercept of the curve at UV dose = 0. A
     separate equation was developed for each UVT condition (lowest and highest). In this test,
     there were two days of testing for MS2, so there were two sets of data.
   2. The  log inactivation (log I) was  calculated for each measured value of N (including zero-
     dose) and the common N0 identified in Step 1 using the following equation:

                                      log I = log(No/N)

                Where:
                    No = The common N0 identified in Step 1 (PFU/mL); and
                    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
     REDmeas 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 quadratic equations were evaluated to determine the best fit
of the data. The  regression  coefficient,  R2,  was  determined for each trend  line and was
considered acceptable  if it was  0.9 or greater and  for "r" ± 0.95 or greater.  The equation
coefficients for each  day  were  also evaluated statistically  to determine which terms were
statistically significant  based on the P factor. A second order polynomial gave the best fit for the
collimated beam dose response curves.

For this validation a  single curve  corresponding to one day's worth of full scale reactor testing
was used  to calculate REDmeas values for that day. The higher UVT dose response curve was
                                           18

-------
used for the high UVT water day with reduced power and the lower UVT dose response curve
was used for the day when the UVT of the test water was lowered with LSA.

3.7.6   Collimated Beam Data Uncertainty
The collimated beam data was fit to a polynomial regression and the uncertainty of the dose
response equation based on a 95% confidence interval (UDR) was calculated as follows:

                               UDR = 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, and
             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 and results are included in Section 4.4

3.8    Full Scale Testing to Validate UV Dose

3.8.1   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.
NeoTech specified  the target flow rates (150, 250, 435 gpm) and UV target intensity levels (7.5,
10.0,  13.0  mW/cm2) 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 UVTs of 91%,
94%, and 97%.

Each set  point represents a given flow rate with testing under two conditions, (1) lowered UVT-
maximum 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 (150 gpm - 7.5 mW/cm2; 250 gpm - 10.0 mW/cm2; 435 gpm -
13.0 mW/cm2) were tested for the set line. All conditions were  performed in duplicate.  The
intensity  targets were based on expected intensity at UVT's of 91%,  94%, and 97%.

The  LT2ESWTR  requires validation of UV reactors  to determine a log  inactivation of
Cryptosporidium or other target pathogens 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 were  run to show an  example of determining the log inactivation of
Cryptosporidium and Giardia by adjusting the REDmeas for uncertainty and RED bias.
                                          19

-------
A reactor control test (MS2 injection with the lamp off) was run at the low flow rate (150 gpm)
and with high UVT water, which demonstrated that there was no inactivation 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 the lamps at full power 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 (REDmeas).

Flow rate, intensity, and UVT data (from the NSF in-line UVT monitor) were collected at each
of the five  sample collection times for all test runs. These data were averaged to determine the
average flow rate, UVT, and intensity for each test condition and it's duplicate.

In addition, samples  for pH, turbidity, temperature, total and residual  chlorine,  E coli, and
heterotrophic plate count (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.
                                           20

-------
                        Table 3-1. Test Conditions for Validation
Validation Test
Condition 1
Condition 2
Condition 3
(reactor control)
Condition 4
(reactor blank)
Flow Rate
150 gpm
250 gpm
435 gpm
150 gpm
250 gpm
435 gpm
150 gpm
150 gpm
UVT (%)
91%
94%
97%
> 97%
> 97%
>97%
>97%
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
3.8.2   Preparation of the Challenge Microorganisms
The challenge microorganism (MS2) used to validate the UV reactor was cultured and analyzed
by NSF's microbiology laboratory as specified in Standard Methods for the Examination of
Water  and Wastewater. NSF microbiological  laboratory personnel  followed the  method for
"Culture of challenge microorganism" in Annex A of NSF/ANSI Standard 55.
                                                                             1 9
Propagation resulted in a highly concentrated stock solution (approximately 1.0x10  PFU/mL)
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 mJ/cm2, the prediction intervals of the data shown in Appendix A of the UVDGM-
2006 are represented by the following equations"
       Upper Bound: log/= -1.4x10  *UVDose2 + 7.6x10 2 *UVDose
       Lower Bound: log/ = -9.6xlO~5 *UVDose2 + 4.5xlO~2 *UVDose

City of Ann Arbor tap water was filtered using activated carbon to remove any residual chlorine
(confirmed by chemical analysis for total  chlorine of the test water),  organic surfactants and
dissolved organic chemicals that may be UV absorbers. The filtered challenge water was then
tested for the following parameters and found acceptable if the result is non-detectable or as
otherwise indicated below:
                                          21

-------
       •  Total Chlorine;
       •  Free Chlorine;
       •  UV254 ;
       •  UVT > 95%;
       •  Total Iron;
       •  Total Manganese;
       •  Turbidity < 0.3 Nephelometric Turbidity Units (NTU);
       •  Total coliform (<1 CFU/lOOmL); and
       •  Heterotrophic Plate Count (<100 CFU/mL).

3.8.3    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 3.6.1. 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 the flow rate, UV sensor measurements
       and the UVT to assure the test water and reactor meet the test conditions. 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 demonstrated 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 two minutes after the injection pump was started.
    2.  MS2 was  injected  into the feed water flow  upstream  of the reactor to achieve a
       concentration  greater than IxlO5 PFU/mL so that  a minimum  of a 4-log inactivation
       could be measured during the runs.
    3.  Sample taps remained open over the duration of the test.
    4.  Samples were collected in  accordance  with  standards of good  practice  as defined by
       Standard Methods Section 9060.
    5.  Five (5) sample pairs were collected during approximately ten to twelve minutes of
       continuous flow at steady conditions. Each set of influent and effluent grab samples were
       collected as close  in time as possible. The five sets  of samples were spread out over the
       10 to 12-minute continuous flow run.
    6.  Samples for assessing the challenge microorganism concentrations  in the influent and
       effluent  were collected in 125 mL bottles.
    7.  Samples were collected in  bottles  that have been  cleaned and sterilized by the NSF
       laboratory.
    8.  Collected samples were delivered directly to the microbiological  lab  located in the same
       building after each sampling period. Sample analysis was generally started immediately,
       but could be stored in the dark and analysis started a few of hours later. All MS2 analysis
       was started within 4-6 hours of the time the samples were collected.
                                           22

-------
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 HPC  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 REDn
validated dose (REDyai) at a each set point.

3.9    Analytical Methods
All laboratory analytical methods for water quality parameters are listed in Table 3-2.

                 Table 3-2. Analytical Methods for Laboratory Analyses
and the


Parameter
Temperature
pH

E. coli 1 Total
Coliform
Iron

Manganese

Turbidity
MS2
Absorbance UV254
Residual chlorine
HPC



Method
SM(2) 2550
SM 4500-H+B

SM 9223

EPA 200.7

EPA 200.8

SM2130
Top Agar
Overlay
SM5910B
SM 4500-C1 D
SM9215B

NSF
Reporting
Limit
-


1CFU
/lOOmL
20 ug/L

1 ug/L

0.1 NTU
1 PFU/mL
NA
0.05 mg/L
1 CFU/mL


Lab Accuracy
(% Recovery)
-
±.1SU
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
-
(3)

24
hours
180
days
180
days
(3)
24
hours
2 days
(3)
24
hours

Sample
Container
-
NA

500 mL
plastic
125 mL
polyethylene
125 mL
polyethylene
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
                                           23

-------
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 at a minimum on
the first day of testing.  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
Spectrophotometer internal QA/QC procedures outlined in the UVDGM-2006 were used to
verify calibration. UV absorbance of solutions used to zero the Spectrophotometer were verified
using reagent grade organic-free water certified by the supplier to have zero UV absorbance.

The measurement uncertainty of the Spectrophotometer must be 10% 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.
                                          24

-------
   •   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 (LEVIS), and were used on the
NSF lab reports for the tests. All challenge organism samples were stored in the dark at 4 + 2 °C
and processed for analysis within 4-6 hours.

3.10   Full Scale Test QA 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 reactor
       UV lamps in the reactor turned off. The change in log concentration from influent to
       effluent should correspond to no more than 0.2 logic.
   •   Reactor blanks - Influent and effluent water samples were collected with no addition of
       challenge microorganisms 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
       microorganisms during transport to the laboratory (in the same building).
   •   Method blanks -  A  sample bottle of sterilized reagent grade  water was  analyzed
       following the challenge microorganism  assay procedure. The concentration  of challenge
       microorganism with the method  blank was non-detectable.
   •   Stability samples - Influent and effluent samples at low and high UVT were collected
       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
       microorganism was added to achieve a concentration of approximately IxlO3 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 h, 8 h and 24 h after time 0.  All analyses were
       performed in triplicate. While stability samples were performed during the test, they are

                                           25

-------
       not directly applicable in this case as all sample analyses for MS2 were started within a 4
       to 6 hours of collection.

3.11   Power Measurements
The voltmeter and ammeter used to measure UV equipment voltage and amperage had traceable
evidence of being in calibration (e.g., have 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% of the true value.  Flow meter accuracy was verified by monitoring the draw
down volume in the supply tanks over time. The supply tanks have been calibrated using the
catch and  weigh technique. The flow meter was within 1.66% of the true value.  Flow meter
calibration data are presented in Section 5.6.

3.13   Evaluation, Documentation and Installation of Reactor
NeoTech provided  technical information on the NeoTech D438™ and basic information  on the
UV  lamps, sensor, and related  equipment. An operating  manual was provided. Additional
information on the  lamp output (confirmation of spectral output) was provided prior to the start
of the validation test. All documentation and equipment data was reviewed prior to the start of
testing. The following documentation was reviewed and found to conform to the GP-2011 and
UVDGM-2006 requirements:

 Reactor Specifications:
    •   Technical description of the reactor's UV dose-monitoring strategy, including the use of
       sensors,  signal processing, and calculations (if applicable);
    •   Dimensions and placement of all wetted components (e.g., lamps, sleeves, UV sensors,
       baffles, and cleaning mechanisms) within the UV reactor;
    •   A technical  description of lamp placement within the sleeve; and
    •   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; and
    •   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;
    •   Materials of construction; and
    •   UVT at 254 nm for the medium pressure (MP) lamp with germicidal sensors.
Specifications for the reference and the duty  UV sensors:

                                           26

-------
    •   Manufacturer and product number; and
    •   Technical description including external dimensions.
Sensor measurement properties:
    •   Working range;
    •   Spectral and angular response;
    •   Linearity;
    •   Calibration factor;
    •   Temperature stability; and
    •   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;
       and
    •   Operation and maintenance manual including cleaning procedures, required spare parts,
       and safety requirements.
                                           27

-------
                                      Chapter 4
                                Results and Discussion
4.1    Introduction
The validation tests to demonstrate a minimum REDmeas of 40 ml/cm were run on August 2 and
3, 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 (91%, 94%, and 97%) and the lamps
were operated at full power. The second day of testing was dedicated to the test conditions and
duplicates where highest UVT feed water (97% target) was used and the  lamp power was
reduced to achieve the target intensity level. The test conditions and detail on the test rig setup,
sampling procedures, and unit operation have been 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 same duty sensor was used for monitoring intensity (irradiance) for all test runs. This sensor
measured the intensity from the two low pressure lamps in the unit. The control panel provided
direct readings of intensity in mW/cm .  This direct reading was based on converting the 4-20 mA
output signal to  intensity based on the calibration set by NeoTech.

The duty sensor was compared against two reference sensors to demonstrate that the duty sensor
was within 10% of the average of the two reference sensors.  This evaluation was conducted
before and after the validation test runs, using the procedure described in Section 3.2. Table 4-1
presents the results of the sensor assessment.  These data  demonstrate that the duty  sensor was
within  10% of  the average of the two  reference sensors.  The two reference  sensors showed a
variance of 2.9% at 100% power and 1.3% and 6.8% at reduced power.
                                          28

-------
                           Table 4.1. Sensor Assessment Data
Sensor
Reference #1
2012-08
Reference #2
2012-09
Average of
Reference Sensor
Duty Sensor
2012-10
Deviation of Duty
Sensor from
Reference average
Intensity at
100% Power
Before testing
90.5% UVT
(mW/cm2)
8.6
9.1
8.85
8.2
7.3
Intensity at
100% Power
After testing
98.2% UVT
(mW/cm2)
17.0
18.0
17.5
18.4
5.1
Intensity with
Power Reduced
Before testing
90.5% UVT
(mW/cm2)
3.4
3.9
3.65
3.9
6.8
Intensity with
Power Reduced
After testing
98.2% UVT
(mW/cm2)
7.8
8.0
7.9
8.6
8.8
4.3    Collimated Beam Dose Response Data
Collimated Beam dose response data were generated for each of the test  days for MS2 in
accordance with the procedures described in Section 3.5. On test Day 1, collimated beam tests
for MS2 were run on the minimum UVT water (91%). On test Day 2, collimated beam tests were
run on maximum UVT water (97%). All collimated beam tests were performed in duplicate.

UV doses for the MS2 tests covered the range  of the targeted RED dose, which in this case was
40mJ/cm2. UV doses were set at 0, 20, 30, 40, 60 and 80 mJ/cm2.

The collimated beam samples  were collected directly from the test rig during the normal testing
runs. Two 1 L bottles of the  seeded influent water (MS2 was injected into the influent  water
during a test run) were collected to provide the two samples for duplicate analyses. Using this
approach,  the dose response data reflect the identical conditions to the biodosimetric flow tests
for sample matrix, UVT, and MS2 concentration.  The collimated beam samples were irradiated
on the same day  as sample collection, and were plated  in triplicate along with the flow test
samples. Therefore, analytical  conditions for the dose response data were also identical to those
for the flow test samples.

The collimated beam results  for MS2 are presented in Tables  4-2 and 4-3. These  data were
calculated as the average of the three individual results obtained at each dose level.

4.4    Development of Dose Response
The development of the UV dose response curves for MS2 for use with flow test data to establish
the REDmeas is a three step process.
                                          29

-------
    1.  For each MS2 collimated beam test and its replicate for each day of testing, the log N
       (PFU/mL) was plotted vs.  UV dose (mJ/cm2). Figures 4-1 and 4-2 show the curves for
       each day.

    2.  A separate equation (second order polynomial) was developed for each UVT condition
       (low and high). In this test there were two days of testing for MS2, so there were two sets
       of data. 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(No/N)

                Where:
                    No = The common N0 identified in Step 1 (PFU/mL); and
                    N = Concentration of challenge microorganisms in the petri dish after
                    exposure to UV light (PFU/mL).

Tables 4-2 and 4-3 show the calculated values for MS2 log inactivation.

Finally, the UV dose as a function of log I was plotted for each day of the testing. Figures 4-3
through 4-4 show the  curves for dose as a function of log  I.  Using regression analysis,  an
equation was derived that best fit the data, forcing the fit through the origin.  The equation used
for the dose response curve 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 sample. REDmeas calculations and  full scale data is presented in the Section
4.5.

A Grubb's test was run to determine if any replicates should be omitted from the development of
the dose response curve. The Grubb's test results show that there was one outlier for replicate 2
for the high UVT water. The calculations were rerun after removing the outlier and that data was
used for the subsequent determination of dose response and the RED calculations.  Table 4-4
shows the  data  with the outlier removed.  Both sets  of data are presented for informational
purposes. There is very little difference in the dose response curve coefficients between the data
set with the outlier included when compared to the one with the outlier removed.

A summary of  the statistics for  uncertainty  for the  collimated beam dose response data is
presented at the end of Tables 4-2, 4-3 and 4-4. The uncertainty (UDR) of the collimated beam
test was slightly higher than 30% at 1 log inactivation for the low UVT water (32.58%). The UDR
for the high UVT water  was 12.25% once the outlier was removed (19.35% with the outlier
included). At 2-log inactivation (dose  of approximately 40 mJ/cm2 RED) the UDR was 15.31%
and 5.53%.  The uncertainty of the collimated beam  results  for Day 1 is  greater than 30%.
Therefore, the uncertainty calculations for the  Cryptosporidium and Giardia log inactivation
calculations presented in Sections 4.7 and 4.8 include the highest UDR value (32.58%) applied to


                                           30

-------
all of the validation factors for the set points. The highest UDR is also used for calculating the
validation factor for calculating the REDVai based on MS2 as the target organism, as shown in
Section 4.9.

Figures 4-3 through 4-4 show the results of the UDR calculations plotted on the dose response
curve.  Also shown in Figures 4-3 and 4-4 are the QC limits  for MS2 taken from the UVDGM-
2006. The results show that the MS2 dose response curves are within the boundaries established
forMS2.

The  polynomial  equation  coefficients  for each day  (Aug 2 and 3,  2012) were evaluated
statistically to determine which terms were statistically significant based on the P factor. All
coefficients were found to be significant except for the x term for the August 2, 2012 day of
testing. The P factor was 0.0539, just above the 0.05 significance test cutoff. When P is greater
than 0.05,  it indicates the dose-response relationship could  be linear rather than  second order.
Both linear and polynomial regressions were evaluated for this data set. The polynomial equation
was selected for the REDmeas calculations because the R2 coefficient was slightly better (0.996
versus 0.994), the standard error was slightly less (3.356 versus  3.755) and the other set of
collimated  beam  data was a second  order polynomial  equation. Also the polynomial equation
gave the  more conservative (lower) results for the REDmeas. If the linear relationship is used, the
REDmeas  is approximately 6% higher than REDmeas calculated using the second order polynomial
equation.

-------
Table 4-2. UV Dose - Response Data from Collimated Beam Tests at 91% UVT
UVT
(%)
90.6
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
20.82
31.25
41.54
62.27
82.56
0.00
20.73
31.12
41.37
62.05
82.74
UV
Dose2
0
433
977
1726
3878
6816
0
430
968
1711
3850
6846
Avg (PFU/mL)
1,820,000
140,000
80,300
18,200
6,970
707
2,190,000
146,000
84,700
23,600
7,030
873
Avg Log(PFU/mL)
6.26
5.15
4.90
4.26
3.84
2.85
6.34
5.16
4.93
4.37
3.85
2.94
Log I
-0.02
1.09
1.33
1.98
2.40
3.39
-0.10
1.07
1.31
1.87
2.39
3.30
Log I2
0.000
1.194
1.780
3.916
5.739
11.488
0.010
1.155
1.719
3.482
5.721
10.876
Predicted
Dose
-0.43
24.05
29.81
46.00
57.09
85.49
-2.07
23.62
29.25
43.09
56.99
82.75
Avg:
SD:
n:
P:
t (95%):
Residual
(mJ/cm2)
0.4
-3.2
1.4
-4.5
5.2
-2.9
2.1
-2.9
1.9
-1.7
5.1
0.0
0.07
3.20
12
0.05
2.228
G
0.1
1.0
0.4
1.4
1.6
0.9
0.6
0.9
0.6
0.6
1.6
0.0
Outlier?
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK

                                                         Grubb's Statistic
                                   32

-------
                      Table 4-2. (continued)
                  Uncertainty of Dose-Response (UDR)
Log I
0.001
0.25
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
3.39
Dose
0.0
5.2
10.6
21.9
33.9
46.6
59.9
74.0
88.8
104.3
85.5
t

2.23
2.23
2.23
2.23
2.23
2.23
2.23
2.23
2.23
2.23
SD

3.20
3.20
3.20
3.20
3.20
3.20
3.20
3.20
3.20
3.20
UDR (%)

136.89
67.31
32.58
21.05
15.31
11.89
9.63
8.03
6.83
8.34
DL
(mJ/cm2/Log T)
20.48
20.83
21.18
21.88
22.58
23.28
23.98
24.68
25.38
26.08
25.22
t - student t test factor SD - standard deviation
Regression Statistics
Multiple R
R Square
Adjusted R Square
Standard Error
Observations
0.99796
0.995924
0.895516
3.356386
12
ANOVA
Regression
Residual
Total
df
2
10
12
ss
27522.43
112.6533
27635.09
MS
13761.22
11.26533

F
1221.555


Significance F
1.0994 1E-11




Intercept
X Variable 1
X Variable 2
Coefficients
0
20.48157
1.398414
Standard Error

1.751277
0.640225
tStat

11.695219
2.1842525
P-value

3.72E-07
0.053857
Lower 95%

16.57948
-0.0281
Upper 95%

24.38365
2.824924
Lower
95.0%

16.57948
-0.0281
Upper 95.0%

24.38365272
2.824924411
                                33

-------
Table 4-3. UV Dose - Response Data from Collimated Beam Tests at 97% UVT
UVT
(%)
97.7
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
20.79
31.09
41.34
61.63
81.83
0.00
20.67
30.85
41.05
61.40
81.78
UV
Dose2
0
432
967
1709
3798
6696
0
427
952
1685
3770
6688
Avg
(PFU/mL
)
3,770,000
173,000
74,300
19,900
3,070
577
3,470,000
128,000
70,000
20,100
2,980
500
Avg
Log(PFU/mL
)
6.58
5.24
4.87
4.30
3.49
2.76
6.54
5.11
4.85
4.30
3.47
2.70
Log I
-0.06
1.28
1.64
2.21
3.03
3.75
-0.03
1.41
1.67
2.21
3.04
3.81
Log I2
0.004
1.626
2.696
4.903
9.156
14.077
0.001
1.976
2.782
4.884
9.235
14.547
Predicte
dDose
-0.90
21.49
28.83
41.31
61.17
81.07
-0.39
24.05
29.37
41.21
61.50
82.87
Avg:
SD:
n:
P:
t (95%):
Grubb's Test for O
P:
t (90%):
Grubb's Statistic
(GCRIT):
Residual
(mJ/cm2)
0.9
-0.7
2.3
0.0
0.5
0.8
0.4
-3.4
1.5
-0.2
-0.1
-1.1
0.07
1.42
12
0.05
2.228
utliers
0.10
3.691
2.412
G
0.6
0.5
1.5
0.0
0.3
0.5
0.2
2.4
1.0
0.2
0.1
0.8
Outlier?
OK
OK
OK
OK
OK
OK
OK
OUTLIER
OK
OK
OK
OK

                                 34

-------
                  Table 4-3. (continued)
              Uncertainty of Dose-Response (UDR)
Regression Statistics
Multiple R
R Square
Adjusted R
Square
Standard Error
Observations
0.999591
0.999183
0.899101
1.488914
12
Log I
0.001
0.25
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
3.81
Dose
0.0
3.7
7.7
16.3
25.9
36.5
48.0
60.5
73.9
88.3
82.9
t

2.23
2.23
2.23
2.23
2.23
2.23
2.23
2.23
2.23
2.23
SD

1.42
1.42
1.42
1.42
1.42
1.42
1.42
1.42
1.42
1.42
UDR (%)

84.88
41.11
19.35
12.18
8.66
6.58
5.22
4.27
3.58
3.81
DL (mJ/cm2/Log
I)
14.41
14.89
15.37
16.33
17.28
18.24
19.20
20.16
21.12
22.08
21.73
t - student t test factor    SD - standard deviation
ANOVA
Regression
Residual
Total
df
2
10
12
ss
27102.05
22.16864
27124.22
MS
13551.02
2.216864

F
6112.7


Significance F
7.94E-15



Intercept
X Variable 1
X Variable 2
Coefficients
0
14.40581
1.919398
Standard Error

0.708567
0.22588
tStat

20.33092
8.497439
P-value

1.83E-09
6.92E-06
Lower 95%

12.82703
1.416107
Upper 95%

15.9846
2.422689
Lower 95.0%

12.82703
1.416107
Upper 95.0%

15.9846
2.422689
                             35

-------
Table 4-4. UV Dose - Response Data from Collimated Beam Tests at 97% UVT with outlier removed
UVT
(%)
97.7
Rep
1
2
Target
UV Dose
(mJ/cm2)
0
20
30
40
60
80
0
30
40
60
80

Actual
UV
Dose
0.00
20.79
31.09
41.34
61.63
81.83
0.00
30.85
41.05
61.40
81.78

UV
Dose2
0
432
967
1709
3798
6696
0
952
1685
3770
6688

Avg
(PFU/mL
)
3,770,000
173,000
74,300
19,900
3,070
577
3,470,000
70,000
20,100
2,980
500

Avg
Log(PFU/mL
)
6.58
5.24
4.87
4.30
3.49
2.76
6.54
4.85
4.30
3.47
2.70

Log I
-0.03
1.30
1.67
2.24
3.06
3.78
0.00
1.70
2.24
3.07
3.84

Log I2
0.001
1.702
2.794
5.034
9.335
14.298
0.000
2.881
5.014
9.414
14.773

Predicte
d Dose
-0.49
22.16
29.48
41.85
61.39
80.84
0.03
30.02
41.75
61.72
82.59

Avg:
SD:
n:
P:
t (95%):
Grubb's Test for O
P:
t (90%):
Grubb's Statistic
(GCRIT):
Residual
(mJ/cm2)
0.5
-1.4
1.6
-0.5
0.2
1.0
0.0
0.8
-0.7
-0.3
-0.8

0.04
0.89
11
0.05
2.262
utliers
0.10
3.751
2.355
G
0.5
1.6
1.8
0.6
0.2
1.1
0.1
0.9
0.8
0.4
1.0

Outlier?
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK
OK


                                          36

-------
                      Table 4-4. (continued)
                  Uncertainty of Dose-Response (UDR)
Log I
0.001
0.25
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
3.84
Dose
0.0
3.8
7.8
16.4
26.0
36.4
47.8
60.0
73.1
87.1
82.6
t

2.26
2.26
2.26
2.26
2.26
2.26
2.26
2.26
2.26
2.26
SD

0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
0.89
UDR (%)

53.32
25.90
12.25
7.75
5.53
4.22
3.36
2.76
2.31
2.44
DL
(mJ/cm2/Log I)
14.68
15.12
15.56
16.45
17.33
18.22
19.11
19.99
20.88
21.77
21.49
t - student t test factor   SD - standard deviation
Regression Statistics
Multiple R
R Square
Adjusted R Square
Standard Error
Observations
0.999851
0.999702
0.888558
0.940025
11
ANOVA
Regression
Residual
Total
df
2
9
11
SS
26689.01
7.952815
26696.97
MS
13344.51
0.883646

F
15101.64


Significance F
4.91681E-15



Intercept
X Variable 1
X Variable 2
Coefficients
0
14.67514
1.772698
Standard Error

0.480973
0.150326
tStat

30.511383
11.792378
P-value

2.14E-10
8.93E-07
Lower 95%

13.5871
1.432638
Upper 95%

15.76317
2.112759
Lower 95%

13.5871
1.432638
Upper 95%

15.76317461
2.112758854
                                37

-------
                        Log N as a function of UV Dose
                                             y = 0.0001 x2-0.048x + 6.2389
                                                     R2 = 0.9871
z
o
      0      10      20      30      40      50      60     70      80      90
                                 UV Dose (mJ/cm2)

 Figure 4-1. MS2 Collimated Beam Dose versus Log N Day 1 08-2-2012 UVT 91%.
                                    38

-------
                           Log N as a function of UV Dose
   O)
   o
                                              y = 0.0002x2-0.0615x +6.5425
                                                      R2 = 0.9988
               10      20      30      40      50      60      70      80
                                     UV Dose (mJ/cm2)
90
Figure 4-2. MS2 Collimated Beam Dose versus Log N Day 2 08-03-2012 97% UVT (outlier
                                    removed).
                                       39

-------
                              Dose-Response Curve
                                       MS2
   100
    so -
    60
IN
E
{j
-^
_§

o>
a   40
    20
                   1.3984x2 + 20.482x
150
120
90
60
30
       0123456

                                 Log Inactivation


          CB Data	UVDGM QC Limits  	Udr(%)   	Dose-Response Curve



     Figure 4-3. Dose response - log I versus dose -UVT 91% Day 1 08-02-2012.
                                    40

-------
     100
      80 -
                                Dose-Response Curve
                                        MS2
                 = 1.7727x2 + 14.675x
150
120
                                                                            30
                    123456
                                   Log Inactivation

            CB Data	UVDGM QC Limits   	Udr{%)   	Dose-Response Curve
Figure 4-4. Dose response - log I versus dose -UVT 97% Day 2 08-03-2012 with outlier
removed.
                                      41

-------
4.5    MS and Operational Flow Test Data
The operational data for each test run (flow rate, UVT, and UV sensor intensity measurements)
are presented in Table 4-5. 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 were taken simultaneously beginning after approximately 2 to 3 minutes of steady state
operation. Subsequent  influent and effluent  samples  were collected simultaneously after  an
additional 2 to 3  minutes of operation, yielding  five sets of samples over a 10 to 12 minute
period. The MS2 concentrations measured during the flow tests are presented in Table 4-6.

For each test condition replicate (i.e.,  each of the three influent and effluent samples), the log
inactivation (log I) was calculated using the following equation:

                                   log I = log (No /N)
       Where:
             No = challenge microorganism concentration in influent sample (PFU/mL); and
             N =  Challenge microorganism concentration in corresponding effluent sample
              (PFU/mL).

Table 4-7 shows the log concentration for the influent (N0) and effluent (N) samples. Table 4-8
presents the log I values for each sample.

The calculated REDmeas results in ml/cm2 are shown in Table 4-9.  For each test condition the
REDmeas for each replicate was determined using the measured log inactivation (log I) and the
collimated beam test UV dose-response curve. The collimated beam regression line, based on the
replicate collimated  beam tests for each day of testing, was used for this calculation. Figures 4-3
and 4-4 present the dose response  curves  and the equations that  were used to calculate the
The replicate REDmeas values were averaged to produce one REDmeas for each test condition and
it's duplicate. All of the flow tests at 150, 250, and 435 gpm, with feed water at 91%, 94%, and
97% UVT or the equivalent reduced power respectively, achieved a  minimum REDmeas of 40
ml/cm .
                                           42

-------
Table 4-5.  NeoTech D438™ Flow Test Operational Data
Test Condition
Lowered UVT Full Power
Lowered UVT-Full Power Dup
High UVT - Lowered Power
High UVT - Lowered Power Dup
Lowered UVT Full Power
Lowered UVT-Full Power Dup
High UVT - Lowered Power
High UVT - Lowered Power Dup
Lowered UVT Full Power
Lowered UVT-Full Power Dup
High UVT - Lowered Power
High UVT - Lowered Power Dup
Test
Day
1
1
2
2
1
1
2
2
1
1
2
2
Run
2
o
5
10
11
4
5
12
13
6
7
14
15
UVT
(%)
91
91
98
98
94
94
98
98
97
96
98
98
Flow
(gpm)
151
151
155
154
251
251
252
252
436
436
434
435
Intensity
(mW/cm2)
7.9
7.9
7.9
7.9
10.0
10.0
10.4
10.4
13.0
13.0
13.1
13.2
                        43

-------
                              Table 4-6. Flow Tests MS2 Concentrations NeoTech D438™
Test Condition
Lowered UVT 91%
Full Power
Lowered UVT 91%
Full Power Dup
High UVT - Lowered
Power
High UVT - Lowered
Power Dup
Lowered UVT 94%
Full Power
Lowered UVT 94%
Full Power Dup
High UVT -Lowered
Power
High UVT -Lowered
Power Dup
Lowered UVT 97%
Full Power
Lowered UVT 97%
Full Power Dup
High UVT -Lowered
Power
High UVT -Lowered
Power Dup
Run
2
o
J
10
11
4
5
12
13
6
7
14
15
Influent (PFU/mL)
Repl
3.53E+05
8.40E+05
2.36E+06
1.07E+06
1.13E+06
1.75E+06
9.57E+05
1.81E+06
3.47E+06
4.37E+06
3.18E+06
2.97E+06
Rep 2
4.53E+05
5.60E+05
1.18E+06
1.55E+06
8.73E+05
1.59E+06
9.60E+05
1.79E+06
1.56E+06
5.60E+06
2.05E+06
2.90E+06
Rep 3
4.27E+05
7.90E+05
1.33E+06
9.10E+05
NA
5.80E+05
2.26E+06
2.24E+06
1.98E+06
2.00E+06
1.94E+06
3.81E+06
Rep 4
4.27E+05
7.37E+05
1.15E+06
9.00E+05
8.97E+05
6.20E+05
1.59E+06
2.56E+06
2.00E+06
1.78E+06
1.57E+06
2.86E+06
Rep 5
3.17E+05
8.53E+05
8.57E+05
8.07E+05
8.17E+05
1.04E+06
2.15E+06
2.04E+06
2.23E+06
2.31E+06
2.49E+06
3.38E+06
Effluent (PFU/mL)
Repl
1.24E+03
4.37E+03
1.40E+03
6.47E+02
3.88E+03
4.20E+03
6.27E+03
3.37E+03
6.33E+03
2.28E+04
9.43E+03
1.08E+04
Rep 2
1.12E+03
3.69E+03
1.07E+03
6.23E+02
3.72E+03
3.92E+03
4.83E+03
4.10E+03
6.17E+03
1.18E+04
1.25E+04
1.50E+04
Rep 3
1.10E+03
4.22E+03
1.19E+03
8.13E+02
NA
4.00E+03
5.30E+03
4.79E+03
5.33E+03
1.42E+04
5.60E+03
7.30E+03
Rep 4
8.87E+02
3.08E+03
1.51E+03
8.67E+02
3.76E+03
4.33E+03
6.97E+03
3.13E+03
1.02E+04
1.63E+04
6.40E+03
1.17E+04
Rep 5
7.80E+02
3.11E+03
1.31E+03
1.57E+03
4.64E+03
4.43E+03
7.50E+03
5.37E+03
7.50E+03
1.53E+04
1.13E+04
7.40E+03
NA- samples not analyzed
                                                         44

-------
                  Table 4-7. MS2 Log Concentration for Influent and Effluent Samples NeoTech D438™
Test Condition
Lowered UVT 91% Full Power
Lowered UVT 91% Full Power Dup
High UVT - Lowered Power
High UVT - Lowered Power Dup
Lowered UVT 94% Full Power
Lowered UVT 94% Full Power Dup
High UVT - Lowered Power
High UVT - Lowered Power Dup
Lowered UVT 97% Full Power
Lowered UVT 97% Full Power Dup
High UVT - Lowered Power
High UVT - Lowered Power Dup
Run
2
o
J
10
11
4
5
12
13
6
7
14
15
Log Influent Concentration
Repl
5.55
5.92
6.37
6.03
6.05
6.24
5.98
6.26
6.54
6.64
6.50
6.47
Rep 2
5.66
5.75
6.07
6.19
5.94
6.20
5.98
6.25
6.19
6.75
6.31
6.46
Rep 3
5.63
5.90
6.12
5.96
NA
5.76
6.35
6.35
6.30
6.30
6.29
6.58
Rep 4
5.63
5.87
6.06
5.95
5.95
5.79
6.20
6.41
6.30
6.25
6.20
6.46
Rep 5
5.50
5.93
5.93
5.91
5.91
6.02
6.33
6.31
6.35
6.36
6.40
6.53
Log Effluent Concentration
Repl
3.09
3.64
3.15
2.81
3.59
3.62
3.80
3.53
3.80
4.36
3.97
4.03
Rep 2
3.05
3.57
3.03
2.79
3.57
3.59
3.68
3.61
3.79
4.07
4.10
4.18
Rep 3
3.04
3.63
3.08
2.91
NA
3.60
3.72
3.68
3.73
4.15
3.75
3.86
Rep 4
2.95
3.49
3.18
2.94
3.58
3.64
3.84
3.50
4.01
4.21
3.81
4.07
Rep 5
2.89
3.49
3.12
3.20
3.67
3.65
3.88
3.73
3.88
4.18
4.05
3.87
NA- samples not analyzed
                                                       45

-------
                              Table 4-8. MS2 Log Inactivation Results NeoTech D438™
Test Condition
Lowered UVT 91% Full Power
Lowered UVT 91% Full Power Dup
High UVT - Lowered Power
High UVT - Lowered Power Dup

Lowered UVT 94% Full Power
Lowered UVT 94% Full Power Dup
High UVT - Lowered Power
High UVT - Lowered Power Dup

Lowered UVT 97% Full Power
Lowered UVT 97% Full Power Dup
High UVT - Lowered Power
High UVT - Lowered Power Dup
Test
Run#
2
3
10
11

4
5
12
13

6
7
14
15
Test
Day
1
1
2
2

1
1
2
2

1
1
2
2
UVT
(%)
90.6
90.6
98.1
98.1

93.9
93.9
98.2
98.2

96.5
96.4
98.1
98.2
Flow
(gpm)
151.1
151.3
154.8
154.1

251.4
250.8
251.8
251.6

436.0
435.5
434.2
435.8
Intensity
(mW/cm2)
7.9
7.9
7.9
7.9

10.0
10.0
10.4
10.4

13.0
13.0
13.1
13.2
Log I
Repl
2.45
2.28
3.23
3.22

2.46
2.62
2.18
2.73

2.74
2.28
2.53
2.44
Rep 2
2.61
2.18
3.04
3.40

2.37
2.61
2.30
2.64

2.40
2.68
2.21
2.29
Rep 3
2.59
2.27
3.05
3.05

NA
2.16
2.63
2.67

2.57
2.15
2.54
2.72
Rep 4
2.68
2.38
2.88
3.02

2.38
2.16
2.36
2.91

2.29
2.04
2.39
2.39
Rep 5
2.61
2.44
2.82
2.71

2.25
2.37
2.46
2.58

2.47
2.18
2.34
2.66
NA- samples not analyzed
                                                        46

-------
                                    Table 4-9. MS2 RED Results NeoTech D438™
Test Condition
Lowered UVT 91%
Lowered UVT 91% Dup
Lowered Power
Lowered Power Dup

Lowered UVT 94%
Lowered UVT 94% Dup
Lowered Power
Lowered Power Dup

Lowered UVT 97%
Lowered UVT 97% Dup
Lowered Power
Lowered Power Dup
Test
Run
2
3
10
11

4
5
12
13

6
7
14
15
Test
Day
1
1
2
2

1
1
2
2

1
1
2
2
UVT
(%)
90.6
90.6
98.1
98.1

93.9
93.9
98.2
98.2

96.5
96.4
98.1
98.2
Flow
(gpm)
151
151
155
154

251
251
252
252

436
436
434
436
Intensity
(mW/cm2)
7.9
7.9
7.9
7.9

10.0
10.0
10.4
10.4

13.0
13.0
13.1
13.2
RED (mJ/cm2)
Repl
58.69
54.07
65.81
65.59

58.96
63.26
40.50
53.28

66.59
54.04
48.43
46.35
Rep 2
62.90
51.33
61.06
70.28

56.41
62.93
43.09
51.10

57.29
64.83
41.20
42.82
Rep 3
62.40
53.76
61.21
61.22

NA
50.80
50.85
51.82

61.87
50.47
48.70
52.97
Rep 4
65.01
56.64
57.01
60.39

56.60
50.66
44.46
57.78

54.30
47.56
45.19
45.16
Rep 5
62.96
58.25
55.37
52.81

53.05
56.41
46.77
49.65

59.21
51.27
44.12
51.57
Average
62.39
54.81
60.09
62.06

56.26
56.81
45.13
52.73

59.85
53.63
45.53
47.77
SD(RED)
2.30
2.69
4.08
6.50

2.43
6.19
3.92
3.11

4.67
6.67
3.13
4.33
USP
10.21
13.63
18.85
29.08

13.76
30.23
24.11
16.40

21.66
34.54
19.12
25.14
NA- not analyzed
                                                        47

-------
4.6   Set Line for REDmeas 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 4-5 shows
the set line.  The unit is validated for a minimum REDmeas of 40 mJ/cm2 for any flow rate -
intensity combination above and to the left of the set line. The maximum flow rate demonstrated
was 434 gpm. A UV system cannot operate above the highest validated flow rate and claim a 40
mJ/cm2 REDmeas. The lowest intensity demonstrating a RED of 40 mJ/cm2 was 7.9 mW/cm2. A
UV system cannot operate below the lowest validated irradiance and claim a REDmeas of 40
mJ/cm .

Set Point 1-151 gpm;  7.9 mW/cm2
Set Point 2-251 gpm; 10.4 mW/cm2
Set Point 3 - 434 gpm; 13.2 mW/cm2
         18
       O
       !/>
          2

          0
                  50     100    150   200    250    300
                                     Flow Rate (gpm)
350
400
450
500
             Figure 4-5. Set Line for 40 mJ/cm  REDmeas - NeoTech D438™.
                                        48

-------
4.7    Deriving the Validation Factor and Log Inactivation 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.

                               W = 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
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:
                                         9             _
                         Sensitivity (ml/cm per log I) = RED/ Log I

Per the GP-2011 and UVDGM-2006, the sensitivity is calculated for each test replicate (five per
test run, 20 samples total per set point). The highest BRED value found among the replicates at a
given set point is then selected for the BRED value for use in the VF calculation per the UVDGM-
2006 requirement. Table 4-10  shows the data for the replicates at each set point. The highest

                                           49

-------
RED bias at each set point is used in the validation factor calculations shown later in Section
4.7.4.
           Table 4-10. RED Bias Factor for Each Set Point for Cryptosporidium
Sample
Number
2-1
2-2
2-3
2-4
2-5
3-1
3-2
3-3
3-4
3-5
4-1
4-2
4-3
4-4
4-5
5-1
5-2
5-3
5-4
5-5
6-1
6-2
6-3
6-4
6-5
7-1
7-2
7-3
7-4
7-5
10-1
10-1
10-1
10-1
10-1
11-1
11-2
11-3
11-4
11-5
12-1
12-2
Test
Run
2
2
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
6
6
6
6
6
7
7
7
7
7
10
10
10
10
10
11
11
11
11
11
12
12
UVT
%
90.6
90.6
90.6
90.6
90.6
90.6
90.6
90.6
90.6
90.6
93.9
93.9
93.9
93.9
93.9
93.9
93.9
93.9
93.9
93.9
96.5
96.5
96.5
96.5
96.5
96.4
96.4
96.4
96.4
96.4
98.1
98.1
98.1
98.1
98.1
98.1
98.1
98.1
98.1
98.1
98.2
98.2
Sensitivity mJ/cm2 per Log I
RED
58.69
62.90
62.40
65.01
62.96
54.07
51.33
53.76
56.64
58.25
58.96
56.41
NA
56.60
53.05
63.26
62.93
50.80
50.66
56.41
66.59
57.29
61.87
54.30
59.21
54.04
64.83
50.47
47.56
51.27
65.81
61.06
61.21
57.01
55.37
65.59
70.28
61.22
60.39
52.81
40.50
43.09
Log I
2.45
2.61
2.59
2.68
2.61
2.28
2.18
2.27
2.38
2.44
2.46
2.37
NA
2.38
2.25
2.62
2.61
2.16
2.16
2.37
2.74
2.40
2.57
2.29
2.47
2.28
2.68
2.15
2.04
2.18
3.23
3.04
3.05
2.88
2.82
3.22
3.40
3.05
3.02
2.71
2.18
2.30
Sensitivity
23.91
24.13
24.10
24.23
24.13
23.68
23.53
23.66
23.81
23.89
23.93
23.80
NA
23.81
23.62
24.15
24.13
23.50
23.50
23.80
24.31
23.84
24.08
23.69
23.94
23.67
24.22
23.49
23.33
23.53
20.40
20.07
20.08
19.78
19.67
20.38
20.69
20.08
20.02
19.48
18.55
18.75
BRED 4-log
Crypto
1.66
1.70
1.70
1.70
1.70
1.66
1.66
1.66
1.66
1.66
1.66
1.66
NA
1.66
1.66
1.70
1.70
1.66
1.66
1.66
1.40
1.38
1.40
1.38
1.38
1.38
1.40
1.38
1.38
1.38
1.18
1.18
1.18
1.17
1.17
1.18
1.18
1.18
1.18
1.17
1.17
1.17
BRED 3.5-log
Crypto
1.79
1.83
1.83
1.83
1.83
1.79
1.79
1.79
1.79
1.79
1.79
1.79
NA
1.79
1.79
1.83
1.83
1.79
1.79
1.79
1.43
1.42
1.43
1.42
1.42
1.42
1.43
1.42
1.42
1.42
1.20
1.20
1.20
1.19
1.19
1.20
1.20
1.20
1.20
1.19
1.19
1.19
BRED 3.0 log
Crypto
1.82
1.85
1.85
1.85
1.85
1.82
1.82
1.82
1.82
1.82
1.82
1.82
NA
1.82
1.82
1.85
1.85
1.82
1.82
1.82
1.42
1.41
1.42
1.41
1.41
1.41
1.42
1.41
1.41
1.41
1.20
1.20
1.20
1.19
1.19
1.20
1.20
1.20
1.20
1.19
1.19
1.19
                                           50

-------
Sample
Number
12-3
12-4
12-5
13-1
13-2
13-3
13-4
13-5
14-1
14-2
14-3
14-4
14-5
15-1
15-2
15-3
15-4
15-5

Test
Run
12
12
12
13
13
13
13
13
14
14
14
14
14
15
15
15
15
15

Maximum BRED
UVT
%
98.2
98.2
98.2
98.2
98.2
98.2
98.2
98.2
98.1
98.1
98.1
98.1
98.1
98.2
98.2
98.2
98.2
98.2

Sensitivity mJ/cm2 per Log I
RED
50.85
44.46
46.77
53.28
51.10
51.82
57.78
49.65
48.43
41.20
48.70
45.19
44.12
46.35
42.82
52.97
45.16
51.57

Log I
2.63
2.36
2.46
2.73
2.64
2.67
2.91
2.58
2.53
2.21
2.54
2.39
2.34
2.44
2.29
2.72
2.39
2.66

Sensitivity
19.34
18.86
19.03
19.51
19.36
19.41
19.84
19.25
19.16
18.60
19.18
18.91
18.83
19.00
18.73
19.49
18.91
19.39

Set Point 151 gpm - 7.9 mW/cm2
Set Point 251 gpm - 10.4mW/cm2
Set Point 434 gpm - 13.2 mW/cm2
BRED 4-log
Crypto
1.17
1.17
1.17
1.17
1.17
1.17
1.17
1.17
1.17
1.17
1.17
1.17
1.17
1.17
1.17
1.17
1.17
1.17

1.70
1.70
1.40
BRED 3.5-log
Crypto
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19

1.83
1.83
1.43
BRED 3.0 log
Crypto
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19

1.85
1.85
1.42
NA - sample not analyzed so RED and bias not determined

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.  Figure 5.4 of the UVDGM-2006, which is specific to a UV intensity set point
approach, was used to determine Uvai in calculating the validated dose.  Per Figure 5.4 in EPA's
UVDGM-2006, any of the following equations may be used to establish the Uvai:
                           Uvai = (USP2 +UDR2)1/2
                                               TT
                                               UDR )
              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; and
                    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. If this objective is met, then it eliminates
                                            51

-------
the need to calculate the Us factor per the GP-2011 and UVDGM-2006,  Section 5.4.4.  The
sensor met the 10% requirement, as shown in Table 4-1, therefore, Us is not used in determining
the uncertainty of validation.

The UVDGM-2006 also shows the formula and calculations for UDR in Appendix C Section C4.

The equation is:

                    UDR = 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; and
             t = t-statistic at a 95% confidence level for a sample size equal to the number of
             test condition replicates used to define the dose-response.


The UDR results are shown in Tables 4-2 and 4-4 for the low and high UVT waters (32.58% and
12.25%,  respectively).  Since the UDR was  > 30% at the UV  dose  corresponding to  1-log
inactivation of the challenge organism, the uncertainty of the dose response (UDR) is included in
the calculation of uncertainty.

The uncertainty in the set point value (Usp) is based on a prediction interval at a 95% confidence
level using the following procedure:

       1. Calculate the average and standard deviation of REDmeas values for each test condition

       2. Calculate the uncertainty of the set point REDmeas using:

                          USP = [(t x SDRED) / (REDmeas)] x  100%

          Where:
             REDmeaS = Average REDmeas value measured for each test condition;
             SDRED =  Standard  deviation of the REDmeas values measured  for  each test
             condition;  and
             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 run 4 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.
                                           52

-------
The USP results based on the REDmeas and standard deviation are shown in Table 4-9. The highest
USP for each set point is 29.08% (151 gpm set point), 30.23% (251 gpm set point), and 34.54%
(434 gpm set point).

The  uncertainty of the validation  is then calculated using the highest UDR (32.58%) from the
2012 dose response data and the highest USP for each test condition using the equation:

                           Uval = (USP2 +UDR2)1/2

Table 4-11 shows the  Uyai values used for determining the uncertainty of the validation at each
set point.

   Table 4-11. Uncertainty of the Validation (Uyai) and BRED Values for Cryptosporidium
Set Point
151 gpm- 7.9mW/cm2
251 gpm- 10.4mW/cm2
434gpm-13.2mW/cm2
Max
UDR
(%)
32.58
32.58
32.58
Max
USP
(%)
29.08
30.23
34.54
UVa,
(%)
43.67
44.44
47.48
Max
BRED
4.0-log
1.70
1.70
1.40
3.5-log
1.83
1.83
1.43
3.0-log
1.85
1.85
1.42
4.7.4   Validated Dose and Set Line for Cryptosporidium
After establishing the Uyai and the RED bias as described above, the VF is calculated using the
equation:
                              W = BRED X[l+(UVai/100)]
       Where:
             VF = Validation Factor;
             BRED = RED bias factor for Cryptosporidium: and
             Uvai = Uncertainty of validation expressed as a percentage.

The validated dose is then calculated as follows:

                         Validated dose (REDVai) = REDmeas / VF

Table 4-12 shows  the calculated VF for various Cryptosporidium log inactivation levels (3.0,
3.5, and 4.0 log inactivation).

Table 4-12 shows the REDyai for Cryptosporidium for each test run using the validation factors
for the various Cryptosporidium log inactivation levels. Table 4-12 also 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 first and third set points (151 and 434 gpm)
show a  validated dose  for  Cryptosporidium that would  achieve  a minimum  of 4.0 log
inactivation.  The  second  set point (middle point) at 251 gpm showed a minimum of 3.5 log
inactivation.
                                           53

-------
The NeoTech D438™ achieved a minimum of 3.5 log inactivation for the low power runs at 251
gpm at 10.4 mW/cm2 and a 4.0-log inactivation for all of the test runs at the set points at 151
gpm - 7.9 mW/cm2 and 434 gpm - 13.2 mW/cm2.   Figure 4-6 shows the operating conditions
based on three set points that achieved the 3-log inactivation oi Cryptosporidium.

The NeoTech D438™ achieved a minimum 3.0-log inactivation for Cryptosporidium over the
range of flow (151 to 434 gpm) and intensity (7.9 to 13.2 mW/cm2) as shown in the set line in
Figure 4-14.

The set points in Figure 4-6 are:

Set Point 1-151 gpm;  7.9 mW/cm2
Set Point 2-251 gpm; 10.4 mW/cm2
Set Point 3 - 434 gpm; 13.2 mW/cm2
            0     50    100    150    200   250    300   350    400   450    500
                                      Flow Rate (gpm)
Figure 4-6. Set Line for Minimum 3-log Cryptosporidium Inactivation for NeoTech D438™.
                                         54

-------
                      Table 4-12. Validation Factors and Validated Dose (REDyai) for Cryptosporidium
Condition
Lowered UVT
Lowered UVT Dup
Lowered Power
Lowered Power Dup

Lowered UVT
Lowered UVT Dup
Lowered Power
Lowered Power Dup

Lowered UVT
Lowered UVT Dup
Lowered Power
Lowered Power Dup
Run
#
2
3
10
11

4
5
12
13

6
7
14
15
Flow
Rate
(gpm)
151.1
151.3
154.8
154.1

251.4
250.8
251.8
251.6

436.0
435.5
434.2
435.8
Intensity
(mW/cm2)
7.9
7.9
7.9
7.9

10.0
10.0
10.4
10.4

13.0
13.0
13.1
13.2
Validation Factor
4.0-log
2.44
2.44
2.44
2.44

2.46
2.46
2.46
2.46

2.06
2.06
2.06
2.06
3.5-log
2.63
2.63
2.63
2.63

2.64
2.64
2.64
2.64

2.11
2.11
2.11
2.11
3.0-log
2.66
2.66
2.66
2.66

2.67
2.67
2.67
2.67

2.09
2.09
2.09
2.09
REDmeas
(mJ/cm2)
62.39
54.81
60.09
62.06

56.26
56.81
45.13
52.73

59.85
53.63
45.53
47.77
REDval
4-log
(mJ/cm2)
22(1)
25.5
22.4
24.6
25.4

22.9
23.1
18.4
21.5

29.0
26.0
22.1
23.1
3.5-log
(mJ/cm2)
15(1)
23.7
20.8
22.9
23.6

21.3
21.5
17.1
19.9

28.4
25.4
21.6
22.7
3.0-log
(mJ/cm2)
12(1)
23.5
20.6
22.6
23.3

21.1
21.3
16.9
19.7

28.6
25.6
21.7
22.8
(1) Required dose for log inactivation validation per the UVDGM-2006 Appendix G.
                                                            55

-------
4.8    Deriving the Validation Factor and Log Inactivation for Giardia

4.8.1   Validation Factor Definition

As described earlier in  Section 4.7.1  on Cryptosporidium, 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 VF for Giardia was determined quantitatively to account for key areas of uncertainty and
variability.  The equation for the VF is shown below.

                               W = 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 below.

4.8.2 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).  As described in Section  4.7.2, a target pathogen  must be selected to
calculate the RED bias factor. In addition to the target pathogen  Cryptosporidium described
previously,  Giardia was  also  selected for evaluation. The  RED  bias tables for Giardia 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:
                                          9              _
                         Sensitivity (ml/cm 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 4-13 shows the data for the replicates at each set point.  The highest
RED bias at each set point is used in the validation factor calculations shown later in Section
4.8.4.
                                           56

-------
Table 4-13. RED Bias Factor for Each Set Point for Giardia
Sample
Number
2-1
2-2
2-3
2-4
2-5
3-1
3-2
o o
J-J
3-4
3-5
4-1
4-2
4-3
4-4
4-5
5-1
5-2
5-3
5-4
5-5
6-1
6-2
6-3
6-4
6-5
7-1
7-2
7-3
7-4
7-5
10-1
10-1
10-1
10-1
10-1
11-1
11-2
11-3
11-4
11-5
12-1
12-2
12-3
12-4
12-5
Test
Run
2
2
2
2
2
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
6
6
6
6
6
7
7
7
7
7
10
10
10
10
10
11
11
11
11
11
12
12
12
12
12
UVT
(%)
90.6
90.6
90.6
90.6
90.6
90.6
90.6
90.6
90.6
90.6
93.9
93.9
93.9
93.9
93.9
93.9
93.9
93.9
93.9
93.9
96.5
96.5
96.5
96.5
96.5
96.4
96.4
96.4
96.4
96.4
98.
98.
98.
98.
98.
98.
98.
98.
98.
98.
98.2
98.2
98.2
98.2
98.2
Sensitivity mJ/cm2
RED
58.69
62.90
62.40
65.01
62.96
54.07
51.33
53.76
56.64
58.25
58.96
56.41
NA
56.60
53.05
63.26
62.93
50.80
50.66
56.41
66.59
57.29
61.87
54.30
59.21
54.04
64.83
50.47
47.56
51.27
65.81
61.06
61.21
57.01
55.37
65.59
70.28
61.22
60.39
52.81
40.50
43.09
50.85
44.46
46.77
Log I
2.45
2.61
2.59
2.68
2.61
2.28
2.18
2.27
2.38
2.44
2.46
2.37
NA
2.38
2.25
2.62
2.61
2.16
2.16
2.37
2.74
2.40
2.57
2.29
2.47
2.28
2.68
2.15
2.04
2.18
3.23
3.04
3.05
2.88
2.82
3.22
3.40
3.05
3.02
2.71
2.18
2.30
2.63
2.36
2.46
per Log I
Sensitivity
23.91
24.13
24.10
24.23
24.13
23.68
23.53
23.66
23.81
23.89
23.93
23.80
NA
23.81
23.62
24.15
24.13
23.50
23.50
23.80
24.31
23.84
24.08
23.69
23.94
23.67
24.22
23.49
23.33
23.53
20.40
20.07
20.08
19.78
19.67
20.38
20.69
20.08
20.02
19.48
18.55
18.75
19.34
18.86
19.03
BRED
4.0-log
Giardia
1.61
1.66
1.66
1.66
1.66
1.61
1.61
1.61
1.61
1.61
1.61
1.61
NA
1.61
1.61
1.66
1.66
1.66
1.66
1.66
1.38
1.36
1.38
1.36
1.36
1.36
1.38
1.36
1.36
1.36
1.17
1.17
1.17
1.16
1.16
1.17
1.17
1.17
1.17
1.41
1.16
1.16
1.16
1.16
1.16
BRED
3.5-log
Giardia
1.79
1.83
1.83
1.83
1.83
1.79
1.79
1.79
1.79
1.79
1.79
1.79
NA
1.79
1.79
1.83
1.83
1.83
1.83
1.83
1.43
1.42
1.43
1.42
1.42
1.42
1.43
1.42
1.42
1.42
1.2
1.2
1.2
1.19
1.19
1.2
1.2
1.2
1.2
1.19
1.19
1.19
1.19
1.19
1.19
BRED
3.0-log
Giardia
.85
.89
.89
.89
.89
.85
.85
.85
.85
.85
.85
.85
NA
.85
.85
.89
.89
.89
.89
.89
.44
.42
.44
.42
.42
.42
.44
.42
.42
.42
.21
.21
.21
1.2
1.2
1.21
1.21
1.21
1.21
1.2
1.20
1.20
1.20
1.20
1.20
                          57

-------
Sample
Number
13-1
13-2
13-3
13-4
13-5
14-1
14-2
14-3
14-4
14-5
15-1
15-2
15-3
15-4
15-5

Test
Run
13
13
13
13
13
14
14
14
14
14
15
15
15
15
15

Maximum BRED
UVT
(%)
98.2
98.2
98.2
98.2
98.2
98.
98.
98.
98.
98.
98.2
98.2
98.2
98.2
98.2

Sensitivity mJ/cm2
RED
53.28
51.10
51.82
57.78
49.65
48.43
41.20
48.70
45.19
44.12
46.35
42.82
52.97
45.16
51.57

Log I
2.73
2.64
2.67
2.91
2.58
2.53
2.21
2.54
2.39
2.34
2.44
2.29
2.72
2.39
2.66

per Log I
Sensitivity
19.51
19.36
19.41
19.84
19.25
19.16
18.60
19.18
18.91
18.83
19.00
18.73
19.49
18.91
19.39

Set Point 151 gpm - 7.9 mW/cm2
Set Point 251 gpm - 10.4 mW/cm2
Set Point 434 gpm - 13.2 mW/cm2
BRED
4.0-log
Giardia
1.16
1.16
1.16
1.16
1.16
1.16
1.16
1.16
1.16
1.16
1.16
1.16
1.16
1.16
1.16

1.66
1.66
1.38
BRED
3.5-log
Giardia
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19
1.19

1.83
1.83
1.43
BRED
3.0-log
Giardia
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20
.20

.89
.89
.44
NA - sample not analyzed so RED and bias not determined

4.8.3 Uncertainty of Validation

As described in Section 4.7.3, the uncertainty of validation (Uvai) addresses many sources of
experimental uncertainty. The same approach to uncertainty calculations described in  Section
4.7.3 apply to calculating the VF for  Giardia and in fact the uncertainty values are the same.
Please  refer to Section 4.7.3 for the equations and discussion of the various uncertainty factors
that are used for determining Uyai at each set point.

The USP results based on the REDmeas and standard deviation are shown in Table 4-9. The highest
USP for each  set point is 29.08% (151  gpm set point), 30.23% (251 gpm set point), and 34.54%
(434 gpm set point).

The uncertainty of the validation is then calculated using the highest UDR (32.58%) from the dose
response data and the highest USP for each test condition using the equation:

                           Uval = (USP2 +UDR2)1/2

Table 4-14 shows the Uyai values used for determining the uncertainty of the validation at each
set point.
                                           58

-------
       Table 4-14. Uncertainty of the Validation (Uyai) and BRED Values for Giardia
Set Point
151 gpm- 7.9mW/cm2
251gpm- 10.4mW/cm2
434gpm-13.2mW/cm2
Max
UDR
0%
32.58
32.58
32.58
Max
USP
(%)
29.08
30.23
34.54
UVa,
%
43.67
44.44
47.48
Max
BRED
4.0-log
1.66
1.66
1.38
3.5-log
1.83
1.83
1.43
3.0-log
1.89
1.89
1.44
4.8.4 Validated Dose for Giardia

After establishing the Uvai and the RED bias as described above, the VF is calculated using the
equation:
                              W = BRED x[l+(UVai/100)]
             Where:
                    VF = Validation Factor;
                    BRED = RED bias factor for Giardia (see Table 4-13); and
                    Uvai  = Uncertainty of validation expressed as a percentage.

The validated dose is then calculated as follows:

                         Validated dose (REDVai) = REDmeas / VF

Table 4-15 shows the calculated VF for various Giardia log inactivation levels (3.0, 3.5, and 4.0
log inactivation).

Table 4-15 shows the  REDyai for Giardia for each test run using the validation factors for the
various Giardia log inactivation levels. Table  4-15 also  shows the Validated Dose for each set
point and a comparison to the dose required for various levels of inactivation of Giardia. As can
be seen, the tests for the first and third set points (151 and 434 gpm) show a validated dose for
Giardia that would  achieve a minimum of 4.0 log inactivation.  The second set point (middle
point) at 251 gpm showed a minimum of 3.5 log inactivation.

The NeoTech D438™ achieved a minimum of 3.5 log inactivation for the low power runs at 251
gpm at 10.4 mW/cm  and  a 4.0 log inactivation  for all of the test runs at the set points at 151
gpm - 7.9 mW/cm2 and 434 gpm - 13.2 mW/cm2.
                                          59

-------
                            Table 4-15 Validation Factors and Validated Dose (REDyai) for Giardia
Condition
Lowered UVT
Lowered UVT Dup
Lowered Power
Lowered Power Dup

Lowered UVT
Lowered UVT Dup
Lowered Power
Lowered Power Dup

Lowered UVT
Lowered UVT Dup
Lowered Power
Lowered Power Dup
Run
#
2
3
10
11

4
5
12
13

6
7
14
15
Flow
Rate
(gpm)
151.1
151.3
154.8
154.1

251.4
250.8
251.8
251.6

436.0
435.5
434.2
435.8
Intensity
(mW/cm2)
7.9
7.9
7.9
7.9

10.0
10.0
10.4
10.4

13.0
13.0
13.1
13.2
Validation Factor
4.0-log
2.38
2.38
2.38
2.38

2.40
2.40
2.40
2.40

1.99
1.99
1.99
1.99
3.5-log
2.63
2.63
2.63
2.63

2.64
2.64
2.64
2.64

2.07
2.07
2.07
2.07
3.0-log
2.72
2.72
2.72
2.72

2.73
2.73
2.73
2.73

2.08
2.08
2.08
2.08
REDmeas
(mJ/cm2)
62.39
54.81
60.09
62.06

56.26
56.81
45.13
52.73

59.85
53.63
45.53
47.77
REDval
4.0-log
(mJ/cm2)
22(1)
26.2
23.0
25.2
26.0

23.5
23.7
18.8
22.0

30.0
26.9
22.8
24.0
3.5-log
(mJ/cm2)
15(1)
23.7
20.8
22.9
23.6

21.3
21.5
17.1
19.9

29.0
26.0
22.0
23.1
3.0-log
(mJ/cm2)
12(1)
23.0
20.2
22.1
22.9

20.6
20.8
16.5
19.3

28.8
25.8
21.9
23.0
(1) Required dose for log inactivation validation per the UVDGM-2006 Appendix G.
                                                              60

-------
4.9    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.   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 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); and
                    Uvai = Uncertainty of validation expressed as a percentage.
The validated dose is then calculated as follows:

                        Validated dose (REDVai) = REDobserved / VF

Table 4-16 shows the REDyai 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. Only some of the test runs achieved a 40 mJ/cm2 validated dose based on MS2. The
lowest flow rate condition (151 gpm) showed that three of the four test runs achieved a validated
dose of 40 mJ/cm .  One run was at 38 mJ/cm , just below the REDyai target of 40 mJ/cm based
on MS2.  A main influence on the validated dose for the lower flow tests is the UDR being just
above the 30% level at 32.58%. If the UDR had been less than or equal to  30%, the validation
factor would be significantly lower. In that case, all of the 151 gpm set point test runs would be >
40 mJ/cm and three of the four 251 gm set point test runs would be above 40 mJ/cm .
                                          61

-------
         Table 4-16. Validation Factors and Validated Dose (REDVai) based on MS2
Condition
Lowered UVT
Lowered UVT Dup
Lowered Power
Lowered Power Dup

Lowered UVT
Lowered UVT Dup
Lowered Power
Lowered Power Dup

Lowered UVT
Lowered UVT Dup
Lowered Power
Lowered Power Dup
Run
2
o
3
10
11

4
5
12
13

6
7
14
15
Flow
Rate
(gpm)
151
151
155
154

251
251
252
252

436
436
434
436
Intensity
(mW/cm2)
7.9
7.9
7.9
7.9

10.0
10.0
10.4
10.4

13.0
13.0
13.1
13.2
Validation
Factor
(1)
1.44
1.44
1.44
1.44

1.44
1.44
1.44
1.44

1.47
1.47
1.47
1.47
REDmeas
(mJ/cm2)
62.39
54.81
60.09
62.06

56.26
56.81
45.13
52.73

59.85
53.63
45.53
47.77
REDva,
(mJ/cm2)
43.4
38.2
41.8
43.2

38.9
39.3
31.2
36.5

40.6
36.4
30.9
32.4
(1) BRED equal to 1.0, therefore all log inactivation levels have the same validation factor.
                                             62

-------
4.10   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 HPC.

The general chemistry and microbiological results are presented in Tables 4-17 through 4-20.
                                           63

-------
          Table 4-17. Temperature and pH Results

Test
Blank
Low UVT (91%) Full power
Low UVT (91%) Full power Dup
Low UVT (94%) Full Power
Low UVT (94%) Full Power Dup
Low UVT (97%) Full power
Low UVT (97%) Full power Dup
Blank
Reactor Control
Reduced Power - High UVT
Reduced Power - High UVT Dup
Reduced Power - High UVT
Reduced Power - High UVT Dup
Reduced Power - High UVT
Reduced Power - High UVT Dup
Run#
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Temperature
(°F)
Influent
72.3
72.4
72.5
72.5
72.3
73.3
72.6
72.0
71.8
71.4
71.3
72.1
72.0
72.2
72.4
Effluent
72.2
72.3
72.3
72.5
72.4
73.2
72.6
72.0
71.8
71.4
71.4
72.1
72.0
72.0
72.4
pH
(S.U.)
Influent
7.39
7.42
7.41
7.52
7.57
8.28
8.43
8.37
8.36
8.36
8.35
8.48
8.51
8.60
8.64
Effluent
7.41
7.38
7.39
7.52
7.55
8.28
7.47
8.35
8.37
8.37
8.37
8.48
8.53
8.58
7.64
Table 4-18. Total Chlorine, Free Chlorine and Turbidity Results
Test
Blank
Low UVT (91%) Full power
Low UVT (91%) Full power Dup
Low UVT (94%) Full Power
Low UVT (94%) Full Power Dup
Low UVT (97%) Full power
Low UVT (97%) Full power Dup
Blank
Reactor Control
Reduced Power - High UVT
Reduced Power - High UVT Dup
Reduced Power - High UVT
Reduced Power - High UVT Dup
Reduced Power - High UVT
Reduced Power - High UVT Dup
Run#
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Total
Chlorine
(mg/L)
Influent
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.03
0.02
0.02
0.02
0.02
0.02
Free
Chlorine
(mg/L)
Influent
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
Turbidity
(NTU)
Influent
0.45
0.41
0.44
0.36
0.34
0.23
0.23
0.22
0.28
0.18
0.19
0.17
0.21
0.14
0.20
                            64

-------
                      Table 4-19. Iron and Manganese Results

Test
Blank
Low UVT (91%) Full power
Low UVT (91%) Full power Dup
Low UVT (94%) Full Power
Low UVT (94%) Full Power Dup
Low UVT (97%) Full power
Low UVT (97%) Full power Dup
Blank
Reactor Control
Reduced Power - High UVT
Reduced Power - High UVT Dup
Reduced Power - High UVT
Reduced Power - High UVT Dup
Reduced Power - High UVT
Reduced Power - High UVT Dup

Run#
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Iron
(mg/L)
Influent
0.03
0.05
0.05
O.02
0.02
O.02
0.72
O.02
O.02
O.02
O.02
O.02
O.02
O.02
<0.02
Manganese
(mg/L)
Influent
O.001
O.001
O.001
O.001
O.001
O.001
0.004
O.001
O.001
O.001
O.001
O.001
0.001
O.001
O.001
UVT(1)
(%)
Influent
91
91
91
93
93
95
95
96
96
96
96
96
96
96
96
Effluent
91
90
90
93
92
95
95
96
96
96
96
96
96
96
96
(l)UVT on grab samples, measured in laboratory after tests; In- line UVT meter used for flow test results
                 Table 4-20. HPC, Total Coliform and K coli Results

Test
Blank
Low UVT (91%) Full power
Low UVT (91%) Full power Dup
Low UVT (94%) Full Power
Low UVT (94%) Full Power Dup
Low UVT (97%) Full power
Low UVT (97%) Full power Dup
Blank
Reactor Control
Reduced Power - High UVT
Reduced Power - High UVT Dup
Reduced Power - High UVT
Reduced Power - High UVT Dup
Reduced Power - High UVT
Reduced Power - High UVT Dup

Run#
1
2
o
J
4
5
6
7
8
9
10
11
12
13
14
15
Total Coliform
(MPN/lOOmL)
Influent
2
8.66E+02
5.94E+01
3.17E+01
6.30E+00
1.60E+01
4.10E+01
<1
1.05E+03
1.57E+02
7.76E+01
1.34E+01
5.20E+00
4.10E+00
1.46E+01
Effluent
<1
<1
<1
<1
<1
<1
<1
<1
1.05E+03
7.76E+01
<1
<1
<1
<1
<1
E. coli
(MPN/lOOmL)
Influent
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
Effluent
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
HPC
(CFU/mL)
Influent
4.76E+03
7.17E+03
3.99E+03
2.80E+03
3.11E+03
1.34E+03
4.65E+02
1.61E+03
3.62E+03
1.75E+03
2.82E+03
9.45E+02
3.70E+02
4.65E+02
7.20E+02
Effluent
3.95E+01
2.35E+00
2.10E+01
1.33E+02
2.60E+01
1.28E+02
4.15E+01
1.15E+02
4.20E+03
4.80E+01
1.85E+01
6.10E+01
1.50E+01
4.70E+01
4.65E+01
                                         65

-------
4.11   Power Measurement
A power monitoring platform was connected to the unit.  This  monitoring platform provided
continuous readout of the volts, amperage, and watts being used by the unit under each test
condition. Volts, amperes, and wattage were recorded during each flow test. The power factor
was also recorded. Table 4-21 presents the power measurements taken during the flow tests.
                        Table 4-21. Power Measurement Results
Test
Blank
Low UVT (91%) Full power
Low UVT (91%) Full power Dup
Low UVT (94%) Full Power
Low UVT (94%) Full Power Dup
Low UVT (97%) Full power
Low UVT (97%) Full power Dup
Blank
Reactor Control
Reduced Power - High UVT
Reduced Power - High UVT Dup
Reduced Power - High UVT
Reduced Power - High UVT Dup
Reduced Power - High UVT
Reduced Power - High UVT Dup
Run#
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Volts
(volts)
117
117
116
117
117
117
117
118
119
117
117
116
NR
116
116
Amperage
(amps)
3.0
3.0
3.0
3.0
3.0
3.0
3.0
2.0
0.0
2.0
2.0
3.0
NR
3.5
3.5
Watts
(watts)
340
340
340
340
340
340
340
230
0.0
230
230
340
NR
390
390
Power Factor
0.97
0.96
0.96
0.97
0.96
0.96
0.97
0.99
—
0.98
0.98
0.97
NR
0.96
0.96
NR-not reported because the data were not recorded on log sheet
4.12   Headloss
Headloss was measured over the flow range of 100 to 450 gpm. Pressure at the inlet and outlet of
the reactor was measured at several flow rates as shown in Table 4-22.
                       Table 4-22. Headloss Measurement Results
Flow Rate
100
150
250
350
450
Inlet (psi)
1.59
1.92
2.87
4.22
6.16
Outlet (psi)
1.20
1.29
1.52
1.88
2.35
Headloss (psi)
0.39
0.63
1.35
2.34
3.81
                                          66

-------
                                       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 TQAP 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.  The peaks at 241.1, 287.6  and 293.5 nm all passed  QC criteria  for peak
intensity and peak nm location. Bichromate standards were also run with each batch of samples
and found to be within 102% 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.
                                          67

-------
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 were also 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.

NSF received reviewer comments about the collimated beam data in the Draft EPA ETV NSF
UV reports in November of 2011.  They identified two issues related to collimated beam data for
NSF to investigate. The two issues were a high  degree of uncertainty with replicates in the
collimated beam data  and the data trending at or below the lower 95% confidence interval for
MS2 UV sensitivity.  The initial investigation revealed no systematic error.  However, further
investigation revealed a  miscommunication  between  the company  that  calibrated NSF's
radiometers and NSF.  It was learned that the radiometers were calibrated at one of two settings
and that NSF used the setting that the radiometer was not calibrated at.  Hence the MS2 had
received about 25% more UV dose than estimated by the collimated beam data. Therefore, the
calculated REDmeas at  each set point was lower than the actual dose  delivered by  the unit. NSF
determined that the best action was to retest all previous tested units. That testing was done in
the summer of 2012 and is reported in this report.  All previous data is not reported herein as it
was deemed to be biased.

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 four  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%


                                          68

-------
The  collimated beam test procedures and key test parameters (radiometer intensity, petri  dish
factor, exposure  time,  suspension  depth, distance from  lamp  to suspension  surface,  and
absorbance/reflectance) were thoroughly reviewed for all days of testing.

The collimated beam data, as presented in Tables 4-3 and 4-4 show that on Day 2 of testing with
the high UVT water,  there was one  outlier in replicate 2 for the collimated beam results.  The
outlier, as determined by the Grubb  statistics presented in  Table 4-3 was eliminated from the
dose response curve and equation. Both sets of data, with and without the outlier are presented in
Tables 4-3 and  4-4.  As can be seen there  was very little  difference in the dose response
relationship between the data set with the outlier included and with it removed.

The dose response curve for the collimated beam data were  within the QC boundaries set in the
GP-2011 and UVDGM-2006.  The boundaries  are shown on Figures 4-3 and 4-4 along with all  of
the collimated beam dose response data. The low UVT water collimated beam data did have two
data  points that were on the  upper boundary line and the  curve was on the high side of the
midpoint. However, the dose response curve was within the established QC boundaries.

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. 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. Therefore, stability samples
were not necessarily needed for  these test runs, as the holding time was very short. However,
stability tests were run  for the times 0, 4, 8, 24 hours for informational  purposes. Table 5-1
shows the stability test results.

Trip  blanks are also normally performed to show that the stock phage solution does not change
during shipment to and from the test site. The phage stock solution was delivered from the
microbiology 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 were  required. However, trip blanks were performed once each
day to demonstrate that no significant change was occurring  during transport and handling of the
samples. Table 5-2 shows the trip blank results.
                                           69

-------
                              Table 5-1. Stability Results



Influent 0 Hour


Influent 4 Hour


Influent 8 Hour


Influent 24 Hour


Effluent 0 Hour


Effluent 4 Hour


Effluent 8 Hour


Effluent 24 Hour


Low UVT Water
MS2
PFU/mL
1.07E+05
1.20E+05
1.24E+05
4.80E+04
4.10E+04
5.50E+04
5.30E+05
4.60E+05
4.50E+05
4.60E+05
5.10E+05
4.40E+05
9.40E+02
8.90E+02
9.60E+02
1.83E+03
8.50E+02
6.00E+02
4.90E+03
2.43E+03
3.10E+03
5.20E+03
3.70E+03
4.20E+03
Average
1.17E+05


4.80E+04


4.80E+05


4.70E+05


9.30E+02


1.09E+03


3.48E+03


4.37E+03


logio
5.07


4.68


5.68


5.67


2.97


3.04


3.54


3.64





Influent 0 Hour


Influent 4 Hour


Influent 8 Hour


Influent 24 Hour


Effluent 0 Hour


Effluent 4 Hour


Effluent 8 Hour


Effluent 24 Hour


High UVT Water
MS2
PFU/mL
1.20E+05
1.21E+05
1.14E+05
6.30E+04
8.30E+04
8.70E+04
5.70E+05
5.40E+05
4.40E+05
7.10E+05
6.80E+05
5.40E+05
1.06E+03
8.80E+02
8.70E+02
4.20E+02
5.10E+02
5.20E+02
2.56E+03
2.04E+03
2.23E+03
6.60E+03
6.50E+03
4.10E+03
Average
1.18E+05


7.77E+04


5.17E+05


6.43E+05


9.37E+02


4.83E+02


2.28E+03


5.73E+03


loglO
5.07


4.89


5.71


5.81


2.97


2.68


3.36


3.76


                             Table 5-2. Trip Blank Results



Held in Micro Lab
Travel to Test Setup

Difference
Day 1 (08-02-12)
MS2
PFU/mL
6.07E+08
5.40E+08


Log™
8.78
8.73

0.05







Day 2 (08-03-12)
MS2
PFU/mL
7.15E+08
7.40E+08


Log™
8.85
8.87

0.02
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
                                         70

-------
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 after the test runs were completed. Calibration was performed at 150, 250,
350, and 450 gpm covering the range of flow rates tested. The calibration of the flow meter
showed accuracy of <1.66% (average  of 1.45%  over the flow  range) and  easily met  the
requirement of ± 5%. Table 5-3 presents the flow meter calibration data.
                         Table 5-3. Flow Meter Calibration Data
Flow Meter Reading
(gpm)
453.5
346.0
254.0
150.3
Average
Actual Calculated
Flow Rate based on tank
level change and time
(gpm)
460.1
351.8
256.0
153.2

Deviation
(%)
1.44
1.66
0.78
1.91
1.45
Reactor control and reactor blanks 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. Reactor blanks
were collected for each day of testing 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 5.74  logic and an average effluent concentration of 5.76
logic showing an increase of 0.02 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.

During  the first day of testing,  the  reactor blanks showed  no detectable MS2 (<1  PFU/ml)
present  in either the influent or  effluent samples. On Day 2, the influent  showed  some MS2
present  at 20 to 23 PFU/mL (1.30 to 1.36  logic) which is above the objective concentration of
less than 0.2 logic. The effluent samples were clean (<1 PFU/mL). The presence of some MS2 in
the influent did not directly impact the data analysis, as the amount of MS2 was small compared
the influent concentration based on an injection rate of > 5.0 logic.

The results for the blank samples for FtPC,  total coliform, and E. coli were presented in Table 4-
20.
                                           71

-------
Table 5-4. Reactor Control and Reactor Blank MS2 Results
Test
Condition
Reactor
Blank
Reactor
Blank

Reactor
Control
Test
Day
1
2

2
Test
Run
1
8

9
UVT
(%)
91
98

98
Flow
(gpm)
150
150

154
Intensity
(mW/cm2)
8.0
7.9

0.0
Influent (PFU/mL)
Repl
<1
23

6.50E+05
Rep 2
<1
22

5.50E+05
Rep 3
<1
20

5.53E+05






Effluent (PFU/mL)
Repl
<1
<1

8.40E+05
Rep 2
<1
<1

4.67E+05
Rep 3
<1
<1

5.47E+05

Test
Condition
Reactor
Blank
Reactor
Blank

Reactor
Control
Test
Day
1
2

2
Test
Run
1
8

9
UVT
(%)
91
98

98
Flow
(gpm)
150
150

154
Intensity
(mW/cm2)
8.0
7.9

0.0
Influent logio
Rep 1
0.00
1.36

5.81
Rep 2
0.00
1.34

5.74
Rep 3
0.00
1.30

5.74

Ave
0.00
1.33

5.76
Effluent Iog10
Rep 1
0.00
0.00

5.92
Rep 2
0.00
0.00

5.67
Rep 3
0.00
0.00

5.74








Ave
0.00
0.00

5.78
                          72

-------
5.7    Documentation

All laboratory activities were documented using specially prepared laboratory bench sheets and
NSF laboratory  reports.  Data from laboratory reports were entered into Microsoft™ Excel®
spreadsheets.  These spreadsheets were used to calculate the means and logic inactivations.  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.

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  ensuredby consistent
execution of the test protocol for each challenge, including timing of sample collection, sampling
procedures,  and  sample preservation.  Representativeness  was  also ensured by using each
analytical method at its optimum capability to provide results that represent the most accurate
and precise measurement it is capable of achieving.

5.9.2   Accuracy
Accuracy was quantified as the percent recovery of the parameter in a sample of known quantity.
Accuracy was measured through use  of  both matrix spikes  of a known quantity,  where
applicable, and certified standards during calibration of an instrument.

The following equation was used to calculate percent recovery:

              Percent Recovery = 100 X [(Xknown - XmeasuredyXknovm]

       where:
                      = known concentration of the measured parameter
                      = measured concentration of parameter

Accuracy of the  bench top  chlorine,  pH, and turbidity meters was checked daily during the
calibration procedures using certified check standards.   The in-line UVT  monitor was calibrated
daily with both a purchased UVT  standard and with DI water at 99.9% UVT, before the flow
tests.
                                           73

-------
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
process.  Precision of duplicate analyses was measured by use  of  the following equation to
calculate Relative Percent Deviation (RPD):
RPD =
                                                 :200
       where:
              Sl = sample analysis result; and
              S2 = 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 verification testing is based on the number of samples collected and analyzed for each
parameter and/or method, as presented in Table 5-5.
                                           74

-------
                        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 hundred percent completeness was achieved for all aspects of this validation except for the
MS2 run samples, power measurements,  and collimated beam analyses. Two samples (one
influent and one effluent - test run 4) were not analyzed out of a total of 120 samples scheduled
during the main test runs. The completeness is calculated as 98.3% complete for the MS2 test run
samples. One power measurement was not recorded out of a total of 15 scheduled measurements
for a completeness of 93.3%.  There was  one  outlier result for replicate 2 from the collimated
beam tests on the high UVT water. A total of 24 samples were scheduled and 23 used for the
dose response relationship. Therefore, completeness was 95.8%. All  planned testing activities
were conducted as scheduled.
                                         75

-------
                              Chapter 6 References

1.  Test/Quality Assurance Plan for The Ultraviolet Sciences Inc. Ultraviolet (UV) Water
   Purification System Model UVS438s-500 Reactor, August 2010.

2.  Generic Protocol for Development of Test/Quality Assurance Plans for Validation of
   Ultraviolet (UV) Reactors, NSF International, 7/2010.

3.  Generic Protocol for Development of Test / Quality Assurance Plans for Validation of
   Ultraviolet (UV) Reactors, 10/01/EPADWCTR, August 2011.

4.  Ultraviolet Disinfection Guidance Manual For the Long Term 2 Enhanced surface Water
   Treatment Rule, Office of Water, US Environmental Protection Agency, EPA 815-R-06-
   007, November 2006.

5.  German    Association    for   Gas   and  Water   (DVGW)   Technical   Standard
   Work Sheet W 294-1,2,3  (June 2006).

6.  Austrian Standards,  ONORM M5873-1, Plants for the disinfection of water using
   ultraviolet  radiation, Requirements and testing, Low pressure mercury lamp plants  (
   March, 2001).

7.  APHA, AWWA, and WEF (1999).  Standard Methods for the Examination of Water and
   Wastewater, 20th Edition.

8.  NSF International (2007). NSF/ANSI Standard 55 -  Ultraviolet Microbiological Water
   Treatment Systems.

9.  NSF International (2011).  NSF/ANSI Standard 50 - Equipment for Swimming Pools,
   Spas, Hot Tubs and Other Recreational Water Facilities.

10. Water Report 113, Safe, Sufficient and Good  Potable Water Offshore:  A guideline to
   design and operation  of offshore potable  water  systems.   2nd edition.  By  Eyvind
   Andersen and Bj0rn E. L0fsgaard, Norwegian Institute of Public Health, November 2009.

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.
                                       76

-------
                                 Attachment 1
            NeoTech D438™ Technical Manual and Specifications
  Contact Mr. Bruce Bartley at 734-769-5148 or bartley@nsf.org for a copy of this document.
This attachment is a large pdf file.

-------
                               Attachment 2







                  Additional 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.

-------
        Attachment 3




NSF Collimated Beam Apparatus

-------
Description of the Collimated Beam Apparatus
(NSF Standard 55  7.2.1.2)

An apparatus shall be assembled in which a small stirred sample can be irradiated in a nearly collimated
beam.  A  radiometer  meeting  specification in 7.2.1.2.1 can  then be  used to measure the incident
irradiance (E0).

A low-pressure  mercury vapor UV  lamp shall  be  wired to a ballast and a voltage regulator (figure 2). A
solution contained in a small dish equal to or smaller in diameter than that of the collimated tube shall be
used. The solution shall be 1 cm deep. E0 shall be measured at the surface of the liquid by removing the
dish  and  stirrer and  placing the  radiometer  at  the corresponding position from  which the  dish  was
removed.  The UV irradiance at each point of the  surface shall be within ± 5% of the average irradiance
across the solution surface.
shield '
jport
nd 	 *-








I 	 1
UV lamp
collimating
tube
*+ — petri dish
o
-* — magnetic stirrer
   NOTE 1 - The collimating tubes shall be a minimum of 53 cm (21 in) in length and the interior shall be painted
   flat black.

   NOTE 2 - The support stand, if used, shall be adjustable to raise or lower the collimating tube to the surface
   of the petri dish.

   NOTE 3 - The petri dish shall be set so the surface of the liquid is at the same level as the radiometer.

   NOTE 4 -  Measurement of the UV dose must be done at the same point at which the petri dish surface is ex-
   posed.

                                Figure 2 - Collimated beam apparatus
7.2.1.2.1   Radiometer specifications

A radiometer with the following specification shall be used:

    -   linearity: ± 0.5%;

-------
-   spectral response: visible-blind detector with narrow band-pass filter centered  at 254 nm, full
width at half maximum = 20 nm or less;

-   spatial response: cosine response ± 5%;

-   calibration: Radiometer  calibration  (including  optics,  transducer and electronics)  shall be
traceable to the National  Institute of Standards and Technology (NIST)  or another national standards
laboratory. Calibration shall be performed annually or at the  intervals specified by manufacturer,
whichever is more frequent;
     Eave=0.98
-   uncertainty:  The  calibration documentation provided  with  each radiometer  (including  optics,
transducer,  and  electronics)  shall include both calibration uncertainties  (transfer  uncertainty to
customer) and the uncertainty associated with the calibration standard.  The NIST (or other national
laboratory) uncertainty is added the transfer uncertainty to customer to yield total uncertainty; and
-   maximum total uncertainty: ± 9 % at 254 nm.

-------
Method for Challenge Microorganism Preparation, Culturing the Challenge Organism and
Measuring its Concentration
(NSF Standard 55 Annex A)


A.1    Summary

MS-2 Coliphage is used as the biological surrogate to determine the average UV dose output of UV water
treatment systems. The methods that are used for suspension preparation,  titration, and analysis of the
challenge organisms for use in the sensitivity calibration and testing are presented in this annex.


A.2    Equipment

   -  autoclave;
   -  radiometer (International light IL-700);
   -  UV collimating beam apparatus and 254 nm photo detector;
   -  incubator, 35 ± 1 °C (95 ± 1 °F);
   -  refrigerator, 5 ± 3 °C (41  ± 3 °F);
   -  water bath 50 ± 1  °C (122 ±1 °F);
   -  freezer;
   -  microwave;
   -  vortex mixer;
   -  UV-vis  spectrophotometer;
   -  pH meter;
   -  hemocytometer;
   -  Colony Counter; and
   -  centrifuge.


A.3    Microorganisms

All organisms shall be obtained from ATCC.

   -  MS-2 Coliphage (ATCC # 15597-BI); and
   -  Escherichia coli host strain (ATCC # 15597).


A.4    Supplies

   -  Petri dishes, 20 x 60 mm and 15 x 100 mm: sterile;
   -  pipettes, 1 ml and 10  ml, sterile;
   -  sterile centrifuge tubes, 10 ml and  50 ml;
   -  sample bottles, 125 ml sterile screw cap;
   -  test tubes, 16 x 125 mm;
   -  sterile inoculating loop;
   -  sterile filtration apparatus;
   -  sterile 0.22 urn polycarbonate membrane filters;
   -  Whatman #1 filter;
   -  chlorine detection kit; and
   -  disposable sterile 250  ml polypropylene container.

-------
A. 5   Reagents

    -  Sterile buffered dilution water (SBDW). This shall be prepared according to the Standard Methods
    for the Examination of Water and Wastewater (dilution water: buffered water);

    -  Phosphate buffer saline (PBS).  A stock  solution shall be prepared by dissolving 80 g sodium
    chloride  (NaCI), 2 g  potassium dihydrogen phosphate (KH2PO4), 29 g hydrated  disodium hydrogen
    phosphate (Na2HPO4-12H2O), and 2 g potassium chloride (KCI) in water to a final volume of 1 L. A
    working  solution shall be prepared from the stock solution by diluting  1 volume  of the stock with 9
    volumes of water. The pH shall  be adjusted using a pH meter to 7.4 with 0.1  N HCI or 0.1 N NaOH
    before use;

    -  Ethylenediaminetetraacetic acid (EDTA), Sigma # ED2SS; and

    -  Lysozyme, Boehringer Mannheim, #1 243004. Store at 2 to 8 °C (35 to 46 °F).


A.6   Safety precautions and hazards

A.6.1  Steam sterilized  samples and equipment shall be handled with  protective gloves when  being
removed from the  autoclave.

A.6.2  Cryogenic culture vials shall be handled with cryoprotective gloves.

A.6.3  Ultraviolet light shall be  used to expose the  organism during calibration. This light can result in
skin cancer and retinal damage; hence personnel must be protected from exposure.

A.6.4  All microbiological samples and contaminated test supplies shall be steam sterilized to 121 ± 1 °C
(250 ±1 °F)  at 15  psi fora minimum of 20 min prior to being discarded.


A.7   Growth  medium

    NOTE 1 - Common bacteriological media may be purchased from bacteriological  medium manufacturers and
    prepared according to the manufacturer's instructions.

    NOTE 2 - The  quality of the growth media shall be monitored by examining growth promotion and sterility prior
    to use.

A.7.2  Formula to be used when MS-2 Coliphage  is chosen for microbiological agent

A.7.2.1 TSB (Tryptic Soy Broth)
Ingredient
tryptone
soytone
dextrose
sodium chloride
dipotassium phosphate
Dl water
pH
Amount
1.7g
0.3 g
0.25 g
0.5 g
0.25 g
100mL
7.3 ±0.2

-------
TSB shall be dissolved by boiling and adjusted to final pH. 8-mL aliquots shall be dispensed into 16 x 150
mm test tubes. TSB shall be autoclaved at 121 ±1 °C (250 ±1 °F) at 15 psi for 20 min. Cooled broth  shall
be stored at 5 ± 1 °C (41 ± 1 °F).

A.7.2.2 1.5% TSA (Tryptic Soy Agar)
Ingredient
tryptone
soytone
sodium chloride
bacto-agar
Dl water
PH
Amount
7.5 g
2.5 g
2.5 g
7.5 g
500 ml
7.3 + 0.2
TSA shall be dissolved by boiling, adjusted to final pH, and autoclaved at 121  ±1 °C (250 ±1  °F) at 15
psi for 20 min. Tempered media shall be poured into sterile petri dishes. Agar plates shall be stored at 5 ±
1 °C (41 ± 1  °F). Plates shall be allowed to come to room temperature before use.

A.7.2.3 Phage top agar 1% ISA (Tryptic Soy Agar)
Ingredient
tryptone
soytone
sodium chloride
agar
Dl Water
PH
Amount
7.5 g
2.5 g
2.5 g
5.0 g
500 ml
7.3 ±0.2
TSA shall be dissolved by boiling, adjusted to final pH, and autoclaved at 121  ±1 °C (250 ±1  °F) at 15
psi  for 20 min. Agar shall be stored at 5 ± 3  °C (41  ±1  °F).  On the day of  testing, the  TSA shall be
liquefied  and placed  in the 45 ±  1 °C  (113 ± 1 °F) water bath. The MS-2 Coliphage top  agar shall be
maintained at 45 ± 1 °C (113 ± 1 °F) to prevent agar solidification.


A.8   Culture of challenge organisms

A.8.2  MS-2 Coliphage

A.8.2.1 Stock culture preparation of MS-2 Coliphage

    NOTE - This  section describes the  propagation  and  harvesting methods for stock  suspensions of MS-2
    Coliphage for use as a challenge suspension for low flow (<  1 gpm) water treatment units. If units possessing a
    flow rate greater than 1 gpm are  to be tested, the stock preparation procedure may  have to be repeated multiple
    times to achieve the required volume of MS-2 Coliphage. This method should also  be repeated when cryogenic
    stocks are low.

    a)  One day prior to preparation of MS-2 Coliphage stock, a cryogenically frozen E. coli host strain
    shall  be thawed.  One TSB tube shall be inoculated with 0.1 ml of the stock suspension. The stock
    suspension shall  be incubated at 35 ± 1°C (95 ±1 °F) for 18 ± 2 h.

    b)  On the day of preparing MS-2 Coliphage stock, 1% TSA shall be liquefied and the media shall be
    tempered in a 45 ± 1 °C (113  ± 1 °F) water bath. 1.5%  TSA plates shall be  room temperature prior to
    use.

-------
    c)  Serial dilutions of MS-2 Coliphage suspension (10~1 to 10~12) shall be made using sterile PBS. 10~5
    to 10~12 dilutions shall be plated in triplicate on 1.5% ISA plates. In a sterile tube, 1 ml of diluted MS-
    2 Coliphage shall be transferred. Then 0.1 ml of E. coli host shall  be added quickly to ~ 5 ml of
    melted 1% ISA. The inoculum and media shall be vortexed and poured  on ISA plates. The plates
    shall  be rocked to spread inoculum evenly. After the 1% TSA layer has solidified, the plates shall be
    inverted and incubated at 35 ± 1°C (95 ±1 °F) for 18 ± 2 h.

    d)  Plates shall be selected that show complete lysis  of host cells by the MS-2  Coliphage. The
    surface of each  plate shallbe flooded  with 3  ml of TSB. The 1%  TSA layer shall be gently removed
    using a cell scraper. The contents shall be poured into two sterile 50 ml centrifuge tubes and the total
    volume brought to 40 ml with TSB. 0.2 g EDTA and 0.026 g lysozyme shall be added to each tube.
    The centrifuge tubes shall be incubated at room temperature for 2  h, mixing every 15 min.

    e)  After the 2 h incubation, the tubes shall be centrifuged at 9280 xg for 5 min, or 2320 xg for 20
    min, at 20 ± 1 °C (68 ± 1 °F). The resulting supernatant shall be removed  while avoiding the pellet. A
    sterile 47-mm filtration assembly shall  be aseptically constructed using a 0.22-|j,m polycarbonate filter.
    The filter shall be  pretreated with 10  ml of TSB broth  just prior to the  filtration to minimize  MS-2
    Coliphage adsorption to the filter. The  supernatant shall be filtered.

    f)  For long-term  storage (greater than 28  d),  V10 volume of  sterile glycerol  shall  be added  to
    suspension, dispensed into 1 ml and  3 ml aliquots in cryovials,  and stored at -70° ±  1 °C (-94 ± 1
    °F).

    g)  The MS-2  Coliphage  suspension shall  be titrated  as in A.8.2.2. The concentration  of  MS-2
    Coliphage should be 10™ to 1012 PFU/mL

A.8.2.2 Enumeration of MS-2 Coliphage plaques

    a)  A cryogenically frozen E. coli host strain shall be thawed. One TSB tube shall be inoculated with
    0.1 ml of the stock suspension. The TSB tube shall be  incubated at 35 ± 1  °C (95 ±1  °F) for 18 ± 2
    h.

    b)  1% TSA shall be liquefied  and the media shall  be tempered in a 45 ± 1 °C  (113  ± 1 °F) water
    bath. 1.5% TSA plates shall be  room temperature prior to use.

    c)  Serial dilutions of MS-2 Coliphage suspension (10~1 to 10~12) shall be made using sterile PBS. 10~7
    to 10~12 dilutions shall be plated in triplicate on 1.5% TSA plates. In a sterile tube, 1 ml of diluted MS-
    2 Coliphage shall be transferred. Then 0.1 ml of E. coli host shall  be added quickly to ~ 5 ml of
    melted 1% TSA. The inoculum and media shall be vortexed and poured  on TSA plates. The plates
    shall  be rocked to spread inoculum evenly. After the 1% TSA layer has solidified, the plates shall be
    inverted and incubated at 35 ± 1°C (95 ±1 °F) for 18 ± 2 h.

    d)  After  incubation, plates  containing  20  - 200  distinct  plaque  forming units (PFU)  shall be
    enumerated using  a Colony Counter.  The MS-2  Coliphage suspension titer shall be  calculated by
    multiplying the number of PFU obtained by the inverse of the dilution factor. The concentration of MS-
    2 Coliphage should be 1010 to 1012 PFU/mL.


A.9   Drinking water treatment unit challenge organism suspension preparation

A.9.1  Determination of the concentration of challenge organism

This determination will be based upon the unit flow rates, injection feed  pump rate, suspension density,
and  the  final challenge organism concentration for the unit challenge. The suspension shall be  of
adequate volume to deliver the challenge organism to two complete on/off cycles at each sample point.

-------
    Example:

               unit flow rate: 1.0 gpm; duplicate units tested so total of 2.0 gpm (7560 mL/min);
               injection rate: 10 mL/min;
               suspension density: 1 x 109/ml_;
               final concentration: 7.0 x 104/ml_; and
               on/off cycle: 10 min /10 min (20 min on for two complete cycles).

    a)  To challenge for 20 min at two  10 min intervals, a total of 200 ml of suspension is  needed to
    challenge 151,200  ml of water (7560 min x 20 min):

               (7.0 x 104/ml_)(151,200 ml) = (injection feed conc.)(200 ml); and
               injection feed concentration = 5.3x107/ml_.

    b)  To prepare this from the stock suspension, combine:

               (200 ml_)(5.3 x 107/ml_) = (ml of suspension density)(1.0 x 109/ml_);
               ml of suspension density = 10.6 ml;  and
               10.6 ml of suspension to 189.4 ml of PBS.

       Once suspension has been made, the suspension shall be mixed using a magnetic stirrer.

       A 10-mL aliquot shall be  removed from the  challenge suspension and set aside  for density
       verification  according to Standard Methods for the Examination of Water and Wastewater.


A. 10  Analysis of influent and effluent samples


A.10.2 Enumeration of MS-2 Coliphage plaques

    a)  Serial dilutions of the influent and effluent samples (10° to 10~5) shall be made using sterile PBS.
    10° to 10~5 dilutions shall be plated in duplicate on 1.5% TSA plates. In a sterile tube, 1 ml of diluted
    MS-2 Coliphage shall be transferred. Then 0.1 ml  of E. coli host shall be added quickly to ~ 5 ml of
    melted 1% TSA. The inoculum and media shall be vortexed and poured on TSA plates. The plates
    shall be rocked to spread inoculum evenly. After the 1% TSA layer has solidified, the plates shall be
    inverted and incubated at 35 ± 1°C (95 ±1  °F) for 18 ± 2 h.

    b)  After  incubation, plates containing 20  - 200 distinct  plaque forming  units (PFU) shall be
    enumerated using  a Colony Counter. The MS-2 Coliphage  suspension titer  shall be  calculated by
    multiplying the number of  PFU obtained  by the  inverse  of the dilution  factor. Results shall be
    expressed as the number of PFU/mL.


A.11  Challenge verification

After the appropriate incubation period for MS-2 Coliphage, the plaques shall be counted on all of the
density determination plates. The mean number of microorganisms per milliliter for plates with 25 to 250
colonies/plaques shall  be calculated. This shall verify that the  challenge organism was  present in the
challenge test water at  the optimum concentration before being added to test apparatus.

-------
                               Attachment 4
       UVT Scans of Influent and Effluent Water at high and low UVT






Contact Mr. Bruce Bartley at 734-769-5148 or bartley@nsf.org for a copy of this document.




This is a large pdf file.

-------