December 2003
                               03/12/WQPC-WWF
                                EPA/600/R-04/035
   Environmental Technology
   Verification Report

   Wet Weather Flow Monitoring
   Equipment

   ADS Environmental Model 3600
   Open Channel Flow Monitor

   Part I - - Laboratory Test Results

               Prepared by
             NSF International
       Under a Cooperative Agreement with
4>EPA U.S. Environmental Protection Agency

-------
         THE ENVIRONMENTAL TECHNOLOGY VERIFICATION
                                     PROGRAM
 U.S. Environmental
 Protection Agency
                                                NSF International
                    ETV Joint Verification Statement
TECHNOLOGY TYPE:
APPLICATION:

TECHNOLOGY NAME:

TEST LOCATION:
COMPANY:
ADDRESS:


WEB SITE:
EMAIL:
        AREA/VELOCITY FLOW METERS
        FLOW METERING IN SMALL- AND MEDIUM-SIZED
        (10- to 42-inch) SEWERS
        ADS ENVIRONMENTAL SERVICES MODEL 3600 OPEN
        CHANNEL FLOW METER

        QUEBEC CITY, QUEBEC, CANADA, AND LOGAN, UTAH
        ADS ENVIRONMENTAL SERVICES
        5030 BRADFORD DRIVE    PHONE: (800)633-7246
        BUILDING 1, SUITE 210    FAX:  (256) 430-6633
        HUNTSVILLE, AL 35805
        http:\\www.adsenv.com
        info@adsenv.com
NSF International (NSF) manages the  Water Quality  Protection Center (WQPC)  under  the U.S.
Environmental Protection Agency's (EPA)  Environmental Technology Verification (ETV)  Program.
NSF evaluated the performance of the Model 3600 Open Channel Flow Meter manufactured by ADS
Environmental Services.  Utah Water Research Laboratory (UWRL) in Logan, Utah, and BPR of Quebec
City, Canada, both NSF-qualified testing organizations, performed the laboratory and field verification
testing, respectively.
EPA created the ETV Program to  facilitate  the deployment of innovative or improved environmental
technologies through performance verification and dissemination of information.  The goal of the ETV
program is to further environmental protection by accelerating the acceptance and use of improved and
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; 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 tests  (as
appropriate), collecting and analyzing data, and 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.
03/12/WQPC-WWF
The accompanying notice is an integral part of this verification statement.

                       VS-i
December 2003

-------
TECHNOLOGY DESCRIPTION
The  following  technology description  is provided  by the vendor  and does  not represent verified
information.
Area/velocity flow meters  are commonly used in wastewater collection, storm sewer, and combined
sewer systems.  The ADS 3600 flow meter utilizes a quad-redundant ultrasonic sensor that measures the
time required for an ultrasonic pulse to travel from the sensor face to the surface of the water and back to
the sensor. The meter converts the travel time to distance by calculating the speed of sound  through air
and adjusting for temperature, which is measured by two sensors inside the ultrasonic sensor head. The
depth of the flow is  then calculated using the  pipe diameter and the  range measured by the ultrasonic
sensor. A pressure-depth sensor is also installed at the bottom of the pipe to measure surcharge levels and
to provide a redundant depth reading when used with the ultrasonic level sensor. Doppler velocity
measurements are made by transmitting an ultrasonic signal upstream using a submerged velocity sensor
and measuring the frequency shift in the sound waves reflected by the moving particles in the water. The
depth and velocity  sensor readings  are stored in  the  flow meter's memory  until the data can be
downloaded to a computer through either  a  voice-grade telephone line  or a cellular network. The
computer software calculates flow rates using the depth and velocity readings.
The ADS 3600 flow  meter system includes the  flow meter unit, sensors, and installation hardware. The
flow meter unit is  housed in a waterproof, marine-grade aluminum housing. The submersible pressure
sensor, ultrasonic level sensor, and velocity sensor are attached to a circular stainless steel band installed
around the inner circumference  of the sewer pipe.  Waterproof cables with sealed connectors convey
power and signals between the flow meter unit and the sensors. The system is battery-powered, and can
power the unit for about one year at a standard 15-minute measurement interval.  The unit is intrinsically
safe. According to  vendor claims, after the unit is installed, minimal operation and maintenance (O&M)
or unit calibration is required; the most common O&M procedure is cleaning the sensors.

VERIFICATION  TESTING DESCRIPTION
Laboratory Test Site
The laboratory  testing was completed at the Utah Water Research Laboratory (UWRL), at Utah State
University in Logan, Utah. The flow meter was installed in three nominal pipe  sizes: 10-inch, 20-inch,
and 42-inch. The straight lengths were sized  so they were at least 40 times the pipe diameter for the 10-
and 20-inch pipes and at least 22 times for the 42-inch pipes. Pipe slopes were  adjustable to allow the
flow meter to be evaluated under different slope conditions. Sluice gates at both ends of the pipes were
used to regulate appropriate flow, head, and  obstruction  during testing. Reference devices were directly
traceable to the National Institute of Standards and Technology (NIST), and were regularly calibrated.
Uncertainty for the reference devices was less than 0.25 percent.
Field Test Site
Field verification testing was conducted in a section of the Quebec Urban Community's sewer network,
located in the City of Sainte-Foy, Quebec,  Canada.  The ADS flow  meter and reference meters were
installed  in a 41.7-inch diameter interceptor pipe, near the downstream end of a straight run of pipe that
had a mean slope of 0.169 percent. The reference devices, which consisted of a bubbler for reference level
measurement,  a reference flow  monitor, and  an  Accusonic 4-path flow  monitor,  were  installed
downstream of the ADS  flow meter. Upstream and downstream sluice gates were used to create the
required flow conditions.
Validation of the reference flow  monitor and bubbler were performed by lithium tracer dye  tests. Flow
rates under the  upstream and downstream gates were calculated using standard hydraulic equations for a
redundant check of flow data.
03/12/WQPC-WWF      The accompanying notice is an integral part of this verification statement.         December 2003

                                              VS-ii

-------
Methods and Procedures

Laboratory evaluation of the flow meters consisted of collecting depth, velocity, and flow data from the
ADS meter and comparing it to the depth, velocity, and flow data from the reference devices. These tests
were performed under normal operating conditions of uniform flow, backwater flow, full pipe (manhole
surcharged), and simulated silt. Water transmission through the pipes, as a ratio of flow depth versus the
pipe diameter (d/D), ranged from 10 to 250 percent (surcharged conditions). Tests were also performed
under the abnormal operating conditions of reverse flow and grease accumulation.
Field evaluation of the ADS flow meter at the Quebec site consisted of a general evaluation of the flow
meter (Test A) and the  performance  of the meter under varying  flow conditions. Testing consisted of
collecting depth,  velocity,  and  flow  data  at regular time intervals and  comparing the data to the
corresponding depth, velocity, and flow data from the reference devices. Four test scenarios were used:
1.  Test B—accuracy under low weather flow (approximately 1.71 million gallons per day [MGD]), with
    back-flow conditions;
2.  Test C—accuracy under wet weather flow (1.71-29.7 MGD), without back-flow conditions;
3.  Test D—accuracy under wet weather flow (1.71-29.7 MGD), with back flow-conditions; and,
4.  Test E—accuracy under short-term (26-day) continuous operation, with various flow rates.
Three conditions were identified during testing  that created an unintended challenge to the  ADS flow
meter:

1.  The  water used in the testing at UWRL did not contain the particulate concentrations of normal
    sewage, so small quantities  of coffee creamer were  added to the water on some  test runs. The
    operating principle  utilized by the ADS flow meter  requires  particles  in the water to serve as
    reflectors for sound  waves. The vendor maintained that the coffee creamer additive provided a level
    of reflectivity, but the particulate concentration in the test water did not approach that of sewage and
    could be a source of measurement error.
2.  During each field test, a portion of the ADS flow meter data collected at one-minute intervals was not
    recorded. ADS  personnel indicated that this happened because the flow  meter was  configured for
    maximum error checking and sensor refiring. They further indicate that the ADS 3600  flow meter can
    be reconfigured to  collect data at one-minute intervals by reducing the level of  real-time  error
    checking.

3.  The field testing results include data in which it appears that standing waves and troughs were  present
    beneath the ADS 3600 flow meter's ultrasonic depth sensor. During portions of the  testing, the depth
    sensor was likely affected by  standing  waves and troughs up to +5 inches. The  ADS flow meter
    measures depth with a downward-looking, narrow-beam ultrasonic sensor mounted on the top of the
    pipe, so depth measurements would be susceptible to influence by waves.  Based on a review of the
    field data, it appears  that waves were most prevalent at higher depths and flow rates.
No editing was allowed  on the metered data during field or laboratory testing. In actual applications, the
flow monitoring service  provider may implement post-monitoring  quality control measures to attempt to
improve the accuracy  of final data. According to ADS, the company typically bundles flow meter sales
with post-monitoring quality control and reporting services.
03/12/WQPC-WWF      The accompanying notice is an integral part of this verification statement.         December 2003

                                             VS-iii

-------
VERIFICATION OF PERFORMANCE
System Operation
The  testing  organizations found the  equipment durable and easy to use, and that it required minimal
maintenance. The  flow meter operation and data retrieval software programs were easy to learn. The
ultrasonic sensors and stainless steel band did not promote accumulation of debris during testing.
Laboratory Testing Results

The  mean deviations and the  95-percent confidence intervals under normal operating conditions (i.e., all
test conditions except grease tests and reverse flow) are presented in Table 1.  The width of the 95-percent
confidence interval is a function of the variation in instrument deviation, and the number of test runs in
each reported category. Categories with a fewer number of runs show wider confidence intervals. The
calculations exclude "abnormal condition" tests,  where grease was applied to the sensors or where
reverse-flow conditions were  created. The mean deviation for the abnormal operating conditions was 1.9
percent for the 0.5-mm grease tests, -100.0 percent for the 2.0-mm grease tests, and -72.7 percent for the
reverse-flow tests.
Table  1. Deviation and 95-Percent Confidence Interval by Test Configuration for Lab Testing
Deviation
Pipe size (inches) (percent)
10
20
42
Pipe slope (percent)
0.1
0.2
0.5
1.25
2.0
Percent full (d/D, percent)
10
30
50
80
150
250
Condition
Free-flow
Back-flow
All conditions
2.8
7.2
1.6

4.9
1.6
5.6
4.2
6.1

3.6
11.1
3.4
1.3
0.9
4.8

2.6
5.8
4.5
95-percent
confidence interval
0.2
2.4-
-1.8

-1.4
-1.8
1.6
-1.6
-3.8

-9.4
-5.4
-12.0
-5.0

-11.2
-5.0
-9.6
-9.9
-16.1

-16.6
6.5- 15.7
0.2
-3.1
0.9
-5.0

-1.4
2.9
2.1
-6.7
-5.6
-2.8
-14.6

-6.7
-8.7
-6.9
The overall accuracy of the ADS flow meter under normal operating conditions is shown in Figure 1. The
meter deviation is segregated into two components—bias and precision. Overall bias was 1.6 percent, as
calculated by the slope of the best-fit line. Precision, as calculated with the correlation coefficient (r2),
was 0.45 percent.
03/12/WQPC-WWF      The accompanying notice is an integral part of this verification statement.         December 2003

                                              VS-iv

-------
        25,000
        22 /BUM
        20,000
        17,500
      5 15,000
      IB
        2,500
      Q 10,000
        ,',500
        5,000
        2,500
                                                   Zero deviation reference line (dashed)
                                                                             y = 0.9841 x
                                                                             R2 = 0.9955
                                                                        Best fit line (solid)
                   2,500     5,000     7,500    10,000    12,500    15,000
                                             Reference Flow (GPM)
                                                                  17,500
                                                                          20,000
                                                                                  22,500
                                                                                          25,000
Figure 1. Laboratory metered flow rate versus reference.

Field Testing Results

Table 2  summarizes the field testing  results in two categories:   mean deviation  and trimmed mean
deviation. The mean deviation is the arithmetic mean of all of the one-minute-interval data. The trimmed
mean deviation is calculated by eliminating values greater than ±99 percent, making it less susceptible to
skewing from large outliers, such as those produced when the ADS flow meter recorded zero velocity.

Table 2. Deviation from Reference Flow: Tests B, C, and D
               Flow regime
                  Mean deviation
                     (percent)
Trimmed mean deviation
        (percent)
TestB
TestC
TestD
Test B-D combined
Simulated low flow
Simulated high flow
Combined flows
-29.7
5.0
-1.0
-5.7
-7.6
-4.3
-5.7
2.4
5.0
4.7
2.1
12.2
-4.1
2.1
Analysis of the data collected during Test B (low flow) revealed that the deviation was -100 percent in
nearly one-third of the samples. This occurred most frequently during back-flow conditions when the
ADS 3600 recorded zero velocity and calculated zero flow. The  data collected during Tests C and D
showed a much lower occurrence of data with deviations exceeding ±99 percent.
03/12/WQPC-WWF
The accompanying notice is an integral part of this verification statement.

                          VS-v
                            December 2003

-------
Test E (not included in Table 2) evaluated the performance of the flow meter over an extended (26-day)
time period. Generally, the data collected during Test E closely correlated with the reference flow monitor
data. Spikes were noted in water level measurements collected toward the end of the test, which may have
been the result of accumulated condensation on the  ultrasonic depth probe. No debris accumulation was
observed on the equipment, and, aside from a thin  film of grease on the probes, the equipment was in
good condition and did not require maintenance.

QUALITY ASSURANCE/QUALITY CONTROL

A complete description of the quality assurance/quality control procedures and findings are included in
the verification reports. Calibration records were maintained by the testing organizations and validation of
the reference flow  devices fell within control limits. NSF completed a data quality  audit of at least 10
percent of the test  data to ensure that the reported data represented the data generated during testing.
Audits of the field and laboratory testing were conducted by NSF with no significant issues noted.
       Original Signed by
       Lee A. Mulkey
         March 31, 2004
    Lee A. Mulkey                  Date
    Acting Director
    National Risk Management Laboratory
    Office of Research and Development
    United States Environmental Protection Agency
Original Signed by
Gordon Bellen    April 26, 2004
                                    Gordon Bellen
                                    Vice President
                                    Research
                                    NSF International
                       Date
    NOTICE:  Verifications are based  on  an evaluation  of technology  performance under  specific,
    predetermined criteria and the appropriate quality assurance procedures. EPA and NSF make no expressed
    or implied warranties as to the performance of the technology, and do not certify that a technology will
    always operate as verified. The end user is solely responsible for complying with any and all applicable
    federal, state, and local requirements. Mention of corporate names, trade names, or commercial  products
    does not constitute endorsement or recommendation for use of specific products. This report is not an NSF
    Certification of the specific product mentioned herein.
        Availability of Supporting Documents
        Copies of the Draft 4.0 - Generic Verification Protocol, Flow Monitors for Wet Weather Flows
        Applications in Small- and Medium-Sized Sewers, September, 2000, the verification statement, and
        the verification report (NSF Report #03/12/WQPC-WWF) are available from:
        ETV Water Quality Protection Center Program Manager (order hard copy)
            NSF International
            P.O. Box 130140
            Ann Arbor, Michigan 48113-0140
        NSF web site: http://www.nsf.org/etv (electronic copy)
        EPA web site: http://www.epa.gov/etv (electronic copy)

        (NOTE: Appendices are not included in the verification report. Appendices are available upon
        request from NSF.)	
03/12/WQPC-WWF
The accompanying notice is an integral part of this verification statement.

                          VS-vi
                          December 2003

-------
     Environmental Technology Verification Report
WET WEATHER FLOW MONITORING EQUIPMENT VERIFICATION

              ADS ENVIRONMENTAL MODEL 3600
               OPEN CHANNEL FLOW MONITOR
             PART I - LABORATORY TEST RESULTS
       UTAH WATER RESEARCH LABORATORY TEST SITE
                          Prepared for:

                        NSF International
                       Ann Arbor, Michigan
                          Prepared by:

                  Utah Water Research Laboratory
                          Logan, Utah
                         December 2003
Under a cooperative agreement with the U.S. Environmental Protection Agency

                 Raymond Frederick, Project Officer
            National Risk Management Research Laboratory
                U.S. Environmental Protection Agency
                         Edison, NJ 08837

-------
                                       Notice

The  U.S. Environmental  Protection  Agency (EPA) through its  Office  of Research  and
Development has financially supported and collaborated with NSF International (NSF) under a
cooperative agreement. This verification effort was supported by the Water  Quality Protection
Center operating  under  the  Environmental  Technology  Verification  (ETV)  Program.  This
document has been peer reviewed and reviewed by NSF and EPA and recommended for public
release.
                                          11

-------
                                      Foreword
The  following is the final report  on an Environmental  Technology Verification (ETV) test
performed for NSF International (NSF) and the United States Environmental Protection Agency
(EPA) by Utah Water Research Laboratory,  in cooperation with ADS Environmental Services
for the Model 3600 Open Channel Flow Monitor. The test protocol for  flow monitors requires
both laboratory and field testing. The final report for this verification is divided into two parts to
address both portions of testing.

This part of the report (Part I) describes the testing and summarizes the data from the laboratory
testing. Part II: Field Test Results, describes the testing and summarizes the  data of the field
testing. Both parts of the report are available on the NSF and EPA websites.

The U.S. Environmental Protection Agency (EPA) is charged by Congress with protecting the
Nation's  land, air,  and water resources. Under a mandate of national environmental laws, the
Agency strives to formulate and implement actions leading to a compatible balance between
human activities and the ability  of natural systems to support and nurture life. To meet this
mandate,  EPA's research program is providing  data  and  technical  support  for  solving
environmental problems today and building a science knowledge base necessary to manage our
ecological resources wisely, understand how  pollutants affect our health, and prevent or reduce
environmental risks in the future.

The  National Risk Management Research Laboratory (NRMRL) is  the Agency's center for
investigation  of  technological  and  management approaches for preventing and reducing risks
from pollution that threaten human health  and the environment. The focus of  the Laboratory's
research  program is on  methods and  their  cost-effectiveness for prevention and control  of
pollution to air, land, water, and subsurface resources; protection of water quality in public water
systems;  remediation of contaminated sites, sediments and ground water; prevention and control
of indoor air pollution; and restoration of  ecosystems. NRMRL  collaborates with both public
and private  sector partners to foster technologies  that reduce the cost of compliance and  to
anticipate emerging problems. NRMRL's research provides solutions to environmental problems
by: developing and promoting technologies that protect and improve the environment; advancing
scientific and engineering information to support regulatory and policy decisions; and providing
the technical support  and information transfer  to ensure implementation of environmental
regulations and strategies at the national, state, and community levels.

This publication  has been produced as part  of the Laboratory's strategic long-term research plan.
It is published and made  available by EPA's Office of Research and Development to assist the
user community and to link researchers with their clients.
                                           in

-------
                                       Contents

Verification Statement	VS-i
Notice	ii
Foreword	iii
Contents	iv
Acronyms and Abbreviations	vii
Acknowledgments	viii
Chapter 1 Introduction	1
  1.1   ETVPurpose and Program Operation	1
  1.2   Testing Participants and Responsibilities	1
    1.2.1   U.S. Environmental Protection Agency	2
    1.2.2   NSF International	2
    1.2.3   Laboratory Testing Organization	2
    1.2.4   Field Testing Organization	3
    1.2.5   Vendor	4
  1.3   Laboratory Verification Testing Site	4
Chapter 2 ADS Equipment Description and Operating Processes	5
  2.1   Equipment Description	5
  2.2   Operating Process	9
    2.2.1   Depth	9
    2.2.2   Velocity	10
Chapters Laboratory Report	11
  3.1   Test Set-Up, Test Equipment, and Procedures	11
    3.1.1   Test Description	11
    3.1.2   Reflectors	19
    3.1.3   Laboratory Test Instrumentation	19
      3.1.3.1   Flow Measurement Tanks and Calibrated Flow Meters	19
      3.1.3.2   Precision Point Gauge	20
      3.1.3.3   Thermometer	20
      3.1.3.4   Timer	21
      3.1.3.5   Precision Calipers	21
    3.1.4   Pretest Procedures	21
    3.1.5   General Test Procedures	21
      3.1.5.1   Set Flow Condition	21
      3.1.5.2   Allow Flow To Stabilize	22
      3.1.5.3   Measure Water Temperature	22
      3.1.5.4   Measure Reference Flow	22
      3.1.5.5   Measure Reference Depth	23
      3.1.5.6   Log Meter Data	24
      3.1.5.7   Measure Reference Flow and Depth (Second Time)	24
      3.1.5.8   Calculate Reference Velocity	24
      3.1.5.9   Record Observations	24
      3.1.5.10  Review Reference Data	25
      3.1.5.11  Download Meter Data	25
                                           iv

-------
    3.1.6   Test Conditions	25
      3.1.6.1    Free Flow and Backwater Tests	25
      3.1.6.2    Full Pipe Tests (Manhole Surcharged)	25
      3.1.6.3    Silt Simulation Tests	26
      3.1.6.4    Grease Build-up Tests	26
      3.1.6.5    Reverse Flow Tests	27
    3.1.7   Data Management and Analysis	30
    3.1.8   Quality Assurance	31
  3.2  Test Results	31
    3.2.1   Preliminary Test Measurements	31
    3.2.2   Test Data	34
      3.2.2.1    Statistical Evaluation of Data	34
      3.2.2.2    Graphical Evaluation of All Flow Data	35
      3.2.2.3    Graphical Evaluation of Flow Data by Test Condition	37
      3.2.2.4    Data Analysis Discussion	38
Appendices	58
  A Laboratory Equipment Calibrations and Information	58
  B  Raw Laboratory Test Notes and Data	58
  C  Operational Procedure and Data Logging Method	58
  D Laboratory Test Data	58
Glossary	59

                                        Tables

Table 3-1. Test Conditions and Sequence:  10-inch Test Pipe	28
Table 3-2. Test Conditions and Sequence: 20-inch Test Pipe	29
Table 3-3. Test Conditions and Sequence: 42-inch Test Pipe	30
Table 3-4. Preliminary ADS 3600 10-inch Pipe Test Measurements	32
Table 3-5. Preliminary ADS 3600 20-inch Pipe Test Measurements	33
Table 3-6. Preliminary Model 3600 42-inch Pipe Test Measurements	33
Table 3-7. Deviation by Test Configuration: Normal Operating Conditions	35

                                        Figures

Figure 2-1. ADS Model 3600 Open Channel Flow Monitor	7
Figure 2-2. ADS flow monitoring sensors (laboratory  20-inch installation)	8
Figure 2-3. Ultrasonic sensor illustration	9
Figure 2-4. Doppler velocity  sensor illustration	10
Figure 3-1. The 10-inch pipe test set-up	13
Figure 3-2. The 20-inch pipe test set-up	14
Figure 3-3. The 42-inch pipe test set-up	15
Figure 3-4. The access hole opening for testing	16
Figure 3-5. View of manhole for 20-inch pipe	17
Figure 3-6. Simulated 20-inch sewer (riser-pipe-manhole shown from right to left)	18
Figure 3-7. Reference depth  point gauge	20
Figure 3-8. ADS flow monitoring sensors (laboratory  10-inch pipe installation)	22
                                           v

-------
Figure 3-9. ADS flow monitoring sensors (laboratory 42-inch pipe installation)	23
Figure 3-10. Silt simulation using a fixed bed (looking downstream)	26
Figure 3-11. ADS depth sensor shown during the 0.5 mm grease test	27
Figure 3-12. Sensor installation locations	32
Figure 3-13. Metered flow rate versus reference flow rate	36
Figure 3-14. ADS 3600 data summary, 10-inch pipe, 0-600 gpm reference flow	39
Figure 3-15. Scatter-graph for 10-inch pipe test at 0.1 percent slope	40
Figure 3-16. Scatter-graph for 10-inch pipe test at 0.5 percent slope	41
Figure 3-17. Scatter-graph for 10-inch pipe test at 1.25 percent slope	42
Figure 3-18. Scatter-graph for 10-inch pipe test at 2.0 percent slope	43
Figure 3-19. Deviation of meter flow to reference flow for 10-inch pipe	44
Figure 3-20. Plot of reference flow versus meter flow in 10-inch pipe	45
Figure 3-21. Scatter-graph for 20-inch pipe test at 0.1 percent slope	46
Figure 3-22. Scatter-graph for 20-inch pipe test at 0.5 percent slope	47
Figure 3-23. Scatter-graph for 20-inch pipe test at 1.25 percent slope	48
Figure 3-24. Scatter-graph for 20-inch pipe test at 2.0 percent slope	49
Figure 3-25. Deviation of meter flow to reference flow for 20-inch pipe	50
Figure 3-26. Plot of reference flow versus meter flow in 20-inch pipe	51
Figure 3-27. Scatter-graph for 42-inch pipe test at 0.2 percent slope	52
Figure 3-28. Scatter-graph for 42-inch pipe test at 0.2 percent slope with silt	53
Figure 3-29. Scatter-graph for 42-inch pipe test at 0.2 percent slope with 0.5mm grease	54
Figure 3-30. Scatter-graph for 42-inch pipe test at 0.2 percent slope with 2.0mm grease	55
Figure 3-31. Deviation of meter flow to reference flow for 42-inch pipe	56
Figure 3-32. Plot of reference flow versus meter flow in 42-inch pipe	57
                                            VI

-------
                          Acronyms and Abbreviations

ADS         ADS Environmental Services, a division of ADS Corporation
BPR         BPR, Quebec, Canada
cfs           Cubic feet per second
ENS         Event Notification System
EPA         United States Environmental Protection Agency
ETV         Environmental Testing Verification
ft            Foot or feet
FTO         Field testing organization
gpm         Gallons per minute
in.           Inch or inches
Ib            Pound
LTO         Laboratory testing organization
mg/L        Milligrams per liter
min          Minimum
NIST        National Institute of Standards and Technology
NSF         NSF International (formerly National Sanitation Foundation)
NTU        Nephelometric turbidity units
QA          Quality assurance
QUC        Quebec Urban Community
SAG        Stakeholders Advisory Group
SCADA      Supervisory control and data acquisition
UWRL       Utah Water Research Laboratory
VTP         Verification test plan
WWF        Wet weather flow
                                         vn

-------
                                 Acknowledgments

The Laboratory Testing Organization, Utah Water Research Laboratory, and the Field Testing
Organization, BPR, Quebec, Canada,  were responsible for all elements in the testing sequence,
including test set-up, calibration and verification of instruments, data collection and analysis,
data management, data interpretation, and the preparation of this report.

       Utah Water Research Laboratory
       Utah State University
       8200 Old Main Hill
       Logan, UT 84322-8200
       Contact Person:  Steven L. Barfuss

       BPR
       4655, blvd Wilfrid-Hamel
       Quebec, Qc, Canada, GIF 4J7
       Contact People:  Denis Simard or Genevieve Pelletier

The manufacturer of the equipment:

       ADS Environmental
       5030 Bradford Dr.
       Bldg. 1,  Suite 210
       Huntsville, AL 35805
       Contact Person:  Eugene C. Cullie

The Utah Water Research Laboratory  wishes to thank the UWRL shop for the many hours spent
preparing the piping for these tests. Also, special thanks to Utah State University Engineering
students Randy  Geldmacher,  Tyler  Smith, John Nunley,  and  Garrett McMullen  for  their
assistance in establishing proper test conditions and collecting test data. It is also necessary to
thank ADS personnel for their support throughout the test program. ADS support was provided
by George Kurz, Pat Stevens, Keith Waites, Christy Kennamer, Heather Hackett, Jeffrey White,
Erica Blanken, Mark MacPherson, and Gillian Woodward.
                                          Vlll

-------
                                      Chapter 1
                                    Introduction

1.1    ETV Purpose and Program Operation

The  U.S. Environmental Protection Agency (EPA)  created  the  Environmental Technology
Verification  (ETV)  Program  to   facilitate  the   deployment  of innovative  or  improved
environmental technologies through  performance verification and dissemination of information.
The goal of the ETV program is to further environmental protection by substantially accelerating
the acceptance and use of improved and more cost-effective technologies. ETV achieves 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; stakeholder
groups, which consist 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 (as appropriate),  collecting and analyzing  data,  and
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.

NSF  International (NSF)  operates  the  Wet Weather  Flow  (WWF)  Technologies  Area in
cooperation with the EPA under the  Water Quality Protection Center. The WQPC evaluated the
performance of the ADS Model 3600 open channel flow monitor, which is an area/velocity flow
meter used to measure flows  in municipal sewers. The performance  claim evaluated during
laboratory testing of the ADS  Model 3600 was  that the instrument is capable of  measuring
depths and velocities in a wide range of pipe sizes and flow conditions. This document provides
the laboratory verification test results for the ADS Model 3600 open channel flow monitor.

1.2    Testing Participants and Responsibilities

The ETV testing of the ADS 3600 Open Channel Flow Monitor was a cooperative effort of the
following participants:

   •   U.S. Environmental Protection Agency
   •   NSF International
   •   Utah Water Research Laboratory
   •   BPR
   •   ADS Environmental

The following is a brief description of the ETV participants and their roles and responsibilities.

-------
1.2.1   U.S. Environmental Protection Agency

The EPA Office of Research and Development, through the Urban Watershed Branch, Water
Supply  and Water Resources Division,  National  Risk Management  Research Laboratory
(NRMRL), provides administrative, technical, and quality assurance guidance and oversight on
all ETV Water Quality Protection Center activities.

Contact Information:
       EPA, NRMRL
       Urban Watershed Management Research Laboratory
       2890 Woodbridge Avenue (MS-104)
       Edison, NJ 08837-3679
       Phone: (732)321-6627
       Fax:  (732)321-6640
       Contact Person: Raymond Frederick, Project Officer
       Email:  Frederick.Rav@epamail.epa.gov

1.2.2   NSF International

NSF is a not-for-profit testing and certification organization dedicated to public health,  safety,
and protection of the environment. Founded in 1946 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. NSF also provides testing and certification services to ensure that
products bearing the NSF Name, Logo and/or Mark meet the organization's standards.

NSF had several different roles in completing this verification.  NSF reviewed the verification of
the verification test plan (VTP) and provided technical oversight of the verification testing,
including  auditing of the  laboratory's analytical, data gathering, and data recording procedures.
NSF also provided review of this verification report.

Contact Information:
       NSF International
       789 North Dixboro Rd.
       Ann Arbor, MI 48105
       Phone: 734-769-8010
       Fax:  734-769-0109
       Contact:  Tom Stevens, Program Manager
       Email:  stevenst@nsf.org

1.2.3   Laboratory Testing Organization

The Utah Water Research Laboratory (UWRL),  a Utah State University hydraulic research
facility, is an NSF-qualified  testing  organization (TO)  for  the  WQPC and  conducted the
verification testing of the ADS 3600 flowmeter. UWRL provided all needed logistical support,
established a communications network, and scheduled and coordinated the activities of all
participants. It was responsible for ensuring that the testing location and feed water conditions

-------
could meet the stated objectives of the verification testing. UWRL prepared the VTP; oversaw
the pilot testing; managed, evaluated, interpreted and reported on the data generated by the
testing; and evaluated and reported on the performance of the technology.

UWRL employees manufactured and prepared the test piping, set test conditions, and measured
and recorded data during the testing. The UWRL's Project Manager provided testing oversight.

Contact Information:
      Utah Water Research Laboratory
      Utah State University
      8200 Old Main Hill
      Logan, UT 84322-8200
      Phone: (435)797-3214
      Fax: (435)797-3663
      Contact Person: Steven L. Barfuss
      Email: Barfuss@cc.usu.edu

1.2.4 Field Testing Organization

BPR is  an NSF-qualified TO for the WQPC and was responsible for  conducting  the  field
verification testing of the ADS 3600 flowmeter.  The testing was conducted on a section of the
Quebec Urban Community's (QUC) western sewer network located in the City of Sainte-Foy,
along the east side of Boulevard Chaundiere,  approximately between Bombardier and Mendel
Streets.

BPR  provided  all needed logistical  support, established  a communications network,  and
scheduled and coordinated the activities of all participants. It was responsible for ensuring that
the testing location and feed water conditions could meet the stated objectives of the verification
testing. It prepared the VTP, oversaw the pilot testing; managed, evaluated, interpreted, and
reported on the data generated by the testing; and  evaluated and reported on the performance of
the technology.

BPR employees prepared the test site, set test conditions, and measured and recorded data during
the testing. BPR's Project Manager provided oversight of the daily tests.

Contact Information:
      BPR-CSO
      4655, Boulevard Wilfrid-Hamel
      Quebec, QC, Canada GIF 4J7
      Phone: (418)871-3414
      Fax: (418)871-9569
      Contact Persons:  Denis Simard or Genevieve Pelletier
      E-mail: dsimard@bpr-cso.com or gepelletier@bpr-cso.com

-------
1.2.5   Vendor

The flow monitoring equipment is manufactured by ADS Corporation. ADS was responsible for
supplying a field-ready open channel flow meter  equipped with all necessary components,
including  an installation and operation manual.  The manufacturer was also responsible for
providing technical  support  personnel. This  individual was responsible  for  installing and
precalibrating the flow monitoring equipment in the simulated sewer for each test series, and was
available during all tests to provide technical assistance as needed.

Contact Information:
       ADS Environmental
       5030 Bradford Dr., Building One, Suite 210
       Huntsville, AL 35805
       Phone: (256)430-3366
       Fax:(256) 430-6633
       Contact Person: Eugene C. Cullie

1.3    Laboratory Verification Testing Site

The Utah Water Research Laboratory, constructed in 1965, is one of the largest water research
laboratories of its kind in the country. The laboratory, which  occupies more than 50,000 square
feet of floor space, contains  a variety of flumes, channels,  pumps, pipelines, equipment, and
instrumentation for conducting hydraulic research, model studies, hydraulic valve testing, and
flow meter calibrations. It is capable of performing an array of hydraulic tests and research
programs. A network of steel piping (18-, 24-, 36-, and 48-inch diameter), located under the floor
of the  lab, provides maximum flexibility in constructing test lines from  1/2-inch to 60-inches.
Under-the-floor channels conduct water from the experiments back to the river, to recirculating
pumps, or to the precise flow measurement facilities.

The primary water supply for the laboratory is  an  85-acre-foot reservoir approximately 500 feet
from the facility. The water supply, which is conveyed to the  laboratory through a 48-inch pipe,
provides a constant 25 feet of head to the main level of the laboratory and 35 feet of head to the
lower level at  shutoff. Flow can also be supplied from either  high-pressure pumps or a constant
level tank within the laboratory.

For direct National Institute of Standards (NIST) traceability, flow rates are measured  using the
laboratory's weigh and  volumetric tanks. These tanks are calibrated regularly and are NIST-
traceable by weight. The UWRL primary flow measurement system consists of one 1,000 Ib
weigh tank, two 30,000 Ib weigh tanks, and a 222,144 Ib (3,560 ft3) volumetric tank. Flow rates
up to  40 cubic feet per second (cfs) can be  accurately  measured  with the  primary  flow
measurement system.

A number of venturi meters—6-, 12-, 20-, 24-, and 48-inch—are  installed  in various  test lines
throughout the laboratory. These master meters are calibrated in place using the UWRL primary
flow measurement system. As a  secondary standard with NIST-traceability, these meters allow
accurate flow measurements to be made for flow rates up to approximately 200 cfs.

-------
                                      Chapter 2
              ADS Equipment Description and Operating Processes

The  information contained in  this chapter is provided by the vendor and does not represent
verified information. The information is intended to provide the reader with a description of the
ADS 3600 flow monitor and to explain  how the technology operates.  The verified performance
characteristics of the ADS 3600 are described in subsequent chapters.

2.1    Equipment Description

The  ADS 3600 flow monitor is specially designed to meet the demands of measuring flow over
long-term periods in open channel applications, such as wastewater collection systems, storm
sewer  systems, and combined sewer  systems. The  battery-powered system acquires  highly
accurate depth and velocity data,  and transmits the data through telemetry to a remote computer.
The  ADS 3600 is certified as Intrinsically Safe (IS) by Factory Mutual (United States), and by
SIRA Certification Services (International). IS  certification confirms that it meets both U.S. and
international standards for electrical apparatus installed in potentially explosive environments.

The ADS 3600 is used for many project  applications, including:

   •  Infiltration/inflow analysis and reduction;
   •  Master plan studies;
   •  Interagency billing networks;
   •  Combined sewer overflow characterization;
   •  Storm sewer monitoring;
   •  Sewer capacity analysis; and
   •  Sewer system performance studies/trending.

Precise pipe dimensions (height  and width)  are measured during installation because they are
critical to both depth and cross-sectional area. ADS  depth sensors measure the distance from the
down-looking sensor to the water surface (range) and depth of flow is obtained by subtracting
the range from the diameter. This is an important step since the inside diameter of most sewer
pipe is  not equal to its nominal  diameter. Velocity sensor confirmation is another critical-to-
quality  activity. While the ADS  3600 velocity sensor will perform in most sites using factory
default  settings, it can be  adjusted by  a trained technician to improve  performance in unique
hydraulic conditions.

The ADS 3600 features:

•    Quad-redundant Ultrasonic  Level  Sensor. This nonintrusive, zero-drift sensing  method
     results in a stable, accurate, and reliable flow depth calculation.
•    Peak Velocity Sensor. Readings from this  sensor are used to calculate average velocity. Its
     miniature size minimizes fouling and flow disruption.
•    Pressure Depth Sensor. This sensor is used to measure surcharge levels or to provide  a
     redundant depth reading when used in conjunction with the ultrasonic level sensor.

-------
•   Water Quality Sampler Interface. A variety of industry standard water quality samplers are
    compatible and easily  interfaced with the Model 3600. Sampling is initiated automatically
    based on a fixed-time or flow proportional basis. The sampler must be certified to maintain
    IS certification.
•   Tipping  Bucket  Interface. Where  installation criteria allow,  the  flow monitor can be
    interfaced to a tipping bucket unit to record rainfall amounts.
•   Supervisory Control and Data Acquisition (SCADA) System Interface. An external modem
    unit can communicate  both with a SCADA system remote computer and via telemetry with
    QuadraScan™ Software.
•   QuadraScan™  Software. This enables a SCADA system to retrieve various data entities,
    including depth of flow, average velocity, flow rate, daily flow totals, etc.
•   Accurate Flow Measurement. This technology has been shown to be highly accurate in both
    laboratory and in-field tests. It measures flow under  open channel, free flow conditions, and
    non-free flow conditions, including surcharge and backwater.
•   QuadraScan™ Software Interface. This complete hydraulics and analysis package provides
    telemetered data collection, trending, reporting, and file management.
•   Automated Data  Collection.  Data can be collected  at user-defined  times, including group
    collections of multiple monitoring units,  increasing the  efficiency  and  flexibility of flow
    data analysis.
•   Event  Notification  System  (ENS)  Capability.  Consisting of  the  ADS  3600  Monitor,
    QuadraScan™ Software, and event reporting software, this versatile system can initiate an
    alarm  to a  remote  location  and/or activate  other instrumentation  activity such  as  water
    quality sampling based on preset conditions.
•   Low Cost Telemetry. The standard telemetry interface allows remote communications and
    diagnostics, resulting in a significant reduction in field labor.
•   Manufacturing Quality Control. Each unit undergoes  factory testing and  certification by
    qualified technicians.
•   Ultra-low Power  Consumption. Batteries can last more than one year under most operating
    and environmental conditions.
•   Fully Water-Resistant Housing. The housing reliably withstands harsh sewer environments,
    even under surcharge conditions.
•   Low Battery Warning. Each unit monitors battery voltages to warn the operator in advance
    of battery failure.
•   Optional Non-IS  Configuration. This unit is suitable for installations where  protection from
    explosive environments is not required.

Figure 2-1 is a view of the  ADS 3600 meter's  canister. Figure 2-2 shows the stainless steel ring
and sensing devices installed inside a pipe.

The ADS meter installation is similar to nearly all depth velocity meters, but  the ADS  meter
differs in that only its velocity sensor is mounted in the flow. The velocity sensor offers a cross-
sectional area of around one-half square inch,  and measures velocity in depths of less than one
inch. Both the depth and velocity sensors are mounted on a stainless-steel ring that is expanded
into place with a hand  crank  and spreader bars.  The meter  is either attached  to rungs in the
manhole or to an attachment hook drilled into place on the manhole wall. The meter is activated

-------
by downloading BASIC  code with monitor and  sensor configuration.  The monitor  clock is
synchronized with the central computer each time data are collected.
Figure 2-1. ADS Model 3600 Open Channel Flow Monitor.

It is important that the  actual pipe dimensions be measured for two reasons: (1) the correct
diameter will be used to calculate cross-sectional area of the pipe; and (2) the depth of flow
equals the actual diameter minus the range measured by the depth sensor. An incorrect diameter
translates directly to an incorrect depth. The inside diameter of most sewer pipe is not equal to its
nominal  diameter. For example, a  15-inch PVC sewer line with SDR 35 wall thickness has a
maximum manufactured diameter of 14.4 inches.

The monitor and software manuals offer  approximately  600 pages of instruction, including
graphics, on installing the meter, activating the meter, collecting data, and processing data.

ADS  provides hardware as well  as flow  monitoring services throughout the United States.
Technical support is provided by its headquarters office  as well as by 16 regional and local

-------
offices. Most offices are staffed with data analysts and operations people who will readily assist
with any hardware, software, or operational problem. Parts can be ordered through any office.
Figure 2-2. ADS flow monitoring sensors (laboratory 20-inch installation).
No special tools are required for installation. Most installations are accomplished with standard
wrenches, screwdrivers, wire ties, and a drill and bits.

Once placed in service, continued  calibrations are not required to obtain valid data, but periodic
site visits are recommended to spot hydraulic changes. The ultrasonic sensor has zero drift, but it
is recommended that the sites with silt be visited regularly to verify silt depth. Silt reduces the
cross-sectional area of flow and creates inaccuracy in the continuity equation. Hydraulic changes
can  also affect the average-to-peak ratio, which should be confirmed  periodically through
velocity profiles.

The unit operates with little routine maintenance. The most  common maintenance needed  is
cleaning the ultrasonic depth  sensor should it become coated with grease during a surcharge
event in  a greasy sewer. In most sewers the ultrasonic  sensors are not affected by  surcharging,
and  they immediately begin reporting depth the moment surcharging subsides. The velocity

-------
sensor is not affected by grease, but can be impaired by heavy silt or debris covering the sensor.
In such cases, the sensor can be rotated off the bottom of the sewer to avoid silt and debris.
Ultrasonic  depth sensor crystals that deteriorate with age can be diagnosed and taken out of
service remotely. A visit to the site can be avoided by  switching to an inactive sensor pair. There
are 12 possible sensor pairs, and the meter will operate with only one pair in service.

2.2    Operating Process

2.2.1   Depth

Depth is measured by a quad-redundant ultrasonic sensor installed at the top of the pipe facing
down toward the water surface. The meter measures the time of travel  of an  ultrasonic pulse
from  the sensor to  the water and back to the sensor. The meter converts the time of travel to
distance by calculating the speed of sound through air and adjusting for temperature, which is
measured by two temperature sensors inside the ultrasonic sensor head.
Each  of the four transducers can operate in either transmit or receive mode,  allowing for 12
possible transmit-listen pairs, as illustrated in Figure 2-3. A crystal cannot transmit and listen for
its  own signal.  The paired sensors  essentially  eliminate  the dead  zone inherent to  a single
ultrasonic crystal that transmits and then listens. The dead zone is usually less than one inch.
Ultrasonic Sensor Pairs
40 kHz crystals
can transmit
and receive.
^
k.

oo
oo


Pair crystals
2 0^1
3 02
4 03
5 1 0
7 1-2
8 1 3
9 20
10 2-1
12 2-3
13 3-0
14 3-1
15 3-2
16 3-3
pages
Figure 2-3. Ultrasonic sensor illustration.
Up to four of the 12 sensor pairs are in operation at any moment, and the operator can remotely
diagnose the operating pairs for strength and quality of signal. The operator can remotely add
and  remove  sensor pairs from operation. When activated  for  a reading, each pair  in  turn
measures  distance 32 successive  times for a total of 128 readings.  Errant readings from each
sensor pair are discarded and four depth readings are recorded. The sensors are designed to have
zero drift, where only extremely noisy or wavy sites affect performance.

-------
2.2.2   Velocity

Doppler  velocity measurements are made by transmitting an ultrasonic  signal upstream and
measuring particle velocity, similar to police radar. The sensor receives echoes from the particles
and records the frequency shift (velocity)  and strength of each echo. The ADS 3600 uses Peak
Velocity  Doppler, which is ADS's third generation velocity technology (V3).  V3 technology
takes advantage  of the fact that the fastest particle in sewage remains constant from moment to
moment,  regardless  of its  size. V3  technology measures  the velocity of the  fastest particle
(particle  C in Figure 2-4) in sewage and converts it to average velocity. The ratio of average to
peak velocity is  around 0.9 in most sewers, and velocity profiles  are used to determine the ratio
in unusual flow.
                         Doppler Sensors Measure Velocity
                            and Echo Strength of Particles
                               Od
                                                Echo Strength
                             top view
                           czoe Od
                          Side view of sensor in sewer
                                                              Velocity
                                                               page 46
Figure 2-4. Doppler velocity sensor illustration.
                                           10

-------
                                      Chapter 3
                                 Laboratory Report

This chapter presents the procedures used in generating the laboratory performance data for the
ADS  3600, along with  the  test data generated from  the verification  testing.  This  chapter
addresses only the laboratory portion  of the verification. Chapter 4, contained in Part II of the
report, addresses the field testing portion of the verification testing.

3.1    Test Set-Up, Test Equipment, and Procedures

3.1.1   Test Description

The laboratory flow meter verification was performed  in three  different nominal  pipe  sizes:
10-inch, 20-inch, and 42-inch. Figures 3-1  through 3-3 show the piping prepared for the tests in
the three respective sizes. Although the schematic drawings are not to scale, the dimensions and
the relative locations of flow meters, risers, valves, and manholes  are accurate. Each set-up was
constructed using steel pipe.

An opening was cut in the top of the pipe near where the vendor test sensors were installed. The
opening provided access to the  sensor location so that a  precision point gauge could be used to
accurately measure flow depths  and to allow test participants to view the installation from above
and to observe flow conditions around the flow meter. The access opening  had a watertight cover
that was  closed during pressurized (full-pipe) tests. Figure 3-4 shows the  20-inch pipe with the
access hole open and closed. The other pipe sizes were similarly constructed.

A manhole was also constructed in each test set-up near the location of the test flow meter. The
size and location of the manhole was  sufficient to provide access  for the  installation of the test
flow meter, and to provide a suitable location  for monitoring surcharge flow conditions. The
manhole  consisted of a  cylindrical steel  tank  with a watertight bottom. Each manhole was
constructed to the dimensions indicated in the figures. The section of pipe passing through the
manhole  had its top removed (down to the  spring-line). A sloping  steel floor was installed in the
bottom of the  manhole, so  that all  water drained to  the pipe  spring-line.  Figure 3-5  is  a
photograph of the outside and inside of the manhole in the 20-inch pipe set-up. The other pipe
size set-ups were similarly constructed.

The test  flow meter sensors were installed in the upstream  pipe  adjacent to the manhole. The
precise location of the meter in the test pipe depended on ADS's specifications for the meter and
its installation requirements. The crown of the pipe was removable near the installation location
for access to the installed flow meter, but this access  location was not used to install the flow
meter. The access hole was covered and sealed for tests requiring surcharged flow conditions.
All cables and wires connected to the sensor were  directed  into the test  pipe through  the
manhole.

In the 10-  and 20-inch  test pipes,  the  length of straight pipe installed  upstream of  the
manhole/flow meter test location was at least 40 times the pipe diameter. This length of straight
pipe was necessary to provide near-uniform flow to the flow monitoring equipment during tests
                                           11

-------
with no backwater effects. In the 42-inch test pipe, the length of straight pipe upstream of the
manhole/flow meter test location was at least 22 times the pipe diameter. This shorter length was
sufficient to provide near-uniform flow at the single pipe slope (0.2 percent).
                                            12

-------
Figure 3-1. The 10-inch pipe test set-up.
                                          13

-------
                                                                                 CL
                                                                                 CD
                                                                                 00
                                                                                 OT
                                                                                 CU
                                                                                 CJ
                                                                                 C

                                                                                 T
                                                                                 O
                                                                            I 8
Figure 3-2. The 20-inch pipe test set-up.
                                            14

-------
                                                                             u
                                                                             tz
Figure 3-3. The 42-inch pipe test set-up.
                                           15

-------
                                    (b) Open pipe.
Figure 3-4. The access hole opening for testing.
                                         16

-------
                                    (a) Outside view.
                                     (b) Inside view.
Figure 3-5. View of manhole for 20-inch pipe.
                                          17

-------
A minimum of  10  diameters of straight pipe was installed downstream of each manhole to
simulate the flow conditions exiting the manhole. Figure 3-6 shows the 20-inch set-up. The other
pipe size set-ups were similarly constructed.
Figure 3-6.  Simulated 20-inch sewer (riser-pipe-manhole shown from right to left).

Each test line was laid at a constant slope along the full length of the pipe.  The test set-up
allowed for the adjustment of pipe slope in the 10- and 20-inch sizes.  The supports for these two
test lines were capable of supporting pipe from a slightly negative slope condition to a maximum
slope of 2.0 percent. The 42-inch pipe set-up was set at a constant slope (0.2 percent) for all tests.
The vertical pipe supports prevented the pipe from  sagging under the weight of the water. The
spacing between vertical pipe supports for the  three steel pipe set-ups was no greater than ten
pipe diameters. No horizontal supports were needed.

A control valve was placed both upstream and downstream of the model piping. The valves were
used to control the rate of flow through the simulated sewer.  The upstream valve was used to
control the rate of flow entering the riser supplying the test line. The downstream valve was used
to impose a  downstream control  to the  test line,  providing  backwater during surcharge test
conditions.

A vertical riser was installed upstream  of each test pipe to  help dissipate the energy of the
incoming flow, to provide a smooth pour-over into the model pipe for open channel tests, and to
                                           18

-------
provide a surcharge stand-pipe during full-pipe tests. The dimensions of each riser are shown in
Figures 3-1 through 3-3.

The test pipe was connected to the vertical supply riser with a flexible coupling that allowed for
the required adjustment of pipe slope. The point of connection was high enough on the riser to
allow the test  pipe  to be set at a maximum slope  of 2.0 percent along its entire length. (The
coupling can be seen on the right side of the photograph in Figure 3-6. In the photograph, flow is
from right to left.) The elevation of the pipe at the coupling was always the same. The adjustable
stands were raised or lowered to set the appropriate slopes for each test  so that the entire test line
rotated about the upstream coupling.

Water was  supplied by a reservoir near the laboratory. This constant head source was capable of
maintaining constant and steady water depths in the test pipe for the duration of each test run. A
turbidity test was performed in the Environmental Quality Laboratory at UWRL and found to be
at a level of 0.58  nephelometric turbidity units (NTU). One NTU is the limit for public drinking
water  supplies.  The  lower limit on the  laboratory  meter  that made the measurement is
approximately 0.1 NTUs. A turbidity measurement of 0.58 NTU  indicates water with high
clarity.

3.1.2   Reflectors

The Doppler shift principle  requires particles in the water to serve as reflectors for sound waves.
The clear water used in the UWRL contains considerably fewer particles than sewage. Since it
was not practical to replicate the particulate concentrations of normal sewage, coffee creamer
was added to the water in small quantities on some test runs. The vendor maintains that although
the coffee creamer additive  did provide a level of reflectivity, the particulate concentration in the
test water did not replicate that of sewage, and thus could be  a source  of measurement error. It
was not necessary to add creamer while the laboratory was making reference measurements.

For test configurations where particulate levels were  especially low, ADS technicians made
compensations  during confirmations  by changing  adjustable parameters on  the  ADS  3600
velocity subsystem. This sensitivity  adjustment was made using the  Fieldscan standard field
configuration software. It allowed the meter to  register the faint echoes from the few particles
that were in the fastest portion of the flow, and adequately compensated for the clarity issue. In
typical sewer installations, this type of adjustment is not  necessary.

3.1.3   Laboratory Test Instrumentation

The laboratory instruments used during the verification  tests are described  in  this section.
Calibration records for the instruments are shown in Appendix  A.

3.1.3.1 Flow Measurement Tanks and Calibrated Flow Meters

Laboratory weigh tanks and master venturi meters were used to determine the reference flow rate
for each test run. The measurement tanks are directly traceable to the National  Institute of
Standards and Technology (NIST) by weight.  The master venturi flow meters are regularly
                                           19

-------
calibrated and are NIST-traceable. Uncertainty for the tanks and flow meters is less than 0.25
percent.

3.1.3.2 Precision Point Gauge

A precision point gauge was used as the reference depth measurement in determining the depth
of flow near the test flow meter focal point (Figure 3-7). The precision point gauge is readable to
the nearest one-thousandth (0.001) of a foot. The point gauge was mounted as close as possible
to the downstream edge of the access hole. Likewise, the ADS depth sensor was also mounted as
close as possible to the downstream edge of the access hole. The reference depth measurement
was compared to the indicated depth measurement from the meter.

3.1.3.3 Thermometer

A calibrated thermometer was used to measure the temperature  of the water flowing through the
test  pipe. The temperature measurement was used  for  spreadsheet  calculations requiring
temperature. The thermometer was calibrated for accuracy before the verification tests.
Figure 3-7.  Reference depth point gauge.
                                          20

-------
3.1.3.4 Timer

A calibrated stopwatch/timer was used to measure the collection time of the water entering the
weigh tanks.  The time measurement was used with the weight reading to calculate the actual
flow rate for  each test in which the master venturi  meters  were not used. The  timer was
calibrated before the verification tests.

3.1.3.5 Precision Calipers

Precision calipers were used to measure the inside diameter of the model piping and to verify the
roundness of the pipe. An accurate measurement of each pipe diameter was necessary for correct
area calculations.

3.1.4   Pretest Procedures

The VTP included a detailed  set of test procedures that was followed during the verification
tests.  The procedures were consistent with the  requirements established  in  the protocol.
Laboratory personnel performed  the following tasks before beginning the verification test on a
specific pipe set-up:

   •   The geometry of the test pipe and location of the vendor instrument within the pipe was
       measured and documented.
   •   A digital photograph was taken  of the installed flow meter sensors in each pipe size
       (Figures 2-2, 3-7, 3-8, and 3-9).
   •   The time required for flow meter installation and set-up by the vendor was recorded.

3.1.5   General Test Procedures

Flow, depth, and velocity data from the test flow meter was logged electronically as recorded by
the  vendor-supplied electronics. Average recorded  values were  entered into a  computer
spreadsheet.

Tables  3-1, 3-2, and  3-3  define  each  run  that  was  tested in 10-,  20-,  and  42-inch  pipe,
respectively.

The  procedures described in  this  section were  conducted for  each set  of test  conditions
established in Tables  3-1  through 3-3.  The procedures were repeated each time the flow
conditions were changed. The following procedural steps were taken for each run.

3.1.5.1 Set Flow Condition

Both the upstream and  downstream control valves were used to set the flow.  Uniform flow (free
flow) tests had no downstream control. Conversely, the downstream control  valve was throttled
for  submerged and backwater  tests. Tables 3-1 through 3-3 list the target flow conditions for
each run.
                                           21

-------
Figure 3-8. ADS flow monitoring sensors (laboratory 10-inch pipe installation).

3.1.5.2 Allow Flow To Stabilize

Test measurements  were not made until  the water had  stabilized in the pipe. The flow was
stabilized  when the depth in the pipe  did not change over time. The precision point  gauge
mounted on the centerline of the flow path acted as a gauge for setting and stabilizing the flow.
The flow depth was set within .02 D (D = diameter of the pipe) of the flow depths specified in
Tables 3-1 through 3-3.

3.1.5.3 Measure Water Temperature

The temperature  of the water in the  test pipe  was measured using a calibrated mercury
thermometer.  This manual  reading was recorded after the mercury  in the thermometer had
stabilized.

3.1.5.4 Measure Reference Flow

The reference flow rate  was measured and recorded utilizing the laboratory instrumentation
before and after each logging period. Laboratory personnel measured the actual flow rate using
the laboratory master flow meters.  The  master  venturi  flow meters are used often and are
regularly calibrated  at the laboratory. Flows too small to  be accurately measured by the master
venturi meters were measured using the laboratory weigh tanks.
                                          22

-------
Figure 3-9. ADS flow monitoring sensors (laboratory 42-inch pipe installation).
3.1.5.5 Measure Reference Depth

The actual  depth was measured and recorded before and after the logging period of each run.
Laboratory personnel measured the actual  flow depth at the centerline of the pipe. This
measurement was made very near the focal point for the sensor depth measurement. The depth
was  determined by  taking the difference between the water surface  measurement and the
reference pipe invert measurement made by the centerline point gauge.

If the water surface was mounded slightly at the measurement  location, a measurement of the
complete cross-sectional water surface profile was necessary to generate the correct flow area for
mean velocity calculations. An evaluation of each water surface profile was performed to decide
the necessity of conducting either a  single centerline depth or a complete cross-sectional water
surface profile. This check was made by comparing the  depths at the centerline and near the wall
of the pipe.
                                          23

-------
If the peak-to-peak difference in depth across the water surface profile was greater than 0.02 D
inches, a complete cross-sectional water surface profile was necessary. A special adjustable point
gauge was used to measure water surface depths at five equally spaced locations across the water
surface. Since the water surface width changes with increased depth, the spacing and position for
the five equally spaced point gauges also varied.

3.1.5.6 Log Meter Data

Average flow, depth, and velocity measurements were electronically logged and recorded from
the ADS output device. ADS provided a laptop computer to retrieve the logged data. The logged
data was manually entered into a desktop computer spreadsheet at the end of each day of testing.

Once the flow stabilized, and after recording the first reference depth and flow measurements,
laboratory personnel logged data from the flow monitoring equipment. Data was recorded in
accordance with the operational procedures provided by the vendor, except that average flow
rate,  depth, and velocity readings were logged and recorded at one-minute intervals over a five-
minute period. Data was reported for each one-minute interval. ADS prepared the instrument so
that each one-minute sample was  independent and  was not averaged with prior samples. The
operational procedure and data-logging method devised and  utilized by  ADS is  given  in
Appendix C.

3.1.5.7 Measure Reference Flow and Depth (Second Time)

After the ADS flow meter completed its  logging  cycle (during the five-minute period),  the
reference flow and reference depth measurements were repeated. This was done to ensure that
the flow conditions had remained constant during the five-minute logging period.

3.1.5.8 Calculate Reference Velocity

Using the measured cross-sectional water surface profile, the flow area was  calculated using a
computer spreadsheet. The mean velocity  for the test was calculated by dividing the flow area
into the reference flow measurement. The mean velocity was based on the average reference
depth and flow measurements for the run.

3.1.5.9 Record Observations

The flow conditions for each run were recorded. Appendix B contains the raw data and notes
documented by laboratory personnel during the verification tests.
                                          24

-------
3.1.5.10      Review Reference Data

Once all pertinent reference data for a single run had been entered into the computer, the results
were reviewed before the flow conditions were changed for the next run. If flow conditions
changed during a logging period, the test run was repeated. A run was repeated if the difference
between the flow measurements taken at the start and end of the logging period were greater than
one percent of the  lesser value. Before repeating  any given  run, the vendor technician and
laboratory personnel came to agreement that flow conditions had changed.

After all measurements for the run had been made, the flow conditions were changed and a new
run was started. Each step listed above was repeated for each run.

3.1.5.11      Download Meter Data

The data was retrieved from the meter at the end of each test  day. The five flow rate, depth, and
velocity readings for each run were manually entered into a computer spreadsheet. The average
of all five one-minute samples was calculated and reported.

3.1.6  Test Conditions

3.1.6.1 Free Flow and Backwater Tests

Free flow  conditions were  established by  setting the  desired depth  in the  pipe  with no
downstream control. Partial backwater conditions were simulated  by closing the downstream
control valve in the pipe. Each  backwater test (nonuniform flow condition)  had a corresponding
uncontrolled flow test (uniform or free flow condition). The flow velocities  for the backwater
tests were  set to approximately one-half the flow velocities for the uncontrolled tests by using the
supply master venturi  device.  No backwater  tests were performed for  slopes greater  than
0.5 percent.

3.1.6.2 Full Pipe Tests (Manhole Surcharged)

For tests in which there was downstream control and the pipe  was full and pressurized, the
opening on the top of the pipe through which the sensor is accessed and point gauge readings
taken was  closed. The access opening was covered with a watertight lid (see Figure 3-7). During
these tests, the depth of water (hydraulic grade line) was not  measured. For runs where the pipe
was full during surcharged conditions, this report indicates the pipe diameter as the flow depth.

The reference flow measurement was divided by the pipe area to determine  mean velocity. Flow
and velocity were compared, but no comparisons of surcharged pressures were made. The ADS
output indicated that the depth of flow in the  pipe was equal to the diameter of the pipe. A
piezometric measurement of the submergence was  made and reported as  an indication of the
magnitude of submergence only, and not as a reference pressure measurement.
                                          25

-------
3.1.6.3 Silt Simulation Tests

A simulated deposit of silt or sediment in the bottom of the pipe was conducted during the
42-inch pipe tests. A rigid flat bed was installed in the bottom of the 42-inch pipe at a depth of
three inches. The simulated fixed sediment bed was constructed of wood, and extended five pipe
diameters at a depth of three inches. A 5:1 smooth sloped transition was installed at the leading
edge of the fixed bed to reduce turbulence.

ADS personnel adjusted installation of the flow  measurement device so that it sat above the
simulated silt bed. Runs were made as indicated in Table 3-3. Figure 3-10 shows the fixed bed in
the bottom of the 42-inch pipe.
Figure 3-10. Silt simulation using a fixed bed (looking downstream).
3.1.6.4 Grease Build-up Tests

Tests were performed with thin layers of grease (Crisco) placed on all submerged components of
the flow monitoring equipment. These  tests evaluated the ability of the meter to continue
functioning  accurately  when  grease  built up on  the  wetted components, and  also  allowed
technicians to observe  and record whether the  grease remained on the components. Both a
                                          26

-------
0.5-mm and 2.0-mm layer of grease was tested under three different flow conditions (six runs).
Runs were made as indicated in Table 3-3. Figure 3-11 shows the depth sensor with a thin layer
of grease applied. The grease thickness in the photograph is 0.5 mm thick. The 2.0-mm tests had
four times the thickness of grease, which was applied manually.  Figure 3-11  shows the flow
sensor after grease was applied.
Figure 3-11. ADS depth sensor shown during the 0.5 mm grease test.
3.1.6.5 Reverse Flow Tests

A reversed pressure flow condition was also tested.  This was accomplished by elevating the test
pipe to a slightly negative slope, installing the flow meter in the test line backwards, submerging
the manhole, and forcing water uphill. The sensor was installed on the downstream side of the
manhole (uphill side) and runs were made as indicated in Tables 3-1 and 3-2. Reverse flow tests
were performed on the 10- and 20-inch pipes, but not the 42-inch pipe.
                                           27

-------
Table 3-1. Test Conditions and Sequence: 10-inch Test Pipe
Run number
1
2
3
4
5
6
7
8
9a
10a
11
12
13
14
15
16
17
18
19a
20a
21
22
23
24
25a
26a
27
28
29
30
31a
32a
33b
34b
Pipe slope
(percent)
0.1
0.1
0.1
0.1
0.1
0.
0.
0.
0.
0.
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
.25
.25
.25
.25
.25
.25
2.0
2.0
2.0
2.0
2.0
2.0
-0.1
-0.1
Water depth
0.1 D
0.3 D
0.5 D
0.8 D
0.1 D
0.3 D
0.5D
0.8 D
-1.5 D
-2.5 D
0.1 D
0.3 D
0.5D
0.8 D
0.1 D
0.3 D
0.5D
0.8 D
-1.5 D
-2.5 D
0.1 D
0.3 D
0.5D
0.8 D
-1.5 D
-2.5 D
0.1 D
0.3 D
0.5D
0.8 D
-1.5 D
-2.5 D
-1.5 D
-2.5 D
Uniform flow
(downstream-
uncontrolled)
Yes
Yes
Yes
Yes






Yes
Yes
Yes
Yes






Yes
Yes
Yes
Yes


Yes
Yes
Yes
Yes




Nonuniform
flow
(downstream-
controlled)




Yes
Yes
Yes
Yes
Yes
Yes




Yes
Yes
Yes
Yes
Yes
Yes




Yes
Yes




Yes
Yes
Yes
Yes
    a Manhole is surcharged.
    bFlow meter installed backwards, pipe slope set negative, surcharged conditions.
                                              28

-------
Table 3-2. Test Conditions and Sequence: 20-inch Test Pipe
Run number
1
2
3
4
5
6
7
8
9a
10a
11
12
13
14
15
16
17
18
19a
20a
21
22
23
24
25a
26a
27
28
29
30
31a
32a
33b
34b
Pipe slope
(percent)
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
1.25
1.25
1.25
1.25
1.25
1.25
2.0
2.0
2.0
2.0
2.0
2.0
-0.1
-0.1
Water depth
0.1 D
0.3 D
0.5 D
0.8 D
0.1 D
0.3 D
0.5D
0.8 D
-1.5 D
-2.5 D
0.1 D
0.3 D
0.5D
0.8 D
0.1 D
0.3 D
0.5D
0.8 D
-1.5 D
-2.5 D
0.1 D
0.3 D
0.5D
0.8 D
-1.5 D
-2.5 D
0.1 D
0.3 D
0.5D
0.8 D
-1.5 D
-2.5 D
-1.5 D
-2.5 D
Uniform flow
(downstream-
uncontrolled)
Yes
Yes
Yes
Yes






Yes
Yes
Yes
Yes






Yes
Yes
Yes
Yes


Yes
Yes
Yes
Yes




Nonuniform flow
(downstream-
controlled)




Yes
Yes
Yes
Yes
Yes
Yes




Yes
Yes
Yes
Yes
Yes
Yes




Yes
Yes




Yes
Yes
Yes
Yes
   1 Manhole is surcharged.
   1 Flow meter installed backwards, pipe slope set negative, surcharged conditions.
                                              29

-------
Table 3-3. Test Conditions and Sequence: 42-inch Test Pipe



Pipe slope
Run number (percent) Water depth
1
2
3
4
5
6
7
8
9a
10a
llb
12b
13b
14C
15C
16C
17d
18d
19d
"Manhole is surcharged.
b Silt simulation test.
0 Grease build-up test with 0.
d Grease build-up test with 2.
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2

5-mm j
,0-mm :
0.1 D
0.3 D
0.5 D
0.8 D
0.1 D
0.3 D
0.5D
0.8 D
-1.5 D
-2.5 D
0.1 D
0.7 D
-1.5 D
0.1D
0.3D
0.8D
0.1D
0.3D
0.8D

grease layer.
grease layer.
Uniform flow
(downstream-
uncontrolled)
Yes
Yes
Yes
Yes









Yes
Yes
Yes
Yes
Yes
Yes


Nonuniform flow
(downstream-
controlled)




Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes








3.1.7  Data Management and Analysis

All raw data, notes,  observations, and test descriptions were recorded using Microsoft Excel.
Reference values for  flow, depth, and velocity were calculated for each test run. These reference
values provided the basis for determining the accuracy of the values from the flow meter output.

The reference flow value for each run was the arithmetic average of the flow at the beginning
and at the end  of logging period, as determined using the calibrated master venturi meter or
weigh tank.

The reference depth value for each run was the arithmetic average of the depth at the beginning
and at the end of the logging period, as determined using precision point gauges.
                                           30

-------
The  reference velocity value for each run was the arithmetic average of the velocity values
calculated from the depth and flow measurements taken at the beginning and end of the logging
period.

Meter output of flow, depth, and velocity was recorded in the Microsoft Excel spreadsheet for
each of the five one-minute samples. When no data was available, the one-minute sample field
was left blank. An average of the five one-minute samples was made to indicate the five-minute
average for the meter.

3.1.8  Quality Assurance

The Utah State University Research Foundation has an active quality control program at UWRL.
Flow meter manufacturers and  nuclear power plants regularly audit the laboratory for quality
assurance. Under the laboratory quality assurance program, equipment is calibrated for accuracy
on a scheduled cycle or occasionally  as-needed. Instrumentation calibrations  are normally
subcontracted to  outside organizations, although  some are performed using the  calibration
facilities at the laboratory itself. Where applicable, calibrations indicate traceability to NIST. All
laboratory instrumentation that was  used during this testing program was calibrated prior to
performing verification tests. Calibration sheets for all applicable instrumentation are provided in
Appendix A. The  laboratory  master venturi  flow  meters  were used  during  most of the test
program as the reference flow measurement.  These venturi meters  have uncertainties less than
0.25  percent, which falls within the desired accuracy for the measurement.

3.2    Test Results

3.2.1   Preliminary Test Measurements

Prior to each test  series the geometry of the test  pipe and  location  of the vendor instrument
within the pipe  were measured and documented. Figure 3-12 shows the size and location of the
access hole in each of the three test set-ups and establishes the location of the ADS stainless steel
mounting  ring  for  each  test  set-up.  Tables 3-4 through  3-6  summarize the  preliminary
information recorded for each pipe size. The reference inside-pipe-diameter measurements were
made using precision calipers at the location in the pipe  where the flow meter sensors were
installed. Measurements were made in four directions and averaged  (every 45 degrees).  Only
three measurements were made in the 10-inch pipe because  it was quite round. The time required
to install and set up the flow meter was also recorded.
                                           31

-------
Flow Direction
                    Dimension A is the distance from the access
                      hole to the opening inside the manhole.
                       Dimension B is the distance from the
                     opening inside the manhole to the leading
                            edge of the sensor band.
                    10-Inch
                    20-Inch
                    42-Inch
                                 Access Hole
                                       Manhole
Access Hole
A (inches)
B (inches)
Access Hole
A (inches)
B (inches)
Access Hole
A (inches)
B (inches)
7 5/8W x 5 7/8L
75/8
21/8
1313/16Wx11 7/8L
13
8 1/4
29 3/4W x 23 7/8L
161/2
11 1/2
                                                                       Meter Sensors
                  (Note: All dimensions are in inches)
    Figure 3-12. Sensor installation locations.
    Table 3-4. Preliminary ADS 3600 10-inch Pipe Test Measurements
      Meter (monitor) serial number:
      Ultrasonic serial number:
      Velocity sensor serial number:
      Pressure transducer serial number:
      Test size:
      Test dates:
      Sensor location U.S.  of manhole opening:
      Straight pipe U.S. of manhole:
      Straight pipe D.S. of manhole:
      Three pipe inside diameter measurement:
      Average pipe inside diameter:
      ADS pipe diameter measurement:
      Pipe invert point gauge reference:
      Time required to install the sensor:
      Time required to set up the sensor:
      Calibration performed by:
      Vendor technicians:
                       02527
                       16405
                       11652
                       8710
                       10-inch
                       5/2/01 to 5/11/01
                       2.125 in.
                       63.9ft
                       28.7ft
                       10.285  in., 10.29 in., 10.27 in.
                       10.282  in.
                       10.250  in.
                       1.731ft
                       15 min
                       90min
                       Steven  Barfuss, Randy Geldmacher, Tyler Smith
                       Keith Waites, Christy Kennamer
                                                  32

-------
Table 3-5. Preliminary ADS 3600 20-inch Pipe Test Measurements
 Meter (monitor) serial number:
 Ultrasonic serial number:
 Velocity sensor serial number:
 Pressure transducer serial number:
 Test size:
 Test dates:
 Sensor location U.S. of manhole opening:
 Straight pipe U.S. of manhole:
 Straight pipe D.S. of manhole:
 Four pipe inside diameter measurement:

 Average pipe inside diameter:
 ADS pipe diameter measurement:
 Pipe invert point gauge reference:
 Time required to install the sensor:
 Time required to set up the sensor:
 Calibration performed by:
 Vendor technicians:
 02527
 16405
 11652
 8710
 20-inch
 5/3 0/01 to 6/4/01
 8.25 in.
 74.0ft
 17.6ft
 19.435 in., 19.555 in., 19.615 in.,
 19.553 in.
 19.535 in.
 19.500 in.
 0.907 ft
 30 min
 120 min
 Steven Barfuss, Randy Geldmacher, Tyler Smith
 Jeffrey White, Heather Hackett
Table 3-6. Preliminary Model 3600 42-inch Pipe Test Measurements
 Meter (Monitor) Serial Number:
 Ultrasonic Serial Number:
 Velocity Sensor Serial Number:
 Pressure Transducer Serial Number:
 Test Size:
 Test Dates:
 Sensor Location U.S. of manhole opening:
 Straight Pipe U.S. of manhole:
 Straight Pipe D.S. of manhole:
 Four Pipe Inside Diameter Measurement:
 Average Pipe Inside Diameter:
 Average Pipe Inside Diameter:
 ADS Pipe Diameter Measurement:
 Pipe Invert Point Gauge Reference:
 Time required to install the sensor:
 Time required to set up the sensor:
 Calibration Performed by:
 Vendor Technicians:
02527
16405
11652
8710
42-nch
5/3 0/01 to 6/4/01
11.5 in.
84.0ft
36.6ft
3.424 ft, 3.455 ft, 3.437 ft, 3.448 ft
3.441 ft
41.292 in.
41.250 in.
-0.100ft
60 min
150 min
Steven L. Barfuss, Randy Geldmacher, Tyler Smith
Erica Blanken, Gillian Woodward
                                             33

-------
3.2.2   Test Data

The data obtained during testing was compiled  in both graphical  and statistical formats.  The
graphs are presented in the following sections, while the statistical tables are included in
Appendix D. The time listed for each run in the statistical tables represents the start time for the
five-minute logging period. This recorded time was critical to the success of the tests since the
logged data was retrieved some time after the run was finished. It was necessary to find the
recorded start time in the logged data to retrieve the corresponding set of five flow, depth, and
velocity readings. For this reason, the times for the UWRL computer and the laptop  computer
used to log the data from the ADS flow meter were synchronized each day.

For each run, the data include the pipe slope, desired flow condition, and the date and time when
the five-minute logging period occurred. Up to five  depth measurements were  made for each
logging period. It was  usually not required to  make  all five  depth  measurements  since the
transverse water surface profiles were normally  quite flat. During submerged tests,  the depth
readings in the tables in Appendix D show "FP," indicating that the pipe was running full.

Select runs were either not practical or not possible, and are indicated by "NA" in the data tables
in Appendix D.  In  some cases, the meter did not record data  during a particular one-minute
interval or  a reference depth reading was not made,  and hyphens  (-)  were placed in the table
when  no data was collected. ADS was allowed to re-test certain runs due to changes in the test
protocol subsequent to the original tests. An "R" following the run  number indicates these runs
in the data tables.

3.2.2.1 Statistical Data Evaluation

Deviation is summarized by the various test configurations in Table 3-7. Deviation is calculated
as the mean of the percentage deviation for each  given category. Depth, velocity, and  flow data
for each test run is outlined in Appendix D.

The tables  also contain  the 95-percent confidence interval for the deviation data. The equation
used to establish the confidence bounds is:
                                            -
                               Confidence = x +  t x
       Where:
       x = sample mean
       s = standard deviation
       n = sample size
       t = Student' s t - distribution with (n - l) degrees of freedom
Therefore, the width of the interval is a function of not only the variation in instrument deviation,
but also the number of test runs in  each reported category. In general, the categories with the
greatest number of runs will show the narrowest confidence intervals.
                                           34

-------
Table 3-7. Deviation by Test Configuration: Normal Operating Conditions

 Pipe size (inches)                Deviation (percent)                 *-_i     •  ^    •
   r	confidence interval
            10                           2.8                           0.2-5.4
            20                           7.2                          2.4-12.0
            42                           1.6                          -1.8-5.0

 Pipe slope (percent)	
            0.1                           4.9                          -1.4-11.2
            0.2                           1.6                          -1.8-5.0
            0.5                           5.6                           1.6-9.6
           1.25                           4.2                          -1.6-9.9
            2.0                           6.1                          -3.8-16.1

 Percent full (d/D)
10
30
50
80
150
250
3.6
11.1
3.4
1.3
0.9
4.8
-9.4
6.5
0.2
-3.1
0.9
-5.0
-16.6
-15.7
-6.7
-5.6
-2.8
-14.6
 Condition
        Freeflow                        2.6                         -1.4-6.7
        Backwater                        5.8                         2.9-8.7

 All normal conditions                   4.5                         2.1 - 6.9
The ADS 3600 was also tested under "abnormal conditions," defined previously as tests where
grease was applied to the depth sensor and reverse flow conditions existed.  During the tests
where 0.5 mm of grease was applied to the ultrasonic depth sensors, the mean deviation was 1.9
percent. When 2.0 mm  of grease was applied to the ultrasonic depth sensors, the meter was
unable to read depth, resulting in zero flow. The mean deviation for reverse flow tests was -72.7
percent. The grease was not removed significantly by the flow during the grease tests.

3.2.2.2 Graphical Evaluation of All Flow Data

The overall  accuracy of the ADS flow meter under normal operating conditions is shown in
Figure 3-13, the plot of measured flow rates versus the laboratory reference. This plot excludes
tests where grease was applied to the sensors and tests with reverse flow.
                                           35

-------
^b.uuu -
22,500 -


O
o
S
CO
<












X







•
X








X








X







Zero deviation reference line (dashed)




















/







/"


/
y = 0.9841x
R2 = 0.9955

Best fit line (solid)














^X









            0      2,500    5,000    7,500    10,000   12,500    15,000    17,500    20,000    22,500   25,000

                                          Reference Flow (GPM)
Figure 3-13. Metered flow rate versus reference flow rate.

Figure 3-13  was generated using  formulas available in Microsoft  Excel for characterizing a
linear trend line. The line is fitted through the origin  (y-intercept  of zero).  The slope of the
regression line is computed as:
                                      slope =
                                                  x
       Where:

       x = reference flow rates
       y = ADS flow rates

The correlation coefficient, r2, is defined as:
                                            1-
SSE
SST
       Where:

       SSE is the sum of squares for the error component
       SST is the sum of squares total
                                            36

-------
With  the  slope  and correlation coefficient, bias and  precision can be  calculated. Bias  and
precision are expressed as functions of the slope and correlation coefficient, respectively, using
the following equations:

                        (   slope of zero deviation reference line")
                  Bias=  1- —^	          	 x 100%
                        V           slope of best fit line        j
                            I    r2 of zero deviation reference line ]
                 Precision =  1-	=—	-—	  x 100%
                            V           r  of best fit line        J

Values of 1.0 for both slope and correlation  coefficient would yield results of zero percent for
both bias and precision, and would indicate a one-to-one relationship between the metered flow
and the reference flow, with changes in the reference flow accounting for all of the variation in
the metered flow.

Evaluation of the ADS 3600 data for the three pipe sizes (10-inch, 20-inch, and 42-inch) under
normal operating conditions yields a bias of 1.6 percent and precision of 0.45 percent.

3.2.2.3 Graphical Evaluation of Flow Data by Test Condition

The flow, depth, and velocity readings, in addition to the average of the readings,  are compared
to the  average reference flow, depth,  and velocity for each run. The average reference flow,
depth,  and velocity are equal to the  mathematical  average of the  initial and final  reference
measurements made for each run.

The graphical evaluation of the flow data by test condition  is provided in Tables 3-15 through
3-32.   The free  flow and backwater data have been differentiated in each figure. The scatter-
graph figures for slopes of 0.1 percent and 0.5 percent have both free flow and backwater data
shown on the same figure. Scatter-graph figures for pipe slopes of 1.25  percent and 2.0 percent
do not have backwater data shown because no backwater data was collected for these slopes. It
should be noted that the free flow curves are not always smooth. Laboratory personnel attempted
to eliminate all  backwater effects  during free flow  tests, but in some  cases were not able to
completely eliminate the effect of the valve at  the downstream end of the test pipe. As an
example, Figure 3-16 shows the free flow curve dropping off on the upper end. This shows that
the velocity was lower than  the desired free flow condition, indicating some backwater effect.
Even though this occurred periodically, it did not  affect the test results, since the verification is a
direct comparison between the reference measurements and the meter data regardless of the flow
condition.

Two sets of plots are presented for each pipe size illustrating the deviation from reference flow.
One series of plots (Figures 3-19,  3-25, and 3-31) show deviation as percent of reference across
the full range of reference flows. Another series of plots (Figures 3-20, 3-26, and 3-32) show
flow meter quantities compared to reference quantities.  The reference line is the point of equality
between the ADS flow meter and the reference measurements, where the deviations between the
                                           37

-------
two would be zero. The slope and correlation coefficient of the best-fit line are measures of bias
and precision, respectively, where 1.0 is the ideal value.

For the tests performed in the 10-inch pipe, Figures 3-15 through 3-18 are scatter-graph plots of
the free flow and backwater tests, Figure 3-19 shows the deviation from the reference flow, and
Figure 3-20 is a plot of reference flow versus meter flow.

For the tests performed in the 20-inch pipe, Figures 3-21 through 3-24 are scatter-graph plots of
the free flow and backwater tests, Figure 3-25 shows the deviation from the reference flow, and
Figure 3-26 is a plot of reference flow versus meter flow for all tests.

For the tests performed in the 42-inch pipe, Figures 3-27 through 3-30 are scatter-graph plots of
the free flow and backwater tests, Figure 3-31 shows the deviation from the reference flow for all
tests, and Figure 3-32 is a plot of reference flow versus meter flow.

3.2.2.4 Data Analysis Discussion

The  protocol outlines  the procedure for characterizing flow meter accuracy by calculating
deviation using the following formula:

                       percent deviation (%D) = (XM-XR) / XR x 100%

       Where:

       XM = mean value recorded by test flow meter
       XR = mean reference value

Calculating deviation as a percent of the  reference value can exaggerate the apparent deviation at
low flows. During low flow conditions,  a low absolute  XM-XR|  deviation results  in a high
percent  (%D)  deviation, creating  a disproportionate  percent  deviation  bias at   low flow
conditions.

For example, during the ADS 3600 tests with the 10-inch pipe, test run number 16 recorded a
test flow meter reading of 99.91 gallons per minute (gpm) and a reference reading of 87.01 gpm.
This results in an absolute deviation of 12.90 gpm but a percent deviation of 14.83 percent. By
contrast, test run number 26, which had the highest reference reading for the  10-inch pipe tests at
1,293.89 gpm, had a corresponding test flow meter reading of 1,274.39 gpm, which computes to
a larger absolute deviation (19.50 gpm) when compared to  test run 16 but a  much lower percent
deviation (-1.51 percent).

This phenomenon can also be represented graphically.  Evaluating the ADS 3600 10-inch pipe
test runs with an arbitrary "low flow" range  of 0-600 gpm, the same data can  be expressed on
two separate plots: absolute deviation and percent  deviation. When  this is done, the absolute
deviation  data shows a  strong correlation in absolute terms  despite the relatively high percent
deviation, as shown in Figure 3-14.
                                           38

-------
     1400

     1200

     1000

      800

      600

      400

      200
                400   600   800   1000

                  Reference Flow(GPM|

400   600   800  1000  1200   1400

 Reference Flow(GPM|
           (a) Absolute deviation                          (b) Percent deviation
Figure 3-14. ADS 3600 data summary, 10-inch pipe, 0-600 gpm reference flow.

The percent deviation data is presented in the ETV verification report because it is a common
statistical method of computing the difference between a measured value and a reference value.
This statistical anomaly is presented to inform the end-users of flow monitoring equipment that
using percent deviation  data when evaluating the performance of flow monitoring equipment at
low flows may result in misleading information. During low flow conditions,  a flow meter can
report flow very close to a reference or actual flow but can be off by a disproportionately high
percent deviation, especially when compared against high flow condition data. This phenomenon
is reflective of a statistical limitation rather than a flow meter limitation.
                                            39

-------
      2.50
                  Free Flow Reference Data
                  Free Flow ADS Data
                  Backwater ADS Data
                  Backwater Reference Data
                                     10-inch pipe at 0.1% slope
      0.00
          0          0.1          0.2          0.3
0.4          0.5
  Depth (d/D)
0.6          0.7         0.8         0.9
Figure 3-15. Scatter-graph for 10-inch pipe test at 0.1 percent slope.
                                                                  40

-------

.oU
.00 •
.50 "
.00 "
"o"
Si ncn
"1 Z.OO
£
.B-
'o
o zo°
.50
.00 "
.oU

Free Flow Reference Data
• Free Flow ADS Data
• Backwater ADS Data
• Backwater Reference
















Data









m jS




•l^^


0 0.1 0.2 0.3 0

Free Flow



	 	 -—
. — ^
Backwater


4 0
Depth (d/D)

er~~




H— —



5 0
10-inch pipe at 0.5% slope
~~-^^^









6 0

^\^ •
•!>



<
~*



7 0





•



8 0
Q
Figure 3-16. Scatter-graph for 10-inch pipe test at 0.5 percent slope.
                                                           41

-------

.uu
.uu
.00
.uu
"o"
i/) o.uu
£
-B-
'o
o 4.00
.UU
.uu
.00

Free Flow Reference Data
• Free Flow ADS Data









.00
0 0



Free








Flow





1 0.2 0









3 0

-
















4 0.5 0
Depth (d/D)









6 0
10-i








7 0
nch pipe at 1








8 0
.25% slope








9

Figure 3-17. Scatter-graph for 10-inch pipe test at 1.25 percent slope.
                                                           42

-------

.uu
8.00 -
.00 •
.00
4> c nn
in 5.00
£.
J
o
- 4.00
>
.UU
.uu
.00

Free Flow Reference Data
• Free Flow ADS Data








0 0


Free





1 0




t Flow





2 0






3 0

•^
















4 0.5 0
Depth (d/D)









10-








6 0.7 0
inch pipe at








8 0
2.0% slope








9

Figure 3-18. Scatter-graph for 10-inch pipe test at 2.0 percent slope.
                                                           43

-------

100%
80%
60%
| 40/o

-------
          1400
          1200 -
          1000 -
       oi   800
       I
       "S   600 -
           400 -
           200 -
                 10-inch Pipe
                             200
                                     Zero deviation reference line (dashed)
                                                                                              ope = 0.987*
                                                                                            Correlation = 0.9991
                                                                                                                      x
                                                                                                 xx'   Best fit line (solid)
400
  600             800
Reference Flow Rate (GPM)
1000
1200
1400
Figure 3-20. Plot of reference flow versus meter flow in 10-inch pipe.
                                                                 45

-------
          4.00
           3.50 -
           3.00
          2.50 -
—O— Free Flow Reference Data
     Free Flow ADS Data
 •  Backwater ADS Data
 •  Backwater Reference Data
        in
        £
        > 2.00
        o
        .o
        
-------
          7.00
          6.00 -
          5.00
       o  4.00
       in
       £
       >
       +j
       1
       a>  3.00
          2.00
          1.00
          0.00
   Free Flow Reference Data
   Free Flow ADS Data
•  Backwater ADS Data
•  Backwater Reference Data
                Backwater
              0          0.1          0.2         0.3
                                           0.4         0.5
                                              Depth (d/D)
                                                                                                 20-inch pipe at 0.5% slope
0.6          0.7          0.8         0.9
Figure 3-22. Scatter-graph for 20-inch pipe test at 0.5 percent slope.
                                                                   47

-------


9.00 -
.uu
.uu
o -00
8!
£
,> 5.00
1
4.UU
.UU
.UU
1.UU
Free Flow Reference Data
• Free Flow ADS Data












s_





.00
0 0.1 0


\

Free Flow





2 0
Depth


>







3 0
(d/D)
20-inch p









4 0
ipe at 1.25% slope









5 0
6
Figure 3-23. Scatter-graph for 20-inch pipe test at 1.25 percent slope.
                                                           48

-------
12.00 i

10.00
8.00
¥

_
> 6.00
8
•s
4.00
2.00


Free Flow Reference Data
• Free Flow ADS Data










/








0 0






1 0




^

/

Free Flow







2 0











3 0



20-inch pipe at 2.0% slope

^











4 0
Depth (d/D)












5 0













6 0


























7 0.8 0










9
Figure 3-24. Scatter-graph for 20-inch pipe test at 2.0 percent slope.
                                                           49

-------

100%
80%
60%
o 40%

-------

OUUU
4oUU
4UUU
GOUU
s
Q_
(3 oUUU
 ZOUU
LL
•a
9)
zuuu
x
^ /
^X





Best
Slope = 1.06!
Correlation =
x
XX
'
x""
Zero d«





Fit line (solid)
S3x
0.9785 >X
j^^ 0
XX
X
X

aviation refere





/
/ X
X
X
X


nee line (dasl





ij>
X



led)






>•
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Reference Flow Rate (GPM)
Figure 3-26. Plot of reference flow versus meter flow in 20-inch pipe.
                                                          51

-------
          8.00
          7.00  -
          6.00
          0.00
 Free Flow Reference Data
 Free Flow ADS Data
 Backwater ADS Data
• Backwater Reference Data
                          0.1
              0.2
0.3
0.4          0.5
  Depth (d/D)
0.6
                                                                                                    42-inch pipe at 0.2% slope
0.7
0.8
0.9
Figure 3-27. Scatter-graph for 42-inch pipe test at 0.2 percent slope.
                                                                   52

-------
       o
       8!
       £
          6.00
          5.00
          4.00
          3.00
          2.00
          1.00
          0.00
                      Free Flow Reference Data

                      Free Flow ADS Data
                               0.2
0.4
                                                                            42-inch pipe at 2.0% slope with simulated silt
    0.6

Depth (d/D)
0.8
1.2
Figure 3-28. Scatter-graph for 42-inch pipe test at 0.2 percent slope with silt.
                                                                    53

-------




¥
to
£
1
1



Free Flow Reference Data
• Free Flow ADS Data








0 0




0.5 mm
Grease
z


1 0
























2 0.3 0.4 0
Depth (d/D)
42-inc







5 0
h pipe at 2.0%







6 0
slope with Q.Z
^







7 0
mm grease







8 0
9
Figure 3-29. Scatter-graph for 42-inch pipe test at 0.2 percent slope with 0.5mm grease.
                                                           54

-------

.uu
.uu
.uu
.uu
¥
i/>
£
£ 4.UU
8
£
.UU
.00
I.UU

— Q— Free Flow Reference Data
• Free Flow ADS Data












X
2mm
Grease

0 0.2 0








42-inch
^
/






4 0.6 0
Depth (d/D)
pipe at 2.0% slopes







rith 2.0 mm grease







8 1 1
2
Figure 3-30. Scatter-graph for 42-inch pipe test at 0.2 percent slope with 2.0mm grease.
                                                           55

-------

100%
80%
60%
| 40/o

-------
           30000
           25000
_ 20000

Q.
(3


I
| 15000

LL
           10000
            5000
               0 -i
                   42-inch Pipe
                                  5000
                                                                                  Zero deviation reference line (dashed)
                                                                                                                          X

                                                                                            Best fit line (solid)


                                                                                          Slope = 0.9757x

                                                                                            orrelation = 0.9956
                                                                                                             X"


                                                                                                                  X

                                                                                                               X

                                           10000              15000              20000

                                                    Reference Flow Rate (GPM)
25000
30000
Figure 3-32. Plot of reference flow versus meter flow in 42-inch pipe.
                                                                   57

-------
                                 Appendices

A  Laboratory Equipment Calibrations and Information
B  Raw Laboratory Test Notes and Data
C  Operational Procedure and Data Logging Method
D  Laboratory Test Data
                                     58

-------
                                       Glossary

Accuracy - a measure of the closeness of an individual measurement or the mean of a number of
measurements to the true value and includes random error and systematic error.

Bias - the systematic or persistent distortion of a measurement process that causes errors in one
direction.

Comparability - a qualitative term that expresses confidence that two data sets can contribute to
a common analysis and interpolation.

Completeness - a qualitative term that expresses  confidence that all  necessary data have been
included.

Precision - a measure of the agreement between replicate measurements of the same property
made under similar conditions.

Protocol - a written document that clearly states the objectives, goals, scope, and procedures for
the study. A protocol shall be used for reference during vendor participation in the  verification
testing program.

Quality Assurance Project Plan - a written document  that describes the implementation of
quality assurance and quality control activities during the life cycle of the project.

Representativeness - a measure of the degree to which data accurately and precisely represent a
characteristic of a  population  parameter  at  a  sampling point,  a  process  condition,  or
environmental condition.

Wet Weather Flows Stakeholder Advisory Group - a group of individuals consisting of any or
all of the following: buyers and users of flow monitoring technologies, developers and vendors,
consulting engineers, the finance and export communities, and permit writers and regulators.

Standard Operating  Procedure -  a written  document containing specific procedures and
protocols to ensure that quality assurance requirements are maintained.

Technology Panel - a group of individuals with expertise  and knowledge of flow monitoring
technologies.

Testing Organization - an independent organization qualified by the Verification Organization
to conduct studies and testing of flow monitoring technologies in accordance with protocols and
Test Plans.

Vendor - a business that assembles or sells flow monitoring equipment.

Verification - to establish evidence on the  performance of flow monitoring technologies  under
specific conditions, following a predetermined study protocol(s) and test plan(s).
                                           59

-------
Verification  Organization -  an  organization qualified  by  EPA to verify environmental
technologies and to issue verification statements and verification reports.

Verification  Report - a  written document containing  all  raw and  analyzed  data, all quality
assurance/quality control  (QA/QC) data sheets, descriptions of all  collected  data, a detailed
description of all procedures and methods used in the verification testing, and all QA/QC results.
The test plan(s) shall be included as part of this document.

Verification  Statement - a document that summarizes the Verification Report reviewed and
approved and signed by EPA and NSF.

Verification Test Plan - a written document prepared to describe the procedures for conducting
a test or study  according  to the verification protocol  requirements for the application of flow
monitoring technology.  At a minimum, the  test plan  shall include detailed instructions  for
sample and data collection, sample handling and preservation, precision,  accuracy, goals, and
QA/QC requirements relevant to the technology and application.
                                           60

-------
                                 December 2003
                               03/13/WQPC-WWF
                                EPA/600/R-04/034
   Environmental Technology
   Verification Report

   Wet Weather Flow Monitoring
   Equipment

   ADS Environmental Model 4000
   Open Channel Flow Monitor

   Part II - - Field Test Results

               Prepared by
             NSF International
       Under a Cooperative Agreement with
4>EPA U.S. Environmental Protection Agency

-------
        THE ENVIRONMENTAL TECHNOLOGY VERIFICATION
                                   PROGRAM
 &EPA
   U.S. Environmental
   Protection Agency
                                            NSF International
                      ETV Joint Verification Statement
    TECHNOLOGY TYPE:
    APPLICATION:

    TECHNOLOGY NAME:

    TEST LOCATION:
    COMPANY:
    ADDRESS:
    WEB SITE:
    EMAIL:
       AREA/VELOCITY FLOW MONITORS
       FLOW METERING IN SMALL- AND MEDIUM
        (10- to 42-inch) SEWERS
       ADS ENVIRONMENTAL SERVICES MODEL 4000 OPEN
       CHANNEL FLOW METER
       QUEBEC CITY, QUEBEC, CANADA, AND LOGAN, UTAH
       ADS ENVIRONMENTAL SERVICES
       5030 BRADFORD DRIVE
       BUILDING 1, SUITE 210
       HUNTSVILLE, AL 35805
       http:\\www.adsenv.com
       info@adsenv.com
PHONE: (800) 633-7246
FAX:    (256) 430-6633
NSF International (NSF) manages the Water  Quality  Protection  Center (WQPC) under the U.S.
Environmental  Protection Agency's (EPA) Environmental Technology Verification  (ETV) Program.
NSF evaluated the performance of the Model 4000 Open Channel Flow Meter manufactured by ADS
Environmental Services.  Utah Water Research Laboratory (UWRL) in Logan, Utah, and BPR of Quebec
City, Canada, both NSF-qualified testing organizations, performed the laboratory and field verification
testing, respectively.
EPA created the ETV Program to facilitate the  deployment of innovative or improved  environmental
technologies through performance verification and dissemination of information.  The goal of the ETV
program is to further environmental protection by accelerating the acceptance and use of improved and
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; 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 tests (as
appropriate), collecting and analyzing data, and  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.
03/13/WQPC-WWF
The accompanying notice is an integral part of this verification statement.
                      VS-i
               December 2003

-------
TECHNOLOGY DESCRIPTION
The  following  technology description is provided  by the vendor  and does  not represent verified
information.
Area/velocity flow meters  are commonly used in wastewater collection, storm sewer, and combined
sewer systems.  The ADS 4000 flow meter utilizes a quad-redundant ultrasonic sensor that measures the
time required for an ultrasonic pulse to travel from the sensor face to the surface of the water and back to
the sensor. The meter converts the travel time to distance by calculating the speed of sound  through air
and adjusting for temperature, which is measured by two sensors inside the ultrasonic sensor head. The
depth of the flow is  then calculated using the  pipe diameter and the  range measured by the ultrasonic
sensor.  A pressure-depth sensor is also installed at the bottom of the pipe to measure surcharge levels and
to provide a redundant depth reading when used with the ultrasonic level sensor. Doppler velocity
measurements are made by transmitting an ultrasonic signal upstream using a submerged velocity sensor
and measuring the frequency shift in the sound waves reflected by the moving particles in the water. The
depth and velocity  sensor readings are  stored  in  the  flow meter's memory  until the data can be
downloaded to a computer through either  a  voice-grade telephone line  or a cellular network. The
computer software calculates flow rates using the depth and velocity readings.
The ADS 4000 flow  meter system includes the flow meter unit, sensors, and installation hardware. The
flow meter unit is  housed in a waterproof, marine-grade aluminum housing. The  submersible pressure
sensor, ultrasonic level sensor, and velocity sensor are attached to a circular stainless steel band installed
around the inner circumference  of the sewer pipe.  Waterproof cables with sealed connectors convey
power and signals between the flow meter unit and the sensors. The system is battery-powered, and can
power the unit  for about one year at a standard  15-minute measurement interval.  According to vendor
claims,  after the unit is installed, minimal  operation and maintenance (O&M) or unit calibration is
required; the most common O&M procedure is cleaning the sensors.

VERIFICATION TESTING DESCRIPTION
Laboratory Test Site
The laboratory  testing was completed at the Utah Water Research Laboratory (UWRL), at Utah State
University in Logan, Utah. The flow meter was installed in three  nominal pipe sizes: 10-inch, 20-inch,
and 42-inch. The straight lengths were sized  so they were at least 40 times the pipe diameter for the 10-
and 20-inch pipes and at least 22 times for the 42-inch pipes. Pipe slopes were  adjustable to allow the
flow meter to be evaluated under different slope  conditions. Sluice gates at both ends of the pipes were
used to  regulate appropriate flow, head, and  obstruction  during testing. Reference devices were directly
traceable to the National Institute of Standards and Technology (NIST), and were regularly calibrated.
Uncertainty for the reference devices was less than 0.25 percent.
Field Test Site
Field verification testing was conducted in a section of the Quebec Urban Community's (QUC) sewer
network, located in the City of Sainte-Foy, Quebec, Canada. The ADS flow meter and reference meters
were installed in a 41.7-inch diameter interceptor pipe, near the downstream of a straight run of pipe that
had an average  slope  of 0.169 percent. The reference devices, which consisted of a bubbler for a reference
level measurement, a reference  flow monitor, and an Accusonic 4-path flow monitor, were installed
downstream of the ADS 4000 flow meter. Upstream and downstream sluice gates were used to create the
required flow conditions.
Validation of the reference flow monitor and bubbler were performed by lithium tracer dye  tests. Flow
rates under the  upstream and  downstream gates were also calculated using standard hydraulic equations
for a redundant check of flow data.
03/13/WQPC-WWF      The accompanying notice is an integral part of this verification statement.         December 2003
                                              VS-ii

-------
Methods and Procedures

Laboratory evaluation of the flow meters consisted of collecting depth, velocity, and flow data from the
ADS meter and comparing it to the depth, velocity, and flow data from the reference devices. These tests
were performed under normal operating conditions of uniform flow, backwater flow, full pipe (manhole
surcharged), and simulated silt. Water transmission through the pipes, as a ratio of flow depth versus the
pipe diameter (d/D), ranged from 10 to 250 percent (surcharged conditions). Tests were also performed
under the abnormal operating conditions of reverse flow and grease accumulation.
Field evaluation of the ADS flow meter at the Quebec site consisted of a general evaluation of the flow
meter (Test A) and the  performance  of the meter under varying  flow conditions. Testing consisted of
collecting depth,  velocity,  and  flow  data  at regular time intervals and  comparing the data to the
corresponding depth, velocity, and flow data from the reference devices. Four test scenarios were used:
1.  Test B—accuracy under dry weather flow (approximately 1.71 million gallons per day [MGD]), with
    back-flow conditions;
2.  Test C—accuracy under wet weather flow (1.71-29.7 MGD), without back-flow conditions;
3.  Test D—accuracy under wet weather flow (1.71-29.7 MGD), with back flow-conditions; and,
4.  Test E—accuracy under short-term (26-day) continuous operation, with various flow rates.
Three conditions were identified during testing  that created an unintended challenge to the  ADS flow
meter:

1.  The  water used in the  testing at UWRL  did not contain the  particulate concentrations of normal
    sewage, so small quantities  of coffee creamer were  added to the water on some  test runs. The
    operating principle  utilized by the ADS  flow meter  requires particles  in the water to serve as
    reflectors for sound  waves. The vendor maintained that the coffee creamer additive provided a level
    of reflectivity, but the particulate concentration in the test water did not approach that of sewage and
    could be a source of measurement error.
2.  During each field test, a portion of the ADS flow meter data collected at one-minute intervals was not
    recorded. ADS  personnel indicated that this happened because the flow  meter was  configured for
    maximum error checking and sensor refiring. They further indicate that the ADS 4000  flow meter can
    be reconfigured to  collect data at one-minute intervals by reducing the level of  real-time  error
    checking.

3.  The field testing results include data in which it appears that standing waves and troughs were  present
    beneath the ADS 4000 flow meter's ultrasonic depth sensor. During portions of the  testing, the depth
    sensor was likely affected by  standing  waves and troughs up to +5 inches. The  ADS flow meter
    measures depth with a downward-looking, narrow-beam ultrasonic sensor mounted on the top of the
    pipe, so depth measurements would be susceptible to influence by waves.  Based on a review of the
    field data, it appears  that waves were most prevalent at higher depths and flow rates.
No editing was allowed  on the metered data during field or laboratory testing. In actual applications, the
flow monitoring service  provider may implement post-monitoring  quality control measures to attempt to
improve the accuracy  of final data. According to ADS, the company typically bundles flow meter sales
with post-monitoring quality control and reporting services.
03/13/WQPC-WWF      The accompanying notice is an integral part of this verification statement.         December 2003
                                             VS-iii

-------
VERIFICATION OF PERFORMANCE
System Operation

The  testing  organizations found the equipment durable and easy to use, and that it required minimal
maintenance. The  flow  meter operation and data retrieval software programs were easy to learn.  The
ultrasonic sensors and stainless steel band did not promote accumulation of debris during testing.
Laboratory Testing Results
The  mean deviation and the 95-percent confidence intervals under normal operating conditions (i.e., all
test conditions except grease tests and reverse flow) are presented in Table 1. The width of the 95-percent
confidence interval is a function of the variation in instrument deviation and of the number of test runs in
each reported category.  Categories with a fewer number of runs show wider confidence intervals.  The
calculations  exclude  "abnormal  condition" tests, where grease was applied to the sensors or where
reverse-flow conditions  were created. The mean deviation for the abnormal operating conditions was 1.3
percent for the 0.5-mm grease tests, -69.5 percent for the 2.0-mm grease  tests, and -62.4 percent for the
reverse-flow tests.
Table 1. Deviation and 95-Percent Confidence Interval by Test Configuration for Lab Testing
Deviation 95-percent confidence interval
Pipe size (inches) (percent) (percent)
10
20
42
Pipe slope (percent)
0.1
0.2
0.5
1.25
2.0
Percent full (d/D, percent)
10
30
50
80
150
250
Condition
Free flow
Backwater
All conditions
4.7
-0.7
-0.9
4.7
-0.9
-0.8
2.3
0.2
-0.1
1.1
5.4
1.9
-4.6
3.3
2.7
0.2
1.2
-6.5-
-7.9-
-10.8-
-4.2-
-10.8-
-10.9-
-10.8-
-30.1-
-22.2 -
-13.5-
-1.6-
-7.4-
-28.8-
-6.7-
-4.3-
-7.5-
-4.0-
15.8
6.6
-9.0
13.5
-9.0
-9.4
15.4
30.4
20.3
15.8
12.5
11.2
19.6
13.2
9.7
8.0
6.5
The  overall accuracy of the ADS  4000 flow meter under normal operating conditions (i.e.,  all test
conditions except grease tests and reverse flow) is shown in Figure 1. The meter deviation is segregated
into two components—bias and precision. Overall bias was 1.6 percent, as calculated by the  slope of the
best-fit line. Precision, as calculated with the correlation coefficient (r2), was 0.74 percent.
03/13/WQPC-WWF      The accompanying notice is an integral part of this verification statement.         December 2003
                                              VS-iv

-------
  25,000


  22,500


  20,000


  17,500


| 15,000


o
[I 12,500
o
o
o
Tf
o 10,000


   7,500


   5,000


   2,500


     0
                                                        Zero deviation reference line [dashed)
            D      2,500     5,000     7,500    10,000    12,500    15,000    17,500    20,000    22,500    25,000
                                             Reference Flow (GPM)
Figure 1. Laboratory-metered flow rate versus reference.

Field Testing Results

Table 2  summarizes the field testing results  in two categories:  mean deviation and  trimmed  mean
deviation. The mean deviation is the arithmetic mean of all of the one-minute-interval data. The trimmed
mean deviation is calculated by eliminating values greater than ±99 percent, making it less susceptible to
skewing from large outliers, such as those produced when the ADS flow meter recorded zero velocity.

Table 2. Deviation from Reference Flow: Tests B, C, and D.
               Flow regime
                                  Mean deviation
                                     (percent)
Trimmed mean deviation
        (percent)
TestB
TestC
TestD
Test B-D combined
Simulated low flow
Simulated wet flow
Combined flows
-14.5
14.0
-0.8
-0.4
0.5
-1.3
-0.4
-0.9
14.5
8.3
3.8
9.5
-1.0
3.8
Analysis of the data collected during Test B (low flow) revealed that in nearly one-fourth of the samples
the deviation was -100 percent. This occurred when the ADS 4000 flow meter recorded zero velocity and
calculated the flow to  be zero. This  occurred most frequently when the pipe experienced back-flow
conditions. The data collected during Tests C and D shows a significantly lower occurrence of data with
deviations exceeding ±99 percent.
03/13/WQPC-WWF
                The accompanying notice is an integral part of this verification statement.
                                          VS-v
                            December 2003

-------
Test E (not included in Table 2) evaluated the performance of the flow meter over an extended (26-day)
time period. Generally, the data collected during Test E closely correlated with the reference flow monitor
data. Spikes were noted in water level measurements collected toward the end of the test, which may have
been the result of accumulated condensation on the ultrasonic depth probe. No debris accumulation was
observed on the equipment, and, aside from a thin film of grease on the probes, the equipment was in
good condition and did not require maintenance.

QUALITY ASSURANCE/QUALITY CONTROL

A complete description of the quality assurance/quality control procedures and findings are included in
the verification reports. Calibration records were maintained by the testing organizations and validation of
the reference flow devices fell within control limits.  NSF completed a data quality  audit of at least  10
percent of the test data to ensure that the reported data  represented the data generated during testing.
Audits of the field and laboratory testing were conducted by NSF with no significant issues noted.
       Original Signed by
       Lee A. Mulkey
         March 31, 2004
    Lee A. Mulkey               Date
    Acting Director
    National Risk Management Laboratory
    Office of Research and Development
    United States Environmental Protection Agency
Original Signed by
Gordon Bellen     April 26, 2004
                                    Gordon Bellen
                                    Vice President
                                    Research
                                    NSF International
                      Date
    NOTICE: Verifications are based on an evaluation of technology performance under specific,
    predetermined criteria and the appropriate quality assurance procedures. EPA and NSF make no expressed
    or implied warranties as to the performance of the technology, and do not certify that a technology will
    always operate as verified. The end user is solely responsible for complying with any and all applicable
    federal, state, and local requirements. Mention of corporate names, trade names, or commercial products
    does not constitute endorsement or recommendation for use of specific products. This report is not an NSF
    Certification of the specific product mentioned herein.
        Availability of Supporting Documents
        Copies of the Draft 4.0- Generic Verification Protocol, Flow Monitors for Wet Weather Flows
        Applications in Small- and Medium-Sized Sewers, September, 2000, the verification statement, and
        the verification report (NSF Report #03/13/WQPC-WWF) are available from:

        ETV Water Quality Protection Center Program Manager (order hard copy)
            NSF International
            P.O. Box 130140
            Ann Arbor, Michigan 48113-0140
        NSF web site: http://www.nsf.org/etv (electronic copy)
        EPA web site: http://www.epa.gov/etv (electronic copy)

        (NOTE: Appendices are not included in the verification report. Appendices are available upon
            request from NSF.)	
03/13/WQPC-WWF
The accompanying notice is an integral part of this verification statement.
                          VS-vi
                          December 2003

-------
     Environmental Technology Verification Report
WET WEATHER FLOW MONITORING EQUIPMENT VERIFICATION

              ADS ENVIRONMENTAL MODEL 4000
               OPEN CHANNEL FLOW MONITOR

                PART II: FIELD TEST RESULTS
            QUEBEC URBAN COMMUNITY TEST SITE
                          Prepared for:

                        NSF International
                       Ann Arbor, Michigan
                          Prepared by:

                             BPR
                    Quebec City, Quebec, Canada
                         December, 2003
Under a cooperative agreement with the U.S. Environmental Protection Agency

                 Raymond Frederick, Project Officer
            National Risk Management Research Laboratory
                U.S. Environmental Protection Agency
                        Edison, NJ 08837

-------
                                       Notice

The  U.S. Environmental  Protection  Agency (EPA) through its  Office  of Research  and
Development has financially supported and collaborated with NSF International (NSF) under a
cooperative agreement. This verification effort was supported by the Water  Quality Protection
Center operating  under  the  Environmental  Technology  Verification  (ETV)  Program.  This
document has been peer-reviewed and reviewed by NSF and EPA and recommended for public
release.

-------
                                      Foreword

The  following is the final report  on an Environmental  Technology Verification (ETV) test
performed for NSF International (NSF) and the United States Environmental Protection Agency
(EPA) by BPR-CSO, in cooperation with ADS Environmental Services for the Model 4000 Open
Channel Flow Monitor.  The test protocol for flow monitors requires both laboratory and field
testing. The final report for this verification is divided into two parts to address both portions of
testing.

This part of the report (Part II) describes the testing and summarizes the data from  the field
testing. Part I: Laboratory Test Results, describes the testing  and summarizes the data of the
laboratory testing. Both parts of the report are available on the NSF and EPA websites.

The U.S. Environmental Protection Agency (EPA) is charged  by Congress with protecting the
Nation's land, air, and water resources. Under a mandate of national environmental laws, the
Agency strives to formulate and implement actions leading to a compatible balance between
human activities and the ability  of natural systems  to  support and nurture life. To  meet this
mandate,  EPA's  research program  is providing  data  and  technical support  for  solving
environmental problems today and  building a science knowledge base necessary to manage our
ecological resources wisely, understand how pollutants affect our health, and prevent or reduce
environmental risks in the future.

The  National Risk Management Research Laboratory  (NRMRL) is  the Agency's center for
investigation  of technological  and  management approaches for preventing and reducing risks
from pollution that threaten human health  and the environment. The focus of the Laboratory's
research program is on methods  and  their cost-effectiveness for prevention and control of
pollution to air, land, water, and subsurface resources; protection of water quality in public water
systems; remediation of contaminated sites, sediments and ground water; prevention and control
of indoor air  pollution; and restoration of ecosystems.  NRMRL collaborates with both public
and private sector partners to foster technologies that reduce the cost of compliance and to
anticipate emerging problems. NRMRL's research provides solutions to environmental problems
by: developing and promoting technologies that protect and improve the environment; advancing
scientific and engineering information to support regulatory and policy decisions; and providing
the technical support and information transfer  to  ensure  implementation of environmental
regulations and strategies at the national, state, and community levels.

This publication has been produced as part  of the Laboratory's strategic long-term research plan.
It is published and made available by EPA's Office of Research and Development to assist the
user community and to link researchers with their clients.

-------
                                       Contents

Verification Statement	VS-1
Notice	i
Foreword	ii
Contents	iii
Tables	vi
Figures	vi
Acronyms and Abbreviations	vii
Chapter 4 Field Report	1
4.1    Description of Test Site	1
   4.1.1      Test Site Infrastructure Description	1
   4.1.2     Pipe Configuration	4
   4.1.3      Waves in the Test Pipe	6
4.2    Description of Test Equipment	11
   4.2.1      Downstream Gate Site	11
     4.2.1.1    Verified Flow Meter: ADS 4000	11
     4.2.1.2    Transit-Time Reference Flow Meter	13
     4.2.1.3    Reference Flow Meter Bubbler	13
     4.2.1.4    Ultrasonic Reference Level Meter	13
     4.2.1.5    Downstream Gate	14
   4.2.2     Upstream Gate Site	14
     4.2.2.1    Electrical Control Room	14
     4.2.2.2    Upstream Gate	14
4.3    Description of Reference Methods	15
   4.3.1      Reference Depth Measurement	15
     4.3.1.1    Manual Depth Measurement	15
     4.3.1.2    Reference Depth Meters	15
   4.3.2     Reference Velocity Measurement	16
     4.3.2.1    Reference Velocity Computed from Tracer Dilution	16
     4.3.2.2    Reference Velocity Meter	16
   4.3.3      Reference Flow Measurement	16
     4.3.3.1    Tracer Dilution Method	16
       4.3.3.1.1  At the Injection Site	17
       4.3.3.1.2  At the Sampling Site	19
     4.3.3.2    Reference Flow Meter	19
       4.3.3.2.1  Flow Under the Gate	19
       4.3.3.2.2  Upstream Gate	20
       4.3.3.2.3  Downstream Gate	20
4.4    Experimental Procedures	20
   4.4.1      General Evaluation (Test A: Flow Meter Software)	21
     4.4.1.1    Procedures	21
     4.4.1.2    Measurements	21
   4.4.2     Performance Evaluation	22
     4.4.2.1    Test B: Accuracy under Dry Weather Flow with Back-Flow Conditions	22
       4.4.2.1.1  Procedures	22
                                           in

-------
       4.4.2.1.2  Measurements	23
     4.4.2.2   Test C:  Accuracy Under Wet Weather Flow Without Back Flow Conditions.... 23
       4.4.2.2.1  Procedures	23
       4.4.2.2.2  Measurements	25
     4.4.2.3   Test D:  Accuracy Under Wet weather flow With Back flow Conditions	25
       4.4.2.3.1  Procedures	25
       4.4.2.3.2  Measurements	26
     4.4.2.4   TestE: Accuracy Under Short-Term Continuous Operation	26
       4.4.2.4.1  Procedures	26
       4.42.42  Measurements	27
4.5    ADS 4000 Evaluation Results	27
  4.5.1    Software Evaluation	27
  4.5.2    Performance Tests	29
     4.5.2.1   Scatter Plots	30
     4.5.2.2   Deviation Distribution Plots	33
       4.5.2.2.1  TestB	38
       4.5.2.2.2  TestC	40
       4.5.2.2.3  TestD	44
       4.5.2.2.4  TestE	46
     4.5.2.3   General Verification Tests	49
       4.5.2.3.1  Installation, Configuration and Calibration	49
       4.5.2.3.2  Operation	52
       4.5.2.3.3  Time for Data Retrieval	52
       4.5.2.3.4  Dismantling	53
       4.5.2.3.5  General Characteristics of the ADS 4000	53
4.6    Reference Meters	54
  4.6.1    Scatter Plots	54
  4.6.2    Deviation Distribution Plots	54
  4.6.3    Tests B, C, and D	56
Chapter 5 Quality Assurance/Quality Control	57
5.1    Audits	57
  5.1.1    Field Audits	58
  5.1.2    Report Audit	59
5.2    Validation of the Lithium Tracer Methods	61
  5.2.1    Sample and Data Handling	62
  5.2.2    Methods Used to Validate Lithium Concentrations	62
     5.2.2.1   Concentrated Samples	62
     5.2.2.2   Blank and Diluted Samples	63
  5.2.3    Validation Results for Test CO	65
  5.2.4    Validation Results for Test C3	66
  5.2.5    Conclusion	68
5.3    Validation of the Reference Flow Meter	68
  5.3.1    Validation Methods	68
     5.3.1.1   Qualitative Analysis	68
     5.3.1.2   Statistical Analysis	69
     5.3.1.3   Reference Flow Deviation Distribution Analysis	69
                                            IV

-------
  5.3.2     Test CO Results	70
     5.3.2.1   Qualitative Analysis	70
     5.3.2.2   Statistical Analysis	70
     5.3.2.3   Reference Flow Deviation Distribution	70
  5.3.3     Test C3 Results	71
     5.3.3.1   Qualitative Analysis	71
     5.3.3.2   Statistical Analysis	72
     5.3.3.3   Reference Flow Deviation Distribution	72
  5.3.4     Test B and D Results	73
  5.3.5     Conclusion	73
5.4    Validation of the Reference Depth Meter (Bubbler)	73
  5.4.1     Validation Methods	74
     5.4.1.1   Qualitative Analysis	74
     5.4.1.2   Statistical Analysis	74
     5.4.1.3   Reference Depth Deviation Distribution	74
  5.4.2     Test B Results	75
     5.4.2.1   Qualitative Analysis	75
     5.4.2.2   Statistical Analysis	75
     5.4.2.3   Reference Depth Deviation Distribution Analysis	76
  5.4.3     Test C Results	76
     5.4.3.1   Qualitative Analysis	76
     5.4.3.2   Statistical Analysis	76
     5.4.3.3   Reference Depth Deviation Distribution Analysis	77
  5.4.4     Test D Results	77
     5.4.4.1   Qualitative Analysis	77
     5.4.4.2   Statistical Analysis	77
     5.4.4.3   Reference Depth Deviation Distribution Analysis	77
  5.4.5     Tests B, C, and D Combined	78
     5.4.5.1   Qualitative Analysis	78
     5.4.5.2   Reference Depth Deviation Distribution Analysis	78
  5.4.6     Conclusion	78
Appendices	80
A Complementary Test Site and Reference Meters Information	80
B Experiment Procedures	80
C Detailed Experimentation Results	80
D Quality Assurance Project Plan: Complementary Information	80
E Field Verification: Reports Generated during Tests	80
F ADS 4000 Flow Meter: Configuration	80
G ADS 4000 Flow Meter Software Functionality/Flexibility Summary	80
H Flow Test Data Figure Summaries	80
Glossary	81

-------
                                        Tables

Table 4-1. Flow Meter Verification Test Overview	21
Table 4-2. Personal Computers Used in Flow Meter Software Evaluation	22
Table 4-3. Deviation from Reference Flow—Tests B, C, andD	30
Table 4-4. Flow Rate Comparison of Data Point Samples for the Three Replicas of Test B	39
Table 4-5. Water Level Comparison of Data Point Samples for Three Replicas of Test B	40
Table 4-6. Flow Rate Comparison of Data Point Samples for Three Replicas of Test C	42
Table 4-7. Water Level Comparison of Data Point Samples for Three Replicas of Test C	43
Table 4-8. Flow Rate Comparison of Data Point Samples for the Three Replicas of Test D	45
Table 4-9. Water Depth Data Point Sample Comparison for the Three Replicas of Test D	46
Table 4-10. Daily Manual Level Measurements - ADS 4000	47
Table 4-11. Time Estimate Required to Install, Operate and Service Flow Meter	50
Table 5-1. Concentrated Samples Validation Chart	63
Table 5-2. Validation Chart for the Blank and Diluted Samples	65
Table 5-3. Validation Chart for Test CO	66
Table 5-4. Validation Chart for Test C3	67

                                       Figures

Figure 4-1. Test site plan view	3
Figure 4-2. Test site profile view of downstream gate site	4
Figure 4-3. Test site pipe invert profile	5
Figure 4-4. Pipe joint conditions  in test system	6
Figure 4-5. Waves in test pipe	7
Figure 4-6. Scatter-plot showing possible standing wave and trough influence	9
Figure 4-7. Flow graph with possible wave and trough influences shown	10
Figure 4-8. Tri-nodal nature of deviation pattern is consistent with presence of waves	10
Figure 4-9. General view of the downstream test site	11
Figure 4-10. ADS 4000 flow meter components	12
Figure 4-11. ADS 4000 flow meter installed in pipe	12
Figure 4-12. Reference flow devices installed in the test pipe	13
Figure 4-13. Upstream gate site arrangements	14
Figure 4-14. Tracer injection diagram	17
Figure 4-15. Lithium injection system	18
Figure 4-16. Blank sampling site	19
Figure 4-17. ADS 4000 system behavior—Test C	31
Figure 4-18. Reference meter system behavior	32
Figure 4-19. ADS 4000 flow vs. reference flow, Tests B, C and D	33
Figure 4-20. Flow deviation distribution—TestB	34
Figure 4-21. Flow deviation distribution—Test C	35
Figure 4-22. Flow deviation distribution—TestD	36
Figure 4-23. Flow deviation distribution—Tests B, C andD	37
Figure 4-24. Flow deviation distribution—simulated dry weather conditions	37
Figure 4-25. Flow deviation distribution—simulated wet weather conditions	38
Figure 4-26. Debris accumulation on ADS 4000's bender	51
                                           VI

-------
                          Acronyms and Abbreviations

ADS         ADS Environmental Services, a division of ADS Corporation
avg          Average
BPR         BPR, Quebec, Canada
cfs           Cubic feet per second
CSO         Combined sewer overflow
EPA         United States Environmental Protection Agency
ETV         Environmental Technology Verification
ft            Foot or feet
gpm         Gallons per minute
in.           Inch
kPa          Kilopascal
Lb           Pound
LCQ         Laboratoire de I'environnement LCQ Inc.
TO          testing organization
mA          milliamps
MGD        Million gallons per day
mg/L        Milligrams per liter
m            Meter
mm          Millimeter
NIST        National Institute of Standards and Technology
NSF         NSF International (formerly National Sanitation Foundation)
PC           Personal computer
PLC         Programmable logic controller
psi           Pounds per square inch
QA          Quality assurance
QAPP        Quality Assurance Project Plan
QUC        Quebec Urban Community
SD           Standard deviation
sec           Second
TO          Testing organization
UWRL       Utah Water Research Laboratory
VTP         Verification test plan
WWF        Wet Weather Flow
WWTP       Wastewater treatment plant
                                         Vll

-------
                                     Chapter 4
                                    Field Report
4.1    Description of Test Site
The test site for the field testing component of the verification of the ADS Model 3600 Open
Channel Flow Monitor (ADS Model 3600) was a section of the Quebec Urban Community's
(QUC) sewer network.  The site is located in the City of Sainte-Foy, along the east side of the
Chaudiere Blvd, approximately between the  Bombardier  and Mendel Streets. Figures 4-1 (a)
through 4-l(c) present a schematic layout of the complete test site and Figure 4-2 presents details
about the portion of the  sewer where the ADS Model 3600 was installed and tested.

4.1.1   Test Site Infrastructure Description

The test site was defined on the upstream side (north) by the Versant-Sud tunnel, which  was
controlled at its downstream end by a regulation chamber (Station #100) equipped with  a sluice
gate. This gate was referred as the "upstream  gate" of the test site. The sluice gate was  used to
control the flow under the gate and the volume of wastewater in the Versant-Sud tunnel when
performing tests under  high flows. A retention volume of 2.9 million gallons (11,000 m3)  was
available from this 96-inch (in.) (2,466 millimeters [mm]) nominal diameter tunnel.

The  test  site was  defined on the downstream side  (south) by the west  junction chamber
(Station #80A), which was the last chamber before the effluent tunnel that leads to QUC's West
Wastewater  Treatment  Plant (WWTP). The  test site's "downstream gate"  was located  just
upstream of the west junction chamber.

As shown in Figures 4-l(a) to (c), the  test site was an interceptor pipe 41.7 in. (1,059  mm) in
diameter. There was approximately 3,620 feet (ft) (1,103 meters (m)) between the downstream
gate  near which the ADS flow meter was installed and the upstream gate where the flow  was
controlled. The interceptor pipe between manholes S-l to  S-2 is 279.1 ft (85.1 m) long on an
average slope of 0.169  percent. The pipe between manholes S-2 and S-3 is 559.9 ft (170.7 m)
long on an average slope of 0.181 percent. The pipe changes direction by 20 degrees at  the S-2
manhole  and by 15 degrees at the S-l manhole. Downstream from the S-l  manhole, the pipe
sloped  10.6 percent over a distance of 82 ft (25 m) before reaching Station #80A.

As shown in Figure 4-l(b), the Craig Collector coming from the southeast joined the interceptor
pipe  at manhole S-5B, approximately 1,968 ft (600 m) upstream of manhole S-l. Manhole S-5B
was the injection site for the lithium dilution tests. The last section of pipe of the Craig Collector
had a 36 in. (910 mm) nominal diameter and was 421 ft (128.4 m) long, with an average slope of
0.417 percent.  The Craig Collector was a small pseudo-sanitary  uncontrolled flow collector, in
which  a  bubbler level  meter was  installed 8.6 ft (2.6 m) upstream of manhole S-5B.  The
electrical  panel  for  this  instrument  is located nearby  on Chaudiere Boulevard and  the
measurements  were transmitted to Station #100  by  a radio  link  to take  into account the
uncontrolled flow in the upstream flow management.

-------
                   SAMPLING SITF I	
                 DOWNSTREAM GATE |  \
                                                  (a)
               :-AUDIERt  BLVD.
                   I
                  \                             ,
                   \            / \  ^UM^mmsm/
                    \
                                                  (b)
Figure 4-1.  Test site plan view.

-------
Figure 4.1 (cont'd). Test site plan view.
As shown on Figure 4-2, the reference bubbler and ultrasonic level meters were installed 4.0 ft
(1.24 m) and 7.52 ft (2.30 m) upstream of the S-l manhole, respectively, and the reference flow
meter was installed in a pipe section from 5.4 ft (1.65 m) to 9.2 ft (2.80 m) upstream of the S-l
manhole. This manhole is made of concrete with a nominal diameter of 6.8 ft (2.1 m). It was
covered with a concrete roof with two circular 28.5-in. (724-mm) diameter access covers.  One
access was downstream  from  the  control gate. The other access was just upstream  and  was
equipped with a trap to facilitate the access to the sewer for  meter installation and operation.

-------
Figure 4-2. Test site profile view of downstream gate site.
There was an electrical room adjacent to the west junction chamber where the transmitters of the
reference flow meter and the reference bubbler level meter were located. The transmitter for the
reference ultrasonic level meter was hung in manhole S-l. This manhole was also the sampling
site for the lithium dilution tests and was the nearest access to the flow meters under test, the
reference meters, and the downstream gate.

Construction of the main interceptor pipe was completed around 1991, with the remainder of the
site constructed between November 1998 and June 2000. The newest parts of the system are the
control gate downstream from manhole S-l and  the ultrasonic level meter at the reference flow
meter site.

4.1.2   Pipe Configuration

The normal flow condition in the interceptor pipe was free surface flow. Neither accumulation of
solids  nor  surface foam was  observed  in the  pipe.  The  flow delay  between the test site's
upstream and downstream control gates was  12  to 15 minutes. Figure 4-3  shows the profile of
the last 13 pipe sections. All sections were 8 ft (2.44 m) long, except the last section, which was
3.7 ft (1.15 m) long. A survey of the invert pipe profile was performed with a 100 ft (30.5 m)
long by 0.5 in. (13 mm) diameter water hose and rulers graduated every 0.039 in. (1 mm). This
method provided an accuracy of approximately ±0.079 in. (±2 mm). As in  all sewers, the slope

-------
of every pipe section was slightly different. Four major elements had a hydraulic impact on the
test site:

•  There was a 2.2 in. (55 mm) bump at the end of the last pipe section due to an excess of
   concrete  at the bottom of manhole S-l. The  consequence  of this  bump,  although not the
   purpose, was that the lowest path of the reference flow meter was always submerged.
•  The last pipe section had a negative slope of approximately 1.2 in. (31 mm).
•  The coupling of the last two pipe sections was not very good and the last pipe has a negative
   elevation of approximately 0.55 in. (14 mm).
•  The large cross-section of the downstream manhole S-l  and its sluice gate could create
   slower wastewater  velocity and  an upstream surge.  The  impact of these  elements  was
   reduced since the velocity increased in the manhole due to the steep slope (10.6 percent) in
   the downstream pipe.






s
4







1






^____ 	 '
2 20 2







B 3
• 	 | 	




_____ — • 	

6 44 5
Distance (ft)







2 6


— - — •




0 6



"•—••.^



8 7







6 8




~~ 	 — •— .


4 9







2 1C
  Pipeslope =  -2.696%  0.202%  1377%  -0.123%  0.328%  1.353%  1.722%  0.574%  1.189% -1.476%  0.123%  -1.517%  0.041%

         -8.190% to bump
           Average slope between S-1 and S-2 manhole = 0.169

           Average slope between peak at 68 ft and bump = 0.335%
Figure 4-3. Test site pipe invert profile.
The difference in elevation between the invert elevation at the ADS 3600 testing site and at the
reference level meter (bubbler) was approximately +1.97 in. (50.0 mm). Since there is a slope of
-3.97  in. (-101 mm) between the invert at the bubbler and the bump described earlier, the back
flow caused by the bump had a hydraulic impact up to approximately 23 ft (7 m) upstream of the
ADS  3600 testing site. For this reason, the hydraulic profile was not exactly the same at the

-------
reference meters location and at the ADS  3600 location. The water level  was lower and the
velocity was higher at the ADS 3600 location than at the reference meters location.

No back flow could occur downstream of the downstream gate, because the maximum level over
the overflow weir  in the downstream west junction chamber is approximately 20 in. (500 mm)
under the invert of manhole S-l. When required during a test, it was possible to create back-flow
conditions by closing the test site's downstream gate.  The water level could then be raised up to
14 in. (350 mm) over the crown of the pipe at the reference level meter location.

4.1.3   Waves in the Test Pipe

Waves in  sewer pipes are common, and result from  a variety of different factors (Figure 4.4).
Waves are commonly caused by the roughness  of the  pipe  walls, poor coupling of the pipe
sections, or surge caused by a change of direction, slope, cross-section or velocity. Because of
these  waves, the measurement of the water level in  sewers is not easy,  even for a technician
measuring  with a  ruler.  The  following  were some  of the  conditions that make  level
measurements difficult:

•  High flows (sometimes present on this test site);
•  High energy slopes (sometimes present on this test site);
•  Location near a change of direction (not significant on this test site);
•  Location near a change of slope (present at some places on this test site);
•  When  pipe couplings are not very good, resulting in open junctions, steps,  uncoupled
   junctions, etc. (present at some places on this test site).
                              Opening of 1
                              inch (25 mm)
                              that can create
                                                                            Pipe
                                                                          junction
                                                                         completely
                                                                           closed.
              (a) Open joint.
Figure 4-4. Pipe joint conditions in test system.
(b) Closed joint.
There were different sizes and kinds of waves on the water surface. Depending on the water level
and flow rate, waves have different behaviors. Three kinds of waves were observed at the test
site (surface disturbances caused by surges are considered waves):

-------
    1.  Short, local dynamic waves: Small waves with lengths of 1 to 3 in. (25 to 75 mm) that
       created small ripples in all directions on the water surface, as on any river or lake. These
       waves were present everywhere. Their amplitude,  approximately  0.5  to  1 in. (13  to
       25 mm), is  a function  of the flow rate and water level (Figure 4-5). They were visual
       signs of surface turbulence.

    2.  Short, local standing waves: Small  waves with lengths 3 to 6 in. (75 to 150 mm) that
       stood at a specific location on the  water surface, without moving forward or  upward.
       Their amplitude, approximately 1 to 3 in.  (25 to 75 mm), was also a function of the flow
       rate and water level.  These waves  were rare (two  or less per pipe  section) and  were
       mostly asymmetric in the pipe (Figure 4-5). They were visual signs of surge. At the test
       site, the negative  slope,  the downstream manhole, or the  reference flow meter probes
       could have caused short local standing waves.

    3.  Long, standing waves: Long  waves with  lengths of several feet.  Their  number and
       amplitude (>2  in.  [>50 mm])  were a function  of the flow rate,  water  level and wave
       length. Most of these waves were symmetric in the pipe.
Figure 4-5. Waves in test pipe.
On June 27, 2001 ADS reviewed data recorded during a high-flow trial the previous day and
noticed  a large and unexpected aberration in the depth-velocity flow relationship in the  pipe.
Analysis in the field conducted by ADS suggested that a hydraulic jump or large waves were the

-------
likely cause of the aberration. Both BPR and NSF were notified of the concern and a request was
made to re-run the high-flow trial to verify if large waves were present.

The high-flow trial was re-run and an ADS field person with a camera was suspended over the
flow at the access manhole. Figure 4-5 was taken during the high-flow trial and it captures an
asymmetrical standing wave in the reference meter section just upstream of the access manhole.
Although  not visible in Figure  4-5, the photographer reported  seeing several similar waves
farther upstream.

Area and velocity flow meters are designed to function in sewers with uniform flow  conditions
such that depth and velocity are constant as flow passes through the metering area. The ADS
flow meter measures depth with a downward-looking, narrow beam ultrasonic sensor installed
on top of the pipe. Based on this configuration, it was expected that depth measurements were
more susceptible to influence by waves. Local wave peaks and troughs would  affect the depth
measurement by the narrow-beam depth sensor when they occurred immediately  below the
sensor. Therefore an important consideration  in selecting a site in which an area and velocity
flow meter is to be installed is to identify a location where the adverse effects of standing waves
are minimized or eliminated.

Determining if standing waves were present  at the ADS meter location was difficult without
visual  confirmation of the  waves. During high flow conditions, it was extremely difficult to
safely  access the test site to make a visual determination of the presence of standing waves.
Based  on the specific characteristics of the test pipe and the photographic documentation of the
standing wave, there exists  the potential for standing waves under certain flow conditions in the
test pipe area.

If there was a standing wave below the flow meter's ultrasonic depth sensor  and the flow meter
was working properly,  the depth readings would be influenced by the presence of the wave. A
comparison of the flow meter data  with the  reference data would  show a larger than normal
deviation between the flow meter depth  readings and the reference depth readings.

Figure 4-6 presents a scatter plot of the velocity deviation (the difference in velocity readings
between the test and reference meters)  versus depth deviation (the difference in depth readings
between the test and reference meters). These observations  are for the period June 28 through
July 4, though the apparent wave conditions were observed throughout the testing period. Three
distinct clusters of observations were  apparent.  The clustering suggests that for certain  flow
conditions the difference in velocities  are relatively constant  (2 ft/sec) while the difference in
depth changes up to 5 in. (-3 to -8 in.). This clustering was repeatable throughout the test, which
suggests the effect was due to standing  waves. The cluster of data points on the right  side of the
plot appears to be generated when  wave peaks influenced  the depth measurement, while the
cluster of data points on the left side of the  plot appears to be generated when wave troughs
influenced the depth measurement.

-------
                 Scatter Plot of Velocity Deviation vs Depth Deviation
                               Data for July 12 through July 18, 2001
           "0 (D
            o >
           4>  0
0 —
            c
            o
              a:
                  -5 -
.'•;JM|
<±3?ir-
Blue do
Red dot

'$
s indica
indicat


e wave
e trough


nfluenc
influen


j
m.

HR


'jOtf'*- •


*





	 r-


                      -12  -10   -8-6-4-20    2   4    6
                        Monitor Depth Minus Reference Depth (In)
Figure 4-6. Scatter-plot showing possible standing wave and trough influence.
The presence of local wave peaks under the depth sensor would result in incorrectly high
measurements and correspondingly high flow rate  calculations. Conversely, the presence of a
local trough under the depth sensor would result in incorrectly low depth measurements and low
flow rate calculations. Figure 4-7 shows that  deviations from reference flow rates appear to
correlate with the influence of waves.

Consistent with the protocol for this field test, accuracy for each of the several flow conditions is
reported as a single value. The reader should be aware that this single value might contain large
deviations resulting from waves. Figure 4-8 displays the deviation distribution of data from June
28 to July 4 and the tri-modal nature of the error pattern is apparent. The bars clustered to the
right of the center cluster represent data under the influence of wave peaks, while the cluster of
bars left of the center cluster represent data under the influence of wave troughs. Together the
waves appear to widen the range of deviation. Because  of the uncertainty associated with these
conditions, ADS  recommends against installing flow meters in  locations with large  standing
waves.

-------
                           ADS 4000 FLOWMETER VERIFICATION - LEVEL MEASUREMENTS

                                          TestE - July 08,2001
                                                                      -——Reference level


                                                                       o  Manual reading


                                                                      	ADS4DOO
                                                     Possible Trough Influences
                                                        1400
                                                                                         1000
                                                                                         EDO
                                                                                       -  400
          0:00    2:00    4:00    6:00    8:00    10:00   12:00    14:00    16:00   18:00   20:00    22:00    0:00

                                             Time (hh:mm)
Figure 4-7. Flow graph with possible wave and trough influences shown.
                            Model 4000 Quantity Deviation Report

                                   Data for July 12 through July 18, 2001
                   600



                   500
               >,  400
               o


               I   300
               cr
               0)

               LL   200
                   100  -
                     0  -
 Wave

Troughs
                        -100


          Blue bars under wave influence-
          Red bars under trough influence.
                     1     Wave Peaks
                 0
             ErrorPct
100
Figure 4-8. Tri-nodal nature of deviation pattern is consistent with presence of waves.
                                                 10

-------
4.2    Description of Test Equipment

4.2.1   Downstream Gate Site

The downstream gate site was the  primary  site for the verification of the flow meters.  The
ADS 4000 and the reference meters were installed at this site.  This is also where the level of
back flow was controlled using the downstream gate, where the diluted  tracer samples were
collected and the manual level readings taken. Figure 4-9 shows  the downstream gate site during
a tracer dilution test, while one technician was  taking manual  level readings and another was
collecting diluted samples.
                            Manual level measurement tape
Figure 4-9. General view of the downstream test site.

4.2.1.1 Verified Flow Meter: ADS 4000

As shown in Figure 4-2, the ADS  4000 flow meter was installed 16.7 ft (5.1 m) upstream of
manhole S-l.  The flow  meter is composed of a transmitter and a pressure level probe, an
ultrasonic level probe and  a velocity probe, as shown in Figure 4-10.  All probes were  pre-
installed on a stainless steel band, which was installed in the sewer using a bender that extends
the diameter of the band to squeeze it on the pipe wall, as shown in Figure 4-11. The transmitter
was hung on the ladder in the manhole. The submersible pressure probe was at the bottom of the
pipe and the velocity probe was close to it, slightly off-center. Wires for the submerged probes
were attached using tie-wraps on the downstream side  of the stainless steel band.  The probes
were small and had a good hydrodynamic profile, causing no discernable disturbance to the flow
pattern.
                                           11

-------
Figure 4-10. ADS 4000 flow meter components.
Figure 4-11. ADS 4000 flow meter installed in pipe.
                                        12

-------
4.2.1.2 Transit-Time Reference Flow Meter

The  reference flow meter transducers (probes) were installed in a pipe section from 5.4 ft
(1.65 m) up to 9.2 feet (2.80 m) upstream of manhole S-l (Figure 4-12(a)). The transmitter was
located in the electrical room of Station #80A approximately 130 ft (40 m) from manhole S-l. A
description of the operating principle  for the reference flow meter is provided in Appendix A,
along with details of its configuration and installation drawings.
  ADS 4000
  flow meter
  Deflector
 installed over
the outlet of the
 bubbler tube
               Reference
               ultrasonic
                             Second path,
                             13" over the
                             invert
                             Reference
                             flow meter
                               probe
                                 path
                             submerged, 4"
                             over the invert
Reference flow meter.
                                                       (b)  Bubbler installation.
Figure 4-12. Reference flow devices installed in the test pipe.

4.2.1.3 Reference Flow Meter Bubbler

The injection pipe for the bubbler level meter was located 16.1  in. (410 mm) downstream from
the reference flow meter transducers described in  Section 4.2.1.2. To avoid a venturi effect due
to high velocities at the outlet of the injection pipe,  a deflector plate was  installed at the end of
the pipe, as shown in Figure 4-12. This instrument transmitted the water level measurements via
the local programmable logic controller (PLC) to  the reference flow meter. A periodic,  100
pound per square inch (psi) air purge lasting three seconds was automatically triggered to keep
the end of the pipe clear. Specifications for the bubbler system are included in Appendix A.

4.2.1.4 Ultrasonic Reference Level Meter

This meter provided redundancy of the level measurement of the reference bubbler level meter.
The meter's probe was located in the 42 in. pipe, 7.52 ft (2.3 m) upstream of manhole S-l, above
all the paths of the reference flow meter. Mounted  on the crown of the pipe, the probe had an
intrinsic  dead band of approximately two in. (50.8  mm),  allowing the meter to measure from
approximately 0 to 40  in. The transmitter was hung in manhole S-l. Additional information
about the meter is provided in Appendix A.
                                            13

-------
4.2.1.5 Downstream Gate

Manhole S-l is equipped with a manual sluice gate to control the water level in the pipe where
the reference flow meter and the ADS 3600 flow meter were installed. The gate could be closed
to create back-flow conditions in the upstream pipe, but the pipe would return to free surface
flow when the gate was fully opened (Figure 4-2). The  manual actuator for the gate permitted
vertical movement of one inch (25 mm) for four turns of crankshaft. The vertical displacement of
the downstream gate  was measured manually using a fixed aluminum ruler graduated every
1/16 in. (2 mm). The ruler was attached vertically to the top of the gate and passed through the
roof of the chamber in a standard floor valve box. The displacement was measured using a fixed
point on the side of the floor valve box.

4.2.2  Upstream Gate Site

4.2.2.1 Electrical Control Room

The tests were directed from the  electrical control room of the upstream gate. The technician at
that site logged all information and measurements taken during the test into the computer (Figure
4-13(a). The technician at the upstream gate site was in continuous telephone communication
with the technician at  the downstream gate and used radio communication with the technician at
the injection site during tracer dilution.
                                                                          Reading on the
                                                                          ruler at the top
                                                                          of the screw rod
                                              Gate opening
                                              measurement
                                                 ruler
                                 Telephone
                                 and
 Laptop with
the test report
 spreadsheet
       (a)  Test operator entering data.
Figure 4-13. Upstream gate site arrangements.
                                                     (b)  Upstream gate actuator.
4.2.2.2 Upstream Gate

The upstream control gate at Station #100 permitted water retention in the Versant-Sud tunnel to
control flows during rainfall events. It was used during the tests to simulate of a wide range of
flow conditions. When fully closed, only flow from the Craig Collector would remain in the pipe
at the flow meter location.
                                           14

-------
The gate was equipped with an electrical actuator, but most gate movements were done with the
hand wheel (Figure 4-13(b)). Flow control was made using  a mathematical model of the flow
under the gate that used the downstream bubbler level meter reading, the average value of the
two upstream ultrasonic level meter readings, and the gate opening reading made directly on a
ruler on the gate actuator.

4.3    Description of Reference Methods

Area-velocity flow meters have two measurable components,  depth and velocity, which are
evaluated independently. Reference measurements were required to evaluate the accuracy of
depth and velocity readings from the flow meter under test.  The flow rates for the flows under
the upstream and downstream gates were also used to verify the accuracy of the reference flow
meter.  A reference instrument that recorded depth, velocity,  or flow in the same interval as the
ADS meter was preferred over manual measurements, since the test and reference data sets were
then comparable in size.

4.3.1   Reference Depth Measurement

4.3.1.1 Manual Depth Measurement

Manual measurements of the water depth were made using a standard Solinst level meter tape
with a stainless steel probe at the location specified  in Figure 4-2. The probe was modified to
ensure that both contacts touched the water surface at the same time to avoid dragging the tape,
which  would result in a higher water level  reading.  A metallic tape, graduated  in millimeters,
was attached to the plastic tape since plastic bends more easily than metal.

During the challenge tests (tests B, C, and D), three manual measurements were taken at the site
of the  reference depth meter, roughly every five minutes. The first measurement was taken 30
seconds before the target time, the second 15 seconds before  the target time, and the third at the
target time. During the passive test (test E), five consecutive manual  measurements were made
on most weekdays at 60-second intervals.

4.3.1.2 Reference Depth Meters

Two reference instruments were used to record depths. The bubbler on the Accusonic meter was
the primary reference meter.  The secondary depth meter was the ultrasonic depth meter. Depth
readings from the ADS Model 3600 were compared with depth measurements obtained from the
Accusonic  bubbler. The ultrasonic depth  meter and  the  manual  depth  readings provided
calibration data for the bubbler. Depth measurements from both instruments were logged at one-
minute intervals, whereas manual readings were taken approximately  every five minutes during
challenge tests. The reference depth meters were calibrated prior to initiation of the verification
testing to ensure their accuracy.
                                           15

-------
4.3.2   Reference Velocity Measurement

4.3.2.1 Reference Velocity Computed from Tracer Dilution

Whenever possible, reference velocities were compared to calculated velocities based on the
tracer dilution flow rates and on the flow sections computed from the bubbler depth readings.

4.3.2.2 Reference Velocity Meter

The reference flow meter used during verification was a 4-path Accusonic flow meter. Because
this instrument did not provide the average velocity on the modbus signal (a serial link to output
data), reference velocities were computed from the reference  flow rates divided by the  flow
sections. The flow sections  were  computed  from  the  bubbler  depth readings. To demonstrate
their accuracy and the fact that they exhibit  no bias, the reference flow meter and depth meter
were tested prior to ADS verification testing.

4.3.3   Reference Flow Measurement

Three reference flow measurement methods were used for testing and evaluation of the test data:

   1.  Tracer dilution was used during two of the three replicas  of Test C.
   2.  A reference flow meter (Accusonic) was used continuously for all tests.
   3.  The flow rate under  the downstream gate  was calculated  during tests B and D, when
       back-flow conditions were present,  and the  flow rate under  the upstream gate was
       calculated during tests CO and C3 (to verify accuracy of reference meters).

4.3.3.1 Tracer Dilution Method

Tracer dilutions performed during the verification were based on the guidelines presented in the
protocol. The major elements were:

•  Due to the variation in suspended solids concentrations in combined sewers, lithium chloride
   was used for the tracer dilution tests.
•  For two of the replicates of Test C, tracer dilution was performed to verify the accuracy of
   the reference meter. The first  tracer dilution  was  done before the first replicate  and the
   second was performed during the third replicate, near the end of the challenge tests.
•  No dilution was performed on the day before a non-working day to ensure faster analysis of
   samples. Samples were brought to the laboratories at the end of the day. Samples from the
   dilution performed Sunday, July 15, 2001 were put in a cooler with ice packs  and brought to
   the laboratories the following morning.
•  Samples were analyzed for lithium content by an accredited laboratory in accordance  with
   Method SM-3111B, from Standard Methods for the Examination of Water and Wastewater
   (American Public Health and Association, 20  Edition, 1998) and MA 200-Method 1.0,  from
   Quebec Environment Ministry-Laboratory Expertise Center.
                                           16

-------
•   To increase the accuracy of the analysis of the concentrated samples, each sample was
    analyzed five times by the laboratories (five separate dilutions) and the laboratories provided
    the results for each analysis.

While laboratories may have good in-house accuracy and low variability results (±3 percent for
diluted samples and ±2 percent for averaged concentrated  samples), deviations of more than ±10
percent between laboratories may occur. This is not a problem as long as the difference for both
laboratories remains constant considering the calculation of the tracer dilution flow rate.

4.3.3.1.1  At the Injection Site

A leveled table with the flow injection control system was installed at the injection site. The flow
injection control system consisted of two cylinders for  flow injection control, two metering
pumps and a set of pipes and valves (Figures 4-14 and  4-15). The cylinders were used one at a
time,  with one in use while the other was refilled  and prepared for the  next flow injection
control. Three quarters of a cylinder was injected over a  period of 15 minutes, then the technician
switched to the other cylinder. One metering pump was  used, with the second being available in
case of failure.  Specifications of the equipment used in the injection control  system  are provided
in Appendix A.
                         TRACER INJECTION SET UP DIAGRAM
                                         (no scale)

                                                Continuous flow
                                                measurement
                                              Cylinders and valves
                                 Boulevard Chaudiere interceptor
Figure 4-14. Tracer injection diagram.
                                            17

-------
                           (a) Lithium injection control system.
                                  Lithium
                                  tank
                                  Lithium
                                  transfer
                                  pump
1  •     -
    Debris
X
                                                  Lithium
                                                  injection
                                                 point - 8 ft
                                                 in the pipe
            (b) Mobile unit.
Figure 4-15. Lithium injection system.
           (c) Injection point inside pipe
The tank containing the concentrated lithium solution was located in the injection mobile unit.
The solution was prepared a few days before the test and agitated continuously during the test
(Figure 4-15(b)). The concentrated lithium  was  injected eight feet (2.4 m) downstream of
manhole S-5B through a 0.5 in. (13 mm) diameter pipe installed along the crown of the pipe
(Figure 4-15(c)). The  injection pipe had a counter-pressure foot valve to minimize the variation
in the injection flow rate under surcharge flow conditions.
                                            18

-------
Blank samples were collected every 30 minutes, with one additional blank collected at the end of
the test. A sample of the concentrated lithium solution was taken every two hours, and at the end
of the test. The concentrated samples were collected with the peristaltic pump from the automatic
water sampler used to refill  the  cylinders. Blank samples were taken using a  sampling rod
(Figure 4-16) on which the bottles were installed. After sampling, preprinted labels were applied
to the bottles.
                                                      Watertight bag
                                                      to avoid bottle
                                                      contamination
Figure 4-16. Blank sampling site.

4.3.3.1.2  At the Sampling Site

Diluted samples were taken every five minutes at the downstream gate sampling site. Samples
were placed into the cooler in a plastic bag to separate them from the blanks and concentrated
samples that were collected at the injection site.

4.3.3.2 Reference Flow Meter

Reference flow measurements were logged at one-minute intervals.

4.3.3.2.1  Flow Under the Gate

The  flow rates under  the upstream  and downstream gates were calculated from a standard
hydraulic equation and were used as a redundant method to verify the accuracy of the reference
meter. The calculation depended on depths upstream and downstream of the gate and the height
of the gate opening. The downstream gate was never submerged because of a steep slope, so the
depth downstream  of the gate was not required for the computation. The following two sections
describe particularities of the application for both gates.
                                           19

-------
4.3.3.2.2  Upstream Gate

The  flow rate under the upstream gate was calculated  for tests  CO and  C3, to provide an
estimation of the flow rate measurement at the downstream gate. Graphs showing the flow rate
measurements for  these tests  are  included in Appendix H.  The  calculation  depends on the
upstream and downstream depths, the gate opening, the flow rate of the Craig Collector, and the
transit time from the upstream gate to the downstream gate. An estimation of the flow rate at the
downstream gate was computed by delaying the summation of the flow rate at the upstream gate
and the flow rate of the Craig Collector by 15 minutes because of the distance between the gates.

The flow rate of the Craig Collector was generally inconsequential  and therefore not measured.
The dry weather flow rate curve was used as the  flow  rate control during active tests. Based on
the observations made during these last tests, the precision on this flow rate is estimated to ±0.23
MGD(±10L/sec).

The  equation of the flow rate under the upstream gate was calibrated for this site  during the
installation  of the real-time  control  system of the QUC. The methods  used  for  these
measurements are discussed in  Section 4.2.1.

4.3.3.2.3  Downstream Gate

The downstream gate was used to create back flow conditions in the section of pipe used for the
verification testing. Prior to the testing,  preliminary calculations were  completed by the TO to
estimate the approximate gate positions needed to achieve different back flow conditions in the
pipe. Because there is a steep slope downstream of the gate, the calculation depends only on the
upstream depth and the gate opening.

4.4    Experimental Procedures

This section describes the tests conducted as part  of the flow meter verification.  The tests can be
classified  as either general evaluations  or performance  evaluations,  and are summarized in
Table 4-1. The general evaluation of the flow meter included:

•  Software evaluation (Test A);
•  Ease of operation and maintenance (Test E);
•  Potential for debris accumulation (Test E); and
•  Data retrieval (Test E).

The performance evaluation for accuracy of depth and velocity  measurements  and accuracy of
the flow calculation included tests:

•  Under a controlled range of flow conditions (Tests B, C, and D);
•  During simulated rainfall events (Tests C and D); and
•  Under short-term continuous operation (Test E).
                                           20

-------
Test F, operation and maintenance under extended operations, was not performed, as noted in the
verification test plan (VTP).

Table 4-1. Flow Meter Verification Test Overview

         „  ,                         ~,.  ,.                          Evaluation
         Test                         Objectives                  „      ,  „  ,.
	General  Performance
 Flow meter software   User-friendliness, functionality, and
                      flexibility                                      ^
 Accuracy under dry    Accuracy of depth and velocity measurements
 weather flow with     and accuracy of the flow calculation under        .          g
 back-flow conditions  controlled range of flow conditions
 Accuracy under wet   Accuracy of depth and velocity measurements
 weather flow          and accuracy of the flow calculation under        _          Q
                      controlled range of flow conditions
 Accuracy under wet   Accuracy of depth and velocity measurements
 weather flow with     and accuracy of the flow calculation under        .          Q
 back-flow conditions  controlled range of flow conditions
 Accuracy under       -  Ease of operation and maintenance
 short-term            -  Potential for debris accumulation
 continuous operation  -  Data retrieval                               E          E
                      -  Accuracy of the flow calculation under
	controlled range of flow conditions	

4.4.1   General Evaluation (Test A: Flow Meter Software)

The objective of this test was to evaluate the software provided by the flow meter manufacturer
for user-friendliness, functionality, innovation,  and compatibility.

4.4.1.1 Procedures

This was a qualitative  evaluation,  parts of which  were  performed in  the  field  during
configuration,  calibration and data collection. Other parts  were completed in the  TO's office
after the flow meter was removed from the sewer.

The two  flow meter software programs were  installed  on a PC  provided by the  testing
organization and  on a PC provided by  ADS. The specifications  for  the PCs used for this
evaluation are included in Table 4-2.

4.4.1.2 Measurements

All of the elements that  the vendor had  indicated  were  available  at the  beginning  of the
verification were tested. The  elements were rated  as "possible" or "not  possible," according to
                                           21

-------
 the  ability of the flow meter software  to process them. Relevant comments concerning the
 evaluated elements were noted by the TO.

 Table 4-2. Personal Computers Used in Flow Meter Software Evaluation


	Testing Organization PC	ADS PC	
 Manufacturer            Toshiba                             Dell
 Processor                Pentium 2-400 MHz                  Pentium
 Hard Disk               6.4 GB, > 1 GB disk free              > 1 GB disk free
 Disk Drives              CD ROM                           CD ROM, 3.5 in. floppy
 Operating System        Windows 95                         Windows 95


 4.4.2  Performance Evaluation

 In the following  procedures,  the waiting period to reach  steady flow and  tracer  concentration
 conditions was established by experimentation.

 4.4.2.1 Test B: Accuracy under Dry Weather Flow with Back-Flow Conditions

 Test B verified the accuracy  of depth and velocity measurements and flow calculations during
 field-simulated dry weather  flow conditions (low flow), and when subjected  to  back-flow
 conditions caused by a downstream obstruction. This test was run three times.

 4.4.2.1.1  Procedures

 For every trial of Test B, the following procedure was applied:

 1.     Adjust the  upstream  gate to  a  position  corresponding to a flow of  approximately
       1.71 MGD (75 L/sec) at the flow meter location.

 2.     Initiate the  reference flow, velocity  and depth  measurements; manually verify the
       reference depth measurements.

 3.     Wait 45 minutes to establish a steady flow at the flow meter location and maintain this
       steady flow condition for 30 minutes.

 4.     Close the  downstream gate as necessary to establish a flow of approximately 0.86 MGD
       (37.5 L/sec); hold the flow for about five minutes to accumulate a head of water equal to
       18 + 2 in.  at the flow meter location.

 5.     Open the downstream gate as necessary to establish  a steady flow of 1.71 MGD
       (75 L/sec) and a head of water equal to 18 + 2 in. at the flow  meter location. A steady
                                           22

-------
       flow and depth was established at the flow meter location approximate 60 minutes after
       step 4; maintain this flow rate for 60 minutes.

6.      Close the downstream gate for about 25 minutes to accumulate a head of water equal to
       36 + 4 in. at the flow meter location.

7.      When a head of 36 in. is reached, open the downstream gate to establish a steady flow of
       1.71 MOD (75 L/sec) and a head of water equal to 36 + 4 in. at the flow meter location.
       A  steady flow and depth was established at the flow meter location approximately 90
       minutes after the beginning of step 6; maintain this flow rate for 60 minutes.

8.      Collect data from the  ADS 3600 and reference meters.

4.4.2.1.2  Measurements

During the testing under Test B, measurements were taken in the following manner:

1.      Log the depth, velocity and flow of the ADS 3600 and the reference meters each minute.

2.      Log every gate manipulation, with the corresponding time and the depths upstream and
       downstream of the gate, for the upstream gate.

3.      Log every gate manipulation for the downstream gate, with the corresponding time and
       depth upstream of the gate.

4.      Record noteworthy events and the corresponding times.

5.      Take manual depth measurements at five minute intervals  for comparison with values
       recorded by the reference depth meter. Each time measurements  are taken, measure the
       depth three times at 15-second intervals, the third measurement to be taken at the exact
       time that the data is reported. For example: the first reading at 10:19:30, the  second at
       10:19:45, and the third at 10:20:00, the mean value of the three readings to be reported at
       10:20:00.

4.4.2.2 Test C: Accuracy Under Wet Weather Flow Without Back Flow Conditions

Test C verified  the accuracy of depth and  velocity measurements and flow calculations during
field-simulated wet weather flow conditions (normal and high flow). This test was  run three
times.

4.4.2.2.1  Procedures

For every  trial of Test C, the following procedure was applied:
                                           23

-------
1.      Adjust the upstream  gate to a position corresponding to a  flow of approximately
       1.71 MGD (75 L/sec) at the  flow meter  location.  Start the  tracer injection when
       applicable.

2.      Initiate reference  flow, velocity,  and depth measurements; manually verify reference
       depth measurements.

3.      Wait 45 minutes to establish a steady flow (and tracer concentration when applicable) at
       the  flow  meter  location and maintain  flow (and  concentration  when applicable)
       conditions for 30 minutes.

4.      Progressively open the upstream gate during a 10-minute period to establish a flow equal
       to 17.1 MGD (750 L/sec) ±1.14 MGD (50 L/sec) at the flow meter location. A steady
       flow (and tracer  concentration, when applicable) was established at the flow meter
       location approximately 45 minutes after beginning this  step. Maintain this flow rate for
       30 minutes.

5.      Close the upstream gate during a  five-minute  period  to establish  a flow equal  to
       8.56 MGD (375 L/sec) ±1.14 MGD (50 L/sec) at the flow meter location. A steady flow
       (and  tracer concentration, when applicable) was established at the flow meter location
       approximately 45 minutest  after beginning this  step.  Maintain  the flow  rate for 30
       minutes.

6.      Open the upstream gate during a ten-minute period to establish a flow equal to 29.7 MGD
       (1,300 L/sec) ± 1.71 MGD  (75 L/sec) at the flow meter location. A steady flow (and
       tracer concentration,  when  applicable)  was established at  the flow  meter location
       approximately 45 minutes after beginning this step. Maintained the flow rate for 30
       minutes.

7.      Close the upstream gate during a  ten-minute  period to establish  a  flow equal  to
       8.56 MGD (375 L/sec) ±1.14 MGD (50 L/sec) at the flow meter location. A steady flow
       (and  tracer concentration, when applicable) was established at the flow meter location
       approximately 45 minutes after beginning  this  step.  Maintain the flow  rate for 30
       minutes.

8.      Close the upstream  gate during a five-minute period to establish a flow of 1.71 MGD
       (75 L/sec) at the  flow meter location. A steady flow  (and tracer concentration, when
       applicable) was established  at the flow meter  location approximately 45 minutes after
       beginning this step. Maintain the flow rate for 30 minutes.

9.      Stop  the tracer injection, when applicable.

10.    Collect data from the ADS 3600 and the reference meters.
                                           24

-------
4.4.2.2.2  Measurements

Measurements made during testing under Test C were taken in the same manner as for Test B.
Refer to Section 4.4.2.1.2 for more detail.

4.4.2.3 Test D:  Accuracy Under Wet weather flow With Back flow Conditions

Test D verified the accuracy of depth and velocity measurements and flow calculations during
simulated wet weather flow (high  flow) when subjected to back-flow conditions  caused by a
downstream obstruction. This test was run three times.

4.4.2.3.1  Procedures

For every trial of Test D, the following procedure was applied:

1.      Adjust the upstream gate  to  a position  corresponding  to  a flow of approximately
       1.71 MOD (75 L/sec) at the flow meter location, according to the calibrated upstream
       gate equation.

2.      Initiate reference  flow, velocity, and  depth measurements; manually  verify reference
       depth measurements.

3.      Wait 45 minutes to establish a steady flow at the flow meter location and maintain this
       steady flow condition for 30 minutes.

4.      Open the  upstream gate during a five-minute period to establish a flow equal to 8.56
       MOD (375 L/sec) ±1.14 MOD (50 L/sec) at the flow meter location,  according  to the
       calibrated  upstream gate equation. A steady flow was established at the  flow  meter
       location approximately 45 minutes after beginning this step. Maintain the flow rate  for 30
       minutes.

5.      Open the upstream gate to establish a flow of approximately 11.4 MOD (500 L/sec) at the
       upstream gate to  accumulate a head of water equal to 54 ± 4  in. at the  flow  meter
       location.

6.      Position the downstream gate at 23/s in. (60 mm) to establish a  head of water equal to 54
       ± 4 in.  at the flow meter location. When the depth of water is close to 54 ± 4 in., operate
       the  downstream gate to maintain a head of water equal to 54  ±  4 in. at the flow  meter
       location.

7.      Three minutes after the beginning of step 5, close the upstream gate as necessary to
       establish a flow equal to 8.56 MOD (375 L/sec) ±1.14 MOD (50 L/sec)  at the flow meter
       location, according to the calibrated upstream gate equation. A steady flow and depth was
       established at the flow meter location  approximately 60 minutes  after  the beginning of
       step 5. Maintain this flow rate for 30 minutes.
                                           25

-------
8.      Open  the  upstream gate during a  five-minute period to establish  a  flow equal to
       17.1 MOD  (750 L/sec)  ±1.14 MOD  (50 L/sec)  at the flow  meter location, while
       adjusting the downstream gate to maintain a head of water equal to 54 + 4 in. at the flow
       meter location.  A steady flow and  depth was established at the flow  meter location
       approximately  60 minutes after  beginning this step. Maintain  this flow rate for 30
       minutes.

9.      Open the upstream gate as necessary  during a ten-minute period to establish a flow equal
       to 29.7 MOD (1,300 L/sec) +  1.71 MOD (75 L/sec) at the flow meter  location, while
       adjusting the downstream gate to maintain a head of water equal to 54 + 4 in. at the flow
       meter location.  A steady flow and  depth was established at the flow  meter location
       approximately 60 minutes after beginning this step.  Maintain  the flow and depth for 30
       minutes.

10.    Maintain the downstream gate  position.  Progressively close  the upstream gate as
       necessary during a 10-minute period to establish a flow equal to 1.71 MGD (75 L/sec) at
       the flow meter  location, according to the calibrated upstream gate  equation.  A steady
       flow was established at the flow meter location approximately 60 minutes  after beginning
       this step. Maintain the flow and depth for 30 minutes.

11.    Collect data from the ADS 3600 and the reference meters.

4.4.2.3.2  Measurements

Measurements made during testing under Test D were taken in the same manner as for Test B.
Refer to Section 4.4.2.1.2 for more detail.

4.4.2.4 Test E: Accuracy Under Short-Term Continuous Operation

Test E verified the accuracy of depth and velocity measurements and flow calculations over a
21-day period of continuous operation.

4.4.2.4.1  Procedures

1.      There was one period following completion of tests B, C, and D, during which flow data
       were not retrieved during seven consecutive days of data collection.

2.      The TO was responsible for the proper operation and maintenance of the flow meter in
       accordance with the operating instructions. Any intervention required to  maintain the
       flow meter in good operating order  was  authorized and done by the vendor. The TO
       recorded the nature and frequency of any operation and maintenance procedures required
       during a 21-day period of continuous  operation.

3.      The TO performed the procedures listed  in Section 4.4.2.4.2 as well as other specific
       procedures  specified in the ADS 3600 operating manual.
                                           26

-------
4.      The personnel required for operation and maintenance procedures were also classified
       according to their qualifications as engineer, technician or general laborer.

5.      The TO noted whether any specialized tool or equipment was required, and whether the
       need for such tools  or  equipment was indicated by the  operation  and  maintenance
       instructions.

4.4.2.4.2  Measurements

The  nature and frequency of operation and maintenance procedures required during a 21-day
period of continuous operation were  noted by the  TO, and are reported in Section 4.5.2.3.
Specifically, the TO completed the following measurements:

1.      The numbers  of hours (rounded to nearest whole hour) for each personnel classification
       were recorded for any field intervention to maintain the flow meter in operating order, in
       accordance with the operation manual or the procedures recommended by the flow meter
       manufacturer.

2.      The time, depth, velocity, and flow were recorded by the flow meter under test and the
       reference meters during all  periods of normal operation,  including the periods covered
       under tests B,  C, and D.

3.      Manual  depth measurements were taken at least once a day at the flow meter under test
       for comparison with values recorded by the verified depth meter.  For each measurement
       time, the depths were measured five times at intervals of 60 seconds. During this visit, the
       depth and  velocity probes were inspected to observe  debris accumulation or  other
       problems. The presence of accumulated debris on the probes was noted and documented
       by photographs.

4.      After dismantling,  the flow meter and probes were inspected for infiltration, broken,
       cracked, or scratched components. The extent and nature of the damage was described.
       Evidence of water infiltration or broken components was noted and  documented with
       photographs.

5.      Any unusual event occurring during the test was noted with its corresponding time.

4.5    ADS 4000 Evaluation Results

4.5.1  Software Evaluation

The ADS 4000 flow meter required two different software programs: Fieldscan™ V.3.1, released
on March 28, 2001, and Profile V.I.4.1, released on April 6, 2001. ADS provided the TO with a
portable computer using a Microsoft Windows 95™ operating  system  with  the two software
programs installed. The TO tested the installation and start-up procedures,  along with the usual
applications of the software as used during the flow meter testing period.
                                           27

-------
TO personnel received training on the flow meter software at the beginning of the test. This six-
hour training, combined with the software user guide, allowed TO personnel to become familiar
with the flow meter software and to accomplish the required tasks for the flow meter verification.
Some complex operations performed afterward with the flow meter software required the help of
the ADS analysts.

Fieldscan™ was  used to set  the  parameters for the on-site installation as well as  for the
configuration and calibration of the flow meter. It can also collect data from the flow meter and
allow real-time viewing of the monitor status and data.

When the Fieldscan™ installation was completed, the Profile  software was used to collect data
from the monitor,  store  it in a database, and view, analyze  and export it. The Profile software
cannot be used to modify set-up parameters  on the flow meter. Fieldscan™ must be used to
perform these tasks.

Fieldscan™ stores parameters and data in files while Profile uses a Microsoft Access database.
Fieldscan™ data and  parameters can be imported into the Profile database but not vice-versa.
Both programs must be used during normal  operation, but since they do not use the same source
files, the user has to keep both sources of information updated.

The manufacturer  first completed  a  checklist of available  features and the TO verified their
availability. A detailed evaluation of the features of both programs is presented in Appendix G.

Both programs were easy to learn, and quick and easy to use. Profile was a powerful analysis
tool with  a large variety of features. The procedure to  collect data was simple and efficient,
taking about 1.5  minutes to record  data collected at one-minute  intervals. A few additional
minutes were required to view data on graphs and to make sure that the collection was complete.
Considering that this task had to be performed frequently under difficult conditions, the features
of Profile were a significant advantage.

The use of a database to store data was also a significant advantage.  It allowed users to easily
access data from several monitoring points  and for any time period. It could also easily export
data in ASCII or Microsoft Excel  files  for other uses. A disadvantage was that two software
programs are needed to perform the required operations. The user had to learn both programs and
keep  both sources of information individually  updated, which  could become a  source of
problems.

One software issue was identified during testing. As tests commenced, the TO noted the program
implemented in the flow meter would occasionally fail to store data points in its memory. The
protocol required the  flow meter to electronically collect mean velocity, ultrasonic depth, and
pressure depth data, and compute flow rate at one-minute intervals. The ADS  flow meter and
laptop computer, as supplied and set up during testing, was not capable of retrieving, computing
and storing all the data within the one-minute interval; it would respond by not storing any of the
data at the particular time interval.  The missing data points would typically be spread randomly
throughout the dataset, with at most four or five consecutive missing data points.
                                           28

-------
True zero readings were represented on the graphs generated by the program while missing data
were not. Data points around the missing data were simply connected. ADS personnel indicated
the program settings could be modified to provide the one-minute readings. The quantity of data
points retrieved was sufficient to provide verification of flow meter effectiveness.

The number of missing data points for each test was noted. The minimum number of missing
data during a 24-hour period was 119, the maximum was 670 and the mean was 389 out of 1,440
data points.  This represents 8 percent,  47 percent, and 27 percent of the daily data collection,
respectively.

A complementary test was  conducted by the TO on  July 24 and 25, 2001 to further investigate
the issue. During  this test, data was  logged at two-minute intervals instead of one-minute
intervals for approximately  24 hours. During this test, no data points were missing from the time
series. It was hypothesized  that the root of the problem was the flow meter processor, and that
the program implemented in the flow meter was not able to process and store the volume and
frequency of data  being provided by the ADS  flow meter  probes  (one pressure probe, four
ultrasonic probes and one velocity probe) or to calculate the flow rate.

The Utah Water Research Laboratory (UWRL) conducted laboratory testing with the same ADS
flow meter equipment. They collected data at one-minute intervals, but did not have the problem
with data collection and storage  experienced at the field testing site.

4.5.2  Performance Tests

The performance tests were completed between June 28 and  July 24, 2001  according to the
procedures outlined in the VTP. Tests were performed on days without precipitation (morning,
afternoon or evening), in no particular sequence except that the first and second to last tests were
two of the three replicates of Test C, for which dilution tests were done. No data were collected
for seven consecutive days  (July 17 to 24) as specified in the protocol. The results from tests B,
C, and D are summarized in Table 4-3.

Results  of all performance tests are presented in graphical form with one-minute data for  all
instruments (Accusonic, bubbler, ADS 3600). Clocks from all instruments were synchronized
and no more than a two-minute difference was observed for the duration of the tests.
                                           29

-------
Table 4-3. Deviation from Reference Flow—Tests B, C, and D

 „,       .                  Average deviation           Trimmed average deviation
 Flow regime                   /•       ^                         /       x\a
 	(percent)	(percent)
TestB
TestC
TestD
Test B-D combined
Simulated dry flow
Simulated wet flow
Combined flows
-14.5b
14.0
-0.8
-0.4
0.5b
-1.3
-0.4
-0.9
14.5
8.3
3.8
9.5
-1.0
3.8
 a  The trimmed average deviation removes the deviations greater than 99 percent or less than -99 percent from the
   mean deviation, and then averages the remaining values. This computation mitigates the skewing caused by
   large outliers.
 b  The ADS 4000 flow meter read zero velocity during low-flow testing, resulting in a deviation of -100 percent
   and a skewing of the data.

The hydraulic conditions at the test site created a condition where water levels  at the reference
site (bubbler) and at the ADS 4000 flow meter were different. These two measuring sites were
approximately 12.7 ft (3.88 m) apart, on a theoretical slope of 0.169 percent (elevation difference
of 0.26 in.,  or 6.6  mm). The two pipe  sections involved  were on  slightly different  slopes,
resulting in a measured elevation  difference  of 0.79 in. (20  mm). Furthermore, there was a
2.17'-in. (55-mm) high concrete bump at the end of the pipe,  approximately four feet (1.2 m)
downstream from the bubbler, which caused a permanent back flow over both the bubbler and
the ADS 4000 sites.

4.5.2.1 Scatter Plots

Two types of scatter plots are presented in this report: a system behavior (velocity versus depth)
plot and a flow rate comparison plot. The system behavior scatter plot is shown in Figure 4-17.
The figure presents the hydraulic  conditions  at the ADS  4000 testing site during the three
replicas of Test C, when there was no back flow. There are 946 data points on this figure, which
includes data from all three replicas: stable (698 data points), ebbing  (178), and rising (70).
There are 407 missing data points compared to the reference data set. There are fewer data points
for the rising water profiles because the transition was more rapid under this condition.
                                            30

-------
                           10
                                      15         20

                                       Level (inches)
                                                           25
                                                                      30
                                                                                35
Figure 4-17. ADS 4000 system behavior—Test C.

Figure 4-18 presents the same scatter plot for the reference flow meter. Three observations can
be made:

    1.  The four stable water profiles of Test C (1.71, 8.56, 17.1, and 29.7 MOD; 75, 375, 750,
       and 1,300 L/sec) are clearly defined;

    2.  The ebbing water profiles data points are generally over the best-fit curve drawn from
       stable conditions data points.

    3.  The rising water profiles data points are generally under the best-fit curve.

These conditions are not observed on Figure 4-17 for the ADS 4000.
                                            31

-------
    10
               Stable
               Ebbing
               Rising
               Best fit curve of stable conditions
                 y = -0.0016X' + 0.2987X -1.6057
                           10
                                     15         20

                                       Level (inches)
                                                           25
                                                                      30
                                                                                35
Figure 4-18. Reference meter system behavior.

Figure 4-19 presents a flow rate comparison plot between the ADS 4000 and the reference flow
meter for all data from tests B,  C, and D. The line on the figure represents a perfect fit between
the two  flow  meters.   The reference  flow meter recorded  3,924  data points (five  true zero
readings and five true negative values) during those tests while the ADS 4000 recorded 3,309
data points  (615 missing data points).  The zero readings and negative values of the reference
flow meter were recorded during a transition period during Test B just after the gate was closed
and there was some counter-flow.
                                            32

-------
                                   Reference flow rate (MGD)

Figure 4-19. ADS 4000 flow vs. reference flow, Tests B, C and D.

Of the 3,309 matching data points between the ADS 4000 and the reference flow meter, the ADS
4000 had 157 true zero readings and no negative readings, while the reference flow meter had
one true zero reading and three negative readings. Most (152) of the 157 true zero readings from
the ADS 4000,  152 were recorded during Test B (low flow) with 36-inch back flow, while the
reference flow meter recorded flow in the range of 1.37 to 2.05 MGD (60 to 90 L/sec). Under
these  conditions,  velocities were in the  range of 0.33 ft/sec (0.1 m/sec). The resolution of the
velocity meter on the ADS 4000 was 0.0394 ft/sec (0.012 m/sec).

4.5.2.2 Deviation Distribution Plots

Figures 4-20 to 4-25 present the ADS 4000 flow deviation distributions. For all these plots, the
+100 percent deviation bar groups all data equal  or greater than 100 percent, whereas the -100
percent deviation bar groups  all  data points equal to zero (the  ADS 4000 does not output
negative  flow rates). The deviation  is presented in two-percent increments. Indicated on the
figures are the :

•  Margin of deviation (±X percent) that contains 95 percent of all measurements (i.e.  standard
   ±2o);
•  Percentage of measurements within a margin of deviation of ±20 percent;
•  Percentage of measurements within a margin of deviation of ±10 percent;
•  Mean deviation of all  measurements (influenced by very large positive deviations);
                                           33

-------
•   Mean deviation of measurements within a margin of deviation of ±99 percent (to exclude
    very large deviations); and
•   Median deviation (gives an indication of the spread).

Figure 4-20 presents the deviation distribution for Test B. It is a two-peak bell-shaped curve,
with the peaks centered at approximately -2 percent and 24 percent. There is no distinct pattern
to explain the two peaks. The high frequency (16.2 percent) of data points with an deviation of
-100 percent is due to ADS 4000 flow rates recording zero  when the velocity meter recorded
velocities of zero (refer to discussion on Test B below). The zero flow data had a large impact on
the mean deviation (-14.5 percent) compared to  the median deviation (-3.5 percent). Two data
points had deviations larger than +100 percent.  These two data points were recorded during a
transitional flow period, and are considered to be an  artifact of testing. The data points were
therefore removed from accuracy calculations.
             95% of measurements within amargin of deviation of ± 100%
             55.2% of measurements within a margin of deviation of ± 20%
             33.9% of measurements within a margin of deviation of ± 10%
                 Mean deviation: -14.50/
Mean deviation of data within ±99% of deviation: -0.9%

                Median deviation : -3.5%
                                        -10%  0%

                                        Deviation (%)
Figure 4-20. Flow deviation distribution—Test B.

Figure 4-21 presents the deviation distribution for Test C. It is a two-peak bell-shaped curve,
with peaks centered at approximately 18 percent and -46 percent,. The small peak (deviations
from -38 percent to -52 percent) was due to the large oscillations on the depth measurements
during stable flow periods. There are seven data points with deviations larger than +100 percent
recorded in the transition period  between 1.71 to 17.1 MGD (75 to 750 L/sec) when there are
large  oscillations in  velocity measurements. These two  data points were  recorded during  a
transitional flow period,  and are considered to be an artifact of testing.  The  data points were
therefore removed from accuracy  calculations.
                                             34

-------
             95% of measurements within a margin of deviation of ±  49%
             55.9% of measurements within a margin of deviation of ± 20%
             18.8% of measurements within amargin of deviation of ± 10%
                  Mean deviation: 14 0°/
 Mean deviationof data within ± 99% of deviation: 14.5^

	 Median deviation: 17.0^
                    -70% -60% -50% -40% -30%  -20% -10% 0%  10% 20%  30% 40%  50% 60%  70% 80% 90% 100%

                                          Deviation (%)
Figure 4-21. Flow deviation distribution—Test C.
Figure 4-22 presents the deviation distribution for Test D. It is roughly a two-peak bell-shaped
curve, with peaks centered at approximately -10 percent and 26 percent. The peak centered at
approximately 26 percent  is due to the ADS 4000 calculating flow rates higher than reference
flow rates during dry weather periods (1.71  MGD; 75 L/sec) while the other peak is due to the
ADS 4000 calculating flow rates lower than reference flow rates during back flow. There are 11
data points with deviations larger than +100 percent (the largest being +388 percent) that were
recorded during  the transition period between  1.71 to  8.56 MGD (75  to 375  L/sec) at the
beginning of the test. These two data points were recorded during a transitional flow period, and
are considered to be an artifact of testing. The data points were therefore removed from accuracy
calculations.

-------
            95% of measurements within a margin of deviation of ±  40%
            64.8% of measurements within a margin of deviation of ± 20% —
             15.3% of measurements within a margin of deviation of ± 10%
                 Mean deviation: -0 8%
Mean deviation of data within ± 99% of deviation: 8.3% -

                Median deviation: -10.0%
                                        Deviation (%)


Figure 4-22. Flow deviation distribution—Test D.


Figure 4-23 presents the deviation distribution for tests B, C, and D combined. It is roughly a
two-peak bell-shaped curve, with peaks centered at approximately  -10  percent and 20 percent.
The peak centered at approximately  20 percent is primarily due to  over-estimations during low
flow testing periods (1.71 MGD; 75  L/sec) and during Test C.  The other peak is mainly due to
underestimations during back flow during Test D. Data points with  an deviation of-100 percent
were due to the zero flows during Test B.

Figure 4-24 presents the deviation distribution for dry weather conditions and Figure  4-25 for
wet weather conditions. The criterion for wet weather flow for  a flow above 1.5 times maximal
dry weather flow (1.71 MGD;  75 L/sec) or. 2.57  MGD (112.5 L/sec). Figure 4-11  shows a bell-
shaped curve centered on a positive value (18 percent) while the wet weather flow plot is a two-
peak bell-shaped curve (centered on  -10 percent  and 20 percent), showing the influence of Test
D and Test C, respectively.
                                             36

-------
       g 4%

       I
      '
              95% of measurements within a margin of deviation of ±  80%
                59.6% of measurements within a margin of deviation of ± 20"^
                21.5% of measurements within amargin of deviation of ± 10%
                       Mean deviation: -0.4°/i
Mean deviation of data within ± 99% of deviation: _ 3.8%
                     Median deviation: 4.2%
          -100% -90% -80% -70% -60% -50% -40% -30% -20% -10%  0%  10%  20%  30%  40%  50%  60%  70%  80%  90% 100%
                                                    Deviation (%)
Figure 4-23. Flow deviation distribution—Tests B, C and D.
                                                                               Median deviation: 12.4%
  **-.
  u
  C
                    -70% -60%  -50% -40%  -30% -20%  -10%  0%  10% 20%  30%  40%  50%  60%  70%  80%  90%  100°/i
                                               Deviation (%)
Figure 4-24. Flow deviation distribution—simulated dry weather conditions.
                                                           37

-------
                                      -10% 0%

                                      Deviation (%)
Figure 4-25. Flow deviation distribution—simulated wet weather conditions.
For tests B, C, D,  and E, figures are presented in Appendix H in the following order:  (1) flow
rates; (2) water depths; and (3) velocities. While flow rates measured by the verified flow meter
and the reference  flow meter can be compared directly,  water depths  and  velocities cannot,
because the flow meters were not installed at the same location in the pipe. Since the pipe profile
was  known, it was possible to  have a good appreciation of the  quality of the water depth
readings. It was much more difficult to evaluate the velocity readings on their own.

4.5.2.2.1   TestB

After some flushing of the sewer  line, this test was conducted at maximal  dry weather flow
(1.71 MGD; 75 L/sec) with three consecutive hydraulic conditions:

   1.  no back flow;
   2.  with a back flow of 18 in. (457 mm); and
   3.  with aback flow of 36 in. (914 mm).

Figures H-l, through H-3  (Appendix H) present flow rates measured by the ADS 4000 and by
the reference flow meter for each replica of Test B. Flow rates measured by both meters were
close except during a 36-inch back flow condition for all replicas. During replicas B2 and B3, the
ADS 4000 sporadically recorded velocities of zero, and thus calculated flow rates  of zero. Table
4-4 gives an appreciation of the deviation between the flow rates (reference/ADS 4000) for
samples of data  points  taken during stable periods (delimited by the indicated  start  and end
times) for each replica and for the three hydraulic conditions of this test. Also shown in this table
                                           38

-------
are the average reference flow rates, the standard deviations of the flow rate differences and the
percentage these differences represent in comparison with the reference flow rates. Without back
flow, flow rates are overestimated by 11 to 12 percent for replicas Bl and B2 and by 23 percent
for replica B3. Under the 18-inch (457 mm) back flow, flow rates were within 5 to 6 percent for
replicas Bl and B3, while they were underestimated by an mean of 31 percent for replica B2.
Under the 36-inch (914 mm) back flow, flow rates were overestimated by an mean of 25 percent
for replica B1.

Table 4-4. Flow Rate Comparison of Data Point Samples for the Three Replicas of Test B
Hydraulic
Condition


Flow rate of 75
L/sec
No back flow



Flow rate of 75
L/sec
Back flow of 18
in.


Flow rate of 75
L/sec
Back flow of 36
in.



Start time
End time
Mean reference flow rate (L/sec)
Mean difference (L/sec)
Standard deviation of difference
Difference/reference (percent)
Start time
End time
Mean reference flow rate (L/sec)
Mean difference (L/sec)
Standard deviation of difference
Difference/reference (percent)
Start time
End time
Mean reference flow rate (L/sec)
Mean difference (L/sec)
Standard deviation of difference
Difference/reference (percent)

Bl
9:40
10:15
79.6
-8.6
4.8
-11
10:30
12:15
74.0
4.6
6.0
6
12:45
14:45
78.2
-19.5
6.3
-25
Replica
B2
9:40
10:15
64.9
-8.1
4.8
-12
10:30
12:15
68.1
21.4
10.9
31
12:45
14:45
77.7
50.1
37.7
64

B3
19:40
20:15
71.1
-16.1
5.5
-23
20:30
22:15
89.1
4.7
9.5
5
22:45
00:45
83.6
35.0
33.8
42
Average
(L/sec)


71.9
-10.9




77.1
10.2




79.9
21.9


Figures H-4 through H-6 present water depths measured by the ADS 4000 and by the reference
depth meter (bubbler) for each replica of Test B. Considering the profile of the test site, a 0.79-
inch (20.0-mm) difference in water depth was expected between the two sites under back-flow
conditions, with the ADS 4000 water depths being lower. The ADS 4000 water depth readings
are always lower with  an  average  1.04  in. (26.4 mm) (see Table C.4). Table 4-5 gives an
appreciation of the difference between the two water depth readings (reference/ADS  4000) for
data points collected during stable periods (delimited by the  indicated start and end times) for
each replica and for the three hydraulic conditions of this test. Also shown in this table are the
average reference water depths and the standard deviations of the water depth differences. The
differences in water depth vary between  1.02 to 1.37 in. (25.8 to 34.9 mm) under back-flow
conditions. Since there is very low disturbance due  to velocity under these conditions, these
values should match the difference in slope, as expected.
                                           39

-------
Table 4-5. Water Level Comparison of Data Point Samples for Three Replicas of Test B
Hydraulic
Condition

Flow rate of 75
L/sec
No back flow

Flow rate of 75
L/sec
Back flow of 18 in.

Flow rate of 75
L/sec
Back flow of 36 in.

Start time
End time
Mean reference level (mm)
Mean difference (mm)
Standard deviation of difference
Start time
End time
Mean reference level (mm)
Mean difference (mm)
Standard deviation of difference
Start time
End time
Mean reference level (mm)
Mean difference (mm)
Standard deviation of difference

Bl
9:40
10:15
273.1
34.3
1.6
10:30
12:15
445.4
34.9
3.5
12:45
14:45
908.2
26.9
2.2
Replica
B2
9:40
10:15
259.6
34.9
2.1
10:30
12:15
464.2
28.3
4.0
12:45
14:45
926.5
25.8
2.2

B3
19:40
20:15
267.5
31.7
2.7
20:30
22:15
513.5
29.6
2.3
22:45
00:45
919.9
25.9
2.2
Average
(mm)


266.7
33.6


474.4
30.9


918.2
26.2
Figures H-7 through H-9 present velocities measured by the ADS 4000 and by the reference flow
meter for each replica of Test B. Considering the difference in water depth, velocities measured
by the ADS 4000 were expected to be higher than those measured by the reference meter. This
was not the case for most of replica B2.  Under the 36-inch (914-mm) back flow, reference
velocities  fell  under 0.49 ft/sec (0.15 m/sec) and the ADS 4000 sporadically recorded zero
velocities for replicas B2 and B3,  although the velocity meter had a resolution of 0.0394 ft/sec
(0.012 m/sec).  The ADS 4000 velocity meter signal had more noise than the reference velocity
meter.

¥.5.2.2.2  TestC

After  some flushing of the sewer line, this test was conducted at six consecutive flow rates with
no back flow:

•  1.71 MOD (75 L/sec);
•  up to 17.1 MOD (750 L/sec);
•  down to 8.56 MOD (375 L/sec);
•  up to 29.7 MOD (1,300 L/sec);
•  back down to 8.56 MOD (375 L/sec); and
•  back to 1.71 MOD (75 L/sec).

Figures H-10 through H-12 present flow rates measured by the ADS 4000 and by the reference
flow meter for each replica of Test C. Flow rates were generally overestimated. There were large
oscillations during the 17.1-MGD (750-L/sec) and second 8.56-MGD (375-L/sec) stable periods
                                           40

-------
of replicas C2 and C3 and during the first 8.56-MGD (375 L/sec) stable period of replica C3 due
to large oscillations on the depth measurements. The ADS 4000 recorded velocities of zero
during the transition between 1.71 MGD (75 L/sec)  and 17.1 MGD (750 L/sec) during all three
replicas,  generating calculated flow  rates of zero. Table 4-6 gives an appreciation  of the
difference between the flow rates (reference versus ADS 4000) for samples of data points taken
during stable periods (delimited by the indicated start and end times) for each replica and for the
six hydraulic conditions of this test. Also shown in this table are the mean reference flow rates,
the standard deviations of the flow rate differences and the percentage these differences represent
in comparison with the reference flow rates. Both sets of maximal dry weather flow (1.71 MGD;
75 L/sec) data points are overestimated by roughly 18 percent, which is also the case at 29.7
MGD (1,300 L/sec). At 8.56 and 17.1  MGD (375 and 750 L/sec), the results are largely affected
by the oscillations.

Figures H-13 through H-15 present  water depths  measured  by the ADS 4000 and by the
reference depth meter (bubbler) for each replica of Test C. Due to  the profile of the test site, the
ADS  4000 water depths are expected to always be lower than the reference, which they are,
except at the highest flow rate (29.7 MGD; 1,300 L/sec). Table 4-7 gives an appreciation of the
difference between  the two water depth readings  (reference versus  ADS 4000) for  samples  of
data points taken during stable periods (delimited by the indicated start and end times) for each
replica and for the three hydraulic conditions of this test. Also shown in this table are the average
reference water depths and the standard deviations of the water depth differences.

Figures H-16 through H-18 present velocities measured by the ADS 4000 and by the reference
flow meter for  each replica of Test C. Velocities measured by the  ADS 4000 are higher than
those  measured by the reference meter as expected. Reference velocities are mostly in the range
of 0.98 to 5.91 ft/sec (0.3 to 1.8 m/sec). Excluding data points at 1.71 MGD (75 L/sec), the span
of reference velocities is greater (3.28 to 5.91 ft/sec;  1.0 to 1.8 m/sec) than for the  ADS 4000
(4.59 to 6.56 ft/sec;  1.4 to 2.0 m/sec). The signal of the ADS 4000 velocity meter has  more noise
than the reference meter.
                                           41

-------
Table 4-6. Flow Rate Comparison of Data Point Samples for Three Replicas of Test C

Hydraulic Condition


Flow rate of 75 L/sec





Flow rate of 750 L/sec





Flow rate of 375 L/sec





Flow rate of 1300
L/sec





Flow rate of 375 L/sec





Flow rate of 75 L/sec





Start time
End time
Mean reference flow rate (L/sec)
Mean difference (L/sec)
Standard deviation of difference
Difference/reference (percent)
Start time
End time
Mean reference flow rate (L/sec)
Mean difference (L/sec)
Standard deviation of difference
Difference/reference (percent)
Start time
End time
Mean reference flow rate (L/sec)
Mean difference (L/sec)
Standard deviation of difference
Difference/reference (percent)
Start time
End time
Mean reference flow rate (L/sec)
Mean difference (L/sec)
Standard deviation of difference
Difference/reference (percent)
Start time
End time
Mean reference flow rate (L/sec)
Mean difference (L/sec)
Standard deviation of difference
Difference/reference (percent)
Start time
End time
Mean reference flow rate (L/sec)
Mean difference (L/sec)
Standard deviation of difference
Difference/reference (percent)

Cl
8:40
9:30
69.9
-11.1
4.8
-16
9:45
10:30
700.0
-178.3
17.9
-25
10:45
11:45
395.9
-105.9
23.8
-27
12:00
13:00
1224.6
-224.9
36.4
-18
13:30
14:30
379.9
-102.2
20.9
-27
15:00
14:45
101.7
-15.8
8.1
-16
Replica
C2
15:40
16:30
82.8
-15.2
5.0
-18
16:45
17:30
704.7
-100.3
16.0
-14
17:45
18:45
404.8
-86.6
15.2
-21
19:00
20:00
1213.3
-144.5
32.0
-12
20:30
21:30
401.0
23.2
153.4
6
22:00
22:45
88.9
-15.8
8.6
-18

C3
7:40
8:30
41.4
-6.5
4.9
-16
8:45
9:30
687.6
-256.4
20.9
-37
9:45
10:45
411.2
41.6
172.5
10
11:00
12:00
1213.9
-164.9
40.5
-14
12:30
13:30
389.1
-53.1
24.7
-14
14:00
14:45
91.3
-18.0
7.1
-20
Average
(L/sec)


64.7
-11.0




697.4
-178.4




404.0
-50.3




1217.3
-178.1




390.0
-44.0




94.0
-16.5


                                       42

-------
Table 4-7. Water Level Comparison of Data Point Samples for Three Replicas of Test C
Hydraulic Condition

Flow rate of 75 L/sec



Flow rate of 750 L/sec



Flow rate of 375 L/sec



Flow rate of 1,300
L/sec



Flow rate of 375 L/sec



Flow rate of 75 L/sec



Start time
End time
Mean reference level (mm)
Mean difference (mm)
Standard deviation of difference
Start time
End time
Mean reference level (mm)
Mean difference (mm)
Standard deviation of difference
Start time
End time
Mean reference level (mm)
Mean difference (mm)
Standard deviation of difference
Start time
End time
Mean reference level (mm)
Mean difference (mm)
Standard deviation of difference
Start time
End time
Mean reference level (mm)
Mean difference (mm)
Standard deviation of difference
Start time
End time
Mean reference level (mm)
Mean difference (mm)
Standard deviation of difference

Cl
8:40
9:30
262.2
28.8
1.7
9:45
10:30
582.7
38.8
6.4
10:45
11:45
475.2
46.9
9.6
12:00
13:00
778.6
-30.8
7.8
13:30
14:30
468.0
48.9
9.2
15:00
15:45
291.1
28.3
2.0
Replica
C2
15:40
16:30
279.5
33.2
2.6
16:45
17:30
585.5
70.5
6.0
17:45
18:45
480.4
73.8
3.7
19:00
20:00
761.8
-52.0
6.7
20:30
21:30
481.5
168.5
100.5
22:00
22:45
285.6
36.3
2.7

C3
7:40
8:30
228.5
35.3
2.6
8:45
9:30
591.3
-9.9
4.9
9:45
10:45
485.2
257.4
3.2
11:00
12:00
760.1
-40.9
9.1
12:30
13:30
476.0
61.8
6.4
14:00
14:45
288.9
35.5
2.6
Average
(mm)

256.8
32.5


586.5
33.1


480.3
126.1



766.8
-41.2


475.1
93.1


288.5
33.4

                                       43

-------
4.5.2.2.3  TestD

Test D was completed with varying flow rates and back flow conditions, as described in Section
4.4.2.3. In summary, the six hydraulic conditions used during the test included:

•   1.71 MOD (75 L/sec) with no back flow;
•   8.56 MOD (375 L/sec) with no back flow;
•   8.56 MOD (375 L/sec) with a back flow of 54 in;
•   17.1 MOD (750 L/sec) with a back flow of 54 in;
•   29.7 MOD (1,300 L/sec) with a back flow of 54 in; and
•   1.71 MOD (75 L/sec) with no back flow.

Figures H-19 through H-21 present flow rates measured by the ADS 4000 and by the reference
flow meter for each replica of Test D. Flow rates were overestimated under conditions without
back flow and underestimated under back-flow conditions. There was some instability associated
with the closing  of the gate to create the back  flow.  Table 4-8 gives an appreciation of the
difference between the  flow rates (reference/ADS 4000) for samples of data points taken during
stable periods (delimited by the indicated start and end times) for each replica and for the six
hydraulic conditions of this test. Also shown in this table are the mean reference flow rates, the
standard  deviations of the flow rate differences and the percentage these differences represent in
comparison with the reference flow rates.  Differences in flow rates at maximal dry weather flow
were higher than in Test C (21 percent compared to 17 percent). Under back-flow conditions,
flow rates were underestimated by 20 percent, 15 percent and 11 percent for the three flow rates
(8.56, 17.1, and 29.7 MOD; 375, 750, and 1,300 L/sec).

Figures H-22 through  H-24  present  water  depths  measured  by  the ADS  4000 and  by the
reference depth meter (bubbler) for each replica of Test C. Due to the profile of the test site, the
ADS 4000 water depths were expected to always be  lower than the reference ones, which they
were, except at the highest flow rate  (29.7 MGD; 1,300 L/sec) during replica Dl. There was
some instability before  closing the gate during replica D2. Table 4-9 gives an appreciation of the
difference between the two water depth readings (reference/ADS  4000) for samples of data
points  taken  during stable periods (delimited by the indicated start and end  times) for each
replica and for the six hydraulic conditions of this test. Also shown in this table are the average
reference water depths  and the standard deviations of the water depth differences. Under back-
flow conditions the mean water depth difference was 0.85 in. (21.5 mm), which was close to the
difference in slope (0.79  in.;  20.0 mm), as expected.  Under maximal dry weather  flow, the
average difference was roughly 1.39 in. (35.3  mm), which was  similar to results obtained at the
same flow rate for tests B (1.32 in.; 33.6 mm) and C (1.30 in.; 33.0 mm).

Figures H-25 through H-27 present velocities measured by the ADS 4000 and by the reference
flow meter for each replica of Test D. Velocities measured by the ADS 4000 are higher than
those measured by the reference meter (as expected) for all conditions without back flow. Under
back flow, ADS 4000 velocities are always below the reference ones, which invariably leads to
underestimation of the  flow rates. The signal of the  ADS 4000 velocity meter has more noise
than the reference meter.
                                           44

-------
Table 4-8. Flow Rate Comparison of Data Point Samples for the Three Replicas of Test D
Hydraulic Condition


Flow rate of 75 L/sec





Flow rate of 375 L/sec





Flow rate of 375 L/sec
Back flow of 36 in.




Flow rate of 750 L/sec
Back flow of 36 in.





Flow rate of 1,300 L/sec
Back flow of 36 in




Flow rate of 75 L/sec




Start time
End time
Mean reference flow rate (L/sec)
Mean difference (L/sec)
Standard deviation of difference
Difference/reference (percent)
Start time
End time
Mean reference flow rate (L/sec)
Mean difference (L/sec)
Standard deviation of difference
Difference/reference (percent)
Start time
End time
Mean reference flow rate (L/sec)
Mean difference (L/sec)
Standard deviation of difference
Difference/reference (percent)
Start time
End time
Mean reference flow rate (L/sec)
Mean difference (L/sec)
Standard deviation of difference
Difference/reference (percent)
Start time
End time
Mean reference flow rate (L/sec)
Mean difference (L/sec)
Standard deviation of difference
Difference/reference (percent)
Start time
End time
Mean reference flow rate (L/sec)
Mean difference (L/sec)
Standard deviation of difference
Difference/reference (percent)

Dl
8:45
9:30
63.2
-11.7
4.9
-19
9:45
10:30
379.2
-107.2
26.3
-28
11:00
15:00
400.1
81.3
19.7
20
12:15
13:30
711.8
104.0
22.1
15
14:00
15:00
1285.3
138.1
19.2
11
15:45
16:00
86.9
-18.6
6.0
-21
Replica
D2
8:45
9:30
68.2
-12.0
8.2
-18
9:45
10:30
382.8
-69.1
12.6
-18
11:00
15:00
413.2
96.8
31.5
23
12:15
13:30
710.7
111.9
23.8
2
14:00
15:00
1272.5
139.2
24.4
11
15:45
16:00
93.8
-20.1
5.4
-21

D3
19:45
20:30
66.7
-16.5
5.5
-25
20:45
21:30
387.4
-123.0
18.8
-32
22:00
2:00
414.4
69.3
17.9
17
23:15
00:30
696.7
88.1
15.4
13
01:00
02:00
1262.5
142.3
22.3
11
02:30
03:30
71.8
-17.5
9.2
-24
Mean
(L/sec)


66.0
-13.4




383.1
-99.8




409.2
82.5




706.4
101.3




1273.4
139.9




84.1
-18.8


                                       45

-------
Table 4-9.  Water Depth Data Point Sample Comparison for the Three Replicas of Test D
Hydraulic Condition

Flow rate of 75 L/sec



Flow rate of 375 L/sec



Flow rates of 375, 750,
and 1,300 L/sec;
Back flow of 54 in


Flow rate of 75 L/sec



Start time
End time
Mean reference level (mm)
Mean difference (mm)
Standard deviation of difference
Start time
End time
Mean reference level (mm)
Mean difference (mm)
Standard deviation of difference
Start time
End time
Mean reference level (mm)
Mean difference (mm)
Standard deviation of difference
Start time
End time
Mean reference level (mm)
Mean difference (mm)
Standard deviation of difference

Dl
8:45
9:30
262.7
34.9
1.9
9:45
10:30
471.4
40.2
10.2
11:00
15:00
1458.5
10.6
12.9
15:45
16:30
280.3
33.5
2.0
Replica
D2
8:45
9:30
269.5
36.7
4.7
9:45
10:30
477.1
56.6
12.6
11:00
15:00
1440.2
36.8
10.4
15:45
16:30
298.8
37.3
3.5

D3
19:45
20:30
270.0
32.4
4.7
20:45
21:30
475.5
29.0
5.5
22:00
02:00
1420.0
17.1
8.1
02:45
03:30
264.8
37.0
3.4
Average
(mm)

267.4
34.6


473.4
42.0



1429.6
21.5


281.3
35.9

4.5.2.2.4  TestE

This test was conducted over a minimum of 21 days, including a period of seven days without
data collection. It comprises results obtained from the three replicas of tests B, C, and D. Test E
was conducted over 26 days (June 28 to July 24, 2001), with no data collection from July 18 to
24, 2001. Flow  rates outside periods of active testing do not vary much  and  are generally
between  6.85 to  15.98 MGD (300 to 700 L/sec). Continuous  operation was useful to identify
shifts in the depth and/or velocity measurements.

Figures H-28 through H-54 (Appendix H) present flow rates measured by the ADS 4000 and by
the reference flow  meter for each 24-hour period of Test E. Flow rates measured by the ADS
4000 are generally  larger than the reference rates. Spikes and shifts  (up and down) in the ADS
4000 signal  can be linked  to the water depth measurements. Flow rates of zero are due to the
zero velocities.

Figures H-55 through H-81 present water depths measured  by the ADS  4000 and  by the
reference depth meter (bubbler) for each 24-hour period of Test E. As expected, considering the
configuration of the test site, ADS 4000  water depths are lower than the reference depths. It is
only periodically that the  ADS  4000 depths are  higher. Manual readings at the verified  flow
meter location are presented on the figures as they were taken (before each active test and almost
daily on working days). Table 4-10 presents all the manual readings taken during the tests.
                                          46

-------
Table 4-10.  Daily Manual Level Measurements - ADS 4000
Verification of level measurement
Site and location
Time
(hh:mm)
Meter
reading
(mm)
Manual
reading
(mm)
Gap
( percent)
Date : 2001/06/28
Reading #1
Reading #2
Reading #3
Reading #4
Reading #5
Extra reading
6:58:10
6:59:10
7:00:10
7:01:10
7:02:10

Average reading :
Comments

226
227
228
229
228

228
240
240
244
242
242

242
-5.9
-5.3
-6.7
-5.2
-6.0

-5.8
Everything seems ok
Date : 2001/06/29
Reading #1
Reading #2
Reading #3
Reading #4
Reading #5
Extra reading
7:23:00
7:24:00
7:25:00
7:26:00
7:27:00

Average reading :
Comments

300
292
284
280


289
315
308
300
295
290

302
-4.8
-5.2
-5.4
-5.1


-4.2
-
Date : 2001/07/03
Reading #1
Reading #2
Reading #3
Reading #4
Reading #5
Extra reading
Average reading :
Comments

7:36:00
7:37:00
7:38:00
7:39:00
7:40:00


248
246
250
248
221

243
260
260
257
260
262

260
-4.7
-5.2
-2.6
-4.7
-15.6

-6.6
Everything seems ok
Date : 2001/07/05
Reading #1
Reading #2
Reading #3
Reading #4
Reading #5
Extra reading
7:39:00
7:41:00
7:42:00
7:43:00
7:44:00
7:40:00
Average reading :
282
265
263
261
241
276
265
298
283
280
277
272
293
284
-5.3
-6.3
-6.2
-5.9
-11.5
-5.8
-6.8
Comments Everything seems ok. We took photo.
Site and location
Time
(hh:mm)
Meter
reading
(mm)
Manual
reading
(mm)
Gap
( percent)
Date : 2001/07/09
Reading #1
Reading #2
Reading #3
Reading #4
Reading #5
Extra reading
6:57:00
6:58:00
6:59:00
7:00:00
7:01:00

Average reading :
Comments

196
198
200
198
188

196
207
207
209
209
209

208
-5.2
-4.2
-4.5
-5.1
-10.1

-5.8
Condensation on Ultrasonic probe.
Everything seems ok
Date : 2001/07/10
Reading #1
Reading #2
Reading #3
Reading #4
Reading #5
Extra reading
18:29:00
18:30:00
18:31:00
18:32:00
18:33:00

Average reading :
Comments

443

404
385


411
480
460
440
420
400

440
-7.7

-8.2
-8.3


-6.6
Little condensation on Ultrasonic probe
(less than usual). Everything seems ok
Date : 200-07/13
Reading #1
Reading #2
Reading #3
Reading #4
Reading #5
Extra reading
Average reading :
Comments

7:39:00
7:40:00
7:41 :00
7:42:00
7:43:00


344
354
362
354
374

358
340
352
361
370
378

360
1.2
0.6
0.3
-4.2
-1.2

-0.7
Condensation on Ultrasonic probe.
Everything seems ok
Date : 2001/07/15
Reading #1
Reading #2
Reading #3
Reading #4
Reading #5
Extra reading
5:47:00
5:48:00
5:49:00
5:50:00
5:51:00

Average reading :
231
228
223
198
196

215
258
253
248
242
238

248
-10.5
-10.0
-10.0
-18.0
-17.6

-13.1
Comments A few drops under the ultrasonic probes
                                      47

-------
Table 4-10 (cont'd)
Verification of level measurement
Site and location
Time
(hh:mm)
Meter
reading
(mm)
Manual
reading
(mm)
Gap
( percent)
Date : 200107/16
Reading #1
Reading #2
Reading #3
Reading #4
Reading #5
Extra reading
7:02:00
7:03:00
7:04:00
7:05:00
7:06:00

Average reading :
Comments

307
311
315
308
316

311
312
318
320
326
325

320
-1.5
-2.2
-1.7
-5.6
-2.9

-2.8
Condensation on Ultrasonic probe. Debris
accumulated on the bender. Everything
seems ok
Date : 2001/07/18
Reading #1
Reading #2
Reading #3
Reading #4
Reading #5
Extra reading
20:01 :00
20:02:00
20:03:00
20:04:00
20:05:00

Average reading :
Comments


431

418
372

407
470
468
460
450
435

457

-8.0

-7.1
-14.4

-10.9
-
Date : 2001/07/20
Reading #1
Reading #2
Reading #3
Reading #4
Reading #5
Extra reading
Average reading :
Comments

7:52:00
7:53:00
7:54:00
7:55:00
7:56:00


533
548
563

592

559
518
532
553
567
580

550
2.8
2.9
1.9

2.1

1.6
2 last photo. Condensation on the
ultrasonic probe
Date : 2001/07/21
Reading #1
Reading #2
Reading #3
Reading #4
Reading #5
Extra reading
7:46:00
7:47:00
7:48:00
7:49:00
7:50:00

Average reading :
248
243
235

202

232
275
272
263
260
255

265
-10.0
-10.6
-10.5

-20.6

-12.4
Comments Two or three drops on the ultrasonic probe
Site and location
Time
(hh:mm)
Meter
reading
(mm)
Manual
reading
(mm)
Gap
( percent)
Date : 2001/07/22
Reading #1
Reading #2
Reading #3
Reading #4
Reading #5
Extra reading
6:39:00
6:40:00
6:41 :00
6:42:00
6:43:00

Average reading :
Comments

193
192
190
177
175

185
212
210
207
205
207

208
-9.1
-8.8
-8.3
-13.6
-15.6

-11.0
No accumulation on the bender.
Date : 2001/07/23
Reading #1
Reading #2
Reading #3
Reading #4
Reading #5
Extra reading
6:41 :00
6:42:00
6:43:00
6:44:00
6:45:00

Average reading :
Comments

215
213
218
194
196

207
235
232
232
230
230

232
-8.7
-8.4
-6.0
-15.6
-14.6

-10.6
Condensation on the ultrasonic probe.
Debris on the bender.
Date : 2001/07/24
Reading #1
Reading #2
Reading #3
Reading #4
Reading #5
Extra reading
Average reading :
Comments

7:22:00
7:23:00
7:24:00
7:25:00
7:26:00


235
233
232

211

228
255
253
250
250
250

252
-7.9
-7.9
-7.0

-15.5

-9.4
-
Date :
Reading #1
Reading #2
Reading #3
Reading #4
Reading #5
Extra reading






Average reading :





















Comments
Periodic air purges to flush debris at the end of the bubbler injection tube can be observed in the
data. Downward spikes (amplitude of roughly five to ten in.) around 11:00, 17:00, and 21:00 on
June 29 represent air purges, which have no impact on the flow rate. Small shifts (up and down)
can be observed almost daily in the ADS 4000 signal. Spikes become very frequent from July 10
                                          48

-------
through 15, after which they become more sporadic until July 19 to the end of the testing period,
when they become  frequent.  According to ADS staff, this is due  to condensation under the
ultrasonic probe. It should be noted that sewers typically have a humidity of 100 percent, and
condensation is likely to occur. There is generally more noise in the signal of the ADS  4000
depth meter than for the reference meter.

Figures H-82 through H-108 present velocities measured by the ADS 4000 and by the reference
flow meter for each 24-hour period of Test E. There are occasional zero readings of the  ADS
4000 velocity meter, some of which occur when the reference velocity increases rapidly. There is
generally more noise in the signal of the ADS 4000 velocity meter than for the reference meter.

Figures H-109 through H-l 17  present comparisons between water depths output by the reference
depth meter (bubbler) and three water depth  readings output by the ADS  4000:  the  pressure
depth, mean ultrasonic depth,  and final depth. The final depth is either the pressure depth or the
average ultrasonic depth. A built-in proprietary algorithm selects the reading to be used, so it is
always superimposed on either curve. The decision is not always a good one: for example, in
Figures 4-113 and 4-114, the pressure depth is more stable than the mean ultrasonic depth but the
final depth is mostly the average ultrasonic reading.

4.5.2.3 General Verification Tests

Test E was composed of two parts: the performance evaluation and the  general evaluation.
Appendix C presents a day-to-day description of the tests.

4.5.2.3.1   Installation,  Configuration and Calibration

ADS  staff, composed of a project manager, a  field manager, a field technician, and an analyst,
completed field installation of the ADS 3600.  The TO helped the ADS field manager with the
underground installation and supervised  all  activities of ADS  during  configuration  and
calibration. The  ADS  team performed almost all the fieldwork. A  complete  crew (except the
project manager) was on-site from June 18  to July 7, 2001; the field analyst who could perform
any sewer intervention stayed on-site July 7-17, 2001.

The installation activities of ADS were performed as follows:

•  The analyst beside the manhole linked a laptop computer to the ADS 4000 transmitter.
•  The project manager beside the manhole managed the safety harness.
•  One ADS technician entered the pipe, supervised  by TO personnel, and installed the
   equipment.
•  A second technician entered the manhole to assist the technician already in the pipe.

The time  spent by the TO staff was not accounted in the  observed number of person-hours
required to install the flow meter shown in Table 4-11. Four individuals could have performed
the same job in the same amount of time during a normal installation.
                                          49

-------
Table 4-11. Time Estimate Required to Install, Operate and Service Flow Meter
        Activities
                  Estimate by     Observed by
 Work class*          ADS         tester team
	(person-hours)   (person-hours)
                                                     Frequency
                                                    (if applicable)
 Installation, Configuration and Calibration
Initialization of the velocity
measurement
Initialization of the level
measurement (Zeroing of
the level probe)
Start-up and trouble
shooting
Replacement of a level
probe

Field Service
Personnel
Field Service
Personnel
Field Service
Personnel
Field Service
Personnel
Total*
1 3
1 2
1 48 hours
1 3
7 65
1
1
1
Every 2-3 yrs
N/A
 Operation, Maintenance and Service
Replacement of desiccant
(SENSOR)
Replacement of on-board
memory battery
Field Service .
Personnel
Field Service N/.
Personnel
Twice a year
Once a year
 Replacement of the input
 board in the transmitter

 Replacement of the output
 board in the transmitter

 Replacement of the main
 board in the transmitter

 Probes maintenance
 recommended by the
 manufacturer
Completed in the
office by technician;        N/A
not done in the field.
Completed in the
office by technician;        N/A
not done in the field.
Completed in the
office by technician;        N/A
not done in the field.

Field Service
Personnel
                                                     Several
                                                    Years apart

                                                     Several
                                                    Years apart

                                                     Several
                                                    Years apart
* The work class: project manager, field manager, field technician, data analyst

The installation, configuration, and calibration took approximately 16 person-hours (four people
during half a day). The only  problem  met during this installation involved the bender on the
stainless steel band. The first two benders moved away from the pipe wall when it  was screwed,
instead of pushing toward it, so the band was secured to the pipe wall at the crown of the pipe.
Since the bender was not close to the pipe wall, debris accumulated (see Illustration 4-9). The
problem was not major, but a bender closer to the pipe wall (pushing on it), would be required in
a permanent installation.
                                              50

-------
Figure 4-26. Debris accumulation on ADS 4000's bender.

The installation took approximately eight person-hours. Configuration and calibration required
another eight person-hours and involved four steps (Appendix F):

•  Configuration and calibration of the ultrasonic depth meter;
•  Configuration and calibration of the pressure depth meter;
•  Configuration and calibration of the Doppler velocity probe;
•  Configuration and calibration of flow computation options, like:
          o  Depth measurement used;
          o  Calibration or the pressure depth probe with the ultrasonic depth probe; and
          o  Range of velocity expected (fast or slow);

The calibration of the reference flow meter required  one specialized tool. A Marsh-McBirney
Model 2000 portable velocity meter was used to measure the peak velocity in the pipe section.
This value was required to calibrate the gain of the Doppler velocity probe.

After calibration with a maximal dry weather flow of  1.71 MGD (75 L/sec), the official test
began June  20 with the test first replica of Test C.  It was discovered after  this test that the
ultrasonic depth sensor was not measuring properly.  It was decided to replace the sensor and
restart the verification. Since tracer  dilution was performed to verify the performance of the
reference meters, the test was considered valid for the reference meters and renamed CO instead
ofCl.

The ultrasonic depth probe was replaced and verification of the transmitter and probes was done
to understand the behavior of the flow meter and try to find a better tuning. This was completed
                                            51

-------
while waiting for the arrival of a spare component required for the other flow meter verified at
the same time. The verification officially began on June 28, 2001.

4.5.2.3.2  Operation

The ADS 4000 operation and maintenance manual includes a chapter about unit maintenance.
ADS recommends a local check at installation, during a monitor site visit and every four months
or during battery replacement. In summary, the following elements have to be verified:

•  Status of the casing;
•  Status of the battery pack;
•  Verification of the desiccant's color (for pressure probe);
•  Verification of the stainless steel ring installation;
•  Verification of grease or scum on the face of probes;
•  Wiping the ultrasonic depth sensor with a clean moist cloth;
•  Verification of the ultrasonic probe with a carpenter's level;
•  Verification of debris accumulation on the ring or probes; and
•  Verification of debris accumulation on cable.

The flow meter was inspected 16 times from June 28 to July 24, 2001 and no problem of debris
accumulation was observed, except with the bender  as shown in Figure 4-26.   Debris on the
bender was  not removed because it had no impact on the measurements; it would have been
removed every four months in a normal installation. Since the probes are thin and have a good
hydraulic profile,  they did not have a tendency to accumulate debris.

Except for a thin  film of grease, the probes were clean at the end  of the test. During verification
in the sewer, drops of condensation were  observed  on the underside  of the ultrasonic depth
probe. Since there is always some condensation in sewers, the ultrasonic probe was not be wiped
during verification.

4.5.2.3.3  Time for Data Re trieval

There was one  period following completion of tests B,  C,  and D,  during which data were not
retrieved for seven consecutive days.  The objective was to verify the time required to retrieve
data during a normal data retrieval period.

For the  period of seven days, it took approximately 1.3 minutes  per day to log  the one-minute
data. This delay was reasonable for a seven-day period, but it may become restrictive for longer
data collection intervals.

For retrieval of two to four days of data, the transmitter took approximately two to three minutes
per day  of one-minute  data  logging. This is probably due to the time required to establish
communication with the transmitter (see Appendix C, Table C-3 for detailed results).
                                           52

-------
4.5.2.3.4  Dismantling

The probes and the transmitter were dismantled, cleaned and photographed on July 25. None of
the components had been damaged or altered during the tests.

4.5.2.3.5  General Characteristics of the ADS 4000

The probes have a good hydraulic profile.  The stainless steel band is well-designed. It is thin
enough to fit the exact shape of the pipe and thick enough to have good strength. However, as
stated, the bender must be pushed onto the pipe wall while securing its position.

The aluminum transmitter was well-designed for sewers and is easy to hang in the  manhole.  It
had a handle on the top to facilitate manipulation and protect connectors.  It is waterproof and
protected against infiltration by pressurizing the casing at approximately 12 psi (82.7 kPa) of air
pressure. There is no connector on the casing for external powering. The internal battery is not
rechargeable, and it is expensive. According to the vendor, the battery can last up  to one year
while collecting 15-minute data, but such an infrequent logging period  is not useful  in many
combined-sewer overflow (CSO) applications, such as modeling and diagnostic, which generally
require a five-minute period. For data validation purposes, it may become useful to perform
monitoring at a higher frequency to eliminate spikes caused by  temporary local disturbances.
When a long data-logging period is used, short local disturbances like standing waves, surge, or
backwash can  result in  two  logging  intervals without  valid  measurements. These  local
disturbances  are more frequent during high flows of a  storm event compared to dry weather
flows.

The ultrasonic depth probe has a specially  designed sliding support to facilitate its installation
and adjustment. Its low dead band (0.5 inch; 13 mm), low depth profile (0.875 inch; 22 mm), and
its four  probes incorporated  in  the same casing are  characteristics that made  this  probe
particularly well-adapted to sewer pipes. This probe is, however, more affected by local waves
than the pressure probe, which performs an integration of the various depths over the probe.

The configuration of the transmitter has various options.  For example,  it is possible to  compute
the flow using the  depth from the ultrasonic depth probe, the pressure probe, or the better of the
two results according to an algorithm programmed into the flow meter.  The ADS 4000 used
during testing was configured using this last configuration.

ADS expressed concern about using a cable longer than that typically  supplied with the Model
4000  (25  ft  [7.6  m]) because of possible  noise  in the longer cable. The cable used during
verification  was the standard 25  ft length.  This should  be a consideration for permanent
installations.  It  might be more convenient to install the transmitter  in a control cabinet close to
the street or in a service building.  The  length  of the  cable's probes, the data acquisition
frequency, the powering options, and the packaging option are not well adapted to CSO real-time
control applications.

The  transmitter is available with  a telephone modem  option.  With  the  modem  and related
software, the logging of many transmitters can be performed using a standard telephone line,
                                           53

-------
instead of going on-site. The ADS 3600 also provides flexibility to connect a rain gauge and an
automatic sampler for sampling proportional to the recorded transmitter flow.

4.6    Reference Meters

4.6.1   Scatter Plots

As with the presentation for the  ADS 4000, three types of scatter plots are presented for the
reference meter data: a system behavior (velocity versus depth) plot, flow rate comparison plots,
and water depth comparison plots. These plots, showing the  data collected for all of the reference
meter runs, are presented in Appendix H.

The system behavior scatter plot was presented in Figure 4-18, showing the hydraulic conditions
at the reference flow meter during Test C  (no back flow).  There are 1,353  data points on this
figure, segregated in three groups: stable (984 data points),  ebbing (261) and rising (108) water
profiles. There are fewer data points for the rising water profiles because the transition is more
rapid  under this condition. Apart from the  stable water profiles, there  is no single relationship
between velocity and water depth.  The best-fit curve polynomial equation drawn from stable
conditions  data points is  indicated on the figure, along with  its  coefficient of determination
(R2 =  0.9938).

The system behavior scatter plot was presented in Figure 4-18, showing the hydraulic conditions
at the reference flow meter testing site during Test C (no back flow). There are 1,353 data points
on this figure, segregated in three groups: stable (984 data points), ebbing (261) and rising (108)
water profiles. There are fewer data points  for the rising water profiles because the transition is
more  rapid under  this condition. Apart from the  stable  water  profiles, there is no  single
relationship between velocity  and water depth. The  best-fit curve polynomial  equation drawn
from  stable conditions data points  is indicated  on the figure, along with its coefficient of
determination (R2 = 0.9938).

Figures H-121 to H-124 present water depth comparison plots between the reference depth meter
(bubbler) and manual readings for tests B, C, and D respectively, and these three tests combined,
along with the perfect fit line.

4.6.2   Deviation Distribution Plots

Figures H-125  to H-130  present the flow rate deviation  distributions (reference  flow meter
compared to the flow rate calculated from the tracer dilutions) for all data of replicas CO and C3.
The deviation is presented in two-percent increments. Indicated on the figures are the:

•  Margin of deviation (±X percent) that contains 95 percent of all measurements (i.e. standard
   ±2o);
•  Percentage of measurements within a margin of deviation of ±8.7 percent (wet-weather racer
   dilution error);
•  Mean deviation of all measurements; and
•  Median deviation (provides an indication of the spread).
                                            54

-------
Figure H-125 presents the flow error distribution for replicas CO and C3. It is a bell-shaped curve
centered on -0.9 percent. The largest negative error is -45.8 percent and the largest positive error
is 20.5 percent, both of which occurred during dry-weather conditions (during replica C3 at 8:30
and 7:45, respectively).  Replica C3 was performed  on  July  15, starting in the  early morning
when the flow rate was roughly 0.80 to 1.03  MGD (35 to 45 L/sec).  At such low flow rates, the
relative  error tends to be  larger since the flow  meter resolution is 0.023 MGD  (1 L/sec).
Excluding these low flows, the largest negative error is -9.5 percent and the largest positive error
is 19.7 percent. Both  data points are during transition periods (during C3 at 10:55 and,  CO at
14:30  respectively). Of the measurements,  84.1  percent are within  the margin of error of 8.7
percent, which is the tracer dilution error for wet-weather conditions,  as indicated in the Protocol
for Flow meters for Wet Weather Flow Applications in Small- and Medium-Sized Sewers (Draft
4.0, September  2000)  (protocol).  The  protocol  referred  to  the  fact  that  95 percent  of
measurements should be within this margin of error. Figures H-126 and H-127 present the same
data for CO  and C3 separately (used for the QAPP). Figures H-128 to H-130 present the same set
of three figures but for stable condition data points only.

Figures H-131 to 4-133 present the water depth error distributions (reference depth  meter
compared to manual readings) for tests B, C, and D, respectively, while Figure H-134 presents
all data from tests B, C, and D combined. These error distributions are presented in inches (mm)
since the protocol referred to an acceptable error of two percent of full scale (full scale is 62.5 in.
[1,586 mm]), which represents 1.25 in. (31.7 mm). The  error is presented in 0.25 in. (6.4 mm)
increments.  Indicated on the figures are the:

•   Margin of error (±Xpercent) that contains 95 percent of all measurements (i.e.  standard ±2o);
•   Percentage of measurements within a margin of error of ±2 percent full-scale;
•   Mean error of all measurements; and
•   Median  error (provides an indication of the spread).

Figure  H-134 presents the water depth  error distribution for  all data of tests B,  C,  and D
combined. It is a bell-shaped curve with a mean error of 0.13 in. The largest negative error is
-3.74 in. recorded during replica D3 in the transition period of between 29.7 MGD (1,300 L/sec)
with a  54-inch back  flow to 1.71  MGD (75 L/sec)  with no back flow (at 2:20). The  largest
positive error is +4.00  in. recorded  during replica Cl in the transition period of between
17.1 MGD (375 L/sec) to 29.7 MGD (1300 L/sec) (at 12:00). The large positive and negative
errors  are due to waves at high flows (17.1 and 29.7 MGD [750 and 1,300 L/sec], respectively)
since the manual readings are more affected by waves than is the bubbler.

Figure  H-135  presents  the water  depth error distribution  (reference  bubbler depth  meter
compared to the reference  ultrasonic depth meter) for  all data  from tests B, C, and D. It is
roughly a centered bell-shaped  curve. There  are  12 data points outside ±5.0 in.  The  larger
positive errors are during transition periods from 1.71 to 17.1  MGD  (75 to 750 L/sec) and from
8.56 to 29.7 MGD (375 to 1,300 L/sec) while the larger negative errors are during the transition
period between 29.7 to 8.56 MGD (1,300 to 375 L/sec). The overestimation at 29.7 MGD (1,300
L/sec) (all replicas) and underestimation at 17.1 MGD (750 L/sec) (replicas CO, Cl and C3) are
                                            55

-------
due to the ultrasonic depth meter being more affected than the bubbler by waves caused by the
high flows.

Figure H-136 presents the velocity error distribution (velocity calculated from the reference flow
rate and water depth compared to velocity calculated from tracer dilutions) for replicas CO and
C3. The largest negative error is -45.8 percent and the largest positive error is 20.5 percent, both
from  replica C3 (at 8:30 and 7:45,  respectively).  Errors are largest  at low flows  and during
transition periods.  The error made on the velocity  is the summation of the error made on flow
rate and the one on water depth since it is the result of a calculation (velocity equals flow rate
divided by area).

4.6.3   Tests B, C, and D

Only  figures that  supply  new information  are  presented  in this section since the bulk of the
reference data is presented with the ADS 3600 flow meter data in Figures H-l to H-108.

Figures H-l37  through H-l39 present flow rates from the reference  flow meter  and from the
flow under the  downstream gate equation for the three replicas  of Test B, while Figures H-l42
through H-l44 present the same information for the three replicas of Test D. Figures H-l40 and
H-141 present flow rates  from the reference flow meter, from tracer dilutions, and from the flow
under  the upstream gate  with  the Craig Collector flow rate added  for  replicas CO and C3,
respectively. On both figures, the largest differences between the reference flow meter and tracer
dilutions are in the transition zones  indicated by the  square markers  (rising and  ebbing water
profiles). As mentioned, replica C3 (Figure H-141)  was performed on a Sunday, with a very low
early-morning flow. From 11:30,  there  was not enough water in the Versant-Sud tunnel to
maintain a stable flow rate of 29.7 MGD (1,300 L/sec) for the prescribed time.

Figures H-l45  through H-l54 present water depths from the reference depth  meter (bubbler),
manual readings and reference ultrasonic depth meter for  the three  replicas  of Test B, four
replicas of  Test C,  and  three  replicas of  Test D, respectively. Water  depths from all  three
instruments  for the three replicas of both tests B  and D (Figures H-l45 through H-l47, and
H-152 through H-154) are practically identical. Water depths  of the four replicas  of Test C
(Figures H-l48 through H-l 51) vary more, principally due to the presence of waves during these
tests.  Replicas CO, Cl, and C3 present the  same pattern: manual readings are higher than the
bubbler and the ultrasonic depths are lower during the 17.1 MGD (750 L/sec) stable period while
the opposite is  found in the 29.7 MGD (1,300  L/sec) stable period. Both manual readings and
ultrasonic depths are affected by waves whereas the bubbler buffers the information over a larger
volume of water. It should be noted that the manual readings are taken very close to the bubbler
site.

Figures H-l55 and H-l56 present velocities calculated from the reference flow meter data and
tracer  dilution data for replicas CO and C3, respectively. Since these velocities are calculated
with the same water depths but different flow rates and because the errors on  the water depths
and flow rates are summed, there is simply more noise in the signal.
                                            56

-------
                                      Chapter 5
                        Quality Assurance/Quality Control
The  assessment  of the accuracy,  precision, and completeness of the flow meters undergoing
testing requires the existence of reference data for comparison and validation. At the beginning
of the  project, it was decided that the reference flow rates would be produced  using a 4-path
Accusonic flow meter and reference water depths from a bubbler. The choice of a flow meter to
assess the performance of other flow meters was  not obvious. However, considering that a 4-path
Accusonic flow  meter has a theoretical uncertainty of ±4  percent for  flows  in open  channel
systems (this is smaller than the uncertainty associated with tracer dilution; refer to Appendix
D-3) and that flow rates can be measured continuously in real time at a low cost,  the choice of a
4-path Accusonic flow meter to generate the reference flow rates was justified.

Two reference depth meters were used: the bubbler and the ultrasonic. The bubbler was preferred
as the primary reference for the following reasons.

•  It exhibits no significant bias when properly installed and operated.
•  It provides measurements for a large flow range.
•  It is more reliable.
•  It is the depth measurement linked with the reference flow meter.

The first step toward the use of the Accusonic flow meter as the reference flow meter was the
validation of its measured  flow rates. Lithium tracer dilutions were used to verify that the flow
rates measured by the reference flow meter were  accurate.

Validation of the data provided by the reference  flow meter and the reference water depth meter
was performed as rigorously as possible. Tight review of all the procedures used to collect and
handle data was performed, and is described in the following sections. Such reviews are essential
to assure that  data quality conforms to protocol specifications. After the validation of the data
collected in the field, described in Section 5.1, the Quality Assurance Project Plan (QAPP) was
designed to verify the representativeness of the lithium concentrations measured by the standard
laboratory. These concentrations were used to validate the flow rates measured by the reference
flow meter. The final objective of the QAPP consisted of verifying that the measurements made
by the reference  devices, a reference flow meter and a bubbler depth meter, were representative
of the true values considering the uncertainty margins associated with each measurement device.
These validations are described in Sections 5.2 through 5.4.

5.1    Audits

Audits of the field testing and the reporting used  in the testing were completed several times over
the course of the  verification testing.
                                           57

-------
5.1.1   Field Audits

During the course of the project, the QAPP representative visited the field several times. These
visits were conducted without prior notice and included the injection site, the sampling site, and
the control gate. The objectives of these visits were to ensure the conformity of the sites with
respect to protocol specifications; and quality measurement procedures.

At each visit, the QAPP representative verified the measurement procedures followed by the
field crew. In particular, the synchronization and frequency of measurements, the configuration
of the measurement sites, the handling of the sampling bottles, and the injection procedure of the
concentrated lithium were verified.

Observations made during field visits  permitted observation of the  methods used by the field
crewmembers.   It was observed  that the steps required  assuring good  quality data  were
meticulously followed according to the VTP. All members knew exactly the tasks at hand, when
to perform them and how to execute them.  The following items were specifically observed:

•  Crewmembers had a thorough knowledge of the hydraulics of the system. They were able to
   assess flow  rates under the gates and had a good understanding of the flow delays between
   the injection and the sampling sites. Therefore, they could accurately control the flows in the
   sewer network and evaluate the time required to reach a steady-state flow.

•  The existing communication procedure between the crewmembers was efficient, no matter
   where they were located. This  guaranteed all members had real-time access to the available
   information. All collected data were quickly saved at a centralized location  notwithstanding
   the fact that some were collected manually.

•  The data collected were well synchronized. This validated the right to use paired sample data
   to assess the accuracy of time-varying measurements provided by different measurement
   devices or different laboratories.

•  The measurement devices were properly configured and installed to provide measurements
   that were as accurate as possible. For example, manual water depth measurements were made
   using a homemade device designed to reduce variations that could be introduced by the flow
   turbulence observed at the sampling site during the wet weather flows of Test D.

•  The procedure followed to label the sampling bottles helped ensure that no error could be
   made on location or on the time when samples were collected.

Field report check-ups conducted  in conjunction with field visits verified that the procedures
followed by the field crewmembers when collecting and handling data met the  requirements of
the QAPP. The  field crewmembers were very careful and attentive to details to assure high
quality data.

The field reports were completed with accuracy and included pertinent  comments. The non-
conformities noted were minor, mainly dealing with non-recorded departure times and the names
                                           58

-------
of technicians. The collected data was appropriately handled, and the data in the spreadsheets
used for validation and flow computations was found representative of the measurements made
in the field.

The only negative observation concerned the handling of the sample bottles at the tracer injection
site.  The same individual collected the blank and the  concentrated samples. This person took
precautions to isolate the blank samples from the concentrated ones (e.g., changing gloves when
manipulating new bottles, storing blank bottles in watertight bags).  Nonetheless, the risk of
contaminating the blank samples was not completely eliminated. It would have been safer to
have a sampling site for the blanks that was geographically distant from the injection site.

5.1.2 Report Audit

The  data collected in the field were transcribed according to the VTP  in different field reports.
One  of the objectives of the audit was to verify if all required reports had been filled out and that
no data was missing. To verify the completeness  of the information gathered in  the field,
verification sheets were specially developed for the QAPP. A verification sheet was filled out for
each report produced during the field tests. These verification sheets included:

Preliminary Tasks

•  Instruments Calibration Report Verification
•  Laboratory Report Verification

Test  Tasks

•  General Report Verification
•  Test Report Verification: Daily General Verification
•  Test Report Verification: Monthly General Verification
•  Test Report Verification: Control Table of Test
•  Injection Site Report Verification
•  Sampling Site Report Verification
•  Sampling Site Measurement Verification
•  Tracing Sheet for Injection Site Verification
•  Tracing Sheet for Sampling Site Verification
•  Laboratory Results Verification
•  Flow Meter Configuration and Calibration Verification
•  Flow Meter Operation and Maintenance Verification
•  Flow Meter Software Checklist Verification: Ease of Use
•  Flow Meter Software Checklist Verification: Functionality and Flexibility
•  Flow Meter Installation Report Verification
•  Flow Meter Dismantling Report Verification

These verification sheets were designed to quickly assess whether the  information contained in
the field or laboratory reports was complete, as specified in the VTP.  These sheets were designed
                                           59

-------
to follow the data collection and handling sequences in chronological order and were divided
into three sections:

Section 1 - General Information

The first section contained information such as the name of the test, the names of the technicians
present during the test, the names of visitors, and the time of their visits.

Section 2 - Field Data

A part of this section was designed to check on the lithium tracer sampling frequencies,  (i.e.,
blanks, concentrated and diluted samples for both the standard and control laboratories) and on
the lithium concentration measured by the laboratories. The chronological sequence of the sheets
enabled one to track all the sampling  bottles collected during the tracer dilution tests (tests CO
and C3) from the collection sites to the laboratories. The injection and sampling  sites verification
sheets  enabled one to verify if all lithium samples had been properly  collected and labeled. The
tracking  sheets  for injection and  sampling sites  confirmed  that  the standard and control
laboratories received all  samples. Finally, the laboratory results verification sheets guaranteed
that all samples sent to the laboratories were analyzed and the  results sent back to the TO for
validation and flow computation.

This section was used to quickly verify that flow rates, velocities and water depths had been
measured using the frequencies and the accuracy specified in the VTP. Particular attention was
paid to the synchronization  of the data saved in the data report spreadsheet since the  data
originated from several  sources.  Data  on the sampling site measurement verification sheets
included:

   •   Bubbler measurements (data transferred from the PLC to an industrial PC using software
       developed by the TO and then to a spreadsheet);
   •   Ultrasonic  depth measurements  (data transferred  from  the ADS  data  logger to  a
       spreadsheet);
   •   Manual measurements (data directly recorded in a spreadsheet);
   •   Reference flow rates  and velocity measurements (data transferred  from the PLC to an
       industrial PC using a software developed by the TO and then to a spreadsheet); and
   •   ADS 3600 flow rate and water depth measurements (data transferred from the  ADS data
       logger to a spreadsheet).

Time series related to flow computations under the upstream and downstream gates, including
water depths and gate positions, were verified on the test report verification sheet, Control Table
of Test.

Due to the large amount of collected data and its various sources, only a fraction of the  data
reported  in the spreadsheet used  for flow, water depth, and  velocity validation were audited.
These random audits indicated proper handling of the field data.  For all testing, the data saved in
the spreadsheets was identical to the  raw measurements and was associated to the exact  time
                                           60

-------
collected. Moreover, the set of data saved in the spreadsheet was complete and respected the
sampling frequency reported in the VTP.

The field data section allowed assessment of the completeness of the information related to the
measurement devices. The information contained in the flow meter configuration and calibration
verification sheet, the flow meter operation and maintenance verification sheet, the flow meter
software checklist verification sheet, the flow meter installation verification sheet, and the flow
meter dismantling verification sheet, verified that all reports related to flow measurement devices
had been properly completed.

Section 3 - Non-Conformity Follow-Up

This section included questions required to complete the general information and the  field data
sections. Answers provided by the project manager or the crewmembers  are also reported. This
permitted good communication between  the QAPP representative and the project  manager in
order to complete the field reports or to explain observations that were not in conformity with the
VTP.

It was during the question-and-answer dialogue that the explanation of the existence  of the CO
test was reported  (this  test was  not planned in the original VTP).  For this particular non-
conformity,  during the  course of the  original  Test Cl, the ADS 4000 flow meter worked
properly, but the ADS 3600 flow meter did not.  Since this test was done in conjunction with a
tracer dilution test, it was decided to keep the dilution results, but to rename the Test CO for the
validation of the reference flow meter. Thereafter, Test  Cl was repeated  without dilution to get
proper measurements with the ADS 3600 flow meter.

5.2    Validation of the Lithium Tracer Methods

Lithium dilution tests were used to validate the reference flow meter. Before using the tracer for
the meter validation, the lithium concentrations provided by the standard laboratory needed to be
validated. Validation of  the  results provided  by the standard  laboratory  (Laboratoire  de
I'environnement LCQ Inc.) consisted of comparing the  standard  lithium concentrations with
those measured by another laboratory,  referred  to as the  control laboratory (Laboratoire de la
qualite du milieu, Centre d'expertise en analyse environnementale du Quebec). This process was
conducted  during the  validation  of  the  reference flow meter.  Results  showed  that the
concentrations measured  by  the standard laboratory  were statistically  equivalent  to those
measured by the control laboratory.  Since the probability that the two laboratories measured the
wrong concentrations is very  low, it was concluded that the standard laboratory was  providing
concentrations that had the level of accuracy needed to validate the reference flow meter.

To analyze  the  performance  of  the verified flow meters,  it was necessary to  confirm the
validation of the standard laboratory analyses previously  conducted during the course of the
validation of the reference flow meter. Two new lithium dilution tests were conducted to verify,
using the same analysis  tools, whether the standard laboratory was still providing concentration
results representative of the real concentrations in the samples.
                                           61

-------
5.2.1  Sample and Data Handling

Before comparing the concentration results provided by the two laboratories, proper handling of
the samples had to  be guaranteed. Therefore, it was verified that all  samples were properly
collected and identified, they had been sent to the laboratories for analysis, and were returned
with the concentration results.

The concentrations  measured  by the two laboratories were recorded in a  spreadsheet for
statistical  analysis. In  order to guarantee that the results used for the statistical analysis were
valid, all  the lithium concentrations recorded  in the spreadsheet were crosschecked with the
concentration results sent by the laboratories. Moreover,  all the equations programmed into the
spreadsheets (analysis results from standard and control laboratories) were validated according to
the theory presented in Appendices D-l and D-2.

5.2.2  Methods Used to Validate Lithium Concentrations

The lithium concentrations collected in the field and needed  to compute the flow rates in the
vicinity of the reference flow meter originate from three sources:

    1.  The  concentrated samples collected in the injection cylinder. The concentrations at this
       site  should match the  lithium  concentration of the prepared  lithium mixture.  For the
       duration of a given test, this concentration was assumed to be constant for the purpose of
       flow computation. The  statistical analysis made for the validation of the concentrations
       measured by  the standard laboratory would reflect this assumption.

    2.  Blank samples collected upstream of the lithium injection point. They were used to verify
       that lithium was conveyed in the QUC sewer network and to compute flow rates near the
       reference flow  meter. These concentrations were assumed to be time varying and were
       treated as such when performing the statistical analyses.

    3.  Diluted samples collected near the reference flow meter, which were treated the same as
       the blank samples when performing hypothesis testing on the lithium concentrations.

5.2.2.1 Concentrated Samples

Two tests were used to validate the accuracy  of the concentrated samples. The first test was a
standard statistical test to compare the means  of two normally distributed populations. In the
literature, the test is referred to as  a "Two-population test of population means for  normal
populations with the same  variance" (Appendix D-l).  This  test compared the means of the
concentrated samples, as measured by the standard and control  laboratories,  and verified whether
their  means differed significantly. A null hypothesis (the mean of the two  populations are
identical)  and  an alternate  hypothesis (the mean of the two  populations are different) were
established,  and a variable called the "test statistic" was computed using the null hypothesis. The
probability of the test statistic being outside  the acceptance range was given by the level of
significance chosen. For this project, a level of significance of 0.05 was chosen to indicate the
likelihood that the test statistic was outside the acceptance range was low (less than five percent).
                                            62

-------
If the test statistic was outside  the acceptance range, the  alternate hypothesis was  accepted.
Conversely, if the test statistic was inside the acceptance range, the null hypothesis was accepted.

Accepting the alternate hypothesis means  that there  was a strong presumption that the  two
laboratories did not measure the  same mean concentrations.  However, it does not mean that the
mean concentration measured by the standard laboratory was not suited to compute flow rates at
the sampling point. If the mean concentrations measured by the two laboratories were inside the
accuracy margins specified by the standard  laboratory, there was no reason not to use the mean
concentration of the standard laboratory.

For the concentrated  samples, the accuracy observed for the average concentration under normal
conditions was ±0.02. The validation variable was defined as:
                                                                               rs-n
       Where:       Csi is the mean concentration of the standard laboratory; and
                    Cciis the mean concentration of the control laboratory.

xv should be between ±0.04, assuming that the accuracy of the concentrations measured by the
standard laboratory satisfied the expected accuracy  (Appendix D-4). The validation chart of the
concentrated samples is summarized in Table 5-1.

Table 5-1. Concentrated Samples Validation Chart


     T.   ix ft.    xi.   • x  x                     Result of accuracy test
     Result of hypothesis test            .„.        .„.               tn/       tn/
_ ^_ _ -4% < jcv > 4% _ -4% >jcv< 4% _

     Null hypothesis accepted           Validation of the             Validation of the
                                     concentrated samples         concentrated samples

   .,.   .  ,    .,   .       .  ,          Validation of the             Rejection of the
   Alternate hypothesis accepted                   ,     ,                      ,     ,
             J                        concentrated samples         concentrated samples
If the  concentrated samples are rejected for the two tests, the lithium test was not used to
compute the flow rates needed to validate the performance of the reference flow meter.

5.2.2.2 Blank and Diluted Samples

For the concentrated samples, two validation tests were used to assess the accuracy of the blank
and diluted  samples measured by the standard laboratory.  The first was a standard  statistical
called  "test on paired-sample data" (Appendix D-2). Each number of a data set was associated
with only one number in the other set. When the differences between the numbers in the pairs
                                           63

-------
were  the meaningful data, the two samples are called  paired  samples.  The idea consists in
verifying if the mean of the paired samples is significantly different from zero with a given level
of significance assuming that the paired data population has a zero mean  (the null hypothesis).
The probability of the test statistic  being  outside the acceptance range is  given by the level of
significance chosen. As for the two-population test statistic, a 0.05 level of significance was
used.  If the test statistic is outside the acceptance range,  since the probability of having such a
result is little (less than five percent), it was concluded that the null hypothesis was wrong and
the alternate hypothesis was accepted. Conversely, if the test statistic was  inside the acceptance
range, the null hypothesis was accepted and the alternate one was rejected.

Accepting  the alternate hypothesis means that there was a strong presumption that the two
laboratories are not measuring the  same blank or diluted  concentrations.  However, it does not
mean that the concentrations measured by the standard laboratory are not  suited to compute the
flow rates at the sampling point. If the mean concentration of the paired samples was very close
to zero and the alternate hypothesis was accepted because  of the small standard deviation of the
paired samples,  there  was no reason not to use the blank and  diluted concentrations of the
standard laboratory since they allow the  computation of flow rates that are similar to  those
computed using the concentrations measured by the control laboratory.

For the diluted samples, the accuracy observed for the measured concentrations  under normal
conditions  was ±0.03 (value specified by  the standard laboratory). The validation variable was
defined as:
                                       =  I
In
       Where:        Csi is the concentration of the standard laboratory; and
                    Cciis the concentration of the control laboratory.

xv should be between ±0.06, assuming that the accuracy of the concentrations measured by the
standard and the control laboratories satisfied the expected accuracy (Appendix D-5).

For the blank samples, the accuracy was given in absolute values and was equal to ±0.01 mg/L
for the standard laboratory and to  ±0.005 mg/L  for  the control laboratory. The validation
variable, as the mean difference of concentrations measured by the two laboratories, is defined
as:
       Where:        Csi is the concentration of the standard laboratory; and
                    Cciis the concentration of the control laboratory.

The  blank samples validation variable should  be  between  ±0.015 mg/L, assuming that the
accuracy of the concentrations measured by the standard and the control laboratories satisfied the
                                            64

-------
expected accuracy (Appendix D-6). Converse to the concentrated samples, for the bland and
diluted samples the validation variable was defined as a mean difference. For this reason, the
acceptance of the measured concentration could not be restricted to the deviation of xv from zero
but also be related to the standard deviation (SD) observed. The concentration measured by the
standard laboratory can be validated only when the mean difference of concentration between the
two laboratories is small and when the hypothesis  test is rejected due to a small variance. An
acceptable value for the standard deviation is 0.05. The validation chart of the blank and diluted
samples is summarized in Table  5-2.

Table 5-2. Validation Chart for the Blank and Diluted Samples

Result
Null hypothesis
accepted
Alternate hypothesis
accepted
Dilute
-0.06  0.06
and SD < 0.05
Validation
Validation
samples
-0.06 >xv< 0.06
or SD > 0.05
Validation
Rejection
Blank
-0.015  0.015
and SD. < 0.05
Validation
Validation
samples
-0.015 >xv< 0.015
or SD > 0.05
Validation
Rejection
If the blank or diluted samples were rejected for the two tests, the lithium test could not be used
to compute the flow rates needed to validate the performance of the reference flow meter.

5.2.3   Validation Results for Test CO

Two  lithium  dilution tests were  conducted during the course of the project.  The first was
performed during Test CO and the second during Test C3. For Test CO, the standard and control
laboratories analyzed 25 and five samples,  respectively.  The 25  samples analyzed by  the
standard laboratory originated from  five different bottles, and each bottle  was analyzed five
times. The five samples analyzed by the control laboratory were produced from a single bottle.

The  data from the 30  concentrated samples  were used to validate  the mean concentration
computed using the results provided  by the standard  laboratory.  The mean  concentrations
measured by the  standard and control  laboratories  were 45,500  mg/L  and  44,900 mg/L,
respectively. The test statistic (Appendix D-l) gives a value of 1.91 using the two-population
test, which was inside the acceptance range. Therefore, the concentrated samples measured by
the standard laboratory were validated for Test CO.

Both laboratories analyzed two blank samples. For these samples, the standard laboratory results
were  lower than the detection limit of 0.01  mg/L, and the control laboratory results were
0.008 mg/L and 0.007 mg/L. Assuming that the blank concentrations measured by the standard
laboratory are null for concentrations lower than 0.01  mg/L (this was the assumption made for
the computation  of flow rates when the concentration measured by the standard laboratory was
lower to 0.01 mg/L), the hypothesis test on paired samples did not permit the validation of the
blank concentrations  (Table  5-3). However,  the test on the mean  differences verified  the
accuracy associated with the blank samples. Therefore, the blank concentrations measured by the
                                           65

-------
standard laboratory during Test CO were judged accurate enough to be used to validate the
reference flow meter.

For the  diluted samples, nine  pairs of samples were  collected for the validation  of the
concentrations measured by the standard laboratory. Results show that the paired samples did not
stand  up to the hypothesis that they have a zero mean assuming a significance level  of five
percent. However, the test on the mean value and standard deviation of the paired  samples
indicated that the concentrations measured by  the two laboratories were within their accuracy
margins. The mean concentration of the paired  samples was equal to 0.06 mg/L with a standard
deviation lower than 0.05 mg/L. Therefore, the diluted concentrations measured by the standard
laboratory were considered accurate enough to be used to validate the reference flow meter.
Table 5-3. Validation Chart for Test CO
              Hypothesis Test (mg/L)
     Accuracy Test (mg/L)
   Samples    Lower    Test    Upper   T     T.  .
              T ...  c. ....   T .  .,   Lower Limit
              Limit  Statistic  Limit
                            	Validation
             Value     Upper Limit    Result
 Concentrated  -2.05     1.91     2.05
-0.04
Blank
Diluted
-12.7
-2.31
-15.0
5.52
12.7
2.31
-0.015 (Mean)
-0.06 (Mean)
-0.008
0.001
0.06
0.032
Mean 0.015
SD 0.05
Mean 0.06
SD 0.05
Accepted
Accepted
5.2.4  Validation Results for Test C3

For Test C3, two analyses with the concentrated samples were done. In the first analysis, the
standard laboratory analyzed 25  samples and  the  control  laboratory analyzed five.  The 25
samples analyzed by the standard laboratory originated from five different bottles, each being
analyzed five times, while the five samples analyzed by the control laboratory were produced
from a single sampling bottle. The mean concentration measured by the standard laboratory was
42,900 mg/L compared to 38,400 mg/L for the control laboratory.

Considering the significant difference between  the mean concentrations measured  by  the two
laboratories, a  second analysis was  conducted. Samples of the five bottles  analyzed by the
standard laboratory were sent to the control laboratory for analysis. From these five bottles, 25
aliquots were created and analyzed. The new results obtained by the control laboratory produced
a mean concentration  of 44,000  mg/L,  and confirmed that the  mean concentration  initially
measured was too low.

Using the lithium concentrations obtained during the second analysis, the test statistic computed
using the two-population test (Appendix D-l) gave a value slightly outside the acceptance limits
                                           66

-------
considering  a level  of significance of 0.05 (Table 5-4).  However, the mean concentration
validation test gives for the validation variable -0.026, a value within the validation limits.
Table 5-4. Validation Chart for Test C3
   Sample      Hypothesis Test (mg/L)
                 Mean Test (mg/L)
               Lower     Test    Upper  T      T.  .,     _, ,
               T •   •<_   O.L .L- .L-   T-  • -   Lower Limit     Value
               Limit   Statistic  Limit
                                   Upper
                                   Limit
                                  Validation
                                    Result
 Concentrated   -2.05      4.55
  Analysis 1
2.05
-0.04
 Concentrated   -2.01     -2.35     2.01        -0.04
  Analysis 2
0.11
                       -0.026
0.04
Rejected
                         0.04       Accepted
Blank
Diluted
-12.71
-2.37
-0.62
3.23
12.71
2.37
Mean -0.015
Mean -0.06
-0.007
0.015
0.04
0.035
Mean 0.015
SD 0.05
Mean 0.06
SD 0.05
Accepted
Accepted
Both  laboratories  analyzed two  blank samples. For  both samples, the standard laboratory
measured a lithium concentration greater than 0.01 mg/L. For Test CO, it was possible to validate
the concentration of the blank samples using a hypothesis test on paired sample data. However,
the reliability of the test was limited due to the very small number of samples analyzed. For the
two, paired samples the degree of freedom of the test was one, which implied a large validation
range for the null hypothesis. Nevertheless, the test statistic was small (-0.62), which indicated
that the mean concentration of the paired samples is close to  zero.

The validity of the blank samples measured by the standard laboratory was also confirmed by the
evaluating whether the mean difference concentration was within  the mean  value validation
limits. For the blank  samples, the mean concentration is -0.007  mg/L, which was inside the
±0.015  mg/L validation  limits.  Therefore, considering the  results obtained with  the two
validation tests,  the blank concentrations measured by the standard laboratory are accepted,
which means they were accurate enough to validate the reference flow meter.

For the diluted samples, nine pairs were collected for the validation analysis of the concentration
measured by the standard laboratory. The first results provided by the control laboratory showed
a mix-up of three sample concentrations, probably due to a handling deviation during the sample
labeling. To ensure that the unexpected results were caused by a handling deviation, the standard
laboratory reanalyzed  the three samples, and the samples were  sent to the control laboratory for
another analysis. The results from the second analysis matched.

Results show that the paired  samples  do not stand up to the hypothesis that the paired sample
data have a zero mean assuming a significance level of five percent. However, the test on the
mean value and standard  deviation of the paired samples indicates  that the concentrations
measured by the two laboratories are within the margin of accuracy.  The mean concentration of
                                           67

-------
the paired samples was equal to 0.04 mg/L with a standard deviation of 0.042 mg/L. Therefore,
the diluted concentrations measured by the standard laboratory were considered accurate enough
to be used to validate the reference flow meter.

5.2.5   Conclusion

In order to use a lithium dilution test to validate the reference flow meter, all three sources of
data  (i.e. concentrated, blank and  diluted) must be accepted according to the validation tests
described  above. If one is not accepted, it is assumed that the flow rates computed using the
lithium concentrations measured by the standard laboratory did not have  the level  of accuracy
necessary to validate the reference flow meter.

In this application, all three sources of  concentrations were  accepted for Tests CO  and C3.
Therefore, all  flow rates computed using the lithium concentrations measured by the  standard
laboratory could be used to validate the reference flow meter.

5.3    Validation of the Reference Flow Meter

This section presents the methods used  and the results obtained to verify and validate the
reference flow meter data. Two methods were used to verify the accuracy of the reference flow
meter and to validate the reference flow data:  the tracer dilution and the flow under the gate.

5.3.1   Validation Methods

To demonstrate that the  reference  flow meter was  accurate and that it exhibited no bias, the
reference flow data was compared to that obtained with the tracer dilution and the flow under the
upstream gate  for replicas CO and C3, and compared with flow under the downstream gate for
Tests B and D.

5.3.1.1 Qualitative Analysis

To assess qualitatively  the  validity of the  reference  flow data,  two types of figures  were
produced:

   1.  Reference meter verification: flow rate measurements:  These figures present, in the case
       of  replicas  CO and C3, the  flow rate output by the reference flow meter, the flow rate
       calculated from the tracer dilution and the flow rate estimated from the equation of the
       flow under the upstream gate (energy balance equation). For replicas CO and C3,  the flow
       rate from the gate equation  is actually the sum of the flow under the upstream gate and
       the flow from the Craig Collector (which is measured at its downstream end), delayed by
       15  minutes to account for travel time between the gate (and Craig Collector) and the
       reference testing site. Therefore, this flow rate is given as a rough approximation only. In
       the case of the three  replicas of Tests B and D, there is no tracer dilution data and the
       flow rate is calculated from the gate equation for the downstream gate without any
       modification.
                                           68

-------
   2.  Reference flow rate comparison: reference meter versus dilution:  These figures present
       the flow rate output by the reference flow meter as a function of the flow rate calculated
       from the tracer dilution (replicas CO  and C3). The y = x (or 1:1)  line is drawn as an
       indication of a  perfect fit between  the  two data sets. The correlation coefficient r
       (between the flow data and the y = x line) is also presented. A perfect fit would generate
       an r of 1.

5.3.1.2 Statistical Analysis

A hypothesis  test  on paired sample data (replicas CO  and C3) was performed  to  assess the
validity of the reference flow meter. In the paired  sample hypothesis test,  a new data set was
calculated, consisting of the differences between the flow rates measured by the reference flow
meter and those computed from tracer dilution. The hypothesis test was then performed as a one-
population test on the differences. This test validated whether the new population made of paired
data has a zero mean. The idea consists in validating whether the mean of the paired samples was
significantly different from zero, with a given level of significance, assuming that the paired data
population has a zero mean (the null hypothesis).

5.3.1.3 Reference Flow Deviation Distribution Analysis

This test consisted in drawing the  bar  chart of the flow deviation  distribution for  which the
deviation is defined as follows:

                                A yo 	    z~> dilution  z~> reference                          (5-4)
                                      preference ^^2 dilution )' ^

The reference  flow deviation distribution test  is the only test described in the protocol for which
a quantitative assessment  is done.  In  order to  validate the measurements produced by the
reference flow meter, 95 percent of the flow deviation  data (AQ) must be included  within the
acceptance range. According to the protocol, the acceptance limits are ±0.095 during dry weather
flow and ±0.087 during wet weather flow. These acceptance  margins were re-evaluated during
the course of the project to better account for uncertainties related to the concentrations measured
by the laboratories. Using  the  concentration accuracy margins  obtained  experimentally, the
lowest acceptance  margins were  -0.1066  and -0.1288 for dry  and wet weather flows,
respectively. Upper acceptance limits were 0.1074  for dry weather flow  and 0.1355 for wet
weather flow (Appendix D-7). The wet-weather flow of 2.57 MGD (112.5 L/sec) was defined as
1.5 times the maximum dry-weather flow of 1.71 MGD (75 L/sec).

The test also included two  qualitative  assessments.  To  accept the flow rates measured by the
reference flow meter, the frequency  distribution of AQ  must  resemble  a bell-shaped curve and
the mean deviation must be close to zero.

The  frequency distribution plots are presented  with  two-percent increments and  additional
information on each figure, including the:
                                            69

-------
•  Margin of deviation (±Xpercent) that contains 95 percent of all measurements (i.e. standard
   deviation ±2o);
•  Percentage of measurements within a margin of deviation of ±8.7 percent (wet weather tracer
   dilution deviation);
•  Percentage of measurements within the re-evaluated margin  of deviation of ±13.55 percent
   (wet weather tracer dilution deviation);
•  Mean deviation of all measurements; and
•  Median deviation (provides an indication of the spread).

5.3.2  Test CO Results

5.3.2.1 Qualitative Analysis

Figure H-142 presents flow rates output by the reference flow meter, flow rates calculated from
the tracer dilution and, as an indication, flow rates estimated from the equation of the flow under
the gate for the upstream gate (with modifications for the Craig Collector and travel time). Tracer
dilution data  points  are presented with two  markers: the dark circles  are  data taken  during
transition periods (rising and ebbing water profiles) and the hollow circles are data taken during
stable flow conditions. This segregation is used for later analysis. The reference meter and tracer
dilution data points were very close, with some minor differences at higher flows  of 17.1 and
29.7 MGD (750 and 1,300 L/sec). The  estimation of the flow under the gate also followed the
reference flow meter curve very closely.

Figure H-120 presents flow rates output by the reference flow meter as a function of flow rates
calculated from the tracer dilution. Again, two markers were used to segregate data points taken
during transition and stable periods. The larger deviations from the y = x (or 1:1) line were the
data points during transition periods and during higher flows. The correlation  coefficients (r)
calculated from all data points and from stable conditions only were very close to  1.

Qualitatively, these two figures show that the flow rates output by the reference flow meter were
very close to those calculated from tracer dilution (and from those estimated from the flow under
the  gate).  There  were no indications  that  the  reference  meter was poorly calibrated or
malfunctioning.

5.3.2.2 Statistical Analysis

The  hypothesis test on paired samples validates whether the new  population made of paired flow
data has a zero mean (null hypothesis test). Since the value of the test statistic (-0.528) is within
the lower and upper limits (±1.990) of the validation range for  the given level  of  significance
(5 percent), the null  hypothesis is accepted.  This result indicates a strong probability for the
paired flow data to belong to a zero mean population.

5.3.2.3 Reference Flow Deviation Distribution

Figure H-127 presents the reference flow  deviation  distribution for replica CO. Two of three
conditions for the test to be  satisfied are met: the frequency  distribution approaches a normal
                                            70

-------
curve and the mean deviation is close to 0 (0.3 percent). The median deviation (-0.1 percent)
provides an indication of the spread of the data set. The condition that was not met is the fraction
of deviation data included in the acceptance margins. Since most of Test CO is conducted under
wet weather conditions, the acceptable margin of deviation for 95 percent of measurements is
±8.7 percent according to the protocol. Only  87.8 percent of measurements are within the ±8.7
percent margin of deviation. However,  as  indicated in Figure  H-127,  97.6 percent  of
measurements are within the reevaluated ±13.5 percent margin of deviation.

The largest positive deviation was 19.7 percent (at 14:30), and occurred during the transition
from 8.56 to 1.71 MGD (375 to 75 L/sec). The largest negative deviation was -13.4 percent (at
15:25) and is observed at the very end of the test during maximum dry weather flow (1.71 MGD;
75 L/sec). All negative deviations larger (in absolute value) than -10.0 percent occurred during
maximum  dry weather  flow while all positive deviations larger than ±10.0 percent occurred
during transition periods.

Figure H-128 presents the reference flow deviation distribution for all data points during stable
conditions  and those during stable conditions but without maximal dry weather flow. In both
cases, the frequency distribution approaches a normal curve and the mean deviation is close to
zero (-0.6 percent and -0.1 percent, respectively). The acceptable  margin of deviation of ±8.7
percent for 95 percent  of the  measurements is met  for  data points  during stable conditions
without maximal dry weather flow (with 100 percent  of measurements within this  margin) and
barely missed for  all data points during stable conditions (with 94.4 percent of measurements
within this margin).

5.3.3   Test C3 Results

5.3.3.1 Qualitative Analysis

Figure H-143 presents flow rates output by the reference flow meter, flow rates calculated  from
the tracer dilution and, as an indication, flow rates estimated from the equation of the flow under
the gate for the upstream gate (with modifications for the Craig Collector and travel time). Tracer
dilution data points are presented with two markers: the dark circles are  data  taken during
transition periods (rising and ebbing water profiles) and the hollow circles are data taken during
stable flow conditions. This segregation is used for later analysis. The reference meter and tracer
dilution data points are very close at maximum dry weather flow of 1.71 MGD (75 L/sec) while
the reference flow rates are always larger for the two intermediate flow rates of 8.56 and 17.1
MGD  (375 and  750 L/sec),  and always smaller  for  the highest  flow rate of 29.7 MGD
(1,300 L/sec).

The higher the flow rate, the lower the lithium concentration. At 29.7 MGD (1,300 L/sec), the
lithium concentration is roughly 0.273 mg/L. A small  deviation on the  blank samples would
result in a large deviation in flow rate estimation for the higher flows. Since blank samples  were
taken every 30 minutes, the in-between concentrations were interpolated from the two known
values. This might explain the mismatch between the reference and dilution data sets.  The curve
of the flow under the gate is in the same vicinity as the reference flow meter curve.
                                           71

-------
Figure H-121 presents flow rates output by the reference flow meter as a function of flow rates
calculated from the tracer dilution.  Again, two markers are used to segregate data points taken
during transition and stable periods. The larger deviations from the y = x (or 1:1) line are the data
points during transition periods  and during  higher flows.  The  correlation coefficients (r)
calculated from all data points and from stable conditions data points only are very close to one.

Qualitatively, these two figures show that the flow rates output by the reference flow meter are
close to those calculated from tracer dilution (and from the ones estimated from the flow under
the  gate).   There  are  no  indications  that the  reference  meter  is poorly  calibrated  or
malfunctioning.

5.3.3.2 Statistical Analysis

The hypothesis test on paired samples validates whether the new population made of paired flow
data has a zero mean (null  hypothesis of the test). Since the value of the test statistic (0.700) is
within the lower and upper limits (±1.990) of the validation  range for the  given level  of
significance (5 percent), the null hypothesis is accepted. This result indicates that there exists a
strong probability for the paired flow data belongs to a zero mean population.

5.3.3.3 Reference Flow Deviation Distribution

Figure H-128 presents the reference flow deviation distribution for replica C3. None of the three
conditions for the test to be satisfied are met.  The frequency distribution  does  not resemble a
normal curve (two peaks),  the mean deviation  is decentered negatively (-2.0 percent) and only
80.5 percent of measurements are within a margin of deviation of ±8.7 percent (as indicated in
the protocol for  wet weather). The median deviation  (-3.8  percent)  is even more negatively
decentered. However, as indicated on Figure H-128, 95 percent of measurements are within the
reevaluated ±13.5 percent margin of deviation.

The largest positive deviation is 20.5 percent (at 7:45) and the largest negative deviation is -45.8
percent (at  8:30). Both occurred at the very beginning of the test during the maximum  dry
weather flow of  1.71 MGD (75 L/sec). All negative and positive deviations larger than ±10.0
percent (in absolute value) occurred during maximum dry weather flow. Negative deviations are
mostly due to the intermediate flows of 8.56 and  17.1 MGD (375 and 750 L/sec) and positive
deviations to the  highest flow of 29.7 MGD (1,300 L/sec). Since there are more data points at
intermediate flows, the curve is decentered negatively.

Figure H-131 presents the reference flow deviation distribution for all data  points during stable
conditions and those during stable conditions but without maximum dry weather flow for replica
C3. In both cases, as for Figure H-128, none of the three conditions for the test to be satisfied are
met. Only data under stable conditions without maximal dry weather flow meet the recalculated
criteria of 95 percent of measurements within a margin of deviation of ±13.5 percent.
                                            72

-------
5.3.4  Test B and D Results

Figures H-138 through H-140 present flow rates from the reference flow meter and from the
flow under the downstream gate equation for the three replicas of Test B, while Figures H-143
through H-145 present the same information for the three replicas of Test D. The flow under the
downstream gate comes from a standard hydraulic equation. At low flows (Test B and beginning
and end of Test D), the reference flow meter and gate equation curves are nearly superimposed.
As the flow rate increases (Test D), the gap between the two curves also increases, with the gate
equation curve always being higher. This is true for all three replicas of Test D. Therefore, the
reference flow meter reacted in the same fashion each time (not a punctual malfunction).

5.3.5  Conclusion

Except  for Test C3,  the qualitative validation analyses show that  the reference  flow meter
provides results representative of the flows computed using tracer dilution tests and an energy
balance equation in the vicinity  of sluice gates. The figures show the flow rates output by the
reference flow meter, the tracer dilution, and the energy balance equation show a  good match
between the three time series data sets. The general behavior can be summarized as follows:

•  For the highest flow rates, the measured reference flows follow closely the flows computed
   by the  energy  balance  equation, but are slightly below  the flows computed using tracer
   dilution. This observation is particularly visible for Test C3. This phenomenon could be due
   to an overestimation  of the blank concentrations caused by contamination of the sampling
   bottles.

•  For dry weather flows and low wet weather flows, the reference flow curve is slightly above
   the tracer dilution flow curve. This observation is particularly apparent for Test C3 and can
   be explained by the fact that the diluted concentrations of the standard laboratory were barely
   accepted. The diluted concentrations measured by the standard laboratory are systematically
   greater by 6  percent than  the  concentrations measured by  the  control laboratory.  This
   constant offset can be the reason why the flow rates measured by the reference flow meter
   are systematically above the flow rates computed using tracer dilution.

For Test C3, the reference flow meter does not pass the qualitative and the quantitative validation
tests for the reference flow deviation distribution. However, it appears that the validation test is
strongly  affected  by potential problems related to the measurement of the blank  and diluted
concentrations. Nevertheless, using the  acceptable  margin of deviation determined using the
updated accuracy values for  lithium concentration measurements, 95 percent of the reference
flow measurements are within ±13.5 percent.

5.4    Validation of the Reference Depth Meter (Bubbler)

This section presents the methods used and results obtained to validate the reference  water-depth
data. Two methods were  used to verify the accuracy of the reference  depth meter (bubbler) and
to validate the reference water-depth data: manual readings and an ultrasonic depth meter.
                                           73

-------
5.4.1   Validation Methods

To demonstrate that the reference depth meter (bubbler) was accurate and exhibited no bias, the
reference water depth data was compared qualitatively and quantitatively to that obtained with
manual readings for Tests B, C, and D, and for all these tests combined. A frequency distribution
analysis was performed with the reference depth meter for Tests B, C, and D combined.

5.4.1.1 Qualitative Analysis

To assess qualitatively the  validity of the reference water depth data, two types of figures were
produced:

   1.  Reference meters verification: depth measurements:  These figures present water depths
       from the reference depth meter (bubbler), from manual readings, and from the ultrasonic
       depth meter for the  three replicas of Test B, four replicas of Test C, and three replicas of
       Test D.

   2.  Reference depth  comparison:  reference  meter versus manual  measurement:   These
       figures present water depths output  by the reference depth meter as a function of the
       manual measurements for Tests B, C, and D, and for all these tests combined. The y = x
       (or 1:1) line is drawn  as an  indication of a perfect fit between the two data sets. The
       correlation coefficient (r) between the reference data and the y = x line is also presented.
       A perfect fit would generate an r of one.

5.4.1.2 Statistical Analysis

A statistical test (a hypothesis test on paired sample depth data) was performed to assess the
validity of the reference depth data. In this test, a new data set was calculated, consisting of the
differences between the depths measured by the  reference  depth meter and those measured
manually. The hypothesis test was then performed as a one-population test on the differences.
This test validates whether  the new population made of paired data has a zero mean, and whether
the mean of the paired  samples was  significantly different  from zero, with a given level of
significance assuming that the paired data population has a zero mean (the null hypothesis).

5.4.1.3 Reference Depth Deviation Distribution

This test consists of drawing the bar chart of the water-depth deviation distribution for which the
deviation is defined as:

                              manual    reference        ultrasonic     reference              T5-5^

The reference depth deviation distribution test is the only test described in the protocol for which
a  quantitative assessment  is done.  In order to  validate the measurements  produced  by the
reference  depth meter,  95  percent of the depth  deviation data (AD) must be included in  an
acceptance range.  According to the protocol, the acceptance limits are ±1.25 in.  (31.7 mm) or
two percent full-scale. The test also includes two qualitative assessments. To accept the water
                                            74

-------
depths measured by the reference depth meter, the frequency distribution of AD must resemble a
bell-shaped curve and the mean deviation must be close to zero.

The  frequency  distribution  plots  are  presented  with  0.25-in.  increments  and  additional
information on each figure, including the:

•  Margin of deviation (±Xpercent) that contains 95 percent of all measurements (i.e., standard
   deviation ±2o);
•  Percentage of measurements within a margin of deviation of ±1.25  in.  (31.7  mm) or two
   percent full scale;
•  Mean deviation of all measurements; and
•  Median deviation (provides an indication of the spread).

Qualitative and statistical  analyses are presented for Test B, C, and D and for these  three tests
combined.

5.4.2    TestB Results

5.4.2.1 Qualitative Analysis

Figures H-146 through H-148 present water depths from the reference depth meter (bubbler),
manual readings, and reference ultrasonic  depth  meter  for the  three replicas of  Test  B,
respectively.  Generally, all three  curves are superimposed except  at the  very beginning of
replicas Bl and  B2 (at 9:00) and during two purges of the bubbler  (at 9:10 and 13:10 during
replica B2).

Figure H-122 presents water depths output by the reference depth meter as a  function of manual
measurements. The  larger deviations from the y = x (or 1:1) line are the data points  during the
transition period from the  18-in. back flow to the 36-in. back flow.  During this transition and
under the 36-in.  back flow, the reference depth meter seems to slightly underestimate the water
depths compared to the manual readings, although the correlation coefficient (r) is very close to
one.

Qualitatively, these  four figures show that the water depths output by the reference depth meter
are very close to the manual  readings and those from the ultrasonic  depth meter. There are no
indications that the reference depth meter was poorly calibrated or malfunctioning.

5.4.2.2 Statistical Analysis

The hypothesis test  on paired samples validates  whether the new  population made of paired data
has a zero mean (null hypothesis  of the test). Since the value of the test statistic (16.39) is outside
the lower and upper limits (±1.972) of the validation range for  the given level  of significance
(five  percent), the  null hypothesis is  rejected.  This  result indicates that there is a strong
probability that the  paired depth  data do not belong to a zero mean population. The two sets of
data come from  populations having very small  standard deviations. Therefore, even though the
depth measurements are almost the same, the intersection covered by the distributions of the two
                                            75

-------
populations is small,  independent  of the values  measured. The statistical test reveals  a
measurement bias between the bubbler and the manual readings.

5.4.2.3 Reference Depth Deviation Distribution Analysis

Figure H-132 presents the reference depth deviation distribution for Test B (bubbler versus
manual measurements).  All  conditions for the test to  be satisfied are met. The frequency
distribution approaches a normal curve, the mean deviation is close to the median deviation,
which is also close to zero (mean deviation 0.39 in.; 9.9 mm);  and more than 95 percent of
measurements are within the acceptable margin of deviation according to the protocol (2 percent
full-scale: 1.25 in.; 31.7 mm).

5.4.3   Test C Results

5.4.3.1 Qualitative Analysis

Figures H-149 through H-152 present water depths  from the reference depth meter (bubbler),
manual readings, and reference  ultrasonic  depth meter for the four replicas of Test C. Water
depths from the three instruments vary more than for Test B, principally due to the presence of
stationary waves during these tests. Replicas CO, Cl, and  C3 present the same pattern: reference
depth measurements are lower  than manual readings but higher that  the ultrasonic  readings
during the 17.1 MGD  (750-L/sec) stable period, while the opposite is found at the 29.7 MGD
(1,300-L/sec). During  the two 8.56 MGD (375-L/sec) stable periods, manual readings  were
usually closer to the ultrasonic  data points. Both manual readings and ultrasonic depths  were
affected by stationary waves (may be  at the crest or trough), whereas  the bubbler buffers the
information over a larger volume of water. It should be noted that the manual readings  were
taken close to the bubbler site.

Figure H-123 presents water depths output by the reference depth  meter as a function of manual
measurements. At lower  depths,  data points are generally on the perfect-fit line and then, as the
depth gets higher, they  oscillate on each  side  of the line.  Despite this fact, the correlation
coefficient (r) is relatively close to one (0.984).

Qualitatively, these five figures show that the water depths output by the reference depth meter
were somewhat different from the manual readings and those from the reference ultrasonic depth
meter test due to the presence of stationary waves at higher flow rates. As previously mentioned,
manual readings and the  ultrasonic depth meter data  were strongly affected by waves (punctual
measurement can be at the crest or trough of waves) whereas the bubbler was less affected.

5.4.3.2 Statistical Analysis

The hypothesis test on paired samples validates whether the new population made of paired data
has a zero mean (null  hypothesis of the test). Since the  value of the test statistic (0.723) was
within the lower and  upper limits (+ 1.967) of the validation  range for the given  level of
significance (five percent) the null hypothesis was  accepted. This indicates that there was a
strong probability for the paired depth data to belong to a zero mean population.
                                            76

-------
5.4.3.3 Reference Depth Deviation Distribution Analysis

Figure H-133 presents the reference depth deviation distribution for Test C  (bubbler versus
manual measurements). Only one condition for the test to be satisfied was  met.  The mean
deviation was very close to zero (0.05 in. (1.3 mm)). The frequency distribution only vaguely
resembled a normal curve and less than 95 percent  of the measurements (77.8 percent) were
within the acceptable margin of deviation according  to the protocol (two percent full-scale, or
1.25 in. (31.7mm)).  The deviations  larger than ±1.25 in. were positive deviations collected
during the 17.1 MGD (750-L/sec) flow period. Since during this period the reference depth meter
did not measure the same depth (due to the distance between the two measurement devices and
the presence of a non-negligible stationary wave), it was legitimate not to consider these data in
the frequency  distribution test. The resulting frequency  distribution  graph resembles a bell-
shaped curve and the quantitative frequency distribution test was satisfied.

5.4.4  Test D Results

5.4.4.1 Qualitative Analysis

Figures H-153 through H-155  present water depths  from the reference depth meter (bubbler),
manual readings  and reference  ultrasonic depth  meter  for  the  three replicas of  Test D,
respectively.  Generally, all three curves are superimposed except under back-flow conditions,
when the ultrasonic becomes out of range (maximum reading of 40.2 in. [1,022 mm]).

Figure H-124 presents water depths output by the reference depth meter as a function of manual
measurements. Most data points were very close to the y = x (or 1:1) line. The larger deviations
were recorded during the transition period from  back flow conditions to maximum dry weather
flow without back flow. The correlation coefficient (r) is one.

Qualitatively, these four figures show that the water depths output by the reference depth meter
are close to the manual readings and those from the reference ultrasonic depth meter. There are
no indications that the reference depth meter was poorly calibrated or malfunctioning.

5.4.4.2 Statistical Analysis

The hypothesis test on paired samples validates whether the new population made of paired data
has a zero mean (null hypothesis of the test). Since the value of the test statistic (1.22) was inside
the lower and upper limits (±1.968) of the validation range for the given level of significance
(five percent), the null hypothesis was accepted. This indicates that a strong probability for the
paired depth data to belong to a zero mean population exists.

5.4.4.3 Reference Depth Deviation Distribution Analysis

Figure H-134 presents the reference depth deviation distribution for Test D  (bubbler versus
manual measurements). The three conditions for the test  to  be  satisfied  are met: the mean
deviation was very close to zero (0.04  in.; 1.0 mm), the frequency  distribution resembled a
                                            77

-------
normal curve and more than 95 percent of the measurements were within the acceptable margin
of deviation (two percent full-scale) noted in the protocol.

5.4.5  Tests B, C, and D Combined

5.4.5.1 Qualitative Analysis

Figure H-125 presents water depths output by the reference depth meter as a function of manual
measurements. Most data points were very close to the y = x (or 1:1) line. The larger deviations
were from Test C. The correlation coefficient (r) was very close to 1.

5.4.5.2 Reference Depth Deviation Distribution Analysis

Figure H-135 presents the reference depth deviation distribution for Tests B, C, and D (bubbler
versus manual measurements). Two of the three conditions for the test to be satisfied were met.
The  frequency distribution resembled a normal  curve and the mean deviation  was very close to
zero (0.13 in.; 3.3 mm). Less than 95  percent of the measurements (88.9 percent) were within the
acceptable  margin of deviation according to the protocol  (two percent  full-scale,  or  1.25 in.
[31.7mm]). This last observation reflects the  measurement differences observed  for Test C
during the presence of stationary waves.

Figure H-136 presents the reference depth deviation distribution for Tests B, C, and D (bubbler
versus reference ultrasonic depth meter).  Only  one  of the three conditions  for the test to be
satisfied was met. The mean deviation was very  close to zero (-0.20 in.; -5.1 mm). The frequency
distribution only  vaguely resembled a  normal  curve  and  less  than 95 percent of the
measurements (79.7  percent)  were within the acceptable margin of deviation  (two percent full-
scale) noted in the protocol.

5.4.6  Conclusion

Except for Test C, the  qualitative validation analyses showed that the reference depth  meter
provides results that  were representative of the manual readings and the depths measured by the
ultrasonic depth meter.  The  figures  showing the water depths output by the reference  depth
meter, the manual readings, and the  ultrasonic  depth meter showed a good match between the
three time series. The general  behavior observed can be summarized as:

•  Under normal conditions, the measured reference  depths closely followed the water depths
   obtained by manual readings and by the ultrasonic depth meter. However, during replicas Bl,
   B2, and B3, a relatively  constant measurement  offset was observed and detected by the
   statistical test. This offset was small, however, and could probably be eliminated by a slight
   readjustment to the calibration of the measurement devices.

•  In the presence of stationary waves (Test C), significant depth differences were observed for
   the three  measurement  devices. However, these  differences  were not  related to bad
   calibration or malfunctioning of the reference depth meter. The distance existing between the
   different depth meters can explain these results.
                                           78

-------
From these observations and the  qualitative and quantitative results obtained for the different
validation tests, the reference depth meter appears to be a valuable measurement device that can
be used to generate reference depth data.
                                             79

-------
                                    Appendices

A Complementary Test Site and Reference Meters Information
B Experiment Procedures
C Detailed Experimentation Results
D Quality Assurance Project Plan: Complementary Information
E Field Verification: Reports Generated during Tests
F  ADS 4000 Flow Meter: Configuration
G ADS 4000 Flow Meter Software Functionality/Flexibility Summary
H Flow Test Data Figure Summaries
                                          80

-------
                                       Glossary

Accuracy - a measure of the closeness of an individual measurement or the mean of a number of
measurements to the true value and includes random error and systematic error.

Bias - the systematic or persistent distortion of a measurement process that causes errors in one
direction.

Comparability - a qualitative term that expresses confidence that two data sets can contribute to
a common analysis and interpolation.

Completeness - a qualitative term that expresses confidence that all necessary data have been
included.

Precision - a measure of the  agreement between replicate measurements of the same property
made under similar conditions.

Protocol - a written document that clearly states the objectives, goals,  scope, and procedures for
the study. A protocol shall be used for reference during vendor participation in the verification
testing program.

Quality Assurance Project Plan - a written document that describes the  implementation of
quality assurance and quality control activities during the life cycle of the project.

Representativeness - a measure of the degree to which data accurately and precisely represent a
characteristic of a  population  parameter  at  a  sampling point,  a process  condition,  or
environmental condition.

Wet Weather Flows Stakeholder Advisory Group - a group of individuals consisting of any or
all of the following: buyers and users of flow monitoring technologies, developers and vendors,
consulting engineers, the finance and export communities, and permit writers and regulators.

Standard Operating  Procedure -  a written document  containing specific procedures and
protocols to ensure that quality assurance requirements are maintained.

Technology Panel - a group  of individuals with expertise and knowledge of flow monitoring
technologies.

Testing Organization - an independent organization qualified by the Verification Organization
to conduct studies and testing  of flow monitoring technologies in accordance with protocols and
Test Plans.

Vendor - a business that assembles or sells flow monitoring equipment.

Verification - to establish evidence on the  performance of flow  monitoring technologies under
specific conditions, following a predetermined study protocol(s) and test plan(s).
                                           81

-------
Verification  Organization -  an  organization qualified  by EPA to verify environmental
technologies and to issue verification statements and verification reports.

Verification  Report -  a  written document containing  all  raw and analyzed  data, all quality
assurance/quality control  (QA/QC) data sheets, descriptions of all collected  data, a detailed
description of all procedures and methods used in the verification testing, and all QA/QC results.
The test plan(s) shall be included as part of this document.

Verification  Statement - a document that summarizes the Verification  Report reviewed and
approved and signed by EPA and NSF.

Verification Test Plan - a written document prepared to describe the procedures for conducting
a test or study  according  to the verification protocol requirements for the  application of flow
monitoring technology.   At a minimum, the  test plan  shall include  detailed instructions  for
sample and data collection, sample handling and preservation, precision, accuracy, goals, and
QA/QC requirements relevant to the technology and application.
                                           82

-------